Cystathionine gamma-lyase-mediated hypoxia inducible factor 1-alpha expression drives clear cell ovarian cancer progression

AME‐N and DGH conceptualized and planned the experiments, conducted histopathology and immunohistochemistry (IHC) scoring, supervised the project, wrote, reviewed, and edited the manuscript, and secured funding. AME‐N additionally performed functional assays, western blotting, immunofluorescence (IF), and formal analysis. YL, BT, YD, LW, GT‐G, JS, AL and CS were responsible for cell culture maintenance, western blotting, RNA extraction, cDNA preparation, and qPCR, as well as contributing to data analysis. ST and YQ were responsible for H&E and IHC staining, include antibodies optimization. SYC and YW conducted the animal studies, including protocol preparation and submission, xenotransplantation, monitoring, and tissue collection. FK, RV and KP managed organoid cultures, including the processing of endometrial tissue samples, cultivation of endometrial cells, and organoid growth, along with the collection and processing of endometriotic cyst content. BL and JMD carried out bioinformatic analyses using publicly available datasets. SL, CSH, GLN and SES provided critical feedback and participated in manuscript review and editing. MMMW provided extensive support for funding applications and manuscript editing. HZ, PJY, GBM and PHS contributed resources and feedback. All authors reviewed the manuscript, discussed the results, and contributed to its final version.

To investigate the potential contribution of CTH to the aggressive phenotype of CCOC, we first generated CTH KO cells using CRISPR/Cas9 in four well‐characterized CCOC cell lines [ 19 , 20 ] (Figure  1A ). No significant morphological changes were observed in cells with or without CTH KO under ambient conditions. However, CTH KO cells exhibited a tendency toward increased proliferation, enhanced cell death (as shown for OVISE cells in supplementary material, Figure  S1A,B ), and reduced viability (Figure  1B ). Moreover, CTH KO cells showed significantly reduced migratory and invasion capacity in vitro as measured by Boyden chamber migration and invasion assays (supplementary material, Figure  S1C,D ). In addition, cells grown on Matrigel 3D matrices exhibited an enhanced outspreading growth pattern in control cells, in contrast to the confined growth pattern observed in CTH KO cells (supplementary material, Figure  S1E ). These findings strongly support a role for CTH in CCOC aggressiveness. To assess the impact of CTH KO on CCOC cell invasiveness, we utilized two murine animal models: subcapsular renal implantation (SRI) and tail‐vein injection. These models were selected based on growing evidence that hematogenous spread plays an important role in epithelial ovarian cancer and in particular CCOC metastasis, which has been previously overlooked [ 21 , 22 ]. The SRI model is a well‐established model of cancer progression [ 17 , 23 , 24 , 25 ]. OVISE cells with or without CTH KO were implanted under the renal capsules of immunocompromised 6‐ to 8‐week‐old NRG female mice to assess local growth and metastasis. In contrast to CTH ‐competent OVISE control cells, CTH KO OVISE cells either did not grow or formed very small tumors in the kidney, which failed to infiltrate intra‐abdominal structures or distantly invade lungs (Figure  1C–F , and supplementary material, Figure  S1F ). CTH KO OVISE xenografts showed no evidence of reduced proliferation; instead, there was clear evidence of elevated cell death, as indicated by increased expression of cleaved caspase‐3 (supplementary material, Figure  S1G,H ). CTH expression in OVISE tumor xenografts was validated using IHC (supplementary material, Figure  S1I , left panels). Moreover, the CCOC marker hepatocyte nuclear factor 1 beta (HNF1β) [ 26 ] showed similar expression in OVISE tumor xenografts with or without CTH KO, suggesting histotypic preservation (supplementary material, Figure  S1I , right panels). Furthermore, we also investigated the effects of CTH loss on the later phase of the metastatic cascade by intravenous tail‐vein injection of RMG‐I cells with or without CTH KO into NRG female mice. In contrast to control cells that formed substantial lung metastases, CTH KO RMG‐I cells formed either small solitary lung lesions or grew no lung metastatic lesions (Figure  1G–J , and supplementary material, Figure  S1J ). As with the OVISE xenografts, CTH and HNF1β expression were validated using IHC (supplementary material, Figure  S1K ). Together, these data highlight CTH's key role in CCOC progression. CTH facilitates CCOC metastatic progression. (A) Western blotting of CTH protein expression in CCOC cell lines transduced with negative‐control sgRNA or two independent sgRNAs targeting CTH . GAPDH was used as a loading control. (B) MTT cell viability assay conducted on OVISE and OVMANA cells, with or without CTH KO ( n  = 3). (C and D) Representative photomicrographs of H&E‐stained tissue sections of primary tumor xenografts and metastatic lesions developed in mice with renal subcapsular implantation (RSI) of OVISE cells, with or without CTH KO, at low (scale bars, 500 μm) and high (scale bars, 50 μm) magnifications. (E) Average tumor size (in millimeters), measured as the maximum dimension, of OVISE tumor xenografts with or without CTH KO described in panels (C and D). (F) Total number of mice bearing xenografts of OVISE CCOC cells, with or without CTH KO, which developed lung metastases, analyzed using Fisher's exact test. (G) H&E‐stained sections of metastatic lesions developed in mice bearing RMG‐I CCOC cells, with or without CTH KO. Scale