Synthetic inhibitor leads of human tropomyosin receptor kinase A (hTrkA)§

Govindan Subramanian*, Rajendran Vairagoundar, Scott J. Bowen, Nicole Roush, Theresa Zachary, Christopher Javens, Tracey Williams, Ann Janssen and Andrea Gonzales
Veterinary Medicine Research & Development, Zoetis, 333 Portage Street, Kalamazoo, MI 49007, USA. E-mail: govindan.subramanian@zoetis.com

Received 25th November 2019 , Accepted 22nd December 2019

First published on 10th January 2020


In silico virtual screening followed by in vitro biochemical, biophysical, and cellular screening resulted in the identification of distinctly different hTrkA kinase domain inhibitor scaffolds. X-ray structural analysis of representative inhibitors bound to hTrkA kinase domain defined the binding mode and rationalized the mechanism of action. Preliminary assessment of the sub-type selectivity against the closest hTrkB isoform, and early ADME guided the progression of select inhibitor leads in the screening cascade. The possibility of the actives sustaining to known hTrkA resistance mutations assessed in silico offers initial guidance into the required multiparametric lead optimization to arrive at a clinical candidate.


Human tropomyosin receptor kinase (Trk) is a phylogenetic member of the wider protein kinase superfamily and specifically belongs to the receptor tyrosine kinase subfamily of the human kinome.1 Three well differentiated Trk isoforms (TrkA, TrkB, TrkC)2 have been characterized to date with the hTrkA subtype implicated in diseases like chronic pain3 and drugged for oncology indications.4 Despite the successful launch of hTrkA kinase inhibitors like Larotrectinib (Vitrakvi®)5 and Entrectinib (Rozlytrek™)6 recently for use in various cancers, the long-term benefits of these drugs are still being investigated due to the reported acquired drug resistance mutations in the hTrkA kinase domain.7,8

Most kinase drugs9 including the launched hTrkA kinase inhibitors and additionally those in clinical trials10 are the classic type 1 inhibitors that bind the catalytically competent active state of the enzyme and are competitive to ATP (co-factor) for the orthosteric kinase ligand binding site.11 Such inhibitors are also known to exhibit kinase promiscuity12 to a varying extent due to ATP binding site residue conservation across the human kinome. Given some of these drawbacks, a paradigm shift is beginning to emerge with the identification of hTrkA inhibitors that are either type 2 (binding to the catalytically incompetent kinase inactive state conformation) or allosteric (binding away from the catalytic site) mechanistically.13

Identification of new or alternate chemical leads for a protein target used to rely on high throughput screening14 of a corporate collection or commercially available compound collection. Modern small molecule discovery research, however, allows the judicious use of known ligands15 and/or protein structure (either in the apo form or a ligand bound complex)16 to apply in silico virtual screening techniques17,18 to prioritize a small subset of the compound collection that has the highest probability of becoming an active synthetic lead. The lead can then be pursued further as part of the medicinal chemistry optimization strategy to arrive at a clinical candidate.

Mining the scientific literature for novel type 2[thin space (1/6-em)]13,19 or allosteric13,20 hTrkA kinase inhibitors resulted only in a handful of compounds with no detailed structure–activity relationship to build a proprietary medicinal chemistry strategy. Consequently, virtual screening was pursued using the purchasable compound collection from the ZINC database21 with the goal of identifying novel inhibitor leads. Initially, the structure-based GLIDE docking and the ligand-based PHASE pharmacophore approach implemented in the MAESTRO molecular modelling application suite (Schrödinger, New York, USA) was used for the new hit finding strategy to identify inhibitors that preferentially bind the inactive state kinase conformation (see ESI Met. 1S). Additionally, a ligand-based ROCS approach22 was undertaken in the pursuit of allosteric inhibitor23 identification (see ESI Met. 2S). These in silico exercises resulted in the identification of multiple chemical series that exhibit type 2 binding mechanism.24,25 In this communication, singleton actives that possess unique kinase structural and binding features are highlighted.

The novel leads 1–3 (Fig. 1) identified from the in silico exercise belong to the substituted heteroarylacetamide scaffold displaying hTrkA kinase domain only IC50's of ∼39 nM, ∼47 nM, and ∼102 nM respectively when screened in a biochemical HTRF based enzymatic assay. The observed nM range activities and diversity of the leads resulting directly from the docking and pharmacophore search exercise clearly attests to the strengths of the in silico virtual screening approach in identifying new chemical matter. It should also be emphasized that the in silico docking screen was pursued using the inactive state kinase conformation (PDB: 3V5Q) and no constraints were imposed to capture the donor–acceptor–donor interaction of the hinge residues with the ligand typically seen in most kinase–ligand complexes.26 Despite this, the prioritized top ranking in silico hits was able to retrieve the actives 1 and 3 with a potential to form the hTrkA kinase domain hinge interaction (Table 1).


