Bondy, Melissa L.

ABSTRACT Background Most genomic studies of adult-type diffuse gliomas have focused on predominantly European ancestry populations, limiting the generalizability of molecular classifications and precision medicine approaches. We assembled a multi-institutional glioma cohort of diverse patients to investigate how germline ancestry, molecular subtypes, and mutational processes shape tumor biology and clinical outcomes. Methods We analyzed 1,102 adults with WHO 2021–classified diffuse gliomas (IDH-mutant, 1p/19q-codeleted oligodendroglioma; IDH-mutant astrocytoma; IDH-wildtype glioma) from seven U.S. institutions. Whole-exome sequencing (WES) of FFPE tumors identified somatic alterations and COSMIC SBS v3.2 mutational signatures. Genetic ancestry was estimated from WES using 1000 Genomes reference populations. Overall survival was assessed using Kaplan–Meier and multivariable models. Results The cohort included 66.9% European (EUR), 21.1% Admixed American/Hispanic (AMR), 10.3% Admixed African (AFR), and 1.6% Asian (AS) ancestry. Survival followed expected molecular hierarchy (median overall survival: oligodendroglioma 15.7 years, astrocytoma 10.6 years, IDH-wildtype glioma 1.9 years). Within oligodendroglioma, AMR patients showed improved survival versus EUR (HR 0.67, 95% CI 0.48–0.94; p=0.011), with similar trends across subtypes. Somatic profiling confirmed canonical subtype-defining alterations and revealed higher ATRX alterations in AFR and AMR IDH-wildtype tumors compared with EUR. ATRX alterations were associated with improved survival only in AFR (p=0.003). Mutational signature analysis identified subtype-specific signatures, including therapy-associated signatures. Chemotherapy-related signatures were more frequent in EUR and AMR than in AFR. Conclusions This ancestrally diverse glioma cohort confirms established molecular classifications and identifies ancestry-associated differences in survival, somatic alterations, and mutational processes, indicating the critical need for broad representation to inform precision neuro-oncology. Key Points A multi-institutional glioma cohort validates subtype and survival patterns across ancestries. Therapy-associated mutational signatures differ by ancestry, suggesting distinct treatment-related mutational processes. Admixed American patients show improved survival, particularly in oligodendroglioma. Importance of the Study Most genomic studies of adult-type diffuse glioma have focused on populations of predominantly European ancestry which limits the ability to examine variation in tumor biology and clinical outcomes across populations. In this study, we assembled one of the largest ancestrally diverse cohorts of molecularly characterized adult diffuse gliomas, integrating germline ancestry inference with tumor whole-exome sequencing and mutational signature analysis. We confirm that established molecular classifications and survival hierarchies remain robust across ancestry groups. However, we also identified ancestry-associated differences in survival within specific tumor subtypes, higher ATRX alteration frequencies in African American and admixed American patients with IDH-wildtype tumors, and variation in therapy-associated mutational signatures across ancestry groups. These findings highlight the importance of incorporating population differences into genomic studies of glioma and provide a resource for future multi-ancestry investigations of glioma risk, tumor evolution, and treatment response, ultimately supporting more inclusive precision neuro-oncology.

