ETHE1 overexpression promotes SIRT1 and PGC1α mediated aerobic glycolysis, oxidative phosphorylation, mitochondrial biogenesis and colorectal cancer

Ethylmalonic Encephalopathy Protein 1 (ETHE1) is a sulfur dioxygenase that regulates cellular H2S levels. We previously demonstrated a significant increase of ETHE1 expression in "single-hit" colon epithelial cells from crypts of patients with Familial Adenomatous Polyposis (FAP). Here, we report elevated levels of ETHE1 expression and increased mitochondrial density occurring in-situ in phenotypically normal FAP colorectal mucosa. We also found that constitutive expression of ETHE1 increased aerobic glycolysis ("Warburg effect"), oxidative phosphorylation, and mitochondrial biogenesis in colorectal cancer (CRC) cell lines, thereby depleting H2S which relieved the inhibition of phosphodiesterase (PDE), and increased adenosine monophosphate (AMP) levels. This led to activation of the energy sensing AMP-activated protein kinase (AMPKp), Sirtuin1 (SIRT1) and peroxisome proliferator-activated receptor γ coactivator 1α (PGC1α), a master regulator of mitochondrial biogenesis. By contrast, shRNA silencing of ETHE1 reduced PDE activity, AMPKp/SIRT1/PGC1α levels and mitochondrial biogenesis. Constitutive expression of ETHE1 accelerated both CRC cell xenograft and orthotopic patient derived xenograft CRC cell growth in vivo. Overall, our data nominate elevated ETHE1 expression levels as a novel biomarker and potential therapeutic target for the prevention of CRC tumorigenesis.


INTRODUCTION
The discovery of tumor suppressor genes (TSGs) and their singular role in the etiology of heritable cancers presents the opportunity to study early biomarkers and potential targets during cancer progression [1]. Previously, we investigated gene expression patterns of phenotypically normal "one-hit" cells before they become hemizygous or homozygous for the inherited mutant TSG that promotes tumor formation in hereditary cancers. This approach enables discovery of biomarkers and molecular targets in clinically-identifiable pre-neoplastic lesions such as polyposis of the colon, wherein numerous lesions appear before progression to adenocarcinomas [1][2][3][4].
Familial Adenomatous Polyposis (FAP) is characterized by hundreds to thousands of adenomatous polyps throughout the colorectum [5,6]. The majority of FAP patients carry a germline mutation in the adenomatous polyposis coli (APC) tumor suppressor, a negative regulator of WNT signaling and regulator of cell proliferation [7]. The high rate of somatic APC mutations in sporadic colorectal cancer is consistent with inactivation of the APC protein playing a critical role in the initiation of colorectal cancers [8]. This highlights the need to improve our understanding of events that occur during the transition of normal colon mucosa to cancer including development of early biomarkers and potential targets to prevent/treat FAP and other cohorts at high risk of CRC [9][10][11][12][13][14][15].
A recent proteomic study showed that phenotypically normal epithelial cells from the colonic crypts of FAP patients, referred to as "one-hit" cells, when compared to normal colon epithelial cells, revealed a significant increase (~27-fold) of the Ethylmalonic encephalopathy protein 1 (ETHE1) which, represented the highest increase in expression among all proteins identified [16]. ETHE1 is a sulfur dioxygenase, a hydrogen sulfide (H 2 S) catabolic enzyme, that is widely expressed in various tissues and is present in the mitochondria, cytosol and nucleus of eukaryotic animals [17]. In the mitochondria, ETHE1 facilitates H 2 S catabolism via oxidation and conversion of sulfide quinone reductase generated glutathione persulfides (GSSH) to sulfite (H 2 SO 3 ) [18,19]. Sulfite is further oxidized to sulfate that is secreted extra-cellularly. Consistent with this role, germline bi-allelic ETHE1 mutations cause ethylmalonic encephalopathy [20][21][22][23], a genetic disease in which H 2 S accumulates in critical tissues and can reach concentrations in the brain that inhibit Cytochrome C Oxidase (COX), blocking mitochondrial respiration, increasing lactic acid accumulation, and inducing encephalopathy [22,24,25]. Here, we focused on the colon since ETHE1 is highly expressed in normal colorectal epithelium wherein it modulates the accumulation of toxic endogenous H 2 S generated by colonic microbiota, dietary sulfur containing compounds and endogenous cellular H 2 S produced by cystathionine beta-synthase (CBS) [19,26].
We found abnormal expression of ETHE1 and increased mitochondria density in phenotypically normal APC +/-FAP intact colorectal mucosa. Furthermore, using constitutively expressed ETHE1 in CRC cells we identified novel mechanisms that link augmented ETHE1 expression with aerobic glycolysis and mitochondria biogenesis. We found that constitutive expression of ETHE1 reduces H 2 S mediated inactivation of phosphodiesterases (PDEs) and increases AMP concurrent with increased AMPK phosphorylation. This in turn activates the NAD-dependent protein deacetylase Sirtuin 1 (SIRT1). AMPK and Sirt1 are key metabolic sensors that regulate mitochondrial respiration and aerobic glycolysis [27][28][29][30][31]. AMPK and SIRT1 directly activate the nuclear receptor Peroxisome Proliferator-Activated Receptor Gamma Coactivator 1 Alpha (PGC1α) through phosphorylation (AMPK) and deacetylation (SIRT1), respectively [32][33][34]. PGC1α then drives mitochondrial biogenesis and oxidative phosphorylation leading to increased CRC proliferation, angiogenesis, and tumor growth in vivo.
Overall our studies provide new insights into the mechanistic role of elevated ETHE1 expression levels during CRC tumorigenesis in phenotypically normal "one-hit" FAP colon, highlighting the utility of the "onehit" model in heritable cancers to facilitate the discovery of potential biomarkers/targets, early during CRC progression.

