Non-enzymatic Rnaseh2c orchestrates proliferating macrophage-driven immunosuppression and HCC progression via Cdk9 proliferation axis and CCL2/CCR2-mediated CD8+ Tex infiltration: a novel therapeutic paradigm with “Rnaseh2c-In1” inhibitor Authors: Zengbin Wang, Ruiying Xiao, Mengxin Liu, Lihua Ceng, Zikun Lin, Xinyu Zheng, Miaoxin Shi, Hong Ye, Linqing Wu, Nanhong Tang Published: 2026 ABSTRACT AbstractBackgroundTumor-associated macrophages are pivotal drivers of hepatocellular carcinoma (HCC) progression. However, the functional contributions of proliferating macrophages (Prolif Ms) within the tumor microenvironment (TME) remain poorly defined.MethodsHere, we integrated single-cell RNA sequencing, Cleavage Under Targets and Tagmentation (CUT&Tag), bulk RNA sequencing, multiparametric immunofluorescence, flow cytometry, and molecular dynamics simulation to dissect the phenotype, functional landscape, and targeting strategy of Prolif Ms in HCC.ResultsWe identified a novel Prolif Ms subset characterized by high expression of proliferation markers (Cdk1, Mki67), distinct from classical M1/M2 macrophages. Mechanistically, ribonuclease H2 subunit C (Rnaseh2c) acts as a non-enzymatic regulator, binding directly to the cyclin-dependent kinase promoter to drive macrophage proliferation while suppressing endocytic and antigen-presenting capabilities. Rnaseh2c further promotes CD8+ exhausted T cell (CD8+ Tex) infiltration via the C-C motif chemokine ligand 2/C-C motif chemokine receptor type 2 (CCL2/CCR2) axis, fostering an immunosuppressive TME. Notably, Rnaseh2c exhibits cancer-promoting effects in both macrophages and hepatocytes. We developed a specific inhibitor “Rnaseh2c-In1”, which inhibits HCC growth by reducing CCL2 secretion and CD8+ Tex infiltration, demonstrating favorable pharmacokinetics and low toxicity.ConclusionsThis study elucidates a non-enzymatic Rnaseh2c-driven mechanism regulating Prolif Ms, providing a novel therapeutic target and candidate compound for HCC treatment. WHAT IS ALREADY KNOWN ON THIS TOPIC It is known that macrophages are crucial components in the hepatocellular carcinoma (HCC) tumor microenvironment, with established roles for M1 and M2 subtypes, and that CD8+ exhausted T cell (Tex) contribute to tumor progression. However, the existence of a distinct proliferating macrophage subset and its specific regulatory mechanisms remains unclear. It is known that macrophages are crucial components in the hepatocellular carcinoma (HCC) tumor microenvironment, with established roles for M1 and M2 subtypes, and that CD8+ exhausted T cell (Tex) contribute to tumor progression. However, the existence of a distinct proliferating macrophage subset and its specific regulatory mechanisms remains unclear. WHAT THIS STUDY ADDS This study identifies a novel proliferating macrophage subset (Prolif Ms) in HCC, driven by the transcription factor Rnaseh2c, and demonstrates that Rnaseh2c promotes HCC progression by enhancing C-C motif chemokine ligand 2/C-C motif chemokine receptor type 2 (CCL2/CCR2) signaling to recruit CD8+ Tex cells, while also developing a targeted inhibitor (Rnaseh2c-In1) that suppresses tumor growth. This study identifies a novel proliferating macrophage subset (Prolif Ms) in HCC, driven by the transcription factor Rnaseh2c, and demonstrates that Rnaseh2c promotes HCC progression by enhancing C-C motif chemokine ligand 2/C-C motif chemokine receptor type 2 (CCL2/CCR2) signaling to recruit CD8+ Tex cells, while also developing a targeted inhibitor (Rnaseh2c-In1) that suppresses tumor growth. HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY These findings highlight Prolif Ms and Rnaseh2c as promising therapeutic targets in HCC, providing a rationale for novel immunomodulatory strategies aimed at disrupting the CCL2/CCR2 axis to alleviate CD8+ T-cell exhaustion and potentially improve patient outcomes. These findings highlight Prolif Ms and Rnaseh2c as promising therapeutic targets in HCC, providing a rationale for novel immunomodulatory strategies aimed at disrupting the CCL2/CCR2 axis to alleviate CD8+ T-cell exhaustion and potentially improve patient outcomes. Introduction Macrophages, as central effector cells of the innate immune system, play critical roles in maintaining tissue homeostasis, regulating immune responses, and eliminating pathogens.1 Their functional versatility allows dynamic phenotypic adaptation to microenvironmental cues, enabling functions ranging from phagocytosis and antigen presentation to cytokine-mediated inflammation or tissue repair. Within the tumor microenvironment (TME), tumor-associated macrophages (TAMs) represent the most abundant immune cell population, and their functional heterogeneity directly influences tumor progression, metastasis, and therapeutic responses. Traditional paradigms classify TAMs into classically activated M1 (pro-inflammatory, antitumor) and alternatively activated M2 (anti-inflammatory, protumor) subtypes,2 with further subdivisions (M2a-d) guiding anti-TAM therapeutic strategies aimed at inhibiting recruitment, depleting local populations, or inducing phenotypic reprogramming.3 Our previous research has revealed a new mechanism for regulating the function of TAMs in hepatocellular carcinoma (HCC) from multiple aspects such as post-translational modifications and metabolic remodeling, elucidating the molecular basis for promoting immune escape.46 However, the actual heterogeneity of TAMs exceeds conventional classification frameworks, often exhibiting continuous phenotypic spectra with unique subpopulations varying across tumor types and even within distinct tumor regions, limiting the efficacy of current targeting approaches. Recent advances in single-cell RNA sequencing (scRNA-seq) have revolutionized the understanding of TAM transcriptional diversity. Studies employing this technology have identified over seven distinct TAM subpopulations across various cancers, including interferon (IFN)-responsive TAMs,7 8 immunoregulatory TAMs,9 inflammatory cytokine-enriched TAMs,10 lipid-associated TAMs,11 angiogenic TAMs,12 and tissue-resident macrophage-like TAMs.13 These findings underscore the transcriptional complexity of TAMs and simultaneously emphasize a critical knowledge gap: the functional characteristics and mechanistic roles of proliferative TAMs (proliferating macrophages (Prolif Ms)) remain undefined. Although