Human pancreatic cancer cell lines AsPC-1 and PANC-1 were obtained from the American Type Culture Collection (ATCC, Rockville, MD) and were maintained in 1640 and DMEM supplemented with 10% fetal bovine serum (FBS). The mouse pancreatic cancer KPC cells were maintained in DMEM supplemented with 10% FBS. All cell lines underwent regular testing to confirm their being free of mycoplasma contamination, using mycoplasma PCR detection (TaKaRa Taq Version 2.0, #R004).

Female C57BL/6 mice, aged 6 to 8 weeks, were utilized to create subcutaneous tumors of pancreatic cancer. KPC cells (1.8 × 10⁶ cells) were suspended in 100 μl PBS and injected subcutaneously into the flanks of the mice. Tumor volume was measured and calculated using the formula: volume = length × width² × 0.5. All animal experimental procedures adhered to the institutional guidelines and approved by the Animal Care and Use Committee of Sun Yat-sen University Cancer Center.

Total RNA was extracted from cultured cells or tumor tissues using the RNA-Quick Purification Kit (EZbioscience, B0004D). Complementary DNA (cDNA) was synthesized through reverse transcription of RNA with the Color Reverse Transcription Kit (A0010CGQ, Ezbioscience, USA). The relative levels of RNA were assessed by quantitative real-time PCR (qRT-PCR) utilizing the Bio-Rad detection system (Bio-Rad, Hercules, CA, USA). The expression levels of CXCL1 and NF-κB p65 were normalized by GAPDH using the comparative Ct method. The primers employed for qRT-PCR are as follows: CXCL1: AGCTTGCCTCAATCCTGCATCC, TCCTTCAGGAACAGCCACCAGT; NF-κB p65:ATGTGGAGATCATTGAGCAGC,CCTGGTCCTGTGTAGCCATT.

Cells were lysed in RIPA buffer containing protease and phosphatase inhibitors (Cat No. P0013B, Beyotime Biotechnology, Shanghai, China). Protein samples were electrophoresed on 10%-15% polyacrylamide gels and transferred to PVDF membrane. Membranes were first incubated in blocking buffer (containing 5% dry milk powder) for 1 h at 24°C and then incubated with primary antibodies overnight at 4°C followed by secondary antibody (24°C, 2 h). Membranes were washed with PBST. Protein bands were detected by chemiluminescence using an ECL detection kit (Tanon, #180-5001). The following antibodies were used in this study: anti-SQLE (Proteintech, #12544-1-AP, dilution 1:1000), anti-CXCL1 (Proteintech, #12335-1-AP, dilution 1:1000), anti-GAPDH (Cell Signaling Technology, #14C10, dilution 1:1000), anti-NF-κB p65 (Cell Signaling Technology, #8242, dilution 1:1000), and anti-Phospho-NF-κB p65 (Ser536) (Proteintech, #80379-2-RR, dilution 1:1000).

Fresh tumors were minced and incubated in RPMI 1640 medium containing 0.1 mg/mL collagenase IV, 0.02 mg/ml DNase I, with 5% FBS at 37°C for 1 hour. The samples were then filtered through a 40 μm cell strainer (Corning) to prepare single-cell suspensions. Cell surface markers were stained with fluorophore-conjugated antibodies in the dark at 4°C for 45 minutes in Cell Staining Buffer (#420201, Biolegend). Flow cytometric analysis was performed using a CytoFLEX flow cytometer (Beckman Coulter). Cells stained with Zombie UV™ dye (#423107, Biolegend, dilution 1:1000), and alive cells were defined as Zombie UV-negative cells. Antibodies used in these analyses included FITC anti-mouse CD3ϵ (#100306, Biolegend, dilution 1:20), Pacific Blue anti-mouse CD4 (#100428, Biolegend, dilution 1:20), PE/Cyanine7 anti-mouse CD8a (#100722, Biolegend, dilution 1:20), APC/Cyanine7 anti-mouse CD45 (#103116, Biolegend, dilution 1:20), APC anti-mouse/human CD11b (#101211, Biolegend, dilution 1:20), Brilliant Violet 510 anti-mouse Ly-6G (#127633, Biolegend, dilution 1:20), Brilliant Violet 650 anti-mouse F4/80 (#123149, Biolegend, dilution 1:20), PE anti-mouse CD206 (#141706, Biolegend, dilution 1:20), Brilliant Violet 786 anti-mouse CD80 (#740888, BD OptiBuild, dilution 1:20). Up to 5×10⁴ CD45+ cells were recorded for further analysis by CytExpert.

