Palmitic acid-triggered B7H3 palmitoylation promotes immune escape

Rao, Zejun; Huang, Changsheng; Wu, Qi; Li, Mao; Liu, Anyi; Zhu, Tong; Yang, Wangshuo; Yin, Lanlan; Zhu, Shengyu; She, Xiaowei; Yu, Chengxin; Liu, Lang; Li, Pengcheng; Bai, Yucong; Zhang, Dongjing; Xie, Tianci; Liu, Xiang; Liu, Lu; Xu, Feng; Wang, Guihua; Hu, Junbo; Luo, Xuelai

doi:10.1038/s41467-026-68525-x

Download PDF

Article
Open access
Published: 19 January 2026

Palmitic acid-triggered B7H3 palmitoylation promotes immune escape

Zejun Rao¹,
Changsheng Huang ORCID: orcid.org/0000-0002-9301-1635¹,
Qi Wu¹,
Mao Li ORCID: orcid.org/0009-0003-9502-8515¹,
Anyi Liu¹,
Tong Zhu¹,
Wangshuo Yang¹,
Lanlan Yin¹,
Shengyu Zhu ORCID: orcid.org/0000-0002-2371-1476¹,
Xiaowei She¹,
Chengxin Yu¹,
Lang Liu¹,
Pengcheng Li¹,
Yucong Bai¹,
Dongjing Zhang¹,
Tianci Xie¹,
Xiang Liu¹,
Lu Liu¹,
Feng Xu¹,
Guihua Wang ORCID: orcid.org/0000-0001-5423-9357^1,2,
Junbo Hu ORCID: orcid.org/0000-0002-3582-999X¹ &
…
Xuelai Luo¹

Nature Communications volume 17, Article number: 1810 (2026) Cite this article

7069 Accesses
4 Altmetric
Metrics details

Subjects

Abstract

Colorectal cancer responds poorly to immune checkpoint blockade in most patients with microsatellite-stable (MSS) tumors, highlighting the need for alternative targets. B7H3 (CD276) is an immune checkpoint protein that is frequently overexpressed in tumors, but how it is maintained at high protein levels is unclear. Here we show that palmitic acid (PA) promotes B7H3 palmitoylation by the palmitoyltransferase ZDHHC24 at cysteine 496. This modification prevents B7H3 from binding to sequestosome 1 (SQSTM1, also called p62), limiting autophagic degradation and stabilizing B7H3, which suppresses CD8⁺ T cell antitumor activity. Disrupting this pathway by mutating cysteine 496, or by deleting Zdhhc24 in a colitis-associated colorectal cancer (CAC) mouse model, enhances CD8⁺ T cell responses. We also develop a cell-penetrant peptide that blocks the ZDHHC24–B7H3 interaction, boosts antitumor immunity, and synergizes with blockade of programmed cell death protein 1 (PD-1). These findings identify B7H3 palmitoylation as a targetable metabolic–immune node for colorectal cancer immunotherapy.

Palmitoylation prevents B7-H4 lysosomal degradation sustaining tumor immune evasion

Article Open access 08 May 2025

Palmitic acid activates c-Myc via dual palmitoylation-dependent pathways to promote colon cancer

Article Open access 17 February 2026

Palmitoylation of GPX4 via the targetable ZDHHC8 determines ferroptosis sensitivity and antitumor immunity

Article 19 March 2025

Introduction

Colorectal cancer is the third most common cancer worldwide¹. Traditionally, surgery has been the primary treatment, but recent advances suggest that immunotherapy may offer durable and potentially curative outcomes². Immune checkpoints, such as programmed death-1 (PD-1), programmed death-ligand 1 (PD-L1), and cytotoxic T-lymphocyte–associated antigen 4 (CTLA-4), play crucial roles in maintaining self-tolerance and regulating antitumor immunity^3,4,5. These checkpoint pathways involve receptor–ligand interactions that deliver co-inhibitory signals to T cells, thereby suppressing their activation and effector function. Immune check point inhibitors (ICIs) targeting PD-1/PD-L1 are effective primarily in microsatellite instability-high (MSI-H) colorectal cancers, which constitute only about 15% of cases^6,7,8,9. In contrast, these treatments show limited success in the majority of colorectal cancers that are microsatellite-stable (MSS)^10,11. Therefore, it is imperative to conduct an intensive search for alternative immune checkpoints to apply in the treatment of MSS tumors.

The B7 family, comprising a group of structurally related transmembrane proteins, including B7-1, B7-2, B7-H1 (also known as PD-L1), B7-H3, and others, interacts with CD28 family receptors to either stimulate or suppress immune response^12,13,14. B7H3 (CD276) has recently emerged as a compelling immunotherapy target due to its role in suppressing cytotoxic T cell and natural killer cell activity^15,16,17,18. B7H3 is highly expressed in various malignancies, and its elevated expression is associated with poor prognosis and adverse clinical outcomes^19,20,21,22. Although B7H3 mRNA is widely detectable across various tissues and cell types, protein expression is comparatively limited, suggesting that posttranscriptional mechanisms tightly regulate its abundance²³. Posttranslational modifications (PTMs) of B7H3 remain underexplored, yet they are likely to play important roles in modulating its function in cancer. While phosphorylation and ubiquitination sites have been predicted, N-glycosylation plays a pivotal role in stabilizing B7H3 and ensuring its stable expression on the plasma membrane, thereby facilitating its immunosuppressive function^24,25,26. However, it remains mostly unknown whether other post-translational modifications apart from glycosylation contribute to regulating the expression or function of B7H3. Elucidating these unexplored modifications and their regulatory mechanisms will be critical for developing more effective B7H3-targeted therapies.

S-acylation (often referred to as S-palmitoylation or simply palmitoylation) is an important post-translational modification in which a thioester bond is formed between the sulfhydryl group of a cysteine residue and, most commonly, a 16-carbon palmitate^27,28,29,30. This PTM leads to an increase in the lipophilicity of the modified target protein, which can affect its stability, function, or subcellular trafficking^{31,32,33,34,35}. Palmitoylation is mainly catalyzed by a relatively conserved family of zinc-finger DHHC (Asp-His-His-Cys) type-containing (ZDHHC) protein acyl-transferases (PATs)^36,37. Accumulating evidence suggests that immune-related molecules such as PD-L1³⁸ and NOD1/2³⁹ can be palmitoylated, indicating that palmitoylation plays an important role in the immune response.

Protein palmitoylation is regulated not only by ZDHHC family enzymes but also by the availability of its active substrate, palmitoyl-coenzyme A. Palmitoyl-coenzyme A, the active metabolite of palmitic acid (PA), reportedly initiates palmitoylation modifications in a variety of proteins^40,41,42. PA is a naturally occurring saturated fatty acid that has been reported to play important roles in immune evasion, metabolism, and inflammation^43,44,45. However, whether PA contributes to the establishment of an immunosuppressive microenvironment in MSS colorectal cancer remains unclear.

In this work, we identify that the palmitoylation of B7H3 triggered by PA is a key driving factor in the formation of the immune-suppressive microenvironment in colorectal cancer. Mechanistically, we show that ZDHHC24 catalyzes B7H3 palmitoylation at cysteine 496 (C496), preventing its interaction with the autophagy receptor SQSTM1/p62 and blocking its autophagic degradation. Biologically, we demonstrate that palmitoylation of B7H3 C496 remodels the immune microenvironment and specifically promotes tumor immune escape by inhibiting the antitumor effects of CD8⁺ T cells. Based on these findings, we design a peptide (E-peptide) that binds competitively to ZDHHC24. Both in vivo and in vitro experiments demonstrate that E-peptide reduces the palmitoylation of B7H3, promotes its autophagic degradation, and ultimately enhances the anti-tumor activity of CD8⁺ T cells. Notably, combining E-peptide with anti-PD-1 therapy produces a synergistic effect, significantly improving therapeutic outcomes. These findings provide important insights into immunotherapy for colorectal cancer and identify B7H3 palmitoylation as a promising therapeutic target.

Results

PA-triggered B7H3 palmitoylation enhances B7H3 protein stability

To identify the key immune checkpoints regulating the tumor microenvironment in MSS colorectal cancer, we performed an immunohistochemical analysis of B7 family members using formalin-fixed paraffin-embedded (FFPE) tumor samples from both MSS and MSI-H colorectal cancer patients (cohort 1, Table S1). The results indicated that the B7H3 protein was markedly overexpressed in MSS tumors (Figs. 1A, B and S1A). As an immune checkpoint molecule, B7H3 is expressed not only on tumor cells but also on various immune cell populations. Immunofluorescence staining showed strong co-localization of B7H3 with the tumor marker epithelial cell adhesion molecule (EPCAM) in tumor tissues (Fig. S1B), indicating prominent expression in tumor cells. Flow cytometry further confirmed that more than 95% of tumor cells are B7H3–positive, compared to about 30% of immune cells (Fig. S1C, D) (cohort 2, Table S2). When normalized to 100,000 total cells, B7H3⁺ tumor cells accounted for approximately 90,000 (Fig. S1E), highlighting tumor cells as the predominant source of B7H3 in MSS tumors. Interestingly, subsequent qPCR analysis revealed no significant alterations in B7H3 mRNA levels in MSS samples compared to MSI-H samples (Fig. S1F) (cohort 3, Table S3), suggesting that PTMs likely play a critical role in regulating B7H3 expression specifically in MSS tumors. Since it has been reported that the FUT8-catalyzed glycosylation of B7H3 can stabilize B7H3 and ensure its stable expression on the plasma membrane²⁵, we next assessed FUT8 expression in MSS and MSI-H colorectal tumor samples via immunohistochemistry. However, no significant difference in FUT8 expression was observed between these groups (Fig. S1G, H). To better understand the underlying mechanisms, we performed RNA sequencing using fresh colon tumor samples (cohort 4, Table S4). Gene Ontology (GO) enrichment analysis revealed that several genes regulating metabolites, such as carboxylic acids, fatty acids, and vitamins, were aberrantly expressed in MSS tumors (Fig. 1C and Supplementary Data 1). Therefore, we speculated that there were underlying metabolic abnormalities in these two types of tumors. We then collected fresh CRC samples for nontargeted metabolite analysis (Figs. 1D and S1I) (cohort 5, Table S5 and Supplementary Data 2) and conducted Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis of the metabolomics data (Fig. 1E). For the metabolites increased in MSS CRC, we observed the enrichment of pathways related to fatty acid degradation (Fig. 1F). Notably, fatty acid metabolism-associated pathways were also enriched in the GSEA results of the RNA-seq data (Fig. S1J–L). The findings suggest that abnormalities in fatty acid metabolism may be characteristic of MSS tumors. On the basis of those findings, we treated RKO cells with five metabolites associated with fatty acid degradation—glutaric acid, palmitoyl carnitine, hexadecanal, Palmitic acid (PA), and coenzyme A—identified in our metabolomic analysis. Among these, only PA significantly increased B7H3 expression (Fig. 1G, H). To further confirm this observation, we performed dose-dependent experiments in RKO cells and found that increasing PA concentrations led to a concentration-dependent increase in B7H3 protein levels (Fig. S1M). Since PA has been widely reported to trigger palmitoylation, we employed an acyl-biotin exchange (ABE) assay⁴⁶ to determine whether B7H3 undergoes palmitoylation (Fig. S1N). Our results demonstrated that B7H3 palmitoylation levels increased in response to higher concentrations of PA (Fig. 1I). Moreover, treatment with 2-bromopalmitate (2-BP), a pan-palmitoylation inhibitor, significantly reduced B7H3 palmitoylation (Figs. 1J and S1O). Like that of another member of the B7 family, B7H1 (PD-L1)³⁸, the protein level of B7H3 was closely related to its palmitoylation status (Fig. S1P). Notably, the qPCR results indicated that neither 2-BP nor palmostatin B (an inhibitor of depalmitoylation) affected B7H3 transcript levels (Fig. S1Q). When 2-BP was applied to inhibit palmitoylation, we observed a dose- and time-dependent decrease in B7H3 expression, whereas palmostatin B treatment led to the opposite effect (Figs. 1K and S1R). Given that palmitoylation can also affect protein membrane association, we conducted cytoplasmic and membrane fractionation assays, which revealed a reduction in B7H3 expression within both the cytoplasmic and membrane fractions following 2-BP treatment (Fig. S1S). Immunofluorescence analysis confirmed this decrease in B7H3 expression (Fig. S1T). Additionally, using a cycloheximide (CHX) chase assay, we assessed the impact of B7H3 palmitoylation on protein stability. Treatment with 2-BP markedly accelerated B7H3 degradation (Fig. 1L, M), whereas inhibition of depalmitoylation with palmostatin B slowed down B7H3 degradation (Fig. S1U, V). Taken together, these findings suggest that the PA-triggered palmitoylation contributes to B7H3 stability.

