Suppression of hypoxia-induced stemness and chemoresistance in ovarian tumors

doi:10.21203/rs.3.rs-3730407/v1

Download PDF

Research Article

Suppression of hypoxia-induced stemness and chemoresistance in ovarian tumors

https://doi.org/10.21203/rs.3.rs-3730407/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 26 Feb, 2024

Read the published version in Journal of Experimental & Clinical Cancer Research →

You are reading this latest preprint version

Background: Hypoxia in solid tumors is an important source of chemoresistance that can determine poor patient prognosis. Such chemoresistance relies on the presence of cancer stem cells (CSCs), and hypoxia promotes their generation through transcriptional activation by HIF transcription factors.

Methods. We used OC cell lines, xenograft models, OC patient samples, transcriptional databases, iPSCs and ATAC-seq.

Results Here, we show that hypoxia induces CSC formation and chemoresistance in ovarian cancer through transcriptional activation of the PLD2 gene. Mechanistically, HIF-1a activates PLD2transcription through hypoxia response elements, and both hypoxia and PLD2overexpression lead to increased accessibility around stemness genes, detected by ATAC-seq, at sites bound by AP-1 transcription factors. This in turn provokes a rewiring of stemness genes, including the overexpression of SOX2, SOX9 or NOTCH1. PLD2 overexpression also leads to decreased patient survival, enhanced tumor growth and CSC formation, and increased iPSCs reprograming, confirming its role in dedifferentiation to a stem-like phenotype. Importantly, hypoxia-induced stemness is dependent on PLD2 expression, demonstrating that PLD2 is a major determinant of de-differentiation of ovarian cancer cells to stem-like cells in hypoxic conditions. Finally, we demonstrate that high PLD2expression increases chemoresistance to cisplatin and carboplatin treatments, both in vitro and in vivo, while its pharmacological inhibition restores sensitivity.

Conclusions. Altogether, our work highlights the importance of the HIF-1a-PLD2 axis for CSC generation and chemoresistance in OC and proposes an alternative treatment for patients with high PLD2 expression.

ovarian cancer

hypoxia

phospholipase D

therapy resistance

stemness

Hypoxia is a common feature of the tumor microenvironment in solid tumors (Harris, 2002). Hypoxia is generated by insufficient oxygen diffusion towards parts of the tumor that are not irrigated, although the highly irregular tumor microvasculature may also generate hypoxic regions. Cancer cells under hypoxic conditions undergo a series of transcriptional changes that are induced by hypoxia-inducible factors (HIF), HIF-1, -2 and − 3, with HIF-1 being the best known. HIFs are heterodimeric transcription factors containing bHLH-PAS domains and consist of an α subunit regulated by oxygen levels and a constitutively expressed β subunit (also called ARNT). The HIFa-ARNT dimers bind DNA at specific sequences known as hypoxia-response elements (HREs) to promote target gene expression (Schito and Semenza, 2016). In normoxia, HIF-α subunits are hydroxylated at specific proline and asparagine residues by prolyl hydroxylase 2 (PHD2) and are recognized and targeted for degradation by the von Hippel‒Lindau (VHL) tumor suppressor. However, hypoxia leads to the inhibition of such hydroxylation and subsequent HIF-α accumulation, heterodimerization with ARNT and transcriptional activation of target genes (Kaelin and Ratcliffe, 2008). The multiple effects of hypoxia on the biology of tumors include preventing apoptosis and promoting proliferation and autophagy, inducing metabolic alterations, and promoting angiogenesis, the epithelial-to-mesenchymal (EMT) transition, invasion and metastasis (Jing et al., 2019; Wilson and Hay, 2011). Indeed, HIF-1α is overexpressed in multiple tumor types and is associated with a poor prognosis in patients and therapy resistance (Chen et al., 2018a; Klemba et al., 2020; Lin and Koong, 2018; Mayer et al., 2008; Simiantonaki et al., 2008; Sun et al., 2007; Ueda et al., 2017; Unruh et al., 2003).

Ovarian cancer (OC) is the most lethal gynecological malignancy (Siegel et al., 2020) mainly due to its nonspecific clinical manifestations, which lead to a late diagnosis and high chemoresistance (Torre et al., 2018). Similar to other solid tumors, hypoxia is a key modulator of the tumor microenvironment in OC, affecting not only the primary tumor but also the ascitic fluid, which is the main route of spread of these malignancies and shows low oxygen levels (Klemba et al., 2020; Munoz-Galvan and Carnero, 2020). Therefore, ovarian tumors are highly hypoxia-dependent and this dependency influence the response to treatment. A major cause of chemoresistance is the persistence of a cancer cell subpopulation known as cancer stem cells (CSCs) or tumor-initiating cells. CSCs are able to recapitulate a new tumor since they possess self-renewal and pluripotency properties similar to those of normal stem cells (Batlle and Clevers, 2017; Beck and Blanpain, 2013). CSCs are resistant to common antitumor therapies, which may indeed cause their enrichment, leading to chemoresistance and relapse (Carnero et al., 2016; Maugeri-Sacca et al., 2011). Several studies indicate that HIF-1 is required for maintaining CSCs and that its activation in hypoxia leads to the increased expression of stem marker genes in multiple cancer types (Mathieu et al., 2011; Peng and Liu, 2015). Therefore, hypoxia may promote the generation of CSCs leading to chemoresistance through HIF factors, but our knowledge of HIF targets that may be responsible for CSC activity is limited.

We previously showed that the PLD2 gene, encoding phospholipase D2, is overexpressed in colorectal cancer and induces stemness in cancer cells through communication with the tumor microenvironment (Munoz-Galvan et al., 2019b). Here we find that PLD2 is also overexpressed in ovarian cancer patients, being associated with poor patient survival. We explore a possible connection of PLD2 with hypoxia in OC and demonstrate that PLD2 expression in OC cells is stimulated by hypoxia and that HIF-1α promotes PLD2 transcription through HREs at its promoter and an intronic enhancer. This leads to increased chromatin accessibility around stemness genes, which are in turn overexpressed provoking an enhancement of tumor growth and formation of CSC-like tumorspheres. This is corroborated by induced pluripotent stem cell (iPSC) reprogramming experiments that confirm the role of hypoxia and PLD2 in dedifferentiation of OC cells to stem-like cells. Importantly, the increase in CSC-like features induced by hypoxia relies on PLD2 expression, indicating that the hypoxia-PLD2 axis is a major contributor to tumor stemness. We confirm these findings in transcriptional databases of OC patients, where high PLD2 expression is associated with a transcriptional rewiring of genes involved in the hypoxia response and in the maintenance of stem cells. Finally, we show that PLD2 overexpression causes resistance to platinum-based compounds and propose a new therapy based on pharmacological inhibition of phospholipases D to suppress such chemoresistance.

PLD2 is overexpressed in OC patients and in ovarian cancer cells under hypoxic conditions.

First, we wondered whether PLD2 was overexpressed in OC patients. For this, we analysed PLD2 expression in 7 OC and one ovarian tissue databases using the R2 platform and found that PLD2 expression was significantly higher in tumoral than in non-tumoral samples (Fig. 1A). This result was confirmed by a comparison between patients and control individuals in 3 databases containing their own controls (GSE18520, GSE4595 and GSE3866) (Fig. 1B). Next, we wondered whether PLD2 expression was associated with patient survival and analysed 4 OC databases with available overall survival (OS) data (GSE13876, GSE19161, GSE23554 and GSE31245). We split patients in low-risk and high-risk groups based on their OS, but the difference in OS between both groups was statistically significant only in one database, GSE19161 (Fig. 1C; Supplementary Figure S1). However, the expression levels of PLD2 were significantly higher in the high-risk group than in the low-risk group in the 4 databases (Fig. 1C; Supplementary Figure S1), indicating that PLD2 is commonly overexpressed in OC patients and may be associated with decreased patient survival.

Similar to other solid tumors, OC shows hypoxic areas, and its main dissemination site, the ascitic fluid, is also characterized by hypoxia (Klemba et al., 2020). Since PLD2 expression is increased under hypoxic conditions in colon cancer cells (Liu et al., 2020) and is involved in tumor stemness through communication with the microenvironment in colorectal cancer (Munoz-Galvan et al., 2019b), we wondered whether hypoxia could increase the expression of PLD2 in OC. For this, we selected SKOV3, OVCAR8 and ES-2 OC cell lines for our studies and analysed the expression levels of PLD2 under hypoxic conditions. We found that PLD2 expression was similar among the three cell lines and that oxygen levels of 3% led to a 2-fold increase in the expression of PLD2 in all of them, while the well-known hypoxia target genes LDHA and VEGFA showed a similar increase, validating the results (Fig. 1D). The hypoxia-induced increase in PLD2 expression was also observed at the protein level by performing immunofluorescence in the three OC cell lines (Fig. 1E). In addition, we further demonstrated these findings by using the HIF-hydroxylase inhibitor DMOG, which increases HIF levels, generating a hypoxia-like phenotype under normoxic conditions (Supplementary Figure S2A-B). Thus, we treated OC cells with DMOG and found that this treatment resulted in a similar increase in hypoxia marker gene expression as that induced by hypoxia, also leading to the observed increase in PLD2 expression at both the mRNA and protein levels (Fig. 1D-E). Therefore, we can conclude that PLD2 expression is promoted by hypoxia.

HIF-1α activates PLD2 transcription through HREs at promoter and hypoxia-specific enhancer regions.

