1 Introduction
Pancreatic ductal adenocarcinoma (PDAC) is among the leading causes of cancer-related mortality worldwide, with an overall survival rate of less than 5%. Gemcitabine-based chemotherapy is the standard treatment for locally advanced or unresectable PDAC cases [1, 2]. However, drug resistance frequently occurs during chemotherapy. As one of the reasons for the occurrence of drug resistance, tumor repopulation plays a critical role in therapeutic failure. It has been reported that tumor cells surviving chemoradiotherapy could repopulate during treatment intervals [3]. Emerging evidence suggests that dying cells can drive the repopulation of pancreatic cancer cells through signaling pathways involving Hedgehog, Wnt, and the microRNA-193a-TGF-β2 signaling axis [4‐6]. Furthermore, caspase-3 regulates prostaglandin E2 (PGE2) to foster the proliferation of surviving tumor cells, suggesting the existence of a cell death-mediated tumor repopulation pathway [7]. Despite these findings, the precise mechanisms underlying this process remain unclear.
The Hippo pathway plays a crucial role in various human cancers, including PDAC. When the Hippo pathway is deactivated, Yes-associated protein 1 (YAP1)/TAZ undergo dephosphorylation and translocate to the nucleus, where they interact with transcriptional enhanced associate domain (TEAD) to drive tumorigenic processes [8, 9]. Alterations in lipid metabolism, particularly lysophosphatidic acid (LPA) receptor-mediated signaling, profoundly influence cancer cell behavior and promote tumor progression [10]. Autotaxin (ATX), a secreted enzyme highly expressed in PDAC, catalyzes conversion of lysophosphatidylcholine (LPC) into LPA [11]. LPA, in turn, can inhibit large tumor suppressor 1/2 (LATS1/2) via G12/13-coupled receptors, thereby activating the transcriptional co-activator YAP1 [12]. However, the role of the ATX-LPA axis in the repopulation of pancreatic cancer remains insufficiently explored.
Anzeige
In this study, we investigated the hypothesis that dying pancreatic cancer cells may provide initial signals that drive tumor repopulation following chemotherapy. Specifically, we demonstrated that gemcitabine induces dying pancreatic cancer cells to release LPA, which subsequently mediates Hippo pathway inactivation. Additionally, we identified a positive feedback loop between LPA and YAP1. Targeting the LPA-YAP1 axis may offer an effective strategy to overcome chemoresistance in PDAC.
2 Results
2.1 Gemcitabine promotes the repopulation of residual tumor cells in vitro
To assess the repopulation of residual PDAC cells following gemcitabine treatment, we first determined the half-maximal inhibitory concentration (IC50) value of gemcitabine using a Cell Counting Kit-8 (CCK-8) assay. The results revealed that the IC50 values for the PDAC cell lines were approximately 1 µM (Fig. 1A).
Subsequently, we investigated whether dying cells could facilitate the repopulation of residual pancreatic cancer cells using both direct and indirect co-culture models (Fig. 1B). Results from these models showed that gemcitabine-treated dying cells significantly enhanced the proliferation activity of reporter cells (Fig. 1C and D). Collectively, our findings indicate that gemcitabine-treated dying cells could promote the repopulation of residual pancreatic cancer cells.
Fig. 1
Gemcitabine promotes the repopulation of residual tumor cells in vitro. A The half-maximal inhibitory concentration (IC50) of gemcitabine for PATU8988T and AsPC-1 cell lines was determined using the CCK-8 assay. B Schematic illustration of the direct and indirect co-culture models used in this study. C Proliferative activity of reporter cells in the direct co-culture model was assessed by fluorescence intensity. In the gemcitabine-treated (GEM) group, dying cells were treated with 1~5 µM gemcitabine, whereas in the negative control (NC) group, pancreatic cancer cells were treated with DMSO. D Proliferative capacity of reporter cells in the indirect co-culture model was evaluated using the CCK-8 assay. All experiments were performed in biological triplicates, and data are presented as the mean ± standard deviation. Statistical significance was determined by *P < 0.05, **P < 0.01, and ***P < 0.001
×
![]()
Anzeige
2.2 Gemcitabine induces LPA release and Hippo pathway inactivation in PDAC
Using the UPLC-MS/MS detection platform, we performed a lipid metabolomics analysis to investigate lipid alterations. Compared to the control group, gemcitabine treatment led to significant upregulation of various lipids, including triglycerides (TG), phosphatidylinositol (PE), cholesterol, and LPA (Fig. 2A). Our previous studies demonstrated that caspase-3 can promote tumor cell regeneration through its effects on Calcium-independent phospholipase A2 (iPLA2). Notably, the role and mechanism of the iPLA2 - AA (arachidonic acid) - PGE2 signal axis have been thoroughly examined in the context of tumor repopulation [7, 13]. However, a second downstream pathway involving iPLA2-LPC remains unexplored in tumor repopulation. Given that LPC was converted to via ATX [11], we selected LPA as the primary focus of this study among the upregulated lipids. Quantitative analysis revealed that while gemcitabine generally suppresses lipid metabolism in pancreatic cancer, it significantly elevates LPA concentrations in the conditioned medium (Fig. 2B and C). ELISA assays further confirmed a dose-dependent rise in LPA levels in conditioned media of pancreatic cancer cells treated with gemcitabine (Fig. 2D). Prior research has demonstrated that LATS1/2 kinases inhibit YAP1 activity by phosphorylating it, causing pYAP1 sequestration in the cytoplasm through binding to 14-3-3 proteins [14]. LPA, however, inhibits LATS1/2 through G12/13-coupled receptors, thus activating YAP1 via dephosphorylation and nuclear translocation [12]. Based on these findings, we investigated whether similar mechanisms are involved in PDAC.