proliferating macrophage subsets have been previously described in other cancer types, including CD68+/PCNA+ proliferating macrophages (promacs) in breast cancer which were associated with high-grade tumors but not directly linked to disease progression,14 the specific transcriptional regulators and functional mechanisms driving such populations in HCC have not been elucidated. HCC, as a malignant tumor with a high global mortality rate, has particularly prominent heterogeneity of TAMs in its TME. Previous studies have shown that TAMs in HCC can promote tumor progression through multiple mechanisms, such as inducing immunosuppression, promoting angiogenesis, and enhancing the ability of invasion and metastasis. However, the existence and functional characteristics of Prolif Ms in HCC and their interaction patterns with tumor cells or other immune cells remain unknown. This study aims to systematically identify the transcriptional characteristics, functional pathways and key regulatory factors of the Prolif Ms subpopulation in HCC, and explore potential therapeutic strategies targeting this subpopulation, expecting to provide new theoretical basis and target selection for the precise immunotherapy of HCC. Materials and methods Additional materials and methods are included in the online supplemental material. Results Identification of the Prolif Ms cells Macrophages play an important role in the HCC microenvironment. We constructed a subcutaneous HCC model and obtained CD45+ immune cells by using CD45+ immunomagnetic beads. The scRNA-seq was used to investigate the distinct infiltrating cell subpopulations (figure 1A). 39,010 transcriptome cells were obtained. Using t-distributed stochastic neighbor embedding clustering, 13 clusters were identified: epithelial cells (cluster 10), fibroblasts (clusters 6, 8, and 12), myeloid cells (clusters 1, 2, 4, and 5), neutrophils (cluster 13), T-natural killer (NK) cells (clusters 3, 7, and 11), and malignant cells (cluster 9) (figure 1B). Subsequently, 23,905 myeloid cells were selected. Uniform Manifold Approximation and Projection (UMAP) was used to identify and display four cell macrophage subpopulations. The macrophage subpopulation could be divided into four clusters, of which three clusters “Monocytes (Monos)”, “M1 macrophages (M1)”, and “M2 macrophages (M2)” can be clearly defined (figure 1C). Macrophages of the three clusters were identified as follows (figure 1D): we identified M1 using Cd86, Cd1415 and Plac816; the identification of M2 was based on the presence of markers Olfml3, Gas6, Ch25h1719 and the cancer-promoting marker Hpgds20; Monos were identified using Cd200r1 and Csf1r.21 Another unique cluster did not highly express the markers of M1 (Cd86, Il6) or M2 (Cd206, Tgfb1), but highly expressed the proliferation-related markers Cdk1 and Mki67. Therefore, we defined this cluster as Prolif Ms as previously described.22 UMAP plots of the top markers of the macrophage subpopulations were shown in online supplemental Figure S1. Additionally, UMAP and violin plots of the markers were shown in figure 1E,F. These results suggested the presence of different macrophage subpopulations in the HCC microenvironment, emphasizing that Prolif Ms may serve as a new macrophage subpopulation. Identification of the cell population of Prolif Ms. (A) A subcutaneous HCC model was established and CD45+ immune cells were obtained. The scRNA-seq was used to explore the distinct infiltrating cell subpopulations. (B) Thirteen clusters were identified using t-SNE. (C) 23,905 myeloid cells were selected. UMAP was used to identify and display four cell macrophage subpopulations. (D) Heatmap showed highly expressed marker genes. (E) and (F) UMAP and violin plots of 6 markers Cd14, Cd68, Cd200r4, Adgre1, Cdk1 and Mki67 were shown. HCC, hepatocellular carcinoma; NK, natural killer; scRNA-seq, single-cell RNA sequencing; t-SNE, t-distributed stochastic neighbor embedding; UMAP, Uniform Manifold Approximation and Projection. We then clustered genes with similar expression patterns in time development trajectories. Prolif Ms was specific to other types of macrophages (online supplemental Figure S2A). The high expression levels of the CD14, Cd68, Mki67, and Cdk1 genes were listed in online supplemental Figure S2B. To verify the presence of Prolif Ms in HCC, we used multiple immunofluorescence to test the expression of F480, CD68, and Ki67 in mouse subcutaneous HCC, mouse orthotopic HCC, and human HCC tissues. Co-localization expression of F480+ Ki67+ cells were observed in mouse subcutaneous HCC and orthotopic HCC (online supplemental Figure S2C). Co-localization expression of CD68+ Ki67+ cells were observed in human HCC (online supplemental Figure S2C), suggesting that Prolif Ms was present in HCC. To clarify the functional heterogeneity of Prolif Ms, we partitioned all the highly variable genes shown in online supplemental Figure S2A into four distinct modules. Prolif Ms mainly focused on module 2, while M2, M1 and Monos, respectively, focus on modules 4, 1 and 3. The heatmap depicted the fluctuations in gene expression during the differentiation process (online supplemental Figure S3A). To demonstrate the altered cellular functions, we performed Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis of the genes in each module. The functions of module 1 (M1) was mainly enriched in DNA replication and spliceosome pathways (online supplemental Figure S3B), while module 3 (Monos) primarily focused on tumor necrosis factor signaling, interleukin (IL)-17 signaling, hypoxia-inducible factor 1 (HIF-1) signaling, glycolysis/gluconeogenesis, and cytokine–cytokine receptor interactions (online supplemental Figure S3D), and module 4 (M2) was strongly associated with Jak-STAT signaling, Nod-like/toll-like receptor signaling, and chemokine pathways (online supplemental Figure S3E); in contrast, module 2 (representing Prolif Ms) stood out as a pivotal functional unit with predominant enrichment in endocytosis, phagosome, and lysosome (online supplemental Figure S3C). These findings suggest that the function of Prolif Ms may differ from that of M1, M2 and Monos, indicating that Prolif Ms may have an impact on the endocytosis of pathogenic microorganisms. Prolif Ms promotes CD8+ Tex infiltration through CCL2/CCR2 signaling To evaluate the association between Prolif Ms and immune cells, we