Macrophage Mtdh deficiency discordantly regulated tumor growth and metastasis through increased thrombospondin-1 production

Abstract

Objective

Metadherin (MTDH) is a recognized oncogene involved in the progression and metastasis of various cancers. However, further studies are needed to elucidate the biological role of MTDH, which is expressed in macrophages during tumor progression.

Methods

Mouse colorectal cancer and melanoma cells were subcutaneously and intravenously injected into myeloid-specific Mtdh knockout mice to evaluate tumor growth and lung metastases. The effects of macrophage with Mtdh knockout on angiogenesis and fibrosis were examined using mass spectrometry, immunofluorescence staining, analyses of data from The Cancer Genome Atlas (TCGA) colon adenocarcinoma and melanoma cohorts, and western blotting. A thrombospondin (TSP-1) blocking peptide was used to inhibit transforming growth factor β1 (TGF-β1) activation for suppression of fibrosis in vivo and in vitro. The molecular mechanisms were investigated using RNA sequencing data from the Gene Expression Omnibus database, ELISA, immunoprecipitation, chromatin immunoprecipitation assay, quantitative real-time PCR, and western blotting.

Results

Mtdh-deficient macrophages suppressed lung metastasis but, unexpectedly, promoted subcutaneous tumor growth of both cancer cell types. This discordant effect was attributed to increased production of TSP-1, an angiogenesis inhibitor that also regulates fibrosis through TGF-β1 activation. Mtdh knockout in macrophages inhibited angiogenesis in both lung metastatic and subcutaneous tumors, whereas enhanced fibrosis was observed only in subcutaneous tumors. In TCGA colon adenocarcinoma data, higher TSP-1 expression correlated with advanced pathological T stage and cancer-associated fibroblasts abundance. Furthermore, Mtdh loss in macrophages induced activation of latent TGF-β1 in tumor cells, promoting fibroblast-to-myofibroblast transition, fibrosis, and unexpected tumor growth through the TSP-1/TGF-β1 axis. Mechanistically, MTDH deficiency led to nuclear retention of murine double minute-2 (MDM2), disruption of the MDM2-p53 interaction, and enhanced p53-dependent TSP-1 transcription.

Conclusions

We found that macrophage Mtdh deficiency discordantly regulates tumor metastasis and growth through either TSP-1–mediated anti-angiogenic effect or a TSP-1/TGF-β1-mediated pro-fibrotic effect. This study, therefore, provides novel insights into the mechanisms underlying the discordance between tumor growth and metastasis.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Metadherin (MTDH) is a recognized oncogene that promotes the progression and metastasis of various cancers. However, additional studies are needed to elucidate the biological role of macrophage-derived MTDH in tumor progression.

WHAT THIS STUDY ADDS

• Macrophage Mtdh deficiency discordantly regulates tumor metastasis and growth through thrombospondin-1 (TSP-1), which inhibits angiogenesis in both lung metastases and subcutaneous tumors. However, the TSP-1/transforming growth factor β1-mediated pro-fibrotic effect contributes to the unexpected tumor growth induced by macrophage Mtdh knockout.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• This study provides novel mechanistic insights into the opposing roles of macrophage Mtdh deficiency in tumor growth and metastasis, which may inform future therapeutic strategies targeting tumor–stroma interactions.

Introduction

The tumor microenvironment (TME) plays a crucial role in regulating tumor growth and metastasis. It consists of various non-malignant cell types, primarily fibroblasts, immune cells (such as macrophages and lymphocytes), and endothelial cells. Numerous studies have demonstrated that stromal cells interact with cancer cells to influence malignant behavior and disease progression. Simultaneously, cancer cells educate stromal cells to promote a pro-angiogenic, fibrotic, chronically inflammatory, and immunosuppressive microenvironment.^{1 2} However, further investigations are needed to elucidate the complex crosstalk between stromal and cancer cells.

Macrophages are the most abundant inflammatory cells within the TME^{3 4} and promote tumor growth, metastasis, and therapy resistance, thereby exerting a pro-tumoral effect.^{2 5 6} In addition to direct interactions with malignant cells, macrophages play a pivotal role in shaping the TME by communicating with other stromal cells, such as stimulating angiogenesis through pro-angiogenic factors,⁷ or fostering an immunosuppressive TME.^{8 9} They also orchestrate oncogenic remodeling of the TME by participating in extracellular matrix remodeling and activating myofibroblasts in both primary tumors and distant metastatic sites.¹⁰ As a critical component of the TME, macrophages regulate the biological properties of tumor and stromal cells. Therefore, macrophages represent promising therapeutic targets in cancer treatment.

Metadherin (MTDH/AEG-1/LYRIC), a metastasis-associated gene, is overexpressed in most cancers studied to date^{11 12} and is a well-established oncogene in malignant cells.^{13 14} MTDH is involved in multiple oncogenic processes, including invasion, metastasis, angiogenesis, tumor cell survival, and therapy resistance, through various signaling pathways.¹⁵ The tumor suppressor gene p53 exerts its functions by transcriptionally regulating numerous target genes, such as p21, BAX, and thrombospondin-1 (TSP-1). Overexpression of MTDH has been shown to inhibit p53 activity through several mechanisms, including interaction with the murine double minute-2 (MDM2) protein, which promotes p53 degradation.¹⁶ Interestingly, MTDH is frequently expressed in macrophages within tumor tissues, as shown in our previous study and others.^{17 18} We previously demonstrated that macrophages with MTDH overexpression promoted invasion and metastasis of head and neck squamous cell carcinoma cells (HNSCC) and were correlated with lymph node metastasis in patients with HNSCC. In contrast, macrophages with MTDH overexpression had no effect on the survival of HNSCC tumor cells.¹⁷ Given the critical and complex role of macrophage-derived MTDH in tumor progression, we generated a macrophage-specific Mtdh knockout (Mac-Mtdh^KO) mouse model to investigate the underlying mechanisms contributing to this discordant behavior.

In this study, we found that macrophage Mtdh knockout inhibited tumor metastasis but enhanced subcutaneous tumor growth of colorectal cancer and melanoma cells, both of which were accompanied by increased TSP-1-mediated inhibition of angiogenesis. We further demonstrated that the TSP-1/transforming growth factor-β1 (TGF-β1) axis was responsible for fibroblast-to-myofibroblast transition (FMT), fibrosis, and subsequent tumor growth induced by Mtdh-deficient macrophages. Moreover, loss of Mtdh in macrophages promoted p53-dependent TSP-1 transcription through nuclear retention of MDM2 and disruption of the MDM2–p53 interaction.

Materials and methods

Subcutaneous tumor model and metastasis model of MC-38 or B16F10 cells

For the subcutaneous tumor model, 3×10⁵ MC-38 or B16F10 cells were implanted subcutaneously into Mac-Mtdh^KO and Mtdh^fl/fl mice. Tumor growth was monitored by measuring the volume using the following formula: Tumor volume=0.5×(larger diameter)×(smaller diameter)². At the end of each experiment, all mice were euthanized and tumors were harvested for further analysis.