Abstract Gliomas are heterogeneous brain tumors with prognostically-significant subtypes and few risk factors. IDH-wildtype glioblastoma, the most common subtype, carries a dismal prognosis. Using the largest GWAS dataset with histological (11,304 cases, 304,523 controls) and molecular (3,418 cases, 8,156 controls) subtypes, we analyzed the plasma and brain proteome to uncover proteins associated with glioma susceptibility. Using protein quantitative trait loci (pQTL) for 1,776 proteins in brain tissue, we identified 21 candidate proteins, including 9 associations with high probability (>70%) of colocalization (shared causal genetic variants for glioma risk and protein abundance). We confirmed known risk genes such as EGFR (p=1.9×10-35) and D2HGDH (p=1.1×10-5). We also identified novel risk proteins, including ENPP6, a choline phosphodiesterase overexpressed in glioma, for IDH-wildtype glioblastoma (OR=0.62, p=7.6×10-5), and galectin-3, a lectin encoded by a tumor fitness gene in CRISPR screens of glioma proliferation and a clinical drug target, for glioma overall (OR=1.37, p=1.3×10-5). Analyses of the plasma proteome using pQTLs for 1,362 proteins from 3 cohorts (N=88,838) identified 9 colocalized associations in whole blood. Most risk signals were subtype-specific, such as CRELD1 (p=2.5×10-5) for triple-negative glioblastoma. We found three candidate risk proteins for IDH-mutant 1p19q-intact glioma: immunoglobulin glycoprotein BCAM (p=9.6×10-8), neurotrophin receptor SorCS2 (p=3.6×10-5) and immune checkpoint PD-1 (p=1.1×10-5). Interestingly, genetically-predicted PD-1 levels had a larger effect on IDH-mutant 1p19q-intact (OR=2.75, 95% CI=1.61-4.69) than IDH-wildtype (OR=0.73, 95% CI=0.51-1.05) tumors, suggesting germline differences in immune checkpoint mechanisms among subtypes. For all colocalized candidates, genetically-predicted protein abundance was correlated with cis-regulated transcriptional activity in brain tissue. For example, genetically-predicted SORCS2 gene expression in brain tissue was associated with plasma SorCS2 levels (Z=16.1, p=1.1×10-68). Our analysis identified candidate proteins involved in immune response and neuronal proliferation, which may enable prioritization of drug targets and provide insight into disease mechanisms for glioma.

Abstract Background Polygenic risk scores (PRS) aggregate the contribution of many risk variants to provide a personalized genetic susceptibility profile. Since sample sizes of glioma genome-wide association studies (GWAS) remain modest, there is a need to efficiently capture genetic risk using available data. Methods We applied a method based on continuous shrinkage priors (PRS-CS) to model the joint effects of over 1 million common variants on disease risk and compared this to an approach (PRS-CT) that only selects a limited set of independent variants that reach genome-wide significance (P < 5 × 10–8). PRS models were trained using GWAS stratified by histological (10 346 cases and 14 687 controls) and molecular subtype (2632 cases and 2445 controls), and validated in 2 independent cohorts. Results PRS-CS was generally more predictive than PRS-CT with a median increase in explained variance (R2) of 24% (interquartile range = 11–30%) across glioma subtypes. Improvements were pronounced for glioblastoma (GBM), with PRS-CS yielding larger odds ratios (OR) per standard deviation (SD) (OR = 1.93, P = 2.0 × 10–54 vs. OR = 1.83, P = 9.4 × 10–50) and higher explained variance (R2 = 2.82% vs. R2 = 2.56%). Individuals in the 80th percentile of the PRS-CS distribution had a significantly higher risk of GBM (0.107%) at age 60 compared to those with average PRS (0.046%, P = 2.4 × 10–12). Lifetime absolute risk reached 1.18% for glioma and 0.76% for IDH wildtype tumors for individuals in the 95th PRS percentile. PRS-CS augmented the classification of IDH mutation status in cases when added to demographic factors (AUC = 0.839 vs. AUC = 0.895, PΔAUC = 6.8 × 10–9). Conclusions Genome-wide PRS has the potential to enhance the detection of high-risk individuals and help distinguish between prognostic glioma subtypes.