In situ expression of ETHE1 in intact colon mucosa tissue specimens of FAP patients
Yeung et al. demonstrated that phenotypically normal, non-transformed APC "single-hit" cells in FAP epithelium showed the highest differential increase in ETHE1 levels among all proteins identified compared to normal colon epithelial cells from control subjects [16]. We further validated these findings using phenotypically normal intact colon mucosa surgical specimens from FAP (N=3) patients carrying APC c.388delA, APC c.1240delC and APC c.2586insCA truncating mutations, and individuals with no history of CRC (N=3) who were matched for location (sigmoid colon) age, gender and ethnicity, collected during screening colonoscopy. Western blot analysis confirmed >160% (p=0.001 unpaired t-test) increase in ETHE1 protein levels in FAP colonic tissues ( Figure 1A and 1B).

ETHE1 expression and activity in CRC cell lines
To study the role of increased ETHE1 levels in CRC, we used human-derived HCT116 and HT29 CRC cell lines that stably constitutively expresses a full-length cDNA encoding ETHE1. Concurrently, gene specific shorthairpin RNA (shRNA) sequences were used to knockdown ETHE1 in these cell lines. Western blot ( Figure  1C and 1D) and densitometry analysis ( Figure 1E and 1F) confirmed increased ETHE1 protein levels (p=0.001 unpaired t-test) in constitutively expressed HCT116 and HT29 ( Figure 1E and 1F) cells.
Next, we tested whether ETHE1 was enzymatically active in CRC cells. ETHE1 facilitates H 2 S catabolism via oxidation of sulfide quinone reductase generated glutathione persulfides (GSSH) to sulfite (H 2 SO 3 ), in a reaction that requires oxygen and water. Sulfite is further oxidized to sulfate and secreted. Several studies have shown that GSSH is the main substrate for ETHE1 [35][36][37][38]. Therefore, we mixed freshly prepared 735 µM glutathione persulfide (GSSH) with 200 µg cell lysate under air-saturated conditions at pH 7.4 and measured the disappearance rate of GSSH using an established monobromobimane (MBB) method with LC/MS [35]. In both ETHE1 overexpressing CRC cell lines, metabolism of GSSH was increased compared to ETHE1 scrambled control cell lysates (HT29-CE p=0.0001; HCT116-CE p=0.007) ( Figure 1G). Conversely, ETHE1 knockdown CRC cells showed reduced rates of GSSH consumption relative to controls. Thus, ETHE1 actively reduces H 2 S metabolites in CRC cell lines.

