CXCL1-overexpressing cancer-associated fibroblasts stimulate hepatocellular carcinoma progression via neutrophil recruitment, NET formation, and immune suppression

Abstract

Background

Hepatocellular carcinoma (HCC) is a highly aggressive malignancy with a prognosis often limited by an immunosuppressive tumor microenvironment (TME). While cancer-associated fibroblasts (CAFs) are known to influence tumor progression, the specific mechanisms by which they modulate immune cell recruitment and function remain poorly defined. This study investigates how CAF-derived CXCL1 orchestrates a pro-tumorigenic stroma through neutrophil manipulation.

Methods

The study employed in vitro co-culture systems and in vivo murine models to evaluate the impact of CXCL1-overexpressing CAFs on HCC progression. The signaling mechanisms were assessed using CXCR2 and STAT3 inhibitors, while immune surveillance was monitored via CD8+ T cell activity and natural killer (NK) cell infiltration assays.

Result

CXCL1 secreted by CAFs was found to activate neutrophils via the CXCR2-STAT3 signaling axis. This activation induces the formation of neutrophil extracellular traps (NETs), which directly facilitate tumor proliferation and metastasis. Furthermore, CXCL1-driven NETs established a potent immunosuppressive TME by significantly impairing CD8+ T cell activity and reducing NK cell infiltration. Conversely, neutralizing CXCL1 or inhibiting the CXCR2-STAT3 pathway successfully suppressed NET formation, restored anti-tumor immune cell functionality, and mitigated tumor growth.

Conclusions

The CXCL1-CXCR2-STAT3 axis is a critical driver of stromal-immune interactions in HCC. By mediating neutrophil recruitment and NET-driven immune evasion, CAFs create a permissive environment for tumor escalation. Targeting this signaling pathway represents a promising therapeutic strategy to disrupt pro-tumorigenic neutrophil mechanisms and reinvigorate the host’s anti-tumor immunity.

Introduction

Hepatocellular carcinoma (HCC) is the most prevalent form of primary liver cancer and remains a major clinical challenge due to its high mortality and limited treatment efficacy [6]. Despite advances in diagnosis and therapy, the prognosis for patients with HCC remains poor, largely because of its high metastatic potential and resistance to conventional treatments [7, 8]. Increasing evidence indicates that the tumor microenvironment (TME) plays a central role in disease progression, with cancer-associated fibroblasts (CAFs) acting as key regulators of tumor-stroma interactions [1, 14, 24, 50]. CAFs, a heterogeneous population of activated fibroblasts, secrete a repertoire of cytokines and chemokines that stimulate oncogenic processes, including tumor growth, invasion, and immune evasion [20]. Among these, CXCL1, a potent chemokine, has garnered significant attention for its multifaceted role in modulating tumor-stroma interactions [44].

CXCL1, a member of the CXC chemokine family, functions at the interface of inflammation and tumorigenesis [10, 15, 16]. By binding to its receptor CXCR2, CXCL1 mediates the recruitment of neutrophils and other immune cells [32]. Beyond its canonical inflammatory role, CXCL1 promotes epithelial-to-mesenchymal transition (EMT), enhances tumor invasiveness, and fosters metastasis in multiple malignancies [18, 28, 51]. Elevated CXCL1 expression has been correlated with poor clinical outcomes, highlighting its role as a prognostic parameter and therapeutic target in cancer [16, 18]. Nonetheless, its specific contribution to the interplay between CAFs and immune cells within the TME of HCC remains inadequately characterized.

Neutrophils, the most abundant circulating leukocytes, can either suppress or promote tumor progression depending on their polarization state [9, 31, 42]. In HCC, neutrophils are frequently co-opted by the TME to adopt a pro-tumorigenic phenotype, thereby facilitating tumor progression [26, 46]. A key mechanism underlying this effect is the formation of neutrophil extracellular traps (NETs)—web-like chromatin structures enriched in cytotoxic enzymes that enhance tumor adhesion, invasion, and immune evasion [36, 48]. Recent studies have implicated the CXCL1-CXCR2 signaling axis in promoting NET formation, linking CXCL1 to the establishment of an immunosuppressive microenvironment [40, 41].

The immunosuppressive activities of CAFs and neutrophils profoundly influence immune surveillance, the process by which cytotoxic CD8+T cells and natural killer (NK) cells recognize and eliminate malignant cells [13]. Cytotoxic CD8 + T cells and natural killer (NK) cells are central to this process, exerting potent anti-tumor activity [5]. Within the HCC TME, CAF-derived factors inhibit the activation and cytotoxicity of these immune cells, thereby promoting immune evasion and metastasis [25, 47]. In parallel, NETs can physically trap and functionally inactivate cytotoxic lymphocytes [27], establishing a feed-forward loop of stromal inflammation, neutrophil activation, and immune suppression. These interdependent interactions between CXCL1, neutrophils, NETs, and immune surveillance mechanisms create a self-reinforcing loop that drives tumor progression and metastasis.

Despite growing recognition of the interactions among CAFs, CXCL1, neutrophils, and immune suppression in various malignancies, their integrated contribution to HCC progression remains insufficiently defined. Although CXCL1’s pro-tumorigenic functions have been demonstrated in colorectal and gastric cancers [10, 53], the mechanistic links between CAF-derived CXCL1, neutrophil recruitment, NET formation, and immune evasion in HCC warrant further investigation. Addressing these gaps is critical for identifying novel therapeutic strategies aimed at reprogramming the TME to restore immune competence.

