Public research tool highlights unexpected off-target interactions and will help with drug R&D.

A kinase interaction map detailing the activity of 178 commercially available kinase inhibitors against a panel of 300 recombinant kinases has been drafted by researchers at the Fox Chase Cancer Center. Jeffrey R. Peterson, Ph.D., and colleagues used a high-throughput enzymatic assay to conduct a large-scale parallel screen, in duplicate, to identify novel inhibitors of kinase targets and investigate the target specificities of both research, clinical-stage, and therapeutically approved kinase inhibitors.

Reporting in Nature Biotechnology, the team says that the dataset is, as far as they are aware, the largest of its type in the public domain, and includes data from over 100,000 independent functional assays measuring pairwise inhibition of a single enzyme by a single compound. As well as identifying kinases that are inhibited by many different compounds and those that are not so easily inhibited by any of the small molecules evaluated, the work also uncovered a number of surprising off-target activities of many kinase inhibitors and highlighted potential leads for orphan kinases.

The results are described in as paper titled “Comprehensive assay of kinase catalytic activity reveals features of kinase inhibitor selectivity.”

Protein kinases represent important therapeutic targets because of their key roles in cell signaling, but designing inhibitors that are specific to a selected kinase or kinases is a major challenge, the authors report. Kinases exhibit an evolutionarily conserved ATP-binding site, which represents a useful target epitope for many kinase inhibitors, but also means that individual compounds are likely to inhibit multiple kinases. Recent technological advances have led to the development of methods to profile kinase target selectivity against large subsets of the 518 known human protein kinases, but these methods often measure the binding of small molecules to kinases, rather than whether they lead to functional inhibition of catalytic activity. As a result, how well these assays actually predict functional inhibition isn’t clear, they note.

Traditional approaches to kinase inhibitor discovery have focused on high-throughput screening of small molecules against a target of interest, and then testing hit compounds for selectivity against a panel of representative kinases. An alternative approach, however, is to screen libraries of compounds in a target-blind manner against a comprehensive panel of recom­binant protein kinases, to reveal the selectivity of each compound, the researchers continue. Compounds demonstrating the desired selectivity can then be chemically optimized. This method should feasibly be able to identify inhibitors that are selective for more than one kinase target.

To test this approach the Fox Chase team developed a high-throughput enzymatic assay to carry out a large-scale parallel screen of 178 known kinase inhibitors (including research, clinical-stage, and FDA-approved compounds) against a panel of 300 protein kinases. The panel covered representatives of all the major human protein kinase families, including the intended targets of 87.6% of the compounds tested. HotSpot, a radiomet­ric assay based on conventional filter-binding assays, was used to directly measure kinase catalytic activity toward a specific substrate.

Each protein kinase and kinase inhibitor combination (kinase-inhibitor pair) was tested in duplicate, and the average sub­strate phosphorylation results expressed as a percent­age of solvent control reactions (i.e., remaining kinase activity). Disparate replicates (only about 0.18% of the dataset) were eliminated.

The researchers’ published paper presents the mean remaining kinase activity for each kinase-inhibitor pair as a heat map, and as a spreadsheet. They have also generated a web-based Kinase Inhibitor Resource (KIR) database ( that allows compound- or kinase-specific queries of the dataset to be down­loaded or analyzed within a browser window.

The overall results were analyzed by two-way hierarchical clustering to group both kinases and inhibitors according to the similarity of their activity patterns: as expected, structurally related compounds generally grouped together, as did kinases with similar sequences, which were inhibited by similar patterns of compounds. Interestingly, exceptions to this were found among the Aurora, PDGFR, and FGFR family kinases, “suggesting the possibility that members of these families can be differentially targeted by small molecules,” the authors write.

One of the questions the Fox Chase team wanted to answer was whether each kinase in the panel was equally likely to be inhibited by any compound, or whether kinases differed in their sensitivity to small molecule inhibitors. To this end they ranked the kinases with respect to a selectivity score (S(50%)), calculated as the fraction of all compounds tested that inhibited the catalytic activity of each kinase by >50%. In fact they found that only 14 kinases in the panel weren’t inhibited by any of the compounds tested. In contrast, a subset of kinases including FLT3, TRKC, and HGK/MAP4K4, were broadly inhibited by large numbers of compounds, suggesting they represent targets that are highly susceptible to chemical inhibition, the authors remark.

