Search

Histone H3.3 G34R/V mutations are drivers of pediatric high-grade glioma (pHGG). However, the mechanism(s) responsible for G34R/V induced tumorigenesis are unclear. We observed that pHGG cells with H3.3 G34 mutations have significantly reduced phosphorylation at H3.3 S31 as compared to H3.3 WT cell lines. And, in vitro, G34 mutation completely blocks H3.3 S31 phosphorylation by Chk1 – the kinase responsible for S31 phosphorylation. During early mitosis, phospho-S31 is enriched in the chromosomal pericentromere; which plays a vital structural role in chromosome segregation. Over-expression of GFP-H3.3 G34R or non-phosphorylatable GFP-H3.3 S31A in chromosomally stable cell lines results in a significant increase in chromosome mis-segregations. Likewise, H3.3 G34V pHGG derived cells showed a higher frequency of mis-segregation than H3.3 WT cells. During cell division, phospho-S31 is lost in late anaphase. However, when chromosome missegregation occurs, phospho-S31 spreads to the chromosome arms where it stimulates p53 accumulation in G1 – thus acting as a sensor to suppress the proliferation of aneuploid cells. Here we show that cells expressing mutant G34 fail to arrest following mis-segregation, despite having WT p53. Taken together this work demonstrates that the H3.3 G34R/V mutations are sufficient to transform normal, diploid cells into proliferative, chromosomally instable cells. To determine if this process contributes to tumorigenesis, we expressed WT H3.3, H3.3-G34R or H3.3-S31A in a mouse model of pHGG. H3.3 WT controls developed low-grade tumors and all survived 80 days. Mice with H3.3-G34R or H3.3-S31A developed low- and high-grade tumors. And 78% and 50% of H3.3-G34R and H3.3-S31A mice, respectively, survived to 80 days. Our work suggests that H3.3 G34R/V drives chromosomal instability through the suppression of H3.3 S31 phosphorylation AND that chromosomal instability is a contributing driver of glioma formation in G34 mutant tumors.

Pediatric glioblastoma and diffuse intrinsic pontine glioma are high-grade gliomas of children (pHGG) with a median overall survival of under 15 months and among the most lethal cancers. Mutations in histone H3.3 and H3.1 occur as an early event in pHGG. H3.3G34R/V-mutations occur in pHGG of cerebral hemispheres, and H3.3K27M mutations occur in midline pHGGs. Post-translational histone modifications (PTMs) serve to regulate gene expression by relaxing or compacting chromatin and by recruiting proteins, with subsequent silencing or activating effects. H3.3 Serine 31 (S31) shows reduced phosphorylation during mitosis in H3.3G34R/V and H3.3K27M mutant cell. Phosphorylation at S31 is restored in wildtype H3.3K27 CRISPR revertants. Serine to alanine (A) mutant H3.3 S31A are nonphosphorylatable in vitro. To study the influence of histone mutations and the role of altered PTM and including the loss of methylation and phosphorylation on tumorigenesis, we have developed an innovative model based on the RCAS/N-TVA mouse model. In this system, the expression of an oncogenic driver is linked to mutant histone expression using a self-cleaving peptide, and tumors develop following viral delivery to neural stem cells in newborn mice. This approach is necessary, as otherwise, clonal selection could prevent tumors from forming with mutations detrimental to growth. To establish the model, N-TVA mice were injected with RCAS H3.3K27M-P2A-PDGFB, RCAS H3.3G34R-P2A-PDGFB, or H3.3WT-P2A-PDGFB. The mean survival of mice injected with H3.3K27M and H3.3S31A was 81 and 68 days, respectively, and 100% of S31A mice developed HGG. In contrast, H3.3WT caused only low-grade tumors in 46% of the mice, and all mice survived until 100 days. In ongoing experiments with H3.3G34R, 23% of mice succumb to tumors by 80 days. These results provide mechanistic insights into the early establishment of pHGGs and established a new mouse model to study the role of histone mutation and PTMs in tumor development.

