Introduction

Tumors evade immune recognition and eradication via several inhibitory checkpoint molecules to hijack the immune system, leading to T-cell exhaustion1. Blockade of these immune checkpoint molecules remarkably reinvigorates T-cell functions and unleashes antitumor immunity for durable antitumor effects in different malignancies2,3. However, not all patients with cancer or types of cancer respond to immune checkpoint blockade (ICB) treatment as expected, as shown by the variable response rates4, suggesting that other intrinsic and extrinsic factors are related to the response to ICB. For example, high PD-L1 expression, tumor mutation burden, and a high frequency of tumor-infiltrating lymphocytes within the tumor microenvironment (TME) have been reported to correlate favorably with the clinical benefits of ICB therapies5,6,7. Therefore, only 10–20% of patients with colorectal cancer (CRC) who are deficient mismatched repair (dMMR) benefit from ICB, and the expression of PD-L1 is detected in 37% of patients with proficient MMR (pMMR)-CRC and in 29% of patients with dMMR-CRC8. These data suggest that other PD-L1-independent immunosuppressive mechanisms may exist to inhibit antitumor immunity, leading to a poor clinical response to ICBs.

Mutation-associated neoantigens (neoAgs) contribute to favorable responses to ICB therapy in cancer types with high mutational burdens, including melanoma, NSCLC, and dMMR-CRC9,10,11. These high mutational burden tumors potentially generate immunogenic neoantigens to elicit effective immune responses12. Currently, several clinical trials have demonstrated the feasibility, safety, and tolerability of neoantigen-based cancer vaccines in patients with melanoma or glioblastoma13,14. Ameliorating the therapeutic efficacy of neoantigen-based cancer vaccines requires immunological interventions, such as radiotherapy, to overcome the immunosuppressive TME15,16,17. In addition to its direct cytotoxic effect, radiotherapy functions as an in vivo tumor vaccine by inducing damage-associated molecular pattern (DAMP)-dependent immunogenic cell death (ICD)18, enhancing the access of immune effector cells to tumor cells19, and upregulating MHC-I expression and tumor antigen cross-presentation20, ultimately resulting in increased numbers and action of tumor-specific CD8+ T cells21,22,23. A recent study revealed that radiation broadens the T-cell repertoire and exposes neoantigens for neoantigen-specific immune responses15,22,23,24, suggesting that radiation can overcome immunosuppression to ameliorate the therapeutic efficacy of neoantigen-based cancer vaccines. Additionally, radiation promotes PD-L1-mediated immune suppression by IFNγ-induced adaptive resistance in infiltrating T cells to inhibit antigen presentation and exhaust CD8+ T-cell function, dampening antitumor immune responses6,25,26,27, suggesting that immune checkpoint proteins may attenuate neoantigen presentation. Indeed, tumor PD-L1 affects neoantigen presentation for acquired resistance28. Our previous studies revealed that the administration of modified adenovirus-associated virus (mAAV), which is engineered with PD-L1-targeting miRNA and a PD1 trap, enhances the neoantigen-specific T-cell response. In combination with radiotherapy, mAAV neoantigen cancer vaccination promotes a neoantigen-specific T-cell response, increasing the therapeutic effect to eradicate tumors and inhibit distant metastasis in poorly immunogenic tumors such as microsatellite-stable colorectal cancer (MSS-CRC) and triple-negative breast cancer (TNBC)16,29. Furthermore, little is known about if the expression of these immune checkpoint proteins in tumors controls T-cell immunity against neoantigens.

B7-H3 (CD276) belongs to the B7 family and shares 20–27% amino acid identity with other B7 family members30. Aberrant B7-H3 expression has been reported in the vast majority of human malignancies, including melanoma, ovarian cancer and CRC31,32,33,34. Aberrant B7-H3 expression promotes tumor progression by inhibiting the functions of CD8+ T cells and NK cells35. Blockade of B7-H3 has potent therapeutic effects by reinvigorating antitumor immunity and synergizing with ICBs in multiple preclinical models33,35. However, no study has evaluated whether B7-H3 mediates immune evasion by attenuating the neoantigen-specific T-cell response.

Here, we found that patients with rectal cancer had higher B7-H3 expression after neoadjuvant chemoradiotherapy (CRT) than before CRT. The status of tumor B7-H3 expression was associated with the response to CRT. Knockdown of tumor B7-H3 expression increases radiotherapy-induced dendritic cell maturation, T-cell infiltration and neoantigen-specific T-cell responses. Moreover, combination with B7-H3 blockade significantly enhances the therapeutic efficacy of neoantigen-based cancer vaccines and radiotherapy. Inhibition of B7-H3 in combination with neoantigen-based AAV cancer vaccines can compensate for ICI-resistant MSS-CRC, resulting in the generation of highly neoantigen-reactive T-cell immunity, suggesting that targeting B7-H3 can increase the therapeutic efficacy of neoantigen vaccines and radiotherapy in ICB-resistant cancers such as MSS-CRC.

Results

Elevation of tumor B7-H3 by chemoradiotherapy is associated with poor survival outcomes in patients with locally-advanced CRC

Previous studies have shown that B7-H3 is aberrantly expressed in CRC and contributes to chemoresistance and radioresistance32,36,37. To further investigate whether the levels of tumor and stromal B7-H3 could be increased by concurrent chemoradiotherapy (CRT), we analyzed 106 pre-CRT biopsies and 119 post-CRT surgical tissues from patients with locally-advanced CRC by IHC staining (Fig. 1a). The results of tumor B7-H3 expression in pair-matched pre-CRT biopsies and post-CRT surgical tissues revealed that the protein expression levels of tumor B7-H3 were much higher in the post-CRT surgical tissues than in the pre-CRT biopsies (Fig. 1a, n = 91, p < 0.001). Additionally, the level of B7-H3 was associated with a poor response to CRT (Fig. 1b, n = 119, p = 0.0339). High B7-H3 tumors were associated with poor disease-free survival (DFS) according to the Kaplan‒Meier method (Fig. 1c, n = 119, log-rank p = 0.0421). There was no significant difference in stromal B7-H3 levels before or after CRT treatment (Supplementary Fig. 1a), and the stromal B7-H3 levels were not associated with survival outcomes (Supplementary Fig. 1b). Additionally, we found that the B7-H3 level in tumors was associated with fewer tumor-infiltrating CD3+ and CD8+ T cells after CRT treatment (Supplementary Fig. 1c-1d, p = 0.0437 and 0.0444, unpaired t test). These results suggest that tumor B7-H3 triggered by CRT may attenuate the therapeutic response and increase the risk of tumor relapse.

