Subnetwork-based analysis of chronic lymphocytic leukemia identifies pathways that associate with disease progression

Aberrant spliceosome activity via elevated intron retention and upregulation and phosphorylation of SF3B1 in chronic lymphocytic leukemia

Splicing factor 3b subunit 1 (SF3B1) is the largest subunit and core component of the spliceosome. Inhibition of SF3B1 was associated with an increase in broad intron retention (IR) on most transcripts, suggesting that IR can be used as a marker of spliceosome inhibition in chronic lymphocytic leukemia (CLL) cells. Furthermore, we separately analyzed exonic and intronic mapped reads on annotated RNA-sequencing transcripts obtained from B cells (n = 98 CLL patients) and healthy volunteers (n = 9). We measured intron/exon ratio to use that as a surrogate for alternative RNA splicing (ARS) and found that 66% of CLL-B cell transcripts had significant IR elevation compared with normal B cells (NBCs) and that correlated with mRNA downregulation and low expression levels. Transcripts with the highest IR levels belonged to biological pathways associated with gene expression and RNA splicing. A >2-fold increase of active pSF3B1 was observed in CLL-B cells compared with NBCs. Additionally, when the CLL-B cells were treated with macrolides (pladienolide-B), a significant decrease in pSF3B1, but not total SF3B1 protein, was observed. These findings suggest that IR/ARS is increased in CLL, which is associated with SF3B1 phosphorylation and susceptibility to SF3B1 inhibitors. These data provide additional support to the relevance of ARS in carcinogenesis and evidence of pSF3B1 participation in this process.

Machine learning and multi-omics data in chronic lymphocytic leukemia: the future of precision medicine?

Chronic lymphocytic leukemia is a complex and heterogeneous hematological malignancy. The advance of high-throughput multi-omics technologies has significantly influenced chronic lymphocytic leukemia research and paved the way for precision medicine approaches. In this review, we explore the role of machine learning in the analysis of multi-omics data in this hematological malignancy. We discuss recent literature on different machine learning models applied to single omic studies in chronic lymphocytic leukemia, with a special focus on the potential contributions to precision medicine. Finally, we highlight the recently published machine learning applications in multi-omics data in this area of research as well as their potential and limitations.

Genes selection using deep learning and explainable artificial intelligence for chronic lymphocytic leukemia predicting the need and time to therapy

Analyzing gene expression profiles (GEP) through artificial intelligence provides meaningful insight into cancer disease. This study introduces DeepSHAP Autoencoder Filter for Genes Selection (DSAF-GS), a novel deep learning and explainable artificial intelligence-based approach for feature selection in genomics-scale data. DSAF-GS exploits the autoencoder's reconstruction capabilities without changing the original feature space, enhancing the interpretation of the results. Explainable artificial intelligence is then used to select the informative genes for chronic lymphocytic leukemia prognosis of 217 cases from a GEP database comprising roughly 20,000 genes. The model for prognosis prediction achieved an accuracy of 86.4%, a sensitivity of 85.0%, and a specificity of 87.5%. According to the proposed approach, predictions were strongly influenced by CEACAM19 and PIGP, moderately influenced by MKL1 and GNE, and poorly influenced by other genes. The 10 most influential genes were selected for further analysis. Among them, FADD, FIBP, FIBP, GNE, IGF1R, MKL1, PIGP, and SLC39A6 were identified in the Reactome pathway database as involved in signal transduction, transcription, protein metabolism, immune system, cell cycle, and apoptosis. Moreover, according to the network model of the 3D protein-protein interaction (PPI) explored using the NetworkAnalyst tool, FADD, FIBP, IGF1R, QTRT1, GNE, SLC39A6, and MKL1 appear coupled into a complex network. Finally, all 10 selected genes showed a predictive power on time to first treatment (TTFT) in univariate analyses on a basic prognostic model including IGHV mutational status, del(11q) and del(17p), NOTCH1 mutations, β2-microglobulin, Rai stage, and B-lymphocytosis known to predict TTFT in CLL. However, only IGF1R [hazard ratio (HR) 1.41, 95% CI 1.08-1.84, P=0.013), COL28A1 (HR 0.32, 95% CI 0.10-0.97, P=0.045), and QTRT1 (HR 7.73, 95% CI 2.48-24.04, P<0.001) genes were significantly associated with TTFT in multivariable analyses when combined with the prognostic factors of the basic model, ultimately increasing the Harrell's c-index and the explained variation to 78.6% (versus 76.5% of the basic prognostic model) and 52.6% (versus 42.2% of the basic prognostic model), respectively. Also, the goodness of model fit was enhanced (χ2 = 20.1, P=0.002), indicating its improved performance above the basic prognostic model. In conclusion, DSAF-GS identified a group of significant genes for CLL prognosis, suggesting future directions for bio-molecular research.

METTL3-Mediated m6A Modification Controls Splicing Factor Abundance and Contributes to Aggressive CLL

RNA splicing dysregulation underlies the onset and progression of cancers. In chronic lymphocytic leukemia (CLL), spliceosome mutations leading to aberrant splicing occur in ∼20% of patients. However, the mechanism for splicing defects in spliceosome-unmutated CLL cases remains elusive. Through an integrative transcriptomic and proteomic analysis, we discover that proteins involved in RNA splicing are posttranscriptionally upregulated in CLL cells, resulting in splicing dysregulation. The abundance of splicing complexes is an independent risk factor for poor prognosis. Moreover, increased splicing factor expression is highly correlated with the abundance of METTL3, an RNA methyltransferase that deposits N6-methyladenosine (m6A) on mRNA. METTL3 is essential for cell growth in vitro and in vivo and controls splicing factor protein expression in a methyltransferase-dependent manner through m6A modification-mediated ribosome recycling and decoding. Our results uncover METTL3-mediated m6A modification as a novel regulatory axis in driving splicing dysregulation and contributing to aggressive CLL.

SIGNIFICANCE

METTL3 controls widespread splicing factor abundance via translational control of m6A-modified mRNA, contributes to RNA splicing dysregulation and disease progression in CLL, and serves as a potential therapeutic target in aggressive CLL. See related commentary by Janin and Esteller, p. 176. This article is highlighted in the In This Issue feature, p. 171.

Seven non-differentially expressed 'dark biomarkers' show transcriptional dysregulation in chronic lymphocytic leukemia

Aim: Transcriptional regulation is actively involved in the onset and progression of various diseases. This study used the feature-engineering approach model-based quantitative transcription regulation to quantitatively measure the correlation between mRNA and transcription factors in a reference dataset of chronic lymphocytic leukemia (CLL) transcriptomes. Methods: A comprehensive investigation of transcriptional regulation changes in CLL was conducted using 973 samples in six independent datasets. Results & conclusion: Seven mRNAs were detected to have significantly differential model-based quantitative transcription regulation values but no differential expression between CLL patients and controls. We called these genes 'dark biomarkers' because their original expression levels did not show differential changes in the CLL patients. The overlapping lncRNAs might have contributed their transcripts to the expression miscalculations of these dark biomarkers.

Biological relevance of ZNF224 expression in chronic lymphocytic leukemia and its implication IN NF-kB pathway regulation

Chronic lymphocytic leukemia (CLL) is a heterogeneous disease, whose presentation and clinical course are highly variable. Identification of novel prognostic factors may contribute to improving the CLL classification and providing indications for treatment options. The zinc finger protein ZNF224 plays a key role in cell transformation, through the control of apoptotic and survival pathways. In this study, we evaluated the potential application of ZNF224 as a novel marker of CLL progression and therapy responsiveness. To this aim, we analyzed ZNF224 expression levels in B lymphocytes from CLL patients at different stages of the disease and in patients showing different treatment outcomes. The expression of ZNF224 was significantly increased in disease progression and dramatically decreased in patients in complete remission after chemotherapy. Gene expression correlation analysis performed on datasets of CLL patients revealed that ZNF224 expression was well correlated with that of some prognostic and predictive markers. Moreover, bioinformatic analysis coupled ZNF224 to NF-κB pathway, and experimental data demonstrated that RNA interference of ZNF224 reduced the activity of the NF-κB survival pathway in CLL cells. Consistently with a pro-survival role, ZNF224 knockdown raised spontaneous and drug-induced apoptosis and inhibited the proliferation of peripheral blood mononuclear cells from CLL patients. Our findings provide evidence for the involvement of ZNF224 in the survival of CLL cells via NF-κB pathway modulation, and also suggest ZNF224 as a prognostic and predictive molecular marker of CLL disease.

