Chemical genetic discovery of targets and anti-targets for cancer polypharmacology

Abstract

The complexity of cancer has led to recent interest in polypharmacological approaches for developing kinase-inhibitor drugs; however, optimal kinase-inhibition profiles remain difficult to predict. Using a Ret-kinase-driven Drosophila model of multiple endocrine neoplasia type 2 and kinome-wide drug profiling, here we identify that AD57 rescues oncogenic Ret-induced lethality, whereas related Ret inhibitors imparted reduced efficacy and enhanced toxicity. Drosophila genetics and compound profiling defined three pathways accounting for the mechanistic basis of efficacy and dose-limiting toxicity. Inhibition of Ret plus Raf, Src and S6K was required for optimal animal survival, whereas inhibition of the 'anti-target' Tor led to toxicity owing to release of negative feedback. Rational synthetic tailoring to eliminate Tor binding afforded AD80 and AD81, compounds featuring balanced pathway inhibition, improved efficacy and low toxicity in Drosophila and mammalian multiple endocrine neoplasia type 2 models. Combining kinase-focused chemistry, kinome-wide profiling and Drosophila genetics provides a powerful systems pharmacology approach towards developing compounds with a maximal therapeutic index.

Figures

Figure 1. Screening for an optimal therapeutic index in a Drosophila MEN2B model yields a polypharmacological kinase inhibitor

a, Suppression of dRetMEN2B-induced developmental block and whole-animal toxicity were scored based on the number of embryos (n) that survived as pupae (x) and adults (y). b, Per cent viability of control- or drug-treated flies determined for pupae (x per n) and adults (y per n). AD57 emerged as the best single-agent hit from the screen. Asterisks indicate significance comparing to control using Student’s t-test (P < 0.05 for adults in AD57 and sorafenib treatments, and P < 0.05 for pupae for the rest). Error bars denote s.e.m. Total n of 200, 75, 98, 54, 91, 280 and 209, from left to right. Soraf., sorafenib; Sunit., sunitinib; Vande., vandetanib. c, ptc > dRetMEN2B adults have notum defects including excessive bristles (asterisks) and scutellum defects (brackets); controls (+ dimethylsulphoxide (DMSO)) died as un-eclosed adults. AD57 strongly suppressed whereas sorafenib (SF) weakly suppressed these defects, yielding fully eclosed adults. Width of each wild-type notum is ~0.75 mm. WT, wild type. d, Structure–activity relationships suggest that dRet inhibition alone is insufficient to rescue MEN2B flies. IC50 values were determined against a purified form of human Ret. e, The AD series of compounds showed broad-spectrum kinase-inhibition profiles. Clinical (asterisks) and known kinase inhibitors are shown for comparison. The number of lipid (PI), tyrosine (Y) and serine/threonine (S/T) kinases tested are shown in the pie chart.

Figure 2. Multiple-pathway inhibition by AD57 mitigates dRet-directed phenotypes

a, z-series confocal images of larval wing epithelia; virtual cross-section through tissue with apical up. Control tissue shows apical phospho-Src (pSrc) expression (red) in the junctions. ptc > dRetMEN2B wing cells (green fluorescent protein; GFP+) shifted basally (arrows) and invaded below adjacent wild-type tissue; phospho-Src emerged at the basal invading front (asterisks). These phenotypes were strongly suppressed by AD57 but not by AD36, AD58 or vandetanib (VD). Apical–basal distance is ~45 µm; imaged with ×63 oil. b, Partial list of signalling pathways activated by oncogenic dRetMEN2B. MMPs, matrix metalloproteinases. c, Per cent in vitro kinase inhibition profiles (left) and relative rescue (right) are shown. Tree indicates similarity of compounds on the basis of hierarchical clustering of per cent kinase inhibition. d, Broad dRetMEN2B expression led to ectopic wing veins (arrows), reflective of hyperactive Ras pathway signalling. The wing defects were suppressed by AD57 and enhanced by AD58. Removal of one functional copy of erk/rolled (erk−/+) enhanced rescue by AD57 and AD58. Quantified in Supplementary Figure 5c.