The selectivity of newly identified kinase inhibitors is often assessed by testing against a limited panel of closely related kinases, based on the assumption that any off-target inter­actions that do occur will most likely impact on kinases with similar amino acid sequence. To test this quantitatively, the researchers compared the fraction of kinase targets within the same kinase subfamily as the primary target that was inhibited by a compound inhibited, with the fraction of inhibited targets that were outside the family of the primary target. The 10 most promiscuous compounds were removed from the analysis.

The overall results suggested that 42% of kinases inhibited by a given compound were actually from a different family than the subfamily of the intended kinase target. Taking inhibi­tors developed against tyrosine kinases as an example, the data indicated that 24% of the off-target hits were serine/threonine kinases. “These results highlight the importance of assessing the selectivity of kinase inhibitors against as broad a panel of kinases as possible,” the investigators stress.

Because kinase inhibitors that are selective for a very limited number of kinase targets represent particularly useful research tools, they then employed Gini coefficient-based calculations to rank their compounds in terms of selectivity. As expected, staurosporine and several of its structural analogs exhibited the lowest Gini scores (i.e., they were the most promiscuous), con­sistent with their known broad target spectrum. Conversely, the most selective compounds included several structurally distinct inhibitors of ErbB family kinases.  

In contrast to highly selective kinase inhibitors used to investigate kinase function in a research setting, clinical success has been achieved with compounds, such as dasatinib (Sprycel) and sunitinib (Sutent), which act as multitargeted kinase inhibitors. Such inhibitors will ideally show selectivity to a limited number of clinically relevant targets, but rational approaches to developing compounds with a defined target spectrum remains technically challenging, the authors note. To demonstrate the utility of their results for identifying potential inhibitors with such features, the authors mined their dataset to identify inhibitors with off-target activities against a limited number of cancer-relevant kinases.

The findings highlighted the ErbB family kinase inhibitor 4-(4-benzyloxyanilino)-6,7-dimethoxyquinazoline, as demonstrating potent inhibition of a few tyrosine kinases beyond ErbB family members and, more surprisingly, potent inhibition of the serine/threonine kinase CHK2, a critical component of the DNA damage checkpoint. “This illustrates how kinase profiling can reveal unanticipated novel scaffolds that show activity against highly divergent kinases of therapeutic interest,” they stress.

The overall screening results indicated that even the the most selective inhibitors generally targeted multiple kinases with similar potency, so the team searched specifically for compounds that inhibited a single kinase more potently than any other in the panel, a characteristic they termed “uni-specificity”. They excluded compounds targeting more than one kinase with similar potency, even if those kinases were closely related isoforms from the same subfamily. Interestingly, only a few compounds were found to display any degree of uni-specificity, and most of these showed only slight potency differences between their primary and secondary targets. In fact just 19 compounds inhibited their primary target with at least 20% higher potency than any other kinase in the panel. Among these were several inhibitors intended to target the epidermal growth factor receptor (EGFR), and the most uni-specific inhibi­tor, a 4,6-dianilinopyrimidine EGFR inhibitor (CAS no. 879127-07-8), inhibited EGFR catalytic activity by over 94%, but inhibited its next most strongly blocked target, MRCKα, but just 22%.

Notably, however, six of the top 17 uni-specific compounds actually more strongly inhbited kinases other than their intended targets, the authors continue. For example, the serine/threonine kinase RIPK2 was found to be a much more sensitive target of the IGF1R tyrosine kinase inhibitor AG1024, which was one of the most uni-specific compounds identified.

“In all cases these more potent off-target hits represent hitherto unknown kinase tar­gets of these compounds,” they write. “Remarkably, in all but one case, that of the compound DMBI, the most potent off-target hit falls outside of the kinase subfamily of the intended target…The uni-specific compounds described here provide new and selective inhibitors for their novel targets and in some cases, starting points for multitargeted kinase inhibition.”

The authors expect that their characterized inhibitor collection will represent a powerful tool to elucidate kinase functions in cell models. “These results have pushed the field closer to finding truly specific inhibitors of the processes that drive cancer,” Dr. Peterson states. “We now have a collection of kinase inhibitors that are more well-characterized and understood than any other library.”

Previous articleOxford BioMedica and Mayo Clinic Partner to Develop Gene Therapy for Glaucoma
Next articleIntellikine Earns $4M from Infinity on Start of Phase I Trials with PI3K Delta and Gamma