TAMs-mediated resistance to oncolytic virus M1 in solid tumors

Abstract

Background

Oncolytic virus M1 (OVM), a naturally occurring alphavirus, has demonstrated potent antitumor activity in various solid tumor models by inducing immunogenic cell death and activating CD8⁺ T cells. However, its in vivo efficacy varies widely, and resistance mechanisms remain poorly understood. Tumor-associated macrophages (TAMs), key immunosuppressive cells within the tumor microenvironment, may limit OVM therapeutic potential.

Methods

We investigated the role of TAMs in OVM resistance using multiple syngeneic mouse tumor models (MC38 colorectal cancer, KPC1199 pancreatic cancer, RM1 prostate cancer, and B16F10 melanoma). TAMs were depleted using clodronate liposomes or CSF1R (Colony Stimulating Factor 1 Receptor) antibodies. Flow cytometry, mass cytometry, quantitative reverse transcription-PCR, and transcriptomic sequencing were employed to assess TAMs infiltration, viral load, and immune responses. CD8⁺ T cells were selectively depleted to determine their functional relevance.

Results

TAMs infiltration was positively correlated with resistance to OVM across tumor models. Depletion of TAMs increased intratumoral viral load and promoted accumulation of GZMB⁺ CD8⁺ T cells. RNA sequencing analysis revealed upregulation of antiviral and T-cell immune pathways in TAMs-depleted tumors. Importantly, the therapeutic benefit of TAMs depletion was abrogated on CD8⁺ T-cell depletion, confirming their essential role in mediating OVM efficacy. In both OVM non-responsive and responsive tumors, TAMs depletion enhanced OVM-mediated tumor suppression and survival.

Conclusions

TAMs, particularly M1-like subsets, play a critical role in mediating resistance to OVM therapy by reducing viral persistence and suppressing CD8⁺ T-cell responses. Targeting TAMs significantly improves the antitumor efficacy of OVM in solid tumors. These findings support the development of TAMs-targeted combination strategies to optimize oncolytic virotherapy.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Oncolytic virus M1 (OVM) has shown selective antitumor activity in solid tumors by inducing tumor cell death and activating CD8⁺ T cells, but its in vivo efficacy varies, and resistance mechanisms remain unclear. Tumor-associated macrophages (TAMs) are known to influence viral therapy outcomes through immunosuppressive and antiviral functions.

WHAT THIS STUDY ADDS

• This study identifies a strong correlation between TAMs infiltration and resistance to OVM therapy across multiple tumor models. Depletion of TAMs enhances intratumoral viral persistence, promotes accumulation of GZMB⁺ CD8⁺ T cells, and reverses resistance to OVM in a CD8⁺ T cell-dependent manner.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• These findings support targeting TAMs—particularly M1-like subsets—as a broadly applicable strategy to improve the efficacy of oncolytic virotherapy and guide the development of combinatorial immunotherapeutic approaches in solid tumors.

Introduction

Cancer remains a major global public health challenge, with its incidence and mortality rates continuing to rise.¹ Oncolytic virus (OV) therapy, an emerging immunotherapeutic approach, has gained considerable attention in recent years. OVs selectively infect and replicate within tumor cells, inducing direct oncolytic effects and activating both innate and adaptive immune responses. Furthermore, OVs can elicit a “bystander effect” through the release of cytokines, leading to the elimination of adjacent uninfected tumor cells.2,5 Several OVs, such as talimogene laherparepvec for melanoma and teserpaturev for glioma, have received clinical approval, with clinical trials demonstrating significantly improved survival rates in certain patient cohorts.6,8

Oncolytic virus M1 (OVM), an alphavirus, has attracted interest due to its pronounced antitumor activity in solid tumor models, including glioma, liver cancer, and colorectal cancer.⁹ OVM achieves targeted infection and replication in tumor cells characterized by high expression of matrix remodeling-associated protein 8 and low expression of zinc finger antiviral protein, which have been identified as potential predictive markers of efficacy.^{10 11} Despite its excellent selectivity in vitro, the in vivo efficacy of OVM varies significantly among diverse tumor models, implying that immunosuppressive factors in the tumor microenvironment (TME) may limit its antitumor activity. Our previous studies confirmed that tumor-associated myeloid-derived cells (TAMCs) impede the antitumor activity of OVM.¹² Tumor-associated macrophages (TAMs), a key constituent of the TME and the main cell type among TAMCs, demonstrate high plasticity: the M1 subtype exerts antitumor effects through phagocytosis and antigen presentation, whereas the M2 subtype promotes immunosuppression and tumor progression.13,16

Recent evidence indicates that TAMs exert a dual influence on oncolytic virotherapy. Activated TAMs can restrict OV replication and delivery by phagocytosing viral particles as shown in glioma models treated with oncolytic herpes simplex virus (HSV) and releasing antiviral cytokines such as type I interferons and tumor necrosis factor (TNF)-α, where TNF-α limited viral persistence.^{17 18} Conversely, oncolytic vesicular stomatitis viruses can reprogram TAMs toward an M1-like state, increasing antigen presentation and pro-inflammatory cytokine production to promote antitumor immunity.¹⁹ These findings underscore the high plasticity of TAMs in shaping therapeutic outcomes. A delineation of the roles of total TAMs and their M1-like and M2-like subpopulations in the context of OVM oncolytic virotherapy will enable the rational design of TAMs-directed strategies to further optimize the antitumor efficacy of OVM.

Our research demonstrates that TAMs are a pivotal factor in the resistance of solid tumors to OVM therapy. Depletion of TAMs across multiple solid tumor models resulted in a significant increase in intratumoral OVM load and enhanced CD8⁺ T cell activation, thereby improving therapeutic outcomes. Moreover, the reversal of OVM therapy resistance through TAMs depletion was dependent on CD8⁺T cells, underscoring the essential role of these immune effector cells in OVM-mediated antitumor immunity. These findings deepen our understanding of the mechanisms underlying solid tumor resistance to OVM therapy and highlight targeting TAMs as a potential strategy to enhance OV efficacy. The consistency of these results across various solid tumor models suggests that TAMs targeting serve as a broadly applicable approach to improve OVM therapy outcomes. This study establishes a foundation for developing combination therapies that integrate TAMs-targeting agents with OVM, offering a promising new direction for optimizing OV therapy and improving treatment outcomes for patients with solid tumors.

