Mutations of H3.3 and H3.1 in a large cohort of gliomas


Ashley Sumrall, Sandeep Mittal, Surasak Phuphanich, Amy Heimberger, Jenny Eschbacher, Michael Korn, Zoran Gatalica, Manjari Pandey, Joanne Xiu


Mutations in the histone genes H3.3 and H3.1 are driver events in pediatric and adult gliomas and carry diagnostic and prognostic importance for tumors originating from midline structures. Patients with tumors affected by these mutations are difficult to treat. We surveyed a large cohort of gliomas for H3-mutations, including H3K27m. Consecutive gliomas submitted for tumor profiling at Caris Life Sciences from 2015- 2019 were analyzed. NextGen sequencing was done on 592 genes; MGMT promoter methylation was tested by pyrosequencing; and EGFRvIII and gene fusions were tested by RNA-sequencing. Of nearly 1800 tumors analyzed, 41 harbored H3F3A alterations, including 33 with the K27M mutation (4 arose from the spinal cord, 1 from cerebellum, 1 from brain stem, 4 from thalamus, and 23 from brain, NOS). Eight G34R mutations were identified. A HIST1H3B-K27M was detected in a tumor from the brain stem. H3 mutations were more prevalent in pediatric tumors, and all H3 mutations seen in pediatric tumors were from grade IV tumors. Among the H3-mutated adult tumors, histology differed. There were 2 grade II tumors, 1 low grade glioma, 1 anaplastic ganglioglioma, and 2 anaplastic astrocytomas. In the investigated cohort, H3-mutations were mutually exclusive of IDH1/2 mutations and EGFR alterations. Significantly higher mutation rates were seen in H3-mutated tumors for TP53, ATRX, NF1, PDGFRA, FGFR1, FBXW7, BLM, and TSC2 compared with H3-WT. The H3-WT tumors were more enriched for MGMT-methylation and PTEN mutation. In H3-mutated tumors that were MGMT-methylated, most H3-mutations seen were G34R while K27M was largely exclusive. There was a heterogeneous distribution of H3 mutations, and the co-occurring molecular alterations seen in H3-mutated tumors further support the hypothesis that these tumors are a distinct molecular entity. By better characterizing these associations, we can develop insight into novel treatment strategies for a class of tumors with historically dismal prognosis.

External Link