Rare cases of medulloblastoma with hypermutation

Abstract Background Medulloblastoma is the most common malignant brain tumor of childhood and is considered a tumor with low mutational burden (~1 Mut/Mb). Therefore, though the medulloblastoma genomes have been extensively characterized in literature, reports on potential hypermutations and underlying mutagenic processes in medulloblastomas are limited. Aim In this report, we studied the landscape of mutational burden in primary and recurrent medulloblastoma. Furthermore, we wanted to understand the differences in underlying mutagenic mechanisms in medulloblastoma with low and high mutational burdens. Methods Fifty‐three primary and recurrent medulloblastoma genomic sequence were downloaded from the European Genome Archive as BAM files. Thirty‐three cases were obtained from formalin‐fixed paraffin‐embedded tissues from pathology diagnostic archives of Spectrum Health and Cooperative Human Tissue Network. Somatic mutations were called using Mutect2, following best practices guidelines for Genome Analysis Toolkit V4. Mutational signatures were analyzed using deconstructSigs. Results We identified nine medulloblastoma cases with high mutational burden (>5 Mut/Mb). Of them, five cases met the criteria of hypermutation (>10Mut/Mb), two of the five tumors had canonical mutations in the POLE proof‐reading domain, where a large proportion of mutations in these tumor genomes contributed to signature 10. The hypermutated cases also demonstrated mutational signatures 14, 15, and 21, indicating the role of mis match repair deficiency in their mutagenesis. Of the four known molecular subgroups in medulloblastoma–SHH, WNT, Group 3, and Group 4—both the POLE‐mutated cases belonged to the SHH subgroup. This report identifies rare cases of hypermutation in medulloblastoma driven by defects in DNA repair mechanisms. Conclusion Hypermutation in medulloblastoma can impact therapeutic decisions, especially at recurrence in otherwise fatal high risk SHH‐medulloblastomas. A defect in DNA repair leading to SHH ‐medulloblastoma is yet another important mechanism that should be further investigated in the genesis of these tumors. Therefore, this report provides important scientific and clinical rationale for future research looking for incidence of hypermutation in large cohorts of medulloblastoma patients.


tional burdens.
Methods: Fifty-three primary and recurrent medulloblastoma genomic sequence were downloaded from the European Genome Archive as BAM files. Thirty-three cases were obtained from formalin-fixed paraffin-embedded tissues from pathology diagnostic archives of Spectrum Health and Cooperative Human Tissue Network.
Somatic mutations were called using Mutect2, following best practices guidelines for Genome Analysis Toolkit V4. Mutational signatures were analyzed using deconstructSigs. that should be further investigated in the genesis of these tumors. Therefore, this report provides important scientific and clinical rationale for future research looking for incidence of hypermutation in large cohorts of medulloblastoma patients.  Whereas most MB genomes (89.5%) had a low TMB, we identified 9 (10.5%) cases as outliers, of which 5 (5.8%) had TMB > 10 Mut/Mb, meeting criteria of hypermutated tumors 10 ( Figure S1A).

Results
We evaluated mutational signatures to establish differences among low and high TMB cases. 8,11 Mutagenesis leaves marks on DNA (e.g., nucleotide substitutions), creating unique signatures. The initial definition of such mutagenic signatures reveals 21 signatures in human cancers. 12 We cataloged 486 078 exonic and intronic mutations by nucleotide context (bases immediately preceding and following it, forming a trinucleotide). Using these cataloged trinucleotides (96 subtypes), we performed linear regression analysis using decon-structSigs 13 to identify fractions of mutations contributing to previously established mutational signatures. 12 Due to few mutations, exomes were not analyzed for mutational signatures.
In T-10, with TMB of 37.5 Mut/Mb, 68% of mutations contributed to signature 10. This signature, characterized by C > A substitution in TpCpT and C > T substitutions in the TpCpG context, is specifically associated with loss-of-function mutations in the exonuclease or proofreading domain of POLE. 8 T-1, with a TMB of 39.5 Mut/Mb, had 9% mutations contributing to signature 10 ( Figure 1(B)). We identified missense mutations p.R821C, p.D391E, and p.V411L in the POLE coding region in T-1 and T-10 (Table 1, Figure 1(C), and Figure S1B).
Presence of POLE mutations and signature 10 in a hypermutated tumor suggested that these POLE mutations were pathogenic. Furthermore, mutation in position 411 that cause amino acid Valine to leucine 10 switch is known to be pathogenic. Therefore, we inferred that T-10 with V411L mutation in the "proof reading" domain of POLE was hypermutated secondary to this mutation. Indirect evidence of pathogenicity of the POLE mutation was determined by  (Figure 1(B)), these signatures has been functionally linked to mismatch repair deficiency (MMRD). 10 We did not identify somatic or germline mutations  of mutations in HAT/HAT complex has been reported as a characteristic feature of SHH-MB. 16 No mutations in CTNNB1 and other genes unique to WNT-MB were identified in the hypermutated cases. Both T-1 and T-10 harbored molecular high-risk features. Mutational analysis in T-1 revealed TP53 mutation (Figure 1(E)), copy number analysis showed a 14q loss in T-10 ( Figure S2) which is known cytogenetic characteristic of SHH-MB and is associated with poor prognosis. 18 In the remaining hypermutated tumors, mutations in T-13 and T-13-R contributed to signatures 1, 5, 12, and 16. Underlying mutagenesis driving signatures 12 and 16 remains unknown. A higher mutation rate in the coding region of a tumor genome is associated with generation of structurally and functionally altered epitopes or possible neoantigens. 22 Neoantigens can trigger a rapid immunologic cytotoxic CD8+ T-cell response often accompanied by several immune checkpoints to attenuate this effect. 16

AUTHOR CONTRIBUTIONS
All authors had full access to the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

CONFLICT OF INTERESTS
The authors have stated explicitly that there are no conflicts of interest in connection with this article.

ETHICAL STATEMENT
Use of participants' tissues in genetic studies, along with waiver of consent and waiver of HIPAA authorizations were approved by Institutional Review Boards of Van Andel Research Institute and Spectrum Health Helen DeVos Children's Hospital. Permission to download whole genome sequencing (WGS) data of the primary and recurrent MBs was obtained from the European Genome Archive (EGA).

DATA AVAILABILITY STATEMENT
The whole genome sequencing data that support the findings of this study are available on request from the corresponding author upon a reasonable request. Whole exome sequencing data cannot be shared due to ethical reasons.