Abstract Background The shared inherited genetic contribution to risk of different cancers is not fully known. In this study, we leverage results from 12 cancer genome-wide association studies (GWAS) to quantify pairwise genome-wide genetic correlations across cancers and identify novel cancer susceptibility loci. Methods We collected GWAS summary statistics for 12 solid cancers based on 376 759 participants with cancer and 532 864 participants without cancer of European ancestry. The included cancer types were breast, colorectal, endometrial, esophageal, glioma, head and neck, lung, melanoma, ovarian, pancreatic, prostate, and renal cancers. We conducted cross-cancer GWAS and transcriptome-wide association studies to discover novel cancer susceptibility loci. Finally, we assessed the extent of variant-specific pleiotropy among cancers at known and newly identified cancer susceptibility loci. Results We observed widespread but modest genome-wide genetic correlations across cancers. In cross-cancer GWAS and transcriptome-wide association studies, we identified 15 novel cancer susceptibility loci. Additionally, we identified multiple variants at 77 distinct loci with strong evidence of being associated with at least 2 cancer types by testing for pleiotropy at known cancer susceptibility loci. Conclusions Overall, these results suggest that some genetic risk variants are shared among cancers, though much of cancer heritability is cancer-specific and thus tissue-specific. The increase in statistical power associated with larger sample sizes in cross-disease analysis allows for the identification of novel susceptibility regions. Future studies incorporating data on multiple cancer types are likely to identify additional regions associated with the risk of multiple cancer types.

Abstract Background Mendelian randomization (MR) studies are susceptible to metadata errors (e.g. incorrect specification of the effect allele column) and other analytical issues that can introduce substantial bias into analyses. We developed a quality control (QC) pipeline for the Fatty Acids in Cancer Mendelian Randomization Collaboration (FAMRC) that can be used to identify and correct for such errors. Methods We collated summary association statistics from fatty acid and cancer genome-wide association studies (GWAS) and subjected the collated data to a comprehensive QC pipeline. We identified metadata errors through comparison of study-specific statistics to external reference data sets (the National Human Genome Research Institute-European Bioinformatics Institute GWAS catalogue and 1000 genome super populations) and other analytical issues through comparison of reported to expected genetic effect sizes. Comparisons were based on three sets of genetic variants: (i) GWAS hits for fatty acids, (ii) GWAS hits for cancer and (iii) a 1000 genomes reference set. Results We collated summary data from 6 fatty acid and 54 cancer GWAS. Metadata errors and analytical issues with the potential to introduce substantial bias were identified in seven studies (11.6%). After resolving metadata errors and analytical issues, we created a data set of 219 842 genetic associations with 90 cancer types, generated in analyses of 566 665 cancer cases and 1 622 374 controls. Conclusions In this large MR collaboration, 11.6% of included studies were affected by a substantial metadata error or analytical issue. By increasing the integrity of collated summary data prior to their analysis, our protocol can be used to increase the reliability of downstream MR analyses. Our pipeline is available to other researchers via the CheckSumStats package (https://github.com/MRCIEU/CheckSumStats).

Abstract Background The etiological basis of glioma is poorly understood. We have used genetic markers in a Mendelian Randomisation (MR) framework to examine if lifestyle, cardiometabolic and inflammatory factors influence the risk of glioma. This methodology reduces bias from confounding and is not affected by reverse causation. Methods We identified genetic instruments for 37 potentially modifiable risk factors and evaluated their association with glioma risk using data from a genome-wide association study of 12,488 glioma patients and 18,169 controls. We used the estimated odds ratio of glioma associated with each of the genetically defined traits to infer evidence for a causal relationship with the following exposures: lifestyle and dietary factors (height, plasma IGF-1, blood carnitine, blood methionine, blood selenium, blood zinc, circulating adiponectin, circulating carotenoids, iron status, serum calcium, vitamin [A1, B12, B6, E and 25-hydroxyvitamin D], fatty acids levels [mono-unsaturated, omega-3 and omega-6] and circulating fetuin-A); cardiometabolic factors (birth weight, HDL cholesterol, LDL cholesterol, total cholesterol, total triglycerides, basal metabolic rate, body fat percentage, body mass index, fasting glucose, fasting proinsulin, HbA1C levels, diastolic and systolic blood pressure, waist circumference, waist-to-hip ratio) were included; inflammatory factors (C-reactive protein (CRP), plasma IL-6 sRa and serum IgE). Results After correction for the testing of multiple potential risk factors and excluding associations driven by one single nucleotide polymorphism (SNP) no significant association with glioma risk was observed (i.e. PCorrected > 0.05). Conclusions This study did not provide evidence supporting any of the 37 factors examined as having a significant influence on glioma risk.

