Transcriptomic and Proteomic Data Integration and ... - ACS Publications

Subscriber access provided by UNIV OF LETHBRIDGE

Article

Transcriptomic and Proteomic data integration and 2D molecular maps with regulatory and functional linkages: Application to Cell Proliferation and Invasion Networks in Glioblastoma Manoj Kumar Gupta, Savita Jayaram, Divijendra Natha Reddy, Ravindra Varma Polisetty, and Ravi Sirdeshmukh J. Proteome Res., Just Accepted Manuscript • DOI: 10.1021/acs.jproteome.5b00765 • Publication Date (Web): 14 Oct 2015 Downloaded from http://pubs.acs.org on October 14, 2015

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Journal of Proteome Research is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

Transcriptomic and Proteomic data integration and 2D molecular maps with regulatory and functional linkages: Application to Cell Proliferation and Invasion Networks in Glioblastoma Manoj Kumar Gupta1,2*, Savita Jayaram1,2*, Divijendra Natha Reddy3, Ravindra Varma Polisetty1, Ravi Sirdeshmukh1,3#

*Both the authors contributed equally to this work 1

Institute of Bioinformatics, International Tech Park, Bangalore-560066, India

2

Manipal University, Madhav Nagar, Manipal-576104, India

3

Neuro-Oncology, Mazumdar Shaw Centre for Translational Research, Mazumdar Shaw

Medical Foundation, Narayana Health, Bangalore-560099, India #

Corresponding author

Institute of Bioinformatics, International Tech Park, Bangalore-560066, India TEL: 0091-9885090963; FAX: 0091-80-28416132 E-mail: [email protected], [email protected]

ABSTRACT Glioblastoma multiforme (GBM) is the most aggressive primary brain tumor characterized by high rates of cell proliferation, migration and invasion. New therapeutic strategies and targets are being continuously explored with the hope for better outcome. By overlaying transcriptomics and proteomic data from GBM clinical tissues, we identified 317 differentially expressed proteins to be concordant with the mRNAs. We used these entities to generate integrated regulatory information at the level of microRNAs and their mRNA and protein targets using prediction programs or experimentally verified miRNA target mode in miRWalk. We observed 60% or more of the miRNA-target pairs to be consistent with experimentally observed inverse expression of these molecules in GBM. The integrated view of these regulatory cascades in the contexts of cell proliferation and invasion networks revealed 2-dimensional molecular interactions with regulatory and functional linkages (miRNAs and their mRNA/protein targets in one dimension; multiple miRNAs associated in 1

ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

a functional network in the second dimension). Twenty-eight of the 35 differentially expressed concordant mRNA-protein entities represented in the proliferation network and 51 of the 59 such entities represented in the invasion network mapped to altered miRNAs from GBM and conformed to inverse relationship in expression with their targets. We believe the 2-dimensional maps of gene expression changes enhance the strength of the discovery datasets derived from omics-based studies for their applications in GBM as well as tumors in general.

Keywords: glioblastoma, microRNA, proteomics, transcriptomics, 2D molecular maps, cell proliferation, cell invasion

INTRODUCTION Malignant gliomas are the most common type of brain cancer and arise from glial cells within the central nervous system (CNS). Gliomas are classified according to histopathological criteria defined by the World Health Organization (WHO) as grade II, III and IV. The Grade IV tumors, also referred to as Glioblastoma multiforme (GBM)1-3, are the most common and aggressive tumors that are characterized by uncontrolled proliferation, areas of necrosis and diffuse infiltration with poor prognosis and a median survival rate of 12 months even with an aggressive treatment1,4. The capacity of GBM cells to disperse from the primary tumor site and infiltrate the brain parenchyma severely limits the effectiveness of surgery as well as radiotherapy. Thus, targeting the invading cells to restrain their migratory capacity is likely to provide a strong alternative to GBM therapy5-7. There are already some efforts in this direction and newer ways of exploring the molecular changes underlying these tumors would be useful to get better molecular insights on their pathogenesis.

