Sirtuin 6 Improves Healthspan In Mice

Previous research has demonstrated that overexpressing a protein known as SIRT6 increases lifespan of mice and offers protection against obesity-induced metabolic abnormalities. In contrast mice that lack SIRT6 suffer from various aging-related problems and die before the age of 4 weeks. SIRT6 is part of a class of proteins known as the sirtuins that have been well-studied for their anti-aging effects. Earlier this year we reported on a small molecule drug that could selectively activate SIRT6.

Now, two papers by the Cohen lab further investigate the effect of SIRT6 on life- and healthspan.

In the first paper the researchers developed a SIRT6 transgenic mouse that has extra copies of the SIRT6 gene (termed MOSES mice). Old MOSES mice had improved insulin sensitivity, reduced fasting glucose levels, increased physical activity, reduced inflammation of their fat tissue, and improved hair growth.

What about SIRT6 knockout?

In the second paper the authors investigate the effect of a complete lack (a so called knockout) of SIRT6. Such work had been done before but all mice died before four weeks of age. The authors wanted to investigate the effect of SIRT6 on health parameters later in the mouse lifespan. By changing the genetic background of the mice that carry the SIRT6 knockout the researchers were able to generate SIRT6 knockout mice that reach adulthood. Average lifespan of males was just 140 days while 80% of female mice were still alive after 200 days. Hence the effect of SIRT6 on lifespan is gender-specific. The authors succeeded in investigating the effect of certain health parameters up to 10 months of age.

By 5-6 months of age SIRT6 knockout mice showed severe corneal injury. The cornea is the transparent outer layer of the eye allowing light to pass into the eye. The age-related thinning of the retina, the eye layer where light is sensed, was accelerated in SIRT6 knockout mice. Furthermore, SIRT6 knockout mice also show degradation of retinal function.

Surprisingly the SIRT6 knockout mice had higher glucose uptake in skeletal muscle compared to normal mice. The authors showed that this glucose uptake was insulin-independent. The interpretation of this observation is unknown at this moment.

A link to neurodegeneration

Finally, in a third paper by another lab the effect of brain-specific SIRT6 knockout was investigated. Mice that lacked SIRT6 in the brain showed learning impairment at 4 months of age. The authors further discovered that such mice had higher levels of phosphorylated tau in their brains, a characteristic of several neurodegenerative diseases including Alzheimer’s disease. Hyperphosphorylated tau is believed to be neurotoxic and indeed mice that lacked SIRT6 had higher levels of cell death and DNA damage in their brains. SIRT6 levels in the brain decrease with age. Finally, the authors showed that Alzheimer’s disease patients had lower levels of SIRT6 protein in their brains.

A NAD+/PARP1/SIRT1 axis in Aging


NAD+ levels decline with age in diverse animals from C. elegans to mice [1] [2] [3] [4] [5] and raising NAD+ levels has apparent anti‐aging effects [3] [4] [2] [6] [5]. Increasing NAD+ levels by dietary supplementation with NAD+ precursors Nicotinamide Riboside (NR) or Nicotinamide Mononucleotide (NMN) improved mitochondrial function [4], and muscle, neural and melanocyte stem cell function in mice. NR supplementation also increased lifespan modestly in mice [5].

Although the mechanisms of gradual NAD+ decline with age remain unclear, beneficial effects have been linked to mitochondrial rejuvenation via restoration of OXPHOS subunits via sirtuin SIRT1/HIF1a/c‐myc pathway [4], and stem cell function via stimulation of the mitochondrial unfolded protein response (UPSmt) [5] and synthesis of prohibitins, a family of mitochondrial stress response proteins linked to senescence of fibroblasts in mammals. Improvement of stem cell function by NAD+ is SIRT1 dependent[5].

NAD+ levels are significant for metabolic homeostasis in that they are important for redox reactions, and as substrates of sirtuins (SIRT1 to SIRT7) and other enzymes including PARPs ([poly(adenosine diphosphate– ribose) polymerases) as well as yet uncharacterized roles through other NAD+ interacting proteins.

Sirtuins have been implicated in aging, as well as in a large number of key cellular processes including mitochondrial biogenesis, stress resistance, inflammation, transcription via epigenomic modulation of histones, apoptosis [7] [8] and SIRT1 can even affect circadian clocks [9].

PARPs are NAD‐dependent pleiotropic enzymes that play a key role in detecting and responding to single‐stranded DNA break repair as well as in inflammation, apoptosis induction and DNA methylation. Increasing PARP activity has been correlated with differential longevity of 13 mammalian species with humans having about 5 fold more PARP activity than rats [10].

Intuitively, increasing PARP activity would seem to be beneficial to organisms and potentially be of benefit to delay aging by maintaining DNA integrity, but PARPs are major consumers of NAD+ and PARP activity can deplete sufficient NAD+ to detrimentally effect mitochondria function. In fact, in apparent contradiction to the species longevity results, PARP1 knockout in mice increases tissue

NAD+ levels, enhances mitochondrial function and blocks inflammation [11]. Moreover, ectopic expression of human PARP1 in mice, an enzyme which has about 2 fold more intrinsic PARP activity than mouse PARP1, induces a set of aging related phenotypes including cardiomyopathy, hepatitis, pneumonitis, kyphosis, nephropathy, adiposity, dermatitis, anemia, increased glucose resistance, and increased incidence of carcinomas.

Paradoxically human PARP1 transgenic mice even have delayed DNA repair [12]. One possible explanation is that human PARP is consuming more NAD+ in mice than can be compensated for metabolically, causing dysfunction.

Clearly decreased NAD+ levels that have been observed in aging animals could be detrimental by multiple mechanisms, which require careful elucidation.

Decreasing NAD+ levels with aging cause DBC1 to bind and inhibit PARP1 allowing accumulation of DNA damage.

Li et al. [13] have recently established that NAD+ levels control protein‐protein interactions of two NAD+ binding proteins: DBC1/CCAR2 and PARP1. At low NAD+ levels DBC1 binds to PARP1 and inhibits PARP activity. In aging mice, DBC1 increasingly binds to PARP1 causing the accumulation of DNA damage, linking reduced NAD+ levels directly to DNA damage.

DBC1 was already known to bind to and inhibit SIRT1, as well as histone deacetylase HDAC3 and methyltransferase SUV39H1. Since SIRT1 is regulated by NAD+, Li et al explored the hypothesis that DBC1 could also bind and inhibit PARP1, another NAD‐dependent enzyme. In HEK293T cells co‐ immunoprecipitation/Western studies showed that PARP1 and DBC1 physically interact. A catalytically inactive PARP1 also interacted with DBC1 in these studies demonstrating that PARP catalytic activity was not necessary for the interaction with DBC1.

However, unlike the DBC1/SIRT1 interaction, the DBC1/PARP1 interaction was sensitive to and broken by NAD+ in the physiological range (100um‐500um), while similar doses of nicotinamide or its structural analog 3‐AB had no effect. NADH and adenine had only weak effects, suggesting that DBC1/PARP1 interaction requires

low levels or the absence of NAD+. A catalytically inactive PARP1 behaved similarly to wild type suggesting that NAD+ cleavage or covalent attachment was not necessary for NAD+ to inhibit the DBC1/PARP1 interaction[13] .

In HEK293T, an inhibitor of NAD+ biosynthesis or depleting NAD+ by genotoxic stress increased the DBC1/PARP1 interaction. Various treatments that increased NAD+, decreased DBC1/PARP1 interaction. These data are strong evidence that NAD+ inhibits the formation of the DBC1/PARP1 complex [13].

In order to determine which domains of DBC1 and PARP1 were interacting, Li et al. expressed a variety of truncated mutants of DBC1 and PARP1. DBC1 lacking a nudix homology domain (NDH) did not bind to PARP1, but a DBC1 carrying only a NDH domain did bind to PARP1. Moreover, DBC1 interacted with a PARP1 mutant consisting only of a BRCT domain, but not with a PARP1 having only a catalytic domain, suggesting that the NDH domain of DBC1 binds to the BRCT domain of PARP1.

Homology modeling based on the known crystal structures of 5 NDH proteins suggested that NAD+ directly binds the DBC1 NHD, and this was confirmed by competition experiments in which radiolabeled or biotin labeled NAD+ bound to DBC1‐NHD and was then competed off with unlabeled NAD+ [13].

Of significance is that PARP1 activity was inhibited by DBC1 interaction in vitro and stimulated in cell culture by siRNA knockdown of DBC1. DBC1 but not DBC1Q391A, a mutant of DBC1 that binds NAD+ poorly and does not interact well with PARP1, reduced PARP1 activity. Reducing DBC1 increased DNA repair after paraquat treatment, as assessed by lower gammaH2AX, reduced DNA fragmentation, increased NHEJ and HR recombination pathway activity and cell viability. NMN. which raises NAD+ levels. had similar effects on these cells. These data support the hypothesis that NAD+ binding to DBC1‐NHD regulates two key pathways of DNA repair through controlling PARP1 activity [13].

DNA repair capability decreases with increasing age [14]. Similarly PARP1 activity has been reported to decrease with age [10], suggesting a possible connection between PARP1 activity and overall DNA repair. Li et al hypothesize that increased DBC1/PARP1 binding occurs with reduced NAD+ levels during aging to reduce DNA repair and increase DNA damage.

In 22‐month old mice, levels of NAD+ were reduced in the liver compared to young mice. Old mice had increased amounts of DBC1‐PARP and increased staining for DNA‐damage associated gammaH2AX. A week of daily NMN treatment (500 mg/kg) intraperitoneally increased NAD+ levels, reduced the number of DBC1‐PARP1 complexes and decreased DNA damage as reported by decreased levels gammaH2AX in old mice. Reduced PARP1 activity in old mice was restored by NMN treatment.

To show that repair was affected, old mice were subjected to gamma irradiation. Irradiated mice were observed to have reduced DNA damage when treated with NMN either before irradiation or even one hour after irradiation[13]. These data support the hypothesis that reduced NAD+ with aging causes increased DBC1/PARP1 inhibiting PARP1‐mediated repair of DNA damage.

The authors reasonably speculate that DBC1 evolved as a buffer to prevent PARP1 from depleting NAD+ levels to cytotoxic levels [15]. They also point out that DBC1 is often mutated or downregulated in cancer and that increased repair capability and protection from radiation are potential mechanisms by which elimination of DBC1 could help cancer cells evade treatment[13]. Another possible mechanism for the tumor suppressor DBC1 is that the known stabilizing interaction of DBC1 with p53 is important[16].

Medical Implications:

Evidence is accumulating that diminished NAD+ during aging results in mitochondrial and stem cell dysfunction, as well as reduced DNA repair activity. Supplementation with NMN or especially NR at doses of 100, 300 or 1000 mg can increase NAD+ levels systemically in humans [17] and may counteract NAD+ aging‐associated phenotypes, although more preclinical and human studies are needed to establish efficacy of NR or NMN as anti‐aging interventions.

 A key question arises, if NAD+ levels and in particular PARP activity levels are so important to maintaining homeostasis and contribute to the biological processes called aging, why did treatment with NR, although beneficial to stem cell function, only increase mean and maximum longevity modestly in one study: about a 5% increase when treatment was initiated in very old mice (24 months) [5].

One point is that the intervention was started very late, which may not allow reversal of aging associated damage or loss of epigenomic information. On the other hand, the effect of increasing NAD+ is significantly less than that achieved by rapamycin, for example[18]. The simplest explanation is that decreased NAD+ represents only one of a set of key regulatory changes that result in the effects of aging.

However, it is possible that the role of PARP is more profound. Given the apparent correlation of PARP activity with mammalian longevity for diverse species, it is quite possible that increasing PARP1 expression by genetic or pharmaceutical means in the context of increasing NAD+ through supplementation may have profound effects on aging.

One interesting experiment would be to supplement the human PARP1 transgenic mice [12], which express a PARP1 2‐fold more active than mouse PARP1, with NR or NMN from birth to overcome the potential detrimental effects of decreased NAD+ levels from the overactive human PARP1. We hypothesize that these mice would not only not be sick, but would thrive and potentially live longer than wild‐type animals.

The significance of PARP1 activity to aging is at least somewhat tied to its DNA repair activity. It is controversial whether that the accumulated effects of DNA mutations alone are sufficient to drive aging [19], normal cells apparently accumulate hundreds to a few thousand mutations over a lifetime of cell divisions[20].

Non‐dividing cells should possess significantly fewer mutations. While some premature aging syndromes and cancer are driven by increased genome instability, it is unclear that accumulation of ~1000 genetic substitutions would sufficiently degrade cell function. On the other hand, because decreased PARP1 activity is expected to also cause increased chromosomal instability through reduced NHEJ and HR DNA repair, PARP1 activity probably plays a significant role in preventing age‐associated cancer.

If PARP1 plays a central role in aging via its role in DNA repair, it could do so indirectly, through effects on the epigenome. PARP1 and other members of the PARP family have been shown to have profound effects on the epigenome through Poly(ADP‐ribosyl)ation (PARylation), affecting DNA methylation/demethylation via effects on DNMT1 and CTCF, histone acetylation, histone methylation, and organizing chromatin domains [21].

DNA repair events initiated by single‐ stranded DNA breaks could change the balance of PARP1 activities changing the epigenome in defined ways. Animals with higher intrinsic PARP1 activity may have less DNA damage and fewer epigenomic alterations over a given amount of chronological time, resulting in a younger biological “age”.

Superimposed upon the effects of decreased NAD+ levels, it does not escape attention that the NAD+/PARP1/SIRT1 axis could contribute to creating the apparent epigenetic aging “clocks” that are reflected in the tissue‐invariant and tissue‐specific DNA methylation patterns observed to strongly correlate biological age in humans, chimpanzees and mice [22][23] [24] through modulation of DNA methylation through DNMT1 and CTCF.


  1. [1]  Braidy N, Guillemin GJ, Mansour H, Chan‐Ling T, Poljak A, Grant R. Age Related Changes in NAD+ Metabolism Oxidative Stress and Sirt1 Activity in Wistar Rats. PLOS ONE 2011;6:e19194. doi:10.1371/journal.pone.0019194.
  2. [2]  Yoshino J, Mills KF, Yoon MJ, Imai S. Nicotinamide Mononucleotide, a Key NAD+ Intermediate, Treats the Pathophysiology of Diet‐ and Age‐Induced Diabetes in Mice. Cell Metab 2011;14:528–36. doi:10.1016/j.cmet.2011.08.014.
  3. [3]  Mouchiroud L, Houtkooper RH, Moullan N, Katsyuba E, Ryu D, Cantó C, et al. The NAD+/Sirtuin Pathway Modulates Longevity through Activation of Mitochondrial UPR and FOXO Signaling. Cell 2013;154:430–41. doi:10.1016/j.cell.2013.06.016.
  4. [4]  Gomes AP, Price NL, Ling AJY, Moslehi JJ, Montgomery MK, Rajman L, et al. Declining NAD+ Induces a Pseudohypoxic State Disrupting Nuclear‐Mitochondrial Communication during Aging. Cell 2013;155:1624–38. doi:10.1016/j.cell.2013.11.037.
  5. [5]  Zhang H, Ryu D, Wu Y, Gariani K, Wang X, Luan P, et al. NAD+ repletion improves mitochondrial and stem cell function and enhances life span in mice. Science 2016;352:1436–43. doi:10.1126/science.aaf2693.
  6. [6]  Bhullar KS, Hubbard BP. Lifespan and healthspan extension by resveratrol. Biochim Biophys Acta BBA ‐ Mol Basis Dis 2015;1852:1209–18. doi:10.1016/j.bbadis.2015.01.012.
  7. [7]  Grabowska W, Sikora E, Bielak‐Zmijewska A. Sirtuins, a promising target in slowing down the ageing process. Biogerontology 2017. doi:10.1007/s10522‐017‐9685‐9.
  8. [8]  Mei Z, Zhang X, Yi J, Huang J, He J, Tao Y. Sirtuins in metabolism, DNA repair and cancer. J Exp Clin Cancer Res CR 2016;35. doi:10.1186/s13046‐016‐0461‐5.
  9. [9]  Masri S, Sassone‐Corsi P. Sirtuins and the circadian clock: bridging chromatin and metabolism. Sci Signal 2014;7:re6. doi:10.1126/scisignal.2005685.

[10] Grube K, Bürkle A. Poly(ADP‐ribose) polymerase activity in mononuclear leukocytes of 13 mammalian species correlates with species‐specific life span. Proc Natl Acad Sci U S A 1992;89:11759–63.

 [11] Bai P, Cantó C, Oudart H, Brunyánszki A, Cen Y, Thomas C, et al. PARP‐1 Inhibition

[12] Mangerich A, Herbach N, Hanf B, Fischbach A, Popp O, Moreno‐Villanueva M, et al.

[13] Li J, Bonkowski MS, Moniot S, Zhang D, Hubbard BP, Ling AJY, et al. A conserved NAD(+) binding pocket that regulates protein‐protein interactions during aging.

[14] Gorbunova V, Seluanov A, Mao Z, Hine C. Changes in DNA repair during aging.

[15] Yang H, Yang T, Baur JA, Perez E, Matsui T, Carmona JJ, et al. Nutrient‐Sensitive

[16] Qin B, Minter‐Dykhouse K, Yu J, Zhang J, Liu T, Zhang H, et al. DBC1 Functions as a

[17] Trammell SAJ, Schmidt MS, Weidemann BJ, Redpath P, Jaksch F, Dellinger RW, et al.

[18] Harrison DE, Strong R, Sharp ZD, Nelson JF, Astle CM, Flurkey K, et al. Rapamycin fed late in life extends lifespan in genetically heterogeneous mice. Nature

[19] López‐Otín C, Blasco MA, Partridge L, Serrano M, Kroemer G. The Hallmarks of Aging. Cell 2013;153:1194–217. doi:10.1016/j.cell.2013.05.039.
[20] Martincorena I, Campbell PJ. Somatic mutation in cancer and normal cells. Science 2015;349:1483–9. doi:10.1126/science.aab4082.
[21] Ciccarone F, Zampieri M, Caiafa P. PARP1 orchestrates epigenetic events setting up chromatin domains. Semin Cell Dev Biol 2017;63:123–34. doi:10.1016/j.semcdb.2016.11.010.
[22] Horvath S. DNA methylation age of human tissues and cell types. Genome Biol 2013;14:3156. doi:10.1186/gb‐2013‐14‐10‐r115.

[23] Hannum G, Guinney J, Zhao L, Zhang L, Hughes G, Sadda S, et al. Genome‐wide methylation profiles reveal quantitative views of human aging rates. Mol Cell 2013;49:359–67. doi:10.1016/j.molcel.2012.10.016.
[24] Stubbs TM, Bonder MJ, Stark A‐K, Krueger F, von Meyenn F, Stegle O, et al. Multi‐tissue DNA methylation age predictor in mouse. Genome Biol 2017;18. doi:10.1186/s13059‐017‐1203‐5.

Rapamycin,rifampicin, Psora-4 cocktail doubles lifespan

Drug combinations work to synergistically extend the life- and healthspan in worms.

The most successful drug combination tested almost doubled lifespan, a feat never reported before. Furthermore, more than half of the worms receiving the drug combination were still in optimal health after all control animals had died! Finally, the authors show that these drugs also extend the lifespan of fruit flies and in fact the same combination that almost doubled worm lifespan was also the most successful one at extending the lifespan of fruit flies.

In a new paper published on the preprint server bioRxiv Jan Gruber and colleagues test various combinations of life extending drugs in worms. Preprint servers allow scientists to publish their article to the wide world for anyone to criticize before submitting it to a classical journal for publication. Furthermore they allow for rapid dissemination of research results as the publication in a peer-reviewed journal can take months to years between submission and publication.

While around a thousand compounds have been found to extend lifespan in at least one study in a model organism few studies have been conducted using combinations of two or more compounds. A rare example includes the combination of rapamycin with metformin which was shown to outperform either drug alone in mice.

The authors started by identifying well known mechanisms for lifespan extension based on a literature review. Then they looked for drugs that influence these mechanisms and had previously been found to extend lifespan leading to the selection of 12 compounds for testing in this study. Next the researchers tested these compounds in roundworms, a common model organism in aging research. When tested in isolation five drugs significantly extended mean and maximal lifespan: Psora-4, rifampicin, rapamycin, metformin, and allantoin.

Image credit: Sven Bulterijs

Next the researchers tested all pairwise combinations of these five compounds. The combination of metformin and rapamycin, both at optimal doses, did not lead to a further increase in lifespan. However if the combination was tested at suboptimal doses than lifespan was further increased. Given that the optimal dose of metformin and rapamycin in mice is unknown the beneficial effect of the combination observed in the mice study mentioned before could possibly be the result of suboptimal concentrations of both drugs. Two combinations, rifampicin + rapamycin and rifampicin + Psora-4 all at optimal concentrations did result in synergistic increases in lifespan.

Next the researchers tested triple combinations of the various compounds. As testing all combinations would be impractical (220 different lifespan tests would be necessary) the researchers decided to first try the combination of the 3 most successful compounds so far (rapamycin + rifampicin + Psora-4). However this combination resulted in a shorter lifespan than the two successful pairwise combinations. Next, the researchers decided to test the two pairwise combinations from before with allantoin added as the third drug. They chose this combination because allantoin shares no mechanistic overlap with the other drugs. Both of these triple combinations resulted in a significant extension of mean and maximal lifespan with the most successful one (rapamycin + rifampicin + allantoin) resulting in a doubling of mean lifespan! The authors remark that this is the largest lifespan extension ever observed by a drug intervention initiated in adult worms.

Abbreviations: Rif = rifampicin; Rap = rapamycin, Allan = allantoin. Image credit: Sven Bulterijs

Certain interventions that result in lifespan extension reduce fertility but neither of the triple drug combinations reduced total fertility and actually slightly extended the fertile period of life. Furthermore, both drug combinations extended the period of life spend in good health (= the healthspan). In fact, more than half of the worms receiving the triple drug combination were still in optimal health after all control animals had died! Old animals on the triple drug combinations were indistinguishable from young control worms when judged by spontaneous movement. Worms really show a significant decrease in spontaneous movement with age. Finally, the two successful triple drug combinations also significantly increased resistance to oxidative and heat stress.

Total mortality is made up of two distinguishable parts: age-dependent and age-independent mortality. The age-dependent mortality is a measure for the rate of aging. So the authors tested which mortality rate was reduced by the drug combinations. Interestingly, the rifampicin + Psora-4 + allantoin combination significantly reduced the rate of age-dependent and age-independent mortality showing that this drug slows down aging in addition to making them more robust at young ages.

Abbreviations: Rif = rifampicin; Rap = rapamycin, Allan = allantoin. Image credit: Sven Bulterijs

The evolutionary distance between fruit flies and worms is larger than between fruit flies and humans. Hence the fact that the drug combinations worked in both fruit flies and worms offers hope that their lifespan extending effect may be conserved in humans. The two drug combinations (Rap + Rif and Rap + Rif + Allan) significantly extended mean and maximum lifespan in fruit flies.

Dessale T et al. (2017). Slowing ageing using drug synergy in C. elegans. bioRxiv: 153205.

Sulforaphane increases metabolism to ameliorate obesity and insulin resistance

Study published here: Glucoraphanin ameliorates obesity and insulin resistance through adipose tissue browning and reduction of metabolic endotoxemia in mice

Low-grade sustained inflammation, triggered by chronically high levels of proinflammatory cytokines and gut microbiota-derived circulatory lipopolysaccharide (LPS), links obesity with comorbidities such as insulin resistance and nonalcoholic fatty liver disease (NAFLD) (1,2).

Although a number of pharmacological treatments for obesity and NAFLD have been tested, few drugs are clinically available owing to the lack of long- term efficacy and safety concerns (3,4). Thus, a novel therapeutic approach that would improve energy metabolism and reduce chronic inflammation in obesity is sorely needed.

Nuclear factor (erythroid-derived 2)-like 2 (Nrf2), a basic leucine zipper transcription factor, is widely expressed in human and mouse tissues, and serves as a defense response against extrinsic and intrinsic stressors (5). Upon exposure to electrophilic and oxidative stress, Nrf2 detaches from its repressor, Kelch-like ECH-associated protein 1-nuclear factor (Keap1), and is translocated from the cytoplasm into the nucleus.

This translocation leads to the transcriptional activation of genes encoding phase 2 detoxifying and antioxidant enzymes (6).

In addition to the ubiquitous induction of cytoprotective genes, Nrf2 regulates a large number of genes involved in glucose and lipid metabolism. In the liver, the constitutive activation of Nrf2 via Keap1 knockdown represses the expression of genes involved in gluconeogenesis (7) and lipogenesis (8), thereby alleviating obesity, diabetes, and hepatic steatosis.

Accordingly, synthetic Nrf2 inducers such as synthetic triterpenoid 2-cyano-3,12-dioxoolean-1,9-dien-28-oic acid (CDDO)-imidazolide (9), CDDO-methyl ester (known as bardoxolone methyl) (10), and dithiolethione analog, oltipraz (11), have been shown to ameliorate high-fat diet (HFD)-induced obesity and diabetes.

These synthetic Nrf2 inducers also decrease liver and adipose tissue lipogenesis, and enhance glucose uptake in skeletal muscles. However, the mechanisms by which Nrf2 enhances energy metabolism in response to a HFD remain largely unknown.

Although enhanced Nrf2 signaling has shown promising results in several animal studies, the synthetic Nrf2 inducers have caused adverse cardiac events and gastrointestinal toxicities in clinical trials (12,13).

These observations prompted us to explore a safer Nrf2 inducer for the treatment of obesity, insulin resistance, and NAFLD.

Sulforaphane, an isothiocyanate derived from cruciferous vegetables, is one of the most potent naturally occurring Nrf2 inducers; this compound exhibits anticancer activity in cancer cell lines and in carcinogen-induced rodent models (14).

Among the cruciferous vegetables, broccoli sprouts are the best source of glucoraphanin, a stable glucosinolate precursor of sulforaphane (15). In both rodents and humans, glucoraphanin is hydrolyzed by gut microbiota-derived myrosinase into bioactive sulforaphane prior to intestinal absorption (16). A recent clinical study demonstrated the safety of orally administered glucoraphanin (17).

In the present study, we examined the dietary glucoraphanin-mediated modulation of systemic energy balance and the mitigation of chronic inflammation, insulin resistance, and NAFLD in diet-induced obese mice.

Research Design and Methods Glucoraphanin preparation
The sulforaphane precursor, glucoraphanin, was prepared as previously described (17) with minor modifications. Briefly, one day after germination from broccoli seeds (Caudill Seed Company, Louisville, KY), sprouts were boiled in water for 30 min.

The water extract was mixed with dextrinized cornstarch and subsequently spray-dried to yield an extract powder containing 135 mg of glucoraphanin per gram (0.31 mmol/g) (18). The total glucoraphanin titer in the resulting powder was determined by HPLC as previously reported (19).

Mice and diets
Male C57BL/6JSlc mice were purchased from Japan SLC (Hamamatsu, Japan) at 7 weeks of age. Nrf2-knockout (Nrf2-/-) mouse strain (RBRC01390; C57BL/6J background) was provided by RIKEN BRC (Tsukuba, Japan) (6). After a week of acclimation, mice were fed normal chow (NC; containing 2.2% dextrinized cornstarch, 10% kcal from fat, Research Diets, New Brunswick, NJ), NC containing 0.3% glucoraphanin (NC- GR; containing 2.2% extract powder), a high-fat diet (HFD; containing 2.2% dextrinized cornstarch, 60% kcal from fat, #D12492; Research diets), or a HFD containing 0.3% glucoraphanin (HFD-GR; containing 2.2% extract powder) for 14 weeks.

Both the NC and the HFD containing cornstarch or glucoraphanin were prepared by Research Diets. All mice studied were maintained on a 12 h light/dark cycle at 24–26°C with free access to water and food. All animal procedures were performed in accordance with the Guidelines for the Care and Use of Laboratory Animals at Kanazawa University, Japan.

Indirect calorimetry
After 3 weeks of feeding, mice were individually housed in an indirect calorimeter chamber at 24–26°C (Oxymax; Columbus Instruments, Columbus, OH). Calorimetry, daily body weight, and daily food intake data were acquired during a 3-day acclimation period, followed by a 2-day experimental period. Oxygen consumption (VO2) and carbon dioxide production (VCO2) were measured in each chamber every 20 minutes. The respiratory exchange ratio (RER = VCO2/VO2) was calculated using Oxymax software. Energy expenditure was calculated as shown below and normalized to the body mass of each subject.
Energy expenditure = VO2 × [3.815 + (1.232 ×RER)]

Metabolic measurements and biochemical analyses
Metabolic parameters, body fat composition, insulin sensitivity, and glucose tolerance were assessed as previously described (20). Plasma LPS levels were analyzed using a Limulus amebocyte lysate assay kit (QCL-1000; Lonza, Allendale, NJ). Plasma LPS binding protein (LBP) levels were determined using an ELISA kit (Enzo Life Sciences, Farmingdale, NY).

Immunoblotting was performed with primary antibodies (Supplementary Table 1) as previously described (20). mRNA expression levels were determined by quantitative real-time PCR using SYBR Green with the primers (Supplementary Table 2) as previously described (20).

Isolation and differentiation of inguinal white adipose tissue-derived primary beige adipocytes
Stromal vascular fractions (SVFs) from inguinal white adipose tissue (WAT) of 7- week-old wild-type and Nrf2-/- mice were prepared as previously reported (21). At confluence, SVF cells were induced for 2 days with differentiation medium containing DMEM/F-12 supplemented with 10% FBS, 20 nM insulin, 1 nM T3, 5 μM dexamethasone, 500 μM isobutylmethylxanthine, 125 μM indomethacin, and 0.5 μM rosiglitazone (all from Sigma-Aldrich, St. Louis, MO).

Induced cells were subsequently cultured in maintenance medium (DMEM/F-12 containing 10% FBS, 20 nM insulin, and 1 nM T3) for 5 days and treated with DMSO or sulforaphane (Toronto Research Chemicals, Toronto, Canada) at the indicated concentrations for 48 h.

Fluorescence-activated cell sorting (FACS)
Cells from the liver and epididymal WAT were prepared as previously described (22). Isolated cells were incubated with Fc-Block (BD Bioscience, San Jose, CA), and subsequently incubated with fluorochrome-conjugated antibodies (Supplementary Table 1).

Flow cytometry was performed using a FACSAria II (BD Bioscience), and the data were analyzed using FlowJo software (Tree Star, Ashland, OR).
Analysis of gut microbiota via pyrosequencing of the 16S rRNA gene
Metagenomic DNA was extracted from mouse cecal content with a QIAamp DNA stool kit (Qiagen, Hilden, Germany). The V1–V2 region of the 16S rRNA gene was amplified using primer sets as previously reported (23).

Mixed samples were prepared by pooling approximately equal amounts of PCR amplicons from each sample and subjected to a GS Junior System (Roche Diagnostics, Basel, Switzerland) for subsequent 454 sequencing.

Pre-processing and taxonomic assignment of sequencing reads were conducted as described previously (23) and separated by unique barcodes. The 16S rRNA sequence database was constructed by retrieving 16S sequences of bacterial isolates (1200–2384 bases in length) from the Ribosomal Database Project Release 10.27.

We used 4000 filter- passed reads of 16S sequences for the operational taxonomic unit (OTU) analysis of each sample. Clustering of 16S sequence reads with identity scores > 96% into OTUs was performed using UCLUST (

Representative sequences with identity scores > 96% for each OTU were assigned to bacterial species using BLAST. Principal component analysis using EZR software ( sct/SaitamaHP.files/statmedEN.html) was applied for assessment of alterations of cecal bacterial phylum associated with diets.

Statistical analyses
Data were expressed as mean ± SEM. P < 0.05 was considered statistically significant. Statistical differences between pairs of groups were determined by a two-tailed Student’s t-test. An overall difference between more than two groups was determined using a one-way ANOVA.

If one-way ANOVAs were significant, differences between individual groups were estimated using a Bonferroni post hoc test All calculations were performed using SPSS Statistics (v19.0, IBM, Armonk, NY).


Glucoraphanin decreases weight gain and adiposity, and increases energy expenditure in high-fat diet (HFD)-fed mice
To investigate the effects of glucoraphanin on systemic energy balance, we examined the body weight of wild-type mice fed normal chow (NC) or a high-fat diet (HFD), supplemented with glucoraphanin or vehicle (i.e., cornstarch only). Glucoraphanin reduced weight gain only in HFD-fed mice without affecting food intake (Fig. 1A and Supplementary Fig. 1A).

This reduction was not accompanied by evidence of gross toxicity. We determined the plasma concentration of sulforaphane in NC-GR and HFD-GR mice, but not in NC or HFD mice, indicating that glucoraphanin was absorbed as a sulforaphane following food consumption (Supplementary Fig. 1B).

The reduction of weight gain in glucoraphanin-treated HFD-fed mice was largely attributed to the decreased fat mass but not lean mass (Fig. 1B). To assess energy expenditure, we placed the mice in indirect calorimetry cages after 3 weeks of feeding, before an evident change in the body mass of HFD-fed mice was observed (HFD: 29.9 ± 0.5 g vs. HFD-GR: 28.5 ± 0.5 g).

Glucoraphanin-treated mice fed the HFD exhibited consistently higher VO2 and VCO2 than vehicle-treated HFD-fed controls (Fig. 1C and D), leading to increased energy expenditure (Fig. 1E); however, they displayed similar RER (Fig. 1F), suggesting that glucoraphanin supplementation enhanced sugar and fat use under HFD conditions. On NC-fed mice, glucoraphanin did not affect these parameters of energy balance (Fig. 1B–F).

Consistent with increased energy expenditure, glucoraphanin increased the core body temperature of HFD-fed mice by approximately 0.5°C (Fig. 1G).

Glucoraphanin improves diet-induced insulin resistance and glucose tolerance
After 14 weeks of feeding, glucoraphanin supplementation did not affect plasma triglycerides, total cholesterol, and free fatty acid (FFA) levels in either NC- or HFD-fed mice (Table 1). On NC, blood glucose levels were not altered by glucoraphanin, but on the HFD, glucoraphanin-treated mice exhibited significantly lower fasted blood glucose compared with vehicle-treated controls (Table 1).

Additionally, glucoraphanin significantly decreased plasma insulin concentrations in HFD-fed mice under both fasted and fed conditions, resulting in lower homeostatic model assessment-insulin resistance (HOMA-IR; Table 1).

During the insulin tolerance test (ITT), glucoraphanin significantly enhanced the reduction in blood glucose levels in HFD-fed mice, but not in NC-fed mice compared with the vehicle-treated controls (Fig. 2A). Glucoraphanin improved glucose tolerance in HFD- fed mice during the glucose tolerance test (GTT), but had no effect on NC-fed mice (Fig. 2B).

Insulin secretion during the GTT was not affected by glucoraphanin (data not shown). In line with increased insulin sensitivity, insulin-stimulated Akt phosphorylation on Ser473 was enhanced by glucoraphanin in the liver, muscle, and epididymal WAT of mice fed the HFD (Fig. 2C).

Glucoraphanin does not exert anti-obesity and insulin-sensitizing effects in Nrf2-/- mice
Although the Keap1-Nrf2 pathway is a well-known target of sulforaphane, this isothiocyanate has also been reported to modulate different biological pathways independent of the Keap1-Nrf2 pathway (24,25).

To determine whether the anti-obesity and insulin-sensitizing effects of glucoraphanin are mediated through Nrf2, the effects of glucoraphanin on energy balance and glucose metabolism were assessed in NC- and HFD- fed Nrf2-/- mice.

Although food intake and plasma concentration of sulforaphane in NC-GR or HFD-GR diet-fed Nrf2-/- nice were comparable that detected in wild-type mice fed NC- GR or HFD-GR diet (Supplementary Fig. 1C and D), the effects of glucoraphanin following HFD feeding on weight gain (Fig. 3A), VO2 (Fig. 3B), VCO2 (Fig. 3C), energy expenditure (Fig. 3D), RER (Fig. 3E), rectal temperature (Fig. 3F), insulin sensitivity (Fig. 3G), and glucose tolerance (Fig. 3H) were abolished by the Nrf2 deficiency.

These data are consistent with comparable plasma metabolic parameters between glucoraphanin-treated and vehicle-treated mice on the HFD. These metabolic parameters include lipids, blood glucose, insulin, HOMA-IR, and liver enzymes such as alanine transaminase (ALT) and aspartate transaminase (AST) (Supplementary Table 3).

Glucoraphanin blocks HFD-induced reduction of Ucp1 expression in WAT of wild- type mice but not in Nrf2-/- mice

The increased energy expenditure and body temperature of glucoraphanin-treated HFD-fed mice suggest an increase in adaptive thermogenesis. However, glucoraphanin supplementation had little effect on the size and number of lipid droplets in the intrascapular brown adipose tissue (BAT) of HFD-fed wild-type mice (Supplementary Fig. 2A).

In addition, the mRNA expression of uncoupling proteins (Ucps), PGC-1α, and deiodinase 2 in BAT and of Ucps in skeletal muscle was not altered by glucoraphanin supplementation in both NC- and HFD-fed wild-type mice (Supplementary Fig. 2B and C). In BAT, HFD increased Ucp1 protein expression, but glucoraphanin did not alter the expression in both wild-type and Nrf2-/- mice (Fig. 4A).