Fig. 1: The upregulation of B7-H3 expression by radiotherapy was associated with poor survival outcomes and correlated with a decreased neoantigen-specific T-cell response.
figure 1

a The status of B7-H3 expression within the tumor microenvironment before and after neoadjuvant chemotherapy (CRT) was evaluated by immunohistochemistry. The level of B7-H3 in tumors before and after CRT was analyzed (paired t test, p < 0.001; n = 91). b The status of tumor B7-H3 post-CRT samples was negatively associated with the response to neoCRT (p = 0.0339, n = 119). c. The level of tumor B7-H3 post-CRT samples was associated with disease-free survival (DFS, log-rank p = 0.0421, n = 119). d Local radiotherapy (5 Gy) significantly elicited a neoantigen-specific T-cell response. A total of 3 × 105 CT26 cells were subcutaneously injected into the right leg of BALB/c mice (n = 7–8). Local RT was administered on day 11 and 14, and splenocytes were collected on day 25. The resected tumors after local radiotherapy treatment were stratified into nonresponder and responder groups on the basis of tumor size. These tumors were analyzed by immunoblotting. e Eight H2-Kd-restricted neopeptides were individually cocultured with splenocyte T cells for 24 h and analyzed via an IFNγ+ ELISPOT assay. ***p < 0.001. f The response to radiotherapy was negatively correlated with the H2-Kd-restricted neoantigen-specific T-cell response (n = 6-7). *p < 0.05 and **p < 0.01. g The levels of immune checkpoint proteins in resected tumors were evaluated by immunoblotting (n = 3). **p < 0.01. h CT26 and HCT116 cells were irradiated with different doses of RT and harvested after 24 h. The levels of immune checkpoint proteins were evaluated by immunoblotting.

Full size image

Radiation promoted the upregulation of immune checkpoint proteins in cancer cells to inhibit antigen presentation, decrease cancer immunogenicity and exhaust CD8+ T-cell function to dampen antitumor immune responses6,38,39. Recent studies have indicated that tumor PD-L1 affects neoantigen presentation for innate immune resistance, suggesting that immune checkpoint proteins may attenuate antigen presentation, including neoantigens, for immune escape28. Therefore, we evaluated whether B7-H3 decreases the therapeutic efficacy of CRT by inhibiting neoantigen presentation for neoantigen-specific T-cell responses. We inoculated BALB/c mice with syngeneic CT26 colon cells (n = 7). As shown in Fig. 1D, the tumor volume was significantly decreased by local radiotherapy. Splenocytes were isolated from PBS- and RT-treated mice and cocultured with well-defined neopeptides (H2-Kd-restricted neoantigens) for 24 h by ex vivo stimulation29,40,41. We found that the number of IFNγ+ T cells was markedly increased in splenocytes from RT-treated mice cocultured with these neopeptides (Fig. 1e). We detected a greater percentage of IFNγ+ T cells in good responders (small tumor volume) than in poor responders (large tumor volume, Fig. 1f) ex vivo when stimulated with neopeptides. In addition, the levels of these immune checkpoint proteins, including B7-H3, markedly increased in surviving tumor cells after radiotherapy. Notably, the upregulation of B7-H3 was significant in poor responder mice (Fig. 1g). Similar results were observed in irradiated colorectal cancer cells (Fig. 1h). We found that the tumor B7-H3 levels were significantly increased by radiotherapy. Taken together, these results suggest that the upregulation of B7-H3 expression via radiotherapy may decrease the neoantigen-specific T-cell response to attenuate the therapeutic efficacy of radiotherapy.

Knockdown of tumor B7-H3 enhances neoantigen-specific T-cell responses to increase the response to radiotherapy