Materials and methods

Cell lines

Cell lines were purchased from Guangzhou Cellcook Biotech (B16F10, MC38, RM1), American Type Culture Collection (BHK-21) and KPC1199 was a gift from Professor Zhigang Zhang at the Shanghai Cancer Institute, Renji Hospital, Shanghai Jiao Tong University School of Medicine. MC38, B16F10, RM1 and KPC1199 cells were cultured in RPMI-1640 medium (Roswell Park Memorial Institute 1640, Gibco, C11875500BT) supplemented with 10% FBS (Fetal Bovine Serum, WISENT, 086-150) 1% penicillin/streptomycin (HyClone, SV30010). All cells were cultured at 37°C with 5% CO₂.

OVM production and quantification

OVM was grown in Vero cells and was provided by Guangzhou Virotech Pharmaceutical Technology. OVM-GFP (Oncolytic virus M1-Green fluorescent protein) is a recombinant derivative of OVM expressing jellyfish green fluorescent protein. Virus titer was determined by CCID₅₀ assay using BHK-21 cells.

Quantitative reverse transcription-PCR

Total RNA was isolated with TRIzol reagent (Thermo Fisher, Waltham, Massachusetts, USA), reverse transcription was performed with oligo dT and RevertAid Reverse Transcriptase (Thermo Fisher) according to the manufacturer’s instructions. Quantitative PCR was performed with SuperReal PreMix SYBR Green (Tiangen biochemical technology, Beijing, China) in the Applied Biosystems 7500 Fast Real-Time PCR System (Life Technologies). Specifically, for OVM virus copy detection in mouse tumor tissues, we use One Step Reverse Transcription Fluorescence Quantitative Kit (Probe Method) (Tiangen biochemical technology, Beijing, China).

Animal models

Animal studies were approved by the Animal Ethical and Welfare Committee of Zhongshan School of Medicine, Sun Yat-sen University (approval ID 2022001429). 6–8 weeks old mice were purchased from GemPharmatech. For evaluation of systemic antitumor effects, 1×10⁶ MC38 cells were inoculated subcutaneously into the hind flank of female C57BL/6J mice. 1×10⁶ KPC1199 cells were inoculated subcutaneously into the hind flank of female C57BL/6J mice. 1×10⁶ RM1 cells were inoculated into male C57BL/6J mice. 1×10⁶ B16F10 cells were inoculated into female C57BL/6J mice. After palpable tumors developed (around 50–100 mm³), the mice were randomized to receive different treatments. The treatments were given as single agents or in combinations with the following agents for each regimen (as shown in the schematics in the related figure legends). OVM 1×10⁷ CCID₅₀/mouse was injected intravenously five times a week when exploring the therapeutic effect of OVM. OVM 1×10⁷ CCID₅₀/mouse was injected intravenously three times a week, total 2 weeks. On the day before administration of OVM, 200 µL of control liposomes (LIPOSOMA, Netherlands, P-025) or clodronate liposomes (LIPOSOMA, Netherlands, C-025) were injected via the tail vein, administered once every 3 days for a total of four times. CSF1R antibody (300 µg/mice, Bio X Cell) and IgG2a isotype control (Bio X Cell, USA, BE0089) were injected intraperitoneally every other day; for the first day, mice were injected 1 mg. CD8α antibody (300 µg/mice, Bio X Cell, A2102) and IgG2b isotype control (300 µg/mice Bio X Cell, USA BE0090) were injected intraperitoneally every other day. Tumor length and width were measured, and the volume was calculated according to the formula length×width²/2. Tumor measurements were performed by researchers blinded to the group allocations. Mice were sacrificed when their tumor’s longest diameter was over 2 cm, and their survival was recorded.

Time curve of the OVM tumor load

Mice tumor tissues were taken at 2 hours, 50 hours, 96 hours, and 144 hours after OVM was given via the tail vein of MC38 tumor-bearing mice. The copy number of OVM RNA in the tumors was analyzed by quantitative reverse transcription-PCR (qRT-PCR). Total RNA was extracted using TRIzol (Life Technologies), OVM qRT-PCR assays were performed using a FastKing One Step Probe RT-qPCR Kit (Tiangen Biotech, China) on Applied Biosystems 7500 Fast Real-Time PCR System (Life Technologies) according to the manufacturer’s protocol. Amplification primers (Thermo Fisher) were:

NSP2 Forward Primers GGGATTCACTACACCTGCTTAGAC

NSP2 Reverse Primers GCTGACTCTGTCTGCGTAACC

NSP2 Probe CTCTCATCAGCAGCGAGCCTCCT

Transcriptome sequencing analysis

Tumor tissues were harvested from three mice per experimental group. Following dissection, the tumor tissues were rinsed with 0.9% saline, dried, and weighed. Each sample was then sectioned to achieve a uniform mass of 50 mg. The prepared samples were snap-frozen in cryovials using liquid nitrogen for 15 min and subsequently transferred to −80°C for shipment to BGI Genomics for transcriptome sequencing and subsequent data analysis.

Flow cytometry

For mouse experiments, dissect the tumor 2 days after OVM injection. Mouse tumor samples were minced with scissors, and single-cell homogenates were prepared using a mouse Tumor Dissociation Kit (Miltenyi Biotec, Germany, 130-096-730) and mashed through a 70 mm strainer (BIOFIL, China, CSS-013–070). Single cells from mouse spleens were isolated by grinding spleens through 40 mm filters (BIOFIL, China, CSS-010–040) before red blood cells were lysed.

