Introduction

Protein arginine methyltransferase 5 (PRMT5) is a highly conserved enzyme that catalyzes the symmetric di-methylation of arginine residues on histones and some non-histone proteins. PRMT5 is involved in a variety of cellular processes, including chromatin remodeling, transcriptional regulation, RNA processing, cell proliferation and differentiation.1 Its diverse effects on RNA metabolism highlight its importance in gene expression regulation at multiple levels. For example, PRMT5-mediated methylation of the splicing regulator SRSF1 modulates its activity and promotes exon skipping.2 Additionally, PRMT5 methylates the RNA-binding protein HNRNPA1, implicated in alternative splicing regulation, messenger RNA (mRNA) stabilization, and translation.3 Furthermore, PRMT5 has been shown to methylate several transcription factors and epigenetic modifiers, including P53, NF-κB, E2F1, DNMT1, and EZH2, thereby regulating their activity.4–8 Dysregulation of PRMT5 activity has been implicated in a variety of human diseases, including neurological disorders such as Alzheimer’s disease and multiple sclerosis,9 10 as well as cancers like leukemia, lymphoma, lung cancer, and breast cancer.2 11–13 Recent research indicated that targeting PRMT5 with small molecule inhibitors holds promise for treating various types of cancer. A number of PRMT5 inhibitors such as PF-06939999 and JNJ-64619178 have been developed and tested in preclinical studies.14 15 Moreover, the second-generation PRMT5 inhibitors like AMG193 and MRTX1719 were safe and demonstrated preliminary antitumor activity in patients with solid tumors with methylthioadenosine phosphorylase (MTAP) deficiency.16 17

PRMT5 also plays a critical role in regulating the immune system. Research from our group and others has demonstrated PRMT5’s involvement in the differentiation, activation, and function of T and B cells. Specific deletion of Prmt5 in mouse T cells induces their terminal differentiation and accelerates tumor progression.18 Conversely, Prmt5 deletion in B cells increases Ccl22 expression, attracting natural killer (NK) cells to the tumor microenvironment (TME) and delaying tumor progression.19 These findings suggest that PRMT5 has distinct effects on different immune cell subsets, influencing tumor dynamics in varied ways. However, the impact of PRMT5 inhibition on myeloid cells and their contribution to tumor progression and immune modulation remains unclear.

Macrophages are key immune cells that play critical roles in both innate and adaptive immune responses. They differentiate from circulating monocytes in response to specific signals20 21 and are broadly categorized into two main statuses: classically activated macrophages (M1) and alternatively activated macrophages (M2). Tumor-associated macrophages (TAMs) are predominantly skewed towards the M2 phenotype, which promotes tumor progression through angiogenesis, immune suppression, and facilitation of tumor cell invasion and metastasis. Reprograming TAMs from a protumor M2-like status to an antitumor M1-like status is a major area of interest for developing novel cancer immunotherapies.

In this study, we first analyzed single-cell RNA sequencing (RNA-seq) data of PRMT5 expression from pan-cancer tumor-infiltrating monocyte-macrophage subsets, identifying high PRMT5 expression in the PPARG-macrophage subset. Additionally, macrophages expressing high levels of PRMT5 correlate with poor survival outcomes. We then employed Lysm-cre-Prmt5fl/fl mice and PRMT5-specific inhibitors to elucidate the role of PRMT5 in macrophage polarization and its effects on the TME. Our results demonstrate that Prmt5 deletion or PRMT5 inhibitor treatment significantly reduces M2 macrophage polarization in vitro, primarily through alterations in cholesterol metabolism via the STAT6-PPARγ pathway. In vivo allograft models showed that Prmt5 deletion in myeloid cells reprograms macrophages and induces an antitumor phenotype. Moreover, Prmt5 deletion in myeloid cells induced programmed death-ligand 1 (PD-L1) expression, and combining PRMT5 inhibition with anti-PD-L1 therapy showed synergistic effects on suppressing tumor growth. Our study unveils PRMT5 as a key regulator of macrophage differentiation and function, highlighting its potential as a therapeutic target not only for directly targeting cancer cells, but also for reshaping the TME to promote an antitumor response.

