CTMP-REV7 axis modulates MHC-I antigen presentation via CDK1-AKT crosstalk in MSI-L/MSS colorectal cancer Authors: Cheng Gu, Rong He, Xiaolan Yin, Jie Wang, Jing Yang, Jie Gao, Rongjia Xie, Yangyang Geng, Jingyi Lu, Zhui Ke, Zhuqing Liu, Qing Xu Published: 2026 ABSTRACT AbstractPurposeMicrosatellite instability-low/microsatellite stable (MSI-L/MSS) colorectal cancer (CRC) exhibits limited responsiveness to immune checkpoint inhibitors due to low tumor mutational burden and insufficient neoantigen generation. Carboxyl-terminal modulator protein (CTMP) is a potential regulatory molecule that may reshape tumor immunogenicity and metabolic patterns, yet its mechanisms in MSI-L/MSS CRC remain poorly understood.MethodsWe systematically investigated CTMP function by integrating clinical sample analysis (GEO database, n = 71 paired tissue microarray, and pretreatment specimens from 11 refractory patients receiving anti-PD-1 plus fruquintinib), in vitro experiments (MSS CRC cells with CTMP knockdown/overexpression, assessed for proliferation, mitochondrial function, lipid metabolism, AKT/PD-L1 signaling, and MHC-I/HLA-B expression), proteomic screening (immunoprecipitation-mass spectrometry to identify CTMP-interacting proteins), structural validation (co-immunoprecipitation, AlphaFold Multimer, GST pulldown with truncation/point mutations), and in vivo studies (immunocompetent BALB/c mice with CT26 syngeneic tumors).ResultsHigh CTMP expression and low HLA-B expression were significantly associated with shortened overall survival in patients with MSI-L/MSS colon adenocarcinoma, as well as with progressive disease in an anti-PD-1/fruquintinib resistance cohort. Mechanistically, CTMP orchestrated an immunosuppressive tumor microenvironment by sustaining AKT/PD L1 signaling, suppressing HLA B expression, and promoting fatty acid metabolism. Proteomic screening identified a previously unreported high-affinity interaction between CTMP and REV7, which facilitated CDK1 mediated G2/M progression and further repressed HLA B expression. CTMP knockdown or REV7 overexpression diminished CDK1 activity, induced G2/M phase arrest, and restored HLA B expression in a cGAS signaling-relevant manner. Furthermore, disruption of the CTMP-REV7 signaling axis modulated CDK1 and AKT activity under IFN-γ-driven immune stress, affecting MHC-I expression. In immunocompetent murine models, CTMP knockdown or REV7 overexpression suppressed tumor growth, upregulated HLA B expression, and enhanced intratumoral CD8⁺ T cell infiltration. Moreover, REV7 overexpression synergized with IFN-γ to promote the release of chemokines (CXCL9, CXCL10, TNF α), leading to marked tumor regression in CTMP-KD tumors.ConclusionsWe delineate that CTMP drives immune evasion in MSI-L/MSS CRC through a novel CTMP–REV7–CDK1 axis that couples metabolic reprogramming, cell cycle control, and MHC-I antigen presentation, with AKT signaling serving as a critical crosstalk node. These findings implicate this integrated axis as a therapeutic target to restore tumor immunogenicity and overcome immunotherapy resistance in immunologically “cold” MSI-L/MSS colorectal cancer.Graphic Abstract Supplementary InformationThe online version contains supplementary material available at 10.1007/s13402-026-01222-5.Keywords: CTMP, REV7, CDK1, AKT, MHC-I Introduction Immunotherapy has revolutionized outcomes for microsatellite instability-high (MSI-H) colorectal cancer (CRC) [1], yet effective treatment options remain elusive for the majority of patients with microsatellite instability-low / microsatellite stable (MSI-L/MSS) tumors. The inherent resistance of these tumors stems from an immunologically quiescent microenvironment, characterized by low tumor mutational burden (TMB) limiting neoantigen diversity [2, 3], impaired antigen presentation, and diminished cytotoxic T lymphocyte (CTL) infiltration [4], rendering them largely unresponsive to immune checkpoint inhibitor (ICI) monotherapy. Overcoming this challenge necessitates combinatorial strategies that co-target tumor-intrinsic immunosuppressive pathways alongside core oncogenic drivers. While recent clinical advances, such as the REGONIVO (regorafenib/nivolumab) [5] and FRESCO-2 (fruquintinib) [6] trials, demonstrate the potential of anti-angiogenic and ICI combinations, and alternative approaches targeting MEK to enhance MHC-I [7], novel immune checkpoints (LAG-3 [8], TIGIT [9]), or adoptive cell therapy [10] are actively explored, the discovery of novel targets capable of intrinsically reprogramming the immunosuppressive landscape of MSI-L/MSS CRC remains a critical priority. Within the context of CRC, reprogrammed lipid metabolism constitutes a critical hallmark that extends beyond its traditional functions in bioenergetics and biosynthesis [11, 12] to actively contribute to an immunosuppressive tumor microenvironment. This dual role is illustrated by the mitochondrial FAO enzyme CPT1A: its overexpression promotes tumor progression and metastasis [13, 14], while its inhibition can induce immunogenic cell death or enhance susceptibility to cytotoxic T lymphocytes [15, 16]. Notably, regulatory elements within both FAO and fatty acid anabolism—such as thioesterases, which limit the accumulation of toxic acyl-CoA species [17–19]—represent underexplored targets with increasing oncogenic significance [20–22]. Among these, the HotDog-fold thioesterase CTMP (carboxyl-terminal modulator protein) [23] has emerged as an immune-metabolic node. Functionally, CTMP participates in lipid metabolism regulation via its C-terminal HotDog fold domain with acyl-CoA thioesterase activity [17, 23, 24], while its N-terminal mitochondrial localization signal targets it specifically to mitochondria [25–27]. Concurrently, CTMP modulates AKT signaling through N-terminal interaction with the AKT C-terminal domain [26, 27]—a known negative regulator of MHC-I antigen presentation [28, 29]. This direct interaction suggests a potential mechanistic connection between lipid metabolism and immunogenicity, although the specific influence of CTMP on AKT phosphorylation (whether promoting or suppressive) remains unresolved [26, 30–32]. Moreover, CTMP has been implicated in the dysregulation of innate immunity under metabolic stress through Toll-like receptor 2 signaling [33, 34], underscoring its role as a context-dependent immune-metabolic regulator. Central to this work is the identification of a novel functional interaction between CTMP and REV7 (MAD2L2), a mitotic checkpoint component preventing early mitosis exit through inhibition of the APC/C activator Cdh1 [35]. Paradoxically, in MSI-L/MSS colorectal cancer—characterized by constitutively high chromosomal instability (CIN) [36]—REV7 overexpression further exacerbates mitotic defects: anaphase bridges are readily detectable in colon cancer cells with elevated REV7 expression [37, 38], a phenotype recapitulated by overexpression of its paralog MAD2 [39] and similarly provoked by enforced expression of other core mitotic checkpoint components (e.g., MAD3, BUB1, YY2/BUB3) [40, 41]. Notably, YY2/BUB3 overexpression induces persistent spindle assembly checkpoint hyperactivation, leading to substantial CIN and CRC cell death [41], supporting the potential therapeutic strategy of exploiting mitotic checkpoint overexpression in CRC treatment. Similarly, mitotic aberrations induced by pharmacological inhibition of Aurora A [42] or MPS1 [43] can upregulate MHC-I and engage cGAS-mediated immune modulation [44]. Thus, it is critical to determine whether REV7 can exploit intrinsic CIN to activate this pathway. Of direct relevance, REV7 directly suppresses CDK1 [45], a kinase whose elevated activity promotes immune evasion [46], while CDK1 inhibition reactivates cGAS signaling [47]. These findings collectively position CDK1 as the critical mechanistic link through which REV7 may convert CIN-driven mitotic stress into antitumor immune regulation. In this study, we uncover a CTMP–REV7–CDK1 axis that connects cell cycle regulation with antitumor immunity, and propose that targeting this axis represents a potential strategy to overcome immunotherapy resistance in MSI-L/MSS colorectal cancer. Materials and methods Bioinformatics analysis Survival analysis was conducted using the KM Plotter online tool (https://kmplot.com/analysis/index.php? p=service%26cancer=colon) across the Affymetrix probe set to assess the prognostic value of target genes in colon cancer [48]. The following parameters were applied: Stratification method: Median expression; Cutoff selection: Auto-selected optimal cutoff; Survival endpoint: Overall survival (OS); Subtype restriction: MSI (stable and low). Statistical significance was determined by the log-rank test, with a hazard ratio (HR) and 95% confidence interval (CI) calculated for each comparison. Pan-cancer and immune correlation analysis of CTMP and REV7 using TIMER3.0 (https://compbio.cn/timer3/). Clinical tissue samples