Knowledge-Based Compact Disease Models: A Rapid Path from High-Throughput Data to Understanding Causative Mechanisms for a Complex Disease

High-throughput profiling of human tissues typically yields the gene lists composed of a variety of more or less relevant molecular entities. These lists are riddle by false positive observations that often obstruct generation of mechanistic hypothesis that may explain complex phenotype. From general probabilistic considerations, the gene lists enriched by the mechanistically relevant targets can be far more useful for subsequent experimental design or data interpretation. Using Alzheimer's disease as example, the candidate gene lists were processed into different tiers of evidence consistency established by enrichment analysis across subdatasets collected within the same experiment and across different experiments and platforms. The cutoffs were established empirically through ontological and semantic enrichment; resultant shortened gene list was reexpanded by Ingenuity Pathway Assistant tool. The resulting subnetworks provided the basis for generating mechanistic hypotheses that were partially validated by mined experimental evidence. This approach differs from previous consistency-based studies in that the cutoff on the Receiver Operating Characteristic of the true-false separation process is optimized by flexible selection of the consistency building procedure. The resultant Compact Disease Models (CDM) composed of the gene list distilled by this analytic technique and its network-based representation allowed us to highlight possible role of the protein traffic vesicles in the pathogenesis of Alzheimer's. Considering the distances and complexity of protein trafficking in neurons, it is plausible to hypothesize that spontaneous protein misfolding along with a shortage of growth stimulation may provide a shortcut to neurodegeneration. Several potentially overlapping scenarios of early-stage Alzheimer pathogenesis are discussed, with an emphasis on the protective effects of Angiotensin receptor 1 (AT-1) mediated antihypertensive response on cytoskeleton remodeling, along with neuronal activation of oncogenes, luteinizing hormone signaling and insulin-related growth regulation, forming a pleiotropic model of its early stages. Compact Disease Model generation is a flexible approach for high-throughput data analysis that allows extraction of meaningful, mechanism-centered gene sets compatible with instant translation of the results into testable hypotheses.

Knowledge-based compact disease models identify new molecular players contributing to early-stage Alzheimer's disease

BACKGROUND

High-throughput profiling of human tissues typically yield as results the gene lists comprised of a mix of relevant molecular entities with multiple false positives that obstruct the translation of such results into mechanistic hypotheses. From general probabilistic considerations, gene lists distilled for the mechanistically relevant components can be far more useful for subsequent experimental design or data interpretation.

RESULTS

The input candidate gene lists were processed into different tiers of evidence consistency established by enrichment analysis across subsets of the same experiments and across different experiments and platforms. The cut-offs were established empirically through ontological and semantic enrichment; resultant shortened gene list was re-expanded by Ingenuity Pathway Assistant tool. The resulting sub-networks provided the basis for generating mechanistic hypotheses that were partially validated by literature search. This approach differs from previous consistency-based studies in that the cut-off on the Receiver Operating Characteristic of the true-false separation process is optimized by flexible selection of the consistency building procedure. The gene list distilled by this analytic technique and its network representation were termed Compact Disease Model (CDM). Here we present the CDM signature for the study of early-stage Alzheimer's disease. The integrated analysis of this gene signature allowed us to identify the protein traffic vesicles as prominent players in the pathogenesis of Alzheimer's. Considering the distances and complexity of protein trafficking in neurons, it is plausible that spontaneous protein misfolding along with a shortage of growth stimulation result in neurodegeneration. Several potentially overlapping scenarios of early-stage Alzheimer pathogenesis have been discussed, with an emphasis on the protective effects of AT-1 mediated antihypertensive response on cytoskeleton remodeling, along with neuronal activation of oncogenes, luteinizing hormone signaling and insulin-related growth regulation, forming a pleiotropic model of its early stages. Alignment with emerging literature confirmed many predictions derived from early-stage Alzheimer's disease' CDM.

CONCLUSIONS

A flexible approach for high-throughput data analysis, the Compact Disease Model generation, allows extraction of meaningful, mechanism-centered gene sets compatible with instant translation of the results into testable hypotheses.