The tumor has been extensively analyzed at the genomic, transcriptomic and proteomic levels which have resulted in large volume of molecular data underlying tumor-associated processes and interaction networks. The repertoire of somatic genomic alterations in GBM have been extensively described by several groups using whole-genome and exome sequencing strategies to identify copy number alterations, complex rearrangements and novel mutated genes that may drive this lethal cancer8,9. Gene expression analysis encompassing both transcriptome and proteome using next generation sequencing (NGS) and mass spectrometry

2


Page 2 of 27

Page 3 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(MS), respectively have been applied to identify differentially expressed mRNA/proteins associated with GBM9-12.

The Cancer genome Atlas (TCGA) network group conducted a large scale study on GBM including copy number variations, mutation frequency analysis, gene expression and microRNA profiling to identify altered genes and pathways in these tumors. The TCGA study involved analysis of somatic mutations and gene expression data from more than 500 tumor tissues8,9,11 and revealed mutations in several key candidate cancer genes such as NF1, ERBB2, TP53, PIK3R1, IDH1 and TERT. The gene expression analysis revealed a large number of altered mRNAs and included altered expression of receptors such as EGFR and PDGFR and cell cycle regulators RB1 and CDK4 involved in growth and proliferation. On the proteomics front, our group has carried out analysis of membrane proteins13 from GBM tumor tissues using high-resolution mass spectrometry and identified differentially expressed proteins. The proteins identified were associated with major cellular processes and known as well as novel tumor associated molecules in GBM10,13 such as, tenascin, glial fibrillary acidic protein, vimentin, nestin and vitronectin, involved in cytoskeletal dynamics related to cell shape, integrity, cell adhesion and cell motility. Many calcium-regulatory proteins were identified and included several members of S100 family proteins (A4, A8, A9, A10, A11), annexins (A1, A2, A4, A5) and integrins (ITGB1, ITGB2, ITGAM, ITGA5). Some other important altered proteins included protease inhibitor SERPINA3, brain specific molecule tenascin-R (TNR) involved in cell adhesion and tripeptidyl-peptidase I (TPP1), a serine protease and tumor-associated proteins like SPARC and MMP912.

These investigations revealed multiple molecular changes in pathways and complex molecular interaction networks underlying the pathogenesis of these tumors. In recent years, investigators are trying to integrate proteomic data with transcriptomic results by exploiting biological complementarity between transcripts and proteins. However, this may be challenging for the following reasons: 1. The steady state levels of mRNA and protein are determined by the balance between their respective synthesis and degradation rates which may be widely different. The factors and mechanisms of regulation of mRNA and protein steady state are independent of each other with some exceptions. In consistence with this, in a recent human proteome analysis by Wilhelm et al.14, it was reported that the levels of mRNA 3



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

and protein expression vary greatly in cells and tissues although their ratios are highly conserved across tissues. 2. Despite advances in analytical platforms and the depth to which a proteome may be accessed, there are still limitations to achieve comparable coverage of the two entities from any biological sample. Given this, large scale linear comparisons between transcript and protein datasets may be difficult to achieve with current analytical platforms. Nevertheless, even if implemented in a limited way, their integration would help in enhancing the understanding of the tumor as well as consolidate the strength of the molecular changes observed. We have attempted to identify altered molecular interactions by integrating gene expression data from transcript and protein level with microRNAs (miRNAs) which are emerging as key players in cellular processes associated with several tumors including GBM15,16 and as novel therapeutic targets17.

In this study, we provide a framework for integrating transcriptomics and proteomics data of GBM for identification of molecular cascades involving miRNA regulators and key modulators of proliferation and invasion - two hallmarks of GBM, revealing important regulatory and functional relationships among them in the context of the tumor.