B7H3 C496 palmitoylation inhibits SQSTM1/p62-mediated autophagic degradation

B7H3, as a membrane protein, can in principle be degraded via the ubiquitin–proteasome system (UPS), autophagy–lysosome pathway, or endocytosis to lysosomes^47,48. We hypothesized that palmitoylation primarily modulates one of these routes to stabilize B7H3. Therefore, we applied pathway-specific modulators together with 2-BP. Pharmacological agents that disrupt distinct stages of the autophagy–lysosome pathway-3-MA (which inhibits class III PI3K and blocks autophagosome initiation), bafilomycin A1(Baf A1) and chloroquine (both block lysosomal degradation)—effectively blocked the 2-BP-induced degradation of B7H3, whereas the proteasome inhibitor MG132 had no such effect (Fig. 2A, B). Moreover, autophagy induction via EBSS significantly accelerated B7H3 degradation, which was further enhanced by 2-BP (Fig. S2A, B). Together, these observations indicate that B7H3 degradation in this system is predominantly mediated by the autophagy–lysosome pathway. The essential autophagy proteins ATG5 and BECN1 are required for autophagosome formation, and B7H3 degradation was markedly reduced in ATG5- or BECN1-knockout (KO) cells (Figs. 2C, D and S2C, D). Impaired autophagy also counteracted the 2-BP-induced reduction in B7H3 expression (Fig. S2E, F). Selective autophagy relies on cargo receptors that transport specific substrates to autophagosomes for degradation^49,50. Among the several cargo receptors tested (SQSTM1, OPTN, NDP52, NBR1, TOLLIP, and NIX), B7H3 exhibited a strong affinity for SQSTM1, as shown by immunoprecipitation assays (Fig. 2E). Additionally, cycloheximide (CHX) chase assays revealed that B7H3 degradation was nearly completely halted in RKO cells lacking SQSTM1 (Fig. 2F, G) and that SQSTM1 knockout prevented the 2-BP-induced degradation of B7H3 (Fig. 2H). These findings collectively demonstrate that B7H3 degradation occurs primarily through SQSTM1-mediated selective autophagy in the cell models examined. To identify the specific palmitoylation site on the B7H3 protein, we examined all 13 of its cysteine residues, five of which were predicted to be palmitoylated using GPS-palm⁵¹ and GPS-lipid⁵² (Fig. 2I). Analysis with the Swiss-Palm database, based on known palmitoyl-proteomes, also revealed a palmitoylation site (Fig. 2I). To validate these predictions experimentally, we generated site-specific cysteine-to-alanine mutants and assessed their palmitoylation levels using ABE assays. Among the mutants, the palmitoylation of C496A was almost undetectable compared with that of wild-type B7H3 (Fig. 2J), highlighting the essential role of C496. C496 is located in the intracellular domain of B7H3 (Fig. S2G) and is highly conserved across species based on multiple sequence alignment analysis performed in Jalview⁵³ (Fig. 2K). Consistently, the C496A mutation at the B7H3 palmitoylation site also showed greater affinity for the SQSTM1/p62 cargo receptor than WT B7H3 (Fig. 2L). To explore the functional role of C496 palmitoylation, we utilized CRISPR/Cas9 to generate B7H3-knockout RKO cells (RKO^KO) (Fig. S2H) and subsequently established stable cell lines expressing the C496A mutant (RKO^C496A) (Fig. S2I). ABE assays confirmed that B7H3 palmitoylation was scarcely detectable in RKO^C496A cells compared with that in wild-type (RKO^WT) cells (Fig. 2M). A CHX assay revealed that B7H3 protein degradation was accelerated in RKO^C496A cells (Fig. 2N, O), similar to the effects of palmitoylation inhibition by 2-BP. Immunofluorescence staining further demonstrated a marked decrease in fluorescence intensity in the RKO^C496A cells (Fig. S2J). In addition, we found that treatment with rapamycin, which inhibits mTORC1 activity and thereby induces autophagy, led to accelerated degradation of B7H3 protein in RKO^WT cells (Fig. S2K). In contrast, treatment with 3-MA, bafilomycin A1, or chloroquine effectively prevented B7H3 degradation in RKO^C496A cells (Fig. S2L). Moreover, confocal microscopy revealed that RKO^C496A cells displayed greater colocalization of B7H3 with LC3 and formed more distinct B7H3–SQSTM1/p62 aggregates compared with RKO^WT cells (Figs. 2P and S2M), indicating that C496 palmitoylation restrains the autophagic degradation of B7H3. In conclusion, these findings demonstrate that the palmitoylation of B7H3 at cysteine 496 reduces its interaction with the cargo receptor SQSTM1/p62, thereby hindering its selective autophagic degradation.

**Fig. 2: B7H3 C496 palmitoylation suppresses SQSTM1/p62-mediated autophagic degradation.**

B7H3 palmitoylation inhibits the antitumor activity of CD8⁺ T cells

To investigate the function of B7H3 palmitoylation in vivo, we generated B7h3-knockout CT26 (CT26^KO) and MC38 (MC38^KO) cell lines of mouse colon cancer (Fig. S3A). Since murine B7H3 comprises only a single pair of IgC-like and IgV-like domains²⁴, we engineered C278A mutants (CT26^C278A and MC38^C278A) that stably express palmitoylation-deficient B7H3, corresponding to the human modification site (Fig. S3B). Palmitoylation in these mutants was barely detectable, as confirmed by ABE assays (Fig. S3C). We then transplanted these modified mouse colon cancer cells subcutaneously into mice. In immunocompetent BALB/c mice, tumor progression was significantly inhibited in the palmitoylation-deficient B7H3 (CT26^C278A) group compared with the CT26^WT group (Figs. 3A, B and S3D). However, in highly immunodeficient NCG (NOD ShiLtJ Gpt Prkdc^em26Cd52 Il2rg^em26Cd22/Gpt) mice, no significant differences in tumor growth were observed between CT26^WT and CT26^C278A, as evidenced by comparable tumor weights and volumes (Figs. 3C, D and S3E). Similar results were obtained with the MC38 model, where there was no significant difference between MC38^WT and MC38^C278A in NCG mice (Fig. S3F–H). These findings suggest that B7H3 palmitoylation does not affect the proliferation of tumor cells directly; rather, the antitumor effects of B7H3 palmitoylation deficiency depend on the host immune system. To investigate the impact of B7H3 palmitoylation on the tumor microenvironment (TME), we conducted a comprehensive analysis of CD45⁺ immune cells infiltrating two types of CT26 tumors in BALB/c mice using spectral flow cytometry. In summary, we identified 14 unique cell subpopulations (Fig. 3E, F). The results revealed that some immunosuppressive cells, including monocytic myeloid-derived suppressor cells (M-MDSCs) and polymorphonuclear myeloid-derived suppressor cells (PMN-MDSCs), were reduced in palmitoylation-deficient (C278A) tumors (Figs. 3G and S3I). In contrast, some cell populations with antitumor effects, including CD8⁺ T, CD4⁺ T, and natural killer (NK) cells, were increased in C278A tumors (Figs. 3G and S3I). These results indicate that the mutation of B7H3 palmitoylation substantially alters the immune landscape within the TME, particularly by promoting CD8⁺ T cell infiltration (Figs. 3G, H and S3I). To test whether CD8⁺ T cells are essential mediators of antitumor effects caused by B7H3 palmitoylation deficiency, we depleted CD8⁺ T cells in vivo using an anti-CD8α antibody (Fig. 3I). Strikingly, depletion of CD8⁺ T cells completely abrogated the tumor growth difference between wild-type B7H3 and the palmitoylation-deficient C278A mutant (Figs. 3J, K and S3J), indicating that the antitumor effect associated with loss of B7H3 palmitoylation is primarily dependent on CD8⁺ T cells. To validate these findings, we subcutaneously transplanted the MC38 B7h3 knockout cell line (MC38^KO) and the stable MC38^WT and MC38^C278A cell lines into C57BL/6J mice. The results revealed that the mutation of the palmitoylation site significantly suppressed tumor growth, an effect that was even more pronounced in the B7h3 knockout (MC38^KO) group (Figs. 3L, M and S3K). An analysis of CD8⁺ T cell infiltration revealed that the MC38^C278A group presented a significantly greater proportion of CD8⁺ T cells than the MC38^WT group did, a trend that was further amplified in MC38^KO tumors (Figs. 3N and S3L). Immunofluorescence staining of tumor tissues corroborated those findings (Fig. 3O, P). Furthermore, the expression of effector molecules of CD8⁺ T cells, such as granzyme B (GZMB) and interferon-gamma (IFN-γ), was higher in MC38^C278A tumors compared with MC38^WT tumors (Figs. 3Q, R and S3M, N). To confirm these observations, we performed a T cell-mediated tumor-killing assay by coculturing activated human primary T cells with the RKO^KO, RKO^WT, and RKO^C496A cell lines. The RKO^C496A mutation markedly enhanced the cytotoxicity of T cells to tumor cells (Figs. 3S and S3O). Consistent with the flow cytometry findings, the lactate dehydrogenase (LDH) cytotoxicity assay yielded similar results (Fig. 3T). Finally, using tissue microarrays of colorectal cancer samples, we examined the correlation between B7H3 expression and CD8⁺ T cell infiltration. Patients with high B7H3 expression had shorter overall survival and significantly fewer CD8⁺ T cells than did those with low B7H3 expression (Fig. S3P, Q). Consistently, Kaplan–Meier Plotter analysis further confirmed that high B7H3 expression is associated with poorer overall survival (Fig. S3R), and TIMER database analysis showed that high B7H3 (CD276) expression correlated with lower CD8⁺ T cell infiltration (Fig. S3S). Furthermore, gene set enrichment analysis (GSEA) based on RNA-seq data from colorectal tumor samples (cohort 4, Fig. 1C) revealed that the gene set negative regulation of T cell activation was significantly enriched in MSS tumors (Fig. S3T). In contrast, pathways related to T cell function, including CD8-positive, alpha-beta T cell differentiation and the T cell receptor signaling pathway, were negatively enriched in MSS tumors (Fig. S3U, V), suggesting impaired cytotoxic T cell functionality. Collectively, these findings suggest that B7H3 palmitoylation enhances its immunosuppressive function by limiting CD8⁺ T cell infiltration and activity, thereby contributing to tumor immune evasion.

ZDHHC24 mediates the palmitoylation of B7H3 at C496 and is correlated with poor survival in CRC patients

To identify which ZDHHC(s) catalyze B7H3 palmitoylation, we co-expressed each of the 23 FLAG-tagged ZDHHC enzymes with HA-B7H3. ZDHHC24 significantly increased the palmitoylation level of B7H3, whereas ZDHHC3, ZDHHC8, and ZDHHC13 only slightly increased the B7H3 palmitoylation level (Fig. S4A). Subsequent transfection of RKO cells with ZDHHC3, ZDHHC8, ZDHHC13, or ZDHHC24 indicated that only ZDHHC24 substantially increased B7H3 expression (Fig. S4B–E). Those findings led us to hypothesize that ZDHHC24 is the primary enzyme responsible for B7H3 palmitoylation. To confirm the interaction between B7H3 and ZDHHC24, we conducted coimmunoprecipitation (co-IP) assays, which revealed a physical interaction at the endogenous level in RKO and SW480 cells (Fig. 4A). Consistent with these findings, we observed the colocalization of B7H3 and ZDHHC24 through immunofluorescence (Fig. 4B). To assess the functional role of ZDHHC24, we used short hairpin RNA (shRNA) to stably knock down ZDHHC24 in human RKO cells and Zdhhc24 in murine MC38 cells (Fig. S4F, G). ABE assays subsequently revealed a significant decrease in B7H3 palmitoylation levels following ZDHHC24 or Zdhhc24 knockdown (Figs. 4C and S4H). In line with the above results, ZDHHC24 or Zdhhc24 knockdown also led to a reduction in endogenous B7H3 protein expression (Fig. 4D, E), an effect further validated by confocal microscopy (Fig. 4F). Notably, the transcript levels of B7H3 showed minimal changes following ZDHHC24 knockdown (Fig. S4I). To confirm this, we performed a CHX assay and found that ZDHHC24 knockdown accelerated the degradation of B7H3 protein (Fig. 4G, H). Additionally, when Flag-ZDHHC24 was transfected into RKO^WT and RKO^C496A cells, an increase in endogenous B7H3 expression was observed exclusively in RKO^WT cells but not in palmitoylation-deficient RKO^C496A cells (Fig. 4I). Moreover, we assessed the palmitoylation levels of RKO^WT and RKO^C496A following ZDHHC24 transfection. The results indicated that ZDHHC24 increased palmitoylation of RKO^WT, but not that of the palmitoylation site-deficient RKO^C496A variant (Fig. 4J). Interestingly, co-IP experiments also revealed that in the palmitoylation-deficient mutant RKO^C496A, B7H3 exhibited a stronger interaction with SQSTM1/p62 proteins compared to RKO^WT, and that ZDHHC24 reduced this interaction in RKO^WT but not in RKO^C496A cells (Fig. 4K). In addition, ZDHHC24 knockdown promoted punctate accumulation and colocalization of B7H3 and p62, as demonstrated by immunofluorescence (Fig. 4L). This enhanced interaction was further confirmed by proximity ligation assay (PLA) (Fig. 4M). An analysis of the Gene Expression Profiling Interactive Analysis (GEPIA) database revealed substantial ZDHHC24 overexpression in multiple tumor types, particularly in colorectal cancer (Fig. S4J). Tissue microarray analysis of colorectal cancer specimens revealed that patients with high ZDHHC24 expression had shorter overall survival (Fig. 4N), a finding further supported by Kaplan–Meier Plotter database analysis, which linked high ZDHHC24 expression to worse relapse-free survival in colon cancer (Fig. S4K). Moreover, ZDHHC24 expression was markedly higher in tumor tissues than in adjacent normal tissues and was positively correlated with clinical tumor stage (Fig. 4O, P). It is worth noting that no significant differences in B7H3 or ZDHHC24 protein levels were detected between MSS and MSI-H subtypes in colorectal cancer cell lines (Fig. S4L). However, patient-derived tumor specimens showed that ZDHHC24 expression was significantly higher in MSS tumors than in MSI-H tumors (Fig. S4M). This suggests that the in vivo tumor microenvironment, which is absent in standard cell culture systems, may modulate ZDHHC24 expression. Importantly, ZDHHC24 expression was positively correlated with B7H3 levels in tissue microarray samples (Fig. 4Q). Together, these findings highlight the critical role of ZDHHC24-mediated B7H3 C496 palmitoylation in maintaining B7H3 protein stability and underscore its association with poor survival outcomes in patients with colorectal cancer.

