Prevalence and Associations of Beta2-Microglobulin Mutations in MSI-H/dMMR Cancers

Abstract

Microsatellite instability (MSI) has emerged as an important predictor of sensitivity for immunotherapy-based strategies. β-2-Microglobulin (B2M) contains microsatellites within the coding regions and is prone to somatic changes in MSI/mismatch repair deficiency (MSI/dMMR) tumors. To delineate prevalence and associations of B2M mutations in MSI-H/dMMR cancers, we investigated the mutational profile of B2M and clinical and pathological features in gastric cancer (GC), colorectal cancer (CRC), and endometrial cancer (EC) with a high incidence of microsatellite instability-high (MSI-H)/dMMR. Formalin-fixed paraffin-embedded (FFPE) tumor tissues along with matched normal tissues were collected from 108 MSI/dMMR patients with GC, CRC, and EC. Genomic profiling of tissue and blood samples were assessed next-generation sequencing (NGS). Immunohistochemistry (IHC) was used to examine the presence or absence of B2M protein. Alternations in the exonic microsatellite regions of B2M were observed at various but high frequencies (57.5% in CRC, 23.9% in GC, and 13.6% in EC) and in different forms. NGS assay revealed that genes involved in chromatin regulation, the PI3K pathway, the WNT pathway, and mismatch repair were extensively altered in the MSI-H cohort. Signature 6 and 26, 2 of 4 mutational signatures associated with defective DNA mismatch repair, featured with high numbers of small insertion/deletions (INDEL) dominated in all 3 types of cancer. Alternations in the exonic microsatellite regions of B2M were observed at various but high frequencies (57.5% in CRC, 23.9% in GC, and 13.6% in EC) and in different forms. Tumor mutational burden (TMB) was significantly higher in the patients carrying MSI-H/dMMR tumors with B2M mutation than that in patients with wild-type B2M (P = .026).The frame shift alteration occurring at the exonic microsatellite sties caused loss of function of B2M gene. In addition, a case with CRC carrying indels in B2M gene resisted the ICI treatment was reported. In conclusion, patients carrying MSI-H/dMMR tumors with B2M mutation showed significantly higher TMB. Prescription of ICIs should be thoroughly evaluated for these patients.

To delineate the prevalence and associations of B2M mutations in MSI-H/dMMR cancers, the mutational profile of B2M in 3 cancer types with a high MSI-H/dMMR prevalence was investigated. This article reports the results.

Implications for Practice.

The mutational profile of B2M to describe the complex interplay between the genetic alteration of B2M and clinical-pathological features of microsatellite instability/mismatch repair deficiency (MSI/dMMR) tumors was investigated in this study. TMB was significantly higher in the patients carrying dMMR/MSI-H tumors with B2M mutation than in ones with wild type B2M. A case with CRC carrying indels in B2M gene resisted the ICI treatment was reported. In conclusionthe mutational status of the B2M gene is a point of consideration in the prescription of immunotherapy for MSI-H/dMMR patients and a potential predictor of the corresponding clinical outcome.

Introduction

With the increasing application of next-generation sequencing to cancer research, microsatellite instability-high (MSI-H)/mismatch repair deficiency (dMMR) has been found in numerous types of cancers at various frequencies. Colorectal cancer (CRC), gastric cancer (GC), and endometrial cancer (EC) are 3 high prevalent types of cancer with MSI-H/dMMR incidences ranging between 15% and 30%.¹ MSI-H/dMMR is known to cause accumulation of somatic mutations in tumor cells and, thus, leads to a range of molecular and biological changes, including high tumor mutation burden (TMB), increased expression of immunogenic neoantigens, high expression of immune checkpoint molecules and the attraction of numerous tumor-infiltrating lymphocytes.^2-4 A deficiency in mismatch repair could cause wide-spreading insertion/deletion (INDEL) and frameshift mutations in the exonic microsatellite repeats, which increases the mutation rate and confers the MSI-H phenotype.^5,6 Specifically, reading frameshifts occur when microsatellites sites locating in the protein-coding region are instable and the subsequent production of chimeric protein generates immune system recognizable neoantigens.⁷ Therefore, the successful treatment of MSI-H tumors with immune checkpoint inhibitors (ICIs) may be due to the recognition of neoantigens by tumor-infiltrating lymphocytes. The MSI/dMMR phenotype is associated with a better overall response rate (ORR) and overall survival (OS) upon treatment with ICI. MSI-H/dMMR has been approved as a predictor for clinical response to ICI therapy.^1,8,9