ETHE1 protein levels correlate with FAP mitochondrial protein enrichment
We performed a re-analysis of 214 differentially expressed two-dimensional gel protein spots identified in morphologically normal, "single-hit" FAP and low CRC risk colon epithelial cells datasets [16]. Using this approach, we identified 66 non-redundant proteins with a threshold of > 1.5X up-regulation (p<0.01) in the "single-hit" FAP cells. The annotation cluster with the highest enrichment (enrichment score 14.95) for ETHE1 upregulation in these cells was identified as "Mitochondrial proteins" (Supplementary Table 1). This subset of upregulated proteomic annotation cluster categories included mitochondrial transit peptide containing proteins (p=1.62E-15), mitochondrial lumen and matrix proteins (p=7.63E-15), as well as envelope (outer membrane) and mitochondrial inner membrane proteins (Supplementary Table 1).
To test further mitochondrial involvement, we compared the ultrastructural characteristic of phenotypically normal, intact FAP colon tissue specimens, with matched colon tissues from control individuals by transmission electron microscopy (EM). Phenotypically normal FAP colon epithelial biopsies showed densely distributed mitochondria throughout the cytoplasm and increased mitochondrial matrix staining, when compared to intact colon controls (Figure 2A and 2B). Next, we examined CRC cell lines with different ETHE1 expression levels for their ultrastructural features. Similar to intact FAP colon biopsies, ultrastructural analysis of constitutively expressing ETHE1 HCT116 ( Figure 2C and 2D) and HT29 cells ( Figure 2F and 2G) showed increased mitochondrial matrix staining. In contrast, silencing of ETHE1 reduced mitochondrial matrix density in HCT116 ( Figure 2E) and HT29 ( Figure  2H) cells, compared with scrambled controls. Overall, we show a strong correlation between elevated ETHE1 levels and increased mitochondrial density and matrix in histologically normal colonic epithelia from FAP patients and in CRC cells.

ETHE1 upregulates oxidative phosphorylation in CRC
Since the mitochondrial matrix plays a critical role to convert energy derived from oxidation/ reduction reactions of electron transport, we examined the expression levels of key mitochondrial oxidative phosphorylation (OXPHOS) related proteins in whole cell homogenates that were derived from phenotypically normal biopsied tissue specimens of FAP patients and control individuals. Phenotypically normal FAP patient lysates exhibited increased (P=0.001; 1-way ANOVA) levels of ATP synthase (ATPase) subunits, succinate dehydrogenase (SDBH), cytochrome c reductase (QH2), cytochrome c oxidase (COX) and ubiquinone oxidoreductase (NADH) ( Figure 2I) protein levels as compared with low CRC risk controls. Previous studies have shown higher COXI expression in CRC cells than in normal colon epithelia, indicating a role for increased cytochrome c oxidase (COX) expression in the transformation of colon epithelia [39]. Consistent with these reports, densitometric analysis of FAP vs sporadic western blot data, showed a ~15 fold increase in COX expression (paired t-test,p=0.001) in the phenotypically normal intact FAP colon biopsies ( Figure 2K).
To investigate whether ETHE1 levels were directly associated with this subset of mitochondrial protein expression, we probed constitutively expressing ETHE1 CRC cell lines for OXPHOS levels. Analysis of ETHE1 CRC cell homogenates showed that constitutive upregulation of ETHE1 in HCT116 and HT29 cells increased OXPHOS protein levels ( Figure  2J). Densitometric analysis revealed that upregulation of ETHE1 in HCT116 ( Figure 2L) and HT29 ( Figure  2M) significantly increased (>2x, p=0.001) expression of COXI, SDBH and NADH. This effect was attenuated by shRNA-mediated silencing of ETHE1 CRC cells, resulting in approximately a 3x to 10x fold reduction (P<0.05) of COXI expression in HT29 and HCT116 ( Figure 2L and 2M), respectively. Taken together, these results are consistent with a direct role of ETHE1 to drive mitochondrial OXPHOS protein expression and mitochondrial content.