Tumor B7-H3 has been reported to suppress antitumor T-cell functions in cold tumors insensitive to ICBs such as ovarian cancer33, which is in accordance with our results. To evaluate whether tumor B7-H3 participates in antigen presentation to decrease antitumor immunity, we performed a global transcriptomic analysis to examine antigen presentation-related gene signatures42. In silico analysis of these databases, which is retrieved from NCBI GEO database (GSE165610), revealed that antigen presentation-related gene signatures, including HLA, TAP and B2M signatures, were significantly increased in HCT116CD276KO cells compared with those in HCT116WT cells (Fig. 2a). These results suggest that tumor B7-H3 may decrease the ability of antigen presentation to reduce cancer immunogenicity for immune evasion. To evaluate the effect of tumor B7-H3 on antitumor immunity in the murine colon cancer cell line CT26 (MSS-CRC, known as a cold tumor type) in vivo, we first generated genetically derived B7-H3 (CD276)-knockdown tumors to examine the response to radiotherapy. By subcutaneous injection of CT26shNC or CT26shCD276, we found that CT26shCD276 was more slowly progressive than was CT26shNC in BALB/c mice, indicating that B7-H3 tumors may inhibit antitumor immunity (Fig. 2b). These findings suggest that B7-H3 expression and tumor antigen presentation in tumor cells directly dampened T-cell-dependent tumor suppression. To understand the relationship between antitumor immunity and B7-H3 expression in tumors treated with radiotherapy, we evaluated the effects of combined treatment with CT26shNC and CT26shCD276 tumor cells in vivo. Tumor regression was more effective in RT-treated CT26shCD276 tumor cell-bearing mice than in RT-treated CT26shNC tumor-bearing mice (Fig. 2b). Due to the enhanced antitumor effect in CT26shCD276 mice, we hypothesized two possibilities: (1) the balance of immune responses was altered in CT26shCD276 mice, or (2) T cells responded to neoantigens that were more common in CT26shCD276-bearing mice. As such, we analyzed the immune responses within the tumor microenvironment. There was a significant difference in the frequency of tumor-infiltrating dendritic cells (MHCII+ CD11c+ DCs, CD86+ MHCII+ CD11c+ DCs and CD86+ MHCII+ CD11c+ DCs) in CT26shCD276-bearing mice compared with that in CT26shNC-bearing mice (Fig. 2c, d and the gating strategies are shown in Supplementary Fig. 2). Moreover, the infiltration of DCs (MHCII+ CD11c+ and CD86+ MHCII+ CD11c+ DCs) was also greater in CT26shCD276 cells treated with local radiotherapy than in those in the CT26shNC/RT subgroup (Fig. 2c and Supplementary Fig. 2b-2d). These findings indicate that tumor B7-H3 expression may directly dampen antigen presentation to decrease cancer immunogenicity. We then analyzed the differences in the number of tumor-infiltrating CD4+ and CD8+ T cells between these groups (Supplementary Fig. 2e-2f). There was no significant difference in the frequency of CD4+ T cells (Fig. 2e). We found that the numbers of tumor-infiltrating CD8+ T cells and effector/memory CD8+ T (CD44+CD62L-CD8+CD3+ TEM) cells were significantly greater in the CT26shCD276/RT subgroup than in the CT26shNC/RT subgroup (Fig. 2f, g). Furthermore, the number of cytotoxic IFNγ+CD8+ T cells was markedly increased in the CT26shCD276/RT subgroup (Fig. 2h and Supplementary Fig. 2g). There was no difference in the infiltration of natural killer (NK) cells (Fig. 2i and Supplementary Fig. 2h). These results indicated that tumor B7-H3 may decrease antigen presentation to attenuate the antigen-specific T-cell response to reduce the response to radiotherapy. Next, we assessed neoantigen-specific CTL function. We analyzed the splenic T cell response to previously identified H2-Kd-restricted neoantigens and mutant (Mut) E2f8, Mtch1 and Glud1 after treatment. In RT-treated CT26shCD276-bearing mice, we detected apparent mE2F8-, mMtch1- and mGlud1-specific CD8+ T cell responses (Fig. 2j), suggesting that RT enhanced the apparent neoantigen-specific T-cell response in mice injected with CT26shCD276. Taken together, these results indicate that tumor B7-H3 may decrease antigen presentation to attenuate neoantigen-specific T-cell responses to reduce the therapeutic benefit of radiotherapy.

Fig. 2: Knockdown of tumor B7-H3 significantly increased its immunogenicity to increase the maturation of tumor-infiltrating dendritic cells (TILs) and T-cell activation.
figure 2

a CD276 knockout significantly increased the cancer immunogenicity gene signature in HCT116 cells, which was retrieved from the NCBI GSE165610 dataset. b Schematic diagram of the treatment regimen. BALB/c mice were subcutaneously injected into the right legs with 3 × 105 CT26shNC or CT26shCD276 cells for 10 days. Local radiotherapy (5 Gy) was administered to the right leg on days 11 and 14. The tumor volume was measured every 2‒3 days (n = 5). *p < 0.05 and ***p < 0.001. c Tumor-infiltrating immune cells were extracted from resected tumors for flow cytometry. The percentage of CD11c+MHC-II+ DCs is shown (n = 3-4). *p < 0.05 and ***p < 0.001 (n = 3). d The percentage of tumor-infiltrating CD86+CD11c+MHC-II+ DCs is shown (n = 3-4). *p < 0.05 and **p < 0.01. e The percentage of tumor-infiltrating CD4+CD3+ T cells is shown (n = 3-4). f The percentage of tumor-infiltrating CD8+CD3+ T cells is shown (n = 3-4). *p < 0.05 and **p < 0.01. g The percentage of tumor-infiltrating CD8+CD44+CD62L-CD3+CD45+ TEM cells is shown (n = 3-4). *p < 0.05 and **p < 0.01. h The percentage of tumor-infiltrating cytotoxic IFNγ+CD8+T cells is shown (n = 3-4). *p < 0.05 and **p < 0.01. i The percentage of tumor-infiltrating CD11b+CD49b+NK cells is shown (n = 5). j The solenocytes were harvested for individual H2-Kd-restricted neopeptide stimulation. Representative images of IFNγ+ ELISPOT from splenocytes are shown. One-way ANOVA. *p < 0.05 and ***p < 0.001.

Full size image

Blockade of B7-H3 increased the therapeutic efficacy of neoantigen-based cancer vaccines in combination with radiotherapy

Neoantigen-based cancer vaccines have been reported to increase neoantigen-specific T-cell responses, increasing the response to radiotherapy15,29. We evaluated whether RT-induced tumor B7-H3 upregulation may attenuate neoantigen presentation to reduce the neoantigen-specific T-cell response, leading to less systemic antitumor immunity. We inoculated CT26 cells and intramuscularly injected them with an AAV-based neoantigen cancer vaccine (AAVax). As shown in Supplementary Fig. 3a and 3b, the administration of low- and high-dose AAVax (1 × 108 and 1 × 109 vg/mouse) significantly delayed tumor growth and decreased tumor weight (Supplementary Fig. 3b and 3c). Moreover, the numbers of tumor-infiltrating CD8+ and CD44+CD8+ T cells were markedly increased (Supplementary Fig. 3d and 3e). The number of regulatory T cells (Foxp3+CD25+CD4+) significantly increased after the administration of a high dose of AAVax (Supplementary Fig. 3d-3e), suggesting that a low dose of AAVax is sufficient to enhance the neoantigen-specific T-cell response. The number of IFNγ+ T cells was also markedly increased in response to these H2-Kd-restricted neoantigens (Fig. S3F), suggesting that vaccination with AAVax elicited a neoantigen-specific T-cell response. We subsequently treated CT26-bearing mice with a low dose of AAVax in combination with local radiotherapy (5 Gy ×2, Fig. 3a). As shown in Fig. 3b, the tumor volume and weight were decreased by approximately 50% in the AAVax- and RT-alone groups. Combined administration of AAVax and RT significantly decreased the tumor volume and tumor weight by approximately 80% (Fig. 3b, c). We further isolated residual tumors to analyze the profile of tumor-infiltrating lymphocytes. We found that the numbers of cytotoxic CD8+ T cells and effector/memory CD44+CD8+ TILs were significantly increased in the AAVax/RT group (Fig. 3d). The numbers of immunosuppressive Foxp3+ Treg cells and Gr1+CD11b+ MDSCs were also increased in the RT-alone and AAVax/RT groups (Fig. 3d). Furthermore, the numbers of cytotoxic CD8+ T cells and effector/memory CD44+CD8+ TILs were significantly increased in the spleens of the AAVax/RT group (Fig. 3e), suggesting that AAVax/RT enhanced the antitumor immune response. To demonstrate the neoantigen-specific immune response elicited by AAVax/RT, splenocytes were isolated and stimulated with individual neoantigens. We found that the neoantigen-specific T-cell immune response was significantly increased in solenocytes from mice that received AAVax/RT (Fig. 3f). However, we found that the infiltration of PD1+ immune cells was significantly decreased and that the infiltration of GzmB+ immune cells was markedly increased in the AAVax/RT group, indicating that combined therapy reshaped tumor immunogenicity and remodeled the TME (Fig. 3g–i). Moreover, proinflammatory cytokines such as Sting, Ifnα2, Ifnβ1, Ifnγ, and GzmB were also increased within the tumor microenvironment in the RT-alone and AAVax-alone groups (Supplementary Fig. 3g-3h). The highest levels of these proinflammatory cytokines were observed in the AAVax/RT group. Additionally, the expressions of immune checkpoint-related genes, including Pdl1, B7-H3 (CD276), Pdl2 and Vista (Fig. 3j and Supplementary Fig. 3i), were significantly increased by RT, suggesting that immune checkpoint proteins may attenuate the therapeutic efficacy of neoAg-based cancer vaccines through compensatory upregulation by RT.