Cancer-associated fibroblasts (CAFs) are a major stromal component of the tumor microenvironment and play a pivotal role in promoting hepatocellular carcinoma (HCC) progression through paracrine signaling, extracellular matrix remodeling, and modulation of immune responses. Among the soluble factors secreted by CAFs, C-X-C motif chemokine ligand 1 (CXCL1) has emerged as a key regulator of tumor inflammation and immune evasion. However, the specific mechanisms by which CAF-derived CXCL1 influences immune cell dynamics and neutrophil activity in HCC remain poorly understood. In this study, we sought to elucidate the functional significance of CXCL1-overexpressing CAFs in modulating the HCC microenvironment. We focused on how CXCL1 secretion affects neutrophil recruitment, neutrophil extracellular trap (NET) formation, and antitumor immune surveillance, particularly the cytotoxic activity of CD8 + T cells and natural killer (NK) cells. We first confirmed that CXCL1 is endogenously upregulated in CAFs isolated from HCC lesions compared with normal fibroblasts (NFs), consistent with previous clinical observations. Building upon this foundation, we established a CXCL1-overexpressing CAF model to mechanistically dissect how elevated CXCL1 signaling within CAFs drives neutrophil infiltration and immune suppression. Furthermore, we validated the physiological relevance of these findings using fibroblast-specific CXCL1 knockout mice, demonstrating that fibroblast-derived CXCL1 is essential for promoting tumor growth, NET formation, and immune evasion in vivo. Collectively, our work identifies CAF-derived CXCL1 as a critical mediator of HCC progression, linking stromal inflammation to immune dysfunction, and provides new insights into targeting the CXCL1—neutrophil—NET axis as a potential therapeutic strategy for liver cancer.

Material and methods

Cell culture

HCC cell lines HepG2 and MHCC97 and CAFs were purchased from the ATCC and cultured in DMEM (Gibco, Thermo Fisher Scientific) plus 10% fetal bovine serum (FBS, Gibco) and 1% penicillin/streptomycin (Gibco) at 37 °C in a 5% CO₂ incubator. Normal fibroblasts (NFs) were purchased from ScienCell Research Laboratories (Cat. #2630, Carlsbad, CA, USA) and maintained under the same conditions as CAFs in DMEM supplemented with 10% FBS and 1% penicillin—streptomycin. For all experiments, cells were used at 70–80% confluence and were authenticated using STR profiling.

Overexpression and knockdown of CXCL1 in CAFs

To overexpress CXCL1, CAFs were transfected with a CXCL1 overexpression plasmid (pCDNA3.1-CXCL1) using Lipofectamine 3000 (Thermo Fisher Scientific). For knockdown experiments, CXCL1 expression was silenced by transfecting CAFs with 100 nm of CXCL1-specific siRNA (si-CXCL1, Thermo Fisher Scientific) or a negative control siRNA (si-NC). After 48 h, successful overexpression and knockdown were confirmed by qPCR and Western blotting.

Co-culture system

To investigate the interplay between CAFs and HCC cells, HepG2 and MHCC97 cells were co-cultured with either wild-type CAFs or CXCL1-overexpressing CAFs in a Transwell co-culture system (Corning, USA). HepG2/MHCC97 cells were seeded in the upper chamber at 1 × 10^5 cells per well, and the lower chamber contained DMEM supplemented with 10% FBS. The co-culture was maintained for 48 h, after which the cells were harvested for downstream assays.

Colony formation assay

HepG2 and MHCC97 cells were plated at 5 × 10^3 cells/well in 6-well plates. After 24 h, the cells were treated with or without cisplatin (5 μM) for 24 h, and then cultured in fresh medium for 7 days to allow colony formation. The colonies were fixed with 4% paraformaldehyde (PFA) for 20 min and stained with 0.5% crystal violet for 10 min. Colonies were counted using ImageJ software.

EdU assay

Cell proliferation was examined utilizing the Click-iT™ EdU Alexa Fluor 488 Imaging Kit (Thermo Fisher Scientific). HepG2 and MHCC97 cells were seeded on glass coverslips and treated with cisplatin (5 μM) for 24 h. Cells were incubated with 10 μM EdU for 2 h, fixed with 4% PFA for 15 min, and then stained using the Click-iT™ EdU Alexa Fluor 488 Reagent. Proliferating cells were analyzed by fluorescence microscopy, and the percentage of EdU-positive cells was calculated.

Quantitative PCR (qPCR)

RNA was extracted with TRIzol reagent (Invitrogen, Thermo Fisher Scientific), and cDNA synthesis was performed using HiScript III RT SuperMix (Vazyme, Nanjing, China). qPCR was implemented using SYBR Green Master Mix (Vazyme) on an ABI 7500 Real-Time PCR System. Gene expression levels were normalized to GAPDH, and the relative expression was gauged using the 2^ − ΔΔCt method. Primer sequences used for qPCR are as follows: CXCL1 forward: 5ʹ-TGGCAAGGAGAGTCTGAGA-3ʹ, reverse: 5ʹ-TGCAGGTTGGCAGGTATG-3ʹ; GAPDH forward: 5ʹ-TGACTTCAACAGCGACACCCA-3ʹ, reverse: 5ʹ-CACCCTGTTGCTGTAGCCAAA-3ʹ. The full list of cytokine primers used in the study is available upon request.