**Fig. 4: ZDHHC24-mediated B7H3 palmitoylation promotes protein stability and is linked to poor colorectal cancer outcomes.**

ZDHHC24 promotes immunosuppression in colorectal cancer

To examine the biological function of ZDHHC24 in vivo, we generated Zdhhc24 floxed (Zdhhc24^fl/fl) mice (Fig. S5A) and crossed them with Villin-Cre transgenic animals (Fig. S5B) to generate intestinal epithelial cell (IEC) specific Zdhhc24-deficient (Zdhhc24^ΔIEC) models (Fig. S5C). The efficacy of Cre recombinase was confirmed by crossing the Villin-Cre line with Rosa26-CAG-LSL-tdTomato reporter (Fig. S5D) mice. As expected, intestinal epithelial cells presented specific red fluorescence under the microscope (Fig. S5E). We further isolated IECs to verify the specific deletion of Zdhhc24 in Zdhhc24^ΔIEC mice (Fig. S5F). Subsequently, we investigated the role of ZDHHC24 in CRC malignancy by establishing a colitis-associated colorectal (CAC) carcinogenesis mouse model (Fig. 5A). During CAC induction, Zdhhc24^ΔIEC mice presented less weight loss than Zdhhc24^fl/fl mice did (Fig. 5B). At the end of the observation period, the Zdhhc24^fl/fl mice had slightly shorter colons than their Zdhhc24^ΔIEC littermates did (Figs. 5C and S5G). Additionally, Zdhhc24^ΔIEC mice developed fewer tumors (Fig. 5D, E) and presented significantly fewer large tumors (> ~1 mm) (Fig. 5F, G). We collected tumor tissues from Zdhhc24^ΔIEC and Zdhhc24^fl/fl and assessed B7H3 protein expression and palmitoylation levels. B7H3 protein expression and palmitoylation were both reduced in Zdhhc24^ΔIEC mice (Fig. S5H, I). Histological examination revealed that Zdhhc24^ΔIEC mice presented reduced colon thickening, inflammation, hyperplasia, and injury in the affected regions (Fig. 5H). Immunofluorescence analysis also revealed increased CD8⁺ T cell infiltration in tumors from Zdhhc24^ΔIEC mice (Fig. 5I). These findings suggest that ZDHHC24 modulates antitumor immunity during colon tumor development. To validate these results, we transplanted Zdhhc24-knockdown mouse colon cancer cells (MC38) subcutaneously into C57BL/6J mice. Tumor growth was markedly reduced in the Zdhhc24-knockdown group compared with the control group, as shown by tumor growth curves and final tumor weights (Figs. 5J, K and S5J). Flow cytometry of single-cell suspensions from these tumors revealed a substantial increase in CD8⁺ T cell infiltration in the Zdhhc24-knockdown group (Figs. 5L and S5K), an effect that was corroborated by immunofluorescence analysis (Fig. S5L, M). Furthermore, the levels of functional markers, including GZMB and IFN-γ, were significantly greater in tumor-infiltrating CD8⁺ T cells in the Zdhhc24-knockdown group than in those in the control group (Figs. 5M, N and S5N, O). We subsequently established an in vitro cytotoxic T-cell killing assay for validation. Compared with control cells, cells with shRNA-mediated ZDHHC24 knockdown displayed increased susceptibility to apoptosis when activated T cells derived from human peripheral blood mononuclear cells were exposed (Fig. S5P, Q). This finding was supported by increased LDH levels in the culture supernatants of the respective cell populations (Fig. S5R). Additionally, IHC analysis of CRC tissue microarrays revealed that patients with high ZDHHC24 expression had lower CD8⁺ T cell counts in tumor tissues (Fig. 5O). A three-dimensional scatter plot further illustrated the correlation among ZDHHC24, B7H3, and CD8⁺ T cells (Fig. 5P). Collectively, these results suggest that ZDHHC24 promotes immunosuppression in colorectal cancer by inhibiting CD8⁺ T cell-mediated antitumor responses.

Fig. 5: ZDHHC24 promotes tumor growth by inhibiting the antitumor activity of CD8+ T cells. — **Fig. 5: ZDHHC24 promotes tumor growth by inhibiting the antitumor activity of CD8⁺ T cells.**

B7H3 palmitoylation is necessary for the ZDHHC24-mediated suppression of tumor immunity

To further investigate the relationship between B7H3 palmitoylation and ZDHHC24 in modulating the tumor immune microenvironment, we stably overexpressed Zdhhc24 in MC38^WT and MC38^C278A cells (Fig. S6A) and subcutaneously implanted these cells into C57BL/6J mice. ZDHHC24 significantly promoted tumor growth in the MC38^WT group; however, this effect was nearly absent in the MC38^C278A group (Figs. 6A, B and S6B). An analysis of immune cell infiltration revealed a greater proportion of CD8⁺ T cells in MC38^C278A tumors, whereas ZDHHC24 overexpression reduced CD8⁺ T cell levels in MC38^WT tumors but had no effect on MC38^C278A tumors (Figs. 6C and S6C). These results were consistent with the immunofluorescence findings (Fig. 6D, E). Furthermore, CD8⁺ T cells expressing GZMB and IFN-γ were more common in MC38^C278A tumors than in MC38^WT tumors (Fig. 6F, H). While ZDHHC24 decreased the levels of GZMB- and IFN-γ-expressing CD8⁺ T cells in MC38^WT tumors, it had no effect on MC38^C278A tumors (Fig. 6F–I). In vitro, cytotoxic T cell assays with RKO^WT and RKO^C496A cells overexpressing ZDHHC24 (Fig. S6D) demonstrated that ZDHHC24 did not affect apoptosis in the absence of T cells (Figs. 6J and S6E). However, in the presence of T cells, ZDHHC24 significantly reduced apoptosis in RKO^WT cells but not in RKO^C496A cells (Figs. 6J, K and S6E). Collectively, these findings confirm that ZDHHC24-mediated tumor immunosuppression depends on B7H3 palmitoylation.

Fig. 6: ZDHHC24 inhibition of antitumor effects on CD8+ T cells requires B7H3 palmitoylation. — **Fig. 6: ZDHHC24 inhibition of antitumor effects on CD8⁺ T cells requires B7H3 palmitoylation.**

Inhibition of B7H3 palmitoylation with synthetic peptides

Given the critical role of ZDHHC24-mediated palmitoylation of B7H3 at C496 in promoting an immunosuppressive microenvironment in colorectal cancer, we further investigated the therapeutic potential of targeting B7H3 palmitoylation. To better understand the binding interaction between ZDHHC24 and B7H3, we analyzed the structural domains of B7H3. As an immune checkpoint protein, the extracellular Ig-like domain of B7H3 likely contributes to antigen recognition and immunomodulation. Thus, we hypothesized that the primary binding sites for ZDHHC24 are located within the transmembrane and intracellular regions. To test this hypothesis, we generated B7H3 constructs with targeted deletions in specific structural domains and conducted immunoprecipitation with ZDHHC24 (Fig. 7A). Our results revealed that amino acids 488–497 in the B7H3 sequence are essential for its interaction with ZDHHC24 (Fig. 7B). Notably, this region is highly conserved in human and mouse and encompasses the palmitoylation site in B7H3, underscoring its functional relevance. We subsequently designed a peptide (E-peptide) corresponding to the B7H3 488–497 amino acid sequence, and a control peptide (C-peptide) in which the same amino acids were randomly scrambled to disrupt the sequence-specific structure and abolish potential binding, following a previously described strategy⁵⁴. Both peptides were fused with an N-terminal transcriptional activator of transcription (TAT) tag to facilitate cellular uptake (Fig. 7C). We assessed the cytotoxicity of the peptides in vitro and determined that cytotoxicity was observed only at concentrations exceeding 50 μM (Fig. S7A). We then attached a fluorescent motif to the first lysine of both the E-peptide and C-peptide (Fig. S7B) and confirmed its cellular entry via fluorescence microscopy (Fig. S7C). To evaluate the functionality of the synthesized peptides, we treated cells with E-peptide and C-peptide. In contrast to C-peptide, E-peptide significantly inhibited the binding of ZDHHC24 to B7H3, suggesting that E-peptide effectively competes with B7H3 for binding to ZDHHC24 (Fig. 7D). In parallel, we employed surface plasmon resonance (SPR) analysis to confirm the direct interaction between the E-peptide and ZDHHC24 (Fig. S7D, E). We further assessed B7H3 palmitoylation levels and found a significant reduction in palmitoylation with E-peptide treatment but not with C-peptide treatment (Fig. 7E). Notably, the decrease in B7H3 protein levels was dose dependent in RKO^WT cells treated with E-peptide, whereas neither peptide affected B7H3 expression in RKO^C496A cells (Fig. 7F). Additionally, E-peptide, but not C-peptide, markedly reduced B7H3 palmitoylation in RKO control cells, with limited effects in ZDHHC24 knockdown cells (Figs. 7G and S7F). Similarly, E-peptide treatment also reduced B7H3 protein levels in RKO control cells, but the decrease was not significant in ZDHHC24 knockdown cells (Fig. 7H, I). Moreover, co-immunoprecipitation analysis showed that E-peptide treatment enhanced the interaction between B7H3 and p62 (Fig. 7J). To determine the impact of synthesized peptides on immune function in vitro, RKO cells were treated with either E-peptide or C-peptide and subsequently coincubated with activated human CD8⁺ T cells. Both flow cytometry and lactate dehydrogenase toxicity assays demonstrated that E-peptide effectively enhanced CD8⁺ T cell killing (Figs. 7K, L and S7G). In conclusion, these data suggest that E-peptide inhibits B7H3 palmitoylation and promotes its degradation by disrupting the interaction between B7H3 and ZDHHC24, thereby effectively enhancing CD8⁺ T cell-mediated cytotoxicity in vitro.

Fig. 7: Targeting B7H3 C496 palmitoylation enhances CD8+ T effects in vitro. — **Fig. 7: Targeting B7H3 C496 palmitoylation enhances CD8⁺ T effects in vitro.**

E-peptide enhances antitumor immunity in vivo

To further evaluate the in vivo efficacy of the synthesized small-molecule peptides, we conducted a toxicity assessment. Mice received five different concentrations of peptides via intraperitoneal injection (Fig. S8A). Only one mouse died at each of the 25 and 50 mg/kg doses, while all others survived throughout the observation period (Fig. S8B). At the end of the experiment, the surviving mice were sacrificed, and peripheral blood and organ tissues were collected for analysis. Hematoxylin–eosin (HE) staining revealed mild hepatic steatosis at the 50 mg/kg dose, but no significant histological abnormalities were observed in other organs (Fig. S8C). Additionally, no notable changes were observed in the levels of creatinine (CRE), alanine aminotransferase (ALT) or aspartate aminotransferase (AST) in the peripheral blood (Fig. S8D–F), indicating minimal systemic toxicity at lower concentrations. To evaluate hemocompatibility, we performed in vitro hemolysis assays and measured prothrombin time (PT) and activated partial thromboplastin time (APTT). The E-peptide did not induce hemolysis (Fig. S8G, H) or prolong PT or APTT (Fig. S8I, J), demonstrating good blood compatibility under the tested conditions. Next, we subcutaneously transplanted MC38 cells into C57BL/6J mice and administered the peptides intraperitoneally (10 mg/kg) after transplantation, with normal saline serving as a control (Fig. 8A). To verify intratumoral peptide infiltration, a separate cohort received 5′-FITC-conjugated peptides via intraperitoneal injection. Immunofluorescence of tumor sections confirmed effective infiltration of the peptides into the tumor, establishing a spatial basis for their biological activity (Fig. S8K). We then assessed the effect of E-peptide treatment on B7H3 expression and its palmitoylation in tumor tissues. E-peptide administration significantly reduced both total B7H3 protein and its palmitoylation levels (Fig. S8L, M). Importantly, compared with NS and C-peptide, E-peptide significantly slowed tumor growth (Fig. 8B–D), and similar results were observed in BALB/c mice (Fig. 8E–G). Flow cytometry analysis of CD8⁺ T cells in BALB/c mice revealed a significant increase in CD8⁺ T cell infiltration within the tumor microenvironment following E-peptide treatment, whereas no such increase was observed with C-peptide (Figs. 8H and S8N). Moreover, the proportion of IFN-γ⁺ CD8⁺ T cells and GZMB⁺ CD8⁺ T cells—key indicators of tumor-killing activity—was significantly elevated after E-peptide treatment (Fig. 8I–L). To further evaluate the therapeutic efficacy of the E-peptide in a human immune context, we employed immune-reconstituted humanized mice bearing subcutaneous SW480 human colorectal tumors. Mice were treated with either the control C-peptide or E-peptide (Fig. 8M). E-peptide treatment significantly suppressed tumor growth (Fig. 8N–P), enhanced CD8⁺ T cell infiltration (Fig. 8Q, R), and increased the expression of cytotoxic effector molecules GZMB and IFN-γ in the tumor microenvironment (Figs. 8S–V). Together, these results demonstrate that the E-peptide effectively enhances CD8⁺ T cell-mediated antitumor immunity in vivo.

**Fig. 8: Synthetic peptides activate antitumor immunity in vivo.**

The combination of E-peptide and anti-PD-1 therapy exerts a significant antitumor effect

A single immune checkpoint inhibitor may be insufficient to address the complexities of the tumor immune microenvironment. Increasing evidence suggested that combination therapies involving immune checkpoint inhibitors yield better clinical outcomes than monotherapies do^55,56. Therefore, we hypothesized that E-peptide synergizes with PD-1 blockade, leading to more effective tumor suppression. We initially subcutaneously injected BALB/c mice with luciferase-expressing CT26 cells, followed by the intraperitoneal administration of anti-PD-1, E-peptide, or a combination of anti-PD-1 and E-peptide, with the IgG group serving as the control (Fig. S9A). The results demonstrated that monotherapies with either E-peptide or anti-PD-1 antibodies had moderate antitumor activity. However, the combination treatment significantly enhanced tumor inhibition, reducing the tumor size to a minimal level (Fig. 9A, B). To validate these findings, we expanded the sample size for each group and conducted a 60-day observation. The results confirmed that, compared to E-peptide or anti-PD-1 monotherapy, the combination therapy significantly inhibited tumor growth and even led to complete tumor regression in one mouse (Fig. 9C–G). Survival data (Fig. 9G) show a mean survival of 18.1 days in the IgG control, versus 24.4 days with anti-PD-1 ( + 34.8%) and 25.2 days with E-peptide (+39.2%). The combination of E-peptide and anti-PD-1 extended mean survival to 40.5 days (~124% over IgG). To assess interaction under the independent-action assumption, we estimated the Bliss-expected combination effect for E-peptide plus anti-PD-1 from their single-agent effects (defined as the normalized improvement in mean survival over IgG). This yielded an expected additive improvement of 60.4%, corresponding to a predicted mean survival of 29.0 days. The observed mean survival (40.5 days) exceeds the Bliss expectation, consistent with a synergistic interaction. Additionally, we validated the findings in C57BL/6J mice, which yielded results consistent with those observed in the BALB/c model (Figs. 9H, I and S9B). Moreover, flow cytometry analysis was performed to assess the impact of different treatments on tumor-infiltrating CD8⁺ T cells. The results demonstrated that the combination of E-peptide and anti-PD-1 antibodies significantly increased the number of CD8⁺ T cells and enhanced their functional activity (Figs. 9J–L and S9C–E). In summary, our findings emphasize the pivotal role of ZDHHC24-mediated B7H3 palmitoylation in promoting immune escape in colorectal cancer and demonstrate that targeting this modification could enhance antitumor immune responses, suggesting a promising strategy for improving the efficacy of immunotherapy in patients with colorectal cancer.