Abstract BACKGROUND: Approximately 5% of gliomas occur in individuals with a family history of glioma, and first-degree relatives of brain tumor cases have a two-fold increase in risk of brain tumor. Recent somatic characterization has shown that tumors from familial cases are indistinguishable from sporadic cases, suggesting that familial cases may arise through similar mechanisms of gliomagenesis, and therefore may be associated with common variants as well as rare mutations. In this analysis, we assessed whether previously identified common risk variants are associated with familial glioma. METHODS: Data were obtained from the Glioma International Case Control (GICC) and Gliogene studies for 448 cases with reported family history, 4,405 cases without reported family history, and 3,288 controls. We assessed 25 risk loci previously identified by glioma GWAS, and odds ratios (OR) and 95% confidence intervals (95%CI) were calculated using an additive genetic logistic regression model adjusted for age, sex, and the first two principal components for familial cases versus unaffected controls, and non-familial cases versus controls. Results were considered significant at p<0.002 (Bonferroni correction for 25 tests). RESULTS: Significant associations were detected at 5/25 loci, including TERT, EGFR, CCDC26, CDKN2B, and RTEL1. The strongest association was at rs55705857 (CCDC26, OR=2.7, p=7.49x10-17). For GBM (222 familial cases), significant associations were detected at 6/26 loci (TERT, EGFR, CDKN2B, TP53 and RTEL1), while in non-GBM (205 familial cases) significant associations were detected at 3/25 loci (LRIG1, CCDC26, PHLDB1). These SNPs were further examined using a case-only approach comparing familial to non-familial cases, and there was no significant difference in allele frequencies by family history status. There was a strong correlation between log(OR) for familial cases only versus non-familial cases (adjusted R2=0.88). CONCLUSIONS: In this analysis we identified a significant association between familial glioma and five common risk loci previously identified by glioma GWAS. This provides further evidence of shared pathways of genetic risk and gliomagenesis between familial and non-familial glioma. Further exploration is necessary to determine the overall contribution of common genetic variation to risk of familial glioma. Citation Format: Quinn T. Ostrom, Georgina Armstrong, Christopher I. Amos, Jonine L. Bernstein, Elizabeth B. Claus, Jeanette E. Eckel-Passow, Dora Il'yasova, Christoffer Johansen, Daniel H. Lachance, Rose K. Lai, Ryan T. Merrell, Sara H. Olson, Joellen H. Schildkraut, Sanjay S. Shete, Richard S. Houlston, Robert B. Jenkins, Margaret R. Wrensch, Beatrice Melin, Jill S. Barnholtz-Sloan, Melissa L. Bondy. Previously identified common glioma risk SNPs are associated with familial glioma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 4173.

AbstractAn association between genetic variants in the vitamin D receptor (VDR) gene and epithelial ovarian cancer (EOC) was previously reported in women of African ancestry (AA). We sought to examine associations between genetic variants in VDR and additional genes from vitamin D biosynthesis and pathway targets (EGFR, UGT1A, UGT2A1/2, UGT2B, CYP3A4/5, CYP2R1, CYP27B1, CYP24A1, CYP11A1, and GC). Genotyping was performed using the custom‐designed 533,631 SNP Illumina OncoArray with imputation to the 1,000 Genomes Phase 3 v5 reference set in 755 EOC cases, including 537 high‐grade serous (HGSOC), and 1,235 controls. All subjects are of African ancestry (AA). Logistic regression was performed to estimate odds ratios (OR) and 95% confidence intervals (CI). We further evaluated statistical significance of selected SNPs using the Bayesian False Discovery Probability (BFDP). A significant association with EOC was identified in the UGT2A1/2 region for the SNP rs10017134 (per allele OR = 1.4, 95% CI = 1.2‐1.7, P = 1.2 × 10−6, BFDP = 0.02); and an association with HGSOC was identified in the EGFR region for the SNP rs114972508 (per allele OR = 2.3, 95% CI = 1.6‐3.4, P = 1.6 × 10−5, BFDP = 0.29) and in the UGT2A1/2 region again for rs1017134 (per allele OR = 1.4, 95% CI = 1.2‐1.7, P = 2.3 × 10−5, BFDP = 0.23). Genetic variants in the EGFR and UGT2A1/2 may increase susceptibility of EOC in AA women. Future studies to validate these findings are warranted. Alterations in EGFR and UGT2A1/2 could perturb enzyme efficacy, proliferation in ovaries, impact and mark susceptibility to EOC.

