An Investigation of Bone Resorption and Dictyostelium discoideum Growth Inhibition by Bisphosphonate Drugs

• Published in: Journal of medicinal chemistry Vol. 45; no. 14; pp. 2894 - 2903
• Main Authors: Szabo, Christina M, Martin, Michael B, Oldfield, Eric
• Format: Journal Article
• Language: English
• Published: Washington, DC American Chemical Society 04-07-2002
• Subjects: Alkyl and Aryl Transferases - antagonists & inhibitors; Animals; Antibiotics. Antiinfectious agents. Antiparasitic agents; Antiparasitic agents; Biological and medical sciences; Bone Resorption - metabolism; Bones, joints and connective tissue. Antiinflammatory agents; Crystallography, X-Ray; Dictyostelium - drug effects; Dictyostelium - growth & development; Diphosphonates - chemistry; Diphosphonates - pharmacology; Enzyme Inhibitors - chemistry; Enzyme Inhibitors - pharmacology; Geranyltranstransferase; Humans; Inhibitory Concentration 50; Medical sciences; Models, Molecular; Pharmacology. Drug treatments; Quantitative Structure-Activity Relationship; Rats; Imidazole derivatives; Geranyltranstransferase; Diphosphonic acid derivatives; Bisphosphonates; Modeling; Phosphorus Organic compounds; Pyridine derivatives; Dictyostelium discoideum; Antiprotozoal agent; Structure activity relation; Parasiticid; Molecular model; Comparative molecular field analysis method; Bone defect; Inhibition; Protozoa; Enzyme; Transferases; Prediction; Enzyme inhibitor; Resorption; Organic phosphonate; X ray diffraction; In vitro; Three dimensional model; Bone; Mycetozoa
• ISSN: 0022-2623; 1520-4804
• DOI: 10.1021/jm010279+

We report the results of 3D-QSAR/CoMFA investigations of the activity of bisphosphonate drugs, farnesyl pyrophosphate synthase (FPPSase) inhibitors, in the inhibition of bone resorption as well as the growth of Dictyostelium discoideum. In the case of D. discoideum, we find an experimental versus QSAR predicted pIC50 R 2 value of 0.94 for 16 bisphosphonates over the 9−1200 μM range of IC50 values, a cross-validated R 2 = 0.90, and a bootstrapped R 2 = 0.94, and we demonstrate that this approach has predictive utility (a 0.18 pIC50 rms error for three test sets of 3 predictions). In bone resorption, we find an experimental versus predicted pLED (lowest effective dose) R 2 = 0.79 for 35 bisphosphonates over the 0.0001−1 mg of P/kg LED range, a cross-validated R 2 = 0.75, and a bootstrapped R 2 = 0.79. Two sets of 31 compounds were used as training sets for the predicted pLED values for two sets of 4 compounds which have an rms error of 0.44, larger than that found with D. discoideum. However, this can be attributed to the rather large uncertainties in the experimental bone resorption data which are almost all reported in decade steps (ΔpLED = 1). The CoMFA predicted (rat) bone antiresorptive pLED values are in agreement with literature (human recombinant) FPPSase inhibition results with an rms error of 0.45 (a factor of 2.8 error in activity prediction). We also report the single-crystal X-ray crystallographic structure of the compound most active in D. discoideum growth inhibition, 2-(3-picolyl)-aminomethylene-1,1-bisphosphonic acid. The structure clearly shows the presence of bond length alternation in the picolyl ring and a planar amino group linked by a very short (1.346 Å) bond to the picolyl group, an amidinium-like structure which is also expected to occur in other highly active species such as minodronate and zoledronate. Overall, these results show that it is now possible to predict the activity of bisphosphonates using 3D-QSAR/CoMFA methods, although bone resorption studies should benefit from additional, accurate information on enzyme inhibition.