bars, 500 μm. (H) Graph illustrating total number of mice that developed lung metastases, analyzed using Fisher's exact test. (I) Box plot showing size of lung metastases, measured as maximum dimension (in millimeters) developed in mice bearing tail‐vein injection of RMG‐I cells, with or without CTH KO. In panel (I), data, mean values, and 10th–90th percentile. (J) Graph representing the average number of lung metastases developed in mice bearing RMG‐I CCOC cells, with or without CTH KO. Data are presented as mean ± SD. Statistical significance was determined using unpaired two‐tailed Student's t ‐test in all panels, except for panel (F) and panel (H) where Fisher's exact test was used. * p  < 0.05, ** p  < 0.005, *** p  < 0.0005; n.s., nonsignificant. See also Supplementary  materials and methods, supplementary material, Figure  S1 . CTH plays a crucial role in de novo cysteine synthesis, which is essential for GSH synthesis [ 16 ]. GSH is critical for cellular adaptation to increased levels of reactive oxygen species (ROS) [ 27 ]. CCOC develops from ovarian endometrioma or endometriotic cysts containing degraded blood and cellular debris, a ROS‐rich microenvironment [ 8 , 28 ]. CTH protein expression increased under oxidative stress induced by the ROS‐inducing agent Piperlongumine (PL) [ 29 ] (Figure  2A ). Consistent with its protective role against elevated ROS, treatment of CCOC cells with or without CTH KO with PL resulted in significant cell death of CTH KO cells (supplementary material, Figure  S2A,B ). This was associated with increased ROS levels as measured by the ROS assay 2’,7’‐dichlorofluorescin diacetate (CM‐H 2 DCFDA) (Figure  2B and supplementary material, Figure  S2C ), as well as a decrease in reduced‐to‐oxidized glutathione ratios (GSH/GSSG) (Figure  2C and supplementary material, Figure  S2D ). Since redox imbalance is significantly associated with ferroptosis [ 30 ], we considered if CTH might protect CCOC cells from ferroptosis. CTH KO cells exhibited increased basal expression of the ferroptosis marker transferrin receptor 1 (TfR1) [ 31 ], indicating increased lipid peroxidation. This effect was intensified by the ferroptosis inducer erastin [ 32 ] and partially reversed by ferrostatin‐1 (Fer‐1) [ 33 ]. Notably, re‐expression of CTH in KO cells, using a previously described shRNA‐resistant plasmid that is also CRISPR‐insensitive and contains the coding sequence (CDS) region while lacking the 3’ and 5’ untranslated regions (UTRs) [ 34 ], resulted in reduced basal ROS levels and a significant decrease in ROS levels upon erastin treatment (Figure  2D,E , and supplementary material, Figure  S2E ). In addition, the BODIPY™ 581/591 C11 (another well‐known lipid peroxidation sensor) [ 35 ] confirmed a significant increase in lipid peroxidation in CTH KO cells (Figure  2F,G , and supplementary material, Figure  S2F,G ). Moreover, IHC analysis of 4‐hydroxy‐2‐nonenal (4‐HNE), an additional indicator of lipid peroxidation [ 36 ], showed higher expression levels in CTH KO OVISE xenografts (supplementary material, Figure  S2H ). These data support a crucial role for CTH in redox regulation and protection against ferroptosis in CCOC. CTH protects CCOC cells against oxidative stress. (A) CTH protein induction in CCOC cells in response to 10 μm Piperlongumine (PL) treatment, determined by immunoblotting analysis, with GRB2 used as a loading control. (B) ROS levels in OVISE cells, with or without CTH KO, treated with vehicle (0.1% DMSO) or with 10 μ m PL for 24 h, assessed using CM‐H2DCFDA with ROS levels normalized to protein content ( n  = 3). (C) OVISE cells, with or without CTH KO, treated with either vehicle or 10 μ m PL for 24 h. Reduced/oxidized glutathione (GSH/GSSG) ratios were measured as a readout for redox stress ( n  = 3). (D) Immunofluorescence (IF) detection of TfR1 expression in OVISE cells, with or without CTH KO, or with CTH KO rescued with CRISPR‐insensitive CTH expressing vector, treated with vehicle alone (0.1% DMSO), 5 μmol/l erastin, or combined erastin and 2 μmol/l Fer‐1 for 12 h. Scale bar, 20 μm. (E) Quantification of 12 fields of view (FOVs) representing n  = 3 using ImageJ software. Data are presented as mean values and 10th–90th percentile. (F) Representative images of C11‐BODIPY™ (581/591) staining for lipid peroxidation in OVISE cells, with and without CTH KO, treated with either vehicle or 2 μm PL for 24 h. Cumene hydroperoxide was used as a positive control. (G) Quantification of lipid peroxidation in OVISE cells using a microplate reader for n  = 2, each performed in quadruplicate. Error bars represent the SD. Scale bar, 20 μm. Statistical analysis was conducted using an unpaired two‐tailed Student's t ‐test. Error bars represent the SD. * p  < 0.05, ** p  < 0.005, *** p  < 0.0005; n.s., nonsignificant. See also Supplementary  materials and methods, supplementary material, Figures  S2 and S3 . Next, we investigated whether the observed decrease in cell viability and motility in CTH KO cells could be rescued by restoring redox balance. Using the cell‐permeable GSH monoethyl ester (GSH‐MEE), as previously described [ 34 ], we found that while GSH‐MEE completely suppressed the induction of NFR2, a key regulator of the antioxidative response, indicative of its activity, it only partially rescued the viability of CTH KO cells under PL stress (supplementary material, Figure  S3A,B ). Since high ROS levels are known to impede tumor migratory and invasive phenotype [ 37 , 38 ], we investigated whether antioxidants could rescue the motility of CTH KO cells. CTH‐deficient OVISE cells demonstrated significantly reduced cell motility