Summary
High tumor mutational burden (TMB) is associated with response to checkpoint blockade in several cancers. We identify pathogenic germline variants associated with increased TMB (GVITMB). GVITMB were found in 7 genes using a pan-cancer approach (APC, FANCL, SLC25A13, ERCC3, MSH6, PMS2, and TP53) and 38 gene sets (e.g., those involved in DNA repair and programmed cell death). GVITMB were also associated with mutational signatures related to the dysfunction of the gene carrying the variant, somatic mutations that further affect the gene or pathway with the variant, or transcriptomic changes concordant with the expected effect of the variant. In a validation cohort of 140 patients with cutaneous melanoma, we found that patients with GVITMB had prolonged progression-free survival (p = 0.0349, hazard ratio = 0.688) and responded favorably (p = 0.0341, odds = 1.842) when treated with immune checkpoint inhibitors. Our results suggest that germline variants can influence the molecular phenotypes of tumors and predict the response to immune checkpoint inhibitors.
Subject areas: Genetics, Genomics, Cancer
Graphical abstract
Highlights
-
•
GVITMB were found in 7 genes and 38 gene sets
-
•
GVITMB influence the somatic mutation and gene expression profiles of tumors
-
•
GVITMB predict immune checkpoint inhibitory efficacy in SKCM
Genetics; Genomics; Cancer
Introduction
The explosion of massively parallel sequencing data has helped to identify rare germline variants that cause or contribute to disease (Sanderson et al., 2019; Vaske et al., 2019). In oncology, it is well-established that patients with germline variants in genes mutated in certain genetic syndromes, such as Lynch syndrome, Li-Fraumeni syndrome, von Hippel-Lindau syndrome, and Fanconi anemia, are at much higher risk of acquiring cancer (Ellrott et al., 2018; Kamps et al., 2017). Although individuals with these pathogenic germline variants are generally screened more aggressively, clinical management of patients with these germline variants is not always differentiated from the management of patients without these pathogenic variants (Ballinger et al., 2017; Lindor et al., 2006; Maher et al., 1990). This has begun to change. For example, patients with Lynch syndrome have pathogenic germline variants in mismatch repair genes, such as MSH2, MSH6, PMS2, and MLH1, and their tumors exhibit higher levels of microsatellite instability. As a consequence, patients with Lynch syndrome are more likely to respond to immune checkpoint inhibitors such as pembrolizumab (Snyder et al., 2014; Van Allen et al., 2015) (Le et al., 2017).
We have reported that germline variants affect tumor progression across a large spectrum of cancers through the analysis of common germline variants with a minor allele frequency greater than 5% in the general population (Chatrath et al., 2019, 2020). In this study, we analyze rare pathogenic germline variants to identify germline variants associated with increased tumor mutational burden (GVITMB) to test whether these germline variants increase the likelihood of a patient responding to immune checkpoint inhibitors (Keenan et al., 2019; Liu et al., 2019; Miao et al., 2018). After identifying the set of pathogenic germline variants predictive of tumor mutational burden (TMB), we demonstrate that they predict responsiveness to immune checkpoint inhibitors in a cohort of 140 patients with skin cutaneous melanoma.
Results
Germline variants can be analyzed using pan-cancer or gene set-level approaches
Huang et al. had previously described 435 rare pathogenic germline variants that were found in the patients in The Cancer Genome Atlas (TCGA) (Huang et al., 2018). Briefly, all somatic variants were scored based on the American College of Medical Genetics and Genomics and the Association for Molecular Pathology (ACMG-AMP) guidelines developed for rare variants in cancer and variants known to be pathogenic in ClinVar and curated databases were labeled as pathogenic. The majority of these pathogenic germline variants were predicted to functionally perturb known tumor suppressor genes or oncogenes. Before identifying which pathogenic germline variants contribute to elevated TMB, we first evaluated whether we were able to identify GVITMB based in individual genes in individual cancers. We set a modest threshold requiring at least five patients in the cancer cohort to have a pathogenic germline variant in a given gene.
We utilized four approaches to identify germline variants associated with increased TMB. (1) We tested individual genes for association with TMB in individual cancers, testing a total of 13 unique genes (Figure 1A). (2) We pooled all the patients in TCGA together, and by doing so we were now able to test 73 total genes for the presence of GVITMB (Figure 1B). (3) We grouped the pathogenic germline variants by gene set to identify gene sets carrying GVITMB in individual cancers (Figure 1C). (4) Finally, we repeated the analysis in (3) but after grouping all cancers together (Figure 1D). Our overall methodology is summarized in Figure 2.
Figure 1.
An overview of the number of genes or gene sets that could be tested with the threshold that the pathogenic germline variants must be present in five or more patients
(A) Number of testable genes in individual cancer types. This analysis was not performed due to the small number of testable genes.