ResultsPRMT5 expressed highly in PPARG-macrophage subset and correlated with poor cancer survival

We first analyzed PRMT5 expression in peripheral blood mononuclear cells (PBMCs) using publicly available data sets. First, single-cell RNA-seq data revealed that monocytes are among the cell types with the high levels of PRMT5 (online supplemental figure 1). Additionally, PRMT5 is upregulated in inflammatory macrophages compared with monocytes (online supplemental figure 1B), based on a separate microarray data set (GSE5099). In mouse allograft colorectal cancer models of MC38, we observed significantly higher PRMT5 expression in TAMs compared with spleen macrophages (figure 1A). Multicolor immunofluorescence analysis of human colorectal cancer tissue further confirmed elevated PRMT5 expression in CD206+ macrophages compared with CD206− macrophage (figure 1B). Furthermore, we analyzed PRMT5 expression in infiltrating monocyte-derived macrophage subsets across various cancers, including lung, liver, breast, colorectal, renal, and esophageal cancers from publicly available pan-cancer single-cell transcriptomic data (GSE154763). Interestingly, analysis revealed that PRMT5 is most highly expressed in PPARG+tumor-infiltrating macrophages, which co-express MRC1 (CD206) (figure 1C), indicating that these cells constitute an M2-related macrophage genes. Moreover, PPARG+macrophages from tumor tissue expressed higher levels of PRMT5 than those from normal tissue (figure 1D). We then examined the relationship between PRMT5 expression in tumor-infiltrating macrophages and survival outcomes across six cancer types (lung adenocarcinoma, colorectal cancer, head and neck cancer, breast cancer, liver cancer, and skin cancer) using the GEPIA database (http://gepia2.cancer-pku.cn/). Higher PRMT5 expression in macrophages correlates with shorter survival times in patients with cancer (figure 1E and online supplemental figure 1C). These data suggested that PRMT5 may be involved in macrophage differentiation, and play a critical role in the function of TAM.

Figure 1

PRMT5 expressed highly in macrophages and was involved in the monocyte-to-macrophage differentiation process. (A) Spleen macrophage (SPM) and tumor-associated macrophage (TAM) were isolated from MC38 tumor-bearing C57BL/6 mice (n=3), Prmt5 expression was determined by qPCR. (B) PRMT5, CD68, CD206 expression in human colorectal tissue by multicolor IHC staining was measured. Representative data from two patients were shown. (C) PRMT5 and related genes in macrophage subpopulations across 15 types of tumors were analyzed from pan-cancer single-cell transcriptomic data (GSE154763). (D) PRMT5 expression in distinct subsets of tumor infiltrating macrophages and the corresponding subsets in normal tissues from pan-cancer single-cell transcriptomic data. (E) The relationship of PRMT5 expression levels in tumor-infiltrating macrophages and survival outcomes in colorectal cancer (COAD) and lung adenocarcinoma (LUAD) using the (http://gepia2.cancer-pku.cn/) database website. (F, G) Monocytes isolated from human PBMC were stimulated with M-CSF for different times. PRMT5 mRNA expression was measured by qPCR (F) (n=5–6); PRMT5 protein expression was measured by western blot (G); the data represented from the two experiments were shown. (H) Monocytes isolated from human PBMC were stimulated with M-CSF and treated with GSK591 or DMSO, CD68 and CD36 expression were measured by flow cytometry (n=6). The mean±SD was graphed. Statistical differences were determined by one-way analysis of variance (F) or the two-tailed unpaired Student’s t-test (A, B, H), *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. IHC, Immunohistochemistry; M-CSF, macrophage colony-stimulating factor; MFI, Mean Fluorescence Intensity; mRNA, messenger RNA; PBMC, peripheral blood mononuclear cell; PRMT5, protein arginine methyltransferase 5; qPCR, quantitative PCR.

To further validate these findings, we isolated human monocytes from healthy donors and stimulated them with macrophage colony-stimulating factor (M-CSF) to induce macrophages. Under M-CSF stimulation, PRMT5 expression was upregulated at both mRNA and protein levels with a surge at mRNA levels at day 1. However, the levels of transcription and protein increase are not temporally consistent (figure 1F,G). Treatment with PRMT5-specific inhibitor GSK591 at the onset of culture reduced PRMT5 activity, as evidenced by decreased dimethyl symmetric arginine (SDMA) level (online supplemental figure 1D). The treatment also reduced the expression of macrophage-related genes like CD36 and CD68 (figure 1H). These in vitro findings support the role of PRMT5 in macrophage differentiation.