We aimed to understand how hypoxia promotes PLD2 expression. HIF-1α is considered the master transcriptional regulator of the cellular response to hypoxia. It forms a heterodimer with ARNT that binds HREs to control the expression of hypoxia-response genes (Schito and Semenza, 2016). To address the possibility that HIF-1α could regulate PLD2 expression at the transcriptional level, we first searched for possible cis-regulatory elements (CREs) near the PLD2 gene that may contain HREs. According to public 3D chromatin conformation experiments (micro-C) in human embryonic stem cells, the PLD2 gene is located within a topologically associating domain (TAD) of 190 kb with a high interaction frequency at the 3D level and is relatively isolated from the neighbouring regions (Fig. 1F). To identify CREs that may regulate PLD2 expression, we focused on a smaller region of 50 kb surrounding the PLD2 gene with a higher interaction frequency with the PLD2 promoter, which we called the PLD2 regulatory region. We scanned the CREs annotated by ENCODE within the PLD2 regulatory region for the presence of the DNA binding motif of HIF1A with a high score (> 90% relative score). We found 15 out of 35 CREs fulfilling this condition, some of which corresponded to gene promoters and others to enhancers, including the PLD2 promoter and an enhancer in PLD2 intron 12 (Fig. 1F). These CREs with high-score HIF1A motifs represent putative HREs.

To assess the regulatory activity of these two CREs in the PLD2 gene (promoter and putative enhancer) containing the HIF1A motif, we cloned both genomic regions in promoter and enhancer reporter vectors controlling the expression of the luciferase gene. We transfected OC cells with these vectors and measured the luciferase activity under normoxia and hypoxia. We found that the PLD2 promoter was able to activate reporter expression in normoxia and that hypoxia led to a significant increase in luciferase activity (8-fold) (Fig. 1G), suggesting that the PLD2 promoter responds to hypoxic conditions by increasing PLD2 transcription. However, the enhancer contained within the PLD2 intron was unable to activate reporter expression in normoxia but led to a 22-fold increase in its expression in hypoxia (Fig. 1G), indicating that this enhancer acts as an HRE in OC cells.

Next, we wanted to validate functionally that the hypoxia-induced upregulation of PLD2 expression was indeed mediated by HIF-1α. Therefore, we depleted HIF1A using a small interfering RNA (siRNA) in the three OC cell lines and found that it suppressed the increase in the PLD2 protein levels induced by hypoxia (Fig. 1H). Additionally, we generated hif1a mutant OC cell lines by transfecting cells with a mutant hif1a allele (Hu et al., 2007) that is unable to be hydroxylated and, therefore, is constitutively active even under normoxic conditions. Interestingly, we observed that the PLD2 levels in hif1a mutant cells were as high as those induced by hypoxia on a HIF1α wild-type background in both normoxia and hypoxia, confirming that HIF-1α activates PLD2 expression (Fig. 1I-J). Altogether, these data indicate that hypoxia induces PLD2 expression in OC cells through transcriptional activation by HIF-1α at HREs in the PLD2 gene.

Hypoxia alters the chromatin landscape of ovarian cancer cells in a PLD2-dependent manner.

We wondered whether PLD2 expression mediated by HIF-1α could have an impact on hypoxia-induced gene regulation. For this, we first analyzed the effect of hypoxia and PLD2 expression in the epigenomic landscape of OC cells through ATAC-seq experiments in SKOV3 cells under normoxia and hypoxia and altered PLD2 expression under normoxia (PLD2 overexpression) and hypoxia (PLD2 depletion). We computationally called open chromatin regions (ATAC peaks) in normoxia and hypoxia and compared both conditions by a differential accessibility analysis; we detected 140 and 102 peaks with increased or decreased accessibility in hypoxia, respectively (Fig. 2A). The heatmaps and aggregate profiles of these differentially accessible regions (DARs) showed that the peaks with increased accessibility in hypoxia were also more open upon PLD2 overexpression in normoxia, although to a lower extent, and vice versa, with the peaks showing decreased accessibility in hypoxia (Fig. 2B), suggesting that PLD2 overexpression in normoxia has a similar effect on chromatin accessibility as hypoxia. Indeed, the DARs of EV- versus PLD2-overexpressing cells under normoxia showed similar changes in accessibility under hypoxia, reinforcing the previous idea (Supplementary Figure S3A-B). Both DARs in normoxia versus hypoxia and control versus PLD2 overexpression were associated with genes enriched in Gene Ontology terms related to the response to hypoxia or well-known functions of PLD2, respectively (Supplementary Figure S3C-D). Moreover, changes in accessibility induced by hypoxia were suppressed by PLD2 depletion (Fig. 2B), suggesting that the effect of hypoxia in chromatin accessibility is mediated by PLD2. Similarly, changes in accessibility upon PLD2 overexpression in normoxia were absent in PLD2-depleted cells in hypoxia (Supplementary Figure S3B), reinforcing the previous idea. Altogether, these data indicate that both hypoxia and PLD2 overexpression induce similar alterations in the chromatin landscape of OC cells and that the effect of hypoxia is mediated by PLD2.

Next, we sought to investigate the possible mechanisms driving the alterations in the chromatin accessibility landscape induced by hypoxia and PLD2 overexpression. For this, we first performed motif enrichment analyses of DARs. We found that DARs in hypoxia were enriched in the motifs of the AP-1, ETS and C2H2 zinc finger transcription factor families in the increased accessibility sites and the C2H2 zinc finger and fork head families in the decreased accessibility sites (Supplementary Figure S4A). Similar enrichments were found in DARs upon PLD2 overexpression (Supplementary Figure S4B), reinforcing the idea of a similar effect under both conditions. To further elucidate the possible TFs involved in hypoxia and PLD2-mediated epigenomic changes, we estimated differential TF binding among the conditions based on footprints in ATAC-seq data. Using this approach, we confirmed the increased chromatin binding of AP-1 family transcription factors, such as FOS and JUN, consistent with their implication in the response to hypoxia (Kunz and Ibrahim, 2003). (Fig. 2C-D). We also found other TF families showing increased TF binding in hypoxia, such as homeobox, paired box or fork head TFs, and the C2H2 zinc finger TF ZBTB32, while TF families with decreased binding in hypoxia included members of the bHLH, NF-Y and ETS families, among others (Supplementary Figure S4C). When we compared these TFs with those showing increased binding upon PLD2 overexpression in normoxia, we found that most of these TFs overlapped (Supplementary Figure S4C). In particular, 25 TF motifs of the AP-1 family showed increased binding under both conditions. Interestingly, these AP-1 and 20 more TF motifs with increased binding in hypoxia showed decreased binding upon PLD2 depletion (Supplementary Figure S4C-E), indicating that they are dependent on PLD2 expression. Altogether, these results suggest a function of the AP-1 family of TFs mediating the alterations in the chromatin landscape mediated by hypoxia and PLD2.

Since we previously connected PLD2 overexpression with tumor stemness in colorectal cancer, we wondered whether alterations in the chromatin accessibility landscape of OC cells induced by hypoxia and PLD2 could result in increased expression of stemness genes. Therefore, we first selected ATAC peaks falling within the putative regulatory landscapes of the genes associated with stem cell maintenance and proliferation. Clustering of these 3,572 peaks revealed 4 groups with different accessibility levels and behaviours (Supplementary Figure S5A). Among them, Cluster 3 corresponded with peaks with increased accessibility in both hypoxic and PLD2-overexpressing cells but decreased accessibility in PLD2-depleted cells. Among the stemness genes associated with this cluster, we found SOX9, PROM1, WNT7A or JAG1 (Fig. 2E; Supplementary Figure S5B). These results indicate that hypoxia promotes chromatin accessibility around stemness genes in a PLD2-dependent manner in OC cells, and suggests that hypoxia and PLD2 could be connected with tumor stemness in OC.

High PLD2 expression in OC patients leads to the transcriptomic rewiring of stemness and hypoxia genes

To determine whether PLD2 expression in OC patients is related to hypoxia and stemness, we analysed the expression of genes related to these functions in the three OC databases with available expression data from control individuals. First, we selected the genes annotated to the Gene Ontology (GO) term “Response to Hypoxia” and whose expression was significantly correlated with that of PLD2 in OC patients (p < 0.05; r > 0.2 or <-0.2). Then, we performed hierarchical clustering of patient and control individuals based on the expression levels of these genes (Fig. 3A; Supplementary Figure S6). In the GSE18520 database, the clustering clearly separated the control individuals (‘non-tumoral’, NT) and a reduced group of patients that we termed ‘Tumoral Cluster 1’ (T1) from most patients who clustered in what we termed ‘Tumoral Cluster 2’ (T2) (Fig. 3A). The patients at T1 showed a transcriptional profile of hypoxia-related genes more similar to NT and clearly different from T2. In contrast, the clustering in the GSE4095 and GSE38666 databases clearly separated clusters of NT and tumoral (T) individuals, who showed distinct transcriptional profiles of hypoxia-related genes (Supplementary Figure S6). These results suggest that there is a switch in the expression of hypoxia-related genes in OC tumors compared with healthy ovaries.

Next, we repeated the hierarchical clustering with the genes annotated to the GO term “Stem cell maintenance” and whose expression was significantly correlated with that of PLD2 in OC patients (p < 0.05; r > 0.2 or <-0.2). Surprisingly, this clustering based on stem-related genes separated the patients and control individuals into the same clusters as the hypoxia-related genes (Fig. 3A; Supplementary Figure S6), suggesting a connection between both groups of genes that supports the model of CSC generation induced by hypoxia. Then, we plotted the expression levels of PLD2 in the 3 clusters obtained from the GSE18520 database and found that PLD2 exhibited significantly increased expression in Cluster T2 compared with that in Cluster NT, while Cluster T1 showed similar levels to NT (Fig. 3B). This observation suggests that a connection exists among PLD2 expression, the hypoxia response and stemness since PLD2 expression is misregulated only in patients showing transcriptional profiles highly different from healthy controls. In addition, we checked the expression of stem-associated and hypoxia-related genes in these clusters. As shown in Fig. 3B, the expression of PAX8, a well-known OC marker, was significantly increased in both patient Clusters T1 and T2, similar to other stemness genes, such as SOX9, SOX17, EPCAM, PROM1, CD24, NOTCH1 and WNT7A. Other genes in this group showed a significant

increase in expression only in Cluster T2, coinciding with higher PLD2 levels, including PAX2, POU5F1 (OCT4), SOX5, SOX11, CD34 and TP63, while the others were unaffected or even significantly reduced, such as KLF4, NANOG or SOX2, although the latter showed a nonsignificant increase in Cluster T2 (Fig. 3B). This finding suggests that there is an OC stemness signature in patients that may be stronger with higher PLD2 expression. In addition, some hypoxia-related genes showed a significant increase in expression in both Clusters T1 and T2, including VEGFA, SLC2A1 (GLUT1), HK2 and NOX4, or only in T2, including SLC2A4, NOS1 and MMP14, and a nonsignificant increase in SLC2A14 (GLUT14) (Fig. 3B). Similar results were obtained in the NT and T clusters in the GSE4095 and GSE38666 databases (Supplementary Figure S6). Altogether, these results suggest that there is transcriptional rewiring of the expression of hypoxia- and stem-related genes in OC patients with PLD2 overexpression.