Enhanced IFN Sensing by Aggressive Chronic Lymphocytic Leukemia Cells

Type I IFN is made by cells in response to stress. Cancer cells exist in a state of stress, but their IFN response is complex and not completely understood. This study investigated the role of autocrine IFN in human chronic lymphocytic leukemia (CLL) cells. CLL cells were found to make low amounts of IFN via TANK-binding kinase 1 pathways, but p-STAT1 and -STAT2 proteins along with IFN-stimulated genes that reflect IFN activation were variably downregulated in cultured CLL cells by the neutralizing IFNAR1 Ab anifrolumab. Patients with CLL were segregated into two groups based on the response of their leukemia cells to anifrolumab. Samples associated with more aggressive clinical behavior indicated by unmutated IGHV genes along with high CD38 and p-Bruton’s tyrosine kinase expression exhibited responses to low amounts of IFN that were blocked by anifrolumab. Samples with more indolent behavior were unaffected by anifrolumab. Hypersensitivity to IFN was associated with higher expression of IFNAR1, MX1, STAT1, and STAT2 proteins and lower activity of negative regulatory tyrosine phosphatases. Autocrine IFN protected responsive CLL cells from stressful tissue culture environments and therapeutic drugs such as ibrutinib and venetoclax in vitro, in part by upregulating Mcl-1 expression. These findings suggest hypersensitivity to IFN may promote aggressive clinical behavior. Specific blockade of IFN signaling may improve outcomes for patients with CLL with higher-risk disease.

Decreased TCF1 and BCL11B expression predicts poor prognosis for patients with chronic lymphocytic leukemia

T cell immune dysfunction is a prominent characteristic of chronic lymphocytic leukemia (CLL) and the main cause of failure for immunotherapy and multi-drug resistance. There remains a lack of specific biomarkers for evaluating T cell immune status with outcome for CLL patients. T cell factor 1 (TCF1, encoded by the TCF7 gene) can be used as a critical determinant of successful anti-tumor immunotherapy and a prognostic indicator in some solid tumors; however, the effects of TCF1 in CLL remain unclear. Here, we first analyzed the biological processes and functions of TCF1 and co-expressing genes using the GEO and STRING databases with the online tools Venny, Circos, and Database for Annotation, Visualization, and Integrated Discovery (DAVID). Then the expression and prognostic values of TCF1 and its partner gene B cell leukemia/lymphoma 11B (BCL11B) were explored for 505 CLL patients from 6 datasets and validated with 50 CLL patients from Henan cancer hospital (HNCH). TCF1 was downregulated in CLL patients, particularly in CD8+ T cells, which was significantly correlated with poor time-to-first treatment (TTFT) and overall survival (OS) as well as short restricted mean survival time (RMST). Function and pathway enrichment analysis revealed that TCF1 was positively correlated with BCL11B, which is involved in regulating the activation and differentiation of T cells in CLL patients. Intriguingly, BCL11B was highly consistent with TCF1 in its decreased expression and prediction of poor prognosis. More importantly, the combination of TCF1 and BCL11B could more accurately assess prognosis than either alone. Additionally, decreased TCF1 and BCL11B expression serves as an independent risk factor for rapid disease progression, coinciding with high-risk indicators, including unmutated IGHV, TP53 alteration, and advanced disease. Altogether, this study demonstrates that decreased TCF1 and BCL11B expression is significantly correlated with poor prognosis, which may be due to decreased TCF1+CD8+ T cells, impairing the effector CD8+ T cell differentiation regulated by TCF1/BCL11B.

Super-Enhancer–Associated nine-gene prognostic score model for prediction of survival in chronic lymphocytic leukemia patients

Chronic lymphocytic leukemia (CLL) is a type of highly heterogeneous mature B-cell malignancy with various disease courses. Although a multitude of prognostic markers in CLL have been reported, insights into the role of super-enhancer (SE)–related risk indicators in the occurrence and development of CLL are still lacking. A super-enhancer (SE) is a cluster of enhancers involved in cell differentiation and tumorigenesis, and is one of the promising therapeutic targets for cancer therapy in recent years. In our study, the CLL-related super-enhancers in the training database were processed by LASSO-penalized Cox regression analysis to screen a nine-gene prognostic model including TCF7, VEGFA, MNT, GMIP, SLAMF1, TNFRSF25, GRWD1, SLC6AC, and LAG3. The SE-related risk score was further constructed and it was found that the predictive performance with overall survival and time-to-treatment (TTT) was satisfactory. Moreover, a high correlation was found between the risk score and already known prognostic markers of CLL. In the meantime, we noticed that the expressions of TCF7, GMIP, SLAMF1, TNFRSF25, and LAG3 in CLL were different from those of healthy donors (p &lt; 0.01). Moreover, the risk score and LAG3 level of matched pairs before and after treatment samples varied significantly. Finally, an interactive nomogram consisting of the nine-gene risk group and four clinical traits was established. The inhibitors of mTOR and cyclin-dependent kinases (CDKs) were considered effective in patients in the high-risk group according to the pRRophetic algorithm. Collectively, the SE-associated nine-gene prognostic model developed here may be used to predict the prognosis and assist in the risk stratification and treatment of CLL patients in the future.

The pyroptosis-related gene signature predicts prognosis and indicates the immune microenvironment status of chronic lymphocytic leukemia

Chronic lymphocytic leukemia (CLL) is the most common leukemia in the Western world with great heterogeneity. Pyroptosis has recently been recognized as an inflammatory form of programmed cell death (PCD) and shares a close relationship with apoptosis. Although the role of apoptosis in CLL was comprehensively studied and successfully applied in clinical treatment, the relationship between pyroptosis genes and CLL remained largely unknown. In this study, eight differentially expressed pyroptosis-related genes (PRGs) were identified between CLL and normal B cells. In order to screen out the prognostic value of differentially expressed PRGs, univariate and multivariate Cox regression analyses were conducted and a risk model with three PRG signatures (GSDME, NLRP3, and PLCG1) was constructed. All CLL samples were stratified into high- and low-risk subgroups according to risk scores. The risk model showed high efficacy in predicting both overall survival (OS) and time to first treatment (TTFT). Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG), and Gene Set Enrichment Analysis (GSEA) showed the dysregulation of immune and inflammatory response in the high-risk group. Single-sample GSEA (ssGSEA) of immune cell infiltration and the activity of immune-related pathways also displayed decreased antitumor immunity in the high-risk group. In conclusion, PRGs are of prognostic value in CLL and may play important roles in tumor immunity, and the underlying relationship between PRGs and CLL needs to be explored further.

Improving Analysis and Annotation of Microarray Data with Protein Interactions

Gene expression microarrays are one of the most widely used high-throughput technologies in molecular biology, with applications such as identification of disease mechanisms and development of diagnostic and prognostic gene signatures. However, the success of these tasks is often limited because microarray analysis does not account for the complex relationships among genes, their products, and overall signaling and regulatory cascades. Incorporating protein-protein interaction data into microarray analysis can help address these challenges. This chapter reviews how protein-protein interactions can help with microarray analysis, leading to benefits such as better explanations of disease mechanisms, more complete gene annotations, improved prioritization of genes for future experiments, and gene signatures that generalize better to new data.

Multi-omics reveals clinically relevant proliferative drive associated with mTOR-MYC-OXPHOS activity in chronic lymphocytic leukemia

Chronic Lymphocytic Leukemia (CLL) has a complex pattern of driver mutations and much of its clinical diversity remains unexplained. We devised a method for simultaneous subgroup discovery across multiple data types and applied it to genomic, transcriptomic, DNA methylation and ex-vivo drug response data from 217 Chronic Lymphocytic Leukemia (CLL) cases. We uncovered a biological axis of heterogeneity strongly associated with clinical behavior and orthogonal to the known biomarkers. We validated its presence and clinical relevance in four independent cohorts (n=547 patients). We find that this axis captures the proliferative drive (PD) of CLL cells, as it associates with lymphocyte doubling rate, global hypomethylation, accumulation of driver aberrations and response to pro-proliferative stimuli. CLL-PD was linked to the activation of mTOR-MYC-oxidative phosphorylation (OXPHOS) through transcriptomic, proteomic and single cell resolution analysis. CLL-PD is a key determinant of disease outcome in CLL. Our multi-table integration approach may be applicable to other tumors whose inter-individual differences are currently unexplained.

