Reply to "Statistical concerns about the GSEA procedure"

A novel bioinformatic approach reveals cooperation between Cancer/Testis genes in basal-like breast tumors

Breast cancer is the most prevalent type of cancer in women worldwide. Within breast tumors, the basal-like subtype has the worst prognosis, prompting the need for new tools to understand, detect, and treat these tumors. Certain germline-restricted genes show aberrant expression in tumors and are known as Cancer/Testis genes; their misexpression has diagnostic and therapeutic applications. Here we designed a new bioinformatic approach to examine Cancer/Testis gene misexpression in breast tumors. We identify several new markers in Luminal and HER-2 positive tumors, some of which predict response to chemotherapy. We then use machine learning to identify the two Cancer/Testis genes most associated with basal-like breast tumors: HORMAD1 and CT83. We show that these genes are expressed by tumor cells and not by the microenvironment, and that they are not expressed by normal breast progenitors; in other words, their activation occurs de novo. We find these genes are epigenetically repressed by DNA methylation, and that their activation upon DNA demethylation is irreversible, providing a memory of past epigenetic disturbances. Simultaneous expression of both genes in breast cells in vitro has a synergistic effect that increases stemness and activates a transcriptional profile also observed in double-positive tumors. Therefore, we reveal a functional cooperation between Cancer/Testis genes in basal breast tumors; these findings have consequences for the understanding, diagnosis, and therapy of the breast tumors with the worst outcomes.

Diptoindonesin G is a middle domain HSP90 modulator for cancer treatment

HSP90 inhibitors can target many oncoproteins simultaneously, but none have made it through clinical trials due to dose-limiting toxicity and induction of heat shock response, leading to clinical resistance. We identified diptoindonesin G (dip G) as an HSP90 modulator that can promote degradation of HSP90 clients by binding to the middle domain of HSP90 (K_d = 0.13 ± 0.02 μM) without inducing heat shock response. This is likely because dip G does not interfere with the HSP90-HSF1 interaction like N-terminal inhibitors, maintaining HSF1 in a transcriptionally silent state. We found that binding of dip G to HSP90 promotes degradation of HSP90 client protein estrogen receptor α (ER), a major oncogenic driver protein in most breast cancers. Mutations in the ER ligand-binding domain (LBD) are an established mechanism of endocrine resistance and decrease the binding affinity of mainstay endocrine therapies targeting ER, reducing their ability to promote ER degradation or transcriptionally silence ER. Because dip G binds to HSP90 and does not bind to the LBD of ER, unlike endocrine therapies, it is insensitive to ER LBD mutations that drive endocrine resistance. Additionally, we determined that dip G promoted degradation of WT and mutant ER with similar efficacy, downregulated ER- and mutant ER-regulated gene expression, and inhibited WT and mutant cell proliferation. Our data suggest that dip G is not only a molecular probe to study HSP90 biology and the HSP90 conformation cycle, but also a new therapeutic avenue for various cancers, particularly endocrine-resistant breast cancer harboring ER LBD mutations.

Biological response of adrenal carcinoma and melanoma cells to mitotane treatment

A previous case report described an adrenal incidentaloma initially misdiagnosed as adrenocortical carcinoma (ACC), which was treated with mitotane. The final diagnosis was metastatic melanoma of unknown primary origin. However, the patient developed rapid disease progression after mitotane withdrawal, suggesting a protective role for mitotane in a non-adrenal-derived tumor. The aim of the present study was to determine the biological response of primary melanoma cells obtained from that patient, and that of other established melanoma and ACC cell lines, to mitotane treatment using a proliferation assay, flow cytometry, quantitative PCR and microarrays. Although mitotane inhibited the proliferation of both ACC and melanoma cells, its role in melanoma treatment appears to be limited. Flow cytometry analysis and transcriptomic studies indicated that the ACC cell line was highly responsive to mitotane treatment, while the primary melanoma cells showed a moderate response in vitro. Mitotane modified the activity of several key biological processes, including ‘mitotic nuclear division’, ‘DNA repair’, ‘angiogenesis’ and ‘negative regulation of ERK1 and ERK2 cascade’. Mitotane administration led to elevated levels of DNA double-strand breaks, necrosis and apoptosis. The present study provides a comprehensive insight into the biological response of mitotane-treated cells at the molecular level. Notably, the present findings offer new knowledge on the effects of mitotane on ACC and melanoma cells.

Adropin Stimulates Proliferation and Inhibits Adrenocortical Steroidogenesis in the Human Adrenal Carcinoma (HAC15) Cell Line

Adropin is a multifunctional peptide hormone encoded by the ENHO (energy homeostasis associated) gene. It plays a role in mechanisms related to increased adiposity, insulin resistance, as well as glucose, and lipid metabolism. The low adropin levels are strongly associated with obesity independent insulin resistance. On the other hand, overexpression or exogenous administration of adropin improves glucose homeostasis. The multidirectional, adropin-related effects associated with the regulation of metabolism in humans also appear to be attributable to the effects of this peptide on the activity of various elements of the endocrine system including adrenal cortex. Therefore, the main purpose of the present study was to investigate the effect of adropin on proliferation and secretory activity in the human HAC15 adrenal carcinoma cell line. In this study, we obtained several highly interesting findings. First, GPR19, the main candidate sensitizer of adrenocortical cells to adropin, was expressed in HAC15 cells. Moreover, GPR19 expression was relatively stable and not regulated by ACTH, forskolin, or adropin itself. Our findings also suggest that adropin has the capacity to decrease expression levels of steroidogenic genes such as steroidogenic acute regulatory protein (StAR) and CYP11A1, which then led to a statistically significant inhibition in cortisol and aldosterone biosynthesis and secretion. Based on whole transcriptome study and research involving transforming growth factor (TGF)-β type I receptor kinase inhibitor we demonstrated that attenuation of steroidogenesis caused by adropin is mediated by the TGF-β signaling pathway likely to act through transactivation mechanism. We found that HAC15 cells treated with adropin presented significantly higher proliferation levels than untreated cells. Using specific intracellular inhibitors, we showed that adropin stimulate proliferation via ERK1/2 and AKT dependent signaling pathways. We have also demonstrated that expression of GPR19 is elevated in adrenocortical carcinoma in relation to normal adrenal glands. High level of GPR19 expression in adrenocortical carcinoma may constitute a negative prognostic factor of disease progression.