Recent studies have shown that a considerable portion of patients with MSI-H cancer are insensitive to ICI treatments. In addition, ICI-sensitive MSI-H tumors present with an initial decrease in tumor volume, but then relapse and exhibit acquired resistance as treatment progresses.¹⁰ The interferon signaling and antigen-presentation pathways have been proposed to be part of the mechanism underlying acquired resistance tumors.^11,12 Moreover, B2M, a gene plays a critical role in antigen presentation to cytotoxic CD8+ T cells and subsequent cell lysis, is an emerging ICI resistance predictive biomarker.^13,14 Human leukocyte antigen (HLA) class I complexes have a characteristic structure of a light chain (beta-2-microglobulin, B2M) overlapped by a heavy chain with an antigen-binding site. The B2M protein forms and stabilizes the trimeric major histocompatibility complex-peptide (MHCp) on the cell surface.^15,16 Once displayed on the cell surface, HLA class I-peptide complexes present tumor antigens to T cell receptors and activate cytotoxic CD8⁺ T cells. The loss of B2M decreases lymphocyte recognition of cancer cells and leads to immune escape in different tumor entities.¹⁷ MSI/dMMR leads to mutations in the exonic microsatellite sites of B2M gene, which results in the synthesis of a truncated B2M protein. In addition, mutations in the B2M gene due to MSI/dMMR can also cause the loss of HLA class I-mediated antigen presentation and impair the recognition of tumor cells by cytotoxic CD8⁺ T cells, resulting in acquired resistance to ICI.^14,17,18 For this, the relationship between the mutation status of B2M and the efficacy of immunotherapy against MSI-H tumors, as well as the clinical implications of this relationship, need full addressing.

To delineate the prevalence and associations of B2M mutations in MSI-H/dMMR cancers, we investigated the mutational profile of B2M in 3 cancer types with a high MSI-H/dMMR prevalence, including CRC, GC, and EC. Next-generation sequencing (NGS) analysis identified exonic microsatellite sites in B2M that allowed the inactivation of B2M in MSI-H/dMMR tumors. A CRC case with MSI-H/dMMR of primary resistance to ICI treatment caused by inactivation of B2M has been reported.

Materials and Methods

Patient Information

We collected 108 patients with stages I-IV GC(46), CRC(40), and EC(22) which were confirmed by histology from Drum Tower Hospital and Margaret Williamson Red House Hospital between 2013 and 2018. Patients with multiple primary tumors were excluded. Formalin-fixed paraffin-embedded (FFPE) of GC, CRC, and EC samples (dMMR determined by IHC and MSI-H determined by PCR) were collected along with paracancerous tissues or peripheral blood.^19,20 All samples underwent comprehensive genomic profiling by NGS. Written informed consent was provided. This study has been approved by the Ethics Committee of Drum Tower Hospital Affiliated to the Medical School of Nanjing University (Approval ID: 2019-037-02).

Next-Generation Sequencing-Based Assessment of Genomic Characteristics

Patients’ information and test details are listed in Supplementary Table S1. Formalin-fixed paraffin-embedded (FFPE) tumor samples with para cancerous tissues or matched peripheral blood were collected and sequenced by a well validated panel based NGS assay²¹ Hematoxylin and eosin (HE) stained slides, with a thickness of 4 μm, underwent pathologist review to ensure that the sample had nucleated cellularity >80% and tumor content >20% for each tissue sample. The unstained FFPE sections (usually 5 slides) underwent DNA extraction, and typically 400 ng of double-stranded DNA was fragmented to ~250 bp through sonication. The sonicated DNA was used for library construction with KAPA Hyper Prep Kit (KAPA Biosystems) for end repair, addition, and adapter ligation. The library was amplified by PCR and quantified by Qubit. Samples yielding <40 ng of extracted DNA or <500 ng of pre-capture library were excluded for further sequencing. A custom hybridization capture panel containing individually synthesized biotinylated DNA oligonucleotides (120 bp) to target ~3 Mb of the human genome mainly 450 cancer-related genes as well as select introns of 38 frequently rearranged genes in cancer. Post-capture libraries were mixed together, denatured and diluted to 1.5-1.8 pM and subsequently sequenced on Illumina NextSeq 500. Paired-end sequencing (2 × 150 bp) was performed following the manufacturer’s protocols.