(B) Number of patients with pathogenic variants in the indicated genes when patients with all cancers were pooled together. The stacked bars show the cancer types color coded as in the key. These patients were analyzed by Approach 1.
(C) Number of testable gene sets in each of the individual cancer types, analyzed by Approach 2.
(D) Number of patients carrying germline variants in the testable gene sets, analyzed by Approach 3. The stacked bars show the cancer types with pathogenic germline variants in a given gene set color coded as in the key.
Figure 2.
A summary of the overall approach employed in this study
Calculation of tumor mutational burden
Overall TMB, nonsynonymous TMB, and clonal nonsynonymous TMB have previously been reported to be associated with favorable response in patients treated with immune checkpoint inhibitors (Keenan et al., 2019; Miao et al., 2018). We, therefore, calculated these three metrics of TMB for each patient in TCGA and normalized them to per megabase (MB) based on the total number of sites in each patient wherein we were sufficiently powered to call a somatic mutation. This normalization accounted for the coverage at each site in the exome and the purity and ploidy of each tumor. All metrics of TMB were highly correlated to each other (Spearman’s rho >0.90 for all pairs, Figure 3A), and we present the normalized distribution of TMB by cancer in Figure 3B. We used clonal nonsynonymous TMB per MB as our dependent variable for this study as it has been shown to have a better association with immune checkpoint inhibitor responsiveness (Keenan et al., 2019; Miao et al., 2018).
Figure 3.
Calculated tumor mutational burden across cancers
(A) All six metrics of tumor mutational burden are highly correlated with each other.
(B) Overall TMB per megabase (MB), nonsynonymous TMB per MB, and clonal nonsynonymous TMB per MB across cancers.
Pan-cancer identification of individual genes associated with TMB
We identified seven genes that when perturbed by a pathogenic germline variant are associated with elevated TMB (Figure 4A, Table 1). Three of these genes (APC, FANCL, and SLC25A13) were determined to be significant after multiple hypothesis testing correction (adjusted p value <0.05). However, later in this study we also characterize the four genes (ERCC3, MSH6, PMS2, and TP53) that did not reach the significance threshold of an adjusted p value < 0.05 even though they crossed the raw p value threshold of <0.05 because they have well-known roles in DNA repair.
Figure 4.
Manhattan plots summarizing the associations with clonal nonsynonymous tumor mutational burden per megabase
(A–C) We identified associations with elevated tumor mutational burden in (A) genes perturbed by pathogenic germline variants using a pan-cancer approach, (B) gene sets perturbed by pathogenic germline variants in individual cancers, and (C) gene sets perturbed by pathogenic germline variants using a pan-cancer approach. For each gene set, the fraction of patients with a particular cancer carrying a pathogenic germline variant is indicated by the color code.
Table 1.
A summary of the associations we found with elevated somatic mutation burden in individual genes using a pan-cancer approach
| Gene | Number of patients with a PGV | Additional clonal nonsynonymous mutations per MB | p value | Adjusted p value |
|---|---|---|---|---|
| APC | 5 | 32.51 | 7.910E-09 | 5.300E-07 |
| FANCL | 8 | 23.77 | 9.500E-08 | 3.180E-06 |
| SLC25A13 | 17 | 11.46 | 1.938E-04 | 4.329E-03 |
| TP53 | 16 | 8.87 | 4.897E-03 | 8.202E-02 |
| ERCC3 | 23 | 6.87 | 9.044E-03 | 1.212E-01 |
| MSH6 | 20 | 6.34 | 2.453E-02 | 2.348E-01 |
| PMS2 | 32 | 5.05 | 2.367E-02 | 2.348E-01 |
PGV, Pathogenic Germline Variants.
Identification of gene sets carrying GVITMB in individual cancers
We identified significant associations of pathogenic germline variants in gene sets and TMB in Colon Adenocarcinoma (COAD), Skin Cutaneous Melanoma (SKCM), and Uterine Corpus Endometrial Carcinoma (UCEC) (Figure 4B, Table 2, list of perturbed genes in Table S1). Although each of the identified gene sets consisted of different and unique gene sets, the genes that empirically contributed to these gene sets sometimes overlapped in this analysis. We have therefore grouped gene sets for which the contributing genes entirely overlapped in this particular analysis. In total, we identified 29 associations (2 in COAD, 11 in SKCM, and 16 in UCEC). The significantly associated gene sets were primarily related to DNA damage and repair and cell cycle control.
Table 2.