Brown-like adipocytes expressing Ucp1, also known as beige cells, exist in various WAT depots and can contribute to thermogenesis (26). Compared with NC, HFD significantly decreased Ucp1 protein levels in epididymal and inguinal WAT of both wild-type and Nrf2-/- mice (Fig. 4A).

Glucoraphanin supplementation restored HFD-induced reduction in Ucp1 protein levels in epididymal and inguinal WAT of wild-type mice but not those in Nrf2-/- mice.

To examine whether the effects of glucoraphanin were fat cell-autonomous and Nrf2-mediated, we tested the effects of sulforaphane, an active metabolite of glucoraphanin, on the expression of brown fat-selective genes in primary beige adipocytes obtained from inguinal WAT of wild-type and Nrf2-/- mice.

In beige adipocytes derived from wild-type mice, treatment with sulforaphane induced the Nrf2 target gene, NAD(P)H:quinone oxidoreductase 1 (Nqo1; Fig. 4B) and antioxidant genes (Supplementary Fig. 3A).

Concurrently, sulforaphane significantly increased the mRNA expression of brown-fat selective genes, including Ucp1, Prdm16, Cidea, and Elovl3 (Fig. 4B).

In contrast, in Nrf2-deficient beige adipocytes, sulforaphane failed to activate Nrf2 as judged by unaltered mRNA expression of the target genes and to promote the expression of brown-fat selective genes (Supplementary Fig. 3B and Fig. 4C).

Importantly, Nrf2-deficient beige adipocytes exhibited less differentiation levels associated with attenuated lipid accumulation (Supplementary Fig. 3C) and lower mRNA expression of fatty acid binding protein 4 (Supplementary Fig. 3C) and brown-fat selective genes compared with wild-type beige adipocytes (Fig. 4C).

Glucoraphanin reduces hepatic steatosis and oxidative stress in HFD-fed mice
The HFD causes hepatic steatosis and inflammation, eventually leading to steatohepatitis. As shown in Fig. 5A, the increase in liver weight caused by the 14-week- HFD was alleviated by glucoraphanin supplementation.

Glucoraphanin also attenuated HFD-induced hepatic steatosis (Fig. 5B). Additionally, compared with the HFD group, the lower levels of plasma ALT, plasma AST, liver triglycerides, and liver FFAs in the HFD- GR group indicate that glucoraphanin alleviated HFD-induced liver damage (Fig. 5C and D).

The reduction in hepatic steatosis was accompanied by the decreased expression of the following lipogenic genes: sterol regulatory element binding protein-1c (Srebf1), fatty acid synthase (Fasn), and peroxisome proliferator-activated receptor gamma (Pparγ) (Fig. 5E).

Additionally, hepatic levels of malondialdehyde, a marker of lipid peroxidation, were increased by the HFD. Glucoraphanin attenuated lipid peroxidation (Fig. 5F) and decreased gene expression of the NADPH oxidase subunits gp91phox, p22phox, p47phox, and p67phox (Fig. 5E).

The HFD led to a compensatory increase in the expression of genes involved in fatty acid β-oxidation (Ppara and Cpt1a) and anti-oxidative stress (Cat, Gpx1, and Sod1) in the liver. However, glucoraphanin did not increase the expression of these genes in the liver of HFD-fed mice further (Fig. 5E).

Glucoraphanin suppresses HFD-induced proinflammatory activation of macrophages in liver and adipose tissue
In response to the HFD, liver-resident macrophages (Kupffer cells) increase the production of proinflammatory cytokines that promote insulin resistance and NAFLD in mice (27). In particular, chemokine (C-C motif) ligand 2 (Ccl2) promotes the recruitment of chemokine (C-C motif) receptor 2 (Ccr2)-positive monocytic lineages of myeloid cells into the liver (28). These recruited cells produce a large amount of proinflammatory mediators and activate a lipogenic program (28).

Here, we found a prominent induction of tumor necrosis factor-α (Tnf-α), Ccl2, and Ccr2 in the liver of HFD-fed mice, which was markedly reduced in glucoraphanin-treated mice (Fig. 6A).

Glucoraphanin significantly suppressed HFD-induced inflammatory pathways such as c-Jun N-terminal kinase (JNK) and extracellular signal-regulated kinase (Erk) (Fig. 6B).

Glucoraphanin tended to decrease levels of p-NF-κB p65 (Ser536) in HFD-fed mice, although this decrease was not statistically significant (Fig. 6B). Of note, in the liver of Nrf2-/- mice, glucoraphanin failed to suppress HFD-induced inflammatory signal pathways (Supplementary Fig. 4).

In addition, glucoraphanin significantly decreased the HFD-induced hepatic expression of macrophage markers, including F4/80, Cd11b, and Cd68 (Fig. 6C). Tissue macrophages are phenotypically heterogeneous and have been characterized according to their activation/polarization state as M1-like proinflammatory macrophages or M2-like anti- inflammatory macrophages (29).

Consistent with the decreased expression of macrophage markers, glucoraphanin prevented macrophage (F4/80+CD11b+ cell) accumulation in the liver of HFD-fed mice (Fig. 6D).

Additionally, glucoraphanin decreased the number of M1- like liver macrophages expressing surface markers (F4/80+CD11b+CD11c+CD206−) (Fig. 6E).

In contrast, glucoraphanin increased the number of M2-like liver macrophages (F4/80+CD11b+CD11c−CD206+), resulting in a predominantly M2-like macrophage population (Fig. 6E).

Moreover, glucoraphanin decreased the mRNA expression of Tnf-α and NADPH oxidase in the epididymal WAT of HFD-fed mice (Supplementary Fig. 5A).

Although the HFD-induced expression of macrophage markers and macrophage accumulation in epididymal WAT were not altered by glucoraphanin (Supplementary Fig. 5B and C), the number of M1-like macrophages was significantly decreased in the epididymal WAT of glucoraphanin-treated HFD-fed mice (Supplementary Fig. 5D).

Glucoraphanin decreases circulating LPS and the relative abundance of Proteobacteria in the gut microbiomes of HFD-fed mice

Gut microbiota-derived LPS induces chronic inflammation that eventually leads to insulin resistance in obesity, termed metabolic endotoxemia (1,2).

Based on our observation that glucoraphanin alleviates inflammation in the liver and epididymal WAT of HFD-fed mice, we subsequently investigated the effects of glucoraphanin on metabolic endotoxemia and gut microbiota.

In accordance with previous studies (1,2), the HFD induced a 2-fold increase in circulatory LPS levels, which was reduced by glucoraphanin supplementation (Fig. 7A). Furthermore, plasma and hepatic levels of the LPS marker, LBP, were significantly elevated by the HFD and were reduced by glucoraphanin supplementation (Fig. 7B).

A principal component analysis distinguished cecal microbial communities based on diet and treatment, revealing that the metagenomes of HFD-fed mice formed a cluster distinct from the cluster formed by NC-fed mice (Fig. 7C). However, samples from glucoraphanin-treated HFD-fed mice formed a cluster that was indistinguishable from that of vehicle- or glucoraphanin-treated NC-fed mice (Fig. 7C).

Importantly, consistent with previous reports (30,31), further analysis at the phylum level demonstrated that the proportion of Gram-negative Proteobacteria was significantly elevated in the gut microbiomes of HFD-fed mice, which was suppressed by glucoraphanin supplementation (Fig. 7D and E).

The increase in the relative abundance of Proteobacteria in HFD-fed mice is mostly explained by an increase in the relative abundance of bacteria from the family Desulfovibrionaceae (Fig. 7F), key producers of endotoxins in animal models of obesity (30).

In fact, the relative abundance of Desulfovibrionaceae was positively correlated with plasma LPS levels (Fig. 7G). Furthermore, plasma LPS levels were significantly and positively correlated with the hepatic mRNA levels of Tnf-α, gp91phox, and F4/80 (Fig. 7H).

Similarly, the liver expression of other marker genes was both significantly and positively correlated with plasma LPS levels and with one another (Supplementary Table 4).

In the present study, we demonstrated that glucoraphanin, a stable precursor of the Nrf2 inducer sulforaphane, mitigated HFD-induced weight gain, insulin resistance, hepatic steatosis, oxidative stress, and chronic inflammation in mice.

The weight-reducing and insulin-sensitizing effects of glucoraphanin were abolished in Nrf2-/- mice. Additionally, glucoraphanin lowered plasma LPS levels in HFD-fed mice, and decreased the relative abundance of Desulfovibrionaceae.

At the molecular level, glucoraphanin increased Ucp1 protein expression in WAT depots, while suppressing the hepatic mRNA expression of genes involved in lipogenesis, NADPH oxidase, and inflammatory cytokines.

Our data suggest that in diet-induced obese mice, glucoraphanin restores energy expenditure and limits gut-derived metabolic endotoxemia, thereby preventing hepatic steatosis, insulin resistance, and chronic inflammation.

Consistent with previous reports demonstrating the anti-obesity effects of synthetic Nrf2 inducers (9–11), we show that the oral administration of glucoraphanin mitigates HFD-induced weight gain (Fig. 1A). The dose of glucoraphanin used in the current study (about 12 μmol/mouse/day) is similar to that used in other experiments investigating its antitumor effects in mice (14,32,33).

Here, we show that the effect of glucoraphanin on whole-body energy expenditure and the protein expression of Ucp1 in WAT were abolished in Nrf2-/- mice (Fig. 3D and 4A).

A recent study using adipocyte-specific PRDM16- deficient mice indicated that adaptive thermogenesis in beige fat also contributes to systemic energy expenditure (26). The mutant mice in the aforementioned study, which exhibited markedly reduced Ucp1 mRNA expression in inguinal WAT and minimal effects on BAT, developed obesity and insulin resistance in response to a HFD.

Thus, we believe that the increased energy expenditure in glucoraphanin-treated HFD-fed mice stems from, at least in part, an increase in beige fat, even though the expression of Ucps in BAT and skeletal muscle is not altered. Further analysis using Ucp1-knockout mice will elucidate the relative contribution of beige fat to the Nrf2-mediated metabolic effects elicited by glucoraphanin.

Several studies indicated that Nrf2-/- mice are partially protected from HFD-induced obesity and associated with milder insulin resistance compared with wild-type counterparts (9,34,35). Recently, Schneider et al. demonstrated that mitigation of HFD-induced obesity in Nrf2-/- mice, which was 25% less body weight than that of wild-type mice after 6 weeks of feeding (35).

They also found that HFD-fed Nrf2-/- mice exhibit a 20–30% increase in energy expenditure that is associated with an approximately 3-fold up-regulation of Ucp1 protein expression in abdominal WAT (35). In the present study, Nrf2-/- mice gained less weight after 6 weeks of HFD feeding compared with HFD-fed wild-type mice (Figure 1A and 3A; Nrf2-/-: 35.9 ± 0.9 g vs. wild-type: 39.0 ± 0.8 g, P < 0.05).

The lower body mass of Nrf2-/- mice raises the possibility that anti-obesity effect of glucoraphanin was completely phenocopied by Nrf2 gene deficiency. However, several observations suggest that Nrf2-/- mice only partially phenocopy the effect of glucoraphanin on weight gain reduction. 1)

In the present study, the weight difference between Nrf2-/- and wild-type mice was only 8%, which is much less than that in the previous study (35). 2) In Nrf2-/- mice, compared with NC, HFD induced significant weight gain (Fig. 3A), glucose intolerance (Fig. 3H), and insulin resistance as judged by increased HOMA-IR (Supplementary Table 3). 3)

Metabolic rate and energy expenditure of Nrf2-/- mice were comparable with those in wild-type mice (Fig. 1C–E and 3B–D). 4) Ucp1 protein levels in both epididymal WAT and inguinal WAT of HFD-fed Nrf2-/- mice were lower than those in HFD-GR-fed wild-type mice (Fig. 4A).

Taken together, these findings suggest that Nrf2 gene deficiency is not sufficient to block HFD-induced obesity by increasing energy expenditure and Ucp1 expression in WAT depots, and to mask the effect of glucoraphanin.

However, we cannot fully exclude the possibility that the effects of glucoraphanin are mediated by Nrf2 independent-mechanisms. Possible reasons for the discordance in metabolic phenotypes of Nrf2-/- between the previous study (35) and ours may be due to differences in knockout mouse lines and experimental conditions (e.g. age of mice at beginning of HFD feeding, composition of HFD, and temperature in the metabolic chamber).

Our in vitro study of primary beige adipocytes revealed that sulforaphane promotes the expression of brown-fat selective genes (Fig. 4B). Importantly, the concentration of sulforaphane used in cell culture (0.2–5 μM) is comparable with that detected in mice fed NC-GR and HFD-GR diet (Supplementary Fig. 1B). Moreover, we determined that Nrf2 acts as a positive regulator of beige adipocyte differentiation (Fig. 4C and Supplementary Fig. 3C).

The less differentiation levels in Nrf2-deficient beige adipocytes are in agreement with previous reports demonstrating Nrf2 induces white adipocyte differentiation through increasing the gene expression of Pparγ (34) and Cebpβ (36), common transcription factors regulating the differentiation of brown, beige, and white adipocytes.

Furthermore, it is noteworthy that Nrf2 has been reported to bind NF-E2-binding sites in the 5′-flanking region of the human and rodent Ucp1 genes (37).

However, we cannot exclude the possibilities that glucoraphanin affects sympathetic nervous activity or that hormonal factors are regulating fat browning (38). In addition, mitochondrial reactive oxidative species facilitate Ucp1-dependent respiration in BAT and whole-body energy expenditure by promoting the sulfenylation of a specific cysteine residue (Cys253) in Ucp1 (39).

The molecular mechanism by which Nrf2 regulates the expression and thermogenic activity of Ucp1 in beige adipocytes requires further investigation.

Glucoraphanin supplementation improved the systemic glucose tolerance and insulin sensitivity of HFD-fed mice. Although the molecular mechanism by which synthetic Nrf2 inducers enhance glucose uptake is unclear, AMP-activated protein kinase (AMPK) activation may mediate this enhancement in mouse skeletal muscle and adipose tissue (9– 11).

The phosphorylation levels of AMPK (Thr172) and acetyl-CoA carboxylase (Ser79) in peripheral insulin target tissues were comparable between glucoraphanin-treated NC- and HFD-fed mice and vehicle-treated controls (Supplementary Fig. 6).

These data suggest that AMPK activation is not necessary for glucoraphanin to exert its insulin-sensitizing effect on HFD-fed mice. Additional studies using the hyperinsulinemic-euglycemic clamp technique are needed to determine which tissues contribute to the insulin-sensitizing effects of glucoraphanin.

The beneficial effects of glucoraphanin on hepatic lipid metabolism were not accompanied by AMPK activation or the increased expression of fatty acid β-oxidation genes (Fig. 5E).

Instead, glucoraphanin mitigated HFD-induced oxidative stress and inflammation in the liver. In obesity, hepatic inflammation mediated by macrophage/monocyte-derived proinflammatory cytokines promotes lipogenesis through the inhibition of insulin signaling and SREBP activation (40,41). In fact, the depletion of Kupffer cells by clodronate liposomes ameliorates hepatic steatosis and insulin sensitivity in HFD-fed mice (27).

Furthermore, Ccl2- or Ccr2-deficient mice are protected from diet- induced hepatic steatosis even though they still become obese (42,43). Moreover, the specific ablation of M1-like macrophages restores insulin sensitivity in diet-induced obese mice (44), while the deletion of Pparδ, which promotes M2 activation, predisposes lean mice to develop insulin resistance (45).

Therefore, decreased hepatic macrophage accumulation and M2-dominant polarization of hepatic and adipose macrophages account, at least in part, for the protection from hepatic steatosis and insulin resistance in glucoraphanin-treated HFD-fed mice.

One of the most important findings of this study is that glucoraphanin decreases the relative abundance of Gram-negative Proteobacteria, particularly family Desulfovibrionaceae, while reducing circulatory LPS levels (Fig. 7).

Recent studies demonstrated that a significant increase in Desulfovibrionaceae, potential endotoxin producers, in the gut microbiomes of both HFD-induced obese mice and obese human subjects compared to lean individuals (30,31,46). We cannot exclude the possibility that other microbiota-derived products, such as bile acids and short chain fatty acids, also mediate the metabolic action of glucoraphanin.

Whether the interaction between sulforaphane and gut microbiota is affected directly or indirectly by altered host physiology remains to be determined.

However, several studies have suggested that sulforaphane can alter the gut microbiota directly, as isothiocyanates (including sulforaphane) have been shown to exhibit antibacterial activity against Proteobacteria (47,48).

This activity may proceed via redox disruption and enzyme denaturation reactions involving the isothiocyanate reactive group, -N=C=S, the thiol group (-SH) of glutathione and proteobacterial proteins (47,48). Additionally, sulforaphane exhibits antibacterial activity against Helicobacter pylori, a member of the phylum Proteobacteria (49).

We are unaware of any previous reports demonstrating that isothiocyanates inhibit the proliferation of Desulfovibrionaceae. The mechanistic underpinnings of this antibacterial activity require elucidation.

In conclusion, the results of the present study indicate that glucoraphanin may be effective in preventing obesity and related metabolic disorders such as NAFLD and type 2 diabetes.

A recent clinical study demonstrated that supplementation with a dietary dose of glucoraphanin (69 μmol/day) for two months significantly decreased the plasma liver enzymes, ALT and AST, although body mass did not change (18).

Long-term treatment with a higher dose of glucoraphanin (800 μmol/day), which can be safely administered without any harmful side effects (17), may be required to achieve an anti-obesity effect in humans.

Enhanced Nutraceutical bioactivity to fight cancer

1. Introduction
Nutraceuticals are biologically active molecules found in foods that may not be essential for maintaining normal human functions, but may enhance human health and wellbeing by inhibiting certain diseases or improving human performance (Gupta, 2016; Wildman & Kelley, 2007). Numerous different classes of nutraceuticals are found in both natural and processed foods including carotenoids, flavonoids, curcuminoids, phytosterols, and certain fatty acids (Gupta, 2016).

Many of these nutraceuticals have the potential to act as anticancer agents, and may therefore be suitable for incorporation into functional or medical foods as a means of preventing or treating certain types of cancer. Nutraceuticals vary considerably in their chemical structures, physiochemical properties, and biological effects (Bagchi, Preuss, & Sarwoop, 2015; Gupta, 2016).

For example, nutraceuticals vary in their molar mass, structure, polarity, charge and functional groups, which influences their chemical reactivity, physical state, solubility characteristics, and biological fate and functions (McClements, 2015b).

Some nutraceuticals are naturally present in whole foods, such as fruits, vegetables, and cereals, and are therefore often consumed in this form. Conversely, other nutraceuticals are isolated from their natural states and converted into additives that can be incorporated into functional foods, dietary supplements, or pharmaceuticals.

In this article, we will mainly focus on the delivery of nutraceuticals using foods, as it has been proposed that increased consumption of foods rich in nutraceuticals may decrease the risk of certain types of chronic diseases, including cancer. However, if consumers are going to benefit from consuming foods containing nutraceuticals, it is important that they have certain characteristics:

There are a number of factors that currently limit the utilization of many types of
anticancer nutraceuticals in functional foods (Gleeson, Ryan, & Brayden, 2016;
McClements, 2015a; McClements, Li, & Xiao, 2015b). Firstly, many of nutraceuticals
cannot easily be incorporated into foods because they have poor-solubility characteristics,
or they cause undesirable changes in appearance, texture, or flavor of foods. Second,
many nutraceuticals are chemically or biochemically unstable and therefore lose their
bioactivity because they are degraded within food products or the human body. Third,
many nutraceuticals have a low bioavailability and therefore only a small fraction of
them are actually absorbed and utilized by the body. Fourth, for some nutraceuticals the
optimum dose has not been established, and therefore it is unclear how much to deliver in
a bioactive form, e.g., the anticancer efficacy of resveratrol actually decreases as the dose
increases (Cai, Scott, Kholghi, Andreadi, Rufini, Karmokar, et al., 2015)
The purpose of this review article is to highlight how food matrices can be designed
to enhance the biological activity of anticancer nutraceuticals. In particular, we focus on
two different approaches that can be utilized for this purpose (Figure 1): very systems: In this approach, nutraceuticals isolated from their original environment are encapsulated within a delivery system that is specifically designed to enhance their bioavailability (Arora & Jaglan, 2016; Gleeson, Ryan, & Brayden, 2016; McClements, Decker, & Weiss, 2007; Sagalowicz & Leser, 2010). An example of this approach would be a carotenoid ingredient isolated from carrots that is encapsulated within an oil-in-water emulsion that could then be added to foods such as soft drinks, yogurts, dressings, or sauces (Salvia- Trujillo, Qian, Martin-Belloso, & McClements, 2013a).

In this approach, a nutraceutical-rich food is co-ingested with an excipient food, which is again specially designed to improve the bioavailability of the nutraceuticals (McClements & Xiao, 2014). An example of this approach would be an oil-in-water excipient emulsion consumed at the same time as carotenoid-rich carrots (Zhang, Zhang, Zou, Xiao, Zhang, Decker, et al., 2016b).
nutraceuticals should initially be present in functional foods at a sufficiently high level to have a beneficial physiological effect. The nutraceuticals should remain stable within the functional foods during manufacturing, storage, and utilization, otherwise they may lose their beneficial health effects.
nutraceuticals should not have an adverse effect on the color, taste, or shelf- life of a food product. After ingestion, the nutraceuticals should be released from the functional foods and delivered to the appropriate site-of-action within the human body.

Initially, a brief overview of some of the most important anticancer nutraceuticals that have been identified in foods is given. Some of the major factors limiting the bioavailability of nutraceuticals is then discussed, and potential approaches to overcome these limitations are highlighted. The design of delivery and excipient systems to enhance the bioavailability of anticancer nutraceuticals is then described.

2. Anticancer nutraceuticals
Accumulating evidence suggests that many nutraceuticals, such as curcumin, resveratrol, tea polyphenols, sulforaphane, anthocyanins, genistein, quercetin and
lycopene, exhibit anticancer activities against various forms of cancer (Arvanitoyannis & van Houwelingen-Koukaliaroglou, 2005; Ullah & Khan, 2008).

One of the important advantages of utilizing nutraceuticals to prevent and treat cancer is that they generally exhibit little or no adverse effects frequently associated with pharmaceutical agents after long-term administration. Nutraceuticals have been found to exert a wide range of cellular effects. The possible mechanisms of action of anticancer nutraceuticals include induction of cell cycle arrest and apoptosis in cancerous cells, detoxification of highly reactive molecules, activation of the host immune system, and sensitization of malignant cells to cytotoxic agents (Kotecha, Takami, & Espinoza, 2016; Pan & Ho, 2008). In this section, we focus on several important anticancer nutraceuticals that have beenintensively investigated with particular emphasis on the clinical evidence supporting the safety and efficacy of these compounds in cancer prevention and treatment.

Curcumin: Curcumin is a polyphenol found in turmeric (Curcuma longa), a member of the ginger family (Zingiberaceae) (Joe, Vijaykumar, & Lokesh, 2004). A large number of in vitro and in vivo studies have shown that curcumin inhibits the development of various cancers by inducing cell cycle arrest and cellular apoptosis, through pleiotropic modulation on several key cancer targets such as Wnt/-catenin, nuclear factor kappa B (NF-  B), cyclooxygenase-2 (COX-2), tumor necrosis factor alpha (TNF- ), STAT-3 and cyclin D1(Perrone, Ardito, Giannatempo, Dioguardi, Troiano, Lo Russo, et al., 2015). Several phase I and phase II clinical trials have been conducted and demonstrated the safety and anticancer effects of curcumin in patients with different malignancies including myeloma, pancreatic and colorectal cancer (Kotecha, Takami, & Espinoza, 2016). Since curcumin is preferentially distributed in the colonic mucosa, in comparison to other tissues, many clinical trials have been focused on its chemo-preventive efficacy on colorectal cancer (Bar-Sela, Epelbaum, & Schaffer, 2010).

In a phase IIa clinical trial, patients taking 4 grams of curcumin daily were found to have a 40% decrease in the number of aberrant crypt foci (ACF) lesions, which are one of the early histologic signs seen in the colon that may lead to cancer (Carroll, 2012). One of the most important factors limiting the bioefficacy of curcumin is its poor bioavailability.

In animal studies, various edible delivery systems have been developed to improve the bioavailability and bioactivities of curcumin, including liposomes, phospholipid complexes, organogel-based nanoemulsions, chitosan-based nanoparticles and self-emulsifying drug delivery systems (Ting, Jiang, Ho, & Huang, 2014). For example, polylactide co-glycolide nanocapsulated curcumin suppressed cell proliferation, induced cancer cell apoptosis and improved pathological structures in a hepatocellular carcinoma model, whereas an identical concentration of free curcumin was found to be ineffective (Ghosh, Choudhury, Ghosh, Mandal, Sarkar, Ghosh, et al., 2012).

In a recent crossover study, liquid micellar formulations of curcumin showed a 185-fold enhancement in bioavailability within 24 hours without increased toxicity in healthy subjects, compared to powdered curcuminoids (Schiborr, Kocher, Behnam, Jandasek, stede, & Frank, 2014).

Resveratrol is a natural phenol produced by many fruits and plants such as grapes, blueberries, raspberries and peanuts (Bhat, Kosmeder, & Pezzuto, 2001; Kundu & Surh, 2008; Smoliga, Baur, & Hausenblas, 2011a). Animal studies have shown protective effects of resveratrol against several types of cancer, such as breast, skin, gastric, colon, prostate and pancreatic cancers, by interfering with multiple stages of carcinogenesis (Shukla & Singh, 2011).

Recently, clinical trials have established the safety and potential anticancer effects of resveratrol as both a single agent and a constituent of foods (Smoliga, Baur, & Hausenblas, 2011b). A phase I pilot study in colorectal cancer patients determined the anticancer effects of freeze-dried grape powder containing a low dose of resveratrol in combination with other bioactive components. The results suggest that dietary intake of the dry grape powder inhibited the Wnt signaling pathway in the colon, which may have contributed to the observed inhibitory effects on colon carcinogenesis (Nguyen, Martinez, Stamos, Moyer, Planutis, Hope, et al., 2009).

The anticancer effect of resveratrol also has been heavily investigated in breast cancer patients. In a randomized, double-blind placebo-controlled trial, twice daily resveratrol supplement (5 or 50 mg) for 12 weeks decreased the methylation of four cancer-related genes in mammary tissue in women with high risk for developing breast cancer (Zhu, Qin, Zhang, Rottinghaus, Chen, Kliethermes, et al., 2012). Research has been conducted to enhance the bioavailability of resveratrol. Using a prostate cancer mice model, liposome encapsulation of resveratrol and curcumin successfully was shown to increase the oral bioavailablity of both nautraceuticals.

Moreover, encapsulation of both nutraeuticals in the same liposome system resulted in a synergistic reduction of prostate cancer incidence (Narayanan, Nargi, Randolph, & Narayanan, 2009). However, as mentioned earlier, the optimum dose for resveratrol has not been established yet, since some studies have shown that higher levels are less effective at inhibiting cancer than low doses (Cai, et al., 2015).

Sulforaphane: Sulforaphane is an isothiocyanate mainly found in cruciferous vegetables, especially abundant in broccoli and broccoli sprouts (Dinkova- Kostova & Kostov, 2012; Houghton, Fassett, & Coombes, 2013). A large number of cell culture and animal studies have shown that sulforaphane is a potent chemo-preventive agent against various type of cancer.

The molecular targets of sulforaphane vary upon cancer type and stage. The major anticancer mechanism by which sulforaphane protects normal cells from carcinogenesis is through Nrf2- mediated induction of phase II antioxidant and detoxifying enzymes. These enzyme systems enhance cell defense against oxidative damage, and facilitate the removal of carcinogens. Sulforaphane also exerts anticancer activities through various mechanisms of action that are involved in regulating cell proliferation, differentiation, apoptosis, and cell cycle progression (Clarke, Dashwood, & Ho, 2008; Juge, Mithen, & Traka, 2007). To date, only a few clinical trials have been conducted on sulforaphane in cancer patients or high-risk populations. In a phase II study, patients who had recurrent prostate cancer were given 200 moles/day of Sulforaphane-rich extracts for a maximum period of 20 weeks.

Although there was no large decline (by ≥50 %) in prostate-specific antigen (PSA), 7 out of 20 patients experienced moderate PSA declines (by <50 %). Moreover, the on- treatment PSA doubling time (PSADT) was significantly lengthened compared to the pre-treatment PSADT (Alumkal, Slottke, Schwartzman, Cherala, Munar, Graff, et al., 2015). In another clinical trial with Helicobacter pylori-infected humans, daily oral intake of sulforaphane-rich broccoli sprouts (70 gram/day; containing 420 moles of sulforaphane precursor) for 2 months reduced the incidence of infection in humans. This trial suggests that sulforaphane may enhance chemo-protection of the gastric mucosa against H. pylori-induced oxidative stress (Yanaka, Fahey, Fukumoto, Nakayama, Inoue, Zhang, et al., 2009). In a melanoma mouse model, sulforaphane-loaded albumin microspheres demonstrated a significantly stronger suppression on tumor growth as compared to non-encapsulated sulforaphane, and no adverse effects were observed (Do, Pai, Rizvi, & D'Souza, 2010). polyphenols: Tea polyphenols are the major bioactive components in tea leaves, and include epigallocatechin-3-gallate (EGCG) and epicatechin-3-gallate (ECG) (Cabrera, Artacho, & Gimenez, 2006; Crozier, Jaganath, & Clifford, 2009; Nijveldt, van Nood, van Hoorn, Boelens, van Norren, & van Leeuwen, 2001). EGCG and ECG can act as potent antioxidants to prevent DNA damage by scavenging reactive oxygen species (ROS), which in turn prevents carcinogenic mutagenesis in normal cells. In pre-clinical studies, tea polyphenols have been found to regulate multiple key cell signaling pathways, resulting in the suppression of angiogenesis, the modulation of the immune system, and the activation of phase II detoxifying enzymes (Yang, Li, Yang, Guan, Chen, & Ju, 2013). Several clinical trials have suggested that teas and tea polyphenols have the potential to prevent multiple types of cancer, including oral leukoplakia, liver, lung and bladder cancer (Kotecha, Takami, & Espinoza, 2016). In a recent phase II randomized presurgical placebo-controlled trial, bladder cancer patients taking 800 or 1200 mg of polyphenon E (a green tea polyphenol formulation mainly consisting of EGCG) showed a dose-dependent tissue accumulation of EGCG in benign and malignant bladder urothelium, which was associated with reduction of cell proliferation and increase of apoptosis (Gee, Saltzstein, Kim, Kolesar, Huang, Havighurst, et al., 2015). However, many clinical studies showed overall inconsistent results regarding the potential anticancer effects of tea polyphenols in different populations, suggesting different types of teas and variable tea preparations may significantly affect the chemo-preventive properties of teas (Sun, Yuan, Koh, & Yu, 2006). Moreover, the food matrix and gastrointestinal tract are known to impact the bioavailability of tea polyphenols (Manach, Williamson, Morand, Scalbert, & Remesy, 2005). flavonoids: Flavonoids are a large group of polyphenolic secondary metabolites of fruits, vegetables, and other plants with a broad-spectrum of bioactivities, including inhibitory effects on a wide range of human cancers (Tapas, Sakarkar, & Kakde, 2008; Weng & Yen, 2012). Epidemiological
evidence suggests a positive correlation between flavonoids-rich diets and low
risk of colon, breast and prostate cancer. Flavonoids can be categorized into
flavones, flavonols, flavanones, flavanols, anthocyanins and isoflavones based on
their structures. Flavonoids can target multiple signaling pathways during
carcinogenesis to inhibit cancer cell proliferation, suppress tumor angiogenesis,
and induce apoptosis in cancer cells (Batra & Sharma, 2013). Quercetin is one of
the most-studied flavones, and is abundant in onions and apples (Chirumbolo,
2013). Many in vitro and in vivo studies have shown the anticancer potential of
quercetin against a variety of human cancers, such as cervical, breast, colon,
prostate, liver and lung cancer (Yang, Song, Wang, Wang, Xu, & Xing, 2015). A
recent phase I study in patients with chronic Hepatitis C, a major causative factor
of liver cancer, reported that quercetin was safe when consumed at levels up to 5
grams per day, and that there was a modest reduction in viral load, suggesting a
potential for liver cancer prevention (Lu, Crespi, Liu, Vu, Ahmadieh, Wu, et al.,
2016). Quercetin normally exhibits a poor oral bioavailability due to its low
absorption. A microencapsulation approach has been shown to enhance the effects
of quercetin in reducing the oxidative damage and attenuating inflammation in a
mouse colitis model, presumably due to increased absorption (Guazelli, Fattori,
Colombo, Georgetti, Vicentini, Casagrande, et al., 2013). Tangeretin is another
well-studied flavonoid mainly found in citrus fruits (Weng & Yen, 2012). Poor
water-solubility limits the oral bioavailability and efficacy of tangeretin. In a
recent study, an emulsion-based delivery system was shown to significantly
enhance the inhibitory activity of tangeretin on colitis-associated colon
carcinogenesis in mice (Ting, Chiou, Pan, Ho, & Huang, 2015).
3. Factors limiting the bioavailability of anticancer nutraceuticals
The design of functional foods to improve the oral bioavailability of anticancer
nutraceuticals relies on an understanding of their fate within the human gastrointestinal
tract (GIT), as well as of the physicochemical and physiological events that typically
restrict their bioavailability. The factors limiting the oral bioavailability (BA) of
anticancer nutraceuticals can be divided into three main categories, as highlighted
schematically in Figure 2 (Arnott & Planey, 2012; Gleeson, Ryan, & Brayden, 2016;
McClements, 2013; McClements, Li, & Xiao, 2015b; Yao, McClements, & Xiao, 2015):
285 BA = B*  A*  T* (1) 286

The parameters B*, A* and T* represent the fractions of the nutraceuticals that are
bioaccessible, absorbed, and biologically active within the GIT, respectively. These
parameters depend on: (i) the molecular and physicochemical characteristics of the
anticancer nutraceuticals; (ii) the composition and structure of the surrounding food
matrices; and (iii) the complex events occurring within the human GIT. Improved
knowledge of the major factors limiting the BA of specific nutraceuticals will lead to a
more rational design of foods with enhanced beneficial health effects.

3.1. Bioaccessibility
The bioaccessibility (B*) of an anticancer nutraceutical represents the fraction of the total
amount of orally consumed nutraceuticals that is in a form that can be readily absorbed
by the GIT (Figure 2):
B*= mB/mI (2) 300
Here, mB is the mass of nutraceutical that is in a form that can be absorbed, and mI is the
total mass of the nutraceutical that is ingested. In the case of lipophilic nutraceuticals, B*
is usually taken to be the fraction of nutraceuticals solubilized in the mixed micelles in
the small intestinal fluids.
The bioaccessibility may be affected by three main factors in the GIT: liberation;
solubility; and interactions (McClements, 2015c; McClements, Li, & Xiao, 2015b).

 3.2. Absorption 

 After an anticancer nutraceutical is released from any structures containing it, and 

 then solubilized within the gastrointestinal fluids, it has to be transported through the GIT 

 contents and then be absorbed by the epithelial cells lining the GIT (Gleeson, Ryan, & 

 Brayden, 2016). The absorption of nutraceuticals is affected by numerous 

 physicochemical and physiological factors

 3.3. Transformation 

 The bioavailability and bioactivity of anticancer nutraceuticals is a result of their 

 precise chemical structures and mole
cular conformation. Changes in the structure or 

 conformation of these nutraceuticals due to chemical or biochemical reactions within the 

 GIT fluids may therefore alter their bio-efficacy. 

 nutraceuticals by using food components that modulate their chemical or 

 biochemical changes within the GIT.

4. Classification of the bioavailability of anticancer nutraceuticals
Foods can be specifically designed to improve the bioavailability profiles of
anticancer nutraceuticals by modulating their bioaccessibility, absorption, and
transformation in the GIT. Nutraceuticals can be classified according to the major factors
that limit their bioavailability, which means that generic food compositions and structures
can often be developed for a broad range of different nutraceuticals. A brief overview of
a recently developed classification schemes for nutraceuticals is given here (McClements,
Li, & Xiao, 2015b). This system is related to the Biopharmaceutical Classification
Scheme (BCS) commonly used to classify the factors limiting the bioavailability of
pharmaceuticals based on their solubility and permeability characteristics (Dahan, Miller,
& Amidon, 2009; Kawabata, Wada, Nakatani, Yamada, & Onoue, 2011; Lennernas &
Abrahamsson, 2005).