Genomic alterations, including single base substitutions, short and long INDELs, copy number variations, and gene rearrangements, were assessed using the OrigiMed-pipeline. The classification of actionable gene variants is guided by OncoKB. Alignment of raw reads to the human genome reference sequence (hg19) was done with the Burrows-Wheeler Aligner, followed by PCR duplicates removal using MarkDuplicates algorithm from Picard. Local realignment and base quality recalibration for SNV was performed using GATK and subsequently called by MUTECT. Short insertion/deletions (S-indels) were calibrated for alignment using ABRA and then called by PINDEL. The raw calls of SNV and S-indel were further selected with a minimum of 5 reads that were required to support alternative calling. To identify CNA, aligned reads were first normalized within each bed by EXCATOR. Log ratio of read depths for each gene from tumor tissue and its matched normal control were calculated. Tumor cellularity was estimated by allele frequencies of sequenced SNPs (single-nucleotide polymorphism). Focal amplifications were called for genes with threshold ≥5 copies for amplification and zero copy for homozygous deletions. For detection of gene rearrangement, aligned reads with abnormal insert size were collected and used as discordant reads, ie, paired-end reads that could not be closely mapped to a genome reference. Next, the discordant reads with the distance less than 500 bp formed clusters that were further assembled by fermi-lite to identify potential rearrangement breakpoints. The breakpoints were double-confirmed by BLAT and the resulted chimeric gene candidates were annotated.

Mutation Signature

Mutation signature was predicted using a public software deconstuctSigs. The profile of each signature is displayed using the 6 substitution subtypes: C>A, C>G, C>T, T>A, T>C, and T>G. Further, each of the substitutions is examined by incorporating information on the bases immediately next to each mutated base generating 96 possible mutation types (6 types of substitution * 4 types of 5’ base * 4 types of 3’ base). The frequency of each was precisely calculated as the characteristic signature of these samples. Mutational signatures were displayed and reported based on the observed trinucleotide frequency of the human genome, ie, representing the relative proportions of mutations generated by each signature based on the actual trinucleotide frequencies of the reference human genome version GRCh37. The signature was then compared with the typical 30 signatures from COSMIC to identify the most similar combination and the percentage of each contributed.¹⁹

Immunohistochemistry

Immunohistochemistry (IHC) was used to confirm the presence or absence of B2M protein. The tissues were fixed in 4% formalin, dehydrated using a graded ethanol series (75%, 85%, 95%, and 100%), placed in xylene, and embedded in paraffin. Next, these tissues were sliced into 5-μm sections, placed on slides, deparaffinized in xylene, and then rehydrated in a graded ethanol series (100%, 95%, 85%, and 75%). All slides were incubated in a citrate and 3% H₂O₂ solution for heat-induced antigen retrieval and to block endogenous peroxidases. Subsequently, the slides were washed 3 times with phosphate-buffered saline and incubated in a blocking solution containing bovine serum albumin for 1 h. An incubation in primary antibody against B2M (Cell Signaling, clone D8P1H) was performed at 4 °C overnight. Samples were visualized using a diaminobenzidine kit to estimate protein expression. After hematoxylin counterstaining of the nuclei, the slides were dehydrated and sealed. All slides were examined under microscope.