Overexpression of ETHE1 enhances CRC mitochondrial respiration and aerobic glycolysis
Our results showed that phenotypically normal, FAP colon biopsies and CRC cell lines contained large numbers of mitochondria with increased mitochondrial matrix staining, consistent with metabolically highly active mitochondria. To investigate the role of elevated ETHE1 levels as a driver of mitochondrial respiration, we used the Seahorse XF96 extracellular flux analyzer on CRC cells with elevated, normal and knocked-down levels of ETHE1. Using this approach, we generated mitochondrial bioenergetics profiles from HCT116 ( Figure 3A) and HT29 ( Figure 3B) cells. Constitutive expression of ETHE1 significantly increased HCT116 basal respiration (~51%), ATP turnover (~57.5%), spare respiratory capacity (~113%) and maximal respiration (~67.7%) ( Figure 3C-3F) compared with scrambled parental control cells. It also increased HT29 basal respiration (~84%), ATP turnover (~93.9%), maximal respiration (~88.1%) and spare respiratory capacity (~95.5%) ( Figure 3G-3J). Notably, all CRC cell lines, regardless of ETHE1 expression levels demonstrated a positive association between increased oxygen consumption rate (OCR) and extra-cellular acidity rate (ECR), indicating the concurrent occurrence of enhanced mitochondrial respiration and aerobic glycolysis ("Warburg effect") ( Figures 3K and 3L). Yet this association appeared to trend most clearly in the case of constitutively enriched ETHE1 cell lines ( Figures 3K and  3L). In this context, shRNA silencing of ETHE1 reduced basal respiration, ATP turnover, spare respiration capacity, and maximal respiration. Altogether, our data demonstrate that upregulation of ETHE1 increased not only protein levels of key mitochondrial oxidative phosphorylation related proteins, but has also contributed to metabolic adaptations associated with tumor malignancies wherein both aerobic glycolysis and respiratory capacity are implicated [40].
First, we therefore determined SIRT1 and PGC1α protein levels in cell homogenates from phenotypically normal, intact FAP colon tissue biopsies. SIRT1 and its target PGC1α protein levels were significantly increased in FAP colon verses normal colon controls ( Figure 4A). Next, we determined PGC1 expression in HCT116 and HT29 scrambled shRNA control cells ( Figure  4B). We then examined constitutively expressing ETHE1 HCT116 ( Figure 4C) and HT29 ( Figure 4D) cells to test whether SIRT1 and PGC1α levels were directly regulated by ETHE1. Similar to intact FAP tissue biopsies, SIRT1 and PGC1α were upregulated in constitutively expressed ETHE1 cell lines. In contrast, shRNA silencing of ETHE1 reduced both PGC1α and SIRT1 protein levels in CRC cells.
In addition, we determined SIRT1 deacetylase activity in high and low ETHE1 expressing CRC cells. We prepared nuclear extracts from constitutively expressing ETHE1 HCT116 and HT29 cells using a SIRT1 activity assay kit ( Figure 4E). The results showed that, consistent with increased SIRT1 levels, SIRT1 functional deacetylase activity was also upregulated in high ETHE1 expressing HCT116 and HT29 cells. In contrast, shRNA silencing of ETHE1 reduced SIRT1 activity ( Figure 4E). Thus, ETHE1 levels can positively drive SIRT activity.