For cell surface antigen staining, cells were washed with phosphate-buffered saline (PBS) containing 1% FBS and preincubated (15 min, 4°C with Purified Rat Anti-Mouse CD16/CD32 (BD Bioscience, USA, 553141), to block non-specific binding and then stained (30 min, 4°C) with 1:100 dilutions of various combinations of the following fluorochrome-conjugated antibodies: BV510 anti-mouse CD45 antibody (BD Bioscience, USA, 563891), Brilliant Violet 421 anti-mouse F4/80 (BioLegend, USA, 123132), FITC anti-mouse/human CD11b (BioLegend, USA, 101206), BV421 anti-mouse/human CD11b (BioLegend, USA, 101236), APC anti-mouse/human CD11b (BioLegend, USA, 101211), PerCP-Cy5.5 anti-mouse F4/80 antibody (Invitrogen, USA, 45-4801-80), PE-Cyanine7 F4/80 monoclonal antibody (Invitrogen, USA, 25-4801-82), FITC anti-mouse/human CD8 (BD Bioscience, USA, 101206), PerCP-Cy5.5 anti-mouse CD3e (Invitrogen, USA, 45-4801-80), APC-eFluor 780 anti-mouse MHC-II antibody (Invitrogen, USA, 47-5321-82), PE anti-mouse I-A/I- (BioLegend, USA, 107608), Fixable Viability Stain 620 (FVS620, BD Bioscience, USA, 564996), Fix Viability Stain 780 (FVS780, BD Bioscience, USA, 565388).

For intracellular staining, add 100 µL of Intracellular fixation buffer (Invitrogen, USA, 88-8824-00) to fix the cells for 30 min. Add 2 mL of permeabilization buffer, centrifuge at 500 g for 5 min at room temperature, and resuspend the cell pellet in 100 µL of permeabilization buffer (Invitrogen, USA, 88-8824-00). Add the recommended amount of antibody Brilliant Violet 605 anti-mouse CD206 antibody (BioLegend, USA, 141721), PE-Cyanine7 CD206 monoclonal antibody (Invitrogen, USA, 25-2061-82) to detect intracellular antigens and incubate at room temperature for 60 min. Flow cytometry analysis was performed by flow cytometry (CytoFLEX, Beckman Coulter) at room temperature.

Mass cytometry

For antibody panel set-up, all mass cytometry antibodies were provided by Polaris Biology, China. The details of the mass cytometry antibodies (CytoAtlas, Polaris Biology, China) are listed in table 1.

Table 1. CytoAtlas mass cytometry antibodies from Polaris Biology.

Category No	Antigen	Clone	Metal
M03111099	Anti-mouse F4/80	BM8	146Nd
M01011189	Anti-mouse CD3e	145–2 C11	209Bi
M03711133	Anti-mouse IFN-γ	XMG1.2	168Er
M03911132	Anti-mouse IL-6	MP5-20F3	167Er
M04411101	Anti-mouse MHC class II	M5/114.15.2	150Nd
M05911140	Anti-mouse CD80	16–10 A1	173Yb
M01111141	Anti-mouse CD4	RM4-5	174Yb
M02521089	Anti-mouse CD95/Fas	SA367H8	139La
M00811115	Anti-mouse CD274/PD-L1	10F.962	155Gd
M04711127	Anti-mouse CD27	LG.3A10	164Dy
M03811116	Anti-mouse IL-10	JES5-16E3	156Gd
M00411112	Anti-mouse CD161/NK1.1	QA19A51	153Eu
M05211134	Anti-mouse CD38	90	170Er
M00511107	Anti-mouse CD19	6D5	1495 m
M01611098	Anti-mouse CD8a	53–6.7	145Nd
M03311094	Anti-mouse Ly-6G	1A8	141Pr
M00311095	Anti-mouse CD11c	N418	142Nd
M01311105	Anti-mouse CD45	30-F11	1475 m
M03621120	Anti-mouse CD279/PD-1	RMP1-14	159Tb
M00211100	Anti-mouse CD11b	M1/70	148Nd
M00611096	Anti-mouse CD223/LAG-3	C9B7W	143Nd
M01411119	Anti-mouse CD62L/L-selectin	MEL-14	160Gd
M04911110	Anti-mouse CD127/IL-7Ra	A7R34	1545 m
M14911143	Anti-mouse CD172a/SIRPa	P84	175Lu
M02311097	Anti-mouse Ly-6C	HK1.4	144Nd
M03211125	Anti-mouse CD366/Tim-3	RMT3-23	162Dy
M00111096	Anti-mouse CD45R	RA3-6B2	143Nd
M01211138	Anti-mouse CD44	IM7	176Yb
M04511117	Anti-mouse CD49b	DX5	157Gd
M08511135	Anti-mouse CD152/CTLA-4	UC10-489	169Tm
M00721109	Anti-mouse CD25/IL-2Ra	3C7	152Sm
M05811124	Anti-mouse CD206	C068C2	161Dy
M07811142	Anti-mouse CD163	S15049l	176Yb
M01713118	Anti-mouse FOXP3	FIK-165	158Gd
M04611111	Anti-mouse TCRgd	UC7-13D5	151Eu
Customized	Anti-mouse TGF-b1	TW7-28611	172Yb
Customized	Anti-mouse granzyme B	REA226	165H0
Customized	Anti-mouse CD47	miap301	163Dy

For sample staining and acquisition, cells were washed with LunaStain cell staining buffer (Polaris Biology, China) and first stained with 10 µL of cisplatin reagent (Polaris Biology, China) at room temperature for 5 min. Cells were then washed with LunaStain cell staining buffer (Polaris Biology, China) and stained with 5 µL of Fc block (BioLegend, USA) for 10 min and the heavy metal-labeled membrane antibody mixtures for 30 min at room temperature. Cells were then washed two times with LunaStain cell staining buffer (Polaris Biology, China) two times and stained with Ir-DNA intercalator reagent (Polaris Biology, China) for 10 min. After staining, cells were washed and adjusted to 1 million cells per milliliter in LunaAcq cell acquisition solution (Polaris Biology, China) together with 20 µL of SureBits element calibration beads (Polaris Biology, China). Cell acquisition was performed at 800 events/second on a mass cytometer (StarionX1, Polaris Biology, China).

For data analysis, after acquisition, mass cytometry data were normalized and converted into standard FSC 3.0 files (StarionX1, Polaris Biology, China). Manual gating was performed using FlowJo (BD Biosciences, USA). Uniform Manifold Approximation and Projection (UMAP) was used to get an overview of the immune compartment. To identify different cell subtypes, FlowSOM clustering and meta-clustering were performed.