image file: c9md00554d-f1.tif
Fig. 1 Structures of experimentally characterized hTrkA kinase domain actives starting from in silico GLIDE docking (1, 3, 6), PHASE pharmacophore (2), and ROCS shape (4) searches.
Table 1 Experimental results for inhibitor leads 1–6
Assay parameter (units) 1 2 3 4 6
a hTrkA kinase domain screened using CisBio HTRF® KinEASE biochemical assay.b hTrkA kinase screened using fluorescent imaging plate reader (FLIPR) cell based assay.c hTrkB kinase screened using FLIPR™ cell based assay.d Neurite outgrowth (NO) assay using rat PC12 cells.e Half-life of compounds determined using dog liver microsomes (DLM).f Half-life of compounds determined using cat liver microsomes (CLM).g MDCK permeability measured from apical to basolateral chamber.h MDCK permeability measured from basolateral to apical chamber.i Measured kinetic solubility.
hTrkA_enz_IC50a (μM) 0.0393 0.0471 0.102 0.005 0.178
hTrkA_cel_IC50b (μM) 3.5 >20 7.5 0.17 >20
hTrkB_cel_IC50c (μM) 7.7 >20 >20 15.9 >20
rPC12_NO_IC50d (μM) >10 >9 >10 0.155  
DLM_T1/2e (min) 120 120 120 81 75
CLM_T1/2f (min) 40 120 120 16 56
Perm_ABg (×10−6 cm s−1) 7.7   4.8 4.1  
Perm_BAh (×10−6 cm s−1) 17.3   6.4 7.6  
Solubilityi (μM) 36.3 15.9 3.5 12.7 1.3


Although the initial biochemical screened actives were promising, the enzyme assay was run under conditions more favorable for active state inhibition. In order to clarify if the identified actives were genuine hTrkA kinase inhibitors that would bind the inactive state kinase more favourably, an additional orthogonal assay was performed on the most active lead, 1, to characterize the binding mechanism. The assay was run such that the kinase exists in either a predominantly inactive or active state. The screen was performed using the Caliper assay technology that measures the kinase activity by monitoring the phosphorylation of the fluorescently-labelled peptide substrate. A complete dose–response clearly revealed 1 to favour the kinase inactive state (IC50 ∼102 nM) by ∼9-fold over the active state (IC50 ∼0.95 μM). In order to confirm unequivocally that the lead 1 is a kinase conformation selective compound, the Caliper dose response was subsequently followed by the binding kinetics. The Ki(app) of 65.84 nM towards the hTrkA inactive state and Ki > 1 μM for the active state follow the same trend and confirm the lead to be a true kinase inhibitor with a preferential binding towards the inactive state.

To confirm the potential binding mechanism of lead 1, an X-ray structure of the bound ligand-hTrkA kinase domain was successfully pursued. The complex structure solved at 2.81 Å final resolution reveals 1 binding at the orthosteric ATP binding site of hTrkA kinase (Fig. 2).


image file: c9md00554d-f2.tif
Fig. 2 Schematic ligand interaction derived from hTrkA-1 X-ray complex. PDB accession code is 6PMB.

More specifically, the ligand carbonyl forms an H-bond with the backbone amide NH functionality of D668 at the start of the kinase activation loop. Similarly, the amide NH group of 1 forms an H-bond with the sidechain carboxylate of E560 from the kinase αC-helix (in conformation). Additionally, the pyridine nitrogen in lead 1 engages in H-bond interaction with the backbone NH of kinase hinge residue, M592 (Fig. 2). Apart from these directed interactions, a network of hydrophobic residues lining the binding pocket stabilize the bound complex and justify its affinity observed in the in vitro assays. Finally, the location of the activation loop along with the disposition of the 668DFG690 residues (DFG-out) characterize the ligand bound hTrkA kinase to be reflective of a classical type 2 binding27 inhibition mechanism.

The effective overlay of the predicted docking pose of 1 to that observed in the complex X-ray structure (see ESI Met. 1S) provided additional support for the in silico results suggesting binding interactions in the orthosteric site. As a result, an X-ray structure was not obtained for 2 since the anticipated interactions are expected to be like 1. An additional possibility of the tetrazole nitrogens in 2 participating in a H-bond with the K544 side chain amine is also contemplated based on the reported X-ray structure (PDB: 4PMM) of a similar ligand in literature19 and used as a query in the pharmacophore search (see ESI Met. 1S).

Despite the confidence in the docking mode of 3 to hTrkA kinase domain, the weaker affinity of ligand in the biochemical HTRF assay and the unusual structure of the ligand with no literature precedence for kinase activity structural biology validation was pursued. As a result, the X-ray structure of 3 bound to hTrkA kinase domain was solved and refined to a final 2.53 Å resolution. The electron density of the ligand bound hTrkA complex structure reveals an unambiguous binding mode and orientation of 3 in the orthosteric ATP binding site (Fig. 3).


image file: c9md00554d-f3.tif
Fig. 3 Ligand interactions schematic derived from hTrkA-3 X-ray complex. PDB accession code is 6PMA.

