2.1 RAC1 expression is significantly upregulated in patients with OC, and its expression is negatively correlated with patient prognosis

First, we detected RAC1 expression at the mRNA and protein levels (immunohistochemistry, IHC) in OC and normal ovarian epithelial samples from our center, and the results revealed that the RAC1 expression level was significantly increased in OC (Figure 1A,B), and the RAC1 mRNA level was consistent with that reported in the Gene Expression Profiling Interactive Analysis (GEPIA) database (Figure S1A).

Furthermore, we divided the IHC results of OC samples into scores of 0, 1, 2, and 3 according to the immunoreactivity score (IRS), among which 0 and 1 were assigned to the low-expression group, and 2 and 3 were assigned to the high-expression group (Figure 1C,D).²¹ The results revealed that patients in the high RAC1 expression group had shorter overall survival (OS) and progression-free survival (PFS; p-OS < 0.001, hazard ratio [HR] = 0.33, 95% confidence interval [CI] = 0.19–0.55; **p-PFS < 0.01, HR = 0.50, 95% CI = 0.31–0.81; Figure 1E,F). This finding is consistent with the results of the Kaplan–Meier (K–M) plotter (kmplot.com) database (Figure S1B).

Univariate regression analysis revealed that RAC1 expression was significantly associated with poor OS (HR = 2.043, 95% CI = 1.497–3.432, **p < 0.01; Figure 1G). Other variables associated with survival included carbohydrate antigen 125 (CA125) level, lymph node (LN) involvement, distant metastasis (M) status, and International Federation of Gynaecology and Obstetrics (FIGO) stage. Multivariate regression analysis revealed that high RAC1 expression was independently associated with poor OS in OC patients (HR = 2.203, 95% CI = 1.110–3.232, *p < 0.05; Figure 1H). In addition, further analysis of the relationships between RAC1 expression and clinicopathological parameters revealed that RC1 expression was significantly correlated with distant metastasis (**p = 0.025; Table S1), which was consistent with most reported results.

Overall, these data further demonstrate that RAC1 expression is significantly associated with the prognosis of OC patients and can be used as a biomarker for predicting the OC prognosis.

2.2 RAC1 promotes OC cell proliferation, cell migration, and invasion in vitro

We detected the expression of RAC1 at the protein and mRNA levels in eight OC cell lines (Figure S2A,B). For the highly expressing HEY-A8 and SKOV3 cell lines, two independent stable RAC1-expressing cell lines were constructed to effectively reduce RAC1 expression. We constructed stable RAC1-overexpressing ES-2 and A2780 cells with low expression (Figure S2C,D). CCK8 and cell cycle assays revealed that RAC1 knockdown significantly inhibited the proliferation of HEYA8 and SKOV3 OC cells and increased the proportion of G0/G1 phase cells (Figure 2A–C). The overexpression of RAC1 significantly promoted the proliferation of ES-2 and A2780 cells and decreased the proportion of cells in G0/G1 phase (Figure 2B–D). Taken together, these data indicate that RAC1 can promote the proliferation of OC cells.

RAC1 has been reported to be associated with migration, invasion and metastasis in OC.^{22, 23} Wound healing and Transwell assays revealed that RAC1 knockdown significantly inhibited the migration and invasion of HEY-A8 and SKOV OC cells and that overexpression of RAC1 significantly enhanced these phenotypes (Figure 2E,F). In conclusion, these phenotypic results suggest that RAC1 can significantly promote OC cell migration and invasion. In addition, an analysis of clinical data from our center showed that RAC1 is significantly associated with LN metastasis and distant metastasis in OC patients.

2.3 The RAC1 inhibitor NSC23766 (NSC) enhances the antitumor effect of PTX in vitro and in vivo

PTX is the first-line chemotherapy agent for OC.²³ We treated the control group and the RAC1-knockdown group with PTX to explore the effect of RAC1 on the antitumor efficacy of PTX, and the results showed that RAC1 knockdown could significantly enhance the antitumor effect of PTX. Similarly, NSC combined with PTX was compared with PTX alone, and the results showed that NSC could enhance the antitumor effect of PTX (Figure 3A,B). These results demonstrate that RAC1 can promote OC progression by promoting PTX resistance.

