MICU1 drives glycolysis and chemoresistance in ovarian cancer

Cancer cells actively promote aerobic glycolysis to sustain their metabolic requirements through mechanisms not always clear. Here, we demonstrate that the gatekeeper of mitochondrial Ca2+ uptake, Mitochondrial Calcium Uptake 1 (MICU1/CBARA1) drives aerobic glycolysis in ovarian cancer. We show that MICU1 is overexpressed in a panel of ovarian cancer cell lines and that MICU1 overexpression correlates with poor overall survival (OS). Silencing MICU1 in vitro increases oxygen consumption, decreases lactate production, inhibits clonal growth, migration and invasion of ovarian cancer cells, whereas silencing in vivo inhibits tumour growth, increases cisplatin efficacy and OS. Mechanistically, silencing MICU1 activates pyruvate dehydrogenase (PDH) by stimulating the PDPhosphatase-phosphoPDH-PDH axis. Forced-expression of MICU1 in normal cells phenocopies the metabolic aberrations of malignant cells. Consistent with the in vitro and in vivo findings we observe a significant correlation between MICU1 and pPDH (inactive form of PDH) expression with poor prognosis. Thus, MICU1 could serve as an important therapeutic target to normalize metabolic aberrations responsible for poor prognosis in ovarian cancer.