Two cohorts of MSI-L/MSS CRC samples were utilized: Commercial tissue microarray cohort: Seventy-one MSI-L/MSS CRC specimens with paired normal colonic mucosa were commercially obtained from Aifang Biotec (Hunan, China). From this cohort, five paired samples were selected based on CTMP expression levels for Western blot quantification. Commercial tissue microarray cohort: Seventy-one MSI-L/MSS CRC specimens with paired normal colonic mucosa were commercially obtained from Aifang Biotec (Hunan, China). From this cohort, five paired samples were selected based on CTMP expression levels for Western blot quantification. Immunotherapy-treated cohort: This cohort consisted of 11 MSI-L/MSS CRC patients treated with combined immunotherapy (anti-PD-1 agents: pembrolizumab or nivolumab) and fruquintinib. Formalin-fixed paraffin-embedded tumor tissues were collected postoperatively at Shanghai Tenth People’s Hospital between January 2018 and November 2024. Based on RECIST v1.1, patients were stratified into partial responders (PR, n = 6) and those with progressive disease (PD, n = 5). Written informed consent was obtained from all participants, and the study protocol was approved by the Ethics Committee of Shanghai Tenth People’s Hospital (Approval No. 25KN307). Immunotherapy-treated cohort: This cohort consisted of 11 MSI-L/MSS CRC patients treated with combined immunotherapy (anti-PD-1 agents: pembrolizumab or nivolumab) and fruquintinib. Formalin-fixed paraffin-embedded tumor tissues were collected postoperatively at Shanghai Tenth People’s Hospital between January 2018 and November 2024. Based on RECIST v1.1, patients were stratified into partial responders (PR, n = 6) and those with progressive disease (PD, n = 5). Written informed consent was obtained from all participants, and the study protocol was approved by the Ethics Committee of Shanghai Tenth People’s Hospital (Approval No. 25KN307). Cell lines and cell culture SW620 (human, MSS, P53 mutant), SW480 (human, MSS, P53 mutant), CT26 (mouse, MSS, P53 mutant), HCT116 (human, MSI-H, P53 wild-type) and tool cell line -HEK293T maintained by our laboratory, were authenticated by Short Tandem Repeat (STR) assay. Cells were cultured in DMEM (Gibco #11965092) supplemented with 10% fetal bovine serum (Dakewe Co., Shenzhen, China) at 37 °C in a humidified incubator with 5% CO₂. CCK-8 assay Cells (2.0 × 10³ per well) were seeded into 96-well plates with 6 replicates. At the indicated time points, the reaction medium, containing 90 µl DMEM and 10 µl CCK-8 reagent, was added and incubated for 2 h. The absorbance at 450 nm was measured using a SpectraMax iD5 microplate reader (Molecular Devices, USA). Colony formation assay Cells were seeded in 6-well plates at 1000 cells/well and cultured for 10 days. The resulting colonies were fixed with 4% paraformaldehyde, stained with 0.1% crystal violet. Colonies containing > 50 cells were considered valid. Quantitative real-time PCR Total RNA was extracted from cultured cells using TRIzol reagent (Vazyme Biotech #R401-01) in accordance with the manufacturer’s instructions, followed by treatment with RNase-free DNase to remove genomic DNA contamination. Reverse transcription was performed on 1000 ng of purified total RNA using the HiScript II One-Step RT-PCR Kit (P611-01, Vazyme). The resulting cDNA was then amplified by real-time PCR with SYBR Green Master Mix on QuantStudio Dx system. Target gene expression levels were normalized to the internal reference gene GAPDH, and relative quantification was calculated employing the comparative threshold cycle method (2 − ΔΔCt). The qPCR primer sequences are listed in the supplementary materials. Acetyl coenzyme A and lactate detection Acetyl-CoA levels were quantified using a commercial ELISA kit (Sangon Biotech, #D751001). Briefly, 10⁶ cells were washed with ice-cold PBS and lysed in PBS containing 1 mM PMSF via three freeze-thaw cycles, followed by centrifugation at 1,500 ×g for 10 min. The supernatant was assayed per the manufacturer’s instructions: samples/standards (100 µl/well) were incubated (37 °C, 90 min), followed by sequential incubation with biotinylated antibody (37 °C, 60 min), HRP-streptavidin (37 °C, 30 min), and substrate (37 °C, 15 min in dark). The reaction was stopped, and absorbance was measured at 450 nm using a SpectraMax iD5 reader. Lactate levels were measured using a commercial assay kit (Sangon Biotech, #D799099). Harvested cells (1 × 10⁶/group) were processed according to the manufacturer’s protocol, and absorbance was read at 570 nm (SpectraMax iD5). Protein recombinant expression and purification The recombinant E. coli BL21(DE3) strain harboring the GST tagged plasmids (Figure S1B) was grown in LB medium with 100 µg/mL ampicillin at 37 °C. Protein expression was induced with 0.1 mM IPTG at an OD₆₀₀ of 0.8–0.9, followed by incubation at 15 °C for 12 h. Cells were harvested (4,000 g, 10 min), resuspended in binding buffer (20 mM Tris-HCl, 500 mM NaCl, 1 mM DTT, pH 8.0), and treated with Benzonase (15 µl per 50 mL culture, 30 min, on ice). After sonication and centrifugation (12,000 ×g, 20 min, 4 °C), the supernatant was mixed with 1 M Tris-HCl (100 µl). For purification, the supernatant was incubated with GSH magnetic beads (30 min, RT, gentle agitation). and the target protein was eluted with 300 µl elution buffer (1 min, RT). The eluate was analyzed by western blot (Figure S1C). Immunofluorescence Cells or tissue chips were fixed with 4% paraformaldehyde (20 min, RT), permeabilized with 0.5% Triton X-100 (5 min) and blocked with 5% BSA in 1× PBST (45 min, RT). Samples were incubated with primary antibody overnight at 4 °C, washed three times with PBS, and incubated with fluorophore-conjugated secondary antibody (60 min, RT, in the dark). Nuclei were counterstained with DAPI (2 min) before mounting. Imaging was performed using a Nikon TI2-E + A1 R laser scanning confocal microscope at excitation wavelengths of 488 nm (green) and 561 nm (red). Colocalization analysis was performed using Fiji (ImageJ) with the JACoP plugin. Pearson’s coefficient and Manders’ overlap coefficients (M1, M2) were calculated after background subtraction using Costes’ automatic thresholding. shRNA transfection RNA-targeting constructs were synthesized by Generay Biotech (Shanghai, China). Lentiviral packaging was performed in 293T cells co-transfected with pLKO.1-shRNA, psPAX2, and pMD2.G plasmids using PEI transfection reagent. Viral supernatants were collected and used to infect target cells at 30% confluence in the presence of 10 µg/mL polybrene (Beyotime, China). Stable transfectants were selected with 5 µg/mL puromycin (Beyotime, China) for 7 days, with knockdown efficiency confirmed by Western blot. The shRNA sequences used were: CTMP-KD: TRCN0000048718 (Human), TRCN0000252302 (Mouse); REV7-KD: TRCN0000006570 (Human), TRCN0000012844 (Mouse). Plasmid construction All plasmid constructs, provided by Generay Biotech (Shanghai, China), comprised the following: Flag-CTMP; CTMP truncation mutants; Flag-tagged CTMP point mutants (D89A, R75A, R8A, R3A, R3A/R8A); and interaction partners (HA-REV7, HA-CDK1, HA-AKT, Flag-REV7) (Figure S1A). All constructs were validated by full-length sequencing of the inserted DNA fragment. Seahorse assay Mitochondrial function was assessed using the Seahorse XF Mito Stress Test Kit (Agilent Technologies#103015-100, USA) to measure the oxygen consumption rate (OCR). The assay medium was prepared by supplementing 97 mL of XF DMEM medium with 1 mL of 200 mM glutamine, 1 mL of 100 mM sodium pyruvate, and 1 mL of 1.0 M glucose. Oligomycin (20 µl, 1 µM; ATP synthase inhibitor), FCCP (22 µl, 1 µM; mitochondrial uncoupler), and Rotenone/Antimycin A (25 µl, 1 µM; ETC inhibitors) were sequentially injected at the indicated time points. Briefly, 2 × 10⁴ cells were seeded per well in XF plates, equilibrated in a CO₂-free incubator at 37 °C for 1 h, and analyzed per the manufacturer’s protocol: baseline (4 cycles) → compound injections (4 measurements per phase). Key parameters included basal respiration, ATP production, and maximal respiratory capacity. Immunohistochemistry Tissue sections were deparaffinized, rehydrated, and subjected to antigen retrieval in citrate buffer (pH 6.0). Endogenous peroxidase was blocked with 3% H₂O₂, and nonspecific binding was blocked with 5% BSA. Sections were incubated with primary antibodies against pAKT-Ser473 (CST #4060), phospho-Histone H1 (Sigma-Aldrich #05-1324), PD-L1 (Abclonal #A19135) and CD8 (CST #98941), followed by HRP-conjugated secondary antibody. Signals were visualized with DAB substrate, and sections were counterstained with hematoxylin. Staining intensity was scored independently by two blinded pathologists on a 0–3 scale (0.2 increments). The average score from at least five fields per sample was used for analysis. Intergroup differences were assessed using the Mann-Whitney U test. Immunoprecipitation Cells from a 10-cm dish at 