Depletion of the Human Ion Channel TRPM2 in Neuroblastoma Demonstrates Its Key Role in Cell Survival through Modulation of Mitochondrial Reactive Oxygen Species and Bioenergetics*

Transient receptor potential melastatin 2 (TRPM2) ion channel has an essential function in modulating cell survival following oxidant injury and is highly expressed in many cancers including neuroblastoma. Here, in xenografts generated from neuroblastoma cells in which TRPM2 was depleted with CRISPR/Cas9 technology and in in vitro experiments, tumor growth was significantly inhibited and doxorubicin sensitivity increased. The hypoxia-inducible transcription factor 1/2α (HIF-1/2α) signaling cascade including proteins involved in oxidant stress, glycolysis, and mitochondrial function was suppressed by TRPM2 depletion. TRPM2-depleted SH-SY5Y neuroblastoma cells demonstrated reduced oxygen consumption and ATP production after doxorubicin, confirming impaired cellular bioenergetics. In cells in which TRPM2 was depleted, mitochondrial superoxide production was significantly increased, particularly following doxorubicin. Ectopic expression of superoxide dismutase 2 (SOD2) reduced ROS and preserved viability of TRPM2-depleted cells, however, failed to restore ATP levels. Mitochondrial reactive oxygen species (ROS) were also significantly increased in cells in which TRPM2 function was inhibited by TRPM2-S, and pretreatment of these cells with the antioxidant MitoTEMPO significantly reduced ROS levels in response to doxorubicin and protected cell viability. Expression of the TRPM2 pore mutant E960D, in which calcium entry through TRPM2 is abolished, also resulted in significantly increased mitochondrial ROS following doxorubicin treatment, showing the critical role of TRPM2-mediated calcium entry. These findings demonstrate the important function of TRPM2 in modulation of cell survival through mitochondrial ROS, and the potential of targeted inhibition of TRPM2 as a therapeutic approach to reduce cellular bioenergetics, tumor growth, and enhance susceptibility to chemotherapeutic agents.

Transient receptor potential (TRP) 3 channels are a superfamily of monovalent and divalent cation-permeable ion channels. Members of the melastatin subfamily, TRPM, have been shown to have important roles in cell proliferation and survival (1)(2)(3)(4)(5). The second member of this subfamily to be cloned, TRPM2, is a 1503-amino acid channel permeable to Ca 2ϩ , Na ϩ , and K ϩ (6) and is widely expressed in many cell types including brain, hematopoietic cells, heart, and endothelial cells (7,8). Its role in normal physiological conditions as well as in oxidative stress is under investigation in many cell types (9). Oxidative stress, TNF␣, and amyloid ␤-peptide are among the key extracellular signals known to activate TRPM2 (10 -13). Stimulation with these extracellular signals results in production of adenosine diphosphate-ribose (ADPR), which activates the channel by binding to the TRPM2 C-terminal NUDT9-H domain (7,(13)(14)(15)(16)(17)(18)(19). A rise in intracellular Ca 2ϩ and calmodulin also positively regulate TRPM2 (20 -22), and acidification can inhibit channel activity (23)(24)(25). TRPM2 functions as a tetramer and six splice variants of full-length TRPM2 (TRPM2-L) have been discovered to date, although their physiological roles in modu-lation of its function are not known (12, 26 -29). TRPM2-S (short, 845 residues) is missing four C-terminal transmembrane domains and the C terminus inhibits calcium influx through the full-length channel, and functions as a dominantnegative isoform (26).
Oxidative stress results from imbalance between the amount of oxidants produced and antioxidants, leading to tissue injury depending on severity and duration (7). Major sources of reactive oxygen species (ROS) are the mitochondrial electron transport chain and NADPH oxidases (NOX), which are key regulators of non-mitochondrial ROS. ROS play a major role in tissue injury through protein oxidation, lipid peroxidation, and DNA oxidation and mutagenesis. A number of chemotherapeutic agents including doxorubicin contribute to cell death by increasing ROS (30,31).
Although early studies supported the paradigm that activation of TRPM2 induced cell death through a sustained increase in intracellular calcium (10,32) or enhanced cytokine production aggravating tissue injury (33), more recent reports suggested that calcium entry through TRPM2 may be protective rather than deleterious in a number of physiological model systems. In pyramidal neurons subjected to oxidant injury, inhibition of TRPM2 enhanced cellular damage (34). In wild type mice challenged with intraperitoneal injection of endotoxin, the survival was five times greater than that of TRPM2 knockout mice due to reduced NOX-mediated ROS production in wild type phagocytes (35). TRPM2 channels also protected the heart from ischemia/reperfusion (I/R) injury, and the mechanism was through maintenance of mitochondrial function and reduced levels of mitochondrial oxidants (36 -38). TRPM2 channels are highly expressed in a number of cancers including melanoma (39), breast cancer, 4 and neuroblastoma (40 -42). Previously, TRPM2 function has been studied in cancer cells with either shRNA down-regulation or expression of the dominant-negative TRPM2-S (40,41). TRPM2-L protected neuroblastoma SH-SY5Y cells from physiological oxidative stress through increased levels of Forkhead box transcription factor 3a (FOXO3a) and superoxide dismutase 2 (SOD2, MnSOD) (40), whereas cells expressing the dominant-negative TRPM2-S isoform had reduced FOXO3a and SOD2 levels and enhanced ROS, increasing susceptibility to oxidant injury. Tumor growth was diminished by expression of dominant-negative TRPM2-S in a xenograft model, and the mechanisms involved reduced HIF-1/2␣ expression and mitochondrial function (41). Reduction of TRPM2 with shRNA may yield inconclusive results if depletion is not complete or sustained, and results depend on the level of endogenous protein required for function. The use of dominant-negative constructs has additional potential pitfalls, which include the possibility of off target effects and the risk of incomplete inhibition. Because the effect of a dominantnegative is competition-based, it can potentially be modified by increased expression of the target protein.