Fig. 2
Gemcitabine induces LPA release and Hippo pathway inactivation in PDAC. A Hierarchical heatmap illustrating differences in lipid classes in conditioned media between the NC and GEM groups, as determined by UPLC-MS/MS. B Total lipid concentrations in conditioned media of the NC and GEM groups. C The concentrations of LPA in conditioned media of the NC and GEM groups. D The ELISA assays measuring LPA levels in conditioned media post-gemcitabine treatment. E LPA induces YAP1 dephosphorylation and suppresses ATX expression in a dose-dependent manner. F LPA enhances YAP1 mRNA expression in a dose-dependent manner. G Gemcitabine reduces YAP1 phosphorylation and downregulates ATX expression in a dose-dependent manner. H Gemcitabine promotes YAP1 mRNA expression in a dose-dependent manner. I LPA (10 μM) induces nuclear translocation of YAP1 in pancreatic cancer cells. J Gemcitabine (10 μM) facilitates nuclear translocation of YAP1 in pancreatic cancer cells. K HA130 (20 μM) inhibits YAP1 protein expression and increases ATX protein expression in pancreatic cancer cells. L HA130 (20 μM) significantly reduces YAP1 mRNA expression in pancreatic cancer cells. M HA130 (20 μM) decreases LPA concentration in conditioned media of pancreatic cancer cells. N The indirect co-culture model demonstrates that HA130 (10 μM) significantly impairs the repopulation ability of reporter cells. Dying cells were treated with DMSO, gemcitabine, HA130, or a combination of gemcitabine and HA130, and co-cultured with reporter cells for 48 h. All experiments were performed at least three times, and results are presented as mean ± standard deviation from three independent experiments. *P < 0.05, **P < 0.01, and ***P < 0.001
×
![]()
To explore the interaction between LPA and the Hippo signaling pathway in PDAC, we initially examined the expression of YAP1 and pYAP1 following LPA treatment. The results showed that LPA significantly reduced YAP1 phosphorylation, promoted YAP1 overexpression, and inhibited ATX expression (Fig. 2E and F). Subsequently, we assessed whether gemcitabine exerts effects comparable to those of LPA. Our findings indicated that gemcitabine similarly suppressed YAP1 phosphorylation, enhanced YAP1 expression, and reduced ATX protein expression (Fig. 2G and H). To investigate the effect of LPA and gemcitabine on YAP1 localization, cytoplasmic and nuclear proteins were isolated separately and analyzed by Western blot. Upon stimulation with LPA or gemcitabine, an increase in YAP1 nuclear localization was observed (Fig. 2I and J). Collectively, these results suggest that gemcitabine induces dying cells to release LPA in a dose-dependent manner, leading to diminished YAP1 phosphorylation, increased YAP1 nuclear translocation, and consequent inactivation of the Hippo pathway.
We also investigated the effects of LPA inhibition on the Hippo pathway. Western blot and RT-qPCR analyses revealed that HA130, an ATX activity inhibitor, significantly reduced YAP1 expression while simultaneously increasing ATX protein expression (Fig. 2K and L). ELISA assay results confirmed that HA130 markedly decreased LPA concentrations in the conditioned medium of pancreatic cancer cells (Fig. 2M). These findings suggest that HA130 inhibits YAP1 expression and enhances ATX protein level via negative feedback mechanisms.
We subsequently investigated the role of LPA in the repopulation of residual tumor cells driven by dying cells. Colony formation and transwell assays showed that LPA significantly enhanced the clonogenic and migratory capacities of pancreatic cancer cells, whereas HA130 significantly suppressed these effects (Fig. S1A-D). The CCK-8 assays revealed that reporter cells in the gemcitabine-treated group exhibited significantly enhanced proliferation compared to the control group, while no significant difference was observed between the HA130-treated and control groups (Fig. 2N). Notably, HA130 effectively reversed the pro-proliferative effects induced by dying cells following chemotherapy. These findings highlight that LPA inhibition serves as an effective strategy to suppress the proliferative potential of residual tumor cells.
In summary, our results demonstrate that gemcitabine induces a dose-dependent release of LPA from dying cells. This release leads to Hippo signaling inactivation, which subsequently enhances the colony formation, migration, and proliferative capacities of residual pancreatic cancer cells.
2.3 LPA promotes the repopulation of residual tumor cells through YAP1
To investigate the regulatory role of YAP1 in PDAC, we initially analyzed transcriptome data from the TCGA database. Compared to adjacent tissues, YAP1 mRNA levels were markedly elevated in PDAC (Fig. 3A). Kaplan-Meier survival analysis of TCGA-PDAC and GSE57495 cohorts revealed that patients with elevated YAP1 expression exhibited significantly poorer overall survival (Fig. 3B). Additionally, Pearson correlation analysis revealed a negative correlation between YAP1 and ENPP2 (the gene encoding ATX, R = -0.15, p = 0.044, Fig. 3C). Western blot and RT-qPCR results showed that YAP1 knockdown led to a pronounced increase in ATX expression (Fig. 3D and E). We further established stable pancreatic cancer cell lines with YAP1 overexpression via lentiviral transfection (Fig. 3F and G). Western blot results revealed that YAP1 overexpression significantly suppressed ATX expression.
