A status-quo in targeted cancer therapy is that out of the thousands of somatic alterations found in cancer, alterations only in driver genes like oncogenes and tumor suppressors determine therapeutic strategy. For example, cancers with deletion/mutation of the driver tumor suppressor gene PTEN and consequently elevated AKT and mTOR signaling are considered rational candidates for PI3K/AKT/mTOR inhibitor therapy. Accordingly, PI3K/mTOR inhibitors are approved or in clinical trials for various cancers with PTEN/PI3K alterations. However, in glioblastoma (GBM) where PTEN loss of function occurs in over 60% of patients, PI3K/AKT/mTOR inhibitors have been largely ineffective. It is often overlooked that when tumor suppressor genes undergo deletion, nearby genes also undergo inadvertent co-deletion. These bystander genes are not necessarily tumor suppressor genes. In fact, many of them are important for cell growth and survival. In some cases, such bystander deletion events create a unique drug sensitivity specifically in cancer cells. For example, deletion of one of the two alleles of POLR2 (a subunit of RNA polymerase II co- deleted as a bystander to P53 deletion) reduces the amount of POLR2 protein and creates high sensitivity of these cells to low dose POLR2 inhibitors. There are several other such bystander deletion events in cancer that causes vulnerability specifically to cancer cells (e.g., PSMC2 deletion due to chromosome 7q22 loss, Enolase 1 deletion due to loss of 1p36 tumor suppressor locus, MAGOHB deletion as part of chromosome 1p loss, and MTAP co-deletion with the tumor suppressor CDKN2A). Searching the TCGA database we have identified that a crucial lipogenic gene is hemizygously deleted as bystander to the tumor suppressor PTEN (on chromosome 10) in glioblastoma, melanoma and prostate cancer. The fatty acid synthesized by the lipogenic enzyme is also present in our diet. Therefore, when we reduced this fatty acid from diet, inhibition of residual activity of the lipogenic gene with specific inhibitors killed glioblastoma and melanoma cells. This subset (subset 1) ultimately acquired drug resistance through a stress response pathway, and were eliminated by a specific pre-clinical grade inhibitor of the stress pathway. During our analysis, we also surprisingly discovered that this lipogenic gene in completely suppressed in a second GBM subset (subset 2) due to a combination of deletion and methylation. Subset 2 lost a gene that is important for growth and proliferation, and yet thrived, through yet unknown alternative mechanisms. Due to loss of the target lipogenic gene, subset 2 lines were completely resistant to the lipogenic enzyme inhibitor. Investigating the mechanism of survival of subset 2 GBM is outside the scope of this application. In this proposal we will use a repertoire of primary GBM lines and test if deletion and methylation status of the lipogenic gene can be used as biomarkers for inhibitor therapy in combination with a custom medicinal diet. Secondly, we will perform molecular, pharmacokinetic/pharmacodynamic and preclinical studies to address the mechanism of acquired resistance of subset 1 GBM. These tests will be performed in a well-established preclinical mouse model of intracranial glioma.
Glioblastoma (GBM) is lethal brain tumor with a median survival of 15 months. Therefore novel therapy is urgently needed. Following analysis of data from The Cancer Genome Atlas, and subsequent experimental validation studies, we have identified a subgroup of GBM patients who can be treated with a novel combination of a specific drug and a research diet. We expect that our studies will shed new insights into a novel glioblastoma therapy.