TCID₅₀

BHK-21 cells were seeded in the 96-well plate, 20 µL of sample diluent was added to each well, and a total of six dilutions of 10⁻¹-10⁻⁶ were set. Each dilution was set to eight duplicate wells. The cytopathic effect and GFP green fluorescence signal were observed daily. After 72 hours of infection, the number of holes with green fluorescence at each dilution was recorded according to the GFP signal. The virus titer was calculated by the Karber method: lgTCID₅₀=L D × (S-0.5) (L: the logarithm of the highest dilution, D: the difference between the logarithm of the dilution, S: the sum of the proportion of positive holes).

Statistical analysis

All statistical analyses were performed with GraphPad Prism V.9.0. All statistical methods and sample sizes are exhibited in the corresponding figure legend. There were no statistical methods used to predetermine the sample size and no generated data excluded. Data were analyzed by Student’s t-test, one-way analysis of variance (ANOVA), and two-way ANOVA were used to test the mean difference between two groups, as indicated in the derived graphs. Survival was analyzed by Kaplan-Meier survival plot and compared using a log-rank test. P value<0.05 was considered statistically significant. The bar charts show the mean±SD.

Results

TAMs hinder the efficacy of OVM in vivo

Our earlier investigations demonstrated that CD45⁺ CD11b⁺ TAMCs substantially attenuate the therapeutic effect of OVM.¹² It is well known that myeloid cells comprise multiple immunoregulatory subsets, including TAMs, myeloid-derived suppressor cells (MDSCs), tumor-associated neutrophils, dendritic cells (DCs) and monocytes (precursors to TAMs).²⁰ For OVM, it is necessary to identify which subset predominantly compromises its therapeutic efficacy. Among these subsets, CD45⁺ CD11b⁺ F4/80⁺ TAMs represent the most abundant myeloid population within the TME and exhibit biological activities such as antiviral responses and suppression of lymphocyte function.¹⁶ Based on these features, we hypothesize that macrophages are the primary subset responsible for limiting OVM efficacy.

To evaluate the antitumor effects of OVM, we employed four subcutaneous tumor models in mice: prostate cancer RM1, melanoma B16F10, colorectal cancer MC38, and pancreatic cancer KPC1199 (figure 1A,B). Treatment commenced on the 5th day post-tumor implantation, with tumor volume measured every 2 days and the relative tumor growth rate (T/C%) recorded. The results revealed that RM1 and B16F10 exhibited significant and partial tumor suppression, respectively, whereas MC38 and KPC1199 showed minimal responses. Tumors with a T/C%≤50%, specifically RM1 and B16F10, were classified as the “responsive group”, while those with a T/C%>50%, namely MC38 and KPC1199, were designated as the “non-responsive group” (figure 1B).

Figure 1. The infiltration of TAMs is positively correlated with tumor resistance to M1 virus therapy in vivo. (A) Mice administration pattern. In four subcutaneous tumor models (RM1, B16F10, MC38, KPC1199), M1 virus was administrated I.V. at 1×10⁷ CCID₅₀/dose, once daily for five consecutive days (Days 6–10). (B) Tumor growth curves and T/C% (n=6). T/C%, relative tumor proliferation rates, TRTV/CRTV×100% (V_t, tumor volume after treatment; V₀, tumor volume before treatment; RTV, relative tumor volume, V_t/V₀; TRTV, RTV in treatment group; CRTV, RTV in control group). P values were calculated using a two-way ANOVA with Tukey’s multiple comparisons test. (C–D) The ratio of CD45⁺ CD11b⁺ F4/80⁺ TAMs in CD45⁺ CD11b⁺ TAMCs (C) and single cells excised from tumors. Tumor tissues excised from B16F10, RM1, MC38 and KPC1199 tumor-bearing mice were harvested and homogenized 7 days after the first M1 virus dose (Day 12), and then the single-cell suspensions were subjected to flow cytometry analysis. ns, no statistical difference. (E–G) Correlation analysis between the tumor infiltration of TAMs (E) /M1-like TAMs (F) /M2-like TAMs (G) and T/C% after M1 virus treatment. ns, no statistical difference. **p<0.01. ***p<0.001. ANOVA, analysis of variance; I.V., intravenously; MHC, major histocompatibility complex; TAM, tumor-associated macrophages; TAMCs, tumor-associated myeloid-derived cells.

Tumors across these models displayed varying degrees of resistance to OVM. This resistance may stem from the presence of inhibitory factors within the TME that suppress the antitumor efficacy of OVM. In previous studies, we found that TAMCs hinder the efficacy of OVM therapy.¹² To further investigate the myeloid cell types exerting immunosuppressive effects in tumor models, we used flow cytometry to assess the proportion of TAMs within TAMCs in MC38 and KPC1199 tumors. Our results indicated that TAMs constituted more than 50% of the TAMCs, presumably serving as the primary immunosuppressive cell group (figure 1C and online supplemental figure S1A). Macrophages are critical regulators of innate immunity and play significant roles in both tumor immunosuppression and antiviral immunity. However, whether TAMs are linked to the differential therapeutic responses to OVM across these tumor models remains unclear.

To investigate a potential correlation between TAMs and OVM treatment resistance, we assessed TAMs infiltration levels across the different models. Although OVM treatment did not alter TAMs infiltration in any model, baseline TAMs infiltration was lower in the responsive group compared with the non-responsive group (figure 1D). The analysis revealed a significant positive correlation between the initial level of TAMs infiltration and the lowest tumor-to-control ratio (T/C%) observed after treatment with OVM. This was evidenced by a high coefficient of determination (R²=0.9618) and a statistically significant p value (p=0.0193), suggesting that TAMs may contribute to OVM resistance (figure 1E).