Fig. 3
LPA promotes the repopulation of residual tumor cells through YAP1. A The differential YAP1 expression between pancreatic cancer and adjacent tissues. B The comparison of overall survival between PDAC patients with high and low YAP1 expression based on TCGA-PDAC and GSE57495 cohorts. C Pearson correlation analysis showing the negative relationship between ENPP2 and YAP1 expression. D Western blot was performed to detect the changes of ATX expression following YAP1 knockdown in pancreatic cancer. E RT-qPCR validation of YAP1 knockdown at the mRNA level. F Western blot was performed to detect the changes of ATX expression following YAP1 overexpression in pancreatic cancer. Stable YAP1-overexpressing pancreatic cancer cell lines were generated using lentiviral transfection. G RT-qPCR validation of YAP1 overexpression at the mRNA level. H ELISA results showing that YAP1 knockdown reduces LPA concentration in conditioned media of pancreatic cancer cells. I ELISA results showing that YAP1 overexpression increases LPA concentration in conditioned media of pancreatic cancer cells. J The indirect co-culture model illustrating that YAP1 knockdown significantly inhibits the repopulation ability of pancreatic cancer reporter cells. Reporter cells with YAP1 knockdown were co-cultured with dying cells treated with DMSO or gemcitabine for 48 h. All experiments were performed in biological triplicates, and data are presented as mean ± standard deviation from three independent experiments. Kaplan-Meier analysis and log-rank test were performed to evaluate overall survival differences between distinct subgroups. Statistical significance was determined as *P < 0.05, **P < 0.01, and ***P < 0.001
×
![]()
Anzeige
ELISA assays were performed to explore whether YAP1 regulates the levels of LPA. The results showed that YAP1 knockdown significantly reduced LPA concentrations in the conditioned medium, whereas YAP1 overexpression resulted in increased LPA levels (Fig. 3H and I). These findings revealed that YAP1 knockdown inhibits ATX activity, thereby suppressing LPA production. Conversely, YAP1 overexpression enhanced LPA synthesis and release, with LPA exerting a negative feedback effect on ATX expression. We also conducted transwell and colony formation assays to evaluate the oncogenic function of YAP1. As expected, YAP1 overexpression promoted both colony formation and cell migration of pancreatic cancer (Fig. S1E-H). Finally, we investigated the role of YAP1 in tumor repopulation through indirect co-culture, and the results showed that after knockdown of YAP1, the difference in proliferation rate between gemcitabine-treated group and control group was nearly negligible (Fig. 3J). In summary, these findings suggest that chemotherapy-induced dying cells release LPA to facilitate YAP1 nuclear translocation. This, in turn, further amplifies LPA release, establishing a positive feedback loop between LPA and YAP1.
2.4 HA130 enhances gemcitabine-induced apoptosis and DNA damage
Gemcitabine primarily functions by incorporating into DNA strands, thereby inhibiting DNA replication and inducing cell death [15]. Therefore, Gene set enrichment analysis (GSEA) was performed to further examine the effect of YAP1 on DNA replication-associated processes. The results indicated that YAP1 notably modulates DNA repair, G2M checkpoint, DNA damage checkpoint, and cell cycle DNA replication (Fig. 4A). These findings suggest that the combination of gemcitabine and HA130 could serve as a promising therapeutic strategy.
Fig. 4
HA130 enhances gemcitabine-induced apoptosis and DNA damage. A Gene set enrichment analysis (GSEA) showing that YAP1 activation enhances pathways related to DNA repair, G2/M checkpoint, DNA damage checkpoint, and DNA replication in pancreatic cancer. B Highest single agent (HSA) synergy score matrix assessing the synergistic effect of gemcitabine and HA130 in AsPC-1 cells. The X-axis represents gemcitabine concentrations, the Y-axis represents HA130 concentrations, and synergy scores are displayed as matrix values. C Dose-effect matrix for the combination of gemcitabine and HA130 in AsPC-1 cells, where gemcitabine concentration is on the X-axis, HA130 concentration is on the Y-axis, and inhibition rates are represented as matrix values. D HSA synergy score matrix evaluating the synergistic effect of gemcitabine combined with HA130 in PATU8988T cells. The X-axis represents gemcitabine concentrations, the Y-axis represents HA130 concentrations, and synergy scores are shown as matrix values. E Dose-effect matrix for gemcitabine and HA130 combination in PATU8988T cells, with gemcitabine concentration on the X-axis, HA130 concentration on the Y-axis, and inhibition rates shown as matrix values. F Combined treatment of gemcitabine with si-YAP1 or HA130 enhances apoptosis in pancreatic cancer cells. Cells in quadrants q2 and q4 are identified as apoptotic cells. G The differences of apoptosis ratio in different treatment groups. H The combination of gemcitabine with si-YAP1 or HA130 promotes DNA damage in pancreatic cancer cells, as assessed by western blot analysis of γH2A.X with GAPDH as the internal control. All experiments were performed in triplicate, and data are presented as mean ± standard deviation from three independent experiments. Statistical significance was determined as *P < 0.05, **P < 0.01, and ***P < 0.001
×
![]()
Based on the Synergyfinder database (https://synergyfinder.org/#!/dashboard), we evaluated potential synergy or antagonism between the two drug combinations. The CCK-8 assays were performed to measure absorbance per well, allowing us to calculate dose-response matrices for each drug concentration and corresponding cell viability [16]. A synergy score below 0 indicates drug antagonism, a score between 0 and 10 suggests additive effects, and a score above 10 indicates synergy between the drug combinations. The results revealed synergy scores of 21.83 and 13.16 for the HA130 and gemcitabine combination in AsPC-1 (Fig. 4B and C) and PATU8988T (Fig. 4D and E), respectively, demonstrating a pronounced synergistic effect between the two drugs.