70–80% confluence were lysed in 600 µl IP lysis buffer supplemented with 0.1 mM PMSF (15 min on ice), followed by centrifugation at 13,000 ×g for 5 min at 4 °C. The supernatant was incubated overnight at 4 °C with gentle rotation using 5 µg/mL specific antibody and 40 µl Protein A/G agarose beads (Beyotime #P2055). Beads were pelleted (5,000 ×g, 1 min) and washed six times with 1 mL IP lysis buffer. Immunoprecipitated complexes were analyzed by mass spectrometry or Western blot. Western blot Cell or tissue lysates were prepared in RIPA lysis buffer (Beyotime #P0013) with 1 mM PMSF (Beyotime #ST505). Protein concentration was determined using a BCA assay kit (Beyotime #P0011). Samples were separated by 10% SDS-PAGE (Beyotime #P0012A) at 120 V for 90 min and transferred to PVDF membranes (Beyotime #FFP20) using a Trans-Blot Turbo system (400 mA, 45 min). Membranes were blocked with 5% non-fat milk in PBST (1 h, RT), incubated with primary antibodies (1:1000, 2 h, RT) and corresponding HRP-conjugated secondary antibodies (Beyotime #A0350/A0352, 1:1000, 1 h, RT). Signals were detected using a Tanon 4600SF imaging system (Tanon, Shanghai, China). Antibodies for western blot Anti-Mad2L2/REV7 (Abcam #ab180579), Anti-HLA-B (Abcam #ab193415), Anti-Akt (CST #2938), Anti-Phospho-Akt(Ser473) (CST #4060), Anti-PD-L1 (Abclonal #A19135), Anti-GAPDH (Beyotime #AF0006), Anti-CDK1 (Proteintech #19532-1-AP), Anti-CTMP (Santa Cruz #sc-390353), Anti-FLAG M2 (Sigma-Aldrich #F1804), Anti-HA (Sigma-Aldrich #H9658), Anti-phospho-Histone H1 (Sigma-Aldrich #05-1324), Anti -Histone H1 (Sigma-Aldrich #05-457),Anti-Hexokinase 2 (Proteintech #22029-1-AP), Anti-LDHA (Proteintech #19987-1-AP), Anti-P21 (Proteintech #28248-1-AP), Anti-TBK1 (Proteintech # 67211-1-Ig), Anti-p-TBK1 (Abcam#ab109272). Anti-IRF3 (Abcam#ab25950), Anti- p-IRF3(S396) (CST, 4947 S). Reagents Ro-3306 (Sigma-Aldrich #SML0569); Human IFN-γ (Sigma-Aldrich #I17001); Murine IFN-γ (Beyotime # P6137); SC-79 (Beyotime #SF2730). Prediction of protein interaction sites The structures of CTMP and REV7 were predicted using the multimer model in AlphaFold v2.3.2 [49, 50] with default parameters. Predicted oligomeric structures were visualized using PyMOL (v2.5.0). Binding affinities and dissociation constants were calculated using PRODIGY (v2.0), and intermolecular interactions were analyzed with Arpeggio. Kinase activity assay Purified GST-REV7, GST-CTMP and GST-NC proteins were serially diluted in kinase assay buffer (50 mM Tris pH 7.5, 10 mM MgCl₂, 0.5 mg/mL BSA) in a 96-well plate. The kinase reaction mixture (25 µl final volume) contained diluted GST-protein, 0.2 mg/mL Histone H1 (Sigma-Aldrich #H4524), 100 µM ATP, and 10 ng CDK1-Cyclin B1 kinase (Sigma-Aldrich #14450D). The plate was incubated at 37 °C for 30 min. Subsequently, 25 µl of ADP-Glo™ Reagent (Promega #V6930) was added to each well to terminate the reaction and deplete residual ATP (30 min incubation at RT). Then, 50 µl of Kinase Detection Reagent was added, and the plate was incubated for 60 min at RT to generate a luminescent signal. Luminescence was recorded using a SpectraMax iD5 reader (Molecular Devices, USA). In vitro binding assay Recombinant GST-tagged protein (Figure S1B-C) was immobilized on glutathione-Sepharose beads (4 °C, 45 min, RT). The protein-bound beads were then incubated with co-immunoprecipitation lysate (1 h, RT). After centrifugation and extensive wash, bound proteins were eluted by boiling in SDS-PAGE sample buffer, detected by immunoblotting with specific antibodies. Flow cytometric analysis of cell cycle Cells were fixed in 1 mL of ice-cold 70% ethanol (4 °C, 30 min), centrifuged at 1,000 ×g for 3 min at 4 °C, and washed with 1 mL ice-cold PBS. Cell pellets were stained with 0.5 mL propidium iodide (PI) solution and incubated in the dark at 37 °C for 30 min. PI fluorescence was measured using a BD FACS Aria III flow cytometer (488 nm excitation; BD Biosciences, USA), and data were analyzed for cell cycle distribution. Luciferase reporter gene assay The tool cell line HEK293T was used for the reporter assay. Cells were seeded in 24-well plates one day before transfection. Transfection was performed the next day using PEI according to the manufacturer’s instructions. Each well contained 20 ng of Renilla luciferase, 100 ng of HLA-B promoter-driven luciferase, and 1 µg of expression plasmids for NLRC5, REV7, shNC, or shCTMP. Luciferase activity was measured at 24 h post-transfection using the Dual-Glo Luciferase Assay System, and relative HLA-B expression was determined by normalizing firefly luminescence to Renilla luminescence. Flow cytometric analysis of HLA-B expression Cells were fixed with 1% paraformaldehyde and subsequently immunostained using an anti-HLA-B antibody (Invitrogen #PA5-35345, diluted 1:50), followed by incubation with a Fluor 488-conjugated secondary antibody (Beyotime #A0428, diluted 1:200, 40 min). A minimum of 10,000 viable single-cell events were acquired per sample. The median fluorescence intensity (MFI) in the FL1 channel was measured for HLA-B-stained and isotype control populations. Free fatty acid detection FFA content was quantified using the Free Fatty Acid Assay Kit (Beyotime #S0215S). Cell /tissue supernatant extracted with isopropyl alcohol or standard was mixed with 50 µl working solution in a microplate. The reaction was incubated at 37 °C for 60 min in the dark. Fluorescence was measured in SpectraMax iD5 at excitation/emission wavelengths of 560/590 nm. TNFα, CXCL9, CXCL10 detection Levels of TNF-α, CXCL9, and CXCL10 in mouse subcutaneous tumor lysates were quantified using commercial ELISA kits (R&D Systems; TNF-α: #MTA00B, CXCL9: #DY392, CXCL10: #DY466) following the manufacturer’s instructions. Briefly, tumor tissues were homogenized in cold lysis buffer, and the clarified supernatants were collected by centrifugation. Total protein concentration was determined for normalization prior to ELISA, and absorbance was measured at 450 nm. In vivo tumorigenesis and treatment Male BALB/c mice (4-week-old, 14–17 g; Shanghai Slac Laboratory Animal Co. Ltd.) were maintained under specific pathogen-free conditions (24 °C, 50 ± 10% humidity, 12 h light/dark cycle) with ad libitum access to food and water. For xenograft establishment, mice received subcutaneous injections of 1 × 10⁶ stably transfected CT26 cells (MSS) into the right hind flank. Tumor volume was measured every week and calculated using the formula V = (length × width²) / 2. Treatment with IFN-γ was initiated when tumors reached a volume exceeding 200 mm³. Mice then received intra-tumoral injections of recombinant murine IFN-γ (2.5 µg in 25 µl) or an equivalent volume of PBS vehicle, administered twice weekly. All animal procedures were approved by the Ethics Committee of Shanghai Tenth People’s Hospital (Approval No. SHDSYY-2023-1740) and conducted in accordance with Tongji University’s guidelines for laboratory animal welfare. Flow cytometric analysis of T cell distribution Single-cell suspensions from CT26 syngeneic tumors were prepared. After washing with ice-cold PBS, cells were resuspended in FACS buffer (PBS, 2% FBS, 0.1% sodium azide) at 1 × 10⁶ cells/mL and stained with fluorochrome-conjugated antibodies—CD45:PerCP-Cy5-5-A (pan-leukocyte), CD3:FITC-A (T cell), and CD8-PE-Texas-Red-A (cytotoxic T cell)—for 30 min at 4 °C in the dark. Cells were washed twice, filtered through a 70-µm strainer, and analyzed on a BD FACS Aria III flow cytometer (BD Biosciences, USA). Lymphocytes were gated based on FSC-A/SSC-A, followed by single cells (FSC-H/FSC-A), CD45⁺ leukocytes, and T cell subsets (CD3⁺CD8⁺). Data were analyzed using FlowJo (v10.8.1, BD Biosciences), and frequencies were expressed as percentages of CD45⁺CD3⁺ lymphocytes. All antibodies were validated for murine samples (Abclonal #RK40007) and used per manufacturer’s instructions. Statistical analysis Data are presented as mean ± SEM. Statistical analyses were performed using GraphPad Prism 10.4.2. Differences between groups were assessed by two-way ANOVA or Student’s t-test, as appropriate. A P-value < 0.05 was considered statistically significant. Result CTMP and HLA-B predict prognosis in MSI-L/MSS CRC Survival analysis in patients with MSI-L/MSS colon adenocarcinoma revealed significantly prolonged overall survival (OS) in the low-CTMP expression cohort compared to the high-expression group (P = 0.015; Fig. 1A). Given the established role of CTMP in binding AKT and modulating its phosphorylation state [26, 30], we concurrently evaluated MHC-I molecules (HLA-A/B/C), which are known downstream targets suppressed by pAKT signaling [28, 29]. Elevated CTMP and reduced HLA-B expression correlate with poor prognosis in MSI-L/MSS CRC. (A-B) Kaplan-Meier overall survival analysis of COAD patients stratified by high versus low expression of (A) CTMP (Affymetrix ID: 1153118_at) and (B) HLA-B (Affymetrix ID: 208729_x_at). Log-rank test p-values are indicated. (C) Top: Representative immunofluorescence images showing inverse correlation between CTMP (green) and HLA-B(orange) expression in MSI-L/MSS CRC tissue microarray (n = 71 cores). Bottom: Waterfall plot quantifying CTMP and HLA-B protein expression levels across the cohort.