A summary of the associations we found with elevated somatic mutation burden in individual gene sets in individual cancers
| Gene set | Cancer | Patients with PGV | Additional clonal nonsynonymous mutations per MB | p value | Adjusted p value |
|---|---|---|---|---|---|
| TP53 regulates transcription of DNA repair genes | UCEC | 14 | 38.78 | 8.712E-31 | 2.004E-29 |
| DNA repair | UCEC | 31 | 25.71 | 7.358E-30 | 8.462E-29 |
| Transcriptional regulation by TP53 | UCEC | 20 | 27.86 | 4.704E-23 | 3.607E-22 |
| Generic transcription pathway | UCEC | 22 | 26.34 | 1.132E-22 | 6.508E-22 |
| Disease | UCEC | 12 | 34.18 | 5.317E-21 | 2.446E-20 |
| Gene expression transcription | UCEC | 24 | 23.87 | 1.828E-20 | 7.009E-20 |
| Mismatch repair, diseases of mismatch repair (MMR) | UCEC | 7 | 34.55 | 3.925E-13 | 1.290E-12 |
| Sumoylation | UCEC | 6 | 31.43 | 9.928E-10 | 2.854E-09 |
| Deubiquitination | UCEC | 6 | 31.15 | 1.396E-09 | 3.568E-09 |
| Fanconi anemia pathway | UCEC | 8 | 25.54 | 1.040E-08 | 2.393E-08 |
| DNA double-strand break response | UCEC | 7 | 26.42 | 2.932E-08 | 6.130E-08 |
| G2 M checkpoints, G2 M DNA damage checkpoint, regulation of TP53 activity, regulation of TP53 activity through phosphorylation | UCEC | 9 | 20.00 | 1.958E-06 | 3.753E-06 |
| Post-translational protein modification | UCEC | 9 | 19.61 | 3.065E-06 | 5.423E-06 |
| Cell cycle checkpoints | UCEC | 10 | 17.62 | 9.955E-06 | 1.635E-05 |
| Disease | COAD | 8 | 20.53 | 4.064E-06 | 7.315E-05 |
| DNA double-strand break repair | UCEC | 15 | 11.48 | 4.291E-04 | 6.580E-04 |
| Cell cycle | UCEC | 15 | 11.37 | 4.835E-04 | 6.950E-04 |
| DNA repair | SKCM | 30 | 7.89 | 6.214E-04 | 1.243E-02 |
| Disease | SKCM | 6 | 16.17 | 1.680E-03 | 1.680E-02 |
| Cell cycle | SKCM | 20 | 8.43 | 2.805E-03 | 1.870E-02 |
| Generic transcription pathway, gene expression transcription | COAD | 15 | 9.86 | 2.467E-03 | 2.220E-02 |
| Cell cycle checkpoints | SKCM | 12 | 9.57 | 8.553E-03 | 2.851E-02 |
| Regulation of TP53 activity | SKCM | 11 | 10.38 | 6.340E-03 | 2.851E-02 |
| DNA double-strand break repair | SKCM | 19 | 7.72 | 7.643E-03 | 2.851E-02 |
| Homology-directed repair (HDR) through homologous recombination (HRR) | SKCM | 17 | 7.85 | 1.027E-02 | 2.903E-02 |
| Resolution of D-loop structures, resolution of D-loop structures through synthesis-dependent strand annealing (SDSA), homologous DNA pairing, and strand exchange | SKCM | 16 | 7.96 | 1.161E-02 | 2.903E-02 |
| Generic transcription pathway, gene expression transcription | SKCM | 18 | 6.90 | 2.029E-02 | 3.568E-02 |
| G2 M checkpoints, G2 M DNA damage checkpoint, regulation of TP53 activity through phosphorylation | SKCM | 10 | 9.52 | 1.696E-02 | 3.568E-02 |
| Transcriptional regulation by TP53 | SKCM | 16 | 7.44 | 1.827E-02 | 3.568E-02 |
Pan-cancer identification of gene sets carrying GVITMB
Last, we identified pathogenic germline variants associated with TMB using a pan-cancer approach in which the pathogenic germline variants were grouped by gene set (Figure 4C, Table 3, list of perturbed genes in Table S2). In total, we identified 12 significant associations. Several of the gene sets were related to Wnt signaling, and the pathogenic germline variants in APC greatly contributed to these associations, as described in our analysis of individual genes. One association was driven entirely by SLC25A13 and had also been described in our previous analysis of individual genes. The other associations were related to apoptosis, cell cycle control, and DNA damage repair.
Table 3.