**Fig. 9: Combined blockade of B7H3 palmitoylation and PD-1 synergistically enhances antitumor effects.**

Discussion

Various mechanisms of immunosuppression and evasion of malignant tumor cells are considered hallmarks of malignancy⁵⁷. Immune checkpoint inhibitors, including anti-PD-1, anti-PD-L1, and anti-CTLA-4 antibodies, have shown durable efficacy in treating and potentially curing many types of tumors, ushering in a new era of tumor treatment⁵⁸. However, the current approval of immune checkpoint inhibitors for the treatment of colorectal cancer is limited to a subset of MSI-H tumors, suggesting the potential involvement of alternative immune checkpoints in regulating immune escape in MSS colorectal cancer⁷. B7H3, an emerging immune checkpoint, is highly expressed across various human malignancies but is expressed at low levels in normal tissues, making it a promising target for the development of future immune checkpoint inhibitors^24,59. Notably, despite low protein expression in normal tissues, B7H3 mRNA is detectable in a wide range of normal tissues and organs, suggesting the presence of significant posttranscriptional regulation²³.

Metabolites not only provide energy for life activities but also regulate the PTM of proteins through derivation into metabolically active products^60,61,62. PA, a long-chain saturated fatty acid, has been extensively reported as a metabolite that facilitates the crosstalk between metabolic processes and posttranslational protein modifications^41,42,63,64. Consistent with the B7 family immunohistochemistry results, we demonstrated that PA triggered the palmitoylation of B7H3, which played a crucial role in the formation of an immunosuppressive microenvironment in MSS colorectal cancer. Although our non-targeted metabolomic profiling revealed elevated levels of PA in MSS tumors compared to MSI-H tumors, the precise mechanisms underlying this difference remain unclear. However, several plausible explanations may account for this observation. First, fatty acid-binding protein (FABP1), a key transporter of long-chain fatty acids, is preferentially lost in MSI-H tumors due to IFN-γ-driven immune signaling, but retained in MSS tumors, potentially facilitating lipid accumulation⁶⁵. In addition, compared with MSS tumors, MSI-H tumors are enriched with metabolically active immune cells, which may consume fatty acids to support their effector functions⁶⁶. Furthermore, fatty acid β-oxidation is suppressed in MSS tumors, leading to reduced lipid degradation and a possible increased intracellular accumulation of saturated fatty acids⁶⁷. Notably, analyses of TCGA and GEO datasets also revealed that fatty acid-related metabolic pathways—including fatty acids, lipid localization, and lipid transport—are significantly downregulated in MSI-H colorectal cancers, consistent with our findings⁶⁶. In our study, we found that PA were elevated in MSS tumors and promoted palmitoylation of B7H3 at the C496 residue. This modification hindered the binding of B7H3 to the autophagy receptor SQSTM1/p62, thereby reducing its autophagic degradation and enhancing B7H3 protein stability. Moreover, we employed an exhaustive screening approach to identify ZDHHC24-a DHHC family member with a conserved DHHC Cys domain³²—as the enzyme that catalyzes B7H3 palmitoylation. Previously, ZDHHC24 has been shown to palmitoylate AKT, enabling membrane anchoring independent of PIP3⁴¹, and to palmitoylate MAVS, promoting its aggregation and mitochondrial localization⁴². In our study, we found that ZDHHC24 catalyzed the palmitoylation of the B7H3 protein and that its expression level was positively correlated with that of B7H3. Furthermore, the expression level of ZDHHC24 is positively associated with advanced cancer grade and serves as a prognostic biomarker for poor clinical outcomes in patients with colorectal cancer.

Multiple types of cancer are categorized as “cold” tumors, characterized by a low tumor mutational burden, minimal immune cell infiltration, and limited expression of PD-L1^11,68, which correlate with poor responses to PD-1/PD-L1 monotherapy⁶⁹. In addition, compensatory mechanisms between immune checkpoints may also limit the efficacy of monotherapy. In many tumors, inhibition of one pathway leads to upregulation of another. For example, Shi et al. found that inhibiting B7H3 leads to compensatory upregulation of PD-L1 on myeloid cells and increases Treg cell infiltration in tumors⁷⁰. Similarly, Altan et al. found that PD-L1-negative breast cancer can evade immune surveillance by upregulating B7H4, and the expression level of B7H4 is often negatively correlated with PD-L1 levels⁷¹. Such feedback emphasizes that single-agent blockade can be undermined by alternative suppressive pathways. Consequently, combination immunotherapies targeting multiple inhibitory pathways have been pursued. Indeed, combined checkpoint blockade has already proven effective: the anti-LAG-3 antibody Relatlimab plus Nivolumab (anti-PD-1) was FDA-approved for advanced melanoma⁷². In addition, in patients with colorectal cancer, adding a CTLA-4 inhibitor to PD-1 blockade helps improve long-term efficacy⁸. In our study, we found that B7H3 palmitoylation inhibited the antitumor activity of CD8⁺ T cells. We synthesized an E-peptide to block B7H3 C496 palmitoylation, which increased intratumoral CD8⁺ T cell infiltration and enhanced their expression of effector molecules such as GZMB and IFN-γ, thereby exerting notable antitumor effects. Anti-PD-1 therapy similarly functions at the effector phase of T cell response by blocking the PD-1/PD-L1 axis and reinvigorating exhausted T cells. Therefore, it is mechanistically reasonable to hypothesize that dual targeting of B7H3 palmitoylation and PD-1 could result in additive—or potentially synergistic—enhancement of CD8⁺ T cell function. Our experimental data support this notion, as we observed significantly improved antitumor efficacy with the combination of E-peptide and anti-PD-1 compared to either monotherapy. Beyond PD-1, other inhibitory receptors, such as CTLA-4, LAG-3, TIM-3, and TIGIT also play critical roles in modulating T cell responses. Prior studies have shown that combining anti-B7H3 with anti-CTLA-4, which primarily reduces Treg-mediated suppression, may be more effective than combinations involving PD-1 blockade in certain tumor contexts⁷⁰. Mechanistically, this may be explained by the distinct immune regulatory roles of CTLA-4 and PD-1: while anti-PD-1 relieves inhibition on already-activated effector T cells, CTLA-4 blockade preferentially reduces the suppressive activity of Tregs, which constitutively express high levels of CTLA-4 and play a central role in maintaining an immunosuppressive tumor microenvironment. Additionally, tumor-derived factors, such as VEGF can upregulate multiple immune checkpoint molecules (including PD-L1, CTLA-4, TIM-3, and LAG-3), suggesting that compensatory mechanisms among these pathways may limit the efficacy of monotherapies^73,74. Thus, combining B7H3 palmitoylation inhibition with agents targeting additional checkpoints, such as LAG-3, TIM-3, or TIGIT may further enhance antitumor immunity. Although we have not yet directly tested combinations of E-peptide with inhibitors of CTLA-4 or other checkpoints, our findings lay a mechanistic foundation for such future explorations. Ongoing and future studies will focus on evaluating such multi-checkpoint strategies in preclinical models to identify synergistic combinations capable of overcoming immune resistance.

In conclusion, our findings suggest that PA-triggered B7H3 palmitoylation may be a key contributor to the immunosuppressive microenvironment in MSS CRC. Specifically, ZDHHC24-induced B7H3 palmitoylation facilitates the ability of malignant cells to evade CD8⁺ T cell attack. We also showed that the pharmacological inhibition of B7H3 palmitoylation using a membrane-permeating peptide effectively enhances the antitumor effects of CD8⁺ T cells both in vitro and in vivo (Fig. 9M). We believe that this study provides valuable insights into the mechanisms of immune escape in colorectal cancer. Taken together, B7H3 palmitoylation represents a critical regulatory process that enhances its immunosuppressive effect, and targeting this modification may offer a promising therapeutic strategy to overcome immune resistance in cancers.

Methods

Cell lines and cell culture

The human colon cancer cell lines RKO and SW480, the human embryonic kidney cell line HEK293T, and the murine colon cancer cell lines MC38 and CT26 were purchased from the American Type Culture Collection (ATCC). RKO and HEK293T cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM; Gibco, USA), whereas SW480 cells were cultured in Leibovitz’s L-15 medium (Gibco, USA). MC38 and CT26 cells were cultured in RPMI-1640 medium (Gibco, USA). All culture media were supplemented with 10% fetal bovine serum (Thermo Fisher, USA). RKO, HEK293T, MC38, and CT26 cells were maintained at 37 °C in a humidified incubator with 5% CO₂, whereas SW480 cells were maintained at 37 °C in a CO₂-free atmosphere. All cell lines were routinely tested for mycoplasma contamination using a mycoplasma PCR detection kit (BMC1040, Abbkine).

Antibodies and chemicals

The following commercially available primary antibodies were used: anti-B7H3 (ab134161, abcam, [WB: 1:200]; ab227679, abcam [IF: 1:100]; 66481-1-Ig, proteintech, [WB: 1:1000; IHC: 1:2500]); anti-ZDHHC24 (ARP55989_P050, aviva systems biology, [WB: 1:1000]; HPA016619, Atlas antibodies, [IHC: 1:50]; and ABIN6743942, antibodies online, [WB: 1:1000]); anti-Flag tag (14793S, cell signaling technology, [WB: 1:1000]); anti-HA tag (3724T, cell signaling technology, [WB: 1:1000]); anti-GAPDH (sc-47724, santa cruz biotechnology, [WB: 1:200] and A19056, ABclonal, [WB: 1:5000]); caveolin-1 (A19006; ABclonal, [WB: 1:5000]); anti-PD-L1 (29122, cell signaling technology, [WB: 1:1000]); horseradish peroxidase (HRP)-streptavidin(A21000, abbkine, [WB: 1:10000]); anti-SQSTM1/p62 (HA721171, HUABIO, [WB: 1:5000; IHC: 1:1000]); anti-CD326/EpCAM (A25344, ABclonal, [IF: 1:200]); anti-ATG5 (ET1611-38, HUABIO, [WB: 1:5000]); anti-Beclin 1 (R381896, zen-bioscience, [WB: 1:1000]). The following secondary anti-bodies were used in Western blot assays: goat anti-rabbit IgG secondary antibody (HRP) (SSA004; Sino Biological Inc), goat anti-mouse IgG secondary antibody (HRP) (SSA007; Sino Biological Inc), HRP-conjugated goat anti-mouse IgG light chain (AS0-62; ABclonal), and HRP-conjugated mouse anti-rabbit IgG light chain (AS061; ABclonal). All Western blot secondary antibodies were used at a dilution of 1:5000. The secondary antibodies used for immunofluorescence (IF) assays were as follows: Cy3, goat anti-rabbit IgG (A22220, Abbkine); Cy3, goat anti-mouse IgG (A22210, Abbkine); D-light 488, goat anti-rabbit IgG (A23220, Abbkine) and DyLight 488, goat anti-mouse IgG (A23210, Abbkine); goat anti-human IgG H&L (FITC) (HY-P83649, MCE); goat anti-human IgG H&L (AF594) (HY-P83648, MCE). All IF secondary antibodies were used at a dilution of 1:200. Antibodies for flow cytometry were listed: Zombie NIR Fixable Viability Kit (423106, BioLegend), BD Pharmingen™ FITC mouse anti-human CD45 (560976, BD), BD OptiBuild™ BV421 mouse anti-human NCAM-1 (CD56) (744217, BD), BD OptiBuild™ BV510 mouse anti-human CD11b (742638, BD), BD OptiBuild™ BV650 mouse anti-human CD14 (740633, BD), PE anti-human CD68 antibody (333808, BioLegend), BD pharmingen™ PerCP-Cy™5.5 mouse anti-human CD3 (552852, BD), BD OptiBuild™ RB780 mouse anti-human CD276 (B7-H3) (755486, BD), BD OptiBuild™ R718 mouse anti-human HLA-DR (752501, BD), BD horizon™ BV605 mouse anti-human CD19 (569363, BD), BD OptiBuild™ BV786 mouse anti-human CD16 (741042, BD), BD pharmingen™ APC mouse anti-human CD326 (EpCAM) (566842, BD), BD OptiBuild™ BV650 Mouse Anti-Human CD8 (743067, BD), APC anti-human/mouse Granzyme B (372204, BioLegend), BD pharmingen™ PE mouse anti-human IFN-γ (554552, BD), brilliant violet (BV)785-conjugated anti-mouse CD45 antibody (103149, BioLegend), BV510-conjugated anti-mouse CD3 antibody (100234, BioLegend), BV785-conjugated anti-mouse CD3 antibody (100232, BioLegend), BV650-conjugated anti-mouse CD8 antibody (100742, BioLegend), PE/cyanine7-conjugated anti-human/mouse Granzyme B recombinant antibody (372214, BioLegend), PerCP/Cyanine5.5-conjugated anti-mouse IFN-γ antibody (505822, BioLegend), FITC-conjugated anti-active Caspase-3(559341, BD), PE hamster anti-mouse γδ T-cell receptor (553178, BD), Alexa Fluor™ 647 rat anti-mouse CD206 (568808, BD), PE-CF594 rat anti-mouse CD49b (562453, BD), BUV661 rat anti-mouse CD93 (741574, BD), BUV737 rat anti-CD11b (612800, BD), BV510 rat anti-mouse F4/80 (743280, BD), PE-Cy7 rat anti-mouse Ly-6G (560601, BD), R718 rat anti-mouse Ly-6C (566987, BD), BUV496 rat anti-mouse CD4 (741050, BD), BUV563 hamster anti-mouse CD3e (741235, BD), BUV805 rat anti-mouse CD45 (748370, BD), BV421 rat anti-mouse Foxp3 (562996, BD), V450 rat anti-mouse CD45R (560472, BD), BV711 rat anti-mouse CD86 (740688, BD), BV650 hamster anti-mouse CD11c (564079, BD), Zombie Yellow™ fixable viability kit (423104, BioLegend), FITC anti-mouse Ly-6G/Ly-6C (Gr-1) antibody (108406, BioLegend), PE/Fire™ 640 anti-mouse I-A/I-E antibody (107664, BioLegend), PE/cyanine7 anti-mouse CD8a antibody (100722, BioLegend). For flow cytometry staining, viability dyes were used at 1:500 (per 2 × 10⁶ cells in 100 μL), and all other fluorochrome-conjugated antibodies were used at 2.5 μL per 1 × 10⁶ cells in 100 μL. The reagents used as follows: phenylmethylsulfonyl fluoride (PMSF) (HY-B0496; MCE), phosphatase inhibitor cocktail I (HY-K0021; MCE), phosphatase inhibitor cocktail II (HY-K0022; MCE), chloroquine (HY-17589A; MCE) 2-bromohexadecanoic acid (HY-111770; MCE), palmostatin B (HY-120911; MCE), N-ethylmaleimide (04260; Millipore Sigma), hydroxylamine solution (467804; Millipore Sigma), 3-MA (HY-19312; MCE), MG-132 (HY-13259; MCE), cycloheximide (HY-12320; MCE), EZ-Link TM BMCC-biotin (21900; Thermo Fisher Scientific), NH₄Cl (HY-Y1269; MCE), anti-fade reagent (with DAPI) (HY-K1048; MCE), PMA (HY-18739; MCE), ionomycin (HY-13434A; MCE), monensin (HY-N0150; MCE), brefeldin A(HY-16592; MCE).

