Passenger deletions generate therapeutic vulnerabilities in cancer

Abstract

Inactivation of tumour-suppressor genes by homozygous deletion is a prototypic event in the cancer genome, yet such deletions often encompass neighbouring genes. We propose that homozygous deletions in such passenger genes can expose cancer-specific therapeutic vulnerabilities when the collaterally deleted gene is a member of a functionally redundant family of genes carrying out an essential function. The glycolytic gene enolase 1 (ENO1) in the 1p36 locus is deleted in glioblastoma (GBM), which is tolerated by the expression of ENO2. Here we show that short-hairpin-RNA-mediated silencing of ENO2 selectively inhibits growth, survival and the tumorigenic potential of ENO1-deleted GBM cells, and that the enolase inhibitor phosphonoacetohydroxamate is selectively toxic to ENO1-deleted GBM cells relative to ENO1-intact GBM cells or normal astrocytes. The principle of collateral vulnerability should be applicable to other passenger-deleted genes encoding functionally redundant essential activities and provide an effective treatment strategy for cancers containing such genomic events.

Figure 1. Homozygous deletions in ENO1 sensitize tumors to molecular targeting of ENO2

a, ENO1 is homozygously deleted in glioblastomas as part of the 1p36 locus. Loss of ENO1 is tolerable to the tumor because ENO2 is still expressed. b, A specific inhibitor of ENO2 should completely eliminate enolase activity in ENO1 null tumor cells (hence blocking glycolysis and ATP synthesis) but leave genomically intact normal tissues unaffected because enolase activity is still present because ENO1 is still expressed.

Figure 2. Homozygous deletion of the 1p36 locus in GBM results in loss of ENO1 expression in primary tumors and cell lines

a, TCGA Affymetrix aCGH data show four primary GBMs with log2 copy number < −1, indicating homozygous deletion of the 1p36 locus. b, DNA copy number correlates with mRNA expression; expression is highest in tumors with n = 2 copies (WT) and lowest in tumors with n = 0 copies (null) of ENO1. c, The D423-MG cell line was identified as homozygously deleted by SNP arrays from the Wellcome Trust Sanger Institute data set. d, The complete absence of enolase 1 protein in D423-MG and Gli56 cells was confirmed by western blotting. e, shRNA knock-down of ENO2 in D423-MG ENO1 null cells ablated intracranial tumorigenesis in vivo (n=4 mice per group).

Figure 3. shRNA ablation of ENO2 affects ENO1-null but not ENO1-WT GBM cells

a, shRNA ablation by two independent doxycycline (Dox)-inducible TRIPZ hairpins against ENO2 (shENO2-3, shENO2-4) resulted in >70% reduction in enolase 2 protein levels in both ENO1 WT (A1207, U87, LN319) and ENO1-null (D423-MG) cell lines. b, Ablation of ENO2 dramatically inhibited growth of ENO1-null but not ENO1 WT cells, whereas non-targeting shRNA against luciferase (shLuc) had no effect in any cell line (n=3 biological replicates, S.E.M, t-test). Representative plates at the last time point of growth for cells infected with shLuc, shENO2-3, or shENO2-4, with or without Dox induction, are shown alongside growth curves for each cell line.

Figure 4. Extreme sensitivity of ENO1-null cells to the pan-enolase inhibitor PhAH

a, D423-MG and Gli56 ENO1-null lines are highly sensitive to PhAH toxicity while ENO1 WT cell lines and normal astrocytes are not. b, The sensitivity of GBM lines to PhAH treatment correlated with their overall enolase activity. Pre-incubation of the lysates with 1 µM PhAH inhibited enolase enzymatic activity by >95% (average, n = 2 technical replicates). c, PhAH minimally affected the growth of ENO1 WT GBM cells and normal astrocytes except at concentrations higher than 50 µM. Low concentrations of PhAH stall the growth of ENO1-null cells, while ENO1 heterozygous cells (D502-MG and U343) showed intermediate sensitivity (n = 4 biological replicates, S.E.M.).