Anzeige
To explore the synergistic induction of apoptosis in pancreatic cancer by combination therapy, we assessed the effects of inhibiting LPA-YAP1 signaling using flow cytometry. Our study revealed that the combination of gemcitabine with HA130 or si-YAP1 significantly increased the proportion of apoptotic cells compared to monotherapy groups (Fig. 4F and G). Additionally, Western blot analysis confirmed that combination treatment markedly enhanced the capacity of gemcitabine to induce DNA damage in pancreatic cancer (Fig. 4H). These findings suggest that targeting the LPA-YAP1 signaling may be an effective strategy to improve chemotherapy efficacy in pancreatic cancer.
2.5 Single-cell sequencing reveals YAP1 may regulate chemokines and cholesterol
Single-cell RNA sequencing analysis (scRNA-seq) of the GSE212966 dataset was employed to investigate the impact of YAP1 on the tumor microenvironment (TME) of pancreatic cancer. The quality control process adhered strictly to the standard protocol outlined in the Materials and Methods section (Fig. S2A-J). Using the “SingleR” package, we identified 13 cell types, including fibroblasts, macrophages, endothelial cells, epithelial cells (tumor cells), NK cells, T cells, monocytes, smooth muscle cells, neurons, and other cell types (Fig. 5A). The single sample GSEA results indicated significant variations in the activity of Hippo signaling across different cell types (Fig. 5B). Consistent with expectations, the median level of Hippo signaling in pancreatic cancer is lower than that observed in adjacent tissues (Fig. 5C). Additionally, YAP1 expression level is significantly higher in pancreatic cancer compared to ENPP2 (Fig. 5D and E). The uniform manifold approximation and projection (UMAP) distribution plot indicated that cell types with high YAP1 expression generally did not express ENPP2, and vice versa (Fig. 5F). To gain a deeper understanding of Hippo pathway-related genes, we also depicted their UMAP distribution (Fig. S3) and analyzed the expression differences between adjacent tissues and pancreatic cancer (Fig. S4).
Fig. 5
Single-cell sequencing reveals YAP1 may regulate multiple chemokines. A The UMAP distribution of different cell types annotated using the “SingleR” package. B The activity of Hippo signaling varies across different cell types. The wilcoxon test was used to compare differences in Hippo signaling activity among different cell types at the single-cell level. C The violin Plot showing differences in Hippo signaling between adjacent tissues (CO) and pancreatic cancer (PC). D, E The expression levels of YAP1 and ENPP2 across different subgroups. F The UMAP distribution diagram showing the expression of ENPP2 and YAP1 across various cell types. G The proportion of epidermal cells expressing high versus low levels of YAP1 across different samples. H Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Ontology (GO) functional enrichment analysis of differentially expressed genes (DEGs) between epidermal cells with high versus low YAP1 expression. I RT-qPCR results indicate that YAP1 knockdown inhibits IL6 mRNA expression, while YAP1 overexpression promotes it. J RT-qPCR results show that YAP1 knockdown reduces IL8 mRNA expression, while YAP1 overexpression enhances it. K, L The ELISA assays confirm that YAP1 positively regulates IL6 and IL8 levels. M Pathway differences identified between epidermal cells with high- and low-YAP1 expression by single-sample gene set enrichment analysis. N GSEA analysis based on the TCGA-PDAC cohort confirms that YAP1 promotes cholesterol homeostasis. O RT-qPCR results showing that YAP1 knockdown decreases SREBP2 mRNA expression, while YAP1 overexpression increases its expression. All experiments were performed in triplicate, and data are presented as mean ± standard deviation from three independent experiments. Statistical significance was determined as *P < 0.05, **P < 0.01, ***P < 0.001 and ****P < 0.0001
×
![]()
We further examined the proportion of epithelial cells (tumor cells) expressing high and low levels of YAP1. The result indicated a higher proportion of epithelial cells with elevated YAP1 levels in pancreatic cancer (Fig. 5G). Differential gene analysis was conducted on epithelial cells based on YAP1 expression levels (Table s1), followed by KEGG and GO enrichment analyses. These analyses revealed that YAP1 primarily regulates pathways involved in chemokine signaling, TNF signaling, and the extracellular matrix (Fig. 5H). Several studies have reported that interleukin 6 (IL6) and interleukin 8 (IL8) are regulated by different signaling axes mediated by YAP1, which enhance the malignancy of various tumors [17, 18]. Pearson correlation analysis of the TCGA-PDAC cohort revealed that YAP1 also demonstrated positive correlations with IL8 and IL6, with correlation coefficients of 0.61 and 0.23, respectively (Fig. S2K and L). RT-qPCR results demonstrated that YAP1 knockdown significantly inhibited IL6 and IL8 expression, whereas YAP1 overexpression notably promoted their upregulation (Fig. 5I and J). ELISA assays further confirmed the positive regulatory effect of YAP1 on IL6 and IL8 (Fig. 5K and L). These findings suggest that gemcitabine-induced overexpression of YAP1 may elevate the levels of several critical chemokines, thereby fostering an inflammatory TME in pancreatic cancer.
