Exploration of the immunologic characteristics of KIT/PDGFRA wild-type gastrointestinal stromal tumor and potential application of neoantigen vaccination

To the Editor: Gastrointestinal stromal tumors (GISTs) are the most common mesenchymal tumors of the gastrointestinal tract. GISTs mostly occur in the stomach or small intestine among patients around 40–60 years of age, and are insensitive to chemotherapy or radiotherapy. Approximately 10% and 85% of adult and pediatric patients, respectively, with GISTs do not harbor a mutation in either KIT or PDGFRA, thereby being defined as KIT/PDGFRA wild type.^[1] According to succinate dehydrogenase subunit B (SDHB) immunohistochemical (IHC) staining, KIT/PDGFRA wild-type-GISTs (WT-GISTs) can be divided into two groups: succinate dehydrogenase (SDH)-deficient GIST, characterized by the absence of SDHB expression and mutations of SDH subunits (SDHX, i.e. SDHA, SDHB, SDHC, and SDHD), especially SDHA and epimutation of SDHC methylation; and SDH non-deficient GIST, characterized by the absence of SDHX mutation and mutations of other genes, mostly involved in the RAS-signaling pathway [Supplementary Figure 1, http://links.lww.com/CM9/C138].

Tyrosine kinase inhibitors (TKIs) targeting KIT/PDGFRA are seldom effective in WT-GISTs, and trials using TKIs, including imatinib, are often done with small samples or negative results.^[2,3] Therefore, it is necessary to explore new methods for the treatment of WT-GISTs, particularly for recurrent or unresectable cases. Immunotherapy has witnessed significant advancements in recent years, and immune checkpoint inhibitors (ICIs) have been reported to achieve durable clinical responses in several cases of WT-GIST.^[4] Additionally, neoantigen vaccines targeting tumor-specific antigens have also achieved predominant efficacy in solid tumors.^[5,6] However, immunotherapy for WT-GISTs is still limited due to a lack of comprehensive understanding regarding the immunological characteristics of WT-GIST. To better clarify the immunologic characteristics of WT-GIST, which will potentially benefit from immunotherapy, we characterized comprehensive genomic features of 142 Chinese GIST patients by next-generation sequencing (NGS).

This study was performed in accordance with the Declaration of Helsinki (as revised in 2013). Written informed consent was obtained from all patients. The study has been reviewed and approved by the Institutional Review Board of Nanjing Drum Tower Hospital, the Affiliated Hospital of Nanjing University Medical School (No. 2021-027-02).

This cohort included 79 males and 63 females, diagnosed between 2008 and 2021. Immunotherapy-related clinicopathological features, mutations, and markers were recorded and retrospectively reviewed in 142 patients [Supplementary Methods, http://links.lww.com/CM9/C138]. The median age of patients was 57 years (range: 26–82 years). Patients were divided into two groups: WT-GISTs (35/142) and GISTs with KIT/PDGFRA mutations (107/142). WT-GIST is more likely to occur in young patients (P = 0.033). Among the eight patients ≤40 years old, five were wild-type, and three of these five young wild-type patients were female [Supplementary Table 1, http://links.lww.com/CM9/C138].

According to the risk stratification of GIST,^[7] 33 (23.24%) patients were classified as high-risk (stage IV). Patients with WT-GISTs (25.71%, 10/35) had higher risk than those with KIT/PDGFRA-mutant GISTs (21.50%, 23/107), but there was no significantly statistical difference (P = 0.251). Most KIT/PDGFRA-mutant GISTs were primarily gastric tumors (31.78%, 34/107), similar to the WT-GISTs (37.14%, 13/35).

Using immunohistochemistry (IHC) staining, we analyzed the cluster of differentiation 117 (CD117), discovered on gastrointestinal stromal tumors protein 1 (DOG1), and Ki67 expression statuses of the patients. All patients with available CD117 expression status were positive. Almost all patients (96.67%, 58/60) with KIT/PDGFRA mutations exhibited positive DOG1 expression, while a majority of WT-GISTs patients (81.25%, 13/16) also showed positive DOG1 expression. In terms of the Ki67, 26.67% (4/15) and 50.00% (30/60) of patients had a Ki67 of ≥10% in WT-GISTs and KIT/PDGFRA-mutant GISTs, respectively (P = 0.080).