Western blotting (WB)

Cell lysates were prepared using RIPA buffer (Beyotime, China), and protein concentration was quantitated using the BCA kit (Beyotime). Equal protein amounts (30 µg) were separated by SDS-PAGE and moved onto PVDF membranes. After blocking with 5% BSA, membranes were probed with primary antibodies: anti-CXCL1 (Santa Cruz Biotechnology), anti-GAPDH (Cell Signaling Technology), and other antibodies as indicated. Membranes were incubated with HRP-conjugated secondary antibodies and developed using the ECL detection system (Thermo Fisher Scientific).

Transwell migration assay

Cell migration was evaluated using Transwell chambers (Corning, USA) with an 8 μm pore size. HepG2 and MHCC97 cells (1 × 10^5 cells) were seeded in the upper chamber, and 10% FBS in DMEM was placed in the lower chamber. After 24 h, the non-migrated cells were eliminated from the upper surface of the membrane, and the migrated cells on the lower surface were stained with crystal violet. The number of migrated cells was counted in five random fields under a microscope.

ELISA

The level of 8-OHdG, a marker of DNA oxidative damage, was quantified in the supernatants of HepG2 and MHCC97 cells using an ELISA kit (Elabscience, USA). Briefly, 50 μL of cell culture supernatant was added to the wells, and absorbance was measured at 450 nm using a microplate reader (BioTek Instruments, USA).

Flow cytometry

Flow cytometric analysis was made to test the proportion of T cells, NK cells, and neutrophils in mouse tissues. After harvesting spleen and lymph node tissues, single-cell suspensions were prepared and incubated with fluorochrome-conjugated antibodies against CD8, CD4, GZMB, CTLA4, Ly6G, and other markers for 30 min at 4 °C. The cells were analyzed in a BD FACSCanto II flow cytometer (BD Biosciences), and data were processed with FlowJo software.

Animal experiments

All animal studies were ratified by the Institutional Animal Care and Use Committee (IACUC) at University of Electronic Science and Technology of China (approval no. 2024012213). C57BL/6 mice (6–8 weeks old, male) from Charles River Laboratories were housed in specific pathogen-free conditions with a 12 h light/dark cycle, at a controlled temperature of 22 °C. Mice were provided with ad libitum access to water and a standard rodent chow diet. All procedures followed the institutional and national guidelines for animal research. For lung metastasis assays, Hepa1–6 cells (5 × 10^5), either alone or co-cultured with wild-type or CXCL1-overexpressing cancer-associated fibroblasts (CAFs) at a 3:1 ratio, were injected into the tail vein of mice (n = 6–8 per group). Lung tissues were collected after 8 weeks, and metastatic nodules were counted. For liver cancer models, CXCL1-overexpressing mice (CXCL1-cKI) and fibroblast-specific CXCL1 knockout mice (Cxcl1-Col1a1-ko, generated by crossing Col1a1-Cre mice with Cxcl1-fl/fl mice) were used. To induce HCC, mice were injected intraperitoneally with 100 mg/kg diethylnitrosamine (DEN) at week 1, followed by carbon tetrachloride (CCl4, 1.5 μL/g body weight, twice weekly) for 8 weeks. Mice were sacrificed at week 12, and liver tissues were collected to assess tumor burden. For immune modulation studies, anti-CD8 antibody (10 mg/kg) or anti-Ly6G antibody (10 mg/kg) was administered intraperitoneally twice weekly to neutralize CD8 T cells or deplete neutrophils, respectively. Lung and liver tissues were fixed in 4% paraformaldehyde for histopathological evaluation, including Masson’s trichrome staining for collagen deposition and immunohistochemistry (IHC) to detect markers such as Ki67, Cleaved-Cas-3, FSP, CD8+T cells, NCR1+NK cells, PD-L1, MPO, and H3cit. Immunofluorescence (IF) was performed to visualize neutrophil extracellular traps (NETs) using H3cit and MPO markers. Immune cell populations, including CD8+T cells, GZMB+CD8+T cells, CTLA4+CD8+T cells, NK cells, and neutrophils, were analyzed by flow cytometry (FACS) using fluorochrome-conjugated antibodies after processing spleens and lymph nodes into single-cell suspensions.

IHC

Formalin-fixed, paraffin-embedded lung and liver tissues were sectioned into 4-µm slices and deparaffinized. Antigen retrieval was achieved by heating slides in citrate buffer (pH 6.0) at 95 °C for 20 min. Sections were blocked with 5% BSA and incubated with primary antibodies: Ki67, Cleaved-Cas-3, FSP, CD8, NCR1, PD-L1, MPO, and others as indicated. After incubation with HRP-conjugated secondary antibodies, sections were visualized with DAB substrate and counterstained with hematoxylin. Images were acquired using a light microscope, and positive cells were quantified.

Immunofluorescence (IF)

Frozen tissue Sects. (10 μm thick) were prepared from tumor tissues, fixed with 4% PFA, and permeabilized with 0.3% Triton X-100. After blocking with 5% BSA, sections were probed with primary antibodies against H3cit, MPO, STAT3, and other markers overnight at 4 °C. After washing, secondary antibodies conjugated with fluorochromes were applied, and the slides were mounted with DAPI-containing medium. Fluorescence was observed using a confocal microscope (Leica), and fluorescence intensity was measured using ImageJ software.