BACKGROUND: Heterozygous mutations in Histone H3.3 (K27M, G34R/V) occur in 70–80% of pediatric high-grade gliomas and 78% of diffuse intrinsic pontine gliomas (DIPG). These mutations are also found in low-grade gliomas that progress to higher grade tumors. Epigenetic dysregulation resulting from altered methylation at or near the mutated sites is believed to drive malignant transformation H3 differs from H3.1 by only 5 amino acids, which includes a unique phosphorylation site, Serine 31. S31 is phosphorylated by Chk1 kinase at pericentromeric heterochromatin in mitosis and on the arms of missegregated chromosomes. Masking S31 phosphorylation by antibody injection or using a non-phosphorylatable S31A substitution prevents cell cycle arrest in response to chromosome missegregation. OBJECTIVE: To determine if H3.3 K27M or G34R/V mutations alter the regulation of S31 phosphorylation to promote chromosome missegregation, aneuploidy, and tumorigenesis. METHODS: Live cell imaging and an established mouse model of glioma. RESULTS: DIPG cells with the K27M mutations show decreased mitotic Ser31 phosphorylation and increased chromosome missegregation. In both model cell lines and normal astrocytes, K27M and G34R mutations abolish S31 phosphorylation. Expression of S31A, K27M, or G34R mutants results in increased chromosome missegregation and aneuploidy, whereas cells expressing phosphomimetic double mutants (K27M/S31E or G34R/S31E) divide normally. Using the RCAS/TVA mouse model, H3.3WT, K27M, and S31A constructs were introduced into the brains of newborn mice along with mouse BRAF(V600E). 0/13 tumors developed in the BRAF(V600E) + H3.3WT cohort; however, 80% (16/20) of H3.3K27M + BRAF(V600E) and 70% (7/10) H3.3S31A + BRAF(V600E) cohorts developed high-grade gliomas within 100 days. CONCLUSION: A single heterozygous H3.3 mutation is sufficient to generate aneuploid cells, allows them to reenter the cell cycle and to promote tumorigenesis. Restoring cell cycle checkpoint activity by modulating S31 phosphorylation may be a rational treatment for DIPG or to prevent the progression of low-grade gliomas with H3.3 mutations.

Diffuse midline gliomas with the H3.3 K27M mutation are lethal brain tumors in children. H3 K27M causes global loss of Lys27 triple methylation (Lys27me3), inducing epigenetic reprograming. Here we show that H3.3 K27M also causes decreased H3.3 Ser31 phosphorylation on mitotic chromosomes. We show that H3.3 K27M DIPG cells have reduced pericentromeric phospho-Ser31 and increased rates of chromosome missegregation compared to normal, diploid human cells. CRISPR-editing K27M to M27K restored phospho-Ser31 to WT levels and dramatically decreased the rate of chromosome missegregation. We confirm that Chk1 is the H3.3 Ser31 kinase: K27M mutant H3.3 protein exhibits ~60% reduced Chk1 phosphorylation of Ser31 in vitro. Chk1 knockdown completely abolishes phospho-Ser31 in cells and these have increased rates of chromosome missegregation. In normal, diploid cells, expression of K27M or an S31A non-phosphorylatable mutant increased chromosome missegregation; this is suppressed by expressing a phosphomimetic double mutant (K27M/S31E) that restores phospho-Ser31. WT cells arrest following chromosome missegregation. However, cells expressing H3.3 K27M or S31A fail to arrest - despite having WT p53. Finally, we expressed H3F3AS31A and PDGFb in an RCAS/TVA mouse model of DIPG and ~80% developed diffuse high-grade brain tumors and show significantly decreased survival. Our results suggest that loss of phospho-Ser31 alone is oncogenic because H3.3 S31A-expressing cells are WT for K27me3. Our results demonstrate that H3.3 K27M inhibits Ser31 phosphorylation both in vitro and in vivo, leading to both chromosome missegregation and loss of subsequent G1 arrest – thus creating diffuse midline gliomas with both dynamic, complex karyotypes and epigenetic reprogramming.