HELQ and EGR3 expression correlate with IGHV mutation status and prognosis in chronic lymphocytic leukemia

Background

IGHV mutation status is a crucial prognostic biomarker for CLL. In the present study, we investigated the transcriptomic signatures associating with IGHV mutation status and CLL prognosis.

Methods

The co-expression modules and hub genes correlating with IGHV status, were identified using the GSE28654, by ‘WGCNA’ package and R software (version 4.0.2). The over-representation analysis was performed to reveal enriched cell pathways for genes of correlating modules. Then 9 external cohorts were used to validate the correlation of hub genes expression with IGHV status or clinical features (treatment response, transformation to Richter syndrome, etc.). Moreover, to elucidate the significance of hub genes on disease course and prognosis of CLL patients, the Kaplan–Meier analysis for the OS and TTFT of were performed between subgroups dichotomized by the median expression value of individual hub genes.

Results

2 co-expression modules and 9 hub genes ((FCRL1/FCRL2/HELQ/EGR3LPL/LDOC1/ZNF667/SOWAHC/SEPTIN10) correlating with IGHV status were identified by WGCNA, and validated by external datasets. The modules were found to be enriched in NF-kappaB, HIF-1 and other important pathways, involving cell proliferation and apoptosis. The expression of hub genes was revealed to be significantly different, not only between CLL and normal B cell, but also between various types of lymphoid neoplasms. HELQ expression was found to be related with response of immunochemotherapy treatment significantly (p = 0.0413), while HELQ and ZNF667 were expressed differently between stable CLL and Richter syndrome patients (p < 0.0001 and p = 0.0278, respectively). By survival analysis of subgroups, EGR3 expression was indicated to be significantly associated with TTFT by 2 independent cohorts (GSE39671, p = 0.0311; GSE22762, p = 0.0135). While the expression of HELQ and EGR3 was found to be associated with OS (p = 0.0291 and 0.0114 respectively).The Kras, Hedgehog and IL6-JAK-STAT3 pathways were found to be associating with the expression of hub genes, resulting from GSEA.

Conclusions

The expression of HELQ and EGR3 were correlated with IGHV mutation status in CLL patients. Additionally, the expression of HELQ/EGR3 were prognostic markers for CLL associating with targetable cell signaling pathways.

Finding disease modules for cancer and COVID-19 in gene co-expression networks with the Core&Peel method

Genes are organized in functional modules (or pathways), thus their action and their dysregulation in diseases may be better understood by the identification of the modules most affected by the disease (aka disease modules, or active subnetworks). We describe how an algorithm based on the Core&Peel method is used to detect disease modules in co-expression networks of genes. We first validate Core&Peel for the general task of functional module detection by comparison with 42 methods participating in the Disease Module Identification DREAM challenge. Next, we use four specific disease test cases (colorectal cancer, prostate cancer, asthma, and rheumatoid arthritis), four state-of-the-art algorithms (ModuleDiscoverer, Degas, KeyPathwayMiner, and ClustEx), and several pathway databases to validate the proposed algorithm. Core&Peel is the only method able to find significant associations of the predicted disease module with known validated relevant pathways for all four diseases. Moreover, for the two cancer datasets, Core&Peel detects further eight relevant pathways not discovered by the other methods used in the comparative analysis. Finally, we apply Core&Peel and other methods to explore the transcriptional response of human cells to SARS-CoV-2 infection, finding supporting evidence for drug repositioning efforts at a pre-clinical level.

Prognostic models for newly-diagnosed chronic lymphocytic leukaemia in adults: a systematic review and meta-analysis

BACKGROUND

Chronic lymphocytic leukaemia (CLL) is the most common cancer of the lymphatic system in Western countries. Several clinical and biological factors for CLL have been identified. However, it remains unclear which of the available prognostic models combining those factors can be used in clinical practice to predict long-term outcome in people newly-diagnosed with CLL.

OBJECTIVES

To identify, describe and appraise all prognostic models developed to predict overall survival (OS), progression-free survival (PFS) or treatment-free survival (TFS) in newly-diagnosed (previously untreated) adults with CLL, and meta-analyse their predictive performances.

SEARCH METHODS

We searched MEDLINE (from January 1950 to June 2019 via Ovid), Embase (from 1974 to June 2019) and registries of ongoing trials (to 5 March 2020) for development and validation studies of prognostic models for untreated adults with CLL. In addition, we screened the reference lists and citation indices of included studies.

SELECTION CRITERIA

We included all prognostic models developed for CLL which predict OS, PFS, or TFS, provided they combined prognostic factors known before treatment initiation, and any studies that tested the performance of these models in individuals other than the ones included in model development (i.e. 'external model validation studies'). We included studies of adults with confirmed B-cell CLL who had not received treatment prior to the start of the study. We did not restrict the search based on study design.

DATA COLLECTION AND ANALYSIS

We developed a data extraction form to collect information based on the Checklist for Critical Appraisal and Data Extraction for Systematic Reviews of Prediction Modelling Studies (CHARMS). Independent pairs of review authors screened references, extracted data and assessed risk of bias according to the Prediction model Risk Of Bias ASsessment Tool (PROBAST). For models that were externally validated at least three times, we aimed to perform a quantitative meta-analysis of their predictive performance, notably their calibration (proportion of people predicted to experience the outcome who do so) and discrimination (ability to differentiate between people with and without the event) using a random-effects model. When a model categorised individuals into risk categories, we pooled outcome frequencies per risk group (low, intermediate, high and very high). We did not apply GRADE as guidance is not yet available for reviews of prognostic models.

MAIN RESULTS

From 52 eligible studies, we identified 12 externally validated models: six were developed for OS, one for PFS and five for TFS. In general, reporting of the studies was poor, especially predictive performance measures for calibration and discrimination; but also basic information, such as eligibility criteria and the recruitment period of participants was often missing. We rated almost all studies at high or unclear risk of bias according to PROBAST. Overall, the applicability of the models and their validation studies was low or unclear; the most common reasons were inappropriate handling of missing data and serious reporting deficiencies concerning eligibility criteria, recruitment period, observation time and prediction performance measures. We report the results for three models predicting OS, which had available data from more than three external validation studies: CLL International Prognostic Index (CLL-IPI) This score includes five prognostic factors: age, clinical stage, IgHV mutational status, B2-microglobulin and TP53 status. Calibration: for the low-, intermediate- and high-risk groups, the pooled five-year survival per risk group from validation studies corresponded to the frequencies observed in the model development study. In the very high-risk group, predicted survival from CLL-IPI was lower than observed from external validation studies. Discrimination: the pooled c-statistic of seven external validation studies (3307 participants, 917 events) was 0.72 (95% confidence interval (CI) 0.67 to 0.77). The 95% prediction interval (PI) of this model for the c-statistic, which describes the expected interval for the model's discriminative ability in a new external validation study, ranged from 0.59 to 0.83. Barcelona-Brno score Aimed at simplifying the CLL-IPI, this score includes three prognostic factors: IgHV mutational status, del(17p) and del(11q). Calibration: for the low- and intermediate-risk group, the pooled survival per risk group corresponded to the frequencies observed in the model development study, although the score seems to overestimate survival for the high-risk group. Discrimination: the pooled c-statistic of four external validation studies (1755 participants, 416 events) was 0.64 (95% CI 0.60 to 0.67); 95% PI 0.59 to 0.68. MDACC 2007 index score The authors presented two versions of this model including six prognostic factors to predict OS: age, B2-microglobulin, absolute lymphocyte count, gender, clinical stage and number of nodal groups. Only one validation study was available for the more comprehensive version of the model, a formula with a nomogram, while seven studies (5127 participants, 994 events) validated the simplified version of the model, the index score. Calibration: for the low- and intermediate-risk groups, the pooled survival per risk group corresponded to the frequencies observed in the model development study, although the score seems to overestimate survival for the high-risk group. Discrimination: the pooled c-statistic of the seven external validation studies for the index score was 0.65 (95% CI 0.60 to 0.70); 95% PI 0.51 to 0.77.