Plasmids

23 Flag-tagged ZDHHC family protein-encoding plasmids (n = 23) were obtained from MiaoLingBio, China (catalog numbers: P44768, P44768, P44770, P41465, P43454, P44771, P43506, P43467, P44775, P44767, P44779, P44777, P44769, P44783, P44773, P44780, P44390, P41826, P43407, P44772, P41821, P41823, P44-776). Expression vectors encoding pcDNA3.1-HA-B7H3, pcDNA3.1-Flag-SQSTM1/p62, pcDNA3.1-Flag-OPTN, pcDNA3.1-Flag-NDP52, pcDNA3.1-Flag-TOLLIP, pcDNA3.1-Flag-NIX and pcDNA3.1-Flag-NBR1 plasmids were constructed by insertion of the indicated DNA into the indicated vector. The pcDNA3.1-HA-B7H3-C496A was performed by using Mut Express II Fast Mutagenesis Kit V2 (C21401 Vazyme) according to the manufacturer’s instructions. The sequences are listed in Table S6 (Supplementary Information). All plasmids were confirmed by sequencing. Plasmids were first transformed into competent DH5α Escherichia coli (TSC-C14, Beijing Tsingke Biotech Co., Ltd.), and plasmid DNA for transfection was prepared using a plasmid miniprep kit (D6904, OMEGA) according to the manufacturer’s instructions. Cells cultured in dishes were starved with serum-free medium for 8 h. The indicated plasmids and transfection reagents (HighGene plus, RM09014P, abclonal) were then mixed well and waited for 20 min before being uniformly transferred to the culture dish. After 8 h, the complete medium containing serum was replaced and waited for 48 h.

RNA sequencing

Total RNA was isolated using TRIzol reagent (Invitrogen, CA, USA) according to the manufacturer’s protocol. The RNA’s purity and concentration were assessed via a NanoDrop 2000 spectrophotometer (Thermo Scientific, USA), while RNA integrity was confirmed using an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). For transcriptome sequencing, libraries were prepared using the VAHTS Universal V6 RNA-seq Library Prep Kit as per the manufacturer’s guidelines. Sequencing was conducted on an Illumina Novaseq 6000 platform, generating 150 bp paired-end reads. Initial raw reads in FASTQ format underwent quality filtering with fast to yield clean reads. Clean reads were aligned to the reference genome using HISAT2, with gene expression levels quantified in fragments per kilobase million (FPKM) and gene read counts obtained through HTSeq-count. Differential expression analysis was performed using DESeq2, identifying differentially expressed genes (DEGs) with a threshold of p < 0.05 and fold change ≥ 2; genes with baseman < 10 were excluded. Functional enrichment analysis of DEGs was performed using a hypergeometric test to identify significantly enriched Gene Ontology (GO) terms.

Non-targeted metabolomics analysis

Non-targeted LC–MS-based metabolomics was performed on 18 colorectal tumor samples (MSS, n = 8; MSI-H, n = 10), each representing an independent biological replicate. In addition, pooled quality-control (QC) samples were prepared by combining equal aliquots of all extracts and were injected at regular intervals to monitor instrument stability by principal component analysis (PCA). All tumor samples used for non-targeted metabolomic analysis were collected under strictly standardized conditions to ensure experimental consistency. Tissues were rapidly excised, snap-frozen in liquid nitrogen, and stored at –80 °C until analysis. To minimize batch-related variability, all samples were processed simultaneously and analyzed within a single LC–MS/MS run. This unified workflow effectively eliminated the need for additional batch effect correction in downstream data analysis. Weigh 30 mg of sample into a 1.5 mL EP tube, followed by the addition of two small steel beads and 400 μL of methanol-water (4:1, v/v) containing an internal standard (4 μg/mL). The sample was pre-cooled at −40 °C for 2 min, then ground at 60 Hz for 2 min. After grinding, the sample underwent ultrasonic extraction in an ice-water bath for 10 min and was then left to stand at −40 °C for 2 h. The mixture was centrifuged for 20 min at 15,000 × g and 4 °C. Subsequently, 150 μL of the supernatant was drawn with a syringe and transferred to an LC injection vial. Liquid chromatography–mass spectrometry (LC–MS) analysis was conducted using a Waters ACQUITY UPLC I-Class plus/Thermo QE HF ultra-high-performance liquid chromatography (UHPLC) tandem high-resolution mass spectrometer (HRMS) at OE Biotech, Inc., Shanghai, China. The system was equipped with a heated electrospray ionization (ESI) source (Thermo Fisher Scientific, Waltham, MA, USA) to analyze metabolic profiles in both positive and negative ESI modes. An ACQUITY UPLC HSS T3 column (1.8 μm, 2.1 × 100 mm) was used for both ionization modes; the column temperature was maintained at 45 °C, and the flow rate was 0.35 mL/min. Gradient elution was performed with solvent A (water with 0.1% formic acid, v/v) and solvent B (acetonitrile with 0.1% formic acid, v/v) using the following program: 0–2 min, 95% A/5% B; 2–4 min, linear to 70% A/30% B; 4–8 min, linear to 50% A/50% B; 8–10 min, linear to 20% A/80% B; 10–14 min, 0% A/100% B; 14–15 min, 0% A/100% B; 15.1–16 min, re-equilibration at 95% A/5% B. Samples were kept at 10 °C during analysis, with an injection volume of 4 μL. Detection was performed over a mass range of 70–1050 m/z, using a primary scan resolution of 60,000 and a secondary scan resolution of 15,000. Collision energies of 10, 20, and 40 eV were applied. The mass spectrometer’s operating conditions were as follows: spray voltage of 3800 V (positive mode) and 3200 V (negative mode); sheath gas flow rate of 35 units; auxiliary gas flow of 8 units; capillary temperature of 320 °C; auxiliary gas heater temperature of 350 °C; and S-lens RF level of 50. Data preprocessing was performed using Progenesis QI v3.0 (Nonlinear Dynamics, Newcastle, UK) for baseline correction, peak identification, integration, retention time alignment, and intensity normalization across samples. Following missing value imputation and filtering, the data were further normalized via log-transformation to approximate a Gaussian distribution and reduce heteroscedasticity, ensuring suitability for downstream statistical analysis. Compound identification was based on accurate mass-to-charge ratios (m/z), secondary fragments, and isotopic distributions, utilizing databases such as the Human Metabolome Database (HMDB), Lipidmaps (V2.3), Metlin, and self-built databases. Peaks with over 50% missing values (ionic intensity = 0) were excluded, and zero values were replaced by half the minimum value. Data from positive and negative ion modes were merged into a single data matrix for further analysis. Differential metabolites between MSS and MSI-H groups were selected using Student’s t test with P < 0.05 and fold change (FC) ≥ 1.2 or ≤ 1/1.2.

Construction of stable cell lines

The lentiviral plasmids pLVX-ZDHHC24 or pLVX-Zdhhc24 were constructed by cloning ZDHHC24 or Zdhhc24 genes into the pLVX vector. We used the pLVX, pMD2.G, and psPAX2 three-plasmid packaging systems to generate the highly expressed viruses. Short hairpin RNA (shRNA) targeting the human ZDHHC24 or murine Zdhhc24 genes was expressed in the pLKO.1 vector to obtain knockdown (KD) clones. The pLKO.1-shRNAs, pMD2.G, and psPAX2 triple packaging systems were then used to construct silencing expression lentiviruses. The sequences are listed in Table S7 (Supplementary Information). Similarly, the three of them were mixed in 500 μL of serum-free medium and 40 μl of Lipofectamine 2000 in a mass ratio of 2:1:2. The plasmid mixture was then transfected into HEK293T cells. Harvested viral supernatant was filtered through a sterilized polyethersulfone PES syringe filter (SFPES025022SI, membrane solutions LLC) and used to infect the indicated cells. Stabilize cell lines by screening with 1 μg/ml puromycin for 48 h or 0.8 mg/mL G418 for 20 days. The CRISPR/Cas9 system was used to generate B7H3, SQSTM1, ATG5, and BECN1 KO cells. The single guide RNAs were designed from a public website, CHOPCHOP, and then cloned into the PX459 plasmid (62988, Addgene) or the PX462 plasmid (62987, Addgene). The sgRNA sequences are listed in Table S7 (Supplementary Information).

Cellular membrane and cytoplasmic protein isolation

Separation of cytosolic and cytoplasmic proteins of the indicated cells was performed using the ExKine™ membrane and cytoplasmic protein extraction kit (KTP3005; Abbkine).

Cell counting kit-8 assay

To assess peptide cytotoxicity in vitro, cells were seeded into 96-well plates (100 μL complete medium) and allowed to adhere overnight. Cells were then treated with E-peptide or C-peptide at the indicated concentrations (2.5, 5, 10, 25, 50, 100, and 200 μM) for 24 h. Subsequently, 10 μL of cell counting kit-8 (C6005, new cell and molecular biotech) reagent was added to each well and incubated at 37 °C for 1–4 h protected from light. Absorbance was measured at 450 nm using a microplate reader. Wells containing medium plus reagent without cells served as blanks, and cell viability was normalized to the vehicle-treated control (set to 100%).

Mouse serum biochemistry

Peripheral blood was collected from mice and allowed to clot at room temperature for 30 min. Serum was separated by centrifugation at 4 °C at 3000 × g for 10 min, and analyzed immediately or stored at −80 °C until use. Serum creatinine (CRE) was measured using the creatinine (CRE) content assay kit (AKNM008M, Beijing Boxbio Science and Technology Co., Ltd). The microplate reader was pre-warmed for ≥30 min and set to 546 nm. Reactions were carried out at 37 °C according to the manufacturer’s protocol, absorbance was recorded, and CRE concentration was calculated against the kit standard/blank using the provided formula. Serum alanine aminotransferase (ALT/GPT) and aspartate aminotransferase (AST/GOT) activities were determined using the GPT activity assay kit (AKAM006M, Beijing Boxbio Science and Technology Co., Ltd) and the GOT activity assay kit (AKAM019M, Beijing Boxbio Science and Technology Co., Ltd), respectively. The instrument was pre-warmed for ≥30 min and set to 505 nm. A matched control and a standard curve were included as instructed. After incubation and color development at 37 °C, absorbance was measured, and enzyme activities were calculated and reported as U/L according to the kit-provided formula.

Immunoprecipitation

Different magnetic beads (HAK21042, HAK21011, HUABIO) were selected according to the experimental needs. If the magnetic beads (HY-K0202, MCE) did not have a coupled antibody, it was necessary to mix the beads with the corresponding primary antibody for half an hour at room temperature or 2 h at 4 °C before harvesting the cells. Subsequently, start lysing the cells to harvest the proteins. Proteins are extracted in NP40 cell lysis buffer containing the protease cocktail inhibitor for 30 min. Next, the magnetic beads that have been coupled to the antibody are transferred to the supernatant into the cell lysate and rotated at 4 °C overnight. The following day, the supernatant was discarded, and the magnetic beads were washed five times with 500 μL of NP40 cell lysis buffer. Between each wash, magnetic separation gently. Finally, the supernatant was discarded, and 40 μl SDS loaded buffer was mixed with magnetic beads and boiled in a water bath for 10 min before being used for Western blotting.

Protein extraction and Western blotting

For cultured cells, discard the medium and wash the cells three times with cold PBS. Cells were lysed for 30 min using NP-40 buffer containing phenylmethylsulfonyl fluoride (HY-B0496, MCE) and cocktail protease inhibitor (HY-K0010, MCE). For tissue samples, total protein was extracted using the Total Protein Extraction Kit (EX1100, Solarbio) according to the manufacturer’s instructions. Lysates were centrifuged at 15,000 × g for 15 min at 4 °C. Supernatants were collected, and protein concentrations were quantified by BCA kit (23227; Thermo Fisher). Subsequently, the protein supernatants acquired were mixed with 5× concentrated sample buffer containing 250 mM Tris-HCl (pH 6.8), 5% β-mercaptoethanol, 10% sodium dodecyl sulfate (SDS), 30% glycerol, and bromophenol blue, followed by heating at 95–100 °C for 10 min. Protein samples were separated by electrophoresis in 8–10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). After electrophoresis, the separated protein bands were transferred to a polyvinylidene fluoride (PVDF) membrane. Subsequently, the membrane was then blocked by immersing in 5% skimmed milk in TBST or rapid blocking buffer (SW162-01, Seven, Beijing, China). Next, the membranes were incubated overnight at 4 °C with primary antibodies and then exposed to HRP-linked secondary antibodies for 2 h at ambient temperature. Finally, the membrane bands were washed three times with TBST and then visualized using an efficient chemiluminescence kit (34096; Thermo Fisher).

