Liver tissue samples were obtained from patients with pathologically confirmed HCC who had not received any anticancer therapy prior to sampling. Individuals with a concurrent autoimmune disease, HIV, or syphilis were excluded. Samples from 48 HCC patients who had complete follow-up data were included to assess overall survival (OS) and recurrence-free survival (RFS) using immunofluorescence staining of GPER1 and CD68. OS was defined as the time between surgery and either death or the last observation for patients who survived. RFS was defined as the time between surgery and either the first recurrence or death, or the last observation for patients without recurrence. The clinical characteristics of these patients are summarized in Supplementary Table S1 . In addition, fresh biopsy specimens from 7 HCC patients were used to isolate tumor-infiltrating leukocytes for flow cytometry analysis. All samples from HCC patients were coded anonymously in accordance with local ethical guidelines (as stipulated by the Declaration of Helsinki). Written informed consent was obtained from all participants prior to study onset. The use of human subjects for this study was approved by the Institutional Review Board of Sun Yat-sen University Cancer Center.

The hepatoma cell lines (HepG2, Huh7, SK-Hep-1and Hepa1-6) were obtained from the American Type Culture Collection (ATCC). All cells were regularly tested for mycoplasma contamination using the single-step polymerase chain reaction (PCR) method. The cells were cultured in DMEM (Gibco) supplemented with 10% FBS (Gibco), 100 U/mL penicillin (Sigma-Aldrich Corp), and 100 µg/mL streptomycin (Sigma-Aldrich Corp) in a humidified 5% CO2 incubator at 37°C. To prepare tumor culture supernatants (TSN), 5 × 10⁶ tumor cells were plated in 10 ml complete medium in 10-cm dishes for 24 hours. Then, the medium was changed to phenol red and serum-free DMEM. After 48 hours, the supernatant was collected, centrifuged, and stored in aliquots at -80°C (14).

Human monocytes were isolated as previously described (20, 21). Briefly, peripheral blood mononuclear cells (PBMCs) were isolated from the buffy coats derived from healthy donors’ blood by Ficoll density gradient centrifugation. CD14⁺ monocytes were then purified from PBMCs using magnetic beads (Miltenyi Biotec) according to the manufacturer’s instructions. The purified monocytes were cultured in DMEM supplemented with 10% human AB serum for 7 days to generate macrophages. Afterward, the culture medium was replaced with phenol red-free DMEM (Procell) containing 2% AB serum and 20% TSN to mimic a relatively nutrient-deficient tumor microenvironment. Meanwhile, in certain experiments as indicated, macrophages were treated with biochemical reagents, including G-1 (10008933, Cayman), G-15 (14673, Cayman) or 17β-estradiol (E8875, Sigma-Aldrich).

All animal experiments were approved by the Institutional Animal Care and Use Committee of Sun Yat-sen University (Guangzhou, China). Male C57BL/6J mice were purchased from the Guangdong Medical Laboratory Animal Center and maintained under specific pathogen-free conditions. The mice used in the experiments were between 6 and 8 weeks old. For the subcutaneous tumor model, a total of 1×10⁶ Hepa1-6 cells were subcutaneously transplanted into the flanks of mice. For the orthotopic tumor model, a total of 1×10⁶ Hepa1-6 cells were suspended in 25 µl of 50% basement membrane extract (354234, corning) and injected into the left lobe of the liver of anesthetized mice. The mice began receiving daily subcutaneous injections of G1 (4mg/kg) dissolved in a solvent containing 10% DMSO (MP), 30% PEG300 (TargetMol), 5% Tween 80 and 55% ddH2O on day 4 (orthotopic tumor model) or day 5 (subcutaneous tumor model). The control mice received a 200 µl dose of the solvent.

Formalin-fixed and paraffin-embedded HCC samples were cut into 4-μm sections and processed for IHC as previously described (14). Following incubation with anti-human CD3 (MA514524, Thermo Fisher Scientific), anti-human CD20 (ab78237, Abcam), anti-human CD15 (ZM-0037, ZS), anti-human CD68 (ZM-0060, ZS), anti-human CD163 (ab182422, Abcam), anti-human CD204 (KAL-KT022) or anti-mouse F4/80 (70076S, CST) antibodies (Abs), the sections were then stained with the corresponding secondary Abs and visualized using 3,3’-diaminobenzidine (Nichirei). An automatic slide scanner (KF-PRO-020) was used to scan the sections, and then positive cells were quantified by HALO image analysis software (Indica Labs).

