Phenotypic modifications in ovarian cancer stem cells following Paclitaxel treatment

Abstract

Epithelial ovarian cancer (EOC) is the most lethal gynecologic malignancy. Despite initial responsiveness, 80% of EOC patients recur and present with chemoresistant and a more aggressive disease. This suggests an underlying biology that results in a modified recurrent disease, which is distinct from the primary tumor. Unfortunately, the management of recurrent EOC is similar to primary disease and does not parallel the molecular changes that may have occurred during the process of rebuilding the tumor. We describe the characterization of unique in vitro and in vivo ovarian cancer models to study the process of recurrence. The in vitro model consists of GFP+/CD44+/MyD88+ EOC stem cells and mCherry+/CD44-/MyD88- EOC cells. The in vivo model consists of mCherry+/CD44+/MyD88+ EOC cells injected intraperitoneally. Animals received four doses of Paclitaxel and response to treatment was monitored by in vivo imaging. Phenotype of primary and recurrent disease was characterized by quantitative polymerase chain reaction (qPCR) and Western blot analysis. Using the in vivo and in vitro models, we confirmed that chemotherapy enriched for CD44+/MyD88+ EOC stem cells. However, we observed that the surviving CD44+/MyD88+ EOC stem cells acquire a more aggressive phenotype characterized by chemoresistance and migratory potential. Our results highlight the mechanisms that may explain the phenotypic heterogeneity of recurrent EOC and emphasize the significant plasticity of ovarian cancer stem cells. The significance of our findings is the possibility of developing new venues to target the surviving CD44+/MyD88+ EOC stem cells as part of maintenance therapy and therefore preventing recurrence and metastasis, which are the main causes of mortality in patients with ovarian cancer.

Keywords: EMT; ovarian cancer stem cells; recurrence; slug.

Figure 1

Paclitaxel induces apoptosis only in CD44−/MyD88− OCC1 cells. Cells were treated with 0.2 μmol/L Paclitaxel for 48 h and allowed to recover for another 72 h. (A) Culture confluence was monitored by the Incucyte live-imaging system. Black arrow indicates the time of treatment initiation. White arrow indicates time of wash out of the treatment. Data is representative of three independent experiments. (B) Morphology assessed at 48 h posttreatment. Note the presence of apoptotic cells in OCC1 treated with Paclitaxel (red arrows). (C) Caspase-3 activity was measured 48 h posttreatment. Note the significant increase in caspase-3 activity in OCC1 following Paclitaxel treatment. *P < 0.005.

Figure 2

Paclitaxel enriches for CD44+/MyD88+ OCSC1 cells. Cocultures of GFP+ OCSC1 and mCherry+ OCC1 were treated with 0.2 μmol/L Paclitaxel for 48 h and allowed to recover for another 72 h. (A) Fluorescence was determined by fluorescence microscopy. (i–ii) Note the overgrowth of mCherry+ OCC1 (red) in relation to GFP+ OCSC1 (green) in the control group; (iii–iv) Mainly GFP+ ovarian cancer stem cell (OCSC)1 (green) are observed in cultures treated with Paclitaxel. (B) Enrichment of GFP+ OCSC1 cells following Paclitaxel treatment determined by flow cytometry. Data are representative of five independent experiments.

Figure 3

Paclitaxel upregulates stemness- and mesenchymal-associated markers in the surviving CD44+/MyD88+ OCSC1 cells. (A) Cells were treated with 0.2 μmol/L Paclitaxel for 24 h and proliferative potential determined by replating the cells. Only OCSC1 are able to repopulate the culture following Paclitaxel treatment. (B) Expression of mRNA stemness- and mesenchymal-associated genes in the surviving OCSC1 following Paclitaxel treatment compared to the nontreated OCSC1 cells. Expression levels were determined by qPCR. *P < 0.05 compared to Control. (C) Western blot analysis for the expression of Klf4 and Slug in the surviving OCSC1 following Paclitaxel treatment compared to the nontreated OCSC1 cells (Control).

Figure 4

Recurrent i.p. ovarian cancer model. (A) Intraperitoneal tumors were established with mCherry+ ovarian cancer stem cell (OCSC1-F2) as described in the Material and Methods section. Tumor burden was monitored for 32 days. Data shown are representative of five independent experiments (n = 10). (B) Correlation between mCherry fluorescent signal obtained from live imaging and actual mouse tumor burden. (i) representative image obtained from in vivo FX system 32 days postinjection of cells; (ii) corresponding photograph of carcinomatosis observed postmortem. (C) Treatment with Paclitaxel was initiated and carried out as described in the text. Note that tumors are undetectable in the treated group after four doses of Paclitaxel (Day 11). Recurrence develops in all mice (Day 37). Recurrent disease was not responsive to the second round of Paclitaxel treatment (Day 56). Data shown are representative of five independent experiments (n = 10 animals per group). (D) Treatment with Cisplatin was carried out as described in the text. (i) plot of region of interest (ROI) tumor area from five representative animals treated with Cisplatin; (ii) representative images comparing Control and Cisplatin-treated mouse. Note that in contrast to Paclitaxel, mice progressed with Cisplatin treatment.

Figure 5

Differential gene expression pattern in Paclitaxel-surviving cells. (A) Differential mRNA expression levels of stemness- and mesenchymal-associated genes between primary tumors (not exposed to Paclitaxel) and recurrent tumors (following Paclitaxel). n = 10 animals per group; blue arrows point to tumors analyzed. (B-i) Flow cytometry analysis comparing CD44 levels in primary and recurrent tumors. Note the enrichment of CD44+ cells in recurrent tumors. (ii) Differential levels of protein expression in primary and recurrent tumors determined by Western blot analysis. Data shown are representative of three independent experiments (n = 10 animals per group). (C) Vimentin expression was compared between Control/primary tumors and Treated/recurrent tumors using immunohistochemistry. Data shown are representative of three independent experiments (n = 10 animals per group). (D) High levels of vimentin were also observed in samples from recurrent epithelial ovarian cancer (EOC) patients. Images are representative of immunohistochemistry staining from two patients. *P < 0.001.

Figure 6

Slug promotes chemoresistance and aggressiveness. (A) Slug or Klf4 were overexpressed in ovarian cancer stem cell (OCSC1) and spheroid formation was quantified by manual counting. (B) Cells overexpressing Slug generated a more aggressive tumor in nude mice which did not respond to Paclitaxel. Images are representative of three independent experiments (n = 10 animals per group). (C) Tumor kinetics comparing vehicle-treated or Paclitaxel-treated mice harboring tumors established from OCSC1 transfected with empty vector or overexpressing Slug (Slug). day 0 designated as beginning of treatment; day 12 is the fourth and final dose of Paclitaxel. *P < 0.05 compared to Control. **, *P < 0.001 compared to control no treatment (n = 10 animals per group).

Figure 7

Depiction of proposed modifications within the tumor during chemotherapy and recurrence. Primary tumor is heterogeneous and composed of at least two types of cancer cells: the inherently chemoresistant tumor-initiating cells, A, and the chemosensitive-differentiated cancer cells, B. Chemotherapy will induce cell death in B but not in A. Due to its plasticity, surviving A undergoes molecular modification to A2, which then differentiates into B2 to rebuild the tumor.