conducted a further investigation of the cell–cell interaction network using the CellPhoneDB. Prolif Ms interacted with Monos, M1 macrophages, M2 macrophages, and T-NK cells (figure 2A). The heatmap showed the ligand-receptor relationships between the Prolif Ms and Monos, M1 macrophages, M2 macrophages, and T-NK cells, where the interactions between Prolif Ms and M1 and M2 involved a significant number of ligand-receptor pairs (figure 2B). scRNA-seq provides transcriptomic analysis of immune cells, while cytometry by time-of-flight (CyTOF) enables examination at the protein level.5 23 These two technologies are therefore complementary and were used in this study for orthogonal validation. To verify the connection between Prolif Ms and M1 and M2 macrophages, we established a subcutaneous HCC model and identified the immune cell composition in HCC using CyTOF (figure 2C). Using self-organizing maps (FlowSOM), we classified various embedded macrophage clusters, including M1, M2, Monos and Prolif Ms (figure 2D). Correlation analysis revealed a significant positive correlation between the Prolif Ms and M1 and M2 (figure 2E). Prolif Ms promotes CD8+ Tex infiltration through CCL2/CCR2 signaling. (A) Cell communication point diagram. Rows represent ligand-receptor pairs with intercellular communication relationships. The list shows the types of cells that undergo cellular communication (receiver cell | ligand cell). The size of the circle indicates the level of significance, and the redder the color of the circle is, the greater the expression value of the interacting cell genes. (B) Heatmap showing the number of interacting pairs between different cell types, with red indicating a greater number of interacting ligand-receptor pairs. (C) Hep53.4 cells were injected subcutaneously into the right side of C57BL/6J mice. On the 20th day after implantation, we stripped the tumor tissue, collected the CD45+ immune cells and performed CyTOF analysis. n=12. (D) A heatmap depicting the median antigen intensity is presented, highlighting only the statistically significant p values. (E) Pearson correlation analysis of the correlation between Prolif Ms and M1, M2 infiltration in HCC. (F) The heatmap shows the number of interacting pairs between different cell types, with red indicating a greater number of interacting ligand-receptor pairs. (G) Cell communication point diagram. (H) Strategies for flow cytometry analysis of CD8+ Tex cells. (I) Comparison of the proportion of CD8+Tex cells between groups. n=6. Data analysis using ANOVA. *p<0.05, **p<0.01, ***p<0.001. ANOVA, analysis of variance; BMDM, bone marrow-derived macrophage; HCC, hepatocellular carcinoma; IL, interleukin; NK, natural killer; Prolif Ms, proliferative macrophages; Tex, exhausted T cell. In addition to their relationship with macrophages, we also analyzed the interaction between Prolif Ms and T cells. According to our previous research,5 we identified the following types of T-cell subsets: Cd4+T (Cd4+), Cd8+Tem (Cxcr3+Cxcr4+Cd44+Ifng+Cd8+), Cd8+Trm (Cd6+Xcl1+Capg+Cd69+Nr4a1+Nr4a2+Cd8+), Cd8+Tex (Mir55hg+Tnfrsf9+Ctla4+Pdcd1+Lag3+Tox+Gzmk+Gzmb+Ifng+Cd8+), Cd8+Tcm (Cd28+Cd27+Prf1+), Cd8+Teff (Ifng+Cxcr3+Tbx21+Eomes+Gzmk+Gzmb+Cd8+), Gzmk−Cd8+Texp (Gzmk−Pdcd-1+Havcr2+Ctla4+Cd8+), Gzmk+Gzmb+Cd8+Texp (Gzmk+Gzmb+Pdcd1+Havcr2+Ctla4+Cd8+), and Gzmk+Gzmb−Cd8+Texp (Gzmk+Gzmb−Pdcd1+Havcr2+Ctla4+Ifng+Cd8+). There was a significant interaction relationship between Prolif Ms and CD8+Tex (figure 2F). The sustained presence of CD8+ exhausted T cell (CD8+Tex) in the immune microenvironment of HCC may lead to the sustained progression or deterioration of HCC.24 Therefore, we speculated that Prolif Ms may potentially interact with CD8+ Tex. We thus evaluated the receptor-ligand interactions between Prolif Ms and CD8+ Tex. The results showed that CCL2/CCR2 signaling and CXCL16/CXCR6 signaling were significantly enriched (figure 2G). Previous studies have shown that these two signals play crucial roles in mediating the CD8+ T-cell immune response.25 26 We obtained bone marrow-derived macrophages (BMDMs) from mouse bone marrow and used IL-4 to induce BMDMs to become Prolif Ms according to the research of Jenkins et al.27 Flow cytometry (FCM) analysis revealed high expression of mKi67 and Cdk1 in BMDMs induced by IL-4 (online supplemental Figure S4A). Moreover, IL-4-induced BMDMs exhibited lower phagocytic ability (online supplemental Figure S4B). We next established a mouse subcutaneous HCC tumor model and found that peritumoral injection of IL-4-induced BMDMs increased the level of CCL2 in mouse serum (online supplemental Figure S4C). The CCR2 and CXCR6 inhibitors were used to evaluate which signal Prolif Ms acts on CD8+ Tex. FCM analysis showed that peritumoral injection of IL-4-induced BMDMs significantly promoted CD8+ Tex aggregation, CCR2 inhibitors reversed this effect, while the effect of CXCR6 inhibitors was not significant (figure 2H,I). The above results suggested that Prolif Ms promotes CD8+ Tex infiltration through CCL2/CCR2 signaling. Rnaseh2c is a specific regulatory factor for Prolif Ms Since the key regulatory mechanism of Prolif Ms was still unclear, we hypothesized that the Prolif Ms harbors crucial regulatory factors that govern this phenomenon. To establish a systematic gene regulatory network and construct corresponding cell maps, we used SCENIC software28 29 to identify coexpressed modules (regulons) between transcription factors and potential target genes, and to calculate the regulon activity score (RAS) of each cell. More importantly, we calculated the specific correspondence between predicted regulons and each individual cell type by determining the regulon specificity score (RSS). Thus, we combined RAS and RSS to screen for the most important regulatory factors. The RAS activity scores indicated that ribonuclease H2 subunit C (Rnaseh2c), transcription regulator protein BACH1 (Bach1), and transcription factor E2F8 (E2f8) acted as regulators of Prolif Ms (figure 3A). However, the RAS scores of Rnaseh2c and E2f8 were relatively low in the Monos, M1 and M2 cells. We further narrowed down the specific regulon of Prolif Ms. The regulon specificity ranking chart revealed that DDIT3 upstream open reading frame protein (Ddit3) had the highest RSS score for M1 and M2, while interferon regulatory factor 2 had the highest score for Monos, and Rnaseh2c