Abstract
Quassinoids are natural product compounds known to possess tumor cytotoxicity and antimalarial activity. Neosergiolide and isobrucein B are two quassinoids previously isolated from roots and stems of Picrolemma sprucei. In screening studies to identify inhibitors that target STAT3, we discovered neosergeolide and isobrucein B as active compounds. Approximately 5000 plant-derived extracts were screened using a cell line that stably expresses a STAT3-dependent luciferase reporter and NPM-ALK, which constitutively induces STAT3 transcriptional activity. Of 25 total hits, a P. sprucei extract was potent and selective for STAT3 inhibition, and bioassay-guided isolation identified neosergeolide and isobrucein B as the inhibitory compounds. Western blot analysis confirmed that neosergeolide and isobrucein B not only inhibit the tyrosine phosphorylation and activation of STAT3 but also decrease total STAT3 protein levels via a mechanism due in part to enhanced proteasome-mediated degradation. Small-molecule proteasome inhibitors such as MG132 and ALLN reversed the ability of the two quassinoids to decrease STAT3 protein levels; furthermore, simultaneous incubation of various hematopoietic malignancy cell lines with either neosergeolide or isobrucein B and MG132 or ALLN antagonized the cytotoxic activity of the quassinoids. Assessment of neosergiolide and isobrucein B antitumor effects using an XTT assay revealed both compounds to possess potent cytotoxic activity across a broad spectrum of hematopoietic malignancies, with T-leukemias/lymphomas being especially responsive. For example, mycosis fungoides (MF)- and Sezary syndrome (SS)-derived cell lines, as well as non-MF/SS cutaneous T-cell lymphoma (CTCL) lines, were potently inhibited by both quassinoids (neosergiolide IC50 values: MAC-1, 11.6 nM; MAC-2A, 6.9 nM; Hut-78, 6.6 nM; HH, 4.3 nM; MJ, 7.0 nM; isobrucein B IC50 values: MAC-1, 31.9 nM; MAC-2A, 72.3 nM; Hut-78, 23.5 nM; HH; 20.3 nM; MJ, 13.5 nM). Non-hematopoietic cell lines representing various solid tumors also exhibited potent cytotoxic responses to the quassinoids (e.g., gastric carcinoma line AGS [neosergiolide IC50: 16.9 nM; isobrucein B IC50: 114.9 nM]). With rare exceptions, the cytotoxicity of the quassinoids against a specific tumor cell line correlated with STAT3 activation status; for example, breast cancer line MCF7 with inactive STAT3 was resistant to both quassinoids even at the maximum concentration tested (6.25 μM), whereas breast cancer lines MDA-MB-468 and MDA-MB-435s with activated STAT3 were inhibited by both compounds at low concentrations (neosergiolide IC50: MDA-MB-435s, 31.3 nM; MDA-MB-468, 29.9 nM; isobrucein B IC50: MDA-MB-435s, 209.3 nM; MDA-MB-468, 356.8 nM). The in vitro antitumor activity of the two quassinoids could also be demonstrated in vivo. For example, isobrucein B (1.0 mg/kg IP once q 3d x 5 doses) could be safely administered and potently inhibited the growth in SCID mice of the CD30+ primary CTCL MAC-1 cell line; mice at treatment day 16 showed average subcutaneous tumor volumes of 3839 ± 863 (s.e.) mm3 in the vehicle-control group and 913 ± 349 (s.e.) mm3 in the isobrucein B group (P=0.008, t-test). These results provide strong support for STAT3 targeting in antitumor drug discovery and suggest that quassinoids may have utility in such an approach.
A Non-Negative Matrix Tri-Factorization based Method for Predicting Antitumor Drug Sensitivity
