Microtubules are cellular organelles with functions that include control of cell division by mitosis, cell morphology, and transport of material within the cell. The anticancer drug paclitaxel (Taxol) promotes accelerated assembly of excessively stable microtubules. Consequently, treated cells tend to become arrested in mitosis. The drug also induces apoptotic cell death in vitro and in vivo. Prior to this study, the relative contributions of mitotic arrest and apoptosis to the in vivo antitumor effect and the relationship between the two factors had not been established; moreover, it is not known whether paclitaxel-induced mitotic arrest inevitably results in cell death. Our aim was to quantify the mitotic arrest and apoptosis induced by paclitaxel in 16 murine tumors in vivo and to correlate these two factors with the drug's antitumor effect. Inbred C3Hf/Kam mice were implanted with one of the following 16 syngeneic tumors: seven adenocarcinomas (MCa-4, MCa-29, MCa-35, MCa-K, OCa-I, ACa-SG, and HCa-I), two squamous cell carcinomas (SCC-IV and SCC-VII), six sarcomas (FSa, FSa-II, Sa-IIa, Sa-NH, NFSa, and Sa-4020), and one lymphoma (Ly-TH). The tumor growth delay induced by paclitaxel (40 mg/kg body weight given intravenously) was measured in 163 control and 163 treated mice, and its significance was assessed by Student's t test. In a separate group of 439 mice, the percentage of cells in mitosis or apoptosis was scored micromorphometrically at various times after paclitaxel administration. The significance of correlations between paclitaxel-induced tumor growth delay and paclitaxel-induced levels of mitosis or apoptosis was determined by simple correlation and Spearman's rank correlation. P values reported represent two-sided tests of statistical significance. Statistically significant tumor growth delays were found in response to paclitaxel treatment of mice for three of four murine mammary carcinomas (all P < or = .010), an ovarian carcinoma (P = .00003), a salivary gland adenocarcinoma (P = .0002), a lymphoma (P = .0002), and two of six sarcomas (both P < or = .034), but not for either of two squamous cell carcinomas or for the hepatocellular carcinoma. Paclitaxel-induced mitotic arrest was apparent in all tumor types, but to various degrees, and was not significantly correlated with growth delay (R2 = .16; P = .124). In contrast, apoptotic cell death in response to paclitaxel was not ubiquitous, but it was strongly correlated with growth delay (R2 = .59; P = .001). The pretreatment level of apoptosis was correlated with both paclitaxel-induced apoptosis (R2 = .71; P = .00004) and tumor growth delay (R2 = .55; P = .001). The antitumor effect of paclitaxel was correlated with paclitaxel-induced apoptosis and base-line apoptosis, but not with mitotic arrest. Apoptosis is an important mechanism of cell death in response to paclitaxel treatment of in vivo murine tumors. An underlying tumor type-specific propensity for apoptosis is implied by the correlation between pretreatment and paclitaxel-induced apoptosis. Both the extent of pretreatment apoptosis and the paclitaxel-induced percentage of apoptosis may be useful predictors of response to the drug.