The Nutraceutical Bioavailability Classification Scheme (NuBACS) uses a three-
coordinate system (B*A*T*) to classify a nutraceutical according to the primary factors
that limit its bioavailability (McClements, Li, & Xiao, 2015a). Here, B* is the
Bioaccessibility, A* is the Absorption, and T* is the Transformation of the nutraceutical.
Each coordinate is designated “(+)” if it does not limit bioavailability, and “(-)” if it does.
Additional insights into the precise mechanisms associated with the poor bioavailability
are provided using subscripts, such as limited liberation from the food matrix (L), poor
solubility in the GIT fluids (S), and high susceptibility to metabolism (M) or chemical
degradation (C) (Table 1). As an example, curcumin, which is a highly hydrophobic
anticancer nutraceutical whose bioavailability is limited by poor solubility in GIT fluids,
metabolism, and chemical degradation is classified as B*(-)S A*(+) T*(-)M,C.
One of the major advantages of using this classification scheme is that a common
strategy may be developed to increase the bioavailability of a group of nutraceuticals
with similar properties. For example, food matrices (delivery or excipient systems)
containing digestible lipids could be used to increase the bioavailability of many
lipophilic nutraceuticals, i.e., B*(-)S.

5. Boosting the bioavailability of anticancer nutraceuticals using food
matrix design

Conventionally, the composition and structure of foods is usually only optimized to
enhance their quality attributes, such as appearance, texture, mouthfeel, and taste. More
recently, foods have been designed to improve their nutritional profiles by reducing the
levels of macronutrients believed to have adverse health impacts (such as saturated fats,
digestible carbohydrates, and salt) or to enrich them with food components that are
believed to bring beneficial health effects (such as vitamins, minerals, dietary fibers or
nutraceuticals). In this section, we focus on two different food-based approaches that
may be utilized to enhance the bioavailability, and therefore bioactivity, of anticancer
nutraceuticals (Figure 1).

5.1. Delivery Systems
Delivery systems are specifically designed to contain the anticancer nutraceuticals to
be delivered (Arora & Jaglan, 2016; Gleeson, Ryan, & Brayden, 2016). If the anticancer
nutraceutical is soluble or dispersible in one of the components (e.g., oil, water, or
powder) used to prepare a traditional food then it can often be simply dissolved or
dispersed in this component prior to preparing the final product. Nevertheless, the food
may still have to be carefully formulated to avoid potential adverse effects, such as
undesirable appearance, taste, or aroma caused by the presence of the nutraceutical.
Moreover, the food matrix may have to be carefully designed to avoid the chemical
degradation of the nutraceutical within the product during food manufacturing, storage,
and utilization. In many cases, an anticancer nutraceutical cannot be simply dissolved or
dispersed into a food matrix. Instead, specialized colloidal delivery systems have to be
fabricated that consist of the anticancer nutraceutical encapsulated within a small particle
(Figure 2).

5.1.1. Particle types
Numerous different types of colloidal particles have been developed that could
potentially be used as delivery systems to encapsulate, protect, and deliver anticancer
nutraceuticals (McClements, 2015c; Yao, McClements, & Xiao, 2015) (Figure 3). Some
of the most promising ones for commercial applications are highlighted here.
Microemulsions: Oil-in-water microemulsions contain colloidal particles dispersed in
water that consist of small clusters of surface active molecules (surfactants) that have
their hydrophobic tails located towards the interior and their hydrophilic heads located
towards the exterior (Figure 3A).

Conversely, water-in-oil microemulsions contain
colloidal particles dispersed in oil where the surfactants are organized so that their
hydrophilic heads form the interior and their hydrophobic tails face toward the exterior.
Both types of microemulsions are thermodynamically stable systems that usually contain
relatively small colloidal particles (5 nm < d < 100 nm). O/W microemulsions are most suitable for encapsulating hydrophobic nutraceuticals, whereas W/O microemulsions are most suitable for encapsulating hydrophilic ones (Flanagan & Singh, 2006; Spernath & Aserin, 2006). Because they are thermodynamically favorable, these systems should form spontaneously when the required components are mixed together. Nevertheless, some energy often has to be applied to ensure thorough mixing of the ingredients and to overcome any kinetic energy barriers. Emulsions: Oil-in-water emulsions (d > 200 nm) or nanoemulsions (d < 200 nm) consist of emulsifier-coated oil droplets dispersed within water (Figure 3B), and are therefore most suitable for encapsulating non-polar nutraceuticals inside the hydrophobic interior of the oil droplets (McClements, 2012; McClements, Decker, & Weiss, 2007). Conversely, water-in-oil emulsions or nanoemulsions consist of emulsifier-coated water droplets dispersed in oil, and are therefore more suitable for encapsulating polar nutraceuticals within the hydrophilic interior of the water droplets. Typically, a relatively hydrophilic emulsifier is required to form an O/W emulsion, whereas a relatively hydrophobic one is required to form a W/O emulsion. In practice, the formation, stability, and functional performance of emulsion-based delivery systems is highly dependent on the nature of the emulsifiers used. Emulsions may be made using a variety of high-energy (high pressure homogenization, microfluidization, and sonication) or low- energy (spontaneous emulsification or phase inversion temperature) methods (McClements & Rao, 2011). The nature of the homogenization method utilized to prepare an emulsion depends on many factors, including the type of ingredients used, the required particle size distribution, and the amount of material that should be produced. Solid lipid nanoparticles: Solid lipid nanoparticles (SLN) are similar to oil-in-water emulsions or nanoemulsions, but the oil phase is crystallized (Figure 3C) (Guri, Guelseren, & Corredig, 2013; Mehnert & Mader, 2012). Crystallization of the oil phase may improve the physical stability of the particles, as well as improving the retention and stability of the encapsulated nutraceuticals. SLNs are usually fabricated by forming an O/W emulsion or nanoemulsion using a high melting point lipid, and then cooling the system to promote crystallization of the lipids. Hydrogel beads: Hydrogel beads (microgels) consist of spherical particles, typically in the range of about 1 to 1000 m, which consist of cross-linked biopolymers that trap water (Figure 3D) (Chen, Remondetto, & Subirade, 2006; Joye & McClements, 2014; Shewan & Stokes, 2013; Zhang, Zhang, Chen, Tong, & McClements, 2015). Filled hydrogel beads also contain other types of colloidal particles dispersed within the beads, such as lipid droplets or liposomes. Hydrogel beads can be formed using a wide range of different methods depending on the nature of the biopolymers and cross-linking agents used. Typically, food-grade proteins (such as whey protein, caseinate, or gelatin) or polysaccharides (such as agar, alginate, carrageenan, pectin, or starch) are used as the biopolymers. The gelation mechanism used is largely determined by the nature of the biopolymer utilized and may involve temperature changes (heating or cooling), desolvation (dehydration or solvent removal), or addition of cross-linking agents (such as ions, polyelectrolytes, or enzymes). Certain types of hydrophilic nutraceutical can be directly incorporated into the hydrogel beads interior by mixing them with the biopolymer solution prior to fabrication. Conversely, hydrophobic nutraceuticals may have to be encapsulated into lipid droplets or liposomes prior to being incorporated into the hydrogel beads. The composition, dimensions, pore size and interactions of hydrogel beads can be controlled to create delivery systems with different properties. Liposomes: This type of colloidal particle typically consists of single or multiple lamellar spherical structures fabricated from phospholipids (Desai & Park, 2005; Fathi, Mozafari, & Mohebbi, 2012; Mozafari, Johnson, Hatziantoniou, & Demetzos, 2008). The phospholipids are assembled into bilayers with the non-polar tails of each layer being in contact (Figure 3E). Liposomes have a hydrophilic interior that can be used to encapsulate polar nutraceuticals, and a hydrophobic region in the bilayer that can be used to encapsulate non-polar ones. Liposomes can be formed using a number of different methods, including microfluidization and rehydration of surface deposited layers (Mozafari, Johnson, Hatziantoniou, & Demetzos, 2008). 5.1.2. Particle properties
The colloidal particles used as delivery systems may vary appreciably in their
properties, which determines their efficacy at encapsulating, stabilizing, and delivering
anticancer nutraceuticals (Figure 4). Some of the most important particle properties that
can be controlled to obtain specific functional attributes in delivery systems are
summarized below (McClements, 2015c):

Composition: Food-grade colloidal particles suitable for encapsulating anticancer
nutraceuticals can be fabricated from lipids, proteins, carbohydrates, minerals, surfactants
and/or water. The nature of the components used to fabricate the particles plays a major
role in determining their ability to encapsulate, stabilize, and release the nutraceuticals.
The amount of an anticancer nutraceutical that can be encapsulated within a colloidal
particle largely depends on its solubility within the particle interior. Hydrophobic
nutraceuticals can be solubilized within particle domains comprised of non-polar
components such as lipids, surfactant tails, phospholipid tails, or hydrophobic
biopolymers (e.g., zein). Conversely, hydrophilic nutraceuticals can be solubilized with
particle domains comprised of polar components such as water, hydrophilic proteins, and
polysaccharides. The chemical stability of certain nutraceuticals can be enhanced by
encapsulating them in colloidal particles that restrict the molecular motion of reactants
(such as solid lipid nanoparticles) or by incorporating components that protect the
nutraceutical from degradation, such as antioxidants. The gastrointestinal fate of
anticancer nutraceuticals can also be controlled by manipulating the composition of the
colloidal particles, i.e., the rate and extent of their digestion in different regions of the
GIT. For example, colloidal particles comprised of starch may be initially digested by
amylases in the mouth, those made by proteins or lipids may be digested by gastric or
pancreatic proteases or lipases in the stomach and small intestine, and those made by
dietary fibers may not be digested until they reach the colon due to the action of colonic

Dimensions: Colloidal particles can be fabricated with a wide range of dimensions,
ranging from around 10 nanometers (surfactant micelles) to a few millimeters (hydrogel
beads). The dimensions of colloidal particles influence many of the properties of
delivery systems, including their chemical stability, release rate, optical properties,
physical stability (to gravitational separation and aggregation), rheology, release rate, and
gastrointestinal fate. Anticancer nutraceuticals that are prone to chemical degradation
when they are dispersed in water (such as curcumin under basic conditions) tend to
breakdown more rapidly when they are in small rather than large lipid particles (Zou,
Zheng, Liu, Liu, Xiao, & McClements, 2015). This effect can be attributed to the fact
that the fraction of nutraceutical molecules in close proximity to the oil-water interface
increases with decreasing particle size. Conversely, the rate of lipid particle digestion
and nutraceutical release tends to increase as the particle size decreases because then
there is more oil-water interface available for digestive enzymes to adsorb to (Salvia-
Trujillo, Qian, Martin-Belloso, & McClements, 2013a). The particle size also affects the
incorporation of colloidal particles into food products. Larger particles tend to cream or
sediment more rapidly than smaller ones because the magnitude of the gravitational
forces is proportional to the diameter squared. Gravitational separation of particles may
be an important consideration in functional foods and beverages that have relatively low
viscosities. The optical properties of foods depend on the particle diameter (d) relative to
the wavelength of light (): colloidal dispersions go from being transparent when d <<  to turbid/opaque when d  , and to visibly distinguishable as separate particles when d >> . The dimensions of the particles will also affect their perception within the
mouth: a colloidal dispersion with particles less than about 50 m in diameter will feel
smooth in the mouth, whereas one with larger particles will feel gritty or lumpy.
Obviously, these factors have to be taken into account when designing delivery systems
intended for application in commercial food products that must be perceived favorably by

Interfacial properties: The colloidal particles used in the food industry as delivery
systems may have different interfacial properties, e.g., composition, thickness, charge,
and hydrophobicity (McClements, 2015c). The interfacial properties also have a major
impact on the chemical stability, physical stability, release rate, optical properties,
rheology, and gastrointestinal fate. For example, the magnitude and sign of the electrical
charge on the surfaces of the colloidal particles determines their ability to interact with
other electrical substances, such as other charged particles, surfaces, or polymers. In
addition, the charge may impact the chemical stability of encapsulated nutraceuticals by
influencing the attraction or repulsion of pro-oxidants to the particle surfaces (such as
cationic transition metals). The surface chemistry and digestibility of the particle surface
influences the ability of digestive enzymes (such as lipases, proteins, or amylases) to
adsorb to them and hydrolyze the particles. Colloidal particles can often be coated with a
substance that is resistant to digestion in one region of the GIT, but that is digested in
another region. This phenomenon can be used to control the release of encapsulated
nutraceuticals in different parts of the GIT, e.g., mouth, stomach, small intestine or colon.

Physical state: Colloidal particles may be liquid, semi-solid, or solid depending on the
materials used to assemble them and the environmental conditions (such as temperature,
pH, and ionic composition). The oils and water phase used to prepare oil-in-water or
water-in-oil emulsions are typically liquid. However, high melting oils can be used to
create solid lipid nanoparticles that have a crystalline interior. The biopolymers (proteins
and polysaccharides) used to prepare hydrogel beads usually form semi-solid gel-like
particles. The physical state of a particle may change appreciably when environmental
conditions are changed within a food or as it passes through the GIT.

5.1.3. Mechanisms of Action
Appropriately designed delivery systems may enhance the oral bioavailability and
bioactivity of anticancer nutraceuticals due to numerous mechanisms of action:
Enhance Dispersibility: Most non-polar nutraceuticals are so hydrophobic that they
cannot simply be dispersed into aqueous-based foods (such as beverages, sauces,
dressings, yogurts, and desserts) because of their extremely poor water-solubility (Donsi,
Sessa, Mediouni, Mgaidi, & Ferrari, 2011). However, if these nutraceuticals can be
encapsulated within colloidal particles that have a hydrophobic core and a hydrophilic
shell (as in oil-in-water emulsions, nanoemulsions, or microemulsions), then they can
easily be dispersed into these aqueous-based foods (Flanagan & Singh, 2006;

McClements, 2012). Conversely, many polar nutraceuticals are so hydrophilic that they
cannot easily be dispersed into oil-based foods (such as butter, margarine, and spreads).
In this case colloidal particles that consist of a hydrophilic core and hydrophobic shell (as
in water-in-oil emulsions, nanoemulsions, or microemulsions) can be used to encapsulate

Enhance Chemical or Biochemical Stability: Many nutraceuticals are highly
susceptible to chemical or biochemical degradation within food products or inside the
GIT (Braithwaite, Tyagi, Tomar, Kumar, Choonara, & Pillay, 2014; Lu, Kelly, & Miao,
2016; Yallapu, Nagesh, Jaggi, & Chauhan, 2015). Delivery systems can be specifically
designed to protect the encapsulated nutraceuticals from degradation by providing a
physical barrier or by containing specific components (such as antioxidants, buffers or
UV-visible absorbers).

As an example, some nutraceuticals (such as curcumin) are more
prone to chemical degradation when they are dissolved in water than when they are
dissolved in oil. Consequently, their chemical stability can be improved by encapsulating
them in delivery systems that have lipophilic domains, such as vesicles, nanoemulsions,
protein nanoparticles, or solid lipid nanoparticles (Zou, Liu, Liu, Xiao, & McClements,
2015; Zou, Zheng, Zhang, Zhang, Liu, Liu, et al., 2016a, 2016b).

Including antioxidants or chelating agents within a delivery system can inhibit the oxidation of unsaturated
lipids (Davidov-Pardo & McClements, 2015; Qian, Decker, Xiao, & McClements, 2012;
Tamjidi, Shahedi, Varshosaz, & Nasirpour, 2014). Delivery systems can be designed so
that chemically unstable nutraceuticals are trapped within colloidal particles with
coatings that physically separate the nutraceuticals from any other components that may
accelerate their degradation.

Enhance Bioaccessibility: Many nutraceuticals have a low oral bioaccessibility
because of their low solubility in aqueous GIT fluids (Arora & Jaglan, 2016;
McClements, Li, & Xiao, 2015b). Delivery systems can be designed to contain digestible
lipids (such as triglycerides) that form free fatty acids and monoacylglycerols that
enhance the solubilization capacity of the mixed micelles for hydrophobic nutraceuticals
(Ozturk, Argin, Ozilgen, & McClements, 2015; Salvia-Trujillo, Qian, Martin-Belloso, &
McClements, 2013b; Salvia-Trujillo, Sun, Urn, Park, & McClements, 2015; Yao,
McClements, & Xiao, 2015).

The type of lipid used is particularly important since it
determines the rates of lipid digestion and mixed micelle formation, as well as the
solubilization capacity of the mixed micelles. The hydrophobic domains in the mixed
micelles (micelles and vesicles) must be large enough to accommodate the nutraceuticals,
so that they can transport them through the mucus layer to the epithelial cells. There is
therefore considerable interest in designing the lipid phase of foods so that they form
mixed micelles in the small intestine that have an appropriate solubilization capacity for
the nutraceuticals being delivered (Yao, McClements, & Xiao, 2015; Yao, Xiao, &
McClements, 2014).

Enhance Absorption: A delivery system may contain specific components that can
enhance the uptake of the anticancer nutraceuticals, such as components that disrupt the
mucus layer, inhibit efflux inhibitors, stimulate active transporters, or increase the
dimensions of tight junctions (Arora & Jaglan, 2016; Gleeson, Ryan, & Brayden, 2016;
McClements, Li, & Xiao, 2015b; Yao, McClements, & Xiao, 2015). Gleeson and co-
workers have recently highlighted the major strategies for increasing the absorption of
nutraceuticals (Gleeson, Ryan, & Brayden, 2016).

Certain components found in foods,
such as bromelain (an enzyme from pineapple) and papain (an enzyme from papaya),
have the potential to disrupt the mucus layer that normally coats the epithelial cells,
which may lead to enhanced absorption of nanoparticles coated with these substances (de
Sousa, Cattoz, Wilcox, Griffiths, Dalgliesh, Rogers, et al., 2015). Other components
present in foods are able to increase the permeability of epithelial cells, including
medium chain fatty acids like caprylic acid (Gleeson, Ryan, & Brayden, 2016). Other
food components may also have the ability to act as permeation enhancers, such as
cationic biopolymers (chitosan) chelating agents (EDTA), surfactants (polyol or sugar
esters of fatty acids) and phytochemicals (such as piperine) (McClements, Li, & Xiao,

These intestinal permeation enhancers are believed to increase absorption by
increasing the fluidity of the phospholipid membranes, increasing the dimensions of the
tight junctions, stimulating active transporters, and/or blocking efflux transporters.
Nevertheless, it is important to realize that there may be adverse health effects associated
with altering the normal absorption mechanisms of nutraceuticals, which should be
considered when designing functional foods based on this principle (McCartney,
Gleeson, & Brayden, 2016).

5.2. Excipient Systems
Unlike delivery systems, excipient systems do not necessarily contain any
nutraceutical components (McClements & Xiao, 2014; McClements, Zou, Zhang, Salvia-
Trujillo, Kumosani, & Xiao, 2015). Instead they are designed to boost the bioavailability
and bioactivity of the nutraceuticals present in other foods that they are co-ingested with,
such as fruits and vegetables. Alternatively, they can be designed to boost the
bioavailability of nutraceuticals in dietary supplements or drugs (Salvia-Trujillo &
McClements, 2016b). Typically, the composition and structure of the excipient food
matrix is carefully controlled so as to improve the bioaccessibility, absorption, or
transformation profile of a nutraceutical-containing substance that is co-ingested with it.
Many of the approaches developed to improve the bioavailability of nutraceuticals using
delivery systems can also be used with excipient systems.

A considerable research effort has recently been carried out with the objective of
developing excipient systems based on emulsions that are suitable for utilization in the
food, dietary supplement, or pharmaceutical industries (McClements & Xiao, 2014;
McClements, Zou, Zhang, Salvia-Trujillo, Kumosani, & Xiao, 2015). These excipient
emulsions consist of small lipid droplets dispersed in an aqueous medium.
composition, size, and interfacial properties of the lipid droplets are optimized based on a
number of factors: (i) the ability to extract and solubilize lipophilic nutraceuticals from
food matrices (such as fruits and vegetables); (ii) the ability to be rapidly digested within
the GIT; (iii) the ability to rapidly form mixed micelles that solubilize the lipophilic
nutraceuticals released from the food matrices; (iv) the ability to protect the
nutraceuticals from chemical or biochemical degradation within the GIT. Additional
components may also be added to the aqueous phase of the excipient emulsions to
improve their efficacy, including antioxidants, chelating agents, buffering agents,
biopolymers, or permeation enhancers.

Recent research has shown that excipient emulsions can be used to increase the
bioavailability of carotenoids in a range of produce, including mangoes (Liu, Bi, Xiao, &
McClements, 2016), carrots (Zhang, Zhang, Zou, Xiao, Zhang, Decker, et al., 2015,
2016a; Zhang, et al., 2016b), yellow peppers (Liu, Bi, Xiao, & McClements, 2015), and
tomatoes (Salvia-Trujillo & McClements, 2016a). The composition and structure of the
excipient emulsions has to be carefully optimized to effectively enhance carotenoid
bioavailability (Figure 5). Studies have shown that excipient emulsions containing long
chain triglycerides (LCT) are more effective than those containing medium chain
triglycerides (MCT) at improving the bioaccessibility of carotenoids (Zhang, et al.,
2015). This effect was attributed to the fact that digestion of LCT by lipase leads to the
generation of long chain fatty acids that enhance the solubilization capacity of the mixed
micelle phase by increasing the dimensions of the hydrophobic domains (Cho, Salvia-
Trujillo, Kim, Park, Xiao, & McClements, 2014; Salvia-Trujillo, Qian, Martin-Belloso, &
McClements, 2013b).

Conversely, the hydrophobic domains formed by the medium
chain fatty acids resulting from MCT digestion are too small to accommodate large
carotenoids. Studies with carrots showed that the bioaccessibility of the carotenoids
increased with increasing fat content in the excipient emulsions co-ingested with them
(Zhang, et al., 2016b). Moreover, decreasing the size of the fat droplets in the excipient
emulsions was found to increase carotenoid bioaccessibility, which was attributed to their
higher surface area and rate of digestion, leading to quicker mixed micelle formation and
nutraceutical solubilization (Zhang, et al., 2016a).

Excipient emulsions have also been used to improve the bioavailability of curcumin
from turmeric powder (Zou, Liu, Liu, Liang, Li, Liu, et al., 2014; Zou, Zheng, Liu, Liu,
Xiao, & McClements, 2015). A higher amount of curcumin reached the simulated small
intestine when the turmeric powder was co-ingested with an excipient emulsion, which
was attributed to a number of effects. First, some of the curcumin was transported into
the hydrophobic interior of the small lipid droplets. Second, the chemical stability of the
curcumin was higher within the hydrophobic interior of the lipid droplets than in the
surrounding aqueous phase. Third, the mixed micelles formed after digestion of the lipid
droplets were able to solubilize the non-polar curcumin molecules in the aqueous GIT

The concept of excipient foods is fairly new, and a considerable amount of research
is still required in this area. Nevertheless, it is a promising approach for improving the
bioavailability characteristics of many anticancer nutraceuticals in functional foods,
dietary supplements, and pharmacological drugs.

6. Conclusions
787 There is growing interest in using specific nutraceuticals commonly found in foods
as chemo-preventative agents. Cell culture and animal studies suggest that ingestion of
these nutraceuticals may inhibit certain types of cancers. Nevertheless, many anticancer
nutraceuticals cannot simply be incorporated into foods because of their poor solubility,
stability, and bioavailability characteristics. There is a considerable potential to create
functional food products designed to overcome these challenges, and therefore increase
the potential efficacy of anticancer nutraceuticals. These foods may be designed to
contain the nutraceuticals themselves (delivery systems) or they may be designed to boost
the bioavailability of nutraceuticals in other foods (excipient systems). Despite their
potential, a considerable amount of research is still needed to demonstrate the efficacy of
nutraceuticals, and the ability of delivery or excipient systems to enhance their bio-

Unlike pharmaceuticals, which are typically taken in a well-defined dose at a
particular time, nutraceuticals are obtained from numerous sources at levels that vary
from person to person and from time to time as part of a complex diet. In addition, the
beneficial effects of nutraceuticals may arise from taking relatively low levels over an
extended period.

Consequently, it is often difficult to carry out clinical studies (double
blind randomized feeding trials) using humans to prove their efficacy. Nevertheless, it
may be possible to gain some insights into their efficacy using long-term animal feedings
studies where the diet is carefully controlled. Having said this, there is also considerable
skepticism about the bioactivity demonstrated by certain nutraceuticals established using
in vitro and in vivo assays.

For instance, recent articles have criticized the evidence
suggesting that curcumin has a high bioactivity due to its ability to interfere with many
types of assays used to measure biological activity (Baker, 2017; Nelson, Dahlin, Bisson,
Graham, Pauli, & Walters, 2017). This may at least partially account for the lack of
efficacy that curcumin has demonstrated in many clinical trials.

Mitochondrial transplantation: From animal models to clinical use in humans

1. Introduction
The importance of the mitochondrion in the maintenance and preservation of cellular homeostasis and function is well established and there is a sufficient body of evidence to show that mitochondrial injury or loss of function is deleterious (Durhuus et al., 2015). The mechanisms leading to mitochondrial dysfunction are varied and include genetic changes occurring at the nuclear or the mitochondrial genome, environmental insult or alterations in homeosis. In all cases, the end result of mitochondrial dysfunction is cellular dysfunction that can limit or severely modulate organ function and ultimately increase morbidity and mortality.

In our research, we have focused on the myocardium, a highly aerobic organ in which mitochondria comprise 30% of cellular volume (Faulk et al., 1995; Faulk et al., 1995a; Toyoda et al., 2000; Toyoda et al., 2001; McCully et al., 2003). The mitochondria supply the energy requirements of the myocardium. This energy is derived through oxidative phosphorylation in the myocardium and is dependent upon the coronary circulation. Under equilibrium conditions the mitochondria within the heart extract greater than 79% of arterial oxygen from the coronary arteries (Fillmore et al., 2013). As heart rate increases or if myocardial workload is increased the oxygen demand is increased and is dependent upon increased coronary flow. Thus, any interruption or impedance in coronary blood flow will significantly limit oxygen delivery to the heart and significantly decrease function and hemostasis (Akhmedov et al., 2015; Doenst et al., 2013; Kolwicz et al., 2013). It is generally accepted that the cessation of coronary blood flow, and thus oxygen delivery, is the initial step in the process leading to myocardial ischemic injury. The sequence of events and the mechanisms associated with this injury are many and are reviewed elsewhere (Lesnefsky et al., 2003; Kalogeris et al., 2012; Ong et al., 2015 Kalogeris et al., 2016 ; Lesnefsky et al., 2017). The end result of ischemia is loss of high energy synthesis and the depletion of high energy stores such that the heart is unable to support hemostasis and maintain function (Rosca et al., 2013 ) .

This reduction of high energy synthesis and stores is rapid. 31Pnuclear magnetic resonance studies have shown that following regional or global ischemia wherein the blood flow to the heart is temporarily ceased, high energy phosphate synthesis and stores are rapidly decreased within 6 minutes and that this decrease continues for at least 60 180 minutes after the restoration of blood flow and is associated with significantly decreased myocardial cellular viability and myocardial function (Tsukube et al., 1997).

Mitochondrial modulations induced by ischemia in the myocardium are many. We and others have demonstrated that following ischemia there are changes in mitochondrial morphology and structure (Rousou et al., 2004; Lesnefsky et al., 2004; McCully et al., 2007). Transmission electron microscopy and lightscattering spectrophotometry have shown that ischemia significantly increases mitochondrial matrix and cristae area and mitochondrial matrix volume (McCully et al., 2007). In addition, there is a decrease in mitochondrial complex activity, cytochrome oxidase I Vmax and a decrease in oxygen consumption and an increase in mitochondrial calcium accumulation (Faulk et al.,1995a).

These changes occur in concert with changes in mitochondrial transcriptomics, with downregulation of annotation clusters for mitochondrion function and energy production and the downregulation of cofactor catabolism, generation of precursor metabolites of energy, cellular carbohydrate metabolism, regulation of biosynthesis, regulation of transcription, and mitochondrial structure and function (enrichment score >2.0, P<.05) ( Black et al., 2012; Masuzawa et al., 2013). In addition, there are changes evident in overall protein synthesis. Proteomic analysis has shown that ischemia significantly alters mitochondrial proteins involved in fatty acid and glucose metabolism, ATP biosynthesis, and oxidoreductase activity (fold change >1.4, P < .05) ( Black et al., 2012; Masuzawa et al., 2013). All these changes are associated with decreased myocardial cellular viability and decreased myocardial function and suggest that the mitochondrion plays a key role in myocardial viability and function following ischemia and repercussion. In total, these data have provided a basis for continued mitochondrial associated investigations into the rescue and preservation of myocardial tissue and myocardial function (Suleiman et al., 2001). The methodologies for these investigations have been many and varied. In general, the approach to cardioprotection has been either associative or indirect with emphasis on a single mechanistic route or complex or the use of an additive or inhibitor, used either as a single therapy or in combination with others. These include, but are not limited to, the use of pharmaceuticals either before or after ischemia, such as antioxidants, the use of calcium channel antagonists, adenosine, adenosine deaminase inhibitors, adenosine transport inhibitors, or a combination of both , adenosine receptor agonists, mitochondrial ATPsensitive potassium channel openers, phosphodiesterase inhibitors, 5' AMPactivated protein kinase activators, metabolic modulators, antiinflammatory agents and procedural approaches including, preischemia, postischemia and remote ischemic preconditioning (Hsiao et al., 2015 ; Madonna et al., 2015; Hausenloy et al., 2016; Laskowski et al., 2016; OrenesPiñero et al., 2015). In some methodologies, therapeutic intervention is required days or months prior to the ischemic event. Unfortunately, clinical trials using these approaches, either alone or in combination, have for the most part been unsuccessful. 2. Mitochondrial transplantation
We rationalized that the therapeutic approach to cardioprotection should be comprehensive and rather than involving a single or multiple mechanistic pathways, intervention should be specific. To this end, we speculated that the replacement or augmentation of mitochondria damaged during ischemia and reperfusion should be the target for therapeutic intervention. We hypothesized that viable mitochondria isolated form the patient’s own body, from a nonischemic area, and then delivered by direct injection into the ischemic organ would replace or augment damaged mitochondria; thus, allowing for the rescue of myocardial cells and restoration of myocardial function. We have termed this therapeutic intervention; mitochondrial transplantation (McCully et al., 2009; Masuzawa et al., 2013; Cowan et al., 2016; Kaza et al., 2016).
2.1 Mitochondrial isolation

Mitochondrial transplantation is based on the delivery of isolated, viable mitochondria to the target organ. The isolation of mitochondria can be performed using a variety of techniques and methodologies. In our initial studies, we used a standard procedure requiring consecutive low and high speed centrifugation to isolate purified mitochondria. This procedure required approximately 90120 minutes to complete. The repetitive centrifugation steps increases the time for mitochondrial isolation and ultimately reduce mitochondrial viability (Graham 2001; Frezza et al., 2007; Pallotti and Lenaz 2007; Wieckowski et al., 2009; FernándezVizarra et al., 2010; Schmitt S, et al., 2013).

In cardiac surgery, and in many other surgical interventions, the interventional time is 4560 minutes and therefore mitochondrial isolation times of 90120 minutes are inappropriate for clinical usage. Mitochondrial isolation time must be short and efficient so that the therapeutic use of mitochondrial transplantation would not extend the surgical time and possibly add to patient morbidity or mortality. To meet the demands and requirements for clinical application, we have developed a rapid methodology for the isolation of autologous mitochondria (Preble et al., 2014; Preble et al., 2014a).

Firstly, two small pieces of autologous tissue are obtained from the patient’s own body during the surgical procedure. The tissue is dissected out using a #6 biopsy punch. The amount of tissue is less than 0.1 gram. The source of tissue is dependent upon the surgical entry point and access. The only requirement being that the tissue source must be free from ischemia and be viable.

In our studies, we have used viable, nonischemic skeletal muscle tissue as a source for isolated mitochondria. The muscle tissue was obtained from the pectoralis major or the rectus abdominis based on standardized minithoracotomy or sternotomy, respectively. Other tissue sources can also be used and we have used liver tissue with excellent results. The tissue, once obtained is immediately used for mitochondria isolation.

The methodology for the isolation of mitochondria for use in mitochondrial transplantation is simple and rapid and can be performed in under 30 minutes. The freshly isolated tissue is homogenized using a commercial automated homogenizer. For our uses we have used the Miletenyi gentleMACS Dissociator (Miltenyi Biotec Inc., San Diego, CA).

This homogenizer was chosen over the more standardized PotterElvehjem homogenizer (glass with Teflon pestle) or bead based homogenizers as it provides a sterilized, disposable unit for homogenization with fixed plastic wings for homogenization and an automated programable homogenization protocol allowing for uniform and consistent homogenization of tissue that is not easily achieved with manual homogenization methods.

The other systems require cleaning and sterilization for each usage and allow for variance in homogenization based on the operator. Bead homogenization systems were not considered due to the possibility of contamination of the mitochondria by bead fragments. Once the tissue is obtained homogenization can be performed in 90120 seconds.

The homogenized tissue is then subjected to brief digestion (10 minutes on ice) with subtilisin A (Protease from Bacillus licheniformis, Type VIII, Sigma, Aldrich, St. Louis, MO) and the digested homogenate filtered through a series of disposable sterile mesh filters. The filtration can all be performed in 23 minutes.

The mitochondria can then be used for direct application or can be concentrated by centrifugation (9000 rpm at 4 oC for 10 minutes). The mitochondrial yield using this methodology is approximately 1 x 109 to 1 x 1010 mitochondria, using the two biopsy tissue samples (< 0.1 g) and provides sufficient mitochondria for quality assurance and quality control assessment ( Preble et al., 2014a). The isolated mitochondria are of the correct size and shape as assessed by electron microscopy and have normal cristae and membranes and show no damage or injury. The isolated mitochondria are pure and have no detectable cytosolic, nuclear or microsomal components. Functional analysis of isolated mitochondria shows that the isolated mitochondria maintain membrane potential and viability and that oxygen consumption and respiratory control index for malateinduced complex I and succinateinduced complex II are equal to that of mitochondria isolated by other methodologies (Rousou et al., 2004; McCully et al., 2009). Once isolated, the mitochondria are immediately used for mitochondrial transplantation. We have found that the isolated mitochondria can be stored on ice for approximately 1 hour but storage beyond this time point greatly reduces efficacy. This is in agreement with previous reports that have shown that mitochondrial bioenergetic function is decreased to <10–15% of normal after the mitochondria were frozen, even when preservatives are used (Wechsler, 1961; Olson et al., 1967). In all our studies, we have used total mitochondria for mitochondrial transplantation. The bioenergetic function of this population includes that of subsarcolemmal and intrafibrillar mitochondria. Previous studies have shown that these mitochondrial subpopulations have differing oxygen consumption and metabolism (Riva et al., 2005; Chen et al., 2008; Kurian et al., 2012). In our studies, we have examined the cardioprotective efficacy of intrafibrillar, subsarcolemmal and total mitochondria. We have found that total mitochondria provide for cardioprotection and that no added cardioprotection is provided using either subsarcolemmal or interfibrillar mitochondrial subpopulations (McCully et al., 2009). It is important that the isolated mitochondria be intact and viable. The use of dead, nonviable mitochondria, mitochondrial proteins or complexes, mitochondrial DNA/RNA or high energy phosphates alone or in combination do not provide cardioprotection (McCully et al., 2009). It has been previously demonstrated that exogenous ATP supplementation and/or ATP synthesis promoters do not restore highenergy phosphate stores and have no beneficial effects on postischemic functional recovery in the heart (McCully et al., 2009). Following determination of mitochondrial number and viability the isolated mitochondria are suspended in 1 ml of respiration buffer containing 250 mmol/l sucrose, 2 mmol/l KH2PO4, 10 mmol/l MgCl2, 20 mmol/l K+HEPES buffer, pH 7.2, 0.5 mmol/l K+EGTA, pH 8.0, 5 mmol/l glutamate, 5 mmol/l malate, 8 mmol/l succinate and 1 mmol/l ADP. Quality control and assurance parameters have been previously described by us (Preble et al., 2014). The isolated mitochondria are then directly injected into the ischemic zone of the heart just prior to reperfusion using a 1 mL tuberculin syringe with a 28or 32gauge needle. The injection volumes are 0.1 mL and contain approximately 1 x 107 mitochondria at each injection site. This volume is optimal and allows for mitochondrial uptake within the myocardium with no backflow leakage of the injected mitochondria. In our studies, we have found that 810 individual injections are sufficient to cover the areaat–risk, although the absolute number of injection sites can be increased. 2.2 Mitochondrial delivery
The direct injection of mitochondria is simple and allows for focal concentration of the injected mitochondria. In our studies the number of mitochondria used for direct injection is 13 x 107 mitochondria. The mitochondria are suspended in 1 mL respiration buffer and injected at 810 sites within the area at risk using a 1 mL tuberculin syringe with a 2832 gage needle (McCully et al., 2009; Masuzawa et al., 2013; Cowan et al., 2016; Kaza et al., 2016). Mitochondrial concentrations > 2 x 108 are not fully suspended in 1 mL of respiration buffer and are therefore not advised for use for mitochondrial transplantation by direct injection.

Fluorescence microscopy has demonstrated that transplanted mitochondria delivered by direct injection are present and viable for at least 28 days following injection into the myocardium, the end point of our animal experiments (Masuzawa et al., 2013, Kaza et al., 2016).