As schematized in Fig. 3, various functional groups in inhibitor 3 make direct H-bond interaction with D668, E560, K544, and M592 residues of hTrkA kinase. Additionally, a solvent water mediated H-bond between the ligand and the kinase domain residues (Fig. 3) is vividly captured in the X-ray structure. Apart from the binding site interactions, evidence for a type 2 binding mechanism can be inferred through the X-ray structural analysis of the key kinase residues and 2° structure orientations (αC-helix in) and dispositions (DFG-out) as described before.27 As anticipated, a striking overlay between the docked pose and the bound ligand in the X-ray, proved once again the strengths of the docking exercise (see ESI Met. 1S).

As part of the in silico screening effort, a shape-based ROCS approach was pursued to identify allosteric kinase inhibitors based on a reported allosteric inhibitor (see ESI Met. 2S).28 The outcome of this exercise resulted in 4 (Fig. 1) as a potential hTrkA kinase inhibitor with an IC50 of ∼5.2 nM in the HTRF biochemical assay. The inhibitor novelty and the surprising sub-nanomolar affinity prompted further investigation to ascertain the inhibition mechanism. As discussed for 1, 4 was also screened using the Caliper technology. The IC50 ∼21 nM and ∼45 nM for the inactive and active hTrkA kinase states respectively suggested only a modest ∼2-fold difference complicating the interpretation as the results could be within standard experimental errors.29 Further experiments leading to the binding Ki determination also resulted in an affinity of ∼1 nM (Ki(app)) and 0.4 nM (Ki) making it difficult to determine the binding mechanism. The in silico screening origin, the non-kinase like structural features of the ligand and the interpretations from the biochemical experiments suggested that 4 could potentially exhibit allosteric binding and so the complex X-ray structure was sought. Unfortunately, preliminary attempts to solve the hTrkA-4 bound complex was futile. As a result, close structural analogs available commercially were sought resulting in the identification of 5 (Fig. 1).

A quick Caliper screen of 5 revealed the inhibitor to be active with a comparable IC50 of ∼94 nM and ∼76 nM respectively for the hTrkA kinase inactive and active state conditions suggesting the structural changes may not have impacted the binding mechanism drastically. Although 5 was slightly less potent than 4, crystallographic studies resulted in the hTrkA-5 kinase complex structure resolved to a final 3.0 Å resolution. Surprisingly, the X-ray structural analysis of the complex does not characterize the ligand (5) to bind allosterically as it occupies the orthosteric ATP binding site.

Unlike the other X-ray structures reported here, three hTrkA-5 monomers with basically the same overall conformation comprised the crystallographic asymmetric unit. The electron density map revealed an unambiguous binding mode and orientation of the inhibitor 5 bound to hTrkA kinase domain. The ligand phenoxy moiety conformation seemed to be a bit more flexible as indicated by the higher B-factor and weaker electron density in this region. Overall, only one specific H-bond between the ligand benzamide carbonyl group and the backbone NH of the M592 kinase residue lock the kinase–ligand directed interaction (Fig. 4). Apart from this, the ligand is shielded by a network of hydrophobic residues. Of the two morpholino groups, the 4-substituted one projects toward the solvent exposed region. Finally, the phenoxy functionality occupies a region conventionally occupied by the phenylalanine in the kinase DFG-out conformation.


image file: c9md00554d-f4.tif
Fig. 4 Ligand interactions schematic derived from hTrkA-5 X-ray complex. PDB accession code is 6PME.

The macro secondary structural features such as the location and orientation of the Gly-rich loop, αC-helix, and the activation loop are suggestive of a classical type-2 kinase conformation. However, the minor 2° structural fold perturbation at the activation loop start, positions F669 sidechain occupying the same region as that of D668 sidechain in 1 reflective of a DFG-in like fingerprint. Notably, 5 does not access the back pocket that is characteristic of traditional type 2 inhibitors, but only occupies the region typically observed for type 1 active state inhibitors. However, additional comparison of hTrkA-5 complex with the hTrkA-AZ-23 active state kinase conformation (PDB: 4AOJ)30 suggests that the 668DFG670 residues are dispositioned differently in both instances, potentially undermining the possibility of classifying the current complex as arising from active state inhibition. These observations are consistent with the Caliper results and support the lack of discrimination between the type 1 and 2 binding mechanism for 5 under the assay conditions that provided discriminatory evidence for 1. The Caliper assay conditions were also able to clearly discriminate hTrkA active state kinase inhibitors like AZ-23 as reported previously23 and hence unlikely to contribute to assay limitations.

Overall, the structural and screening analyses thus far provide supporting evidence for 5 to bind both the active and inactive kinase states effectively. Assigning a unique binding classification for 5 is difficult but suggests the kinase conformation to be representative of a potential transition state between the two DFG conformational states.27 This represents yet another novel binding mechanism for hTrkA that takes advantage of active state inhibition while retaining features of the inactive state that provide opportunities to garner mechanistic selectivity (e.g. 4 in Table 1).

