br L induces apoptosis in a caspase
3.2. L80 induces apoptosis in a caspase-dependent manner
L80-induced apoptosis was evaluated by annexin V assay. A sig-nificant increase in early and late apoptotic LY-294002 in the presence of L80 (5–10 μM) was observed (p < 0.05, Fig. 2A). This effect was accom-panied by caspase-3 activation. The pretreatment with the pan-caspase inhibitor Z-VAD-fmk (20 μM, 1 h) significantly attenuated L80-induced caspase-3 activity (Fig. 2B and C). We additionally examined the effects of L80 and its parent drug
deguelin on apoptosis in normal human mammary epithelial MCF10A, normal murine mammary gland NMuMG and normal human embryonic kidney HEK293 cells. Annexin-V/PI staining analysis revealed that there were no statistically significant differences between CTL and L80 treatment (5 μM, 72 h) in MCF10A, NMuMG and HEK293 cells, in-dicating that L80 exhibits minimal cytotoxicity in normal cells (NS; not significant). Deguelin (0.5 μM, 72 h) elicited a significant effect on apoptosis in normal cells, which was observed at a 10-fold lower con-centration (p < 0.01, Fig. 2D).
3.3. L80 inhibits phosphorylation of AKT, MEK/ERK and JAK2/STAT3 signaling
AKT, MEK, JAK2, and STAT3 are the major client proteins of HSP90, and their function and activation are ultimately governed by HSP90 action [7,15,21]. We observed that L80 treatment not only down-regulated AKT, MEK, and ERK protein content, but also markedly re-duced their phosphorylation. Quantitative data further showed a sig-nificant reduction in the activation of these HSP90 client kinases and HSP70 expression in the presence of L80 (p < 0.05, http://journals. plos.org/plosone/article?id=10.1371/journal.pone.0049194 Fig. 3A). Of particular note, STAT3 is constitutively activated in all breast cancer sub-types but it is most often associated with TNBC and its cytosolic and nuclear activations are mediated by HSP90 activity [21,30,31]. Ex-posure to L80 treatment was found to downregulate phospho-JAK2 and phospho-STAT3 (Tyr705) in MDA-MB-231 and 4T1 cells (p < 0.05, http://journals.plos.org/plosone/article?id=10.1371/journal.pone.
0049194 Fig. 3B). This phenomenon was accompanied by down-regulation of STAT3 downstream factors including survivin and cyclin D1, as evidenced by significant reductions in their mRNA abundance (p < 0.01, Supplementary Fig. S1) and protein content (p < 0.05, http://journals.plos.org/plosone/article?id=10.1371/journal.pone. 0049194 Fig. 3B) in vitro.
3.4. L80 targets BCSC-like properties in TNBC cells
A significant relationship between constitutive activation of STAT3 and BCSC-like features was observed [28,30,32,33]. Since L80 sig-nificantly inhibits STAT3 activation, we next examined whether L80
regulates BCSC-like properties and assessed the stem cell surface mar-kers CD44high/CD24low and the progenitor marker aldehyde dehy-
0049194 Fig. 4B) were observed in MDA-MB-231 cells following L80 treatment. Mammospheres are known to be highly enriched with pu-tative cancer stem cells harboring self-renewal capacity . Treat-ment with L80 also impaired mammosphere-forming ability, as evi-denced by a marked decrease in the number and volume of mammospheres derived from MDA-MB-231 and 4T1 cells (p < 0.001, Fig. 4C). Consistent with the in vitro data, exposure to L80 significantly reduced the number and volume of mammospheres derived from 4T1 allograft tumors in vivo (p < 0.001, Fig. 4D). A recent study has shown that HSP90 directly interacts with Oct4 and Nanog and prevents them
from degrading via the ubiquitin proteasome pathway, suggesting that they are potential novel HSP90 client proteins . We observed that L80 significantly downregulates Nanog and Oct4 protein content (p < 0.01, Supplementary Fig. S2) in MDA-MB-231 and 4T1 cells,
suggesting that downregulation of these pluripotent transcription fac-tors by L80 could contribute to the suppression of breast cancer stem cell-like properties such as mammosphere-forming ability.
(caption on next page)
Fig. 5. L80 impairs BCSC-enriched TNBC tumors. (A) 8 × 104 cells from mammosphere cultures were orthotopically injected into the duct of the fourth mammary gland of BALB/c female mice. Mice were administered intraperitoneally with L80 (20 mg/kg, every other day) or control solvent for 27 days (n = 8/each group). (B–D) After exposure to L80 or control solvent in allograft mice, tumor growth (B, ***p < 0.001), tumor weight (C, **p < 0.01) and body weight (D, NS, not significant) were evaluated. (E) Effect of L80 on nuclear Ki-67 expression. Tissue sections were immunostained for Ki-67 (red) or normal rabbit IgG (as a negative control) with DAPI (blue); images are shown at magnification ( × 500). The graph represents the percentage of Ki-67-positive cells (bottom panel, ***p < 0.001).