- 3:15 PM
582a

Computational Approaches for Prediction of Cross Reactivity in Human Protein Kinome

Sridhar Maddipati1, Ariel Fernandez2, and Sangtae Kim1. (1) School of Chemical Engineering, Purdue University, 480 Stadium Mall Dr., West Lafayette, IN 47906, (2) Bioenegineering, Rice University, 6100 Main Street, Houston, TX 77005

Protein kinases are central targets for drug-based treatment of diseases such as cancer, diabetes and arthritis. However, recent high throughput screening data [1] reveals that most kinase inhibitors of pharmacological relevance exhibit high cross reactivity. This cross reactivity might be desirable in particular therapeutic contexts, i.e., when the inhibitory impact spreads exclusively over targets of clinical interest, but it frequently leads to toxic side effects since kinases play a critical role in many cell signaling events. Progress in drug development thus faces challenges due to undesirable cross-reactivity and difficulties in modulating selectivity, both consequences of fold conservation. Kinase inhibitors target the ATP binding pocket in the kinases, which is highly conserved across the entire human kinome. Hence there is a need for assessing apriori the cross reactivity of different kinases based on their structure and sequence attributes.

Here we present a structure-based predictor of cross reactivity and validate it against affinity fingerprinting of the kinases. The predictor assesses protein environments of binding pockets, compares patterns of packing defects, in the form of hydrogen bonds which are not sufficiently wrapped by the hydrophobic groups and are thus prone to attack by water. The packing defects are sticky and are typically not conserved across homologs, making them ligand-anchoring sites potentially important to enhance selectivity [2]. We introduced a packing distance in kinase space based on the differences in the arrangement of packing defects and this packing based metric is conclusively shown to be equivalent to pharmacological distance generated by comparing affinity fingerprintings [3]. We also perform a hierarchical clustering of PDB-reported kinases and partition the kinase space according to their packing differences. Further, we show some experimental evidence for drug redesign of imatinib (Gleevec®) to sharpen and redirect its inhibitory impact. Thus, this tool should prove useful to target clinically relevant regions of the pharmacokinome, as our experimental assays reveal. Since the computation of packing based metric relies on the availability of the protein structures, we are currently investigating the generalization of packing metric to an equivalent sequence based metric to infer cross reactivity across kinases for which structures are not available and thus extend our predictor to the entire human kinome.

1. Fabian, M.A. et al. A small molecule kinase interaction map for clinical kinase inhibitors. Nat. Biotechnol. 23, 329–336 (2005).

2. Ariel Fernandez, Incomplete Protein Packing as a Selectivity Filter in Drug Design. Structure. 13, 1829-1836 (2005).

3. Ariel Fernández, Sridhar Maddipati, The a-priori inference of cross reactivity for drug targeted kinases. (Accepted in J. Med. Chem.).