Statistical analysis

Data are delineated as mean ± standard deviation (SD). Normality of data distribution was assessed using the Shapiro–Wilk test where feasible before selecting parametric tests. Statistical analyses were made in GraphPad Prism 9.0. Pairwise comparisons were done using the Student's t-test, and multi-group comparisons were done using one-way ANOVA followed by Tukey's test. A p-value of < 0.05 implied statistical significance.

Results

High CXCL1 expression in CAFs enhances HCC growth

Initially, we co-cultured CAFs with HCC cell lines HepG2 and MHCC97. Consistent with previous reports, CAFs significantly promoted HCC cell proliferation and conferred resistance to chemotherapy (Fig. 1A, B). CXCL1 secretion was higher in CAFs than in normal fibroblasts (NFs) (Fig. 1C). Previous studies have highlighted that high CXCL1 in CAFs promotes EMT in colon cancer cells [43]. To further investigate the impact of CXCL1-expressing CAFs on HCC, we overexpressed CXCL1 in CAFs (Fig. 1D, E) and co-cultured them with HCC cells. We then introduced an anti-CXCL1 neutralizing antibody (aCL1) into the culture system. Our results demonstrated that the co-culture of HCC cells with CXCL1-overexpressing CAFs visibly promoted both tumor growth and metastasis (Fig. 1F–H). Moreover, this co-culture further exacerbated the resistance of HCC cells to cisplatin. Given that cisplatin primarily exerts its anti-cancer effects by inducing DNA damage, we assessed DNA repair in HCC cells. Notably, co-culturing with CXCL1-overexpressing CAFs significantly reduced γH2AX signal following cisplatin treatment (Fig. 1I, J), suggesting that CXCL1-CAFs protect HCC cells from cisplatin-induced DNA damage. We acknowledge that γH2AX reflects double-strand DNA lesions rather than repair activity. These findings indicate a general pro-survival effect of CXCL1-CAFs under genotoxic stress but do not define a chemoresistance mechanism, which remains outside the primary scope of this study. Besides, an effect that was largely abrogated by aCL1 treatment (Fig. 1F–J). These findings suggest that CXCL1 overexpression in CAFs can promote HCC cell growth in vitro and enhance DNA repair after cisplatin-induced DNA damage. To explore whether tumor cells could modulate CXCL1 expression in CAFs, we co-cultured CAFs with murine hepatic stellate cells (HSCs) or Hepa1-6 hepatocellular carcinoma cells. As shown in Fig. S1A, B, co-culture with Hepa1-6 cells markedly increased both CXCL1 mRNA and secreted protein levels in CAFs compared with CAFs alone or with HSCs. These data suggest that tumor-derived cues can further enhance CXCL1 production by CAFs, reinforcing the CXCL1-dependent stromal activation observed in vivo.

Fig. 1.

CXCL1-overexpressing CAFs promote HCC cell growth and activity. A HepG2 and MHCC97 cells were co-cultured with CAFs, and colony formation assays were performed to assess the number of colonies formed by HCC cells with or without cisplatin treatment. B EdU staining was used to evaluate cell viability of liver cancer cells with or without cisplatin treatment. C qPCR analysis of the mRNA expression levels of secreted cytokines TGF-β, CCL2, CXCL1, HGF, VEGF, PDGF, FGF2, IGF1, IGF2, WNT3A, WNT5A, CCL3, CCL4, CXCL2, CXCL5, CXCL9, and CXCL10 in normal fibroblasts and cancer-associated fibroblasts (CAFs). D–E CXCL1 was overexpressed in CAFs, and qPCR and WB were performed to measure mRNA and protein expression levels of CXCL1 in CAFs. F HepG2 and MHCC97 cells were co-cultured with CXCL1-overexpressing CAFs and treated with or without anti-CXCL1 neutralizing antibody. Colony formation assays were performed to assess the number of colonies formed by HCC cells in the presence or absence of cisplatin. G EdU staining was used to assess cell viability. H Transwell assays were conducted to evaluate the migration capacity of HCC cells. I ELISA was performed to measure the levels of 8-OHdG in HCC cells with or without cisplatin treatment. J Immunofluorescence (IF) was used to detect the fluorescence intensity of γH2AX. Cell experiments were repeated 3–5 times, and statistical analysis was performed using ANOVA, followed by Tukey’s multiple comparison test (P < 0.05 was considered statistically significant)

CXCL1-overexpressing CAFs enhance tumor progression by suppressing T cell activity and reducing NK cell numbers