For the lung metastasis model, 3×10⁵ MC-38 or B16F10 cells were injected into the tail vein of Mac-Mtdh^KO and Mtdh^fl/fl mice. Approximately 2 weeks later, lung tissues were collected and processed for subsequent analyses.

Peritoneal macrophages and coculture systems

Peritoneal macrophages (PMs) were isolated from thioglycollate (211716, BD Biosciences)-elicited mice and purified by adherence to plastic culture dishes, as described previously.¹⁹ The adherent macrophages were cultured with or without LSKL peptide (an inhibitor of TSP-1-mediated activation of TGF-β1, HY-P0299A, MedChemExpress) or pifithrin-β (PFT-β, HY-16702A, MedChemExpress) for 48 hours. The culture supernatants were collected from Mac-Mtdh^KO and Mtdh^fl/fl PMs and used for coculture with other cell types for 48 hours.

Histopathological analysis

H&E staining, immunohistochemistry (IHC) assay, and immunofluorescence (IF) staining were performed. Subcutaneous tumors and lungs were dissected from Mac-Mtdh^KO and Mtdh^fl/fl mice. Tissues and cells were prepared for H&E staining and IHC as described previously.²⁰ IHC was carried out using the Universal Two-Step Test Kit (Mouse/Rabbit Enhanced Polymer Test System, Cat. No. PV-9000, Zsbio) following the manufacturer’s instructions. Sections were counterstained with hematoxylin (Cat. No. H8070, Solarbio).

For IF staining, frozen tumor tissue sections from Mac-Mtdh^KO and Mtdh^fl/fl mice were incubated with primary antibodies at 4°C overnight. After washing, the sections were incubated with appropriate secondary antibodies and counterstained with DAPI solution (Cat. No. C0065, Solarbio).

For Masson’s trichrome staining, collagen deposition was evaluated as previously described using the Masson Trichrome Stain Kit (Cat. No. G1340, Solarbio) according to the manufacturer’s protocol. Images were visualized and captured under an upright fluorescence microscope (Olympus).

LC–MS/MS (Liquid Chromatography-Tandem Mass Spectrometry) analysis

MC-38 subcutaneous tumors from Mac-Mtdh^KO and Mtdh^fl/fl mice were lysed in RIPA buffer containing protease inhibitors (Cat. No. P6730, Solarbio) to obtain protein extracts. The proteins were digested into peptides, which were subsequently desalted using spin desalting columns (Cat. No. 89852, Thermo Fisher Scientific) and redissolved in 0.1% formic acid. The peptide mixtures were analyzed using an Orbitrap mass spectrometer (Q Exactive Plus, Thermo Fisher Scientific) coupled to a nano-HPLC system. The resulting mass spectra were searched against the mouse UniProt database using Proteome Discoverer software (V.2.4). The Liquid Chromatography-Tandem Mass Spectrometry (LC–MS/MS) proteomics data have been deposited in the ProteomeXchange Consortium (https://proteomecentral.proteomexchange.org) via the iProX partner repository^{21 22} under the dataset identifier PXD065975.

Immune infiltration estimations of stromal and immune cells based on the LC-MS/MS data were obtained using the “ESTIMATE” module of the TIMER database (https://cistrome.shinyapps.io/timer/).²³ The ESTIMATE analysis was performed using the immunedeconv package with mMCP-counter, seqImmuCC, DCQ, and BASE algorithms.

RNA extraction and quantitative real-time PCR

Total RNA was extracted using TRIzon Reagent (CW0580S, CWBIO) and reverse transcribed into cDNA with the HiFiScript gDNA Removal cDNA Synthesis Kit (W2582M, CWBIO), as described previously.¹⁷ Quantitative real-time PCR (qRT-PCR) was performed using UltraSYBR Mixture (Cat. No. CW0957M, CWBIO) according to the manufacturer’s instructions. Relative mRNA expression levels were calculated using the 2^-(ΔΔCt) method, normalized to β-actin expression. The primer sequences were as follows:

Mouse Thbs1: Forward 5′-CTAGGTGTCCTGTTCCTGTTG-3′, Reverse 5′-AAGGAAGCCAGGAAGATGAAG-3′. Mouse Cdkn1a (p21): Forward 5′-GTCCAATCCTGGTGATGTCC-3′, Reverse 5′-GTTTTCGGCCCTGAGATGT-3′. Mouse BAX: Forward 5′-ATGAAGACAGGGGCCTTTTTGCTA-3′, Reverse: 5′-TCAGCCCATCTTCTTCCAGATGGT-3′. Mouse Collagen type I, Alpha1 (Col1a1): Forward 5′-TTCTCCTGGCAAAGACGGACTCAA-3′, Reverse 5′-AGGAAGCTGAAGTCATAACCGCCA-3′. Mouse β-actin: Forward: 5′-GGCTGTATTCCCCTCCATCG-3′, Reverse 5′-CCAGTTGGTAACAATGCCATGT-3′.

Western blot

Tissues and cells were lysed in ice-cold RIPA lysis buffer (P0013B, Beyotime) to extract total proteins. Cytoplasmic and nuclear fractions were isolated using a commercial Nuclear and Cytoplasmic Protein Extraction Kit (P0027, Beyotime). The extracted proteins were subjected to western blotting as described previously.²⁰

ELISA

Culture supernatants were analyzed by ELISA to quantify TSP-1 levels using the Mouse TSP-1 ELISA Kit (Cat. No. E-EL-M3083, Elabscience). Following treatment or non-treatment with acid activation, TGF-β1 concentrations were measured using the Human/Mouse TGF-β1 ELISA Kit (Cat. No. EK981, MultiSciences). Briefly, cell culture supernatants were added to the wells, followed by sequential incubation with detection antibodies, streptavidin–HRP conjugate, and TMB substrate. Absorbance was recorded at 450 nm using a multifunctional microplate reader (Molecular Devices). ELISA data were quantified using the ELISACalc software.

RNA sequencing analysis

For this study, RNA sequencing (RNA-seq) data from bone marrow-derived macrophages (BMDMs) of Mtdh^fl/fl and Mac-Mtdh^KO mice, originally published by Devanand Sarkar et al,¹⁸ were reanalyzed. The raw sequencing data were downloaded from the Gene Expression Omnibus (GEO) database (accession number: GSE107691; available at https://identifiers.org/geo:GSE107691). Differentially expressed genes (DEGs) were identified, and Gene Ontology (GO) term enrichment and volcano plot analyses were performed using the R statistical software packages.

Cell migration assay

For the migration assay, NIH3T3 cells (5×10⁴ cells per insert) were suspended in conditioned medium with or without the addition of LSKL and seeded into the upper chamber of a 24-well Transwell insert (membrane pore size, 8 µm; Corning). DMEM medium supplemented with 10% FBS was added to the lower chamber. After 24 hours of incubation, migrated cells were fixed with 4% paraformaldehyde and stained with crystal violet for 15 min. Images were captured using an upright fluorescence microscope (Olympus).