A summary of the associations we found with elevated somatic mutation burden in individual gene sets using a pan-cancer approach
| Gene set | Number of patients with PGV | Additional clonal nonsynonymous mutations per MB | p value | Adjusted p value |
|---|---|---|---|---|
| Degradation of β-catenin by the destruction complex | 5 | 32.51 | 7.907E-09 | 8.223E-07 |
| β-catenin phosphorylation cascade, disassembly of the destruction complex and recruitment of axin to the membrane, signaling by WNT in cancer, phosphorylation site mutants of CTNNB1 are not targeted to the proteasome by the destruction complex | 5 | 32.51 | 7.907E-09 | 8.223E-07 |
| Ovarian tumor domain proteases | 22 | 13.70 | 3.466E-07 | 2.403E-05 |
| Deactivation of the β-catenin transactivating complex | 7 | 22.42 | 2.517E-06 | 1.309E-04 |
| Programmed cell death | 28 | 11.03 | 3.729E-06 | 1.551E-04 |
| Regulation of kit signalling | 5 | 23.99 | 2.086E-05 | 7.231E-04 |
| Apoptotic cleavage of cellular proteins, apoptotic execution phase | 11 | 14.55 | 1.299E-04 | 3.378E-03 |
| Signaling by WNT, TCF-dependent signaling in response to WNT | 10 | 15.31 | 1.241E-04 | 3.378E-03 |
| Mitochondrial protein import, gluconeogenesis, glucose metabolism, aspartate and asparagine metabolism, protein localization | 17 | 11.46 | 1.938E-04 | 4.480E-03 |
| Disease | 211 | 3.17 | 2.993E-04 | 6.226E-03 |
| Mismatch repair | 63 | 5.63 | 4.116E-04 | 7.782E-03 |
| Diseases of mismatch repair (MMR) | 62 | 5.62 | 4.615E-04 | 7.999E-03 |
GVITMB influence somatic events
We next sought to characterize the somatic events associated with GVITMB. Several studies have suggested that germline variants influence somatic events (Carter et al., 2017; Chatrath et al., 2019, 2020; Chirita-Emandi et al., 2020). We found that patients with GVITMB in mismatch repair genes exhibited enrichment of mutational signatures associated with mismatch repair gene dysfunction, suggesting exome-wide evidence of the dysfunction of these genes (Table 4).
Table 4.
Mutational signature results concordant with the expected effects of the pathogenic germline variants
| Gene or gene set | Cancer | Mutational signature | Fold enrichment | p value |
|---|---|---|---|---|
| MSH6 | Pan-cancer | 44 | 3.83 | 3.11E-03 |
| Mismatch repair | UCEC | 20 | 2.16 | 2.90E-02 |
| Mismatch repair | Pan-cancer | 20 | 2.16 | 2.13-03 |
| Mismatch repair | Pan-cancer | 26 | 1.58 | 3.48E-02 |
| Mismatch repair | Pan-cancer | 44 | 2.89 | 8.38E-06 |
We next tested whether the genes and gene sets perturbed by GVITMB were associated with somatic mutations in these same genes or gene sets. We controlled for TMB in all analyses to account for the general increase in somatic mutations in tumors with the GVITMB, along with controlling for tumor type and demographic factors. Patients with GVITMB in the mismatch repair gene PMS2 were much more likely to exhibit somatic mutations in PMS2 than patients without the GVITMB in PMS2 (beta = 3.05, p value = 5.86E-5, adjusted p value=4.1E-4). We found that GVITMB in ERCC3 or TP53 were associated with an increased incidence of somatic mutations in gene sets that include ERCC3 or TP53, respectively (Table S3). In addition, patients with SKCM with GVITMB in the disease gene set (a compilation of genes associated with human diseases) were more likely to acquire somatic mutations in other genes of the same gene set (beta=20.2, p value = 4.12E-6, adjusted p value=1.73E-4).
Finally, we tested for up- or downregulation of gene expression consistent with the expected effects of the GVITMB. We found that patients with GVITMB in genes regulating the G2-M checkpoint in UCEC exhibited upregulation of E2F target genes, suggesting upregulation of cell cycle activity (p value = 0.013).
GVITMB predict immune checkpoint inhibitory efficacy in SKCM
To test whether patients with SKCM with pathogenic germline variants in the gene sets that we had found to be associated with TMB in the TCGA dataset (Table 2) responded better to immune checkpoint inhibitors, we analyzed sequencing data from 140 patients with SKCM treated with either nivolumab or pembrolizumab (Liu et al., 2019). Given the relatively small sample size, we were not sufficiently powered to test individual gene sets for association with outcome. Of all the gene sets that contained GVITMB in SKCM (Table 2), only the disease gene set was sufficiently powered to detect an association with progression-free survival. Patients with pathogenic germline variants in the disease gene set exhibited prolonged progression-free survival (p = 0.0245, hazard ratio [HR] = 0.662) (Figures 5A and 5B) and were more likely to show a response to immune checkpoint inhibitors based on Response evaluation criteria in solid tumors (RECIST) criteria (p = 0.0393, odds = 1.781, ordering of categories was progressive disease, stable disease, partial response, and then complete response) (Figure 5C). Although patients with pathogenic germline variants had a higher median number of overall mutations, nonsynonymous mutations, and clonal nonsynonymous mutations, this difference was not statistically significant (Table S4, top three rows).