Results

Characteristics and Genomic Profiling of Patients with MSI-H Tumors

Among the 108 cases, the patients with MSI-H tumors presented with relatively low rates of advanced stages of diseases, where 5% of patients with EC, 22.5% with CRC, and 15% with GC were classified as clinical stage IV. NGS analysis revealed elevated TMB, ranging from 17 to 421 muts/MB, across all the types of cancer (Table 1). Genes involved in chromatin regulation, the PI3K pathway, the WNT pathway, and mismatch repair were extensively altered in the MSI-H cohort (Fig. 1). The most frequent genomic alterations in this cohort were found in KMT2D (69%), ARID1A (67%), ACVR2A (63%), KMT2C (56%), PIK3CA (54%), TGFBR2 (47%), CREBBP (44%), NRG1 (44%), APC (42%), and ATM (40%). Two major processes involved in chromatin regulation, histone modification and SWItch/Sucrose Non-Fermentable (SWI-SNF) chromatin remodeling, were altered extensively in the MSI-H cohort. The SWI-SNF chromatin remodeling complexes include PBRM1, ARID1A, and SMARCA4, regulate chromosome architecture, and impact responses to ICIs. Reading frameshift alteration was the most frequent gene alteration type among the top mutations, consisting of 43%, 46%, and 61% of alterations in KMT2D, ARID1A, and ACVR2A, respectively. Additionally, the mutation rate of PBRM1 was at 53% in the MSI-H CRC cohort. Although the overall level of genomic alteration was high in the MSI-H cohort, there was a lack of actionable gene variants for target therapy.

Table 1.

Characteristic of EC, CRC, and GC MSI-H patients.

Variables	EC	CRC	GC
Number	22	40	46
Gender	Female/male: 22/0	Female/male: 18/22	Female/male: 22/24
Age, years, median (range)	51 (35-81)	48 (23-75)	68 (43-86)
TMB, median (range)	27 (17-121.5)	74 (35-421)	58 (23-117)
Stage (I/II/III/IV)	8/8/5/1	5/18/8/9	10/17/12/7
B2M mutated sample, no. (%)	3 (14)	23 (58)	11 (24)

Figure 1.

The genomic mutational landscape of MSI-H cancers. Top panel, TMB of each tumor enrolled in this study. Middle panel, the frequent somatic mutations of MSI-H tumors grouped vertically by functional group and colored by the type of alteration and germline variants associated with Lynch syndrome. The total number of alterations in each gene and the percentage of individuals affected were list at right. Genes associated with histone modification, SWI/SNF, TGF-b, and WNT pathways are among the most altered genes. Truncations caused by indels in the coding region of genes have been frequently found in ACVR2A, TGFBR2, and B2M genes. Bottom panel, Key clinicopathological characteristics including tumor type, stage and age.

Different mutational processes generate unique combinations of mutation types, termed “Mutational Signatures”. Mutational signatures 6, 15, 20, and 26 featured with high numbers of small INDELs (shorter than 3bp) were associated with defective mismatch repair. Signature 6 and 26 were 2 common mutation signatures dominated in all 3 types of cancer. The attribution of signature 20, another defective mismatch repair associated feature, were high in EC (20%) but not in CRC (0%) or GC (8%) (Fig. 2a). Germline mutations of MSH2, MLH1, and MSH6, mainly truncations, associated with Lynch syndrome were identified in matching normal tissues in all 3 types of cancer. Patients with CRC possessed highest prevalence of Lynch syndrome with the following distribution of mutually excluded germline variants, MSH2 (n = 6, 15%), MLH1 (n = 6, 15%), and MSH6 (n = 1, 2.5%) (Fig. 1). Three MSH2 truncations (14%) and 1 MSH6 truncation (4.5%) were identified in EC, and only 2 MSH2 truncations (4%) were identified in GC cohort. MSI/dMMR and other genomic alterations reportedly associated with ICI efficacy may be promising biomarkers, if used in combination, to identify ICI sensitive patients.

Figure 2.

Mutation signature of MSI-H tumors and the feature of B2M alteration. (a) The fraction of mutational signature 6 and 26 which associated with defective mismatch repair system were found high in all 3 types of cancers. (b) Three of 5 microsatellite sites have been found frequently altered and leading to the production of B2M truncations as L15Ffs*41, V69Wfs*34, and T93Hfs*2. The incidence of such alterations was color coded by tumor type. (c) The odds ratio of key clinicopathological characteristics associated with B2M alteration. (d) The odds ratio of gene alteration associated with B2M (x-axis) is plotted against the negative log-transformed P value (y-axis) for each gene mutation. (e) The illustration of B2M bi-allelic mutation identified by IGV. (f) The mono-allelic and bi-allelic variants of B2M gene leading to the decreased expression of B2M protein identified by IHC.

