본문으로 건너뛰기
← 뒤로

Deubiquitinating enzyme JOSD2 modulates cGAS to facilitate immune evasion in colorectal cancer.

1/5 보강
Oncoimmunology 📖 저널 OA 100% 2025: 71/71 OA 2026: 27/27 OA 2025~2026 2025 Vol.14(1) p. 2590245
Retraction 확인
출처

PICO 자동 추출 (휴리스틱, conf 2/4)

유사 논문
P · Population 대상 환자/모집단
환자: high microsatellite instability (MSI-H), highlighting the critical need to identify key molecular drivers of immune evasion in CRC
I · Intervention 중재 / 시술
추출되지 않음
C · Comparison 대조 / 비교
추출되지 않음
O · Outcome 결과 / 결론
we demonstrated both and that the inhibition of JOSD2 activated the cGAS-STING signaling pathway, leading to robust anti-tumor effects in CRC.

Du J, Fan Y, Fu J, Ge F, Zhu H, Yang S

📝 환자 설명용 한 줄

Colorectal cancer (CRC) remains a formidable clinical challenge, many patients exhibit limited responses to conventional chemotherapy and targeted therapies.

🔬 핵심 임상 통계 (초록에서 자동 추출 — 원문 검증 권장)
  • 95% CI 3.6-7.96
  • HR 4.79

이 논문을 인용하기

↓ .bib ↓ .ris
APA Du J, Fan Y, et al. (2025). Deubiquitinating enzyme JOSD2 modulates cGAS to facilitate immune evasion in colorectal cancer.. Oncoimmunology, 14(1), 2590245. https://doi.org/10.1080/2162402X.2025.2590245
MLA Du J, et al.. "Deubiquitinating enzyme JOSD2 modulates cGAS to facilitate immune evasion in colorectal cancer.." Oncoimmunology, vol. 14, no. 1, 2025, pp. 2590245.
PMID 41351298 ↗

Abstract

Colorectal cancer (CRC) remains a formidable clinical challenge, many patients exhibit limited responses to conventional chemotherapy and targeted therapies. Although immunotherapy has demonstrated potential, its efficacy is largely restricted to a subset of patients with high microsatellite instability (MSI-H), highlighting the critical need to identify key molecular drivers of immune evasion in CRC. Through comprehensive bioinformatic analysis, we identified the deubiquitinating enzyme JOSD2 as a key player in CRC progression, with elevated expression correlating with poor prognosis in MSI-H patients (HR = 4.79, 95% CI: 3.6-7.96) and dysregulation of multiple immune-related pathways. Mechanistically, we discovered that JOSD2 suppresses cGAS enzymatic activity by removing K27-linked ubiquitination, thereby promoting M2 polarization of macrophages, a process critical for immunosuppression in the tumor microenvironment. Furthermore, using the JOSD2 catalytic inhibitor HY041004, we demonstrated both and that the inhibition of JOSD2 activated the cGAS-STING signaling pathway, leading to robust anti-tumor effects in CRC. These findings not only uncover a novel immunomodulatory mechanism in colorectal cancer but also provide a therapeutic rationale for the development of JOSD2-targeted anticancer strategies.

🏷️ 키워드 / MeSH 📖 같은 키워드 OA만

같은 제1저자의 인용 많은 논문 (5)

📖 전문 본문 읽기 PMC JATS · ~68 KB · 영문

Introduction

Introduction
Colorectal cancer ranks as the third most common malignancy worldwide.1-4 Therapy resistance and adverse effects of standard treatments cause poor prognosis in CRC.5-8 Immunotherapy has emerged as a transformative therapeutic paradigm, demonstrating unprecedented success in CRC patients with MSI-H or deficient mismatch repair (dMMR).9,10 However, such cases represent a mere 5% of all CRC patients.9,11 Despite combinatorial regimens are applied to enhance immunotherapy efficacy, clinical outcomes remain suboptimal.12 These challenges underscore the critical need to identify novel molecular mechanisms underpinning immune evasion of CRC, develop innovative immunotherapy approaches.
Deubiquitinating enzymes (DUBs) play pivotal roles in tumor immunology by regulating protein stability and localization through substrate-specific deubiquitination, modulating immune checkpoint activation, cytokine signaling, and antigen presentation.13 For instance, USP7 and CSN5 sustain PD-L1 by preventing its proteasomal degradation, thereby suppressing T cell-mediated antitumor immunity.14,15 Conversely, CYLD inhibits NF-κB signaling by cleaving TRAF2/6 polyubiquitin chains, maintaining inflammatory homeostasis,16 while USP21 stabilizes Foxp3 to preserve regulatory T cell immunosuppressive activity.17 The enzymatic dependency of DUBs on conserved cysteine residues further renders them pharmacologically tractable, offering a rationale for DUB-targeted immunotherapies.
JOSD2, a member of the Machado-Joseph disease (MJD) domain DUB family,18 exhibits context-dependent oncogenic roles in cancer through diverse mechanisms: stabilizing glycolytic complexes to fuel metabolic reprogramming in lung adenocarcinoma;19 promoting SMAD4 stabilization and nuclear translocation to accelerate breast cancer metastasis;20 and inactivating LKB1 by removing K6-linked ubiquitin chains in NSCLC, effectively subverting its tumor-suppressive function.21 Despite these advances, JOSD2’s influence on the tumor immune microenvironment-particularly its potential role in promoting immune evasion or modulating immune response-remains unexplored, presenting a critical knowledge gap in immuno-oncology.
In our model, colorectal tumors in JOSD2 knockout (KO) mice exhibited significantly reduced growth compared to wild-type (WT) controls. This effect coincided with macrophage-intrinsic activation of the cGAS-STING pathway, suggesting that JOSD2 plays a role in shaping an immunosuppressive microenvironment. Cyclic GMP-AMP synthase (cGAS), a cytosolic DNA sensor, functions as a pattern recognition receptor for exogenous or tumor-derived double-stranded DNA (dsDNA).22 Upon dsDNA binding, cGAS catalyzes the synthesis of cyclic GMP-AMP (cGAMP), which activates the stimulator of interferon genes (STING). This triggers a phosphorylation cascade involving TANK-binding kinase 1 (TBK1) and interferon regulatory factor 3 (IRF3), culminating in the production of type I interferons (IFN-β), tumor necrosis factor-alpha (TNFα), and other cytokines.23 Under physiological conditions, cGAS-STING signaling bolsters innate immune surveillance by enhancing tumor cell recognition and elimination while simultaneously priming adaptive T-cell responses against malignancies.24,25 However, chronic STING activation paradoxically fosters immunosuppression by inducing T-cell exhaustion and pro-tumorigenic inflammation.26 Given these complexities, elucidating the regulatory impact of JOSD2 on this pathway and its broader immunomodulatory effects, emerges as a critical next step.
In this study, we observed that elevated JOSD2 expression is correlated with poor prognosis in CRC patients and associated with key immune evasion signatures. JOSD2 KO mouse models displayed significantly reduced tumor growth and enhanced cGAS-STING activation in macrophages, suggesting JOSD2’s involvement in CRC immune evasion. Mechanistically, we demonstrated that JOSD2 antagonizes cGAS-STING activation by selectively cleaving K27-linked polyubiquitin chains from cGAS without altering its protein stability, thereby suppressing its enzymatic activity. This post-translational inhibition skews macrophage polarization toward an M2 immunosuppressive phenotype, facilitating tumor immune escape. Collectively, our findings reveal a previously unrecognized oncogenic role of JOSD2 in tumor immunology, delineate a novel ubiquitin-dependent regulatory axis for cGAS activity, and identify JOSD2 as a druggable target for enhancing immunotherapy efficacy in CRC.