The transplanted mitochondria are widely distributed from the epicardium to the subendocardium >2–3 mm from the injection site (McCully et al., 2009; Masuzawa et al., 2013). The majority of injected mitochondria are found initially within the interstitial spaces between cardiomyocytes. Within 1 hour postdelivery, the transplanted mitochondria are detectable within cardiomyocytes residing near the sarcolemma between Zlines of the sarcomeres and in clusters around endogenous damaged mitochondria as well as near the nucleus. Enumeration of injected mitochondria has shown that 43.52% ± 4.46 (mean ± SEM) of the injected mitochondria are attached to or found within cardiomyocytes (Cowan et al., 2016).

We have speculated that the mechanisms through which the transplanted mitochondria distribute within the myocardium following direct injection into the myocardium may be associated with alterations in myocardial structure that occur after myocardial ischemia. Our studies have demonstrated that following ischemia there is a significant increase in myocardial interfibrillar space that provides for both longitudinal and transverse myocardial interfibrillar separations (Tansey et al., 2006).

These structural changes are not associated with alterations in conduction velocity anisotropy or in tissue edema but occur coincident with significant decreases in postischemic functional recovery and increased myocardial apoptosis and necrosis. We have hypothesized that these interfibrillar separations allow for the distribution of injected mitochondria within the myocardium.

While the delivery of mitochondria by direct injection is practical for many applications, it does not allow for global distribution of the transplanted mitochondria.

Multiple injections are needed for global distribution within the heart and require organ manipulation to access posterior and lateral aspects. To allow for global distribution of mitochondria we have recently demonstrated that mitochondria can also be delivered to the target organ by vascular infusion (Cowan et al., 2016).

In preliminary investigations, we have found that for optimal distribution 1 x 109 mitochondria in 5 mL respiration buffer is efficacious and multiple injections can be performed. Using this protocol we have demonstrated that vascular delivery of mitochondria through the coronary arteries results in the rapid and widespread distribution of exogenous mitochondria throughout the heart, within 10 minutes and provides for cardioprotection.

To demonstrate uptake and distribution of the transplanted mitochondria we have labeled isolated mitochondria with 18Frhodamine 6G and magnetic iron oxide nanoparticles. The use of these labels allowed for image analysis using positron emission tomography, computed tomography, and magnetic resonance imaging. Our results show that the transplantation of mitochondria by vascular delivery is rapid and effective Decaycorrected measurements of 18Frhodamine 6G using a dose calibrator revealed most of the 18Frhodamine 6G labeled mitochondria remained contained within the injected hearts throughout reperfusion (77.3% ± 5.5, mean ± SEM).

Positron emission tomography and computed tomography revealed that the transplanted mitochondria were distributed from the heart apex to the base. Quantitative assessment of perfused mitochondrial position in the heart tissue demonstrated that 24.76% ± 2.50 (mean ± SEM) and 23.64% ± 2.42 (mean ± SEM) of the transplanted mitochondria were associated with cardiomyocytes and blood vessels, respectively (Cowan et al., 2016).

To demonstrate the efficacy of vascular delivery of mitochondria to the heart, we delivered unlabeled mitochondria to the regionally ischemic heart. In these experiments we demonstrated that vascular delivery of mitochondria through the coronary arteries following transient ischemia and reperfusion significantly decreased myocardial infarct size and significantly enhanced postischemic functional recovery. The reductions in infarct size and the enhancement of regional functional recovery achieved with vascular delivery of mitochondria are not significantly different from that obtained by direct injection of mitochondria (McCully et al., 2009; Masuzawa et al., 2013; Cowan et al., 2016). These data indicate that autologous mitochondrial transplantation is efficacious as a cardioprotective therapy whether these organelles ae delivered by directly injection or delivered by vascular infusion through the coronary arteries.

2.3 Localization to endorgan by vascular delivery
In all our studies, we have noted that the distribution of mitochondria following delivery by direct injection or by vascular infusion remains within the heart and is not detectable in other organs. This finding is important as the delivery of mitochondria by vascular infusion provides for localized therapy without crosscontamination to other endorgans. In the heart we deliver mitochondria by injection into the coronary arteries to avoid systemic distribution; however, we have also delivered mitochondria to the lung by vascular infusion through the pulmonary artery.

In these studies, the mitochondria were labeled with 18Frhodamine 6G as above. Positron emission tomography and computed tomography showed that the mitochondria were localized in the lung and were not detectable in any other areas of the body.

At present, we do not have a mechanism for this “endorgan homing”, where the transplanted mitochondria are retained by the immediate downstream organ, but suggest that this observation may play an important therapeutic role in future studies and applications using mitochondrial transplantation.

2.4 Mechanisms of mitochondrial uptake and internalization
In previous studies, we have investigated a variety of mechanisms that may be associated with mitochondrial uptake and internalization following mitochondrial. These studies were performed using well established pharmacological blockers of clathrin mediated endocytosis, actinmediated endocytosis, macropinocytosis, and tunneling nanotubes (Le et al., 2000; BereiterHahn Jet al., 2008; Lou et al., 2012; Islam et al., 2012; Huang et al., 2013; Kitani et al., 2014). Our studies demonstrated that autologous mitochondria delivered by direct injection are internalized by actindependent endocytosis (Pacak et al., 2015).

Mitochondrial uptake by vascular delivery appears to be more complex. The rapid and widespread uptake of mitochondria when delivered by vascular infusion would suggest that mechanisms allowing for the rapid passage of mitochondria through the vascular wall are involved. Previous studies support the concept that cells can routinely escape from the circulation. It has been previously shown that certain cardiac and mesenchymal stem cells appear to be actively expelled from the vasculature in a process different from diapedesis (Cheng et al., 2012; Allen et al., 2017).

Transmigration of stem cells through the vascular wall requires extensive remodeling of the endothelium (Allen et al., 2016). Cheng et al., 2012 have suggested that this occurs in three stages 1) adhesion of infused cells to the microvessel lining; 2) pocketing of infused cells by endothelial projections; 3) breakdown of the adjacent vascular wall, releasing cells into the interstitium.

The first two steps require integrindependent interactions between transplanted cells and host endothelium, while matrix metalloproteinases mediate the subsequent breakdown of the microvessel wall. These steps are unlikely to occur rapidly or to be associated with the mitochondrion.

Another possible mechanism for mitochondrial uptake may be diapedesislike. Previous studies have shown that some cells routinely escape from the circulation. For example, leukocyte extravasation (i.e. diapedesis) between venous endothelial cells is a wellunderstood process that involves cell adhesion proteins.

It is unlikely, however, that isolated, exogenous mitochondria use a similar mechanism as leukocytes to move through the wall of blood vessels as they do not express the array of proteins involved in diapedesis on their outer membrane (Pagliarini et al., 2008; Calvo et al., 2010).

Whether mitochondria pass between or through endothelial cells and the region of the vasculature at which this process occurs remain to be determined. The size of the isolated mitochondria and the rapidity of their uptake by vascular perfusion would not appear to support any currently defined mechanisms. At present, we are investigating several possible mechanisms of mitochondrial uptake by vascular infusion; but, have yet to conclusively identify a definitive mechanistic route.

3. Mitochondrial transplantation for cardioprotection

3.1 Animal models
In the isolated perfused rabbit heart model and subsequent studies in the in vivo rabbit and pig heart models, we have investigated the use of mitochondrial transplantation as a cardioprotective therapy (McCully et al., 2009; Masuzawa et al., 2013; Cowan et al., 2016; Kaza et al., 2016). In these studies, the myocardium was made temporarily ischemic by ligating the left anterior descending artery with a snare. This temporary ligation results in the attenuation or cessation of coronary blood flow such that oxygen delivery to the myocardium is insufficient to meet oxygen demand. The resulting injury is termed ischemia/reperfusion injury and is characterized by the loss of myocardial cell viability and decreased contractile function within the area at risk. In general, we subject 2530% of the left ventricular mass to ischemia, resulting in a loss in cell viability of approximately 30% and a reduction in contractile force of approximately 25% based on regional systolic shortening. These decrements in cell viability and function are reproducible between large animal models and mimic events occurring in human acute myocardial infarction or surgical intervention during cardiopulmonary bypass. Following 30 minutes of temporary ligation the snare is released and coronary blood flow to the region is reestablished. To show the efficacy of mitochondrial transplantation, we typically deliver autologous mitochondria by direct injection or by vascular delivery, just prior to or at the very start of reperfusion. Control hearts receive vehicle alone (respiration buffer with no mitochondria).

3.2 Effects of mitochondrial transplantation on arrhythmogenicity
The transplantation of mitochondria has similarities in methodology with stem cell transplantation. In stem cell transplantation using skeletal muscle myoblasts, it has been reported that skeletal muscle myoblasts, clustering of the cells can occur, resulting in arrhythmia with postoperative episodes of sustained tachycardia due to alterations in electrical coupling (Macia et al., 2009). To demonstrate that mitochondrial transplantation is not proarrhythmic we have used serial 12 lead electrocardiogram (ECG) analysis and optical mapping (Masuzawa et al., 2013). Our results show that there is no proarrhythmogenicity associated with mitochondrial transplantation. We observed no ventricular tachycardia, bradycardia, fibrillation, or conduction system defects or repolarization heterogeneity associated with mitochondrial transplantation. In addition, there were no changes in serial ECG, QRS duration or corrected QT interval. There was no evidence of any changes associated with ventricular wall motion disturbances, left ventricle hypertrophy, valve dysfunction, fibrosis, or pericardial effusion either at the time of injection, or at any time up to 4 weeks following transplantation of autogeneic mitochondria (Masuzawa et al., 2013).
To confirm these findings, we also performed optical mapping (Masuzawa et al., 2013). In these studies, a 400fold increase in mitochondria, 8.4 x 107/gram tissue wet weight as compared to 2 x 105/gram tissue wet weight, was directly injected into the heart. The number of mitochondria injected per site was significantly greater than that used in the in situ heart (4.2 × 108 vs. 1.2 × 106), so that any acute arrhythmogenic responses could be observed. Our results showed that even with this highly increased mitochondrial load, sequential isopotential maps from the left ventricles injected with mitochondria showed no detectable abnormal impulse propagation on the myocardial surface associated with mitochondrial transplantation.

3.3 Effects of mitochondrial transplantation on immune and autoimmune response
In all our studies, we have used autologous mitochondria isolated from the patient’s own body for cardioprotection. The use of autologous tissue was based on the assumption that this approach would not cause an immune response and avoid the need for antirejection therapy required for nonautologous cellbased therapies (Kofidis et al., 2005). To confirm these expectations serial analysis immune and inflammatory markers was performed. The effects of mitochondrial transplantation were also investigated using multiplex (42plex) analysis of cytokines and chemokines (Masuzawa et al., 2013). The results from these studies confirmed that there was no immune or inflammatory response associated with autologous mitochondrial transplantation (Masuzawa et al., 2013). In addition there was no upregulation of cytokines associated with the immune response that is seen in patients with acute heart transplantation rejection (Rose, 2011).
The possibility of autoimmune response due to the presence of increased mitochondrial number in the myocardium was also investigated. The need for these studies was based on previous studies indicating that oxidative modification of E2 subunits of mitochondria pyruvate dehydrogenase, branched chain 2oxoacid dehydrogenase, and 2oxoglutarate dehydrogenase is a critical step leading to the induction of an autoimmune response in the liver as demonstrated by the presence of antimitochondrial antibodies (Leung et al., 2007). Indirect immunofluorescence was used to test for the presence of antimitochondrial antibodies (AMA) in serial blood samples. AMAs were not detected in the serum of any animals treated with autologous mitochondrial transplantation indicating that the transplantation of mitochondria does not induce an autoimmune response (Masuzawa et al., 2013).
3.4 Effects of mitochondrial transplantation on cellular viability and function

Our results have demonstrated that direct injection or vascular delivery of mitochondria to the heart rescues cell function and myocardial contractile function following ischemia and reperfusion. Creatine kinaseMB isoenzyme (CKMB) and cardiac troponin–I (cTnI) are specific and sensitive markers of myocardial injury, and elevated levels indicate myocardial injury (Pourafkari et al., 2015). Mitochondrial transplantation significantly decreased serum CKMB and cTnI indicating that myocardial injury following 30 minutes of transient ischemia was decreased (Masuzawa et al., 2013; Kaza et al., 2016). These effects were confirmed by infarct analysis using triphenyl tetrazolium chloride analysis to determine necrosis and terminal deoxynucleotidyl transferasemediated dUTP nickend labeling (TUNEL) and caspase activity to determine apoptosis in hearts treated with mitochondrial transplantation and those treated with vehicle alone (Masuzawa et al., 2013). Our results demonstrated that mitochondrial transplantation significantly decreased myocardial injury, including both necrosis and apoptosis, resulting from transient ischemia.
Increased myocardial cellular viability would be expected to correlate with enhanced myocardial function and our studies confirm this assumption. Our results show that 10 minutes following mitochondrial transplantation myocardial function is significantly enhanced as compared to hearts receiving injection of respiration media (vehicle) alone and that this function remains enhanced for at least 28 days – the end point of our studies (Masuzawa et al., 2013; Kaza et al., 2016). This is in contrast to hearts receiving vehicle alone that had persistent left ventricular hypokinesis.
3.5 Mechanisms of mitochondrial transplantation
The mechanisms through which mitochondrial transplantation provides cardioprotection have yet to be fully elucidated. At present, our studies have shown that the transplanted mitochondria act both extraand intracellularly. Once transplanted, the mitochondria increase total tissue ATP content and ATP synthesis. This increase in high energy acts rapidly, as early

as 10 minutes following delivery of the mitochondria to the heart, to enhance cardiac function as determined by echocardiography and pressure volume measurement. The mitochondria then act to upregulate proteomic pathways for the mitochondrion and the generation of precursor metabolites for energy and cellular respiration (P < 0.05, Enrichment Score > 2.0) (Masuzawa et al., 2013).
At 10 minutes to one hour following transplantation the transplanted mitochondrial are internalized into cardiomyocytes, by actindependent endocytosis (Pacak et al., 2015). Once internalized, the transplanted mitochondria further increase cardiomyocyte ATP content and upregulate cardioprotective cytokines (Masuzawa et al., 2013). These cytokines have been shown to be associated with enhanced cardiac function by stimulating cell growth, proliferation, and migration, enhancing vascularization, providing protection against cardiomyocyte apoptosis and improving functional cardiac recovery and cardiac remodeling independent of cardiac myocyte regeneration (Masuzawa et al., 2013).
We have recently shown that transplanted mitochondria also act at the mitochondrial genomic level. In previous experiments, we have shown that following ischemia there is damage to mitochondrial DNA resulting in the reduction of ATP synthesis (Levitsky et al., 2003). These alterations were associated with poor recovery following cardiac surgery in humans. Our studies demonstrate that mitochondrial transplantation replaces damaged mitochondrial DNA with intact mitochondrial DNA and rescues myocardial cell function (Pacak et al., 2015).
The transplanted mitochondria maintain viability and function for at least 28 days, the limit of our studies (Masuzawa et al., 2013; Cowan et al., 2016; Kaza et al., 2016). This is in contrast to xenoand allotransplanted cells that are rapidly rejected leading to the loss of transplanted cells, despite the use of antirejection pharmaceuticals (Yau et al., 2003; Hamamoto et al., 2009).

4 Human application
Premised upon these in vivo studies demonstrating the efficacy of mitochondrial transplantation for cardioprotection we have recently performed the first clinical application of mitochondrial transplantation (Emani et al., 2017). The study was performed in pediatric patients who suffered myocardial ischemiareperfusion injury. All procedures were performed under an Innovative Therapies process developed by the Boston Children’s Hospital Institutional Review Board. An individual review of the proposed therapy for each patient was provided by two independent physicians, not involved in the patient’s care. Families were extensively counseled regarding the experimental nature of the procedure and a separate Innovative Therapies consent form was signed.
Five pediatric patients in critical condition who were unable to be weaned off extracorporeal membrane oxygenation (ECMO) support due to myocardial dysfunction related to ischemia and reperfusion were treated with autologous mitochondria. Patient diagnosis included dextrotransposition of the great arteries (4 days and 25 days of age ), hypoplastic left heart syndrome (6 days), left ventricular outflow tract obstruction (6 months of age) and tricuspid atresia 1B (2 years of age) (Emani et al., 2017).
The cause of ischemia was coronary artery obstruction that was relieved in 4 patients, and LV distension with subendocardial ischemia in one patient. The autologous mitochondria were isolated from the patient’s rectus abdominis muscle. The patients received 10 mitochondrial injections, 100 uL each containing 1 x 107 ± 1 x 104 mitochondria. The mitochondria were delivered to the myocardium by direct injection with a 1 mL tuberculin syringe (28gauge needle). All injections were delivered to the area affected by ischemiareperfusion that was identified by epicardial echocardiography as being hypokinetic. Following mitochondrial transplantation, all 5 patients had significant improvement in their myocardial systolic function. Epicardial echocardiography showed moderate to severe systolic function ventricular dysfunction with regional hypokinesis prior to treatment (Emani et al., 2017). Ventricular function was improved to mildmoderate to normal systolic function at 46 days following autologous mitochondrial transplantation and improved to mild dysfunction in one patient and normal systolic function with no regional hypokinesia detected in any patient at 10 days following autologous mitochondrial transplantation (Emani et al., 2017).

All but one patient were successfully weaned off ECMO support by the 2nd day post mitochondrial transplantation. The single patient who was unable to wean off ECMO support suffered irreversible multiorgan failure despite the recovery of myocardial function following mitochondrial transplantation. This patient was on ECMO support for 15 days prior to treatment with mitochondrial transplantation. There were no adverse complications such as arrhythmia, intramyocardial hematoma or scarring with mitochondrial transplantation, in agreement with our animal studies. This case study demonstrates for the first time the potential role mitochondrial transplantation to improve ventricular dysfunction following ischemiareperfusion injury in humans.

While this is the first clinical usage of mitochondrial transplantation in humans, the same protocol can be used in adults and in other settings of ischemiareperfusion injury (Emani et al., 2017). The ability to use mitochondrial transplantation for clinical intervention in situations such as the stunned myocardium is enhanced by the fact that mitochondrial harvest and isolation can be performed within 2030 minutes during the same procedure and involves minimal manipulation of muscle tissue.

5 The timing of mitochondrial transplantation delivery
Studies by us and by others have shown that the mechanisms associated with ischemic myocardial injury converge on the mitochondrion. Rapidly following the onset of ischemia there are alterations in mitochondrial structure, complex activity, oxygen consumption, high energy synthesis and changes in mitochondrial transcriptomics and proteomics.

All these changes occur during ischemia and persist following reestablishment of coronary blood flow (reperfusion). For practical efficacy we therefore deliver the transplanted mitochondria just prior to reperfusion or during early reperfusion in order to limit the effects of ischemia on the transplanted mitochondria. We reasoned that if the mitochondria were transplanted at the start of ischemia they too would be damaged. This approach has been shown to be efficacious in our animal studies. In the human trial described above the delivery of mitochondria was many days after the ischemic insult. In these cases, the patient’s heart was unable to provide sufficient contractile force to be weaned from ECMO (Emani et al., 2017).

We believe that the role of mitochondrial transplantation is the rescue of cells and cellular function. This is premised on our animal studies. The results obtained in the clinical studies in humans are most likely the result of reversal of myocardial stunning. Stunning “describes the mechanical dysfunction that persists after reperfusion despite the absence of myocellular damage and despite the return of normal or nearnormal perfusion (Kloner et al., 1998). Mentzer, 2011, has noted that myocardial stunning, is a frequent consequence after heart surgery and is characterized by a requirement for postoperative inotropic support despite a technically satisfactory heart operation. In the stunned heart the myocardial cells remain alive but there is a prolonged depression of cardiac contractility after reperfusion. This depression in contractility can last days but with extended time results in patient mortality.
In the human cases in which we have used mitochondrial transplantation it is likely that we have rescued the stunned myocardium. The mechanisms most likely involve those we have demonstrated in our animal models, namely, restoration of high energy synthesis and replacement of damaged mitochondrial DNA. The veracity of these mechanisms remains to be elucidated.

6 Mitochondrial transplantation for the rescue of other tissues
While our studies have focussed on the heart and ischemia and reperfusion, other groups have shown that mitochondrial transplantation can be used to enhance drug sensitivity in human breast cancer cells (Elliott et al., 2012); to rescue cell function in cells harboring the mitochondrial DNA mutation (Chang et al., 2013); Parkinson’s disease (Chang et al., 2016); liver ischemia/reperfusion injury (Lin et al., 2013); in cellular studies demonstrating that isolated mitochondria rescue mitochondrial respiratory function and improved the cellular viability in cardiomyocytes (Kitani et al., 2014) and neurorecovery after stroke ( Hayakawa et al., 2016).

These studies demonstrate the potential of mitochondrial transplantation in a variety of diseases. The utility in other organ related diseases and syndromes remains to be investigated. However, diabetes, Alzheimer’s disease and dementia, posttraumatic stress disorder, concussion and others have all been shown to be associated with alterations occurring at or affecting mitochondrial function. Included in these pathological disorders are ischemiareperfusion events affecting pulmonary, renal, hepatic, cerebral, ocular and skeletal muscles. We expect that the application and usage of mitochondrial transplantation will provide for a simple and efficacious therapeutic approach to many of these disease and will significantly ameliorate morbidity and mortality.

Akhmedov, A.T., Rybin, V., MarínGarcía, J., 2015. Mitochondrial oxidative metabolism and uncoupling proteins in the failing heart. Heart Fail. Rev. 20, 227249.
Allen, T.A., Gracieux, D., Talib, M., Tokarz, D.A., Hensley, M.T., Cores, J., Vandergriff, A., Tang, J., de Andrade, J.B., Dinh, P.U., Yoder, J.A., Cheng, K., 2017. Angiopellosis as an Alternative Mechanism of Cell Extravasation. Stem Cells. 35,170180.
BereiterHahn, J., Vöth, M., Mai, S., Jendrach, M., 2008. Structural implications of mitochondrial dynamics. Biotechnol. J. 3, 765780.
Black, K.M., Barnett, R., Bhasin, M.K., Daly, C., Dillon, S.T., Libermann, T.A., Levitsky, S., McCully, J.D., 2012, Microarray and Proteomic Analysis of Cardioprotection in the Mature and Aged Male and Female. Physiol. Genomics 44,10271041.
Calvo, S.E., Mootha, V.K., 2010, The mitochondrial proteome and human disease. Annu Rev Genomics Hum. Genet. 11, 2544.
Chang, J.C., Liu, K.H., Li, Y.C., Kou, S.J., Wei, Y.H., Chuang, C.S., Hsieh, M., Liu, C.S., 2013, Functional recovery of human cells harbouring the mitochondrial DNA mutation MERRF A8344G via peptidemediated mitochondrial delivery. Neurosignals. 21,160173.
Chang, J.C., Wu, S.L., Liu, K.H., Chen, Y.H., Chuang, C.S., Cheng, F.C., Su, H.L., Wei, Y.H., Kuo, S.J., Liu, C.S., 2016, Allogeneic/xenogeneic transplantation of peptidelabeled mitochondria in Parkinson’s disease: restoration of mitochondria functions and attenuation of 6hydroxydopamine–induced neurotoxicity. Translational Research. 170, 4056.
Chen, Q., Moghaddas, S., Hoppel, C.L., Lesnefsky, E.J., 2008, Ischemic defects in the electron transport chain increase the production of reactive oxygen species from isolated rat heart mitochondria. Am. J. Physiol. Cell. Physiol. 294, C460C466.
Cheng, K., Shen, D., Xie, Y., Cingolani, E., Malliaras, K., Marbán, E., 2012, Brief report: Mechanism of extravasation of infused stem cells. Stem Cells. 30, 28352842.
Cowan, D.B., Yao, R., Akurathi, V., Snay, E.R., Thedsanamoorthy, J.K., Zurakowski, D., Ericsson, M., Friehs, I., Wu, Y., Levitsky, S., del Nido, P.J., Packard, A.B., McCully, J.D., 2016, Intracoronary Delivery of Mitochondria to the Ischemic Heart for Cardioprotection. PLoS One. e0160889. doi: 10.1371/journal.pone.0160889. eCollection 2016.
Doenst, T., Nguyen, T.D., Abel, E.D., 2013, Cardiac metabolism in heart failure: implications beyond ATP production. Circ. Res. 113, 709724.
Durhuus, J.A., Desler, C., Rasmussen, L.J., 2015, Mitochondria in health and disease 3rd annual conference of society for mitochondrial research and medicine 1920 December 2013 Bengaluru, India. Mitochondrion. 20, 712.
Elliott, R. L., Jiang, X. P., Head, J. F., 2012, Mitochondria organelle transplantation: introduction of normal epithelial mitochondria into human cancer cells inhibits proliferation and increases drug sensitivity. Breast Cancer Res. Treat. 136, 347–354.

Emani, S.M., Piekarski, B.L., Harrild, D., del Nido, P.J., McCully, J.D., 2017, Autologous Mitochondria Transplantation for Ventricular Dysfunction following Myocardial IschemiaReperfusion Injury. In press J. Thorac. Cardiovasc. Surg. DOI:
Faulk, E.A., McCully, J.D., Tsukube, T., Hadlow, N.C., Krukenkamp, I.B., Levitsky, S., 1995, Myocardial mitochondrial calcium accumulation modulates nuclear calcium accumulation and DNA fragmentation. Annals Thorac. Surg. 60,338344.
Faulk, E.A., McCully, J.D., Hadlow, N.C., Tsukube, T., Krukenkamp, I.B., Federman, M., Levitsky, S., 1995a, Magnesium cardioplegia enhances mRNA levels and the maximal velocity of cytochrome oxidase I in the senescent myocardium during global ischemia. Circulation 92, 405412.
FernándezVizarra, E., Ferrín, G., PérezMartos, A., FernándezSilva, P., Zeviani, M., Enríquez, J.A., 2010, Isolation of mitochondria for biogenetical studies: An update. Mitochondrion. 10, 53262.
Fillmore, N., Lopaschuk, G.D., 2013, Targeting mitochondrial oxidative metabolism as an approach to treat heart failure. Biochim. Biophys. Acta. 1833, 857865.
Frezza, C., Cipolat, S., Scorrano, L., 2007, Organelle isolation: functional mitochondria from mouse liver, muscle and cultured fibroblasts. Nat. Protoc. 2, 287295.
Graham, J.M., 2001, Isolation of mitochondria from tissues and cells by differential centrifugation. Curr. Protoc. Cell Biol. Chapter 3, Unit 3.3.
Hausenloy, D.J., Barrabes, J.A., Bøtker, H.E., Davidson, S.M., Di Lisa,F., Downey, J., Engstrom, T., Ferdinandy, P., CarbreraFuentes, H.A., Heusch, G., Ibanez, B., Iliodromitis, E.K., Inserte, J., Jennings, R., Kalia, N., Kharbanda, R., Lecour, S., Marber, M., Miura, T., Ovize, M., PerezPinzon, M.A., Piper, H.M., Przyklenk, K., Schmidt, M.R., Redington, A., RuizMeana, M., Vilahur, G., VintenJohansen, J., Yellon, D.M., GarciaDorado, D., 2016, Ischaemic conditioning and targeting reperfusion injury: a 30 year voyage of discovery. Basic Res. Cardiol. 111,70.
Hsiao, F.C., Tung, Y.C., Chou, S.H., Wu, L.S., Lin, C.P., Wang, C.L., Lin, Y.S., Chang, C.J., Chu, P.H., 2015, FixedDose Combinations of ReninAngiotensin System Inhibitors and Calcium Channel Blockers in the Treatment of Hypertension: A Comparison of Angiotensin Receptor Blockers and AngiotensinConverting Enzyme Inhibitors. Medicine (Baltimore). 94:e2355
Huang, X., Sun, L., Ji, S., Zhao, T., Zhang, W., Xu, J., Zhang, J., Wang, Y., Wang, X., FranziniArmstrong, C., Zheng, M., Cheng, H., 2013, Kissing and nanotunneling mediate intermitochondrial communication in the heart. Proc. Natl. Acad. Sci. U. S. A. 110, 28462851.
Hamamoto, H., Gorman, J.H. 3rd, Ryan, L.P., Hinmon, R., Martens, T.P., Schuster, M.D., Plappert, T., Kiupel, M., St JohnSutton, M.G., Itescu, S., Gorman, R.C., 2009, Allogeneic mesenchymal precursor cell therapy to limit remodeling after myocardial infarction: the effect of cell dosage. Ann. Thorac. Surg. 87, 794801.
Hayakawa, K., Esposito, E., Wang, X., Terasaki, Y., Liu, Y., Xing, C., Ji, X., Lo, E.H., 2016, Transfer of mitochondria from astrocytes to neurons after stroke. Nature. 535, 551555.

Islam, M.N., Das, S.R., Emin, M.T., Wei, M., Sun, L., Westphalen, K., Rowlands, D.J., Quadri, S.K., Bhattacharya, S., Bhattacharya, J., 2012, Mitochondrial transfer from bonemarrowderived stromal cells to pulmonary alveoli protects against acute lung injury. Nat. Med. 18, 759765.
Kalogeris ,T., Baines ,C.P., Krenz, M., Korthuis, R.J., 3012, Cell biology of ischemia/reperfusion injury. Int. Rev. Cell. Mo.l Biol. 298:229317.
Kalogeris, T., Baines, C.P., Krenz, M., Korthuis, R,J., 2016, Ischemia/Reperfusion. Compr. Physiol. 7:113170.
Kaza, A.K., Wamala, I., Friehs, I., Kuebler, J.D., Rathod, R.H., Berra, I., Ericsson, M., Yao, R., Thedsanamoorthy. J.K., Zurakowski, D., Levitsky, S., del Nido, P.J., Cowan, D.B., McCully, J.D., 2016, Myocardial Rescue with Autologous Mitochondrial Transplantation in a Porcine Model of Ischemia/Reperfusion. In press J. Thorac. Cardiovas. Surg. DOI:
Kitani, T., Kami, D., Matoba, S., Gojo, S., 2014, Internalization of isolated functional mitochondria: involvement of macropinocytosis. J. Cell. Mol. Med. 18, 16941703.
Kloner, R.A.,, Bolli, R., Marban, E., et al., 1998, Medical and cellular implications of stunning, hibernation and preconditioning. An NHLBI Workshop. Circulation. 97,1848–1867.
Kofidis, T., Weissman, I., Fedoseyeva, E., Haverich, A., Robbins, R.C., deBruin, J.L., Tanaka, M., Zwierzchoniewska, M., 2005, They are not stealthy in the heart: embryonic stem cells trigger cell infiltration, humoral and Tlymphocytebased host immune response. Eur J. Cardiothorac. Surg. 28, 461466.
Kolwicz, S.C. Jr, Purohit, S., Tian, R. 2013, Cardiac metabolism and its interactions with contraction, growth, and survival of cardiomyocytes. Circ. Res. 113, 603616.
Kurian, G.A., Berenshtein, E., Kakhlon, O., Chevion, M., 2012, Energy status determines the distinct biochemical and physiological behavior of interfibrillar and subsarcolemmal mitochondria. Biochem. Biophys. Res. Commun. 428,376382.
Laskowski, M., Augustynek, B., Kulawiak, B., Koprowski, P., Bednarczyk, P., Jarmuszkiewicz, W., Szewczyk, A., 2016, What do we not know about mitochondrial potassium channels? Biochim. Biophys. Acta. 1857,2471257.
Le, P.U., Benlimame, N., Lagana, A., Raz, A., Nabi, I.R., 2000, Clathrinmediated endocytosis and recycling of autocrine motility factor receptor to fibronectin fibrils is a limiting factor for NIH3T3 cell motility. J. Cell. Sci. 113, 32273240.
Lesnefsky, E.J., Hoppe,l C.L., 2003, Ischemiareperfusion injury in the aged heart: role of mitochondria. Arch. Biochem. Biophys. 420:287297.
Lesnefsky, E.J., Chen, Q., Slabe, T.J., Stoll, M.S., Minkler, P.E., Hassan, M.O., Tander. B., Hoppel, C,L.. 2004,Ischemia, rather than reperfusion inhibits respiration through cytochrome oxidase in the isolated perfused rabbit heart: role of cardiolipin. Am. J. Phys. Heart Circ. Phys. 287, H258H267.

Lesnefsky EJ, Chen Q, Tandler B, Hoppel CL. Mitochondrial Dysfunction and Myocardial IschemiaReperfusion: Implications for Novel Therapies. Annu Rev Pharmacol Toxicol. 2017 6;57:535565
Leung, P.S., Rossaro, L., Davis, P.A., Park,O., Tanaka, A., Kikuchi, K., Miyakawa, H., Norman, G.L., Lee, W., Gershwin, M.E., 2007, Antimitochondrial antibodies in acute liver failure: implications for primary biliary cirrhosis. Hepatology 46, 14361442.
Levitsky, S., Laurikka, J., Stewart, R.D., Campos, C.T., Lahey, S.J., McCully, J.D., 2003, Mitochondrial DNA deletions in coronary artery bypass grafting patients. European Society for Surgical Research International Proceedings 38,149153.
Lin, H. C., Liu, S. Y., Lai, H. S., Lai I. R., 2013, Isolated mitochondria infusion mitigates ischemiareperfusion injury of the liver in rats. Shock. 39, 304310.
Lou, E., Fujisawa, S., Morozov, A., Barlas, A., Romin, Y., Dogan, Y., Gholami, S., Moreira, A.L., ManovaTodorova, K., Moore, M.A., 2012 Tunneling nanotubes provide a unique conduit for intercellular transfer of cellular contents in human malignant pleural mesothelioma. PLoS One. 2012;7:e33093.
Macia, E., Boyden, P.A., 2009, Stem cell therapy is proarrhythmic. Circulation 119, 18141823.
Madonna, R., Cadeddu, C., Deidda, M., Giricz, Z., Madeddu, C., Mele, D., Monte, I., Novo, G., Pagliaro, P., Pepe, A., Spallarossa, P., Tocchetti, C.G., Varga, Z.V., Zito, C., Geng, Y.J., Mercuro, G., Ferdinandy, P., 2015,Cardioprotection by gene therapy: A review paper on behalf of the Working Group on Drug Cardiotoxicity and Cardioprotection of the Italian Society of Cardiology. Int. J. Cardiol. 191, 203210.
Masuzawa, A., Black, K.M., Pacak, C.A., Ericsson, M., Barnett, R.J., Drumm, C., Seth, P., Bloch, D.B., Levitsky, S., Cowan, D.B., McCully, J.D., 2013, Transplantation of autologouslyderived mitochondria protects the heart from ischemiareperfusion injury. Am. J. Phys. Heart Circ. Physiol. 304, H966H982.
McCully, J.D., Levitsky, S., 2003, The Mitochondrial KATP Channel and Cardioprotection. Ann. Thorac. Surg. 75, S667S673.
McCully JD, Rousou AJ, Parker RA, Levitsky S. Age and gender differences in mitochondrial oxygen consumption and free matrix calcium during ischemia/reperfusion and with cardioplegia and diazoxide. Ann. Thorac. Surg. 2007;83 11021109.
McCully, J.D., Cowan, D.B., Pacak, C.A., Levitsky, S., 2009, Injection of Isolated Mitochondria During Early Reperfusion for Cardioprotection. Am. J. Phys. Heart Circ. Physiol. 296, 94105.
Mentzer, R.M. Jr., 2011, Myocardial protection in heart surgery J. Cardiovasc. Pharmacol. Ther. 16, 290297.
Olson, M.S., Von Korff, R.W., 1967, Changes in endogenous substrates of isolated rabbit heart mitochondria during storage. J. Biol. Chem. 242,325332.
Ong, S.B., Samangouei, P., Kalkhoran, S.B., Hausenloy, D.J., 2015 The mitochondrial permeability transition pore and its role in myocardial ischemia reperfusion injury. J. Mol. Cell. Cardiol. 78:2334.