Formalin-fixed and paraffin-embedded HCC sections were processed as described previously (14). For human tumor sections, anti-human CD68, anti-human Ki67 (ZM-0167, ZS) and anti-human GPER1 (PA5-109319, Thermo Fisher Scientific) Abs were used. For mouse tumor sections, anti-mouse F4/80 (700767, CST) and anti-mouse Ki67 (ab15580, Abcam) Abs were used. Immunofluorescence signals were amplified by a tyramide signal amplification kit (PANOVUE) as instructed by the manufacturer for visualization. Sections were scanned using the Polaris Fully Automatic Digital Slide Scanner (Akoya Biosciences), and then positive cells were quantified using HALO image analysis software (Indica Labs).

For immunofluorescence staining of cultured cells, the cells growing on coverslips were fixed with tissue/cell fixation buffer for 15 minutes at room temperature. Afterward, they were rinsed with PBS and permeabilized and blocked with PBS containing 5% BSA and 0.3% Triton X-100 for 1 hour at room temperature. The cells were then incubated overnight at 4°C with the primary antibody against human Ki67. This was followed by exposure to Alexa Flour 488-conjugated anti-mouse IgG. For double staining of Ki67 and GPER1 or ERα, the cells were simultaneously incubated with primary Abs against human Ki67 and GPER1 (PA5-109319, Thermo Fisher Scientific), or ERα (ab16660, Abcam). They were then exposed to Alexa Flour 488-conjugated anti-rabbit IgG and Alexa Flour 555-conjugated anti-mouse IgG. Nuclei were counterstained with DAPI. The immunofluorescence staining images were visualized using a high-resolution confocal laser microscope (LSM880 with fast airyscan, Zeiss).

Human monocyte-derived macrophages were either left untreated or treated with 20% Huh7-TSN for 48 hours, in the presence or absence of G-1 (1 μM), and then the cells were cultured with EdU (5-ethynyl-2’-deoxyuridine) at a final concentration of 1 μM for 5 hours. Afterward, the cells were fixed, permeabilized, and dyed following the manufacturer’s instructions (C0071S, Beyotime). Images were then visualized using an inverted fluorescence microscope (Nikon ECLIPS Ti2).

Tumor-infiltrating leukocytes were obtained from fresh tissue samples as described previously (20). Briefly, fresh biopsy specimens from HCC patients were cut into small pieces and digested in RPMI 1640 medium supplemented with 0.002% DNase I (Roche), 0.05% collagenase IV (Sigma-Aldrich) and 10% FBS for 45 minutes at 37°C. The dissociated cells were passed through a 70-μm cell strainer and then erythrocytes were lysed and removed. The remaining cells were thoroughly rinsed and resuspended in PBS supplemented with 1% FBS (Gibco) for flow cytometry analysis.

Flow cytometry was performed as previously described (20). Before antibody staining, the cells were incubated with the Zombie Fixable Viability reagent for 15 minutes at room temperature. For surface staining, the cells were stained with fluorochrome-conjugated Abs for 30 minutes. For intracellular staining, the cells were fixed using the Fix/Perm solution (eBioscience), washed with the Perm/Wash buffer (eBioscience), and then incubated with fluorochrome-conjugated Abs for 30 minutes. For GPER1 staining, cells were incubated with GPER1 antibody before being stained with Alexa Flour 488-conjugated anti-rabbit IgG. Data was acquired using Cytoflex flow cytometer (Beckman Coulter) and analyzed using CytExpert software. Representative plots were created using Flowjo software 10 (Tree Star). The reagents and Abs used for flow cytometry are listed as follows: Zombie NIR™ (423105, Biolegend), Zombie Violet™ (423113, Biolegend); anti-human Abs including CD45-PE (304008, Biolegend), CD14-AF700 (557923, Biolegend), Ki67-PE (556027, Biolegend), Ki67-APC (350514, Biolegend), PD-L1-PC7 (558017, Biolegend) and GPER1 (PA5-109319, Thermo Fisher Scientific); anti-mouse Abs including CD45-BV605 (103140, Biolegend), CD11b-FITC (101206, Biolegend), Ly-6G-ECD (562700, BD Biosciences), F4/80-PE (123110, Biolegend), PD-L1-PC7 (124314, Biolegend) and Ki67-APC (652405, Biolegend).