MATERIALS AND METHODS Data sources - Proteomic, transcriptomic and miRNA data in GBM The proteomic data used in this study was obtained by our group through analysis of differentially expressed membrane proteins13 from GBM tumor tissues, using high-resolution mass spectrometry. The study identified 1834 proteins, 710 of them were differentially expressed (≥1.4 fold change). About 60% of them were known membrane proteins associated with major tumor associated cellular processes. The differentially expressed transcript data generated by TCGA group on GBM9 used here were imported from the analysis of Dong et al.18 and Anduril resource19 (http://csbi.ltdk.helsinki.fi/anduril/tcga-gbm). Using primary TCGA data9, Dong et al.18 have arrived at a subset of 1697 differentially expressed transcripts with unified fold change values through statistical analysis for inter patient variations. Global profiling of miRNAs by the TCGA group resulted in the identification of over 500 altered miRNAs from 240 GBM tumor tissues9, 199 of them with ≥1.2 fold change value as per TCGA Anduril resource19. Interestingly, 149 of them were also found to be significant based on independent co-expression analysis of the TCGA data by Dong et al.18. We have taken the set of 199 miRNAs for this analysis. Since our objective has been to map miRNA to their 4


Page 4 of 27

Page 5 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


mRNA and protein target pairs (see below), the transcript and protein datasets were compared to examine the concordance between them and the concordant subset was used for further target analysis.

Identification of regulatory miRNA and target cascades We explored the miRNA–target cascades for the dataset of differentially expressed miRNAs and concordant mRNA and protein entities discussed above. For mapping differentially expressed miRNAs to their target mRNA and protein entities, we applied two strategies for identification of the targets. In Strategy 1, using miRWalk database20 (http://umm.uniheidelberg.de/apps/zmf/mirwalk) that utilizes multiple prediction programs available on the web, we identified proteins and mRNA pairs in the dataset that could be targets of the 199 altered miRNAs identified in GBM as discussed above. The criteria for selection of the predicted targets applied were as follows: restricted search of miRNA binding sites within 3’UTR the longest length of the transcript of the gene of interest, minimum number of nucleotides of miRNA sequence (seed length) as 7 and the probability score (p-value) was set to 0.05. Targets were further screened for their prediction by at least 4 prediction programs and the results were then assessed for their inverse relationship against the GBM expression data.

In Strategy 2, we first constructed functional interaction networks (cell proliferation and invasion) in the context of GBM as described below. We then examined all the entities in the functional network to map to miRNAs and target cascades. For this we selected experimentally validated miRNA-target pair mode in miRWalk instead of the prediction program and used it as reference to pair up the network entities with respective miRNA regulators together with inverse correlation in expression observed in GBM datasets. Each of these miRNA-target pairs was supported by at least one PubMed research article. Those which did not show inverse correlation in the experimental data, were not considered further as targets although present in the network. Construction of functional interaction networks We used the list of all the mRNA and protein pairs found to be concordant in their altered expression in GBM datasets. Gene Ontology annotations for these concordant mRNA and protein pairs were extracted using Human Protein Reference Database (http://www.hprd.org) 5



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

in the context of cell proliferation and cell adhesion and migration – the latter two being integral parts of the tumor invasion process. Further, the cell proliferation and cell invasion entities from the analysis were used to construct their respective functional interaction networks in the context of GBM using STRING (version 9.1; Jan, 2015) web resource. The confidence of the network was assessed by p-values observed in the STRING output. It is to be noted that the protein data used by us is largely membrane-centric. Therefore, we further expanded the networks with additional key entities (8 each) present in mRNA data with recognized role in cell proliferation and invasion for which proteomic information was drawn from other published resources8,21-25.

RESULTS AND DISCUSSION mRNA- protein overlap and regulating microRNAs in GBM Mapping the expression changes observed at transcript and protein level to biological functions enhances the functional perspective of the high throughput data. Integrating information on miRNAs and their target mRNAs and proteins brings in yet another relational element for the biological processes associated with tumors. When the differentially expressed mRNA and proteins (TCGA data and our proteomic analysis) were compared, we observed that 327 of the 710 differentially altered proteins (≥1.4 fold) overlapped with mRNA data. Since the depth of the proteome accessed is usually smaller than that of the transcriptome, such comparison is generally limited by the number of differentially expressed proteins. Although the number of the overlapping mRNA and protein entities was much lower than their number in individual data sets, we observed concordance in the expression trend of more than 95% of them (317 out of 327; Supplementary Table 1).

