Design of a multi-signature ensemble classifier predicting neuroblastoma patients' outcome

Adaptive phenotype drives resistance to androgen deprivation therapy in prostate cancer

Background

Prostate cancer (PCa), the second most common cancer affecting men worldwide, shows a broad spectrum of biological and clinical behaviour representing the epiphenomenon of an extreme heterogeneity. Androgen deprivation therapy is the mainstay of treatment for advanced forms but after few years the majority of patients progress to castration-resistant prostate cancer (CRPC), a lethal form that poses considerable therapeutic challenges.

Methods

Western blotting, immunocytochemistry, invasion and reporter assays, and in vivo studies were performed to characterize androgen resistant sublines phenotype in comparison to the parental cell line LNCaP. RNA microarray, mass spectrometry, integrative transcriptomic and proteomic differential analysis coupled with GeneOntology and multivariate analyses were applied to identify deregulated genes and proteins involved in CRPC evolution.

Results

Treating the androgen-responsive LNCaP cell line for over a year with 10 μM bicalutamide both in the presence and absence of 0.1 nM 5-α-dihydrotestosterone (DHT) we obtained two cell sublines, designated PDB and MDB respectively, presenting several analogies with CRPC. Molecular and functional analyses of PDB and MDB, compared to the parental cell line, showed that both resistant cell lines were PSA low/negative with comparable levels of nuclear androgen receptor devoid of activity due to altered phosphorylation; cell growth and survival were dependent on AKT and p38MAPK activation and PARP-1 overexpression; their malignant phenotype increased both in vitro and in vivo. Performing bioinformatic analyses we highlighted biological processes related to environmental and stress adaptation supporting cell survival and growth. We identified 15 proteins that could direct androgen-resistance acquisition. Eleven out of these 15 proteins were closely related to biological processes involved in PCa progression.

Conclusions

Our models suggest that environmental factors and epigenetic modulation can activate processes of phenotypic adaptation driving drug-resistance. The identified key proteins of these adaptive phenotypes could be eligible targets for innovative therapies as well as molecules of prognostic and predictive value.

Electronic supplementary material

The online version of this article (10.1186/s12964-017-0206-x) contains supplementary material, which is available to authorized users.

Artificial neural network classifier predicts neuroblastoma patients' outcome

Background

More than fifty percent of neuroblastoma (NB) patients with adverse prognosis do not benefit from treatment making the identification of new potential targets mandatory. Hypoxia is a condition of low oxygen tension, occurring in poorly vascularized tissues, which activates specific genes and contributes to the acquisition of the tumor aggressive phenotype. We defined a gene expression signature (NB-hypo), which measures the hypoxic status of the neuroblastoma tumor. We aimed at developing a classifier predicting neuroblastoma patients’ outcome based on the assessment of the adverse effects of tumor hypoxia on the progression of the disease.

Methods

Multi-layer perceptron (MLP) was trained on the expression values of the 62 probe sets constituting NB-hypo signature to develop a predictive model for neuroblastoma patients’ outcome. We utilized the expression data of 100 tumors in a leave-one-out analysis to select and construct the classifier and the expression data of the remaining 82 tumors to test the classifier performance in an external dataset. We utilized the Gene set enrichment analysis (GSEA) to evaluate the enrichment of hypoxia related gene sets in patients predicted with “Poor” or “Good” outcome.

Results

We utilized the expression of the 62 probe sets of the NB-Hypo signature in 182 neuroblastoma tumors to develop a MLP classifier predicting patients’ outcome (NB-hypo classifier). We trained and validated the classifier in a leave-one-out cross-validation analysis on 100 tumor gene expression profiles. We externally tested the resulting NB-hypo classifier on an independent 82 tumors’ set. The NB-hypo classifier predicted the patients’ outcome with the remarkable accuracy of 87 %. NB-hypo classifier prediction resulted in 2 % classification error when applied to clinically defined low-intermediate risk neuroblastoma patients. The prediction was 100 % accurate in assessing the death of five low/intermediated risk patients. GSEA of tumor gene expression profile demonstrated the hypoxic status of the tumor in patients with poor prognosis.

Conclusions

We developed a robust classifier predicting neuroblastoma patients’ outcome with a very low error rate and we provided independent evidence that the poor outcome patients had hypoxic tumors, supporting the potential of using hypoxia as target for neuroblastoma treatment.

Electronic supplementary material

The online version of this article (doi:10.1186/s12859-016-1194-3) contains supplementary material, which is available to authorized users.