Data analysis

Experimental data are presented as mean±SEM or mean±SD of at least three independent experiments. Statistical significance between two groups was determined using a two-tailed Student’s t-test. One-way analysis of variance was performed for multiple-group comparisons. Spearman’s rank correlation analysis was used to evaluate the association between MTDH and THBS1 expression in colon adenocarcinoma (COAD) and skin cutaneous melanoma (SKCM). The relationship between THBS1 expression and clinicopathological parameters in COAD and SKCM was assessed using the χ² test. Differences were considered statistically significant at *p<0.05, **p<0.01, ***p<0.001.

Results

Macrophages with Mtdh knockout suppressed metastasis but promoted subcutaneous tumor growth of mouse colorectal cancer and melanoma cells

To establish a macrophage-specific Mtdh knockout mouse model, we bred Mtdh^fl/fl mice with LysM-Cre mice to delete Mtdh in the myeloid compartment (figure 1A), including macrophages. PMs isolated from Mtdh^fl/fl mice and Mac-Mtdh^KO mice were analyzed by western blotting and IF assays. Western blot results showed that the Mtdh protein level in PMs from Mac-Mtdh^KO mice was markedly lower than that in PMs from Mtdh^fl/fl mice (figure 1B,C), whereas lymphocytes and platelets retained normal physiological Mtdh expression levels compared with those from Mtdh^fl/fl mice (figure 1D). Moreover, IF analysis confirmed a significant reduction in Mtdh protein expression in PMs from Mac-Mtdh^KO mice (figure 1E,F), consistent with the previously reported ∼80% deletion efficiency in the macrophage-specific Mtdh knockout mice.¹⁸

Figure 1. Mac-MTDH^KO inhibits metastasis but enhances tumor growth in mouse colorectal cancer and melanoma models. (A) Schematic illustration of the generation of Mac-Mtdh^KO mice. (B) Representative western blot images and (C) quantitative analysis of Mtdh protein levels in peritoneal macrophages (PMs) isolated from Mtdh^fl/fl and Mac-Mtdh^KO mice. Protein expression was normalized to GAPDH. (D) Representative western blot analysis of Mtdh expression in platelets and lymphocytes from Mac-Mtdh^KO and Mtdh^fl/fl mice. (E) Representative confocal images showing Mtdh (green) and F4/80 (red) expression in PMs from Mac-Mtdh^KO and Mtdh^fl/fl mice. Scale bar: 25 µm. (F) Quantitative analysis showing the average number of Mtdh–positive macrophages in PMs from Mac-Mtdh^KO and Mtdh^fl/fl mice. (G) Representative gross images of metastatic lung nodules in Mtdh^fl/fl and Mac-Mtdh^KO mice intravenously injected with MC-38 cells. (H) H&E staining of lung metastatic nodules corresponding to panel (G). Scale bar: 500 µm. (I) Quantification of metastatic lung nodules shown in panel (G). (J) Representative gross images of metastatic lung nodules in Mtdh^fl/fl and Mac-Mtdh^KO mice intravenously injected with B16F10 cells. (K) H&E staining of lung metastatic nodules corresponding to panel (J). Scale bar: 500 µm. (L) Quantification of metastatic lung nodules shown in panel (J). (M) Tumor growth curve of subcutaneous MC-38 tumors implanted in Mtdh^fl/fl and Mac-Mtdh^KO mice. Tumors excised on day 17 were weighed. Statistical significance was calculated by the two-way ANOVA. (N) Quantitative analysis of tumor weight from subcutaneous MC-38 tumors shown in panel (M). (O) Representative images of subcutaneous MC-38 tumors corresponding to panel (M) (n=6 mice per group). (P) Quantitative analysis of tumor volume from subcutaneous B16F10 tumors implanted in Mtdh^fl/fl and Mac-Mtdh^KO mice. Tumor volumes were measured by caliper on day 17. (Q) Quantitative analysis of tumor weight from subcutaneous B16F10 tumors shown in panel (P). Tumors extracted on day 17 were weighed. (R) Representative images of subcutaneous B16F10 tumors corresponding to panel (P) (n=5 mice per group). Statistical significance was calculated by a two-tailed Student’s t-test. Data are presented as the mean±SEM. *p<0.05, **p<0.01, ***p<0.001. ANOVA, analysis of variance; PMs, peritoneal macrophages.

We next investigated whether macrophage-specific deletion of Mtdh affected tumor metastasis. To establish murine lung metastasis models, colorectal cancer (MC-38) and melanoma (B16F10) cells were intravenously injected into Mac-Mtdh^KO or Mtdh^fl/fl mice. Lung metastases were significantly reduced in Mac-Mtdh^KO mice compared with Mtdh^fl/fl controls injected with either MC-38 or B16F10 cells via the tail vein (figure 1G–L).

To evaluate subcutaneous tumor growth, MC-38 and B16F10 cells were implanted subcutaneously into Mac-Mtdh^KO and Mtdh^fl/fl mice. Subcutaneous tumors derived from MC-38 (figure 1M–O) or B16F10 (figure 1P–R) cells were markedly larger and heavier in the Mac-Mtdh^KO group than in their respective control group.

The effect of macrophage-specific Mtdh knockout on the tumor metastasis was consistent with our previous study, which demonstrated that macrophages with MTDH overexpression significantly promoted tumor invasion and metastasis.¹⁷ Unexpectedly, mice with macrophage-specific deletion of Mtdh exhibited accelerated subcutaneous tumor growth. This divergence between tumor growth and metastasis prompted us to investigate how conditional Mtdh knockout in macrophages produced such opposing effects.

Immune infiltration in tumor and lung tissues from subcutaneous and lung metastasis models

Because macrophages are among the most abundant immune cells in tumors,^{3 4} we compared the tumor immune microenvironment between Mac-Mtdh^KO and Mtdh^fl/fl mice in both the subcutaneous and lung metastasis models.

MC-38 subcutaneous tumors from Mac-Mtdh^KO and Mtdh^fl/fl mice were subjected to LC–MS/MS analysis and the composition of immune cell populations within the tumor samples was estimated using the TIMER database. The principal effectors of antitumor immunity—including T cells, natural killer cells, dendritic cells, and macrophages—play essential roles in directly killing tumor cells.^{24 25} However, no significant difference was observed in the relative abundance of these immune cells between Mac-Mtdh^KO and Mtdh^fl/fl tumors (online supplemental figure S1A,B). IF staining further confirmed that the numbers of infiltrating macrophages (F4/80⁺) (online supplemental figure S1C) and CD8⁺ T cells (online supplemental figure S1D) were comparable between the two groups. The immune microenvironment was also examined in lung tissues from the MC-38 metastasis model. Consistent with the subcutaneous findings, no difference was detected in the numbers of F4/80⁺ cells within metastatic lesions (online supplemental figure S1E) or in the surrounding lung stroma (online supplemental figure S1F) between Mac-Mtdh^KO mice and Mtdh^fl/fl mice. Moreover, IF staining with an anti-CD8 antibody showed that the number of CD8⁺T cells infiltrating metastatic sites was similar between the two groups (online supplemental figure S1G), whereas the lung stroma of Mac-Mtdh^KO mice displayed markedly greater CD8⁺ T-cell infiltration (online supplemental figure S1H).