Expression of a majority of human genes is regulated by one or more miRNAs and the regulation occurs by repression at the post transcriptional or translational level15-17. MicroRNAs are thus emerging as key players in tumorigenesis with both oncogenic and tumor suppressive roles. Therefore, global profiling of dysregulated miRNAs is important to understand their involvement in GBM26. The altered miRNAs are expected to inversely correlate with the expression of their corresponding transcript and protein targets and may thus offer a key link for integrating the expression changes in terms of regulatory cascades involving the three different levels. We therefore further explored if the concordance relationship observed for mRNAs and proteins could be extended to altered regulatory 6


Page 6 of 27

Page 7 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


microRNAs. Our analysis was thus carried out within the frame work of 317 differentially expressed proteins, their corresponding transcripts and 199 miRNAs deregulated in GBM.

The miRNA and mRNA-protein target analysis carried out using multiple prediction programs for the GBM datasets (Strategy 1) resulted in identification of 16,253 predicted miRNA-target pairs which covered 312 of the 317 differentially expressed mRNA-proteins and 194 of the 199 microRNAs used for the analysis. When we examined these interactions in expression data of GBM, more than 60% of the miRNA and target pairs (n=9817) were found to show inverse relationship in their expression with the remaining 40% (n=6436) not conforming to the inverse relationship (Supplementary Table 2). In Figure 1, we show 12 representative miRNAs that were significantly altered and their mRNA and protein targets with an inverse relationship in their expression in GBM9,13,18,26 forming regulatory cascades. For the miRNAs over expressed in GBM, presumably oncogenic miRNAs, we observed under expression of their corresponding mRNA and protein targets (Figure 1A). On the other hand, for under expressed miRNAs, presumably tumor suppressors, the corresponding targets were shown to be upregulated (Figure 1B). A schematic representation of this analysis and the results are shown in Supplementary Figure 1. It was interesting to observe this correlation between the miRNAs and many of their mRNA and protein targets from two completely independent investigations with potential of being extendable to many more molecules. A plausible explanation for this could be the dysregulation of whole regulatory cascades and not single entities involved in the tumor related events. Their identification would be important as multiple miRNAs and their targets reveal 2-dimensional molecular maps (2D Maps) – with regulatory linkage in one dimension and different miRNAs with plausible biological/functional linkages in the second dimension (Figure 2). They may be applied to view tumor related networks and processes in the 2D Map perspective.

Functional network analysis and 2D molecular maps To support our interpretations as above and the 2D maps in the context of GBM, we used the list of mRNA and protein pairs found to be concordant in their altered expression in GBM datasets (n=317) along with their Gene Ontology annotations (Human Protein Reference Database; http://www.hprd.org) to view them in the context of cell proliferation and cell invasion process (adhesion and migration). For construction of GBM related networks and 2D maps, we could identify 45 and 70 genes/proteins from the GBM data for proliferation 7



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

and invasion network, respectively and used them initially in STRING. The STRING output showed about 60 interactions (p-value = 1.37e-8) for proliferation and about 280 interactions (p-value = 1.00e-14) for invasion network. It is to be noted that the protein data used by us is largely membrane-centric. Therefore, we further expanded the networks with additional key entities (8 each) as described in the ‘Materials and Method’ section. Thus the total number of genes/proteins entities used for constructing these two functional networks were 53 and 78, respectively. This added about 35 more interactions in proliferation module and about 70 more interactions in the invasion module and increased the confidence score further. In the proliferation network, the increase was observable and the p-value improved further (p-value = 1.20e-10). In the case of invasion network, the p-value was already much above the significance threshold so the increase was not observable after incorporation of additional entities. The final network outputs generated by STRING captured 38 (of 53) interacting entities in proliferation and 69 (of 78) in invasion module (Figure 3A and 4A) and were used for identifying regulatory cascades. All the entities from the two modules along with their fold changes are listed in the Supplementary Table 3 and 4, respectively.