Mouse tumor tissues were harvested, fixed in 4% formalin overnight, and subsequently embedded in paraffin. The paraffin-embedded tissues were sliced into 4μm-thick sections, and the endogenous peroxidase activity was quenched by incubation with 3% H₂O₂ for 15 minutes at room temperature. For antigen retrieval, the samples were boiled in an antigen-retrieval buffer (pH 8.0) for 5 minutes. After blocking with 3% BSA for 1 hour, the samples were then incubated overnight with the primary antibody at 4°C, and after washing, incubated with an HRP-conjugated secondary antibody for 1 hour at room temperature. DAB solution was then applied to visualize staining. The antibodies used were F4/80 (Proteintech, #28463-1-AP, IHC: 1:4000), CD206 (Invitrogen, #MR5D3, IHC: 1:200), and CXCL1 (Proteintech, #12335-1-AP, IHC: 1:75).

PANC-1 cells were treated with or without squalene for various time points as indicated, and the samples were subjected to RNA-seq analysis (Shanghai Majorbio Bio-Pharm Technology Co., Ltd) to identify differentially expressed genes (DEGs) in cells with or without squalene treatment. KEGG pathway enrichment analysis was performed to determine the pathways enriched with all differentially expressed genes. Fisher test was conducted to assess significantly enriched gene functions and KEGG pathways associated with DEGs. Enrichment analyses of KEGG pathways for DEGs were performed using the Python Scipy software package (https://scipy.org/install/). Pathways exhibiting a corrected p-value of less than 0.05 were considered significantly enriched.

Genotype-Tissue Expression (GTEx) (https://xenabrowser.net) and the Cancer Genome Atlas (TCGA) (https://www.cancer.gov/ccg/research/genome-sequencing/tcga) databases were used to obtain RNA-seq data for pancreatic cancer tissues (n=178) and normal pancreatic tissues (n=171). SQLE expression levels in the normal and tumor tissues were compared using the Wilcoxon rank-sum test.

Statistical difference between two groups was assessed using the Student’s t-test. The relationship between two variables was evaluated using the Spearman’s correlation method. These statistical analyses were conducted with GraphPad Prism software (version 8.0, CA, USA). Gene expression levels in the normal and tumor tissues from public databases were compared using the Wilcoxon rank-sum test. A p-value of less than 0.05 was considered statistically significant. Data are expressed as mean ± S.D. as detailed in the figure legends.

To objectively evaluate the potential clinical relevance of SQLE expression, we first used public datasets from the GTEx and TCGA databases to compared the expression of SQLE in pancreatic cancer tissues (n=178) and normal tissues (n=171). Our analysis revealed that SQLE expression was significantly elevated in pancreatic tumor tissues (p<0.0001, Figure 1A ). We then established an immune-competent mouse model (C57BL/6) bearing mouse pancreatic cancer tumors (KPC cells) with or without stable SQLE-knockdown (SQLE-KD) to evaluate the potential impact of SQLE-mediated metabolism on pancreatic cancer growth and immune functions. As shown in Figures 1B, C , shRNA-mediated silencing of SQLE expression led to a significant decrease in tumor growth compared to the control group. Some of the KPC cells with SQLE-KD were unable to form tumor in mice, while the control KPC cells were highly tumorigenic ( Figure 1C ). Since pancreatic cancer is immunologically “cold” tumor often infiltrated with immunosuppressive cells including tumor-associated macrophages and MDSCs consisted of aberrant monocytes and immature neutrophils (19), we then evaluated the impact of SQLE expression in pancreatic cancer cells on various immune cells in the tumor tissues. Flow cytometry analysis revealed that silencing of SQLE caused a major decrease in CD11b+/Ly6G+ MDSCs among the CD45+ leukocytes in the tumor tissues ( Figure 1D ). Quantitative analysis of multiple tumor tissues showed that over 60% of CD45+ cells were CD11b+/Ly6G+ MDSCs in the tumors of the control group, whereas the MDSCs decrease to less than 20% in the SQLE-KD group ( Figure 1D , lower panel). There was also a significant decrease in CD11b+ macrophages in the SQLE-KD tumor tissues ( Figure 1E ). Interestingly, within the macrophage population, there was a significant decrease in CD206+ subpopulation (primarily consisting of M2 subtype) in the SQLE-KD tumor tissues ( Figure 1F ). Conversely, there was a substantial increase in CD8+ T cells in the SQLE-KD tumors (average 28%) compared to that in the shControl tumors (average 7%, Figure 1G ). These data together suggest that SQLE might play an important role in promoting an immunosuppressive phenotype, and that its knockdown could mitigate such immunosuppression.