as assessed by scratch assays (supplementary material, Figure  S3C,D ), which could not be fully rescued by the commonly used antioxidant N‐acetyl cysteine (NAC) [ 39 , 40 ] or Trolox [ 41 ]. Together, these data suggest that CTH plays a critical role in CCOC that extends beyond redox regulation. Among epithelial ovarian cancer subtypes, CCOC is distinguished by a strong hypoxic signature and elevated expression of HIF1α protein, which drives its aggressive behavior [ 42 ]. However, the link between CTH, hypoxia, and HIF1α in CCOC remains insufficiently explored. We first examined CTH protein expression under hypoxia, which revealed enhanced expression (Figure  3A ), in line with previous studies indicating hypoxia‐induced upregulation of CTH [ 43 , 44 ]. Unlike CCOC cells, CTH protein expression levels remained unchanged under hypoxia in non‐CCOC cells, including the high‐grade carcinoma (HGCS) cells, OVCAR3, and HEY (Figure  3B ), highlighting the histotype‐specific role of CTH in CCOC. Additionally, we investigated whether other TSS pathway enzymes, including cystathionine beta‐synthase (CBS) and 3‐mercaptopyruvate sulfurtransferase (MPST), were altered under hypoxia, similar to CTH. Indeed, neither CBS nor MPST was induced under hypoxia in CCOC (Figure  3C ), highlighting that this response was specific to CTH. Next, we examined the effects of HIF1α inhibition on CTH protein expression. Silencing HIF1A using siRNAs was observed to have no effect on CTH protein expression in CCOC cells (supplementary material, Figure  S4A ). CTH facilitates HIF1α expression. (A) Western blotting shows CTH protein induction in OVISE and OVMANA cells incubated under hypoxia (1% O 2 , 2 h), with GRB2 used as a loading control. (B) Western blotting showing CTH protein expression in nonclear ovarian cancer cell lines OVCAR3 and HEY, incubated under 1% O 2 for 2 h. Vinculin was used as a loading control. (C) Western blotting shows effects of hypoxia (1% O 2 , 2 h) on protein expression levels of cystathionine beta‐synthase (CBS) and 3‐mercaptopyruvate sulfurtransferase (MPST) in CCOC cell lines. Vinculin was used as a loading control. (D) Western blotting shows effects of CTH KO on HIF1α protein expression levels in OVISE cells, incubated under −/+ hypoxia (1% O 2 , 4 h). GAPDH was used as a loading control. (E) Left panels: IHC of HIF1α expression in lung metastases of RMG‐I cells with or without CTH KO. Right panel: quantification of HIF1α staining intensity for six fields of view (FOVs), representing n  = 3 lungs with metastases per group. (F) Left panels: IHC of HIF1α expression in tumor xenografts of OVISE cells with or without CTH KO. Right panel: quantification of HIF1α staining intensity in 12 FOVs representing n  = 3 tumors per group. (G) Left panels: IHC of VEGFA in OVISE tumor xenografts with or without CTH KO. Right panels: quantification of staining intensity in six representative images ( n  = 3 tumors per group) was performed using ImageJ with the color deconvolution plug‐in. (H) Left panels: IHC of CD31 expression in tumor xenografts of OVISE cells with or without CTH KO. Right panel: quantification of tumor microvessel density in 12 FOVs representing n  = 3 tumors per group. Scale bar, 50 μm for panels (E–H). (I) CTH and HIF1A mRNA expression in indicated cells determined by RT‐qPCR. Data were normalized against GAPDH and expressed as fold change ± SD, in n  = 3 experiments. (J) Hydrogen sulfide (H 2 S) levels in OVISE cells, with or without CTH KO, or with CTH KO rescued with CRISPR‐insensitive CTH ‐expressing vector, detected using the H 2 S fluorescence probe, P3 ( n  = 3 experiments, each performed in quadruplicate). (K) Western blotting analysis of CTH, CBS, and MPST protein expression levels in CTH KO OVISE cells transduced with either shCTRL or shRNA targeting CBS . Vinculin was used as a loading control. (L) H 2 S levels, detected using H 2 S fluorescence probe HSip‐1, in CTH KO OVISE cells transduced with either shCTRL or shRNA targeting CBS , normalized to protein content. Data are presented as mean ± SD of n  = 3 experiments. (M) CBS activity, measured using CBS assay kit in OVISE cells with and without CTH KO, normalized to protein content. Data are presented as mean ± SD of n  = 3 experiments. (N) Western blotting shows HIF1α protein restoration in CTH KO OVISE cells transfected with plasmid expressing arginine (R) to alanine (A) CTH mutant (R62A). GAPDH was used as loading control. Statistical analysis was conducted using unpaired two‐tailed Student's t ‐test. Data are presented as mean ± SD. a.u. = arbitrary units, kd = knockdown. * p  < 0.05, ** p  < 0.005, *** p  < 0.0005. See also Supplementary  materials and methods, supplementary material, Figure  S4 . Examination of HIF1α expression in cells with or without CTH KO revealed that hypoxia‐induced HIF1α upregulation was significantly attenuated in CTH KO cell lines (Figure  3D ). Consistent with in vitro data, HIF1α protein levels were also significantly downregulated in CTH KO RMG‐I lung metastases, as assessed by IHC (Figure  3E ), and in CTH KO OVISE xenografts (Figure  3F ). In the latter, this downregulation was associated with reduced VEGFA expression and decreased blood vessel formation (Figure  3G,H ). To investigate whether other enzymes in the TSS pathway contributed to HIF1α regulation, we used siRNAs to target CBS (three individual siRNAs) and MPST (two individual siRNAs). As shown in supplementary material, Figure  S4B , silencing of CBS or MPST did not affect HIF1α protein expression under hypoxia in OVISE and RMG‐I cells, highlighting a specific link between CTH and HIF1α in CCOC. To