Materials and methods

Materials and methods

Mice
Josd2-deficient C57BL/6J mice were purchased from GemPharmatech Co., Ltd. (Jiangsu, China). Genomic DNA from tail biopsies was analyzed via PCR to confirm successful gene knockout. Using the CRISPR/Cas9 technology, a deletion was generated in exon 2 of Josd2-201 (ENSMUSG00000038695.14), a region that encompasses the ATG start codon and the cysteine residue at position 24. This cysteine residue is critical for its deubiquitinase activity. This deletion ensures the disruption of the protein’s functional domain while leaving the expression of other genes unaffected. Two pairs of genotyping test primers were designed flanking the KO segment:
Josd2 KO Genotype test 1 F: GCGAGAACCCTGGTTATTGAGGA,
R: TCGAGCTGCTCAGCATGTCCCT;
Josd2 KO Genotype test 2 F: AGAGCTGTGTGCTGTCCATGCT,
R: GCTCAGATAAGAGAGGTTGAGGCC.
For homozygous wild-type mice, the PCR product of test 1 is 1325 bp, and the PCR product of test 2 is 265 bp; for homozygous JOSD2 knockout mice, the PCR product of test 1 is 255 bp, and the PCR product of test 2 is 0 bp.
All the mice were housed and maintained in a specific pathogen-free (SPF) facility. All experimental mice were euthanized by exposure to CO2 at the experimental endpoint. The animal experiments were approved by the Institutional Animal Care and Use Committee (IACUC) of the Innovation Institute for Artificial Intelligence in Medicine of Zhejiang University (Hangzhou, China) with ethical approval numbers: DW202208291501, DW202212051051, DW202406281329, and DW202409191037. Each mouse was subcutaneously injected with 400,000 cells in the axillary region to establish the MC38 syngeneic tumor model. All the procedures were conducted in accordance with the IACUC. All animal experiments were performed in accordance with the ARRIVE guidelines.

Bone marrow-derived macrophages (BMDMs)
Femurs were dissected from 6- to 8-week-old wild-type or JOSD2 knockout mice and flushed with DMEM to isolate bone marrow cells. The samples were subsequently centrifuged to collect primary myeloid cells. The cells were cultured in DMEM with 10% fetal bovine serum, 100 U/mL penicillin-streptomycin, and 30 ng/mL macrophage colony-stimulating factor in a humidified incubator at 37 °C with 5% CO₂ to induce differentiation. On day 5, the entire medium was replaced. By day 7, the BMDMs were ready for use.

Cells culture
The human cervical cancer HeLa cell line and the human embryonic kidney cell line HEK-293T were purchased from the Cell Bank of the China Science Academy. The murine colorectal cancer cell line MC38 was purchased from Nanjing Cobioer Biosciences Co., Ltd. It is characterized by a high tumor mutation burden (TMB) and is highly responsive to immunotherapy, as it shares features with human MSI-H CRC. The cells were cultured in DMEM (Gibco, Grand Island, NY, USA) supplemented with 10% FBS at 37 °C in a 5% CO2 humid atmosphere. Mycoplasma contamination was routinely monitored every six months.