Fig. 3: Neoantigen cancer vaccines in combination with RT reshaped tumor immunogenicity and promoted immune checkpoint upregulation.
figure 3

a Scheme and regimen of AAV vaccination and radiotherapy. A total of 3 × 105 CT26 cells were subcutaneously inoculated into the right leg of BALB/c mice for 8 days, and AAV (1 × 108 vg) particles were intramuscularly injected on days 8, 14, 21 and 25. Local radiotherapy (5 Gy) of the right leg was conducted on days 11 and 18. The immune profile of the tumors was analyzed on day 30 (n = 6–7/group). b After AAV injection, the tumor volume was measured every 3 days (n = 5–6/group). *p < 0.05 and **p < 0.01. c The resected tumors were weighed on day 30 (n = 6–7/group). **p < 0.01 and ***p < 0.001. One-way ANOVA. d The CD45+CD8+ T cells, CD45+CD8+CD44+cells (effector CD8+ T cells), CD25+Foxp3+CD4+ regulatory T cells, and Gr1+MDSCs within the tumors were analyzed by flow cytometry (n = 3–5). *p < 0.05, **p < 0.01***p < 0.001. One-way ANOVA. e The numbers of CD45+CD8+ T cells and CD45+CD8+CD44+ cells (effector CD8+ T cells) within the spleen were analyzed by flow cytometry (n = 3–5). *p < 0.05 and **p < 0.01. One-way ANOVA. f The solenocytes were harvested for individual H2-Kd-restricted neopeptide stimulation. Representative images of IFNγ+ ELISPOT from splenocytes are shown. *p < 0.05 and **p < 0.01. One-way ANOVA. g The infiltration of PD1+ and GzmB+ immune cells was evaluated via immunofluorescence microscopy (n = 3). h The quantification of PD1+ immune cells within tumors is shown (n = 3). *p < 0.05 and **p < 0.01. i The quantification of GzmB+ immune cells within tumors is shown (n = 3). **p < 0.01. j The mRNA levels of the immune checkpoint proteins PD-L1 and B7-H3 were evaluated via qRT‒PCR (n = 3‒4/group). **p < 0.01 and ***p < 0.001.

Full size image

B7-H3 is also constitutively expressed on antigen-presenting cells (APCs) to promote tumor progression by inhibiting the functions of CD8+ T cells and NK cells35,43,44. Blockade of B7-H3 on APCs has potent therapeutic effects and can synergize with anti-PD-1 therapy in multiple mouse tumor models33,35. Therefore, we evaluated whether blockade of B7-H3 can increase the extent of the neoantigen-specific T-cell response in combination with radiotherapy to eradicate murine MSS-CRC cancer cells. As shown in Fig. 4a, b, blockade of B7-H3 significantly increased the therapeutic efficacy of AAVax and RT in delaying tumor growth. Moreover, the number of apoptotic cancer cells was markedly increased in combination with anti-B7-H3 antibodies (Fig. 4c, d). Additionally, the density of tumor-infiltrating CD11c+ DCs and GzmB+ immune cells markedly increased after AAVax/RT treatment in combination with anti-B7-H3 antibodies (Fig. 4c and e, f). The mRNA levels of GzmB, Ifna2 and Ifnb1 were also increased after AAVax/RT treatment in combination with anti-B7-H3 antibodies (Fig. 4g–i). Furthermore, the mRNA levels of the exhausted T-cell markers Tim3 and Tigit were reduced in this subgroup (Fig. 5a, b), suggesting that blockade of B7-H3 reshaped the tumor microenvironment to increase the antitumor immunity of neoantigen cancer vaccines in combination with radiotherapy.