Consistently, analysis of the relationship between tumor sizes and immune cell infiltrations revealed a positive correlation between tumor sizes and the degrees of infiltrations of MDSCs and TAMs in the tumor tissues ( Figures 2A–C ). In contrast, there was a negative correlation between tumor sizes and CD8+ T cell infiltration, with larger tumors having fewer CD8+ T cells in the tumor microenvironment ( Figure 2D ). Interestingly, these correlations were observed when the data from the tumors of both groups (shControl and SQLE-KD) were pooled and plotted on the same charts, suggesting the critical role of the infiltrated immune cells in affecting the overall tumor growth. Of note, the SQLE-KD tumor samples contained less immunosuppressive cells (red dots in Figures 2A–C ) and more cytotoxic CD8+ cells (red dots in Figure 2D ), compared to that in the shControl tumors (blue dots).

Since SQLE is the key enzyme that catalyzes the conversion of squalene to oxidosqualene (13), we postulated that suppression of SQLE expression would cause an accumulation of squalene due to a stagnation of the metabolic flow ( Figure 3A ). We indeed observed that a stable knockdown of SQLE in pancreatic cancer cells (AsPC-1) by shRNA led to an 8-folds increase in cellular squalene ( Figure 3B ). Thus, we tested the effect of squalene on tumor growth and immune cells by direct administration of squalene to immunocompetent C57BL/6 mice bearing pancreatic cancer tumors (KPC cells). The results showed that in vivo treatment with squalene was able to significantly retard tumor growth in the KPC syngeneic model ( Figure 3C ). Flow cytometry analysis of immune cells from the tumor tissues revealed that squalene could reduce the percentage of macrophages (CD11b+ F4/80+) in the CD45+ cell population ( Figure 3D ) and also decreased the proportion of M2 (CD206+) macrophages ( Figure 3E ), similar to that observed in the SQLE-KD experiments. Consistently, the degrees of TAM infiltration appeared correlated with the tumor sizes, which were substantially smaller in the squalene-treated mice ( Figure 3F , red dots) compared with the untreated mice (blue dots). Of note, the proportion of MDSCs (CD11b+ Ly6G+) in the CD45+ cell population in the tumor tissues showed a tendency of decrease after squalene treatment ( Figure 3G ), although such a reduction did not reach statistical significance likely due to the small sample size.

The ability of squalene to inhibit the infiltration of TAMs was further confirmed by immunohistochemical (IHC) staining of the tumor tissues from the control mice and squalene-treated mice, as evidenced by a substantial decrease in F4/80+ cells ( Figure 4A ) and CD206+ macrophages ( Figure 4B ) in the squalene-treated group. These data ( Figures 3 , 4 ) together showed that squalene was able to inhibit the infiltration of immunosuppressive cells into the tumor tissues, and suggest that it might likely mediate the immune-modulating effect observed in SQLE-KD tumor. Interestingly, we also observed that squalene treatment caused a substantial decrease in expression of CXCL1 (a chemokine known to recruit target cells such as MDSCs and TAMs (20–24)), in the tumor tissues ( Figure 4C ). These findings prompted us to further investigate the possibility that CXCL1 might be an important down-stream target of squalene as described below.