Using experimentally validated miRNA-target mode in miRWalk, we constructed the regulatory cascades of miRNA and targets for all the entities within the cell proliferation and cell invasion networks, as described under ‘Materials and Methods’. The resulting 2dimensional molecular maps (2D maps) included miRNAs and target entities that are biologically/functionally linked in one dimension with their regulatory linkages in the second dimension. We observed majority of the respective network entities conformed to the 2D maps and are detailed below. For the remaining entities of the network either experimentally validated miRNA-target information was not available in miRWalk or the expected inverse relationship was not observed in GBM data.

Cell proliferation network Dysregulation of signal transduction pathways governing cell proliferation is a hallmark feature of most GBMs. Many studies including ours have reported the activation of multiple proteins and pathways in aggressive GBMs, specifically overexpression or activation of EGFR, PDGFR and FGFR receptor tyrosine kinases leading to activation of their downstream effectors, RAS and PI3K/AKT1. EGFR is overexpressed in almost 40-50% of GBMs while PDGFR subclass accounts for 25-30% of GBMs and they are known to 8


Page 8 of 27

Page 9 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


contribute to uncontrolled proliferation and survival of glioma cells27,28. Activation of AKT1 is seen in 70% of GBMs and constitutive activation of RAF1 was shown to induce glioma formation in mice models21. Interestingly, DNM1, which is required for internalization and degradation of EGFR and is part of this network, was found to be down regulated, assisting the enhancement of the RTK signalling pathway. The proliferation networks consist of cell cycle kinases, CDK4 and CDK6 and their downstream signalling molecules involving CHI3L1, RAC2 and PLG. CDK4 is also the most frequently amplified gene that is part of the chromosome 12q13-15 amplification cluster in gliomas and concordantly associated with high protein and transcript levels implying its important role in glioma malignant progression23,29,30. Matrix metalloproteinases, MMP9 and cathepsin B participate in tumor cell proliferation, angiogenesis and invasion and inhibiting both was found to result in the reduction of tumor cell growth and invasion31. All these molecules are overexpressed in both proteomic and transcriptomic data, leading to increased cell cycle progression and cell survival. Putting them together using STRING tool followed by manual curation, we arrive at an interaction network shown in Figure 3A. As described above the network contains 38 concordant mRNA and protein entities having more than 95 interactions (p-value = 1.20e10). We further mapped the miRNA that may target these entities using experimentally validated miRNA target mode of miRWalk and integrated them with inverse relationship in their expression observed in experimental GBM data, as described above. The analysis scheme, the number of miRNA-target interactions and those conforming to inverse relationship as per GBM data are shown in Supplementary Figure 2A. We identified experimentally validated miRNA regulators for 35 of the 38 entities present for the network and observed the expected inverse relationship for 28 of them accounting for 80% of the target entities in the network which are shown in Supplementary Table 3. Figure 3B shows the 2D map of some of the key miRNAs and their targets included in the proliferation network. MicroRNA miR-218 is downregulated in GBM resulting in overexpression of its RTK signalling targets, such as EGFR and PDGFR32,33, CDK634, cathepsin B35 and MMP936. Similarly, miR-737 and miR-12838 also target EGFR and its downstream mediators, PIK3CA, AKT1 and RAF1. These miRNAs collectively regulate the growth factors, EGFR and PDGFR, and their downstream targets, PIK3CA and AKT1 and their interactors such as CDK6, RAF1 and MMP9 that together are involved in the proliferative signalling processes inside the cell.

9



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Cell Invasion network Cancer invasion is a cell- and tissue-driven process which involves changes in the dynamics of cell adhesion and migration through transient formation of focal adhesion complexes, degradation of ECM components via MMPs and detachment from ECM as the cell advances under the influence of migratory factors. The resulting reciprocal reprogramming of both the tumor cells and the surrounding tissue structures guide a series of molecular changes representing a shift toward mesenchymal traits39. Invasion, which is thus central to GBM progression is a multi-step cell- and tissue-driven process wherein different factors including growth factors, cell adhesion molecules and ECM proteins contributed by glioma initiating cells (GICs) with stem cell features, mature astrocytes and cells of myeloid origin, mainly microglia and macrophages within the brain microenvironment40-43.