(D) Western blot analysis of CTMP and HLA-B protein levels in five paired CTMP-high-expressing MSI-L/MSS CRC tumors (T) and adjacent normal tissues (N).GAPDH serves as loading control. (E) Longitudinal changes in tumor size (RECIST v1.1 criteria) during clinical follow-up in patients receiving anti-PD-1 plus fruquintinib therapy. (F) Left: Representative IF images depicting CTMP (green) and HLA-B (orange) expression in baseline tumor biopsies from MSI-L/MSS CRC patients treated with anti-PD-1 plus fruquintinib. Right: quantification of IF Scores (Range:1-100). Patient response categories: PR (partial response, n = 6), PD (progressive disease, n = 5). Mean ± SEM (Student’s t-test). (G) Top: Western blot analysis of CTMP and HLA-B expression in pre-treatment tumor samples from patients undergoing anti-PD-1 plus fruquintinib therapy. Bottom: Densitometric quantification of protein levels normalized to GAPDH. Mean ± SEM (Student’s t-test) Consistent with previous studies demonstrating that loss of MHC-I correlates with poor clinical outcomes in multiple cancer types including melanoma, CRC, and breast cancer [51–54], elevated HLA-B expression was significantly associated with improved OS in this cohort (p = 0.037; Fig. 1B). In contrast, neither HLA-A nor HLA-C expression showed a significant association with survival outcomes (Figure S2A-B). For tissue-based validation, we performed a comprehensive analysis using a 71-pair MSI-L/MSS CRC tissue microarray (clinical characteristics detailed in Table S1). Immunofluorescence (IF) analysis revealed significantly higher CTMP expression in tumor tissues compared to adjacent normal mucosa. Conversely, HLA-B expression was downregulated in tumor compartments (Fig. 1C). Of note, the delineated area (white dashed line) indicated higher constitutive HLA-B expression in normal tissues than in tumors, predominantly localized to immune cells within the lamina propria mucosae, with minimal detection in enterocytes. This pattern suggests enhanced antigen processing and presentation capacity in normal mucosa. Quantitative assessment of differential protein expression, represented as IF score differentials (Normal minus Tumor, N-T), was visualized using a waterfall plot (Fig. 1C). Subsequent western blot validation in five matched tumor-normal pairs with dichotomous CTMP expression (defined by tumor IF score ≥ 70 and normal IF score ≤ 30) confirmed a consistent inverse correlation between CTMP upregulation and HLA-B downregulation (Fig. 1D). Given the established positive correlation between MHC-I expression and tumor-local immune cytolytic activity [55], as well as the association between deficient MHC-I-restricted neoantigen presentation and impaired efficacy of immune checkpoint inhibitor (ICI) therapy [56], we investigated the prognostic value of CTMP and HLA-B expression in ICI-based treatment for metastatic MSI-L/MSS CRC. In a cohort of refractory MSI-L/MSS CRC patients (n = 11) receiving third-line anti-PD-1 plus fruquintinib therapy (Fig. 1E)—all of whom had progressed after two prior lines of standard systemic therapy (fluorouracil-, oxaliplatin-, or irinotecan-based chemotherapy combined with anti-VEGF agents; baseline clinicopathological parameters in Table S2)—IF analysis of pretreatment tumor specimens revealed significantly reduced CTMP expression (P < 0.0001) and elevated HLA-B expression (P = 0.0041) in partial responders (PR, n = 6) compared to patients with progressive disease (PD, n = 5) (Fig. 1F). Specifically, enhanced HLA-B in the PR group showed a high-consistent localization within the enteroid morphology, suggesting a local functional HLA-B. Quantitative analysis of normalized fluorescence intensity corroborated these findings: CTMP levels were substantially lower in the PR group than in the PD group, whereas HLA-B expression was markedly upregulated in the PR group, showing an approximately 2-fold increase compared to the PD group (Fig. 1G). Collectively, our integrated analysis identifies CTMP overexpression and concomitant HLA-B downregulation as independent prognostic biomarkers in MSI-L/MSS CRC. Furthermore, these molecular signatures may serve as potential exploratory markers for therapeutic response to anti-PD-1 plus fruquintinib combination therapy in third-line refractory MSI-L/MSS CRC patients. CTMP coordinates metabolic and immunological programs in CRC cells To delineate the role of CTMP in colorectal carcinogenesis, we established stable CTMP-knockdown (CTMP-KD) and CTMP-overexpressing (CTMP-OE) CRC cell lines. CTMP depletion markedly attenuated proliferative capacity, as evidenced by CCK-8 viability and colony formation assays (Fig. 2A, B). Conversely, CTMP-OE cells exhibited enhanced proliferation and clonogenicity. CTMP regulates proliferation, immunogenicity, and metabolism in CRC cells. (A) Cell proliferation. CCK-8 assay in control, CTMP-KD, and CTMP-OE SW620/SW480 cells at indicated time points. Mean ± SEM, n = 3(**P < 0.01, Student’s t-test).(B) Colony formation via crystal violet staining after 10-day culture. Mean ± SEM, n = 3(**P < 0.01, Student’s t-test).(C) Western blot analysis of indicated proteins in shNC, shCTMP, and CTMP-OE cells. GAPDH as loading control. (D) MHC-I gene expression. qPCR analysis of HLA-B, B2M, TAP1, and TAP2 in shNC /CTMP-KD cells. Normalized to GAPDH. Mean ± SEM, n = 3(*P < 0.05, Student’s t-test). (E) Metabolic intermediates. Intracellular acetyl-CoA and lactate levels in shNC /CTMP-KD cells. Mean ± SEM, n = 3(**P < 0.01, Student’s t-test). (F) Lipid metabolism. Left: Cell viability after palmitic acid treatment (100 µM, 24 h) via CCK-8, normalized to untreated control. IC₅₀ indicated by dashed lines. Right: Oil Red O staining in cells treated with 100 µM palmitic acid for 48 h. Scale bar, 100 μm. Images were converted to grayscale, a fixed threshold was applied to select Oil Red O-positive areas, and the stained area percentage relative to the total area was quantified. n = 3(**P < 0.01, Student’s t-test).(G) qPCR analysis of HLA-B in shNC /CTMP-KD cells after palmitic acid treatment (100 µM, 48 h). Normalized to GAPDH. Mean ± SEM, n = 3(*P < 0.05, Student’s t-test). (H) Mitochondrial respiration. Left: OCR parameters in shNC /CTMP-KD cells. Right: Quantification of basal and maximal OCR. Mean ± SEM, n = 6 (**P < 0.01, Student’s t-test) Immunophenotypic analysis revealed a dual immunoregulatory function of CTMP in CRC (Fig. 2C). CTMP-KD concurrently upregulated HLA-B and downregulated PD-L1, while CTMP-OE elicited the opposite phenotype. Given its known interaction with the C-terminus of AKT, we examined CTMP’s impact on AKT signaling—a subject of conflicting reports [26, 30]. In our model, CTMP-KD suppressed intrinsic AKT phosphorylation at Ser473, accompanied by reduced expression of the glycolytic enzymes HK2 and LDHA, indicating compromised glycolytic flux. Notably, CTMP-KD also consistently elevated P21 expression, a marker of cell cycle arrest. In contrast, while CTMP-OE augmented AKT phosphorylation, the lack of further changes in p21, HK2, and LDHA (Fig. 2C) may reflect near-maximal activation of these downstream effectors under the experimental conditions. Furthermore, analogous effects were observed in the MSI-H HCT116 cell line (Figure S3), indicating that the observed functions of CTMP are not restricted to a particular microsatellite instability background. Subsequent qPCR analysis confirmed a broader upregulation of MHC-I-related genes in CTMP-KD cells, including the antigen-presenting determinants HLA-B and B2M, as well as the antigen-processing transporters TAP1 and TAP2 (Fig. 2D). Given the established role of thioesterases in fatty acid metabolism [17, 23], we next performed metabolic profiling of CTMP-KD cells. CTMP depletion resulted in a > 40% reduction in intracellular acetyl-CoA levels and a > 35% decrease in lactate secretion (Fig. 2E), reflecting disrupted acetyl-CoA homeostasis and impaired glycolytic flux. To assess fatty acid metabolism, we determined the IC50 for palmitic acid. CTMP-KD cells exhibited a reduced IC50 and decreased Oil Red O staining (Fig. 2F), indicating impaired incorporation of exogenous palmitic acid into lipid droplets—a phenotype reminiscent of ACOT13 deficiency [57]. Interestingly, palmitic acid attenuated HLA-B expression in both shNC and CTMP-KD cells, supporting a functional link between fatty acid metabolism and MHC-I antigen presentation (Fig. 2G). Nevertheless, the regulatory effect of palmitic acid is expected to be even more limited under physiological concentrations. Furthermore, Seahorse extracellular flux analysis revealed severe mitochondrial dysfunction in CTMP-KD cells (Fig. 2H), characterized by