OrenesPiñero, E., Valdés, M., Lip, G.Y., Marín, F., 2015, A comprehensive insight of novel antioxidant therapies for atrial fibrillation management. Drug Metab. Rev. 47, 388400.
Pacak, A.P., Preble, J.M., Kondo, H., Seibel, P., Levitsky, S., del Nido, P.J., Cowan, D.B., McCully, J.D., 2015, ActinDependent Mitochondrial Internalization in Cardiomyocytes: Evidence for Rescue of Mitochondrial Function. Biol. Open. 4, 622626.
Pagliarini, D.J., Calvo, S.E., Chang, B., Sheth, S.A., Vafai, S.B., Ong, S.E., Walford, G.A., Sugiana, C., Boneh, A., Chen, W.K., Hill, D.E., Vidal, M., Evans, J.G., Thorburn, D.R., Carr, S.A., Mootha, V.K., 2008, A mitochondrial protein compendium elucidates complex I disease biology. Cell. 34, 1223.
Pallotti, F., Lenaz, G., 2007, Isolation and subfractionation of mitochondria from animal cells and tissue culture lines. Methods Cell Biol. 80, 344.
Pourafkari, L., Ghaffari, S., Afshar, A.H., Anwar, S., Nader, N.D., 2015, Predicting outcome in acute heart failure, does it matter? Acta. Cardiol. 70, 653663.
Preble, J.M., Pacak, C.A., Kondo, H., McKay, A.A., Cowan, D.B., McCully, J.D., 2014, Rapid Isolation and Purification of Mitochondria for Transplantation. J. Vis. Exp. 91: e51682. doi: 10.3791/51682.
Preble, J.M., Kondo, H., Levitsky, S., McCully, J.D., 2014a, Quality Control Parameters for Mitochondria Transplant in Cardiac Tissue. JSM Biochem. Mol. Biol. 2014 2(1): 1008.
Riva, A., Tandler, B., Loffredo, F., Vazquez, E., Hoppel, C., 2005, Structural differences in two biochemically defined populations of cardiac mitochondria. Am. J. Physiol. Heart. Circ. Physiol. 289, H868H872.
Rosca, M.G., Hoppel, C.L., 2013, Mitochondrial dysfunction in heart failure. Heart Fail. Rev. 18, 607622.
Rose, N.R., 2011, Critical cytokine pathways to cardiac inflammation. J. Interferon Cytokine Res. 31, 705710.
Rousou, A.J., Ericsson, M., Federman, M., Levitsky, S., McCully, J.D., 2004, Opening of mitochondrial KATP enhances cardioprotection through the modulation of mitochondrial matrix volume, calcium accumulation and respiration. Am. J. Physiol. Heart Circ. Physiol. 287, H1967H1976.
Schmitt, S., Saathoff, F., Meissner, L., Schropp, E.M., Lichtmannegger, J., Schulz, S., Eberhagen, C., Borchard, S., Aichler, M., Adamski, J., Plesnila, N., Rothenfusser, S., Kroemer, G., Zischka, H.,2013, A semiautomated method for isolating functionally intact mitochondria from cultured cells and tissue biopsies. Anal. Biochem. 443, 6674.
Suleiman, M.S., Halestrap, A.P., Griffiths, E.J., 2001, Mitochondria: a target for myocardial protection. Pharmacol. Ther. 89, 2946.
Tansey, E.E., Kwaku, K.F., Hammer, P.E., Cowan, D.B., Federman, M., Levitsky, S.,. McCully, J.D., 2006, Reduction and Redistribution of Gap and Adherens Junction Proteins Following Ischemia/Reperfusion. Ann. Thorac. Surg. 82, 14721479.

Toyoda, Y., Friehs, I., Parker, R.A., Levitsky, S., McCully, J.D., 2000, Differential role of sarcolemmal and mitochondrial KATP channels in adenosine enhanced ischemic preconditioning. Am. J. Phys. Heart Circ. Physiol. 279, H2694H2703.
Toyoda, Y., Levitsky, S., McCully, J.D., 2001, Opening of mitochondrial ATPsensitive potassium channels enhances cardioplegic protection. Ann. Thorac. Surg. 71, 12811289.
Tsukube, T., McCully, J.D., Metz, R.M., Cook, C.U., Levitsky, S., 1997, Amelioration of ischemic calcium overload correlates with high energy phosphates in the senescent myocardium. Am. J. Physiol. (Heart Circ. Physiol.) 273, H418H427.
Wechsler, M.B., 1961, Studies on oxidative phosphorylation and ATPase activity of fresh and frozen brain mitochondria. Arch. Biochem. Biophys. 95,494498.
Wieckowski, M.R., Giorgi, C., Lebiedzinska, M., Duszynski, J., Pinton, P., 2009, Isolation of mitochondriaassociated membranes and mitochondria from animal tissues and cells. Nat. Protoc. 4, 15821590.
Yau, T.M., Tomita, S., Weisel, R.D., Jia, Z.Q., Tumiati, L.C., Mickle, D.A., Li, R.K., 2003, Beneficial effect of autologous cell transplantation on infarcted heart function: comparison between bone marrow stromal cells and heart cells. Ann. Thorac. Surg. 75, 169176.

Sulforaphane a powerful tool to fight cancer, aging, and other inflammatory health issues

Sulforaphane has been shown to be an effective antioxidant, antimicrobial, anticancer, anti-inflammatory, anti-aging, neuroprotective, and anti-diabetic (R). It also protects against cardiovascular and neurodegenerative diseases (R).

Many test-tube and animal studies have found sulforaphane to be particularly helpful for suppressing cancer development by inhibiting enzymes that are involved in cancer and tumor growth (R, R, R).

According to some studies, sulforaphane may also have the potential to stop cancer growth by destroying cells that are already damaged (R, R).

Sulforaphane appears to be most protective against colon and prostate cancer but has also been studied for its effects on many other cancers, such as breast, leukemia, pancreatic and melanoma (R).

Sulforaphane may also help reduce high blood pressure and keep arteries healthy — both major factors in preventing heart disease (R).

Recent research shows that Sulforaphane can help control blood glucose levels in type 2 diabetic patients as effectively as the most commonly used prescription medicine Metformin (R).

Sulforaphane and Broccoli Sprouts

Sulforaphane (SFN) is an isothiocyanate. It is derived from glucoraphanin, found in cruciferous vegetables such as broccoli, cabbage, cauliflower, brussels sprouts and kale (R, R).

Glucoraphanin is stable, but when the vegetables are cut or chewed, it comes in contact with the enzyme myrosinase, that also occurs naturally in these vegetables, and Sulforaphane is formed(RR).

Unlike the glucoraphanin, sulforaphane degrades quickly (R).

The quantity of glucoraphanin varies greatly in different plants. In general, levels of glucoraphanin and sulforaphane are highest in broccoli sprouts (R), but 3 day-old sprouts can contain up to 100 times more glucoraphanin than in mature plants (R).

Sulforaphane  is rapidly absorbed, reaching peak concentration after 1-3 hours. (RR). Levels are back to baseline within 72 hours (RRR).

Daily consumption of cruciferous vegetables can maintain levels of Sulforaphane in the body, if properly prepared (R), however most people find  it is easier to take a daily supplement.

Sulforaphane activation of  AMPK pathway is key to wide range of benefits

There are many hundreds of studies of Sulforaphane’s  effect in fighting various disease and metabolic problems.  Activation or Inhibition of several different genes are described in how it manages to impact such a wide range of health problems.

As with vitamin antioxidants the notion that supplements act as “antioxidants” in human cells is called into question []. Emerging evidence suggests that the most effective supplement exert their intracellular effects not as direct “antioxidants” per se but as modulators of signaling pathways.

Compared with widely used phytochemical-based supplements like curcumin, silymarin, and resveratrol, sulforaphane more potently activates Nrf2 – which researchers call the “Master Regulator” of Cell Defense (R).

A list of genes and enzymes Sulforaphane influences is at the bottom of this page.

Perhaps even more meaningful is that like Metformin and Berberine (R), Sulforaphane strongly activates AMPK, which raises the intracellular NAD+ concentrations and activates SIRT1 which has been shown to have numerous disease fighting and anti-aging potential(R).

It’s possible many of the health benefits are at least partially related to this AMPK/NAD+/SIRT activity.

Here is a  short list of studies showing Sulforaphane activating AMPK to fight Cancer(R),Diabetes(R),Obesity(R),Neurological disease (R),Heart disease (R),HIV (R),Colitis(R).

Sulforaphane helps prevent and can even kill cancer

3-5 servings per week of Cruciferous vegetables decrease the risk of cancer by 30-40% (R).

Even ONE serving of cruciferous vegetables per week significantly reduced the risk of pharynx, colorectal, esophageal, kidney and breast cancer (R).

In vitro, Sulforaphane has been demonstrated to kill breast cancer cells (R),oral squamous cell carcinoma cells (R), colorectal cancer cells (R),  cervical, liver, prostate, and  leukemia cancer cells (RR), while having little to no effect on healthy cells (R)

Sulforaphane combats cancer by multiple mechanisms:

  • Sulforaphane reduces inflammation by inhibiting the NF-κB pathway(R).
  • Sulforaphane induces cancer cell death (R).
  • SFN inhibits Phase I enzymes that enable cancer cell growth (R).
  • SFN induces Phase II enzymes that clear  DNA damaging chemicals (R).
  • Sulforaphane thereby inhibits cancer cell proliferation  (R)

In addition to the numerous cancer fighting mechanism of Sulforaphane, it is also very effective at enhancing commonly used anti-cancer drugs such as    cisplatin, gemcitabine, doxorubicin, and 5-fluorouracil    , allowing for smaller dosages and limiting toxicity to healthy cells (R).

Sulforaphane helps lower Cholesterol

Clinical studies with humans has shown eating broccoli reduces LDL cholesterol.

Twelve healthy subjects that consumed 100 grams per day of broccoli sprouts lowered LDL cholesterol, increased HDL cholesterol, and improved maarkers for oxidative stress  (R).

Sulforaphane May Help Parkinson’s, Alzheimers, Huntingtons

In mouse models of Parkinson’s disease, Sulforaphane increased dopamine levels in the brain to alleviate loss of motor coordination(RRRR).

A buildup of amyloid beta ( Aβ ) peptides are thought to be the cause of Alzheimer’s disease.  Broccoli sprouts were shown to prevent amyloid beta buildup and cell death (RR).

Sulforaphane has also bee shown to reduce Aβ plaque, and lessens cognitive impairment in mouse models of Alzheimer’s disease (RRR).

Sulforaphane activates a protein that slows huntingtins disease in mice (R).

Sulforaphane Prevents and Combats Heart & Cardiovascular Disease

Observational studies in humans has shown those who eat 3-5 servings of cruciferous vegetables a week have a significantly decreased risk of cardiovascular disease (R).

Research with mice shows Sulforaphane decreases blood pressure (RR).

Rats that were given Sulforaphane after heart attack exhibited  reduced heart damage  (R).

Sulforaphane helps prevent atherosclerosis (R) and minimizes inflammation caused by hardening of arteries in mice (R).

Sulforaphane reduces formation of  blood clots and platelet aggregation in humans (R).

Lastly, Sulforaphane has proven beneficial in minimizing damage from strokes, with decreased brain tissue damage (R), and loss of neurological function (R).

Sulforaphane helps control Diabetes and fight  Obesity

In humans, eating Broccoli sprouts  increased  HDL cholesterol, and lowered  triglycerides, insulin, insulin resistance,oxidative stress, and C-Reactive Proteins (RRR).

Sulforaphane decreases incidence and severity of the following diabetes complications in mice (RR)

  • vascular complications .
  • diabetes-induced heart dysfunction
  • heart damage in mice
  • nephropathy
  • tissue damage

Mice fed a high fat diet to induce obesity that were subsequently treated with sulforaphane for 3 weeks had significantly less weight gain, and improved insulin resistance, glucose and cholesterol levels (R,R).

Sulforaphane is Antiviral

Eating Broccoli sprouts increase the bodies natural anti-virus response and reduce influenza (R, R).

In vitro, Sulforaphane  combats  influenza, HIV, Epstein-Barr virus (R) and hepatitis C virus (R).

Sulforaphane Combats Bacterial and Fungal Infections

Human β-defensin-2 (HBD-2) is a key part of our defense against bacterial invasion. Sulforaphane increases HBD-2 in response to  to 23 of 28 bacterial species (R).

Cystic fibrosis patients have increased levels of Mycobacterium abscessus.  Treatment with Sulforphane of such macrophages  significantly decreased bacterial burden (R).

Sulforaphane Combats Inflammation

Nuclear Factor Kappa-B (NF-kB) is a well known driver of inflammation. Sulforaphane greatly decreases NF-kB activity (R).

As previously mention, Sulforaphane very strongly activates Nrf2, which lowers inflammation (RR).

Sulforaphane May Combat Depression and Anxiety

Inflammation has been recognized as one of the causes of depression. By reducing inflammation, sulforaphane can help combat depression.

Repeated SFN administration reverses depression– and anxiety-like behaviors in chronically stressed mice, likely by inhibiting the hypothalamic-pituitary-adrenal (HPA) axis and inflammatory responses to stress (RR).

In another study, it was shown that Nrf2 deficiency in mice results in depressive-like behavior, while the induction of Nrf2 by sulforaphane has antidepressant-like effects (R).

Also, dietary intake of glucoraphanin during the juvenile and adolescent periods in mice prevents the onset of depression-like behaviors at adulthood (R).

Sulforaphane Protects the Brain and Restores Cognitive Function

Sulforaphane increases neuronal BDNF in mice, a factor that supports the survival of existing neurons and encourages the formation of new neurons and synapses (R).

SFN reduces brain inflammation in various animal models of pathogen-induced neuroinflammation and neurodegenerative disease (RRRR).

Sulforaphane promotes microglia differentiation from pro-inflammatory M1 to anti-inflammatory M2 state. This reduces brain inflammation and restores spatial learning and coordination in rats (R).

Sulforaphane is beneficial in various pathological conditions:

  • SFN improves cognitive performance and reduces working memory dysfunction in rats after traumatic brain injury (R).
  • SFN attenuates cognitive deficits in mouse models of psychiatric disease. Also, the intake of glucoraphanin during the juvenile and adolescent periods prevents the onset of cognitive deficits at adulthood (R).
  • SFN alleviates brain swelling in rats, by attenuating the blood-brain barrier disruption, decreasing the levels of pro-inflammatory cytokines, and inhibiting NF-κB (R). SFN also increases AQP4 (a water channel protein) levels, thereby reducing brain swelling (R).
  • SFN prevents memory impairment and increases the survival of hippocampal neurons in diabetic rats (R).

Sulforaphane recovers memory in mice and rats with chemically induced memory impairment (RR R).

SFN exerts positive effects against brain damage induced by acute COpoisoning in rats (R).

Sulforaphane protects human neurons against prion-mediated neurotoxicity (R).

Insufficient NRF2 activation in humans has been linked to neurodegenerative diseases such as Parkinson’s disease, Alzheimer’s disease and amyotrophic lateral sclerosis (R). SFN, as a potent Nrf2 activator, may help in the treatment of these diseases.

Sulforaphane Improves Symptoms of  Autism

Sulforaphane activates several genes that lower inflammation and protect cells from oxidative stress and DNA damage, which are much higher in those with Autism (R).

In a clinical trial of 29 young men with  moderate to severe autism (age 13-27), Sulforphane treatment over 13 weeks resulted in a 35% improvement in disruptive behavior(R).

Sulforaphane relieves Gastrointestinal inflammation, colitis, and ulcers

Aspirin and NSAID’s are very effective at relieving pain, but can damage stomach lining and even cause ulcers.  Sulforaphane has been shown to protect agains such damage in mice (RR,R).

Sulforaphane has also been shown to increase Nrf2 and decrease inflammation in mice with colitis (R,R).

Sulforaphane May be Beneficial in Airway Inflammation and Asthma

Sulforaphane received airway inflammation and asthma symptoms in mice  (R)

Broccoli sprout extract  relieved airway inflammation in humans exposed to vehicle exhaust levels similar to those on a Los Angeles freeway (R).

But in other studies,  broccoli sprouts did not alleviate symptoms of asthma (R),  COPD (R) or ozone-induced airway inflammation (R).

Sulforaphane Can Be Beneficial in Arthritis

We previously pointed out that Sulforaphane strongly activates Nrf2, which relieves inflammation in many conditions.  In additions, sulforaphane was found to inhibit metalloproteinases that cause osteoarthritis and cartilage destruction (R).

Sulforaphane also decreases inflammatory cytokines, reducing symptoms of arthritis in mice (R).

Sulforaphane Protects the Eyes

Sulforaphane protects photoreceptor cells from excessive light exposure damage  (R) and  degeneration (R) in humans, and light-induced retinal damage in mice (R).

Sulforaphane protects human retinal cells and delays onset of cataracts (R), and helps prevent complications after cataract surgery (R).

In mice, sulforaphane helped maintain vitamin A and C levels in retinal cells to prevent damage from oxidative stress (RRR).

Negative Side Effects

Possible liver Toxicity at extreme dosages

There has been a single report of liver toxicity in one individual that consumed over 800 ml per day of broccoli juice for 4 weeks, but function returned to normal wishing 15 days or discontinuing the juice (R).

Note this individual was making juice from mature broccoli plants which have many different active substance and are not recommended for source of glucoraphanin/sulforaphane as the young sprouts have up to 100 times higher glucoraphanin levels.

Maximizing Bioavailability

Glucoraphanin Sources

Broccoli has the highest amounts of glucoraphanin of any vegetable, but it is also found in Brussel sprouts, Kale, Cabbage, Bok Choy, and several others (R).

As previously mentioned, young broccoli sprouts have up to 100 times more glucoraphanin than mature broccoli, making them ideal sources if you want to grow or purchase them (RR).

Myrosinase also required

Remember, Sulforaphane is only formed when it comes in contact with the enzyme myrosinase, by chewing or chopping or other processing.

Myrosinase is a fragile enzyme that is quickly damaged by heating (boiling over 1 minute), or freezing (R).

Many Supplements do not provide active Myrosinase

Broccoli has long been known to provide many health benefits and as such, broccoli sprout supplements are not new to the market.  Unfortunately, many of these were developed before researchers realized the Myrosinase + Glucoraphanain = Sulforaphanin  equation required great care in processing to protect the Myrosinase.

As a result, most broccoli supplements do not provide ANY active myrosinase  (RR).

Sulforaphane was found to be 7 times greater in fresh broccoli sprouts vs supplements with inactive myrosinase  (R).

Glucoraphanin powder with inactive tyrosinase can be combined with a good source of tyrosinase such as broccoli sprouts or mustard seeds to greatly increase Sulforaphane absorption (R).

Mustard seed

The myrosinase found in broccoli is quite fragile and inactivated by freezing or heat. A much more robust form of myrosinase  is found in Mustard seeds. The addition of powdered mustard seed to heat processed broccoli dramatically increases sulforphane (R).


We’ve looked at all the leading brands sold on Amazon. There are several that are good, and many that are garbage. The best quality we found is this one by Avmacol.


Sulforaphane activates genes and enzymes that stimulate antioxidant production:

  • inhibits Phase I enzymes CYP1A1, CYP1A2, CYP1B1, CYP2B2 and CYP3A4 (RR).
  • activates (RR). SFN reacts with Keap1, thereby releasing Nrf2 from Keap1 binding (R).
  • increases other Phase II enzymes: NQO1, GSTA1, and HO-1 (RRRR).
  • blocks SXR (R).

Sulforaphane inhibits inflammation:

  • inhibits NfkB (RRR).
  • inhibits TNF-α (RR), NLRP3, IL-1β, IL-18 (R), IFN-gamma and IL6 (RRR).
  • inhibits IL-17 (RR).
  • inhibits TGF-β/Smad (R).
  • increases IL-10 (RRR), IL-4, Arg1, and YM-1 (R).
  • inhibits NO, iNOS and COX-2 (RRR).
  • silences Th17/Th1  (RR).
  • inhibits IL-23 and IL-12 (R).
  • inhibits MMP-9 (RR).
  • inhibits LDH and PGE2 (R).

Sulforaphane changes gene expression:

  • Sulforaphane inhibits DNMT1 and DNMT3A (R)
  • SFN is one of the most potent (histone deacetylase) HDAC inhibitors found to date (R).
  • SFN inhibits HDAC1, HDAC2, HDAC3, and HDAC4 (RR).
  • SFN decreases miR-21 and TERT (R).

Sulforaphane induces cell death (apoptosis) in cancer:

  • SFN activates caspase-3, caspase-7, caspase-8, caspase-9 (RR).
  • SFN decreases anti-apoptotic Bcl-2 (R) and Bcl-XL (R).
  • SFN increases pro-apoptotic Bax (R).
  • SFN induces p21 (CDKN1A) (R) and p53 (R).
  • SFN inactivates PARP (R).
  • SFN decreases HIF1A (R).
  • SFN decreases β-catenin (CTNNB1) (R).

More on Sulforaphane and Cancer

The mechanisms of SFN effects on cancer cells have been well studied. It suppresses the proliferation of cancer cells via diverse mechanisms including cell-cycle arrest, apoptosis induction, ROS production, and manipulation of some signaling pathways (166). SFN inhibits proliferation of PC-3 cells in culture in concentrationand time-dependent manner. Singh et al. (167) showed that oral administration of SFN led to >50% reduction in PC-3 xenograft tumor volume in SFN-treated mice in 10 days and more than 70% reduction in 20 days after starting treatment with no effect on body weight.

They also reported that SFN changes the Bax: Bcl-2 ratio, activates caspases 3, 8, and 9, and cleaves and inactivates PARP protein. The authors proposed that SFN induces apoptosis in PC-3 xenograft tumors in a p53-independent manner through cytoplasmic and mitochondrial pathways. Liquid chromatography–mass spectrometry (LC-MS) analyses performed by Rose et al. (17) showed the presence of 7-methylsulphinylheptyl isothiocyanates in watercress (Rorippa nasturtium aquaticum) extract and 4-methylsulfinylbutyl nitrile and 4-methylsulfinylheptyl isothiocyanates in the broccoli extract. Their investigations showed that these compounds contribute to the inhibitory effects of broccoli and watercress extracts on the invasion of MDA-MB-231 cancer cells through suppression of MMP-9 activity.

Treatment of HEK293 cells with different concentrations of SFN with and without TSA, as a HDAC1 inhibitor, leads to the increase in TOPflash reporter activity without affecting b-catenin protein levels. Further studies showed that this increase is due to the decrease in HDAC activity and consequently the increase in histone acetylation following SFN treatment (168).

It has been demonstrated that mamosphere formation in breast cancer cells is dependent on E-cadherin expression (168). It is showed that SFN could target breast cancer stem cells. The mammosphere formation test on two cancer cell lines, MCF7 and SUM195, indicated that SFN could reduce the proportion of cell with stem cell properties, and this was further supported by ALDEFLUOR assay. In vivo examination results of SFN effects on xenograft SUM159 cells in NOD/SCID mice were consistent with the in vitro results. More importantly, cells derived from SFN-treated primary tumors could not produce secondary tumors, while cells derived from the nontreated primary tumors rapidly produced the secondary tumors in the contralateral mammary fat pad of the same mice (168).

Aldehyde dehydrogenase activity is a stem cell marker for enriching tumorigenic stem/progenitor cells (169,170). Five mmol/L of SFN led to >80% reduction of ALDH-positive SUM159 cells in vitro, and daily treatment of xenograft of SUM159 tumors with 50 mg/kg of SFN for 2 weeks led to 50% reduction in tumor size through the reduction in ALDH-positive SUM159 cells by 50%, with no effect on body weight (171). ApcMin/C mice consumed SFN in their diet have fewer tumors with lower sizes in comparison with a control group, albeit, immunohistochemical (IHC) staining revealed that the b-catenin expression was not affected by SFN consumption (172).

Furthermore, the effect of SFN treatment on selfrenewal contributing to signaling pathway, Wnt pathway, was examined by analysis of b-catenin and some other downstream genes at mRNA and/or protein levels (171).
Treatment of T24 bladder cancer cells with SFN results in induction of miR-200c expression (173).

Previous studies demonstrated that miR-200c targets the E-cadherin repressors ZEB1 and ZEB2. Ectopic expression of miR-200c resulted in upregulation of E-cadherin in cancer cells (174). Therefore, treatment of T24 cells with SNF led to E-cadherin induction and EMT suppression (173). However, it seems that these results depend upon cell type and treatment conditions. Although clinical trials seem necessary, there is a large body of investigations about anticancer effects of SFN, and the explicit point is that SFN inclusion into the diet promises a safe and confident strategy.

Another active ingredient of broccoli and other cruciferous vegetables is Indole-3-carbinol (I3C) that has anticancer effects too. Meng et al. (175,176) reported despite a somehow prohibiting effect of I3C on cell attachment in vitro, and I3C could also suppress the invasion and motility of cells. The effect of I3C on cellular metastasis was also evaluated by injecting treated cells into the tail vein of mice and tracing surface metastasis in the lung of the sacrificed animal. Their results indicated that I3C treatment reduced the metastatic capability of the cells.

Bioavailability and new biomarkers of cruciferous sprouts consumption

There are epidemiological evidences of the benefit of consuming cruciferous foods on the reduction of cancer risk (Royston & Tollefsbol, 2015), degenerative diseases (Tarozzi, et al., 2013) and the modulation of obesity-related metabolic disorders (Zhang, et al., 2016), after cruciferous intake. Cruciferous sprouts are especially rich in bioactive compounds compared to the adult plants, due to their young physiological state, being an excellent choice for consuming healthy fresh vegetables (Pérez-Balibrea, Moreno, & García-Viguera, 2011; Vale, Santos, Brito, Fernandes, Rosa, & Oliveira, 2015). The highest benefit of cruciferous foods occurs when are consumed fresh, as young sprouts, avoiding degradation of the enzyme myrosinase by cooking, which is necessary to hydrolyse their characteristic sulphur and nitrogen compounds, the glucosinolates (GLS), to the bioactive isothiocyanates (ITC) and indoles. In case of sprouts, the degradation of GLS after consumption occurs during chewing, in presence of the plant’s enzyme myrosinase, and also is mediated by β-thioglucosidases in the gut microbiota (Angelino & Jeffery, 2014).

There is growing evidences that ITC, such as sulforaphane (SFN) and sulforaphene (SFE), as well as the indole-3-carbinol (I3C), play antioxidant, anti-inflammatory and multi-faceted anticarcinogenic activities in cells (Stefanson & Bakovic, 2014), through the in vivo inhibition of the activation of the central factor of inflammation NF-κB (Egner, et al., 2011), and the induction of the Keap1/Nrf2/ARE pathway related with antioxidant genes and detoxifying enzymes, such as glutathione S-transferases (GST) (Baenas, Silván, Medina, de Pascual-Teresa, García-Viguera, & Moreno, 2015; Myzak, Tong, Dashwood, Dashwood, & Ho, 2007), and also blocking carcinogenic stages in vitro and in vivo by induction of apoptosis, cell cycle arrest and inhibition of histone deacetylases, among others finally, metabolised in the liver with N-acetyl-L-cysteine (-NAC). During the last years, some conjugated ITC, such as SFN-NAC, and other secondary compounds, such as 3,3’diindolylmethane (DIM), which is released by I3C in acid medium (i.e. the stomach), have been used as biomarkers of cruciferous intake (Angelino & Jeffery, 2014; Fujioka, et al., 2016a). However, the bioavailability of the GLS glucoraphenin (GRE) and its isothiocyanate SFE, from radish sprouts, which only differ from SFN in a double bond between the third and fourth carbon (see Figure 2 in section 3), have not been yet investigated.

To our concern, there are no publications studying the bioavailability of radish sprouts compounds, and it is unknown if SFE is metabolised also by the mercapturic acid pathway. Furthermore, there are no commercially available conjugated metabolites of SFE, which would be needed for study its bioavailability by a rapid and sensitive UHPLC-QqQ-MS/MS method, stablishing their appropriate ionization conditions and MRM transitions.

On the other hand, the presence of SFN has been described in radish (Pocasap, Weerapreeyakul, & Barusrux, 2013), maybe by the hydrolysis of GRE, or through a modification of SFE once formed to SFN, suggesting its possible transformation by the mercapturic acid pathway. Therefore, the aim of this study was to evaluate and compare the bioavailability and metabolism of GRA and GRE, from broccoli and radish sprouts, respectively, and the study of possible different biomarker profiles of broccoli and radish consumption for the first time. Also the urine profile evolution of ITC, indoles and conjugated metabolites, after consumption of both broccoli and radish sprouts, were evaluated in a 7 days-cross-over trial with 14 healthy adult women.

2. Material and methods
2.1. Plant material
Broccoli (Brassica oleracea var. italica) and radish (Raphanus sativus cv. Rambo) 8-day-old sprouts were supplied by Aquaporins & Ingredients, S.L. (Murcia, Spain). These sprouts were bioestimulated during production (4 days previous to delivery) with the natural compound methyl jasmonate 250 μM (Baenas, Villaño, García-Viguera, & Moreno, 2016), in order to obtain cruciferous sprouts up to 2-fold richer in bioactive compounds. Three trays of sprouts

(n=3) were collected once a week during the study, then, samples were flash frozen and lyophilised prior analysis of GLS and ITC (see 2.3. section), through an hydro-methanolic (Baenas, Garcia-Viguera, & Moreno, 2014) and aqueous extraction (Cramer & Jeffery, 2011), respectively, as previously described.

2.2. Human subjects and study design. 

A total of 14 women, aged 27-36 years, non-smokers with stable food habits and not receiving medication, during the experimental procedure, were selected to participate in the study. Due to the sex-related disparities in pharmacokinetics and bioavailability results reported by different authors (Soldin, Chung, & Mattison, 2011; Soldin & Mattison, 2009), the choice of a single genus was chosen to avoid a high dispersion in the data, and the availability of healthy young adult female volunteers for the study. Written informed consent was obtained from all subjects. The present study was conducted according to guidelines and procedures approved by the CSIC Committee of Bioethics for the AGL-2013-46247-P project. Subjects were randomly assigned to a seven-by-seven cross-over design (Figure 1), one group receiving broccoli sprouts and the second receiving radish sprouts. Nobody dropped out of the study. A list of foods containing glucosinolates was given to all the participants in order to avoid consumption during the study. Experimental doses (7 trays of broccoli or radish sprouts of 20 grams each) were given at once, on Friday. Subjects were instructed to ingest 1 tray per day, at 10 a.m., according to the cross- over design and to keep trays refrigerated (4 °C) at home. The first day of the study, the urine samples were collected from 0 to 12 h, and from 12 to 24 h after ingestion. From day 2 to 7, the urine samples were collected in 24 h periods. All urine samples were kept refrigerated during collection and were frozen upon reception in the laboratory.

2.3. Metabolites analysis 

The quantitative analysis of GLS in sprouts was carried out by HPLC-DAD 1260 Infinity Series (Agilent Technologies, Waldbronn, Germany), according to UV spectra, and order of elution already described for similar acquisition conditions (Baenas, et al., 2014). For samples preparation, briefly, freeze-dried sprouts (50 mg) were extracted with 1mL of methanol 70% V/V in a US bath for 10 min, then heated at 70°C for 30 min in a heating bath and centrifuged (17500 xg, 15min, 4°C). Supernatants were collected, and methanol was completely removed using a rotary evaporator. The dry material obtained was dissolved in 1 mL of ultrapure water and filtered through a 0.22 μm Millex-HV13 filter (Millipore, Billerica, MA, USA). Measurement of metabolites in sprouts and urine (GRA, SFN, SFN-GSH, SFN-CYS, SFN- NAC) was performed following their MRM transition by a rapid, sensitive and high throughput UHPLC-QqQ-MS/MS (Agilent Technologies, Waldbronn, Germany) method, with modifications of the protocol of Dominguez-Perles et al. (2014), for the optimization of new compounds: GRE, SFE, glucobrassicin (GB), I3C, DIM; assigning their retention times, MS fragmentation energy parameters and preferential transitions (Supplemental File 1). The urine samples were centrifuged (11,000g, 5 min) and the supernatants (400 μL) were extracted using SPE Strata-X cartridges (33u Polymeric Strong Cation), following the manufacturer’s instructions (Phenomenex, Inc., Madrid, Spain), and the slight modifications of Dominguez- Perles et al., (2014). Briefly, the cartridges were conditioned and then aspirated until dryness. The target analytes were eluted with 1 mL of MeOH/formic acid (98:2, v/v) and dried completely using a SpeedVac concentrator (Savant SPD121P; Thermo Scientific, Waltham, MA). The extracts were reconstituted with 200 μL of mobile phase solvent A/B (90:10, v/v) previously to their UHPLC/MS/MS analyses. GRA and GRE were obtained from Phytoplan (Diehm & Neuberger GmbH, Heidelberg, Germany) and ITC and indoles were obtained from Santa Cruz Biotechnology (Santa Cruz, CA, USA). No commercially available conjugated metabolites of SFE from radish were found.

2.4. Statistical analysis 

All assays were conducted in triplicate. Data were processed using SPSS 15.0 software package (LEAD Technologies Inc., Chicago, IL, USA). First of all, data were tested by Shapiro-wilk normality test, as these values do not follow a normal distribution (non-parametric data), statistical differences were determined by the Wilcoxon signed-rank test when comparing two samples and by the Friedman test when comparing multiple samples. Values of P<0.05 were considered significant.

3. Results and discussion
3.1. Bioactive compounds present in cruciferous sprouts
Broccoli and radish sprouts from each week of the study were characterised in GLS and ITC (Table 1). Results are presented as commercial serving dose (20 grams of fresh weight (F.W.)). The amount of cruciferous sprouts consumed daily by the participants was considered a normal serving according to the EFSA (2009), which consider that there is no a perfect way of measuring habitual intake, however, the portion size should be convenient for the use in the context of regular dietary habits.
In broccoli sprouts, glucoraphanin (GRA), in the hydromethanolic extract, and its hydrolysis compound sulforaphane (SFN), in the aqueous extract were the predominant compounds, according to previous findings (Angelino & Jeffery, 2014; Cramer & Jeffery, 2011).
Results showed that radish sprouts presented glucoraphenin (GRE) and glucoraphasatin (GPH) as predominant GLS (Table 1), which were hydrolysed to sulforaphene (SFE) and raphasatin (RPS), respectively. Only SFE was detected in the aqueous extract, as RPS is very unstable and rapidly degraded to less bioactive compounds in aqueous media, such as raphanusanins, E- and Z-3-methylsulfanylmethylene-2-pyrrolidinethiones and E-4-methylsulfanyl-3- butenyldithiocarbamate (Kim, Kim, & Lim, 2015; Montaut, Barillari, Iori, & Rollin, 2010). GRA was not detected in radish sprouts, but the hydrolysis product SFN was present in the aqueous extract (Pocasap, et al., 2013). This could be due to the formation of SFN derived from SFE, losing its double bond, or directly hydrolysed from GRE (Figure 2). Forthcoming evaluations of GRE hydrolysis under different conditions would provide more information about its possible transformations.

3.2. Bioavailability and metabolism of GLS/ITC 

After ingestion of a serving portion of broccoli sprouts (20 g F.W.), GRA (64 μmol) was hydrolysed, absorbed and metabolised, through the mercapturic acid pathway, by a 12 % on average. SFN and its conjugated metabolites, with glutathione (-GSH), cysteine (-CYS) and N- acetyl-L-cysteine (-NAC) (Angelino & Jeffery, 2014), were found in urine (~7.6 μmol/ 24 h as the sum of SFN, SFN-GSH, SFN-CYS and SFN-NAC) (Figure 3A), considered markers of bioavailability.

In the case of radish sprouts, the metabolism of SFE, the predominant ITC, has not been described to the present date, and there are not any conjugated SFE metabolites commercially available to evaluate its bioavailability, by optimization of MRM-transitions in UHPLC-QqQ- MS/MS system. Therefore, it was hypothesised that conjugated SFN metabolites could be found also in urine after radish sprouts consumption, being also possible biomarkers of intake.

Results showed that GRE (61 μmol in 20 g F.W.) was metabolised in SFE, SFN and SFN metabolites (SFN-NAC, SFN-CYS, SFN-GSH), corresponding to 8% on average (~4.9 μmol/24 h) of the GRE consumed (Figure 3B). Therefore, SFN metabolites (SFN-NAC, SFN- CYS, SFN-GSH) could act also as biomarkers of radish consumption. In this sense, this analysis method allowed us to evaluate radish compounds bioavailability, in addition to differentiate it from the bioactives in broccoli, as well as other cruciferous foods, finding intact SFE metabolite as characteristic biomarker of radish consumption.

The values of bioavailability ranged from 9 to 100 % according to different GLS/ITC profiles of the cruciferous vegetables administered, and the consumption as raw or cooked foods and the influence of the microbiota (Shapiro, Fahey, Wade, Stephenson, & Talalay, 1998; Vermeulen, Klopping-Ketelaars, van den Berg, & V aes, 2008). In our case, after broccoli sprouts consumption, the SFN-NAC was the predominant metabolite found in urine (~80 %), followed by SFN-CYS (~11 %), SFN (7.5 %), and SFN-GSH (~0.9 %) (Figure 3A), as previously found (Clarke, et al., 2011; Dominguez-Perles, et al., 2014). In the case of radish sprouts, the SFE was excreted in higher amounts (~65 %) (Figure 3B), followed by the conjugated metabolites: SFN- NAC (~19 %), SFN-CYS (~4%), SFN (~1.1%) and SFN-GSH (~0.7%). SFN and SFE present the isothiocyanate group (−N=C=S), which central carbon is highly electrophilic and actively interacts with cellular nucleophilic targets; such as, the GSH and/or cysteine residues (Kim, Kim, & Lim, 2010). Little information is available about SFE bioactivity and, only recently, Byun, et al., (2016) showed that SFE reduced the cellular GSH levels in vitro, which could indicate its conjugation with GSH. Thus, the synthesis of SFE conjugates with GSH, CYS or NAC could help to generate knowledge on the metabolism of this compound.

On the other hand, the higher excretion of pure SFE after radish consumption compared to pure SFN after broccoli ingestion, may suggest that this compound could follow a different transformation pathway. Contrary to SFN, SFE contains a double bond between the third and fourth carbon, which could result in a decrease in the electrophilicity of the –N=C=S group (Kim, et al., 2010), and different excretion kinetics and transformation.