Figure 5.
Pathogenic germline variants predict immune checkpoint inhibitor efficacy in an independent cohort of 140 patients with skin cutaneous melanoma treated with immune checkpoint inhibitors
(A–C) Patients with pathogenic germline variants in the (A) disease gene set exhibit (B) prolonged progression-free survival and (C) are more likely to respond to immune checkpoint inhibitors.
(D–F) We (D) pooled all gene sets with GVITMB in SKCM together and found that patients with germline variants in these gene sets exhibited (E) prolonged progression-free survival and are (F) more likely to respond to immune checkpoint inhibitors.
Abbreviations: PD, progressive disease; SD, stable disease; PR, partial response; CR, complete response.
We were better powered to detect such an association by pooling all pathogenic germline variants found in the genes of the gene sets that we found to be associated with elevated TMB in SKCM the TCGA dataset (Figures 5D and Table 2). When tested, we found that patients with pathogenic germline variants in these genes exhibited favorable outcome and were less likely to progress (Figure 5E, p = 0.0349, HR = 0.688). Similarly, patients with pathogenic germline variants in these genes were more likely to exhibit a response to immune checkpoint inhibitors based on RECIST criteria (Figure 5F, p = 0.0341, odds = 1.842). Turning to TMB, we found that the median number of total mutations, nonsynonymous mutations, and clonal nonsynonymous mutations was greater in patients with pathogenic germline variants in genes in our gene set than patients without these pathogenic germline variants, although these differences were also not statistically significant (Table S4, lower three rows). Thus the GVITMB have a more significant effect on responsiveness than can be expected from the differences in TMB alone.
Discussion
The widespread collection of sequencing data has enabled detailed study of rare genetic syndromes (Kamps et al., 2017; Sylvester et al., 2018). Although patients with pathogenic germline variants are often screened more aggressively for cancer, clinical guidelines for these patients have only changed in a few circumstances (Le et al., 2017; Lindor et al., 2006). We had previously identified common germline variants associated with differences in patient outcome across a multitude of cancers, suggesting that germline variation contributes not only to cancer risk but also to tumor progression (Chatrath et al., 2019, 2020). In this study, we have identified pathogenic germline variants associated with TMB. Some of these associations were expected and confirmed existing hypotheses (e.g., mutations in known DNA repair genes such as MSH6 and PMS2), whereas other associations (e.g., mutations in SLC25A16) are more surprising and can motivate future hypotheses. We identified molecular fingerprints of the effects of some of the pathogenic germline variants by analyzing RNA sequencing data and somatic mutation profiles. Our findings suggest that these pathogenic germline variants remain relevant after a patient has been diagnosed with cancer and may contribute to the molecular differences in tumors collected from patients with and without pathogenic germline variants.
After identifying the set of pathogenic germline variants associated with TMB in skin cutaneous melanoma, we showed that patients with these pathogenic germline variants exhibit prolonged progression-free survival and increased responsiveness to immune checkpoint inhibitors. Given the relatively small size of the validation cohort, our validation study had limited resolution because we were not adequately powered to test individual genes or gene sets. As the total amount of sequencing data from patients treated with immune checkpoint inhibitors continues to increase, our ability to identify individual genes and gene sets predictive of responsiveness will improve. In this study, we identify pathogenic germline variants associated with TMB as a proxy for immune checkpoint inhibitory efficacy, although determining the extent to which TMB is predictive of immune checkpoint inhibitor efficacy across all cancers is still an active area of research (Wood et al., 2020).
Tumors from patients with pathogenic germline variants in the mismatch repair genes MSH6 and PMS2 and in the mismatch repair pathway exhibit elevated TMB. We found enrichment in the contribution to these patients’ somatic mutation profiles from COSMIC signatures related to mismatch repair. Germline mismatch repair deficiency has previously been associated with microsatellite instability and increased responsiveness to immune checkpoint inhibitors, and so these findings served as an important positive control in our study (Le et al., 2017).
Tumors with pathogenic germline variants in the nucleotide excision repair gene ERCC3 were associated with elevated TMB in our study. Although a previous study showed that somatic mutations in the nucleotide base excision repair gene ERCC2 likely contributes to increased TMB, no previous study has demonstrated an association between nucleotide excision repair gene perturbation and immune checkpoint inhibitor efficacy (Van Allen et al., 2014). We did not find a significant association between nucleotide excision repair pathway perturbation by pathogenic germline variants and TMB at the pathway level, suggesting that the contribution to TMB may be limited to select genes in the pathway.