Here, we utilized a fundamentally different approach to study the role of TRPM2 in cancer cell growth, survival, and bioenergetics, namely, depletion of TRPM2 with targeted CRISPR/Cas9 technology. TRPM2 was knocked out in neuroblastoma cells as a model to explore its role in tumor growth and cell survival after doxorubicin. Our major findings are: 1) depletion of TRPM2 significantly inhibits tumor growth and increases chemotherapy sensitivity; 2) TRPM2 depletion results in decreased levels of HIF-1/2␣ and downstream proteins that modulate oxidant injury and mitochondrial bioenergetics; 3) mitochondrial ROS levels are increased and play a key role in reduced cell survival in neuroblastoma tumor cells in which TRPM2-L is depleted or inhibited; and 4) calcium entry through TRPM2 plays an important role in modulation of cell viability, mitochondrial bioenergetics, and mitochondrial ROS levels. A recent search of the OncoGenomics database (Khan lab, Center for Cancer Research, National Cancer Institute) complements these findings and showed that low TRPM2 expression in human neuroblastoma tumors was associated with significantly better patient survival than high expression. The experimental findings presented here strongly support the conclusion that targeted inhibition of TRPM2 function in cancer cells has a significant impact on tumor growth and response to chemotherapy through modulation of mitochondrial ROS and cellular bioenergetics.

Results
Generation and Characterization of TRPM2-depleted Neuroblastoma Cells-To study the function of TRPM2 in neuroblastoma tumor growth and chemotherapy sensitivity, we generated SH-SY5Y neuroblastoma cells in which TRPM2 was depleted with CRISPR technology. The area chosen for deletion of TRPM2 genomic DNA encoded the first 40 amino acids and the remaining TRPM2 sequence was frameshifted. Deletion of this segment was confirmed by real-time reverse transcriptasepolymerase chain reaction (RT-PCR), and by Western blotting of TRPM2-L protein (Fig. 1, A and B).
Cationic current was characterized in SH-SY5Y cells in which TRPM2 was depleted with CRISPR/cas9 technology and in scrambled SH-SY5Y control cells. Under our experimental conditions, intracellular application of ADPR (300 M) elicited large inward and outward cation currents in scrambled SH-SY5Y control cells expressing endogenous TRPM2, but not in TRPM2-depleted cells (Fig. 1C). ADPR-activated current displayed the characteristic TRPM2 linear I-V relationship with reversal potential close to 0 mV (43,44). Omission of ADPR in pipette solutions resulted in nearly absent current in scrambled control cells, similar to TRPM2-depleted cells exposed to ADPR. Together, these data demonstrate that (i) ADPR-activated current is mediated through endogenous TRPM2, and (ii) TRPM2 current activated by ADPR was successfully reduced in SH-SY5Y cells in which TRPM2 was depleted with CRISPR. Reduced current in TRPM2-depleted cells was similar to that observed in cells expressing the dominant-negative isoform TRPM2-S; TRPM2-S does not have the calcium pore and inhibits calcium entry through TRPM2-L (41).
TRPM2 Depletion Reduced Cell Proliferation and Increased the Sensitivity of Neuroblastoma Cells to Doxorubicin-To evaluate the impact of TRPM2 depletion on SH-SY5Y cell growth, two individual clones in which TRPM2 was depleted and two scrambled control clones were cultured. By 2,3-bis(2-methoxy-4-nitro-5-sulfophenyl)-2H-tetrazolium-5-carboxanilide (XTT) analysis, cells in which TRPM2 was depleted grew slower than scrambled control cells (also grown in puromycin), and had significantly greater sensitivity to doxorubicin at 24 h ( Fig. 2A). Because XTT measures cell viability through detection of cellular metabolic activity, cell number was also quantified by trypan blue exclusion. Live cell number was variably greater in scrambled controls cells compared with TRPM2-depleted cells; doxorubicin treatment resulted in a significant reduction in live TRPM2-depleted cells compared with live scrambled control cells (Fig. 2B). These in vitro studies demonstrate that cell proliferation and viability after doxorubicin are significantly decreased when TRPM2 is depleted with CRISPR technology.
To rule out the possibility that these observations were due to secondary off-target effects occurring during CRISPR/Cas9 treatment or subsequent selection, SH-SY5Y cells in which TRPM2 was depleted (KO) were transfected with empty vector, wild type TRPM2, or the TRPM2 inactive pore mutant E960D. The E960D construct was previously generated by our labora-tory and the loss of function has been authenticated (38,45). Expression of wild type TRPM2 but not E960D in the KO reconstituted cell viability at or close to the level observed in the scrambled control (Fig. 2C), demonstrating that the reduced viability in the KO was not due to off-target effects, and the critical role of TRPM2-mediated calcium entry.