PDE mediated stimulation of pAMPK/SIRT1 is linked to ETHE1 levels
H 2 S inhibits phosphodiesterase activity, affecting the rate of c-AMP conversion to AMP [34,43,44]. Building upon our findings that overexpression of ETHE1 increased the rate of GSSH catabolism in CRC cell lysates ( Figure 1G), we then determined whether H 2 S levels can alter phosphodiesterase activity. We first measured the levels of total GSSH in constitutively expressing ETHE1 HCT116 and HT29 using the GSSG-Glo assay. Consistent with increased ETHE1 -mediated H 2 S catabolism, H 2 S levels were decreased in high ETHE1 expressing HCT116 ( Figure 5A) and HT29 ( Figure 5B) cells, as determined by GSSG levels. In contrast, shRNAmediated silencing of ETHE1 CRC upregulated GSSH levels. Next, we determined PDE activity in high and low ETHE1 expressing cells using a cAMP-Glo assay. These results showed that cAMP levels were decreased in high ETHE1 expressing HCT116 ( Figure 5C) and HT29 cells ( Figure 5D), consistent with increased PDE activity and conversion of cAMP to AMP.
To study the mechanism through which ETHE1 decrease of H 2 S promotes the AMPK/SIRT1 axis, including increased mitochondrial biogenesis and function, we used a Cystathionine-β-synthase (CBS) inhibitor, aminooxyacetic acid (AOAA). CBS is a H 2 S producing enzyme, shown to be overexpressed in CRC cells [26,45]. We, therefore, prepared total cell extracts from constitutively expressing ETHE1 HCT116 and HT29 cells treated with AOAA (1-10mM) and measured AMPKp and SIRT1 protein levels. We found that inhibition of CBS activity resulted in increased AMPKp, SIRT1 and PGC1α expression in these cells ( Figure 5E-5H). Overall, these results are consistent with increased ETHE1 catabolism of endogenous H 2 S, leading to increased PDE activity and AMPKp/SIRT1 expression. Since Cystathionine-βsynthase (CBS) is an important contributor to intracellular H 2 S production (24), we also examined whether low or high ETHE1 levels altered CBS expression in CRC cells. Western blot analysis showed that changes in ETHE1 levels had no effect on CBS expression in HT29 (Supplementary Figure 1A) or HCT116 cells (Supplementary Figure 1B). Therefore, increased ETHE1 levels do not appear to notably affect the CBS pathway.

ETHE1 stimulates oncogenesis of CRC cells in vitro and in vivo
Increased expression of SIRT1 and PGC1α increase cell proliferation and angiogenesis [46,47]. Since ETHE1 upregulated SIRT1 and PGC1α, we tested whether cell proliferation was affected by ETHE1 expression levels using BrdU incorporation assays. The results show that high expressing ETHE1 HCT116 ( Figure 6A) and HT29 ( Figure 6B) cells exhibit increased cell proliferation in vitro.
Next, we performed xenograft studies to test the effects of increased ETHE1 expression on angiogenesis. Male NOD/Scid mice were injected in either flank (S.C) with constitutively expressing ETHE1 HT29/HCT116 or scrambled control cells. The percentage of KI67 positive tumor cell nuclei was counted in 4 × random images (magnification 20x) ( Figure 6C, 6D). The results show that, in contrast to scrambled controls, high ETHE1 expression significantly increased HT29 and HCT116 Ki67 + (122-194%) staining ( Figure 6E and 6F). In addition, we found that high ETHE1 expression significantly increased the number of vascular endothelial CD31+ cells ( Figure 6G and 6H) in xenograft tumors by >2 fold. Taken together, these data are consistent with a role for ETHE1 to increase not only CRC cell proliferation, aerobic glycolysis and OxPhos, but also angiogenesis in vivo.
To assay the effects of high ETHE1 expression on CRC tumorigenicity, we injected (via lateral tail vein) male non-obese diabetic/severe combined immunodeficient (NOD/SCID) mice with constitutively expressing ETHE1 HCT116 (1x10 6 ) or HT29 (1x10 6 ) cells. High ETHE1 expression increased intestinal and extra-colonic HCT116 and HT29 tumor multiplicity and tumor size (4x fold) compared with scrambled control parental cells ( Figure 6I-6L). Importantly, mice injected (via lateral tail vein) with high expressing ETHE1 HCT116 or HT29 cells showed reduced overall survival ( Figure 6M and 6N). These results are consistent with ETHE1 levels as a mechanistic driver of colorectal cancer growth in vivo.