PLD2 promotes tumorigenesis and CSC-like features in ovarian cancer cells.

The ability of hypoxia to induce a CSC-like phenotype in OC cells has been previously observed in several cancer types (Liang et al., 2012; Mathieu et al., 2011; Peng and Liu, 2015). We first aimed to validate these results in our OC cell lines SKOV3, OVCAR8 and ES-2 and found that hypoxia led to significant increases in the number of tumorspheres, which were generated by growing the cells under low-attachment conditions, and in the percentage of holoclones, both of which were used as a proxy for CSCs (Supplementary Figure S7A-B). Next, we analysed the expression of stem cell markers in OC cell lines grown under hypoxic conditions and detected an increase in the mRNA levels of NANOG, CD44, SOX2 and EPCAM and the percentage of cells containing the surface CSC marker CD133 (Supplementary Figure S7C-D). These results confirm that hypoxia induces a CSC-like phenotype in OC cells.

Next, we wondered whether increased PLD2 expression led to an increase in the CSC population in OC cells, as suggested by the chromatin accessibility and gene expression data. Therefore, we first established OC cell lines expressing ectopic PLD2 cDNA or depleted of PLD2 using a short hairpin RNA (shRNA). The expression of PLD2 under these conditions was assessed at the mRNA and protein levels (Fig. 4A-B and Supplementary Figure S8). We observed that the overexpression of PLD2 in OC cells led to a significant increase in the number of clones generated by the three cell lines, while a significant decrease was detected in the OVCAR8 cells upon PLD2 depletion (Fig. 4C), suggesting that PLD2 promotes tumor growth. To address this question, we analysed the growth of these cell lines and found that the enhanced PLD2 expression led to a significant increase in proliferation, while the PLD2 depletion generated the opposite effect with statistical significance in all cell lines (Fig. 4D). This effect was further confirmed in vivo by generating xenograft models of OC cells overexpressing or depleted of PLD2, showing an increase or decrease in the tumor volume, respectively, 50 days after transplantation (Fig. 4E). Altogether, these results indicate that PLD2 expression promotes tumorigenesis.

Next, we wondered whether PLD2 expression was related to the formation of ovarian CSCs. First, we analysed the formation of different types of colonies, including holoclones, meroclones and paraclones, which are considered stem cells, transit-amplifying cells and differentiated cells, respectively (Barrandon and Green, 1987). We found a significant increase in the percentage of holoclones and a significant decrease in the percentage of paraclones in the three cell lines overexpressing PLD2 (Fig. 4F). A significant decrease in holoclone formation was also observed in the three OC cell lines upon PLD2 depletion. Furthermore, we measured the formation of tumorspheres under low attachment conditions in OC cells overexpressing or depleted of PLD2. We found that PLD2 overexpression led to a significant increase in the number of tumorspheres, while PLD2 depletion generated the opposite effect (Fig. 4G), although we did not observe changes in the size of such tumorspheres. These data indicate that PLD2 expression promotes the formation of ovarian CSCs.

Finally, we analysed the expression of pluripotency and CSC marker genes in our OC cell lines overexpressing or depleted of PLD2. We found that the expression of

SOX2, CD44 and EPCAM was significantly increased in the ES-2 and SKOV3 cells overexpressing PLD2, while NANOG was only increased in ES-2 cells (Fig. 4H). In OVCAR8 cells, the expression of these genes was increased in the same trend, although in a nonsignificant manner. Then, we measured the expression levels of these pluripotency genes in the tumorspheres extracts. First, we found that PLD2 was highly expressed in the tumorspheres compared with that in the total extracts transfected with only the empty vector (Fig. 4H), confirming that CSCs indeed have higher expression levels of PLD2. The expression of stemness genes was also increased in the tumorspheres compared with that in the total extracts, as expected, while they were further upregulated in most cases upon PLD2 overexpression, or downregulated upon PLD2 depletion (Fig. 4H). Altogether, these data indicate that PLD2, whose expression is induced by hypoxia, is an important oncogene in OC and that its overexpression leads to increased tumor stemness.

The hypoxia-induced stemness of ovarian cancer cells partially depends on PLD2 expression.

Thus far, we have showed that both hypoxia and PLD2 expression led to an increase in CSCs in OC cells (Fig. 4, Supplementary Figure S7) and that PLD2 expression was increased under hypoxic conditions in a HIF-1α-dependent manner (Fig. 1). Thus, we wondered whether both phenomena were connected and whether the hypoxia-induced increase in CSC-like features was dependent on PLD2 expression. To address this, we first analysed the expression levels of stemness genes by RT‒qPCR using custom TaqMan Array plates containing probes against a selection of these genes in OC cells. We observed that either hypoxia or PLD2 overexpression in normoxia led to the increased expression of many stemness genes, while PLD2 depletion largely suppressed this increase (Fig. 5A). Indeed, hierarchical clustering of the four analysed conditions showed that the samples corresponding to EV hypoxic cells and PLD2-overexpressing cells clustered together, while EV normoxic cells and PLD2-depleted cells clustered separately (Fig. 5B). We confirmed these result by RT‒qPCR of individual representative genes, including SOX2, NANOG, CD44 and EPCAM, showing that either hypoxia or PLD2 overexpression in normoxia lead to increased expression of stemness genes, while combination of both conditions further increased their expression (Fig. 5C). PLD2 depletion partially suppressed the hypoxia-induced enhancement of expression, with a lower non-statistically significant effect in normoxia, and rescue experiments confirmed the specificity of PLD2 depletion (Fig. 5C).

Next, we analyzed the formation of tumorspheres in normoxia and hypoxia with altered PLD2 expression. We observed that either PLD2 overexpression or hypoxia led to a similar increase in the formation of tumorspheres, with only a slightly higher nonsignificant increase when both conditions were combined (Fig. 5D). However, PLD2 depletion caused a partial suppression of the increase in tumorsphere formation under hypoxic conditions in SKOV3 and OVCAR8 cells, which was rescued by overexpressing back PLD2 in shRNA-transfected cells (Fig. 5D and Supplementary Figure S9A), suggesting that PLD2 is partially responsible for the hypoxia-induced stemness. Then, we measured the formation of holoclones, meroclones and paraclones under hypoxic conditions upon PLD2 overexpression or depletion and found that the increase in the percentage of holoclones induced by hypoxia was further enhanced by the PLD2 overexpression, while it was suppressed by PLD2 depletion and rescued back by expressing PLD2 after its depletion (Supplementary Figure S9B).

Then, we measured the protein levels of the pluripotency factors Sox2, Sox17, Sox9 and Notch1 (found to correlate with PLD2 in OC patients, Fig. 3B) by immunofluorescence in tumorspheres to determine whether PLD2 could influence their expression in CSCs. We validated PLD2 protein levels in tumorspheres (Supplementary Figure S9C) and found that either hypoxia or PLD2 overexpression led to an increase in the levels of Sox2, Sox9 and Notch1, while only hypoxia led to an increase in Sox17 protein levels (Fig. 5E). In addition, PLD2 depletion led to a partial suppression of the hypoxia-induced expression of Sox2, Sox9 and Notch1 that was rescued by expressing back PLD2 in these cells (Fig. 5E), suggesting that PLD2 plays a role in the generation of CSCs in hypoxia through these genes. These observations were confirmed at the mRNA level by RT‒qPCR (Fig. 5F). Altogether, these data indicate that PLD2 plays a major role in the induction of the CSC phenotype in hypoxia, promoting the expression of specific stem-related genes, such as SOX2, SOX9 or NOTCH1.

Finally, we extended the gene expression analyses to EMT genes using TaqMan Arrays to assess whether PLD2 expression may have a role in tumor invasion and metastasis. We found that either hypoxia or PLD2 overexpression in normoxia led to an increase in the expression of many of these genes, but PLD2 depletion was unable to suppress such increase (Supplementary Figure S9A). Indeed, hierarchical clustering of the samples did not exhibit the pattern observed in stemness genes (Supplementary Figure S9B), and results were further validated by RT‒qPCR of particular EMT genes (Supplementary Figure S9C). Consistently, invasiveness assays using Boyden’s chamber showed that both PLD2 overexpression and hypoxia were able to increase invasion, but PLD2 depletion did not have any effect (Supplementary Figure S9D). These results indicate that while the increased expression of stemness genes induced by hypoxia relies on PLD2 overexpression, this is not the case for EMT genes and suggests that PLD2 is a specific mediator of the increase in CSCs induced by hypoxia in OC cells.

Hypoxia-mediated reprogramming to induced pluripotent stem cells is dependent on PLD2.