To evaluate the effects of TRPM2 depletion on xenograft tumor growth, athymic female mice were injected in one flank with either two different clones of TRPM2-depleted SH-SY5Y cells, two scrambled control clones, or parental wild type (WT) SH-SY5Y cells. Tumor size was measured twice weekly for 6  NOVEMBER 18, 2016 • VOLUME 291 • NUMBER 47

TRPM2 Depletion Inhibits Mitochondrial Function and Survival
weeks. Tumors from TRPM2-depleted cells demonstrated significantly reduced tumor volume (Fig. 3A) and tumor weight ( Fig. 3B) compared with scrambled control cells, which were similar to SH-SY5Y parental wild type cells. Differences in tumor sizes were visually apparent (Fig. 3C). These in vivo studies demonstrate that tumor growth is significantly decreased when TRPM2 is depleted. HIF-1/2␣ and Downstream Signaling Pathways Are Downregulated in TRPM2-depleted Xenografts-To determine the mechanisms responsible for the reduced growth of TRPM2depleted tumors, Western blotting was performed on lysates of xenograft tumors harvested 6 weeks after SH-SY5Y cell injection. A statistically significant decrease in expression of the transcriptional regulators HIF-1␣ and HIF-2␣ was observed in tumors from cells in which TRPM2 was depleted. Scrambled control cells expressing endogenous TRPM2-L demonstrated more HIF-1␣ and -2␣ (Fig. 4) and formed larger tumors (Fig. 3), consistent with reports that HIF-2␣ promotes a more aggressive tumor phenotype in patients with neuroblastoma (46,47). Western blotting confirmed modulation of expression of proteins downstream of HIF-1␣ including reduced expression of antioxidants (SOD1/2, FOXO3a) (48,49), mitochondrial proteins (Bcl2/adenovirus E1B 19-kDa interacting protein 3, BNIP3), proteins involved in mitochondrial electron transport chain activity (complex IV, cytochrome oxidase 4.1/4.2, COX4.1/4.2) (50 -55), and in glycolysis (enolase 2, ENO2) in TRPM2-depleted cells. By contrast, there were no differences in Nox2 or Nox4 between knock-out and scrambled cells, suggesting that TRPM2 is unlikely to be involved in regulation of Nox activities. These data show that TRPM2 plays an important role in regulation of HIF-1/2␣ expression and downstream signaling transducers, many of which are involved in regulation of oxidant stress, mitochondrial function, and bioenergetics that modulate cell survival.
Mitochondrial O 2 Consumption Is Lower in TRPM2-depleted Neuroblastoma Cells-Because of reduced mitochondrial protein expression in TRPM2-depleted SH-SY5Y neuroblastoma cells, we examined whether mitochondrial complex activities are also affected. Basal and maximal oxygen consumption rates (OCR) were significantly reduced in TRPM2-depleted cells compared with scrambled control cells or wild type parental cells (Fig. 5). These findings suggest that when TRPM2 expression or function is reduced, the mitochondrial oxygen consumption rate is decreased. Because doxorubicin reduces viability of TRPM2-depleted cells, the effect of doxorubicin on OCR was examined. Basal and maximal OCRs were measured in TRPM2-depleted and scrambled control cells with and without doxorubicin treatment. Doxorubicin treatment significantly reduced OCRs in all cells, but the OCR of TRPM2depleted cells was significantly less than scrambled after doxorubicin. Results of a representative experiment using one knock-out and scrambled control clone are shown (Fig. 6A).
Neuroblastoma Cells Depleted of TRPM2 Have Lower ATP Levels-Because TRPM2-depleted SH-SY5Y cells demonstrated reduced OCR, the bioenergetic capacity was further examined by measurement of ATP. ATP levels were measured in TRPM2-depleted or scrambled control cells at 24 h, and 48 h after doxorubicin (0.3 M) treatment. ATP levels progressively declined with time after doxorubicin treatment in all cells, but the rate of decline was significantly faster in TRPM2-depleted compared with scrambled control cells (Fig. 6B). These findings demonstrate that endogenous TRPM2 is important in mainte- nance of cellular bioenergetics after doxorubicin treatment. SH-SY5Y cells in which TRPM2 was depleted (KO) were transfected with empty vector, wild type TRPM2, or the TRPM2 inactive pore mutant E960D (Fig. 6C). Expression of wild type TRPM2 but not E960D in the KO-reconstituted ATP production, at or close to the level observed in the scrambled control, demonstrated that reduced ATP production in the KO was not due to secondary effects.
Reduced TRPM2 Results in Increased Mitochondrial ROS after Doxorubicin-To determine the level and origin of ROS in SH-SY5Y cells in which TRPM2 is depleted (40), mitochondrial superoxide was quantitated with MitoSOX Red at baseline and following doxorubicin using confocal microscopy. At baseline, TRPM2-depleted SH-SY5Y cells loaded with MitoSOX Red showed a small but significantly greater level of mitochondrial ROS compared with scrambled control cells (Fig. 7). In TRPM2-depleted cells treated with doxorubicin, a further significant increase in mitochondrial ROS was seen compared with scrambled cells. Cellular ROS was not quantitated in TRPM2-depleted cells because of interference with 5,6-chloromethyl-2Ј,7Ј-dichlorodihydrofluorescein diacetate (CM-H 2 DCF-DA) by GFP in the CRISPR vector.