The contribution by Chakraborty and colleagues is an original and interesting piece of work suggesting that dysregulation of mitochondrial Ca2+ homeostasis by the overexpression of the negative regulator of the mitochondrial Ca2+ uniporter MICU1 in ovarian carcinomas is responsible for diverting pyruvate from the action of pyruvate dehydrogenase (PDH) triggering in this way the upregulation of glycolysis and cellular chemoresistence to cisplatin. In other words, emphasizes another mechanism by which pyruvate dehydrogenase kinase (PDK) could be playing a role in metabolic rewiring in cancer cells (see also McFate et al., JBC 283, 22700 (2008), Pate et al., EMBO J. 33, 1454EMBO J. 33, (2014). Cell-death resistance to cisplatin-induced Ca2+ release from intracellular stores is ascribed to the protection exerted by MICU1 for mitochondrial Ca2+ entry and the subsequent unlashing of the apoptotic response. In addition, and consistent with the work presented, they provide relevant information regarding the clinical outcome for ovarian cancer patients with overexpression of MICU1 mRNA and protein and the expression of the inactive (phosphorylated) PDH. Moreover, they show that tumor xenographs in nude mice bearing silenced MICU1 are much less aggressive than tumors derived from the parental cells. The work, for the most part of it, is clearly presented and documented and accompanied by appropriate controls. However, the paper fails to demonstrate the role of MICU1 expression level in controlling intramitochondrial Ca2+ concentrations.
MAJOR POINTS: 1.-Intramitochondrial Ca2+ levels need to be documented in response to MICU1 regulated expression with appropriate Ca2+ sensitive proteins targeted to mitochondria because a major claim of the paper is based on the role played by intramitochondrial Ca2+ in metabolic dysregulation. Nowadays, Rhod-2 fluorescence is not reliable enough as a mitochondrial Ca2+ sensitive probe.
2.-The observed changes in maximum respiratory rate (FCCP stimulated) in response to the regulated expression of MICU1 need some explanation and additional investigation. Is it possible that MICU1 could be primarily affecting the expression/assembly of respiratory complexes (RC)? Analysis of RC in BN-gels could help in this regard. Minor points: 1.-On page 6, it is mentioned that mitochondrial copy number remains unaltered. The authors have determined the expression of a mitochondrial gene and assumed that the copy number of mtDNA is not affected. This should be clarified in the text or alternatively the abundance of a mitochondrial gene relative to the abundance of a nuclear one should be determined.
2.-On page 8, the Ferrick reference has not been converted. The two c's of Cytochrome C should be in lower case.
3.-On page 10, the abbreviation Pyr6 should be spelled out.
4.-The effect of FCCP on the TMRM Fluorescence Intensity should be incorporated in Figs. 3A,B. In addition, since in the Y-axis it is being represented the Relative Fluorescence Intensity the scale can be simplified representing data in the 0 to 20 range.
5.-The claimed association/dissociation of pyruvate dehydrogenase phosphatase (PDP) with PDH in response to the silencing/overexpression of MICU1 (Fig. 6) needs quantification and statistical evaluation.
Reviewer #2 (Calcium metabolism in Cancer) (Remarks to the Author): The paper of Chakraborty et al. describes new mechanism of chemoresistance in ovarian cancer cells, it demonstrates that MICU1 drives aerobic glycolysis in ovarian cancer by modulating the pyruvate dehydrogenase (PDH). The authors suggest that MICU1 could serve as a novel therapeutic target to normalize metabolic aberrations. Although the results are interesting and the main message really important, I have a number of major points to be clarified before the paper becomes acceptable for Nature Communications.
Major points -Most of the experiments described in this manuscript were performed in CP20 and OV90 cells. It is however unclear as to why the authors chose these cell lines among all those they tested. Their choice should be explained.
-Throughout the manuscript, the authors have used only one test for proliferation/migration/invasion. Confirmation of these preliminary results with different assays would go a long way to strengthen the authors' conclusions.
-The authors chose cisplatin as their reference chemotherapeutic drug. Are similar results obtained with a different drug, or are they exclusive to this compound? -In Fig. 3F, the authors use mitoTEMPO to scavenge ROS. A control experiment showing that this is actually the case in their models is missing. They could use MitoSOX dye, as shown in Fig. 3E.
- Fig. 4A: while the spike in Ca2+m induced by cisplatin seems indeed different between siCTL and siMICU1-treated cell, this difference disappears within 400s. Are the authors suggesting that cell fate will be drastically altered by these initial 400s? They should carefully study the kinetics of their responses, and show results proving long-lasting differences between their conditions. - Fig. 4E: "...indicating absence of Ca2+ release from ER without cisplatin treatment." This point is very important, and need to be clearly demonstrated. Are ER Ca2+ stores really identical in shRNA-treated cells? The authors should measure them in their models, using ionomycin in Fura2loaded cells, or any other method available to them.
-page 12: they authors hypothesize that MICU1 will lead to "Ca2+ overload in the mitochondria". They should show evidence of this: dyes are available for that kind of measurement, and they need to substantiate the fact that calcium homeostasis is indeed modified by MICU1 silencing.