We aimed to obtain additional evidence of the contribution of PLD2 to dedifferentiation or reprogramming events mediated by hypoxia that may generate ovarian CSCs from normal OC cells. Therefore, we performed reprogramming experiments of mouse embryonic fibroblasts (MEFs) to induced pluripotent stem cells (iPSCs) in normoxia and hypoxia and upon alteration of PLD2 expression levels (overexpression or depletion). We used a previously published protocol (Yoshida et al., 2009) in which MEFs were infected using a HEK293T cell-derived virus that provides OSKM genes and Nanog reporter retroviruses and then cocultured on SNL feeder cells that produce LIF. Then, the samples were incubated with or without hypoxia for 7 days, and the efficiency of iPSC generation was measured for additional 5 days. Cell reprogramming and the acquisition of pluripotency were assessed by colony morphology, alkaline phosphatase and nanog promoter-driven GFP expression analyses (Fig. 6A) to assess the effect of PLD2 and hypoxia on the efficiency of the reprogramming process and the acquisition of stem cell-like properties. Using this protocol, we found that, as expected, hypoxia led to a significant increase in the generation of iPSCs (Fig. 6B). Furthermore, we found that PLD2 overexpression in normoxia provoked a similar increase in iPSC generation, consistent with its effect on the generation of CSCs, and that the combination of both hypoxia and PLD2 overexpression further increased iPSC formation (Fig. 6B). This confirms that high PLD2 expression leads to dedifferentiation processes. Finally, PLD2 depletion in hypoxia suppressed the increased iPSC production induced by hypoxia, and this was recovered by expressing back PLD2 in PLD2-depleted cells (Fig. 6B), suggesting that PLD2 is an important mediator in the activation of pluripotency by hypoxic conditions.

Overexpression of PLD2 leads to chemotherapy resistance in ovarian tumors.

Since we showed that PLD2 overexpression leads to an increase in CSC-like cells in OC and CSCs were previously proposed to be responsible for chemotherapy resistance and tumor relapse, we wondered whether PLD2 overexpression could cause resistance to conventional therapy in ovarian tumors. Therefore, we first analysed the expression levels of PLD2 in our own cohort of OC patients. The immunohistochemistry analyses showed that PLD2 protein levels were higher in tumors than in healthy tissue (Fig. 7A), and RT‒qPCR revealed that PLD2 mRNA was significantly more abundant in OC patients than in control non-tumoral samples (Fig. 7B), thus confirming the results observed in the transcriptomic databases (Fig. 1A-B). Then, we separated our patient samples into those who were sensitive or resistant to platinum-based chemotherapy (without or with tumor relapse within the next 6 months after chemotherapy, respectively) and analysed PLD2 expression levels. Importantly, we found that the resistant patients showed significantly higher expression of PLD2 than the sensitive patients (Fig. 7C), suggesting that PLD2 overexpression may contribute to resistance to platinum-based therapy. Resistant patients in our cohort showed reduced OS and PFS (Fig. 7D-E), consistent with the reduced survival of patients with high PLD2 expression (Fig. 7C).

Next, we analysed the effect of PLD2 expression on resistance to platinum compounds in OC cells in vitro. First, cells overexpressing or depleted of PLD2 were treated with increasing concentrations of cisplatin and carboplatin, and the IC₅₀ values were calculated in each case. We found that PLD2 overexpression led to a significant increase in the IC₅₀ values, while PLD2 depletion led to only a weak nonsignificant reduction (Fig. 7F). This finding suggests that higher PLD2 expression causes resistance to platinum-based compounds. We repeated these experiments under hypoxic conditions and found that hypoxia also led to increased IC₅₀ values (Fig. 7F). However, PLD2 depletion reduced the hypoxia-induced increase in IC₅₀ values, suggesting that enhanced resistance to cisplatin and carboplatin in OC cells under hypoxic conditions relies on PLD2 expression.

Finally, we performed in vivo analyses to validate our findings by establishing xenograft models from SKOV3 and OVCAR8 cells overexpressing or depleted of PLD2 and analysing tumor growth upon treatment with cisplatin. The control tumors from cells transfected with the empty vector were sensitive to the cisplatin treatment, significantly reducing tumor growth in xenografts from both SKOV3 and OVCAR8 cells (Fig. 7H). However, tumors overexpressing PLD2 showed higher tumor growth that was not reduced upon cisplatin treatment, indicating both a higher aggressiveness of PLD2-overexpressing tumors and the resistance of these tumors to cisplatin. However, PLD2 depletion resulted in significantly reduced tumor growth that was further reduced upon cisplatin treatment (Fig. 7H). Importantly, cisplatin treatment in control tumors led to increased survival, while PLD2-overexpressing tumors did not exhibit improved survival, and PLD2-depleted tumors exhibited increased survival independent of cisplatin treatment (Fig. 7I). These results indicate that the overexpression of PLD2 causes resistance to platinum-based chemotherapy in OC tumors.

Combination treatment with cisplatin and a PLD inhibitor suppresses chemotherapy resistance in ovarian cancer.

Finally, we wondered whether the increased therapy resistance to platinum-based compounds induced by PLD2 overexpression and hypoxia could be suppressed by the pharmacological inhibition of PLD2. For this, we used the PLD inhibitor (PLDi) 5-Fluoro-2-indolyl des-chlorohalopemide (FIPI), which inhibits the catalytic activity of phospholipases D (Ganesan et al., 2015). First, we tested this possibility in vitro by calculating the IC₅₀ in OC cells treated with cisplatin, PLDi and their combination in normoxia and hypoxia with altered levels of PLD2. We found that the higher IC₅₀ to cisplatin in OC cells overexpressing PLD2 or in hypoxia was suppressed by the PLDi (Fig. 7G). Next, we validated these results in vivo by establishing xenografts of OC cells expressing EV or PLD2 and treating mice with cisplatin, PLDi or their combination. We found that the increased tumor growth provoked by PLD2 overexpression was reduced upon treatment with PLDi (Fig. 7H). Moreover, although treatment with cisplatin did not reduce the higher tumor growth induced by PLD2 overexpression, its combination with PLDi led to a significant reduction in tumor growth that was stronger than that following treatment with PLDi alone (Fig. 7H). This finding was confirmed in xenografts from two OC cell lines (OVCAR8 and SKOV3) and led to an increase in survival (Fig. 7I). Altogether, these results indicate that chemotherapy resistance to cisplatin caused by PLD2 overexpression can be overcome by the pharmacological inhibition of PLD2, suggesting that combined treatment with cisplatin and PLDi is a promising alternative treatment for patients with high PLD2 expression levels.

We show here that the HIF-1α-PLD2 axis is a major player in the chemoresistance of OC by promoting the generation of CSCs under hypoxic conditions. On the one hand, PLD2 expression is regulated by HIF-1α through HREs in its promoter and a hypoxia-specific enhancer; on the other hand, PLD2 expression is required for the full transcriptional and epigenomic rewiring promoted by hypoxia as evidenced using gene expression and chromatin accessibility analyses. In particular, the expression of stem-related genes, the opening of enhancers in the vicinity of these genes, the generation of CSCs induced by hypoxia, and the reprogramming of normal cells to iPSCs rely on normal PLD2 expression levels. These findings indicate that PLD2 is a major player in the response to hypoxia in cancer cells that leads to increased stem cell properties resulting in higher chemoresistance.

Hypoxia is a feature of the tumor microenvironment in regions with low oxygen supply that is known to increase the stemness features of cancer cells in several types of cancer (Carnero and Lleonart, 2016; Li et al., 2009b; Mathieu et al., 2011; Peng and Liu, 2015; Schwab et al., 2012; Wang et al., 2011), including OC in which hypoxia has been shown to increase the stem-like properties of cancer cells (Liang et al., 2012). HIF-ARNT can regulate the expression of many genes that promote the hypoxic response (Kaelin and Ratcliffe, 2008; Schito and Semenza, 2016). Therefore, HIF factors may exert their effect of increasing CSCs via multiple mechanisms. In OC, HIF factors contribute to the upregulation of pluripotency factor genes, such as SOX2 or OCT3/4 (Liang et al., 2012; Seo et al., 2016), proliferation pathways, such as Notch or Wnt (Chau et al., 2013; Seo et al., 2016), or epigenetic modulation by affecting chromatin modifiers, such as SIRT1 (Qin et al., 2017). Here we show that the OC cell lines ES-2, SKOV3 and OVCAR8 under hypoxia show upregulated expression of PLD2, which encodes phospholipase D2, in these cells in a Hif-1α-dependent manner (Fig. 1), as recently reported in colon cancer (Liu et al., 2020). Using DNA binding motif analyses and reporter assays, we also demonstrate that PLD2 expression is regulated at the transcriptional level by HREs located within the PLD2 promoter and a hypoxia-specific enhancer that activates PLD2 transcription in hypoxia, thus providing a mechanistic explanation of Hif-1α-mediated PLD2 overexpression. Therefore, it is likely that the HIf-1α-PLD2 axis works in other solid tumors under hypoxic conditions.

The expression of PLD2 is elevated in several cancer types (Carnero et al., 1994; Frankel et al., 1999; Henkels et al., 2013; Munoz-Galvan et al., 2019b; Saito et al., 2007; Song et al., 1991; Zheng et al., 2006), and here, using public patient databases and our cohort of patients, we show that this is also the case in OC in which it may be related to decreased OS (Fig. 1). Furthermore, consistent with the results in OC cell lines, clustering analyses of OC patient gene expression revealed that PLD2 expression is correlated with the rewiring of transcriptomic programs of the response to hypoxia and stem cell maintenance (Fig. 3). Indeed, we found two different clusters of patients, one of which showed gene expression patterns more different from healthy controls coinciding with higher PLD2 expression. Although the expression of the OC marker PAX8 and other stemness- and hypoxia-related genes was enhanced in both clusters, other markers showed increased expression only in the patient cluster with high PLD2 expression, suggesting that PLD2 may promote stemness through specific genes or pathways as we show in SOX2, SOX9 and NOTCH1, but not SOX17. These data indicate that a correlation exists between PLD2 expression and highly altered transcriptomic programs of the response to hypoxia and stemness. In addition, we provide evidence that PLD2 is important for the alteration in the epigenomic landscape provoked by hypoxia since both hypoxia and PLD2 overexpression in normoxia lead to similar alterations in chromatin accessibility that are counteracted by PLD2 depletion, including the opening of enhancers in the proximity of genes related to the stem fate, which we show were upregulated (Fig. 2). These changes connect hypoxia with stemness gene expression and likely occur through the activation of CREs bound by the AP-1 family of TFs. In a previous study analysing OC tumors, solid metastasis and effusions, higher PLD2 expression was described in effusions rather than solid tumors and metastasis (Harel-Dassa et al., 2017). This finding is consistent with our findings since OC effusions, most of which are peritoneal, are characterized by low oxygen levels and a high content in CSCs (Munoz-Galvan and Carnero, 2020). Therefore, the hypoxia-PLD2 axis seems to play a major role in the generation of ovarian CSCs.