significantly reduced oxygen consumption rates (OCR) under both basal and stressed conditions (P < 0.05). Taken together, CTMP drives colorectal cancer progression by coordinating metabolic and immunological programs. CTMP knockdown suppresses proliferation, impairs glycolysis and mitochondrial function, while concurrently upregulating MHC-I molecules and downregulating PD-L1. These findings establish CTMP as a key regulator linking cellular metabolism to immune recognition in CRC cells. CTMP interacts with REV7 in CRC cells To systematically identify CTMP-interacting partners, co-IP coupled with proteomic analysis in SW620 cell lysates (anti-CTMP antibody) identified multiple candidates (Fig. 3A). Given evidence linking CDK1 inhibition to cGAS re-activation [58] and cGAS-mediated antitumor immunity [47], we prioritized REV7—a mitotic checkpoint regulator and CDK1 suppressor [35, 45]—for validation. Notably, profile of REV7 levels in immune modulation was accessed (Figure S4 B-D) and showed positive correlations with MHC-I components across most tumors, including COAD, with HLA-A, HLA-B, HLA-C, and TAP1 exhibiting modest to moderate positive correlations (Figure S4D). CTMP interacts with REV7 in CRC cells. (A) Proteomic profiling of CTMP-interacting proteins. Co-IP with anti-CTMP antibody versus IgG control, followed by proteomic analysis. REV7 is highlighted as a top-scoring interactor. (B) Validation of endogenous CTMP-REV7 interaction. Co-IP with anti-REV7 or anti-CTMP antibody in SW620 cell lysates, probed for CTMP or REV7. (C) Subcellular co-localization analysis. Left: Representative confocal immunofluorescence micrographs depicting the subcellular distribution of endogenous CTMP (green) and REV7 (red) in SW620 cells. Nuclei were counterstained with DAPI (blue). Right: Line-scan fluorescence intensity profiles for CTMP and REV-7 along the transect indicated by the white arrow (left panel). Profiles were quantified and plotted using ImageJ software (version 1.53). (D) De novo structural modeling of the CTMP-REV7 complex. Predicted interaction interface generated by AlphaFold Multimer (v2.3.2). CTMP(green) includes the core enzyme activity area- hotdog fold structure of α-helix (pink) and β-sheet (red). (E) Co-IP assays of CTMP trucations interaction with REV7. (F) Interaction and structural mapping of Flag-CTMP point mutants with GST-REV7. Left: Coomassie Brilliant Blue staining of the interaction between GST-REV7 and indicated CTMP point mutants; Middle: Structural mapping of CTMP interface residues. Close-up view of computationally identified CTMP binding residues (stick representation) within the REV7 binding pocket (surface representation). Right: Coomassie Brilliant Blue staining of the interaction between GST-REV7 and indicated CTMP R3A/R8A double mutant. (G) Exogenous Co-IP in SW620 cells transfected with FLAG-CTMP point mutants and HA-REV7 overexpressed cells. Top: CTMP (D89A, R75A, R8A, R3A); Bottom: CTMP R3A/R8A double mutant. (H) CCK-8 assay of CTMP-WT, D89A, R75A, R8A, R3A, R3A/R8A cells at 24, 48, 72 h post-treatment. Mean ± SEM, n = 3 (*P < 0.05, ** P < 0.01, ***P < 0.001, two-way ANOVA). (I) qPCR analysis of HLA-B in Vec, CTMP-WT, CTMP-R3A/R8A cells. Normalized to GAPDH. Mean ± SEM, n = 3(*P < 0.05, Student’s t-test) We next validated the CTMP-REV7 interaction. Endogenous co-IP assays confirmed complex formation between CTMP and REV7 (Fig. 3B). To assess its clinical relevance, we analyzed the TCGA COAD cohort and found significant co-upregulation of CTMP and REV7 transcripts in tumor tissues compared to matched normal controls (P < 0.001; Figure S4A). This coordinated elevation suggests a potential dependency on the CTMP-REV7 axis in colon adenocarcinoma pathogenesis. Immunofluorescence microscopy delineated distinct subcellular compartmentalization patterns (Fig. 3C). CTMP (green) exhibited predominantly cytoplasmic localization with concomitant nuclear presence, where pronounced co-localization with REV7 (red) was specifically observed within the nuclear compartment (merged yellow signal). Quantitative line-scan fluorescence intensity profiling along the designated axis (white arrow) further corroborated the spatial correlation between CTMP and REV7 protein distributions. CTMP is a 240-amino acid protein comprising a C-terminal PaaI-type thioesterase domain (residues 119–231) that adopts a canonical hot-dog fold and harbors the catalytic center [24] (Fig. 3D). To delineate its REV7-binding region, we performed systematic truncation assays and identified the N-terminal segment (residues 1–119) as both necessary and sufficient for robust interaction, while the catalytic thioesterase domain showed no detectable binding (Fig. 3E). This establishes that CTMP engages REV7 independently of its enzymatic module. We modeled the CTMP–REV7 complex using AlphaFold Multimer (v2.3.2), which predicted an ultra-tight interface with a PRODIGY-calculated ΔG of − 17.2 kcal/mol—well below the − 12 kcal/mol threshold associated with exceptionally stable complexes. Interface analysis across 451 residues identified 23 distinct electrostatic nodes (Table S3). Guided by the strong binding mapped to residues 1–119 and the structural model, we selected four charged residues—R3, R8, R75, and D89—for alanine-scanning mutagenesis (Figure S1A). The model suggests that R3 and R8 form a high-density, dual-anchored salt-bridge network with D175 and D178 of REV7, whereas D89 and R75 engage in complementary electrostatic contacts with R45 and E59 (Table S3), collectively shaping the conformational landscape of the complex. GST pulldown (Fig. 3F) and co-IP (Fig. 3G) assays consistently demonstrated that R3A and R8A substitutions severely compromised binding, while D89A and R75A retained substantial interaction. The R3A/R8A double mutant abolished complex formation in both assays, underscoring the essential role of this arginine pair. Functionally, wild-type CTMP strongly promoted SW620 cell viability (P < 0.001; Fig. 3H). Single mutants (R75A, R8A, R3A) induced moderate but significant reductions in proliferation, whereas the R3A/R8A double mutant provoked a profound defect (P < 0.001), directly linking the structural integrity of the CTMP–REV7 interface to cellular proliferation. To further assess whether disrupting the CTMP–REV7 complex impacts antigen presentation, qPCR analysis revealed that the CTMP-R3A/R8A mutant partially reversed the CTMP-WT-mediated downregulation of HLA-B (Fig. 3I), suggesting a role for this complex in MHC-I regulation. In summary, CTMP interacts with REV7 via its N-terminal domain (residues 1–119), with R3 and R8 serving as critical residues for complex formation. Disruption of this interface impairs proliferation and modulates HLA-B expression, suggesting that the CTMP–REV7 axis links cell cycle regulation to MHC I machinery in CRC. CTMP-REV7 governs MHC-I/HLA-B expression via CDK1 activity modulation and is functionally linked to cGAS signaling Given the established interaction between REV7 and CDK1 [45], a master kinase governing the G2/M transition [59], we hypothesized that the proliferative defect associated with the CTMP R3A/R8A mutant (Fig. 3H) may arise from disrupted regulation of the REV7–CDK1 axis. To dissect the molecular role of CTMP within this pathway, we first mapped its protein-protein interaction profile. Exogenous binding assays confirmed that CTMP directly engages REV7 and AKT, but not CDK1 itself (Fig. 4A). CTMP-REV7 regulates MHC-I/HLA-B expression by modulating CDK1 activity and is relevant to cGAS. (A) Exogenous Co-IP in SW620 cells transfected with Flag-CTMP and HA-tagged Constructs. Mixed transfection (Flag-AKT, -REV7, -CDK1) with CTMP served as a positive control. (B) Western bolt of pHistone H1, pAKT(S473), and HLA-B expression in REV7-KD/REV7-OE cells. Mean ± SEM, n = 3 (*P < 0.05, Student’s t-test). (C) Endogenous Co-IP with anti-REV7 antibody in CTMP-KD vs. control cells. (D) In vitro binding assay: To assess competitive disruption, the anti-REV7 precipitations of the REV7-CDK1 complex were incubated with GST-CTMP or its mutant form (R3A/R8A) for 60 min at 25 °C, the washed immunoprecipitation was incubated with anti-CDK1 and anti-REV7 antibody. (E) The suppressive effect of REV7 on CyclinB-CDK1 activity was quantified using an in vitro kinase assay. CDK1-Cyclin B1 kinase was incubated with GST-NC (2 parts), GST-REV7 + GST-NC (1:1 ratio), GST-REV7 + GST-CTMP (1:1 ratio), and kinase activity was determined by measuring Histone H1 phosphorylation. Mean ± SEM, n = 6 (**P < 0.01, two-way ANOVA). (F) Cell cycle redistribution upon modulation of CTMP-REV7 axis: Propidium iodide (PI) flow cytometry analysis of controls, CTMP-KD or REV7-OE cells, Data quantified as % cells in G2-M phases. Mean ± SEM, n = 3 (**P < 0.01, Student’s t-test). (G) Representative flow cytometry histograms depicting HLA-B expression in tumor cells from the Control, CTMP-KD, REV7-OE, and REV7-OE + Ro-3306 groups. Ro-3306, a