The novel lead bis(1H-2-indolyl)methanone inhibits autophosphorylation of platelet-derived growth factor (PDGF) receptor tyrosine kinase in intact cells. Various substituents in the 5- or 6-position of one indole ring increase or preserve potency, whereas most modifications of the ring structures and of the methanone group as well as substitution at both indoles result in weak or no activity. An ATP binding site model, derived by homology from the FGFR-1 tyrosine kinase crystal structure suggesting hydrogen bonds of one indole NH and the methanone oxygen with the backbone carbonyl and amide, respectively, of Cys684, explains why only one indole moiety is open for substitution and locates groups in the 5- or 6-position outside the pocket. The hitherto most active derivatives, 39, 53 and 67, inhibit both isoforms of the PDGF receptor kinase in intact cells, with IC(50) of 0.1-0.3 microM, and purified PDGFbeta-receptor in vitro, with IC(50) of 0.09, 0.1, or 0.02 microM, respectively. PDGF-stimulated DNA synthesis is inhibited by these derivatives with IC(50) values of 1-3 microM. Kinetic analysis of 53 showed an ATP-competitive mode of inhibition. The compounds are inactive or weakly active toward a number of other tyrosine kinases, including the FGF receptor 1, EGF receptor, and c-Src kinase, as well as toward serine-threonine kinases, including different PKC isoforms and GRK2, and appear therefore selective for PDGF receptor inhibition.

A new class of simple synthetic antimitotic compounds based on 2-aroylindoles was discovered. (5-Methoxy-1H-2-indolyl)-phenylmethanone (1) as well as analogous 3-fluorophenyl- (36) and 3-methoxyphenyl (3) derivatives displayed high cytotoxicity of IC(50) = 20 to 75 nM against the human HeLa/KB cervical, SK-OV-3 ovarian, and U373 astrocytoma carcinoma cell lines. The inhibition of proliferation correlated with the arrest in the G2/M phase of the cell cycle. In in vitro assays with tubulin isolated from bovine brain, in general antiproliferative activity correlated with inhibition of tubulin polymerization. Thus, the antimitotic activity of 2-aroylindoles is explained by interference with the mitotic spindle apparatus and destabilization of microtubules. In contrast to colchicine, vincristine, nocodazole, or taxol, 1 did not significantly affect the GTPase activity of beta-tubulin. Interestingly, selected compounds inhibited angiogenesis in the chorioallantoic membrane (CAM) assay. In xenograft experiments, 1 was highly active after oral administration at 200 mg/kg against the human amelanocytic melanoma MEXF 989 in athymic nude mice. We conclude, that 2-aroylindoles constitute an interesting new class of antitubulin agents with the potential to be clinically developed for cancer treatment.

FLT3 receptor tyrosine kinase is aberrantly active in many cases of acute myeloid leukemia (AML). Recently, bis(1H-indol-2-yl)methanones were found to inhibit FLT3 and PDGFR kinases. To optimize FLT3 activity and selectivity, 35 novel derivatives were synthesized and tested for inhibition of FLT3 and PDGFR autophosphorylation. The most potent FLT3 inhibitors 98 and 102 show IC50 values of 0.06 and 0.04 microM, respectively, and 1 order of magnitude lower PDGFR inhibiting activity. The derivatives 76 and 82 are 20- to 40-fold PDGFR selective. Docking at the recent FLT3 structure suggests a bidentate binding mode with the backbone of Cys-694. Activity and selectivity can be related to interactions of one indole moiety with a hydrophobic pocket including Phe-691, the only different binding site residue (PDGFR Thr-681). Compound 102 inhibited the proliferation of 32D cells expressing wildtype FLT3 or FLT3-ITD similarly as FLT3 autophosphorylation, and induced apoptosis in primary AML patient blasts.