had the highest score for Prolif Ms (figure 3B). The heatmap generated for RSS revealed a distinct correlation between Rnaseh2c and Prolif Ms, as depicted in figure 3C. These results indicate that Rnaseh2c is a specific regulatory factor in Prolif Ms. Through pseudotemporary gene expression curve analysis, we found that Rnaseh2c was more highly expressed in the Prolif Ms than in Monos, M1 and M2 (figure 3D). To evaluate the spatial localization of Rnaseh2c and Prolif Ms, we constructed two animal models (subcutaneous HCC and orthotopic HCC) in vivo. The spatial colocalization of Rnaseh2c and Prolif Ms was detected in mouse subcutaneous and orthotopic HCC tissues, respectively (figure 3E). We also observed the spatial colocalization of Rnaseh2c and Prolif Ms in human HCC tissues. Compared with para-carcinoma, there were more such colocalized cells around the carcinoma (figure 3F). Rnaseh2c is a specific regulatory factor for Prolif Ms. (A) The RAS activity heatmap of regulons in each cell group. Rows represent different regulons, while columns represent different cell groups. The color changes from blue to red, indicating a decrease in RAS activity score from low to high. The higher the RAS score, the stronger the activity of regulon in this cell group. (B) RSS chart. The horizontal axis represents ranking, and the vertical axis represents RSS score. The higher the RSS is, the more specific the regulator may be related to the cell population. (C) RSS specific heatmap of regulons in each cell group. Rows represent different regulons, and columns represent different cell populations. The color changes from blue to red, indicating a decrease in RSS specificity score from low to high. The higher the RSS score, the stronger the specificity of regulon in this cell group. (D) Pseudotemporal gene expression curve. (E) Multiple immunofluorescence detection of the Rnaseh2c+ Prolif Ms (Rnaseh2c+ F480+ Ki67+) in subcutaneous and orthotopic HCC. (F) Multiple immunofluorescence detection of the Rnaseh2c+ Prolif Ms (Rnaseh2c+ CD68+ Ki67+) in human HCC. HCC, hepatocellular carcinoma; Prolif Ms, proliferative macrophages; RAS, regulon activity score: Rnaseh2c, ribonuclease H2 subunit C; RSS, regulon specificity score. These results suggest that Rnaseh2c-mediated Prolif Ms may play an important role in the progression of HCC. Rnaseh2c exacerbates HCC progression through CCL2/CCR2 signaling-mediated CD8+ Tex infiltration We further explored the key role of Rnaseh2c-mediated Prolif Ms in CD8+ Tex infiltration. We sorted the Macrophage (F480+Cd11b+) and Prolif Ms (F480+Cd11b+Rnaseh2c+) from the HCC subcutaneous tumors using FCM (figure 4A). We injected macrophages and Prolif Ms around the subcutaneous HCC tumor and found that Prolif Ms significantly promoted the growth of HCC (figure 4B). We also found that injecting Prolif Ms around the orthotopic HCC tumor significantly worsened mouse prognosis (online supplemental Figure S4D) and enhanced the infiltration of CD8+ Tex (online supplemental Figure S4E). These results indicate that Rnaseh2c enhances the progression of HCC. Furthermore, the messenger RNA (mRNA) expression levels of Rnaseh2c and Mki67 in Prolif Ms were detected using quantitative reverse transcription polymerase chain reaction (qRT-PCR) in macrophage. In both subcutaneous HCC and subcutaneous melanoma (Mm), correlation analysis suggested a positive correlation between the mRNA levels of Rnaseh2c and Mki67 (figure 4C). Moreover, the mRNA expression level of Rnaseh2c was positively correlated with tumor growth (figure 4D). To further confirm the HCC promoting effect of Rnaseh2c, we constructed the BMDMs/overexpression (OE) Rnaseh2c and BMDMs/OE negative control (NC) cell line. BMDMs/OE Rnaseh2c and BMDMs/OE NC were subcutaneously injected into the HCC tumor area. Compared with that in the BMDMs/OE NC group, the growth of tumor in the BMDM/OE Rnaseh2c group was significantly accelerated (figure 4E). Overexpression of Rnaseh2c significantly promoted the secretion of CCL2 in mouse serum (online supplemental Figure S4F). To clarify whether Rnaseh2c also promotes CD8+ Tex infiltration by influencing CCL2/CCR2, the CCR2 and CXCR6 inhibitors were used. We performed peritumoral injection of the Macrophage and Prolif Ms obtained from figure 4A. FCM analysis demonstrated that Prolif Ms enhanced the infiltration of CD8+ Tex in the TME. This effect was significantly suppressed by the CCR2 antagonist 4, whereas the CXCR6 antagonist ML339 showed no measurable impact on CD8+ Tex recruitment (figure 4F). Consistent with these findings, multiplex immunofluorescence assays in a murine subcutaneous HCC model confirmed comparable patterns of immune cell modulation across treatment groups (figure 4G). These results indicate that Rnaseh2c recruits CD8+ Tex through the CCL2/CCR2 signaling pathway to promote HCC. Rnaseh2c exacerbates HCC progression through CCL2/CCR2 signaling-mediated CD8+ Tex infiltration. (A) CD45, F480, Cd11b and Rnaseh2c antibodies were used for Rnaseh2c+Prolif Ms sorting. (B) The tumor tissue after cell therapy was collected. The difference in tumor weight between the two groups was compared. n=6. Data were mean±SD and analyzed by t-test. (C) Hep53.4 and B16/F10 cells were used to construct subcutaneous HCC and Mm models, respectively, n=20. qRT-PCR was used to detect the mRNA expression levels of Mki67 and Rnaseh2c in macrophage. Correlation analysis of mRNA expression levels of Mki67 and Rnaseh2c. (D) Correlation analysis between the mRNA expression level of Rnaseh2c and tumor weight. (E) The tumor tissue after cell therapy was collected. The difference in tumor weight between the two groups was compared. n=6. Data were mean±SD and analyzed by unpaired t-test. (F) Comparison of the proportion of CD8+ Tex cells between groups. n=4. Data analysis using one-way ANOVA. (G) Multiple immunofluorescence detection of CD8 (red) and PD-1 (green) expression in subcutaneous HCC. n=4. Data analysis using one-way ANOVA. *p<0.05, **p<0.01, ***p<0.001. ANOVA, analysis of variance; BMDM, bone marrow-derived macrophage; FCM, flow cytometry; HCC, hepatocellular carcinoma; mRNA, messenger RNA; PD-1, programmed cell death protein 1; Prolif Ms, proliferative macrophages; Rnaseh2c, ribonuclease H2 subunit C; Tex, exhausted T cell. Rnaseh2c enhances macrophage proliferation by binding to the promoter of Cdk9 To deeply evaluate