MSI-H GI and Endometrial Tumors Show High Immunogenicity With Frameshift Mutations

Nina Bhardwaj, MD, PhD

Microsatellite instability–high (MSI-H) tumors in patients with endometrial, colorectal, and stomach carcinomas were found to be enriched in recurrent shared, immunogenic, tumor-specific frameshifts expressed on RNA and protein levels, according to a recent study published in Cell.

Moreover, these frameshifts can elicit impressive T-cell responses and encode multiple epitopes that originate from indel mutations. These epitopes occur in a large proportion of patients with MSI-H disease and tumor subclones and can bind to the most frequent major histocompatibility complex (MHC) alleles.

Additionally, neoantigens that arise from the frameshift mutations are distinct and differentiated from self and viral antigens. This suggests that these neoantigens are highly immunogenic and, notably, may be suitable for use in developing “off-the-shelf” cancer vaccines.

“These classes of antigens are very immunogenic and can be used potentially for the design of common off-the-shelf cancer vaccines and eventually for T-cell therapy,” said senior author of the study Nina Bhardwaj, MD, PhD, director of the Immunotherapy Program at the Tisch Cancer Institute at Mount Sinai.

“We need to determine in larger cohorts whether these antigens are broadly recognized, if T cells that are specific for them gain entry into the tumor site, whether they are responsible for controlling tumor growth and whether they can be combined with standard-of-care checkpoint blockade to further induce long-term tumor control,” Bhardwaj explained.

Whole-exome sequencing data from The Cancer Genome Atlas (TCGA) database was utilized to evaluate the relationship between MSI status and frameshift load. About 20% to 30% of patients with endometrial, colorectal, and stomach cancers within the TCGA are considered MSI-H (n = 338) rather than microsatellite stable (MSS) or status unknown.

Stratification of somatic mutation load showed that the majority of patients with endometrial, colorectal, or stomach cancer and high frameshift load, defined by frameshift count per patient, were also MSI-H. Moreover, endometrial, colorectal, and stomach cancers that are MSI-H tend to share frameshift events compared with cell lines derived from other tissues.

From these frameshift mutations, immunogenic epitopes can be derived. A frameshift neoantigen-calling pipeline developed for this study analyzed the distribution of frameshift mutations, peptides, and corresponding epitopes across MSI-H endometrial, colorectal, and stomach cancer cohorts of patients included in TCGA.

In the microsatellite (MS) regions of the tumors, multiple genes were commonly mutated via indels. About 80% of patients with MSI-H colorectal and stomach cancers shared frameshift genes, including ACVR2A, MIKI67, and RPL22. More than 50% of patients with MSI-H endometrial and colorectal cancers shared frameshift mutations, such as CASP5, MUC6, and KMT2C.

Shared frameshift peptides, defined as mutation events derived from the exact same frameshift mutation, were observed at lower frequencies across cohorts. For example, few frameshift peptides, such as RPL22and SETD1B, were shared among more than 40% of patients with MSI-H endometrial carcinoma. Similarly, frameshift peptides, such as SGOL1, SEC31A, and ACVR2A, in 50% of patients with MSI-H colorectal cancer and more than 50% of patients with MSI-H stomach cancer, were not frequently observed.

Notably, the average frequency of shared frameshift-gene, frameshift-peptide, and frameshift-neoepitope events in MSI-H colorectal and stomach cancers was double that in MSI-H endometrial cancer, suggesting that the rate of MS-related mutagenesis is influenced by the tumors’ varying tumorigenesis pathways and rates of cell growth.

A mutation ranking system was developed to identify immunogenic frameshift peptides based on 4 parameters: somatic score for each frameshift mutation, distribution of frameshift-peptide lengths, MHC allele groups, and total amount of peptide-MHC interactions per frameshift.

Utilizing these parameters to frameshift mutations shared by at least 20% of MSI-H patients identified 37 colorectal, 23 stomach, and 9 endometrial shared frameshift-peptides that encode poly-epitopes. Moreover, 5 unique peptide sequences SLC35F5, SEC31A, TTK, SETD1B, and RNF43, were shared among all patients analyzed irrespective of tumor type.

Specifically, in MSI-H endometrial cancer, the combination of neoepitope-yielding frameshift-peptides varied between patients. However, SEC31A and ASTE1 were frequently found together and were predicted to bind to frequently occurring and less frequently occurring MHC class 1 alleles. Prior to this finding, the characterization of frameshift-derived neoantigens was not well explored in endometrial cancer.

Additionally, in MSI-H endometrial cancer, allele frequencies were 30% to 40% higher on tumor specimens vs normal tissue specimens, which could decrease the probability that shared frameshift peptides are correctly translated. No correlation between RNA expression and shared frameshift load was observed, suggesting that frameshift genes were not selectively epigenetically silenced in tumors. However, further analyses showed significant and robust frameshift-allele expression in MSI-H patients vs MSS patients. Overall, the study authors concluded that frameshift mutations have the potential to be correctly expressed within endometrial cancer tumors.

These findings were further applied to enhance the understanding of immunotherapy in these patients. Although MSI-H tumors are characterized by high tumor mutational burden load, not all patients respond to immunotherapy with checkpoint inhibitors. This suggests that differences within the tumor play a role in immunotherapy responsiveness.

Analyses revealed that the shared frameshift neoantigen–load was higher in patients with MSI-H disease who were responsive to PD-1 inhibitors compared with unresponsive patients. Moreover, patients with MSS disease had undetectable frameshift neoantigens, whereas these were widely expressed in MSI-H disease. This confirms that combining anti–PD-1/PD-L1 therapy with a shared MSI-H vaccine could be beneficial for patients who are not responsive to traditional immunotherapy.

“Altogether, our data show that MSI-H patients have an increased frequency of high-quality T-cell epitopes derived from shared frameshift peptides, binding to a broad spectrum of MHC alleles, which are capable of inducing CD8+ T-cell responses,” the study authors wrote. “By applying such tailored vaccines for MSI-H endometrial, colorectal, and stomach carcinomas, one can potentially achieve immunological responses against existing neoplasms or develop preventive memory T-cell responses in high-risk patient populations, like those with Lynch syndrome.”

Reference

Roudko V, Bozkus CC, Orfanelli T, et al. Shared immunogenic poly-epitope frameshift mutations in microsatellite unstable tumors. Cell. 2020;183(6). doi:10.1016/j.cell.2020.11.004