Oncolytic adenovirus-mediated short hairpin RNA targeting MYCN gene induces apoptosis by upregulating RKIP in neuroblastoma

The amplification of MYCN is a typical characteristic of aggressive neuroblastomas, whereas acquired mutations of p53 lead to refractory and relapsed cases. We had previously examined the applicability of the replication-competent oncolytic adenovirus, ZD55-shMYCN, to deliver a short hairpin RNA targeting MYCN gene for p53-null and MYCN-amplified neuroblastoma cell line LA1-55N. Our data have shown that ZD55-shMYCN has an additive tumor growth inhibitory response through shRNA-mediated MYCN knockdown and ZD55-mediated cancer cell lysis. In this regard, ZD55-shMYCN can downregulate MYCN and perform anticancer effects, thereby acquiring significance in the administration of MYCN-amplified and p53-null neuroblastomas. Hence, we further investigated the anticancer properties of ZD55-shMYCN in neuroblastomas. Our data showed that ZD55-shMYCN induced G2/M arrest via decreasing the levels of cyclin D1 and cyclin B1 irrespective of p53 status. ZD55-shMYCN effectively induced apoptosis in neuroblastomas through activation of caspase-3 and enhancing PARP cleavage. Furthermore, ZD55-shMYCN could downregulate phosphoinositide 3-kinase and pAkt and upregulate RKIP levels. Similarly, pro-apoptosis was revealed by the histopathologic examination of paraffin-embedded section of resected tumors of mice xenograft. In vitro and in vivo studies, we elucidate the apoptosis properties and mechanisms of action of ZD55-shMYCN, which provide a promising approach for further clinical development.

A combined gene signature of hypoxia and notch pathway in human glioblastoma and its prognostic relevance

Hypoxia is a hallmark of solid tumors including glioblastoma (GBM). Its synergism with Notch signaling promotes progression in different cancers. However, Notch signaling exhibits pleiotropic roles and the existing literature lacks a comprehensive understanding of its perturbations under hypoxia in GBM with respect to all components of the pathway. We identified the key molecular cluster(s) characteristic of the Notch pathway response in hypoxic GBM tumors and gliomaspheres. Expression of Notch and hypoxia genes was evaluated in primary human GBM tissues by q-PCR. Clustering and statistical analyses were applied to identify the combination of hypoxia markers correlated with upregulated Notch pathway components. We found well-segregated tumor—clusters representing high and low HIF-1α/PGK1-expressors which accounted for differential expression of Notch signaling genes. In combination, a five-hypoxia marker set (HIF-1α/PGK1/VEGF/CA9/OPN) was determined as the best predictor for induction of Notch1/Dll1/Hes1/Hes6/Hey1/Hey2. Similar Notch-axis genes were activated in gliomaspheres, but not monolayer cultures, under moderate/severe hypoxia (2%/0.2% O₂). Preliminary evidence suggested inverse correlation between patient survival and increased expression of constituents of the hypoxia-Notch gene signature. Together, our findings delineated the Notch-axis maximally associated with hypoxia in resected GBM, which might be prognostically relevant. Its upregulation in hypoxia-exposed gliomaspheres signify them as a better in-vitro model for studying hypoxia-Notch interactions than monolayer cultures.

Individual patient risk stratification of high-risk neuroblastomas using a two-gene score suited for clinical use

Several gene expression-based prognostic signatures have been described in neuroblastoma, but none have successfully been applied in the clinic. Here we have developed a clinically applicable prognostic gene signature, both with regards to number of genes and analysis platform. Importantly, it does not require comparison between patients and is applicable amongst high-risk patients. The signature is based on a two-gene score (R-score) with prognostic power in high-stage tumours (stage 4 and/or MYCN-amplified diagnosed after 18 months of age). QPCR-based and array-based analyses of matched cDNAs confirmed cross platform (array-qPCR) transferability. We also defined a fixed cut-off value identifying prognostically differing subsets of high-risk patients on an individual patient basis. This gene expression signature independently contributes to the current neuroblastoma classification system, and if prospectively validated could provide further stratification of high-risk patients, and potential upfront identification of a group of patients that are in need of new/additional treatment regimens.

HOXD-AS1 is a novel lncRNA encoded in HOXD cluster and a marker of neuroblastoma progression revealed via integrative analysis of noncoding transcriptome

Background

Long noncoding RNAs (lncRNAs) constitute a major, but poorly characterized part of human transcriptome. Recent evidence indicates that many lncRNAs are involved in cancer and can be used as predictive and prognostic biomarkers. Significant fraction of lncRNAs is represented on widely used microarray platforms, however they have usually been ignored in cancer studies.