gain further insights into how CTH facilitates HIF1α expression, we compared total HIF1A mRNA levels in OVISE cells with or without CTH KO grown under hypoxia by reverse transcription quantitative PCR (RT‐qPCR). Total HIF1A mRNA levels in hypoxic OVISE cells were relatively higher in CTH KO cells, though this was not statistically significant (Figure  3I ), arguing against a transcriptional mechanism for CTH‐mediated HIF1α expression. Furthermore, HIF1A mRNA decay assay, using a well‐known transcriptional inhibitor actinomycin D [ 45 ] showed no significant difference between OVISE cells with and without CTH KO (supplementary material, Figure  S4C ). In addition, investigation of four publicly available datasets [ 8 , 46 , 47 , 48 ] did not reveal a correlation between HIF1A mRNA and CTH mRNA expression (supplementary material, Figure  S4D ). Together, these data suggest a post‐transcriptional mechanism is involved in the regulation of CTH‐mediated HIF1α protein expression. Next, we investigated whether CTH loss affected HIF1α protein stability, as HIF1α stability is a critical post‐transcriptional regulatory mechanism and a key step in cellular adaptation to hypoxia [ 49 , 50 ]. Neither the proteasome inhibitor MG132 [ 51 ] nor the prolyl hydroxylase inhibitor DMOG [ 52 ] was able to restore HIF1α in CTH KO cells to comparable levels to CTH ‐competent cells (supplementary material, Figure  S4E,F ), arguing against enhanced degradation being the mechanism involved. We further assessed the effects of CTH on HIF1α protein stability by treating OVISE cells with or without CTH KO with cycloheximide (CHX) to block translation [ 53 ] and measuring HIF1α protein degradation rates. As shown in supplementary material, Figure  S4G,H , the degradation rates were almost comparable between CTH‐competent and CTH‐lacking cells, arguing against CTH regulating HIF1α protein stability. Together, these data strongly suggest that a post‐transcriptional process, neither involving degradation nor stability mechanisms, contributes to CTH‐mediated HIF1α expression. Traditional views of CTH as a key TSS pathway enzyme attribute its effects to cysteine and hydrogen sulfide (H 2 S) production [ 54 , 55 ]. H 2 S was previously reported to play an important role in non‐small cell lung cancer (NSCLC) angiogenesis through the activation of HIF1α [ 56 ]. Furthermore, multiple studies have linked H 2 S to HIF1α expression, with evidence suggesting that H 2 S can either enhance [ 57 , 58 , 59 ] or inhibit HIF1α expression [ 60 ]. To explore this further, we measured H 2 S levels using a previously described P3 probe [ 61 ] in OVISE cells with and without CTH KO. Unexpectedly, CTH KO cells exhibited relatively higher H 2 S levels compared to control cells, which were reduced upon rescue with CRISPR‐insensitive CTH ‐expressing plasmid (Figure  3J ). This unexpected increase in H 2 S levels may be attributed to elevated CBS protein expression, likely a compensatory response, observed in CTH KO OVISE cells (Figure  3K ). Silencing CBS in these CTH KO OVISE cells using shRNA significantly reduced H 2 S levels, as assessed by H 2 S fluorescence probe HSip‐1, as previously described [ 62 , 63 ] (Figure  3L ). Additionally, the increased CBS protein levels in CTH KO cells (Figure  3K ) were associated with enhanced CBS activity, as measured using a CBS activity assay (Figure  3M ). These findings suggest that CBS plays a predominant role in H 2 S production in CCOC cells under conditions of CTH deficiency. Next, we investigated whether the reduced HIF1α protein levels observed in CTH KO cells could be attributed to elevated H 2 S, given previous reports that H 2 S inhibited HIF1α expression [ 60 ]. Treatment of CCOC cells with a high dose (1 m m ) of the slow‐releasing H 2 S donor GYY4137 or a sublethal dose (10 μ m ) of the fast‐releasing H 2 S donor sodium sulfide (Na 2 S) did not alter HIF1α expression, ruling out this possibility (supplementary material, Figure  S4I,J ). Finally, arginine 62 (R62) has been reported to be essential for the enzymatic activity of CTH [ 64 ]. Therefore, we hypothesized that this activity might be essential for CTH‐driven HIF1α expression. Introduction of arginine (R) to alanine (A) CTH mutant (R62A) into CTH KO OVISE cells significantly restored HIF1α expression (Figure  3N ), providing strong evidence for potential nonenzymatic functions of CTH. Taken together, our findings suggest that CTH regulates HIF1α expression through an H 2 S‐independent post‐transcriptional mechanism. We next assessed whether HIF1α was important for CTH‐driven invasion and metastasis of CCOC cells. Similar to CTH KO, HIF1A knockdown (kd) with two independent siRNAs in CTH‐competent OVISE cells significantly inhibited in vitro cell motility. Furthermore, rescuing CTH KO cells with a previously reported WT HIF1A ‐expressing plasmid [ 17 ] effectively restored cell motility (supplementary material, Figure  S5A–D ). Notably, these effects were observed under normoxic conditions, aligning with previous reports that support the role of HIF1α in normoxia [ 65 , 66 , 67 ]. We then expressed WT HIF1A in CTH KO RMG‐I cells and used tail‐vein injection to monitor lung metastatic progression, which revealed that the metastatic burden of CTH KO cells was completely rescued, highlighting HIF1α as a critical contributor to CTH‐mediated CCOC metastasis (Figure  4A,B ). Expression levels of HIF1α and CTH were determined by IHC, showing reduced HIF1α in CTH KO lung mets and its restoration in mets formed by RMG‐I cells transfected with a HIF1A ‐expressing plasmid. CTH loss was maintained in both CTH KO and CTH