selective CDK1 inhibitor, was used at 2 µM, a concentration at which it exhibits > 10-fold selectivity for CDK1 over other CDKs with reference to Vassilev et al. (PNAS, 2006). Cells were treated for 24 h. (H) HLA-B reporter activity was measured in HEK293T cells transfected with the indicated plasmids. NLRC5-OE was used as a positive control, n = 3. (*P < 0.05, **P < 0.01, Student’s t-test) (I) qPCR analysis of HLA-B, B2M, TAP1, and TAP2 in the indicated groups, normalized to GAPDH. Mean ± SEM, n = 3(*P < 0.05, **P < 0.01, one-way ANOVA). (J) Western blot detects cGAS sinaling(p-TBK1(S172), p-IRF3(S396)) Functionally, REV7-OE suppressed CDK1 activity—as measured by histone H1 phosphorylation—and reduced AKT phosphorylation, while concurrently upregulating HLA-B expression. Conversely, REV7-KD reversed these effects (Fig. 4B). These observations positioned REV7 as a key node linking CTMP to CDK1 regulation. We next examined whether CTMP physically influences the REV7–CDK1 interaction. Co-IP revealed that CTMP-KD enhanced the association between REV7 and CDK1 (Fig. 4C), indicating that endogenous CTMP limits their interaction. Supporting this, an in vitro binding assay demonstrated that the CTMP R3A/R8A mutant—which is deficient in REV7 binding—preserved a significantly larger pool of the REV7–CDK1 complex compared to wild-type CTMP (Fig. 4D). Collectively, these findings suggest that CTMP functions as a dynamic regulator: its loss—whether by knockdown or mutation—liberates REV7 to more effectively engage and modulate CDK1 activity. Direct enzymatic evidence showed that 200 nM GST-REV7 suppressed CDK1-cyclin B1 activity by over 50% (P < 0.01; Fig. 4E), and the addition of GST-CTMP partially rescued CDK1 activity in the presence of REV7 (P < 0.01),accordingly, both CTMP-KD and REV7-OE induced significant G2/M phase arrest (P < 0.05; Fig. 4F), a hallmark of impaired CDK1 function and REV7-mediated cell cycle dysregulation. To test whether the CTMP-REV7-CDK1 axis regulated MHC-I expression could be interfered by CDK1 activity, we performed the following experiments. Both CTMP knockdown and REV7-OE significantly increased surface HLA-B levels (P < 0.05; Fig. 4G). Notably, the CDK1 inhibitor Ro-3306(IC 50 determined in Figure S5A) further enhanced HLA-B expression in REV7-OE cells, indicating a cooperative effect between REV7-mediated modulation and CDK1 activity. Direct HLA-B promoter activity assays confirmed that REV7 overexpression drives promoter activation, an effect further potentiated by Ro-3306. Consistently, CTMP-KD similarly augmented HLA-B promoter activity (Fig. 4H). In line with these findings, REV7 knockdown reduced MHC-I transcript levels, an effect partially reversed by Ro-3306, whereas REV7 re-expression significantly increased MHC-I transcripts (Fig. 4I). To investigate the mechanism by which CDK1 inhibition modulates REV7-regulated MHC-I output, we performed bioinformatic analysis, which revealed positive correlations between REV7 and the innate immune mediators cGAS and IRF3 across most tumor types, with a moderate-to-strong correlation specifically for IRF3 (Figure S4E). This suggests a conserved regulatory axis linking REV7 expression to the cGAS-STING pathway. Functionally, REV7-KD led to increased CDK1 activity and reduced phosphorylation of TBK1 and IRF3, key events in cGAS signaling. These effects were reversed by the CDK1 inhibitor Ro-3306, indicating that REV7 sustains cGAS signaling by suppressing CDK1 activity. Conversely, REV7-OE enhanced Ro-3306-mediated cGAS activation (Fig. 4J), further supporting that REV7 primes the pathway for CDK1 inhibition-driven amplification. Mechanistically, CDK1 supresses cGAS signaling [47, 58]. REV7, as a mitotic checkpoint component, negatively regulates CDK1 activity (Fig. 4B), consequently, REV7-OE synergistically enhances cGAS activation (Fig. 4J), promotes G2/M arrest (Fig. 4E), then drives downstream immune outputs. In colorectal cancer, MSS and MSI-L tumors are collectively characterized by high levels of chromosomal instability (CIN) [36], our findings suggest that the REV7-CDK1 axis may be particularly relevant in CIN-high MSS/MSI-L CRC. Together, our findings demonstrate that CTMP binds REV7 and limits its interaction with CDK1. Loss of CTMP or REV7 OE suppresses CDK1 activity, upregulates HLA-B expression, and promotes cGAS signaling. These results identify the CTMP-REV7-CDK1 axis as a regulator of MHC-I expression and innate immune pathways. CTMP-REV7 modulates IFN-γ-driven MHC-I expression via crosstalk of CDK1 and AKT IFN-γ exerts context-dependent dual roles in tumor immunity, promoting antitumor responses via enhanced antigen presentation and ferroptosis while driving adaptive resistance through the STAT1-IRF1 axis [60]. To determine whether the CTMP–REV7–CDK1 module integrates into this signaling landscape, we performed RNA-seq on IFN-γ-stimulated SW620 cells. Transcriptomic analysis (Figure S6) confirmed canonical upregulation of MHC-I molecules (HLA-A/B/C) and antigen processing machinery (TAP1/2, B2M), and further revealed a metabolic reprogramming characterized by upregulation of lipolytic genes (e.g., ADRB2, ADORA1), consistent with IFN-γ-induced lipolysis [61–63]. Within this context, REV7 was induced while CDK1 was downregulated, with CTMP expression remaining stable. Co-IP and IF analyses confirmed that IFN-γ promotes the physical association and subcellular co-localization of CTMP with REV7 (Figs. 5A-B). Notably, while IFN-γ also increased CTMP–AKT binding, this interaction was competitively reduced upon REV7-OE (Fig. 5D), suggesting that REV7 remodels the IFN-γ-induced signaling complex by competing with AKT for CTMP engagement. CTMP-REV7 governs IFN-γ-induced MHC-I expression through the CDK1-AKT signaling crosstalk. (A) CTMP interacts with REV7 in an IFN-γ-dependent manner. SW620 cells were treated with IFN-γ (10 or 20 ng/mL,24 h). Cell lysates were immunoprecipitated with an anti-CTMP antibody, followed by immunoblotting with an anti-REV7 antibody. (B) IFN-γ enhances CTMP-REV7 colocalization. Left: Representative immunofluorescence images showing CTMP (green) and REV7 (red) in SW620 cells treated with or without IFN-γ (10 ng/mL, 24 h). Scale bar, 10 μm. Right: Quantification of CTMP-REV7 colocalization per cell. Mean ± SEM, n = 3. **P < 0.01 (Student’s t-test). (C) Western blot analysis of pHistone H1, pAKT (S473), and HLA-B in REV7-OE cells following treatment with IFN-γ (10 ng/mL, 24 h), Ro-3306 (2 µM, 24 h), or Sc-79 (5 µM, 24 h). Mean ± SEM, n = 3. *P < 0.05; ns, not significant (Student’s t-test). (D) REV7-OE displaces CTMP from AKT. Cells were treated with IFN-γ (10 ng/mL, 24 h) and subjected to immunoprecipitation with anti-CTMP antibody, followed by immunoblotting with an anti-AKT antibody. (E) CTMP mutation promotes the interaction between REV7 and CDK1. HA-REV7 wild-type (WT) stable SW620 cells were transfected with CTMP-WT or CTMP-MT, treated with or without IFN-γ (10 ng/mL, 24 h). Cell lysates were immunoprecipitated with anti-HA antibody, followed by immunoblotting with anti-CDK1 antibody to detect the REV7-CDK1 interaction. (F) REV7 attenuates the interaction between CDK1 and AKT. HA-CDK1 wild-type (WT) stable SW620 cells were transfected with REV7-WT or vector, treated with or without IFN-γ (10 ng/mL, 24 h). Cell lysates were immunoprecipitated with anti-HA antibody, followed by immunoblotting with anti-AKT antibody to detect the CDK1-AKT interaction. (G) REV7 overexpression enhances IFN-γ-induced MHC-I expression, which is attenuated by AKT activation. Vector, REV7-OE, REV7-OE + Sc-79 (5 µM, 24 h) cells were treated with IFN-γ (10 ng/mL, 24 h). qPCR analysis for HLA-B, B2M, TAP1, TAP2. Normalized to GAPDH. Mean ± SEM, n = 3(**P < 0.01, *P < 0.05, Student’s t-test). (H) Summary scheme.Normal: CTMP binds AKT and REV7. IFN-γ: Enhances CTMP-REV7/AKT binding. IFN-γ + CTMP-KD: Reduces AKT signaling, relieves REV7. IFN-γ + REV7-OE: Suppresses CDK1 and disrupts CTMP-AKT binding To determine whether REV7 modulates HLA-B through CDK1-AKT crosstalk, we investigated the functional interplay between these kinases. We first established that AKT activation with SC-79 attenuated IFN-γ-induced HLA-B expression (Figure S5B). Subsequently, in REV7-OE cells, Ro-3306 upregulated HLA-B—an effect reversed by SC-79 both with and without IFN-γ stimulation—while SC-79 also restored p-AKT and p-histone H1 levels suppressed by Ro-3306 (Fig. 5C, S5D). These results indicate that REV7 enhances HLA-B expression through coordinated suppression of CDK1 and AKT, and that AKT activation can override the effect of CDK1 inhibition. We next dissected the internal regulatory logic of this axis. Co-IP revealed that exogenous CTMP attenuated the REV7–CDK1 interaction (Fig. 5E, lane 3 vs. 2). This disruptive effect required functional CTMP, as the REV7-binding-deficient mutant (R3A/R8A) failed to disrupt the complex and instead enhanced REV7–CDK1 association