Prediction of Response to Paclitaxel by ChemoFx assay Correlates with Estrogen Receptor Status and Changes in Apoptosis Pathway in Human Breast Cancers.

Abstract

Background: Paclitaxel belongs to the taxane family of therapeutics, which have emerged as critically important drugs for breast cancer treatment. In addition to inhibiting cell growth by interfering with microtubule disassembly, its mechanism of action also includes induction of apoptosis. Recent studies suggest that besides being a key predictor for endocrine therapy response, Estrogen Receptor (ER) status also influences sensitivity of breast cancer to paclitaxel, with ER negative tumors being more responsive to the drug.Methods: The ChemoFx live cell chemoresponse assay was performed on 25 breast cancer cell lines (10 ER+ and 15 ER-). These cells were treated with a range of 10 doses of paclitaxel for 72 hours before DAPI staining of nuclei and counting. AUC (Area Under Curve) values were calculated and additional statistical analysis was performed on the resulting dose-response curves. Differential gene expression analysis was conducted to compare ER+ (n=82) and ER- (n=51) breast cancer patients using a public Microarray database. In addition, 2 of the 25 breast cancer cell lines, T47D (ER+) and SKBR3 (ER-), were treated with paclitaxel, lysed, and analyzed with Western blotting to detect cleaved caspase-3 and cleaved PARP expression, with beta-actin employed as a normal control.Results: The ChemoFx assay results revealed that none of the ER+ cells were categorized as R (responsive) to paclitaxel, with seven NR (non-responsive) and one IR (intermediate responsive). On the contrary, of the 15 ER- cell lines, three were categorized as R, only four were categorized as NR, and eight were categorized as IR. Statistical analysis suggested that paclitaxel responsiveness based on ChemoFx assay correlates with ER status (Chi-square test, p<0.05), with ER- breast tumors being more responsive to paclitaxel. Microarray analysis revealed differential expressions of genes implicated in the apoptosis pathway (q< 0.05) in ER+ and ER- breast cancers. Western blot analysis showed that paclitaxel induced cleaved caspase-3 and cleaved PARP expressions, both of which are indicators of activation of apoptosis, in SKBR3 cells (ER-), but not in T47D cells (ER+).Conclusions: ER status appears to predict in part, the response of breast cancer cells to paclitaxel as determined by the ChemoFx assay. ER-negative breast cancer cells are more likely to be responsive, which is consistent with established clinical findings. Our assay also distinguishes between NR/IR and R to paclitaxel within the ER- population. Similar ChemoFx assays are being performed on primary cultures from ER+ and ER- breast cancer patient specimens. Results from RNA microarray and Western blot analyses indicate that differences in gene expression in the apoptosis pathway, and in activation of apoptosis pathway, namely changes in expressions of cleaved PARP and cleaved caspase-3 in response to paclitaxel, may explain differences in the responsiveness of ER+ and ER- breast cancers to paclitaxel. This also suggests a potential role of cleaved PARP and cleaved caspase-3 as biomarkers in addition to ER for prediction of paclitaxel responsiveness in breast cancer. Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 2028.

From Patient-Specific Mathematical Neuro-Oncology to Precision Medicine

“…The PI model characterizes the net rates of growth and invasion of the glioma cells contributing to this overall profile, a sum of individual cell behaviors. Swanson et al have demonstrated that D and ρ can be calculated on a patient-specific basis and can vary widely, even for patients within the same histological grade (Harpold et al, 2007; Swanson et al, 2008a; Szeto et al, 2009b; Wang et al, 2009; Rockne et al, 2010). (B) A simulation of the reaction-diffusion mathematical model on an anatomically accurate brain phantom (Cocosco et al, 2004) with differential motility in gray and white matter as proposed by Swanson (1999).…”

“…The MRI-detectable edge of the lesion is superimposed as a dark gray contour emphasizing the extent of invasion well beyond the threshold of detection. From Wang et al (2009) with permission from Cancer Research.…”

“…The PI model can further incorporate differential motility of glioma cells through gray and white matter of the brain, providing predictions of diffuse tumor invasion through the regions of the brain that are specific to the patient’s tumor (Figure 2). This simple model has served as a foundation for patient-specific MNO and provided numerous insights into clinical behaviors such as survival outcome (Pallud et al, 2006; Swanson et al, 2008b; Wang et al, 2009; Rockne et al, 2010), hypoxia development (Szeto et al, 2009b), response to surgical resection (Swanson et al, 2008b), chemo- and radiation therapies (Rockne et al, 2010), biological aggressiveness (Szeto et al, 2009a; Ellingson et al, 2010b), and to date is the single most applied patient-specific clinical scale model for glioma growth and response to therapy. Extensions to this model include consideration of anisotropic growth in white matter tracts (Jbabdi et al, 2005).…”

“…The authors are only aware of work based on such an outline in the context of gliomas by the Swanson group (Swanson et al, 2002a, 2008b; Szeto et al, 2009b; Wang et al, 2009; Neal and Kerckhoffs, 2010; Rockne et al, 2010; Baldock et al, 2012a,b; Gu et al, 2012; Neal et al, 2012, 2013). But that is not to say that other efforts are not informative or useful.…”

“…In a study of 32 newly diagnosed glioblastoma patients, Wang et al (2009) used the PI model to find relationships between the Patient-specific model (PSM) parameters for glioma cell net dispersal ( D ), proliferation (ρ), and prognosis. As illustrated in Figure 1, patient-specific estimates D and ρ combine with the patient’s MRI to yield a map of the diffuse gradient of glioma cells that is expected to lie beyond thresholds visible to imaging (Figure 1).…”