Through animal experiments, we found that PTX or NSC alone could significantly inhibit the growth of subcutaneously transplanted tumors and NSC combined with PTX could enhance this inhibitory effect (Figure 3C,D). Further IHC analysis of the tumors revealed that the combined treatment significantly reduced the Ki-67 expression level (Figure 3E). The above experiments indicate that the RAC1 inhibitor NSC can increase the antitumor effect of PTX in OC.

2.4 RAC1 can promote pyroptosis inhibition by mediating PTX resistance

During the treatment of OC, cancer cells gradually undergo apoptosis and are resistant to PTX chemotherapy.²⁴ However, whether this resistance mechanism is involved in pyroptosis is unclear.

Lactate dehydrogenase (LDH) assays can be used to detect pyroptosis.²⁵ LDH release tests revealed that a high dose of PTX-induced pyroptosis in the OC cell lines HEY-A8 and SKOV3 WT in a dose-dependent manner. However, PTX-resistant (PTXR) cell lines did not show dose dependence of PTX-induced pyroptosis (Figure 4A,B). Next, we compared LDH release in WT ES2 and A2780 cells after PTX was administered to the control group and the group transduced with the lentivirus containing the RAC1 overexpression plasmid, and the results indicated that LDH release was distinctly reduced in the RAC1 overexpression group (Figure 4C). The above experiments proved that RAC1 inhibited PTX-induced LDH release. We knocked down RAC1 in PTXR HEY-A8 and SKOV3 cells and observed a significant increase in LDH release (Figure 4D). High-resolution confocal imaging was further applied to analyze pyroptosis in cells subjected to the aforementioned treatments, and the trend of the results was consistent with that of the LDH release experiment (Figure 4E,F).

These findings suggest that knocking down RAC1 can reverse the inhibition of pyroptosis induced by PTX resistance.

2.5 Chemotherapy-induced GSDMD-mediated OC pyroptosis is driven by RAC1-controlled caspase-1-associated inflammasome activation

WT HEY-A8 cells and SKOV3 cells and PTXR HEY-A8 cells and SKOV3 OC cells were treated with different concentrations of PTX (0.08 nM, 0.4 nM, 2 nM, 10 nM, or 50 nM). A high dose of PTX increased the expression levels of the pyroptosis-related molecular markers cleaved caspase-3, cleaved caspase-1, and gasdermin D (GSDMD)-N. The expression levels of pyroptosis-related molecular markers and the release of the inflammatory factors IL-1β/IL-18 and the nucleotide oligomerization domain-like receptor family pyrin domain-containing 3 (NLRP3) inflammasome in PTXR OC cells were significantly decreased (Figure 5A). Therefore, we concluded that high doses of PTX can cause pyroptosis in WT OC cell lines.

In WT ES-2 and A2780 cells, RAC1 overexpression significantly reduced the levels of cleaved caspase-1, cleaved caspase-3, GSDMD-N, the inflammatory factor IL-1β/IL-18 and the NLRP3 inflammasome after the administration of 10 nM PTX (Figure 5B). In PTXR HEY-A8 and SKOV3 cells, RAC1 knockdown significantly increased the levels of cleaved caspase-1, cleaved caspase-3, GSDMD-N, the inflammatory factor IL-1β/IL-18 and the NLRP3 inflammasome (Figure 5C). Therefore, we concluded that RAC1 can mediate pyroptosis in OC cells through GSDMD-mediated activation of the caspase-1 inflammasome.

2.6 RAC1 drives GSDMD-mediated pyroptosis in OC through the PAK4/MAPK pathway

Transcriptome sequencing and gene set enrichment analysis (GSEA) were performed on the OC cell lines in the HEY-A8/RAC1-knockdown group and the NC no-load group, and the mitogen-activated protein kinase (MAPK) pathway was highly enriched in the RAC1-knockdown group (Figure 6A and Table S2). Through further verification of the expression of MAPK pathway-related molecular markers, RAC1 depletion significantly inhibited the enrichment of the MAPK signaling pathway, whereas RAC1 overexpression significantly promoted the enrichment of the MAPK signaling pathway (Figure 6B).