Plasmids, antibodies, and common reagents
The lentiviral packaging plasmids pCMV-p8.9 and pMDG-VSVG were kindly provided by Dr. Cohen at the Children’s Hospital of Los Angeles. The empty vector pCDH and pLKO.1 were purchased from Addgene (Watertown, MA, USA). The targeted fragments or gene-specific shRNAs were subsequently cloned into an empty vector plasmid by CloneEZ PCR cloning kit (GeneScript, Nanjing, China). The relevant sequences for the shRNAs used were as follows:
shJOSD2#1: CCAGGTGGACGGTGTCTACTA;
shJOSD2#2: CACCGGCAACTATGATGTCAA.
The antibodies used for immunoblotting were as follows: anti-P-TBK1 (#5483), anti-TBK1 (#3504), anti-P-STING (#72971), anti-STING (#50494), and anti-cGAS-mouse (#31659) antibodies were from Cell Signaling Technology (Danvers, MA, USA); anti-RNF185 (#A16158), anti-P-STAT6 (#AP0456), anti-STAT6 (#A0755), and anti-CD206 (#A11192), antibody were from ABclonal (Wuhan, China); anti-cGAS-human (#sc−11384) antibody was from Zenbio (Chengdu, China); anti-JOSD2 (#SAB2103354) was purchased from Sigma–Aldrich (St. Louis, MO, USA); anti-Flag (#db7002), anti-GAPDH (#db106), anti-ACTIN (#db11112), anti-HA (#db2603), anti-rabbit and anti-mouse secondary antibodies (db10002/db10003) were purchased from Diagbio (Hangzhou, China); anti-ARG1 (#ET1605) was purchased from Huabio (Hangzhou, China); anti-CD8 (#Ab237709) and anti-CD209(#Ab37220) were purchased from Abcam (Abcam plc, Cambridge, United Kingdom). The anti-FLAG resin beads (#L00425) and SureBeadsTM Protein G (#161-4023) for co-immunoprecipitation (Co-IP) were obtained from GenScript, Biomake (Houston, TX 77014, USA) and Bio-Rad (Hercules, CA, USA), respectively. For dual immunofluorescence, DAPI (#EF704) was purchased from Dojindo Molecular Technologies (Tokyo, Japan); TSA dual immunofluorescence kit (#RK05902), anti-CD163 (#A26411PM), and anti-F4/80 (#A23788) were purchased from ABclonal (Wuhan, China). For flow cytometry analysis, anti-mouse CD16/32 (#156604), PerCP/Cyanine5.5 anti-mouse CD206 (#141716, clone C068C2) and APC anti-mouse F4/80 (#123116, clone BM8) were purchased from Biolegend (San Diego, USA). The proteasome inhibitor MG132, anti-mouse PD-L1 (B7-H1)-InVivo (#A2115) and rat IgG2b isotype control-InVivo (#A2116) were obtained from Selleck Chem (Houston, TX, USA). HT-DNA (#D6898) was purchased from Sigma–Aldrich (St. Louis, MO, USA). 2’3’-cGAMP was purchased from MedChemExpress (New Jersey, USA). Prestained protein marker was from Vazyme (#MP102, Najing, China).

Gene transfection
For transient protein overexpression, 1 μg of plasmid DNA was mixed with 2 μL of jetPRIME reagent (Polyplus, France; #114-15) in 200 μL of jetPRIME buffer, incubated, and delivered into target cells. To establish stable cell lines, lentiviral particles were generated by PEI-mediated co-transfection of the target plasmid with the packaging plasmids pCMV-p8.9 and pMDG-VSVG, followed by harvesting of viral supernatants 48 h post-transfection. The target cells were infected for 24 h with lentiviral particles in 6  mg/mL polybrene (1:1000 dilution; Solarbio, Beijing, China) and then cultured in fresh medium for 48 h post-infection. Empty vector-transfected cells served as controls.

Co-immunoprecipitation (Co-IP)
After plasmids transfection, the cells were lysed with 1% NP-40 or 4% SDS. Part of the lysate was reserved as the “input” sample, and the rest was incubated overnight with antibody-conjugated beads. After washing, the beads were collected as the “IP” sample. Western blotting was used to assess protein‒protein interactions or ubiquitination levels. For endogenous immunoprecipitation, anti-cGAS antibody was pre-incubated with SureBeads to immobilize it before use.

Quantitative reverse transcription polymerase chain reaction (qRT-PCR)
BMDMs subjected to different stimulations were lysed using the FastPure RNA Isolation Kit (#RC112, Vazyme, China) to extract mRNA, which was reverse transcribed with the HiScript III All-in-one RT SuperMix Perfect for qPCR kit (#R333, Vazyme, China). Real-time PCR was performed with Taq Pro Universal SYBR qPCR Master Mix (#Q712, China) to assess target gene expression. The relative expression determined by qPCR was calculated using the 2^–ΔΔCt method, with Actin as the reference gene. Primer information for the analysis was as follows:
Actin, forward: ATGCCACAGGATTCCATACCCAAGA;
Actin, reverse: CTCTAGACTTCGAGCAGGAGATGG;
Ifnβ, forward: TCCGAGCAGAGATCTTCAGGAA;
Ifnβ, reverse: TGCAACCACCACTCATTCTGAG;
Isg15, forward: GGCCACAGCAACATCTATGA;
Isg15, reverse: ACTGGGGCTTTAGGCCATAC;
Il6, forward: TAGTCCTTCCTACCCCAATTTCC;
Il6, reverse: TTGGTCCTTAGCCACTCCTTC;
Ym1, forward: GCTAAGGACAGGCCAATAGAA;
Ym1, reverse: GCATTCCAGCAAAGGCATAG;
Arg1, forward: ACAGCAAAGCAGACAGAACTA;
Arg1, reverse: GAAAGGAACTGCTGGGATACA.

ELISA assay
TNF-α and IFN-β ELISA kits use a sandwich ELISA method (#RK04875 and #RK00420, ABclonal, China). Tumor tissues (0.05 g) were homogenized in 400 μL of PBS at 60 Hz for 10 min and centrifuged. The supernatants and standards were incubated in cytokine antibody-coated wells (37 °C: 2 h for capture, 1 h for biotinylated detection antibody, 0.5 h for streptavidin-HRP). The cGAMP assay kit used a competitive ELISA method (#501700, Cayman, USA). The cell lysates were mixed with 2’3’-cGAMP-HRP tracer and antiserum, followed by overnight incubation (4 °C, dark). The TMB substrate reactions (20 min, 37 °C, dark) were terminated, and the absorbance (450 nm) was measured. Cytokine concentrations were calculated using standard curves adjusted for dilution factors.