AUTHORS' CONCLUSIONS

Despite the large number of published studies of prognostic models for OS, PFS or TFS for newly-diagnosed, untreated adults with CLL, only a minority of these (N = 12) have been externally validated for their respective primary outcome. Three models have undergone sufficient external validation to enable meta-analysis of the model's ability to predict survival outcomes. Lack of reporting prevented us from summarising calibration as recommended. Of the three models, the CLL-IPI shows the best discrimination, despite overestimation. However, performance of the models may change for individuals with CLL who receive improved treatment options, as the models included in this review were tested mostly on retrospective cohorts receiving a traditional treatment regimen. In conclusion, this review shows a clear need to improve the conducting and reporting of both prognostic model development and external validation studies. For prognostic models to be used as tools in clinical practice, the development of the models (and their subsequent validation studies) should adapt to include the latest therapy options to accurately predict performance. Adaptations should be timely.

Biomolecular Databases and Subnetwork Identification Approaches of Interest to Big Data Community: An Expert Review

Next-generation sequencing approaches and genome-wide studies have become essential for characterizing the mechanisms of human diseases. Consequently, many researchers have applied these approaches to discover the genetic/genomic causes of common complex and rare human diseases, generating multiomics big data that span the continuum of genomics, proteomics, metabolomics, and many other system science fields. Therefore, there is a significant and unmet need for biological databases and tools that enable and empower the researchers to analyze, integrate, and make sense of big data. There are currently large number of databases that offer different types of biological information. In particular, the integration of gene expression profiles and protein-protein interaction networks provides a deeper understanding of the complex multilayered molecular architecture of human diseases. Therefore, there has been a growing interest in developing methodologies that integrate and contextualize big data from molecular interaction networks to identify biomarkers of human diseases at a subnetwork resolution as well. In this expert review, we provide a comprehensive summary of most popular biomolecular databases for molecular interactions (e.g., Biological General Repository for Interaction Datasets, Kyoto Encyclopedia of Genes and Genomes and Search Tool for The Retrieval of Interacting Genes/Proteins), gene-disease associations (e.g., Online Mendelian Inheritance in Man, Disease-Gene Network, MalaCards), and population-specific databases (e.g., Human Genetic Variation Database), and describe some examples of their usage and potential applications. We also present the most recent subnetwork identification approaches and discuss their main advantages and limitations. As the field of data science continues to emerge, the present analysis offers a deeper and contextualized understanding of the available databases in molecular biomedicine.

B-cell-specific IRF4 deletion accelerates chronic lymphocytic leukemia development by enhanced tumor immune evasion

Chronic lymphocytic leukemia (CLL) is a heterogenous disease that is highly dependent on a cross talk of CLL cells with the microenvironment, in particular with T cells. T cells derived from CLL patients or murine CLL models are skewed to an antigen-experienced T-cell subset, indicating a certain degree of antitumor recognition, but they are also exhausted, preventing an effective antitumor immune response. Here we describe a novel mechanism of CLL tumor immune evasion that is independent of T-cell exhaustion, using B-cell-specific deletion of the transcription factor IRF4 (interferon regulatory factor 4) in Tcl-1 transgenic mice developing a murine CLL highly similar to the human disease. We show enhanced CLL disease progression in IRF4-deficient Tcl-1 tg mice, associated with a severe downregulation of genes involved in T-cell activation, including genes involved in antigen processing/presentation and T-cell costimulation, which massively reduced T-cell subset skewing and exhaustion. We found a strong analogy in the human disease, with inferior prognosis of CLL patients with low IRF4 expression in independent CLL patient cohorts, failed T-cell skewing to antigen-experienced subsets, decreased costimulation capacity, and downregulation of genes involved in T-cell activation. These results have therapeutic relevance because our findings on molecular mechanisms of immune privilege may be responsible for the failure of immune-therapeutic strategies in CLL and may lead to improved targeting in the future.

Transcriptomics-based validation of the relatedness of heterogeneous nuclear ribonucleoproteins to chronic lymphocytic leukemia as potential biomarkers of the disease aggressiveness

OBJECTIVES

To use independent transcriptomics data sets of cancer patients with prognostic information from public repositories to validate the relevance of our previously described chronic lymphocytic leukemia (CLL)-related proteins at the level of transcription (mRNA) to the prognosis of CLL.  Methods: This is a validation study that was conducted at Majmaah University, Kingdom of Saudi Arabia between January-2017 and July-2018. Two independent data sets of CLL transcriptomics from Gene Expression Omnibus (GEO) with time-to-first treatment (TTFT) data (GSE39671; 130 patients) and information about overall survival (OS) (GSE22762; 107 patients) were used for the validation analyses. To further investigate the relatedness of a transcript of interest to other neoplasms, 6 independent data sets of cancer transcriptomics with prognostic information (1865 patients) from the cancer genomics atlas (TCGA) were used. Pathway-enrichment analyses were conducted using Reactome; and correlation analyses of gene expression were performed using Pearson score. Results: Nine of the CLL-related proteins exhibited transcript expression that predicted TTFT and 7 of the CLL-related proteins showed mRNA levels that predicted OS in CLL patients (p≤0.05). Of these transcripts, 8 were different types of heterogeneous nuclear ribonucleoproteins (HNRNPs); and 2 (HNRNPUL2 and HIST1C1H) retained prognostic significance in the 2 independent data sets. Furthermore, genes that enriched CLL-related pathways (p≤0.05; false discovery rate [FDR] ≤0.05) were found to correlate with the expression of HNRNPUL2 (Pearson score: ≥0.50; p lessthan 0.00001). Finally, increased expression of HNRNPUL2 was indicative of poor prognosis of various types of cancer other than CLL (p less than 0.05). Conclusion: The cognate transcripts of 14 of our CLL-related proteins significantly predicted CLL prognosis.

Investigating epigenetic effects of activation-induced deaminase in chronic lymphocytic leukemia

Activation induced deaminase (AID) has two distinct and well defined roles, both relying on its deoxycytidine (dC) deaminating function: one as a DNA mutator and another in DNA demethylation. In chronic lymphocytic leukemia (CLL), AID was previously shown to be an independent negative prognostic factor. While there is substantial impact on DNA mutations, effects of AID on gene expression by promoter demethylation of disease related target genes in leukemia has not been addressed. To shed light on this question, we aimed at determining genome wide methylation changes as well as gene expression changes in response to AID expression in CLL. Although we found minor differences in individual methylation variable positions following AID expression, we could not find recurrent methylation changes of specific target sites or changes in global methylation.

Inhibition of maternal embryonic leucine zipper kinase with OTSSP167 displays potent anti-leukemic effects in chronic lymphocytic leukemia

TP53 pathway defects contributed to therapy resistance and adverse clinical outcome in chronic lymphocytic leukemia (CLL), which represents an unmet clinical need with few therapeutic options. Maternal embryonic leucine zipper kinase (MELK) is a novel oncogene, which plays crucial roles in mitotic progression and stem cell maintenance. OTSSP167, an orally administrated inhibitor targeting MELK, is currently in a phase I/II clinical trial in patients with advanced breast cancer and acute myeloid leukemia. Yet, no investigation has been elucidated to date regarding the oncogenic role of MELK and effects of OTSSP167 in chronic lymphocytic leukemia (CLL). Previous studies confirmed MELK inhibition abrogated cancer cell survival via p53 signaling pathway. Thus, we aimed to determine the biological function of MELK and therapeutic potential of OTSSP167 in CLL. Herein, MELK over-expression was observed in CLL cells, and correlated with higher WBC count, advanced stage, elevated LDH, increased β2-MG level, unmutated IGHV, positive ZAP-70, deletion of 17p13 and inferior prognosis of CLL patients. In accordance with functional enrichment analyses in gene expression profiling, CLL cells with depletion or inhibition of MELK exhibited impaired cell proliferation, enhanced fast-onset apoptosis, induced G2/M arrest, attenuated cell chemotaxis and promoted sensitivity to fludarabine and ibrutinib. However, gain-of-function assay showed increased cell proliferation and cell chemotaxis. In addition, OTSSP167 treatment reduced phosphorylation of AKT and ERK1/2. It decreased FoxM1 phosphorylation, expression of FoxM1, cyclin B1 and CDK1, while up-regulating p53 and p21 expression. Taken together, MELK served as a candidate of therapeutic target in CLL. OTSSP167 exhibits potent anti-tumor activities in CLL cells, highlighting a novel molecule-based strategy for leukemic interventions.