KO/HIF1α rescue groups (Figure  4C,D and supplementary material, Figure  S5E ). Together, these data highlight HIF1α as a significant contributor to CTH‐mediated metastatic progression of CCOC cells. To further assess whether CTH regulates HIF1α in human CCOC, a small tissue microarray (TMA) consisting of 84 CCOC tumors was immunostained for CTH and HIF1α. CTH expression was found to be significantly correlated with HIF1α expression (Figure  4E,F ). These data provide strong evidence that CTH positively regulates HIF1α protein expression in CCOC. HIF1α restoration in CTH KO cells restores their metastatic phenotype. (A) H&E‐stained sections of metastatic lesions developed in mice bearing RMG‐I CCOC cells, with or without CTH KO or with CTH KO combined with HIF1α rescue. Scale bar, 500 μm. (B) Graphical representation of metastatic burden in the three groups described in panel (A), calculated as percentage of lung occupied by metastases, using the following formula: (lung metastasis area / total lung area) × 100. Values are presented as mean ± SD. (C) Representative low‐ (scale bar, 100 μm) and high‐magnification (scale bar, 50 μm) images of IHC staining for HIF1α expression in lung metastases of RMG‐I cells, with or without CTH KO or with CTH KO combined with HIF1α rescue. (D) Quantification of HIF1α staining intensity for nine fields of view, representing n  = 3 lungs with metastases per group. Staining intensity was quantified with ImageJ using the color deconvolution plug‐in. (E) Representative images of IHC analysis of CTH and HIF1α proteins in clinical samples, shown at low (×4) and high magnification (×20). The 20× objective images are subsets of the 4× objective images. Scale bar, 100 μm. (F) Correlation between CTH and HIF1α protein expression in CCOC primary tumors using IHC on tissue microarrays ( n  = 84 cases). Pearson correlation estimate (R) is shown at the top of the right panel graph (with 95% confidence intervals shown in brackets). The black line shows a linear regression model fit. (G) MTT cell viability assay shows the effects of cisplatin treatment (10 −9  mg/ml, 96 h) on OVISE and OVMANA cells, with or without CTH KO, for n  = 3 experiments. CTH KO increases CCOC cell sensitivity to cisplatin chemotherapy. Statistical analysis was conducted using an unpaired two‐tailed Student's t ‐test. Data are presented as mean ± SD. * p  < 0.05, ** p  < 0.005, *** p  < 0.0005. See also Supplementary  materials and methods, supplementary material, Figures  S5 and S6 . To test whether CTH‐mediated HIF1α expression occurs in other systems, we assessed Ewing sarcoma (EwS) xenografts, with or without CTH kd, since CTH was recently implicated in EwS metastasis [ 34 ]. CTH‐deficient EwS xenografts, generated using two independent shRNAs [ 34 ], showed significant downregulation of HIF1α, VEGF, and CD31, similar to CCOC (supplementary material, Figure  S6A–E ), all of which were rescued by CTH re‐expression using shRNA‐resistant CTH ‐expressing plasmid. We noted that CTH‐deficient EwS xenografts showed reduced expression of carbonic anhydrase IX (CAIX), a well‐established HIF1α transcriptional target [ 68 ], but CTH re‐expression significantly rescued its expression [supplementary material, Figure  S6A (right panels) and S6F ]. In contrast, CAIX was not expressed in either control or CTH KO RMG‐I xenografts (supplementary material, Figure  S6G ). This suggests that hypoxia response components are tailored to specific cell context. Furthermore, the clear cell renal cell carcinoma cell line RCC4, which is known to stabilize HIF1α expression due to inactivation of ECV E3 ligase through loss‐of‐function mutations in VHL [ 49 , 69 ], still showed a significant reduction of HIF1α expression upon CTH inhibition using siRNA. This reduction occurred despite low CTH protein expression levels (supplementary material, Figure  S6H ), suggesting that CTH‐mediated HIF1α expression is not through protein stability but instead highlights a general role for CTH in regulating HIF1α expression. Lastly, our finding that lack of CTH induces cell death, notably under stress, raised the question as to whether these features could be exploited therapeutically. OVISE and OVMANA cells with or without CTH KO treated with cisplatin for 96 h showed significantly enhanced cell death in CTH KO cells (Figure  4G ), consistent with other reports [ 70 , 71 ], highlighting the therapeutic utility of co‐targeting CTH in CCOC. Endometriosis of the ovary is associated with an increased risk of transformation and development of CCOC. We previously showed that CTH was highly expressed in CCOC and adjacent endometriosis [ 7 ]. Therefore, to investigate whether CTH played a critical role in normal endometrial cells, we examined normal Müllerian tract‐derived cells, including endometrial cells. This study revealed that these cells expressed low levels of CTH, except for the ciliated cells, where CTH was abundantly expressed (Figure  5A,B ). Similarly, CTH expression levels were high in the remnant ectopic endometrial epithelium of ovarian endometriotic cyst walls (Figure  5C ). Using an organoid modeling system, established from benign endometrial tissues obtained from patients who underwent surgery for noncancerous lesions (e.g. fibroids) [ 7 , 14 , 72 ], we found that CTH was induced under hypoxia (Figure  5D ), similar to CCOC, further highlighting the critical role for CTH in adaptation to hypoxia. Next, endometriotic cyst content, known to contain elevated levels of ROS [ 28 ], was obtained from ovarian endometrioma patients who underwent surgery and processed as previously described [ 73 ], with minor