Another intriguing aspect of 5 is the ligand sub-structural features that closely mimic the key ligand motifs of the launched type-1 inhibitor, Entrectinib (Fig. 5).31 Despite the overall ligand structural difference, the placements of the key ligand motifs (benzyl vs. phenoxy; pyran vs. 2-morpholino; and piperazine vs. 4-morpholino) in the binding site and very similar. However, the superposition of the two ligand bound X-ray complexes clearly demonstrate that the 2° structure dispositions of the Gly-rich loop and the fully resolved A-loop are placed differently (see ESI Fig. 1S).


image file: c9md00554d-f5.tif
Fig. 5 Ligand interactions schematic derived from hTrkA-Entrectinib X-ray complex. Literature PDB accession code is 5KVT.

Although the type 2 inhibitor leads identified via the docking exercise in this study do interact with the kinase hinge residues interaction, the initial in silico effort was intended to find active leads that would allow for structural modifications to extend out to the kinase hinge residues subsequently. Indeed, the diaryl urea scaffold exemplified by 6 (Fig. 1) serve as a potential representative. The docking pose of 6 in hTrkA kinase domain and its experimental HTRF determined IC50 ∼178 nM clearly supports the avenues for affinity improvements. To validate if 6 is a good starting point for kinase conformation specific inhibition, the inhibitor was subjected further to the Caliper screen. Surprisingly, 6 was inactive with IC50 >10 μM under both the active and inactive conditions tested.

Puzzled at the contradiction between the two assays, additional clarity was sought through structural biology. The hTrkA-6 kinase domain complex was solved at a final resolution of 2.2 Å. The complex structure revealed an unambiguous binding mode of 6 occupying the orthosteric ATP binding site and accessing the back pocket as well. The overall structural inference suggests 6 to be a classical type 2 inhibitor that bind the inactive kinase in an αC-helix in and DFG-out conformation.27

As shown in Fig. 6, the major stabilizing factor for the complex is the urea NH functionality interacting with the sidechain carboxylate of E560 from the αC-helix, and the urea carbonyl H-bonding to the backbone NH of D668 at the start of the activation loop. The benzamide moiety in 6 occupies the ATP base sub-pocket and is positioned appropriately to take advantage of the kinase hinge residue interaction through subtle functional group modifications of the benzamide aryl group for improved kinase activity.


image file: c9md00554d-f6.tif
Fig. 6 Schematic of the ligand interactions derived from hTrkA-6 kinase domain X-ray complex. PDB accession code is 6PMC.

Finally, the susceptibility of the newly identified inhibitors (Fig. 1) to known acquired resistance mutations8 in the ligand binding site were evaluated using the resistance scan option in MOE. Although this exercise provides qualitative guidance on the susceptibility potential (Table 2; also see ESI Table 2S), more rigorous methods32 and evaluations are required to fully quantitate the mutational effects. However, the preliminary in silico modelling interpretations suggest that mutating the gate-keeper residue F589 seem to affect the inhibitor binding to varying extent. None of the other mutations in the binding site are expected to alter the inhibitor affinity drastically that might impede the inhibitor potential. More strikingly, lead 5 demonstrates a much better mutational profile among the chemotypes studied here as well as that reported for Entrectinib24 and paves the way for a differential and promising starting point to initiate medicinal chemistry optimization efforts.

Table 2 Δaffinity (kcal mol−1) between wildtype and mutant hTrkA⋯inh kinase domain complex
Mutation 1 3 5 6
F589L 0.438 0.287 0.086 0.176
G595R −0.185 −0.357 −0.118 −0.076
G667C −0.032 0.466 −0.067 0.271
G667S 0.024 0.266 −0.045 0.166


The enzyme assays and the structural biology efforts thus far investigated only the kinase domain of the hTrkA. To further characterize the performance of the identified inhibitors (Fig. 1) in a full-length hTrkA kinase, cellular dose–response was pursued using FLIPR® (Table 1). While most inhibitors seem to exhibit weaker activity, 4 demonstrated reasonably good potency with a good selectivity window against the closest hTrkB isoform. Further evaluation of 4 in the neurite outgrowth assay under a more physiologically relevant rat neuronal PC12 cells (American Type Culture Collection [ATCC CRL-1721], Rockville, MD) also revealed an IC50 ∼155 nM underscoring the potential of an unconventional kinase lead with a plausible heterogeneous binding mechanism that has not yet been reported for hTrkA in the scientific literature.