Flow cytometry analysis
For single or multi-color staining, 1 × 106 primary cells were added to 100 μL of PBS supplemented with 3% BSA, and Fc receptors were blocked on ice for 20 min. For surface staining, antibodies were added at the recommended dilutions, incubated at room temperature in the dark for 30 min, and then washed with PBS. For intracellular staining, the cells were fixed with 4% PFA (0.2 mL, RT, dark, 20 min) and washed with permeabilization buffer (1 mL). The samples were resuspended in 100 μL of permeabilization buffer, incubated with intracellular antibodies at RT in the dark for 30  min, and then washed with permeabilization buffer (1 mL). Finally, the cells were resuspended in 0.5 mL of PBS and analyzed by flow cytometer. Live cells were first gated based on forward scatter (FSC-A) and side scatter (SSC-A) to exclude debris and dead cells (P1, 50,000 live cells). Within the live cell population, a secondary gate (P2) was applied using FSC-A vs. FSC-H to exclude doublets. The resulting singlet population was then analyzed for surface expression of F4/80 (APC, R660 channel) and CD206 (PerCP-Cy5.5, B690 channel), M2-polarized macrophages were defined as F4/80⁺CD206⁺ cells. Data were acquired on a Beckman Coulter, CytoFLEX and analyzed using FlowJo software (v10.8.1).

Dual immunofluorescence
Paraffin-embedded tissue sections were dewaxed using xylene and graded ethanol. Antigen retrieval was performed in citrate buffer. The sample areas were outlined, followed by incubation with peroxidase blocking reagent (20 min, RT, dark). After washing, the sections were blocked with 3% BSA and incubated overnight at 4 °C with appropriately diluted anti-CD163 primary antibody. The fluorescent-conjugated secondary antibody was added and incubated (50 min, RT), followed by TSA-enhanced green fluorescent probe incubated for 5 min. The sections were then subjected to antigen retrieval in citrate buffer to strip the antibodies (95 °C, 30 min). Anti-F4/80 antibody incubation and red fluorescent labeling were repeated. DAPI was added for nuclear staining (10 min, RT). After washing, the sections were mounted with anti-fade mounting medium and imaged by fluorescence microscope.

RNA-seq sample processing and analysis
The tumor samples were collected and sent to LC Biotechnology Co., Ltd. (Hangzhou, China) for RNA sequencing. The data were processed and analyzed using R (V4.4.1). Principal Component Analysis (PCA) was performed to assess sample variability. Differential gene expression analysis was conducted using the DESeq2 and limma packages. Subsequently, Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Set Enrichment Analysis (GSEA) were applied to identify enriched signaling pathways and functional categories.

Statistical analysis
JOSD2 expression across multiple cancers was analyzed using the Ualcan (http://ualcan.path.uab.edu). Its association with the immunotherapy response in CRC patients was evaluated using the Kaplan‒Meier (KM) plotter (https://kmplot.com). The correlation between JOSD2 and immune-related target genes were analyzed via TISIDB (http://cis.hku.hk/TISIDB/index.php). The Western blot band intensities were quantified using ImageJ. All the statistical analyses were performed using GraphPad Prism. The data presented are from three independent experiments. Two-tailed Student’s t-test (TTEST) was used for two-group comparisons, and one-way ANOVA followed by Tukey’s multiple comparisons test was performed for multiple groups. Significance was defined as follows: P < 0.05 (*), P < 0.01 (**), and P < 0.001 (***). Non-significant differences are indicated as n.s. or P ≥ 0.05.

Results

Results

JOSD2 plays a critical role in the tumor immune microenvironment of colorectal cancer
To systematically evaluate the clinical relevance of JOSD2 and validate its potential function as a proto-oncoprotein, we conducted multi-dimensional bioinformatics analyses. First, leveraging The Cancer Genome Atlas (TCGA) pan-cancer data and the UALCAN database, we systematically assessed the expression profile of JOSD2 across multiple malignancies. JOSD2 was expressed at significantly higher levels in various solid tumors compared to normal tissues (Figure 1A). We subsequently used the KM plotter to evaluate the impact of JOSD2 expression on the survival of CRC patients. The results revealed that patients with high JOSD2 expression presented shorter overall survival (OS) (Figure 1B).
Notably, CRC patients with MSI-H/dMMR accumulate extensive genetic mutations and DNA double-strand breaks, resulting in "hot tumor" characteristics with abundant neoantigens. This contrasts with the immunogenicity of microsatellite-stable (MSS) CRC.27 Strikingly, high JOSD2 expression correlated significantly with reduced OS exclusively in the MSI-H subgroup (Figure 1C), whereas no significant differences were detected between the MSS and MSI-L groups (Figure S1A). These findings suggest that the pro-tumorigenic effects of JOSD2 may depend on an immunologically active tumor microenvironment. Further investigation of the relationship between JOSD2 expression and immunotherapy response demonstrated that among 121 pan-cancer patients treated with anti-CTLA−4 therapy, those with high JOSD2-expressing tumors had significantly shorter median survival (Figure 1D).
Furthermore, we conducted an analysis using the TISID database to systematically investigate the correlation between JOSD2 expression and key immune signaling molecules. As illustrated in Figure 1E, JOSD2 expression was significantly positively correlated with the expression of immunosuppressive molecules, including ADORA2A,28
PVRL2,29 and LGALS9.30 In stark contrast, JOSD2 showed pronounced negative correlations with immunostimulatory factors such as inducible ICOS,31
CD86,32 and IL-6R.33 This dual regulatory mechanism reshaped the tumor immune microenvironment and collectively drove immune evasion in colorectal cancer.
We subsequently established syngeneic tumor models by subcutaneously inoculating MC38 colorectal cancer cells into both WT and JOSD2 KO mice to investigate tumor proliferation. As demonstrated in Figure 1F, MC38 tumor growth was significantly attenuated in JOSD2 KO mice, with the mean tumor volume at the experimental endpoint markedly reduced compared to WT controls. To characterize alterations in the tumor immune microenvironment alterations, RNA sequencing was performed on two tumor specimens from each group. PCA of the global gene expression profiles revealed clear segregation between the two groups (Figure S1B). Differential gene expression analysis revealed upregulated expression of macrophage-associated proteins (e.g., Lyz2, Cd74, Mpeg1, Mmp12) and MHC class II antigens (H2-Aa, H2-Eb1) involved in antigen presentation and the interferon response in JOSD2 KO tumors (Figure S1C), suggesting enhanced macrophage-mediated antigen processing. KEGG pathway enrichment analysis highlighted significant enrichment in cytokine‒receptor interactions, osteoclast (macrophage) differentiation, T-cell differentiation pathways, and inflammatory bowel disease-related signaling (Figure S1D). Furthermore, GSEA revealed pronounced upregulation of gene sets associated with the immunoglobulin complex (P < 0.001) and antigen binding (P < 0.001) in JOSD2 KO tumors (Figure S1E).
Integrated analyses of multi-omics databases and in vivo syngeneic tumor experiments demonstrated that JOSD2 not only directly drove malignant tumor cell proliferation by modulating oncogenic substrate proteins but also orchestrated tumor immune evasion through macrophage-mediated mechanisms. However, the precise molecular pathways by which JOSD2 regulates macrophage functionality require further in-depth investigation.