Anzeige
The “AUCell” and “UCell” algorithms were used to perform GSEA at the single-cell level. The results indicated that YAP1 significantly enhanced cholesterol homeostasis (Fig. 5M), which was corroborated by GSEA conducted on the TCGA-PDAC cohort (Fig. 5N). Numerous studies have emphasized the role of YAP1 in cholesterol synthesis across various types of tumors [19‐21]. The accumulation of cholesterol and oxysterols inactivates the sterol regulatory element-binding protein-2 (SREBP2) pathway by promoting the retention of the SREBP Cleavage-Activating Protein-SREBP2 complex in the endoplasmic reticulum, which in turn downregulates cholesterol biosynthesis and uptake [19]. RT-qPCR results demonstrated that YAP1 overexpression significantly upregulated SREBP2 expression, while YAP1 knockdown resulted in a decrease in SREBP2 levels (Fig. 5O). Given that cholesterol metabolism serves as a potential target for modulating cancer-related processes [20, 21], these findings imply that interventions in the Hippo signaling pathway may offer an effective strategy to improve the TME following chemotherapy in pancreatic cancer.
3 Discussion
PDAC is a highly malignant tumor of the digestive system, projected to become the second leading cause of cancer-related death in the United States by 2030 [22]. Patients with pancreatic cancer are often diagnosed at an advanced stage, resulting in a grim prognosis. Gemcitabine is widely used as the first-line chemotherapy for PDAC [23, 24]. However, its response rate remains remarkably low, with most patients rapidly developing resistance to the treatment. Emerging evidence suggests that metabolic reprogramming constitutes a novel mechanism of chemotherapy resistance [25]. Lipid metabolism reprogramming has been recognized as a hallmark of malignancies. Tumor-associated alterations in lipid metabolism primarily encompass increased lipogenesis, enhanced lipid uptake from the extracellular microenvironment, and elevated lipid storage and mobilization within intracellular lipid droplets [26]. Tumor repopulation, characterized by the accelerated growth of residual tumor cells following chemoradiotherapy, also contributes to resistance against these treatments [27, 28]. Both lipid metabolism and tumor repopulation are critical contributors in chemotherapy failure; however, the underlying correlations and mechanisms remain to be fully elucidated. In this study, direct and indirect co-culture models were used to mimic the in vivo environment, demonstrating that gemcitabine-induced dying cells can promote the regeneration of residual pancreatic cancer cells through the release of LPA.
Ample evidence indicates that abnormal lipid metabolism-encompassing membrane formation, lipid synthesis and degradation, and lipid-driven signaling plays a crucial role in various cancers [29]. Fatty acid synthase (FASN), a key lipogenic enzyme in de novo fatty acid synthesis, is involved in the regulation of plasma membrane structure and palmitoylation of proteins post-translationally [30]. In pancreatic cancer, FASN overexpression is associated with adverse reactions to gemcitabine, primarily due to the upregulation of pyruvate kinase M2 [31]. Transcriptionally, FASN is primarily regulated by the SREBP family, which is influenced by PI3K/AKT and MEK/ERK pathway [32]. Previous studies have highlighted that lipid not only plays a crucial role in the formation of biological membranes and energy metabolism, but also functions as second messengers regulating tumor progression [29, 33]. Specifically, phospholipases generate several second messengers, such as LPA and AA, which activate multiple carcinogenic pathways [34, 35]. Emerging evidence underscores the physiological role of LPA in regulating tumor progression, angiogenesis, and metastasis [36]. LPA is a bioactive phospholipid that engages with at least six receptors, namely LPAR1-6, each coupled to a specific G-protein [37]. LPA is produced by hydrolysis of LPC mediated by ATX (LPA synthetase). Our previous research has confirmed the critical role of ATX in driving the malignancy of pancreatic cancer [11]. It has been reported that ATX also serves as a biomarker for early diagnosis [38]. Consequently, we focused on the impact of gemcitabine on lipid metabolism, particularly investigating whether dysregulated LPA metabolism following chemotherapy contributes to drug resistance and repopulation in pancreatic cancer. As expected, while gemcitabine generally suppresses lipid metabolism in pancreatic cancer cells, it aberrantly induces a significant increase in LPA levels.
Hippo is a crucial anti-cancer pathway in vivo, wherein mammalian Sterile 20-like kinase 1/2 (MST1/2) and LATS1/2 are activated through various mechanisms [39]. By phosphorylating the transcription factors YAP1/TAZ, LATS1/2 facilitates the binding of YAP1 to 14-3-3 proteins in the cytoplasm, thereby exerting a cancer-inhibiting effect. When LATS1/2 activity decreases, YAP1/TAZ translocates to the nucleus in an unphosphorylated state, promoting tumor proliferation, invasion, and the maintenance of stem-like properties [40]. Our research group was the first to report that TAZ, an effector of the Hippo pathway, is highly expressed in pancreatic cancer and facilitates metastasis by enhancing epithelial-mesenchymal transition [41]. Based on existing evidence and mass spectrometry analysis, our study proposes the hypothesis that gemcitabine induces the release of LPA from dying pancreatic cancer cells, which subsequently regulates the nuclear translocation of YAP1 in residual tumor cells, thereby promoting tumor repopulation.