Conclusions

A multitude of in silico virtual screening approaches used to identify new hTrkA kinase domain inhibitors resulted in novel and diverse leads, 1–6. Most of the sub-structural motifs in the chemical leads have not been reported in the kinase literature to date and presents new scaffolds for exploration across other kinome targets as well. Systematic evaluation of the binding mechanism achieved through elaborate in vitro enzymatic assays of the kinase domain, cellular screens using the full length hTrkA and hTrkB (for addressing isoform selectivity) in recombinant cells, rat neuronal PC12 cells that closely mimics the physiologically relevant native environment outlines the needed biology before initiating and investing in a resource intensive medicinal chemistry effort. Validation of the kinase conformational state dependent binding mode of the leads interpreted through X-ray crystallography lend strong support for pursuing the inhibitors in a lead optimization phase within preclinical discovery research with the added optimal ADME property advantages (Table 1). The unique and novel leads 4 and 5 with a mixed binding mechanism profile offers avenues to evaluate the resistance effects of TrkA inhibition33 and to mitigate the current acquired resistance apprehensions of the drugged inhibitors of hTrkA kinase domain.

Abbreviations

Secondary
ADMEAbsorption, distribution, metabolism, excretion
ATPAdenosine triphosphate
FLIPRFluorescent imaging plate reader
H-BondHydrogen bond
HTRFHomogeneous time resolved fluorescence
MDCKMadin–Darby canine kidney cells
MOEMolecular operating environment
PC12Pheochromocytoma
PDBProtein Data Bank
ROCSRapid overlay of chemical structures

Conflicts of interest

All authors are/were employees of Zoetis.

Acknowledgements

The authors thank the Veterinary Medicine Research & Development leadership, Zoetis, for funding this work. GS thanks G. Bainbridge, S. Kamerling, C. Haines, M. Cox, for additional support throughout the research work. GS also thanks A. Wittwer and J. Hirsch from Confluence Discovery Technologies for the Caliper assays and Stefan Steinbacher from Proteros Biostructures GmbH, Germany for the structural biology work reported here.