JOSD2 modulates the cGAS-STING signaling pathway
Macrophages participate in diverse immune-related pathophysiological34 and employ multiple signaling pathways to sense and respond to extracellular stimuli.35,36 Notably, macrophages predominantly rely on the cGAS-STING pathway, a nearly exclusive mechanism, for direct detection of cytosolic DNA. Within the tumor microenvironment, activation of the cGAS-STING axis enhances macrophage-mediated anti-tumor immunity by promoting antigen presentation and pro-inflammatory cytokine production.37
In this context, we analyzed RNA-seq data from JOSD2 KO tumors and observed significant upregulation of cGAS-STING pathway genes – including Sting1, Irf3, Tnf, Stat1, and Ccl5 –in the KO group (Figure 2A). Consistent with transcriptional activation, Western blot analysis of tumor lysates revealed elevated phosphorylation levels of the downstream signaling components IRF3 and TBK1 in the KO group (Figure 2B). We subsequently analyzed 459 colorectal cancer samples via the TISID database and identified inverse correlations between JOSD2 expression and the expression of key cGAS-STING-dependent chemokines CCL2, CXCL9, CXCL10, and CXCL13 (Figure S2A). These chemokines are critical for recruiting T cells, NK cells, and B cells to inflammatory or tumor microenvironments, thereby amplifying antiviral and antitumor immune responses.
To precisely evaluate whether cGAS-STING signaling alterations primarily originate from macrophage modulation within the tumor microenvironment, we isolated bone marrow cells from the femurs of WT and JOSD2 KO mice. Monocytic myeloid cells were differentiated into BMDMs and transfected with herring testes DNA (HT-DNA) to mimic cytoplasmic DNA activation following phagocytosis of tumor cell debris. Western blot analysis revealed increased phosphorylation of p-TBK1 and p-STING in WT BMDMs post-HT-DNA stimulation, with JOSD2 KO BMDMs exhibiting further increases in the levels of these phosphorylated forms (Figure 2C). Consistent with phosphoactivation dynamics, qRT-PCR and ELISA analyses showed synchronized upregulation of canonical cGAS-STING downstream effectors – Ifnβ, Isg15, Il6 transcripts and the enzymatic product cGAMP – in JOSD2 KO BMDMs (Figure 2D and 2E). We subsequently overexpressed JOSD2 or the empty vector control in HeLa cell lines and subjected them to HT-DNA stimulation. JOSD2 overexpression markedly attenuated the HT-DNA-induced phosphorylation of TBK1 and STING across multiple timepoints, indicating impaired DNA-sensing capacity (Figure S2B). Intriguingly, neither JOSD2 overexpression nor knockdown altered cGAS protein levels, suggesting that its regulatory role is specific to post-translational signaling events rather than cGAS stability (Figure S2C and S2D). Notably, stimulation with cGAMP induced comparable phosphorylation (p-TBK1/p-STING) and downstream cytokine (Ifnβ, Isg15) responses in WT and JOSD2 KO BMDMs (Figure 2F and S2E), which contrasts with the distinct effects observed upon HT-DNA stimulation.
Collectively, these results demonstrate that JOSD2 suppresses antitumor immunity by inhibiting the cGAS–STING signaling axis.