modifications (Figure  5E ). The post‐centrifugation supernatants, free from blood and cellular debris that could interfere with the culture process, were used for treatment. CCOC cells treated with cyst supernatants showed enhanced CTH expression, suggesting a link between cyst content and CTH functionality (Figure  5F ). To further investigate the effects of CTH on endometrial organoid growth under ambient conditions or stress conditions, primary endometrial cells were transduced with Cas9‐expressing lentiviral vectors co‐expressing egfp and either a nontargeting single guide RNA (sgRNA) (NTCA1) or CTH sgRNAs ( CTH KO). Cells were then grown in 3D cultures with Matrigel to produce organoids, as previously described [ 7 ]. After 1 week, organoids were exposed to vehicle (0.1% DMSO) or cyst content treatments for a further week. In contrast to well‐formed, CTH ‐competent organoids, CTH KO organoids displayed a stunted growth pattern and insufficiently viable morphology (Figure  5G–I ). Cyst content treatment of control organoids enriched cells with enhanced CTH expression, as assessed by IHC, whereas CTH KO organoids showed stressed and crumbled morphologies (Figure  5J–L ). Induction of CTH under cyst content treatment (Figure  5K ), further supports its role in stress adaptation. In this system, CTH was expressed in all epithelial cells, not just ciliated cells, a pattern described previously by our team in atypical epithelial cells within endometriosis adjacent to CCOC [ 7 ]. CTH role in CCOC's precursor cells. (A) Left panels: H&E‐stained human fallopian tube (FT) and endometrial tissue sections. Scale bar, 50 μm. Right panels: IHC staining of CTH shows high expression in Mullerian tract‐derived ciliated cells. Scale bars, 50 and 10 μm, respectively. (B) Western blotting of CTH expression in the CCOC cells, OVISE and OVMANA, compared to primary endometrial cells. GAPDH was used as a loading control. (C) H&E‐stained and CTH‐stained ovarian endometriosis tissue sections show endometrial gland surrounded with endometrial stroma within ovarian tissue positively stained for CTH. Scale bars, 50 μm for left and middle panels and 10 μm for right panels. (D) Left top panels: brightfield (BF) photomicrographs of patient‐derived endometrial organoids grown under normoxia for 7 days, then subjected to normoxia or hypoxia conditions for 7 days. Scale bar, 200 μm. Left middle panels: H&E‐stained sections of organoids grown under normoxia or hypoxia, as described in top panels. Left lower panels: IHC detection of CTH. For left middle and lower panels, scale bar, 50 μm. Right panels: quantification of CTH staining in left lower panels of (D) was assessed in 12 fields of view per condition using ImageJ software and presented graphically. In panel (D), data are expressed as mean values and 10th–90th percentile. (E) Schematic representation of processing of endometriotic cyst content to produce the supernatant used for treatment of endometrial and CCOC cells grown in 2D or 3D cultures. Created with BioRender.com . (F) Western blotting of CTH expression in CCOC OVMANA cells treated with different samples of endometriotic cyst content, collected from different patients (cases 1–4), with vinculin and GAPDH used as loading controls. (G) Organoids derived from normal endometrium with −/+ CRISPR‐mediated CTH loss (labeled with Enhanced Green Fluorescent Protein) as detected using fluorescent microscopy. Scale bar, 200 μm. (H) Left panels: H&E‐stained sections of organoids described in (G). Right panels: IHC staining of CTH in organoids with or without CTH KO. Organoids with CTH loss displayed a significantly stunted formation pattern (asterisk). Scale bar, 50 μm. (I) Graph representing size of organoids with or without CTH KO, described in panel (G) ( n  = 51–52), assessed using ImageJ. (J) Organoids derived from normal endometrium with −/+ CRISPR‐mediated CTH loss, grown for 1 week and then treated with vehicle (0.1% DMSO), are shown in panels (G) and (H), or processed endometriotic cyst content for an additional week. Organoids were then visualized and assessed using fluorescent microscopy. Scale bar, 200 μm. (K) Left panels: H&E‐stained sections of organoids as described in panel (J). Right panels: IHC staining of CTH in organoids with or without CTH KO. CTH expression showed upregulation in control organoids [compared to panel (H), top right]. (L) Graph representing size of organoids with or without CTH KO, as described in panel (J) ( n  = 49–76), assessed using ImageJ. For panels (I) and (L), data represent mean values and 10th–90th percentiles, with statistical analysis performed using unpaired two‐tailed Student's t ‐test. * p  < 0.05, *** p  < 0.0005. Taken together, these findings suggest that CTH plays a crucial role in the development and aggressiveness of CCOC by regulating redox balance through GSH synthesis and enhancing HIF1α expression to promote angiogenesis. This in turn facilitates metastatic growth, ensuring cell survival in potentially lethal microenvironments (Figure  6 ). Proposed mechanism of CTH‐mediated progression in CCOC. Schematic depicting critical roles of CTH in CCOC. CTH is induced under stress of endometriotic cyst content, replenishing cellular GSH pool to protect against oxidative stress and ferroptosis. Independent of H 2 S production, CTH enhances HIF1α protein expression, defining CCOC cell state manifested by hypoxic signature, which in turn enhances VEGF expression and promotes angiogenesis and cancer progression. Created with Adobe Illustrator and Adobe Photoshop (Adobe Inc., https://www.adobe.com ), and Jmol ( http://jmol.sourceforge.net/ ).