Frameshift INDELs Caused Loss-of-Function of B2M Protein

Next, we further assessed somatic alterations in the B2M gene in MSI-H cancers. The most frequently altered microsatellite site was c.43-44del (16 incidences) for all 3 types of tumor, followed by c.204del with 10 incidences, and c.276dup with 4 incidences (Fig. 2b). Among the 3 cancer types, the B2M gene was altered at an exceptionally high frequency in CRCs (57.5%, 23/40), followed by GCs (23.9%, 11/46), and then ECs (13.6%, 3/22). To conclude whether Lynch Syndrome makes the B2M susceptible to mutate, we investigated the association between B2M variants and germline/somatic MMR genes alterations. It was demonstrated that B2M alterations were significantly associated with the somatic MSH6 mutation (P = .03) but not with the germline MSH2, MSH6, or MLH1 mutations. TMB was significantly higher in the patients carrying MSI-H/dMMR tumors with B2M mutation than in ones with wild-type B2M (P = .026), which indicates the mutation of B2M facilitating the accumulation of mutations.

Genomic alterations, dominantly truncated variants caused by INDELs at the microsatellite, occurring on ACVR2A, TGFBR2, SOX9 and TCF7L2 were significantly associated with B2M mutation, POLD1 and KDM6A genomic alterations, mainly SNV, were also significantly associated with B2M mutations (Fig. 2d). Biallelic INDELs causing non-functional protein expression were also discovered. The support reads covering the 2 mutation sites within a single Illumina sequencer read length (~150 bp) were manually inspected using the integrative genomics viewer.²² Alterations of the B2M gene were confirmed to be in trans using the integrative genomics viewer based on the distribution of the 2 mutations being exclusively separate or on different alleles (Fig. 2e). Furthermore, the B2M mutated cases with available tissues were stained with IHC and the result demonstrated that the expression of B2M of 60% samples with genomic alteration in this gene was impaired (Supplementary Table 2). IHC assays showed expression of B2M protein was impaired due to INDEL mutations, while wild type B2M protein was expressed (Fig. 2f). B2M loss-of-function mutations have long been reported to be associated with resistance to ICIs.¹¹ We next profiled the cancer genome from a patient with ICI-resistant cancer.

Primary Resistance of MSI-H CRC with B2M Mutation to ICI Treatment

Case

A 27-year-old female presenting with Stage IIIB (pT3N2aM0) CRC received surgery in August 2017 (Fig. 3). The resected CRC tumor tissue was diagnosed as dMMR by IHC. The tumor relapsed in March 2018 despite the patient remaining on 6 cycles of adjuvant chemotherapy (XELOX) after surgery. Six months after second-line therapy (FOLFIRI and bevacizumab) and the palliative surgery removing the bowel obstruction, PD-1 inhibitor was prescribed based on MSI-H/dMMR as a biomarker for ICIs in November 2018.^10,23 However, the tumor progressed after 2 months of anti-PD-1 ICI treatment, thus demonstrating primary resistance. NGS testing of the biopsy tumor tissue and matched normal tissue confirmed MSI-H and high TMB at 60 muts/MB. A MSH2 variant (p.Q337*) was identified in DNA from the tumor tissue. At the meantime, the profiling of the matched normal tissue revealed this MSH2 variant (p.Q337*) was a germline mutation with a VAF of 44.6%, indicating this patient had Lynch syndrome. All genomic variants were further analyzed to understand the intrinsic ICI resistance. An INDEL mutation (c.43_44del) in the B2M gene exonic microsatellite site was found and this led to the expression of chimeric B2M protein p.L15Ffs*41. The deleterious mutation in the B2M gene very likely causes resistance to ICI as has been previously reported.^24,25

Figure 3.

A case of ICI primary resistant patient. The diagram of a CRC patient who failed multiple lines of treatments resisted ICI primarily. The CT/MR scans have demonstrated that the tumor lesion had been observed at baseline before treatment and metastatic lesions were discovered after the ICI treatments.