Notes and references

  1. G. Manning, D. B. Whyte, R. Martinez, T. Hunter and S. Sudarsanam, The Protein Kinome Complement of the Human Genome, Science, 2002, 298, 1912–1934 CrossRef CAS PubMed.
  2. S. D. Skaper, The neurotrophin family of neurotrophic factors: An overview, Methods Mol. Biol., 2012, 846, 1–12 CrossRef CAS PubMed.
  3. F. F. Hefti, A. Rosenthal, P. A. Walicke, S. Wyatt, G. Vergara, D. L. Shelton and A. M. Davies, Novel class of pain drugs based on antagonism of NGF, Trends Pharmacol. Sci., 2006, 27, 85–91 CrossRef CAS PubMed.
  4. A. M. Lange and H.-W. Lo, Inhibiting TRK proteins in clinical cancer therapy, Cancers, 2018, 10, E105 CrossRef PubMed.
  5. L. J. Scott, Larotrectinib: First global approval, Drugs, 2019, 79, 201–206 CrossRef CAS PubMed.
  6. D. Liu, M. Offin, S. Harnicar, B. T. Li and A. Drilon, Entrectinib: An orally available, selective tyrosine kinase inhibitor for the treatment of NTRK, ROS1, and ALK fusion-positive solid tumors, Ther. Clin. Risk Manage., 2018, 14, 1247–1252 CrossRef CAS PubMed.
  7. A. Drilon, R. Nagasubramanian, J. F. Blake, N. Ku, B. B. Tuch, K. Ebata, S. Smith, V. Lauriault, G. R. Kolakowski, B. J. Brandhuber, P. D. Larsen, K. S. Bouhana, S. L. Winski, R. Hamor, W.-I. Wu, A. Parker, T. H. Morales, F. X. Sullivan, W. E. DeWolf, L. A. Wollenberg, P. R. Gordon, D. N. Douglas-Lindsay, M. Scaltriti, R. Benayed, S. Raj, B. Hanusch, A. M. Schram, P. Jonsson, M. F. Berger, J. F. Hechtman, B. S. Taylor, S. Andrews, S. M. Rothenberg and D. M. Hyman, A next-generation TRK kinase inhibitor overcomes acquired resistance to prior TRK kinase inhibition in patients with TRK fusion-positive solid tumors, Cancer Discovery, 2017, 7, 963–972 CrossRef CAS PubMed.
  8. E. Cocco, M. Scaltriti and A. Drilon, NTRK fusion-positive cancers and TRK inhibitor therapy, Nat. Rev. Clin. Oncol., 2018, 15, 731–747 CrossRef CAS PubMed.
  9. R. Roskoski Jr., Properties of FDA-approved small molecule protein kinase inhibitors, Pharmacol. Res., 2019, 144, 19–50 CrossRef PubMed.
  10. S. Klaeger, S. Heinzlmeir, M. Wilhelm, H. Polzer, B. Vick, P. A. Koenig, M. Reinecke, B. Ruprecht, S. Petzoldt, C. Meng, J. Zecha, K. Reiter, H. Qiao, D. Helm, H. Koch, M. Schoof, G. Canevari, E. Casale, S. R. Depaolini, A. Feuchtinger, Z. Wu, T. Schmidt, L. Rueckert, W. Becker, J. Huenges, A. K. Garz, B. O. Gohlke, D. P. Zolg, G. Kayser, T. Vooder, R. Preissner, H. Hahne, N. Tonisson, K. Kramer, K. Gotze, F. Bassermann, J. Schlegl, H. C. Ehrlich, S. Aiche, A. Walch, P. A. Greif, S. Schneider, E. R. Felder, J. Ruland, G. Medard, I. Jeremias, K. Spiekermann and B. Kuster, The target landscape of clinical kinase drugs, Science, 2017, 358, eaan4368 CrossRef PubMed.
  11. Z. Zhao and P. E. Bourne, Overview of current type I/II kinase inhibitors, 2018, arXiv:1811.09718 [q-bio.MN] Search PubMed.
  12. M. W. Karaman, S. Herrgard, D. K. Treiber, P. Gallant, C. E. Atteridge, B. T. Campbell, K. W. Chan, P. Ciceri, M. I. Davis, P. T. Edeen, R. Faraoni, M. Floyd, J. P. Hunt, D. J. Lockhart, Z. V. Milanov, M. J. Morrison, G. Pallares, H. K. Patel, S. Pritchard, L. M. Wodicka and P. P. Zarrinkar, A quantitative analysis of kinase inhibitor selectivity, Nat. Biotechnol., 2008, 26, 127–132 CrossRef CAS PubMed; P. P. Zarrinkar, R. N. Gunawardane, M. D. Cramer, M. F. Gardner, D. Brigham, B. Belli, M. W. Karaman, K. W. Pratz, G. Pallares, Q. Chao, K. G. Sprankle, H. K. Patel, M. Levis, R. C. Armstrong, J. James and S. S. Bhagwat, Blood, 2009, 114, 2984–2992 CrossRef PubMed; M. I. Davis, J. P. Hunt, S. Herrgard, P. Ciceri, L. M. Wodicka, G. Pallares, M. Hocker, D. K. Treiber and P. P. Zarrinkar, Comprehensive analysis of kinase inhibitor selectivity, Nat. Biotechnol., 2011, 29, 1046–1051 CrossRef PubMed.
  13. W. Yan, N. R. Lakkaniga, F. Carlomagno, M. Santoro, N. Q. McDonald, F. Lv, N. Gunaganti, B. Frett and H. Y. Li, Insights into current tropomyosin receptor kinase (TRK) inhibitors: development and clinical application, J. Med. Chem., 2019, 62, 1731–1760 CrossRef CAS PubMed.
  14. R. Macarron, M. N. Banks, D. Bojanic, D. J. Burns, D. A. Cirovic, T. Garyantes, D. V. S. Green, R. P. Hertzberg, W. P. Janzen, J. W. Paslay, U. Schopfer and G. S. Sittampalam, Impact of high-throughput screening in biomedical research, Nat. Rev. Drug Discovery, 2011, 10, 188–195 CrossRef CAS PubMed.