Prmt5 deletion impaired the macrophage differentiation and migration

To further determine whether PRMT5 similarly impacts monocyte differentiation in mice, we isolated mouse bone marrow monocytes and cultured them with M-CSF. Consistently, Prmt5 expression is upregulated when mouse monocytes are stimulated with M-CSF (figure 2A). To assess the role of PRMT5 in monocyte-macrophage differentiation in vivo, we crossed Lysm-cre transgenic mice with Prmt5fl/fl mice to generate Lysm-cre-Prmt5fl/fl conditional knockout mice (Prmt5 cKO) (figure 2B). Prmt5 mRNA expression in bone marrow-derived macrophages (BMDMs) of Prmt5 cKO mice was significantly reduced compared with their wild type littermates, confirming the successful construction of the cKO model (figure 2C).

Figure 2

PRMT5 regulates macrophage differentiation and migration. (A) Bone marrow monocytes were isolated from C57BL/6 (n=6–7) and stimulated with M-CSF with differentiation time, Prmt5 expression was determined by qPCR. (B) Schematic of Prmt5 genomic locus targeting strategy from the European Conditional Mouse Mutagenesis Program. Exon 7 of Prmt5 was flanked by loxP sites and then crossed with Lysm-Cre mice to produce Lysm-cre-Prmt5fl/fl mice (Prmt5 cKO) and littermate control mice. (C) BMDMs isolated from Prmt5 cKO and control mice, Prmt5 expression was measured by real-time qPCR (n=6). (D) Flow cytometry analysis of Ly6C+Ly6G- monocytes in control (n=7) and Prmt5 cKO mice (n=5). The mean±SD was graphed. (E) The frequencies of CD11b+F4/80+cells gated in Ly6C+Ly6G− cells from control (n=6) and Prmt5 cKO (n=6) mouse spleens. (F) The frequencies of CD11b+F4/80+cells from control (n=6) and Prmt5 cKO (n=5) mouse peritoneal cavity. (G) The frequencies of CCR2 expression in Ly6C+cells from control (n=9) and Prmt5 cKO (n=6) mouse blood. (H–J) Control (n=7) and Prmt5 cKO mice (n=7) were injected with thioglycolate for 3 days, the percentages of B cells, T cells and CD11b+myeloid cells in CD45+cells were determined by flow cytometry (H); the percentages of F4/80+macrophages, Ly6C+monocytes and Ly6G+neutrophils in peritoneal cavity (I); absolute cell number of macrophages in peritoneal cavity (J). The mean±SD was graphed. Statistical differences were determined by one-way analysis of variance (A) or the two-tailed unpaired Student’s t-test (C–J), *p<0.05, **p<0.01, ****p<0.0001. cKO, conditional knockout; BMDM, bone marrow-derived macrophages; FACS, Fluorescence - activated Cell Sorting; M-CSF, macrophage colony-stimulating factor; PRMT5, protein arginine methyltransferase 5; qPCR, quantitative PCR.

We then examined the phenotype of Prmt5 cKO mice. Analysis of CD11b+myeloid cells and monocytes in the bone marrow showed no significant differences between Prmt5 cKO and control groups (online supplemental figure 2A, B). Meanwhile, percentages of CD11b+myeloid cells and dendritic cells (DC) in spleens showed no differences between Prmt5 cKO and control groups (online supplemental figure 2C, D). However, we observed a decreased frequency of monocytes (CD11b+Ly6C+Ly6G–) and macrophages (F4/80+) in the spleens and peritoneal cavities of Prmt5 cKO mice, compared with controls (figure 2D–F). No significant difference between the two groups was observed on neutrophils, T and B cells (online supplemental figure 2E–G).