-page 13: "a direct consequence of MICU1 driven lowering of Ca2+m". Could the authors indicate where they have shown that MICU1 overexpression decreases Ca2+m? Does not this statement contradict results of Fig. 4D? -The in-vivo results are quite puzzling: first, Fig. 7B seems to indicate that all 10 animals injected with shCTL cells died on the same day. Is it really that synchronized? Second, the conclusion that cells without MICU1 are more sensitive to cisplatin is invalidated by these experiments. Indeed, on both populations (shCTL and shMICU1 injected mice), cisplatin treatment gave the same extra 7 days of survival. The conclusion from this experiment should rather be that MICU1 and cisplatin act through distinct pathways, as also illustrated in Fig. 7A where tumor growth inhibition by cisplatin seems to be proportionally identical between shCTL and shMICU1.
Reviewer #3 (mechanisms of chemotherapy resistance) (Remarks to the Author): The MS by Chakraborty and colleagues is a follow up of the authors' early observation that small interfering RNA-mediated silencing of the mitochondrial uniporter MICU1 is able to sensitize ovarian cancer cells to positively charged gold nanoparticles. This observation suggested a potential role of MICU1 in drug resistance.
It is known that MICU1 resides in the inner mitochondrial membrane (IMM) and required for Ca2+ uptake through mitochondrial Ca2+ 'uniporter' (MCU), a calcium-selective ion channel responsible for low-affinity calcium uptake into the mitochondrial matrix. It has been shown that MCU expression is correlated with cancer progression in sporadic studies (prostate, colon and breast). However, the functional role for MICU1 in cancer is still unknown. In the present MS the authors suggesting new function of MICU1 that regulates metabolic fate and confers chemoresistance in ovarian cancer. Moreover, they tried to make a link between obtained observation and cancer prognosis. Although the message of this MS is interesting and the quality of submitted MS is high, presented data raise several questions that the authors have to address.
1. Work from Rizzuto's lab revealed that MICU1 and MICU2 finely tune the mitochondrial Ca2+ uniporter by exerting opposite effects on MCU activity. It is also known that MICU1 and MICU2, which, both in purified lipid bilayers and in intact cells, stimulate and inhibit MCU activity, respectively. Keeping that in mind inhibition of MICU1 should even stronger influence MCU activity. Downregulation of MICU1 does not influence expression level of MCU and EMRE, but MICU2 showed modest decrement. How MICU2 should act in these conditions? What is the level of expression of MICU2 in ovarian cancers?
2. The authors suggested that removal of MICU1 leads to mitochondrial calcium accumulation. If MICU2 is acting properly, then the balance in calcium transport that regulated by MICU1 and MICU2 is disturbing. Data presented in Fig. 4a should be explained, keeping in mind even small changes in expression of MICU2. Moreover, removal of MICU1 on OV90 cells did not changed significantly level of calcium. It is important to note that cisplatin causes similar release of [Ca2+]ER in presence or absence of MICU1 in OV90 cells ( fig. 4), but appearance of calcium in cytosol is dangerous. This should be explained and stronger evidence that MICU1 restricts chemotherapy-stimulated entry of Ca2+ into the mitochondria, even when large concentration of Ca2+ is released from ER, thus conferring chemoresistance of ovarian cells should be presented.
3. Indeed, the silencing of MICU1 influenced anchorage-independent clonal growth. However, in two out of three significant decrease (up to 50%) in colony formation was observed even after transfection of control siRNAs (see Fig. 2 a-c), which questioning the response of cells to more specific silencing of MICU1. Moreover, as the authors stated the role of calcium on clonal growth and differentiation has been well established in human bronchial, leukemic and epidermal cells. Several calcium channels are involved in this process. How elimination of one protein that involves in the regulation of function of just one channel can so drastically influence the process of clonal growth? This should be explained and discussed more carefully.
4. The authors suggested that experiments with short incubation with cisplatin mimic clinical cisplatin therapy schedule, which is not correct. To draw this conclusion the authors should show pumping out of cisplatin from the cells, the effect that can simply be observed in clinical settings. 5. Cisplatin-induced apoptosis is not always perpetuates with drop of MMP. Release of cytochrome c and other intermembrane-located proteins might occur via formation of specific Bax/Bak-induced pores that lead to activation of caspase-9, etc, the process independent on Ca2+. Indeed, upon treatment with cisplatin caspase-9 is activating in cells expressing MICU1 (Fig. 3c) and removal of this protein is just accelerate this process. It is unclear why in cells with active caspase-9 PARP is not cleaved? It also raises the question concerning the role of calcium in this process. If removal of MICU1 just accelerates cisplatin-induced death, then the role of MICU1 in regulation of sensitivity to treatment is not the primary.
6. The authors concluded that expression level of PDK and pPDH in ovarian cancers is higher than in normal cells. However, data from Fig 6a contradict it. Level of PDK in OSE cells is similar to OVCAR2 and OV90 cells. On the other hand, level of pPDH is not correlated with level of PDK. This discrepancy should be explained. 7. Unfortunately, the total level of PDH in cells with siMICU1 is increasing (see Fig. 6C). In this case decreasing of the level of pPDH should be explained.
8. Why shCTL-implantation group formed statistically significant larger tumors? 9. Data in Fig. 7a showed that treatment with cisplatin is equally efficient as removal of MICU1, which again raises the questions whether this removal sensitizes cells to treatment?
11. Discussion is too long and unclear.