  15. Y. Tanrikulu and G. Schneider, Pseudoreceptor models in drug design: Bridging ligand- and receptor-based virtual screening, Nat. Rev. Drug Discovery, 2008, 7, 667–677 CrossRef CAS PubMed.
  16. E. Lionta, G. Spyrou, D. K. Vassilatis and Z. Cournia, Structure-based virtual screening for drug discovery: Principles, applications and recent advances, Curr. Top. Med. Chem., 2014, 14, 1923–1938 CrossRef CAS PubMed.
  17. S. Ekins, J. Mestres and B. Testa, In silico pharmacology for drug discovery: Methods for virtual ligand screening and profiling, Br. J. Pharmacol., 2007, 152, 9–20 CrossRef CAS PubMed.
  18. S. P. Leelananda and S. Lindert, Computational methods in drug discovery, Beilstein J. Org. Chem., 2016, 12, 2694–2718 CrossRef CAS PubMed.
  19. P. Albaugh, Y. Fan, Y. Mi, F. X. Sun, F. Adrian, N. X. Li, Y. Jia, Y. Sarkisova, A. Kreusch, T. Hood, M. Lu, G. X. Liu, S. L. Huang, Z. S. Liu, J. Loren, T. Tuntland, D. S. Karanewsky, H. M. Seidel and V. Molteni, Discovery of GNF-5837, a selective TRK inhibitor with efficacy in rodent cancer tumor models, ACS Med. Chem. Lett., 2012, 3, 140–145 CrossRef CAS PubMed; S. J. Stachel, J. M. Sanders, D. A. Henze, M. T. Rudd, H. P. Su, Y. Li, K. K. Nanda, M. S. Eqbertson, P. J. Manley, K. L. Jones, E. J. Brnardic, A. Green, J. A. Grobler, B. Hanney, M. Leitl, M. T. Lai, V. Munshi, D. Murphy, K. Rickert, D. Riley, A. Krasowska-Zoladek, C. Daley, P. Zuck, S. A. Kane and M. T. Bilodeau, Maximizing diversity from a kinase screen: Identification of novel and selective pan-Trk inhibitors for chronic pain, J. Med. Chem., 2014, 57, 5800–5816 CrossRef PubMed; S. E. Skerratt, M. Andrews, S. K. Bagal, J. Bilsland, D. Brown, P. J. Bungay, S. Cole, K. R. Gibson, R. Jones, I. Morao, A. Nedderman, K. Omoto, C. Robinson, T. Ryckmans, K. Skinner, P. Stupple and G. Waldron, The discovery of a potent, selective, and peripherally restricted pan-Trk inhibitor (PF-06273340) for the treatment of pain, J. Med. Chem., 2016, 59, 10084–10099 CrossRef PubMed; H. Tanaka, H. Sase, T. Tsukaguchi, M. Hasegawa, H. Tanimura, M. Yoshida, K. Sakata, T. Fujii, Y. Tachibana, K. Takanashi, A. Higashida, K. Hasegawa, Y. Ono, N. Oikawa and T. Mio, Selective TRK inhibitor CH7057288 against TRK fusion-driven cancer, Mol. Cancer Ther., 2018, 17, 2519–2529 CrossRef PubMed; S. K. Bagal, M. Andrews, B. M. Bechle, J. Bian, J. Bilsland, D. C. Blakemore, J. F. Braganza, P. J. Bungay, M. S. Corbett, C. N. Cronin, J. J. Cui, R. Dias, N. J. Flanagan, S. E. Greasley, R. Grimley, K. James, E. Johnson, L. Kitching, M. L. Kraus, I. McAlpine, A. Nagata, S. Ninkovic, K. Omoto, S. Scales, S. E. Skerratt, J. Sun, M. Tran-Dube, G. J. Waldron, F. Wang and J. S. Warmus, Discovery of potent, selective, and peripherally restricted pan-Trk kinase inhibitors for the treatment of pain, J. Med. Chem., 2018, 61, 6779–6800 CrossRef PubMed.
  20. H.-P. Su, K. Rickert, C. Burlein, K. Narayan, M. Bukhtiyarova, D. M. Hurzy, C. Stump, X. Zhang, J. Reid, A. Krasowska-Zoladek, S. Tummala, J. M. Shipman, M. Kornienko, P. A. Lemaire, D. Krosky, A. Heller, A. Achab, C. Chamberlin, P. Saradjian, B. Sauvagnat, X. Yang, M. R. Ziebell, E. Nickbarg, J. M. Sanders, M. T. Bilodeau, S. S. Carroll, K. J. Lumb, S. M. Soisson, D. A. Henze and A. J. Cooke, Structural characterization of nonactive site, TrkA-selective kinase inhibitors, Proc. Natl. Acad. Sci. U. S. A., 2017, 114, E297–E306 CrossRef CAS PubMed; N. Furuya, T. Momose, K. Katsuno, N. Fushimi, H. Muranaka, C. Handa, T. Ozawa and T. Kinoshita, The juxtamembrane region of TrkA kinase is critical for inhibitor selectivity, Bioorg. Med. Chem. Lett., 2017, 27, 1233–1236 CrossRef PubMed; S. K. Bagal, K. Omoto, D. C. Blakemore, P. J. Bungay, J. Bilsland, P. J. Clarke, M. S. Corbett, C. N. Cronin, J. J. Cui, R. Dias, N. J. Flanagan, S. E. Greasley, R. Grimley, E. Johnson, D. Fengas, L. Kitching, M. L. Kraus, I. McAlpine, A. Nagata, G. J. Waldron and J. S. Warmus, Discovery of allosteric, potent, subtype selective, and peripherally restricted TrkA kinase inhibitors, J. Med. Chem., 2019, 62, 247–265 CrossRef PubMed.
  21. J. J. Irwin, T. Sterling, M. M. Mysinger, E. S. Bolstad and R. G. Coleman, ZINC: A free tool to discover chemistry for biology, J. Chem. Inf. Model., 2012, 52, 1757–1768 CrossRef CAS PubMed.