CCL2 has been reported to interact with CCR2 on mono/macrophages, mediating their migration to sites of immune response.22 We found a significant decrease in CCR2 expression in Ly6C+cells from the blood of Prmt5 cKO mice compared with control mice (figure 2G). To further investigate macrophage recruitment and migration, we induced peritonitis in mice using thioglycolate.23 24 3 days post-injection, the percentage of CD11b+cells in the peritoneal cavity was significantly lower in Prmt5 cKO mice compared with controls, suggesting impaired myeloid cell migration (figure 2H). Moreover, both the percentage and absolute number of macrophages in the peritoneal lavage fluid were significantly lower in Prmt5 cKO mice compared with controls (figure 2I,J). However, no differences in apoptosis or proliferation were observed in peritoneal macrophages between the two groups (online supplemental figure 2H, I), suggesting that the reduction in macrophages within the peritoneal cavity was primarily due to impaired myeloid cell migration.

PRMT5 promotes macrophage polarization towards the M2 phenotype

We further investigated the role of PRMT5 in macrophage polarization. BMDMs were isolated from Prmt5 cKO and control mice and stimulated towards either M1 or M2 polarization for 48 hours. Interestingly, while M1-associated genes such as Il6, Tnf, and Inos showed no significant differences between the two groups, the co-stimulatory molecule CD80 was significantly upregulated in Prmt5 cKO mice (figure 3A and online supplemental figure 3A). Conversely, M2-related genes Arg1, Ym1, and Mrc1 were significantly downregulated in Prmt5 cKO mice, correlating with lower CD206 expression as measured by flow cytometry (figure 3B,C). Moreover, SDMA levels were significantly reduced in the M2 inducing system in Prmt5 cKO mice (figure 3D) compared with the control mice, suggesting that PRMT5 plays a more prominent role in M2 differentiation than M1 differentiation.

Figure 3

Prmt5 promotes macrophage polarization towards the M2 phenotype. (A–D) Bone marrow cells were isolated from Prmt5 cKO and control mice, then induced to differentiate into macrophages and subsequently stimulated towards M1 or M2 polarization. CD80 expression was assessed by flow cytometry (A); M2 markers including Arg1, Ym1, Mrc1 expressions were measured by real-time qPCR (B); CD206 expression was measured by flow cytometry (C); Prmt5 and SDMA expression in M0, M1, M2 macrophages were determined by western blot. The data represented from the two experiments were shown (D). (E–G) Bone marrow-derived macrophages isolated from C57BL/6 mice were stimulated towards M1 or M2 polarization treated with DMSO or GSK591. CD80 expression was assessed by flow cytometry (E); M2 markers including Arg1, Ym1, and Mrc1 expressions were measured by real-time qPCR (F); CD206 expression was measured by flow cytometry (G). (H) Human monocyte-derived macrophages from healthy donors’ blood treated with GSK591 during M1 and M2 polarization, CD206 and CCL18 expressions were measured by real-time qPCR. The mean±SD was graphed. Statistical differences were determined by one-way analysis of variance, *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. cKO, conditional knockout; mRNA, messenger RNA; MFI, Mean Fluorescence Intensity; PRMT5, protein arginine methyltransferase 5; qPCR, quantitative PCR; SDMA, dimethyl symmetric arginine.

To further explore this, PRMT5-specific inhibitor GSK591 was applied in the M1/M2 polarization experiments. We generated three types of macrophages—mouse BMDMs, THP-1-derived macrophages, and human Mo-macs. Cells were treated with or without GSK591 for 2 days, followed by stimulation with specific cytokines to induce M1 or M2 polarization (online supplemental figure 3B). Consistently, GSK591 treatment during BMDM polarization significantly increased CD80 expression, but not other M1-related molecules such as Tnf and Inos (figure 3E and online supplemental figure 3C). M2-related genes Arg1, Ym1, and Mrc1 were significantly downregulated in GSK591-treated M2-polarized macrophages, with reduced CD206 expression confirmed by flow cytometry (figure 3F,G). Additionally, human Mo-macs isolated from healthy donors and treated with GSK591 during M1 and M2 polarization exhibited downregulated CD206 and CCL18 expressions in M2-polarized macrophages, while M1-related genes were unaffected (figure 3H and online supplemental figure 3D). Similarly, THP-1 cells stimulated into macrophages with PMA (Phorbol 12 - Myristate 13 - Acetate) and then polarized into M1 or M2 showed that GSK591 treatment reduced the expression of M2-related markers without affecting M1-related genes (online supplemental figure 3E, F). These findings indicate that PRMT5 is essential for macrophage polarization towards the M2 phenotype.