Discussion

Immunotherapy blocking immune checkpoints has proven efficacy in approximately 20% of populations for several solid cancers.²⁶ For patients with the same diseases, ICI has different efficacy.^27,28 Therefore, biomarkers predicting patient sensitivity to ICI are urgently required to avoid ineffective treatment. The tumors with MSI-H have been confirmed to be sensitive to immune checkpoint blockades and approved as a biomarker for ICI therapy by the FDA.^10,29 Despite high response rate being associated with MSI-H/dMMR cancers, there are studies suggesting that a large percent of this cohort remains resistant to an immune checkpoint blockade.³⁰ According to our research, histone modification and SWI-SNF chromatin remodeling, both of which are major processes involved in chromatin regulation, were altered extensively in the MSI-H cohort. The SWI-SNF chromatin remodeling complexes include PBRM1, ARID1A, and SMARCA4, regulate chromosome architecture, and impact responses to ICIs. One of the mechanisms underlying this is the interruption of antigen-presentation on the cell surface via silencing or altering of B2M gene. The human B2M gene is enriched with dinucleotide repeats in its coding region, which are hotspots for frameshift indels.³¹ In our study, we investigated the mutational profile of B2M across cancers that were highly prevalent for MSI/dMMR, including CRC, GC, and EC. We found that the mutation rate of B2M was significantly high in MSI-H/dMMR CRCs (57.5%, 23/40), followed by GCs (23.9%, 11/46), and then ECs (13.6%, 3/22). A few studies had detected B2M mutations in MSI-H CRCs and the frequencies range from 20.8~24% in MSI-H CRCs^32-34 and 17~50% in Lynch Syndrome-associated MSI CRCs.³⁵ Biallelic disruption of B2M gene were found in 11% of untreated MSI-H CRCs here. Even so, data of B2M mutations on MSI-H GCs and ECs remain insufficient.³⁶ Considering the discrepancy of genomic variants in different signal pathways across 3 types of cancer, the various frequencies of B2M alteration may be caused by the endogenous and exogenous factors associated with the different tumors.

We found recurring INDELs in microsatellite sites of the B2M gene in these samples at various frequencies (Fig. 1). Furthermore, a subset of the patients with MSI-H tumors had biallelic B2M mutations. These biallelic frameshift mutations can cause a loss of functional B2M protein or failure to transcribe, which leads to failure to present the antigen and, thus, tumor survival. Some studies have demonstrated that germline mutations of MLH1 or MSH2 are related with higher rates of B2M mutations in MSI-H cancers.^18,33 Here, we also found that more than half of the germline MLH1 and MSH2 are associated with somatic B2M mutations in CRC but not in EC or GC, which reindicated the difference of the biology in these tumors. In addition, we revealed that B2M alterations were significantly associated with the somatic MSH6 mutation, but not with the germline MSH2, MSH6 or MLH1 mutations across the cancers. This result was partially caused by the low frequency of germline mutations in MSH2, MSH6 and MLH1 genes in EC and GC.

Our result demonstrated that MSI-H/dMMR tumors with B2M mutations had a large significance of association with high TMB, which indicated that the immune system might spare the tumor cells with impaired antigen presentation. However, whether the B2M mutation status is related to the different sensitivities to ICI immunotherapy is yet to still be undetermined. A rational hypothesis for this is that patients with a single allelic INDEL in the B2M gene may acquire resistance during ICI treatment, while those with biallelic mutations have effects causing loss-of function of B2M and display primary resistance to ICI therapy. This B2M gene mutation can be accumulated as cancer progresses. An early occurrence of B2M mutation, whether or not leading to actual resistance in patients, favors the survival of aggressive tumor cells that have defects in antigen presentation and indicate a hypermutating status caused by immune editing.³⁶ Such warnings urges close monitoring of ICI efficacy or to proceed in considering other treatments.