Studies have shown that RAC1 can target PAK4-mediated pyroptosis in tumor cells.¹² After the Myc-RAC1 or NC plasmid was transfected into HEK293T cells, protein coimmunoprecipitation (Co-IP) and mass spectrometry (MS) were performed. The results revealed significantly higher expression of PAK4 in the Myc-RAC1 group than in the NC group (Figure 6C, Tables S3 and S4). Therefore, we can infer that RAC1 can bind to the targeted binding protein PAK4; a positive correlation was observed between the expression levels of the two proteins. We then further verified the forward and reverse Co-IP results in the OC cell line HEY-A8 and found that RAC1 could bind the PAK4 protein (Figure 6D). When RAC1 was knocked down or overexpressed, PAK4 protein expression decreased or increased, respectively (Figure 6E).

We added cyclohexane (CHX) to RAC1-knockdown or RAC1-overexpressing cell lines, and the proteins were collected at different time points (0, 4, 8, and 12 h) to further explore the mechanism by which RAC1 regulates PAK4.²¹ The experimental results revealed that the degradation rate of the PAK4 protein increased after RAC1 knockdown. The overexpression of RAC1 slowed the degradation rate of the PAK4 protein (Figure 6F,G). These results indicated that RAC1 could inhibit the degradation of the PAK4 protein. The methods of protein degradation vary, but ubiquitination has been reported to be more likely. Subsequent ubiquitination experiments revealed that the ubiquitination level in the RAC1-overexpressing group decreased, whereas that in the RAC1-knockdown group increased (Figure 6H,I). Therefore, we concluded that RAC1 can inhibit the ubiquitination and degradation of PAK4.

2.7 PAK4 is associated with the patient prognosis

These studies suggest that PAK4 can be positively regulated by RAC1 in vitro. Here, we performed PAK4 IHC staining on the same batch of TMAs and divided the degree of staining into a low RAC1 expression group and a high RAC1 expression group according to the aforementioned IRS scoring method. The results revealed that 71.43% of patients in the low RAC1 expression group had low PAK4 expression. PAK4 was highly expressed in 73.08% of patients in the high RAC1 expression group (Figure S4A,B). The results showed that patients with high PAK4 protein expression had a poor prognosis (***p-OS < 0.01, HR = 0.48, 95% CI = 0.28–0.80; *p-PFS < 0.001, HR = 0.32, 95% CI = 0.20–0.53; Figure S4C,D). In conclusion, the above results show that PAK4 is positively correlated with RAC1 expression at the protein level and negatively correlated with the patient prognosis.

2.8 RAC1 regulates OC progression by regulating the MAPK signaling pathway through the targeting of PAK4

Many studies indicate that PAK4 can promote resistance to chemotherapy drugs such as platinum-based chemotherapies, but its effect on sensitivity to PTX has not been reported.²⁶ Therefore, we knocked down PAK4 expression in the OC cell lines HEY-A8 and SKOV3 and found that PAK4 knockdown attenuated PTX resistance (Figure 7A). Similarly, the overexpression of PAK4 in the OC cell lines ES-2 and A2780 promoted PTX resistance (Figure 7B). These results indicate that PAK4 can promote PTX resistance in OC cell lines. In the PTXR HEY-A8 and SKOV3 cell lines, PAK4 knockdown promoted LDH release (***p-PTXR-HEY-A8 < 0.001, ***p-PTXR-SKOV3 < 0.001; Figure 7C). The overexpression of PAK4 significantly inhibited LDH release from the WT ES-2 cell line and WT A2780 cell line (***p-WT-ES2 < 0.001, ***p-WT-A2780 < 0.001; Figure 7D). These results indicate that PAK4 can inhibit LDH release. Collectively, these results suggest that PAK4 may promote PTX resistance in OC cell lines by inhibiting pyroptosis.