SOD2 Restores Cell Viability and Reduces ROS Levels in TRPM2-depleted Cells in Vitro-
To evaluate the role of elevated mitochondrial ROS in modulation of cell viability, TRPM2-depleted SH-SY5Y cells and scrambled control cells were stably transfected with empty vector, SOD1, SOD2, or both. Expression of SOD1 and SOD2 was confirmed by Western blotting (Fig. 8D). Viability of untreated cells or cells treated with doxorubicin for 24 h was analyzed by XTT. Scrambled control cells expressing empty vector, SOD1, SOD2, or both had similar viability by XTT analysis at 24 h after doxorubicin (Fig. 8A). TRPM2-depleted cells transfected with empty vector had significantly lower viability compared with scrambled control cells. However, expression of SOD1, SOD2, or both significantly increased the survival of TRPM2-depleted cells after doxorubicin. In fact, the viability of TRPM2-depleted cells transfected with SOD2 was not significantly different from that of scrambled control cells at 24 h (Fig. 8A, p ϭ 0.07). Levels of mitochondrial superoxide were quantitated at baseline and following doxorubicin with MitoSOX Red fluorescence and confocal microscopy. Without doxorubicin treatment, TRPM2depleted SH-SY5Y cells transfected with empty vector showed significantly greater levels of mitochondrial ROS compared  NOVEMBER 18, 2016 • VOLUME 291 • NUMBER 47 with scrambled control cells (Fig. 8B). In contrast, TRPM2depleted cells expressing SOD1, SOD2 or both showed similar levels of ROS compared with scrambled control cells stably expressing the same SOD constructs. Cells expressing SOD2 had significantly lower levels of ROS compared with empty vector or SOD1 expressing cells. After doxorubicin, TRPM2-depleted cells transfected with empty vector had significantly greater levels of superoxide than scrambled control cells. In TRPM2-depleted cells treated with doxorubicin, the lowest levels of ROS were found in SOD2-transfected cells. TRPM2-depleted cells transfected with SOD2 had ROS levels similar to scrambled control. ROS levels in cells expressing SOD1 were slightly reduced compared with vector-transfected TRPM2-depleted cells, but results did not reach statistical significance. These results confirm that higher levels of mitochondrial ROS in TRPM2-depleted cells play a major role in reduced viability and increased sensitivity to doxorubicin. Of note, the decline in ATP production was significantly greater in TRPM2-depleted cells compared with scrambled cells (Fig. 8C). TRPM2-depleted cells transfected with SOD1 or both SOD1 and -2 showed modest increases in ATP levels compared with cells transfected with empty vector, but ATP production in all of the TRPM2-depleted cell lines remained significantly reduced compared with scrambled control cells transfected with the same constructs.  Whereas expression of SOD2 reduced mitochondrial ROS and maintained viability of TRPM2-depleted cells in culture, the defect in ATP production was not rescued by SOD2 expression.

TRPM2 Depletion Inhibits Mitochondrial Function and Survival
MitoTEMPO Reduces Mitochondrial ROS and Preserves Viability of TRPM-S Expressing Cells-To verify that the increase in ROS in TRPM2-depleted cells was from the mitochondria, a second approach to reduce TRPM2 function was utilized, inhibition with TRPM2-S. Expression of TRPM2-S was previously shown to inhibit TRPM2-L channel gating (26) and suppress cell viability, mitochondrial function, and ATP production (41). Here, superoxide levels were measured with MitoSOX Red in SH-SY5Y cells stably transfected to express vector (V), fulllength TRPM2 (L), or TRPM2-S (S) (Fig. 9A). No differences in baseline superoxide levels were observed between V, L, or S expressing cells. However, a significant increase in mitochondrial ROS was seen in TRPM2-S expressing cells following doxorubicin treatment (Fig. 9B). Cellular ROS levels were also measured with CM-H 2 DCF-DA (DCF). In cells expressing S, cellular ROS levels were slightly increased at baseline (Fig. 9D) and dramatically after doxorubicin treatment (Fig. 9E) compared with TRPM2-L (L) or empty vector (V)-transfected cells.
SH-SY5Y cells expressing L, S, or V were then pretreated with the mitochondrial superoxide scavenger MitoTEMPO, and mitochondrial and cellular ROS levels were determined after doxorubicin. Mitochondrial ROS levels quantitated with MitoSOX were reduced in all groups by MitoTEMPO, but greater in TRPM2-S expressing cells than in V or L expressing cells (Fig. 9C). Following pretreatment with MitoTEMPO, the increase in cellular ROS measured with DCF was also significantly reduced in S expressing cells compared with V or L (Fig.  9F). This supports the conclusion that the major source of the ROS increase in cells following TRPM2 inhibition was mitochondrial, rather than secondary to increased Nox activity. This is consistent with the finding that Nox2 and Nox4 expression are not increased in TRPM2-depleted cells (Fig. 4). To further examine the source of ROS, cells expressing TRPM2-S were pretreated with MitoTEMPO, the Nox inhibitor diphenyleneiodium chloride (DPI), or vehicle DMSO, and then treated with doxorubicin. MitoTEMPO significantly reduced the rise in both mitochondrial and cellular ROS levels following treatment with doxorubicin, but pretreatment with DPI reduced neither (Fig. 10, A and B). The importance of the ROS increase in reducing viability of cells in which TRPM2 is inhibited was then studied. MitoTEMPO pretreatment significantly increased the survival of TRPM2-S expressing cells treated with doxorubicin ( Fig. 10, C and D), demonstrating the important role of mitochondrial ROS in reducing cell viability following TRPM2 inhibition.
Ca 2ϩ Entry via TRPM2 Is Required for Reduction of Mitochondrial ROS Post-doxorubicin-To establish whether calcium influx through TRPM2 is involved in maintaining low levels of mitochondrial ROS, the TRPM2 pore mutant E960D was expressed in SH-SY5Y wild type parental cells and ROS measured in untreated cells and following doxorubicin exposure. No differences in mitochondrial ROS were observed in untreated E960D expressing cells compared with cells expressing the L or S isoforms (Fig. 10E). However, following doxorubicin, mitochondrial ROS levels were much greater in cells expressing TRPM2-S or E960D compared with V or L (Fig.  10F). These data support the conclusion that calcium entry through activated TRPM2 is involved in maintaining lower mitochondrial ROS after doxorubicin. This is consistent with the ability of TRPM2-L but not E960D to restore cell viability and ATP production in the KO.

