In this study, we generated a new KRASG12V leukemia model in Drosophila by expressing human oncogenic KRASG12V in fly hemocytes, which led to a nearly 100-fold increase in hemocyte proliferation and adult fly lethality. Using this easy-to-score phenotype, we designed highly efficient genetic and drug screens to identify genes and compounds able to inhibit oncogenic KRAS. Fortuitously, the parallel genetic and drug screens converged on the same pathways, as they identified both the key genes involved in the hypoxia pathway (sima and tgo), as well as the hypoxia inhibitor echinomycin as potent oncogenic KRAS antagonists. This discovery has provided strong evidence supporting an essential role for the hypoxia pathway in the oncogenic KRAS-induced cancer phenotype and has identified echinomycin as a promising hypoxia-targeting drug.

The genetic screen identified sima and tgo as mediators of the KRASG12V pathomechanism. The two genes encode the fly homologs of human HIF1A and ARNT, respectively. HIF activation is a well-established feature of cancer cell survival in solid tumors, as cells positioned in the tumor interior are subject to severe oxygen deprivation, i.e. hypoxia. Activated HIFs lead to changes in the microenvironment that support tumor survival and growth, such as angiogenesis (Bousquet et al., 2016; Talks et al., 2000). In addition to the role of HIFs in solid tumors, HIF-mediated signaling has been found to play a crucial role in leukemia (Deynoux et al., 2016; Szymczak et al., 2018; Wang et al., 2011b). In addition, activated KRAS has been shown to induce HIF1A and ARNT target gene expression in human colon cancer cells (Chun et al., 2010). Echinomycin treatment has demonstrated therapeutic effect on acute myeloid leukemia cells with TP53 mutation in xenograft mouse models (Wang et al., 2020), but our study has shown, for the first time, that echinomycin is effective for oncogenic RAS in Drosophila, mouse and human cell systems. Our data also indicated significantly increased expression of sima and tgo in human KRASG12V hemocytes (Fig. 5F), suggesting that uncontrolled proliferation of hemocytes within the larval hemocoel (body cavity), in our in vivo screen, could produce a hypoxic environment in which HIF gene function and HIF pathway activation are essential for cancer cell survival. In this scenario, HIF gene function would sustain progression and maintenance of leukemia in the fly, ultimately leading to death during the pupal stage. Consistent with this interpretation, HIF1A or ARNT gene silencing in hemocytes completely rescued adult fly development (Fig. 5E). To our knowledge, this is the first time HIF1 gene(s) have been identified in a KRAS genetic screen and illustrates the value of an in vivo Drosophila model expressing a human oncogene for genetic screening purposes.

Furthermore, our fly leukemia model can be directly applied in an in vivo drug screen (Aritakula and Ramasamy, 2008; Dar et al., 2012; Gladstone and Su, 2011; Markstein et al., 2014; Willoughby et al., 2013). Testing the effect of multiple drugs at multiple concentrations in fly larvae revealed echinomycin as a strong inhibitor of KRASG12V-induced hemocyte overproliferation. Interestingly, echinomycin is an inhibitor of the HIF1 pathway (Kong et al., 2005; Wang et al., 2011a, 2014). It prevents HIF1 binding to DNA target sites and thus disrupts HIF1 activation of target genes in response to hypoxic stress. Remarkably, echinomycin treatment was as effective as silencing either of the sima or tgo HIF pathway genes (Fig. 6). Moreover, echinomycin showed no toxicity to the fly at a concentration that completely reversed the leukemia phenotype (Fig. S3). Thus, these results have demonstrated another unique advantage of in vivo drug testing in Drosophila to identify potentially beneficial anticancer compounds of low organismal toxicity. Furthermore, we showed that echinomycin was particularly effective in multiple oncogenic RAS human leukemia lines compared to non-RAS cell lines and demonstrated its effect in a mammalian in vivo model for oncogenic RAS (xenograft mouse model using human THP1 cells with oncogenic NRASG12D; Fig. 7). In addition, our recent data showed that echinomycin also effectively inhibited HIF1A oncoprotein and regressed lung tumor cell growth based on multiple RAS cell lines, including NCI-H727 (KRASG12V), NCI-H1944 (KRASG13D) and Calu-1 (KRASG12C) (Huang et al., 2021). Taken together, these findings demonstrate the potential of echinomycin in treating RAS-induced cancer phenotypes. These findings are in line with previous studies that showed that echinomycin suppressed the growth of multiple AML cell lines and T-lymphoblastic leukemia cell lines (Yonekura et al., 2013) and effectively treated a mouse Mll-Flt3 AML model (Wang et al., 2011b, 2014). However, our study provides the first in vivo evidence that echinomycin treatment can attenuate oncogenic KRAS-induced leukemia in both Drosophila and mouse xenograft models.

Functionally, the KRASG12V-expressing hemocytes showed excessive proliferation, which ultimately resulted in complete mortality at the pupal stage (i.e. no adult flies emerged). Furthermore, they displayed impaired innate immune capabilities as evidenced by significantly reduced phagocytic activity (which could be partially rescued by HIF1 silencing; Fig. 5C,D), increased sensitivity to bacterial infection and reduced immune-related gene expression (Fig. 3D). These observations stand in contrast with previous reports that found that Drosophila strains carrying activated fly Ras85DG12V in hemocytes displayed no changes in phagocytosis activity compared to wild-type flies (Arefin et al., 2017). This possibly reflects the differential effects of human KRASG12V versus fly Ras85DG12V when expressed in hemocytes. Furthermore, a closer look at the expanded number of hemocytes caused by KRASG12V-induced overproliferation revealed a substantially altered composition of circulating hemocytes. Composition was notable for an expanded population of hemocytes that expressed Wg, an early hemocyte lineage surface marker (Fig. 2A), and a reduced P1+ plasmatocyte population, which is typically the most abundant and shows phagocytic activity (Fig. 2C). Interestingly, increased HIF1A has been shown to mediate cancer stem cell maintenance in leukemia, a highly proliferative cell type (Wang et al., 2011b, 2014). KRASG12V expression also drove expansion of the normally very small population of Lz+ crystal cells (Fig. 2B), which contribute to melanization reactions linked to innate immunity (Stofanko et al., 2010; Williams, 2007). It is possible that a subgroup of the Wg+ early hemocyte population also expresses the more mature crystal cell and/or plasmatocyte markers, as it has been previously reported that oncogenic RAS can lead to reprogramming and dedifferentiation (Ischenko et al., 2013). Owing to antibody limitations (all used here are mouse monoclonal), we cannot rule out this possibility at this time. In a recent study of single-cell profiling of Drosophila blood, we identified new hemocyte cell types (Fu et al., 2020). We are currently using this information to investigate how KRASG12V expression changes the composition of the hemocyte population.

Future mechanistic studies to elucidate the pathways underlying the differential susceptibility of oncogenic RAS leukemic cells to HIF inhibition by echinomycin will be of great interest. These studies could identify additional drug targets to inhibit HIF pathway signaling. Testing the efficacy of these and additional drug candidates, such as those shown to inhibit oncogenic KRAS and HIF pathways in a colorectal cancer screen (Bousquet et al., 2016), in treating oncogenic RAS in model systems like our fly are urgently needed to identify valuable treatment options in the clinic. In conclusion, the data presented here demonstrate the significant role of hypoxia signaling in oncogenic RAS leukemias and identify echinomycin as a viable candidate for treatment. They also establish the potential of our Drosophila model for use in conducting genetic screens and drug screens for human cancers.