Next, we further verified the roles of RAC1 and PAK4 in pyroptosis in OC cell lines through a rescue experiment. We knocked down PAK4 in the RAC1-overexpressing OC cell lines ES-2 and A2780 and found that PAK4 knockdown reversed the tumor-promoting effects of RAC1 overexpression. The effects were similar when the RAC1-overexpressing ES-2 and A2780 cell lines were treated with PAK4 inhibitors (Figure 7E,F).

We further validated the protein expression of classical caspase-1-dependent pyroptosis markers.²⁷ Moreover, knocking down PAK4 or administering PAK4 inhibitors reversed the downregulation of pyroptosis-related molecular markers induced by RAC1 overexpression (Figure 7G,H). Therefore, these results suggest that RAC1 can mediate PTX resistance by targeting PAK4, thereby inhibiting the classical caspase-1-dependent pyroptosis pathway.

2.9 In animal models and clinical specimens, RAC1 can promote PTX resistance in OC cells by targeting the PAK4/MAPK signaling pathway and mediating GSDMD-mediated pyroptosis

The above studies demonstrated that RAC1 can target PAK4 and mediate PTX resistance through the MAPK pathway, ultimately leading to caspase-1/GSDMD-mediated pyroptosis. Next, we further verified this mechanism at the animal level and in clinical specimens.

We performed an axillary injection of RAC1-overexpressing A2780 cell lines or empty vector-containing cell lines on day 0 and administered a placebo, PTX alone, a PAK4 inhibitor (PF3758309, PF) alone, or PF combined with PTX beginning on day 7 until week 4 to investigate the role of PAK4 in RAC-1-mediated PTX resistance in vivo.²⁸ At the beginning of the fourth week, the mice were sacrificed, and the results revealed that PF inhibited the growth of subcutaneous OC tumors. Compared with PF alone or PTX alone, PF combined with PTX had a greater inhibitory effect on the growth of subcutaneous OC tumors. PF reversed the RAC1 overexpression-mediated increase in tumor growth; moreover, PTX combined with PF increased the effects of PF alone on the RAC1 overexpression group (Figure 8A,B). In addition, compared with PF alone or PTX alone, PF combined with PTX significantly inhibited tumor growth in the PTXR model group (Figure 8C,D). IHC staining of tumors from the animals showed that PF promoted the expression of pyroptosis-related proteins and that PF combined with PTX increased the expression of pyroptosis-related proteins (Figure S5).

We further verified the relationship between RAC1/PAK4 expression and the patient prognosis in clinical samples. The group with high expression of these two genes at the same time had a worse prognosis than the other groups did, which was consistent with the results of the cell and animal experiments described above.

3.1 Limitations

Our study has several limitations. Although we demonstrated that pyroptosis in OC cells plays a crucial role in PTX resistance in vitro, the role of PTX-induced pyroptosis in OC patient samples remains to be further explored. In addition, this study did not assess any models of resistance to multiple chemotherapeutic drugs.

5.1 Patient dataset

All cell and animal experiments and clinical samples used in this study were authorized by the Ethics Committee of the Shanghai Cancer Center of Fudan University (FUSCC) and approved by the Ethics Committee of the World Health Organization. All patient samples, including samples from 120 patients with a normal ovarian epithelium and 120 OC patients, were obtained from our center and prepared into tissue chips. Moreover, the clinical data of the patients were collected. Seventy-five patients with OC and 62 patients with a normal ovarian epithelium were randomly selected for the mRNA analysis.

5.2 Statistical analysis

In this study, the K–M method was used for the survival analysis. The Cox risk model was used for univariate and multivariate analyses. Photoshop, GraphPad Prism 8, and Biorender were used for image generation and synthesis. IBM SPSS 23.0 and GraphPad Prism 8 were used for statistical analyses. The results from all statistical analyses are presented as two-sided p values, which are shown as the means ± SDs (*p < 0.05, **p < 0.01, and ***p < 0.001).