Fig. 4: Blockade of B7-H3 enhanced systemic antitumor immunity to eradicate tumors in combination with a neoantigen cancer vaccine and local radiotherapy.
figure 4

a Scheme and regimen of AAV vaccination, radiotherapy and anti-B7-H3 antibodies. A total of 3 × 105 CT26 cells were subcutaneously inoculated into the right leg of BALB/c mice, and AAV (1 × 108 vg) particles were intramuscularly injected on days 5, 12, 19 and 24. Local radiotherapy (5 Gy) of the right leg was conducted on days 11, 14 and 17. Anti-B7-H3 antibodies (100 μg/mouse) were administered intraperitoneally on days 7, 10, 13 and 16. The tumor volume was measured every 3 days (n = 5–6/group). *p < 0.05. One-way ANOVA. b The relative tumor volume was measured on day 30 (n = 6-7/group). *p < 0.05, **p < 0.01 and ***p < 0.001. One-way ANOVA. c Representative images of apoptotic cells (cleaved caspase-3), CD11c+ DCs and GzmB+ immune cells within the tumor microenvironment were evaluated via immunofluorescence microscopy. d The quantification of apoptotic cells (cleaved caspase-3+ cells) is shown (n = 3). *p < 0.05 and **p < 0.01. One-way ANOVA. e The quantification of CD11c+ DCs is shown (n = 3). *p < 0.05 and **p < 0.01. One-way ANOVA. f The quantification of GzmB+ immune cells is shown (n = 3). **p < 0.01 and ***p < 0.001. One-way ANOVA. g The mRNA levels of the cytotoxic T-cell marker GzmB in resected tumors are shown (n = 3). *p < 0.05 and **p < 0.01. One-way ANOVA. h The mRNA levels of Ifnα2 in resected tumors are shown (n = 3). *p < 0.05 and **p < 0.01. One-way ANOVA. i The mRNA levels of Ifnβ1 in resected tumors are shown (n = 3). *p < 0.05 and **p < 0.01. One-way ANOVA.

Full size image
Fig. 5: Blockade of B7-H3 enhances the extent of the neoantigen-specific T-cell response to neoantigen-based cancer vaccines and local radiotherapy.
figure 5

a The mRNA levels of the exhausted T-cell marker Tim3 in resected tumors are shown (n = 3). *p < 0.05. One-way ANOVA. b The mRNA levels of the exhausted T-cell marker Tigit in resected tumors are shown (n = 3). *p < 0.05. One-way ANOVA. c The percentage of CD4+CD45+CD3+ T cells within tumors was analyzed by flow cytometry (n = 3-5). *p < 0.05. One-way ANOVA. d The percentage of CD8+CD45+CD3+ T cells within tumors was analyzed by flow cytometry (n = 3–5). *p < 0.05. One-way ANOVA. E. The percentages of CD4+CD44+CD62L-CD45+ cells (effector/memory CD4+ T cells) and CD8+CD44+CD62L-CD45+ cells (effector/memory CD8+ T cells) within tumors were analyzed by flow cytometry (n = 3-5). *p < 0.05. One-way ANOVA. f The percentage of NKp46+CD19-CD3-CD45+ cells (NKp46+ NK cells) within tumors was analyzed by flow cytometry (n = 3–5). *p < 0.05. One-way ANOVA. g The percentage of CD27+CD11b+NKp46+CD19-CD3-CD45+ cells (cytotoxic NK cells) within tumors was analyzed by flow cytometry (n = 3–5). *p < 0.05. One-way ANOVA.

Full size image

We then examined the profile of tumor-infiltrating lymphocytes, including CD4, effector/memory CD4, CD8, effector/memory CD8 and NK cells (Supplementary Fig. 4a-4b). We found significant differences in the numbers of CD4 T cells, effector/memory CD4 T cells, CD8 T cells, and effector/memory CD8 T cells (Fig. 5c–e). Additionally, there was a significant difference in the density of NKp46+NK cells (Fig. 5f) and activated CD27+CD11b+NKp46+NK cells (Fig. 5g). We then examined whether blockade of B7-H3 enhances the splenic T cell response to H2-Kd-restricted neoantigens after treatment. We detected apparent mE2F8-, mMtch1- and mGlud1-specific IFNγ+CD8+ T cell responses (Fig. 6a–c). Moreover, the splenocytes from AAVax/RT/anti-B7-H3 antibody-treated mice also had greater cytotoxic effects on CT26 cells (Fig. 6d), suggesting that the inhibition of B7-H3 enhanced the apparent neoantigen-specific T-cell responses induced by AAVax in combination with RT. Taken together, these results indicate that the blockade of B7-H3 may increase antigen presentation to augment neoantigen-specific T-cell responses, increasing the therapeutic benefit of radiotherapy and neoantigen-based cancer vaccines.

Fig. 6: Blockade of B7-H3 enhanced systemic antitumor immunity to inhibit lung metastasis.
figure 6

a The splenocytes were harvested for H2-Kd-restricted mE2f8 neopeptide stimulation. Representative images of IFNγ+ ELISPOT from splenocytes are shown. *p < 0.05 and ***p < 0.001. One-way ANOVA. b The splenocytes were harvested for H2-Kd-restricted mMtch1 neopeptide stimulation. Representative images of IFNγ+ ELISPOT from splenocytes are shown. *p < 0.05 and ***p < 0.001. One-way ANOVA. c The splenocytes were harvested for H2-Kd-restricted mGlud1 neopeptide stimulation. Representative images of IFNγ+ ELISPOT from splenocytes are shown. *p < 0.05 and ***p < 0.001. One-way ANOVA. d The splenetic T cells were harvested and incubated with CT26 cells for 9 h. The cytotoxicity ability was evaluated by a CCK8 assay. *p < 0.05. e Scheme and regimen of AAV vaccination, radiotherapy and anti-B7-H3 antibodies. A total of 3 × 105 CT26 cells and 1 × 105 CT26 were subcutaneously and intravenously inoculated into the right leg of BALB/c mice, and AAV (1 × 108 vg) particles were intramuscularly injected on days 5, 12, 19 and 24. Local radiotherapy (5 Gy) of the right leg was conducted on days 11, 14 and 17. Anti-B7-H3 antibodies (100 μg/mouse) were administered intraperitoneally on days7, 10, 13 and 16. f Representative images of lung metastatic tumors were evaluated. The quantification of the lung weight is shown (n = 3). *p < 0.05. One-way ANOVA. g The quantification of the lung metastasis index is shown (n = 3). *p < 0.05. One-way ANOVA.