SK-N-DZ and SK-N-FI Neuroblastoma Cells Also Demonstrate Decreased Cell Viability and Bioenergetics and Increased ROS Production after Doxorubicin when TRPM2 Is Inhibited-
To confirm these findings in additional neuroblastoma cell lines, SK-N-DZ and SK-N-FI cells were transiently transfected with empty vector, wild type TRPM2, or TRPM2-S. In both of these cell lines, cell viability measured by XTT assay was significantly reduced in cells expressing TRPM2-S compared with vector or TRPM2-L expressing cells, both before and after doxorubicin treatment (Fig. 11A). Whereas SK-N-DZ cells expressing TRPM2-L had greater viability at 24 and 48 h than vector-expressing cells, viability of SK-N-FI cells expressing TRPM2-L was slightly reduced at 0 and 24 h compared with vector-expressing cells, although still significantly greater than TRPM2-S-expressing cells. This may be due to the high level of TRPM2-L overexpression. Mitochondrial ROS levels and ATP levels in empty vector, TRPM2-L, or TRPM2-S expressing cells were not significantly different in untreated cells. After doxorubicin, mitochondrial ROS levels in SK-N-DZ and SK-N-FI cells expressing TRPM2-S were significantly increased com-  NOVEMBER 18, 2016 • VOLUME 291 • NUMBER 47 JOURNAL OF BIOLOGICAL CHEMISTRY 24455 pared with levels in empty vector or TRPM2-L expressing cells, which were similar to each other (Fig. 11B). After doxorubicin, ATP production was significantly reduced in TRPM2-S expressing cells compared with empty vector or TRPM2-Ltransfected cells, which were not statistically different (Fig.  11C). These experiments demonstrate that TRPM2 inhibition reduces viability, and influences mitochondrial function and ROS in multiple neuroblastoma cell lines.

Discussion
TRPM2 has an important role in cell survival following oxidative stress or ischemic injury (35)(36)(37)40). Recent studies have shown that TRPM2 is highly expressed in a number of malignancies including melanoma, lung, neuroblastoma, and breast cancer (39,41,42), suggesting that it plays a role in promoting tumor growth, and that inhibition may be a novel therapeutic approach. Here, TRPM2 was depleted in neuroblastoma cells with CRISPR/Cas9 technology as a model to explore its role in tumor growth and cell survival after doxorubicin. These studies convincingly demonstrated the significant impact of TRPM2 inhibition on reduction of mitochondrial function, impairment of cellular bioenergetics, enhancement of mitochondrial ROS production, and its potential role as a novel therapeutic approach in cancer. This is very consistent with past and pres-ent work that elucidated the importance of mitochondrial energy reprogramming in cancer growth and the recognition that drugs that target mitochondrial function may have great clinical potential as potent anti-cancer agents (56).
The first major finding of this work is that TRPM2 depletion reduces tumor growth in xenografts and significantly increases neuroblastoma sensitivity to doxorubicin. The mechanism involved a reduction in HIF-1/2␣, and proteins involved in glycolysis (ENO2), oxidative stress (FOXO3a, SOD 1/2), and mitochondrial function downstream of HIF. HIFs are frequently up-regulated in cancer, important in tumor progression, and mediate expression of a number of genes involved in many processes promoting tumor growth including oxidative stress, mitochondrial function, and ROS production (50,54,57). Our findings are consistent with the reports that HIF-1/2␣ promotes a more aggressive phenotype including disseminated disease and poor outcome in neuroblastoma patients (46,47,58). They are also consistent with a recent report that targeting the TRPM2 channel fostered cell death in T cell leukemia (59). The second important finding here is that mitochondrial function is significantly impaired by TRPM2 depletion; mitochondrial proteins including BNIP3 and COX4.1/4.2, oxygen consumption rate, and ATP levels after doxorubicin were all significantly reduced in the knock-out. A proposed scheme for the impact of TRPM2 inhibition on mitochondrial function, ROS production, and cell viability is shown in Fig. 12. The significant decrease in expression of COX4.1/4.2, involved in complex IV activity (50), may contribute significantly to the decrease in ATP production and to increased ROS due to reaction of electrons with oxygen prior to complex IV (55). Reduced BNIP3, a key regulator of mitophagy, may further impact mitochondrial function in TRPM2-depleted cells through impaired removal of dysfunctional mitochondria. When TRPM2 was  NOVEMBER 18, 2016 • VOLUME 291 • NUMBER 47 inhibited with a different approach, use of dominant-negative TRPM2-S, mitochondrial defects were also found (41). In these cells, a HIF-1␣ target gene NDUFA4L2 (NADH dehydrogenase (ubiquinone) 1␣ subcomplex, 4-like 2) was decreased. NDUFA4L2 can attenuate mitochondrial oxygen consumption through inhibition of Complex I activity, limiting ROS production under lowoxygen (50). Together, these data confirm that TRPM2 is essential for normal mitochondrial function and cellular bioenergetics, particularly under oxidant stress.

TRPM2 Depletion Inhibits Mitochondrial Function and Survival
The third key finding of this report is that TRPM2 inhibition with two fundamentally different approaches, depletion or inhibition with a dominant-negative, both resulted in significant increases in mitochondrial ROS. MitoSOX measures mitochondrial superoxide anion, which is dismutated in the mitochondria to hydrogen peroxide by SOD2. SOD1 catalyzes superoxide anion conversion to hydrogen peroxide primarily in the cytoplasm. Increased expression of SOD2 had a significant impact on reducing superoxide production and increasing cell viability in TRPM2-depleted cells, confirming the importance of mitochondria as the source of ROS. However, expression of SOD2 affected ATP production minimally. A possible explanation is that the defect in mitochondrial function (the electron transport chain) resulting in increased superoxide production, and reduced ATP production was not repaired by increased SOD2 (Fig. 12). The mitochondrial antioxidant MitoTEMPO also inhibited the increase in mitochondrial ROS and preserved cell viability when TRPM2-inhibited cells were treated with doxorubicin, further confirming the important contribution of mitochondrial ROS in reduced cell survival. In contrast, pretreatment with DPI, an inhibitor of NADPH oxidase, did not significantly inhibit the cellular rise in ROS after doxorubicin. These data, together with the observation that Nox2 and Nox4 levels were not altered in TRPM2-inhibited cells, strongly support the conclusion that the increase in cellular ROS predominantly originates in mitochondria and is associated with significantly reduced mitochondrial function.