We then compared the differences in the most frequent genomic alterations (GAs, ≥10%) between the two groups. Of the 512 mutations detected in KIT/PDGFRA mutant group, the most frequent were KIT (90.65%, 97/107), CDKN2A (21.50%, 23/107), CDKN2B (17.76%, 19/107), and PDGFRA (11.21%, 12/107). The most frequent GAs in the WT-GIST group were TP53 (28.57%, 10/35), OBSCN (20.00%, 7/35), MUC16 (14.29%, 5/35), FANCA (11.43%, 4/35), FAT3 (11.43%, 4/35), LRP1B (11.43%, 4/35), and ROS1 (11.43%, 4/35) [Supplementary Table 2, http://links.lww.com/CM9/C138].

The SDHB expression status was also explored using IHC staining, and SDHX mutations were detected by NGS. Among WT-GIST patients, nine (25.71%) patients with available SDHB IHC statuses positively expressed SDHB were categorized as having SDH-non-deficient GIST. Among these patients, one harbored NF1 and one harbored the BRAF p.V600E mutation. Of the 26 patients whose SDHB expression was unknown, only one patient had SDHA mutation, and the remaining patients harbored no SDHX mutations; however, one case was found to positively express NF1 mutations; mutations in KRAS and PIK3CA were also detected in this patient. Additionally, we identified two cases of ETV6/NTRK3 fusion and one case of KRAS amplification [Supplementary Figure 2 and Supplementary Table 3, http://links.lww.com/CM9/C138].

Programmed cell death-ligand 1 (PD-L1) expression, microsatellite instability (MSI) status, and tumor mutation burden (TMB) are generally recognized biomarkers related to the effectiveness of immunotherapy in solid tumors. WT-GISTs showed a slightly higher PD-L1 positive rate (33.33%, 2/6) than those with KIT/PDGFRA mutations (19.44%, 7/36) (P = 0.818). Notably, both WT-GIST patients with positive PD-L1 expression belonged to the SDH non-deficient subtype, and one patient harbored the NF1 mutation, while the other was sporadic wild-type, with other mutations, including TP53 [Supplementary Figure 2, http://links.lww.com/CM9/C138]. Microsatellite stable (MSS) was observed in all cases with available data. Median TMB of all 142 cases was 1.5 muts/Mb (range 0–15.5 muts/Mb). The median TMB value in WT-GISTs was 2.3 muts/Mb (range 0–10.5 muts/Mb), which was slightly higher than that in the KIT/PDGFRA mutant GISTs (1.2 muts/Mb, range 0–15.5 muts/Mb) [Supplementary Figure 3, http://links.lww.com/CM9/C138].

We also analyzed other potential biomarkers that have been shown to correlate positively or negatively with the effectiveness of immunotherapy and hyper-progressive disease (HPD)-related markers [Supplementary Table 4, http://links.lww.com/CM9/C138]. Generally, compared to KIT/PDGFRA mutant GISTs, WT-GISTs had higher gene mutation frequency in positive markers of immunotherapy (6.20% vs. 2.73%, for WT-GISTs vs. KIT/PDGFRA mutant GISTs, P = 0.001) and fewer gene mutations frequency in negative markers of immunotherapy (5.00% vs. 12.27%, for WT-GISTs vs. KIT/PDGFRA mutant GISTs, P = 0.044). A lower proportion of mutated HPD-related markers (1.67% vs. 18.94%, P = 0.001) were also found in WT-GISTs [Supplementary Table 1 and Supplementary Figure 4, http://links.lww.com/CM9/C138].