Glioma incidence is highest in non‐Hispanic Whites, and to date, glioma genome‐wide association studies (GWAS) to date have only included European ancestry (EA) populations. African Americans and Hispanics in the US have varying proportions of EA, African (AA) and Native American ancestries (NAA). It is unknown if identified GWAS loci or increased EA is associated with increased glioma risk. We assessed whether EA was associated with glioma in African Americans and Hispanics. Data were obtained for 832 cases and 675 controls from the Glioma International Case–Control Study and GliomaSE Case–Control Study previously estimated to have <80% EA, or self‐identify as non‐White. We estimated global and local ancestry using fastStructure and RFMix, respectively, using 1,000 genomes project reference populations. Within groups with ≥40% AA (AFR ≥0.4 ), and ≥15% NAA (AMR ≥0.15 ), genome‐wide association between local EA and glioma was evaluated using logistic regression conditioned on global EA for all gliomas. We identified two regions (7q21.11, p = 6.36 × 10 −4 ; 11p11.12, p = 7.0 × 10 −4 ) associated with increased EA, and one associated with decreased EA (20p12.13, p = 0.0026) in AFR ≥0.4 . In addition, we identified a peak at rs1620291 ( p = 4.36 × 10 −6 ) in 7q21.3. Among AMR ≥0.15 , we found an association between increased EA in one region (12q24.21, p = 8.38 × 10 −4 ), and decreased EA in two regions (8q24.21, p = 0. 0010; 20q13.33, p = 6.36 × 10 −4 ). No other significant associations were identified. This analysis identified an association between glioma and two regions previously identified in EA populations (8q24.21, 20q13.33) and four novel regions (7q21.11, 11p11.12, 12q24.21 and 20p12.13). The identifications of novel association with EA suggest regions to target for future genetic association studies.

Abstract Genome-wide association studies (GWAS) have so far identified 25 loci associated with glioma risk, with most showing specificity for either glioblastoma (GBM) or non-GBM tumors. The majority of these GWAS susceptibility variants reside in noncoding regions and the causal genes underlying the associations are largely unknown. Here we performed a transcriptome-wide association study to search for novel risk loci and candidate causal genes at known GWAS loci using Genotype-Tissue Expression Project (GTEx) data to predict cis-predicted gene expression in relation to GBM and non-GBM risk in conjunction with GWAS summary statistics on 12,488 glioma cases (6,183 GBM and 5,820 non-GBM) and 18,169 controls. Imposing a Bonferroni-corrected significance level of P &lt; 5.69 × 10−6, we identified 31 genes, including GALNT6 at 12q13.33, as a candidate novel risk locus for GBM (mean Z = 4.43; P = 5.68 × 10−6). GALNT6 resides at least 55 Mb away from any previously identified glioma risk variant, while all other 30 significantly associated genes were located within 1 Mb of known GWAS-identified loci and were not significant after conditioning on the known GWAS-identified variants. These data identify a novel locus (GALNT6 at 12q13.33) and 30 genes at 12 known glioma risk loci associated with glioma risk, providing further insights into glioma tumorigenesis. Significance: This study identifies new genes associated with glioma risk, increasing understanding of how these tumors develop.