The limitations of simple gene set enrichment analysis assuming gene independence

Since its first publication in 2003, the Gene Set Enrichment Analysis method, based on the Kolmogorov-Smirnov statistic, has been heavily used, modified, and also questioned. Recently a simplified approach using a one-sample t-test score to assess enrichment and ignoring gene-gene correlations was proposed by Irizarry et al. 2009 as a serious contender. The argument criticizes Gene Set Enrichment Analysis's nonparametric nature and its use of an empirical null distribution as unnecessary and hard to compute. We refute these claims by careful consideration of the assumptions of the simplified method and its results, including a comparison with Gene Set Enrichment Analysis's on a large benchmark set of 50 datasets. Our results provide strong empirical evidence that gene-gene correlations cannot be ignored due to the significant variance inflation they produced on the enrichment scores and should be taken into account when estimating gene set enrichment significance. In addition, we discuss the challenges that the complex correlation structure and multi-modality of gene sets pose more generally for gene set enrichment methods.

Gene-Set Local Hierarchical Clustering (GSLHC)--A Gene Set-Based Approach for Characterizing Bioactive Compounds in Terms of Biological Functional Groups

Gene-set-based analysis (GSA), which uses the relative importance of functional gene-sets, or molecular signatures, as units for analysis of genome-wide gene expression data, has exhibited major advantages with respect to greater accuracy, robustness, and biological relevance, over individual gene analysis (IGA), which uses log-ratios of individual genes for analysis. Yet IGA remains the dominant mode of analysis of gene expression data. The Connectivity Map (CMap), an extensive database on genomic profiles of effects of drugs and small molecules and widely used for studies related to repurposed drug discovery, has been mostly employed in IGA mode. Here, we constructed a GSA-based version of CMap, Gene-Set Connectivity Map (GSCMap), in which all the genomic profiles in CMap are converted, using gene-sets from the Molecular Signatures Database, to functional profiles. We showed that GSCMap essentially eliminated cell-type dependence, a weakness of CMap in IGA mode, and yielded significantly better performance on sample clustering and drug-target association. As a first application of GSCMap we constructed the platform Gene-Set Local Hierarchical Clustering (GSLHC) for discovering insights on coordinated actions of biological functions and facilitating classification of heterogeneous subtypes on drug-driven responses. GSLHC was shown to tightly clustered drugs of known similar properties. We used GSLHC to identify the therapeutic properties and putative targets of 18 compounds of previously unknown characteristics listed in CMap, eight of which suggest anti-cancer activities. The GSLHC website http://cloudr.ncu.edu.tw/gslhc/ contains 1,857 local hierarchical clusters accessible by querying 555 of the 1,309 drugs and small molecules listed in CMap. We expect GSCMap and GSLHC to be widely useful in providing new insights in the biological effect of bioactive compounds, in drug repurposing, and in function-based classification of complex diseases.

Gene set enrichment analysis for non-monotone association and multiple experimental categories

BACKGROUND

Recently, microarray data analyses using functional pathway information, e.g., gene set enrichment analysis (GSEA) and significance analysis of function and expression (SAFE), have gained recognition as a way to identify biological pathways/processes associated with a phenotypic endpoint. In these analyses, a local statistic is used to assess the association between the expression level of a gene and the value of a phenotypic endpoint. Then these gene-specific local statistics are combined to evaluate association for pre-selected sets of genes. Commonly used local statistics include t-statistics for binary phenotypes and correlation coefficients that assume a linear or monotone relationship between a continuous phenotype and gene expression level. Methods applicable to continuous non-monotone relationships are needed. Furthermore, for multiple experimental categories, methods that combine multiple GSEA/SAFE analyses are needed.

RESULTS

For continuous or ordinal phenotypic outcome, we propose to use as the local statistic the coefficient of multiple determination (i.e., the square of multiple correlation coefficient) R2 from fitting natural cubic spline models to the phenotype-expression relationship. Next, we incorporate this association measure into the GSEA/SAFE framework to identify significant gene sets. Unsigned local statistics, signed global statistics and one-sided p-values are used to reflect our inferential interest. Furthermore, we describe a procedure for inference across multiple GSEA/SAFE analyses. We illustrate our approach using gene expression and liver injury data from liver and blood samples from rats treated with eight hepatotoxicants under multiple time and dose combinations. We set out to identify biological pathways/processes associated with liver injury as manifested by increased blood levels of alanine transaminase in common for most of the eight compounds. Potential statistical dependency resulting from the experimental design is addressed in permutation based hypothesis testing.

CONCLUSION

The proposed framework captures both linear and non-linear association between gene expression level and a phenotypic endpoint and thus can be viewed as extending the current GSEA/SAFE methodology. The framework for combining results from multiple GSEA/SAFE analyses is flexible to address practical inference interests. Our methods can be applied to microarray data with continuous phenotypes with multi-level design or the meta-analysis of multiple microarray data sets.