To confirm that CAFs modulate HCC progression via CXCL1, we injected Hepa1-6 cells co-cultured with CXCL1-overexpressing CAFs (3:1 ratio) into mice via tail vein injection. After 8 weeks, we collected lung tissue for metastatic nodule analysis. The number of metastatic nodules in mice injected with CXCL1-overexpressing CAFs was significantly higher (Fig. 2A). Furthermore, collagen deposition was markedly increased (Fig. 2B), while Ki67-positive cells greatly reduced (Fig. 2C), and C-caspase-3 positivity was notably elevated (Fig. 2D), suggesting that CXCL1-expressing CAFs promote tumor growth in vivo. Previous studies have shown that myeloid-derived suppressor cells (MDSCs) promote T cell exhaustion through CXCL1 secretion, thereby facilitating gastric cancer progression. Thus, we next analyzed the tumor-infiltrating CD8+and CD4+T cell populations. In the tumors co-injected with CXCL1-overexpressing CAFs, CD8+T cells notably reduced, whereas CD4+T cells remained largely unchanged (Fig. 2E, F). Additionally, analysis of GZMB+or CTLA+CD8+T cells revealed that CXCL1-overexpressing CAFs substantially suppressed the cytotoxic activity of T cells (Fig. 2G, H). Interestingly, NK cell numbers were also significantly reduced in the co-injection group with CXCL1-overexpressing CAFs and Hepa1-6 cells (Fig. 2I). To further confirm that CXCL1 in fibroblasts promotes the growth of hepatocellular carcinoma cells, we employed fibroblast-specific CXCL1 knockout mice (Cxcl1-Col1a1-ko). Hepa1-6 cells were injected via the tail vein into either Cxcl1-Col1a1-ko or Cxcl1^fl/fl control mice. We observed that tumor growth of Hepa1-6 cells was markedly attenuated in Cxcl1-Col1a1-ko mice (Fig. S2A). Moreover, neutrophil activation was significantly reduced in the tumor tissues (Fig. S2B), while the numbers of CD8+T cells and NK cells were notably increased (Fig. S2C–D).

Fig. 2.

CXCL1-overexpressing CAFs promote tumor progression by inhibiting T cell function and reducing NK cell numbers. A Hepa1-6 cells were injected into mice via tail vein, co-cultured with CXCL1-overexpressing CAFs at a 3:1 ratio. After 8 weeks, lung tissues were collected to analyze the number of metastatic nodules. B MTS staining was used to assess collagen deposition in lung metastatic nodules. C–D Immunohistochemistry (IHC) was performed to evaluate the staining intensity of Ki67 and Cleaved-Cas-3 in lung metastatic lesions. E–F Immunofluorescence analysis of the number of CD4+ or CD8+ positive cells in lung metastatic lesions. G–H Flow cytometry was used to measure the percentage of GZMB+CD8+ or CTLA4+CD8+ T cells in lymph node tissues from mice. I IHC was performed to analyze the staining intensity of NCR1 in lung metastatic lesions. Each group contained 6–8 mice, and statistical analysis was performed using ANOVA, followed by Tukey’s multiple comparison test (P < 0.05 was considered statistically significant)

CXCL1-overexpressing CAFs induce neutrophil infiltration and NET formation, promoting tumor growth

CXCL1, a multifaceted cytokine, is known to recruit neutrophils. We assessed neutrophil infiltration in tumor tissues formed by co-culturing CXCL1-overexpressing CAFs with Hepa1-6 cells. CD11b + Ly6G + double-positive cells significantly increased in tumor tissues (Fig. 3A). Recent studies have implied the CXCL1-CXCR2 axis in NET formation [40, 41]. In line with these findings, NET formation and activated neutrophils were notably elevated in the tumors co-cultured with CXCL1-overexpressing CAFs and Hepa1-6 cells (Fig. 3B–C). To further study the role of neutrophils in mediating the effects of CXCL1-CAFs on tumor growth, we administered anti-Ly6G to deplete neutrophils in mice four weeks after injection of CXCL1-overexpressing CAFs and Hepa1-6 cells. Anti-Ly6G treatment significantly reduced Ki67 positivity (Fig. 3D) and increased C-caspase-3 staining (Fig. 3E). Furthermore, neutrophil depletion also reduced PD-L1 expression, suggesting a reduction in tumor immune evasion (Fig. 3F). It is noteworthy that preliminary assays confirmed minimal metastatic formation in mice injected with Hepa1-6 cells alone, indicating that the enhanced tumor burden observed here arises primarily from CAFs-dependent microenvironmental activation. To further determine whether the reduction of NETs after anti-Ly6G treatment was associated with restoration of T-cell function, we analyzed tumor-infiltrating CD8+T-cell subsets by flow cytometry. The proportions of GZMB+CD8+and IL-2+CD8+effector T cells were significantly increased, while PD-1+CD8+and TIM-3+CD8+exhausted T cells were markedly decreased following neutrophil depletion (Fig. 3G-H). These findings indicate that the inhibition of NETs alleviates T-cell exhaustion and reactivates cytotoxic responses, thereby contributing to reduced tumor growth rather than being a mere consequence of smaller tumor size.

Fig. 3.

