ID #324 Trans-species analysis of DNA replication-repair deficient, IDH mutant gliomas unravels unique genomic and immune landscape, leading to successful translation to combined immune and targeted therapy
Anirban Das, Vienna Mazzoli, Nicholas Fernandez, Olha Kos, Nuno Nunes, Katharine O’ Flaherty, Hope Friedman, Olfat Ahmad, Abigail Suwala, Kevin Bielamowicz, Gadi Abebe-Campino, Shani Caspi, Per Nyman, Richard Graham, John Kim, Mari Wilhelmsson, Mette Jorgensen, Orli Michaeli, Maria Baro, Alyssa Reddy, Nicholas Llosa, Amanda Li, Adrian Levine, Lucie Stengs, Logine Negm, Vanessa Bianchi, Melissa Edwards, Birgit Ertl-Wagner, Julie Bennett, Stefan Pfister, Peter Dirks, Eric Bouffet, Cynthia Hawkins, Uri TaboriAbstract
Background
IDH mutant (mIDH) gliomas are the most common primary brain tumors in age <50 years. Checkpoint-inhibitors have historically failed to demonstrate responses after their inevitable malignant transformation. We identified that primary DNA replication-repair deficiency (RRD) drives >60% of malignant mIDH malignant gliomas in age <40 years and have an exceptionally poor outcome. We aimed to decipher their biology and identify novel vulnerabilities.
Methods
Multi-omic analyses were performed on mIDH RRD-glioma registered to the International RRD Consortium and a novel immunocompetent mouse model was generated.
Results
mIDH RRD-glioma was frequent in older children and young adults with CMMRD and Lynch syndrome, majority were high-grade, with diffuse frontal lobe or multifocal distribution. Mutation burden was lower than IDH-wildtype RRD-glioma (TMB, 113 vs 28 mutations/Mb; p < 0.05). Loss of TP53, ATRX and CDKN2A/2B, and global hypomethylation were observed. CD8-T-cell infiltration (immunohistochemistry) and tumor inflammation score (transcriptome) were lower than IDH-wildtype RRD-glioma (p < 0.05), contributing to extremely poor survival, even after anti-PD1 monotherapy (p < 0.05). A novel mouse model (Olig2Cre+/Msh2LoxP/LSL-Idh1R132H) demonstrated similar lower TMB, diffuse involvement and low immune infiltrate, mimicking human disease. Cell lines demonstrated high levels of 2-hydroxyglutarate, an immuno-suppressive oncometabolite, that was reduced following IDH-inhibition. Adding a 1st-generation IDH-inhibitor (ivosidenib) after progression on anti-PD1 demonstrated objective responses and prolonged survival in refractory human patients (n = 13; p = 0.01).
Conclusion
Tumors in patients with hypermutant mIDH RRD-glioma that do not respond to anti-PD1 monotherapy harbor an immune-suppressed microenvironment that can be overcome by IDH-inhibition. Additional data demonstrating superior efficacy and CNS-penetration of vorasidenib (2nd-generation IDH1/IDH2 inhibitor), and higher sensitivity of RRD-gliomas to combined PD1+LAG3 inhibition, have now provided the rationale for developing a prospective international clinical trial in adolescents and young adults with mIDH RRD-glioma spanning five countries.