JOSD2 interacts with cGAS proteins
We investigated whether JOSD2 modulates cGAS-STING signaling through direct binding to pathway components. To test this hypothesis, we exogenously overexpressed JOSD2-HA in HEK-293T cells alongside Flag-tagged key components of the cGAS-STING cascade (cGAS, STING, TBK1, IRF3). Co-IP assays revealed specific interaction between JOSD2 and cGAS (Figure 3A). Reciprocal Co-IP experiments with Flag-tagged JOSD2 and HA-tagged cGAS confirmed this interaction (Figure 3B). To validate semi-endogenous binding, we overexpressed cGAS-Flag or JOSD2-Flag and performed Co-IP against their endogenous counterparts. Exogenous cGAS bound endogenous JOSD2 (Figure 3C), and vice versa (Figure 3D). Interestingly, endogenous Co-IP in untransfected cells confirmed the physiological interaction between cGAS and JOSD2 (Figure 3E). Additionally, we enriched the prokaryotically purified GST-JOSD2 protein using glutathione-tagged agarose beads and then incubated it with HEK-293T cell lysate for in vitro co-immunoprecipitation. The results, as shown in Figure 3F, indicate that JOSD2 binds to cGAS in the in vitro system.
To map the binding interface, we generated Flag-tagged cGAS truncation mutants spanning key domains: the N-terminal domain (NTD, 1–160), the nucleotidyltransferase domain (NTase, 161–330), and the C-terminal domain (CTD, 331−522). NTD plays a critical role in regulating the liquid‒liquid phase separation of cGAS (Figure 3G).37 NTase is the core catalytic domain, which stabilizes the cGAS–DNA complex. The CTD contains a zinc-finger motif, which participates in the regulation of cGAS function.38 Co-IP experiments showed that full-length cGAS and NTD-containing truncations (1−160, 1−330) robustly bound JOSD2, whereas constructs lacking NTD (161−330, 161−522, 331−522) failed to interact (Figure 3H), identifying the NTD as the critical binding region.
Combined with the above experimental results, stimulation with cGAMP induced comparable phosphorylation (p-TBK1/p-STING) and downstream cytokine (Ifnβ, Isg15) responses in WT and JOSD2 KO BMDMs (Figure 2F and Figure S2E), confirming that JOSD2 specifically targets cGAS itself rather than downstream signaling components.

JOSD2 removes K27-linked ubiquitin chains from cGAS.
Ubiquitin is a 76-amino acid protein containing an N-terminal methionine and seven lysine residues (K6, K11, K27, K29, K33, K48, K63). These lysines serve as linkage sites for the formation of distinct ubiquitin chain types that regulate substrate protein stability, cellular localization, and functionality.39 To investigate this, we constructed HA-tagged ubiquitin mutants retaining single lysine residues and exogenously co-expressed them with cGAS-Flag in HEK-293T cells. K27-linked ubiquitin chains were enriched predominantly on cGAS under these conditions (Figure 4A). We then expressed wild-type JOSD2 in this system and observed the specific removal of K27-linked polyubiquitin chains from cGAS, whereas the catalytically inactive mutant JOSD2-C24A exhibited no such activity (Figure 4B). To validate enzymatic activity, we performed an in vitro cGAS deubiquitination assay. Strikingly, purified JOSD2 efficiently removed K27-linked ubiquitin chains from cGAS in vitro (Figure 4C).
RNF185 is currently the sole reported E3 ubiquitin ligase capable of mediating K27-linked ubiquitination of cGAS.40 We thus hypothesized that JOSD2 might act as a deubiquitinase to counteract RNF185-mediated K27 ubiquitination of cGAS.
RNF185 overexpression robustly intensified K27-linked ubiquitin signals on cGAS, while co-expression of JOSD2 coexpression reversed this signal accumulation (Figure 4D). To further validate whether JOSD2 targets these specific ubiquitination sites, we generated a cGAS-2KR mutant (K173R/K384R) via molecular cloning to abrogate ubiquitin modification. These results confirm that mutation of these sites impairs cGAMP synthesis (Figure 4E), which is consistent with previous literature reports. The in vitro deubiquitination assay confirmed that JOSD2 selectively removes RNF185-deposited K27 ubiquitin chains from wild-type cGAS but not from the 2KR mutant (Figure 4F).
In summary, we identified JOSD2 as a cGAS-specific deubiquitinating enzyme that antagonizes RNF185-mediated K27-linked ubiquitination at residues K173 and K384, thereby suppressing cGAS enzymatic activity (Figure 4G).

JOSD2 knockout enhances the anti-tumor efficacy of anti-PD-L1 antibody
Recent studies have elucidated the association between the cGAS-STING pathway and macrophage polarization: STING activation-induced IFN-β upregulates M1 markers via STAT1, while suppressing M2 markers.41 M1 polarization relies on glycolysis, and STING activation promotes glycolytic enzyme expression through HIF-1α, driving the transition from M2 to M1 phenotypes.42 To investigate whether JOSD2 modulates macrophage polarization by regulating cGAS enzymatic activity, we stimulated BMDMs to induce M2 polarization. Flow cytometry revealed that IL-4-stimulated JOSD2 KO BMDMs exhibited a significantly lower proportion of F4/80+CD206+ cells compared to WT controls (Figure 5A). Further analysis using Western blotting demonstrated robust early phosphorylation of STAT6 and rapid induction of CD206 and ARG1 protein expression in WT BMDMs following IL-4 stimulation, whereas JOSD2 KO cells exhibited attenuated responsiveness to M2 polarization (Figure 5B). Prolonged IL-4 treatment and subsequent qRT-PCR analysis confirmed that transcript levels of Ym1 and Arg1 remained significantly higher in the WT BMDMs than in the KO group (P < 0.001, Figure 5C).
JOSD2 knockout in macrophages activates the cGAS-STING pathway and suppressed M2 polarization, reshaping the inflammatory immune landscape of the tumor microenvironment. cGAS-STING agonists activate the immune system, while anti-PD-L1 antibodies restore the cytotoxic activity of suppressed immune cells. The combination of these two agents is considered a highly promising therapeutic strategy. We established MC38 subcutaneous tumor models in WT and JOSD2 KO mice. These mice were randomized into two groups and treated with either 15 mg/kg anti-PD-L1 or isotype IgG2b control antibodies every 3 d for four cycles. The results demonstrated that JOSD2 KO mice exhibited significantly smaller mean tumor volumes and weights than WT controls. Furthermore, anti-PD-L1 therapy exhibited superior antitumor efficacy in JOSD2 KO mice (Figure 5D and 5E), with no significant impact on body weight changes in the treated mice (Figure S3A). ELISA quantification of IFNβ and TNF-α levels in tumor lysates revealed significantly higher cytokine concentrations in JOSD2 KO mice after anti-PD-L1 treatment compared to WT controls (Figure 5F and 5G). These findings indicate that JOSD2 deficiency not only amplifies cGAS-STING pathway activation and T cell-mediated cytotoxicity but also synergizes with anti-PD-L1 therapy to further potentiate cGAS-driven antitumor immunity. Notably, immunohistochemical staining of tumor samples revealed a modest increase in the expression of CD8 and the dendritic cell (DC) marker CD209 in JOSD2 knockout tumors compared to wild-type tumors, particularly following combination therapy with an anti-PD-L1 antibody therapy (Figure S3B). These findings suggest that DC may also play a role in JOSD2-regulated tumor immunity. This mechanism provides a novel strategy to increase the therapeutic efficacy of immune checkpoint inhibitors.