We found patients with pathogenic germline variants in APC, which binds to beta-catenin and leads to its degradation, and genes involved with beta-catenin degradation to be associated with elevated somatic mutation burden. Aberrations to the Wnt signaling pathway are linked to the formation of many cancers (Anastas and Moon, 2013). Spranger et al. showed that non-T cell inflamed tumors exhibited high β-catenin signaling activity and reduced response to immune checkpoint blockade (Spranger et al., 2015). Further work is necessary to predict whether pathogenic germline variants in APC and genes involved with β-catenin degradation will be associated with increased or decreased response to immunotherapy, as the elevated TMB would be expected to increase efficacy, whereas the elevated β-catenin signaling would be expected to decrease efficacy.
Tumors from patients with pathogenic germline variants in SLC25A13 exhibited elevated somatic mutation burden. This gene codes for a mitochondrial aspartate/glutamate transporter. Pathogenic germline variants in this gene are associated with the urea cycle disorder type II citrullinemia and neonatal intrahepatic cholestasis (Song et al., 2013). Lee et al. have previously shown that tumors exhibiting urea cycle dysfunction generate nitrogen metabolites, resulting in DNA damage and ultimately better response to immune checkpoint blockade (Lee et al., 2018). Lee et al.’s analysis focused on somatic urea cycle dysfunction, whereas our work suggests that germline urea cycle dysfunction may also be a marker for improved immune checkpoint blockade response.
FANCL is the E3 ubiquitin ligase subunit within the FA core complex that enhances the efficiency of FANCD2 monoubiquitination. FANCD2 participates in DNA damage recognition and repair. As the pathogenic germline mutations in FANCL associated with TMB are predicted to be loss-of-function mutations, we hypothesize that they lower the efficacy of interstrand crosslink repair, affecting TMB.
High TMB has been associated with response to checkpoint blockade in several malignancies. However, the degree to which TMB changes over time, across anatomical sites, and with intervening treatment is still not clear. Studies have noted that tumor sampling from different anatomical sites may be associated with greater discrepancies in TMB calculations (Smithy et al., 2019). Efforts are ongoing to standardize TMB evaluation, which is needed to ensure reliability, reproducibility, and clinical utility (Galuppini et al., 2019). Compared with TMB, germline variants are relatively simpler to detect, annotate, score, and classify (Huang et al., 2018). Furthermore, they do not change during the course of the disease. It remains to be evaluated if they have additional value as a biomarker beyond that is provided by TMB, but our analyses suggest that they should be viewed as a biomarker candidate that can provide a robust and reproducible signal.
Overall, the results of our analysis suggest that understanding the germline contribution to somatic events could inform clinical therapy decisions (Carter et al., 2017; Menden et al., 2018). In this study, we have shown that pathogenic germline variants inform TMB and that these sets of pathogenic germline variants can be used to predict immune checkpoint inhibitor efficacy in patients with skin cutaneous melanoma. Future studies of germline variants in cancer will likely continue to illuminate areas in which clinical management can be further personalized based on an understanding of a patient's germline variants.
Limitations of the study
In this study, we used the TCGA data to identify pathogenic germline variants that are associated with increased tumor mutation burden (GVITMB). More than 80% of the patients in TCGA are of European ancestry, so it remains to be seen whether these associations will be replicated in a more diverse cohort. For the association analysis, we collapse the pathogenic variants in genes and gene set with the assumption that all pathogenic germline variants contribute toward increased TMB. It is likely that using adaptive burden association tests could increase our power to determine the associations, but that would come at the expense of interpretability. Using a second SKCM dataset, we were able to show that GVITMB have prognostic value, but it still needs to be determined whether GVITMB offer additional prognostic value beyond TMB. However, GVITMB do offer some advantages, as we highlight in the discussion, and should be considered as possible biomarker candidates in future studies.
Resource availability
Lead contact
Further information and questions should be directed to and will be fulfilled by the lead contact, Anindya Dutta (ad8q@virginia.edu).
Materials availability
This study did not generate new unique reagents.
Data and code availability
All scripts used for analyses are available at https://github.com/achatrath/GermlineSomaticMutationBurden.
Methods
All methods can be found in the accompanying Transparent Methods supplemental file.
Acknowledgments
This work was supported by grants from theNIHR01CA166054, R01CA60499, and T32GM007267 (A.C.).
Author contributions
Conceptualization, A.C. and A.D.; methodology, A.C., A.R., and A.D.; software, formal analysis, investigation, writing – original draft, visualization, and data curation, A.C.; resources and funding acquisition, A.D.; writing – review & editing, all authors; supervision and administration, A.D. and A.R.
Declaration of interests
As required by our employer, the University of Virginia, we have declared the invention to the University, based on which the University has applied for a provisional patent application. All three authors of this manuscript will be the inventors declared in the patent application.