Importantly, PLD2 depletion partially suppresses the effect induced by hypoxia (Fig. 5), suggesting that PLD2 is an important mediator of hypoxia-induced stemness. This finding was also corroborated by reprogramming experiments of MEFs to iPSCs in which PLD2 was required for the increased reprogramming induced by hypoxia. Indeed, PLD2 might also potentiate the response to hypoxia by a positive feedback loop. This hypothesis can be supported since the combination of hypoxia and PLD2 overexpression additively enhanced the stem properties of OC cells. The activation of HIF-1α expression or activity by PLD2 has been reported in endothelial, glioma and renal cancer cells (Ghim et al., 2014; Han et al., 2014; Toschi et al., 2008), but there is evidence of the opposite effect in HEK293 cells (Park et al., 2015), suggesting that this feedback loop may work in specific contexts or cell types. Nevertheless, whether the effect of PLD2 on stemness is an autocrine or a paracrine effect remains to be elucidated. In this regard, we previously showed that exosomes from PLD2-overexpressing colorectal cancer cells induced senescence in stromal fibroblasts (Munoz-Galvan et al., 2019b), which, in turn, induced WNT pathway activation and increased stemness in tumor cells. In OC, hypoxia-induced exosomes have been involved in increased tumorigenic properties and chemoresistance by several mechanisms. These include exosome-containing oncogenic proteins, such as STAT3 and FAS (Dorayappan et al., 2018), microRNAs that altered tumor-associated macrophages (Chen et al., 2018b; Zhu et al., 2019), as well as plasma gelsolin, which induces the conversion of chemosensitive OC cells to chemoresistant cells (Asare-Werehene et al., 2020). Finally, a role of PLD1 and PLD2 inducing exosome secretion in OC cells has also been recently proposed (Onallah et al., 2022), suggesting that PLD2 may also influence the tumor microenvironment in OC.

Several mechanisms have been described to promote chemotherapy resistance under hypoxia in OC, including the upregulation of the ABCG2 transporter gene, which increases drug efflux (He et al., 2019; Wang et al., 2016), c-KIT overexpression (Chau et al., 2013) and high cysteine levels (Nunes et al., 2018). CSCs are responsible for chemotherapy resistance (Batlle and Clevers, 2017; Beck and Blanpain, 2013), and ovarian CSCs were identified sixteen years ago and reported to be chemoresistant (Bapat et al., 2005; Hu et al., 2010). In agreement with this idea, we previously found several markers linking ovarian CSCs and chemoresistance (Munoz-Galvan et al., 2019a; Munoz-Galvan et al., 2020). Here, using OC cells and xenograft models, we show that the overexpression of PLD2 leads to resistance to platinum-derived compounds, including cisplatin and carboplatin (Fig. 7). Moreover, PLD2 expression is higher in patients resistant to platinum-based chemotherapy than in sensitive patients, confirming our results in cell lines and mouse models. How PLD2 provokes such resistance is an intriguing issue, although it is likely that its enzymatic product PA, an important molecule acting as a second messenger in multiple cellular functions (Jang et al., 2012), might play some relevant role in avoiding chemotherapy-induced cell damage. Indeed, we previously showed that PA administration has similar effects as PLD2 overexpression (Munoz-Galvan et al., 2019b). This is particularly relevant in ovarian tumors, which are typically treated with platinum-based compounds and show high rates of chemoresistance, with frequent metastasis in hypoxic environments, such as abdominal ascites.

Our findings suggest a model in which hypoxia leads to the transcriptional overexpression of PLD2 in OC, which, in turn, generates PA and induces the generation of chemoresistant ovarian CSCs. Therefore, we propose an alternative treatment based on a combination of cisplatin and the pharmacological inhibition of PLD2. Our in vitro and in vivo results demonstrate that this combined treatment may be useful for patients with high PLD2 expression, who are resistant to conventional therapy with cisplatin alone.

Cell culture. Cells were cultured according to the manufacturer's recommended procedures in McCoy (ES-2 line) or RPMI (SKOV3 and OVCAR8 lines) and incubated at 37°C in 5% CO₂ in a humidified atmosphere.

Gene transfer. The gene transfer was performed as previously described (Munoz-Galvan et al., 2020). The PLD2 overexpression plasmid was described in (Munoz-Galvan et al., 2019b). The shRNA against PLD2 was provided by Origene.

Proliferation assay. The proliferation assay was performed as previously described (Munoz-Galvan et al., 2020).

Cytotoxic assay. ES-2, SKOV3 or OVCAR8 cells were seeded and then treated with platinum drugs and/or the PLD inhibitor 5-Fluoro-2-indolyl des-chlorohalopemide (FIPI) at 300nM concentration 24 hours later. After 96 hours, cells were stained with 0.5% crystal violet. Then, the crystal violet was solubilized in 20% acetic acid and quantified at 595 nm absorbance to measure the cell viability.

Maintenance of mouse colonies. All experiments involving animals received expressed approval from the IBIS/HUVR Ethical Committee for the Care and Health of Animals. The mice were maintained in the IBIS animal facility according to the facility guidelines, which are based on the Real Decreto 53/2013 and were sacrificed by CO₂ inhalation either using a planned procedure or as a human endpoint when the animals showed significant signs of illness.

Colony formation assay and clonal heterogeneity analysis This analysis was performed as previously described in (Munoz-Galvan et al., 2020). Briefly, in total, 10³ cells were seeded onto 10 cm plates, and every condition was evaluated in triplicate. The medium was replaced every 3 days for 12 days, and the colonies were fixed, stained and counted. The values are expressed as the number of observed colonies among the 10³ seeded cells. To analyze the clonal heterogeneity, 10² random colonies were classified in triplicate as having the following phenotypes: holoclone, meroclone and paraclone (Li et al., 2008), which are considered stem cells, transit-amplifying cells and differentiated cells, respectively (Barrandon and Green, 1987).

Sphere-forming assay. In total, 2x 10³ cells were resuspended in 1 ml of complete MammoCultTM Basal Medium (Stemcell Tech) and seeded in ultralow attachment plates. The cultures were imaged, the tumorspheres were counted, and their diameters were quantified using CellSenseDimension software on Days 2, 3 and 4.

Western blot analyses and immunofluorescence were performed according to standard procedures. Information of antibodies and dilutions is shown in Supplementary Table S1.

RT–qPCR. The total RNA was isolated using an RNeasy kit (Qiagen), and cDNA was generated from 1 µg of RNA with MultiScribe Reverse Transcriptase (Applied Biosystems). qPCR was performed using a TaqMan Assay (Applied Biosystems) with probes. The relative mRNA expression was calculated as 2^−∆Ct relative to the ACTB gene. Information of probes is shown in Supplementary Table S1.

Fluorescence-activated cell sorting For FACS staining, live cells were incubated with antibodies for 30 minutes at dilutions specified in the manufacturer's protocols.

iPSCs protocol

Briefly, mouse embryonic fibroblasts (MEFs) were infected with 4 different retroviral vectors encoding the Yamanaka factors (pMXs-Oct3/4, pMXs-Sox2, pMXs-Klf4, and pMXs-cMyc), an additional lentiviral vector that expresses GFP in cells where the Nanog promoter/enhancer is active (mNanog-pGreenZeo), and the corresponding plasmid overexpressing PLD2, carrying shPLD2 or carrying Ev. Seventy-two hours after the infection, MEFs were seeded on top of an SNL feeder layer in the presence of ES media, and the media was renewed every 24 h. Cell reprogramming and the acquisition of pluripotency were assessed by colony morphology, GFP expression and alkaline phosphatase activity assays to assess the effect of the gene of interest/condition of interest* on the efficiency of the reprogramming process and the acquisition of stem cell-like properties.

ATAC-seq. ATAC-seq assays were performed using standard protocols (Buenrostro et al., 2013; Fernandez-Minan et al., 2016), with minor modifications. Briefly, 70,000 ovarian cancer cells overexpressing PLD2 carrying shPLD2 or carrying Ev growing under normoxic or hypoxic conditions were collected by centrifugation for 5 min at 500 g 4°C. The supernatant was removed, and the cells were washed with PBS. Then, the cells were lysed in 50 µl of lysis buffer (10 mM Tris-HCl pH 7.4, 10 mM NaCl, 3 mM MgCl₂, 0.1% NP-40, 1x Roche Complete protease inhibitors cocktail) by pipetting up and down. The whole cell lysate was used for TAGmentation, which was centrifuged for 10 min at 500 g 4°C, resuspended in 50 µl of the Transposition Reaction containing 2.5 µl of Tn5 enzyme and TAGmentation Buffer (10 mM Tris-HCl pH 8.0, 5 mM MgCl2, 10% w/v dimethylformamide), and incubated for 30 min at 37°C. Immediately after TAGmentation, DNA was purified using a Minelute PCR Purification Kit (Qiagen) and eluted in 20 µl. Libraries were generated by PCR amplification using NEBNext High-Fidelity 2X PCR Master Mix (NEB). The resulting libraries were multiplexed and sequenced in a HiSeq 4000 paired-end lane, producing 100 M 49-bp paired-end reads per sample.