Mechanistic studies have revealed that both gemcitabine and LPA inhibit YAP1 phosphorylation in a dose-dependent manner, ultimately leading to the inactivation of the Hippo pathway. In the context of pancreatic cancer, we observed a negative correlation between ATX and YAP1. Both gemcitabine and LPA also enhanced YAP1 translocation and reduced ATX protein expression through a complex feedback mechanism. Additionally, YAP1 overexpression exhibited effects similar to those induced by gemcitabine and LPA. While several studies have reported that various factors can activate YAP1 to enhance the malignancy of solid tumors [18, 42], the relationship between YAP1 and post-chemotherapy repopulation in pancreatic cancer remains unclear. Our findings suggest the presence of an LPA-YAP1 positive feedback loop in the TME following chemotherapy (Fig. 6). The results from the indirect co-culture model further confirmed that YAP1 suppression in reporter cells abrogated the pro-proliferative effect of dying cells. Classical molecular biology experiments suggest that targeting the LPA-YAP1 axis may serve as a potential strategy to enhance chemotherapy efficacy.
Fig. 6
Mechanisms of post-chemotherapy repopulation in pancreatic cancer. GPCRs, G-protein-coupled receptors; TEAD, transcriptional enhanced associate domain
×
![]()
The upregulation of the ATX-LPA signaling in tumor cells can effectively induce lymphocyte entry into secondary lymphoid organs, indicating the complex regulatory role of LPA in the TME [43]. Therefore, the potential impact of the LPA-YAP1 signal on TME warrants further exploration. The results of single-cell sequencing analysis revealed that YAP1 might regulate the levels of various chemokines and cholesterol, highlighting a significant role of YAP1 in metabolic reprogramming. Furthermore, GO analysis suggests that YAP1 may be involved in T cell activation and immune receptor activity, implying that targeting LPA-YAP1 signaling could enhance the potential for immunotherapy response in pancreatic cancer patients.
We humbly acknowledge the limitations of our study. Primarily, our research focused solely on in vitro mechanistic studies and lacked in vivo validation. Currently, a stable mouse model for chemotherapy-induced repopulation has not yet been established. Furthermore, additional investigation is necessary to optimize the dosage and administration intervals of gemcitabine in vivo.
Altogether, we demonstrated that gemcitabine induces the release of LPA from dying pancreatic cancer cells, which in turn inhibits YAP1 phosphorylation and promotes its nuclear translocation. On the one hand, YAP1 overexpression facilitates the release of LPA, thereby forming a positive feedback loop between LPA and YAP1. On the other hand, YAP1 elevates chemokines and cholesterol levels, fostering a TME that supports pancreatic cancer survival. These mechanisms underscore the potential of targeting the LPA-YAP1 signaling axis to overcome chemoresistance and reduce recurrence in pancreatic cancer.
4 Methods
4.1 Materials and reagents
The components and reagents used in this study were sourced from the following companies: RPMI 1640 culture medium (Roswell Park Memorial Institute, Gbico, USA), Penicillin-Streptomycin solution (100 ×) (Hangzhou Jino Company), 10 × PBS (Biyuntian Company), DMSO and Polybrene (Sigma), fetal bovine serum (FBS) (Hyclone, USA), 0.25% trypsin-EDTA, Lipofectamine 3000, and puromycin (Gibco), and D-luciferin substrate (Promega).
4.2 Cell culture
Human pancreatic carcinoma cell lines, PATU8988T and AsPc-1, were obtained from the American Type Culture Collection (Manassas, VA, USA). These cells were cultured at 37 °C in a humidified incubator with 5% CO2, using RPMI 1640 medium supplemented with 10% FBS. Short tandem repeat analysis was performed to verify the identities of the cell lines, all of which were confirmed to be free of mycoplasma contamination.
4.3 Tumor repopulation cell model
Pancreatic cancer cells were treated with 1 ~ 5 µM gemcitabine (Cat# HY-17026, MedChem Express, NJ, USA) for 48 to 72 h, washed twice with 1 × PBS, and the supernatant was discarded. The medium was then replaced with 2% serum, and the cells were incubated for an additional 24 h. The adherent dying cells were collected and referred to as the experimental group (gemcitabine treatment). The term “dying cell” broadly refers to cells undergoing programmed cell death, which can be characterized by morphological changes, DNA fragmentation, activation of caspase family proteases, and decreased metabolic activity [44]. Classical morphological features indicative of impending cell death, such as cell contraction, detachment from neighboring cells, and altered membrane integrity, can be observed under the microscope. The control group consisted of dying cells treated with DMSO for 48 to 72 h, with subsequent treatment procedures identical to those of the experimental group.
In the direct co-culture model, 1 × 105 dying cells were evenly seeded in 24-well plates. After attachment, 1000 GFP-labeled pancreatic cancer cells (reporter cells) were uniformly seeded on top. The medium containing 2% serum was replaced every 48 h. Images were captured every 48 h using a fluorescence microscope to monitor the proliferation activity of the reporter cells.
In the indirect co-culture model, 1 × 105 normal pancreatic cancer cells were uniformly seeded in 24-well plates. Once the cells adhered, transwell inserts (Cat# 725122, NEST, Wuxi, China) containing 3 × 105 dying cells with a 0.4 μm pore size were placed in each well, ensuring that the lower part of the transwell inserts was submerged in medium containing 2% serum. After 48 h, the adherent pancreatic cells (reporter cells) in the lower chamber were collected, and their proliferation activity was assessed using a CCK-8 assay (Cat# C0037, Beyotime, Shanghai, China).