Published: March 19, 2021
Footnotes
Supplemental information can be found online at https://doi.org/10.1016/j.isci.2021.102248.
Supplemental information
References
- Anastas J.N., Moon R.T. WNT signalling pathways as therapeutic targets in cancer. Nat. Rev. Cancer. 2013;13:11–26. doi: 10.1038/nrc3419. [DOI] [PubMed] [Google Scholar]
- Ballinger M.L., Best A., Mai P.L., Khincha P.P., Loud J.T., Peters J.A., Achatz M.I., Chojniak R., Balieiro da Costa A., Santiago K.M. Baseline surveillance in Li-fraumeni syndrome using whole-body magnetic resonance imaging: a meta-analysis. JAMAOncol. 2017;3:1634–1639. doi: 10.1001/jamaoncol.2017.1968. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Carter H., Marty R., Hofree M., Gross A.M., Jensen J., Fisch K.M., Wu X., DeBoever C., Van Nostrand E.L., Song Y. Interaction landscape of inherited polymorphisms with somatic events in cancer. Cancer Discov. 2017;7:410–423. doi: 10.1158/2159-8290.CD-16-1045. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chatrath A., Kiran M., Kumar P., Ratan A., Dutta A. The germline variants rs61757955 and rs34988193 are predictive of survival in lower grade glioma patients. Mol. Cancer Res. 2019;17:1075–1086. doi: 10.1158/1541-7786.MCR-18-0996. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chatrath A., Przanowska R., Kiran S., Su Z., Saha S., Wilson B., Tsunematsu T., Ahn J.H., Lee K.Y., Paulsen T. The pan-cancer landscape of prognostic germline variants in 10,582 patients. Genome Med. 2020;12:15. doi: 10.1186/s13073-020-0718-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chirita-Emandi A., Andreescu N., Zimbru C.G., Tutac P., Arghirescu S., Serban M., Puiu M. Challenges in reporting pathogenic/potentially pathogenic variants in 94 cancer predisposing genes - in pediatric patients screened with NGS panels. Sci. Rep. 2020;10:223. doi: 10.1038/s41598-019-57080-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ellrott K., Bailey M.H., Saksena G., Covington K.R., Kandoth C., Stewart C., Hess J., Ma S., Chiotti K.E., McLellan M. Scalable open science approach for mutation calling of tumor exomes using multiple genomic pipelines. Cell Syst. 2018;6:271–281.e277. doi: 10.1016/j.cels.2018.03.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Galuppini F., Dal Pozzo C.A., Deckert J., Loupakis F., Fassan M., Baffa R. Tumor mutation burden: from comprehensive mutational screening to the clinic. Cancer Cell Int. 2019;19:209. doi: 10.1186/s12935-019-0929-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Huang K.L., Mashl R.J., Wu Y., Ritter D.I., Wang J., Oh C., Paczkowska M., Reynolds S., Wyczalkowski M.A., Oak N. Pathogenic germline variants in 10,389 adult cancers. Cell. 2018;173:355–370.e314. doi: 10.1016/j.cell.2018.03.039. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kamps R., Brandao R.D., Bosch B.J., Paulussen A.D., Xanthoulea S., Blok M.J., Romano A. Next-Generation sequencing in oncology: genetic diagnosis, risk prediction and cancer classification. Int. J.Mol.Sci. 2017;18:308. doi: 10.3390/ijms18020308. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Keenan T.E., Burke K.P., Van Allen E.M. Genomic correlates of response to immune checkpoint blockade. Nat. Med. 2019;25:389–402. doi: 10.1038/s41591-019-0382-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Le D.T., Durham J.N., Smith K.N., Wang H., Bartlett B.R., Aulakh L.K., Lu S., Kemberling H., Wilt C., Luber B.S. Mismatch repair deficiency predicts response of solid tumors to PD-1 blockade. Science. 2017;357:409–413. doi: 10.1126/science.aan6733. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lee J.S., Adler L., Karathia H., Carmel N., Rabinovich S., Auslander N., Keshet R., Stettner N., Silberman A., Agemy L. Urea cycle dysregulation generates clinically relevant genomic and biochemical signatures. Cell. 2018;174:1559–1570.e1522. doi: 10.1016/j.cell.2018.07.019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lindor N.M., Petersen G.M., Hadley D.W., Kinney A.Y., Miesfeldt S., Lu K.H., Lynch P., Burke W., Press N. Recommendations for the care of individuals with an inherited predisposition to Lynch syndrome: a systematic review. JAMA. 2006;296:1507–1517. doi: 10.1001/jama.296.12.1507. [DOI] [PubMed] [Google Scholar]