Quantification and statistical analysis. All statistical analyses were performed using GraphPad Prism 4. The distribution of the quantitative variables among different study groups was assessed using parametric (Student’s t test) or nonparametric (Kruskal–Wallis or Mann–Whitney) tests as appropriate. The experiments were performed a minimum of three times and were performed in independent triplicates each time. The survival data from the patient databases were analyzed by a log-rank Mantel‒Cox statistical test.

Analyses of cancer patient databases. We performed meta-analyses using the R2 Genomics analysis and visualization platform (http://hgserver1.amc.nl) to analyze the PLD2 expression levels in tumor and nontumor ovarian samples from the databases. The statistical significance of the tumor versus normal samples was assessed (P < 0.05). Patient survival was analyzed using an R2 Genomics analysis and visualization platform (http://hgserver1.amc.nl), which was developed by the Department of Oncogenomics of the Academic Medical Center (Amsterdam, Netherlands). Kaplan‒Meier plots showing patient survival were generated using the databases with available survival data with the scan method, which searches for the optimum survival cut-off based on statistical analyses (log-rank test), thereby identifying the most significant expression cut-off.

ATAC-seq data analyses. ATAC-seq reads were aligned to the GRCh38 (hg38) human genome assembly using Bowtie2 2.3.5 (Langmead and Salzberg, 2012), and pairs separated by more than 2 kb were removed. For ATAC-seq, the Tn5 cutting site was determined as position − 4 (minus strand) or + 5 (plus strand) from each read start, and this position was extended 5 bp in both directions. Reads below 150 bp were considered nucleosome free. The conversion of the SAM alignment files to BAM was performed using SAMtools 1.9 (Li et al., 2009a). The conversion of BAM to BED files and peak analyses, such as overlaps or merges, were carried out using the Bedtools 2.29.2 suite (Quinlan and Hall, 2010)⁠. The conversion of BED to BigWig files was performed using the genomecov tool from Bedtools and the wigToBigWig utility from UCSC (Haeussler et al., 2019). For ATAC-seq, peaks were called using the MACS2 2.1.1.20160309 algorithm (Zhang et al., 2008) with an FDR < 0.05 for each replicate and merged into a single pool of peaks that was used to calculate the differentially accessible sites with the DESeq2 1.18.1 package in R 3.4.3 (Love et al., 2014); a value < 0.05 was set as the cut-off for statistical significance of the differential accessibility. For visualization purposes, reads were extended 100 bp for ATAC-seq. For the data comparison, all ATAC-seq experiments used were normalized using reads falling into peaks to counteract differences in background levels between experiments and replicates (Santos-Pereira et al., 2019).

Heatmaps, average profiles and k-means clustering of the ATAC-seq data were generated using the computeMatrix and plotHeatmap tools from the Deeptools 3.5 toolkit (Ramirez et al., 2016). The TF motif enrichment was calculated using XSTREME (Grant and Bailey, 2021) with the standard parameters. For the gene assignment to ATAC peaks, we used the GREAT 3.0.0 tool (Hiller et al., 2013)⁠, with the basal plus extension association rule and the standard parameters (5 kb upstream, 1 kb downstream, and 1 Mb maximum extension). For the footprinting analyses, we used TOBIAS 0.12.9 (Bentsen et al., 2020). First, we performed bias correction using ATACorrect and calculated the footprint scores with ScoreBigwig, both with the standard parameters. Then, we used BINDetect to determine the differential TF binding of all vertebrate motifs in the JASPAR database (Fornes et al., 2020). We considered motifs with a linear fold-change ≥ 15% between conditions differentially bound. Aggregated ATAC-seq signals in the footprints were visualized using PlotAggregate. ATAC-seq data generated in this study is available through the Gene Expression Omnibus (GEO) accession number GSE210599.

In vivo xenograft studies. Tumor growth was assayed following the subcutaneous injection of 4x10⁶ SKOV3 or OVCAR8 cells that were transfected with a plasmid carrying PLD2 or shRNA against PLD2 in cohorts of five nude mice each that were analyzed weekly. The tumors were measured using callipers. All mice were sacrificed once the growth experiment was completed.

In vivo xenograft treatment. Tumors were harvested when they reached 1500 mm³, cut into 2×2×2 mm pieces and reimplanted. Mice were randomly allocated to the drug-treated and control-treated (solvent only) groups, and once the tumor reached 20 mm³, the mice received the appropriate treatment for 4 weeks (2 doses/week). The mice were monitored daily for signs of distress and weighed twice a week. The tumor size was measured, and the size was estimated according to the following equation: tumor volume = [length x width²]/2. The experiments were terminated when the tumor reached 350 mm³ or when the clinical endpoint was reached. The drugs cisplatin and carboplatin were obtained from Pharmacy HUVR and were freshly prepared and administered by intraperitoneal injection. We used higher doses in mice, assuming a 70 kg average weight for humans (125 mg/dose in humans) (Munoz-Galvan et al., 2020). We administered two doses per week as follows: 3.5 mg/kg of cisplatin (equivalent to 7 mg/kg, averaging 25 g body weights of each mouse) with or without 3 mg/kg of FIPI. We did not observe signs of toxicity.

In vivo xenografts from tumorspheres. This assay involved the subcutaneous injection of 1×10³ cells grown as tumorspheres into the hind legs of 4-week-old female athymic nude mice. The animals were treated as previously described, examined twice a week, incubated for 4 more weeks, and killed, and the tumors were extracted. The tumors were measured using callipers.

Patient cohort. The entire procedure was approved by the local ethical committee of the HUVR (CEEA O309-N-15). A cohort of paraffin-embedded tissue samples from 25 patients with ovarian cancer was obtained from the biobank of Hospital Universitario Virgen del Rocío-Instituto de Biomedicina de Sevilla (Sevilla, Spain) for the RNA expression studies and the evaluation of the correlation of the clinicopathological features (see Supplementary Table S2). The samples were obtained from biopsies of patients who were subjected to platinum treatment and who were evaluated for their response according to the RECIST criteria; normal tissue, platinum-resistant tumor samples and platinum-sensitive tumor samples were obtained. The tumor samples were sent to the pathology laboratory for diagnosis and were prepared for storage with formalin fixation and paraffin embedding. The samples were stained with haematoxylin/eosin, and RNA was extracted from the tumor tissue.

Ethics approval and consent to participate: All methods were performed in accordance with the relevant guidelines and regulations of the Institute for Biomedical Research of Seville (IBIS) and University Hospital Virgen del Rocio (HUVR). All animal experiments and the entire procedures of the patient cohort were performed according to the experimental protocol approved by HUVR Animals Ethics (CEI 0309-N-15).

Consent for publication: Written consent for publication was obtained from all patients involved in our study

Availability of data and materials: The datasets used and/or analysed during the current study are available from the corresponding author upon reasonable request. ATAC-seq data generated in this study is available through the Gene Expression Omnibus (GEO) accession number GSE210599.

Competing interests: The authors declare that they have no other competing financial interests.

Funding: This research was funded by Grants RTI2018-097455-B-I00 and PID2021-122629OB-I00 funded by MCIN/AEI/10.13039/501100011033 and by “ERDF A way of making Europe”, by the “European Union”. Additional grants from CIBER de Cáncer (CB16/12/00275), from Consejeria de Salud (PI-0397-2017) and Project P18-RT-2501 from 2018 competitive research projects call within the scope of PAIDI 2020—80% co-financed by the European Regional Development Fund (ERDF) from the Regional Ministry of Economic Transformation, Industry, Knowledge and Universities. Junta de Andalucía. Special thanks to the AECC (Spanish Association of Cancer Research) Founding Ref. GC16173720CARR for supporting this work. SMG was funded by a grant from the Fundación AECC. EMV-S and JMS-P are funded by postdoctoral fellowships from Junta de Andalucía (DOC_01655 and DOC_00512, respectively).

Acknowledgements: The authors thank the donors and the HUVR-IBiS Biobank (Andalusian Public Health System Biobank and ISCIII-Red de Biobancos PT17/0015/0041) for the human specimens that were used in this study.

Author contributions: SMG and AC conceived and designed this study. SMG and EMVS performed the experiments; JMSP analysed the NGS data; PEG collected the clinical data; and SMG and AC analysed and interpreted the data and drafted and edited the manuscript. All authors revised the manuscript