Subtype assignment of CLL based on B-cell subset associated gene signatures from normal bone marrow - A proof of concept study

Diagnostic and prognostic evaluation of chronic lymphocytic leukemia (CLL) involves blood cell counts, immunophenotyping, IgVH mutation status, and cytogenetic analyses. We generated B-cell associated gene-signatures (BAGS) based on six naturally occurring B-cell subsets within normal bone marrow. Our hypothesis is that by segregating CLL according to BAGS, we can identify subtypes with prognostic implications in support of pathogenetic value of BAGS. Microarray-based gene-expression samples from eight independent CLL cohorts (1,024 untreated patients) were BAGS-stratified into pre-BI, pre-BII, immature, naïve, memory, or plasma cell subtypes; the majority falling within the memory (24.5-45.8%) or naïve (14.5-32.3%) categories. For a subset of CLL patients (n = 296), time to treatment (TTT) was shorter amongst early differentiation subtypes (pre-BI/pre-BII/immature) compared to late subtypes (memory/plasma cell, HR: 0.53 [0.35-0.78]). Particularly, pre-BII subtype patients had the shortest TTT among all subtypes. Correlates derived for BAGS subtype and IgVH mutation (n = 405) revealed an elevated mutation frequency in late vs. early subtypes (71% vs. 45%, P < .001). Predictions for BAGS subtype resistance towards rituximab and cyclophosphamide varied for rituximab, whereas all subtypes were sensitive to cyclophosphamide. This study supports our hypothesis that BAGS-subtyping may be of tangible prognostic and pathogenetic value for CLL patients.

Alterations in DNA methylation/demethylation intermediates predict clinical outcome in chronic lymphocytic leukemia

Cytosine derivative dysregulations represent important epigenetic modifications whose impact on the clinical outcome in chronic lymphocytic leukemia (CLL) is incompletely understood. Hence, global levels of 5-methylcytosine (5-mCyt), 5-hydroxymethylcytosine (5-hmCyt), 5-carboxylcytosine (5-CaCyt) and 5-hydroxymethyluracil were tested in purified B cells from CLL patients (n = 55) and controls (n = 17). The DNA methylation 'writers' (DNA methyltransferases [DNMT1/3A/3B]), 'readers' (methyl-CpG-binding domain [MBD2/4]), 'editors' (ten-eleven translocation [TET1/2/3]) and 'modulators' (SAT1) were also evaluated. Accordingly, patients were stratified into three subgroups. First, a subgroup with a global deficit in cytosine derivatives characterized by hyperlymphocytosis, reduced median progression free survival (PFS = 52 months) and shorter treatment free survival (TFS = 112 months) was identified. In this subgroup, major epigenetic modifications were highlighted including a reduction of 5-mCyt, 5-hmCyt, 5-CaCyt associated with DNMT3A, MBD2/4 and TET1/2 downregulation. Second, the cytosine derivative analysis revealed a subgroup with a partial deficit (PFS = 84, TFS = 120 months), mainly affecting DNA demethylation (5-hmCyt reduction, SAT1 induction). Third, a subgroup epigenetically similar to controls was identified (PFS and TFS > 120 months). The prognostic impact of stratifying CLL patients within three epigenetic subgroups was confirmed in a validation cohort. In conclusion, our results suggest that dysregulations of cytosine derivative regulators represent major events acquired during CLL progression and are independent from IGHV mutational status.

High-level ROR1 associates with accelerated disease progression in chronic lymphocytic leukemia

ROR1 is an oncoembryonic orphan receptor found on chronic lymphocytic leukemia (CLL) B cells, but not on normal postpartum tissues. ROR1 is a receptor for Wnt5a that may complex with TCL1, a coactivator of AKT that is able to promote development of CLL. We found the CLL cells of a few patients expressed negligible ROR1 (ROR1^Neg), but expressed TCL1A at levels comparable to those of samples that expressed ROR1 (ROR1^Pos). Transcriptome analyses revealed that ROR1^Neg cases generally could be distinguished from those that were ROR1^Pos in unsupervised gene-expression clustering analysis. Gene-set enrichment analyses demonstrated that ROR1^Neg CLL had lower expression and activation of AKT signaling pathways relative to ROR1^Pos CLL, similar to what was noted for leukemia that respectively developed in TCL1 vs ROR1xTCL1 transgenic mice. In contrast to its effect on ROR1^Pos CLL, Wnt5a did not enhance the proliferation, chemotaxis, or survival of ROR1^Neg CLL. We examined the CLL cells from 1568 patients, which we randomly assigned to a training or validation set of 797 or 771 cases, respectively. Using recursive partitioning, we defined a threshold for ROR1 surface expression that could segregate samples of the training set into ROR1-Hi vs ROR1-Lo subgroups that differed significantly in their median treatment-free survival (TFS). Using this threshold, we found that ROR1-Hi cases had a significantly shorter median TFS and overall survival than ROR1-Lo cases in the validation set. These data demonstrate that expression of ROR1 may promote leukemia-cell activation and survival and enhance disease progression in patients with CLL.

Network-based pathway enrichment analysis with incomplete network information

MOTIVATION

Pathway enrichment analysis has become a key tool for biomedical researchers to gain insight into the underlying biology of differentially expressed genes, proteins and metabolites. It reduces complexity and provides a system-level view of changes in cellular activity in response to treatments and/or in disease states. Methods that use existing pathway network information have been shown to outperform simpler methods that only take into account pathway membership. However, despite significant progress in understanding the association amongst members of biological pathways, and expansion of data bases containing information about interactions of biomolecules, the existing network information may be incomplete or inaccurate and is not cell-type or disease condition-specific.

RESULTS

We propose a constrained network estimation framework that combines network estimation based on cell- and condition-specific high-dimensional Omics data with interaction information from existing data bases. The resulting pathway topology information is subsequently used to provide a framework for simultaneous testing of differences in expression levels of pathway members, as well as their interactions. We study the asymptotic properties of the proposed network estimator and the test for pathway enrichment, and investigate its small sample performance in simulated and real data settings.

AVAILABILITY AND IMPLEMENTATION

The proposed method has been implemented in the R-package netgsa available on CRAN.

CONTACT

jinma@upenn.eduSupplementary information: Supplementary data are available at Bioinformatics online.

Big data to smart data in Alzheimer's disease: Real-world examples of advanced modeling and simulation

Many disease-modifying clinical development programs in Alzheimer's disease (AD) have failed to date, and development of new and advanced preclinical models that generate actionable knowledge is desperately needed. This review reports on computer-based modeling and simulation approach as a powerful tool in AD research. Statistical data-analysis techniques can identify associations between certain data and phenotypes, such as diagnosis or disease progression. Other approaches integrate domain expertise in a formalized mathematical way to understand how specific components of pathology integrate into complex brain networks. Private-public partnerships focused on data sharing, causal inference and pathway-based analysis, crowdsourcing, and mechanism-based quantitative systems modeling represent successful real-world modeling examples with substantial impact on CNS diseases. Similar to other disease indications, successful real-world examples of advanced simulation can generate actionable support of drug discovery and development in AD, illustrating the value that can be generated for different stakeholders.

Identifying network-based biomarkers of complex diseases from high-throughput data

In this work, we review the main available computational methods of identifying biomarkers of complex diseases from high-throughput data. The emerging omics techniques provide powerful alternatives to measure thousands of molecules in cells in parallel manners. The generated genomic, transcriptomic, proteomic, metabolomic and phenomic data provide comprehensive molecular and cellular information for detecting critical signals served as biomarkers by classifying disease phenotypic states. Networks are often employed to organize these profiles in the identification of biomarkers to deal with complex diseases in diagnosis, prognosis and therapy as well as mechanism deciphering from systematic perspectives. Here, we summarize some representative network-based bioinformatics methods in order to highlight the importance of computational strategies in biomarker discovery.

A network-based method for identifying prognostic gene modules in lung squamous carcinoma

Similarities in gene expression between both developing embryonic and precancerous tissues and cancer tissues may help identify much-needed biomarkers and therapeutic targets in lung squamous carcinoma. In this study, human lung samples representing ten successive time points, from embryonic development to carcinogenesis, were used to construct global gene expression profiles. Differentially expressed genes with similar expression in precancerous and cancer samples were identified. Using a network-based greedy searching algorithm to analyze the training cohort (n = 69) and three independent testing cohorts, we successfully identified a significant 22-gene module in which expression levels were correlated with overall survival in lung squamous carcinoma patients.