4.4 Identification of differential lipids using mass spectrometry analysis
Lipidomic profiling of supernatants from pancreatic cancer cells was performed using liquid chromatography-tandem mass spectrometry (LC-MS/MS) in both positive and negative ion modes to detect a range of lipid molecules. The analysis was conducted in the following steps: (1) The samples underwent lipid extraction using the Bligh and Dyer method, ensuring integrity of lipid components. The extracted lipids were dried and dissolved in a dichloromethane/methanol mixture for subsequent analysis. Reverse-phase liquid chromatography was used for chromatographic separation, with an optimized gradient elution program to achieve effective categorization of lipid classes. (2) Full scan and data-dependent acquisition modes were utilized to obtain detailed lipid molecular and fragment information. The analysis system comprised ultra-performance liquid chromatography (UPLC, ExionLC™ AD and tandem mass spectrometry (MS/MS, QTRAP® 6500+ (https://sciex.com.cn/). Raw mass spectrometry data were preprocessed using mass spectrometry-data independent analysis software, including peak extraction, alignment, and noise reduction. (3) Quantification results were normalized using methods such as total ion intensity normalization and principal component analysis. Lipid molecules were subsequently identified and annotated based on the human metabolome database (https://hmdb.ca/). All detection and analyses were provided by Wuhan Metware Biotechnology Co., Ltd.
4.5 Flow cytometry for apoptosis detection
Pancreatic cancer cells (1 × 105 cells/mL) were allowed to grow for 24 h in a 6-well plate containing RPMI 1640 medium supplemented with 10% FBS. Subsequently, the cells were treated with gemcitabine (1 µM). After 48 h, the cells were harvested and washed with cold PBS buffer. They were then resuspended in a binding buffer and stained in the dark with 5 µL of Annexin V-FITC and 5 µL of propidium iodide (Cat# C1383S, Beyotime, Shanghai, China) at 4 °C for 30 min. The stained cells were washed three times with binding buffer to remove excess dye and then resuspended in 500 µL of binding buffer. Untreated cells served as negative controls. The percentage of apoptotic cells was analyzed within 1 h using flow cytometry (BD, FACSCalibur, USA).
4.6 CCK-8 assay for the detection of proliferation
Cell viability was assessed using the CCK-8 following the manufacturer’s instructions. Cells were seeded in a 96-well microplate at a density of 3 × 103 cells per well in 100 µL of culture medium. Every 24 h, 10 µL of CCK-8 reagent was added to each well, followed by a 1-hour incubation. The absorbance was measured at 450 nm using a microplate reader (Bio-Rad, Hercules, CA, USA).
4.7 Transwell migration assay
Complete medium (1 mL) was added to the lower chamber of the transwell invasion system. Subsequently, pancreatic cancer cells were resuspended in serum-free medium and seeded into the upper chamber at a density of 5 × 104 cells per well. After 24 h of incubation, we quantitatively assessed the number of cancer cells that traversed the membrane using ImageJ software. The images were converted to grayscale, and threshold adjustments were made to separate the cells from the background. The analyze particle function was used to calculate the number of migrated cells. We repeated the experiments multiple times to reduce variability and calculated the average.
4.8 Colony formation assay
Reporter cells were resuspended in RPMI 1640 medium containing 10% FBS and incubated at 37 °C in a 5% CO2 environment for 7 to 14 days. A total of 3000 cells were seeded per well to facilitate colony formation. After incubation, the plates were washed with cold PBS, and the colonies were fixed with 4% paraformaldehyde (Cat# P0099-3 L, Beyotime, Shanghai, China) at room temperature. The fixed colonies were then stained with 1% crystal violet for 30 min at room temperature. Colonies containing more than 100 cells were counted using ImageJ software.
4.9 Lentivirus transduction
1 × 106 pancreatic cancer cells were inoculated in the six-well plate, and the medium was discarded after 24 h when the cell fusion reached about 80%. Next, 1 mL of viral culture medium was mixed with 1 mL of fresh culture medium containing Polybrene (final concentration 1:1000) and added to the cells for continued cultivation. After 24 h of infection, GFP expression in the tumor cells was assessed under a fluorescence microscope. Cells exhibiting uniform GFP brightness were collected and transferred to 10 cm culture dishes for continued cultivation.
To generate stable cells expressing the exogenous gene GFP-Luc, puromycin (1 µg/mL) was added 24 h post-infection for selection. The culture medium was replaced every 3 days, with fresh puromycin added for selection over 2 weeks. This process yielded double-labeled cells, namely AsPC-1/GFP-Luc and PATU8988T/GFP-Luc. The expression of GFP in the cells was accessed and photographed under a fluorescence microscope. The labeled cells were expanded, and corresponding cell lines were stored for future use.
4.10 Western blot
Pancreatic cells were collected by centrifugation and washed twice with cold PBS. The cell pellets were suspended in 100 µL of cold RIPA lysis buffer and incubated on ice for 30 min, with vortexing every 10 min. Cytoplasmic and nuclear protein was separated and extracted using the NE-PER™ nuclear and cytoplasmic extract reagent kit (Cat# 78833, Thermo Fisher, MA, USA). The samples were then centrifuged at 15,000 g for 30 min. Immunoblot analysis was performed using 10 µg of sample proteins on a 12% SDS-PAGE gel. Electrophoresis was conducted at 120 V for 60 min to separate the proteins, followed by transfer to a PVDF membrane in cold transfer buffer using a wet transfer system at 300 mA for 90 min.