Rewiring of tumor cell metabolism is central to all steps in cancer progression, sustaining an increased growth rate and adaptation to microenvironmental stresses such as hypoxia and oxidative stress [ 74 ]. This may be particularly important for cancers derived from oxidative stress and hypoxia‐rich environments such as the cancers that arise from endometriotic cysts [ 10 ]. The TSS pathway enables cells to deal with such stresses through the generation of vital molecules such as cysteine, an essential GSH precursor required for redox homeostasis, and the gaseous transmitter H 2 S; both playing critical roles in health and disease [ 54 , 75 ]. CTH, a key enzyme in the TSS pathway [ 16 ], has been increasingly implicated in cancer progression across various cancer types, including prostate cancer [ 76 ], breast cancer [ 77 ], and glioblastoma [ 78 ]. Further, CTH's role in preserving redox status has been shown to be pivotal for Ewing sarcoma metastasis [ 34 ]. Recently, our group reported CTH as a defining feature of CCOC and its precursor cells within endometriosis, but not other ovarian cancer subtypes [ 15 ]. CCOC is distinguished from other ovarian cancers by its strong hypoxic signature, which drives tumor aggressiveness [ 79 ]. However, a connection between CTH, HIF1α, and CCOC progression has not so far been established. Using in vitro molecular analyses of CCOC cells and in vivo studies, we showed that CTH loss in CCOC was associated with increased ROS levels and ferroptosis, consistent with its established role in redox regulation [ 80 , 81 ]. Notably, HIF1α was significantly downregulated in CTH KO CCOC cells under normoxia and hypoxia conditions. Importantly, this regulation occurs via an H 2 S‐independent post‐transcriptional mechanism. Our data show that CTH facilitates CCOC cell invasion and metastasis by enhancing HIF1α protein expression. The in vitro findings are supported by in vivo data, where CTH KO tumor xenografts exhibited a marked reduction in HIF1α and its transcriptional target, VEGFA, alongside reduced vascularity, potentially explaining the reduced metastatic capacity of these tumors. Forced expression of HIF1α in CTH KO cells significantly rescued tumor cell motility in vitro and fully rescued the metastatic burden in vivo . Furthermore, we presented data supporting clinically relevant findings demonstrating that CTH deficiency enhanced sensitivity to cisplatin treatment. This underscored the therapeutic potential of co‐targeting CTH in CCOC. Finally, the effects of CTH on HIF1α expression were evident in EwS xenografts and ccRCC4 cells, suggesting that the underlying mechanism may have broader implications in cancer biology. It remains to be determined how CTH regulates HIF1α expression independent of H 2 S. This is particularly relevant in light of emerging evidence suggesting that enzymes in the TSS pathway may have functions beyond H 2 S and cysteine production. For example, MPST was recently shown to regulate apoptosis by directly interacting with AKT and reducing its phosphorylation in an H 2 S‐independent manner [ 82 ]. Similarly, CTH protects endothelial cells from senescence by interacting with p53. This interaction sequesters p53 in the cytosol, preventing its acetylation and activation, and occurs independently of p53 S‐sulfhydration by H 2 S [ 83 ]. Collectively, these findings highlight the diverse and complex roles of TSS pathway enzymes in cellular processes, extending beyond their traditional functions in de novo cysteine synthesis and H 2 S production. In summary, we have put forth a model whereby CTH confers broad cell plasticity to CCOC and its precursor cells, enabling them to survive the hypoxic, ROS‐rich microenvironment of endometrial cysts through redox regulation and enhanced HIF1α expression. Given the strong expression of CTH in ciliated cells in endometrium [ 8 ], atypical endometriosis [ 7 ], and endometrium exposed to high hormonal levels during pregnancy (the Arias Stella reaction) [ 9 ], we postulate that the elevated CTH expression of CCOC reflects a cell state that promotes survival in endometriotic cysts, which is carried forward during transformation and cancer progression. The dependency of CCOC on CTH expression may present a therapeutic opportunity for this challenging cancer by targeting this transsulfuration enzyme.