Response to the reviewers' comments
Response: At the outset we would like to mention that we appreciate the editor for giving us the opportunity to revise the paper. We took the reviewers' recommendations to heart and worked diligently to address all of the concerns raised by performing additional experiments, providing new data and making appropriate changes in the manuscript.
Below is our point-by-point response to the reviewers' concerns:

Reviewer#1:
The contribution by Chakraborty and colleagues is an original and interesting piece of work suggesting that dysregulation of mitochondrial Ca2+ homeostasis by the overexpression of the negative regulator of the mitochondrial Ca2+ uniporter MICU1 in ovarian carcinomas is responsible for diverting pyruvate from the action of pyruvate dehydrogenase (PDH) triggering in this way the upregulation of glycolysis and cellular chemoresistance to cisplatin. In other words, emphasizes another mechanism by which pyruvate dehydrogenase kinase (PDK) could be playing a role in metabolic rewiring in cancer cells (see also McFate et al., JBC 283, 22700 (2008), Pate et al., EMBO J. 33, 1454EMBO J. 33, (2014). Cell-death resistance to cisplatin-induced Ca2+ release from intracellular stores is ascribed to the protection exerted by MICU1 for mitochondrial Ca2+ entry and the subsequent unlashing of the apoptotic response. In addition, and consistent with the work presented, they provide relevant information regarding the clinical outcome for ovarian cancer patients with overexpression of MICU1 mRNA and protein and the expression of the inactive (phosphorylated) PDH. Moreover, they show that tumor xerographs in nude mice bearing silenced MICU1 are much less aggressive than tumors derived from the parental cells. The work, for the most part of it, is clearly presented and documented and accompanied by appropriate controls. However, the paper fails to demonstrate the role of MICU1 expression level in controlling intramitochondrial Ca 2+ concentrations.

Response:
We appreciate the reviewer's encouraging comments and thoughtful suggestions. In agreement with the reviewer, we performed additional experiments to demonstrate modulation of intramitochondrial Ca +2 by MICU1 and its role in glycolysis and chemoresistance. Briefly, in addition to chemical probes, as suggested by the reviewers, we have now used protein based mitochondrial Ca 2+ indicator GCaMP2-mt to determine the regulation of mitochondrial Ca +2 ([Ca 2+ ] m ) by MICU1 and its role in drug sensitivity in ovarian cancer models (Figures 4D, F, Supporting Figures S10, S11). Further to investigate a cause for increased oxygen consumption rate (OCR) in MICU1 silenced cells, we determined the protein expression levels of the key ETC components. . However, we did not observe any appreciable changes in the expression levels of the ETC components (Supporting Figure  S15). We next investigated individual mitochondrial complex activity from MICU1 silenced cells and observed a significant increase in Complex III activity that is in agreement with the observed increase in mitochondrial ROS (mtROS) (Supporting Figure S16). Future investigations will be focused on how MICU1 regulates Complex III activity. Furthermore, as suggested by the reviewer, we have now quantified the immunoblots demonstrating PDH-PDP interactions using NIH Image J and presented as a bar graph (Figures 6 C, E, H, K). We have also corrected all the minor issues as suggested by the reviewer. Please find our point-by-point response to the reviewer's comments in the following section.

Reviewer 1: MAJOR POINTS:
Reviewer 1: 1. Intramitochondrial Ca2+ levels need to be documented in response to MICU1 regulated expression with appropriate Ca2+ sensitive proteins targeted to mitochondria because a major claim of the paper is based on the role played by intramitochondrial Ca2+ in metabolic dysregulation. Nowadays, Rhod-2 fluorescence is not reliable enough as a mitochondrial Ca2+ sensitive probe.