Asare-Werehene, M., K. Nakka, A. Reunov, C.T. Chiu, W.T. Lee, M.R. Abedini, P.W. Wang, D.B. Shieh, F.J. Dilworth, E. Carmona, T. Le, A.M. Mes-Masson, D. Burger, and B.K. Tsang. 2020. The exosome-mediated autocrine and paracrine actions of plasma gelsolin in ovarian cancer chemoresistance. Oncogene 39:1600-1616.
Bapat, S.A., A.M. Mali, C.B. Koppikar, and N.K. Kurrey. 2005. Stem and progenitor-like cells contribute to the aggressive behavior of human epithelial ovarian cancer. Cancer Res 65:3025-3029.
Barrandon, Y., and H. Green. 1987. Three clonal types of keratinocyte with different capacities for multiplication. Proc Natl Acad Sci U S A 84:2302-2306.
Batlle, E., and H. Clevers. 2017. Cancer stem cells revisited. Nat Med 23:1124-1134.
Beck, B., and C. Blanpain. 2013. Unravelling cancer stem cell potential. Nat Rev Cancer 13:727-738.
Bentsen, M., P. Goymann, H. Schultheis, K. Klee, A. Petrova, R. Wiegandt, A. Fust, J. Preussner, C. Kuenne, T. Braun, J. Kim, and M. Looso. 2020. ATAC-seq footprinting unravels kinetics of transcription factor binding during zygotic genome activation. Nat Commun 11:4267.
Buenrostro, J.D., P.G. Giresi, L.C. Zaba, H.Y. Chang, and W.J. Greenleaf. 2013. Transposition of native chromatin for fast and sensitive epigenomic profiling of open chromatin, DNA-binding proteins and nucleosome position. Nat Methods 10:1213-1218.
Carnero, A., A. Cuadrado, L. del Peso, and J.C. Lacal. 1994. Activation of type D phospholipase by serum stimulation and ras-induced transformation in NIH3T3 cells. Oncogene 9:1387-1395.
Carnero, A., Y. Garcia-Mayea, C. Mir, J. Lorente, I.T. Rubio, and L.L. ME. 2016. The cancer stem-cell signaling network and resistance to therapy. Cancer Treat Rev 49:25-36.
Carnero, A., and M. Lleonart. 2016. The hypoxic microenvironment: A determinant of cancer stem cell evolution. Bioessays 38 Suppl 1:S65-74.
Chau, W.K., C.K. Ip, A.S. Mak, H.C. Lai, and A.S. Wong. 2013. c-Kit mediates chemoresistance and tumor-initiating capacity of ovarian cancer cells through activation of Wnt/beta-catenin-ATP-binding cassette G2 signaling. Oncogene 32:2767-2781.
Chen, A., J. Sceneay, N. Godde, T. Kinwel, S. Ham, E.W. Thompson, P.O. Humbert, and A. Moller. 2018a. Intermittent hypoxia induces a metastatic phenotype in breast cancer. Oncogene 37:4214-4225.
Chen, X., J. Zhou, X. Li, X. Wang, Y. Lin, and X. Wang. 2018b. Exosomes derived from hypoxic epithelial ovarian cancer cells deliver microRNAs to macrophages and elicit a tumor-promoted phenotype. Cancer Lett 435:80-91.
Dorayappan, K.D.P., R. Wanner, J.J. Wallbillich, U. Saini, R. Zingarelli, A.A. Suarez, D.E. Cohn, and K. Selvendiran. 2018. Hypoxia-induced exosomes contribute to a more aggressive and chemoresistant ovarian cancer phenotype: a novel mechanism linking STAT3/Rab proteins. Oncogene 37:3806-3821.
Fernandez-Minan, A., J. Bessa, J.J. Tena, and J.L. Gomez-Skarmeta. 2016. Assay for transposase-accessible chromatin and circularized chromosome conformation capture, two methods to explore the regulatory landscapes of genes in zebrafish. Methods Cell Biol 135:413-430.
Fornes, O., J.A. Castro-Mondragon, A. Khan, R. van der Lee, X. Zhang, P.A. Richmond, B.P. Modi, S. Correard, M. Gheorghe, D. Baranasic, W. Santana-Garcia, G. Tan, J. Cheneby, B. Ballester, F. Parcy, A. Sandelin, B. Lenhard, W.W. Wasserman, and A. Mathelier. 2020. JASPAR 2020: update of the open-access database of transcription factor binding profiles. Nucleic Acids Res 48:D87-D92.
Frankel, P., M. Ramos, J. Flom, S. Bychenok, T. Joseph, E. Kerkhoff, U.R. Rapp, L.A. Feig, and D.A. Foster. 1999. Ral and Rho-dependent activation of phospholipase D in v-Raf-transformed cells. Biochem Biophys Res Commun 255:502-507.
Ganesan, R., M. Mahankali, G. Alter, and J. Gomez-Cambronero. 2015. Two sites of action for PLD2 inhibitors: The enzyme catalytic center and an allosteric, phosphoinositide biding pocket. Biochim Biophys Acta 1851:261-272.
Ghim, J., J.S. Moon, C.S. Lee, J. Lee, P. Song, A. Lee, J.H. Jang, D. Kim, J.H. Yoon, Y.J. Koh, C. Chelakkot, B.J. Kang, J.M. Kim, K.L. Kim, Y.R. Yang, Y. Kim, S.H. Kim, D. Hwang, P.G. Suh, G.Y. Koh, Y.Y. Kong, and S.H. Ryu. 2014. Endothelial deletion of phospholipase D2 reduces hypoxic response and pathological angiogenesis. Arterioscler Thromb Vasc Biol 34:1697-1703.
Grant, C.E., and T.L. Bailey. 2021. XSTREME: Comprehensive motif analysis of biological sequence datasets. bioRxiv
Haeussler, M., A.S. Zweig, C. Tyner, M.L. Speir, K.R. Rosenbloom, B.J. Raney, C.M. Lee, B.T. Lee, A.S. Hinrichs, J.N. Gonzalez, D. Gibson, M. Diekhans, H. Clawson, J. Casper, G.P. Barber, D. Haussler, R.M. Kuhn, and W.J. Kent. 2019. The UCSC Genome Browser database: 2019 update. Nucleic Acids Res 47:D853-D858.
Han, S., J. Huh, W. Kim, S. Jeong, S. Min do, and Y. Jung. 2014. Phospholipase D activates HIF-1-VEGF pathway via phosphatidic acid. Exp Mol Med 46:e126.
Harel-Dassa, K., S. Yedgar, C.G. Trope, B. Davidson, and R. Reich. 2017. Phospholipase D messenger RNA expression and clinical role in high-grade serous carcinoma. Hum Pathol 62:115-121.
Harris, A.L. 2002. Hypoxia--a key regulatory factor in tumour growth. Nat Rev Cancer 2:38-47.
He, M., H. Wu, Q. Jiang, Y. Liu, L. Han, Y. Yan, B. Wei, F. Liu, X. Deng, H. Chen, L. Zhao, M. Wang, X. Wu, W. Yao, H. Zhao, J. Chen, and M. Wei. 2019. Hypoxia-inducible factor-2alpha directly promotes BCRP expression and mediates the resistance of ovarian cancer stem cells to adriamycin. Mol Oncol 13:403-421.
Henkels, K.M., G.P. Boivin, E.S. Dudley, S.J. Berberich, and J. Gomez-Cambronero. 2013. Phospholipase D (PLD) drives cell invasion, tumor growth and metastasis in a human breast cancer xenograph model. Oncogene 32:5551-5562.
Hiller, M., S. Agarwal, J.H. Notwell, R. Parikh, H. Guturu, A.M. Wenger, and G. Bejerano. 2013. Computational methods to detect conserved non-genic elements in phylogenetically isolated genomes: application to zebrafish. Nucleic Acids Res 41:e151.
Hu, C.J., A. Sataur, L. Wang, H. Chen, and M.C. Simon. 2007. The N-terminal transactivation domain confers target gene specificity of hypoxia-inducible factors HIF-1alpha and HIF-2alpha. Mol Biol Cell 18:4528-4542.
Hu, L., C. McArthur, and R.B. Jaffe. 2010. Ovarian cancer stem-like side-population cells are tumourigenic and chemoresistant. Br J Cancer 102:1276-1283.
Jang, J.H., C.S. Lee, D. Hwang, and S.H. Ryu. 2012. Understanding of the roles of phospholipase D and phosphatidic acid through their binding partners. Prog Lipid Res 51:71-81.
Jing, X., F. Yang, C. Shao, K. Wei, M. Xie, H. Shen, and Y. Shu. 2019. Role of hypoxia in cancer therapy by regulating the tumor microenvironment. Mol Cancer 18:157.
Kaelin, W.G., Jr., and P.J. Ratcliffe. 2008. Oxygen sensing by metazoans: the central role of the HIF hydroxylase pathway. Mol Cell 30:393-402.
Klemba, A., L. Bodnar, H. Was, K.K. Brodaczewska, G. Wcislo, C.A. Szczylik, and C. Kieda. 2020. Hypoxia-Mediated Decrease of Ovarian Cancer Cells Reaction to Treatment: Significance for Chemo- and Immunotherapies. Int J Mol Sci 21:
Kunz, M., and S.M. Ibrahim. 2003. Molecular responses to hypoxia in tumor cells. Mol Cancer 2:23.
Langmead, B., and S.L. Salzberg. 2012. Fast gapped-read alignment with Bowtie 2. Nat Methods 9:357-359.
Li, H., X. Chen, T. Calhoun-Davis, K. Claypool, and D.G. Tang. 2008. PC3 human prostate carcinoma cell holoclones contain self-renewing tumor-initiating cells. Cancer Res 68:1820-1825.
Li, H., B. Handsaker, A. Wysoker, T. Fennell, J. Ruan, N. Homer, G. Marth, G. Abecasis, R. Durbin, and S. Genome Project Data Processing. 2009a. The Sequence Alignment/Map format and SAMtools. Bioinformatics 25:2078-2079.
Li, Z., S. Bao, Q. Wu, H. Wang, C. Eyler, S. Sathornsumetee, Q. Shi, Y. Cao, J. Lathia, R.E. McLendon, A.B. Hjelmeland, and J.N. Rich. 2009b. Hypoxia-inducible factors regulate tumorigenic capacity of glioma stem cells. Cancer Cell 15:501-513.
Liang, D., Y. Ma, J. Liu, C.G. Trope, R. Holm, J.M. Nesland, and Z. Suo. 2012. The hypoxic microenvironment upgrades stem-like properties of ovarian cancer cells. BMC Cancer 12:201.
Lin, S.H., and A.C. Koong. 2018. Breathing New Life Into Hypoxia-Targeted Therapies for Non-Small Cell Lung Cancer. J Natl Cancer Inst 110:
Liu, M., K. Du, B. Jiang, and X. Wu. 2020. High Expression of PhospholipaseD2 Induced by Hypoxia Promotes Proliferation of Colon Cancer Cells through Activating NF- kappa Bp65 Signaling Pathway. Pathol Oncol Res 26:281-290.
Love, M.I., W. Huber, and S. Anders. 2014. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol 15:550.
Mathieu, J., Z. Zhang, W. Zhou, A.J. Wang, J.M. Heddleston, C.M. Pinna, A. Hubaud, B. Stadler, M. Choi, M. Bar, M. Tewari, A. Liu, R. Vessella, R. Rostomily, D. Born, M. Horwitz, C. Ware, C.A. Blau, M.A. Cleary, J.N. Rich, and H. Ruohola-Baker. 2011. HIF induces human embryonic stem cell markers in cancer cells. Cancer Res 71:4640-4652.
Maugeri-Sacca, M., P. Vigneri, and R. De Maria. 2011. Cancer stem cells and chemosensitivity. Clin Cancer Res 17:4942-4947.
Mayer, A., M. Hockel, A. Wree, C. Leo, L.C. Horn, and P. Vaupel. 2008. Lack of hypoxic response in uterine leiomyomas despite severe tissue hypoxia. Cancer Res 68:4719-4726.
Munoz-Galvan, S., and A. Carnero. 2020. Targeting Cancer Stem Cells to Overcome Therapy Resistance in Ovarian Cancer. Cells 9:
Munoz-Galvan, S., B. Felipe-Abrio, M. Garcia-Carrasco, J. Dominguez-Pinol, E. Suarez-Martinez, E.M. Verdugo-Sivianes, A. Espinosa-Sanchez, L.E. Navas, D. Otero-Albiol, J.J. Marin, M.P. Jimenez-Garcia, J.M. Garcia-Heredia, A.G. Quiroga, P. Estevez-Garcia, and A. Carnero. 2019a. New markers for human ovarian cancer that link platinum resistance to the cancer stem cell phenotype and define new therapeutic combinations and diagnostic tools. J Exp Clin Cancer Res 38:234.
Munoz-Galvan, S., B. Felipe-Abrio, E.M. Verdugo-Sivianes, M. Perez, M.P. Jimenez-Garcia, E. Suarez-Martinez, P. Estevez-Garcia, and A. Carnero. 2020. Downregulation of MYPT1 increases tumor resistance in ovarian cancer by targeting the Hippo pathway and increasing the stemness. Mol Cancer 19:7.
Munoz-Galvan, S., A. Lucena-Cacace, M. Perez, D. Otero-Albiol, J. Gomez-Cambronero, and A. Carnero. 2019b. Tumor cell-secreted PLD increases tumor stemness by senescence-mediated communication with microenvironment. Oncogene 38:1309-1323.
Nunes, S.C., C. Ramos, F. Lopes-Coelho, C.O. Sequeira, F. Silva, S. Gouveia-Fernandes, A. Rodrigues, A. Guimaraes, M. Silveira, S. Abreu, V.E. Santo, C. Brito, A. Felix, S.A. Pereira, and J. Serpa. 2018. Cysteine allows ovarian cancer cells to adapt to hypoxia and to escape from carboplatin cytotoxicity. Sci Rep 8:9513.
Onallah, H., S.T. Mannully, B. Davidson, and R. Reich. 2022. Exosome Secretion and Epithelial-Mesenchymal Transition in Ovarian Cancer Are Regulated by Phospholipase D. Int J Mol Sci 23:
Park, M.H., S.S. Bae, K.Y. Choi, and S. Min do. 2015. Phospholipase D2 promotes degradation of hypoxia-inducible factor-1alpha independent of lipase activity. Exp Mol Med 47:e196.
Peng, G., and Y. Liu. 2015. Hypoxia-inducible factors in cancer stem cells and inflammation. Trends Pharmacol Sci 36:374-383.
Qin, J., Y. Liu, Y. Lu, M. Liu, M. Li, J. Li, and L. Wu. 2017. Hypoxia-inducible factor 1 alpha promotes cancer stem cells-like properties in human ovarian cancer cells by upregulating SIRT1 expression. Sci Rep 7:10592.
Quinlan, A.R., and I.M. Hall. 2010. BEDTools: a flexible suite of utilities for comparing genomic features. Bioinformatics 26:841-842.
Ramirez, F., D.P. Ryan, B. Gruning, V. Bhardwaj, F. Kilpert, A.S. Richter, S. Heyne, F. Dundar, and T. Manke. 2016. deepTools2: a next generation web server for deep-sequencing data analysis. Nucleic Acids Res 44:W160-165.
Saito, M., M. Iwadate, M. Higashimoto, K. Ono, Y. Takebayashi, and S. Takenoshita. 2007. Expression of phospholipase D2 in human colorectal carcinoma. Oncol Rep 18:1329-1334.
Santos-Pereira, J.M., L. Gallardo-Fuentes, A. Neto, R.D. Acemel, and J.J. Tena. 2019. Pioneer and repressive functions of p63 during zebrafish embryonic ectoderm specification. Nat Commun 10:3049.
Schito, L., and G.L. Semenza. 2016. Hypoxia-Inducible Factors: Master Regulators of Cancer Progression. Trends Cancer 2:758-770.
Schwab, L.P., D.L. Peacock, D. Majumdar, J.F. Ingels, L.C. Jensen, K.D. Smith, R.C. Cushing, and T.N. Seagroves. 2012. Hypoxia-inducible factor 1alpha promotes primary tumor growth and tumor-initiating cell activity in breast cancer. Breast Cancer Res 14:R6.
Seo, E.J., D.K. Kim, I.H. Jang, E.J. Choi, S.H. Shin, S.I. Lee, S.M. Kwon, K.H. Kim, D.S. Suh, and J.H. Kim. 2016. Hypoxia-NOTCH1-SOX2 signaling is important for maintaining cancer stem cells in ovarian cancer. Oncotarget 7:55624-55638.
Siegel, R.L., K.D. Miller, and A. Jemal. 2020. Cancer statistics, 2020. CA Cancer J Clin 70:7-30.
Simiantonaki, N., M. Taxeidis, C. Jayasinghe, U. Kurzik-Dumke, and C.J. Kirkpatrick. 2008. Hypoxia-inducible factor 1 alpha expression increases during colorectal carcinogenesis and tumor progression. BMC Cancer 8:320.
Song, J.G., L.M. Pfeffer, and D.A. Foster. 1991. v-Src increases diacylglycerol levels via a type D phospholipase-mediated hydrolysis of phosphatidylcholine. Mol Cell Biol 11:4903-4908.
Sun, H.C., Z.J. Qiu, J. Liu, J. Sun, T. Jiang, K.J. Huang, M. Yao, and C. Huang. 2007. Expression of hypoxia-inducible factor-1 alpha and associated proteins in pancreatic ductal adenocarcinoma and their impact on prognosis. Int J Oncol 30:1359-1367.
Torre, L.A., B. Trabert, C.E. DeSantis, K.D. Miller, G. Samimi, C.D. Runowicz, M.M. Gaudet, A. Jemal, and R.L. Siegel. 2018. Ovarian cancer statistics, 2018. CA Cancer J Clin 68:284-296.
Toschi, A., J. Edelstein, P. Rockwell, M. Ohh, and D.A. Foster. 2008. HIF alpha expression in VHL-deficient renal cancer cells is dependent on phospholipase D. Oncogene 27:2746-2753.
Ueda, S., T. Saeki, A. Osaki, T. Yamane, and I. Kuji. 2017. Bevacizumab Induces Acute Hypoxia and Cancer Progression in Patients with Refractory Breast Cancer: Multimodal Functional Imaging and Multiplex Cytokine Analysis. Clin Cancer Res 23:5769-5778.
Unruh, A., A. Ressel, H.G. Mohamed, R.S. Johnson, R. Nadrowitz, E. Richter, D.M. Katschinski, and R.H. Wenger. 2003. The hypoxia-inducible factor-1 alpha is a negative factor for tumor therapy. Oncogene 22:3213-3220.
Wang, W.J., H. Sui, C. Qi, Q. Li, J. Zhang, S.F. Wu, M.Z. Mei, Y.Y. Lu, Y.T. Wan, H. Chang, and P.T. Guo. 2016. Ursolic acid inhibits proliferation and reverses drug resistance of ovarian cancer stem cells by downregulating ABCG2 through suppressing the expression of hypoxia-inducible factor-1alpha in vitro. Oncol Rep 36:428-440.
Wang, Y., Y. Liu, S.N. Malek, P. Zheng, and Y. Liu. 2011. Targeting HIF1alpha eliminates cancer stem cells in hematological malignancies. Cell Stem Cell 8:399-411.
Wilson, W.R., and M.P. Hay. 2011. Targeting hypoxia in cancer therapy. Nat Rev Cancer 11:393-410.
Yoshida, Y., K. Takahashi, K. Okita, T. Ichisaka, and S. Yamanaka. 2009. Hypoxia enhances the generation of induced pluripotent stem cells. Cell Stem Cell 5:237-241.
Zhang, Y., T. Liu, C.A. Meyer, J. Eeckhoute, D.S. Johnson, B.E. Bernstein, C. Nusbaum, R.M. Myers, M. Brown, W. Li, and X.S. Liu. 2008. Model-based analysis of ChIP-Seq (MACS). Genome Biol 9:R137.
Zheng, Y., V. Rodrik, A. Toschi, M. Shi, L. Hui, Y. Shen, and D.A. Foster. 2006. Phospholipase D couples survival and migration signals in stress response of human cancer cells. J Biol Chem 281:15862-15868.
Zhu, X., H. Shen, X. Yin, M. Yang, H. Wei, Q. Chen, F. Feng, Y. Liu, W. Xu, and Y. Li. 2019. Macrophages derived exosomes deliver miR-223 to epithelial ovarian cancer cells to elicit a chemoresistant phenotype. J Exp Clin Cancer Res 38:81.

MunozGalvansupplementary.docx

Download PDF

Journal Publication

published 26 Feb, 2024

Read the published version in Journal of Experimental & Clinical Cancer Research →

Editorial decision: Major revision
21 Dec, 2023
Reviewers agreed at journal
12 Dec, 2023
Reviewers invited by journal
12 Dec, 2023
Editor assigned by journal
11 Dec, 2023
First submitted to journal
11 Dec, 2023

You are reading this latest preprint version

Suppression of hypoxia-induced stemness and chemoresistance in ovarian tumors

Status:

Journal Publication

Version 1

Abstract

Figures

BACKGROUND

RESULTS

DISCUSSION

CONCLUSIONS

METHODS

Declarations

References

Supplementary Files

Status:

Journal Publication

Version 1