Acyl-biotin exchange (ABE) assay

The ABE assay was performed based on established protocols^33,39,46, and the key experimental steps are detailed below. 50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1 mM MgCl2, 1% NP-40, 10% glycerol, protease inhibitor, and phosphatase inhibitor were added to NP-40 to configure the lysis buffer. Fresh N-ethylmaleimide (NEM) solution was prepared and added to the lysis buffer to a final concentration of 50 mM. The indicated cells were lysed in a buffer containing 50 mM NEM for 1 h on ice. Collect the cell suspension and mix it with magnetic beads precoated with the appropriate antibody for 3 h at 4°C. After 5 washes with lysis buffer (pH = 7.5), the beads were washed again 3 times with lysis buffer (pH = 7.2). Fresh hydroxylamine (HAM) solution was prepared in lysis buffer at pH 7.2 to a final concentration of 1 M. The beads were then incubated with lysis buffer containing 1 M HAM for 1 h at room temperature. Wash 3 times with Lysis Buffer (pH = 6.2) to remove residual HAM. Subsequently, the beads were incubated with lysis buffer (pH = 6.2) containing Biotin-BMCC (5 μM) at 4 °C for 1 h. The beads were washed twice with buffer (pH = 6.2) and then three times with lysis buffer (pH = 7.5). Excess buffer was removed and added to 50 μL sample buffer (5% SDS, 5% glycerol, 125 mM Tris-HCL (pH = 6.8), 0.01% bromophenol blue) containing 5 mM DL-Dithiothreitol (DTT). Samples were boiled at 80 °C for 10 min and then cooled to room temperature for 15 min.

Immunofluorescence assays

Cells were inoculated onto slides of 24-well culture plates, allowed to adhere to the wall, and then subjected to various treatments. The slides were rinsed three times with precooled PBS, followed by fixation in 4% paraformaldehyde for 20 min at room temperature. After removing the paraformaldehyde, the slides were washed three times with PBS. Added 0.3% Triton-X-100 to the wells and permeabilized for 15 min at room temperature. The slides were blocked with 3% BSA for 1 h at room temperature after washing three times with PBS. Subsequently, they were incubated with primary antibodies at 4 °C overnight. The following day, fluorochrome-labeled secondary antibodies were applied in the dark for 2 h at room temperature, and the samples were then stained with DAPI for 10 min. Tissue samples were treated similarly as described above. Finally, we obtained representative images using a Japanese Olympus FLUOVIEW FV1000 confocal scanning microscope or a Japanese Nikon ECLIPSE Ti2 fluorescence microscope.

Tumor-infiltrating lymphocyte isolation and flow cytometry analysis

For tumor-infiltrating immune cells, all mice were isolated immediately after sacrifice at the indicated times, and tumors were placed on ice and carefully cut into small pieces of approximately 1 mm. These fragmented tumor tissues were then digested in serum-free 1640 medium containing collagenase IV (1 mg/ml, 17104019; Invitrogen), deoxyribonuclease I (100 μg/ml, HY-108882; MCE), and hyaluronidase (250 μg/ml, BS171; Biosharp) for 90 min at 37 °C. The cell suspension was then filtered using a cell strainer. A Percoll gradient cell separation solution (17089101, Cytiva) was used to isolate tumor cells and immune cells infiltrating the tumor. For cytokine analysis, the obtained immune cells were incubated with PMA, Brefeldin A, monensin, and ionomycin for 6 h and then stained with the corresponding antibodies. Cytofix/Cytoperm™ fixation and permeabilization solution (554722, BD) was used for intracellular staining. Transcription factor buffer set (562574, BD) used for nuclear staining. Raw FCS files were analyzed using FlowJo software (v10.8.1).

T-cell killing assay

The experimental methodology was adapted from published protocols^75,76, and the main procedures are described below. Peripheral blood mononuclear cells (PBMCs) were isolated from anticoagulated whole blood using human lymphocyte separation medium (7111011, Dakewe). Briefly, whole blood was diluted 1:1 with PBS and carefully layered onto the separation medium. Samples were centrifuged at room temperature at 800 × g for 30 min using a horizontal rotor with low acceleration and deceleration settings. After centrifugation, the opaque mononuclear cell layer at the plasma-medium interface was carefully collected into a new tube. Cells were washed 2 times with RPMI-1640 by centrifugation at 20 °C and 250 × g for 10 min, and the cell pellet was resuspended in normal saline or appropriate culture medium for downstream isolation. Human CD8⁺ T cells and human DCs were isolated from PBMCs using the CD8⁺ T cell isolation kit, human (130-096-495, Miltenyi Biotec), and the Dynabeads® Human DC Enrichment Kit (11308D, Thermo Fisher Scientific), respectively. The isolated DCs were first co-incubated with the indicated tumor cells, and then the pretreated DCs were then used to activate CD8⁺ T cells. To evaluate the tumor-killing effect mediated by T cells, the preconditioned CD8⁺ T cells were reactivated with anti-CD3 antibody (2 μg/ml), anti-CD28 antibody (2 μg/ml), and IL-2 (100 IU/ml). The activated CD8⁺ T cells were then co-incubated with the indicated tumor cells at a ratio of 10:1 for 24–48 h. Detection of T-cell mediated cytotoxicity using caspase-3 cleavage assay and lactate dehydrogenase (LDH) release assay. After centrifugation, the supernatant of each well was collected, and the amount of LDH released in the culture medium was determined using the LDH Cytotoxicity Assay Kit (HY-K1090, MCE). Cell clusters obtained after centrifugation were washed three times with PBS and then stained with FITC-coupled activated caspase-3 antibody, followed by flow cytometric analysis.

Immunohistochemistry (IHC)

The embedded tissue specimens were cut into sections at a thickness of 4 μm. Following sequential processes of dewaxing, rehydration, antigen retrieval, and blocking, the sections were treated with the specified primary antibody overnight and then incubated with the secondary antibody for 1 h at room temperature. Immunohistochemical results were quantified by the density and intensity of positive cells. The density of positive staining was categorized into 4 grades based on the intensity of positive staining as follows: 1 (<5%), 2 (5–30%), 3 (30–70%), and 4 (>70%). The intensity of positive staining was categorized into 4 grades, 0 (unstained), 1 (weak), 2 (moderate), and 3 (strong). The density and intensity scores were then multiplied to obtain the IHC score, which ranged from 0 to 12. To evaluate the CD8 staining score, three random fields of view were chosen from each tissue section and analyzed at a magnification of ×400. The mean number of positively stained cells observed in these fields was calculated to represent each case. IHC staining was independently scored by two experienced pathologists, and the final IHC score for each case was defined as the average of the two scores.

Quantitative reverse transcription polymerase chain reaction (qRT-PCR)

RNA was initially extracted from cells using AG RNAex Pro RNA extraction reagent (AG21101, Accurate Biotechnology (Hunan) Co., Ltd) following the manufacturer’s protocol. The extracted RNA was then reverse-transcribed into cDNA using the Reverse Transcription Kit (HiScript III RT SuperMix for qPCR, Vazyme). Finally, qPCR was performed using the SupRealQ Ultra Hunter SYBR qPCR Master Mix(U+) (Q713-02, Vazyme Biotech Co., Ltd) reagent on QuantStudio 3 instrument (Thermo Fisher). All primer sequences are provided in Table S8.

Animal study

The mice used in this study (Mus musculus) included C57BL/6J, BALB/c, NCG, Zdhhc24 ^fl/fl, Villin-cre, and Rosa26-CAG-LSL-tdTomato, which were purchased from GemPharmatech (Nanjing, China). The Zdhhc24 ^fl/fl mouse line was generated using CRISPR/Cas9 on a C57BL/6J genetic background (gRNA sequences are provided in Table S9) and maintained in our facility. The huHSC-C-NKG-ProF mice were generated by Cyagen Biosciences (Suzhou, China). All animals used in this study were female. Unless otherwise indicated in the figure legends, mice were 6 weeks old at the start of each experiment, and the number of animals used for each experiment (n) and the corresponding strain/substrain are reported in the figure legends. Sex was not included as a biological variable because all experiments were performed using female mice to reduce variability and maintain consistency across cohorts; therefore, no sex-based comparisons were performed, and this choice was made solely for experimental design and reproducibility considerations and does not reflect any sex bias or discrimination. All animal experiments were conducted with approval from the Animal Care and Use Committee of Tongji Hospital (TJH-202303006). Mice were housed under specific pathogen-free conditions with a 12 h light/12 h dark cycle, an ambient temperature of 22–24 °C, and relative humidity of 40–60%, with food and water available ad libitum.

For xenograft tumor assays, cells were digested, washed three times with PBS, and mixed with Matrigel (abs9493, Absin) at a 1:1 ratio to a final density of 2 × 10⁶ cells/mL. A total of 100 μL of cell suspension was injected subcutaneously into 6-week-old female mice. Tumor volume was measured at the indicated time points and calculated as V = L × D²/2 (L: long axis, D: short axis). In the approved protocol, the maximal tumor burden was a tumor volume of 2000 mm³, and mice were euthanized when tumors reached this limit or when other humane endpoints were met. In addition to the predefined maximal tumor burden (tumor volume of 2000 mm³), mice were euthanized if any of the following humane endpoint criteria were met: (i) ulceration, necrosis, or bleeding at the tumor site; (ii) impaired mobility, inability to access food or water, or persistent abnormal posture/behavior; (iii) rapid weight loss (≥20% of baseline) or severe dehydration; (iv) signs of systemic distress such as labored breathing, hypothermia, or marked lethargy; or (v) any other condition judged by the animal care staff/veterinarian to require euthanasia to prevent undue suffering. Mice were monitored at least every 1–2 days, and more frequently when tumors approached the endpoint. Overall, the study did not substantially exceed the permitted limit. In the CT26 survival experiment (Fig. 9C–G), a small number of mice showed calculated tumor volumes slightly above 2000 mm³ at the scheduled measurement time points, reflecting caliper measurement variability and the estimated nature of the volume formula, together with rapid tumor growth between two consecutive monitoring time points; mice were euthanized immediately upon detection in accordance with the approved protocol. For therapeutic agents, anti-PD-1 antibody (10 mg/kg), E-peptide, and C-peptide (10 mg/kg) were injected into mice by intraperitoneal injection at the indicated time intervals. For Zdhhc24^fl/fl mice, CRISPR/Cas9 technology was used to modify the Zdhhc24 gene (gRNA sequences are provided in Table S9). Briefly, the CRISPR/Cas9 system and donor were microinjected into fertilized eggs of C57BL/6 J mice. Transplantation of fertilized eggs yielded positive F0 mice, which were confirmed by PCR and sequencing. Positive F0 generation mice were backcrossed with C57BL/6 J for at least two generations to obtain a stable Zdhhc24^fl/fl mouse model. Villin-cre and Rosa26-CAG-LSL-tdTomato are well-established commercial mice. Validation of PCR primers for Zdhhc24^fl/fl and Villin-cre mice is provided in Table S10.

Hemolysis assay

Approximately 0.6 mL of blood was collected from anesthetized mice via cardiac puncture into a centrifuge tube pre-coated with sodium citrate as an anticoagulant. The sample was centrifuged at 2500 × g for 10 min to remove the plasma. The remaining red blood cells (RBCs) were washed repeatedly with normal saline until the supernatant became colorless. After discarding the supernatant, the RBCs were resuspended in 1 mL of normal saline to obtain an RBC suspension. Then, 100 μL of RBC suspension was added to 1 mL of E-peptide solution (dissolved in saline) at various concentrations. Saline alone was used as a negative control, and distilled water was used as a positive control. The mixtures were incubated at 37 °C for 2 h, followed by centrifugation at 2500 × g for 10 min. The optical density (OD) of the supernatant was measured at 541 nm. The hemolysis rate was calculated using the following formula:

$${{{\mathrm{Hemolysis}}}}\,(\%)= ({OD}\_{{sample}}-{OD}\_{{negative}})/\\ ({OD}\_{{positive}}-{OD}\_{{negative}})\times 100\%$$

Coagulation assays: prothrombin time (PT) and activated partial thromboplastin time (APTT)

PT and activated partial thromboplastin time APTT were measured using commercial assay kits (Bioswamp Life Science Lab, Wuhan, China) according to the manufacturer’s instructions. Briefly, whole blood was centrifuged at 2500 × g for 10 min to obtain platelet-poor plasma. Two hundred microliters of plasma was mixed with 10 μL of E-peptide solution (dissolved in normal saline) to achieve final peptide concentrations of 5, 10, 25, and 50 μM. An equal volume of normal saline was added to the plasma as the control. For PT testing, 70 μL of peptide-treated or control plasma and 140 μL of PT reagent were pre-incubated separately in a 37 °C water bath for 3 min. The two components were then mixed rapidly, and the clotting time was recorded upon cessation of flow during gentle tilting of the reaction tube. For APTT testing, 100 μL of peptide-treated or control plasma was mixed with 100 μL of APTT reagent (containing ellagic acid) and incubated at 37 °C for 5 min. Subsequently, 100 μL of 25 mM CaCl₂ solution (pre-warmed to 37 °C) was added to initiate coagulation. The tube was gently agitated in a 37 °C water bath, and the time to visible fibrin formation was recorded as the APTT value.