Response:
We appreciate the reviewer's thoughtful suggestions. As suggested we performed our [Ca 2+ ] m measurements using the highly sensitive and selective protein-based mitochondrial Ca 2+ probe GCaMP2-mt [Kirichok et al., Nature 427, 360-364 (2004) and Doonan et al., The FASEB Journal vol. 28 no. 11, 4936-4949 (2014)]. We have repeated our experiments in different cell lines expressing GCaMP2-mt to further confirm regulation of [Ca 2+ ] m by MICU1 and its effect on cisplatin treatment (Figures 4D, F, Supporting Figures S10, S11). In brief, stable cell lines shCTL-OV90 and shMICU1-OV90, were transfected with GCaMP2-mt and mitochondrial Ca +2 response monitored after cisplatin exposure ( Figure 4D). Compared to scrambled control, MICU1 knockdown increased [Ca 2+ ] m by ~5 fold upon cisplatin treatment. Similar effects were observed in siMICU1-CP20 cells compared to siCTL-CP20 cells (Supporting Figure S10). Ectopic expression of MICU1 in nonmalignant FTE-188 cells having low endogenous MICU1 levels, inhibited cisplatin induced increase in [Ca 2+ ] m ( Figure 4F). Taken together, these additional experiments further confirm our hypothesis that MICU1 functions as a gatekeeper preventing [Ca 2+ ] m overload in response to cisplatin thereby conferring resistance to cisplatin. In accordance, two recent reports demonstrated that MICU1 serves as a molecular gate keeper preventing [Ca 2+ ] m overload, increasing survival in post-natal life [Antony et al., Nature Communications 7:10955 (2016) and Liu et al., Cell Reports 16(6), 1561-73 (2016 Reviewer 1: 2.The observed changes in maximum respiratory rate (FCCP stimulated) in response to the regulated expression of MICU1 need some explanation and additional investigation. Is it possible that MICU1 could be primarily affecting the expression/assembly of respiratory complexes (RC)? Analysis of RC in BN-gels could help in this regard.

Response:
We appreciate the reviewer's thoughtful suggestions. As suggested we have performed immunoblotting for the components of the mitochondrial complexes. Using commercially available kits, we performed these experiments from isolated mitochondria of both OV90 and CP20 cells with or without MICU1 silencing. We did not observe any significant differences in the protein expression of the components of the respiratory complexes (Supporting Figure S15). Next, measuring individual mitochondrial Complex activity, we observed that silencing MICU1 in CP20 or OV90 cells significantly increased Complex III activity (Supporting Figure S16), while those for Complex I, II and IV remained unchanged (Supporting Figure S16) 1827, Issues 11-12 (2013] and corroborates our results demonstrating higher mtROS in MICU1 silenced cells. Since higher ROS levels render cells vulnerable to cytotoxic stress [Fulda et al., International Journal of Cell Biology Volume 2010, 214074 (2010], a similar mechanism could be envisioned for MICU1 silenced cells being sensitized to cisplatin. Future investigations will be focused on how MICU1 regulates Complex III activity. We have incorporated these findings in the results and discussion section of the revised manuscript.

Reviewer 1: Minor points:
Reviewer 1:1. On page 6, it is mentioned that mitochondrial copy number remains unaltered. The authors have determined the expression of a mitochondrial gene and assumed that the copy number of mtDNA is not affected. This should be clarified in the text or alternatively the abundance of a mitochondrial gene relative to the abundance of a nuclear one should be determined.

Response:
We appreciate the reviewer's suggestions. We have now repeated this experiment using real-time PCR and normalized the abundance of mitochondrial gene ND1 relative to the abundance of the nuclear gene version. Our conclusions remain unchanged in that there were no significant differences in relative mRNA expression between siMICU1 or shMICU1 cells compared to their respective controls. We have revised the results and methods section with this new information.
Reviewer 1:2. On page 8, the Ferrick reference has not been converted. The two c's of Cytochrome C should be in lower case.

Response:
We regret these mistakes. We have now corrected these errors in the revised manuscript.