Due to the correlation of TMB and MSH6 to B2M mutations, whether TMB and MSH6 could be surrogate biomarkers for B2M mutations still needs to be elucidated. The deficiency of the mismatch repair (dMMR) system causes the accumulating single nucleotide DNA variants and small indels to be unrepaired in the genome, this leads to the high tumor mutational burden (TMB). The dMMR is commonly caused by epigenetic inactivation of the MLH1 gene and mutation in MMR genes such as MLH1, MSH2, MSH6, and PMS2 in sporadic CRCs, and the remainder of dMMR tumors that are associated with Lynch syndrome which is caused by germline mutations in the aforementioned genes. Among the 3 cancer types which are all TMB-H described here, the B2M gene was altered at a high frequency in CRCs (57.5%, 23/40), followed by GCs (23.9%, 11/46), and then ECs (13.6%, 3/22). Therefore, TMB-H can not be the surrogate biomarker for the B2M mutation. In terms of MSH6, it is one of the causes of dMMR involving many genetic and epigenetic. Though the somatic mutation of MSH6 is significantly associated with B2M mutation, it does not make for a good surrogate biomarker for the B2M mutation.

Here, we report a patient with MSI-H CRC displaying primary resistance to ICI and carrying a frameshift mutation in one of the exonic microsatellite sites (c.43_44del, p.L15Ffs*41) of B2M. It is rational to speculate that patients with MSI-H tumors that have both a high TMB and are experiencing loss-of-function of the B2M gene which would fail to respond to ICI. Recently, several reports have supported that B2M mutations may be a key mechanism of resistance to ICI in MSI-H cancers. Le et al detected B2M mutations in brain metastases from 2 patients with dMMR CRCs that were resistant to pembrolizumab.³⁷ Gurjao et al reported a patient with metastatic dMMR CRC progressing on to pembrolizumab after combined chemotherapy. They detected a somatic frameshift deletion of the B2M gene(p.V47Afs*6) and a loss of heterozygosity in the patient’s tumor, resulting in complete loss of B2M expression in the tumor cells. Pitifully, only 108 patients, including those with GC, CRC, or EC, were collected in the recent 5 years. From these results, we will improve on critiques within the successive clinical trials as it is necessary to consider these factors when utilizing immunotherapy. This work needs to be validated using a larger cohort and assessing several different types of cancers.

Although most studies have proved that MSI-H/dMMR cancers with B2M mutations may cause resistance to ICI, some studies revealed that B2M loss did not appear to affect response to ICI in MSI-H/dMMR CRCs.^32,38,39 These contradictory findings indicated that differences in protein expression caused by type and different mutation sites of B2M mutations may influence response to immunotherapy. This urges us to further explore the B2M mutation and acquire more evidence before or during treatment.

As a limitation of this study, it was difficult to expand the cohort to a larger one. Since the first ICI was launched in 2018 in China, there were fewer advanced cases received ICI that were qualified for the study. Further investigation is required in successive clinical trials.

In summary, this study discovered that patients with the somatic MSH6 mutation were likely to display B2M alterations. Also, patients carrying MSI-H/dMMR tumors with B2M mutations showed significantly higher TMB. The mutational status of the B2M gene is a point of consideration in the prescription and application of immunotherapy for MSI/dMMR patients and serves as a worthy predictor of the corresponding clinical outcomes.

Funding

This study was supported financially by grants from the Ministry of Science and Technology of the People’s Republic of China (2019ZX09301-150) and grants from Department of Science and Technology of Jiangsu Province (BK2019001). The funding source had no role in the study design or in the collection, analysis, and interpretation of data.

Ethics Approval and Consent to Participate

All patients provided written informed consent. The study was approved by the ethics committee of Drum Tower Hospital Affiliated to the Medical School of Nanjing University (Approval ID: 2019-037-02).

Conflict of Interest

Huan Wu, Jiaochun Shi, and Weifeng Wang are employees of OrigiMed. The other authors indicated no financial relationships.

Author Contributions

Conception/design: J.W. Provision of study material or patients: F.Z., H.W. Collection and/or assembly of data: F.L., K.C., J.S. Data analysis and interpretation: N.W., Y.F., Y.W., X.Q., X.F., W.W. Manuscript writing: F.L., K.C. Final approval of manuscript: All authors.

Data Availability

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.