the relationship between Rnaseh2c and Prolif Ms, we also divided the macrophages obtained from scRNA-seq into two groups: high expression and low expression of the Rnaseh2c gene. By comparing the differentially expressed genes (DEGs) between the two groups (online supplemental Table S1), we detected increased Mki67 gene expression in cells with high Rnaseh2c expression (figure 5A). Gene Ontology (GO) enrichment analysis showed that the cell cycle, DNA binding and chromatin signaling pathways were significantly enriched in the group with high Rnaseh2c expression (figure 5B). Cell cycle analysis also showed that Prolif Ms were mainly distributed in the G2M and S phases, while there was no such effect in Monos, M1, or M2 cells, indicating that Prolif Ms were in a highly proliferative state (figure 5C). Rnaseh2c enhances macrophage proliferation by binding to the promoter of Cdk9. (A) Gene expression heatmap. (B) GO enrichment analysis. (C) Cell cycle analysis. (D) Differential gene volcano map. n=3. (E) GO enrichment analysis. (F) CFSE staining. n=3. (G) Difference peaks. (H) Signal distribution map and peak heat map. (I) DLR assay detecting Cdk9 promoter activity. n=4. Data were mean±SD and analyzed by t-test. (J) Western blot analysis detecting protein expression of Cdk9. n=3. Data were mean±SD and analyzed by t-test. (K) CFSE staining. n=3. Data analysis using one-way ANOVA. (L) We used the EdU proliferation detection kit to detect the number of EdU positive cells, n=6. Bar=50 µm. Images were representative of three experiments, and analyzed by one-way ANOVA. *p<0.05, **p<0.01, ***p<0.001. ANOVA, analysis of variance; Cdk9, cyclin-dependent kinase 9; DLR, dual-luciferase reporter assay; FC, fold change; Rnaseh2c, ribonuclease H2 subunit C; UMAP, Uniform Manifold Approximation and Projection. We constructed mice with macrophage-specific knockout of Rnaseh2c (Rnaseh2cCKO). We sorted the macrophages in the orthotopic HCC model and performed RNA-seq. Compared with the Ctrl group, 582 genes were upregulated and 529 genes were downregulated in the Rnaseh2cCKO group (figure 5D). GO enrichment analysis revealed that knockout of Rnaseh2c decreased DNA replication, nuclease activity and endonuclease activity (figure 5E). Specifically, carboxyfluorescein diacetate succinimidyl ester (CFSE) staining suggested that BMDM cells overexpressing Rnaseh2c demonstrated enhanced proliferation rates (figure 5F). To further analyze how Rnaseh2c affects the proliferation of Prolif Ms, we employed the CUT&Tag technique. The results uncovered significant differences in peaks between the OE NC and OE Rnaseh2c group (figure 5G). When Rnaseh2c was overexpressed in BMDMs, the abundance of gene promoters bound to it was high (figure 5H). Depth analysis revealed that Rnaseh2c bound to the promoter of the proliferation cyclin-dependent kinase 9 (Cdk9). The dual-luciferase reporter assay also found that overexpression of Rnaseh2c activated the Cdk9 gene in BMDMs (figure 5I). Overexpression of Rnaseh2c significantly promoted Cdk9 expression (figure 5J). In addition, CFSE staining displayed that the cell proliferation induced by Rnaseh2c overexpression was significantly reversed by si Cdk9 (figure 5K). We next also constructed lentiviral vectors for overexpressing of Rnaseh2c (OE Rnaseh2c) in mouse macrophage cell lines (RAW264.7). Overexpression of Rnaseh2c-induced 5-ethynyl-2'-deoxyuridine (EdU)-positive cells was significantly reversed by si Cdk9 (figure 5L). These results indicate that Rnaseh2c significantly enhances the macrophage proliferation progression via binding to the Cdk9 promoter. Rnaseh2c inhibits the endocytosis and antigen presentation of Prolif Ms Macrophage phagocytosis plays a pivotal role in the host defense against invading microorganisms.30 Ki67+ macrophages possess strong endocytic activity.31 The above KEGG analysis results have indicated that we should pay attention to the phagocytic function of Prolif Ms. Thus, we used the Cell Meter Fluorometric Phagocytosis Assay to investigate the impact of Rnaseh2c on the phagocytic ability. Our results demonstrated that knockdown of Rnaseh2c significantly promoted the phagocytic ability of BMDMs and RAW264.7 cells, while overexpression of Rnaseh2c restored this effect (figure 6A,B). In addition, gene set enrichment analyses (GSEA) analysis revealed that knockout of Rnaseh2c significantly increased its endocytosis ability (figure 6C). The endocytosis ability of macrophages is closely related to their antigen-presenting ability. Furthermore, our previous studies confirmed that Rnaseh2c decreased the expression of major histocompatibility complex II and also inhibited the antigen-presenting ability of macrophages.32 We next used immunoblotting to detect the expression levels of Cdk1, Cdk9 and Mki67 related to cell proliferation, respectively. Our data showed that the expression of Cdk1, Cdk9 and Mki67 were decreased in the BMDMs/sh Rnaseh2c and RAW264.7/sh Rnaseh2c groups, but increased in the BMDMs/OE Rnaseh2c and RAW264.7/OE Rnaseh2c groups (figure 6D,E). These results suggested that Rnaseh2c can enhance the proliferation ability of Prolif Ms, but inhibit its phagocytic ability. Rnaseh2cCKO inhibited HCC growth and improved prognosis compared with wild-type mice in the orthotopic (figure 6F) and subcutaneous HCC models (online supplemental Figure S5A-C). The BMDMs from Rnaseh2cCKO mice demonstrated stronger phagocytic ability (figure 6G). Furthermore, FCM analysis showed that the proportion of CD8+ Tex in the HCC microenvironment of Rnaseh2cCKO mice was significantly reduced in the orthotopic (figure 6H) and subcutaneous HCC models (online supplemental Figure S5D). ELISA demonstrated that the serum CCL2 level in Rnaseh2cCKO mice was markedly inhibited (figure 6I). These results indicate that Rnaseh2c interferes with the endocytosis and antigen presentation of macrophages. Rnaseh2c interferes with the endocytosis and antigen presentation of Prolif Ms. (A) and (B) The phagocytic rate of macrophages was determined with a Cell Meter Fluorometric Phagocytosis Assay Kit. Green represents macrophages. n=4. Bar=100 µm. Data were mean±SD and analyzed by ANOVA: ***p<0.001. (C) GSEA. (D) and (E) Immunoblotting was used to detect the expression of the Cdk1, Cdk9 and Mki67 proteins. n=3. Blots were each representative of three experiments. Data were mean±SD and analyzed by