Reviewer 1: 4. The effect of FCCP on the TMRM Fluorescence Intensity should be incorporated in Figs. 3A,B. In addition, since in the Y-axis it is being represented the Relative Fluorescence Intensity the scale can be simplified representing data in the 0 to 20 range.
Response: We appreciate the reviewer's suggestions. As suggested we modified figures 3A and 3B and included FCCP control by repeating the experiments. Furthermore, as suggested, we have simplified the Relative Fluorescence Intensity scale representing data within the 0 to 20 range.
Reviewer1: 5. The claimed association/dissociation of pyruvate dehydrogenase phosphatase (PDP) with PDH in response to the silencing/overexpression of MICU1 (Fig. 6) needs quantification and statistical evaluation.

Response:
We appreciate the reviewer's suggestions. As suggested we have quantified the immunoblots (from three independent experiments) in Figure 6 pertaining to the association/dissociation of pyruvate dehydrogenase phosphatase (PDP) with PDH in response to the silencing/overexpression of MICU1, using NIH-Image J and graphically represented as mean + SEM. The revised and re-interpreted results based on the quantitative analysis are now incorporated in Figures 6 C, E, H, K.

Reviewer#2:
The paper of Chakraborty et al. describes new mechanism of chemoresistance in ovarian cancer cells, it demonstrates that MICU1 drives aerobic glycolysis in ovarian cancer by modulating the pyruvate dehydrogenase (PDH). The authors suggest that MICU1 could serve as a novel therapeutic target to normalize metabolic aberrations. Although the results are interesting and the main message really important, I have a number of major points to be clarified before the paper becomes acceptable for Nature Communications.

Response:
We appreciate the reviewer's encouraging remarks and thoughtful suggestions. We have revised the manuscript as suggested by the reviewer. Please find our point-by-point response to the reviewer's comments below; Reviewer 2: 1. Most of the experiments described in this manuscript were performed in CP20 and OV90 cells. It is however unclear as to why the authors chose these cell lines among all those they tested. Their choice should be explained.

Response:
We appreciate the reviewer's thoughtful suggestions. Though the expression of MICU1 was significantly higher in most of the ovarian cancer cell line, we selected CP20 and OV90 for our studies because we wanted to examine the role of MICU1 in chemoresistant and aggressive cancer cell lines. The CP20 cell line was developed by treating cells with increasing concentrations of cisplatin and presents a chemoresistance cell line model [Sood AK et al., The American journal of pathology 158, 1279-1288(2001]. The OV90 cell line was derived from ascites of a grade 3, stage IIIC ovarian cancer patient and hence provides a relevant model to demonstrate proof-of-concept role of MICU1 in ovarian cancer [Provencher DM et al., In vitro cellular & developmental biology. Animal 36, 357-361 (2000)]. The rationale has now been explained in the result section.

Reviewer2: 2. Throughout the manuscript, the authors have used only one test for proliferation/migration/invasion. Confirmation of these preliminary results with different assays would go a long way to strengthen the authors' conclusions.
Response: We appreciate the reviewer's suggestions. As suggested, we have now expanded our studies with additional tests for proliferation (BrDU incorporation assay), migration/invasion (Boyden Chamber based Calcein-AM assay and gelatin degradation assay). Similar to the CyQUANT assay, the BrDU incorporation assay also demonstrated a dose dependent reduction in BrDU incorporation upon cisplatin treatment in the MICU1 silenced cells, compared to the control, indicating sensitization of tumor cells to cisplatin upon MICU1 silencing. Furthermore, similar to the CyQUANT assay, MICU1 silencing alone did not inhibit proliferation as evidenced by the absence of changes in BrDU incorporation upon MICU1 silencing. (Supporting Figures S4 A,B) We next performed the Boyden-Chamber assay and quantified cell migration or invasion by determining Calcein-AM fluorescence to further support a role of MICU1 in cell migration and invasion ( Figures  2D,E). Silencing MICU1 in CP20 or OV90 cells significantly reduced migration (~72% in CP20 and 63% in OV90 cells) and invasion (~76% in CP20 and 74% in OV90 cells). In addition, we performed the gelatin-degradation assay to further support a role for MICU1 in cellular invasion (Figures 2F,G). Extent of degradation of the gelatin matrix provides a measure of the invasive potential of cancer cells and our results demonstrate that MICU1 silenced CP20 or OV90 cells have significantly compromised gelatin-degrading capabilities. These new experiments and the results have been now been incorporated in the methods, results and discussion sections of the revised manuscript.
Reviewer2: 3. The authors chose cisplatin as their reference chemotherapeutic drug. Are similar results obtained with a different drug, or are they exclusive to this compound?