Human monocytes were seeded in 96-well plates with a density of 12,500 cells per well, and cultured in DMEM supplemented with 10% human AB serum for 7 days. Then the monocyte-derived macrophages were either left untreated or treated with 20% Huh7-TSN for 48 hours, in the presence or absence of G-1 (1 μM). Afterward, the cells were incubated with the CCK-8 (CK-04, KYD bio) solution for 2 hours, and absorbance was measured at 450 nm.

Total RNA was extracted using TRIzol reagent, and then 1μg RNA was used to synthesize cDNA with Color Reverse Transcription Kit (A0010CGQ, EZBioscience). Sequences of the primers used are listed as follows: ESR1 (ERα), Forward: 5′- GCTTACTGACCAACCTGGCAGA -3′, Reverse: 5′- GGATCTCTAGCCAGGCACATTC -3′; ESR2 (ERβ), Forward: 5′- AGAGTCCCTGGTGTGAAGCAAG, Reverse: 5′- GACAGCGCAGAAGTGAGCATC-3′; GPER1, Forward: 5′- TCTAAACTGCGGTCAGATGTGGC-3′, Reverse: 5′- TGTGAGGAGTGCAAGGTGACCAG-3′; AR, Forward: 5′- GACGACCAGATGGCTGTCATT-3′, Reverse: 5′- GGGCGAAGTAGAGCATCCT-3′. QPCR was performed in triplicate according to a standard protocol using Color SYBR Green qPCR Master Mix (A0012-R2, EZBioscience) with the LightCycler 480 System (Roche). To determine the levels of the analyzed RNA, their expression was normalized relative to human GAPDH.

human monocyte-derived macrophages were either left untreated or treated with 20% Huh7-TSN for 48 hours, in the presence or absence of G-1(1 μM). Then the protein was extracted for immunoblotting analysis. Immunoblotting was performed as described previously (14). Primary Abs used are listed as follows: anti-human Cyclin E1 (4129T, CST), Cyclin D1 (2978T, CST), CDK2 (2546T, CST), CDK4 (12790T, CST), p-AKT (13038S, CST), AKT (4685S, CST), p-Erk1/2 (4370T, CST), Erk1/2 (4695T, CST), β-actin (4970S, CST). HRP-linked anti-rabbit/mouse IgG Abs were purchased from CST.

Fresh biopsy specimens from HCC patients were weighed and fully ground to generate tissue homogenate. The cells were then lysed completely using an ultrasonic crusher, followed by centrifugation to obtain clarified supernatant. Methanol was added to the supernatant, and the mixture was incubated at room temperature for 10 minutes. After centrifugation, the supernatant was transferred to a clean tube and evaporated to dryness using centrifugal concentration drying system (Eppendorf). Assay buffer was added to reconstitute the precipitate, and then the estradiol content was immediately measured using an ELISA kit (501890, Cayman), following the manufacturer’s instructions.

The gene transcription expression data of HCC was obtained from The Cancer Genome Atlas (TCGA) database (22). Then the data was used to estimate the accumulation levels of various immune cells in 371 HCC tumor samples. This was achieved using the “IOBR” R package, employing the “Estimating the Proportion of Immune and Cancer cells” (“EPIC”) and the “xCell” methods (23).

The statistical tests used are indicated in the figure legends. Two-tailed Student’s t-test or one-way ANOVA with Tukey’s multiple comparisons test was used to compare the means of two or multiple groups, respectively. Survival curves were calculated using the Kaplan-Meier method and analyzed with the log-rank test. The statistical analyses mentioned above were performed using GraphPad Prism 6. Univariate and multivariate analyses were performed using the Cox proportional hazards model (SPSS Statistics 21, IBM). P < 0.05 was considered statistically significant for all tests.

