5-azacytidine reduces methylation, promotes differentiation and induces tumor regression in a patient-derived IDH1 mutant glioma xenograft

Abstract

Somatic mutations in Isocitrate Dehydrogenase 1 (IDH1) are frequent in low grade and progressive gliomas and are characterized by the production of 2-hydroxyglutarate (2-HG) from α-ketoglutarate by the mutant enzyme. 2-HG is an "oncometabolite" that competitively inhibits α-KG dependent dioxygenases resulting in various widespread cellular changes including abnormal hypermethylation of genomic DNA and suppression of cellular differentiation. Despite the growing understanding of IDH mutant gliomas, the development of effective therapies has proved challenging in part due to the scarcity of endogenous mutant in vivo models. Here we report the generation of an endogenous IDH1 anaplastic astrocytoma model which rapidly grows in vivo, produces 2-HG and exhibits DNA hypermethylation. Using this model, we have demonstrated the preclinical efficacy and mechanism of action of the FDA approved demethylating drug 5-azacytidine in vivo. Long term administration of 5-azacytidine resulted in reduction of DNA methylation of promoter loci, induction of glial differentiation, reduction of cell proliferation and a significant reduction in tumor growth. Tumor regression was observed at 14 weeks and subsequently showed no signs of re-growth at 7 weeks despite discontinuation of therapy. These results have implications for clinical trials of demethylating agents for patients with IDH mutated gliomas.

Figure 1. Characteristic histological and genetic features of the IDH1 (R132H) anaplastic astrocytoma model

(A) H&E sections of the primary tumor and subsequent flank xenografts show histopathological similarity to the original tumor including a typical astrocytic morphology, gemistocytic cells and mitotic figures. (B) Immunohistochemical staining specific for the IDH1 (R132H) mutant protein shows robust staining in the original tumor and all subsequent xenografts. (C) Sequencing of exon 4 of IDH1 shows an initial heterozygous G/A mutation in the original patient tumor which converts to a hemizygous genotype when the wild type copy is lost in the xenograft.

Figure 2. Growth in vivo and 2-HG production in the IDH mutant model

(A) Subcutaneously implanted flank xenografts grow to maximum size (2.0 cm³) in approximately 7 weeks. (B) IDH mutant xenograft produces high levels of 2-HG as measured by LC/MS. Error bars=SEM.

Figure 3. JHH-273 shows characteristic DNA hypermethylation which can be reversed with 5-azacytidine treatment in vivo

(A) Pyrosequencing shows that the original patient tumor exhibits high levels of DNA methylation in several target genes. This hypermethylated phenotype is maintained in the xenograft and is characteristic of IDH1 mutant gliomas. In contrast, IDH1 wild type glioma is not hypermethylated. (B) 5-azacytidine treatment for 1 cycle reverses methylation at several targets in a dose specific manner. * p<0.05, Error bars=SEM.

Figure 4. Treatment strategy for 5-azacytidine in the IDH1 mutant flank model

Flank tumors were treated with 5-azacytidine starting 5 days after implantation until maximum tumor size was reached (Cycle 1). In order to extend the relative time of treatment, the flank tumors were individually passaged and 5-azacytidine treatment resumed immediately following implantation (Cycle 2). Tumor passaging was repeated a third time into two groups, one which immediately resumed 5-azacytidine treatment or one which had treatment withdrawn (Cycle 3).

Figure 5. Long term treatment with 5-azacytidine reduces tumor growth in an IDH1 mutant model

(A) Mice bearing IDH1 mutant flank tumors and treated with 5-azacytidine show a reduction in tumor burden (B) Cycle 2 of 5-azacytidine treatment significantly reduces tumor growth compared to untreated tumors (C) Pre-treatment with 5-azacytidine arrests tumor growth, even after treatment is withdrawn. Cycles 1 and 2: Five mice (10 tumors) per treatment group, Cycle 3: Four mice (4 tumors) per treatment group. * p<0.01, Error bars=SEM,

Figure 6. Treatment with 5-azacytidine induces differentiation and reduces the proliferative index in an in vivo IDH1 (R132H) glioma model

(A) 5-azacytidine treatment causes loss of DNMT1 expression in vivo following one treatment cycle. (B) GFAP expression is restored following one passage of 5-azacytidine treatment and is maintained. (C) Immunohistological staining of GFAP shows significant increase of protein expression in the cytoplasm of 5-azacytidine treated cells (D) Ki67 staining shows a decrease in the proliferative index of 5-azacytidine treated cells in a time dependent manner.