The numbers of tumor-infiltrating immune cells in MC-38 subcutaneous tumors from Mac-Mtdh^KO and Mtdh^fl/fl mice were further analyzed by flow cytometry (online supplemental figure S2A). There were no significant differences in either the proportion or absolute number of CD8⁺ T cells (online supplemental figure S2B) between Mac-Mtdh^KO and Mtdh^fl/fl tumors. Compared with Mtdh^fl/fl subcutaneous tumors, the proportion of CD11b⁺F4/80⁺ macrophages in Mac-Mtdh^KO subcutaneous tumors was significantly increased (online supplemental figure S2C). We further performed flow cytometry to determine whether Mtdh knockout regulates the phenotype of tumor-associated macrophages (TAMs) in vivo. The proportion and absolute count of CD206⁺ macrophages did not differ significantly between the two groups (online supplemental figure S2D). The proportion and absolute number of M1-specific costimulatory molecule MHC-II (online supplemental figure S2E), as well as a functional protein iNOS (online supplemental figure S2F), was significantly lower in Mac-Mtdh^KO subcutaneous tumors compared with Mtdh^fl/fl tumors. No significant differences were observed in either the proportions or absolute number of Ly-6C⁺ macrophages (online supplemental figure S2G) between Mac-Mtdh^KO and Mtdh^fl/fl tumors. These findings indicate that Mtdh-deficient macrophages inhibit macrophage differentiation towards a pro-inflammatory phenotype without affecting the infiltration of CD8⁺ T cells.

Together, these data suggest that alterations in antitumor immunity are unlikely to account for the enhanced tumor growth observed in Mac-Mtdh^KO mice.

Knockout of Mtdh upregulated the expression and secretion of TSP-1 in macrophages

Robertson et al¹⁸ generated a mouse model with macrophage-specific deletion of Mtdh and performed RNA-seq analysis on BMDMs. We reanalyzed their RNA-seq dataset from the GEO database (GSE107691). Differential expression analysis revealed 54 DEGs between BMDMs from Mac-Mtdh^KO or Mtdh^fl/fl mice (figure 2A). GO enrichment analysis indicated that these genes were significantly associated with the regulation of angiogenesis (figure 2B). Among the DEGs, Thbs1 (encoding TSP-1), a natural inhibitor of angiogenesis, showed the highest basemean expression (figure 2C). To validate these findings, we examined TSP-1 expression in mouse macrophages and tumor tissues. qRT-PCR confirmed that Thbs1 mRNA levels were markedly increased in Mac-Mtdh^KO PMs compared with those from Mtdh^fl/fl mice (figure 2D). Consistently, western blot analysis showed that TSP-1 protein levels were also increased in Mac-Mtdh^KO PMs (figure 2E,F). ELISA further demonstrated that Mtdh knockout enhanced both TSP-1 expression and secretion in PMs (figure 2G). Similarly, elevated TSP-1 was observed in TAMs from Mac-Mtdh^KO tumors relative to Mtdh^fl/fl controls (figure 2H,I). Moreover, double IF staining for TSP-1 and the macrophage marker F4/80 revealed that increased TSP-1 expression in the lung tissues of Mac-Mtdh^KO mice was localized exclusively within macrophages lacking Mtdh expression (figure 2J,K).

Figure 2. Knockout of Mtdh upregulates the expression and secretion of TSP-1 in macrophages. (A) Volcano plot showing 54 DEGs between Mac-Mtdh^KO and Mtdh^fl/fl BMDMs from GEO database GSE107691. Red dots represent upregulated genes, and purple dots represent downregulated genes in Mac-Mtdh^KO BMDMs. (B) GO enrichment analysis showing that DEGs were significantly associated with the regulation of angiogenesis pathway. (C) Basemean expression values of selected DEGs, indicating that TSP-1 was the most highly upregulated gene. (D) qRT-PCR analysis of TSP-1 mRNA expression in PMs isolated from Mac-Mtdh^KO and Mtdh^fl/fl mice. (E) Representative western blot images and (F) corresponding quantification of TSP-1 protein levels in PMs from Mtdh^fl/fl and Mac-Mtdh^KO mice. Protein expression was normalized to GAPDH. (G) ELISA-based quantification of TSP-1 concentrations in culture supernatants from Mac-Mtdh^KO and Mtdh^fl/fl PMs. (H) Representative western blot images and (I) quantitative analysis of TSP-1 and Mtdh protein levels in tumor-associated macrophages (TAMs) isolated from subcutaneous MC-38 tumors in Mtdh^fl/fl and Mac-Mtdh^KO mice. Protein expression was normalized to GAPDH. (J) Representative confocal images showing co-staining of TSP-1 (green) and F4/80 (red) in lung tissues of Mac-Mtdh^KO and Mtdh^fl/fl mice. White arrows indicate macrophages. Scale bar: 25 µm. (K) Quantification of TSP-1-positive macrophages per field in lung tissues from Mac-Mtdh^KO and Mtdh^fl/fl mice. (L) Platelet counts in peripheral blood from Mac-Mtdh^KO and Mtdh^fl/fl mice. (M) Percentage of CD62P⁺/CD41⁺ platelets in whole blood from Mac-Mtdh^KO and Mtdh^fl/fl mice. Statistical significance was calculated by a two-tailed Student’s t-test. Data are presented as the mean±SD. *p<0.05, **p<0.01, ***p<0.001. BMDMs, bone marrow-derived macrophages; DEGs, differentially expressed genes; GEO, Gene Expression Omnibus; GO, gene ontology; qRT-PCR, quantitative real-time PCR; TSP-1, thrombospondin-1.

TSP-1 is a glycoprotein secreted by platelets on activation and participates in multiple interactions that regulate platelet function and thrombus formation.²⁶ We performed flow cytometry to assess platelet counts and activation status in peripheral blood samples from Mtdh^fl/fl and Mac-Mtdh^KO mice. In Mac-Mtdh^KO mice, platelet counts in peripheral blood were lower than those in Mtdh^fl/fl mice (figure 2L). However, no significant differences in platelet activation were observed between the two groups (figure 2M). These findings suggest that macrophages lacking Mtdh influence platelet production without altering platelet activation. This observation is consistent with previous studies showing that TSP-1 inhibits platelet production but does not affect platelet activation.27,29

Knockout of Mtdh in macrophages diminished angiogenesis in both subcutaneous tumors and lung metastases

TSP-1, a matricellular protein, acts as an endogenous inhibitor of angiogenesis.^{30 31} We first assessed angiogenesis in tumor tissues from the subcutaneous tumor model. As shown in online supplemental figure S3A–C), Mac-Mtdh^KO mice developed larger tumors with reduced angiogenesis, as indicated by decreased CD31 protein expression in subcutaneous tumors derived from MC-38 cells. Western blot analysis of the MC-38 subcutaneous tumors (figure 3A,B) yielded results consistent with those from the IHC and IF assays (online supplemental figures S2D and S3A–C). We next evaluated angiogenesis in lung tissues from the MC-38 metastasis model. Reduced CD31-positive staining was observed in both metastatic foci (online supplemental figure S3D,E) and the surrounding lung stroma (online supplemental figure S3F,G) of Mac-Mtdh^KO mice intravenously injected with MC-38 cells. These findings were further corroborated by western blot analysis for CD31 and CD8 expression in lung tissues from Mac-Mtdh^KO and Mtdh^fl/fl mice (figure 3C,D). Collectively, these results demonstrated that Mtdh loss in macrophages suppressed angiogenesis in both subcutaneous tumor and lung metastatic tissues.