ANOVA. *p<0.05, **p<0.01, ***p<0.001. (F) DEN induced primary mouse HCC (left) and survival probability (right). n=10. Data were mean±SD and analyzed by Kaplan-Meier: ***p<0.001. (G) The phagocytic rate of macrophages was determined with a Cell Meter Fluorometric Phagocytosis Assay Kit. Green represents macrophages. n=6. Bar=100 µm. Data were mean±SD and analyzed by t-test: ***p<0.001. (H) Comparison of the proportion of CD8+ Tex cells between groups. n=6. Data were mean±SD and analyzed by t-test: ***p<0.001. (I) ELISA was used to analyze the secretion level of CCL2 in mouse serum. n=6. Data were mean±SD and analyzed by t-test: ***p<0.001. ANOVA, analysis of variance; BMDM, bone marrow-derived macrophage; Cdk9, cyclin-dependent kinase 9; HCC, hepatocellular carcinoma; Prolif Ms, proliferative macrophages; Rnaseh2c, ribonuclease H2 subunit C; Tex, exhausted T cell. Virtual screening and molecular dynamics simulation for the identification of inhibitor “Rnaseh2c-In1” targeting Rnaseh2c We then proceeded to screen for inhibitors targeting Rnaseh2c. The Rnaseh2c protein structure was selected from a previously reported protein (PDB: 3KIO). Through pocket detection on this protein, three binding sites were identified (figure 7A). Among them, Site1 scored the highest, and therefore we considered it suitable as the binding pocket for virtual screening. Subsequently, we performed multiple rounds of virtual screening via molecular docking against a compound library containing 1,620,219 molecules, resulting in 1,267 candidate compounds. Computational filtering was applied to remove PAINS structural motifs. With the criteria of quantitative estimation of drug-likeness (QED) >0.5 and a docking score <−8 kcal/mol, 28 small molecules were retained. Based on the analysis of binding poses with Rnaseh2c and structural diversity of the compounds, five potential candidates were finally selected (figure 7B and online supplemental Figure S6A-E). To further identify effective inhibitors of Rnaseh2c, we treated Hepa1-6 cells with these five compounds at a concentration gradient. The results showed that compound “8019-9386” at 8 µM exhibited the most pronounced inhibitory effect on Rnaseh2c (figure 7C,D). Therefore, we selected “8019-9386” as the optimal inhibitor of Rnaseh2c and designated it as “Rnaseh2c-In1”. Moreover, CCK-8 assays demonstrated that Rnaseh2c-In1 significantly inhibited the proliferation of mouse hepatocellular carcinoma, lung cancer, gastric cancer, and colon cancer cells (figure 7E). Virtual screening and molecular dynamics simulations identify the Rnaseh2c-targeting inhibitor “Rnaseh2c-In1”. (A) Pocket scoring of the Rnaseh2c protein (PDB: 3KIO). (B) Five potential candidate compounds. (C) and (D) Western blot analysis of Rnaseh2c protein expression. Hepa1-6 cells were treated with the compound at concentrations of 0, 2, 4, and 8 µM for 48 hours. Proteins were extracted and subjected to western blotting to detect Rnaseh2c expression. Data were mean±SD and analyzed by ANOVA. *p<0.05, **p<0.01, ***p<0.001. (E) CCK-8 assay analyzing the effect of Rnaseh2c-In1 (8 µM) on cell viability. Data were mean±SD and analyzed by two-way ANOVA. ***p<0.001. (F) ADMET prediction results for compound Rnaseh2c-In1. The radar chart illustrates the ADMET properties of Rnaseh2c-In1, evaluating lipophilicity, size, polarity, solubility, saturation, and flexibility. The area between the red and blue polygons represents the ideal range; the green polygon indicates the predicted values for Rnaseh2c-In1 (left panel). Physicochemical properties, drug-likeness, and lipophilicity of Rnaseh2c-In1. The area inside the red polygon indicates the ideal range; the blue polygon shows the predicted values for Rnaseh2c-In1 (middle panel). Toxicity predictions for Rnaseh2c-In1 (including hepatotoxicity, neurotoxicity, cytotoxicity, mutagenicity, immunotoxicity, carcinogenicity). The area inside the red polygon is predicted as “inactive”, with probabilities approaching 100% toward the center. The area outside the red polygon is predicted as “active”, with probabilities approaching 100% toward the periphery (right panel). (G) RMSD trajectory of the Rnaseh2c-Rnaseh2c-In1 complex. (H) RMSD trajectory of Rnaseh2c-In1. (I) Rg trajectory of the Rnaseh2c-Rnaseh2c-In1 complex. (J) SASA trajectory of the Rnaseh2c-Rnaseh2c-In1 complex. (K) RMSF profile of the Rnaseh2c protein. (L) Hydrogen bond variation during the simulation of the Rnaseh2c-Rnaseh2c-In1 complex. (M) Gibbs free energy landscape of the Rnaseh2c-Rnaseh2c-In1 complex. (N) Binding free energy histogram of the Rnaseh2c-Rnaseh2c-In1 complex. ANOVA, analysis of variance; Rnaseh2c, ribonuclease H2 subunit C. To evaluate the pharmacokinetic properties (ADME: absorption, distribution, metabolism, and excretion) and drug-likeness of the compound, predictions were performed using the SwissADME web tool. The results indicated that Rnaseh2c-In1 exhibits favorable ADME characteristics (figure 7F) and drug-like properties (figure 7F). Toxicity prediction using ProTox V.3.0 showed that Rnaseh2c-In1 falls within acceptable ranges for hepatotoxicity, neurotoxicity, cytotoxicity, mutagenicity, immunotoxicity, and carcinogenicity (figure 7F). These results suggest that Rnaseh2c-In1 possesses promising ADMET properties. To further verify the reliability of the screening and predict the binding stability between Rnaseh2c and Rnaseh2c-In1, 100-ns molecular dynamics (MD) simulations were performed following molecular docking. The root mean square deviation(RMSD) analysis indicated that the Rnaseh2c-Rnaseh2c-In1 complex underwent rapid conformational equilibration and subsequently reached a stable state (figure 7G). Meanwhile, the small molecule Rnaseh2c-In1 remained stably positioned within the binding pocket (figure 7H). Both Rg (figure 7I) and solvent accessible surface area (SASA) (figure 7J) analyses demonstrated that Rnaseh2c maintained a compact conformation on binding to Rnaseh2c-In1. Root mean square fluctuation (RMSF) analysis revealed no substantial fluctuations in the active-site region of Rnaseh2c (figure 7K), indicating stable binding with Rnaseh2c-In1. Throughout the simulation, Rnaseh2c-In1 consistently formed 3–4 hydrogen bonds with Rnaseh2c (figure 7L), suggesting persistent and crucial polar interactions. Gibbs free energy landscape analysis showed that the Rnaseh2c-Rnaseh2c-In1 complex