DISCUSSION
Previous studies have shown that germline ETHE1 mutations in ethylmalonic encephalopathy lead to an accumulation of H 2 S and inhibition of Cytochrome C Oxidase (COX) which is essential for mitochondrial respiration, affecting severe dysfunction in cellular energy metabolism [18,48,49]. While the effects of ETHE1 deficiency have been thoroughly investigated, the consequences of increased ETHE1 expression in CRC tumorigenesis are poorly understood. In hepatocellular carcinoma (HCC), ETHE1 is overexpressed [50], where it is associated with inhibition of caspase 9 activation, suppressing DNA-damage induced apoptosis by binding to the RelA p65 subunit of NFkB, and increasing its export from the nucleus [50]. However, in our initial studies we did not see activation of this pathway in phenotypically normal intact colon mucosa from FAP patients (cancer initiation), indicating that increased ETHE1 expression involves different mechanisms in different cancer types.
Similar to previous proteomic studies on "singlehit" FAP/APC normal colon epithelial cells [16], our data show a significant increase in f ETHE1 expression in phenotypically normal intact mucosa tissue from FAP patients, including an associated increase in mitochondrial content/matrix, as compared with normal colon mucosa from control subjects. The impact of ETHE1 overexpression in early in situ CRC tumorigenesis on mitochondrial proteins was further illustrated through re-analysis of differentially expressed two-dimensional gel protein spots identified in phenotypically normal, "single-hit" FAP epithelial cells [16]. Using this approach, we identified 66 non-redundant proteins in "single-hit"FAP epithelial cells in which the annotation cluster with the highest enrichment (enrichment score  Table 1), attesting to the strong correlation between elevated ETHE1 levels and increased mitochondrial proteins in "single-hit" FAP cells. These data are consistent with our current observations regarding ETHE1 over expression and increased mitochondrial biogenesis seen in phenotypically normal FAP mucosa specimens occurring in situ.
Accordingly, we investigated the metabolic impact of ETHE1 expression in CRC cell lines at baseline and after knockdown or constitutive overexpression. Increased ETHE1 expression in CRC cells stimulated mitochondrial biogenesis/bioenergetics and increased mitochondrial matrix staining, consistent with an overall increased mitochondrial activity. Thus, constitutively upregulated ETHE1 in CRC cells share phenotypic similarities with alterations seen in FAP patient normal mucosa epithelium in situ. By contrast, mitochondrial matrix, OXPHOS protein expression and respiration levels were reduced following shRNA silencing of ETHE1 in CRC cells. These data are consistent with a novel mechanistic role for ETHE1 function in which, in addition to H 2 S catabolism, if abnormally expressed, it promotes mitochondrial biogenesis and cellular respiration.
Mitochondria are key metabolic and biosynthetic regulators that are intimately involved in cancer cell growth. SIRT1, AMPK and PGC1α are important regulators of mitochondrial biogenesis. The NAD+ dependent deacetylase SIRT1 and AMPK have been shown to upregulate PGC1α activity [27,29,31,32,43]. Thus, our data point to a basic pathway that underlies ETHE1 overexpression in CRC cells. Our data support a process whereby increased H 2 S catabolism lowers H 2 S levels, which increases PDE activity. This leads to increased AMP, which promotes phosphorylation of AMPK, leading AMPKp and SIRT1 directly to stimulate PGC1α. Together, AMPKp and SIRT1 increase PGC1α and promote CRC mitochondrial biogenesis and respiration, including aerobic glycolysis (the "Warburg