The Dohner fluorescence in situ hybridization prognostic classification of chronic lymphocytic leukaemia (CLL): the CLL Research Consortium experience

This study revisited the Dohner prognostic hierarchy in a cohort of 1585 well-documented patients with chronic lymphocytic leukaemia. The duration of both time to first treatment (TTFT) and overall survival (OS) were significantly longer than observed previously, and this is at least partly due to improved therapeutic options. Deletion 13q remains the most favourable prognostic group with median TTFT and OS from fluorescence in situ hybridization (FISH) testing of 72 months and >12 years, respectively. Deletion 11q had the poorest median TTFT (22 months) and 17p deletion the poorest median OS (5 years). The percentages of abnormal nuclei were significantly associated with differential TTFT for the trisomy 12, 13q and 17p deletion cohorts but not for the 11q deletion cohort. From the date of the first FISH study, patients with >85% 13q deletion nuclei had a notably shorter TTFT (24 months). Patients with ≤20% 17p deletion nuclei had longer median TTFT and OS from the date of the first FISH study (44 months and 11 years), and were more likely to be IGHV mutated.

Drugs that reverse disease transcriptomic signatures are more effective in a mouse model of dyslipidemia

High-throughput omics have proven invaluable in studying human disease, and yet day-to-day clinical practice still relies on physiological, non-omic markers. The metabolic syndrome, for example, is diagnosed and monitored by blood and urine indices such as blood cholesterol levels. Nevertheless, the association between the molecular and the physiological manifestations of the disease, especially in response to treatment, has not been investigated in a systematic manner. To this end, we studied a mouse model of diet-induced dyslipidemia and atherosclerosis that was subject to various drug treatments relevant to the disease in question. Both physiological data and gene expression data (from the liver and white adipose) were analyzed and compared. We find that treatments that restore gene expression patterns to their norm are associated with the successful restoration of physiological markers to their baselines. This holds in a tissue-specific manner—treatments that reverse the transcriptomic signatures of the disease in a particular tissue are associated with positive physiological effects in that tissue. Further, treatments that introduce large non-restorative gene expression alterations are associated with unfavorable physiological outcomes. These results provide a sound basis to in silico methods that rely on omic metrics for drug repurposing and drug discovery by searching for compounds that reverse a disease's omic signatures. Moreover, they highlight the need to develop drugs that restore the global cellular state to its healthy norm rather than rectify particular disease phenotypes.

Trends in survival of chronic lymphocytic leukemia patients in Germany and the USA in the first decade of the twenty-first century

Background

Recent population-based studies in the United States of America (USA) and other countries have shown improvements in survival for patients with chronic lymphocytic leukemia (CLL) diagnosed in the early twenty-first century. Here, we examine the survival for patients diagnosed with CLL in Germany in 1997–2011.

Methods

Data were extracted from 12 cancer registries in Germany and compared to the data from the USA. Period analysis was used to estimate 5- and 10-year relative survival (RS).

Results

Five- and 10-year RS estimates in 2009–2011 of 80.2 and 59.5 %, respectively, in Germany and 82.4 and 64.7 %, respectively, in the USA were observed. Overall, 5-year RS increased significantly in Germany and the difference compared to the survival in the USA which slightly decreased between 2003–2005 and 2009–2011. However, age-specific analyses showed persistently higher survival for all ages except for 15–44 in the USA. In general, survival decreased with age, but the age-related disparity was small for patients younger than 75. In both countries, 5-year RS was >80 % for patients less than 75 years of age but <70 % for those age 75+.

Conclusions

Overall, 5-year survival for patients with CLL is good, but 10-year survival is significantly lower, and survival was much lower for those age 75+. Major differences in survival between countries were not observed. Further research into ways to increase survival for older CLL patients are needed to reduce the persistent large age-related survival disparity.

Integrated Genomic and Network-Based Analyses of Complex Diseases and Human Disease Network

A disease phenotype generally reflects various pathobiological processes that interact in a complex network. The highly interconnected nature of the human protein interaction network (interactome) indicates that, at the molecular level, it is difficult to consider diseases as being independent of one another. Recently, genome-wide molecular measurements, data mining and bioinformatics approaches have provided the means to explore human diseases from a molecular basis. The exploration of diseases and a system of disease relationships based on the integration of genome-wide molecular data with the human interactome could offer a powerful perspective for understanding the molecular architecture of diseases. Recently, subnetwork markers have proven to be more robust and reliable than individual biomarker genes selected based on gene expression profiles alone, and achieve higher accuracy in disease classification. We have applied one of these methodologies to idiopathic dilated cardiomyopathy (IDCM) data that we have generated using a microarray and identified significant subnetworks associated with the disease. In this paper, we review the recent endeavours in this direction, and summarize the existing methodologies and computational tools for network-based analysis of complex diseases and molecular relationships among apparently different disorders and human disease network. We also discuss the future research trends and topics of this promising field.

Regulatory network reconstruction reveals genes with prognostic value for chronic lymphocytic leukemia

Background

The clinical course of chronic lymphocytic leukemia (CLL) is highly variable; some patients follow an indolent course, but others progress to a more advanced stage. The mutational status of rearranged immunoglobulin heavy chain variable (IGVH) genes in CLL is a feature that is widely recognized for dividing patients into groups that are related to their prognoses. However, the regulatory programs associated with the IGVH statuses are poorly understood, and markers that can precisely predict survival outcomes have yet to be identified.

Methods

In this study, (i) we reconstructed gene regulatory networks in CLL by applying an information-theoretic approach to the expression profiles of 5 cohorts. (ii) We applied master regulator analysis (MRA) to these networks to identify transcription factors (TFs) that regulate an IGVH mutational status signature. The IGVH mutational status signature was developed by searching for differentially expressed genes between the IGVH mutational statuses in numerous CLL cohorts. (iii) To evaluate the biological implication of the inferred regulators, prognostic values were determined using time to treatment (TTT) and overall survival (OS) in two different cohorts.

Results

A robust IGVH expression signature was obtained, and various TFs emerged as regulators of the signature in most of the reconstructed networks. The TF targets expression profiles exhibited significant differences with respect to survival, which allowed the definition of a reduced profile with a high value for OS. TCF7 and its targets stood out for their roles in progression.

Conclusion

TFs and their targets, which were obtained merely from inferred regulatory associations, have prognostic implications and reflect a regulatory context for prognosis.

Electronic supplementary material

The online version of this article (doi:10.1186/s12864-015-2189-6) contains supplementary material, which is available to authorized users.

Clustering of Expression Data in Chronic Lymphocytic Leukemia Reveals New Molecular Subdivisions

Although the identification of inherent structure in chronic lymphocytic leukemia (CLL) gene expression data using class discovery approaches has not been extensively explored, the natural clustering of patient samples can reveal molecular subdivisions that have biological and clinical implications. To explore this, we preprocessed raw gene expression data from two published studies, combined the data to increase the statistical power, and performed unsupervised clustering analysis. The clustering analysis was replicated in 4 independent cohorts. To assess the biological significance of the resultant clusters, we evaluated their prognostic value and identified cluster-specific markers. The clustering analysis revealed two robust and stable subgroups of CLL patients in the pooled dataset. The subgroups were confirmed by different methodological approaches (non-negative matrix factorization NMF clustering and hierarchical clustering) and validated in different cohorts. The subdivisions were related with differential clinical outcomes and markers associated with the microenvironment and the MAPK and BCR signaling pathways. It was also found that the cluster markers were independent of the immunoglobulin heavy chain variable (IGVH) genes mutational status. These findings suggest that the microenvironment can influence the clinical behavior of CLL, contributing to prognostic differences. The workflow followed here provides a new perspective on differences in prognosis and highlights new markers that should be explored in this context.