Ligand-directed profiling: applications to target drug discovery in cancer

BACKGROUND

Generation of targeted therapy remains a major challenge in medicine. The development of drugs that can discriminate between tumor cells and non-malignant cells would improve efficacy and reduce general side effects. Phage display allows identification of specific supramolecular complexes that can target therapeutic compounds or imaging agents, both in vitro and in vivo. The use of phage display to identify molecules expressed on the surface of human cancer cells without bias, as well as to provide initial steps toward identification of a ligand/receptor-based map of the human microvasculature, has broad implications for drug discovery in general, especially for cancer therapy.

OBJECTIVE/METHOD

In this review, we discuss the use of phage display technology as a ligand-directed targeting strategy and its applications to drug discovery.

CONCLUSION

Compared to other existing drug discovery platforms, phage display technology has the advantage to provide valuable clues pointing to target proteins in an unbiased biological context. The result from various display library screenings indicates that in many cases the selected peptide motifs mimic biological ligands. Analysis of peptide motifs targeting a receptor provides a basis for rational drug design of targeted peptidomimetics.

The use of volumetric CT as an imaging biomarker in lung cancer

RATIONALE AND OBJECTIVES

Lung cancer is the leading cause of cancer death in the United States. Mortality outcomes have improved only modestly over the past 30 years. There is intense focus on the development of better treatments for lung cancer. Major issues include the cost and time duration of the clinical trials required to establish the utility of a drug so that it can be formally approved by regulatory agencies. In clinical settings, biomarkers that accelerate assessments of responses to treatment could benefit patients by providing earlier diagnoses of progressive disease, particularly when there are multiple options for treatment, and the effects of toxicity from one treatment tend to limit the ability to administer the next line of therapy.

MATERIALS AND METHODS

Quantifying longitudinal changes in tumor volumes using computed tomography could eventually become a more useful surrogate endpoint for assessing tumor responses or progression events than simple unidimensional measurements.

RESULTS

The authors review the historical development of response measurements in lung cancer, set out the medical context for specifying volumetric imaging requirements and goals, compare volumetric technique to conventional methods, and identify the imaging profiles being pursued.

CONCLUSION

The Quantitative Imaging Biomarkers Alliance is investigating volumetric computed tomographic acquisition and analytic methods to increase the analytic power per subject enrolled in clinical trials to reduce the number of total subjects needed or shorten the length of time an individual needs to be followed to reliably establish drug response.

A multistep validation process of biomarkers for preclinical drug development

Biomarkers that can be measured in preclinical models in a high-throughput, reproducible manner offer the potential to increase the speed and efficacy of drug development. Development of therapeutic agents for many conditions is hampered by the limited number of validated preclinical biomarkers available to gauge pharmacoefficacy and disease progression, but the validation process for preclinical biomarkers has received limited attention. This report defines a five-step preclinical biomarker validation process and applies the process to a case study of diabetic retinopathy. By showing that a gene expression panel is highly reproducible, coincides with disease manifestation, accurately classifies individual animals and identifies animals treated with a known therapeutic agent, a biomarker panel can be considered validated. This particular biomarker panel consisting of 14 genes (C1inh, C1s, Carhsp1, Chi3l1, Gat3, Gbp2, Hspb1, Icam1, Jak3, Kcne2, Lama5, Lgals3, Nppa, Timp1) can be used in diabetic retinopathy pharmacotherapeutic research, and the biomarker development process outlined here is applicable to drug development efforts for other diseases.

Pathway analysis using random forests with bivariate node-split for survival outcomes

MOTIVATION

There is great interest in pathway-based methods for genomics data analysis in the research community. Although machine learning methods, such as random forests, have been developed to correlate survival outcomes with a set of genes, no study has assessed the abilities of these methods in incorporating pathway information for analyzing microarray data. In general, genes that are identified without incorporating biological knowledge are more difficult to interpret. Correlating pathway-based gene expression with survival outcomes may lead to biologically more meaningful prognosis biomarkers. Thus, a comprehensive study on how these methods perform in a pathway-based setting is warranted.