hypothesis") ( Figure 7). In fact cellular respiration and aerobic glycolysis have been shown to occur concurrently in many tumors, and although aerobic glycolysis is (A) and HT29 (B) cells expressing low (KD), scrambled control (WT) or high ETHE1 (OE) were used to measure H 2 S, as determined by GSSG levels (GSH/GSSG Promega), ETHE1 high expressing cells showed reduced GSSG, suggesting reduced H 2 S levels. shRNA silencing of ETHE1 showed increased GSSH levels; N=3; unpaired T-test. PDE activity determined by cAMP-Glo Assay (Promega), (cAMP levels are inversely proportional to luminescence); HCT116 (C) and HT29 (D) ETHE1 overexpressing cells (seeded 30K in 96 well plates) show reduced cAMP levels (high luminescence), suggesting increased PDE activity, shRNA silencing of ETHE1 increased cAMP levels, as compared with wild type control. N=3; unpaired T-test. Total protein extracts from HCT116 (E) and HT29 (F) CRC cells treated with the CBS inhibitor aminooxyacetic acid (AOAA), showed increased activation of AMPK (pAMPK) and Sirt1 expression. PGC1α levels in HCT116(G) and HT29(H) ETHE1 KD/WT/OE CRC cells treated with AOAA. www.oncotarget.com  an inefficient way to generate ATP, most cancer cells depend on it [40]. Supporting this notion are recent studies showing that proliferating cells require metabolic pathways that incorporate nutrients into biomass, and that cancer-associated mutations enable cancer cells to support proliferation rather than efficient ATP production [40,[51][52][53]. We showed a positive association between increased aerobic glycolysis and cell respiration that is further impacted by ETHE1. Thus, ETHE1 alters the metabolism of cancer cells in order to facilitate increased biomass and cell growth, increased cellular respiration notwithstanding.
Concordant with our results in vitro, in vivo studies showed that constitutive expression of ETHE1 in CRC cells increased KI67 + /CD31 + staining in NOD/SCID xenograft mouse models. Additionally, mice injected (via lateral tail vein) with constitutively expressing ETHE1 HCT116 (1x10 6 ) or HT29 (1x10 6 ) cells show reduced median survival time, increased tumor incidence, tumor burden, multiplicity and metastasis in vivo. Significantly, ETHE1 null (KO) CRC cells showed much reduced tumor cell growth and increased median survival time in vivo compared to WT or ETHE1 overexpressing CRC cells. These data provide evidence that upregulation of ETHE1 plays an important role in driving pathogenesis of CRC in vivo.
Previously, ETHE1 protein levels were shown to be elevated in FAP normal colonic mucosa (15). Our findings here confirm these original findings in additional FAP colon biospecimens and extend our current mechanistic understanding of the role of ETHE1 in FAP CRC tumorigenesis with both in vitro and in vivo studies, leading to the following salient conclusions: First, because ETHE1 promotes CRC cell growth, and is upregulated in phenotypically normal APC +/-FAP colon epithelium, high ETHE1 levels may be useful as an early biomarker of CRC premalignant risk. Second, ETHE1 may be a potential chemoprevention target for FAP and high-risk sporadic CRC patients. Third, there has been a significant amount of interest regarding the potential influence of gastrointestinal microbiota on CRC tumorigenesis (50)(51)(52)(53)(54). Specifically, microbial species associated with CRC risk include fusobacteria, a genus of anaerobic gram-negative bacteria, and in particular fusobacterium nucleatum which produces H 2 S [54][55][56][57][58]. Thus, paracrine H 2 S production by colonic epithelium adherent Fusobacteria nucleatum driving ETHE1 upregulation may constitute an additional mechanism of CRC tumorigenesis for this important bacterium.