Since CXCL1 is a chemokine known to recruit target cells such as MDSCs and TAMs and contribute to an immunosuppressive microenvironment (20–25), we thus tested the possibility that squalene might affect CXCL1 expression and thus mediate the immune-modulating effect of SQLE. Quantitative RT-qPCR analysis showed that squalene could significantly inhibit the expression of CXCL1 mRNA in both AsPC-1 cells ( Figure 5A ) and PANC-1 cells ( Figure 5B ). Western blotting further demonstrated that squalene was able to substantially reduce CXCL1 protein the two pancreatic cell lines ( Figure 5C ). This Western blot analysis also revealed that treatment of pancreatic cells with squalene caused a decrease in the protein level of p65, a key component of the NF-κB signaling pathway known to regulate the expression of cytokines and chemokines including CXCL1 (26–28). The phosphorylation of p65 at S536 was also substantially reduced by squalene treatment ( Figure 5C ). These data suggest a possibility that squalene might regulate CXCL1 expression through the NF-κB pathway via affecting p65. Interestingly, RT-qPCR analysis showed that squalene did not cause any significant change in p65 mRNA level ( Figure 5D ), suggesting that squalene affected p65 at the post-transcriptional level.

Genetic silencing SQLE by shRNA, which caused squalene accumulation ( Figures 3A, B ), also led to an inhibition of CXCL1 expression consistently observed in both PANC-1 and AsPC-1 cells ( Figure 5E ). This was also associated with a decrease in p65 protein and a reduced phosphorylation at S536 ( Figure 5E ). These data were very similar to that observed with squalene treatment, and thus suggest that the effect of SQLE-KD on CXCL1 expression was likely mediated by the accumulation of its metabolic substrate (squalene). Of note, analysis of pancreatic cancer and normal tissue datasets from TCGA and GTEx databases revealed a positive correlation between SQLE and CXCL1 expression in clinical samples ( Figure 5F ), indicating that the regulation of CXCL1 by SQLE/Squalene likely occurred in vivo.

To further characterize the cellular processes in response to squalene treatment, we employed RNA-sequencing technology (RNA-Seq) to analyze the gene expression profiles in pancreatic cancer cells treated with or without squalene. PANC-1 cells were treated with 200 μM squalene for 6 and 24 h, and their gene expression profiles were characterized by RNA sequencing. Molecular analysis of the RNA-seq data using the Kyoto Encyclopedia of Genes and Genomes (KEGG) webtools revealed that squalene induced multiple changes in a variety of cellular processes, including metabolism, signal transduction, cellular processes, and other pathways ( Figure 6 ). The squalene-induced changes in gene expression at the early time point (6 h, Figure 6A ) and the next day (24 h, Figure 6B ) were similar. Interestingly, “signal transduction” and “immune system” are among the pathway categories with most numbers of genes whose expression was altered (indicated by red color in Figures 6A, B ). Genes related to “signal transduction” included FOS, RELB, SREBF1, MAP2K6, TLR2, TLR4, CSF2, CXCL1, CXCL8, CXCL3. Genes related to “immune system” were IFNE, LSP1, MAP2K6, TNFSF10, FCGR2A, CCR6, CCL20, CXCL1, CXCL8, CXCL3, and TLR4. Further characterization using KEGG pathway enrichment analysis showed that NF-κB signaling is the most prominent pathway with most changes induced by squalene ( Figure 7A ). The enriched genes in the NF-κB signaling pathways included TRAF1, RELB, GADD45A, TNFAIP3, TLR4, CXCL1, CXCL3 and CXCL8. These data were consistent with the Western blot data showing a decrease of p65 protein and inhibition of its phosphorylation by squalene ( Figure 5C ). Other pathways enriched in the squalene-treated cells included IL-17 signaling pathway, TNF signaling pathway, transcriptional misregulation, Toll-like receptor signaling, and neutrophil extracellular trap formation ( Figure 7A ), consistent with alterations of immune functions. Further reactome enrichment analyses revealed multiple pathways that were enriched in the squalene-treated cells ( Supplementary Figure S1A ). Among these enriched pathways, “signal transduction” pathways related to cytokine signaling and interleukin signaling in immune system were consistently identified (indicated by the red arrows in Supplementary Figure S1A ). The names of the specific genes enriched in the relevant pathways are provided in Supplementary Figure S1B .