Less stringent mining of the literature using the term DDR

Less stringent mining of the literature using the term ‘DDR1’ in potent web search tools such as Thomson Integrity [83] yields an impressive number of results for both scientific publications and patents (respectively 1219 and 267) in multiple indications (Fig. 4B). Unfortunately, the majority of these 1219 publications only briefly mention DDR1, and most of the 267 published patents report off-target activity on DDR1 (Google Patents search is referred here [84]). Despite this mass of scientific publications, no selective DDR1 inhibitor has entered clinical development (note that mining of clinical trial descriptions can be very misleading as marketed pan-tyrosine kinase inhibitors such as dasatinib and nilotinib commonly cite marginal off-target effects on DDR1). Since the average time for a NME to enter clinical studies is around 10–15 years [85], absence from the clinical arena (note that DDR1 was cloned 25 years ago [20]) indicates the challenge in effectively and safely drugging this receptor. Selectivity for DDR1 over the kinome, particularly its close analog DDR2, appears to be a challenging hurdle, and reported small molecule DDR1 inhibitors [86,87] generally lack selectivity. The challenge in discovering a selective DDR1 inhibitor partly resides in the limited chemical diversity used to derive new molecules. New chemical starting matter is generally recycled from existing PHA-848125 sale collections historically enriched with known kinase inhibitor motifs. However, existing kinase inhibitors are challenging starting points toward the development of selective DDR1 inhibitors, likely because ATP binding pocket homology is highly conserved across all RTKs [88]. For example, past DDR1 inhibitors have been derived from 1.) existing tyrosine kinase inhibitors such as imatinib, nilotinib and dasatinib (see structures 1.1–1.3 in Fig. 4C) [64,87,89,90] 2.) repurposed focused compound collections originally constructed to target other kinases (see 1.4 in Fig. 4C) [87] and, 3.) commercially available compound collections necessarily enriched with known ATP binding motifs (1.5 in Fig. 4C) [91]. Our recent discovery (M.P., H.R., A.S.) of a highly selective and in vivo active DDR1 small-molecule inhibitor provides evidence that DDR1 is a druggable pharmaceutical target, and some details of our efforts are provided below. To avoid the repurposing of known kinase inhibitor structural motifs, we explored the chemical space offered by DNA encoded libraries (DELs). DELs employ split-and-pool combinatorial chemistry to generate large compound collections, and the technology is increasingly used by the pharmaceutical industry for the discovery of novel chemical starting points [92,93]. The effective numerical size of DEL compound collections is likely larger than that of traditional pharmaceutical-industry compound collections, and the chemical space contained within is known to be different from published chemical space [94]. Lastly, DELs are generally constructed in a target agnostic manner, and therefore unlikely to be heavily enriched with promiscuous ATP binding pocket moieties. We screened both DDR1 and DDR2, and prioritized those hit clusters that selectively enriched against DDR2; this provided us with an initial chemical starting point that was inherently selective for DDR1 over the rest of the kinome (see 2a in Fig. 4C).