The gene expression analysis by the TCGA group showed altered expression of several annexins, cadherins, and collagenases namely, COL4A1, COL3A1, COL5A2, that are involved in cell adhesion to the extracellular matrix and actin cytoskeleton dynamics. Regulation of glioblastoma invasion is mediated by receptor-initiated signaling events involving key elements such as the Rho family GTPases, including RAC, RhoA and CDC426. These GTPases stimulate cell movement by regulating cell morphology and actin dynamics. Integrins, which are a large family of cell surface receptors, mediate the interaction of tumor cells with their microenvironment and along with other receptors and adhesion molecules regulate cell-cell and cell-matrix interactions. As a result, they bring about the activation of diverse intracellular signalling pathways which include PI3K, Akt, mTOR and MAP kinases. In the process, glioma cells also secrete proteolytic enzymes that assist in cleaving cell surface adhesion molecules, including CD44 and L1, in order to detach from the tumor mass6. Some of the important proteases produced by glioma cells include uPA, ADAMs and MMPs. We see several integrins like ITGB1, ITGB2, ITGB4, ITGB5, ITGB8, ITGAV, ITGA5, ITGAM that are altered in GBM tissues along with a variety of neural cell adhesion molecules such as NCAM1, NCAM2, L1CAM and NRCAM, membrane proteins, such as CD47 and NEGR144,45 and ARGHDIA, a regulator of Rho GTPases46 with known role in cell invasion and migration. Interestingly, Reyes et al. showed that αvβ8 integrin drives GBM cell invasion via association with ARGHDIA47. Several ECM components, such as tenascin (TNC), vitronectin (VTN), osteopontin (SPP1) and SPARC, as well as matrix metalloproteinases and collagenases required for degradation and remodelling of integrin 10


Page 10 of 27

Page 11 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


network or actin cytoskeletal components have been found to be upregulated at the edge of the advancing tumor39. Apart from these, elevated levels of proteinases/peptidases such as cathepsin B (CTSB)48 and MMP949,50, collagenases (COL3A1, COL18A1, COL6A3), laminins (LAMA4, LAMB1, LAMC1), filamins (FLNA, FLNB, FLNC), actin (ACTN1) and fibronectin (FN1) have also been observed. Many of the adhesion mediators involved in cellcell contacts were found to be concordantly down regulated in both proteomic and TCGA transcriptomic GBM datasets9,13,18, 25.

These proteins together may modulate cell adhesion and migration through interactions with ECM components via binding to integrins and with adjacent cells via members of the cadherin family51, resulting in disruption of cell-cell contacts, changes in adhesion properties and detachment of cells from the cytoskeletal matrix and cell morphology39,52. The dynamic regulation of this process is revealed by the integrin adhesome network proposed by ZaidelBar et al.53, consisting of adaptors, actin-binding proteins, integrins and the actin cytoskeleton. This network includes both overexpressed and underexpressed modulators underlying invasion process and contains 69 mRNA and protein concordant entities (out of 78 entities used initially) having more than 350 interactions (p-value = 1.00e-14). Further, we mapped the miRNAs dysregulated in GBM that may target these genes and examined miRNA and target pairs with inverse relationship in their expression for all the network entities as described under cell proliferation network. The analysis scheme, the number of miRNA-target interactions and those conforming to inverse relationship as per GBM data are shown in Supplementary Figure 2B. We identified experimentally validated miRNA-target pairs for 59 of the 69 entities used in the network, 51 of them with the expected inverse relationship accounting for >80% of the target entities included in the network as also observed in the proliferation network described above. Details of the miRNAs and their mRNA and protein targets in the invasion network are shown in Supplementary Table 4. Representative miRNA and target cascades for key interactors of the invasion network representing the 2D Map is shown in Figure 4B. The overexpression of miRNAs such as, miR-15554, miR-1618 and miR-2155 are seen to reciprocally target some of the key downregulated cell adhesion and migration factors such as DNM1, MAP2K1, ARHGDIA, and PPP2CA involved in the invasion process. On the other hand, several miRNAs - mir12456, mir-29b18 and mir-157 were found to be down regulated in GBM targeting proteins such as ITGB1, ITGAM, CD44, SPARC, FN1 and VIM that are overexpressed in GBM. 11