Figure 3. Knockout of Mtdh in macrophages diminishes angiogenesis in tumor and lung tissues from subcutaneous and metastatic mouse models. (A) Representative western blot images and (B) quantitative analysis of CD31 and CD8 protein levels in subcutaneous MC-38 tumors from Mtdh^fl/fl and Mac-Mtdh^KO mice. Protein expression was normalized to GAPDH. (C) Representative western blot images and (D) corresponding quantification of CD31 and CD8 protein levels in lung tissues from MC-38 metastasis model. Protein expression was normalized to GAPDH. (E) Representative phase-contrast images showing tube formation of bEnd.3 cells 6 hours after treatment with conditioned medium from peritoneal macrophages (PMs) isolated from Mtdh^fl/fl or Mac-Mtdh^KO mice. (F) Quantification of tube formation numbers shown in panel (E). (G) Representative images of bEnd.3 cell migration 36 hours after treatment with conditioned medium from PMs of Mtdh^fl/fl and Mac-Mtdh^KO mice. (H) Quantitative analysis of wound closure in the scratch assay shown in panel (G). (I) Representative images showing tube formation of bEnd.3 cells 6 hours after treatment with conditioned medium from tumor-associated macrophages (TAMs) derived from Mtdh^fl/fl or Mac-Mtdh^KO subcutaneous tumors. (J) Quantification of tube formation numbers shown in panel (I). (K) Representative images of bEnd.3 cell migration 36 hours after treatment with conditioned medium from TAMs derived from Mtdh^fl/fl or Mac-Mtdh^KO subcutaneous tumors. (L) Quantitative analysis of wound closure in the scratch assay shown in (K). Statistical significance was calculated by a two-tailed Student’s t-test. Data are presented as the mean±SD. *p<0.05, **p<0.01.

To confirm that macrophages with Mtdh deficiency contribute to the inhibition of angiogenesis, we examined the effect of conditioned medium derived from PMs on the tube formation of bEnd.3 cells cultured on Matrigel in vitro. Only sparse tube-like structures were observed when bEnd.3 cells were treated with conditioned medium from Mtdh^fl/fl PMs. In contrast, conditioned medium from Mac-Mtdh^KO PMs resulted in a significant reduction in the number of vessel-like structures (figure 3E,F). As shown in figure 3G,H, conditioned medium derived from Mac-Mtdh^KO PMs also markedly attenuated the migratory capacity of bEnd.3 cells. To further validate the anti-angiogenic role of macrophages lacking Mtdh, we evaluated the effects of TAMs on bEnd.3 endothelial tube formation in vitro. While Mtdh^fl/fl TAMs supported robust tube-like network formation, Mac-Mtdh^KO TAMs significantly reduced the number of vessel-like structures formed by bEnd.3 cells (figure 3I,J). Consistently, wound-healing migration assays showed that Mac-Mtdh^KO TAMs markedly suppressed the migration of bEnd.3 cells (figure 3K,L). Collectively, these findings indicate that Mtdh deficiency in macrophages suppresses angiogenesis in vitro.

Knockout of Mtdh in macrophages promoted fibrosis in subcutaneous tumors but not in lung metastases

Another well-known function of TSP-1 is its ability to activate TGF-β1, thereby regulating fibrosis and contributing to the pathogenesis of fibrotic disease.32,34 Previous studies have reported that high concentrations of TSP-1 induce the activation of latent TGF-β1^{35 36} and that activation of the TSP-1/TGF-β1 axis promotes both fibrosis and inflammation.^{37 38} We next examined the extent of fibrosis in tumor and lung tissues from the subcutaneous tumor and lung metastasis models. A significant increase in fibrosis was observed in subcutaneous MC-38 tumors from Mac-Mtdh^KO mice compared with those from Mtdh^fl/fl mice, as evidenced by Sirius Red and Masson’s trichrome staining (figure 4A). Moreover, Mac-Mtdh^KO mice exhibited a marked elevation in α-SMA expression—a marker for myofibroblast—in subcutaneous tumors (figure 4B,C). Consistently, enhanced fibrosis was also detected in subcutaneous B16F10 tumors from Mac-Mtdh^KO mice relative to those from Mtdh^fl/fl mice (figure 4D). In contrast, the lung stroma of Mac-Mtdh^KO mice displayed fibrosis levels comparable to those observed in Mtdh^fl/fl controls (figure 4E). Collectively, these results indicate that Mtdh deficiency in macrophages promotes fibrosis in subcutaneous tumor tissues but does not affect fibrotic progression in lung metastases.

Figure 4. Knockout of Mtdh in macrophages augments fibrosis in tumor tissues of the subcutaneous tumor model. (A) Representative Sirius Red and Masson’s trichrome staining of subcutaneous MC-38 tumors from Mtdh^fl/fl and Mac-Mtdh^KO mice. Scale bar: 100 µm. (B) Representative confocal images showing a-SMA expression in subcutaneous MC-38 tumors from Mtdh^fl/fl and Mac-Mtdh^KO mice. Scale bar: 40 µm. (C) Quantification of α-SMA-positive cells per field in subcutaneous MC-38 tumors. Statistical significance was calculated by a two-tailed Student’s t-test. (D) Representative Masson’s trichrome staining of subcutaneous B16F10 tumors. Scale bar: 100 µm. (E) Representative Sirius Red staining of lung stroma from the MC-38 lung metastasis model. Scale bar: 100 µm. (F) Correlation between THBS1 expression and pathological T and M stages in colon adenocarcinoma (COAD) from the TCGA dataset. Statistical significance was calculated by the χ² test. (G) Correlation between THBS1 expression and pathological T and M stages in skin cutaneous melanoma (SKCM) from the TCGA dataset. Statistical significance was calculated by χ² test. (H) Correlation between THBS1 expression and CAF infiltration levels in COAD, determined using TIMER. ρ indicates Spearman’s correlation coefficient. Correlation was calculated by Spearman’s rank correlation. (I) Correlation between THBS1 expression and CAF infiltration levels in SKCM, determined using TIMER. ρ indicates Spearman’s correlation coefficient. Correlation was calculated by Spearman’s rank correlation. *p<0.05. CAF, cancer-associated fibroblast; COAD, colon adenocarcinoma; TCGA, The Cancer Genome Atlas. *p<0.05.