To further examine whether the function of macrophages was altered after knockdown of PRMT5, first, we established a macrophage-T cell co-culture system. We found that the proportion of naïve T cells, central memory T cells and Tregs (Regulatory T cells) showed no significant difference between the control and Prmt5 cKO group, while effector memory T cells were elevated in the Prmt5 cKO group (online supplemental figure 4A–D). Meanwhile, reducing PRMT5 expression in BMDMs had little effect on T-cell cytotoxicity, since the expression of interferon (IFN)-γ and tumor necrosis factor (TNF)-α in T cells showed no significant difference between the two groups (online supplemental figure 4E,F). To further investigate the impact of macrophages on tumor cells, we established a co-culture system of BMDMs and MC38 cells in transwell chambers for 3 days and found that control M2 macrophages promote MC38 cell proliferation, compared with M0 macrophages, while no significant difference was exhibited in M0 and M2 BMDMs from the Prmt5 cKO group (online supplemental figure 4G).

Prmt5 regulates lipid metabolism in BMDMs and promotes M2 polarization via the pSTAT6/PPARγ axis

To further investigate the mechanism by which Prmt5 regulates macrophage polarization, we performed RNA-seq with BMDMs isolated from Prmt5 cKO and control mice. Transcriptomic analysis revealed 1146 upregulated and 166 downregulated genes in Prmt5 cKO BMDMs (online supplemental figure 5A), indicating a role for Prmt5 in gene transcriptional repression, consistent with previous reports.4 25 26 Gene Set Enrichment Analysis (GSEA) demonstrated that the downregulated genes were enriched in lipid metabolism pathways, including PPAR signaling, arachidonic acid metabolism, linolenic acid metabolism, arginine biosynthesis (online supplemental figure 5B,C), all of which are associated with M2 differentiation.27

Next, we stimulated BMDMs toward M2 differentiation and performed RNA-seq. In this context, 643 genes were upregulated while 318 genes were downregulated in Prmt5 cKO BMDMs compared with controls (figure 4A). GSEA indicated that the downregulated genes were enriched in cholesterol metabolism pathways (figure 4B,C). Additionally, we used the STRING database to investigate the gene regulatory networks of differentially expressed genes of control and Prmt5 cKO M2 BMDMs. As a result, we found that genes related to macrophage polarization and lipid metabolism were included, such as Chil3, Csf1, Tnf, Cd80, Apoe. Besides, chemokines and chemokine receptors, such as Cxcr6, Cxcl14, Ccr3, were also involved in the regulatory network (online supplemental figure 5D). A heat map further highlighted that M2-related genes, including Mrc1, Chil3, Csf1, and Pparg, were downregulated in Prmt5 cKO M2 BMDMs (figure 4D). Lipid and cholesterol metabolism are key regulators of macrophage activation and function. ATP binding cassette subfamily A member 1 (Abca1) and subfamily G member 1 (Abcg1) are crucial genes involved in cholesterol efflux and cholesterol metabolism, along with PPAR-γ.28 We analyzed the expression of these genes in our RNA-seq analysis of M2 BMDMs from control and Prmt5 cKO mice. Consistently, cholesterol metabolism-related genes, including Abca1, Abcg1, Eepd1, Cyp27a1, and Scd1, were significantly reduced in Prmt5 cKO BMDMs (figure 4E). We verified the downregulation of these genes by quantitative PCR (qPCR) in both Prmt5 cKO M2 BMDMs and GSK591-treated M2 BMDMs (figure 4F,G). These results indicated that Prmt5 regulates lipid metabolism during M2 differentiation.