TAMs comprise two main subtypes: classically activated M1-like and alternatively activated M2-like macrophages. M1-like TAMs are pivotal in antiviral and antitumor immunity, whereas M2-like TAMs promote tumor immunosuppression by depleting CD8⁺T cells and recruiting regulatory T (Treg) cells. To further explore this, we used flow cytometry to examine the distribution of M1-like and M2-like TAMs within the TME and analyzed their association with OVM treatment resistance. Correlation analysis revealed a strong positive correlation between OVM treatment resistance and M1-like TAMs infiltration (R²=0.9345), while the correlation with M2-like TAMs did not reach statistical significance (R²=0.7782), though it still suggests a potential role in the resistance mechanism (figure 1F,G and online supplemental figure S1B,C).

Reducing TAMs increases viral load of OVM in tumor and effectively induces upregulation of antitumor immune signaling pathways

OVs, owing to their inherent characteristics, must penetrate the TME to replicate, thus directly exerting oncolytic effects and stimulating immune responses.⁹ Given the established role of macrophages in viral clearance, we hypothesized that TAMs may restrict the intratumoral viral load of OVM. To test this hypothesis, we examined the impact of TAMs depletion on OVM load within the TME.

In our experiments, immunocompetent mice were subcutaneously inoculated with MC38 colorectal cancer cells, and the treatment regimen was administered as depicted (figure 2A). 3 days after the tail vein injection of clodronate liposomes, the efficiency of TAMs depletion was evaluated using flow cytometry. The infiltration ratio of TAMs was significantly reduced, confirming successful depletion (figure 2B).

Figure 2. The impact of TAMs depletion on M1 virus viral load and on immune-activating transcriptional signals in tumor. Mice administration pattern. After MC38-derived tumor volumes reached 50–100 mm³, mice were randomly grouped. On Day 0 and Day 3, clodronate liposomes were administered via tail vein injection. On Days 1–3, M1 virus was administered via tail vein injection (I.V.). (B) Intratumoral infiltration of TAMs following clodronate liposome treatment. MC38-derived tumors were harvested 4 days after the first M1 virus dose (Day 10) and processed for flow cytometric analysis. Vehicle, control liposome. Clodronate, clodronate liposome. *p<0.05. (C) Viral copies of M1 virus in tumor. For virus copies testing, tumors were harvested at 2, 50, 96, and 144 hours after the first M1 dose, and were subject to qPCR. ns, no statistical difference. *p<0.05. **p<0.01. (D–F) Volcano plot of differentially expressed genes (D) and KEGG pathway enrichment signaling pathway map for differentially expressed genes (E) or for anti-pathogen immune pathways (F). MC38-derived tumors were harvested 4 days after the first M1 virus dose (Day 10) and processed for RNA sequencing. (G–K) GSEA analysis for T-cell receptor immune pathways (G), Th1 and Th2 cell immune pathways (H), PD-1 and PD-L1 immune checkpoint pathways (I), NK cell cytotoxic immune pathways (J) and apoptosis signaling pathways (K). ns, no statistical difference. *p<0.05. **p<0.01. GSEA, Gene Set Enrichment Analysis; I.V., intravenously; KEGG, Kyoto Encyclopedia of Genes and Genomes; NK, natural killer; PD-1, programmed cell death protein-1; PD-L1, programmed death-ligand 1; qPCR, quantitative PCR; TAM, tumor-associated macrophage.

Tumor samples were collected at 2, 50, 96, and 144 hours following OVM administration, and intratumoral viral RNA copy numbers were quantified via quantitative PCR. In the OVM group, the viral load peaked at 2 hour post-administration, declined to low levels by 50 hours, and became undetectable at 96 and 144 hours, indicating rapid viral clearance. In contrast, the OVM combined with clodronate liposome group exhibited detectable viral RNA at 2 hours, with the viral load peaking at 50 hours, significantly decreasing by 96 hours, and becoming undetectable by 144 hours. Compared with the OVM group, the OVM combined with clodronate liposome group demonstrated significantly higher intratumoral viral loads at 2 hours, 50 hours, and 96 hours. These findings indicate that TAMs infiltration in TME directly influences the dynamics of OVM viral load within the tumor, and depleting TAMs enhances the persistence and abundance of OVM in the tumor (figure 2C).

Building on the observation that TAMs depletion elevates OVM load in the TME, we further explored transcriptomic changes in tumor tissues using the MC38 colorectal cancer model. Transcriptomic sequencing identified 65 differentially expressed genes in the OVM combined with clodronate liposome group compared with the OVM group, with 18 genes upregulated and 47 downregulated (figure 2D). Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis revealed that the OVM combined with clodronate liposome group was enriched in pathways associated with antiviral immunity, T-cell and natural killer (NK)-cell signaling, and apoptosis, whereas the OVM group showed fewer enriched genes (figure 2E). Gene Set Enrichment Analysis based on KEGG pathways further demonstrated significant upregulation of antipathogen immune signaling in the OVM combined with the clodronate liposome group, likely linked to the increased OVM viral load following TAMs depletion (figure 2F). Additionally, pathways related to T-cell receptor signaling, Th1/Th2 differentiation, PD-1/PD-L1 signaling, and NK-cell cytotoxicity were significantly enhanced, suggesting activation of both innate and adaptive immunity. The apoptosis pathway was also notably enriched, indicating increased immune-mediated tumor cell death (figure 2G–L).

TAMs depletion increased the viral load of OVM within the TME, underscoring the antiviral role of TAMs. RNA sequencing analysis indicated that the depletion of TAMs, in conjunction with OVM treatment, resulted in a substantial upregulation of antipathogenic immune signaling. Specifically, this combination activated T-cell and NK-cell pathways and facilitated the induction of apoptosis. These findings imply that an increased viral load of OVM potentiates immune activation and enhances antiviral responses.

Thus, TAMs serve as critical mediators of OVM clearance. Reducing TAMs significantly increases intratumoral OVM load and effectively upregulates T-cell and NK-cell antitumor immune signaling pathways.