We further examined the relationship between THBS1 expression and clinicopathological features associated with tumor growth and metastasis in patients with COAD and SKCM using data from The Cancer Genome Atlas (TCGA) database. It was shown that higher THBS1 expression levels were detected in tumors with advanced pathological T stages in COAD (figure 4F,G). Using the TIMER database,²³ we also found a positive correlation between THBS1 expression and cancer-associated fibroblast (CAF) infiltration in both COAD and SKCM (figure 4H,I). A statistically significant but weak positive correlation was observed between MTDH and THBS1 expression levels in COAD (R²=0.01475, online supplemental figure S4A) and SKCM (R²=0.04874, online supplemental figure S4B). However, the biological significance of this association remains ambiguous because of the limited explanatory capacity of the model.

Collectively, these findings suggest that macrophage-specific Mtdh ablation promotes increased fibrosis in subcutaneous tumors but not in lung metastases.

Knockout of Mtdh in macrophages promoted TGF-β1 activation and FMT

Within the TME, cancer cells and activated fibroblasts are major sources of TGF-β1. Latent TGF-beta1 is secreted and stored as a biologically inactive form in the extracellular matrix, while active TGF-beta1 is the form that can exert biological effects and plays a pivotal role in fibrosis and tumor progression.^{39 40} ELISA results showed that the concentration of active TGF-β1 was comparable between Mac-Mtdh^KO and Mtdh^fl/fl PMs (figure 5A). Similarly, no significant difference was observed in the total TGF-β1 concentration between MC-38 and B16F10 supernatants (figure 5B). To assess TGF-β1 activation in the TME, tumor cells (MC-38 or B16F10) and NIH3T3 fibroblasts were cocultured with conditioned medium from PMs. The results demonstrated that conditioned medium derived from Mac-Mtdh^KO PMs significantly enhanced TGF-β1 activation in the coculture supernatants of both MC-38 and B16F10 cells (figure 5C,D). However, when macrophages were cocultured with NIH3T3 fibroblasts, no significant difference was observed in the concentration of active TGF-β1 between the two groups (figure 5E).

Figure 5. Supernatants from Mac-Mtdh^KO-coculture promote fibroblast-to-myofibroblast transition via the TSP-1/TGF-β1 signaling pathway. (A) ELISA-based quantification of active TGF-β1 levels in culture supernatants from Mac-Mtdh^KO and Mtdh^fl/fl peritoneal macrophages (PMs). (B) ELISA-based quantification of total TGF-β1 levels in culture supernatants from MC-38 and B16F10 cells. (C) ELISA-based quantification of active TGF-β1 in supernatants from cocultures of Mac-Mtdh^KO or Mtdh^fl/fl PMs with MC-38 cells. (D) ELISA-based quantification of active TGF-β1 in supernatants from coculture of Mac-Mtdh^KO or Mtdh^fl/fl PMs with B16F10 cells. (E) ELISA-based quantification of active TGF-β1 in supernatants from cocultures of Mac-Mtdh^KO or Mtdh^fl/fl PMs with NIH3T3 cells. (F) Representative confocal images showing co-staining of α-SMA (green) and F4/80 (red) in subcutaneous MC-38 tumors from Mac-Mtdh^KO and Mtdh^fl/fl mice. Scale bar: 25 µm. (G) Graphs represent Manders’ overlap coefficient showing mean±SD indicated by colored dots in panel (F). (H) Representative confocal images showing co-staining of α-SMA (red) and CD31 (green) in subcutaneous MC-38 tumors from Mac-Mtdh^KO and Mtdh^fl/fl mice. Scale bar: 25 µm. (I) Graphs represent Manders’ overlap coefficient showing mean±SD indicated by colored dots in panel (H). (J) Representative images of Transwell migration assays for NIH3T3 fibroblasts treated with supernatants from cocultures of Mac-Mtdh^KO or Mtdh^fl/fl PMs with MC-38 cells. Scale bar: 50 µm. (K) Quantification of migrated NIH3T3 fibroblasts in panel (J). (L) Representative western blot images and (M) quantitative analysis of α-SMA and fibronectin protein levels in NIH3T3 fibroblasts treated with supernatants from Mac-Mtdh^KO or Mac-Mtdh^fl/fl cocultures. Protein levels were normalized to GAPDH. (N) qRT-PCR analysis of COL1A1 mRNA expression in NIH3T3 fibroblasts treated with supernatants from Mac-Mtdh^KO or Mac-Mtdh^fl/fl cocultures. Statistical significance was calculated by a two-tailed Student’s t-test. Data are presented as the mean±SD from three independent experiments. *p<0.05, **p<0.01, ***p<0.001. PMs, peritoneal macrophages; TGF-β1, transforming growth factor β1.

It is well established that active TGF-β1 can induce the myofibroblast transition in various cell types, including fibroblasts,⁴¹ vascular endothelial cells,⁴² and macrophages.⁴³ Using IF staining, no co-localization of F4/80 with α-SMA (figure 5F,G), or CD31 with α-SMA (figure 5H,I) was observed in subcutaneous MC-38 tumors, indicating that Mtdh loss in macrophages did not affect the transition of macrophages or endothelial cells into myofibroblasts. We next stimulated mouse embryonic fibroblasts (NIH3T3) with supernatants derived from cocultures of tumor cells and Mac-Mtdh^KO or Mtdh^fl/fl PMs (referred to Mac-Mtdh^KO-coculture and Mtdh^fl/fl-coculture, respectively). Under treatment with Mac-Mtdh^KO-coculture-conditioned medium, NIH3T3 cells exhibited enhanced migration capacity (figure 5J,K). Furthermore, Mac-Mtdh^KO-coculture supernatants upregulated the protein expression of α-SMA and fibronectin—two well-established myofibroblast markers—in NIH3T3 cells (figure 5L,M). The expression of Col1a1 (another FMT marker) at the mRNA level was also significantly upregulated in NIH3T3 fibroblasts treated with Mac-Mtdh^KO-coculture supernatants compared with control cells (figure 5N). No significant differences were observed in the expression of the other markers between the two groups. Collectively, these results indicate that Mtdh-deficient PMs markedly promoted FMT in mouse fibroblasts through elevated levels of active TGF-β1 secreted by tumor cells.

We further cocultured mouse fibroblasts with supernatants derived from TAMs and evaluated their migratory capacity and TGF-β1-driven FMT. Compared with control cells, NIH3T3 fibroblasts cocultured with Mac-Mtdh^KO TAMs exhibited significantly enhanced migration (online supplemental figure S5A,B). This was accompanied by increased expression of α-SMA and fibronectin—key markers of FMT (online supplemental figure S5C,D). These findings were consistent with the effects of Mac-Mtdh^KO PMs on TGF-β1-induced FMT, further validating the role of PMs in this process.

Blockade of TSP-1 with the TSP-1/TGF-β1 inhibitor suppressed fibrosis, subcutaneous tumor growth, and FMT induced by Mtdh-deficient macrophages

To determine whether macrophages lacking Mtdh contribute to tumor development through regulation of the TSP-1/TGF-β1 axis in vivo, we treated mice bearing subcutaneous MC-38 tumors with the LSKL peptide. LSKL, a short peptide sequence, is known to inhibit TSP-1-mediated activation of TGF-β1. Blockage of the TSP-1/TGF-β1 axis by LSKL administration markedly inhibited tumor growth and reduced tumor weight (figure 6A–C), as well as decreased the number of α-SMA⁺ fibroblasts in subcutaneous MC-38 tumors from Mac-Mtdh^KO mice (figure 6D,E). These results indicate that LSKL treatment effectively attenuated the increased fibrosis and tumor growth induced by macrophages with Mtdh knockout.