Figure 4

Prmt5 regulates lipid metabolism in BMDMs and promotes M2 polarization through the pSTAT6–PPARγ axis. Bone marrow cells were isolated from Prmt5 cKO and control mice, then induced to differentiate into macrophages and polarized into M2 phenotype, then for RNA sequencing. (A) Volcano plot showing the upregulated and downregulated genes in M2 BMDMs between two groups. (B) M2 BMDMs gene expression profiles were analyzed by Gene Set Enrichment Analysis, upregulated and downregulated pathways were shown by the normalized enrichment score. (C) Enrichment plots of cholesterol processes in M2 BMDMs between the two groups. (D) Heatmap of M2-related gene expression in M2 BMDMs from control and Prmt5 cKO mice. (E) Heatmap of genes associated with cholesterol metabolism in M2 BMDMs from control and Prmt5 cKO mice. (F) Eepd1, Cyp27a1 and Scd1 expression in M2 BMDMs from control and Prmt5 cKO mice by qPCR. (G) M2 BMDMs were induced from C57BL/6 mice, and treated with GSK591 or DMSO; Eepd1, Cyp27a1 and Scd1 expression were measured by qPCR. (H, I) M2 BMDMs were induced from control and Prmt5 cKO mice, Pparg expression was measured by qPCR (H); PPARγ and ARG1 expression were measured by western blot (I). (J, K) M2 BMDMs were induced from C57BL/6 mice, and treated with GSK591 or DMSO; Pparg expression was measured by qPCR (J); PPARγ and ARG1 expression were measured by western blot (K). (L) BMDMs were generated from control and Prmt5 cKO mice and stimulated with IL-4 and IL-13 for 12 hours, p-STAT6 expression was measured by western blot. (M) BMDMs were generated from C57/BL6 mice and stimulated with IL-4 and IL-13 for 12 hours treated with GSK591 or DMSO, p-STAT6 expression was measured by western blot. (N) Human Mo-macs from healthy donors’ blood treated with GSK591 and stimulated with IL-4 and IL-13 for 12 hours, p-STAT6 expression were measured by western blot. (O–P) TCGA database to analyze the correlation between PRMT5 and PPARG expression, PRMT5 and STAT6 expression in colorectal cancer (COAD) (O) and lung adenocarcinoma (LUAD) (P). The mean±SD was graphed. Statistical differences were determined by unpaired Student’s t-test (F, G) or one-way analysis of variance (H, J), *p<0.05, **p<0.01, ***p<0.001. One of the three similar representative data was shown. BMDM, bone marrow-derived macrophages; cKO, conditional knockout; IL, interleukin; mRNA, messenger RNA; PRMT5, protein arginine methyltransferase 5; qPCR, quantitative PCR;TCGA, The Cancer Genome Atlas.

PPARγ has been reported to promote M2 polarization through lipid metabolism, which is essential for TAM protumor activation.29–31 We verified the expression of PPARγ using RT-PCR (Reverse Transcription - Polymerase Chain Reaction) and western blot, confirming that its expression was indeed decreased in the Prmt5 cKO group (figure 4H,I). Similarly, GSK591 treatment led to a significantly decreased PPARγ expression at both the transcriptional and protein levels (figure 4J,K). Moreover, the M2-related ARG1 was downregulated in response to Prmt5 deletion or inhibition (figure 4I,K). It has been reported that the pSTAT6/PPARγ signaling pathway is involved in M2 polarization.32 We first observed that p-STAT6 was downregulated in M2-polarized BMDMs from Prmt5 cKO mice compared with control mice (figure 4L). Consistent with this, STAT6 phosphorylation was decreased in GSK591-treated BMDMs, human Mo-macs, and THP-1-derived macrophages (figure 4M,N and online supplemental figure 5E). Moreover, suppressor of cytokine signaling 1 (SOCS1), the transcription factor which is a negative regulator of phosphorylation of STAT633 was also detected. Data showed that SOCS1 expression was upregulated on interleukin (IL)-4 and IL-13 stimulation, while Prmt5 cKO BMDMs displayed higher SOCS1 expression levels than the control group (online supplemental figure 5F, G). Additionally, we analyzed RNA-seq data from the TCGA (The Cancer Genome Atlas) database to explore the relationship between the expression of PRMT5 and PPARG or STAT6 in tumor-infiltrating macrophages. The analysis revealed a positive correlation between PRMT5 and both PPARG and STAT6 expression across several tumor types, such as Lung Adenocarcinoma (LUAD) and colorectal cancer (figure 4O,P). These findings suggest that PRMT5 promotes macrophage M2 polarization by regulating cholesterol metabolic reprogramming through PPARγ expression.