RESULTS

In this article, we describe a pathway-based method using random forests to correlate gene expression data with survival outcomes and introduce a novel bivariate node-splitting random survival forests. The proposed method allows researchers to identify important pathways for predicting patient prognosis and time to disease progression, and discover important genes within those pathways. We compared different implementations of random forests with different split criteria and found that bivariate node-splitting random survival forests with log-rank test is among the best. We also performed simulation studies that showed random forests outperforms several other machine learning algorithms and has comparable results with a newly developed component-wise Cox boosting model. Thus, pathway-based survival analysis using machine learning tools represents a promising approach in dissecting pathways and for generating new biological hypothesis from microarray studies.

AVAILABILITY

R package Pwayrfsurvival is available from URL: http://www.duke.edu/~hp44/pwayrfsurvival.htm.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Expression variation: its relevance to emergence of chronic disease and to therapy

Background

Stochastic fluctuations in the protein turnover underlie the random emergence of neural precursor cells from initially homogenous cell population. If stochastic alteration of the levels in signal transduction networks is sufficient to spontaneously alter a phenotype, can it cause a sporadic chronic disease as well – including cancer?

Methods

Expression in >80 disease-free tissue environments was measured using Affymetrix microarray platform comprising 54675 probe-sets. Steps were taken to suppress the technical noise inherent to microarray experiment. Next, the integrated expression and expression variability data were aligned with the mechanistic data covering major human chronic diseases.

Results

Measured as class average, variability of expression of disease associated genes measured in health was higher than variability of random genes for all chronic pathologies. Anti-cancer FDA approved targets were displaying much higher variability as a class compared to random genes. Same held for magnitude of gene expression. The genes known to participate in multiple chronic disorders demonstrated the highest variability. Disease-related gene categories displayed on average more intricate regulation of biological function vs random reference, were enriched in adaptive and transient functions as well as positive feedback relationships.

Conclusions

A possible causative link can be suggested between normal (healthy) state gene expression variation and inception of major human pathologies, including cancer. Study of variability profiles may lead to novel diagnostic methods, therapies and better drug target prioritization. The results of the study suggest the need to advance personalized therapy development.

Genomics screen in transformed stem cells reveals RNASEH2A, PPAP2C, and ADARB1 as putative anticancer drug targets

Since the sequencing of the human genome, recent efforts in cancer drug target discovery have focused more on the identification of novel functions of known genes and the development of more appropriate tumor models. In the present study, we investigated in vitro transformed human adult mesenchymal stem cells (MSC) to identify novel candidate cancer drug targets by analyzing the transcriptional profile of known enzymes compared with non-transformed MSC. The identified enzymes were compared with published cancer gene expression data sets. Surprisingly, the majority of up-regulated enzymes are already known cancer drug targets or act within known druggable pathways. Only three enzymes (RNASEH2A, ADARB1, and PPAP2C) are potentially novel targets that are up-regulated in transformed MSC and expressed in numerous carcinomas and sarcomas. We confirmed the overexpression of RNASEH2A, PPAP2C, and ADARB1 in transformed MSC, transformed fibroblasts, and cancer cell lines MCF7, SK-LMS1, MG63, and U2OS. In functional assays, we show that small interfering RNA knockdown of RNASEH2A inhibits anchorage-independent growth but does not alter in vitro proliferation of cancer cell lines, normal MSC, or normal fibroblasts. Knockdown of PPAP2C impaired anchorage-dependent in vitro growth of cancer cell lines and impaired the in vitro growth of primary MSC but not differentiated human fibroblasts. We show that the knockdown of PPAP2C decreases cell proliferation by delaying entry into S phase of the cell cycle and is transcriptionally regulated by p53. These in vitro data validate PPAP2C and RNASEH2A as putative cancer targets and endorse this in silico approach for identifying novel candidates.