Gene signatures have failed to predict responses to breast cancer therapy in patients to date. In this study, we used bioinformatic methods to explore the hypothesis that the existence of multiple drug resistance mechanisms in different patients may limit the power of gene signatures to predict responses to therapy. In addition, we explored whether substratification of resistant cases could improve performance. Gene expression profiles from 1,550 breast cancers analyzed with the same microarray platform were retrieved from publicly available sources. Gene expression changes were introduced in cases defined as sensitive or resistant to a hypothetical therapy. In the resistant group, up to five different mechanisms of drug resistance causing distinct or overlapping gene expression changes were generated bioinformatically, and their impact on sensitivity, specificity, and predictive values of the signatures was investigated. We found that increasing the number of resistance mechanisms corresponding to different gene expression changes weakened the performance of the predictive signatures generated, even if the resistance-induced changes in gene expression were sufficiently strong and informative. Performance was also affected by cohort composition and the proportion of sensitive versus resistant cases or resistant cases that were mechanistically distinct. It was possible to improve response prediction by substratifying chemotherapy-resistant cases from actual datasets (non-bioinformatically perturbed datasets) and by using outliers to model multiple resistance mechanisms. Our work supports the hypothesis that the presence of multiple resistance mechanisms in a given therapy in patients limits the ability of gene signatures to make clinically useful predictions. Cancer Res; 74(11); 1–16. ©2014 AACR.

Taxane antineoplastic agents are extensively taken up into hepatocytes by OATP1B-type transporters before metabolism and excretion. Because the biodistributional properties imposed upon these agents by different solubilizers drive clinically important pharmacodynamic endpoints, we tested the hypothesis that the in vitro and in vivo interaction of taxanes with OATP1B transporters is affected by the choice of drug delivery system. Transport of paclitaxel, docetaxel, and cabazitaxel was studied in vitro using various cell lines transfected with OATP1B1, OATP1B3, or the rodent equivalent OATP1B2. Pharmacokinetic studies were done in wild-type and OATP1B2-knockout mice in the presence or absence of polysorbate 80 (PS80) or Kolliphor EL (formerly Cremophor EL; CrEL). Paclitaxel and docetaxel, but not cabazitaxel, were transported substrates of OATP1B1, OATP1B3, and OATP1B2, and these in vitro transport processes were strongly reduced in the presence of clinically relevant concentrations of PS80 and CrEL. When paclitaxel was administered without any solubilizers, deficiency of OATP1B2 in mice was associated with a significantly decreased systemic clearance because of a liver distribution defect (P = 0.000484). However, this genotype dependence of paclitaxel clearance was masked in the presence of PS80 or CrEL because of significant inhibition of OATP1B2-mediated hepatocellular uptake of the drug (P < 0.05). Our findings confirm the importance of OATP1B-type transporters in the hepatic elimination of taxanes and indicate that this process can be inhibited by PS80 and CrEL. These results suggest that the likelihood of drug–drug interactions mediated by these transporters is strongly dependent on the selected taxane solubilizer. Cancer Res; 74(11); 1–9. ©2014 AACR.

PRL-3, an oncogenic dual-specificity phosphatase, is overexpressed in 50% of acute myelogenous leukemia (AML) and associated with poor survival. We found that stable expression of PRL-3 confers cytokine independence and growth advantage of AML cells. However, how PRL-3 mediates these functions in AML is not known. To comprehensively screen for PRL3-regulated proteins in AML, we performed SILAC-based quantitative proteomics analysis and discovered 398 significantly perturbed proteins after PRL-3 overexpression. We show that Leo1, a component of RNA polymerase II–associated factor (PAF) complex, is a novel and important mediator of PRL-3 oncogenic activities in AML. We described a novel mechanism where elevated PRL-3 protein increases JMJD2C histone demethylase occupancy on Leo1 promoter, thereby reducing the H3K9me3 repressive signals and promoting Leo1 gene expression. Furthermore, PRL-3 and Leo1 levels were positively associated in AML patient samples (N = 24; P < 0.01). On the other hand, inhibition of Leo1 reverses PRL-3 oncogenic phenotypes in AML. Loss of Leo1 leads to destabilization of the PAF complex and downregulation of SOX2 and SOX4, potent oncogenes in myeloid transformation. In conclusion, we identify an important and novel mechanism by which PRL-3 mediates its oncogenic function in AML. Cancer Res; 74(11); 1–11. ©2014 AACR.

Cancer-associated fibroblasts (CAF) stimulate tumor growth and metastasis. Signals supporting CAF function are thus emerging as candidate therapeutic targets in the tumor microenvironment. The chemokine CXCL14 is a potent inducer of CAF protumorigenic functions. This study is aimed at learning how the protumoral functions of CXCL14-expressing CAF are maintained. We found that the nitric oxide synthase NOS1 is upregulated in CXCL14-expressing CAF and in fibroblasts stimulated with CXCL14. Induction of Nos1 was associated with oxidative stress and occurred together with activation of NRF2 and HIF1α signaling in CXCL14-expressing CAF. Genetic or pharmacologic inhibition of NOS1 reduced the growth of CXCL14-expressing fibroblasts along with their ability to promote tumor formation following coinjection with prostate or breast cancer cells. Tumor analysis revealed reduced macrophage infiltration, with NOS1 downregulation in CXCL14-expressing CAF and lymphangiogenesis as a novel component of CXCL14-promoted tumor growth. Collectively, our findings defined key components of a signaling network that maintains the protumoral functions of CXCL14-stimulated CAF, and they identified NOS1 as intervention target for CAF-directed cancer therapy. Cancer Res; 74(11); 1–12. ©2014 AACR.

To reveal biologic mechanisms underlying clinical outcome of high-grade serous (HGS) epithelial ovarian carcinomas (EOC), we evaluated the association between tumor epigenetic changes and time to recurrence (TTR). We assessed methylation at approximately 450,000 genome-wide CpGs in tumors of 337 Mayo Clinic (Rochester, MN) patients. Semi-supervised clustering of discovery (n = 168) and validation (n = 169) sets was used to determine clinically relevant methylation classes. Clustering identified two methylation classes based on 60 informative CpGs, which differed in TTR in the validation set [R vs. L class, P = 2.9 × 10−3, HR = 0.52; 95% confidence interval (CI), 0.34–0.80]. Follow-up analyses considered genome-wide tumor mRNA expression (n = 104) and CD8 T-cell infiltration (n = 89) in patient subsets. Hypomethylation of CpGs located in 6p21.3 in the R class associated with cis upregulation of genes enriched in immune response processes (TAP1, PSMB8, PSMB9, HLA-DQB1, HLA-DQB2, HLA-DMA, and HLA-DOA), increased CD8 T-cell tumor infiltration (P = 7.6 × 10−5), and trans-regulation of genes in immune-related pathways (P = 1.6 × 10−32). This is the most comprehensive assessment of clinical outcomes with regard to epithelial ovarian carcinoma tumor methylation to date. Collectively, these results suggest that an epigenetically mediated immune response is a predictor of recurrence and, possibly, treatment response for HGS EOC. Cancer Res; 74(11); 1–8. ©2014 AACR.