Depletion of TAMs favors the intratumoral accumulation of GZMB⁺ CD8⁺ T cells following OVM treatment

OVM selectively infects and replicates within tumor cells, inducing immunogenic cell death that lyses tumor cells and releases tumor antigens into the TME, thereby recruiting immune cells to exert antitumor effects.⁹ Among these, CD8⁺T cells, as the core effector cells in antitumor immunity, kill tumor cells by secreting cytotoxic molecules such as interferon-γ and granzyme B, making them a critical population for OVM therapy to activate immune responses.²¹ Given the virus-depleting properties of TAMs, we hypothesized that depleting TAMs could enhance OVM efficacy by increasing viral load and reshaping the immune cell landscape within the TME. To test this hypothesis, we employed mass cytometry (CyTOF, Cytometry by Time-Of-Flight) to analyze changes in the immune cell profile within the TME and investigated the effects of clodronate liposomes combined with OVM treatment on immune cell distribution in immunocompetent mouse models (figure 3A).

Figure 3. Changes in immune cell profile in the TME after depletion of TAMs combined with OVM treatment. Mice administration pattern. MC38-derived tumors were harvested 4 days after the first M1 virus dose (Day 10) and processed for CyTOF analysis of the immune cell profile. (B) t-SNE plot of CD45⁺ immune cells. (C) Heatmap of normalized marker expression for each cell clusters. (D) Proportion of each immune cell cluster within the CD45⁺ population. (E–M) Proportion of CD4⁺ T cells (E), CD8⁺ T cells (F), effector CD8⁺ T cells (G), CD8⁺ T_em (H), CD8⁺ T_cm (I), naïve CD8⁺ T cells (J), GZMB⁺ CD8⁺ T cells (K), M1-like TAMs (L) and M2-like TAMs (M) among CD45⁺ immune cells. (N–O) Expression of PD-L1 (N) and TGF-β (O) on M2-like TAMs. ns, no statistical difference. *p<0.05. **p<0.01. ***p<0.001. ****p<0.0001. TGF, Transforming Growth Factor; CyTOF, Cytometry by Time-Of-Flight; NKT, Natural Killer T cells; DC, dendritic cell; GZMB, granzyme B; I.V., intravenously; MDSC, myeloid-derived suppressor cell; OVM, oncolytic virus M1; PD-L1, programmed death-ligand 1; TAM, tumor-associated macrophage; T_cm, T central memory cell; T_em, T effector memory cell; TME, tumor microenvironment; t-SNE, t-distributed stochastic neighbor embedding.

Using UMAP to analyze CyTOF data (figure 3B,C and online supplemental figure S2), we identified eight major immune cell types including T cells, NK cells, NKT cells, B cells, DCs, macrophages, MDSCs, and monocytes. In comparison to the solvent group, both the clodronate liposome group and the combination treatment significantly attenuated TAMs infiltration, thereby confirming the effective depletion of TAMs (figure 3D). The combination treatment group showed significantly increased infiltration proportions of NK cells, NKT cells, and T cells compared with the control group (figure 3D). Relative to the OVM group, the combination group displayed a significant increase in NKT cell infiltration and a trend toward increased NK cell and T-cell infiltration. Within CD45⁺ cells, the proportion of CD8⁺ T cells was significantly higher in the combination group compared with both the control and OVM groups (p<0.05), while the proportion of CD4⁺ T cells remained unchanged.

We further examined changes in CD8⁺ T-cell subsets (online supplemental figure S2). As shown in figure 3E–G, compared with the control group, the proportion of CD8⁺ T and effector CD8⁺ T cells in the OVM group showed no significant change, whereas both the clodronate liposome monotherapy and combination groups exhibited a significant increase (p<0.05). Although the proportions of effector memory CD8⁺ T cell and central memory CD8⁺ T cells trended higher in the combination group compared with the OVM group, these differences did not reach statistical significance (figure 3H,I). The proportion of naïve CD8⁺ T cells was significantly elevated in the combination group relative to both the solvent and OVM groups (figure 3J). Notably, the proportion of granzyme B-secreting CD8⁺ T cells (GZMB⁺ CD8⁺ T) was significantly increased in the combination group compared with the other three groups (p<0.05) (figure 3K). These findings indicate that clodronate liposomes combined with OVM treatment reshaped the TME immune profile, significantly promoting the intratumoral accumulation of GZMB⁺ CD8⁺ T cells and thereby driving the transition of CD8⁺ T cells toward an effector phenotype.

Given the correlation between M1-like TAMs and resistance to OVM treatment, we further analyzed changes in TAMs subsets. The results revealed that the infiltration of M1-like TAMs was significantly reduced in the combination treatment group compared with the control group (p<0.05), while M2-like TAMs infiltration showed no significant differences across groups, suggesting a potential key role for M1-like TAMs in treatment resistance (figure 3L,M). To exclude the influence of functional changes in M2-like TAMs, we assessed the expression levels of their immunosuppressive molecules PD-L1 and TGF-β, finding no significant differences between groups, indicating that the suppressive function of M2-like TAMs remained unaffected (figure 3N,O).

In summary, clodronate liposomes, by reducing the depletion of OVM by M1-like TAMs, increased the viral load within the TME, thereby promoting the infiltration and activation of NK cells, NKT cells, and CD8⁺ T cells. Consequently, TAMs, particularly M1-like TAMs, impede the antitumor immune response by restricting OVM load.

Depletion of TAMs effectively reverses M1 virus treatment resistance in a CD8⁺ T-dependent manner

Here we established the MC38 colorectal cancer model in immunocompetent mice and employed two independent strategies to deplete TAMs: clodronate liposomes and CSF1R antibodies (CSF1R-Ab) (figure 4A,B). Experimental results revealed that clodronate liposomes significantly reduced the infiltration of TAMs within the tumor and macrophages in peripheral blood; similarly (figure 4C), CSF1R-Ab decreased the infiltration of TAMs and peripheral monocytes (figure 4D). Although neither OVM alone nor any single TAMs depletion method significantly inhibited tumor growth, the combination of OVM with either clodronate liposomes or CSF1R antibodies markedly slowed tumor progression and extended mouse survival (figure 4E–J). To elucidate the role of CD8⁺ T cells, we depleted CD8⁺ T cells within the TME using Cluster of Differentiation 8 alpha chain-specific antibodies (CD8α-Ab) (figure 4K,L). The depletion of CD8⁺ T cells completely abrogated the antitumor effects of TAMs depletion combined with OVM treatment, resulting in accelerated tumor growth and increased tumor weight, comparable to levels observed in the control group. In contrast, dual treatment with OVM and clodronate liposomes significantly suppressed tumor progression, whereas the addition of CD8α-Ab in a triple treatment abolished this effect, underscoring the indispensable role of CD8⁺ T cells in this process (figure 4M–O). In conclusion, this study demonstrates that, in the mouse colorectal cancer model, neither OVM alone nor TAMs depletion alone is sufficient to effectively suppress tumor growth; however, their combination significantly reverses OVM treatment resistance, an effect dependent on CD8⁺ T-cell mediation. These findings elucidate the mechanism by which TAMs depletion enhances OVM efficacy and provide a theoretical foundation and strategic support for the combined targeting of TAMs and OVs in the treatment of colorectal cancer.