remained predominantly within one major conformational cluster (figure 7M), reflecting high conformational stability. We calculated the binding free energy of the complex using molecular mechanics/poisson-boltzmann surface area (MM/PBSA), and the results showed that the binding process was thermodynamically significantly favorable (ΔGbind=−29.89 kcal/mol). Rnaseh2c and Rnaseh2c-In1 have good spatial complementarity and charge matching (figure 7N). To evaluate the binding specificity of Rnaseh2c-In1, we selected compound “8019-3662” with the highest structural similarity to Rnaseh2c-In1 based on Tanimoto similarity coefficient as the negative control (online supplemental Table S2), and performed MD simulations on its interaction with Rnaseh2c. The results showed that compared with the Rnaseh2c-8019-3662 complex system, the Rnaseh2c-Rnaseh2c-In1 complex system exhibited stronger binding affinity and more stable interactions (online supplemental Figure S7A-I). These results demonstrate stable binding between Rnaseh2c and Rnaseh2c-In1. Rnaseh2c-In1 significantly suppresses HCC progression To analyze the clinical significance of Rnaseh2c, we further investigated the relationship between Rnaseh2c expression and the characteristics of HCC. Analysis of the Cancer Genome Atlas-Liver Hepatocellular Carcinoma (TCGA-LIHC) dataset using R revealed that Rnaseh2c expression was significantly elevated in tumor tissues compared with adjacent tissues (figure 8A). Patients in the TCGA-LIHC cohort were then stratified into high and low Rnaseh2c expression groups. Survival analysis indicated that high Rnaseh2c expression was associated with poorer overall survival (figure 8B). Furthermore, Rnaseh2c exhibited minimal insertion/deletion mutations or single nucleotide polymorphisms (figure 8C). We further used the tumor immune dysfunction and exclusion (TIDE) algorithm to predict the response of Rnaseh2c expression to immune checkpoint blockade (ICB) therapy. The results showed that the TIDE score was higher in the group with high expression of Rnaseh2c (online supplemental Figure S8A), indicating that patients with high expression of Rnaseh2c had poor ICB efficacy and shorter survival after ICB treatment. In addition, compared with the low expression group of Rnaseh2c, the high expression group of Rnaseh2c had higher expression levels of immune checkpoints (ITPRIPL1, TIGIT, PDCD1, CTLA4, and LAG3) (online supplemental Figure S8B). These results indirectly suggest that Rnaseh2c may be related to the immunosuppressive microenvironment and potential resistance to immunotherapy. Based on these findings, we performed immunohistochemistry to assess Rnaseh2c protein expression in HCC tissues. Consistent with the mRNA data, Rnaseh2c protein levels were significantly higher in carcinoma tissues than in par_carcinoma tissues (figure 8D). Correlation analysis of clinical features from patients with HCC showed that Rnaseh2c expression levels positively correlated with both tumor size and alpha-fetoprotein values (figure 8E). Additionally, we established an Hep53.4 cell line overexpressing Rnaseh2c and generated an orthotopic HCC mouse model. Rnaseh2c overexpression significantly promoted tumor growth and worsened mouse prognosis (figure 8F). These results demonstrate that Rnaseh2c exerts protumorigenic effects not only in macrophages but also in hepatocytes. Rnaseh2c-In1 significantly suppresses HCC progression. (A) Analysis of Rnaseh2c gene expression in the TCGA-LIHC dataset using R. G1, G2, and G3 represent pT stages T1, T2, and T3, respectively. “Normal” denotes adjacent tissue samples from the TCGA database. (B) Kaplan-Meier survival curves for high and low Rnaseh2c expression groups in the TCGA dataset. (C) Mutation profile of Rnaseh2c mRNA in different sample groups. (D) Immunohistochemical detection of Rnaseh2c expression in par_carcinoma and carcinoma tissues. Bar=50 µM Data were mean±SD and analyzed by paired t-test. ***p<0.001. (E) Correlation analysis between Rnaseh2c expression levels and tumor volume or AFP values. (F) Effect of Rnaseh2c overexpression on HCC growth and prognosis in mice. Data analysis was conducted using Kaplan-Meier. (G) H&E staining showing the histopathological structure of mouse liver, kidney, and spleen after Rnaseh2c-In1 treatment. (H) and (I) Effect of Rnaseh2c-In1 on the growth of subcutaneous HCC in mice. The data of the growth curve was analyzed using two-way ANOVA. Tumor weight analysis was performed using the t-test. ***p<0.001. (J) TUNEL staining analysis of apoptosis levels. Data were mean±SD and analyzed by t-test. (K) ELISA analysis of serum CCL2 secretion levels in mice. Data were mean±SD and analyzed by t-test. ***p<0.001. (L) FCM analysis of CD8+ Tex cell infiltration levels. Data were mean±SD and analyzed by t-test. ***p<0.001. (M) The pattern diagram illustrates the mechanism by which Rnaseh2c promotes HCC progression. This image was drawn by Figdraw. AFP, alpha-fetoprotein; ANOVA, analysis of variance; FCM, flow cytometry; HCC, hepatocellular carcinoma; mRNA, messenger RNA; PBS, phosphate-buffered saline; PD-1, programmed cell death protein 1; Prolif Ms, proliferative macrophages; Rnaseh2c, ribonuclease H2 subunit C; Tex, exhausted T cell. To evaluate the impact of targeting Rnaseh2c on HCC progression, we first assessed the in vivo safety profile of the inhibitor Rnaseh2c-In1. No histological damage to the liver, kidney, or spleen was observed in mice treated with Rnaseh2c-In1 at doses ranging from 0 to 100 mg/kg (figure 8G). We next established a subcutaneous HCC model using Hep53.4 cells. Treatment with Rnaseh2c-In1 (10 mg/kg) significantly inhibited HCC growth (figure 8H,I). Immunoblot analysis confirmed that Rnaseh2c-In1 reduced the protein expression levels of Cdk1, Cdk9 and Mki67 (online supplemental Figure S9A). Terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL) staining indicated that Rnaseh2c-In1 markedly promoted apoptosis (figure 8J). ELISA analysis showed that Rnaseh2c-In1 suppressed serum CCL2 secretion levels in mice (figure 8K). FCM revealed that Rnaseh2c-In1 decreased the infiltration level of CD8+ Tex cells (figure 8L), but did not affect the infiltration levels of myeloid-derived suppressor cells (MDSCs), regulatory T Cells (Tregs), and neutrophils (online supplemental Figure S9B). These results indicate that Rnaseh2c-In1 also significantly suppresses