To investigate the sex-related differences in the immune microenvironment of HCC tissues, we compared the density of various immune cells, determined by IHC, in tumor tissues derived from male and female patients in our previous cohort (24). While there were no significant differences in the density of CD3⁺ T cells, CD20⁺ B cells or CD15⁺ neutrophils between male and female patients, we observed a higher accumulation of CD68⁺ macrophages in the tumor tissues of male patients compared to female patients ( Figures 1A–C ; n _male = 381; n _female = 52). We also found higher levels of the markers CD204 and CD163, which are associated with a pro-tumor phenotype of macrophages in HCC (25, 26), in the tumor tissues of male patients compared to female patients ( Figures 1D, E ). These differences were not observed in the non-tumor regions. Additionally, we examined the immune cell infiltrations in HCC tumor samples from TCGA dataset using the “xCell” and the “EPIC” methods, and found that the level of macrophages, particularly M2-like macrophages, also showed the same sex discrepancy ( Supplementary Figures S1A, B ; n _male = 250; n _female = 121).

Self-replicating macrophages are enriched in HCC tumors and serve as an important mechanism for macrophage accumulation (14). Therefore, we set out to investigate whether sex difference was involved in the variation in macrophage self-replication. We conducted double immunofluorescence staining of Ki67 and CD68 in HCC tumor tissues from our previous cohort (14), and compared the level of Ki67⁺CD68⁺ cells between male and female patients ( Figure 1F ). The results revealed that both the percentage and the number of proliferating macrophages in tumor tissues were significantly higher in male patients compared to female patients ( Figure 1G ).

These results collectively reveal that there are sex disparities in macrophage proliferation and accumulation in HCC tumor tissues.

To investigate the mechanism behind this phenomenon, we established an in vitro model by incubating human monocyte-derived macrophages with tumor culture supernatants (TSN) to induce macrophage proliferation, as previously described (14) ( Figure 2A ). Then, qPCR was utilized to examine the expression of sex hormone receptors, specifically estrogen receptor α (ERα), estrogen receptor β (ERβ), GPER1 and androgen receptor (AR), which are known to play roles in sex-related differences in physiological and pathological conditions. The results showed that both GPER1 and ERα were significantly downregulated by TSN, whereas ERβ and AR were not affected ( Figure 2B ). Next, we compared the protein-level expression of GPER1 or ERα between proliferating and non-proliferating macrophages. Flow cytometry and immunofluorescence staining revealed that proliferating macrophages exhibited significantly lower expression of GPER1 when compared to non-proliferating cells in the presence of TSN ( Figures 2C–E ). However, no significant difference was observed in the expression of ERα between proliferating and non-proliferating macrophages ( Figure 2F and Supplementary Figure S2 ).

The association between GPER1 expression and macrophage proliferation was examined in human HCC tumor tissues. Flow cytometry analysis of fresh tumoral leukocytes isolated from HCC patients showed that the proliferation level was significantly lower in CD14⁺GPER1⁺ cells compared to CD14⁺GPER1^- cells ( Figure 2G ). GPER1 expression in macrophages was also detected in situ using confocal microscopy. As shown in Figure 2H , the GPER1 positive signal was less prominent in Ki67⁺ macrophages compared to Ki67^- macrophages. Statistical analysis revealed that patients with a higher frequency of GPER1⁺ macrophages exhibited a significantly lower level of macrophage proliferation in tumor tissues. Additionally, the levels of 17β-estradiol (E2), the primary natural ligand for GPER1 (27, 28), were measured in HCC tumor tissues, and the result indicated that tumor tissues from female patients tended to exhibit a higher level of E2 compared to those from male patients ( Figure 2I ). Therefore, there is a negative correlation between the expression of GPER1 in macrophages and their proliferation, both in vitro and in HCC tumor tissues.

Next, we set out to investigate the impact of GPER1 signaling on macrophage proliferation using G-1, a GPER1 specific agonist (29). The results showed that G-1 significantly suppressed the proliferation of macrophages induced by TSN from various hepatoma cell lines ( Figures 3A–D ). Moreover, the inhibitory effect of G-1 on TSN-induced macrophage proliferation was observed to be dose-dependent ( Figures 3C, D ). Immunofluorescence staining confirmed that G-1 treatment resulted in a notable decrease in the percentage of Ki67⁺ macrophages ( Figures 3E, G ). To determine DNA synthesis in macrophages, we conducted an EdU incorporation assay, which revealed an elevated frequency of EdU⁺ cells in macrophages exposed to TSN. However, this frequency was markedly reduced by the treatment of G-1 ( Figures 3F, H ). Additionally, the impact of GPER1 signaling on macrophage accumulation was examined using a CCK-8 assay. The results indicated that TSN increased the number of macrophages, while G-1 significantly decreased it ( Figure 3I ).