Figure 4. The effect of TAMs depletion on reversing M1 virus treatment resistance and its dependence on CD8⁺ T cells. (A–B) Mice administration pattern for clodronate liposome treatment (A) and CSF1R-Ab treatment (B). After MC38-derived tumor volumes reached 50–100 mm³, mice were randomly grouped (n=6). On Day 6, 7, 9, 12 and 15, clodronate liposomes or CSF1R-Ab were respectively administered via tail vein injection (I.V.) or I.P. injection. (C–D) The ratio of CD45⁺ CD11b⁺ F4/80⁺ TAMs in CD45⁺ cells following clodronate liposome treatment. Vehicle, control liposome (C) and CSF1R-Ab treatment (D). MC38-derived tumors and peripheral blood were harvested 4 days after the first M1 virus dose (Day 10) and processed for flow cytometric analysis. (E–G) Tumor growth curve (E), T/C (%) (F) and Kaplan-Meier survival curves (G) following clodronate liposome treatment. (H–J) Tumor growth curve (H), T/C (%) (I) and Kaplan-Meier survival curves (J) following CSF1R-Ab treatment. (K) Mice administration pattern for CD8α-Ab treatment (n=5). On Day 7, 10, 12 and 14, CD8α-Ab were administered via (I.P.). (L) The ratio of CD3⁺ CD8⁺ T cells in CD45⁺ cells following CD8α-Ab treatment. MC38-derived tumors, blood and spleen were harvested 5 days after the first M1 virus dose (Day 12) and processed for flow cytometric analysis. (M–O) Tumor growth curve (M), T/C (%) (N) and tumor weights (O) following clodronate liposome treatment and/or CD8α-Ab treatment. ns, no statistical difference. *p<0.05. **p<0.01. ***p<0.001. ****p<0.0001. CD8α-Ab, CD8α-specific antibodies; CSF1R-Ab, CSF1R antibodies; I.V., intravenously; I.P., intraperitoneal: TAM, tumor-associated macrophage.

Depleting TAMs effectively enhances the antitumor effect of M1 in three other tumor types

Our data show that, in the murine MC38 colorectal cancer model, depletion of TAMs overcomes resistance to OVM (figure 4). Whether TAM depletion similarly abrogates OVM resistance across other solid tumor models remains to be determined. Thus, we employed two TAM depletion strategies, clodronate liposome and CSF1R-Ab, and validated their effects in mouse models of OVM-non-responsive pancreatic cancer KPC1199 and OVM-responsive prostate cancer RM1 and melanoma B16F10.

The experimental design was as follows: KPC1199, RM1, and B16F10 cells were inoculated into immunocompetent mice, which were then randomly assigned to receive either control solvent (PBS or control liposome), OVM, clodronate liposome, or CSF1R-Ab as monotherapies, or combination treatments of OVM with clodronate liposome or CSF1R-Ab (figure 5A,L). Results revealed that clodronate liposome significantly reduced TAMs across all three models, while CSF1R-Ab also effectively reduced TAMs (figure 5B,M). In the OVM-non-responsive KPC1199 model, neither OVM nor TAMs depletion alone suppressed tumor growth; however, combining OVM with clodronate liposome or CSF1R-Ab significantly delayed tumor progression and prolonged survival (figure 5C–E and N–P). In the OVM-responsive RM1 and B16F10 models, OVM monotherapy already exhibited antitumor effects, which were further enhanced when combined with clodronate liposome or CSF1R-Ab (figure 5F–K and Q–V), resulting in reduced tumor weight and extended survival. The consistency of outcomes across both TAMs depletion strategies suggests that TAMs suppress OVM efficacy in a variety of solid tumors.

Figure 5. Depletion of TAMs enhances the therapeutic effect of M1 oncolytic virus in multiple tumor models. Mice administration pattern for clodronate liposome treatment. After tumor volumes reached 50–100 mm³, mice were randomly grouped (n=6 in KPC1199, n=7 in RM1, B16F10). On Day 6, 9, 13 and 15, clodronate liposome was administered via I.V. injection. Vehicle, control liposome. (B) The ratio of CD45⁺ CD11b⁺ F4/80⁺ TAMs in CD45⁺ cells following clodronate liposome treatment. Tumors were harvested 4 days after the first M1 virus dose (Day 10) and processed for flow cytometric analysis. (C–K) Tumor growth curve (C, F, I), T/C (%) (D, G, J) and Kaplan-Meier survival curves (E) or tumor weight (H, K) following clodronate liposome treatment. (L) Mice administration pattern for CSF1R-Ab treatment. After tumor volumes reached 50–100 mm³, mice were randomly grouped (n=6). On Day 6, 9, 13 and 15, CSF1R-Ab were administered via I.P. injection. (M) The ratio of CD45⁺ CD11b⁺ F4/80⁺ TAMs in CD45⁺ cells following CSF1R-Ab treatment. Tumors were harvested 4 days after the first M1 virus dose (Day 10) and processed for flow cytometric analysis.(N–V) Tumor growth curve (N, Q, T), T/C (%) (O, R, U) and Kaplan-Meier survival curves (P, S, V) following CSF1R-Ab treatment. n.s., not significant; *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. CD8α-Ab, CD8α-specific antibodies; CSF1R-Ab, CSF1R antibodies; I.V., intravenously; I.P., intraperitoneal: TAM, tumor-associated macrophage.