JOSD2 inhibitor suppresses CRC growth by enhancing the enzymatic activity of cGAS.
To further validate the molecular mechanism and its tumor-suppressive effects, we conducted pharmacological targeting studies using HY041004 – a JOSD2 enzymatic inhibitor previously identified by our group.21,43 We first assessed the impact of HY041004 on JOSD2-mediated molecular mechanisms at the cellular level. Co-IP assays demonstrated that HY041004 treatment rescued JOSD2-mediated cleavage of K27-linked ubiquitin chains on cGAS (Figure 6A). Using BMDMs models, ELISA quantification revealed that HY041004-treated BMDMs produced significantly higher cGAMP levels compared to DMSO controls following HT-DNA stimulation (Figure 6B). Western blot analysis confirmed the pronounced accumulation of p-TBK1 and p-STING in HY041004/HT-DNA-treated cells relative to that in the DMSO/HT-DNA group. Notably, this effect was abolished in JOSD2 KO BMDMs, confirming target specificity (Figure 6C and Figure S4A). Concurrently, qRT-PCR analysis showed markedly elevated transcript levels of Ifnβ and Il6 in HY041004-treated cells compared to DMSO controls (Figure 6D). We next investigated the impact of HY041004 on macrophage M2 polarization. Wild-type BMDMs were divided into two groups and treated with either DMSO or HY041004 for 6 h, followed by IL-4 stimulation to induce polarization. Western blot analysis revealed robust phosphorylation of STAT6 at 2 h in the DMSO group, whereas HY041004 treatment resulted in delayed STAT6 activation (Figure 6E). Furthermore, both the protein and transcript levels of ARG1 were significantly higher in the DMSO group compared to the HY041004-treated group after 6 h of stimulation (Figure 6F).
Next, we evaluated the anti-tumor potential of HY041004 in vivo. Wild-type mice bearing MC38 subcutaneous tumors were randomized into two groups: one group received 0.5% CMC-Na as a control, while the other was treated with 200 mg/kg HY041004 via oral gavage twice daily. At the experimental endpoint, HY041004-treated mice exhibited significantly smaller mean tumor volumes and weights compared to the control group (Figure 6G and 6H), with no significant impact on body weight changes in the treated mice (Figure S4B). Western blot analysis of tumor tissues revealed markedly elevated p-TBK1 and p-STING in HY041004-treated tumors (Figure 6I). Similarly, ELISA quantification demonstrated significantly increased intratumoral IFNβ and TNFα levels in the HY041004 group, indicating a pro-inflammatory immune microenvironment (Figures 6J and S4C). Paraffin-embedded tumor sections subjected to dual immunofluorescence staining showed abundant yellow-labeled M2-polarized macrophages in control tumors, whereas HY041004-treated tumors presented markedly reduced yellow-stained areas, reflecting diminished M2 macrophage infiltration (Figure 6K).
In summary, through both in vitro and in vivo studies, we not only demonstrated the efficacy of HY041004 as a JOSD2 enzymatic inhibitor but also revealed that it suppressed M2 macrophage polarization by enhancing cGAS activity, thereby exerting anti-colorectal cancer effects.