We further used a GPER1 specific antagonist G-15 to inhibit GPER1 signaling (30) that was activated by G-1. The results displayed that the suppression of macrophage proliferation mediated by G-1 was attenuated by G-15 treatment ( Figure 3J and Supplementary Figure S3 ). In addition, we also examined the effect of E2 on macrophage proliferation, and the results showed that E2 inhibited TSN-induced macrophage proliferation, which was rescued by G-15 ( Supplementary Figure S4 ).

Taken together, these data suggest that activation of GPER1 signaling restricts macrophage proliferation.

The PI3K/AKT and MEK/ERK pathways play crucial roles in regulating macrophage proliferation (14). Therefore, we investigated whether these signaling pathways are involved in the GPER1-mediated downregulation of macrophage proliferation. As expected, the levels of p-ERK and p-AKT significantly increased in macrophages exposed to TSN. G-1 treatment reversed the level of p-ERK, while the level of p-AKT was not significantly affected ( Figures 4A, B ). This indicates that GPER1 activation restricts macrophage proliferation by inhibiting the MEK/ERK pathway.

The ERK-mediated cyclin-dependent pathway is involved in cell-cycle progression and proliferation (31). Considering the reduced activity of DNA synthesis in TSN-treated macrophages after GPER1 activation ( Figures 3F, H ), we examined the expression levels of cyclins and cyclin-dependent kinases that regulate the entry of the DNA synthesis phase. Immunoblotting results showed that G-1 treatment significantly decreased the TSN-induced upregulation of cyclin D1, cyclin E1 and CDK4 ( Figures 4C, D ). These data suggest that GPER1 activation may restrict macrophage proliferation by downregulating the MEK/ERK/cyclin signaling pathway.

Next, we investigated whether the activation of GPER1 signaling with G-1 affects macrophage proliferation and accumulation in mouse models of HCC. We found that G-1 treatment significantly suppressed tumor growth in a subcutaneous Hepa1-6 tumor model ( Figures 5A–C ). Flow cytometry analysis displayed a significant reduction in macrophage proliferation in G-1-treated tumors ( Figure 5D ). Immunofluorescence staining of tumor tissues showed that G-1 treatment significantly decreased both the proliferation and the density of macrophages compared to the control group ( Figures 5E, F ). This inhibition of macrophage proliferation and accumulation by G-1 was also confirmed in an orthotopic Hepa1-6 tumor model ( Figure 5G ). Furthermore, we found that the macrophages in G-1-treated tumors tended to exhibit a lower level of PD-L1 in both the subcutaneous ( Figure 5H ) and the orthotopic HCC model ( Figure 5I ). The decrease in PD-L1 expression was also observed on human macrophages cultured in vitro when treated with G-1 ( Figure 5J ). Collectively, these results suggest that activating GPER1 signaling by G-1 restricts the proliferation and accumulation of macrophages in HCC mouse models.

To investigate the clinical significance of GPER1⁺ macrophages, percentages of GPER1⁺CD68⁺ cells in HCC tumor tissues were determined using immunofluorescence staining and confocal microscopy analysis. The patients were divided into two groups based on the median percentage of GPER1⁺CD68⁺ cells. As shown in Figures 6A, B , patients with a higher percentage of GPER1⁺ macrophages exhibited significantly longer OS and RFS compared to those with a lower level. Multivariate analysis indicated that the percentage of GPER1⁺ macrophages was an independent prognostic factor for both OS and RFS ( Figures 6C, D , Supplementary Table S2 ). There was no obvious association between GPER1⁺ macrophages and the clinical characteristics of the patients ( Supplementary Table S3 ).

These findings suggest that the presence of GPER1⁺ macrophages in HCC tumors predicts a better prognosis for HCC patients.