Peptide synthesis

The C-peptide (YGRKKRRQRRRRKWEISQCKC) and E-peptide (YGRKKRRQ-RRRCWRKIKQSCE) were synthesized by QYAOBlO (ChinaPeptides Co., Ltd.) using standard Fmoc-based solid-phase peptide synthesis (SPPS) protocols on 2-chlorotrityl chloride resin. The synthesis followed a stepwise procedure involving resin swelling in DMF, sequential coupling of Fmoc-protected amino acids using HBTU and DIEA, and Fmoc deprotection with 20% piperidine in DMF. Coupling efficiency was monitored by Kaiser tests. Upon completion of chain assembly, the peptide was cleaved from the resin using a TFA-based cleavage cocktail (TFA/H₂O/EDT/TIS, 95:2:2:1, v/v) and precipitated in cold ether. The crude product was purified by preparative reversed-phase high-performance liquid chromatography (RP-HPLC) on a C18 column. Final peptide identity and purity were confirmed by mass spectrometry (MS) and analytical HPLC. The purified peptide was lyophilized and stored at −20 °C until use.

Proximity ligation assay (PLA)

Proximity ligation assay was detected using the Duolink® in situ red starter kit mouse/rabbit (Sigma-Aldrich) following the manufacturer’s protocol. Briefly, fixed and permeabilized samples on glass slides were blocked with Duolink® Blocking Solution at 37 °C for 60 min, then incubated overnight at 4 °C with primary antibodies: anti-CD276 (Abcam, ab134161) and anti-SQSTM1/p62 (CST, 88588). After washing in 1× Wash Buffer A, PLUS and MINUS PLA probes (1:5 dilution) were applied for 1 h at 37 °C, followed by ligation (1:40 ligase, 30 min at 37 °C) and amplification (1:80 polymerase, 100 min at 37 °C, in the dark). Slides were washed, mounted with Duolink® in situ mounting medium containing DAPI, and imaged using a confocal microscope (LSM900 Airyscan).

Surface plasmon resonance (SPR)

Binding measurements between ZDHHC24 and E-peptide were performed by surface plasmon resonance (SPR) on a Biacore 8 K instrument (GE Healthcare/Cytiva) using CM5 sensor chips, based on established protocols^77,78 with all key experimental parameters provided below. In brief, experiments were conducted at 25 °C with a blank activated/deactivated reference flow cell. ZDHHC24 was immobilized on CM5 via amine coupling using 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide/N-hydroxysuccinimide (EDC/NHS), and residual esters were blocked with ethanolamine. The running/immobilization buffer was 1× phosphate-buffered saline with surfactant P20 (PBS-P+, pH 7.4). Interaction analyses were carried out in PBS-P+ (pH 7.4) supplemented with 5% (v/v) dimethyl sulfoxide (DMSO); an eight-point solvent-correction curve bracketing 5.0% DMSO (prepared by mixing 4.5% and 5.8% DMSO stocks) was used for bulk-refractive index correction. E-peptides were injected as analytes from low to high concentration (0.3125–10 µM) at a flow rate of 30 µL min⁻¹ with a 150 s association phase, followed by surface regeneration with 10 mM glycine–HCl (pH 2.0) for 5 min between injections. Binding parameters were obtained by global fitting of double-referenced, solvent-corrected sensorgrams to a 1:1 Langmuir model in Biacore Insight (Cytiva), yielding the equilibrium dissociation constant (K_d).

Statistics and reproducibility

All assays were performed in at least three independent experiments. Sample sizes were determined based on prior experience or published literature. No statistical method was used to predetermine sample size. No data were excluded from the analyses. Human specimens diagnosed with colorectal cancer were randomly selected from eligible hospital patients. Animals were randomly assigned to control and experimental groups using a simple randomization method. Investigators were blinded to group allocation during data collection and analysis, including outcome assessment. Data are presented as mean ± SD. Statistical analyses were performed using GraphPad Prism (version 9.5) and Origin (version 2021). Before statistical analysis, tests for normality and homogeneity of variance were performed. Two-tailed unpaired or paired Student’s t tests were used for comparisons between two groups. For multiple-group comparisons, one-way ANOVA with Tukey’s or Dunnett’s multiple-comparisons tests, or two-way ANOVA with Tukey’s multiple-comparisons test were used. Categorical variables were analyzed using chi-square tests. Correlations were assessed using Pearson’s correlation (two-tailed). Survival analyses were performed using the Kaplan–Meier method and compared by the log-rank test. Statistical significance was set at p < 0.05.

Ethics statement

The clinical specimens were obtained from the Department of Gastrointestinal Surgery, Tongji Hospital. All patients were informed about the study and provided written informed consent prior to surgery. Consent was obtained for the acquisition, reporting, and sharing of patient clinical information at the individual level. No participant compensation was provided. Patients’ microsatellite status (MSS or MSI-H) and clinical characteristics were obtained from their medical records. Patient information, including age and sex, is shown in Tables S1–S5 (Supplementary Information). Tissue microarrays of colorectal tumors and paracancers were purchased from the Shanghai Outdo Biotech Co. Ltd (HColA180Su21). For these commercial tissue microarrays, demographic information was provided by the vendor and is reported in the Source Data. Sex- and/or gender-stratified analyses were not performed because the study was not designed to detect sex/gender differences, and the human analyses primarily focused on tumor molecular subtype (MSS vs MSI-H). The study was approved by the Ethics Committee of Tongji Hospital (TJ-IRB20220723). All animal experiments were conducted with approval from the Animal Care and Use Committee of Tongji Hospital (TJH-202303006).

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

The analysis of CD8⁺ T cell infiltration levels based on CD276 (B7H3) mRNA levels was performed using data from the TIMER database (https://cistrome.shinyapps.io/timer/). Kaplan–Meier analysis of overall survival (OS) and relapse-free survival (RFS) based on B7H3 and ZDHHC24 mRNA levels was conducted using the Kaplan–Meier Plotter database (https://kmplot.com/analysis/). The gene expression profile across all tumor samples and paired normal tissues based on ZDHHC24 mRNA levels was obtained from the Gene Expression Profiling Interactive Analysis (GEPIA) database (http://gepia.cancer-pku.cn/index.html). RNA-sequencing data for MSS and MSI-H tumors have been deposited in the Sequence Read Archive (SRA) under BioProject accession PRJNA1233603. Non-targeted metabolomics sequencing data for MSS and MSI-H tumors have been deposited in the MetaboLights database under accession MTBLS12300. All data in this study are provided in the supplementary table and source data. Source data are provided with this paper.