Discussion

Discussion
Conventional therapies offer limited benefits for CRC patients.44-46 Immunotherapy has shown transformative efficacy only in CRC patients with MSI-H.47 Therefore, identifying key regulators of immune evasion remains critical for increasing immunotherapy sensitivity and increasing therapeutic eligibility. This study delineates a novel biological role of the deubiquitinating enzyme JOSD2 in regulating CRC immune evasion. We demonstrated that the inhibition of JOSD2 exerts antitumor effects by activating the cGAS-STING pathway and suppressing M2 macrophage polarization.
Previous studies demonstrated that JOSD2, which is highly expressed in KRAS-mutant CRC, stabilizes KRAS protein levels. Concurrently, mutant KRAS suppresses E3 ubiquitin ligase-mediated ubiquitination and degradation of JOSD2 via a positive feedback loop, thereby driving malignant proliferation in CRC.43 Moreover, Cui et al. reported that in patients with very early-onset inflammatory bowel disease (VEO-IBD) harboring the novel NLRP3 R779C mutation, the NLRP3-R779C variant exhibited reduced ubiquitination levels compared to wild-type NLRP3 (NLRP3-WT) due to deubiquitination by BRCC3 and JOSD2 in macrophages.48 This enhances inflammasome activity, exacerbating pro-inflammatory cytokine release and pyroptosis. Our database analyses further revealed that JOSD2 expression is significantly correlated with poor prognosis in MSI-H CRC patients and is associated with multiple immune regulators. Collectively, these findings suggest that JOSD2 has potential as a novel biomarker and therapeutic target, particularly from both the tumor cell-intrinsic and immune microenvironment perspectives.
JOSD2 differentially affects CRC subtypes, potentially through variations in tumor-derived immunogenicity and the DNA damage response. MSI-H tumors are characterized by high levels of cytosolic DNA, leading to constitutive cGAS-STING activation. In this context, the role of JOSD2 in suppressing this pathway is a critical determinant of immune evasion and patient prognosis. Conversely, in MSS tumors with lower innate immunogenicity, the immunomodulatory function of JOSD2 is less pivotal. Additionally, our previous research evidence links JOSD2 to DNA damage response (DDR) pathways, and its expression correlates with DDR activation.49 Since MSI-H tumors inherently harbor greater genomic instability, their increased dependency on JOSD2 for DDR could further explain why its high expression is selectively associated with poor outcomes in this subtype.50
Further molecular mechanism studies revealed that JOSD2 interacts with cGAS and cleaves ubiquitin chains in an enzymatic activity-dependent manner without affecting cGAS protein stability but critically regulates its enzymatic function. Notably, regarding cGAS ubiquitination, previous studies have identified the roles of several E3 ligases:51 TRIM56 mediates monoubiquitination at K355, promoting cGAS dimerization and DNA-binding activity to increase cGAMP production during antiviral responses;52 TRIM41 similarly enhances cGAS activity by ubiquitination but the specific site remains unknown;53 RNF178 inhibits cGAS activity by linking K63-linked polyubiquitin chains at K411;54 RNF185 conversely promotes cGAS enzymatic activity by mediating K27-linked polyubiquitination at K173 and K384.40 These findings underscore that ubiquitin chain length, linkage type, and modification sites differentially regulate cGAS functionality.
DUBs such as UAF1-USP1, USP14 (targeting cGAS-K414), USP27X, and USP29 (targeting cGAS-K271) cleave K48-linked polyubiquitin chains to prevent cGAS degradation, amplifying downstream signaling via protein stabilization.55-58 Notably, Jiang et al. identified OTUD3 as a cGAS-interacting DUB via proximity labeling. OTUD3 enhances cGAS protein stability by cleaving both K27- and K48-linked ubiquitin chains and increasing cGAS-DNA interactions through its direct DNA-binding capacity, thereby increasing enzymatic activity. However, in vitro reversal experiments have shown that OTUD3 cannot cleave RNF185-deposited K27 chains but targets K27 chains from an unidentified E3 ligase.59 Given the heterogeneity between tumor and immune cells, the enzymes regulating cGAS ubiquitination differ across cell types. The current understanding of the E3 ligases responsible for the K48-linked polyubiquitination of cGAS and DUBs that regulate cGAS enzymatic activity remains incomplete, hindering the establishment of a comprehensive regulatory network.
DUBs are highly attractive drug targets because of their well-defined cysteine catalytic pockets, which enable precise small-molecule intervention. Additionally, transcription factors or proteins that are challenging to target directly can be indirectly modulated by regulating their associated DUBs.60,61 Currently, the most advanced small-molecule candidate is KSQ-4279, a USP1 inhibitor. In combination with the PARP inhibitor olaparib, KSQ-4279 has demonstrated significant efficacy against BRCA1/2-mutated breast cancer, and a phase I clinical trial for advanced solid tumors is underway.62 Using the JOSD2 inhibitor HY041004 in an MC38 colorectal cancer subcutaneous syngeneic model, we validated that this inhibitor enhances cGAS enzymatic activity, suppresses M2 macrophage polarization, and reprograms the anti-tumor immune microenvironment. However, despite the proven efficacy of HY041004 as a lead compound, its low bioavailability limits its in vivo therapeutic performance, necessitating chemical optimization to improve its pharmacokinetics. Future efforts will focus on redesigning JOSD2 inhibitors to overcome bioavailability limitations and evaluating their synergy with immune checkpoint inhibitors, aiming to establish a novel and potent therapeutic strategy for colorectal cancer.
While this study delineates a novel role for JOSD2 in promoting immune evasion, several limitations should be acknowledged. First, our investigation primarily utilized the MC38 syngeneic model, which is derived from a highly immunogenic and mutated cell line. Although this model was instrumental in revealing the immunomodulatory function of JOSD2, it may not fully recapitulate the immunological characteristics of the more common MSS colorectal cancer subtypes. Second, the mechanistic insights into macrophage polarization were largely derived from in vitro studies using BMDMs. Although BMDMs provide a controlled system, they may not entirely mirror the complex phenotypic and functional states of tumor-associated macrophages in vivo. A limitation of this study is the lack of direct assessment of NK cells, CD8⁺ T cells infiltration and PD-1 expression within the tumor microenvironment to evaluate immune remodeling. Additionally, long-term follow-up or tumor re-challenge experiments are required to evaluate long-term durability and immune memory, thereby demonstrating the clinical potential of targeting JOSD2. Finally, our study focused on macrophages as key mediators of the JOSD2-cGAS axis; however, other innate immune cells, particularly conventional type 1 dendritic cells (cDC1), which are crucial for T cell priming and also depend on the cGAS-STING pathway,63 were not examined. The potential contribution of JOSD2 regulation in other immune cells to the observed anti-tumor immunity remains an important area for future investigation.
In conclusion, our study identified JOSD2 as a critical regulator of immune evasion in colorectal cancer. Mechanistically, JOSD2 specifically reverses RNF185-mediated K27-linked ubiquitination at residues K173/K384 of cGAS, thereby suppressing its enzymatic activity and driving M2 macrophage polarization. This discovery reveals a novel ubiquitination-dependent mechanism of immune evasion. JOSD2 in macrophages bridges a critical gap between innate immune sensing and adaptive anti-tumor immunity. Importantly, targeting JOSD2 pharmacologically reverses this suppression, and its synergy with anti-PD-L1 therapy offers a promising strategy to potentiate cancer immunotherapy, particularly for patients with currently limited treatment options.

Supplementary Material

Supplementary Material

Supplementary material
Supplementary_Figures.

출처: PubMed Central (JATS). 라이선스는 원 publisher 정책을 따릅니다 — 인용 시 원문을 표기해 주세요.

🏷️ 같은 키워드 · 무료전문 — 이 논문 MeSH/keyword 기반

🟢 PMC 전문 열기