- Liu D., Schilling B., Liu D., Sucker A., Livingstone E., Jerby-Amon L., Zimmer L., Gutzmer R., Satzger I., Loquai C. Integrative molecular and clinical modeling of clinical outcomes to PD1 blockade in patients with metastatic melanoma. Nat. Med. 2019;25:1916–1927. doi: 10.1038/s41591-019-0654-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Maher E.R., Yates J.R., Harries R., Benjamin C., Harris R., Moore A.T., Ferguson-Smith M.A. Clinical features and natural history of von Hippel-Lindau disease. Q. J. Med. 1990;77:1151–1163. doi: 10.1093/qjmed/77.2.1151. [DOI] [PubMed] [Google Scholar]
- Menden M.P., Casale F.P., Stephan J., Bignell G.R., Iorio F., McDermott U., Garnett M.J., Saez-Rodriguez J., Stegle O. The germline genetic component of drug sensitivity in cancer cell lines. Nat.Commun. 2018;9:3385. doi: 10.1038/s41467-018-05811-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Miao D., Margolis C.A., Vokes N.I., Liu D., Taylor-Weiner A., Wankowicz S.M., Adeegbe D., Keliher D., Schilling B., Tracy A. Genomic correlates of response to immune checkpoint blockade in microsatellite-stable solid tumors. Nat. Genet. 2018;50:1271–1281. doi: 10.1038/s41588-018-0200-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Sanderson S.C., Hill M., Patch C., Searle B., Lewis C., Chitty L.S. Delivering genome sequencing in clinical practice: an interview study with healthcare professionals involved in the 100 000 Genomes Project. BMJ Open. 2019;9:e029699. doi: 10.1136/bmjopen-2019-029699. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Smithy J.W., Hwang D.H., Li Y.Y., Spurr L., Cherniack A.D., Sholl L.M., Awad M.M. Changes in tumor mutational burden in serially biopsied non-small cell lung cancer. J. Clin. Oncol. 2019;37:e14286. [Google Scholar]
- Snyder A., Makarov V., Merghoub T., Yuan J., Zaretsky J.M., Desrichard A., Walsh L.A., Postow M.A., Wong P., Ho T.S. Genetic basis for clinical response to CTLA-4 blockade in melanoma. N.Engl. J. Med. 2014;371:2189–2199. doi: 10.1056/NEJMoa1406498. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Song Y.Z., Zhang Z.H., Lin W.X., Zhao X.J., Deng M., Ma Y.L., Guo L., Chen F.P., Long X.L., He X.L. SLC25A13 gene analysis in citrin deficiency: sixteen novel mutations in East Asian patients, and the mutation distribution in a large pediatric cohort in China. PLoS One. 2013;8:e74544. doi: 10.1371/journal.pone.0074544. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Spranger S., Bao R., Gajewski T.F. Melanoma-intrinsic beta-catenin signalling prevents anti-tumour immunity. Nature. 2015;523:231–235. doi: 10.1038/nature14404. [DOI] [PubMed] [Google Scholar]
- Sylvester D.E., Chen Y., Jamieson R.V., Dalla-Pozza L., Byrne J.A. Investigation of clinically relevant germline variants detected by next-generation sequencing in patients with childhood cancer: a review of the literature. J. Med. Genet. 2018;55:785–793. doi: 10.1136/jmedgenet-2018-105488. [DOI] [PubMed] [Google Scholar]
- Van Allen E.M., Miao D., Schilling B., Shukla S.A., Blank C., Zimmer L., Sucker A., Hillen U., Foppen M.H.G., Goldinger S.M. Genomic correlates of response to CTLA-4 blockade in metastatic melanoma. Science. 2015;350:207–211. doi: 10.1126/science.aad0095. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Van Allen E.M., Mouw K.W., Kim P., Iyer G., Wagle N., Al-Ahmadie H., Zhu C., Ostrovnaya I., Kryukov G.V., O'Connor K.W. Somatic ERCC2 mutations correlate with cisplatin sensitivity in muscle-invasive urothelial carcinoma. Cancer Discov. 2014;4:1140–1153. doi: 10.1158/2159-8290.CD-14-0623. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Vaske O.M., Bjork I., Salama S.R., Beale H., Tayi Shah A., Sanders L., Pfeil J., Lam D.L., Learned K., Durbin A. Comparative tumor RNA sequencing analysis for difficult-to-treat pediatric and young adult patients with cancer. JAMANetw. Open. 2019;2:e1913968. doi: 10.1001/jamanetworkopen.2019.13968. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wood M.A., Weeder B.R., David J.K., Nellore A., Thompson R.F. Burden of tumor mutations, neoepitopes, and other variants are weak predictors of cancer immunotherapy response and overall survival. Genome Med. 2020;12:33. doi: 10.1186/s13073-020-00729-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
All scripts used for analyses are available at https://github.com/achatrath/GermlineSomaticMutationBurden.