CXCL1-overexpressing CAFs enhance neutrophil infiltration and NET formation, thereby promoting HCC progression in vivo. A FACS analysis of the number of CD11b+Ly6G+double-positive cells in tumor tissues. B Immunofluorescence analysis of the fluorescence intensity of H3cit in tumor tissues. C IHC analysis of MPO staining intensity in lung metastatic lesions. D After 4 weeks of CXCL1-overexpressing CAFs and Hepa1-6 cell injection, anti-Ly6G was used to deplete neutrophils in mice. After 8 weeks, mice were sacrificed, and lung tissues were analyzed for the number of metastatic nodules. E–F IHC analysis of Ki67 and Cleaved-Cas-3 staining intensity in lung metastatic lesions. G IHC analysis of PD-L1 staining intensity in lung metastatic lesions. H, I FCS analysis of GZMB+, IL-2, PD-1 and TIM3 positive CD8 T cells. Each group contained 6–8 mice, and statistical analysis was performed using ANOVA, followed by Tukey’s multiple comparison test (P < 0.05 was considered statistically significant)

Conditional deletion of CXCL1 in fibroblasts suppresses hepatocellular carcinoma growth by enhancing CD8+T cell activity and reducing neutrophil activation

Building on our previous findings that CXCL1-overexpressing CAFs promote HCC growth in vivo, we generated Col1a1-Cre;Cxcl1-fl/fl mice to conditionally knock out CXCL1 in fibroblasts. After tail vein injection of Hepa1-6 cells into Cxcl1-Col1a1-ko and Cxcl1-fl/fl mice, a significant reduction was noted in lung metastatic nodules in the Cxcl1-Col1a1-ko mice (Fig. 4A). Additionally, the intensity of FSP staining, a marker of activated CAFs, was reduced (Fig. 4B), and collagen deposition in lung metastatic lesions was markedly diminished as indicated by Masson’s trichrome staining (Fig. 4C). Furthermore, Ki67 staining intensity was significantly lower, while C-caspase-3 expression was notably higher in Cxcl1-Col1a1-ko mice (Fig. 4D–E). Importantly, depletion of CXCL1 in CAFs increased the number of CD8 + and NK cells in metastatic lesions (Fig. 4F–G), while neutrophil infiltration, activation, and NET formation were significantly suppressed (Fig. 4H–J).

Fig. 4.

Conditional knockout of CXCL1 in fibroblasts inhibits HCC progression by promoting CD8+ T cell activity and reducing neutrophil activation. A Col1a1-Cre and fibroblast-specific Cxcl1-fl/fl mice were generated to create Cxcl1-Col1a1-ko mice with fibroblast-specific CXCL1 knockout. Hepa1-6 cells were injected into Cxcl1-Col1a1-ko and Cxcl1-fl/fl mice via tail vein, and after 8 weeks, lung tissues were collected to analyze the number of metastatic nodules. B IHC analysis of the staining intensity of activated CAFs marker FSP in lung metastatic lesions. C MTS staining to assess collagen deposition in lung metastatic nodules. D–E IHC analysis of Ki67 and Cleaved-Cas-3 staining intensity in lung metastatic lesions. F–G IHC analysis of CD8 + or NCR+ cells in lung metastatic lesions. H FACS analysis of CD11b+Ly6G+ double-positive cells in tumor tissues. I Immunofluorescence analysis of the fluorescence intensity of H3cit in tumor tissues. J IHC analysis of MPO staining intensity in lung metastatic lesions. Each group contained 6–8 mice, and statistical analysis was performed using ANOVA, followed by Tukey’s multiple comparison test (P < 0.05 was considered statistically significant)

Depletion of NK or CD8 + T cells impairs the efficacy of CXCL1 deletion in suppressing HCC progression

Although depletion of CD8 + or NK cells reduced their respective populations, it promoted the formation of metastatic nodules (Fig. 5A–C). However, the activation of CAFs and collagen deposition in metastatic lesions were not significantly altered (Fig. 5D–E). In contrast, CD8 or NK cell depletion resulted in significantly increased Ki67 positivity and decreased C-caspase-3 staining in metastatic nodules (Fig. 5F–G). Interestingly, neither CD8 + nor NK cell depletion had a substantial impact on neutrophil infiltration or activation (Fig. 5H–J).

Fig. 5.

Depletion of NK or CD8+ T cells impairs the efficacy of CXCL1 deletion in suppressing HCC progression. A Cxcl1-Col1a1-ko mice were treated with IgG, anti-CD8, or anti-NK1.1 antibodies at week 4. At week 8, lung tissues were analyzed for the number of metastatic nodules formed by Hepa1-6 cells. B–C IHC analysis of CD8+ or NCR + cells in lung metastatic lesions. D IHC analysis of the staining intensity of activated CAFs marker FSP in lung metastatic lesions. E MTS staining to assess collagen deposition in lung metastatic nodules. F–G IHC analysis of Ki67 and Cleaved-Cas-3 staining intensity in lung metastatic lesions. H FACS analysis of CD11b+Ly6G + double-positive cells in tumor tissues. I Immunofluorescence analysis of the fluorescence intensity of H3cit in tumor tissues. J IHC analysis of MPO staining intensity in lung metastatic lesions. Each group contained 6–8 mice, and statistical analysis was performed using ANOVA, followed by Tukey’s multiple comparison test (P < 0.05 was considered statistically significant)

CXCL1 overexpression promotes tumorigenesis in vivo

To evaluate the consequence of stromal CXCL1 overexpression, we used CXCL1 knock-in (cKI) mice in a DEN-CCl4 primary HCC model. To further explore the impact of CXCL1 overexpression on tumor progression, we generated CXCL1 overexpressing mice (cKI) and subjected them to a DEN-CCl4-induced HCC model. Tumor numbers in the liver were significantly higher in cKI mice (Fig. 6A). Additionally, activated FSP and collagen deposition in metastatic lesions were markedly increased (Fig. 6B–C). Ki67 staining intensity was elevated, accompanied by a reduction in C-caspase-3 expression (Fig. 6D–E). Moreover, CD8+ and NK cell numbers were significantly decreased in cKI mice (Fig. 6F–G), and NET formation was notably exacerbated (Fig. 6H–J).