We further analyzed the correlation between TMB and neoantigen mutation burden (NAB) of patients available. The number of neoantigens predicted from mutations detected in WT-GISTs ranged from 0 to 90 and was positively correlated with TMB [Supplementary Figure 5, http://links.lww.com/CM9/C138], indicating that patients with WT-GISTs and high TMB may potentially be ideal candidates for neoantigen vaccination. We further predicted potential neoantigen peptides with strong, major histocompatibility complex (MHC) binding ability, i.e., %rank below 0.5 or binding affinity less than 50 nmol/L. Candidate neoantigen peptides were then prioritized and screened based on both the mutational frequency and the number of human leukocyte antigen (HLA) alleles to which they could bind. Among the GAs that are widely acknowledged as driving mutations in WT-GIST, one sample harboring PIK3CA mutations was predicted to have neoantigens with high binding affinity. The corresponding peptide was MALNLWPVL, predicted from PIK3CA P449L and targeting HLA-C*03:04, HLA-C*08:01, and HLA-B*13:01. We also identified two other peptides, MNRRPTIITL and NRRPTIITL, targeting HLA-C*06:02 and HLA-C*06:08, in a patient carrying the TP53 I251_L252 deletion mutation [Supplementary Tables 5 and 6, http://links.lww.com/CM9/C138].

For patients with advanced or unresectable GIST, the application of TKIs targeting KIT and/or PDGFRA has significantly improved clinical outcomes. KIT and PDGFRA have received more attention; however, other genes with lower mutation rates are non-negligible. In this study, we focused on the unique characteristics of KIT/PDGFRA-wild-type GISTs.

A total of 35 patients with WT-GIST were analyzed in our cohort, accounting for 24.65% of all the enrolled adult patients with GIST. WT-GISTs in our cohort were more likely to occur in young females, which generally had higher stages, and tended to be located primarily in the stomach. All the nine patients with WT-GIST, whose SDHB IHC statuses were available, were categorized as having SDH-non-deficient GIST. As SDH deficiency is closely related to younger female patients with primary gastric tumors, higher chances of nodal involvement or metastasis, and thus, higher stages,^[8] the clinical characteristics of our cohort may be attributed to the dominance of SDH-deficient GISTs.

Mutations in TP53, OBSCN, MUC16, FANCA, FAT3, LRP1B, and ROS1 are most frequently found in WT-GISTs. Detection of GAs provides new indications for the treatment of WT-GIST. There are also opportunities for therapies other than TKIs to target these potential targets. We identified several patients with ETV6/NTRK3 fusion, KRAS amplification, and BRAF p.V600E mutation. One patient harbored a coexisting mutation in NF1, KRAS, and PIK3CA, which has rarely been reported. In additional, the application of novel, small-molecule inhibitors targeting BRAF^[9] and NTRK mutations^[10] has rarely been reported. Although, the PIK3CA inhibitors—alpelisib or buparlisib—have been administered, no additional benefits have been observed.^[11,12]

Although the application of targeted therapy was limited in WT-GISTs, a series of ongoing clinical trials have indicated the potential efficacy of immunotherapy, especially ICIs.^[4] However, few studies have focused on the immune microenvironment of WT-GIST because of its low incidence, even though patients with WT-GIST urgently need new therapeutics. Limited data show that PD-L1 is heterogeneously expressed in GIST.^[13,14] In our study, WT-GISTs showed high PD-L1 expression rates and TMB. For other markers related to the immunologic response, WT-GISTs expressed more positive, predictive markers, fewer negative markers, and fewer HPD-related markers, which implies that WT-GISTs may potentially benefit more from immunotherapy than KIT/PDGFRA-mutant compartments.

Due to the low incidence of WT-GIST, the small sample size of this study failed to thoroughly discuss the immune characteristics of the different subtypes of WT-GISTs. Moreover, we focused mainly on gene mutations based on the data collected from NGS, and the relationship between molecular characteristics and the efficacy of immunotherapy in clinical practice requires further exploration.

Through NGS and neoantigen prediction, we identified three neoantigen peptides with potentially high mutation frequency and binding affinity, including MALNLWPVL predicted from PIK3CA P449L, MNRRPTIITL, and NRRPTIITL predicted from the TP53 I251_L252 deletion. Identification of these neoantigen peptides may contribute to the clinical application of neoantigen vaccines in future cases of WT-GISTs.

Acknowledgements

The authors are grateful to all patients who participated in our study.

Funding

This research was supported by the Natural Science Foundation of China (Nos. 82203178; 82272852; and 82103687), Jiangsu Provincial Key Research and Development Program (No. BE2020619), Natural Science Foundation of Jiangsu Province (No. BK20220191), and Nanjing Medical Science and Technology Development Foundation (No. YKK22086).

Conflicts of interest

None.

Supplementary Material

cm9-137-2627-s001.docx ^{(1.1MB, docx)}