Expression of TRPM2-L restored cell viability and cellular bioenergetics in the TRPM2 KO, but E960D did not, demonstrating the role of TRPM2-mediated calcium entry in maintenance of cell viability and ATP production. Also, experiments with the TRPM2 pore mutant E960D and TRPM2-S showed that calcium influx through TRPM2 is involved in TRPM2mediated suppression of mitochondrial ROS. One potential pathway through which TRPM2 may influence cell viability and energetics is calcium-mediated regulation of HIF-1/2␣ expression (60 -63). Stabilization of HIF-1/2␣ expression by calcium activation of calcineurin (60) is a potential mechanism responsible for the greater expression of HIF-1/2 and the mitochondrial genes it regulates including BNIP3 and NDUFA4L2 (53,64) in TRPM2 full-length expressing cells. This would help sustain mitochondrial function and reduce ROS in the presence of oxidative stress. Alternatively, low level mitochondrial Ca 2ϩ uptake has been shown to have a role in regulation of cell bioenergetics (65)(66)(67). Mitochondrial function may be impaired by the very low calcium uptake in TRPM2-depleted cells. Although evidence suggests that TRPM2 is not present on mitochondrial membranes, TRPM2 could influence mitochondrial calcium by its localization in the ER (41), for example,  Scheme of the influence of TRPM2 inhibition on mitochondrial function, ROS production, and cell survival. Inhibition of TRPM2 results in reduced intracellular calcium entry and mitochondrial calcium uptake in response to TRPM2 activation. Decreased intracellular calcium results in reduced HIF-1/2␣ expression as well as genes transcriptionally regulated by HIF-1/2␣ including FOXO3a, which modulates SOD1/2 expression, and mitochondrial genes including BNIP3 and COX 4.1/4.2 (Complex IV). Decreased mitochondrial calcium uptake together with decreased expression of mitochondrial genes impairs mitochondrial function including the electron transport chain. This contributes to reduced ATP production and increased mitochondrial superoxide production, particularly with reduced SOD1/2, and subsequently results in a rise in cellular ROS. Decrease in BNIP3 contributes to impaired mitophagy and accumulation of dysfunctional/damaged mitochondria, which further results in increased ROS and susceptibility to chemotherapeutic agents, and reduced cell survival and tumor growth. NOVEMBER 18, 2016 • VOLUME 291 • NUMBER 47 through the close spatial relationship between ER and mitochondria (ER-mitochondrial tethering) (68,69). TRPM2mediated Ca 2ϩ influx could also have a role in regulation of the mitochondrial calcium uniporter (70). The mechanisms through which calcium influx through TRPM2 regulates mitochondrial ROS are currently under investigation. The decreased proliferation of untreated SK-N-DZ and SK-N-FI cells expressing TRPM2-S (Fig. 11A) may be due to reduced calcium influx leading to lower levels of mitochondrial proteins including HIF-1/2␣ and mitochondrial function, as shown for SH-SY5Y cells expressing TRPM2-S (40,41).

TRPM2 Depletion Inhibits Mitochondrial Function and Survival
Previous work has shown that ROS are often increased in malignant cells, possibly as a result of the increased metabolism required to support accelerated cell growth (71), and play a key role in maintaining signaling pathways that promote proliferation and survival. However, when ROS in cancer cells increase above a certain threshold, cells become more susceptible to oxidative stress-induced cell death (72,73). Recent evidence suggests that antioxidant treatment may be harmful and actually increase tumor burden (74 -77), whereas raising ROS levels may induce cancer cell death (78 -80). Our experiments here with inhibition of TRPM2, which is highly expressed in a number of cancers, support this conclusion. After exposure to doxorubicin, which increases oxidative stress, depletion or inhibition of TRPM2 resulted in a large increase in ROS levels, enhancing cell death significantly. In neuroblastoma xenograft experiments shown here and in ones utilizing the breast cancer cell line MCF-7, 4 we observed that tumor growth was significantly reduced in cells in which TRPM2 was inhibited. A theoretical concern with this mechanistic approach is that if ROS levels are not sufficiently raised, tumorigenesis could be promoted (81); this was not observed in our experiments with TRPM2 inhibition thus far. In all of our experiments with neuroblastoma cell lines, treatment of TRPM2-depleted or inhibited cells with doxorubicin resulted in reduced cancer cell survival compared with TRPM2 expressing cells.
In summary, we demonstrate a dramatic effect on in vitro survival and in vivo xenograft growth of TRPM2 depletion, resulting from reduced calcium entry, mitochondrial dysfunction, compromised cellular bioenergetics, and increased ROS. Future research efforts will focus on understanding the pathways through which reduced calcium entry decreases HIF-1/2␣ expression, impairs mitochondrial function, and enhances mitochondrial and cellular ROS levels. Because TRPM2 is up-regulated in many types of cancer (42), exploration of the functional role of TRPM2 in cell survival and development of an effective inhibitor will have broad applicability. TRPM2 inhibition has great promise as a novel therapeutic approach, in particular for designs that aim to kill cancer cells through increasing ROS levels or impairment of mitochondrial function.