Full size image

Blockade of B7-H3 increased the protective antitumor immunity of neoantigen-based cancer vaccines and radiotherapy to inhibit lung metastasis

Our previous studies demonstrated that a neoantigen cancer vaccine in combination with local radiotherapy can induce systemic antitumor immunity to inhibit tumor relapse and distant metastasis in murine colon and triple-negative breast cancer models16,29. Therefore, we aimed to evaluate whether the inhibition of B7-H3 could enhance systemic antitumor immunity to inhibit distant metastasis and improve long-term survival. Local radiotherapy, B7-H3 blockade and neoantigen-based cancer vaccines were administered on the indicated days after subcutaneous and metastatic tumor inoculation (Fig. 6e). To determine whether the inhibition of B7-H3-induced strong systematic or long-term antitumor immunity, resected lungs were used to estimate the strength of the immune response after treatment. Three days after local tumor inoculation with 3 × 105 CT26 cells, 1 × 105 CT26 cells were intravenously infused. Lung metastatic nodules were counted on day 30 after tumor inoculation. All control mice developed tumor nodules in the lungs (Fig. 6f). A substantial number of lung tumor nodules were observed in the AAVax/RT group. In contrast, no lung metastasis was detected in the AAVax/RT/anti-B7H3 group (Fig. 6g), indicating that B7-H3 inhibition generated a strong systematic antitumor immune response and long-term antitumor immunity that inhibited lung metastasis.

Discussion

Despite advances in immunotherapy, particularly the approval of ICB therapies for several types of tumors, MSS-CRC remains generally resistant to immunotherapies due to its immunosuppressive TME45. In our study, B7-H3 was highly expressed in patients with advanced CRC after neoCRT treatment, which was negatively correlated with the tumor regression grade (TRG) to neoCRT. Tumor B7-H3 promoted tumor progression by attenuating the infiltration of dendritic cells and the cytotoxicity of CD8+ T cells by decreasing the presentation of antigens, including neoantigens. Knockdown of tumor B7-H3 significantly increased the extent of neoantigen-specific T-cell responses induced by radiotherapy. Therapeutically, B7-H3 blockade augments the neoantigen-specific T-cell response to neoantigen cancer vaccines in combination with radiotherapy, leading to delayed tumor growth and reduced lung metastasis. Taken together, these results highlight B7-H3 as a potential target to increase the clinical benefit of immunotherapy, such as neoantigen cancer vaccines, in cold tumors such as MSS-CRC.

Several intrinsic factors within the tumor microenvironment may constrain the process of antigen presentation, such as mutations in TAPs and loss of MHC-I expression. The modulation of DCs by immune checkpoint proteins also influences their antigen presentation ability. High expression of PD1/PD-L1 and CTLA-4 on host DCs attenuates their capacity for antigen presentation to stimulate an effective antitumor response26,27,46,47. PD-L1 on DCs is upregulated during antigen presentation to prevent T-cell-mediated cytotoxicity27. B7-H3 is also expressed on host DCs to attenuate their antigen presentation ability to promote tumor progression by inhibiting the functions of CD8+ T cells35,43. In addition, tumor-intrinsic PD-L1 weakens antigen presentation for acquired resistance to immunotherapy28, suggesting that tumor PD-L1 can increase tumor immune tolerance by decreasing neoantigen presentation. Knockout of tumor PD-L1 enhances robust CTL activation and produces a better neoantigen-specific T-cell response to eradicate tumor cells28, suggesting that immune checkpoint proteins play critical roles in regulating antigen presentation.

High expression of tumor B7-H3 has been reported to be associated with a reduced number of TILs48, a poor prognosis, and poor clinical outcomes49,50,51,52. In our study, we further demonstrated that robust upregulation of B7-H3 by radiotherapy in human colorectal cancer was greater than that of other family members, specifically PD-L1, PD-L2 and VISTA. We also found that the expression of tumor B7-H3 was greater than that of stromal B7-H3 in the TME after neoCRT treatment and was associated with the response to neoCRT, the infiltrating density of CD8+ T cells and survival outcome. Our results indicate that tumor B7-H3 elicited by radiotherapy may increase the tumor immune tolerance by reducing neoantigen presentation, DC maturation and T-cell activation in a murine animal model. Knockdown of tumor B7-H3 increased the response to radiotherapy and slowed tumor growth. Moreover, the density of activated CD80+ DCs, CD86+ DCs, CD8+ T cells, effector/memory CD8+ T cells and cytotoxic IFNγ+ CD8+ T cells in the tumor microenvironment was markedly increased by radiotherapy in low-B7-H3 tumors. Moreover, the extent of the neoantigen-specific T-cell response elicited by radiotherapy was also greater in low-B7-H3 tumors than in high-B7-H3 tumors, suggesting that B7-H3 tumors may increase the tumor immune tolerance by decreasing neoantigen presentation for immune evasion. In support of our findings, immunosuppressive cytokines such as IL-10 and TGF-β1 promote B7-H3 upregulation to inhibit the activity of multiple immune cells, including CD8+ T cells, γδ T cells, Vδ2 T cells, macrophages and dendritic cells53. Tumor B7-H3 binds to unidentified receptors on macrophages and dendritic cells to inhibit their antigen presentation ability53,54. Additionally, B7-H3 expression on DCs reduces the number of MHC–peptide complexes, leading to impaired T-cell stimulatory function and promoting tumor growth in multiple malignancies33,35,43. Moreover, the tumor B7-H3 level was also positively correlated with the frequency of exhausted T cells (PD1+TIM3+CD8+ T cells), suggesting that B7-H3 blockade can disrupt tumor immune tolerance to enhance the NeoAg-specific T-cell response and decrease T-cell exhaustion to effectively eliminate tumor cells. Notably, we did not observe differences in the infiltration of CD4, effector/memory CD4 or NK cells in low-B7-H3 tumor-bearing mice. These data indicate that in the CT26 MSS-CRC model, tumor-expressed B7-H3 plays a predominant role in CD8+ T cell-mediated cytotoxicity. However, we did not exclude out other mechanisms that B7-H3 might be participated in the tumor microenvironment. B7-H3 is expressed on various type of cells, acting as a driver of immune evasion55,56,57. High B7-H3 expression in tumor-associated macrophages (TAMs) stimulates aberrant angiogenesis to hamper intra-tumoral infiltration of CD8+ T cells58. Moreover, increased B7-H3 activity promotes macrophage polarity toward pro-tumor immunosuppressive macrophage (M2) phenotype. Cancer stem cells (CSCs) also dampen immune responses and promote tumor escape from immune surveillance by expressing B7-H356,59,60. These studies indicated that targeting B7-H3 as a desired approach in cancer immunotherapy to overcome the immunosuppressive TME.