Figure 6. Blockade of TSP-1 with a TSP-1/TGF-β1 inhibitor suppresses fibrosis, subcutaneous tumor growth, and FMT induced by macrophages with Mtdh loss. (A) Tumor growth curves of subcutaneous MC-38 tumors from Mac-Mtdh^KO or Mtdh^fl/fl mice treated with or without LSKL (30 mg/kg, intraperitoneally once daily). Statistical significance was calculated by the two-way ANOVA. Phosphate buffer saline (PBS) was used as the solvent to dissolve LSKL. (B) Representative tumor images from the four experimental groups (n=5 mice per group). (C) Quantification of tumor weight in subcutaneous MC-38 tumors from the four groups. Statistical significance was calculated by one-way ANOVA. (D) Representative confocal images showing α-SMA expression in subcutaneous MC-38 tumors from the four groups. Scale bar: 40 µm. (E) Quantification of α-SMA-positive cells per field in subcutaneous MC-38 tumors. Statistical significance was calculated by one-way ANOVA. (F) Representative images of Transwell migration assays for NIH3T3 fibroblasts treated with LSKL (2.25 µg/mL, 48 hours) and supernatants from Mtdh^fl/fl or Mac-Mtdh^KO cocultures. Scale bar: 50 µm. (G) Quantification of migrated NIH3T3 fibroblasts shown in panel (F). Statistical significance was calculated by one-way ANOVA. (H) Representative western blot images and (I) quantitative analysis of α-SMA and fibronectin protein levels in NIH3T3 fibroblasts treated with LSKL (2.25 μg/mL, 48 hours) and supernatants from Mtdh^fl/fl or Mac-Mtdh^KO cocultures. Protein levels were normalized to GAPDH. Statistical significance was calculated by one-way ANOVA. Data are presented as the mean±SD from at least three independent experiments. *p<0.05, **p<0.01. TGF-β1, transforming growth factor β1. ANOVA, analysis of variance; FMT, fibroblast-to-myofibroblast transition; TGF-β1, transforming growth factor-β1; TSP-1, thrombospondin-1.

We next investigated whether Mtdh deletion in macrophages promoted FMT through the TSP-1/TGF-β1 pathway in vitro. Transwell migration assays (figure 6F,G) showed that LSKL treatment inhibited the enhanced fibroblast migration induced by supernatants from Mac-Mtdh^KO-cocultures. Additionally, in the presence of Mac-Mtdh^KO-coculture supernatants, LSKL reduced the protein expression levels of α-SMA and fibronectin (figure 6H,I). Moreover, pharmacological inhibition of the TSP-1/TGF-β1 signaling axis using LSKL abolished both the increased migratory capacity (online supplemental figure S5E,F) and the upregulation of α-SMA and fibronectin (online supplemental figure S5G,H) of NIH3T3 cells induced by Mac-Mtdh^KO TAMs. Under the treatment with the supernatants from Mac-Mtdh^KO TAMs, the addition of LSKL also decreased the protein expression level of active TGF-β1 (online supplemental figure S5I,J). Collectively, these findings demonstrate that Mtdh-deficient macrophages promote FMT in mouse fibroblasts via a TSP-1/TGF-β1-dependent mechanism in vitro.

Mtdh knockout or knockdown upregulated p53-dependent TSP-1 transcription via nuclear retention of MDM2 in macrophages

To elucidate the mechanism by which Mtdh regulates TSP-1 expression in macrophages, we focused on p53, a well-known transcription factor for TSP-1.^{44 45} We isolated PMs from Mac-Mtdh^KO and Mtdh^fl/fl mice to assess the transcriptional activity of p53. qRT-PCR analysis revealed that TSP-1 (figure 2D), and the p53 target gene CDKN1A (p21) were significantly upregulated at the mRNA level in Mac-Mtdh^KO PMs compared with Mtdh^fl/fl PMs (figure 7A). To confirm the dependence of this effect on p53 activity, we treated Mac-Mtdh^KO and Mtdh^fl/fl PMs with PFT-β, a reversible inhibitor of p53-mediated transcriptional activation. Notably, Mtdh knockout failed to induce a significant increase in TSP-1, p21, and BAX expression when PFT-β was added to the culture medium (figure 7B–D). Potential p53-binding sites within the TSP-1 promoter regions were predicted using the JASPAR database (http://jaspar.genereg.net/), and the four motifs with the highest scores were selected for validation (figure 7E). ChIP assay results further confirmed that p53 was significantly enriched at the TSP-1 promoter region, with greater enrichment observed in Mac-Mtdh^KO PMs than in Mtdh^fl/fl controls (figure 7F). Collectively, these findings demonstrate that Mtdh knockout enhances TSP-1 expression in macrophages through p53-dependent transcriptional activation.

Figure 7. Mtdh deficiency transcriptionally upregulated the expression of TSP-1 via nuclear retention of MDM2 and disruption of the MDM2–p53 interaction in macrophages. (A) qRT-PCR analysis of p21 mRNA expression in PMs from Mtdh^fl/fl and Mac-Mtdh^KO mice. Statistical significance was calculated by a two-tailed Student’s t-test. (B) qRT-PCR analysis of TSP-1 mRNA expression in Mtdh^fl/fl and Mac-Mtdh^KO PMs treated with or without PFT-β (10 µM, 48 hours). Statistical significance was calculated by one-way ANOVA. (C) qRT-PCR analysis of p21 mRNA expression in Mtdh^fl/fl and Mac-Mtdh^KO PMs treated with or without PFT-β (10 µM, 48 hours). Statistical significance was calculated by one-way ANOVA. (D) qRT-PCR analysis of BAX mRNA expression in Mtdh^fl/fl and Mac-Mtdh^KO PMs treated with or without PFT-β (10 µM, 48 hours). Statistical significance was calculated by one-way ANOVA. (E) Predicted p53-binding motifs generated using JASPAR database, with schematic representation of potential p53-binding sites in the TSP-1 promoter region. (F) ChIP-qRT-PCR analysis showing p53 enrichment in the TSP-1 promoter region in PMs from Mtdh^fl/fl and Mac-Mtdh^KO mice. Statistical significance was calculated by a two-tailed Student’s t-test. (G) Co-immunoprecipitation assay of PM lysates from Mtdh^fl/fl and Mac-Mtdh^KO mice using anti-p53 antibody, followed by western blotting with anti-p53 and anti-MDM2 antibodies. (H) Representative confocal images showing MDM2 (green) localization in PMs from Mtdh^fl/fl and Mac-Mtdh^KO mice. Scale bar: 25 µm. The inset showed a zoom on one cell. Scale bar: 5 µm. (I) Quantification of cells exhibiting nuclear localization of MDM2. Statistical significance was calculated by a two-tailed Student’s t-test. (J) Representative western blot images and (K) quantification of cytoplasmic and nuclear MDM2 levels in Mac-shMTDH and Mac-shRNA; protein levels were normalized to Lamin B1. Statistical significance was calculated by a two-tailed Student’s t-test. Data are presented as the mean±SD. *p<0.05, **p<0.01, ***p<0.001. ANOVA, analysis of variance; MDM-2, murine double minute-2; PMs, peritoneal macrophages; TSP-1, thrombospondin-1;