even under IFN-γ stimulation (Fig. 5E, lane 5 vs. 4). Downstream, REV7 diminished the CDK1–AKT interaction [64] (Fig. 5F, lane 3 vs. 2), an effect that persisted in the presence of IFN-γ (Fig. 5F, lane 4 vs. 5). Functionally, REV7-OE-mediated MHC-I upregulation was partially reversed by SC-79 (Fig. 5G), indicating that REV7 modulates immunogenicity in part through AKT. These findings establish that REV7 integrates IFN-γ signaling with cell cycle control and metabolic cues to modulate MHC-I expression, partially through CDK1-AKT crosstalk. This positions the CTMP-REV7 interaction as a potential therapeutic target to potentiate IFN-γ-driven antigen presentation in tumors. Disrupting CTMP–REV7 axis activates anti-tumor immunity in vivo Building on our in vitro findings, we established in vivo models via subcutaneous implantation of CT26 cells into immunocompetent BALB/c mice. Both CTMP -KD and REV7-OE markedly attenuated tumor growth by Day 28th post-implantation (Fig. 6A and C), with no discernible impact on body weight throughout the study period (Fig. 6B). Modulation of CTMP-REV7 axis enhances anti-tumor immunity in vivo. (A) Tumor appearance and (B) Body weights of immunocompetent BALB/c mice were monitored weekly throughout the 4-week treatment period. Mean ± SEM, n = 4 (N.S. not significant, two-way ANOVA). (C) Growth kinetics of subcutaneous xenografts derived from control, CTMP-KD, or REV7-OE CT26 cells, Mean ± SEM, n = 4 (**P < 0.01, two-way ANOVA). (D) CTMP-REV7 axis modulates CDK1/AKT signaling and enhance CD8 + T cell infiltration in vivo. Left: Immunoblot analysis of pHistone H1, pAKT (Ser473) expression in individual subcutaneous tumor samples. Right: Representative immunohistochemical images and corresponding quantification of pHistone H1, pAKT (Ser473), PD-L1 expression, and CD8⁺ T-cell infiltration in tumor sections. Mean ± SEM, n = 4, (*P < 0.05, Student’s t-test). (E) Analysis of FFA and acetyl-CoA levels in tumor tissues. Quantification of FFA and acetyl-CoA in syngeneic tumors from CTMP-KD and REV7-OE groups. Mean ± SEM, n = 4 (*P < 0.05, **P < 0.01, Student’s t-test). (F) CTMP-REV7 axis regulates MHC-I antigen presentation in vivo. Total RNA was extracted from control, CTMP-KD, and REV7-OE tumors and analyzed by qPCR for H2-K1, B2M, TAP1, and TAP2. mRNA levels were normalized to GAPDH. Mean ± SEM, n = 4 (**P < 0.01, Student’s t-test). (G) Tumor-infiltrating lymphocyte profiling by flow cytometry demonstrating gating strategy for CD8+/CD3+ T cells and quantitative analysis of infiltrating CD8+/CD3+ T-cell populations ratio. Mean ± SEM, n = 4(*P < 0.05, Student’s t-test) Immunoblotting of tumor lysates confirmed that both CTMP knockdown and REV7 overexpression reduced CDK1 and AKT activity in vivo (Fig. 6D, left). These findings were corroborated by IHC analysis, which showed lower levels of phosphorylated histone H1 and phosphorylated AKT in tumor sections from the same intervention groups (Fig. 6D, right). Notably, IHC also revealed decreased PD-L1 expression and increased CD8⁺ T lymphocyte infiltration within the tumor microenvironment following both genetic modifications (Fig. 6D, right). Flow cytometry further quantified a significant increase in the proportion of CD8⁺ T cells among the CD3⁺ T-cell population (P < 0.05; Fig. 6G). Concurrent metabolic profiling showed that both interventions reduced intratumoral FFA and acetyl-CoA levels (Fig. 6E). Specifically, FFA decreased by > 35% in CTMP-KD tumors and ~ 40% in REV7-OE tumors, while acetyl-CoA declined by approximately 50% and 30%, respectively (Fig. 6E). Transcriptomic analysis of the same tumors revealed significant upregulation of MHC-I-related genes, including H2-K1, B2M, TAP1, and TAP2 (Fig. 6F). Together, these in vivo data demonstrate that CTMP-KD and REV7-OE attenuate CRC progression in immunocompetent mice, correlating with reduced proliferative signaling, enhanced intratumoral CD8⁺ T-cell infiltration, and a reprogrammed metabolic landscape. REV7 quantity determines the synergistic efficacy of CTMP knockdown with IFN-γ in tumor regression To determine whether REV7 mediates the synergistic anti-tumor effect of CTMP knockdown under IFN-γ-induced chronic immune stress, we established an in vivo model using CT26 cells. Cells were engineered with stable CTMP-KD, followed by secondary perturbation of REV7 expression (REV7-KD or REV7-OE). Control cells were transduced with shNC and empty vector. To model immune stress, murine IFN-γ (2.5 µg [65]) was administered via intra-tumoral injection twice weekly beginning on Day 18, at which point tumors had reached a comparable size exceeding 200 mm³ (Fig. 7A). REV7 modulates CTMP knockdown-enhanced anti-tumor immunity and MHC-I antigen presentation in IFN-γ-treated tumors. (A) Schematic diagram of the syngeneic mouse tumor model and experimental timeline. BALB/c mice were subcutaneously implanted with CT26 cells with indicated genetic modifications on Day 0, followed by intratumoral IFN-γ (2.5 µg) injections as indicated. Tumors were harvested on Day 39 for analysis. (B) Body weights of mice were measured weekly throughout the 39-day treatment period. Mean ± SEM, n = 6, No significant differences were observed among groups (two-way ANOVA). (C) Growth kinetics of subcutaneous xenografts derived from control (shNC+Vector), CTMP-KD+REV7-KD, CTMP-KD+Vector, CTMP-KD+REV7-OE. Mean ± SEM, n = 6(*P < 0.05, **P < 0.01, two-way ANOVA). (D) REV7 is required for CTMP knockdown-enhanced MHC-I antigen presentation under IFN-γ treatment. Left: Mice bearing control or CTMP-KD tumors received intratumoral injections of PBS or IFN-γ as indicated. Tumor volumes were measured at harvest on Day 39. Data are mean ± SEM, n = 6 (**P < 0.01, *P < 0.05, two-way ANOVA). Right: qPCR analysis of H2-K1, B2M, TAP1, TAP2 in tumors treated with IFN-γ. mRNA levels were normalized to GAPDH. Mean ± SEM, n = 6(**P < 0.01, *P < 0.05, Student’s t-test). (E) REV7 modulates CTMP knockdown-enhanced CD8⁺ T cell infiltration. Left: Representative immunohistochemical images and quantification of CD8⁺ T cell infiltration in tumor sections from indicated groups. Scale bar, 50 μm. Mean ± SEM, n = 6 (*P < 0.05, *P < 0.05, Student’s t-test). Right: Representative flow cytometry plots showing gating strategy for CD8⁺/CD3⁺ T cells (right) and quantification (middle) of CD8⁺/CD3⁺ T cell frequency among total CD3⁺ tumor-infiltrating lymphocytes. Mean ± SEM, n = 6(*P < 0.05, Student’s t-test). (F) Tumor lysates from indicated groups were analyzed by ELISA for TNF-α, CXCL9, and CXCL10 protein levels. Mean ± SEM, n = 6(**P < 0.01, *P < 0.05, N.S. not significant, Student’s t-test) By Day 39, CTMP-KD alone induced significant tumor regression compared to controls (P < 0.01, Figs. 7C–D). Notably, this anti-tumor effect was partially abrogated by concurrent REV7 knockdown (P < 0.05), indicating that REV7 is required for the full efficacy of CTMP deficiency. Conversely, CTMP-KD combined with REV7-OE triggered pronounced tumor regression, achieving a greater than 50% reduction in tumor volume relative to CTMP-KD alone (P < 0.01). These results demonstrate a REV7 quantity-dependent enhancement of anti-tumor efficacy, wherein REV7-OE maximally potentiates the effect of CTMP loss. Mechanistic interrogation revealed that CTMP-KD synergized with REV7-OE to orchestrate a phenotypic shift toward enhanced immunogenicity. This was evidenced by significantly elevated MHC-I expression (Fig. 7D, right) and increased intra-tumoral infiltration of CD8⁺ T cells, accompanied by a higher CD8⁺/CD3⁺ T-cell ratio (P < 0.05; Fig. 7E). Consistent with these immunological changes, analysis of tumor lysates demonstrated substantial upregulation of key anti-tumor chemokines in CTMP-KD combined with REV7-OE group—including TNF-α, CXCL9, and CXCL10 (Fig. 7F)—a signature characteristic of an immunologically “hot” tumor microenvironment. Collectively, these findings establish REV7 as a critical determinant of IFN-γ synergy with CTMP knockdown. REV7 overexpression maximally potentiates anti-tumor immunity through the coordinated enhancement of antigen presentation, cytotoxic T-cell recruitment, and chemokine-mediated immune activation. Discussion In this study, we identified CTMP as a clinically actionable biomarker in MSI-L/MSS colorectal cancer. Elevated CTMP expression, coupled with diminished HLA-B levels, robustly predicts poor patient outcomes across multiple independent cohorts (Fig. 1A-D). This prognostic significance extends to therapeutic response: in a small cohort of MSI-L/MSS patients receiving anti-PD-1 plus fruquintinib therapy, CTMP downregulation correlated with HLA-B upregulation and treatment efficacy (Fig. 1E-G). Although limited by sample size (n = 11)—a constraint inherent to the low response rate of MSI-L/MSS CRC to such regimens—these findings warrant validation in larger, independent cohorts to establish CTMP as a stratifying biomarker for combined immune checkpoint and