Figure 3. Feedback downregulation of the Ras pathway through the anti-target Tor

a, Reducing dTor gene dosage decreased per cent viability of AD57- nd AD58-treated dRetMEN2B flies (P < 0.05 comparing pupae (AD57, AD58) or adults (AD57)). Conversely, reducing erk gene dosage enhanced survival of both (P < 0.05 comparing pupae (AD58) or adults (AD57)). Treatment with a specific MEK inhibitor alone, AZD6244, in control (ptc > dRetMEN2B) or erk−/+ flies did not rescue viability compared to AD57-treated flies, suggesting its level of Ras pathway suppression is close to optimal (P>0.05 comparing AZD6244 pupae). Reducing S6K (S6k−/+) partially mitigated toxicity from AD58 treatment. Column bars represent the mean of three separate experiments. Total n of 214, 58, 63, 130, 254, 66, 118, 52, 59, 190, 93, 114 and 96 from left to right. Error bars denote s.e.m. b, Decreased viability of wild-type flies by AD58 was mitigated by co-administration of sorafenib or AZD6244. Total n of 118, 47, 51 and 28 from left to right. c, Reducing dTor strongly enhanced AD58-mediated invasion (asterisks, arrow) and excess proliferation. Top panel, a lateral reconstruction; compare with Fig. 2a. Bottom panels represent an apical view, constructed as a z-series overlay of confocal images spanning the full depth of the wing disc epithelia. It shows how cells migrate from the ptc domain to distant sites (for example, arrow, asterisks), a phenotype strongly enhanced in the presence of dTor−/+ plus AD58. Top panel (GFP+) width is ~150 µm; imaged with ×63. Bottom right panel width (GFP+) is ~150 µm; both bottom panels were imaged with ×40. d, Wing defects in ptc > dRetMEN2B dTor−/+ adults were further enhanced by AD58. e, Quantification of ptc > dRetMEN2B phenotypes. Invasion was established by scoring for single or groups of GFP-labelled cells that relocated away from the ptc boundary (Fig. 3c, asterisks). Basal migration was scored as indentation of the apical surface (see Fig. 2a, arrows). Proliferation was scored as significant widening of the ptc boundary. The number of wings analysed under each condition is indicated in brackets. Reduced dTor increased proliferation in the presence of AD58 as well as reducing survival; all aspects were improved by feeding AD57 whereas increased invasion by feeding AD36 did not translate to reduced survival. f, Migration of dRetMEN2B-transformed cells was blocked by co-treatment with AD58 plus sorafenib (bottom). Treatment with similar doses of AD58 (top) or sorafenib alone (not shown) did not suppress migration. Arrow indicates constriction of apical cell surface and asterisk indicates basal invading front. Apical–basal distance is ~45 µm; imaged with ×63.

Figure 4. Balanced kinase polypharmacology provides optimal efficacy and toxicity

a, Chemical structures of the AD57 derivatives AD80 and AD81 and percentage inhibition of relevant targets at 1 µM. Unlike AD57 and AD58, both lack significant inhibitory activity against mTOR. b, AD80 and AD81 showed improved rescue relative to AD57. *P < 0.05, significance compared to AD57 in a two-tailed Student’s t-test. Total n of 214, 58, 109 and 99 from left to right. Error bars denote s.e.m. c, Basal migration (arrow) of dRetMEN2 cells and basal phospho-Src (pSrc; asterisk) were blocked by AD80. Apical–basal distance is ~45 µm; imaged with ×63. d, Reducing erk gene dosage enhanced survival of AD57 (P < 0.05 for adult flies compared across genotypes) but not AD80 (P>0.5 for adults compared across genotypes), suggesting that the Tor feedback loop was not altered by AD80 and that Erk was optimally suppressed in flies. Total n of 214, 43 and 109 from left to right. Error bars denote s.e.m. e, 765 > dRetMEN2B-dependent extra wing vein phenotype was fully rescued by AD80. f, AD80 and vandetanib (VD) reduced tumour progression 3.1- and 1.9- fold, respectively, relative to vehicle-treated nude mice transplanted with TT cells. Change in tumour volume was calculated per mouse and shown are the median per group. Twenty vehicle- and ten drug-treated mice were analysed for each treatment group. g, Corresponding body-weight measurements.

Figure 5. Differential polypharmacology and outcomes from the AD compounds

Models to explain the AD series of compounds in dRetMEN2B transgenic flies. Pathway components blocked by inhibitors have been boxed, with resulting flux indicated by orange lines and arrows, and the incoherent feed-forward (i.f.f.) loop highlighted in blue. Grey dashed lines indicate loss of dTor inhibition. Targets in black boxes contribute to efficacy whereas inhibition of the anti-target dTor (red) leads to hyperactivation of the Ras pathway, causing high toxicity in the MEN2 model. The polypharmacological profile of AD80 best addresses the three key pathways, providing high drug efficacy and optimal therapeutic index.