Clear cell ovarian carcinoma (CCOC) accounts for 5%–11% of ovarian cancers in North America, with a higher frequency reported in cohorts from Japan [ 1 ]. Advanced stage CCOC has the poorest prognosis among ovarian carcinoma histotypes as it is inherently resistant to standard platinum/taxane chemotherapy [ 2 , 3 ]. CCOC possesses a distinct genetic profile, with 40%–50% harboring ARID1A mutations, typically resulting in a loss of protein function [ 4 , 5 ]. These mutations usually co‐occur with PIK3CA mutation or gene amplification, leading to PI3K/AKT pathway hyperactivation [ 6 ]. CCOC cells also express some proteins associated with a ciliated cell lineage [ 7 ] and share similarities with secretory phase endometrial epithelium [ 8 ] or the Arias Stella reaction, a benign, hormone‐induced endometrial change with glandular hypertrophy, vacuolization, and nuclear atypia, typically from prolonged progesterone exposure in pregnancy or hormone therapy [ 9 ]. Molecular analysis has indicated that CCOC arises directly from endometriosis [ 10 ]. Moreover, an extensive body of evidence has demonstrated the role of hypoxia and oxidative stress in the pathogenesis of endometriosis and CCOC [ 11 , 12 , 13 ]. Recently, we identified CTH, a key enzyme in the transsulfuration (TSS) pathway, as a marker of CCOC of both the ovary and uterus [ 7 , 14 , 15 ]. We previously reported that CTH was highly expressed in both CCOC and the morphologically atypical epithelial cells of adjacent or contiguous endometriosis [ 7 ]. The TSS pathway enhances cancer cell survival by maintaining redox balance through de novo cysteine synthesis for glutathione (GSH) production [ 16 ]. However, its role in CCOC adaptation to harsh microenvironments and metastasis remains largely unexplored. In this study, we used in vitro and in vivo approaches to investigate the role of CTH in CCOC, revealing novel H 2 S‐independent functions, including HIF1α regulation, beyond its role in redox balance and ferroptosis protection. We propose that targeting CTH alone or in combination with chemotherapy could offer new treatment options for CCOC and other cancers.

Studies involving primary endometrial cells, endometriotic cyst content, cyst wall, and immunohistochemical (IHC) studies on human tissue samples were approved by the UBC Research Ethics Board (UBC REB Number H18‐1652‐A027), and all patients provided written consent at their local institution. All animal studies were conducted under the UBC animal care certificate A22‐0005. The murine in vivo renal subcapsular implantation model was established as previously described [ 17 ]. In brief, OVISE cells (1 × 10 6 cells) with or without CTH KO were admixed with 6 μl of fibrinogen (20 mg/ml; bovine; Millipore Sigma, F8630‐1G, Burlington, MA, USA) and 4 μl of Thrombin (40 U/ml; bovine; Millipore Sigma, 605,157‐1KU) to establish xenograft cell blocks. These cell blocks were then engrafted in 6‐ to 8‐week‐old NRG female mice (eight mice per group). The mice were obtained from the Animal Resource Centre at BC Cancer Research Centre and were maintained according to the University of British Columbia (UBC) Animal Care Committee (ACC) regulations. Tumor growth was monitored for 10–12 weeks. For tail‐vein injection, 1 × 10 6 RMG‐I cells with or without CTH KO were injected into 12‐week‐old NRG female mice (eight mice per group), as previously described [ 18 ], and tumors were allowed to grow for 12 weeks. Histological analysis of xenografts was performed on 4‐μm‐thick sections stained with hematoxylin–eosin (H&E) (Biocare Medical, CAT Hematoxylin, Catalogue No.: CATHEM, Concord, CA, USA; Leica Biosystems, Eosin, Catalogue No.: 3801602, Concord, ON, Canada) using a Zeiss Axioplan 2 fluorescence microscope (Zeiss Microscopy, Oberkochen, Germany). For HIF1α rescue experiments, the same tail‐vein injection protocol was followed as described earlier. Detailed information on cell lines, antibodies, primers, and assay methods are provided in Supplementary materials and methods. Statistical analyses were performed using GraphPad Prism 8.4.3 (Dotmatics, Bishop's Stortford, Herts, UK). A Student's two‐tailed t ‐test was used for all comparisons, unless otherwise specified. Data are presented as mean ± SD, with p values <0.05 considered statistically significant. Image analysis and staining quantification was performed using the WCIF ImageJ software with a color deconvolution plugin feature ( https://imagej.net/ij/ ).

Supplementary materials and methods Figure S1. CTH‐mediated CCOC progression Figure S2. CTH‐mediated stress adaptation in CCOC Figure S3. Restoring redox balance partially rescues viability and motility in CTH KO cells Figure S4. CTH‐driven HIF1α expression in CCOC Figure S5. HIF1α is a potential contributor to CTH‐driven CCOC progression Figure S6. CTH‐mediated HIF1α expression in other cancers Figure S7. Full western blotting gels for main manuscript figures Figure S8. Full western blotting gels for figures in Supplementary  materials and methods