Results

We developed a computational pipeline to annotate lncRNAs on popular Affymetrix U133 microarrays, creating a resource allowing measurement of expression of 1581 lncRNAs. This resource can be utilized to interrogate existing microarray datasets for various lncRNA studies. We found that these lncRNAs fall into three distinct classes according to their statistical distribution by length. Remarkably, these three classes of lncRNAs were co-localized with protein coding genes exhibiting distinct gene ontology groups. This annotation was applied to microarray analysis which identified a 159 lncRNA signature that discriminates between localized and metastatic stages of neuroblastoma. Analysis of an independent patient cohort revealed that this signature differentiates also relapsing from non-relapsing primary tumors. This is the first example of the signature developed via the analysis of expression of lncRNAs solely. One of these lncRNAs, termed HOXD-AS1, is encoded in HOXD cluster. HOXD-AS1 is evolutionary conserved among hominids and has all bona fide features of a gene. Studying retinoid acid (RA) response of SH-SY5Y cell line, a model of human metastatic neuroblastoma, we found that HOXD-AS1 is a subject to morphogenic regulation, is activated by PI3K/Akt pathway and itself is involved in control of RA-induced cell differentiation. Knock-down experiments revealed that HOXD-AS1 controls expression levels of clinically significant protein-coding genes involved in angiogenesis and inflammation, the hallmarks of metastatic cancer.

Conclusions

Our findings greatly extend the number of noncoding RNAs functionally implicated in tumor development and patient treatment and highlight their role as potential prognostic biomarkers of neuroblastomas.

An investigation of biomarkers derived from legacy microarray data for their utility in the RNA-seq era

BACKGROUND

Gene expression microarray has been the primary biomarker platform ubiquitously applied in biomedical research, resulting in enormous data, predictive models, and biomarkers accrued. Recently, RNA-seq has looked likely to replace microarrays, but there will be a period where both technologies co-exist. This raises two important questions: Can microarray-based models and biomarkers be directly applied to RNA-seq data? Can future RNA-seq-based predictive models and biomarkers be applied to microarray data to leverage past investment?

RESULTS

We systematically evaluated the transferability of predictive models and signature genes between microarray and RNA-seq using two large clinical data sets. The complexity of cross-platform sequence correspondence was considered in the analysis and examined using three human and two rat data sets, and three levels of mapping complexity were revealed. Three algorithms representing different modeling complexity were applied to the three levels of mappings for each of the eight binary endpoints and Cox regression was used to model survival times with expression data. In total, 240,096 predictive models were examined.

CONCLUSIONS

Signature genes of predictive models are reciprocally transferable between microarray and RNA-seq data for model development, and microarray-based models can accurately predict RNA-seq-profiled samples; while RNA-seq-based models are less accurate in predicting microarray-profiled samples and are affected both by the choice of modeling algorithm and the gene mapping complexity. The results suggest continued usefulness of legacy microarray data and established microarray biomarkers and predictive models in the forthcoming RNA-seq era.

A digital repository with an extensible data model for biobanking and genomic analysis management

Motivation

Molecular biology laboratories require extensive metadata to improve data collection and analysis. The heterogeneity of the collected metadata grows as research is evolving in to international multi-disciplinary collaborations and increasing data sharing among institutions. Single standardization is not feasible and it becomes crucial to develop digital repositories with flexible and extensible data models, as in the case of modern integrated biobanks management.

Results

We developed a novel data model in JSON format to describe heterogeneous data in a generic biomedical science scenario. The model is built on two hierarchical entities: processes and events, roughly corresponding to research studies and analysis steps within a single study. A number of sequential events can be grouped in a process building up a hierarchical structure to track patient and sample history. Each event can produce new data. Data is described by a set of user-defined metadata, and may have one or more associated files. We integrated the model in a web based digital repository with a data grid storage to manage large data sets located in geographically distinct areas. We built a graphical interface that allows authorized users to define new data types dynamically, according to their requirements. Operators compose queries on metadata fields using a flexible search interface and run them on the database and on the grid. We applied the digital repository to the integrated management of samples, patients and medical history in the BIT-Gaslini biobank. The platform currently manages 1800 samples of over 900 patients. Microarray data from 150 analyses are stored on the grid storage and replicated on two physical resources for preservation. The system is equipped with data integration capabilities with other biobanks for worldwide information sharing.