Network-based sub-network signatures unveil the potential for acute myeloid leukemia therapy

Although gene expression profiling studies of acute myeloid leukemia (AML) patients have provided key insights into potential diagnostic and prognostic markers and therapeutic targets, it is not clear that the patterns of molecular heterogeneity affect the tumor biology and respond to the treatment. We hypothesized that network-based gene expression signatures of AML represent the mechanistically important genes and may improve the predicted performance of prognosis and clinical outcome. We provided the random walk with restart (RWR) analysis to discover the sub-network of genomic alterations. The RWR approach integrates the signature genes derived from the random forest (RF) analysis as "seeds" to identify genes critical to the AML recurrence phenotype. To test whether the 81-gene biomarkers could predict AML recurrence, we developed Survival Support Vector Machine (SSVM) models using a gene expression dataset and test on an independent dataset. The random forest classifier was built based on 81-gene biomarkers to separate the AML patients into "recurrence" and "non-recurrence" groups. The 81-gene biomarkers showed significant enrichment related to cancer pathophysiology and provided good coverage of sub-network biomarkers and AML-related signaling pathways. The SSVM-based score was significantly associated with overall survival (hazard ratio [HR], 2.16; 95% confidence interval [CI], 1.18-3.97; p = 0.01). Similar results were obtained with reversed training and testing datasets (hazard ratio [HR], 1.6; 95% confidence interval [CI], 1.08-2.37; p = 0.02). The 81-gene biomarker based RF classifier improved classification performance. Overall, 81-gene biomarkers might be useful prognostic and predictive molecular markers to predict the clinical outcome of AML patients.

Systems biology approach to studying proliferation-dependent prognostic subnetworks in breast cancer

Tumor proliferative capacity is a major biological correlate of breast tumor metastatic potential. In this paper, we developed a systems approach to investigate associations among gene expression patterns, representative protein-protein interactions, and the potential for clinical metastases, to uncover novel survival-related subnetwork signatures as a function of tumor proliferative potential. Based on the statistical associations between gene expression patterns and patient outcomes, we identified three groups of survival prognostic subnetwork signatures (SPNs) corresponding to three proliferation levels. We discovered 8 SPNs in the high proliferation group, 8 SPNs in the intermediate proliferation group, and 6 SPNs in the low proliferation group. We observed little overlap of SPNs between the three proliferation groups. The enrichment analysis revealed that most SPNs were enriched in distinct signaling pathways and biological processes. The SPNs were validated on other cohorts of patients, and delivered high accuracy in the classification of metastatic vs non-metastatic breast tumors. Our findings indicate that certain biological networks underlying breast cancer metastasis differ in a proliferation-dependent manner. These networks, in combination, may form the basis of highly accurate prognostic classification models and may have clinical utility in guiding therapeutic options for patients.

The cancer cell map initiative: defining the hallmark networks of cancer

Progress in DNA sequencing has revealed the startling complexity of cancer genomes, which typically carry thousands of somatic mutations. However, it remains unclear which are the key driver mutations or dependencies in a given cancer and how these influence pathogenesis and response to therapy. Although tumors of similar types and clinical outcomes can have patterns of mutations that are strikingly different, it is becoming apparent that these mutations recurrently hijack the same hallmark molecular pathways and networks. For this reason, it is likely that successful interpretation of cancer genomes will require comprehensive knowledge of the molecular networks under selective pressure in oncogenesis. Here we announce the creation of a new effort, The Cancer Cell Map Initiative (CCMI), aimed at systematically detailing these complex interactions among cancer genes and how they differ between diseased and healthy states. We discuss recent progress that enables creation of these cancer cell maps across a range of tumor types and how they can be used to target networks disrupted in individual patients, significantly accelerating the development of precision medicine.

Proteins interaction network and modeling of IGVH mutational status in chronic lymphocytic leukemia

Background

Chronic lymphocytic leukemia (CLL) is an incurable malignancy of mature B-lymphocytes, characterized as being a heterogeneous disease with variable clinical manifestation and survival. Mutational statuses of rearranged immunoglobulin heavy chain variable (IGVH) genes has been consider one of the most important prognostic factors in CLL, but despite of its proven value to predict the course of the disease, the regulatory programs and biological mechanisms responsible for the differences in clinical behavior are poorly understood.

Methods

In this study, (i) we performed differential gene expression analysis between the IGVH statuses using multiple and independent CLL cohorts in microarrays platforms, based on this information, (ii) we constructed a simplified protein-protein interaction (PPI) network and (iii) investigated its structure and critical genes. This provided the basis to (iv) develop a Boolean model, (v) infer biological regulatory mechanism and (vi) performed perturbation simulations in order to analyze the network in dynamic state.

Results

The result of topological analysis and the Boolean model showed that the transcriptional relationships of IGVH mutational status were determined by specific regulatory proteins (PTEN, FOS, EGR1, TNF, TGFBR3, IFGR2 and LPL). The dynamics of the network was controlled by attractors whose genes were involved in multiple and diverse signaling pathways, which may suggest a variety of mechanisms related with progression occurring over time in the disease. The overexpression of FOS and TNF fixed the fate of the system as they can activate important genes implicated in the regulation of process of adhesion, apoptosis, immune response, cell proliferation and other signaling pathways related with cancer.

Conclusion

The differences in prognosis prediction of the IGVH mutational status are related with several regulatory hubs that determine the dynamic of the system.

Electronic supplementary material

The online version of this article (doi:10.1186/s12976-015-0008-z) contains supplementary material, which is available to authorized users.

ID helix-loop-helix proteins as determinants of cell survival in B-cell chronic lymphocytic leukemia cells in vitro

BACKGROUND

Members of the inhibitor of DNA-binding (ID) family of helix-loop-helix proteins have been causally implicated in the pathogenesis of several types of B-cell lineage malignancy, either on the basis of mutation or by altered expression. B-cell chronic lymphocytic leukemia encompasses a heterogeneous group of disorders and is the commonest leukaemia type in the Western world. In this study, we have investigated the pathobiological functions of the ID2 and ID3 proteins in this disease with an emphasis on their role in regulating leukemic cell death/survival.

METHODS

Bioinformatics analysis of microarray gene expression data was used to investigate expression of ID2/ID3 in leukemic versus normal B cells, their association with clinical course of disease and molecular sub-type and to reconstruct a gene regulatory network using the 'maximum information coefficient' (MIC) for target gene inference. In vitro cultured primary leukemia cells, either in isolation or co-cultured with accessory vascular endothelial cells, were used to investigate ID2/ID3 protein expression by western blotting and to assess the cytotoxic response of different drugs (fludarabine, chlorambucil, ethacrynic acid) by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay. ID2/ID3 protein levels in primary leukemia cells and in MEC1 cells were manipulated by transduction with siRNA reagents.

RESULTS

Datamining showed that the expression profiles of ID2 and ID3 are associated with distinct pathobiological features of disease and implicated both genes in regulating cell death/survival by targeting multiple non-overlapping sets of apoptosis effecter genes. Consistent with microarray data, the overall pattern of ID2/ID3 protein expression in relation to cell death/survival responses of primary leukemia cells was suggestive of a pro-survival function for both ID proteins. This was confirmed by siRNA knock-down experiments in MEC1 cells and in primary leukemia cells, but with variability in the dependence of leukemic cells from different patients on ID protein expression for cell survival. Vascular endothelial cells rescued leukemia cells from spontaneous and cytotoxic drug-induced cell death at least in part, via an ID protein-coupled redox-dependent mechanism.

CONCLUSIONS

Our study provides evidence for a pro-survival function of the ID2/ID3 proteins in chronic lymphocytic leukemia cells and also highlights these proteins as potential determinants of the pathobiology of this disorder.

Pathway activity inference for multiclass disease classification through a mathematical programming optimisation framework

BACKGROUND

Applying machine learning methods on microarray gene expression profiles for disease classification problems is a popular method to derive biomarkers, i.e. sets of genes that can predict disease state or outcome. Traditional approaches where expression of genes were treated independently suffer from low prediction accuracy and difficulty of biological interpretation. Current research efforts focus on integrating information on protein interactions through biochemical pathway datasets with expression profiles to propose pathway-based classifiers that can enhance disease diagnosis and prognosis. As most of the pathway activity inference methods in literature are either unsupervised or applied on two-class datasets, there is good scope to address such limitations by proposing novel methodologies.

RESULTS

A supervised multiclass pathway activity inference method using optimisation techniques is reported. For each pathway expression dataset, patterns of its constituent genes are summarised into one composite feature, termed pathway activity, and a novel mathematical programming model is proposed to infer this feature as a weighted linear summation of expression of its constituent genes. Gene weights are determined by the optimisation model, in a way that the resulting pathway activity has the optimal discriminative power with regards to disease phenotypes. Classification is then performed on the resulting low-dimensional pathway activity profile.