Experimental Procedures
Depletion of TRPM2 with CRISPR and Generation of Stably Transfected Neuroblastoma Cell Lines-The neuroblastoma cell line SH-SY5Y was purchased from the American Type Culture Collection (ATCC, Manassas, VA). SH-SY5Y cells were cultured in 50% DMEM and 50% Ham's F-12 (DMEM/F-12 50/50) medium supplemented with 10% heat-inactivated FBS. To deplete cells of endogenous TRPM2 with CRISPR (Clustered regularly interspaced short palindromic repeats) technology, TRPM2 gRNA in pCas9-guide vector (target sequence GAGGAAAGCTGGCTCGGAGC), negative scrambled control vector, and donor vector containing 600-bp left and right homologous arms to the TRPM2 gene with GFP with puromycin markers, were purchased from OriGene Technologies Inc. (Rockville, MD). Either TRPM2 gRNA vector or scrambled vector was cotransfected with donor vector into SH-SY5Y (wild type) neuroblastoma cells using the Neon Transfection System (Invitrogen), following the manufacturer's instructions. Cells were passed up to 8 passages with 0.5 g/ml of puromycin selection. TRPM2 genomic DNA was verified by RT-PCR at passages 2 and 4. Depletion of TRPM2 protein expression was confirmed by Western blot analysis at passages 6, 7, and 8. Cells at passage 8 were then subjected to single clone selection in the presence of 0.5 g/ml of puromycin. All clones were subsequently screened by real-time PCR and Western blot analysis, and three TRPM2 knock-out clones and three scrambled clones were selected for further analysis. In general, puromycin was removed from cultures 7-14 days before experiments were performed.
To modulate TRPM2 isoform expression, SH-SY5Y cells at 90% confluence were transfected with pcDNA3.1/V5-His TOPO expressing L, S, or empty vector for 48 h using Lipofectamine 2000 (Invitrogen) (26). Stably transfected cell lines were selected using 600 g/ml of G418 (Geneticin, an analogue of neomycin) (Gemini Bio-Products, West Sacramento, CA) and cell cultures were maintained in the presence of 250 g/ml of G418 for a minimum of two months before cells were used for experiments.
In some experiments, SH-SY5Y cells in which TRPM2 was depleted with CRISPR or scrambled control, cells were transfected with wild type TRPM2 or the E960D pore mutant constructs subcloned into pcDNA3.1/V5-His TOPO using the Neon Transfection System (Invitrogen) following the manufacturer's instructions. SOD1 (Addgene, Cambridge, MA) and SOD2 (gift of Dr. Paul M. Hwang, NHLBI, NIH) constructs subcloned into pcDNA3.1/V5-His TOPO were also transfected into SH-SY5Y cells using the Neon Transfection System. Stably transfected cells were generated and maintained by culture in 0.5 g/ml of puromycin and/or 250 g/ml of G418 as appropriate. For doxorubicin studies, plates were treated when cells were 70 -80% confluent and harvested at the time points indicated.
In other experiments, SK-N-DZ and SK-N-FI cells were transiently transfected with empty vector, TRPM2-L, or TRPM2-S subcloned in pcDNA3.1/V5-His TOPO and transfected with the Neon Transfection System. Cells were studied at 48 -96 h after transfection and expression was confirmed with Western blotting.
RT-PCR of TRPM2 in Neuroblastoma Cell Lines to Confirm TRPM2 Depletion-RNA was prepared from neuroblastoma cell lines to confirm TRPM2 depletion using RNeasy Minikit (Qiagen, Valencia, CA). First-strand cDNA synthesis was performed from 500 to 2000 ng of RNA using SuperScript First Strand kit (Invitrogen by Life Technologies). The cDNA was then subjected to the quantitative real-time PCR using 4.5 ml of the first-strand cDNA reaction (ϫ10 diluted) and TaqMan gene expression assays (Applied Biosystems, Foster City, CA). TRPM2 TaqMan assay was custom designed to detect part of Exon1, which was targeted by CRISPR/cas9 technology: forward primer, 5Ј-CTCAGCCCTGAGGAAAGC; reverse primer, 5Ј-CGGAGATTGGAGACCATCC, reporter, 5Ј-TGC-CCAGAAGGGTCACTGACCT. Predesigned TaqMan assays were used for reference genes: GAPDH (Applied Biosystems catalog number 4333764) and TBP (Applied Biosystems catalog number 4333769). Real-time PCR was performed using Quant-Studio 12KFlex with 384-well real-time PCR system (Thermo-Fisher, Waltham, MA) or StepOne plus with 96-well real-time PCR system (Applied Biosystems). Reactions were run in three or four replicates. The PCR results were analyzed using Expres-sionSuite software ( Life Technologies) as relative mRNA levels of cycle threshold (C T ) value using wild type of neuroblastoma cell line as a calibrator.
Generation of E960D TRPM2 Mutant Expressing Cells with Retroviral Gene Transfer and Cell Sorting-Wild-type human TRPM2-L, S, and E960D mutant (HA-tagged) cDNA were cloned using NotI and EcoRI sites into the pMSCV bicistronic retroviral vector (MIG vector). This vector contains internal ribosome entry site, and enhanced green fluorescent protein (EGFP) (82). Retroviruses were produced by transient transfection in amphotropic packaging 293 cell lines. SH-SY5Y cells were cultured in retroviral supernatant with 10 g/ml of Polybrene and centrifuged at 1,400 ϫ g at 32°C for 2 h. Cells were re-suspended in fresh DMEM/F-12 medium with 10% FBS, at 37°C, 5% CO 2 for 3 days. The GFP positive cells were sorted using a BD Biosciences FACSAria SORP.