anti-angiogenic therapy. Mechanistically, CTMP exerts its pro-tumorigenic effects primarily through AKT signaling. AKT activation dampens antitumor immunity via multiple convergent mechanisms: downregulation of MHC-I [28, 29], upregulation of PD-L1 [66], and promotion of immunosuppressive cell recruitment (Tregs and MDSCs) [67, 68], collectively shaping an immunosuppressive tumor microenvironment (TME). Notably, CTMP exhibits context-dependent functional duality in regulating AKT phosphorylation. Although initially characterized as an AKT inhibitor [26, 27], subsequent studies—including our own (Fig. 2C)—demonstrate its ability to enhance pAKT levels in cancer cells, consistent with pro-tumorigenic roles reported in other malignancies [31, 32]. Importantly, CTMP knockdown directly counteracts AKT-driven immune evasion by simultaneously inhibiting AKT/PD-L1 signaling and enhancing HLA-B expression (Fig. 2C), achieving synergistic advantage through coordinated enhancement of antigen presentation and relief of immunosuppression. The tissue-specific overexpression of CTMP in colon cancer (Figure S4A) further reinforces its suitability as a tumor-intrinsic therapeutic target. While AKT inhibitor monotherapy is limited by systemic toxicity, its combination with immune checkpoint inhibitors is under active investigation [69], underscoring the translational relevance of targeting the CTMP–AKT axis. Beyond AKT modulation, CTMP functions as a master metabolic regulator. By virtue of its hotdog-fold-dependent acyl-CoA thioesterase activity [25], CTMP dynamically coordinates acyl-CoA availability with fatty acid metabolism [23, 24]. Functionally, CTMP deficiency reduced cellular tolerance to palmitic acid (Fig. 2F) and suppressed FAO and OXPHOS flux (Fig. 2H). These observations align with CTMP upregulation under high-fat diet or free fatty acid stress [33, 34], positioning it as a metabolic rheostat that fine-tunes FAO during nutrient surplus. Mechanistically, CTMP knockdown may impair acetyl-CoA conversion to free CoASH, elevating the acetyl-CoA/CoASH ratio [70], which serves as a negative feedback signal to further suppress acetyl-CoA content and FAO. These findings integrate CTMP into the broader metabolic regulatory network, linking its thioesterase activity to acetyl-CoA homeostasis and FAO efficiency. The metabolic functions of CTMP prompted us to investigate its broader regulatory network. Central to this effort, we identified a previously uncharacterized interaction between CTMP and REV7, a mitotic checkpoint component. In MSI-L/MSS CRC—where chromosomal instability (CIN) is a constitutive hallmark [36]—our data show that REV7 overexpression enhances cGAS signaling while REV7-KD conversely suppresses it (Fig. 4J). Mechanistically, REV7 enhances MHC-I expression through at least three convergent pathways: suppression of CTMP–AKT signaling (Fig. 5), modulation of CDK1 activity, and direct regulation of cGAS signaling (Fig. 4). From a translational perspective, the REV7–CDK1 axis represents an attractive strategy for treating MSI-L/MSS CRC. High baseline CIN, further amplified by REV7 overexpression, may cooperate with CDK1 inhibition (Fig. 4B) to activate cGAS signaling, enhance MHC-I presentation (Fig. 4I). Unlike direct CDK1 inhibition—which faces clinical barriers including on-target toxicity in proliferating normal tissues and limited tumor selectivity—modulating the REV7 node achieves coordinated cell cycle arrest and immune reactivation. However, given REV7’s dual role in DNA repair, its overexpression necessitates rigorous safety assessment—including genomic stability assays, DNA repair functional tests, and long-term in vivo tumorigenicity studies—to evaluate potential risks of genomic instability and chemotherapy resistance before clinical application. These findings collectively position the CTMP–REV7 axis as a previously unrecognized immune-metabolic checkpoint linking mitotic stress to immunogenic activation in CIN-high colorectal cancer. The functional significance of the CTMP–REV7 axis is further contextualized within IFN-γ signaling, which exhibits a well-established dichotomy: it augments MHC-I yet induces immunosuppressive PD-L1 via JAK-STAT1 and PI3K-AKT [71]. In CRC, chronic PD-L1 overexpression predicts poor survival across MSI subtypes (Figure S2C), making baseline PD-L1 reduction a rational strategy for non-responsive MSI-L/MSS tumors. As noted above, CTMP knockdown directly counteracts this antagonism by inhibiting AKT/PD-L1 while enhancing HLA-B (Fig. 2C), thereby reprogramming IFN-γ into a unified antitumor response. Notably, the CTMP–REV7 interaction responds to IFN-γ stimulation (Fig. 5A), and even in its presence, modulating CTMP or REV7 significantly affects downstream CDK1–AKT signaling, underscoring the pathway’s robustness. In summary, this study identifies CTMP as a prognostic biomarker and therapeutic target in MSI-L/MSS CRC. Mechanistically, CTMP drives immune evasion via AKT/PD-L1 and MHC-I modulation through a novel CTMP–REV7 axis that integrates metabolism, cell cycle control, and antigen presentation. Disruption of this axis enhances MHC-I even under IFN-γ stimulation, offering a strategy to combat immune evasion in immunologically “cold” MSI-L/MSS colorectal cancer. Electronic Supplementary Material Below is the link to the electronic supplementary material. Supplementary Figure: Figure S1 Supplementary Figure: Figure S2 Supplementary Figure: Figure S3 Supplementary Figure: Figure S4 Supplementary Figure: Figure S5 Supplementary Figure: Figure S6 Supplementary Table: Table S1 Supplementary Table: Table S2 Supplementary Table: Table S3 Supplementary Figure: qPCR of Figure 2G Supplementary Figure: qPCR of Figure 2D Supplementary Figure: qPCR of Figure 5G Supplementary Figure: qPCR of Figure 7D Supplementary Figure: qPCR of Figure 4I Supplementary Figure: qPCR of Figure S5D Supplementary Figure: qPCR of Figure 6F Supplementary Material: Original WB & gels Supplementary Figures S1–S6 legend Acknowledgements We thank Prof. Haowen Zhang (Institute of Raidiation Medicine, Soochow University), Dr. Zhiqing Liu (Department of Orthopedics, Shanghai Tenth People’s Hospital, Tongji University), Dr. Jinkai Feng (Naval Medical University), Dr. Xu Wang (Shanghai University of Traditional Chinese Medicine) for generously providing academic guidance for this study. Abbreviations Immune checkpoint inhibitor Microsatellite Instability-Low Microsatellite stable Colorectal cancer Colon adenocarcinoma Tumor-infiltrating lymphocytes Major histocompatibility complex class I Fatty acid β-oxidation Carboxyl-terminal modulator protein Cyclic GMP-AMP Synthase Chromosomal instability Author contributions Q.X., C.G., J.W. and R.H. conceived and designed the study; Z.L., X.Y., J.L., J.G., R.X. and Z.K. collected the clinical samples; C.G., R.H., X.Y., J.W. and J.Y. performed the in vivo and in vitro assays; R.H., Y.G. and Z.K. performed the data analysis; C.G. and J.W. wrote the main manuscript text; J.Y., Z.L. and X.Y. revised the manuscript. All authors reviewed and approved the final manuscript. Funding This study was funded by National Natural Science Foundation of China (T2495260); Shanghai Clinical Research Center for Cell Therapy (23J41900100). Data availability All data generated or analyzed during this study are included in this published article and its supplementary information files. Declarations Ethics approval and consent to participate All human studies involving CRC tissue samples were performed after obtaining written informed consent from each participant and with the approval of the Ethics Committee of Shanghai Tenth People’s Hospital (Approval No. 25KN307). All animal experiments were approved by the same Ethics Committee (Approval No. SHDSYY-2023-1740) and carried out in compliance with Tongji University’s guidelines for laboratory animal welfare. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Cheng Gu, Rong He and Xiaolan Yin contributed equally to this work. Contributor Information Zhuqing Liu, Email: liuzhuqing521@126.com. Qing Xu, Email: xuqingmd@tongji.edu.cn. References Associated Data Supplementary Materials Supplementary Figure: Figure S1 Supplementary Figure: Figure S2 Supplementary Figure: Figure S3 Supplementary Figure: Figure S4 Supplementary Figure: Figure S5 Supplementary Figure: Figure S6 Supplementary Table: Table S1 Supplementary Table: Table S2 Supplementary Table: Table S3 Supplementary Figure: qPCR of Figure 2G Supplementary Figure: qPCR of Figure 2D Supplementary Figure: qPCR of Figure 5G Supplementary Figure: qPCR of Figure 7D Supplementary Figure: qPCR of Figure 4I Supplementary Figure: qPCR of Figure S5D Supplementary Figure: qPCR of Figure 6F Supplementary Material: Original WB & gels Supplementary Figures S1–S6 legend Data Availability Statement All data generated or analyzed during this study are included in this published article and its supplementary information files.