Response:
We appreciate the reviewer's insightful comments. We have now investigated the effect of other standard chemotherapeutics such as topotecan, paclitaxel and doxorubicin on ovarian cancer cell proliferation upon MICU1 silencing. We observed that MICU1 silencing sensitized ovarian cancer cell to topotecan, paclitaxel as well as doxorubicin (Supporting Figures S5). Interestingly, prior reports suggest that doxorubicin [Biochim Biophys Acta. 1813(6):1144-52.(2011] and paclitaxel [Kidd et al., The Journal of Biological Chemistry 277, 6504-6510 (2002)] affect calcium homeostasis in cancer cells to exert their cytotoxic effects and further support our hypothesis that MICU1 imparts resistance to their cytotoxicity by serving as a gate keeper preventing [Ca 2+ ] m overload.
Reviewer2: 4. In Fig. 3F, the authors use mitoTEMPO to scavenge ROS. A control experiment showing that this is actually the case in their models is missing. They could use MitoSOX dye, as shown in Fig.  3E.

Response:
We appreciate the reviewer's thoughtful suggestions. As suggested we performed control experiments showing mitoTEMPO mediated scavenging of mtROS as evident by reduced staining with the MitoSOX dye (Supporting Figure S7). Pre-incubation of the MICU1 silenced CP20 cells with the mitochondrial ROS scavenger mitoTEMPO (10M) resulted in quenching of mtROS as determined by a reduction in MitoSOX staining compared to untreated control (Supporting Figure  S7). Extending these investigations, the fold change in of apoptosis in cisplatin treated MICU1 silenced cells significantly decreased in the presence of mitoTEMPO ( Figure 3F). These results suggest that silencing MICU1 generates mtROS in ovarian cancer cells facilitating a cisplatinmediated apoptotic response.
Reviewer2: 5. Fig. 4A: while the spike in Ca2+m induced by cisplatin seems indeed different between siCTL and siMICU1-treated cell, this difference disappears within 400s. Are the authors suggesting that cell fate will be drastically altered by these initial 400s? They should carefully study the kinetics of their responses, and show results proving long-lasting differences between their conditions.  et al., Cell 151, 630-644 (2012)]. We posit the rapid change in [Ca 2+ ] m results from influx of the cisplatin-induced release of Ca 2+ from endoplasmic reticulum (ER) , but nevertheless were sufficient enough to activate dehydrogenases and signaling cascades [Rizzuto et al., Science. 262:744-747 (1993) and Denton et al., Biochim Biophys Acta 1787, 1309-1316(2009].

Response
Reviewer2: 6. Fig. 4E Figure S12). Reviewer2: 7. page 12: they authors hypothesize that MICU1 will lead to "Ca2+ overload in the mitochondria". They should show evidence of this: dyes are available for that kind of measurement, and they need to substantiate the fact that calcium homeostasis is indeed modified by MICU1 silencing.