According to Holst & Williamson (2004), human studies about Phase I metabolism may contribute considerably to understand the biotransformation of ITC and, consequently, the limit of their bioavailability and health-promoting effects. Therefore, further studies about SFE metabolism and bioactivity are needed to support the health-promoting activities of SFE, now insufficiently studied. For instance, Myzak, et al, (2006) showed differences in the bioactivity of SFN conjugated metabolites, as being SFN-CYS and SFN-NAC, significantly active as HDAC inhibitors, with cancer therapeutic potential in vitro and in vivo, but not for SFN and SFN-GSH (Myzak, Ho, & Dashwood, 2006). On the other hand, pure SFN and SFE have shown bioactivities, inhibiting growth of several colon cancer cells (Byun, et al., 2016). Therefore, whether SFE is metabolised by the mercapturic acid pathway or acting in the cell without transformation, this ITC may provide health-promoting effects through induction of detoxification enzymes and antioxidant activity, which did not appear to be affected by their hydrophilicity or other structural factors (La Marca, et al., 2012).

3.3. Urine profile evolution of SFN, SFE and their metabolites. 

The values of individual metabolites excreted, analysed in urine samples, were represented according to two different criteria: 1) the levels of excretion during 24 h were normalised, first from 0 to 12 h after consumption and then from 12 to 24 h, using creatinine as an index to which refers the results, since the creatinine excretion is a relatively constant value between subjects; 2) when comparing the daily excretion during 7 days, the data were represented in volumes collected every 24 h.

The excretion of SFN and its conjugated metabolites, as well as SFE after radish sprouts ingestion was higher during the first 12 h after consumption of sprouts (Figure 4). It has been described that urinary excretion of SFN metabolites after consumption of fresh broccoli reaches peak concentration 3-6 h after consumption, but it could be delayed until 6-12 h (Vermeulen, et al., 2008). After broccoli sprouts ingestion, non-significant differences were found (Figure 4A) between the two periods. In contrast, significant differences were shown in the values of excretion after radish sprouts consumption in all metabolites except for SFN-GSH and SFN- CYS (Figure 4B). The delayed excretion of metabolites, after broccoli sprouts consumption, could be related to a saturation of the membrane transporters (such as P-glycoprotein) in the cells, as SFN is conjugated with GSH and cysteinylglycine in the cells and exported after protein binding, being available for metabolism and excretion (Hanlon, Coldham, Gielbert, Sauer, & Ioannides, 2009). SFE is mostly excreted during the first 12 h after ingestion of radish sprouts (Figure 4B), suggesting that it might be not subjected to the mercapturic acid metabolism. Nevertheless, the biological activity of the pure ITC has shown to be similar than the N -acetylcysteine conjugates (Tang, Li, Song, & Zhang, 2006), being of great interest either if are metabolised or not.

The levels of excretion after 7 days of ingestion (Supplemental files 2 and 3) were also studied and no differences were found in the median daily excretion of SFE, SFN and its metabolites in both broccoli and radish studies. This suggests that repetitive dosing of sprouts should not produce accumulation of any metabolite in the body, as any factor that increases the metabolites amount in the body will cause a decrease in its excretion (Hanlon, et al., 2009). Also, there is a high interindividual variation in excretion values related to human bioavailability studies, which could be explained by different factors, such as the intensity of chewing, where myrosinase enzymes come into contact with intact GLS, gastric pH, intestinal transporters and the activity of the microbiota, where one subject could metabolize three times more GLS into ITC than another, and also the polymorphisms of GST enzymes may affect ITC metabolism (Clarke, et al., 2011; Egner, et al., 2011; Fujioka, Fritz, Upadhyaya, Kassie, & Hecht, 2016b). Low amounts of intact GRA and GRE, on average 0.011 and 0.04 μmol/24 h, respectively, were also recovered in the urine.

One of several challenges in the design of clinical trials is the selection of the appropriate dosage. In this work, commercial trays of sprouts were used, facilitated and quality certified by the company. The average amount of sprouts (~ 20 g per tray) was chosen as one serving, representing a realistic dietary supply. Different specific dietary intervention studies have estimated that the consumption of 3-5 servings per week of cruciferous foods (broccoli, red cabbage, Brussels sprouts, among others) may produce upregulation of detoxification enzymes, responsible for clearance of chemical carcinogens and ROS (Jeffery & Araya, 2009). Therefore, the daily consumption of cruciferous sprouts may result in potential effects decreasing the risk for cancer, even though further epidemiological trials and in vivo studies testing broccoli and radish sprouts are necessary to further understand these effects.

3.4. Bioavailability, metabolism and urine profile evolution of GLS/indoles 

Glucobrassicin (GB) present in broccoli and radish sprouts is an indole GLS derived from tryptophan and releases bioactive indole-3-carbinol (I3C) upon hydrolysis. This bioactive compound requires acid modification in the stomach to form 3,3’-diindolylmethane (DIM) and other condensates to optimize activity, increasing levels of Phase II enzymes, related to detoxification against lung, colon and prostate cancers (Egner, et al., 2011), and the antiproliferative effects on estrogenic-sensitive tumours (Fujioka, et al., 2016a). In particular, DIM has been associated with the suppression of epigenetic alterations related to carcinogenesis, by suppression of DNA methylation and aberrant histone modifications (Fujioka, et al., 2016b).

Additionally, the induction of Phase I enzymes, including the CYP 1 family, catalysing the oxidation of xenobiotics may also be responsible of the action (Ebert, Seidel, & Lampen, 2005; Watson, Beaver, Williams, Dashwood, & Emily, 2013).
Because of the rapid hydrolysis of I3C to DIM in vivo, high stability of DIM, and the strong correlation between GB intake and the amount of DIM excreted, this compound has been described as a biomarker of cruciferous vegetables consumption (Fujioka, et al., 2014). Other condensated compounds such as indol-[3,2-b]-carbazole and related oligomers, were in non- quantifiable concentrations in other studies (Reed, et al., 2006).

Little is known about the bioavailability of other indole GLS present in broccoli and radish sprouts, such as hydroxyglucobrassicin (HGB), methoxyglucobrassicin (METGB) and neoglucobrassicin (NEOGB), which might be also hydrolysed leading indolyl-3-methyl isothiocyanates, unstable and hydrolysed to their corresponding carbinols (Agerbirk, De Vos, Kim, & Jander, 2008; Hanley & Parsley, 1990).

Additional studies are required to confirm if these indole GLS could be hydrolysed also in I3C and DIM, as well as to evaluate the possible health-promoting effects of their hydrolysis and condensate metabolites.
Results demonstrated that broccoli and radish sprouts content in GB were ~11.4 and ~7.7 μmol/20 g F.W, respectively. After ingestion of broccoli sprouts, 49 % of GB was suitably metabolised and excreted as hydrolysis metabolites, calculated as the sum of I3C and DIM (~5.57 μmol /24 h). Following radish ingestion, the percentage of GB hydrolysed and absorbed was 38 % (~2.92 μmol /24 h). It is remarkable that the DIM excreted correspond to over 99 % of these total metabolites. These results of bioavailability contrast with the extremely low percentage (< 1 %) of GB excreted as DIM after consumption of Brussels sprouts and cabbage in a previous study (Fujioka, et al., 2014). Nevertheless, results show relevant bioavailability of GB and the successful use of DIM as biomarker of cruciferous intake. Further studies about conversion of other indole GLS to I3C and DIM are needed to know more about bioavailability of these compounds, as there is no information in literature.

When urine samples were collected in two periods after the ingestion of the sprouts, higher values of excretion of I3C from 12 to 24 h than from the first period were detected (Figure 5), although non-statistically different. Regarding excretion values of DIM, no differences among results were found from 0 to 12 h and from 12 to 24 h (Figure 5). Even if a previous study in humans has shown that the majority of DIM was excreted in the first 12 h (Fujioka, et al., 2016b), other authors have detected DIM in plasma at 12 and 24 h post ingestion (Reed, et al., 2006). Therefore, the excretion of this compound might be longer than for the ITC, which were almost totally excreted during the first 12 h. According to these results, in vivo evidences show that I3C condensation products were absorbed, preferentially targeting the liver, and were detected within the first hour in urine. However, the amount increased significantly between 12 and 72 h, implying effects on the xenobiotic metabolism (WHO, 2004).

No statistical differences were found in the 24 h urine excretion values of I3C and DIM after 7- days of consumption of sprouts (Supplemental File 3). The high variability between subjects has been described before within a low dose level of GB consumed (50 μmol), which was considerably higher than in this study (Fujioka, et al., 2016a).

Furthermore, the results showed no accumulation of metabolites after 7 days of intervention, an important result for safe consumption, also proven with hyper-doses of GB (400-500 μmol) in humans (Fujioka, et al., 2016a).
Low amounts of intact GB (~0.011 and ~0.04 μmol/24 h, after broccoli and radish ingestion, respectively) were also recovered in the urine, but the biological activities of I3C, administered orally to humans, cannot be attributed to the parental compound but rather to DIM and other oligomeric derivatives (Reed, et al., 2006). Therefore, the evaluation of DIM in urine after broccoli and radish consumption provided a susceptible tool to design future clinical trials.

4. Conclusions 

The measurement of ITC, indoles and conjugated metabolites are useful biomarkers of dietary exposure to cruciferous foods. The SFN-NAC is not the only metabolite present in urine and, as along with DIM, could be used as biomarker of the consumption of cruciferous vegetables.

After ingestion of radish sprouts, SFE, together with SFN-NAC and DIM, could be considered as biomarkers, however, metabolites of SFE are not commercially available yet and, consequently, understudied. Repeated dosing of sprouts does not lead to accumulation or higher urine levels of metabolites over time. Furthermore, human short-term pharmacokinetics (e.g. 48 – 72 h) as well as long-term intervention studies (e.g. 10 – 12- weeks) using different doses, and collecting urine and plasma (several time points), as well as faecal samples (microbial metabolites and potential beneficial changes in the gut/colonic microbiota), are strongly recommended for future research in this area.

NMNAT: It’s an NAD+ synthase, chaperone, AND neuroprotector

Nicotinamide mononucleotide adenylyl transferases (NMNATs) are a family of highly conserved proteins indispensable for cellular homeostasis. NMNATs are classically known for their enzymatic function of catalyzing NAD+ synthesis, but also have gained a reputation as essential neuronal maintenance factors. NMNAT deficiency has been associated with various human diseases with pronounced consequences on neural tissues, underscoring the importance of the neuronal maintenance and protective roles of these proteins.

New mechanistic studies have challenged the role of NMNAT-catalyzed NAD+ production in delaying Wallerian degeneration and have specified new mechanisms of NMNAT’s chaperone function critical for neuronal health. Progress in understanding the regulation of NMNAT has uncovered a neuronal stress response with great therapeutic promise for treating various neurodegenerative conditions under question [3]. Given the essential role for NAD+ in cellular metabolism, it is not surprising that the enzyme is required for the survival of all living organisms, from archaebacteria to humans.

The discovery of the remarkable neuroprotective function of NMNAT pro- teins sparked a burst of investigations on NMNAT in the nervous system. More recently, NMNAT mutations have been identified to cause a severe form of retinal degener- ation and NMNAT deficiency has been associated with complex neurological diseases. In this review, we focus on the function of NMNAT in the nervous system and discuss the recent advances in understanding the regulatory mechanisms of neuronal maintenance that are relevant for neuroprotective therapies against neuro- degenerative conditions. For previously published reviews that focus on the neuroprotective effects of NMNAT, particularly in axon degeneration and injury, we refer readers to [1,4–6].

Genetic links between NMNAT and diseases of the nervous system
The only known monogenetic disease associated with NMNAT proteins is Leber congenital amaurosis (LCA), one of the most common forms of inherited blindness in children. Compound heterozygous or homozygous muta- tions in NMNAT1 cause LCA9 (OMIM 608553), an auto- somal recessive condition characterized by severe early- onset and rapid progression of vision loss and retinal degeneration [7]. To date more than 30 mutations spread across the NMNAT1 gene have been reported, including missense, nonsense, and splicing mutations (Figure 1) [8–11]. Importantly, most LCA9 patients have reported normal physical and mental health, suggesting a specific requirement of NMNAT1 for maintenance of the neural tissue in the retina [9].

So far no human diseases have been shown to be directly caused by mutations in NMNAT2 or NMNAT3, though several studies have signified a putative contribution of NMNAT deficiency to the progression of complex neu- rological diseases. Microarray studies have shown that NMNAT2 mRNA levels are reduced in various neurode- generative diseases, including Alzheimer’s disease (AD), Parkinson’s disease, and Huntington’s disease [12–14]. In AD, NMNAT2 transcript levels negatively correlate with cognitive dysfunction and AD pathology [15 ]. Genome- wide association studies have indicated a SNP (rs952797) located 126 kb downstream of the NMNAT3 gene that is associated with late-onset AD [16]. Although these asso- ciation studies do not demonstrate a causal relationship,



Humans have three NMNAT genes that produce three NMNAT protein isoforms with distinct tissue expression patterns and subcellular localizations [1]. NMNAT1 is ubiquitously expressed and is enriched in the nucleus. NMNAT2 is predominantly expressed in the brain and is localized to the cytosol and enriched in membrane com- partments. NMNAT3 is also widely expressed but is highest in liver, heart, skeletal muscles, and red blood cells [2]. NMNAT3 is reported to have two splice variants, encoding mitochondrial localized FKSG76 and cytosolic NMNAT3v1, though the expression and function of these endogenous protein variants are still they do recommend NMNAT as an important contributor to neuronal health.

Molecular functions of NMNAT necessary for neuronal maintenance and protection 

LCA9 and other neurological disease phenotypes associ- ated with NMNAT deficiency are consistent with its neural maintenance function, yet the neuroprotective capacity of NMNAT had already emerged with the serendipitous discovery of the slow Wallerian degenera- tion mutant (Wlds) mouse [17,18]. In the Wlds mouse, degeneration of the distal axon following axotomy (Wallerian degeneration) is remarkably delayed by a dominant mutation causing overexpression and redistri- bution of Nmnat1 to the cytoplasmic compartment of the axon [4]. The neuroprotective role of NMNAT is effica- cious against not only axonal injury, but also models of various neurodegenerative conditions such as toxic neuropathy, glaucoma, spinocerebellar ataxia, tauopathy, and Huntington’s disease [1,5,19–25]. Furthermore, NMNAT homologues across species, including archae- bacteria, yeast, fly, mouse, and human, exhibit a conserved cytoprotective function, even when expressed in different model organisms [1,19]. While these studies chronicle the potent and conserved protection NMNAT proteins confer against a variety of neuronal insults, the precise mechanism by which NMNAT engenders this protection has been hard to pinpoint.

NMNAT proteins have two distinct functions that can bestow neuronal maintenance and protection: NAD synthase activity and chaperone activity [26]. NMNAT catalyzes the reversible conversion of NMN (nicotin- amide mononucleotide) to NAD+ in the final step of both the de novo biosynthesis and salvage pathways. NAD+ is a vital coenzyme important for metabolism and redox biology, as well as a substrate in multiple signaling processes [27]. Thus, NMNAT enzymes play a straight- forward role in neuronal maintenance via balancing NAD+ consumption with production. A molecular chap- erone function has also been demonstrated by a number of NMNAT proteins tested, including Drosophila Nmnat, mouse Nmnat2, and human NMNAT3 [15,19,28]. In vitro these NMNAT proteins bind to client proteins and prevent thermal-stress induced unfolding [15,19]. Thus, NMNAT chaperones can contribute to neuronal protein homeostasis.

There is still considerable controversy surrounding the mechanism of NMNAT-mediated neuroprotection. The prevailing hypothesis is that NMNAT overexpression pro- vides continuous enzyme activity in injured neurons, thus preventing the consequent decrease in NAD+ and the accumulation of the precursor NMN [6]. However, recent work revealed that NAD+ depletion following axon injury is due to a dramatic increase in consumption by the pro- degenerative SARM1 (sterile alpha and TIR motif-con- taining 1), but that overexpression of cytoplasmic Nmnat1 blocks SARM1-dependent NAD+ consumption without increasing NAD+ synthesis [29 ]. Thus, NMNAT-medi- ated axon protection hinges on its ability to block pro- degenerative SARM1 signaling, but this does not rely on enzymatic conversion of NMN to NAD+. One alternative is that NMNAT provides neuroprotection via an enzyme- independent function, leading to the hypothesis that NMNAT mediated protection is chaperone-dependent. Recently, it has been shown that Nmnat2’s enzyme activity is dispensable for relieving the toxic phosphorylated tau burden in a model of fronto-temporal dementia and parkinsonism 17 (FTDP-17), and that Nmnat2 complexes with the classical HSP90 chaperone, possibly to promote refolding of toxic tau [15 ]. In another study, overexpres- sion of cytoplasmic Nmnat1 partially preserved neuronal function in a model of early-onset FTDP-17 by decreasing insoluble tau aggregates, without altering phosphorylated tau [30]. Similarly, overexpression of yeast homologs of NMNAT, NMA1 and NMA2, suppresses proteotoxicity in yeast models of polyglutamine- and a-synuclein-induced neurodegeneration also by enhancing clearance of mis- folded proteins [22].

Though the precise mechanism may vary by disease model, it is clear that NMNAT over- expression alleviates proteotoxic stress in neurodegenera- tive conditions associated with protein misfolding, consis- tent with a chaperone function. Further supporting the chaperone function of NMNAT is the identification of its endogenous ‘client’ protein in Drosophila synapses, where Nmnat maintains active zone structure by directly interacting with the active zone protein Bruchpilot (BRP) in an activity-dependent manner [28]. Perhaps a source of resistance to this hypothesis is that enzyme- inactive mouse Nmnat proteins fail to protect axons [31– 33], although enzyme-inactive Drosophila Nmnat is suffi- cient to protect against activity-induced retinal degenera- tion, Wallerian degeneration, and axonal degeneration induced by loss of JNK (c-Jun N-terminal kinase); in these conditions NMNAT chaperone-dependent protection is perhaps less implicit since here the client for such a chaperone role is not apparent [34–36]. Continued advances in NMNAT overexpression-mediated protection in neurodegenerative models will be invaluable for identi- fying not only the underlying mechanisms of NMNAT’s protection, but also by extension, the primary triggers of neurodegeneration.

Mouse models of NMNAT1-dependent LCA recapitu- late several aspects of human disease and further delin- eate disease progression; importantly, degeneration occurs after retinal development, is observed first within the photoreceptors, and is only observed in inner retinal neurons and RPE (retinal pigment epithelia) cells in advanced stages [37]. The sensitivity of the retina, there- fore, may reflect a twofold requirement of NMNAT: biosynthesis of NAD+ and chaperone activity for meeting the high metabolic and protein turnover demands of photoreceptors, respectively. Although LCA mutations are widely distributed across the NMNAT1 gene, most NMNAT1 mutant proteins characterized exhibit several common features. NAD+ synthase activity is reduced by most mutations in vitro, yet patient fibroblasts exhibit normal basal NAD+ levels, suggesting that enzyme deficiency is not a primary cause of disease [10,38 ]. Importantly, ectopic expression of several LCA9 NMNAT1 mutants is sufficient to protect cultured neu- rons from Wallerian degeneration, indicating that LCA9- causing mutations do not disrupt the qualitative neuro- protective properties of NMNAT proteins [38 ]. How- ever in vitro, NMNAT1 mutant proteins are more sus- ceptible to stress-induced unfolding compared to wild type protein, which may amplified in vivo to cause disease [38 ]. Interestingly, Drosophila photoreceptors are also particularly vulnerable to loss of Nmnat, but Nmnat’s enzyme activity is completely dispensable for mainte- nance of retinal health and function in this model [34]. So while LCA9 may be primarily attributed to a loss of NMNAT1 protein stability, more studies are needed to determine whether retinal degeneration arises from a consequential loss of NAD+ synthesis and/or loss of the molecular chaperone function (Figure 1).

Regulation of NMNAT during stress and disease
Because of the essential neuronal maintenance role of NMNAT and its capacity for neuroprotection, under- standing the regulation of NMNAT levels becomes key to unlocking its therapeutic potential. It is critical that neurons maintain sufficient levels of NMNAT2; NMNAT2 loss is considered an initiating event in Wal- lerian degeneration and also appears to be consistent in progression of neurodegenerative conditions [39]. At the protein level, NMNAT2 is constitutively degraded by the ubiquitin-proteasome system, the exact players of which are beginning to emerge. Several ubiquitin ligases, including Phr1 (Highwire in Drosophila), Skp1a, and Fbxo45, have been identified, though direct evidence of NMNAT ubiquitination is lacking [40–42]. Genetic interference of these ligases increases NMNAT levels and subsequently enhances neuronal protection. In addi- tion to protein regulation, two functional cAMP-response elements (CREs) have been identified in the Nmnat2 promoter region that influence expression at the tran- scriptional level. In an FTDP-17 tauopathy model, a pathological reduction of phospho-CREB (CRE binding protein) levels and binding with the Nmnat2 promoter resulted in decreased Nmnat2 transcript and protein [24].


In contrast to mammals that use three NMNAT genes to produce three different NMNAT protein isoforms, Dro- sophila has only one Nmnat gene that is alternatively spliced to generate two variants: Nmnat-RA mRNA variant encodes the nuclear Nmnat-PC protein isoform and Nmnat-RB variant encodes the cytosolic Nmnat-PD isoform. In a Drosophila model of spinocerebellar ataxia 1 (SCA1), overexpression of Nmnat-PD significantly improved behavioral and morphological defects by reduc- ing neuronal mutant human Ataxin-1 (hAtx1-[82Q]) aggregates, yet surprisingly, overexpression of Nmnat- PC increased aggregate size and exacerbated neurode- generation [43 ]. Importantly, cytoplasmic targeted Nmnat-PC is also ineffective in relieving hAtx-[82Q] proteotoxicity. This study revealed a functional diver- gence underlying Drosophila Nmnat isoform-specific neuroprotection. Therefore, alternative splicing in Dro- sophila uses one gene to produce two functionally exclu- sive mRNA variants: one housekeeping variant, Nmnat- RA, that is stably expressed under normal conditions, and one stress response variant, Nmnat-RB, that can be quickly induced under stress conditions for neuronal protection (Figure 3). Taken together with the previous finding that the transcription of Drosophila Nmnat pre- mRNA is increased under various stress conditions [44], the observation that stress also drives post-transcriptional alternative splicing to preferentially generate Nmnat-RB sheds light on the complex neuronal stress response to achieve self-protection [43 ].

Regulation of Drosophila Nmnat under neuronal stress may provide insight into regulation of NMNAT2 in mammals in neurodegenerative conditions. Notably, the transcript level of the neuroprotective Drosophila variant Nmnat-RB is reduced in late stages of models of tauopathy and SCA1, however it is increased in the early stages, suggesting an upregulation to combat pathogenic protein aggregation during disease progression that is perhaps overwhelmed at more advanced stages of neuro- nal degeneration and cell death [43 ]. Similarly, although many studies show a decrease of NMNAT2 transcript postmortem in patients with neurodegenerative diseases, there is also evidence of increased levels at the early stages of disease. For example, in mild cognitive impairment and early stages of AD, NMNAT2 mRNA levels are higher in the frontal cortex of patients exhibit- ing high levels of oxidative stress-induced neuronal DNA damage, compared to those with lower levels of DNA damage [45]; at this stage the nervous system is mounting a stress response as HSP90 is also upregulated in patient brains with high oxidative damage. In a genome-wide gene-expression study of another mouse model of FTDP- 17, Nmnat2 transcript is significantly upregulated in the hippocampus during early disease progression, but is downregulated at end stages of disease when neurofibril- lary tangle burden is heavy [46]. Here the early rise of Nmnat2 is likely a proteotoxic stress response, while the late decrease is consistent with observed downregulation of synaptic genes corresponding to synaptic loss and upregulation of inflammatory and apoptotic genes indic- ative of neuronal cell death. Since endogenous upregula- tion of NMNAT under stress conditions seems to be insufficient to maintain neural integrity in the long term, NMNAT and its transcriptional and post-transcriptional regulation can serve as a therapeutic target for pharma- cological intervention to slow the progression of neuro- degenerative diseases.

Concluding remarks

The long life of neurons, up to a century in humans, makes neuronal maintenance an important challenge. It is conceivable that compromised maintenance would result in degeneration in an age-dependent manner. The emer- gence of the chaperone function of NMNAT, an NAD+ synthase, specifically in the nervous system for neuronal maintenance, exemplifies an evolutionarily conserved

strategy of ‘repurposing’ (or ‘moonlighting’) housekeep- ing enzymes. Neurons have developed transcriptional and post-transcriptional regulatory mechanisms to bal- ance the metabolic activity and stress response role of NMNAT. Thus, understanding neuronal requirements for NMNAT the NAD+ synthase and NMNAT the chaperone under various stress and disease conditions, in addition to dissecting the regulatory mechanisms that direct the activity of NMNAT, has huge implications for neuroprotective therpapies. Furthermore, identifying druggable targets to enhance expression and/or reduce protein degradation would be a pharmacologically approachable way to achieve higher NMNAT protein levels and therefore confer neuroprotection.

Nicotinamide mononucleotide inhibits JNK activation to reverse Alzheimer disease.

Study published here


  •   NMN improved behavioral measures of cognitive impairments in AD-Tg mice.
  •   NMN decreased β-amyloid production, amyloid plaque burden, synaptic loss, and inflammatory responses in AD-Tg mice.
  •   NMN reduced JNK activation in AD-Tg mice.
  •   NMN regulated the expression of APP cleavage secretase in AD-Tg mice.


Amyloid-β (Aβ) oligomers have been accepted as major neurotoxic agents in the therapy of Alzheimer’s disease (AD). It has been shown that the activity of nicotinamide adenine dinucleotide (NAD+) is related with the decline of Aβ toxicity in AD. Nicotinamide mononucleotide (NMN), the important precursor of NAD+, is produced during the reaction of nicotinamide phosphoribosyl transferase (Nampt). This study aimed to figure out the potential therapeutic effects of NMN and its underlying mechanisms in APPswe/PS1dE9 (AD-Tg) mice. We found that NMN gave rise to a substantial improvement in behavioral measures of cognitive impairments compared to control AD-Tg mice. In addition, NMN treatment significantly decreased β-amyloid production, amyloid plaque burden, synaptic loss, and inflammatory responses in transgenic animals. Mechanistically, NMN effectively controlled JNK activation. Furthermore, NMN potently progressed nonamyloidogenic amyloid precursor protein (APP) and suppressed amyloidogenic APP by mediating the expression of APP cleavage secretase in AD- Tg mice. Based on our findings, it was suggested that NMN substantially decreases multiple AD- associated pathological characteristically at least partially by the inhibition of JNK activation.


As a chronic neurodegenerative disorder, Alzheimer’s disease (AD) is clinically featured by progressive pattern of cognitive deficits and memory impairment. Disturbed energy metabolism in the brain and oxidative stress are two potential factors leading to neural degeneration and cognitive impairments [1]. Aβ oligomers are found to be associated with the pathology of AD [2]. Recent studies indicates that Aβ oligomers inhibit synaptic transmission prior to neuronal cell death [3] and LTP (long-term potentiation), an experimental model for synaptic plasticity and memory [4]. In addition, Aβ oligomers are also found to be relevant to the producing of the free oxygen radical. So far, there is no curative treatment for AD [5]. Considering the varied and well-defined pathologies of AD, new therapies with the functions of reducing pathologies are needed to prevent or slow disease progression.

Nicotinamide adenine dinucleotide (NAD), oxidized (NAD+) or reduced (NADH), plays a key role in many metabolic reactions, for both forms of NAD regulate transfer of hydrogens metabolic reactions, oxidative or reductive [6], as well as mitochondrial morphological dynamics in brain [7]. Among these two forms, oxidized NAD is particularly important to mitochondrial enzyme reactions and cellular energy metabolism [8, 9]. In normal conditions, as people ages, the level of NAD+ drops [6], inhibiting cellular respiration and further causing decreased mitochondrial ATP and possibly cellular death. NAD+ serves a substrate for enzymes that depend on NAD+, such as ADP-ribosyl cyclase (CD38), poly(ADP-ribose) polymerase 1 (PARP1), and Sirtuin 1 (SIRT1) [10].

To treat neurodegenerative diseases, NAD+ depletion and cellular energy deficits need to be prevented for protecting nerves [10]. There are four pathways synthesizing NAD+ in mammals. The salvage pathway (primary route) way is to use nicotinamide, nicotinic acid, nicotinamide riboside, or the de novo pathway with tryptophan [11]. As an essential precursor of NAD+, Nicotinamide mononucleotide (NMN) is produced during the reaction of nicotinamide phosphoribosyltransferase (Nampt). Nampt is essential to regulating NAD+ synthesis [12], for it stimulates phosphoribosyl components to separate from phosphoribosyl pyrophosphates and to combine with nicotinamides. In this way, NMN is generated and with NMN adenylyltransferase, NMN is converted to NAD+. However, the potential therapeutic effects of NMN on AD remain unclear.

c-Jun N-terminal kinases (JNKs) are a family of protein kinases that play a central role in stress signaling pathways implicated in gene expression, neuronal plasticity, regeneration, cell death, and regulation of cellular senescence [13]. Activation of JNK has been identified as a key element responsible for the regulation of apoptosis signals and therefore, it is critical for pathological cell death associated with neurodegenerative diseases and, among them, with Alzheimer’s disease (AD) [14].

As suggested, NAD+ may be essential to brain metabolism and might influence memory and learning. According to recent studies, the stimulation of NAD level is relevant to the reduced amyloid toxicity in AD animal models [15]. Therefore, in this study, the potential therapeutic effects of NMN and the mechanisms of its action regulated in JNK in APPswe/PS1dE9 mice with AD were investigated.

Materials and Methods


The Institutional Animal Experiment Committee of Tongji University, China, approved all procedures conforming to the Animals’ Use and Care Policies. APPswe/PS1dE9 transgenic mice (6 months old) were purchased from Beijing Bio-technology, China. All animals were maintained in an environment that was pathogen-free. During the experimental period, water and food were accessible to all mice, and the body weight of mice and the intake of food and water were identified at the beginning of the study and then on a weekly basis. In addition, all mice that receive the treatment were observed for their general health. APPswe/PS1dE9 transgenic mice (AD-Tg) and their nontransgenic wild-type mice (NTG) were randomly assigned into four groups with six mice in each group, and each type was treated by NMN and vehicle, respectively Subcutaneous adiministration of NMN (100 mg/kg, Sigma N3501) in sterile (Phosphate Buffered Saline) PBS (200 μl) was applied to each mouse of NMN-treated groups every other day for 28 days. Each mouse with vehicle treatment subcutaneously received sterile PBS (200 μl) every other day for 28 days.

Behavioral Tests

Behavioral tests were carried out by 2 experimenters who were blinded to the treatments twelve weeks after the treatments.

Memory and spatial learning test

To evaluate the memory and spatial learning of all animals, a Morris water navigation task was performed as described previously [16]. Generally, a tracking system (Water 2020; HVS Image, Hampton, UK) was utilized to monitor the trajectory of all mice. During the training trials, a platform with the diameter of 5cm was hidden 1.5 cm below the surface of water and maintained at the same quadrant. In every trial, all mice had at most 1 minute to find the hidden platform and climb onto it. If one mouse cannot find the platform within 1 minute, the experimenters would manually guide the mouse to the platform and kept it there for 10 seconds. The trial was carried out 4 times daily for 6 days. The escape latency referring to the time that a mouse spent in finding the platform is considered as spatial learning score. Following the last training trial, the probe trial was carried out for spatial memories by allowing animals to take a free swim in the pool with the platform removed for 1 minute (swim speeds are equal). The time that each animal took to reach the previously platform-contained quadrant was measured for spatial memories.

Measurement of Passive Avoidance

To assess contextual memories, passive avoidance test was carried out, which was described in the previous studies [17]. Briefly speaking, a two-compartment apparatus with one brightly lit and one dimly lit was used. During the training trial, the animal was put into the light lit compartment. After 60 seconds, the door between the two compartments was opened. The acquisition latency refers to the first latency time of mice to ran into the dimly lit compartment. After coming into the lit compartment, mice were exposed to a mild foot shock (0.3mA) for 3 seconds with the door closed. After 5 seconds, the animals were taken out of the compartment. One day later after the acquisition trial, the mice underwent a retention test. Like in the acquisition test, the latency to go into the dark compartment without foot shock was regarded as retention latency to test retention memory. Longer latency indicates better retention.

Tissue Preparation

Following the two behavioral tests, 24 mice were first anesthetized and then infused with icy normal saline in a transcardial way. The brains were taken out and cut into 2 hemibrains along the midsagittal plane. One of the hemispheres was kept in PBS with 4% paraformaldehyde. Following the xylene treatment, the other fixed hemisphere was maintained in the paraffin for immunohistochemical tests. Then the cerebral cortex and the hippocampus were separated quickly from the hemisphere on the ice. For biochemical tests, they were maintained at −80°C following the separation. The hippocampus, brain cortex, and as well as the whole brain were weighed, respectively.


Immunohistochemical staining was carried out as described [16]. Briefly speaking, 10 μm brain slices were deparaffinized and rehydrated. To retrieve antigens, proteinase K (200μg/ml) was treated for the staining of Aβ, and sodium citrate (0.01M, pH 6.0) was for the staining of microglia and astrocyte. Sections were blocked through incubation with fetal bovine serum (2%) and Triton X-100 (0.1%) for nonspecific binding. For immunohistochemical analysis, the section was incubated at 4°C for a night with anti-Aβ1-16 monoclonal antibody (1:600; Cell Signaling Technology, Massachusetts, USA) and monoclonal antibody anti-Iba1 (1:1,000; Osaka Wako Pure Chemical Industries, Japan) for rabbits and also monoclonal anti-Aβ antibody (1:200; Billerica, MA) for mouse.

Olympus (Tokyo, Japan) microscope with a connection to a digital microscope camera was applied to capture the images for quantitative analyses. The plaques in μm2s and the proportion of area kept by plaques positive to Aβ1–16 respectively, microglia positive to Iba1 were obtained with imaging software (Bethesda Media Cybernetics, MD). The mean value of every parameter was obtained from 6 sections with an equidistant interval of 150μm through the hippocampal region of each mouse in all groups. All measurements were blindedly conducted.

Enzyme-Linked Immunosorbent Assay (ELISA)

As described before, soluble Aβ fractions and insoluble ones were obtained from both the cortex and hippocampi of brain homogenates of mouse using RIPA (Radioimmunoprecipitation assay buffer) buffer and formic acid, respectively [18]. The levels of both the insoluble and soluble Aβ were identified using the ELISA kits (Camarillo Invitrogen, CA). Besides, concentrations of oligomeric Aβ of brain homogenates treated with RIPA were obtained employing an ELISA kit for amyloid β oligomer (Gunma Immuno-Biochemical Laboratories, Japan).

Proinflammatory Cytokines Measurement

As described, mouse brain proinflammatory cytokine was evaluated [19]. The expressions of TNFα, IL-6, and IL-1β were identified with immunoassay kits (Minneapolis R&D Systems, Minnesota, USA) which is for measuring these factors in mouse.

Western blotting (WB) analysis

The cortex and hippocampus tissue was homogenized with icy PBS and the lysate was for Western blot. At first SDS-PAGE (Sodium Dodecyl Sulfate Polyacrylamide Gel Electrophoresis) was applied to divide the proteins on NuPage Bis-Tris gel (12%, Invitrogen). The separated protein was subsequently transferred to nitrocellulose membrane which was blocked with 5% nonfat milk and probed overnight at 4°C with anti-p-JNK, anti-JNK, anti-APP, anti-sAPPα, anti- sAPPβ, anti-phosphorylated APP (p-APP, Thr668), anti-ADAM10, anti-BACE1 (CA Santa

Cruz Biotechnology, USA), anti-CDK5, anti–p-CDK5, anti–p-GSK3β, anti-GSK3β, anti-SYP, anti–postsynaptic density-95, anti–β-actin Abcam (Cambridge, MA, USA). The membrane was cleaned with TBS/0.05% Tween-20 and incubated at room temperature with secondary antibodies conjugated with horseradish peroxidase for 60 minutes, following incubation with primary antibodies. Enhanced chemiluminescence reagents (Pierce, Rockford, IL, USA) were used for detecting signals.

Statistical Analysis

The data were expressed as mean ± SD. The comparisons in the speed of swimming and escape latency between the groups during the test of memory and spatial learning were made employing two-way ANOVA with repeated measures. Next, post hoc least significant difference (LSD) test was used for multiple comparison. Before post hoc LSD test or Student t test, 1-way ANOVA was employed for the rest data. Statistical analyses were carried out with Prism version 5. A P < 0.05 was considered statistically significant.


NMN Treatment Rescues Cognitive impairments in AD-Tg Mice

The test of memory and spatial learning has shown that 1-year-old AD-Tg mice have experienced impairment in memory and spatial learning [16]. In our study, comparied to the vehicle-/NMN-treated wild-type (WT) mice, the vehicle-treated AD-Tg mice had a longer escape latency, which showed severe impairment of spatial learning in the test (Fig. 1A). However, with shorter escape latency, NMN treatment greatly improved the impairment of spatial learning in vehicle-treated AD-Tg animals (Fig. 1A). Besides, it was also identified that compared with two wild-type groups, vehicle-treated AD-Tg animals spent less time in the target quadrant during the probe trial (p < 0.01), suggesting severe spatial memory impairment. AD-Tg mice treated with NMN spent longer time in the target quadrant, which indicates marked alleviation of the spatial learning impairments present in AD-Tg mice treated with vehicle (p<0.01) (Fig. 1B).