Fig. 6.

CXCL1 overexpression promotes tumorigenesis in vivo. A CXCL1 overexpression mice (cKI) were injected with DEN at week 1 and treated with CCl4 for 8 weeks. Tumor numbers in liver tissues were analyzed at week 12. B IHC analysis of the staining intensity of activated CAFs marker FSP in lung metastatic lesions. C MTS staining to assess collagen deposition in lung metastatic nodules. D–E IHC analysis of Ki67 and Cleaved-Cas-3 staining intensity in lung metastatic lesions. F–G IHC analysis of the number of CD8+ or NCR + cells in lung metastatic lesions. H FACS analysis of CD11b+ Ly6G + double-positive cells in tumor tissues. I Immunofluorescence analysis of the fluorescence intensity of H3cit in tumor tissues. J IHC analysis of MPO staining intensity in lung metastatic lesions. Each group contained 6–8 mice, and statistical analysis was performed using ANOVA, followed by Tukey’s multiple comparison test (P < 0.05 was considered statistically significant)

CXCL1-CXCR2-STAT3 axis mediates neutrophil recruitment and activation in HCC progression

Previous studies have demonstrated that CXCL1 recruits macrophages into the TME via CXCR2. Given this, we explored the potential ligand-receptor interaction between CXCL1 and CXCR2 in neutrophils. Moreover, CXCL1-CXCR2-STAT3 signaling has been implicated in neutrophil activation. In the absence of CXCL1, MPO+STAT3+double-positive cells significantly reduced but markedly increased in cKI mice (Fig. 7A). To investigate the role of STAT3 in mediating the CXCL1-CXCR2 axis in neutrophil activation and HCC progression, we treated cKI and wild-type (wt) mice with the STAT3-specific inhibitor NSC-74859. Treatment with NSC-74859 significantly reduced tumor numbers in both wt and cKI mice (Fig. 7B). Furthermore, NSC-74859 treatment did not significantly alter CAF activation, as indicated by FSP and aSMA staining (Fig. 7C–E), but it significantly reduced NET formation and exhibited a higher therapeutic efficacy in cKI mice compared to wt mice (Fig. 7F–H). Additionally, NSC-74859 treatment reduced Ki67 positivity and increased C-caspase-3 expression (Fig. 7I, J), while increasing CD8 + and NK cell numbers in tumor tissues (Fig. 7K, L).

Fig. 7.

CAFs recruit and activate neutrophils through CXCL1-CXCR2-STAT3 signaling. A Immunofluorescence analysis of MPO+ STAT3 + double-positive cells in tumor tissues from CXCL1 knockout or CXCL1 cKI mice. B CXCL1-overexpressing mice (cKI) were injected with DEN at week 1 and treated with CCl4 for 8 weeks. From week 6, mice were treated with 5 mg/kg NSC-74859 until the endpoint. Tumor numbers in liver tissues were analyzed at week 12. C–D IHC analysis of the staining intensity of activated CAFs marker FSP and alpha-SMA in lung metastatic lesions. E MTS staining to assess collagen deposition in lung metastatic nodules. F FACS analysis of CD11b + Ly6G + double-positive cells in tumor tissues. G Immunofluorescence analysis of the fluorescence intensity of H3cit in tumor tissues. H IHC analysis of MPO staining intensity in lung metastatic lesions. I–J IHC analysis of Ki67 and Cleaved-Cas-3 staining intensity in lung metastatic lesions. K–L IHC analysis of the number of CD8 + or NCR + cells in lung metastatic lesions. Each group contained 6–8 mice, and statistical analysis was performed using ANOVA, followed by Tukey’s multiple comparison test (P < 0.05 was considered statistically significant)

Discussion

This study identifies fibroblast-derived CXCL1 as a central mediator of the immunosuppressive microenvironment that drives hepatocellular carcinoma (HCC) progression. CXCL1 secreted by cancer-associated fibroblasts (CAFs) activates neutrophils through the CXCR2—STAT3 signaling axis, inducing the formation of neutrophil extracellular traps (NETs) that promote tumor growth and metastasis. In parallel, CXCL1-driven NETs compromise immune surveillance by suppressing CD8⁺ T-cell cytotoxicity and diminishing NK-cell infiltration. The complementary gain-and loss-of-function approaches used here—including CXCL1-overexpressing CAFs and fibroblast-specific CXCL1 knockout (Cxcl1-Col1a1-ko) mice—establish both the sufficiency and necessity of fibroblast-derived CXCL1 in orchestrating stromal and immune crosstalk within the HCC microenvironment.

CAFs are well recognized as key effectors in the tumor microenvironment, promoting oncogenesis through paracrine secretion of cytokines and chemokines [29, 39]. CXCL1 produced by CAFs facilitates EMT and metastasis in colorectal cancer [22, 43] and correlates with inflammatory remodeling and poor outcomes in breast and gastric cancers [17, 37]. Our data extend these findings to HCC, showing that CXCL1-overexpressing CAFs enhance tumor-cell proliferation and migration and recruit neutrophils via CXCL1—CXCR2 signaling, consistent with observations in lung cancer [23]. Moreover, hepatoma cells further upregulate CXCL1 in CAFs, supporting a bidirectional tumor-stroma feedback loop that amplifies CXCL1 signaling and promotes a collagen-rich, immunosuppressive niche. Within this context, activation of the CXCR2STAT3 pathway in neutrophils drives NET formation and metastatic potential. The observed protection of HCC cells from cisplatin-induced damage reflects a general pro-survival phenotype of CXCL1-CAFs rather than a defined chemoresistance mechanism, which lies beyond the main scope of this study.