References

Sung, H. et al. Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J. Clin. 71, 209–249 (2021).
PubMed Google Scholar
Dekker, E., Tanis, P. J., Vleugels, J. L. A., Kasi, P. M. & Wallace, M. B. Colorectal cancer. Lancet 394, 1467–1480 (2019).
PubMed Google Scholar
Sharma, P. et al. Immune checkpoint therapy-current perspectives and future directions. Cell 186, 1652–1669 (2023).
ADS CAS PubMed Google Scholar
Topalian, S. L., Taube, J. M., Anders, R. A. & Pardoll, D. M. Mechanism-driven biomarkers to guide immune checkpoint blockade in cancer therapy. Nat. Rev. Cancer 16, 275–287 (2016).
CAS PubMed PubMed Central Google Scholar
Sun, Q. et al. Immune checkpoint therapy for solid tumours: clinical dilemmas and future trends. Signal Transduct. Target. Ther. 8, 320 (2023).
CAS PubMed PubMed Central Google Scholar
Le, D. T. et al. Mismatch repair deficiency predicts response of solid tumors to PD-1 blockade. Science 357, 409–413 (2017).
ADS CAS PubMed PubMed Central Google Scholar
Overman, M. J. et al. Nivolumab in patients with metastatic DNA mismatch repair-deficient or microsatellite instability-high colorectal cancer (CheckMate 142): an open-label, multicentre, phase 2 study. Lancet Oncol. 18, 1182–1191 (2017).
CAS PubMed PubMed Central Google Scholar
André, T. et al. Nivolumab plus low-dose ipilimumab in previously treated patients with microsatellite instability-high/mismatch repair-deficient metastatic colorectal cancer: 4-year follow-up from CheckMate 142. Ann. Oncol. J. Eur. Soc. Med. Oncol. 33, 1052–1060 (2022).
Google Scholar
Zhao, P., Li, L., Jiang, X. & Li, Q. Mismatch repair deficiency/microsatellite instability-high as a predictor for anti-PD-1/PD-L1 immunotherapy efficacy. J. Hematol. Oncol. 12, 54 (2019).
PubMed PubMed Central Google Scholar
Le, D. T. et al. PD-1 blockade in tumors with mismatch-repair deficiency. N. Engl. J. Med. 372, 2509–2520 (2015).
CAS PubMed PubMed Central Google Scholar
Galon, J. et al. Type, density, and location of immune cells within human colorectal tumors predict clinical outcome. Science 313, 1960–1964 (2006).
ADS CAS PubMed Google Scholar
Schildberg, F. A., Klein, S. R., Freeman, G. J. & Sharpe, A. H. Coinhibitory pathways in the B7-CD28 ligand-receptor family. Immunity 44, 955–972 (2016).
CAS PubMed PubMed Central Google Scholar
Ceeraz, S., Nowak, E. C. & Noelle, R. J. B7 family checkpoint regulators in immune regulation and disease. Trends Immunol. 34, 556–563 (2013).
CAS PubMed PubMed Central Google Scholar
Burke, K. P., Chaudhri, A., Freeman, G. J. & Sharpe, A. H. The B7:CD28 family and friends: unraveling coinhibitory interactions. Immunity 57, 223–244 (2024).
CAS PubMed PubMed Central Google Scholar
Lee, Y. H. et al. Inhibition of the B7-H3 immune checkpoint limits tumor growth by enhancing cytotoxic lymphocyte function. Cell Res. 27, 1034–1045 (2017).
CAS PubMed PubMed Central Google Scholar
Peuker, K. et al. Microbiota-dependent activation of the myeloid calcineurin-NFAT pathway inhibits B7H3- and B7H4-dependent anti-tumor immunity in colorectal cancer. Immunity 55, 701–717.e7 (2022).
CAS PubMed Google Scholar
Tyagi, A. et al. Evidence supporting a role for the immune checkpoint protein B7-H3 in NK cell-mediated cytotoxicity against AML. Blood 139, 2782–2796 (2022).
CAS PubMed PubMed Central Google Scholar
Wang, C. et al. CD276 expression enables squamous cell carcinoma stem cells to evade immune surveillance. Cell Stem Cell 28, 1597–1613 (2021).
CAS PubMed PubMed Central Google Scholar
Ingebrigtsen, V. A. et al. B7-H3 expression in colorectal cancer: nuclear localization strongly predicts poor outcome in colon cancer. Int. J. Cancer 131, 2528–2536 (2012).
CAS PubMed Google Scholar
Zang, X. et al. Tumor-associated endothelial expression of B7-H3 predicts survival in ovarian carcinomas. Mod. Pathol. 23, 1104–1112 (2010).
CAS PubMed PubMed Central Google Scholar
Zang, X. et al. B7-H3 and B7x are highly expressed in human prostate cancer and associated with disease spread and poor outcome. Proc. Natl. Acad. Sci. USA 104, 19458–19463 (2007).
ADS CAS PubMed PubMed Central Google Scholar
Ingebrigtsen, V. A. et al. B7-H3 expression in colorectal cancer: associations with clinicopathological parameters and patient outcome. BMC Cancer 14, 602 (2014).
PubMed PubMed Central Google Scholar
Picarda, E., Ohaegbulam, K. C. & Zang, X. Molecular pathways: targeting B7-H3 (CD276) for human cancer immunotherapy. Clin. Cancer Res. J. Am. Assoc. Cancer Res. 22, 3425–3431 (2016).
CAS Google Scholar
Getu, A. A. et al. New frontiers in immune checkpoint B7-H3 (CD276) research and drug development. Mol. Cancer 22, 43 (2023).
CAS PubMed PubMed Central Google Scholar
Huang, Y. et al. FUT8-mediated aberrant N-glycosylation of B7H3 suppresses the immune response in triple-negative breast cancer. Nat. Commun. 12, 2672 (2021).
ADS CAS PubMed PubMed Central Google Scholar
Chen, R. et al. Glycoproteomics analysis of human liver tissue by combination of multiple enzyme digestion and hydrazide chemistry. J. Proteome Res. 8, 651–661 (2009).
ADS CAS PubMed Google Scholar
Jiang, H. et al. Protein lipidation: occurrence, mechanisms, biological functions, and enabling technologies. Chem. Rev. 118, 919–988 (2018).
CAS PubMed PubMed Central Google Scholar
Chamberlain, L. H. & Shipston, M. J. The physiology of protein S-acylation. Physiol. Rev. 95, 341–376 (2015).
CAS PubMed PubMed Central Google Scholar
Mesquita, F. S. et al. Mechanisms and functions of protein S-acylation. Nat. Rev. Mol. Cell Biol. 25, 488–509 (2024).
Google Scholar
Resh, M. D. Covalent lipid modifications of proteins. Curr. Biol. 23, R431–R435 (2013).
CAS PubMed PubMed Central Google Scholar
Rocks, O. et al. The palmitoylation machinery is a spatially organizing system for peripheral membrane proteins. Cell 141, 458–471 (2010).
CAS PubMed Google Scholar
Linder, M. E. & Deschenes, R. J. Palmitoylation: policing protein stability and traffic. Nat. Rev. Mol. Cell Biol. 8, 74–84 (2007).
CAS PubMed Google Scholar
Zhang, M. et al. A STAT3 palmitoylation cycle promotes TH17 differentiation and colitis. Nature 586, 434–439 (2020).
ADS CAS PubMed PubMed Central Google Scholar
Hao, J. W. et al. CD36 facilitates fatty acid uptake by dynamic palmitoylation-regulated endocytosis. Nat. Commun. 11, 4765 (2020).
ADS CAS PubMed PubMed Central Google Scholar
Huang, X. et al. S-acylation of p62 promotes p62 droplet recruitment into autophagosomes in mammalian autophagy. Mol. Cell 83, 3485–3501 (2023).
CAS PubMed PubMed Central Google Scholar
Rana, M. S. et al. Fatty acyl recognition and transfer by an integral membrane S-acyltransferase. Science 359, eaao6326 (2018).
PubMed PubMed Central Google Scholar
Elliot Murphy, R. & Banerjee, A. In vitro reconstitution of substrate S-acylation by the zDHHC family of protein acyltransferases. Open Biol. 12, 210390 (2022).
CAS PubMed PubMed Central Google Scholar
Yao, H. et al. Inhibiting PD-L1 palmitoylation enhances T-cell immune responses against tumours. Nat. Biomed. Eng. 3, 306–317 (2019).
CAS PubMed Google Scholar
Lu, Y. et al. Palmitoylation of NOD1 and NOD2 is required for bacterial sensing. Science 366, 460–467 (2019).
ADS CAS PubMed Google Scholar
Chen, S. et al. Palmitoylation-dependent activation of MC1R prevents melanomagenesis. Nature 549, 399–403 (2017).
ADS CAS PubMed PubMed Central Google Scholar
Bu, L. et al. High-fat diet promotes liver tumorigenesis via palmitoylation and activation of AKT. Gut 73, 1156–1168 (2024).
CAS PubMed Google Scholar
Bu, L. et al. Targeting APT2 improves MAVS palmitoylation and antiviral innate immunity. Mol. Cell 84, 3513–3529 (2024).
CAS PubMed Google Scholar
Tall, A. R. & Yvan-Charvet, L. Cholesterol, inflammation and innate immunity. Nat. Rev. Immunol. 15, 104–116 (2015).
CAS PubMed PubMed Central Google Scholar
Pascual, G. et al. Dietary palmitic acid promotes a prometastatic memory via Schwann cells. Nature 599, 485–490 (2021).
ADS CAS PubMed Google Scholar
Zheng, S. et al. ZDHHC5-mediated NLRP3 palmitoylation promotes NLRP3-NEK7 interaction and inflammasome activation. Mol. Cell 83, 4570–4585 (2023).
CAS PubMed Google Scholar
Brigidi, G. S. & Bamji, S. X. Detection of protein palmitoylation in cultured hippocampal neurons by immunoprecipitation and acyl-biotin exchange (ABE). J Vis. Exp. 50031, https://doi.org/10.3791/50031 (2013).
Pohl, C. & Dikic, I. Cellular quality control by the ubiquitin-proteasome system and autophagy. Science 366, 818–822 (2019).
ADS CAS PubMed Google Scholar
Cullen, P. J. & Steinberg, F. To degrade or not to degrade: mechanisms and significance of endocytic recycling. Nat. Rev. Mol. Cell Biol. 19, 679–696 (2018).
CAS PubMed Google Scholar
Gatica, D., Lahiri, V. & Klionsky, D. J. Cargo recognition and degradation by selective autophagy. Nat. Cell Biol. 20, 233–242 (2018).
CAS PubMed PubMed Central Google Scholar
Vargas, J. N. S., Hamasaki, M., Kawabata, T., Youle, R. J. & Yoshimori, T. The mechanisms and roles of selective autophagy in mammals. Nat. Rev. Mol. Cell Biol. 24, 167–185 (2023).
CAS PubMed Google Scholar
Ning, W. et al. GPS-palm: a deep learning-based graphic presentation system for the prediction of S-palmitoylation sites in proteins. Brief. Bioinform. 22, 1836–1847 (2021).
CAS PubMed Google Scholar
Xie, Y. et al. GPS-lipid: a robust tool for the prediction of multiple lipid modification sites. Sci. Rep. 6, 28249 (2016).
ADS CAS PubMed PubMed Central Google Scholar
Waterhouse, A. M., Procter, J. B., Martin, D. M. A., Clamp, M. & Barton, G. J. Jalview version 2—a multiple sequence alignment editor and analysis workbench. Bioinform. Oxf. Engl. 25, 1189–1191 (2009).
CAS Google Scholar
Zhang, W. et al. Lactate is a natural suppressor of RLR signaling by targeting MAVS. Cell 178, 176–189.e15 (2019).
CAS PubMed PubMed Central Google Scholar
Wang, F. et al. Combined anti-PD-1, HDAC inhibitor and anti-VEGF for MSS/pMMR colorectal cancer: a randomized phase 2 trial. Nat. Med. 30, 1035–1043 (2024).
CAS PubMed Google Scholar
Ahmed, K. A. et al. Clinical outcomes of melanoma brain metastases treated with stereotactic radiosurgery and anti-PD-1 therapy, anti-CTLA-4 therapy, BRAF/MEK inhibitors, BRAF inhibitor, or conventional chemotherapy. Ann. Oncol. J. Eur. Soc. Med. Oncol. 27, 2288–2294 (2016).
CAS Google Scholar
Hegde, P. S. & Chen, D. S. Top 10 challenges in cancer immunotherapy. Immunity 52, 17–35 (2020).
CAS PubMed Google Scholar
Topalian, S. L., Drake, C. G. & Pardoll, D. M. Immune checkpoint blockade: a common denominator approach to cancer therapy. Cancer Cell 27, 450–461 (2015).
CAS PubMed PubMed Central Google Scholar
Zhou, W. T. & Jin, W. L. B7-H3/CD276: an emerging cancer immunotherapy. Front. Immunol. 12, 701006 (2021).
CAS PubMed PubMed Central Google Scholar
Diskin, C., Ryan, T. A. J. & O’Neill, L. A. J. Modification of proteins by metabolites in immunity. Immunity 54, 19–31 (2021).
CAS PubMed Google Scholar
Qin, W., Yang, F. & Wang, C. Chemoproteomic profiling of protein-metabolite interactions. Curr. Opin. Chem. Biol. 54, 28–36 (2020).
CAS PubMed Google Scholar
Shang, S., Liu, J. & Hua, F. Protein acylation: mechanisms, biological functions and therapeutic targets. Signal Transduct. Target. Ther. 7, 396 (2022).
CAS PubMed PubMed Central Google Scholar
Peng, F. et al. Oncogenic fatty acid oxidation senses circadian disruption in sleep-deficiency-enhanced tumorigenesis. Cell Metab. 36, 1598–1618 (2024).
CAS PubMed Google Scholar
Du, A. et al. Fatty acids in cancer: metabolic functions and potential treatment. MedComm – Oncol. 2, e25 (2023).
CAS Google Scholar
Wood, S. M. et al. Fatty acid-binding protein 1 is preferentially lost in microsatellite instable colorectal carcinomas and is immune modulated via the interferon γ pathway. Mod. Pathol. 30, 123–133 (2017).
CAS PubMed Google Scholar
Lin, A., Zhang, J. & Luo, P. Crosstalk between the MSI status and tumor microenvironment in colorectal cancer. Front. Immunol. 11, 2039 (2020).
CAS PubMed PubMed Central Google Scholar
Akbari, V., Kallhor, M., Mollashahi, B. & Abolfazl, A. Transcriptome and network dissection of microsatellite stable and highly instable colorectal cancer. Asian Pac. J. Cancer Prev. 20, 2445–2454 (2019).
CAS PubMed PubMed Central Google Scholar
Galon, J. & Bruni, D. Approaches to treat immune hot, altered and cold tumours with combination immunotherapies. Nat. Rev. Drug Discov. 18, 197–218 (2019).
CAS PubMed Google Scholar
Herbst, R. S. et al. Predictive correlates of response to the anti-PD-L1 antibody MPDL3280A in cancer patients. Nature 515, 563–567 (2014).
ADS CAS PubMed PubMed Central Google Scholar
Shi, W. et al. Immune checkpoint B7-H3 is a therapeutic vulnerability in prostate cancer harboring PTEN and TP53 deficiencies. Sci. Transl. Med. 15, eadf6724 (2023).
CAS PubMed PubMed Central Google Scholar
Altan, M. et al. Association of B7-H4, PD-L1, and tumor infiltrating lymphocytes with outcomes in breast cancer. NPJ Breast Cancer 4, 40 (2018).
PubMed PubMed Central Google Scholar
Tawbi, H. A. et al. Relatlimab and nivolumab versus nivolumab in untreated advanced melanoma. N. Engl. J. Med. 386, 24–34 (2022).
CAS PubMed PubMed Central Google Scholar
Griffioen, A. W., Damen, C. A., Martinotti, S., Blijham, G. H. & Groenewegen, G. Endothelial intercellular adhesion molecule-1 expression is suppressed in human malignancies: the role of angiogenic factors. Cancer Res. 56, 1111–1117 (1996).
CAS PubMed Google Scholar
Wang, R., Shi, T. & Chen, W. Combination therapies with anti-angiogenesis and B7-H3 blockade in cancers. J. Cancer Immunol. 2, 17–21 (2020).
Google Scholar
Huang, C. et al. PD-L1 methylation restricts PD-L1/PD-1 interactions to control cancer immune surveillance. Sci. Adv. 9, eade4186 (2023).
CAS PubMed PubMed Central Google Scholar
Zhu, Q. et al. O-GlcNAcylation promotes tumor immune evasion by inhibiting PD-L1 lysosomal degradation. Proc. Natl. Acad. Sci. USA 120, e2216796120 (2023).
CAS PubMed PubMed Central Google Scholar
Devlin, J. R. et al. Structural dissimilarity from self drives neoepitope escape from immune tolerance. Nat. Chem. Biol. 16, 1269–1276 (2020).
CAS PubMed PubMed Central Google Scholar
Chandran, S. S. et al. Immunogenicity and therapeutic targeting of a public neoantigen derived from mutated PIK3CA. Nat. Med. 28, 946–957 (2022).
CAS PubMed PubMed Central Google Scholar

Download references

Acknowledgements

We acknowledge the equipment support from the Experimental Medicine Center of Tongji Hospital, Tongji Medical School, Huazhong University of Science and Technology, and the technical support from the Medical Subcenter of HUST Analytical and Testing Center. This work was supported by the National Natural Science Foundation of China (no. 82472829 (X. Luo), no. 82425041 (G. Wang), no. 82330084 (G. Wang)), the National Key Research and Development Program of China (2022YFA1105303 (G. Wang), 2023YFC3402100 (G. Wang)), the Basic Research Support Program of Huazhong University of Science and Technology (2023BR036 (G. Wang)), and the Noncommunicable Chronic Diseases—National Science and Technology Major Project (no. 2024ZD0520600 (J. Hu)).

Author information

Authors and Affiliations

GI Cancer Research Institute, Tongji Hospital, Huazhong University of Science and Technology, Wuhan, China
Zejun Rao, Changsheng Huang, Qi Wu, Mao Li, Anyi Liu, Tong Zhu, Wangshuo Yang, Lanlan Yin, Shengyu Zhu, Xiaowei She, Chengxin Yu, Lang Liu, Pengcheng Li, Yucong Bai, Dongjing Zhang, Tianci Xie, Xiang Liu, Lu Liu, Feng Xu, Guihua Wang, Junbo Hu & Xuelai Luo
State Key Laboratory for Diagnosis and Treatment of Severe Zoonotic Infectious Diseases, Huazhong University of Science and Technology, Wuhan, Hubei, China
Guihua Wang

Authors

Zejun Rao
View author publications
Search author on:PubMed Google Scholar
Changsheng Huang
View author publications
Search author on:PubMed Google Scholar
Qi Wu
View author publications
Search author on:PubMed Google Scholar
Mao Li
View author publications
Search author on:PubMed Google Scholar
Anyi Liu
View author publications
Search author on:PubMed Google Scholar
Tong Zhu
View author publications
Search author on:PubMed Google Scholar
Wangshuo Yang
View author publications
Search author on:PubMed Google Scholar
Lanlan Yin
View author publications
Search author on:PubMed Google Scholar
Shengyu Zhu
View author publications
Search author on:PubMed Google Scholar
Xiaowei She
View author publications
Search author on:PubMed Google Scholar
Chengxin Yu
View author publications
Search author on:PubMed Google Scholar
Lang Liu
View author publications
Search author on:PubMed Google Scholar
Pengcheng Li
View author publications
Search author on:PubMed Google Scholar
Yucong Bai
View author publications
Search author on:PubMed Google Scholar
Dongjing Zhang
View author publications
Search author on:PubMed Google Scholar
Tianci Xie
View author publications
Search author on:PubMed Google Scholar
Xiang Liu
View author publications
Search author on:PubMed Google Scholar
Lu Liu
View author publications
Search author on:PubMed Google Scholar
Feng Xu
View author publications
Search author on:PubMed Google Scholar
Guihua Wang
View author publications
Search author on:PubMed Google Scholar
Junbo Hu
View author publications
Search author on:PubMed Google Scholar
Xuelai Luo
View author publications
Search author on:PubMed Google Scholar

Contributions

X. Luo and J. Hu conceived and designed the study. Z. Rao performed most experiments and analyzed the data. C. Huang, Q. Wu, and M. Li performed the animal experiments. A. Liu, T. Zhu, and W. Yang collected biological samples. L. Yin, S. Zhu, X. She, and C. Yu performed data analysis. Lang Liu, P. Li, Y. Bai, and D. Zhang curated and compiled patients’ clinical and pathological data. T. Xie and X. Liu reviewed and edited the manuscript. Lu Liu and F. Xu provided experimental support. G. Wang provided project-level guidance, contributed to study design and data interpretation, and critically revised the manuscript with constructive suggestions. Z. Rao and X. Luo drafted the manuscript and revised it. X. Luo and J. Hu supervised the project. All authors reviewed the manuscript and approved the final version.

Corresponding authors

Correspondence to Junbo Hu or Xuelai Luo.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks Sérgio Veloso and the other anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplement Information (download PDF )

Description of Additional Supplementary Files (download PDF )

Supplementary Data 1 (download XLSX )

Supplementary Data 2 (download XLSX )

Reporting Summary (download PDF )

Peer Review File (download PDF )

Source data

Source Data (download XLSX )

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Rao, Z., Huang, C., Wu, Q. et al. Palmitic acid-triggered B7H3 palmitoylation promotes immune escape. Nat Commun 17, 1810 (2026). https://doi.org/10.1038/s41467-026-68525-x

Download citation

Received: 05 March 2025
Accepted: 08 January 2026
Published: 19 January 2026
Version of record: 18 February 2026
DOI: https://doi.org/10.1038/s41467-026-68525-x