Myeloid cell-specific deletion of Prmt5 inhibits tumor progression

Our previous study indicated that Prmt5 promotes M2 macrophage differentiation in vitro (figure 3). To investigate the impact of myeloid cell-specific deletion of Prmt5 on tumor development, we implanted syngeneic MC38 colorectal cancer cells subcutaneously into Prmt5 cKO mice and littermate controls. As expected, Prmt5 cKO mice developed smaller tumors than the control group (figure 5A,B). We analyzed the tumor-infiltrating CD45+immune cells, using the gating strategy shown in online supplemental figure 6A, B. No significant differences were observed in the proportions of CD3+T cells, NK cells, or CD4+Foxp3+ Treg cells between the two groups (online supplemental figure 6C, D). However, we observed a decrease of macrophages and an increase in neutrophils in the tumor tissue of Prmt5 cKO mice (figure 5C and online supplemental figure 6E). Additionally, CD206 expression was decreased in Prmt5 cKO macrophages (figure 5D), while TNF-α expression was elevated (figure 5E and online supplemental figure 6F), suggesting that the reduction of M2 macrophages in Prmt5 cKO tumors may contribute to the inhibition of tumor progression. We further found reducing CCR2 expression in macrophages from the Prmt5 cKO group (online supplemental figure 6G). Moreover, we implanted syngeneic LLC (Lewis Lung Carcinoma) lung cancer cells subcutaneously into Prmt5 cKO mice and their littermate controls. Consistent with the MC38 tumor models, Prmt5 cKO in host suppressed tumor growth compared with the control hosts (figure 5F,G), and the percentage of TAMs was also reduced (figure 5H and online supplemental figure 6H) while TNF-α expression was elevated in Prmt5 cKO macrophages (figure 5I and online supplemental figure 6I). Regarding T cells, neutrophils and monocytes showed no significant changes between the two groups (online supplemental figure 6J). To further validate these findings, we used another mouse tumor model B16, and found that Prmt5 cKO mice developed smaller tumors than the control group (online supplemental figure 6K and figure 5J). Consistent with previous findings, we observed a decrease of F4/80+macrophages in Prmt5 cKO mice (figure 5K and online supplemental figure 6L). Additionally, CD206 expression was decreased in Prmt5 cKO macrophages (figure 5L). And TNF-α expression had a trend of increase in Prmt5 cKO macrophages (figure 5M and online supplemental figure 6M), while tumor-bearing Prm5 cKO mice exhibited an increase in T cells (online supplemental figure 6N). These data indicated that despite the distinct immune microenvironments across different tumor types, prmt5 deficiency in macrophages consistently reprograms macrophages toward an antitumor phenotype and showed a consistent trend of tumor reduction.

Figure 5

Myeloid cell-specific deletion of Prmt5 protected tumor progression. (A–B) MC38 cancer cells were s.c. implanted into Prmt5 cKO (n=7) and littermate control mice (n=8), images of resected tumors (A) and tumor weights (B) in the two groups were shown. (C) The frequency of CD11b+F4/80+TAM cells from Prmt5 cKO (n=6) and control (n=6) was determined by flow cytometry. (D) The frequency of CD206+cells in F4/80+TAM from Prmt5 cKO (n=7) and control (n=7) was determined by flow cytometry. (E) The frequency of TNF-α+ cells in F4/80+TAM from Prmt5 cKO (n=7) and control (n=7) was determined by flow cytometry. (F–I) LLC cancer cells were s.c. implanted into Prmt5 cKO (n=6) and littermate control mice (n=6), image of resected tumors (F) and tumor weights (G) in the two groups were shown. (H) The frequency of CD11b+F4/80+TAM cells from Prmt5 cKO and control was determined by flow cytometry. (I) The frequency of TNF-α+cells in F4/80+TAM from Prmt5 cKO and control was determined by flow cytometry. (J–M) B16 cancer cells were s.c. implanted into Prmt5 cKO (n=11) and littermate control mice (n=9). (J) Tumor weights in the two groups were shown. (K) The frequency of CD11b+F4/80+TAM cells from Prmt5 cKO and control was determined by flow cytometry. (L) The frequency of CD206+cells in F4/80+TAM cells from Prmt5 cKO (n=11) and control (n=9) was determined by flow cytometry. (M) The frequency of TNF-α+ cells in F4/80+TAM cells from Prmt5 cKO and control was determined by flow cytometry. The mean±SD was graphed. Statistical differences were determined by Student’s t-test, *p<0.05, **p<0.01. cKO, conditional knockout; PRMT5, protein arginine methyltransferase 5; s.c, subcutaneous Injection; TAM, tumor-associated macrophage; TNF, tumor necrosis factor.