NET formation is a hallmark of neutrophil activation and a potent driver of tumor dissemination [33]. NETs enhance tumor-cell adhesion, shield malignant cells from cytotoxic lymphocytes, and facilitate metastatic colonization [2, 38]. While intrinsic drivers of NETosis, such as hypoxia and reactive oxygen species (ROS), have been extensively documented [3], our study uniquely highlights the extrinsic influence of stromal-derived CXCL1 in promoting this phenomenon. Although anti-Ly6G administration reduced neutrophil numbers and NETs, previous reports indicate incomplete depletion in BALB/c mice. Future studies will employ CXCR2 inhibitors or conditional genetic ablation to verify that CXCL1 promotes tumor progression primarily through CXCR2-dependent neutrophil activation. By establishing the CXCL1CXCR2STAT3 cascade as a mechanistic link between CAFs and NET formation, this work positions neutrophil activation within a broader stromal regulatory framework and underscores the therapeutic rationale for targeting NETs in HCC.

The immunosuppressive consequences of NETs are equally significant. NETs physically trap cytotoxic lymphocytes and release arginase-1 and reactive oxygen species that induce T-cell exhaustion [19]. NETs physically ensnare cytotoxic lymphocytes, including CD8 + T cells and NK cells, and release immunosuppressive mediators such as arginase-1 and ROS, which induce T cell exhaustion [12, 35]. Consistent with these reports, CXCL1-induced NETs in our model suppressed CD8⁺ T-cell cytotoxicity and reduced NK-cell infiltration, thereby compromising immune surveillance. Similar mechanisms have been described in gastric cancer, where CAF-derived CXCL1 promotes MDSC-mediated T-cell suppression [21, 34]. Our results extend this paradigm by directly linking stromal CXCL1, neutrophil activation, and adaptive-immune dysfunction, demonstrating how fibroblast-driven inflammation shapes immune evasion.

The CXCL1-CXCR2-STAT3 axis has emerged as a pivotal pathway in multiple cancers [11]. In lung and esophageal tumors, pharmacologic CXCR2 blockade reduces neutrophil infiltration and tumor burden [4, 52]. In HCC, we similarly observed that CXCR2 or STAT3 inhibition not only curtailed neutrophil recruitment and NET formation but also restored CD8⁺ T—and NK-cell activity, revealing a dual mechanism of action that dampens inflammation while re-engaging immune surveillance. These results highlight the therapeutic promise of targeting the CXCL1—CXCR2STAT3 axis to simultaneously suppress pro-tumorigenic myeloid functions and reinvigorate antitumor immunity.

Beyond CXCL1, other CAF-derived chemokines—including CCL2 [45] and CXCL5 [49], may cooperate to orchestrate neutrophil recruitment and immune suppression. The intersection between NET-induced immunosuppression and immune-checkpoint signaling also remains poorly understood [30]. Future studies integrating CXCR2 blockade, NET-targeting agents, and immune-checkpoint inhibitors may provide synergistic therapeutic benefits.

Fibroblast-derived CXCL1 promotes HCC progression by recruiting and activating neutrophils through the CXCR2-STAT3 pathway, inducing NET formation, and suppressing immune surveillance (Fig. 8). These findings establish the CXCL1-CXCR2-STAT3 axis as a central regulator of stromal—immune interactions in HCC and a compelling therapeutic target. While CAF heterogeneity and the precise CXCL1-producing subpopulations require clarification through single-cell transcriptomics and conditional genetic models, the integrated gain—and loss-of-function, pharmacologic, and immune-profiling data presented here provide robust mechanistic evidence that CXCL1-driven neutrophil activation constitutes a key driver of HCC aggressiveness.

Fig. 8.

CXCL1-overexpressing CAFs orchestrate a pro-tumorigenic TME by activating neutrophils and impairing immune surveillance via the CXCL1-CXCR2-STAT3 axis. Targeting this pathway represents a promising therapeutic strategy to disrupt NET-mediated tumor progression and restore anti-tumor immunity in HCC

Author contribution

Bin Bi: Conceptualizatio, Data curation, Project administration, Writingoriginal draft, Writingreview and editing. Liqin Tang: Resources, Project administration, Resources; Rritingreview and editing. Ranxi Liang: Data curation, Validation, Supervision. Haijing Wu: Data curation, Validation, Writingreview and editing. Qilan Huang: Data curation, Conceptualization, Project administration. Li Pan: Data curation, Validation, Writingreview and editing.

Funding

Not applicable.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Conflict of interest

The authors declare no conflict of interest.

Consent for publication

Not applicable.

Ethical approval

All animal experiments conducted in this study were performed in compliance with the ethical standards and regulations set forth by Institutional Animal Care and Use Committee (IACUC) of University of Electronic Science and Technology of China (approval no. 2024012213).