Previous synthetic lethal screening efforts Cox et
Previous synthetic lethal screening efforts (Cox et al., 2014) have mainly used RNAi as a means of identifying potential targets (Barbie et al., 2009, Kim et al., 2016, Luo et al., 2009, Scholl et al., 2009), although a few screens (Shaw et al., 2011, Steckel et al., 2012) have been performed with small chemical libraries. In all cases, these screens have relied on adherent monolayer (2D) cell cultures, and in most cases have used only a matched pair of mutant/wild-type (WT) cell lines. To improve the probability of finding translatable targets, it might be beneficial to use more physiologically relevant cellular models and to screen larger panels of cell lines (Cox et al., 2014).
In an effort to discover translatable mutant KRAS synthetic lethal targets, we have performed a 3D clonogenic synthetic lethal screen with a diverse small-molecule library of 280,000 compounds across six mutant KRAS-dependent cell lines and four KRAS-independent cell lines. We report here the results from this screen, and the identification and characterization of dihydroorotate dehydrogenase (DHODH) as the target of one of the most KRAS mutant-selective compounds. As expected, DHODH inhibition prevents de novo pyrimidine biosynthesis, which appears to be a particularly important pathway in the growth/survival of KRAS mutant cells. DHODH inhibition also substantially decreases cellular levels of glutamine and glutamate, suggesting a connection to the previously established glutamine dependence of KRAS mutant BINA mg (Son et al., 2013, Weinberg et al., 2010). We also show that brequinar, a DHODH inhibitor that has previously failed to demonstrate efficacy in clinical trials as an anticancer agent (Moore et al., 1993), exhibits strong in vivo antitumor activity in a KRAS mutant pancreatic tumor xenograft model. The links observed in vitro between KRAS status and a requirement for metabolic flux through the de novo pyrimidine biosynthetic pathway suggest new strategies for the clinical application of potent DHODH inhibitors against KRAS mutant cancers.
Discussion The current study integrates results from both a new screen in 3D culture and more conventional screens in 2D culture. Screening in 3D culture has the advantage of sensitizing KRAS mutant cells to Raf-MEK-ERK pathway inhibition (Figure 1D), so that the phenotypes observed may be more directly translatable to in vivo systems. Empirically, primary screening in 3D culture enabled the discovery of KRAS synthetic lethal mechanisms that had not previously been reported. However, screening in large numbers of different cell lines proved to be more technically tractable for 2D cell cultures. While results from both kinds of screens were broadly consistent with each other, responses of 2D-cultured lines to synthetic lethal perturbation varied widely (see, e.g., Figure 2A). These results suggest that 3D culture systems are useful for studying KRAS mutant phenotypes, and that results from small numbers of cell lines cultured in 2D may or may not be predictive of results from 3D cultures and/or in vivo biology. The primary finding of the current work is that KRAS mutant cells exhibit a synthetic lethal sensitivity to inhibitors of DHODH. However, it is clear that inhibition of DHODH has pleiotropic effects in cancer cells. Multiple downstream consequences of DHODH inhibition may thus contribute to the selective effect of DHODH inhibitors. Results in cells and in vivo (Figures 5, 6A, and 6B) indicate that, as expected, DHODH inhibition leads to decreased flux through the de novo pyrimidine biosynthetic pathway. In addition, cell growth inhibition by DHODH inhibitors is rescued by uridine (Figure 4D), suggesting that inhibition of de novo pyrimidine biosynthesis contributes to cytotoxicity. The latter observation further suggests that de novo pyrimidine biosynthesis is required for growth and survival in KRAS mutant cells. Previous studies have also linked Ras-Raf-MEK-ERK signaling with increased de novo pyrimidine biosynthesis. Decreased expression of an inducible mutant KRAS downregulates pyrimidine biosynthetic genes at the transcriptional level (Ying et al., 2012). In addition, MAPK signaling abolishes feedback inhibition of carbamoyl phosphate synthetase, which catalyzes the rate-limiting step in de novo pyrimidine biosynthesis (Graves et al., 2000). Notably, glutamine deprivation in KRAS-driven cancer cell lines has recently been shown to interfere with deoxynucleotide biosynthesis, leading to S-phase arrest and replicative stress (Patel et al., 2016). Deprivation of pyrimidine nucleotides is thus likely to be a major contributor to the differential effect of DHODH inhibitors on cell viability in KRAS mutant cell lines.