Prmt5 deficiency in myeloid cells reprograms macrophages and induces an antitumor phenotype in Prmt5 cKO mice

To access the impact of myeloid cell-specific deletion of Prmt5 on the immune landscape of the TME, we performed single-cell RNA-seq of CD45+immune cells infiltrating tumors in Prmt5 cKO and control mice. We revealed major immune cell subsets, including T cells, B cells, monocytes/macrophages, neutrophils, DCs, and NK cells (figure 6A and online supplemental figure 7A). In tumors from Prmt5 cKO mice, the proportion of infiltrating neutrophils and DCs increased, while the proportion of monocytes/macrophages decreased; however, the proportions of NK cells, B cells, and T cells remained unchanged, consistent with the flow cytometry results (figure 6B). To further analyze the monocyte-macrophage population, we performed unsupervised clustering on these cells and identified six clusters: Arg1-Macro, Pparg-Macro, Tnf-Macro, C1qc-Macro, Mki67-Macro and Ccr2-Mono (figure 6C). The corresponding marker genes were shown (figure 6D). GSEA revealed that the Tnf-Macro subset was enriched for M1-related genes, compared with other macrophage/monocyte subsets (figure 6E). Enrichment analysis revealed that the Pparg-Macro cluster was enriched in pathways related to negative regulation of immune system processes, suggesting the inflammatory modulation of this subset (figure 6F). Notably, Prmt5 cKO mice showed significant decreases in the Pparg-Macro and Mki-67-Macro subsets, alongside increases in the Tnf-Macro and C1qc-Macro subsets (figure 6G). M2-related genes, including Chil3, Ctsc, and Csf1r, were downregulated in TAMs from Prmt5 cKO mice compared with control mice, while the M1-related genes, such as Tnf and Cd86, were upregulated (figure 6H). Similarly, the expression of monocyte-macrophage migration receptor genes Ccr2 was downregulated (figure 6H). Conversely, chemokines Cxcl2, Ccl2, and Ccl5 were upregulated in Prmt5 cKO mice, with these chemokines being highly expressed in Tnf-Macrophages, potentially promoting neutrophil migration consistent with the increased proportion of neutrophils observed in Prmt5 cKO mice (figure 6H,I). Importantly, KEGG (Kyoto Encyclopedia of Genes and Genomes) enrichment analysis revealed that the TNF signaling pathway, Fc gamma R-mediated phagocytosis, and antigen processing and presentation were all enriched in neutrophils (figure 6J), suggesting an antitumor phenotype of the infiltrating neutrophils.34 We further re-clustered T cells into ten subtypes, including naïve and activated T cells, Treg cells, and Th17 cells (online supplemental figure 7B). The proportions of T-cell subclusters showed no significant difference between control and Prmt5 cKO mice (online supplemental figure 7C). Additionally, there were not significant differences in Gzmb expression or cytotoxic score in the CD8+T cells between the two groups (online supplemental figure 7D, E), suggesting that myeloid cell-specific deletion of Prmt5 may not directly impact T cells function in the TME. Lastly, we assessed NOS2 expression, a marker of the pro-inflammatory phenotype of macrophages.35 Flow cytometry analysis revealed a significantly increase in NOS2+macrophages in TAMs of Prmt5 cKO mice (figure 6K). Moreover, real-time RT-PCR showed a decrease in PPARγ expression in TAM isolated from Prmt5 cKO mice compared with control mice (figure 6L). Our results indicate that Prmt5 deficiency in myeloid cells reprograms macrophages and ind