A set of <i>meta</i>-substituted 3-arylisoquinolinones have been identified that show substantial cytotoxicity in breast, liver, lung and colon cancer cell lines; these are up to 700-fold more active than the corresponding <i>para</i> analogues. These compounds were initially proposed as inhibitors of <i>N</i>-ribosyl dihydronicotinamide (NRH): quinone oxidoreductase 2 (NQO2) but were found to be inactive against the enzyme. Instead, COMPARE analysis suggested that 6-fluoro-3-(<i>meta</i>-fluorophenyl)isoquinolin-1(2<i>H</i>)-one (<b>4</b>) could mimic colchicine and interact with microtubules, a recognized target for cancer therapy. Subsequent docking, molecular dynamics simulations, and free energy analysis further suggested that compound <b>4</b> bound well into the colchicine-binding pocket of tubulin. Indeed, <b>4</b> suppressed tubulin polymerization, caused <i>G</i><sub>2</sub>/<i>M</i> cell cycle arrest, and induced apoptosis. Also, <b>4</b> inhibited the formation of endothelial cell capillary-like tubes and further disrupted the structure of preestablished tubes; the effects were not observed with <i>para</i> analogue <b>5</b>. In accordance with this, the computed free energy of binding of <b>5</b> to tubulin was lower in magnitude than that for <b>4</b> and appeared to arise in part from the inability of the <i>para</i> substituent to occupy a tubulin subpocket, which is possible in the <i>meta</i> orientation. In conclusion, the antiproliferative potential of the novel 3-arylisoquinolinones is markedly influenced by a subtle change in the structure (<i>meta</i> versus <i>para</i>). The <i>meta</i>-substituted isoquinolinone <b>4</b> is a microtubule-destabilizing agent with potential tumor-selectivity and antiangiogenic and vascular disrupting features.