Response:
We appreciate the reviewer's thoughtful suggestions. As suggested we performed our [Ca 2+ ] m measurements using the highly sensitive and selective protein-based mitochondrial Ca 2+ probe GCaMP2-mt [Kirichok et al., Nature 427, 360-364 (2004) and Doonan et al., The FASEB Journal vol. 28 no. 11, 4936-4949 (2014)]. We have repeated our experiments in different cell lines expressing GCaMP2-mt to further confirm regulation of [Ca 2+ ] m by MICU1 and its effect on cisplatin treatment (Figures 4D, F, Supporting Figures S10, S11). In brief, stable cell lines shCTL-OV90 and shMICU1-OV90, were transfected with GCaMP2-mt and mitochondrial Ca +2 response monitored after cisplatin exposure ( Figure 4D). Compared to scrambled control, MICU1 knockdown increased [Ca 2+ ] m by ~5 fold upon cisplatin treatment. Similar effects were observed in siMICU1-CP20 cells compared to siCTL-CP20 cells (Supporting Figure S10) Figure  4E). However, ectopic expression of Flag-MICU1 in OSE cells significantly prevented cisplatinmediated [Ca 2+ ] m overload as compared to the EV-expressing control cells (Figure 4E, F), confirming the role of MICU1 as a negative regulator of mitochondrial calcium uptake. Importantly, OSE cells harboring Flag-MICU1 demonstrate higher pPDH (ser293) expression (Supporting Figure S17) and decrease in PDH activity as compared to the control cells ( Figure 6I). These results further support that though ectopic expression of MICU1 in absence of any stimulation decreases [Ca 2+ ] m marginally, but this small decrement is significant enough to inhibit PDP-PDH interaction ( Figure 6G) leading to pPDH accumulation. Indeed MICU1 silencing alone in Hela cells did not affect the [Ca 2+ ] m of unstimulated cells yet possessed significantly different phosphoPDH levels [Mallilankaraman et al., Cell 151, 630-644 (2012)].
Reviewer2: 9. The in-vivo results are quite puzzling: first, Fig. 7B seems to indicate that all 10 animals injected with shCTL cells died on the same day. Is it really that synchronized? Second, the conclusion that cells without MICU1 are more sensitive to cisplatin is invalidated by these experiments. Indeed, on both populations (shCTL and shMICU1 injected mice), cisplatin treatment gave the same extra 7 days of survival. The conclusion from this experiment should rather be that MICU1 and cisplatin act through distinct pathways, as also illustrated in Fig. 7A where tumor growth inhibition by cisplatin seems to be proportionally identical between shCTL and shMICU1.

Response:
We appreciate the reviewer's concern and regret prior lack of clarification. Before the initiation of treatment, the animals were randomized in a blinded fashion. We observed that all the animals injected with shCTL cells (tumor growth study as well as survival study) developed large tumors (~1.469g), were moribund, appeared in distress and hence were euthanized along with the experimental groups for tumor growth study on the same day to avoid undue suffering according to the animal protocol approved by the institutional Animal Care and Use Committee (IACUC). We used same IACUC approved criteria to euthanize animals for all other groups in survival studies. We have now corrected and included the median survival values for the different groups ( Figure 7B). It is evident from the table that cisplatin treatment in shCTL group yielded 2 extra days of survival while animals in shMICU1 group survived for 43 days. In addition, cisplatin therapy further prolonged their survival by 10 more days. It is also important to note that though tumor growth inhibition by cisplatin appears to be proportionally identical between shCTL and shMICU1 groups; in the shMICU1 group, 4 out of 5 animals showed negligible tumor formation upon cisplatin treatment (80% of the sample size) as compared to 1 out of 5 animals in shCTL+cisplatin group (20% of the sample size), supporting a role of MICU1 in drug resistance. However, we agree that there may be other unknown pro-oncogenic function of MICU1 that will evolve over time.

Reviewer3:
The MS by Chakraborty and colleagues is a follow up of the authors' early observation that small interfering RNA-mediated silencing of the mitochondrial uniporter MICU1 is able to sensitize ovarian cancer cells to positively charged gold nanoparticles. This observation suggested a potential role of MICU1 in drug resistance. It is known that MICU1 resides in the inner mitochondrial membrane (IMM) and required for Ca2+ uptake through mitochondrial Ca2+ 'uniporter' (MCU), a calcium-selective ion channel responsible for low-affinity calcium uptake into the mitochondrial matrix. It has been shown that MCU expression is correlated with cancer progression in sporadic studies (prostate, colon and breast). However, the functional role for MICU1 in cancer is still unknown. In the present MS the authors suggesting new function of MICU1 that regulates metabolic fate and confers chemoresistance in ovarian cancer. Moreover, they tried to make a link between obtained observation and cancer