MDM2, a critical negative regulator of the p53 tumor suppressor, binds to p53 and inhibits its transcriptional activity by promoting p53 ubiquitination and proteasomal degradation.^{46 47} MDM2 is predominantly localized in the nucleus, where nuclear MDM2 plays a central role in regulating p53 transcriptional function.^{48 49} Compared with Mtdh^fl/fl PMs, Mtdh knockout markedly impaired the interaction between MDM2 and p53 in Mac-Mtdh^KO PMs (figure 7G). Furthermore, IF assays revealed that while MDM2 was primarily localized in the nucleus of Mtdh^fl/fl PMs, it became largely redistributed to the cytoplasm following Mtdh knockout (figure 7H,I). A similar reduction in nuclear MDM2 localization was also observed in Mtdh^fl/fl TAMs (online supplemental figure S6A,B) and pulmonary macrophages (online supplemental figure S6C,D) compared with their respective controls. Consistent with these findings, Co-IP assays demonstrated that silencing MTDH in macrophages disrupted the MDM2–p53 interaction (online supplemental figure S6E). Additionally, IF analysis showed that macrophages transduced with shMTDH (Mac-shMTDH) displayed markedly decreased nuclear localization of MDM2 compared with shRNA controls (Mac-shRNA) (online supplemental figure S6F). Western blot analysis further confirmed that nuclear MDM2 protein levels were significantly reduced in Mac-shMTDH compared with Mac-shRNA controls (figure 7J,K). Ubiquitination assays showed that although total cellular p53 ubiquitination levels were slightly increased following MTDH silencing (online supplemental figure S6G), nuclear p53 expression was markedly higher in 293T-shMTDH cells than 293T-shRNA controls (online supplemental figure S6H). Taken together, Mtdh knockout diminished nuclear MDM2 abundance and disrupted the MDM2–p53 interaction, thereby enhancing the transcriptional activation of p53-dependent target genes such as TSP-1. These data underscore that nuclear MDM2 localization—rather than total p53 ubiquitination—is the critical determinant of p53-dependent TSP-1 transcriptional regulation under MTDH deficiency or silencing.

Discussion

Previous studies have consistently reported elevated MTDH expression in multiple cancer types and its critical involvement in tumor progression and metastasis. However, the specific contribution of macrophage-derived MTDH to tumor development has remained largely unexplored. In our previous work, we demonstrated that MTDH overexpression in macrophages promoted the migration, invasion, and vascular mimicry of HNSCC cells, while exerting minimal influence on cancer cell survival.¹⁷ In the present study, we generated a macrophage-specific Mtdh knockout mouse model and discovered that macrophages lacking Mtdh discordantly regulated tumor growth and metastasis through distinct TSP-1–mediated mechanisms. Specifically, Mtdh deficiency in macrophages inhibited angiogenesis via the anti-angiogenic function of TSP-1, while simultaneously promoting tumor fibrosis and growth through the TSP-1/TGF-β–dependent pro-fibrotic axis. Mechanistically, MTDH knockdown reduced nuclear MDM2 levels and disrupted the MDM2–p53 interaction, leading to enhanced p53-dependent transcription of TSP-1 in macrophages. Collectively, these findings provide novel mechanistic insights into the paradoxical relationship between tumor growth and metastasis, emphasizing the multifaceted role of macrophage-derived MTDH in shaping the TME.

Macrophages are well-established contributors to tumor angiogenesis^{50 51} primarily through the secretion of a wide array of pro-angiogenic factors as well as angiogenesis inhibitors such as TSP-1. Moreover, several macrophage depletion strategies using liposome-encapsulated clodronate have been shown to markedly reduce angiogenesis in osteosarcoma⁵² and transplanted tumor models.⁵³ These findings collectively highlight that angiogenesis is essential for sustaining tumor growth and that macrophages play a central role in vascular remodeling as tumors progress to advanced carcinoma stages.⁵⁴ Our previous work demonstrated that MTDH overexpression in macrophages increased VEGFA secretion; however, this did not significantly influence angiogenesis or tumor growth in a subcutaneous head and neck cancer model.¹⁷ In the current study, although Mtdh-deficient macrophages inhibited angiogenesis in both metastatic and subcutaneous tumors, Mac-Mtdh^KO mice developed significantly larger tumors than Mtdh^fl/fl controls. Taken together, these findings suggest that while angiogenesis may be closely associated with tumor metastasis, it is not directly linked to the enhanced tumor growth induced by macrophage-derived MTDH deficiency.

High concentrations of TSP-1 have been reported to activate latent TGF-β1 and promote fibrotic signaling in both fibrotic diseases³⁵ and several types of cancer.³⁶ In the present study, macrophages lacking Mtdh exhibited increased expression and secretion of TSP-1, leading to a substantial elevation in active TGF-β1 levels within cancer cells. Moreover, inhibition of the TSP-1/TGF-β1 axis significantly attenuated FMT, fibrosis, and the enhanced tumor growth induced by Mtdh loss in macrophages in subcutaneous tumor models. Analysis of TCGA datasets further revealed that elevated THBS1 expression was associated with advanced pathological T stage in COAD and positively correlated with CAF infiltration in both COAD and SKCM. Collectively, these findings underscore the pivotal role of TSP-1-mediated FMT and fibrosis in promoting subcutaneous tumor growth in Mac-Mtdh^KO mice and suggest that concurrent targeting of macrophage Mtdh and the TSP-1/TGF-β1 pathway may represent an improved therapeutic strategy. Recent advances in macrophage-targeted therapies have demonstrated considerable promise in enhancing the clinical efficacy of cancer treatments.^{55 56} Given the critical role of macrophage-derived MTDH in regulating tumor growth and metastasis, a combined therapeutic approach targeting MTDH in macrophages alongside established modalities such as chemotherapy, radiotherapy, or immunotherapy could substantially improve therapeutic outcomes.

TSP-1 is encoded by the THBS1 gene, whose promoter can be directly activated by p53.⁴⁴ The ubiquitination of p53 by MDM2 represents the principal post-translational mechanism that facilitates p53 degradation.^{57 58} MDM2-mediated ubiquitination of p53 occurs in both the nucleus and cytoplasm,⁵⁹ and intranuclear ubiquitination of p53 represses its transcriptional activity.^{60 61} In this study, Mtdh knockout or knockdown markedly upregulated TSP-1 expression through nuclear retention of MDM2 and disruption of the MDM2–p53 interaction in macrophages. In contrast, nuclear localization of p53 was significantly increased in cells with MTDH silencing, despite a slight elevation in total p53 ubiquitination. These findings underscore the critical role of nuclear MDM2 in regulating the transcriptional activation of TSP-1 induced by MTDH deficiency in macrophages.

Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information.