Conclusions

Our data model enables users to continuously define flexible, ad hoc, and loosely structured metadata, for information sharing in specific research projects and purposes. This approach can improve sensitively interdisciplinary research collaboration and allows to track patients' clinical records, sample management information, and genomic data. The web interface allows the operators to easily manage, query, and annotate the files, without dealing with the technicalities of the data grid.

Use of Attribute Driven Incremental Discretization and Logic Learning Machine to build a prognostic classifier for neuroblastoma patients

BACKGROUND

Cancer patient's outcome is written, in part, in the gene expression profile of the tumor. We previously identified a 62-probe sets signature (NB-hypo) to identify tissue hypoxia in neuroblastoma tumors and showed that NB-hypo stratified neuroblastoma patients in good and poor outcome 1. It was important to develop a prognostic classifier to cluster patients into risk groups benefiting of defined therapeutic approaches. Novel classification and data discretization approaches can be instrumental for the generation of accurate predictors and robust tools for clinical decision support. We explored the application to gene expression data of Rulex, a novel software suite including the Attribute Driven Incremental Discretization technique for transforming continuous variables into simplified discrete ones and the Logic Learning Machine model for intelligible rule generation.

RESULTS

We applied Rulex components to the problem of predicting the outcome of neuroblastoma patients on the bases of 62 probe sets NB-hypo gene expression signature. The resulting classifier consisted in 9 rules utilizing mainly two conditions of the relative expression of 11 probe sets. These rules were very effective predictors, as shown in an independent validation set, demonstrating the validity of the LLM algorithm applied to microarray data and patients' classification. The LLM performed as efficiently as Prediction Analysis of Microarray and Support Vector Machine, and outperformed other learning algorithms such as C4.5. Rulex carried out a feature selection by selecting a new signature (NB-hypo-II) of 11 probe sets that turned out to be the most relevant in predicting outcome among the 62 of the NB-hypo signature. Rules are easily interpretable as they involve only few conditions.

CONCLUSIONS

Our findings provided evidence that the application of Rulex to the expression values of NB-hypo signature created a set of accurate, high quality, consistent and interpretable rules for the prediction of neuroblastoma patients' outcome. We identified the Rulex weighted classification as a flexible tool that can support clinical decisions. For these reasons, we consider Rulex to be a useful tool for cancer classification from microarray gene expression data.

Logic Learning Machine creates explicit and stable rules stratifying neuroblastoma patients

BACKGROUND

Neuroblastoma is the most common pediatric solid tumor. About fifty percent of high risk patients die despite treatment making the exploration of new and more effective strategies for improving stratification mandatory. Hypoxia is a condition of low oxygen tension occurring in poorly vascularized areas of the tumor associated with poor prognosis. We had previously defined a robust gene expression signature measuring the hypoxic component of neuroblastoma tumors (NB-hypo) which is a molecular risk factor. We wanted to develop a prognostic classifier of neuroblastoma patients' outcome blending existing knowledge on clinical and molecular risk factors with the prognostic NB-hypo signature. Furthermore, we were interested in classifiers outputting explicit rules that could be easily translated into the clinical setting.

RESULTS

Shadow Clustering (SC) technique, which leads to final models called Logic Learning Machine (LLM), exhibits a good accuracy and promises to fulfill the aims of the work. We utilized this algorithm to classify NB-patients on the bases of the following risk factors: Age at diagnosis, INSS stage, MYCN amplification and NB-hypo. The algorithm generated explicit classification rules in good agreement with existing clinical knowledge. Through an iterative procedure we identified and removed from the dataset those examples which caused instability in the rules. This workflow generated a stable classifier very accurate in predicting good and poor outcome patients. The good performance of the classifier was validated in an independent dataset. NB-hypo was an important component of the rules with a strength similar to that of tumor staging.

CONCLUSIONS

The novelty of our work is to identify stability, explicit rules and blending of molecular and clinical risk factors as the key features to generate classification rules for NB patients to be conveyed to the clinic and to be used to design new therapies. We derived, through LLM, a set of four stable rules identifying a new class of poor outcome patients that could benefit from new therapies potentially targeting tumor hypoxia or its consequences.

Bioinformatics in Italy: BITS2011, the Eighth Annual Meeting of the Italian Society of Bioinformatics

The BITS2011 meeting, held in Pisa on June 20-22, 2011, brought together more than 120 Italian researchers working in the field of Bioinformatics, as well as students in Bioinformatics, Computational Biology, Biology, Computer Sciences, and Engineering, representing a landscape of Italian bioinformatics research.

This preface provides a brief overview of the meeting and introduces the peer-reviewed manuscripts that were accepted for publication in this Supplement.