The PVDF membrane was blocked for 1 h and then incubated with primary antibodies (diluted 1:1000) overnight at 4 °C. After washing, the membrane was incubated with secondary antibodies (diluted 1:4000) at room temperature for 1.5 h. Protein bands were visualized using an ECL protein blot detection system (Tanon 4200), and band intensities were analyzed using Quantity One software. The antibodies against YAP1 (Cat# 4912 S) and pYAP1 (Ser 397, Cat# 13619T) were obtained from Cell Signaling Technology, Inc. (Boston, USA). The antibody against ATX (Cat# abs116120) was purchased from Absin Bioscience Inc. (Shanghai, China). The antibodies against Lamin B (Cat# ab16048), γH2A.X (Ser 139, Cat# ab26350) and GAPDH (Cat# ab263962) were sourced from Abcam, Inc. (Cambridge, USA). Secondary antibodies against rabbit (Cat# BA1060) or mouse (Cat# BM2002) were purchased from Boster Biotechnology (Wuhan, China).
4.11 Real-time quantitative reverse transcription polymerase chain reaction
According to the manufacturer’s instructions, cDNA was generated using the RevertAid First Strand cDNA Synthesis Kit (Thermo Scientific). Reverse transcription quantitative polymerase chain reaction (RT-qPCR) was performed using PowerUp SYBR Green Master Mix (Thermo Fisher). β-actin was used as the internal control. The primer sequences are listed in Table s2.
4.12 ELISA
Following the manufacturer’s instructions, the levels of LPA in the conditioned medium were assessed using an ELISA kit from Bioo Scientific, Inc. (Cat# EH11121M). Additionally, the concentrations of IL6 and IL8 in the conditioned medium were determined using ELISA kits obtained from eBioscience Inc. (Cat# EH2IL6) and BioLegend Inc. (Cat# 431504), respectively.
4.13 Single-cell sequencing analysis and RNA-sequencing analysis
The pancreatic cancer scRNA-seq data, GSE212966 [45], was downloaded from the Gene Expression Omnibus database (GEO, https://www.ncbi.nlm.nih.gov/geo/). A total of 44 Hippo pathway-related genes were curated from the MSigDB database (https://www.gsea-msigdb.org/gsea/msigdb/human/search.jsp). The detailed processing of the scRNA-seq data was conducted as follows: (1) The scRNA-seq data was preprocessed using the “Seurat” package. The “PercentageFeatureSet” function was employed to determine the proportion of mitochondrial genes, followed by correlation analysis to investigate the relationship between sequencing depth and the proportion of mitochondrial genes or total intracellular sequences. (2) Each gene was required to be expressed in at least 3 cells. (3) Cells were included if they expressed between 300 and 5000 genes, had mitochondrial content below 20%, and the unique molecular identifier (UMI) count exceeding 1000. (4) After filtering, the dataset was normalized using the “LogNormalize” method to ensure consistency in gene expression values.
We analyzed a total of 38,981 cells from six scRNA-seq samples, of which 34,208 cells were retained after quality control (Table s3). As shown in Fig. S2A and B, there is a strong correlation between UMI count and mRNA (r = 0.85). However, no significant correlation was observed between mRNA counts and mitochondrial or ribosomal gene content. The top 2000 highly variable genes were identified for subsequent dimensionality reduction analysis (Fig. S2C). We used the principal component analysis (PCA) to estimate available dimensions and depicted the raw distribution of different samples. The result showed no substantial differences among all cells (Fig. S2D and E). The top 30 most unique principal components were selected for further investigation (Fig. S1F). After removing batch effects, we visualized the combined distribution of the CO (adjacent tissues) and PC (pancreatic cancer) samples (Fig. S2G and H). The gene expression patterns of the top 15 principal components were further examined (Fig. S1I). Guided by the clustering tree, we set the resolution parameter to 1, resulting in the identification of 34 distinct clusters. (Fig. S1J).
The RNA-sequencing expression data and clinical features of PDAC were downloaded from the Cancer Genome Atlas database (TCGA, https://portal.gdc.cancer.gov/). The RNA-sequencing expression profiles in the GSE57495 dataset were also retrieved from the GEO database [46]. The detailed processing of the microarray data was as follows: (1) Gene symbols were obtained by converting probe IDs. (2) Probes that corresponded to multiple genes were removed. (3) When multiple probes corresponded to a single gene, the average value of these probes was considered as the gene expression level.
4.14 Statistical analysis
Statistical analysis was conducted using R (version 4.2.3) and GraphPad Prism 7 (CA, USA). The wilcoxon test was used to compare differences in Hippo signaling activity among various cell types at the single-cell level. Pearson correlation analysis was employed to assess the linear relationships between different genes. Kaplan-Meier analysis and log-rank test were performed to evaluate survival differences between distinct subgroups. Student’s t-test and analysis of variance (ANOVA) were employed to compare quantitative data between groups. All experiments were performed using representative data in triplicate and independently repeated at least three times. Results are presented as the mean ± standard deviation from three or more experimental replicates. P < 0.05 was considered statistically significant.
Acknowledgements
We are grateful for the support of lipidomics analysis provided by Wuhan Metware Biotechnology Co., Ltd.
Declarations
Competing interests
The authors declare no competing interests.
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