HCC progression in vivo. Discussion In this study, we identified Prolif Ms subpopulation in the HCC microenvironment. Rnaseh2c encodes a scaffolding subunit of the ribonuclease H2 (RNase H2) enzyme. Mutations in Rnaseh2c can lead to an autosomal recessive neurological autoinflammatory disorder called Aicardi-Goutières syndrome.33 Multiple studies have shown an association between RNase H2 gene expression and various types of cancer.3436 The shRNA-mediated knockdown of Rnaseh2c expression led to a decrease in the cell survival rate and an increase in the percentage of apoptotic cells.37 We demonstrated that Prolif Ms exhibit distinct expression of proliferation markers including Cdk1 and Mki67, which functionally contribute to tumor progression. This finding stands in sharp contrast to Campbell et al’s research in breast cancer, where they reported that promacs were associated with high-grade, hormone receptor-negative tumors, but did not show a direct association with tumor size, stage, or lymph node metastasis.14 Our findings further reveal that Rnaseh2c drives a dual regulatory program within Prolif Ms: it promotes cellular proliferation by directly binding to the promoter of Cdk9, while simultaneously impairing key antitumor effector functions such as endocytosis and antigen presentation. Furthermore, we also showed that Prolif Ms recruits CD8+ Tex cells into the TME through the CCL2/CCR2 axis, thereby accelerating HCC progression. Finally, we developed a novel small-molecule inhibitor, Rnaseh2c-In1, which effectively targets this axis and suppresses tumor growth in vivo. The regulatory role of Rnaseh2c in this process is particularly noteworthy. Although Rnaseh2c is a subunit of the RNase H2 complex involved in nucleic acid metabolism and genomic stability, our data suggest a non-canonical, enzymatic-activity-independent function in transcriptional regulation. Similar non-enzymatic roles have been observed for other ribonucleases; for example, RNase1 can modulate cytotoxic T lymphocyte (CTL) responses by interfering with STAT1 nuclear translocation38; the downregulation of Rnaseh2c in breast cancer can trigger CTL-mediated immune responses.39 Here, Rnaseh2c acts as a transcriptional regulator that directly activates Cdk9, a key cyclin-dependent kinase involved in cell cycle progression. Concurrently, Rnaseh2c suppresses genes related to phagocytosis and antigen presentation, effectively crippling the ability of Prolif Ms to engage in antitumor immunity. This “pro-proliferation & anti-function” duality presents a sophisticated immune evasion mechanism, wherein the macrophage subset expands numerically but loses its capacity to perform tumor-suppressive duties. The functional output of Prolif Ms is mediated through the CCL2/CCR2 chemokine axis. We demonstrate that Rnaseh2c-high Prolif Ms secrete CCL2, which recruits CCR2+ CD8+ Tex cells into tumors. CD8+ Tex cells are pivotal mediators of T-cell dysfunction and contribute to immune tolerance in HCC.40 41 This mechanistic link between Prolif Ms and T-cell exhaustion provides a plausible explanation for the poor prognosis associated with high Rnaseh2c expression in TCGA-LIHC data. Importantly, genetic ablation or pharmacological inhibition of CCR2, but not CXCR6, reversed CD8+ Tex accumulation and tumor growth, underscoring the specificity of this pathway. These observations position the Rnaseh2c-CCL2-CCR2 axis as a promising target for disrupting the immunosuppressive niche in HCC. From a translational perspective, our development of Rnaseh2c-In1 demonstrates the therapeutic potential of targeting this pathway. Virtual screening followed by MD simulations identified a stable inhibitor that binds Rnaseh2c with high affinity. In vivo, Rnaseh2c-In1 treatment reduced Prolif Ms expansion, decreased CD8+ Tex infiltration, lowered serum CCL2 levels, and significantly suppressed HCC growth without observable organ toxicity. This inhibitor, therefore, represents a prototype for a novel class of agents that selectively modulate macrophage subpopulations within tumors. Given the central role of TAMs in resistance to current immunotherapies,42 43 strategies to eliminate specific TAM subsets have gained considerable interest. Clinical trials combining CSF1R inhibitors with immune checkpoint blockers are underway,44 45 and chimeric antigen receptor (CAR)-macrophage therapies are entering early-phase testing.4648 A previous study showed that targeting proliferating macrophages (F480+Ki67+) improves CD8+ T-cell infiltration and enhances a CD8+T cell-dependent antitumor effects in mouse pancreatic ductal adenocarcinoma models.49 Our work suggests that targeting proliferating macrophages via Rnaseh2c could complement these approaches, particularly in HCC where this subset appears to drive immunosuppression. Several limitations and future directions should be noted. First, while we validated the presence of Prolif Ms in human HCC tissues, larger multicenter cohorts are needed to definitively correlate their abundance with clinical stage, recurrence, and response to therapy. Second, the precise upstream signals that induce Rnaseh2c expression in macrophages within the HCC niche remain to be elucidated. Third, given the potential functional plasticity of Prolif Ms, it will be important to investigate whether these cells can be repolarized toward an antitumor phenotype. Finally, the synergistic effect between Rnaseh2c-In1 and existing immune therapies (such as programmed cell death protein 1/programmed death-ligand 1 inhibitors) deserves systematic evaluation in preclinical models and further validation in subsequent cohorts receiving immune checkpoint inhibitor treatment. In conclusion, our study delineates a novel Prolif Ms subpopulation in HCC that is governed by the transcriptional regulator Rnaseh2c. Through a dual mechanism that enhances self-renewal while inhibiting immune function, and by recruiting CD8+ Tex cells via CCL2/CCR2 signaling, Prolif Ms foster an immunosuppressive microenvironment conducive to tumor progression. The development of Rnaseh2c-In1 provides a proof of concept for therapeutic targeting of this pathway. These findings advance our understanding of macrophage heterogeneity in liver cancer and highlight a promising strategy for restoring antitumor immunity by intercepting the proliferative-immune axis within the TME. Supplementary material