Comprehensive germline genomic profiles of children, adolescents and young adults with solid tumors

Abstract

Compared to adult carcinomas, there is a paucity of targeted treatments for solid tumors in children, adolescents, and young adults (C-AYA). The impact of germline genomic signatures has implications for heritability, but its impact on targeted therapies has not been fully appreciated. Performing variant-prioritization analysis on germline DNA of 1,507 C-AYA patients with solid tumors, we show 12% of these patients carrying germline pathogenic and/or likely pathogenic variants (P/LP) in known cancer-predisposing genes (KCPG). An additional 61% have germline pathogenic variants in non-KCPG genes, including PRKN, SMARCAL1, SMAD7, which we refer to as candidate genes. Despite germline variants in a broad gene spectrum, pathway analysis leads to top networks centering around p53. Our drug-target analysis shows 1/3 of patients with germline P/LP variants have at least one druggable alteration, while more than half of them are from our candidate gene group, which would otherwise go unidentified in routine clinical care.

Fig. 1. Germline alterations and clinical outcomes in the Cleveland Clinic series.

a Genes with germline pathogenic/likely pathogenic (P/LP) variants in known cancer-predisposing genes (KCPG) and candidate genes and their type of alterations in children, adolescents, and young adult (C-AYA) patients with solid tumors. b Oncoplots of top mutated genes with P/LP variants in KCPG and candidate genes based on the age group. Each column represents one patient and its affected genes. c Two examples of copy number variations (CNVs) found in C-AYA patients with solid tumors. d The number of patients with germline alterations, both single-nucleotide variations (SNVs) and CNVs, in each tumor type. e Clinical outcome comparison between two groups of C-AYA patients with solid tumors, with and without germline alterations. Gray color represents the number of patients with the specified clinical outcome in each group. Two-sided Fisher's exact test was implemented, P = 0.31, OR = 2.14, 95% CI = 0.6–8.0. ACT adrenocortical carcinoma, CNS central nervous system, EWS Ewing sarcoma, GCT germ cell tumor, HGG high-grade glioma, LGG low-grade glioma, LMPRT low malignancy potential renal tumor, NBL neuroblastoma, NM non-malignant tumor, OS osteosarcoma, RCC renal cell carcinoma, RHB rhabdomyosarcoma, STS soft tissue sarcoma, WLM Wilms tumor, Del deletion, Ins insertion.

Fig. 2. Distribution of germline pathogenic/likely pathogenic (P/LP) mutations in children, adolescents, and young adults (C-AYA) with solid tumors.

a Top mutated genes with germline P/LP variants in KCPG (left panel) and candidate genes (right panel) and their type of alterations. b Oncoplots of top mutated genes with P/LP variants in known cancer-predisposing genes (KCPG) (top panel) and candidate genes (lower panel) based on the age group. Each column represents one patient and its affected genes. c Most frequently mutated genes with P/LP variants in KCPG (left panel) and candidate genes (right panel) based on their affected tumor types. ACT adrenocortical carcinoma, BCC basal cell carcinoma, CA carcinoma, CNS central nervous system, EWS Ewing sarcoma, GCT germ cell tumor, GICT giant cell tumor, HGG high-grade glioma, LGG low-grade glioma, LMPRT low malignancy potential renal tumor, NBL, neuroblastoma, NM non-malignant tumor, OS osteosarcoma, OST other solid tumors, PGL paraganglioma, RB retinoblastoma, RCC renal cell carcinoma, RHB rhabdomyosarcoma, STS soft tissue sarcoma, WLM Wilms tumor, Del deletion, Ins insertion.

Fig. 3. Germline genomic signatures of children, adolescents, and young adults (C-AYA) with 12 types of solid tumors.

a, b Germline gene cloud signatures of the C-AYA patients with solid tumors based on their altered known cancer-predisposing genes (KCPG) (a) or candidate genes (b). The size of the genes is proportional to their pathogenic/likely pathogenic (P/LP) variant frequency in that tumor type, colors do not specify any meaning. c, d Heat maps of top altered KCPGs (c) or candidate genes (d). Two-sided Fisher´s Exact test implemented in R statistical software. P values were adjusted for multiple testing with Bonferroni correction considering 593 tests. FDR threshold of 0.05 considered a significant event. Scale refers to log10 (frequency of P/LP variants in specified genes in each tumor type). Blue rectangles specify significant corrected P values in comparison to non-TCGA ExAC database. ACT adrenocortical carcinoma, CNS central nervous system, EWS Ewing sarcoma, GCT germ cell tumor, HGG high-grade glioma, LGG low-grade glioma, NBL neuroblastoma, OS osteosarcoma, RB retinoblastoma, RHB rhabdomyosarcoma, STS soft tissue sarcoma, WLM Wilms tumor.

Fig. 4. Pathway analysis of altered genes with germline pathogenic/likely pathogenic (P/LP) variants in children, adolescents, and young adults (C-AYA) with solid tumors.

a Affected pathways based on altered genes with P/LP germline variants. Top panel: only known cancer-predisposing genes (KCPG), lower panel: a combination of all KCPGs and candidate genes. Size of the circles increases as the fraction affected increases. b Genes mutated in TP53 (top panel) and RAS–RTK (lower panel) pathways, and the number of patients affected in our cohort. Red font: tumor suppressor genes; blue font: oncogenes. c Top network, predicted by Ingenuity Pathway Analysis (IPA), based on all the KCPG (green color) and candidate genes (salmon color) with at least four P/LP variants in our C-AYA patients with solid tumors (right-tailed Fisher’s exact test P = 1 × 10⁻⁴²). d Eukaryotic Translation Initiation Factor 4 Gamma 1 (EIF4G1, B–H corrected P = 1.39 × 10⁻³) and I kappa b kinase (IκB kinase, B–H corrected P = 1.39 × 10⁻³) predicted to be the top upstream regulators/causal network based on our IPA analysis. Right-tailed Fisher’s exact test was used, and Benjamini–Hochberg (B–H) P value correction performed to reduce the false discovery rate (FDR).

Fig. 5. Drug–target network analysis in children, adolescents, and young adults (C-AYA) with solid tumors.

Known cancer-predisposing genes (KCPG) and candidate genes with germline pathogenic/likely pathogenic (P/LP) variants in C-AYA patients with solid tumors that have existing FDA-approved antineoplastic and immunomodulating-related compounds, in regard to their affected tumor types. ACT adrenocortical carcinoma, CNS central nervous system, EWS Ewing sarcoma, GCT germ cell tumor, HGG high-grade glioma, LGG low-grade glioma, NBL neuroblastoma, NM non-malignant tumor, OS osteosarcoma, OST other solid tumors, PGL paraganglioma, RB retinoblastoma, RHB rhabdomyosarcoma, STS soft tissue sarcoma, WLM Wilms tumor.