Tumor B7-H3 inhibits T-cell-mediated antitumor immunity, leading to insensitivity to PD-1 blockade therapy in ovarian cancer33. Moreover, Wang et al. reported that irradiation increases the level of B7-H3 in different cancer cell lines, including pancreatic cancer, colorectal cancer and acute myelocytic leukemia. Preconditioned irradiation significantly enhances the ability of B7-H3-CAR-T cells to infiltrate and kill tumors and enhances CAR-T-cell tumor-killing ability on the nonirradiated side, confirming the abscopal effect of irradiation in combination with CAR-T-cell therapy61. These results of studies were consistent with our findings that elevated tumor B7-H3 was observed in post-CRT surgical tissues and was associated with the response to CRT. However, little is known regarding whether B7-H3 expression in tumors controls T-cell immunity against neoantigens, which may partly determine the response to neoantigen cancer vaccines. A recent elegant study by Okada revealed that somatic-mutation immunogenic neoantigens enhanced CTL responses and that the number of TCR repertoires relevant to the neoantigens increased due to PD-L1 deficiency28, which resulted in better therapeutic efficacy of multiple neopeptide-pulsed DC vaccines. These results support our findings that these immune checkpoint proteins may not only attenuate T-cell-mediated antitumor immunity but also decrease the antigen presentation ability, including that of somatic mutation-derived neoantigens.

Taken together, our findings provide tumor immunological evidence that the level of T-cell response to neoantigens and B7-H3 expression in tumors determine the positive and negative cancer immune cycles, and therefore, may shape immunoediting during tumor occurrence, which must be optimally targeted for clinical response. Therefore, targeting B7-H3 may increase either the therapeutic efficacy of radiotherapy or neoantigen-based cancer vaccines, providing alternative therapeutic strategies for poor immunologic, immune-resistant cancers, such as MSS-CRC.

Materials and methods

Mice

Six-week-old female BALB/c mice were purchased from Biolasco. Taiwan Co., Lt. All animal experimental procedures were approved by the China Medical University animal care committee (CMUIACUC-2018-182) and were conducted in accordance with the Guide for the Care and Use of Laboratory Animals published by the National Research Council (NIH Publication No. 85-23, revised 1996).

Cell culture

The CT26 cell line, BALB/c mouse undifferentiated colon carcinoma, and human colorectal cancer HCT116 cell line were purchased from ATCC17. The cells were cultured in complete RPMI1640, supplemented with 10% fetal bovine serum, 2 mM L-glutamine, and 1% (v/v) penicillin/streptomycin and maintained at 37 °C in 5% CO2. These cell lines were not further authenticated but were cultured for a limited number of passages. The cell lines were routinely tested for the absence of mycoplasma contamination by PCR. For X-irradiation experiments, the medium was replaced prior to irradiation in a horizontal position. Cells were exposed under (0, 2.5 and 5 Gy) X-rays separately (6 MV X-ray with 400 MU/min, TrueBeam, Varian).

Construction of a tissue microarray (TMA) and immunohistochemistry

Tissue microarrays were constructed from 106 pre-CRT biopsies and 119 post-CRT surgical tissue samples from patients with rectal cancer, and other samples were not available (data not suitable for IHC) as previously described in refs. 62,63. This study was reviewed and approved by the Internal Review Board (IRB) of China Medical University Hospital [Protocol number: CMUH105-REC2-072].

Immunohistochemistry (IHC) was performed on 3-μm-thick histological TMA sections as previously described in refs. 45,62,64. TMA slides were stained individually with horseradish peroxidase-conjugated avidin biotin complex (ABC) using the Vectastain Elite ABC Kit (Vector Laboratories, Burlingame, CA, USA) and DAB (Vector Laboratories) and counterstained with hematoxylin. The following antibodies were used in this study: CD8 (ab4055, Abcam, Cambridge, UK) and B7-H3 (ab134161, Abcam). The stained tissue sections were scored according to the histo-score (H-score), which involves a semiquantitative assessment of both the intensity of staining (graded as 0, nonstaining; 1, weak; 2, medium; or 3, strong) and the percentage of positive tumor cells. The H-scores ranged from 0 to 300. The level of tumor B7-H3 was categorized as low or high according to the median H-score.

Intratumoral CD8+ T lymphocyte staining was positive when detected in the cytoplasm or at the cell membrane of tumor-infiltrating lymphocytes (TILs) and was evaluated using microscopy (OLYMPUS BX53, Tokyo, Japan) according to the intensity of the TILs. These tissues were reviewed at 40× magnification, and the area with the highest density of TILs adjacent to the malignant cells was counted (number of TILs/mm2). The average number of TILs in five high-power fields was included in the evaluation and categorized as low or high according to the median density of TILs/mm2 65.

Design of the AAV-based cancer vaccine (AAVax)

Codon-optimized CT26 neoantigen cassette sequences were fused by RERK linkers, synthesized, cloned and inserted into a CMV-driven pAAV-CMV expression vector as previously described16,29. The neopeptide sequences were optimized and linked to the amino acid spacer REKR, which is recognized by the furin protease as previously described40,41. The IL12 signal peptide sequence was included in front of the neoantigen cassette to increase neoantigen secretion. The control AAV vector was the empty AAV-CMV vector. The corresponding peptides for these neoepitopes were synthesized as peptide controls for all of the experiments.

Generation and purification of the AAV-neoAg cancer vaccines

All of the viruses were produced by the triple transfection, in 293 T cells, including AAV-Vec (AAV-Vector), AAVax (AAV2-CT26-neoAg), and the helper plasmids pRC2-miR342 and pHelper. Seventy-two hours after transfection, the cells were collected by centrifugation, and recombinant AAV2 vectors were produced and purified using an AAVpro purification kit (6232, Takara, Japan). AAV2 titration was performed by quantitative PCR (qPCR) of the vector genomes.