ETHE1 overexpressing CRC lines
Constitutive expression of ETHE1 in HT29 and HCT116 cells, shRNA knockdown and scrambled shRNA control cells were established using a Lentivector-based system according to the manufacturer's instructions (GeneCopoeia Inc., Rockville, MD, USA). For lentiviralmediated gene transfer, Lv-101 T298) expression vector and transfected into HEK293 cells for pseudoviral packaging. Supernatants were collected and purified viral particles were titered and used to infect cells.

FAP tissue specimens
Phenotypically normal sigmoid colonic epithelium from FAP patients matched patient age, gender and ethnicity (European ancestry) with patients having screening colonoscopy with no personal, first or second degree history of CRC were obtained from the New York Presbyterian Hospital Center for Advanced Digestive Care Biobank under an IRB approved protocol (0908010582).

Metabolic flux assay
Cells seeded in triplicate in 100 μl DMEM at a density of 1 × 10 4 /well in XF96 96-well cell culture plate (Seahorse Bioscience, North Billerica, MA) and incubated at 37°C for 24hrs. The XF96 sensor cartridge (Seahorse Bioscience) was prepared by incubation of each sensor pair in 1 ml of Seahorse Bioscience XF96 Calibrant pH 7.4 (Seahorse Bioscience) at 37°C without CO 2 for 24 h. Prior to analysis, the medium was removed and the wells washed with 1 ml of pre-warmed Seahorse DMEM supplemented with 5 mM glucose. A final volume of 180 μl assay medium was added to each well and the plate incubated at 37°C without CO 2 for 1 hr. Either 1.5 μM carbonyl cyanide-p-trifluoromethoxyphenylhydrazone (FCCP), 5 μM oligomycin (Sigma-Aldrich, St. Louis, MO) or 5 μM antimycin was added to injection ports of the XF96 sensor cartridge in assay medium, equilibrated at least 15 min prior to analysis at 37°C without CO 2. The XF96 sensor cartridge was calibrated in the Seahorse XF96 analyzer (Seahorse Bioscience), which was pre-warmed to 37°C. Following calibration, the cell culture plate was placed in the analyzer and the extracellular acidification rate (ECAR) and the oxygen consumption rates (OCR) were simultaneously measured via the following protocol: three cycles of mix (2 min), delay (2 min); seven cycles of measure (4 min), mix (2 min), delay (2 min); port A injection; six cycles of mix (2 min), delay (2 min), measure (4 min). Post-drug injection values were normalized to respective basal value and expressed as either mpH/ min/10 4 cells (ECAR) or pmoles/min/10 4 cells (OCR).

Animals
All animal studies and procedures done in accordance with protocols approved by WCMC. NOD/ SCID mice purchased from Jackson Laboratory. CCIC 707 or 823 (1 × 10 6 ) injected directly into the tail vein of 7to 9-week-old male NOD/SCID mice. Tumor burden was quantified by manually counting nodules visible on the intestinal epithelia surface. 8 week-old NOD/SCID female mice were used for ETHE1 CRC variant xenograft studies.

Fixation & embedding for electron microscopy
Cells were washed with serum-free media or appropriate buffer, fixed with modified Karmovsky's; 2.5% glutaraldehyde, 4% parafomaldehye and 0.02% picric acid in 0.1M sodium caocdylate buffer at pH 7.2. Secondary fixation in 1% osmium tetroxide, 1.5% potassium ferricyanide samples were dehydrated through a graded ethanol series, and embedded in an epon analog resin. Ultrathin sections were cut using a Leica UltractuS ultramicrotome (Leica, Vienna, Austria) Sections collected on copper grids and contrasted with lead citrate and viewed on a JEM 1400 electron microscope (JEOL, USA, Inc., Peabody, MA) operated at 120 kV. Images were recorded with a Veleta 2K x2K digital camera (Olympus-SIS, Germany).

Q-TOF LC-MS and LC-MS/MS
Mass Spectroscopy (MS) was performed as previously performed in [59] Proliferation assay Cells were seeded in 96 well clear plates at 50,000 cells/well. After 24 hrs fresh cell culture media containing 1x Brdu solution (cell biosciences) was added to media. Cell number was determined by measuring absorption at 450nM.

GSSG and cAMP Assay
Cells were seeded in 96 well white plates 50,000 cells/well. After 24 hrs, cells are lysed to release cAMP or total GSSG (Promega). cAMP or GSSH was determined by Luminesce using a FLUOstar Omega Multi-Mode Microplate Reader with CCD-based Spectrometer.

Proteomic and ingenuity pathway reanalysis
Reanalysis of the 214 protein spots that were identified as changed in intensity in FAP compared with the control group with p<0.01 and fold changes each > half fold. The David Bioinformatics Gene Ontology at NIH is one of the best accepted software for this analysis. The UNIPROT_ACCESSION names of the up or down proteins in spreadsheet "up vs down in 214" were submitted as gene list to David Bioinformatics site (http://david.abcc.ncifcrf.gov/summary.jsp) for "functional annotation clustering" analysis, using the human genome as background. The software calculates how much of the data fits into each protein category, according to Gene Ontology (GO terms) or KEGG protein pathway database.

Author contributions
Anthony Yeung and Levy Kopelovich made the original discovery regarding ETHE1 overexpression in FAP, leading to this conceived work and subsequent submission. Mavee Witherspoon, Davinder Sandu and Changyuan Lu acquired data. All authors played an important role interpreting the results. Mavee Witherspoon, Steven Lipkin and Levy Kopelovich drafted the manuscript. Steven Lipkin planned the project as principal investigator. All authors approved the final version.