To further identify the alleviation of memory deficits by NMN treatment in AD-Tg mouse, contextual memories were evaluated employing the measurement of passive avoidance [17]. As illustrated in Fig. 1C, retention latency was decreased compared with two wild-type mice groups (p < 0.01), suggesting impaired contextual memories in the AD-Tg animals treated by vehicle. In contrast, NMN-treated mice exhibited longer retention latency compared with those treated with vehicles (p < 0.01), demonstrating outstanding reversal of NMN in contextual memories. All these data indicate that NMN treatment markedly improves cognitive impairments in AD-Tg animals.

NMN Suppresses JNK Phosphorylation in AD-Tg Mice

JNK, also called a protein kinase activated by stress, is said to play a role in a couple of pathophysiological processes in AD [13]. Therefore, in this study, we tested the inhibitory effects of NMN on the activation of JNK through Western blotting. It was revealed by quantitative analysis that p-JNK level was significantly grown in hippocampus and cerebral cortex in the vehicle-treated AD-Tg mice when contrasting to two wild-type groups (Fig. 2A; p < 0.01), whereas NMN gave rise to a sharp decline in p-JNK in hippocampus and cerebral cortex with a comparison to the vehicle-treated AD-Tg mice (Fig. 2B; P< 0.01). Both reductions symbolized a reverse to the wild-type level. But the whole expression of JNK kept unchanged in all the 4 groups. Conclusively, all data indicate that NMN treatment has an inhibitory effects on JNK activation in AD-Tg mice.

NMN Treatment Decreases the Level of Aβ and Deposition in AD-Tg Mice

The role of reduced activation of JNK in the changes of the Aβ level and deposition was studied in AD-Tg mice through employing histological and biochemical analyses. As presented in Figs. 3A-D, it was found that NMN-treated AD-Tg mice had a sharp reduction in the levels of Aβ when comparing to the vehicle treatment group (p < 0.01). Comparing to the vehicle treatment group, NMN treatment gave rise to a marked decrease in Aβ oligomers (p < 0.01) (Fig. 3E). Immunohistochemical staining identified this observation, indicating the lessened diffuse plaques and also the shrinked area taken by diffuse plaques in AD-Tg mice treated by NMN compared to the vehicle treatment group (Figs. 4A-D). Thus, on the basis of the findings, it was demonstrated that the generation of Aβ in the brain of AD-Tg mice is effectively decreased by the inhibited activation of JNK with NMN treatment.

NMN Treatment Changes the Processing of APP in AD-Tg Mice

To study the mechanism of inhibition on the production of Aβ and deposition, the effects of NMN on the processing of APP were examined by Western blotting. As presented in Figs. 5A-C, the level of full-length APP expression was greatly increased in the brain of AD-Tg mouse treated with vehicle compared with wild-type ones (p < 0.01). However, they kept unaltered between the group treated with vehicle and that with NMN. Importantly, it was found that NMN treatment remarkably lowered the increased levels of p-APP in the AD-Tg mice treated by vehicle (p < 0.01). Besides, α-secretase cleaved sAPPα and β-secretase cleaved sAPPβ in the brain tissues of Tg mice were tested via Western blotting. It was shown by quantitative analyses that NMN treatment led to a remarkable elevation of sAPPα (p < 0.01) and a marked decline in sAPPβ (p < 0.01) compared with the transgenic mice treated by vehicle (Figs. 5D-F). Based on these data, it was indicated that NMN treatment is strongly effective in suppressing the phosphorylation of APP, improving cleaving of APP by α-secretase, and decreasing the cleaving of APP by β-secretase in AD-Tg mice brains.

NMN Treatment Improves Inflammatory Responses in AD-Tg Mice

Since JNK activation is indicated to play a role in the inflammatory response induced by Aβ in previous studies [20], whether reduced activation of JNK influences neural inflammation in AD- Tg animals was investigated. The role of NMN on the neural inflammation was identified by measuring proinflammatory cytokines that were in the lysates of cortical tissues. It was found that the level of IL-6, IL-1β, and TNFα were sharply declined in the AD-Tg mice treated by NMN relative to those by vehicle (Figs. 6A-C). According to these findings, NMN treatment is indicated to be potently effective in the amelioration of neural inflammation in AD-Tg mice brains.

NMN Treatment Ameliorates Synaptic Loss in AD-Tg Mice

The loss of synapse is an important pathological characteristic of AD and said to be relevant to the cognitive impairments of AD [21]. The changes in SYP (presynaptic marker) level and PSD- 95 level (postsynaptic marker) were investigated via Western blotting. It was showed by quantitative analysis that SYP levels and the levels of PSD-95 expression substantially reduced in hippocampus and brain cortex of AD-Tg mice treated with vehicle relative to WT ones

(p < 0.01), whereas NMN treatment significantly elevated SYP levels and the levels of PSD-95 expression in hippocampus and brain cortex relative to AD-Tg mice treated with vehicle (p < 0.01) (Fig. 6D-F). This finding suggest that NMN treatment sharply ameliorates the loss of synapse in AD-Tg mice brains.


In the present study, it is mainly found that NMN treatment substantially improves primary pathological characteristics of the AD-modeled AD-Tg mice, including cognitive impairments, neuroinflammation, Aβ pathology, and synaptic loss, which consistent with a recent study [22]. It was also found that NMN treatment inhibited JNK activation and amyloidogenic processing of APP by mediating the expression of APP-cleavage secretase, and also facilitated APP processing in AD-Tg mice. The data prove that NMN treatment greatly reduces multiple AD-associated pathological characteristics, at least partially by the inhibition of JNK activation.

Numerous studies have reported the increase of abnormal activation of JNK in both the transgenic AD mice models and the AD patients [23-25]. Conforming to the above previous studies, we also found that the level of phosphorylated JNK in AD-Tg mice treated by vehicle was higher than that in the wild-type group, but NMN treatment in AD-Tg mice potently suppressed the phosphorylation of JNK to the basic level of WT groups. The controlled activation of JNK through NMN gave rise to a substantial decrease of Aβ pathology in AD-Tg animals. According to the studies before, active JNK is proved to engage in BACE1 expressions and PS1 expressions [26, 27]. In addition, the increased BACE1 and PS1 in AD-Tg mice treated by vehicle were found to be greatly suppressed by NMN to the basic level of WT groups (data not shown). More interestingly, it was also observed that NMN treatment led to substantially elevated sAPPα and reduced sAPPβ. It was notable that according to the previous studies, APP phosphorylation at the site of Thr668 is proved to promote the β-secretase cleavage of APP to grow Aβ generation in vitro [28]. In present study, we also found that the administration of NMN in AD-Tg mice significantly declined the elevated phosphorylation of APP to the primary level of WT controls, indicating an in vivo inhibition mechanism of Aβ pathology through NMN treatment. Collectively, all these findings indicate that the potent effects of NMN on the marked decrease in Aβ pathology in the brains of AD-Tg mice may be responsible for its enhancement of nonamyloidogenic APP processing. What we found is consistent to a recent study demonstrating that genetic depletion of JNK3 in 5XFAD mice is attributed to a significant decrease in the levels of Aβ and the total plaque loads [29]. Recently numerous studies suggested energy failure and accumulative intracellular waste also play a causal role in the pathogenesis of several neurodegenerative disorders and Alzheimer’s disease (AD) in particular regulated by potential role of several metabolic pathways Wnt signaling, 5′ adenosine monophosphate-activated protein kinase (AMPK), mammalian target of rapamycin (mTOR), Sirtuin 1 (Sirt1, silent mating-type information regulator 2 homolog 1), and peroxisome proliferator-activated receptor gamma co- activator 1-α (PGC-1α) [30, 31]. It will be warrant to study if NMN also participate in regulation of these signaling pathways.

Some recent studies indicate that enhanced neuroinflammation is essential to the development of AD [32-34]. In our study, a marked decline in the proinflammatory cytokines levels (IL-6, IL-1β, and TNFα) proved that NMN treatment effectively controlled the neuroinflammatory responses of the brain of AD-Tg mouse. Considering the key role of oligomeric and fibrillar Aβ for activation of microglia cells and astrocytes with the subsequent generation of proinflammatory cytokines [34], the reduction in neuroinflammatory responses may be less important to the substantial reduction in Aβ pathologies presented in the AD-Tg mice treated by NMN. Several previous studies had proved that JNK represents an important mediator for activation of glial cell and proinflammatory cytokines [35, 36]. Thus, the favorable effects of NMN on lowered inflammatory responses in AD-Tg groups can be largely responsible for its direct control of inflammation by inhibiting JNK activation. Based on the previous reports, it was proved that some proinflammatory cytokines (ie, IL-1β, interferon gamma, and TNFα) may elevate the expression of β-secretase and γ-secretase to ameliorate amyloidogenic APP processing and Amyloid-β production by an in vitro JNK-mediated pathway [33]. Hence, we have reasons to believe that the reduced proinflammatory cytokines through NMN treatment may be effective in reducing the production of Aβ in vivo. Moreover, in our study, it was demonstrated that NMN treatment ameliorates cognitive impairments in AD-Tg mouse models. An increasing evidence has proved that grown Aβ levels, neuroinflammation, synaptic dysfunction and loss are closely related to the cognitive dysfunction in AD [37]. In addition, our data confirms the finding of a recent research, which revealed that genetic down-regulation of JNK3 gives rise to a remarkable amelioration of cognitive impairments in 5XFAD mice [29]. Collectively, our findings, along with all the previous research, demonstrate that the inhibited JNK activaty by NMN is potently effective in ameliorating AD-associated cognitive deficits.

Synaptic loss is a major pathological change of AD and is tightly associated with AD-related cognitive impairments [37]. It was presented that PSD-95, a biomarker of postsynaptic density, is essential to synapse maturation and synaptic plasticity [38], and that SYP, a presynaptic protein, also acts as an integral membrane protein in the synapse and it plays a key role in plasticity of synapses [39]. Therefore, it can be soundly supposed that the greatly lowered expression of PSD- 95 and SYN presented in the study may suggest the impairment of synaptic integrity and  plasticity in AD-Tg mice treated by vehicle. Intriguingly, the treatment of NMN in AD-Tg animals substantially elevated the lowered PSD-95 and SYN expression level back to the primary level of WT controls. Since it was demonstrated by several studies that Amyloid-β- induced synaptic loss and dysfunction are regulated through the JNK activation [40, 41], the possible mechanisms behind NMN treatment leading to the elevated expression of PSD-95 and SYN in AD-Tg animals may be responsible for its inhibitory effects on JNK activation. Thus, it is possible that the treatment of NMN may ameliorate the impaired synaptic plasticity which is caused by toxic Aβ species in AD-Tg mice.

In summary, this study provides essential preclinical evidences that NMN takes effects in reversing cognitive deficits and substantially lowering the burden of amyloid plaque, neuroinflammation, cerebral amyloid-β concentrations, and loss of synapse in middle-aged AD- Tg mice, at least partially by the inhibition of JNK activation. According to our findings, NMN could be a new target for disease-modifying treatments of AD.


  1. Kapogiannis, D. and M.P. Mattson, Disrupted energy metabolism and neuronal circuit dysfunction in cognitive impairment and Alzheimer’s disease. Lancet Neurol, 2011. 10(2): p. 187-98.
  2. Gong, Y., et al., Alzheimer’s disease-affected brain: presence of oligomeric A beta ligands (ADDLs) suggests a molecular basis for reversible memory loss. Proc Natl Acad Sci U S A, 2003. 100(18): p. 10417-22.
  3. Hardy, J. and D.J. Selkoe, The amyloid hypothesis of Alzheimer’s disease: progress and problems on the road to therapeutics. Science, 2002. 297(5580): p. 353-6.
  4. Lesne, S., et al., A specific amyloid-beta protein assembly in the brain impairs memory. Nature, 2006. 440(7082): p. 352-7.
  5. Tayeb, H.O., et al., Pharmacotherapies for Alzheimer’s disease: beyond cholinesterase inhibitors. Pharmacol Ther, 2012. 134(1): p. 8-25.
  6. Imai, S. and L. Guarente, NAD+ and sirtuins in aging and disease. Trends Cell Biol, 2014. 24(8): p. 464-71.
  7. Long, A.N., et al., Effect of nicotinamide mononucleotide on brain mitochondrial respiratory deficits in an Alzheimer’s disease-relevant murine model. BMC Neurol, 2015. 15: p. 19.
  8. Kristian, T., et al., Mitochondrial dysfunction and nicotinamide dinucleotide catabolism as mechanisms of cell death and promising targets for neuroprotection. J Neurosci Res, 2011. 89(12): p. 1946-55.
  9. Owens, K., et al., Mitochondrial dysfunction and NAD(+) metabolism alterations in the pathophysiology of acute brain injury. Transl Stroke Res, 2013. 4(6): p. 618-34.
  10. Liu, D., M. Pitta, and M.P. Mattson, Preventing NAD(+) depletion protects neurons against excitotoxicity: bioenergetic effects of mild mitochondrial uncoupling and caloric restriction. Ann N Y Acad Sci, 2008. 1147: p. 275-82.
  11. Yamamoto, T., et al., Nicotinamide mononucleotide, an intermediate of NAD+ synthesis, protects the heart from ischemia and reperfusion. PLoS One, 2014. 9(6): p. e98972.
  12. Imai, S., Dissecting systemic control of metabolism and aging in the NAD World: the importance of SIRT1 and NAMPT-mediated NAD biosynthesis. FEBS Lett, 2011. 585(11): p. 1657-62.
  13. Mehan, S., et al., JNK: a stress-activated protein kinase therapeutic strategies and involvement in Alzheimer’s and various neurodegenerative abnormalities. J Mol Neurosci, 2011. 43(3): p. 376-90.
  14. Yarza, R., et al., c-Jun N-terminal Kinase (JNK) Signaling as a Therapeutic Target for Alzheimer’s Disease. Front Pharmacol, 2015. 6: p. 321.
  15. Kim, D., et al., SIRT1 deacetylase protects against neurodegeneration in models for Alzheimer’s disease and amyotrophic lateral sclerosis. EMBO J, 2007. 26(13): p. 3169- 79.
  16. Zhang, W., et al., Multiple inflammatory pathways are involved in the development and progression of cognitive deficits in APPswe/PS1dE9 mice. Neurobiol Aging, 2012. 33(11): p. 2661-77.
  1. Ishrat, T., et al., Amelioration of cognitive deficits and neurodegeneration by curcumin in rat model of sporadic dementia of Alzheimer’s type (SDAT). Eur Neuropsychopharmacol, 2009. 19(9): p. 636-47.
  2. Li, J.G., et al., Homocysteine exacerbates beta-amyloid pathology, tau pathology, and cognitive deficit in a mouse model of Alzheimer disease with plaques and tangles. Ann Neurol, 2014. 75(6): p. 851-63.
  3. Chu, J. and D. Pratico, Involvement of 5-lipoxygenase activating protein in the amyloidotic phenotype of an Alzheimer’s disease mouse model. J Neuroinflammation, 2012. 9: p. 127.
  4. Vukic, V., et al., Expression of inflammatory genes induced by beta-amyloid peptides in human brain endothelial cells and in Alzheimer’s brain is mediated by the JNK-AP1 signaling pathway. Neurobiol Dis, 2009. 34(1): p. 95-106.
  5. Querfurth, H.W. and F.M. LaFerla, Alzheimer’s disease. N Engl J Med, 2010. 362(4): p. 329-44.
  6. Wang, X., et al., Exendin-4 antagonizes Abeta1-42-induced suppression of long-term potentiation by regulating intracellular calcium homeostasis in rat hippocampal neurons. Brain Res, 2015. 1627: p. 101-8.
  7. Braithwaite, S.P., et al., Inhibition of c-Jun kinase provides neuroprotection in a model of Alzheimer’s disease. Neurobiol Dis, 2010. 39(3): p. 311-7.
  8. Sclip, A., et al., c-Jun N-terminal kinase regulates soluble Abeta oligomers and cognitive impairment in AD mouse model. J Biol Chem, 2011. 286(51): p. 43871-80.
  9. Thakur, A., et al., c-Jun phosphorylation in Alzheimer disease. J Neurosci Res, 2007. 85(8): p. 1668-73.
  10. Guglielmotto, M., et al., Amyloid-beta(4)(2) activates the expression of BACE1 through the JNK pathway. J Alzheimers Dis, 2011. 27(4): p. 871-83.
  11. Rahman, M., et al., Intraperitoneal injection of JNK-specific inhibitor SP600125 inhibits the expression of presenilin-1 and Notch signaling in mouse brain without induction of apoptosis. Brain Res, 2012. 1448: p. 117-28.
  12. Colombo, A., et al., JNK regulates APP cleavage and degradation in a model of Alzheimer’s disease. Neurobiol Dis, 2009. 33(3): p. 518-25.
  13. Yoon, S.O., et al., JNK3 perpetuates metabolic stress induced by Abeta peptides. Neuron, 2012. 75(5): p. 824-37.
  14. Godoy, J.A., et al., Signaling pathway cross talk in Alzheimer’s disease. Cell Commun Signal, 2014. 12: p. 23.
  15. Killick, R., et al., Clusterin regulates beta-amyloid toxicity via Dickkopf-1-driven induction of the wnt-PCP-JNK pathway. Mol Psychiatry, 2014. 19(1): p. 88-98.
  16. Hong, H.S., et al., Interferon gamma stimulates beta-secretase expression and sAPPbeta production in astrocytes. Biochem Biophys Res Commun, 2003. 307(4): p. 922-7.
  17. Liao, Y.F., et al., Tumor necrosis factor-alpha, interleukin-1beta, and interferon-gammastimulate gamma-secretase-mediated cleavage of amyloid precursor protein through aJNK-dependent MAPK pathway. J Biol Chem, 2004. 279(47): p. 49523-32.
  18. Morales, I., et al., Neuroinflammation in the pathogenesis of Alzheimer’s disease. A rational framework for the search of novel therapeutic approaches. Front Cell Neurosci,2014. 8: p. 112.
  19. Kim, S.H., C.J. Smith, and L.J. Van Eldik, Importance of MAPK pathways for microglialpro-inflammatory cytokine IL-1 beta production. Neurobiol Aging, 2004. 25(4): p. 431-9.
  1. Waetzig, V., et al., c-Jun N-terminal kinases (JNKs) mediate pro-inflammatory actions of microglia. Glia, 2005. 50(3): p. 235-46.
  2. Marcello, E., et al., Synaptic dysfunction in Alzheimer’s disease. Adv Exp Med Biol, 2012. 970: p. 573-601.
  3. El-Husseini, A.E., et al., PSD-95 involvement in maturation of excitatory synapses. Science, 2000. 290(5495): p. 1364-8.
  4. Janz, R., et al., Essential roles in synaptic plasticity for synaptogyrin I and synaptophysin I. Neuron, 1999. 24(3): p. 687-700.
  5. Costello, D.A. and C.E. Herron, The role of c-Jun N-terminal kinase in the A beta- mediated impairment of LTP and regulation of synaptic transmission in the hippocampus. Neuropharmacology, 2004. 46(5): p. 655-62.
  6. Sclip, A., et al., c-Jun N-terminal kinase has a key role in Alzheimer disease synaptic dysfunction in vivo. Cell Death Dis, 2014. 5: p. e1019

Mitochondrial dysfunction and the inflammatory response

The .pdf version of this research can be found at

1. Introduction
Inflammation, a basic biological response against invading pathogens, contributes to repairing damage and preventing further tissue or cell injury (Medzhitov, 2008). However, when the inflammatory response cannot be appropriately attenuated, multi-organ failure or chronic inflammation can develop (Medzhitov, 2008).

In this sense, increasing levels of pro-inflammatory cytokines have been associated with disease evolution and unfavorable outcomes in several pathological scenarios, such as chronic heart failure, peritoneal dialysis, pancreatitis, and cancer (Berres et al., 2011; Gravante et al., 2009; Hiller et al., 2011; Nevado et al., 2006; Yndestad et al., 2012).

While each of these disorders has a specific pathogenesis and disease evolution, they all involve inflammatory processes. In addition, the inflammatory process also has a role in mediating the aging process via changes in redox status and oxidative stress-induced inflammatory responses (Gilca et al., 2007); this process has been called inflammaging (Salminen et al., 2012). In fact, aging and low-grade inflammation involves many of the same molecular pathways, including oxidative stress and DNA damage pathways.

The molecular inflammation hypothesis of aging postulates that chronic low-grade inflammation in aging causes cumulative damage to biological macromolecules through the formation of oxygen radicals (Chung et al., 2009).

This process would eventually render cells more susceptible to damage and more sensitive to catabolism, processes associated with an exacerbated response to age-related diseases (Chung et al., 2009; Jian et al., 2011; Loeser, 2009). Hence, the risk of degenerative diseases and poor outcomes of acute processes rise exponentially with age.

This is also supported by studies showing that healthy centenarians exhibit low levels of inflammatory biomarkers (Franceschi, 2007).

Altered mitochondrial function is linked to several acute and chronic inflammatory diseases (Naik and Dixit, 2011). Additionally, mitochondria have long been proposed to play a key role in aging (Green et al., 2011).

As a consequence of their central role in ATP formation via the mitochondrial respiratory chain (MRC), mitochondria are the major source of reactive oxygen species (ROS) and are thus highly involved in oxidative stress processes (Brookes et al., 2004). However, mitochondria are also targets of these molecules (Brookes et al., 2004).

Under physiological conditions, about 1–3% of molecular oxygen is incompletely reduced during redox reactions in the MRC, and this in turn leads to production of the ROS superoxide anion (O2−) as a by-product of these reactions. In this scenario, complex interactions in antioxidant defense systems repress oxidative stress within mitochondria (Kienhöfer et al., 2009).

However, under pathological conditions, an excess of O2− ions are produced, which causes the activation of redox-sensitive transcription factors, such as the key regulator of tissue inflammation, nuclear factor-κB (NF-κB), and a subsequent increase in the expression of cytokines, chemokines, eicosanoids, inducible nitric oxide synthase (iNOS), and adhesion molecules (Chung et al., 2009).

In addition to alteration of signaling pathways regulated by ROS, the excess of ROS also drives oxidation of membrane lipids, proteins, and mitochondrial DNA (mtDNA) (Henze and Martin, 2003).

In particular, mtDNA seems to be especially sensitive to oxidative damage, and therefore susceptible to mutations, because of its close proximity to the site of ROS production and the lack of protective histones. In fact, aged individuals, whose cells have accumulated a high level of oxidative damage, exhibit an increased rate of mutagenesis in their mtDNA (Kujoth et al., 2005).

Moreover, superoxide anions can combine with nitric oxide (NO) to produce peroxynitrite (ONOO-), a powerful oxidant capable of affecting mitochondrial integrity (Brown, 2003; Escames et al., 2011).

Mitochondrial dysfunction drives mitochondrial genome mutagenesis, affecting genes encoding respiratory chain complexes and compromising the efficiency of oxidative phosphorylation (OXPHOS), which may lead to further mtDNA mutations and even mutations in the nuclear genome, thereby causing increased cell damage (Escames et al., 2011).

In this review, we aim to highlight the significance of mitochondrial dysfunction and to discuss the contribution of mitochondria to the development of inflammatory human diseases and the aging process.

2. Mitochondrial dysfunction may modulate inflammatory processes

It is well known that mitochondria are not only the primary source of ATP, but also participate in many other cell signaling events, such as regulation of calcium (Ca2 +) homeostasis, orchestration of apoptosis, and production of ROS, as mentioned above (Brookes et al., 2004).

In this sense, ROS are the major host defense mechanism against infection and harmful agents (Chen et al., 2012), and mitochondria are the principal mediators of inflammation (Tschopp, 2011).

During viral infection the pattern recognition receptors RIG-I and MDA5 attach to viral RNA allowing its interaction with a mitochondrial polypeptide adaptor, MAVS, which finally drives the production of type I interferon (Saitoh and Akira, 2010).

Interestingly, pathogens have recently been shown to have developed their own reciprocal defense systems; for example, Leishmania infection has been shown to cause marked upregulation of uncoupling protein 2 (UCP2), a negative regulator of mitochondrial ROS generation, potentially preventing ROS-mediated inactivation to suppress macrophage defense mechanisms and facilitate parasite survival (Basu Ball et al., 2011).

However, under dysregulated processes, when mitochondria are compromised by damage or mutations, excess O2− ions are produced, and cellular stress cannot be effectively resolved. Thus, mitochondria are the principal arbitrators of the pro-inflammatory status; they act through modulating innate immunity via redox-sensitive inflammatory pathways or direct activation of the inflammasome, a group of protein complexes whose activation results in immediate activation of caspase-1, thereby allowing for cleavage and subsequent activation of the inactive precursors IL-1β and IL-18 (Strowig et al., 2012).

Additionally, both of these pathways, i.e., the redox-sensitive inflammatory pathway and the inflammasome pathway, may work together to activate inflammatory cytokines, leading to an overstimulation of the inflammatory response (Escames et al., 2011).

Taken together, these recent studies have suggested the unexpected supplementary role of mitochondria as drivers of the inflammatory process.

2.1. Activation of redox-sensitive inflammatory pathways by mitochondrial impairment
Cellular systems that protect against oxidants involve antioxidative defense enzymes (superoxide dismutase [SOD], glutathione peroxidase [GPx], and catalase) (Kienhöfer et al., 2009), oxidant scavengers (vitamin E, vitamin C, carotenoids, uric acid, and polyphenols) and mechanisms to repair oxidant-induced damage to lipids, proteins, or DNA (Karger et al., 2012; Ugarte et al., 2010).

Despite these protective mechanisms, uncontrolled generation of ROS can overwhelm the capacity of cellular antioxidant protection causing mitochondrial dysfunction. In this sense, in vivo studies in transgenic mice showed that the overexpression of catalase targeted to mitochondria reduces age-associated pathologies and the mice have an extension in lifespan (Schriner et al., 2005).

Additionally, several studies have demonstrated that mitochondrial dysfunction may generate low-grade inflammatory and matrix degradation responses in several cell types via mitochondrial Ca2 + exchange, ROS generation, and NF-κB activation (Amma et al., 2005; Cillero-Pastor et al., 2008; Ichimura et al., 2003).

Thus, the flow of mitochondrial Ca2 + plays a crucial role in the evolution of cell signals and in the regulation of mitochondrial activity. Moreover, excessive mitochondrial Ca2+ accumulation has been shown to be critical for disease development (Dada and Sznajder, 2011; Tanaka et al., 2004).

For example, increased mitochondrial Ca2+ levels in cardiomyocytes contribute to cardiac inflammation and dysfunction after burn injury or sepsis (Maass et al., 2005).

Additionally, when cells are treated with inhibitors of mitochondrial Ca2 + exchange, ROS levels decrease significantly, leading to subsequent reductions in the levels of pro-inflammatory mediators. The rise in mitochondrial Ca2 +, a consequence of importing cytosolic Ca2 +, augments the mitochondrial production of ROS by different mechanisms, including 1) Ca2+ stimulation of the tricarboxylic acid cycle, which would enhance electron flow into the respiratory chain; 2) Ca2 + stimulation of NO synthase, elevating NO levels, which would in turn inhibit respiration at complex IV and enhance ROS generation; and 3) Ca2+ binding to cardiolipin (Brookes et al., 2004).

Thus, ROS production by mitochondrial electron transport could represent an intermediate step in pro-inflammatory gene expression/NF-κB signaling.
Mitochondrial dysfunction has also been shown to increase the responsiveness of several types of cells to cytokine-induced inflammatory responses, resulting in a significant amplification of the inflammatory response through ROS generation and NF-κB activation (Vaamonde-García et al., 2012; Nakahira et al., 2011).

When mitochondrial dysfunction was induced in normal synovial fibroblasts by an inhibitor of mitochondrial ATP synthase, oligomycin, an otherwise less-efficient concentration of IL-1β was as effective as a 10-times greater concentration of IL-1β would be in the absence of pretreatment with the mitochondrial inhibitor (Valcárcel-Ares et al., 2010).

In lung epithelial cells, pre-existing mitochondrial dysfunction induced by antisense oligonucleotides to complex III increased mitochondrial ROS generation, resulting in marked potentiation of ragweed pollen extract-induced accumulation of inflammatory cells in the airways (Aguilera-Aguirre et al., 2009).

In other cell types, mitochondrial impairment also increased the generation of ROS, resulting in potentiation of cytotoxicity or inflammatory cell accumulation (Bulua et al., 2011; Schulze-Osthoff et al., 1993).

Moreover, several studies have reported enhanced sensitivity to NF-κB activation in the context of mitochondrial dysfunction, suggesting that NF-κB binding sites may be essential for inflammatory gene expression (Ungvari et al., 2007; Vaamonde-García et al., 2012).

As expected, this inflammatory response is modulated by antioxidants. Additionally, these in vitro studies are supported by in vivo results. Thus, the effects of in vivo NF-kB inhibition (with pyrrolidinedithiocarbamate, PDTC or resveratrol) in aged rats significantly attenuated inflammatory gene expression and inhibited monocyte adhesiveness in vessels (Ungvari et al., 2007).

2.2. Mitochondrial modulation of the NLRP3 inflammasome.
Recent studies have shown mitochondria as important players in NLRP3 inflammasome activation (Green et al., 2011; Kepp et al., 2011; Nakahira et al., 2011; Shimada et al., 2012; Zhou et al., 2011).

Cell damage can initiate the accumulation of damage-associated molecular patterns (DAMPs), molecules recognized by the innate immune system and able to develop an inflammatory response (Krysko et al., 2011b) through the framework of the inflammasome and subsequent caspase-1 activation and pro-inflammatory cytokines release. Mitochondria have been recently identified as key sources of DAMPs (mito-DAMPs) playing a role in DAMP-modulated inflammation in different disorders, such as systemic inflammatory response syndrome (SIRS), rheumatoid arthritis (RA), cirrhosis, cancer, and heart diseases (Green et al., 2011; Zhang et al., 2010; Zitvogel et al., 2012), as well as the aging process (Salminen et al., 2012).

DAMPs activate the same receptors that detect pathogen-associated molecular patterns (PAMPs) (Krysko et al., 2011), such as Toll-like membrane receptors (TLRs) and cytoplasmic nucleotide-binding oligomerization domain (NOD)-like receptors (NLRs).

One of the best-studied NLRs, because of its solid association with several inflammatory diseases, is NLRP3. NLRP3, a cytosolic receptor, is one of 22 described human NLR family members; it functions to monitor the cytosol for stressful situations and has a crucial role in regulating immune responses (Kepp et al., 2011; Schroder and Tschopp, 2010; Shimada et al., 2012; Tschopp, 2011).

Once activated, the NLRP3 receptor induces NLRP3 oligomerization and recruitment of the adaptor protein apoptosis-associated speck-like protein containing a caspase recruitment domain (ASC) and procaspase-1, generating a multiprotein platform called the NLRP3 inflammasome. Then, NLRP3 activates caspase-1 and in turn drives the maturation of pro-inflammatory cytokines, such as IL-1β and IL-18. The NLRP3 inflammasome also conducts caspase-1-mediated cleavage of other substrates such as cytoskeletal proteins, glycolytic enzymes and caspase-7 (Tschopp, 2011; Zitvogel et al., 2012).

Mitochondria through ROS generation act as integrators across different stimuli activating the NLRP3 inflammasome (Kepp et al., 2011).

Increased levels of ROS are essential for NLRP3 activation; therefore, scavenging ROS with different chemicals suppresses inflammasome activation (Tschopp, 2011). Initially, Tschopp’s group, performing experiments in the human acute monocyte leukemia cell line, THP-1, suggested that NLRP3 inflammasome-activating ROS were generated by an NADPH oxidase stimulated upon particle phagocytosis, since in their model, ROS production was inhibited by NADPH oxidase inhibitors (Dostert et al., 2008).

However, later experiments performed by the same group (Zhou et al., 2011) and others showed that this result was ambiguous since immune cells from chronic granulomatous disease (CGD) patients, which have defective NADPH activity and thereby lack NADPH-dependent ROS generation, achieved normal levels of active caspase-1 and, in turn, active IL-1β.

Thus, these studies provided evidence of the existence of an alternative source of ROS (Meissner et al., 2010; van Bruggen et al., 2010; van de Veerdonk et al., 2010).

Since mitochondria are the main source of ROS, it is not surprising that mitochondria could also be the principal initiator of NLRP3 inflammasome activation. For example, the subcellular localization of NLRP3 suggests an important link between mitochondria and NLRP3. Under unstimulated conditions, most NLRP3 and ASC proteins are associated with the endoplasmic reticulum (ER).

Once the inflammasome is activated, NLRP3 redistributes to the perinuclear space and colocalizes with mitochondria (Zhou et al., 2011). Additionally, inhibition of voltage-dependent anion channels (VDACs), which are regulated by Bcl-2 family members through a decrease in mitochondrial Ca2 + levels and the resulting ROS production, significantly decreases NLRP3 inflammasome activation (Zhou et al., 2011).

Intriguingly, VDAC is also required for pro-apoptotic activity by Bax.
Other danger signals, specifically mtDNA and ATP, have been implicated in NLRP3 inflammasome activation; these signals can be released following different types of cell death (discussed below).

Indeed, it has been reported that mtDNA in the plasma of patients with femur trauma reaches values several thousand times higher than that of healthy volunteers, with functionally important immune repercussions (Zhang et al., 2010).

Additionally, mtDNA has been detected in RA synovial fluid and induces in vivo and in vitro inflammatory responses (Collins et al., 2004). Also, recent results have identified degraded mtDNA as a new DAMP subtype and a possible trigger of neurodegeneration (Mathew et al., 2012).

With regard to ATP, studies have shown that shedding of ATP from necrotic hepatocytes activates the NLRP3 inflammasome to generate an inflammatory microenvironment, alerting circulating neutrophils to adhere within liver sinusoids (McDonald et al., 2010).

Also, it has been shown that the release of extracellular ATP during apoptosis could be required for the successful immunogenic tumor cell death elicited by cancer treatment (Hahn et al., 2012).

Moreover, ATP can decrease intracellular K+ concentrations, a necessary event for NLRP3inflammasome activation, and mitochondria could also have a crucial role in modulating this process since they possess several K+ channels and can modulate intracellular K+ levels (Heinen et al., 2007).

Finally, NFPs is also considered a crucial mito-DAMP in several pathologies, including SIRS; this molecule can attract and stimulate neutrophils through high-affinity formyl peptide receptors (Zhang et al., 2010).

Interestingly, DAMPs and mito-DAMPs can activate antigenpresenting cells (APCs), as well as other nonimmune cell types, such as mesenchymal stem cells and astrocytes (Mathew et al., 2012; Pistoia and Raffaghello, 2011). In one study, degraded mtDNA strongly induced pro-inflammatory IL-1β, IL-6, MCP-1, and TNFα in mouse primary astrocytes, demonstrating the role of degraded mtDNA in inflammasome activation (Mathew et al., 2012).

The NLRP3 inflammasome and NF-κB pathways could work together to activate inflammatory cytokines, leading to overstimulation of the inflammatory response (Escames et al., 2011).

Once constituted, inflammasomes activate inflammatory caspases, mainly caspase-1, and subsequently induce IL-1β and IL-18 secretion. Then, pro-inflammatory cytokines can also activate NF-κB, generating a vicious cycle that increases the level and duration of the inflammatory response. The activation of inflammasomes may represent a possible explanation for this synergistic effect between mitochondrial impairment and cytokines.

Thus, studies investigating the role of mitochondria in inflammasome activation are revealing novel concepts in the development of inflammatory response, and these mechanisms may be of help in explaining the usual association of mitochondria with several inflammatory disorders and the aging process.

2.2.1. Role of mitochondria in the inflammatory response modulated by cell death.
Depending on the type of cell death, the dying cells could release and present at their surface specific signals that could modulate the type of immune response (Zitvogel et al., 2010).

Mitochondria have a key role in cell death (Kepp et al., 2011). Cell death can result from a variety of processes, including the apoptotic pathway and the necrotic pathway, 2 well-described extremes.

Apoptosis involves a series of orchestrated events requiring protein synthesis and ATP consumption. The dying cell forms vesicles that fragment into several apoptotic bodies; these are phagocytosed by macrophages or adjacent cells, thus avoiding the inflammatory reaction. Interestingly, contradictory recent data show that this paradigm is not always true, and depending on the inflammatory scenery, apoptosis could elicit a powerful inflammatory response (Shimada et al., 2012).

In particular, this study provides a link between apoptosis and inflammasome activation via binding of cytosolic oxidized mtDNA to the NLRP3 inflammasome. In this sense, Nakahira et al. also find that cyclosporine A, which modulates intrinsic apoptosis through inhibition of mitochondrial permeability transition, prevents the release of IL-1β by apoptotic stimuli (Nakahira et al., 2011).