CONCLUSIONS

The model was evaluated through a variety of published gene expression profiles that cover different types of disease. We show that not only does it improve classification accuracy, but it can also perform well in multiclass disease datasets, a limitation of other approaches from the literature. Desirable features of the model include the ability to control the maximum number of genes that may participate in determining pathway activity, which may be pre-specified by the user. Overall, this work highlights the potential of building pathway-based multi-phenotype classifiers for accurate disease diagnosis and prognosis problems.

Identification of a 20-gene expression-based risk score as a predictor of clinical outcome in chronic lymphocytic leukemia patients

Despite the improvement in treatment options, chronic lymphocytic leukemia (CLL) remains an incurable disease and patients show a heterogeneous clinical course requiring therapy for many of them. In the current work, we have built a 20-gene expression (GE)-based risk score predictive for patients overall survival and improving risk classification using microarray gene expression data. GE-based risk score allowed identifying a high-risk group associated with a significant shorter overall survival (OS) and time to treatment (TTT) (P ≤ .01), comprising 19.6% and 13.6% of the patients in two independent cohorts. GE-based risk score, and NRIP1 and TCF7 gene expression remained independent prognostic factors using multivariate Cox analyses and combination of GE-based risk score together with NRIP1 and TCF7 gene expression enabled the identification of three clinically distinct groups of CLL patients. Therefore, this GE-based risk score represents a powerful tool for risk stratification and outcome prediction of CLL patients and could thus be used to guide clinical and therapeutic decisions prospectively.

miR-150 influences B-cell receptor signaling in chronic lymphocytic leukemia by regulating expression of GAB1 and FOXP1

We examined the microRNAs (miRNAs) expressed in chronic lymphocytic leukemia (CLL) and identified miR-150 as the most abundant, but with leukemia cell expression levels that varied among patients. CLL cells that expressed ζ-chain-associated protein of 70 kDa (ZAP-70) or that used unmutated immunoglobulin heavy chain variable (IGHV) genes, each had a median expression level of miR-150 that was significantly lower than that of ZAP-70-negative CLL cells or those that used mutated IGHV genes. In samples stratified for expression of miR-150, CLL cells with low-level miR-150 expressed relatively higher levels of forkhead box P1 (FOXP1) and GRB2-associated binding protein 1 (GAB1), genes with 3' untranslated regions having evolutionary-conserved binding sites for miR-150. High-level expression of miR-150 could repress expression of these genes, which encode proteins that enhance B-cell receptor signaling, a putative CLL-growth/survival signal. Also, high-level expression of miR-150 was a significant independent predictor of longer treatment-free survival or overall survival, whereas an inverse association was observed for high-level expression of GAB1 or FOXP1 for overall survival. This study demonstrates that expression of miR-150 can influence the relative expression of GAB1 and FOXP1 and the signaling potential of the B-cell receptor, thereby possibly accounting for the noted association of expression of miR-150 and disease outcome.

Principles and methods of integrative genomic analyses in cancer

Combined analyses of molecular data, such as DNA copy-number alteration, mRNA and protein expression, point to biological functions and molecular pathways being deregulated in multiple cancers. Genomic, metabolomic and clinical data from various solid cancers and model systems are emerging and can be used to identify novel patient subgroups for tailored therapy and monitoring. The integrative genomics methodologies that are used to interpret these data require expertise in different disciplines, such as biology, medicine, mathematics, statistics and bioinformatics, and they can seem daunting. The objectives, methods and computational tools of integrative genomics that are available to date are reviewed here, as is their implementation in cancer research.

Integration of transcription and flux data reveals molecular paths associated with differences in oxygen-dependent phenotypes of Saccharomyces cerevisiae

BACKGROUND

Saccharomyces cerevisiae is able to adapt to a wide range of external oxygen conditions. Previously, oxygen-dependent phenotypes have been studied individually at the transcriptional, metabolite, and flux level. However, the regulation of cell phenotype occurs across the different levels of cell function. Integrative analysis of data from multiple levels of cell function in the context of a network of several known biochemical interaction types could enable identification of active regulatory paths not limited to a single level of cell function.

RESULTS

The graph theoretical method called Enriched Molecular Path detection (EMPath) was extended to enable integrative utilization of transcription and flux data. The utility of the method was demonstrated by detecting paths associated with phenotype differences of S. cerevisiae under three different conditions of oxygen provision: 20.9%, 2.8% and 0.5%. The detection of molecular paths was performed in an integrated genome-scale metabolic and protein-protein interaction network.

CONCLUSIONS

The molecular paths associated with the phenotype differences of S. cerevisiae under conditions of different oxygen provisions revealed paths of molecular interactions that could potentially mediate information transfer between processes that respond to the particular oxygen availabilities.

Skin sensitizers differentially regulate signaling pathways in MUTZ-3 cells in relation to their individual potency

Background

Due to the recent European legislations posing a ban of animal tests for safety assessment within the cosmetic industry, development of in vitro alternatives for assessment of skin sensitization is highly prioritized. To date, proposed in vitro assays are mainly based on single biomarkers, which so far have not been able to classify and stratify chemicals into subgroups, related to risk or potency.

Methods

Recently, we presented the Genomic Allergen Rapid Detection (GARD) assay for assessment of chemical sensitizers. In this paper, we show how the genome wide readout of GARD can be expanded and used to identify differentially regulated pathways relating to individual chemical sensitizers. In this study, we investigated the mechanisms of action of a range of skin sensitizers through pathway identification, pathway classification and transcription factor analysis and related this to the reactive mechanisms and potency of the sensitizing agents.

Results

By transcriptional profiling of chemically stimulated MUTZ-3 cells, 33 canonical pathways intimately involved in sensitization to chemical substances were identified. The results showed that metabolic processes, cell cycling and oxidative stress responses are the key events activated during skin sensitization, and that these functions are engaged differently depending on the reactivity mechanisms of the sensitizing agent. Furthermore, the results indicate that the chemical reactivity groups seem to gradually engage more pathways and more molecules in each pathway with increasing sensitizing potency of the chemical used for stimulation. Also, a switch in gene regulation from up to down regulation, with increasing potency, was seen both in genes involved in metabolic functions and cell cycling. These observed pathway patterns were clearly reflected in the regulatory elements identified to drive these processes, where 33 regulatory elements have been proposed for further analysis.

Conclusions

This study demonstrates that functional analysis of biomarkers identified from our genomics study of human MUTZ-3 cells can be used to assess sensitizing potency of chemicals in vitro, by the identification of key cellular events, such as metabolic and cell cycling pathways.

Characterization of glycoproteins in pancreatic cyst fluid using a high-performance multiple lectin affinity chromatography platform

Currently, pancreatic cancer is the fourth cause of cancer death. In 2013, it is estimated that ∼38 460 people will die of pancreatic cancer. Early detection of malignant cyst (pancreatic cancer precursor) is necessary to help prevent late diagnosis of the tumor. In this study, we characterized glycoproteins and nonglycoproteins on pooled mucinous (n = 10) and nonmucinous (n = 10) pancreatic cyst fluid to identify "proteins of interest" to differentiate between mucinous cyst from nonmucinous cyst and investigate these proteins as potential biomarker targets. An automated multilectin affinity chromatography (M-LAC) platform was utilized for glycoprotein enrichment followed by nano-LC-MS/MS analysis. Spectral count quantitation allowed for the identification of proteins with significant differential levels in mucinous cysts from nonmucinous cysts of which one protein (periostin) was confirmed via immunoblotting. To exhaustively evaluate differentially expressed proteins, we used a number of proteomic tools including gene ontology classification, pathway and network analysis, Novoseek data mining, and chromosome gene mapping. Utilization of complementary proteomic tools revealed that several of the proteins such as mucin 6 (MUC6), bile salt-activated lipase (CEL), and pyruvate kinase lysozyme M1/M2 with significant differential expression have strong association with pancreatic cancer. Furthermore, chromosome gene mapping demonstrated coexpressions and colocalization of some proteins of interest including 14-3-3 protein epsilon (YWHAE), pigment epithelium derived factor (SERPINF1), and oncogene p53.