In conclusion, depleting TAMs reverses treatment resistance in OVM-non-responsive tumors, while in OVM-responsive tumors, it amplifies the antitumor effect. These findings indicate that TAMs are critical inhibitory immune cells limiting OVM efficacy and may serve as a synergistic target to enhance the clinical potential of OVM in solid tumor therapy.

Discussion

This study systematically elucidates the pivotal role of TAMs in resistance to OVM therapy and its underlying mechanisms. Our findings reveal a significant positive correlation between the extent of TAMs infiltration in various solid tumors and resistance to OVM treatment, suggesting that TAMs may serve as a critical determinant limiting the antitumor efficacy of OVM. We demonstrate that TAMs substantially reduce the viral load of OVM within the TME, then suppressing the proliferation and activation of effector CD8⁺ T cells. Additional experiments indicate that targeted depletion of tumor-localized TAMs, particularly M1-like TAMs—using clodronate liposomes or CSF1R-Ab effectively reverses treatment resistance in solid tumors, such as colorectal and pancreatic cancer. Notably, this reversal effect is entirely contingent on the participation of CD8⁺ T cells. These insights provide a robust theoretical foundation for optimizing OV therapy and delineate a promising direction for the combined application of TAMs-targeted strategies with OVM.

The dual role of macrophages in OV therapy is well-documented. On one hand, TAMs can bolster antitumor immunity by phagocytosing tumor cells and presenting antigens; on the other, their antiviral activities such as type I interferon secretion and direct phagocytosis of viral particles curtail OV efficacy.²² For example, in glioma models, macrophage depletion significantly elevates the intratumoral titer of oncolytic HSV.¹⁸ This study corroborates these observations and, through subgroup analysis, unveils the functional heterogeneity of TAMs. Our data reveal that M1-like TAMs exhibit a significantly stronger association with OVM treatment resistance than M2-like TAMs. In OVM-non-responsive models, such as the MC38 colorectal cancer mouse model, clodronate liposome treatment predominantly reduced M1-like TAMs, while the proportion of M2-like TAMs and expression levels of immunosuppressive markers (eg, PD-L1 and TGF-β) remained largely unchanged. This finding suggests that M1-like TAMs primarily drive treatment resistance by clearing OVM rather than directly suppressing T cells.

This study substantiates the role of targeted TAMs depletion in overcoming OVM treatment resistance. In multiple solid tumor models, standalone TAMs depletion did not yield significant antitumor effects, indicating that TAMs are not direct exerting factors that promote tumor progression. However, when combined with OVM, TAMs depletion markedly enhanced therapeutic efficacy, achieving reversal of resistance in OVM-non-responsive tumors. Mechanistically, TAMs depletion increased intratumoral OVM load and critically facilitated the accumulation of GZMB⁺ CD8⁺ T cells following OVM administration. This effect was completely abolished in vivo CD8⁺ T cells depletion experiments, affirming that CD8⁺ T cells are the central effector cells mediating resistance reversal through TAMs depletion.

These results align with the immune activation paradigm of OV therapy, wherein OVs trigger immunogenic cell death and antigen release to stimulate CD8⁺ T cell-mediated antitumor immunity. However, TAMs indirectly impede this process by limiting viral replication. Our study demonstrates that intratumoral infiltration of M1-like TAMs is inversely associated with therapeutic response to OVM. Compared with M2-like TAMs, M1-like TAMs typically display stronger antiviral programs.¹⁷ In the context of oncolytic virotherapy, a plausible mechanism is that M1-like TAMs curtail intratumoral viral replication via phagocytosis of virions and infected cells and the production of type I interferons and pro-inflammatory mediators, thereby indirectly attenuating OVM-induced CD8⁺ T cell priming and effector function. This working model requires further validation in dedicated mechanistic studies.

The heterogeneity of TAMs underpins their complex therapeutic roles. M1-like TAMs, with their pronounced phagocytic capacity, are more inclined to clear viruses, whereas M2-like TAMs suppress CD8⁺ T cell function through immunosuppressive pathways such as PD-1/PD-L1 or CD47/SIRPα.²³ Through correlation analysis and subgroup characterization, this study confirms the predominant role of M1-like TAMs in OVM treatment resistance, with no significant direct immunosuppressive effects observed from M2-like TAMs in this model. This insight suggests that precision targeting of M1-like TAMs could be a key strategy for enhancing OVM efficacy.

Nonetheless, TAMs-targeted strategies as monotherapies have demonstrated limited clinical success. For instance, CSF1R inhibitors (eg, PLX3397 and AMG820) exhibited modest antitumor activity in phase II clinical trials, likely due to non-specific depletion of TAMs subgroups and tumor heterogeneity.^{24 25} Similarly, in this study, TAMs depletion alone produced no notable efficacy; however, when paired with OVM, it elicited a synergistic effect. This indicates that the therapeutic potential of TAMs-targeted approaches may be maximized in combination with OVs, where reducing the antiviral activity of M1-like TAMs elevate viral load, thereby amplifying CD8⁺ T cell-mediated antitumor immunity.

This study highlights the central role of CD8⁺ T cells in OVM antitumor effects and elucidates the mechanism by which TAMs indirectly suppress immune activation through restricted viral replication. These findings offer theoretical support for the development of combination therapies integrating TAMs-targeted agents with OVM. Future investigations could prioritize: (1) devising precision strategies to target M1-like TAMs while minimizing interference with other immune functions; (2) exploring the feasibility of concurrently mitigating M2-like TAMs immunosuppressive effects to further enhance OVM efficacy; (3) refining combination therapy regimens through the integration of preclinical and clinical data.

In summary, this study clarifies that TAMs impair OVM’s antitumor efficacy in solid tumors by reducing viral load, thus inhibiting OVM to exert antitumor effects in a CD8⁺ T cell-dependent manner. Targeted TAMs depletion effectively reverses treatment resistance, an effect reliant on CD8⁺ T cells. These discoveries not only advance our understanding of TAMs mechanisms in OV therapy but also pave the way for novel research and clinical translation of combined OVM and TAMs-targeted therapeutic approaches.

Data availability statement

Data are available upon reasonable request.