Cell Proliferation Assay-Cells from stably transfected cell lines or CRISPR depletion were seeded on 96-well plates and cultured in media with 250 g/ml of G418 or 0.5 g of puromycin, respectively, for 96 h. Cell proliferation was assessed by measurement at A 490 nm/690 nm using XTT cell proliferation assay (Trevigen Inc., Gaithersburg, MD) following the manufacturer's instructions (84). In some experiments, cells were treated with doxorubicin (0.3 M) for specified durations during cell culture. Viability after doxorubicin was quantitated with trypan blue exclusion and/or XTT assay.
TRPM2-depleted Xenograft Tumors-To determine the role of TRPM2 in tumor growth, athymic Nude-FOXn1 nu female mice (Harlan Laboratories, Inc., Indianapolis, IN) were injected in one flank with 1.5 ϫ 10 7 SH-SY5Y cells depleted of TRPM2 with CRISPR, or scrambled CRISPR or wild type parental control cells. Approximately 10 -15 mice per group were used in each of two experiments. Tumor length and width were measured twice weekly with a caliper to determine tumor growth over 6 -7 weeks, and the tumor area was calculated. At the completion of each experiment, tumors were harvested, weighed, and frozen for analysis. All protocols and procedures applied to the mice in this study were approved by the Institutional Animal Care and Use Committee of the Pennsylvania State University College of Medicine.
Measurement of Mitochondrial O 2 Consumption and ATP Levels-Oxygen consumption rate was measured in neuroblastoma SH-SY5Y cells at 37°C in an XF96 extracellular flux analyzer (Seahorse Bioscience). Neuroblastoma cells were sequentially exposed to oligomycin, FCCP, and rotenone, using the XF Cell Mito Stress Kit (Seahorse Bioscience) according to the manufacturer's instructions. Preliminary experiments were performed to select optimal seeding density (10 5 cells/well) and compound concentrations, according to the manufacturer's instructions. To measure ATP levels, neuroblastoma SH-SY5Y cells or cells isolated from neuroblastoma xenografts were lysed and ATP (luminescence) levels were measured using a Cell Titer-Glo Luminescent Cell Viability Assay kit (Promega, Madison, WI) according to the manufacturer's instructions.

Measurement of Intracellular and Mitochondrial ROS Using
Confocal Microscopy-SH-SY5Y cells were cultured in glass bottom dishes (MatTek, Ashland, MA) with DMEM/F-12 (Corning, Manassas, VA) containing 10% fetal bovine serum (Gibco by Life Technologies, Eugene, OR). MitoSOX (Molecular Probes by Life Technologies) was used as a marker for mitochondrial superoxide measurement. Cells were loaded with MitoSOX at a final concentration of 5 M for 20 min, at 37°C, according to the manufacturer's recommendations. Intracellular ROS was detected by staining with the fluorescent probe CM-H 2 DCF-DA (Molecular Probes by Life Technologies), which although often used to measure cytosolic ROS is less specific for cellular compartment or type of ROS. CM-H 2 DCF-DA was added to the cells in a final concentration of 1 M for 20 min at 37°C. Cells in glass bottom dishes were washed 3 times after 20 min CM-H 2 DCF-DA or MitoSOX loading and used for confocal microscopy imaging immediately. Cells were untreated or treated with 0.3 M doxorubicin for 24 h before loading and microscopy. In some experiments, cells were also pretreated with 25 M MitoTEMPO (Enzo Life Sciences, Farmingdale, NY), 0.5-5 M DPI (Sigma) or DMSO for 60 min before doxorubicin and during the experiment. For live cell ROS measurements, confocal microscopy (Leica AOBS SP8 laser scanning confocal microscope; Leica Microsystems, Heidelberg, Germany) was used along with a heated live cell stage equipped with a humidified 5% CO 2 perfusion system (TOKAI HIT). Images were captured at 1024 ϫ 1024 pixels using the ϫ63/1.2 high numerical aperture apochromatic water-immersion objective. For three-dimensional image data set acquisition, the excitation beam was first focused at the maximum signal intensity focal position within the sample and the photon counting mode available within the hybrid detector (HyD) was used to register the fluorescence signals or emitted photons/pixels. Later on, the beginning and end of the three-dimensional stack (i.e. the top and the bottom optical sections) were set based on the signal level degradation. A series of two-dimensional images for a selected three-dimensional stack volume were then acquired with a ϳ0.3 m section thickness (z axis). For each cell volume reported here, z-section images were compiled and finally the three-dimensional image restoration was performed using Imaris software (Bitplane, Switzerland). The Surface Module available within the Imaris was used to compute fluorescence intensities from three-dimensional cell volumes. A Gaussian noise removal filter was used. Sum of all the fluorescence intensities above the lower threshold level was then determined and finally the fluorescence intensity per cell was obtained by dividing the sum of all fluorescence intensities by the number of cells. Similar quantitation protocol was applied to all three-dimensional image volume datasets generated using similar setting parameters.
Statistics-All results are expressed as mean Ϯ S.E. For analysis of TRPM2 current as a function of group and voltage, twoway analysis of variance (ANOVA) was used. For analysis of protein expression levels, O 2 consumption, XTT, trypan blue exclusion, ATP levels, and ROS levels, one-way ANOVA or Student's t test was used. In all analyses, p Յ 0.05 was taken to be statistically significant.
Author Contributions-L. B. performed and designed experiments involving ROS and bioenergetics measurements, generation of E960D pore mutant, and those involving SOD1/2 shown in Figs. 5-11; S. J. C. designed and established the TRPM2 CRISPR knockout and scrambled SH-SY5Y cell lines, assisted with in vitro and xenograft experiments, and worked with L. B. in preparation of figures; K. C. provided technical assistance and performed in vitro cell viability experiments (Fig. 10), assisted with xenografts, and Western blotting analysis shown in Fig. 4; K. K. performed xenograft experiments in Fig. 3