  22. P. C. D. Hawkins, A. G. Skillman and A. Nicholls, Comparison of shape-matching and docking as virtual screening tools, J. Med. Chem., 2007, 50, 74–82 CrossRef CAS PubMed.
  23. G. Subramanian, P. D. Johnson, T. Zachary, N. Roush, Y. Zhu, S. J. Bowen, A. Janssen, B. A. Duclos, T. Williams, C. Javens, N. D. Shalaly, D. M. Molina, A. J. Wittwer and J. L. Hirsch, Deciphering the allosteric binding mechanism of the human tropomyosin receptor kinase A (hTrkA) inhibitors, ACS Chem. Biol., 2019, 14, 1205–1216 CrossRef CAS PubMed.
  24. G. Subramanian, S. J. Bowen, Y. Zhu, N. Roush, T. Zachary, C. Javens, T. Williams, A. Janssen and A. Gonzales, Type 2 inhibitor leads of human tropomyosin receptor kinase (hTrkA), Bioorg. Med. Chem. Lett., 2019, 29, 126624 CrossRef PubMed.
  25. G. Subramanian, Y. Zhu, S. J. Bowen, N. Roush, J. A. White, D. Huczek, T. Zachary, C. Javens, T. Williams, A. Janssen and A. Gonzales, Lead identification and characterization of hTrkA type 2 inhibitors, Bioorg. Med. Chem. Lett., 2019, 29, 126680 CrossRef PubMed.
  26. A. J. Kooistra, G. K. Kanev, O. P. J. van Linden, R. Leurs, I. J. P. de Esch and C. de Graaf, KLIFS: a structural kinase-ligand interaction database, Nucleic Acids Res., 2016, 44, D365–D371 CrossRef CAS PubMed.
  27. R. Rahman, P. M.-U. Ung and A. Schlessinger, KinaMetrix: A web resource to investigate kinase conformations and inhibitor space, Nucleic Acids Res., 2019, 47, D361–D366 CrossRef CAS PubMed.
  28. B. J. Brandhuber, Y. Jiang, G. R. Kolakowski and S. L. Winski, Pyrazolyl urea, thiourea, guanidine and cyanoguanidine compounds as TrkA kinase inhibitors, WO2014/078417 (example 3), 2014.
  29. C. Kramer, G. Dahl, C. Tyrchan and J. Ulander, A comprehensive company database analysis of biological assay variability, Drug Discovery Today, 2016, 21, 1213–1221 CrossRef PubMed.
  30. T. Wang, M. L. Lamb, D. A. Scott, H. Wang, M. H. Block, P. D. Lyne, J. W. Lee, A. M. Davies, H. J. Zhang, Y. Zhu, F. Gu, Y. Han, B. Wang, P. J. Mohr, R. J. Kaus, J. A. Josey, E. Hoffmann, K. Thress, T. Macintyre, H. Wang, C. A. Omer and D. Yu, Identification of 4-aminopyrazolylpyrimidines as potent inhibitors of Trk kinases, J. Med. Chem., 2008, 51, 4672–4684 CrossRef CAS PubMed.
  31. A. Drilon, S. Siena, S. I. Ou, M. Patel, M. J. Ahn, J. Lee, T. M. Bauer, A. F. Farago, J. J. Wheler, S. V. Liu, R. Doebele, L. Giannetta, G. Cerea, G. Marrapese, M. Schirru, A. Amatu, K. Bencardino, L. Palmeri, A. Sartore-Bianchi, A. Vanzulli, S. Cresta, S. Damian, M. Duca, E. Ardini, G. Li, J. Christiansen, K. Kowalski, A. D. Johnson, R. Patel, D. Luo, E. Chow-Maneval, Z. Homby, P. S. Multani, A. T. Shaw and F. G. de Braud, Safety and antitumor activity of the multitargeted pan-Trk, ROS1, and ALK inhibitor Entrectinib: Combined results from two Phase I Trials (ALKA-372-001 and STARTRK-1), Cancer Discovery, 2017, 7, 400–409 CrossRef CAS PubMed.
  32. A. E. Brankin and P. W. Fowler, Predicting resistance is (not) futile, ACS Cent. Sci., 2019, 5, 1312–1314 CrossRef CAS PubMed; M. Aldeghi, V. Gapsys and B. L. de Groot, Predicting kinase inhibitor resistance: Physics-based and data-driven approaches, ACS Cent. Sci., 2019, 5, 1468–1474 CrossRef PubMed.
  33. E. Cocco, A. M. Schram, A. Kulick, S. Misale, H. H. Won, R. Yaeger, P. Razavi, R. Ptashkin, J. F. Hechtman, E. Toska, J. Cownie, R. Somwar, S. Shifman, M. Mattlar, S. D. Selcuklu, A. Samoila, S. Guzman, B. B. Tuch, K. Ebata, E. de Stanchina, R. J. Nagy, R. B. Lanman, B. Houck-Loomis, J. A. Patel, M. F. Berger, M. Ladanyi, D. M. Hyman, A. Drilon and M. Scaltriti, Resistance to TRK inhibition mediated by convergent MAPK pathway activation, Nat. Med., 2019, 25, 1422–1427 CrossRef CAS PubMed.

Footnotes

Atomic coordinates of the kinase domain of hTrkA. Ligand have been deposited in Protein Data Bank (PDB) under PDB codes 1: 6PMB; 3: 6PMA; 5: 6PME; 6: 6PMC.
The manuscript was written by G. S. All authors have given approval to the final version of the manuscript.
§ All experimental procedures are available as part of the ESI in ref. 23–25.
Electronic supplementary information (ESI) available: Computational workflow and X-ray data collection and structure determination details. See DOI: 10.1039/c9md00554d

This journal is © The Royal Society of Chemistry 2020