In vivo tumor growth of the AAV-based cancer vaccines

Adult female BALB/c mice (6 weeks old) were used. Animals were randomly assigned to different treatment subgroups. All experiments were approved by the China Medical University Institutional Animal Care and Use Committee [Protocol No. CMUIACUC-2022-244-1]. Briefly, 100 μL of 50% Matrigel (Corning, CA, USA) containing 3 × 105 fresh CT26 cells was inoculated subcutaneously into the right leg of each mouse17. After 8 days, when the tumor volume reached 70–100 mm3, the mice were randomly assigned to different groups. For radiotherapy, the mice were anesthetized with 300 μL of PBS with ketamine (140 mg/kg) and xylazine (3 mg/kg) by intraperitoneal injection prior to irradiation. Before radiotherapy, dosimetry data on the irradiated square (8.5 cm × 8.5 cm, depth 5 cm) were collected to validate the dosage of irradiation (293.2 ± 4.4 cGy/300 MU). The local tumors were then subjected to radiotherapy (6 MV X-ray with 400 MU/min, TrueBeam, Varian) according to different protocols. The tumor volume was measured every 3 days throughout the study. The longest and shortest diameters (L and W, respectively) of the tumors were measured using Vernier calipers (Sata, Shanghai, China) every 3 days, and the tumor volume (V) was calculated using the following formula: V = (L ×W2)/2. The mice were euthanized and sacrificed at the termination of the experiments by displacement of air with 100% CO2 for 5 min in the cage, and the tumor tissues were collected for lysis and subjected to immunoblotting analysis and immunohistochemistry. AAV-Vec and AAVax cancer vaccines were administered via intramuscular injections in the quadriceps by delivering a volume of 50 μl per side at 1 × 108 vg on the indicated days. The mice were sacrificed when any side of the tumor reached 20 mm.

Combined treatment with anti-B7-H3 blockade and an AAV-based cancer vaccine

Briefly, 100 μL of 50% Matrigel (Corning, CA, USA) containing 3 × 105 fresh CT26 cells was inoculated subcutaneously into the right leg of each mouse. After 7 days, when the tumor volume reached 70–100 mm3, the mice were randomly assigned to different groups. Local irradiation (5 Gy) was performed on days 11, 14 and 17. The anti-B7-H3 antibodies were administered on days 7, 10, 13 and 17 (100 μg/mouse, intraperitoneal injection, 4 times with 3-day intervals between injections; Bio×Cell clone MJ18; NH, USA). The AAVax cancer vaccines were administered via intramuscular injections in the quadriceps by delivering a volume of 50 μl per side at 5 × 109 vg on the indicated days. The tumor volume was measured every 3 days throughout the study. The tumors were harvested on day 30 for immunohistochemistry, western blotting and qRT–PCR.

Ex vivo immune analysis

IFN-γ ELISpot assays were performed on single-cell suspensions of spleen cells according to the manufacturer’s manual as previously described in ref. 29. Mouse splenocytes were plated in duplicate at two different cell densities and stimulated overnight with single 25-mer peptides at a final concentration of 1 μg/ml. The peptide diluents dimethyl sulfoxide (Sigma-Aldrich) and concanavalin A (Sigma- Aldrich) were used as negative and positive controls, respectively. The plates were subsequently incubated with a biotinylated anti-mouse IFN-γ antibody, conjugated with streptavidin–alkaline phosphatase and 5-bromo-4-chloro-3-indoyl-phosphate/nitro blue tetrazolium 1-step solution (Thermo Fisher Scientific). An automated enzyme-linked immunosorbent-spot assay video analysis system automated plate reader was used to analyze the plates (Immunospot S6 LITE M2). ELISpot data are expressed as the number of IFN-γ SFCs per million splenocytes.

Evaluation of immune cell infiltration in vivo

The tumors were isolated from the mice, weighed, and placed in Petri dishes containing blank RPMI media at room temperature to prevent dehydration17. The tumors were minced into small pieces (1–2 mm) with a beaver blade, filtered through a 70 μm strainer, spun down, and then resuspended in blank RPMI media. Thereafter, the cell suspensions were layered over Ficoll-Paque media and centrifuged at 1025 × g for 20 min. The layer of mononuclear cells was transferred into a conical tube and 20 ml of complete RPMI media was added before the mixture was gently mixed and centrifuged at 650 × g for 10 min twice. Finally, the supernatant was removed, and the TILs were resuspended in complete RPMI media.

Then, the TILs were resuspended in 500 μL of staining buffer (2% BSA and 0.1% NaN3 in PBS). The cells were stained with a surface marker panel containing CD8a (551162, BD PharMingen, CA, USA), CD44 (E-AB-F1100C, Elabscience, Texas, USA), and CD45 (E-AB-F1136D, Elabscience, Texas, USA) and their isotypes. A PE rat IgG2b, κ isotype control was included for CD45 and FITC rat IgG2b (E-AB-F09842D, Elabscience, Texas, USA), and FITC rat IgG2b κ isotype control was used for CD44 (E-AB-F09842C, Elabscience, Texas, USA). The samples were analyzed using a Guava® easyCyte™ Flow Cytometer (Luminex, CA, USA).

For immunohistochemical analysis, 3 μm-thick histological tissue sections were stained individually with horseradish peroxidase-conjugated avidin biotin complex (ABC) using the Vectastain Elite ABC Kit (Vector Laboratories, Burlingame, CA) and DAB chromogen (Vector Laboratories) and counterstained with hematoxylin. The following antibodies were used in this study: granzyme B (ab4059, Abcam), CD11c (ab219799, Abcam), PD-1 (ab214421, Abcam) and cleaved caspase-3 (ab32042, Abcam).

Statistical analysis

IBM SPSS statistical software version 26 was used to perform the statistical analyses. All tests reported a two-sided p value with a significance level set at 0.05. Student’s t test, the Pearson chi-square test and Fisher’s exact test were used for group comparisons. The Kaplan‒Meier method was used to estimate ten-year disease-free survival (DFS). Survival time was defined as the time from surgery until the event, including local recurrence, distant metastasis and death. Between-group comparisons were performed using an unpaired t-test and ordinary one-way ANOVA (including Dunnett’s and Tukey’s multiple comparison tests), and the two-sided p value was reported for all tests using GraphPad Prism 7 statistical software (GraphPad Software, San Diego, CA, USA).