Moreover, in vitro studies showed that silencing of antiapoptotic protein Bcl-2 increases IL-1β secretion while Bcl-2 overexpression resulted in the contrary (Shimada et al., 2012). In contrast, necrosis results from a rapid loss of cell homeostasis as intracellular ATP stores become depleted.

Cells increase their volume, leading to loss of cell membrane integrity and leakage of cellular components, thus inducing inflammation. In this study by Shimada above, necrotic stimuli did not secrete IL-1β (Shimada et al., 2012).

Mitochondrial DAMPs can also trigger autophagy. Autophagy is usually considered a type of cell death, although this process is thought to be an attempt to adapt and survive during periods of stress, such as nutrient deprivation, hypoxia, etc. (Edinger and Thompson, 2004).

Autophagy can control inflammation through macrophage-mediated clearing of apoptotic corpses or by inhibiting NLRP3 inflammasome activation by removing permeabilized or ROS-producing mitochondria (mitophagy). Thus, the inhibition of mitophagy results in spontaneous inflammasome activation (Goldman et al., 2010; Levine and Kroemer, 2008; Nakahira et al., 2011). Besides, mitochondria also play an important role in autophagia as they supply membranes for the biogenesis of autophagosomes (Hailey. 2010). Decreases in the expression of several autophagy-related genes, as well as the presence of mutations in these genes, have been found both in different inflammatory diseases, including neurodegenerative diseases, cardiovascular diseases, and Crohn’s disease (Gegg and Schapira, 2011; Travassos et al., 2010), and in the aging process (Salminen et al., 2012). For example, alterations in PINK1 and Parkin genes, which cooperate to identify and label damaged mitochondria for selective degradation via autophagy or mutations in ATG16 and IRGM, respectively, have been described. Additionally, the in vivo relevance of autophagic regulation of caspase-1 mediated inflammatory responses has been shown in two animal models of sepsis (the endotoxic shock model and the cecal ligation and puncture model of polymicrobial sepsis) (Nakahira et al., 2011). Furthermore, autophagy also declines during aging, generating the inflammatory condition via activation of inflammasomes, which in turn can actually accelerate the aging process (Salminen et al., 2012).

Mitochondria are triggers for the apoptotic pathway; several soluble pro-apoptotic proteins, including cytochrome c, are released from the mitochondrial intermembranous space after mitochondrial outer membrane permeabilization (MOMP). This leads to the formation of the apoptosome and the subsequent activation of the caspase cascade, which will trigger the execution phase of apoptosis. On the other hand, the mitochondrial permeability transition (MPT), which depends on the mitochondrial matrix protein cyclophilin D (Nakagawa et al., 2005), results in the instantaneous dissipation of the mitochondrial transmembrane potential and cessation of OXPHOS, thus rapidly increasing ROS, depleting ATP levels, and triggering necrotic cell death (Green et al., 2011; Hahn et al., 2012; Sasi et al., 2009).

As mentioned before, mtDNA can directly bind NLRP3; however, the current results do not rule out that mtDNA may bind to other members of the NLRP3 complex. ATP released from the dying cell constitutes a “find me” signal that acts as a chemo-attractant for monocytes and macrophages. Apoptosis releases much higher levels of cellular ATP related to necrosis (Hahn et al., 2012). It is therefore possible that ATP binding would be necessary for mtDNA binding (Shimada et al., 2012). DAMPs are often modified by processes of proteolysis and oxidation, which are related to cell death mechanisms. In this sense, HMGB1, a member of the high mobility group box family of DNA, can be released by necrotic cells, triggering a powerful immune response. During apoptosis, the generation of ROS by mitochondrial oxidized HMGB1 in turn enabled its immunostimulatory capacity (Hahn et al., 2012; Krysko et al., 2011).

Overall, depending on the specific circumstances and environment, cells can suffer from different lethal stress pathways promoting specific inflammatory responses, i.e., sterile inflammation or pathogen-induced inflammation. A better understanding of these processes is necessary for the development of new successful therapies to mitigate inflammatory diseases.

3. Pro-inflammatory mediators impair mitochondrial activity
A great number of inflammatory mediators, including the cytokines TNFα and IL-1β and the reactive nitrogen intermediate NO, may induce mitochondrial damage (López-Armada et al., 2006a; Maneiro et al., 2005; Stadler et al., 1992; Zell et al., 1997). TNFα and IL-1β decrease the activity of MRC complex I, ATP production, and mitochondrial membrane potential (Δψm). These mediators also induce the accumulation of significant amounts of ROS (Guidarelli et al., 2007; Kim et al., 2010). Interestingly, complex I together with complex III have been suggested to be major sources of ROS (Guzy et al., 2005), and the activities of complexes II and IV were also decreased in other cell types (Biniecka et al., 2011; Maneiro et al., 2003). Moreover, in adipocytes, the changes induced by TNFα cause pronounced morphological changes in the mitochondria, presumably due to variations in the levels of several mitofusion proteins (Chen et al., 2010). Another important study by Rowlands et al. showed that the severity of inflammation in mouselung microvessels is modulated by the inhibitory effect of TNFR1 ectodomain shedding by mitochondrial Ca2+ (Rowlands et al., 2011). Thus, mitochondrial Ca2+ and ROS are primary mediators of TNFαmediated inflammatory responses (Dada and Sznajder, 2011).

NO is another particularly important mediator in the pathophysiology of inflammatory processes. A variety of NO donors have been shown to suppress mitochondrial energy production in different cell types. Moreover, most of the catabolic effects of NO are potentially related to the ability of NO to combine with the superoxide anion to generate peroxynitrite, which, as mentioned earlier, is a powerful oxidant capable of inhibiting important enzymes and affecting mitochondrial integrity (Fermor et al., 2010; Johnson et al., 2000; Maneiro et al., 2005). In particular, NO induced a significant, reversible decrease in the activity of complex IV as well as a reduction in Δψm (Tomita et al., 2001). NO can also irreversibly inhibit respiration, most likely through ATP synthase and the strong oxidant peroxynitrite, which inactivates all respiratory complexes. Other studies demonstrated that, in intact U937 cells, peroxynitrite enhances the formation of superoxide ions in the mitochondria through a Ca2+-dependent mechanism that involves the inhibition of complex III, which then dismutates to H2O2 (Guidarelli et al., 2007).

Mitochondria play an important role during apoptosis. As mentioned above, severe mitochondrial damage can even lead to cell death by necrosis or by activation of the apoptotic signaling pathway and related caspases (Cillero-Pastor et al., 2011; Irrinki et al., 2011; López-Armada et al., 2006b; Maneiro et al., 2005). Indeed, the classic signs of cell death are preceded by mitochondrial alterations, which include changes in MRC activities, loss of Δψm, decreases in energy production, and/or induction of the mitochondrial permeability transition (MPT), and all of these alterations are modulated by both cytokines and NO. Thus, such pro-inflammatory mediators can modulate cell survival by exerting their effects on mitochondria. In addition, mito-DAMPs can be generated as consequence of these cell death processes.

To summarize, mitochondrial complexes could suffer permanent alterations caused by mutations in DNA or inflammatory mediators, such as reactive nitrogen species (RNS). Other stimuli, including cytokines and NO, induce functional and reversible modification of mitochondrial complexes at physiological concentrations. The combination of both effects (permanent and functional alteration of mitochondrial complexes) may cause a more substantial modification of mitochondrial activity, establishing continual, sublethal mitochondrial damage and thereby generating a vicious cycle of increasing inflammation that is difficult to break. Therefore, preserving mitochondrial function could reduce excessive oxidative stress and may represent a novel therapeutic advantage for patients with inflammatory diseases and chronic low-grade inflammation in aging.

4. Physiological function of ROS in the regulation of cell signaling
Although ROS has been traditionally linked to disease states, its crucial roles in normal physiological processes (vascular tone, oxygen sensing, and skeletal muscle physiology) and protective mechanisms, such as the above-mentioned host defense system and control of inflammation and immune responses, are becoming obvious. In this sense, there is a tight balance between appropriate redox states and oxidative stress. The final balance results from the magnitude and class of ROS, as well as the persistence of ROS production. Thus, high ROS concentrations may drive unspecific damage to nucleic acids, proteins, carbohydrates, and lipids or overtake the signaling pathways regulated by ROS.

In contrast, low to moderate levels of ROS lead to physiological regulation of cell signaling pathways (protein phosphorylation, ion channels, and transcription factors) and subsequently control cellular processes, such as differentiation, apoptosis, and migration.

A brief description of the most relevant physiological roles of ROS is as follows. ROS participate in immune system, at both innate and acquired levels. These molecules produced by phagocytes represent one of the early actions of defense against pathogen invasion. Notably, some pathogens have been shown to downregulate mitochondrial ROS generation and thereby prevent macrophage defense mechanisms, favoring pathogen persistence (Basu Ball et al., 2011). An example of the importance of ROS in immune system signaling is reflected in chronic granulomatous disease (CGD), which is caused by a lack of ROS-generating phagocytes. In this manner, CGD patients inefficiently kill invader bacteria by ROS, suffering recurrent infections (Mauch et al., 2007). Interestingly, CGD patients and mice defective for ROS production also show evidence of noninfectious inflammatory states, supporting the role of ROS in controlling inflammation and immune responses (van de Veerdonk et al., 2010). With respect to acquired immunity, ROS participates in signal transduction cascades within T lymphocytes (Alfadda and Sallam, 2012). Furthermore, as detailed earlier, recent evidence highlights a role for ROS in immune responses through NRPL3 activation (Nakahira et al., 2011; Shimada et al., 2012; Zhou et al., 2011).

Other illustrations of the crucial role of ROS in normal physiological processes are the key regulation of vascular system, the role of ROS in oxygen sensing through stabilization of hypoxia-inducible factor-1 (HIF-1) and the subsequent stimulation of angiogenesis and bioenergetic responses to rescue redox homeostasis, the role of ROS in skeletal muscle physiology regulation through glucose uptake during contraction, and the role of ROS in regulating gene stability, transcription, and signal transduction (Alfadda and Sallam, 2012).
Finally, increasing evidence that antioxidants may be harmful highlights the physiological role of ROS. Specifically, Ristow et al. reported that antioxidant supplementation may prevent increased insulin sensitivity by reducing the expression of ROS-sensitive transcriptional regulators of insulin sensitivity associated with exercise training in humans (Ristow et al., 2009). Also, the nuclear factor-erythroid 2-related factor 2 (Nrf2), a redox stress-sensitive transcription factor that induces several antioxidant and detoxification genes, is kept inactive in the absence of redox stress by its binding to Kelch-like ECH-protein 1 (Keap1). Thus, disruption of Nrf2 signaling impaired angiogenesis and microvascular rarefaction in aging (Valcarcel-Ares et al., 2012). In this sense, it is worth bearing in mind that antioxidant strategies will likely not be successful because their activity is too unspecific, too insufficient, and too delayed; however, it is also plausible that the ineffectiveness of antioxidant strategies may arise from their inhibition of essential cell functions that require ROS.

However, excessive levels of ROS cause not only oxidative damage to macromolecules, but can also affect the physiological signaling pathways regulated by ROS. Among other consequences, increased ROS production will lead to augmented activation of the transcription factors HIF-1, NF-κB, and activator protein-1 (AP-1) as well as the NLRP3 inflammasome. This will increase the release of pro-inflammatory cytokines that in turn will enhance ROS production, and hence, feed-back in a self-stimulatory manner, thereby amplifying the inflammatory response (Dröge, 2002).

5. Mitochondrial dysfunction and inflammatory responses in chronic human pathologies
Extensive research over the past decade has revealed that continued mitochondrial dysfunction can lead to chronic inflammation, which in turn may mediate most chronic diseases, including rheumatoid, cardiovascular, neurological, or metabolic diseases.

5.1. Rheumatoid diseases
Osteoarthritis (OA) and RA are the most common rheumatoid pathologies. In both of these conditions, the mitochondrial characteristics, including MRC activity, ATP synthesis, and Δψm, are adversely affected in several cell types. In particular, ex vivo studies have revealed that OA chondrocytes, as compared with normal chondrocytes, showed significant decreases in the activities of MRC complexes II and III and Δψm (Johnson et al., 2000; Liu et al., 2010; Maneiro et al., 2003). In RA and

systemic juvenile idiopathic arthritis, studies have demonstrated that there is a deficiency in a subunit of the MRC complex IV in the synovium (Biniecka et al., 2011; Ishikawa et al., 2009). Furthermore, human RA synoviocytes (Da Sylva et al., 2005; Ospelt and Gay, 2005) and normal synoviocytes under arthritis-like conditions, such as hypoxia (Biniecka et al., 2011), which causes a significant increase in Δψm in synovial cells, have been shown to exhibit significantly higher mutation rates than normal-aged or OA synoviocytes. These data support the conclusion that mitochondrial mutagenesis is correlated with the local inflammatory environment in arthritis (Harty et al., 2011). Other cells, such as T lymphocytes, show mitochondrial impairment in patients with RA and systemic lupus erythematosus (Gergely et al., 2002; Moodley et al., 2008).

Oxidative stress can contribute to the activated phenotype of synoviocytes in inflammatory arthritis and OA chondrocytes (Davies et al., 2008; Filippin et al., 2008; Henrotin et al., 2005). Interestingly, studies have shown that inhibition of complex III or V induces ROS production and NF-κB activation in human articular chondrocytes (CilleroPastor et al., 2008; Milner et al., 2007) and synovial cells (Valcárcel-Ares et al., 2010) in vitro. Additionally, inhibition of complex III or V in chondrocytes/synoviocytes induces low-grade inflammatory and matrix degradation processes through the synthesis of pro-inflammatory stimuli, including prostaglandin E2 (PGE2), the chemokines IL-8 and monocyte chemotactic protein-1, and several matrix metalloproteinases (MMPs) (Cillero-Pastor et al., 2010; Cillero-Pastor et al., 2008; VaamondeGarcía et al., 2012).

Specifically, it has been shown that the inhibition of MRC activity induces cyclooxygenase-2 (COX-2) expression and PGE2 production through redistribution of mitochondrial Ca2+, generation of ROS, and activation of the transcription factor NF-κB (Biniecka et al., 2010; Cillero-Pastor et al., 2008). Mitochondrial dysfunction can also induce PGE2 liberation in human OA chondrocytes through 4-hydroxy-2nonenal (4-HNE) (Vaillancourt et al., 2007), a lipid peroxidation byproduct that is associated with a higher frequency of mtDNA mutations, which are often increased in articular tissues of patients with inflammatory arthritis (Biniecka et al., 2010; Vaillancourt et al., 2007).

In addition to the above processes, the decline in mitochondrial function can also increase the inflammatory responsiveness of both normal human chondrocytes and synoviocytes to cytokines, promoting mechanisms that may contribute to joint destruction and pain (Valcárcel-Ares et al., 2010; Vaamonde-García et al., 2012). Interestingly, the hypoxic state of arthritic joints in vivo, which correlates with increased disease activity and mtDNA mutagenesis rates, exacerbates the inflammatory response in synoviocytes by increasing COX-2 expression and the release of MMPs in response to IL-1β (Demasi et al., 2004). In fact, inflammasome activation may explain these synergistic phenomena. NF-κB signaling and the NLRP3 inflammasome pathway could work together to activate inflammatory cytokines, thereby leading to overstimulation of the inflammatory response (Escames et al., 2011). As additional support for the critical role of the inflammasome in the pathogenesis of OA, the NLRP3 inflammasome could also mediate the pathological effects of hydroxyapatite crystals in vitro and in vivo (Jin et al., 2011).

Other DAMPs/mito-DAMPs are also highly secreted in tissues of rheumatoid patients. In fact, oxidized mtDNA, which has immunostimulatory properties and is capable of inducing arthritis in mice, can be detected in the synovial fluid of RA patients, but not in that of control subjects (Collins et al., 2004). Some studies have also considered cytochrome c a mitoDAMP that can induce arthritis (Pullerits et al., 2005); however, recent data suggest that this implication is not mediated by a direct role in NLRP3 activation (Shimada et al., 2012). Moreover, as mentioned earlier, autophagy prevents the accumulation of dysfunctional mitochondria. In this regard, OA cartilage has been shown to be deficient in autophagic processes, and the pharmacological activation of autophagy may be an effective therapeutic approach for OA (Caramés et al., 2012). Together, these findings support the critical role of mitochondrial dysfunction in rheumatoid disorders.

5.2. Cardiovascular diseases
Abnormalities in mitochondrial functions are increasingly recognized in association with cardiomyopathies. In fact, the myocardium is a highly energy-demanding tissue, with mitochondria supplying greater than 90% of ATP. Consequently, mitochondria are essential for cardiomyocyte function and viability (Dutta et al., 2012). Since ROS are constantly generated during mitochondrial respiration, it is not surprising that a dysregulation in the production of free radicals can cause oxidative damage in the heart. This oxidative damage may cause mtDNA mutations that result in mitochondrial dysfunction. Moreover, increased mitochondrial Ca2 + levels modulate myocardial inflammation and dysfunction in states of injury, such as sepsis and burn trauma (Maass et al., 2005). Another important study showed that escape of mtDNA from autophagy-mediated degradation leads to TLR9-mediated inflammatory responses in cardiomyocytes and may induce myocarditis and dilated cardiomyopathy (Oka et al., 2012).

With advanced age, mitochondrial ROS production significantly increases both in the heart (Judge et al., 2005) and vasculature (Ungvari et al., 2007), making the vascular endothelium and smooth muscle susceptible to ROS oxidation. In fact, mitochondria-derived ROS likely contribute to the development of chronic low-grade vascular inflammation in aging by activating redox signaling pathways (Dai et al., 2012). This low-grade vascular inflammation actives the redox-sensitive transcription factor NF-kB, stimulating endothelial activation by production and secretion of cytokines, growth factors, and proteases in the vascular wall, thereby promoting atherosclerosis. This damage is augmented by excessive ROS generation in the heart and the vasculature, as well as defective oxidant scavenging. These mechanisms disturb endothelial homeostasis (i.e., induce endothelial dysfunction), a characteristic feature of patients with coronary atherosclerosis (Munzel et al., 2010).

Traditionally, it has been hypothesized that cardiovascular diseases are precipitated by atherosclerosis due to arterial blockage from fatty deposits. However, in recent years, diagnostic and therapeutic strategies have been based not only on control of cholesterol levels, healthy diet, and smoking, but also taking into account other factors, such as inflammatory markers and mitochondria (Garg, 2011; Marchant et al., 2012). In particular, the roles of inflammasomes in the development of cardiovascular diseases are currently being studied in depth (Garg, 2011). Finally, pathways that improve mitochondrial function and attenuate mitochondrial oxidative stress (i.e., antioxidant therapies) show efficacy in various animal models and promising results for the treatment of cardiovascular diseases.

5.3. Neurological diseases
Increasing evidence has demonstrated the role of mitochondrial dysfunction in neurological and neurodegenerative diseases, including Alzheimer’s disease (AD), Parkinson’s disease (PD), dementia, multiple sclerosis, ataxia, and encephalomyopathies (Monje et al., 2003; Morán et al., 2012; Witte et al., 2010). A common feature in neurodegenerative diseases is the presence of neuroinflammation and oxidative stress. In both AD and PD, oxidative stress activates inflammatory signaling pathways through several mechanisms, including exacerbating the production of ROS, inducing mitochondrial dysfunction, activating microglia and astrocytes to release pro-inflammatory cytokines, and others (Schapira, 2010; von Bernhardi and Eugenín, 2012; Whitton, 2007). In fact, animal models of PD, induced by the mitochondrial inhibitor rotenone, exhibit increased production of IL-1β in the hypothalamus (Yi et al., 2007).

Furthermore, pro-inflammatory cytokines released in activated glia have been shown to stimulate increased ROS production in the brain, contributing to other mechanisms that amplify ROS production and lead to deterioration of mitochondrial function. In addition to exacerbating disease progression, this process may also suppress neurogenesis. Severe mitochondrial damage may even lead to cell death by necrosis or by activation of apoptotic signaling and activation of caspases (Voloboueva and Giffard, 2011).

Because neurodegenerative diseases are accompanied by chronic inflammation, the induction of mitophagy may have beneficial effects by removing permeabilized or ROS-producing mitochondria (Green et al., 2011). Moreover, as mentioned earlier, alterations in several genes involved in autophagy, such as PINK1 or Parkin, have been described in PD. Furthermore, downregulation of mitochondrial ROS may also decrease astrocyte activation (Voloboueva and Giffard, 2011). Finally, it has been recently described in mouse primary astrocytes how degraded mtDNA could be a new identified subtype of DAMP inducing inflammasome activation and ensuing neurodegeneration (Mathew et al., 2012).

5.4. Metabolic diseases
Insulin resistance, pancreatic islet β-cell failure, and elevated plasma free fatty acid (FFA) levels are features commonly associated with the development of several metabolic diseases, including obesity, type 2 diabetes mellitus, and metabolic syndrome. Accumulating evidence indicates that mitochondrial dysfunction is a central contributor or mediator of these processes (Coletta and Mandarino, 2011; Ma et al., 2012; Martins et al., 2012; Vial et al., 2010).

Diverse sources of inflammation can contribute to the insulin resistance observed in obesity and type 2 diabetes. Elevated plasma FFA levels play an important role in the development of skeletal muscle insulin resistance through different mechanisms, including oxidative stress, inflammation, and mitochondrial dysfunction (Coletta and Mandarino, 2011; Martins et al., 2012). The inflammation induced by FFAs may induce impairment in mitochondrial function in adipocytes (Ji et al., 2011; Youssef-Elabd et al., 2012). Furthermore, FFAs can activate these inflammatory signaling pathways directly through interactions with members of the TLR family and inflammasome activation (Reynolds et al., 2012; Wen et al., 2011; Youssef-Elabd et al., 2012).
Since mitochondrial dysfunction drives NLRP3 inflammasome activation, it is not surprising that this pathway can underlie cellular metabolic disorders. Specifically, a large body of evidence demonstrates a pathological role for the association of NLRP3 with a nonmitochondrial protein, thioredoxin-interacting protein (TXNIP), in type 2 diabetes and obesity. For example, TXNIP deficiency impairs activation of the NLRP3 inflammasome, improving glucose tolerance and insulin sensitivity in β-cells and adipocytes and decreasing fat deposits in the liver (Vandanmagsar et al., 2011; Zhou et al., 2010). Interestingly, high levels of oxidized DNA have been related with diabetes (Simone et al., 2008). It is possible that the elevated levels of damaged DNA detected in diabetes could be related with the activation of NLRP3 (Shimada et al., 2012). Consequently, mitochondria may represent novel pharmacological targets for therapeutic interventions in metabolic disorders.

5.5. Cancer and immunogenic cell death
Mitochondrial processes play an important role in tumor initiation and progression. Otto Warburg made the first association between mitochondrial dysfunction and cancer in the 1930s. He observed increased rates of aerobic glycolysis in a variety of tumor cell types, and he hypothesized that the cell’s choice of the glycolytic route to produce ATP instead of the more productive oxidative phosphorylation route may be due to the impaired respiratory capacity of these cells (Warburg, 1956). Years after this discovery, a number of cancer-related mitochondrial defects have been identified. In this sense, altered expression and activity of respiratory chain subunits as well as mtDNA mutations have been associated with human tumors. For example, MRC Complex I dysfunction has been associated with the pathogenesis of Hürtle cell tumors of the thyroid (Máximo et al., 2005). In the same way, low activity of Complex III has been found in breast cancer (Putignani et al., 2008) and Hürtle cell tumors (Bonora et al., 2006; Stankov et al.,

The aggressiveness of renal cell tumors is related to lower Complex II, III, and IV activities (Simonnet et al., 2002). As we know, dysfunction in MRC complexes is associated with mtDNA mutations. In fact, mutations in mtDNA have been reported in a variety of cancers, including ovarian, thyroid, salivary, kidney, liver, lung, colon, gastric, brain, bladder, head and neck, and breast cancers, as well as leukaemia (Modica-Napolitano and Singh, 2004).

In the context of mitochondrial dysfunction, we found that ROS is a key player in tumor formation and progression. It is now recognized that ROS plays an important role as a signaling molecule, mediating changes in cell proliferation, differentiation, migration, and invasiveness, as well as large-scale changes in gene transcription (Weinberg and Chandel, 2009). It is well known that mitochondrial dysfunction increases ROS levels, and tumor cells generally exhibit higher levels of ROS than normal cells (Weinberg and Chandel, 2009). One signaling pathway that may be particularly important to ROS-mediated tumorigenesis involves the activation and stabilization of HIF. HIF is a key transcription factor in cancer progression that regulates the enhancement of both glucose metabolism and angiogenesis. HIF activation occurs under hypoxic conditions, but ROS also appears to stabilize HIF under conditions of normal oxygen concentrations, leading to aberrant activation of HIF and promoting tumorigenesis (Chandel et al., 1998).

In addition, it has been demonstrated that tumors lacking the tumor-suppressor succinate dehydrogenase (Complex I) and fumarate hydratase exhibit impaired HIF degradation (King et al., 2006), once again linking mitochondrial impairment with oncongenesis.
As was discussed earlier, mitochondrial ROS production results in activation of the NLRP3 inflammasome with the consequent release of pro-inflammatory cytokines (Zitvogel et al., 2012). In vivo experiments suggest that inflammasome products, such as IL-1β, can directly drive oncogenesis or suppress immunosurvellaince mechanisms, thereby facilitating tumor development (Zitvogel et al., 2012). In the same way, many human cancers are etiologically linked to chronic inflammatory processes, such as those that occur with gastric cancer, hepatic cancer, and colorectal cancer (Coussens and Werb, 2002). However, the role of the inflammasome in carcinogenesis is controversial. Yet, in sharp contrast, the inflammasome or its products, depending on the specific tumor and its grade, can reduce tumorogenesis by causing cell death and promoting antitumor immune responses (Zitvogel et al., 2012).

In this context, recently developed therapies have been able to induce cell death by triggering ROS production. These include a specific group of mitochondrial-targeted anticancer drugs (mitocans) (Hahn et al., 2012). An example of these mitocans is the vitamin E succinate. This pro-oxidant selectively targets cancer cell mitochondria and is a potent inducer of ROS, leading to apoptosis in cancer cells but not in related normal cells. This is because cancer cells are under much more oxidative stress than normal cells and thus are more vulnerable to further damage by ROS-generating agents. Interestingly, these types of therapies have immune-enhancing properties, meaning that they are able to modulate antitumor immune responses, stimulating the immune system to cause immunogenic tumor cell death by upregulating the release of DAMP molecules. Of note, the activation or inhibition of the inflammasome as cancer therapy will likely depend on the specific cancer and its grade (Hahn et al., 2012).

6. Mitochondrial dysfunction and inflammatory responses in acute human pathologies
Generally, the acute inflammatory response to infection and tissue damage should be able to reduce injury to the organism, preventing further damage to the cells or tissues. However, as mentioned above, when the acute inflammatory responses cannot be effectively resolved, progressive damage and multi-organ failure can occur. Sepsis, a representative acute inflammatory disease, is unfortunately the main cause of death in critical care units. Mitochondrial dysfunction and impaired oxygen consumption play a role in sepsis, and the severity of sepsis has

been shown to correlate with mitochondrial damage in both humans and experimental models (Garrabou et al., 2012; Kung et al., 2012). Thus, changes to the ultrastructure of the mitochondria and significant inhibition of mitochondrial complex activities have been described in several cell types isolated from septic patients. Also, during overwhelming sepsis, increased Ca2+ influx into the mitochondria leads to mitochondrial dysfunction, resulting in the release of cytochrome c and cell death (Dada and Sznajder, 2011).

In this context, as mentioned before, mitochondrial damage may release several mito-DAMPs, including mtDNA. Consistent with this, septic plasma samples have shown significantly increased amounts of mtDNA and inflammatory cytokines, with corresponding NF-κB activation, and a correlation between adverse outcomes in sepsis and levels of extracellular mtDNA has been demonstrated (Garrabou et al., 2012). Since mtDNA can activate the NLRP3 inflammasome, its participation in the pathogenesis of sepsis is not surprising.

Interestingly, NLRP3 polymorphisms may be used as relevant risk estimates for the development of sepsis, which highlights the importance of NLRP3 in the pathogenesis of sepsis (Zhang et al., 2011). As expected, mitochondrial antioxidants have been shown to alleviate oxidative and nitrosative stress in several in vitro and in vivo model of sepsis (Apostolova et al., 2011; Zang et al., 2012). Taken together, these findings have led to a better understanding of the pathophysiological
processes of acute diseases and will be fundamental in improving the outcome of sepsis-related disorders.

7. Mitochondrial protection attenuates inflammation
Excessive mitochondrial oxidative stress plays a central role in triggering the deleterious cascade of events associated with inflammatory diseases and chronic low-grade inflammation in aging. In this regard, several approaches to preserve mitochondria are under investigation, including the use of antioxidant compounds, development of specific mitophagy-inducing therapies, and the pharmacological manipulation of mitochondrial sirtuins (SIRTs).

Strategies that control ROS production could limit either the redox-sensitive inflammatory pathway or the direct activation of the inflammasome. Thus, several meta-analyses have supported that dietary antioxidant intake is associated with a lower incidence of inflammatory diseases (Alissa and Ferns, 2012; Demetriou et al., 2012; Lahiri et al., 2012; Patelarou et al., 2011). In particular, resveratrol, a natural antioxidant found in grape skin and red wine, improves mitochondrial function by preventing oxidative stress and subsequent inflammation (Catalgol et al., 2012). In vitro, this molecule attenuates mitochondrial oxidative stress in several cell types, i.e., by restoring mitochondrial

complex III activity, which is considered to be the major source of mitochondrial oxidative stress (Xu et al., 2012). Moreover, resveratrol pretreatment efficiently prevents inhibition of mitochondrial membrane depolarization and ATP depletion, preserving mtDNA content and inhibiting COX-2 activity through a decrease in NF-κB activation in different cells (Dave et al., 2008; Lin et al., 2011; Ungvari et al., 2009; Xu et al., 2012). In animal models, resveratrol protects from developing age-related and acute diseases by improving mitochondrial function (Elmali et al., 2007; Jian et al., 2012; Lin et al., 2012; Ungvari et al., 2010).
As described earlier, an imbalance between mitophagy and mitochondrial biogenesis could be involved in inflammatory pathology. Autophagy/mitophagy decreases ROS production and mtDNA release after MPT opening, in turn affecting the NF-κB pathway and inflammasome activation. Thus, pharmacological induction of autophagy could mitigate inflammatory reactions. For example, activation of autophagy by rapamycin reduces the severity of experimental OA (Caramés et al., 2012). Interestingly, resveratrol also improves mitochondrial function through induction of mitochondrial biogenesis (Csiszar et al., 2009; Ungvari et al., 2009).

Recent studies revealed that SIRTs or class III histone deacetylases, have an essential role in mitochondrial protection against oxidative stress (Pereira et al., 2012). Three of the 7 SIRTs described in humans, SIRT3, SIRT4, and SIRT5, are localized inside mitochondria, which lends support to the potential role of SIRTs in mitochondrial biology (Hirschey et al., 2010). SIRTs can indirectly regulate the expression of several inflammatory genes through deacetylation of several signaling proteins, including the transcription factors FOXO (Salminen et al., 2008) and NF-κB (Natoli, 2009) and the peroxisome proliferator-activated receptor gamma coactivator-1α (PGC-1α) (Sugden et al., 2010). Deacetylation of FOXO by SIRT1 or SIRT3 upregulates the expression of SOD2 and catalase, thereby decreasing cellular ROS levels. Also, deacetylation of PGC-1α by both SIRT1 and SIRT3 stimulates mitochondrial biogenesis (Pereira et al., 2012). As mentioned above, several studies have demonstrated that resveratrol can protect mitochondrial functions through activation of SIRTs (Shinmura et al., 2011).

While the role of mitochondria in the development of the inflammatory response is now beginning to be understood, there is still a lot to be discovered. Understanding the mechanisms that induce inflammation through mitochondrial impairment may help to identify relevant therapeutic targets for the treatment of multiple degenerative and acute disorders as well as the aging process. For these reasons, it is plausible that, in few years, specific mitochondrial-targeting pharmacological treatments of inflammatory responses may become a reality. Indeed, some pharmaceutical companies have already begun designing new molecules that are the equivalent of super-antioxidants and have demonstrated exciting results in several preclinical studies in vitro and in vivo (McManus et al., 2011; Mukhopadhyay et al., 2012).

8. Limitations and future research directions
The use of effective strategies to increase our knowledge about the role of mitochondrial dysfunction in the inflammatory components of multiple degenerative and acute diseases, as well as in the aging process, is necessary for the development of new successful therapies aimed at improving human health. Thereby, an appropriate way to avoid contradictory results is the use of unmanipulated cells from healthy and diseased donors. Most in vitro studies are conducted in isolated cells kept in standard media with high glucose concentrations and under aerobic conditions, quite different from effects in tissues in a live organism, which could lead cells to obtain their energy predominantly from anaerobic glycolysis. Thus, the hypoxic nature of rheumatoid joints in vivo, which is correlated with increased mtDNA mutagenesis rate and disease activity, increases the inflammatory response in synoviocytes by increasing COX-2 expression and MMPs release in response to IL-1β (Demasi et al., 2004). For this reason, it is likely that the effects of mitochondrial impairment observed in cultures

maintained in normal atmosphere will even be more crucial in damaged tissue, where mitochondrial activity is probably lower due to genetic predisposition, somatic mutations in the mtDNA of the MRC as a consequence of the elevated levels of pro-inflammatory mediators (notably ROS), or the direct effects of pro-inflammatory mediators on the MRC. Future in vitro studies must take this into account. In addition, an important tool in these studies could be the transmitochondrial cybrid system, which allows the study of the effects of mitochondrial dysfunction and inflammatory responses with a common nuclear background, thereby avoiding the large number of inflammatory properties regulated by nuclear-encoded genes (Kaipparettu et al., 2010).

The majority of studies to date were conducted using in vitro or ex vivo strategies. Thus, it is necessary to translate the results to in vivo designs to support a clear role for mitochondrial dysfunction in the development of the inflammatory response. Animal models are also useful tools in medical science research. In this sense, different studies have reported that transgenic expression of mitochondrial-targeted catalase in mouse models of chronic diseases, such as cancer or Alzheimer’s disease, reduced primary tumor invasiveness and decreased metastasis incidence and burden (Goh et al., 2011) and several typical events of the disease (Mao et al., 2012).

Finally, in transgenic mouse lines, the overexpression of catalase targeted to mitochondria reduced age-associated pathologies and led to an extension in lifespan (Schriner et al., 2005). Overall, these in vivo studies support that mitochondria-targeted molecules may be an effective therapeutic approach to treat or prevent inflammation linked to numerous degenerative and acute diseases. In addition, other interesting models to be kept in mind are represented by mice with increased mtDNA mutations that exhibit pathology associated to human aging (Trifunovic, 2006) Analysis of the inflammasome has not been performed in these animals, but could be a useful model in which to study the role of mitochondria in NLRP3 inflammasome activation.

However, despite our increasing knowledge on the role of mitochondria in the development of the inflammatory response, there is still much to learn. Several key questions remain to be addressed. For example, which mechanisms determine whether mitochondria induce the inflammasome or apoptosome? Along these lines, recent research has suggested that oxidized mtDNA released during apoptosis results in activation of the NLRP3 inflammasome, providing a link between apoptosis and inflammasome activation (Shimada et al., 2012).

Interestingly, VDAC, whose activity is modulated by Bcl-2 family members, is decisive for inflammasome activation, and also in apoptosis induction by Bax (Yamagata et al., 2009; Zhou et al., 2011). However, what is the exact mechanism through which mtDNA activates the NLRP3 inflammasome? It also remains to be determined the importance of inflammasome in non-immune cell types.

Furthermore, other factors, such as tissue specificity, the effects of environmental factors, and age, may determine the final outcome. Clearly, more work is necessary to shed light on how mitochondrial dysfunction modulates the inflammatory response associated with distinct degenerative and acute diseases as well as the aging process.

9. Conclusions

Both acute and chronic inflammatory diseases, as well as the aging process, have been linked to accumulation of ROS and RNS, which could potentially be a main source of mitochondrial genomic instability leading to respiratory chain dysfunction. Indeed, mitochondrial impairment could modulate innate immunity through both redox-sensitive inflammatory pathways and direct activation of the inflammasome. Mitochondria could integrate these 2 pathways, leading to an overstimulation of the inflammatory response (Fig. 1). In conclusion, there is substantial evidence supporting the hypothesis that a decline in mitochondrial function is essential to the development of the inflammatory phenotype observed in multiple human degenerative or acute diseases as well as in advanced age.

All of these findings support mitochondria as new pharmacological targets. The preservation of mitochondrial function could reduce oxidative stress and may represent a novel therapeutic advantage for patients with degenerative or acute diseases as well as functional decline in aging. Despite these advances in our knowledge of inflammatory alterations, it is necessary to keep in mind that the best treatment is prevention; this is true for both inflammatory-related diseases and in healthy aging. Although we are becoming increasingly aware of the importance of a healthy diet, there is still a long way to go. Adequate nutrition could in fact be the best way to prevent aging declines and inflammatory degenerative or acute diseases. Thus, as Hippocrates said 25 centuries ago, “Let food be the medicine and medicine be the food”