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The network and the 2D maps shown in Figure 3 and Figure 4 are restricted to molecules present in the proteomics data which is membrane centric, their interactors and miRNA regulators. Many other miRNA regulators and their targets (not represented in the figures) have been reported in other studies (see reviews39,43,52) particularly for invasion process. The networks and 2D maps may be extended to these molecules as well. A generalized scheme to generate 2D maps by integrating the expression data acquired at RNA/protein level and mapping their regulatory and functional relationships is shown in Figure 5 and can be applied to additional entities and processes important in the context of GBM and other cancers.

Implications of 2D maps The roadmap of cancer biomarker and target development has two distinct phases – first, an unbiased ‘Discovery Phase’ and second, a targeted ‘Validation Phase’ to translate leads from the first phase towards a specific clinical application. Selection of single molecular entities for verification through functional assays or molecular panels for experimental validation using expanded sample cohorts is challenging. High confidence molecular panels emerging from the discovery phase that can be extended to the second phase are thus important for their successful clinical applications. Stable as well as transient protein-protein interactions form the basis of many functional processes and regulatory mechanisms in the cells. These functional interactions are dynamic and regulated by many internal and external factors, therefore understanding their perturbations is important to decipher the changes related to disease processes including cancer. There is a rising interest in targeting protein-protein interaction networks with their associated signaling hubs and nodes that govern the acquisition of tumor hallmarks. The large volume of omics data has further enhanced the scope of protein-protein interaction (PPI) networks and their application for identification of biomarkers and targets58,59. Unlike stable interacting complexes which can be studied experimentally, transient interactions can only be predicted. The strength of these interactions is determined by several factors such as the functional and spatial proximity of genes60, their co-expression and co-occurrence with direct experimental support61. Improvement in protein interaction network in quality and coverage as well as regulatory linkages would add further value. We therefore hypothesized that the identification of 2D map possibility discussed

12


Page 12 of 27

Page 13 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


above would not only add strength to the existing network entities but also expand the landscape of interactions as regulatory interplay of miRNAs is further unraveled.

The cell proliferation and invasion modules along with miRNA-linked 2D Maps as discussed above offer a way of further consolidation of network data. For instance, invasion modulators, particularly integrins that are a combination of alpha and beta subunits, are being explored as therapeutic targets in GBM (Reviewed in62,63). Cilengitide, a pentapeptide and integrin antagonist that selectively and potently inhibits the activation of αVβ3 (ITGAV and ITGB3) and αVβ5 (ITGAV and ITGB5) integrins, is in Phase III clinical trials in GBM63-65. ITGAV and ITGB3 form the main hubs in the invasion network whereas ITGB5 integrins was not present in our datasets and not included in the network although there is evidence of its significance in GBM66. Similarly α5β1 integrin (consisting of ITGA5 and ITGB1 that are part of the invasion network) - a fibronectin receptor, is associated with worse prognosis and is being targeted by antibody drug Volociximab and peptide antagonist ATN-161 that is in phase I clinical trials67. These targets are overexpressed in several solid tumors including our proteomic analysis and TCGA analysis of GBM and also conform to 2D Maps. This can be expanded and more such entities that form the main hubs or take key positions in the functional networks and 2D maps discussed here may be explored further as targets for developing new therapies.

The 2-dimensional molecular maps with regulatory and functional linkages that reside within the interaction networks underlying tumor-associated processes would thus add a higher degree of strength and confidence for the entities in the network. Candidate molecules that have the most interactors with two-dimensional regulatory and functional linkages may score higher biological significance in the subsequent investigations for functional assays and clinical applications. There is also possibility that one miRNA can regulate other miRNAs in the nucleus suggesting cooperativity among the miRNAs68. 2D maps would reveal leads to predict and explore this kind of possible connectivity among miRNAs in the network. In addition, there are instances where miRNAs may exclusively target specific splice variants69 or effect the expression of splicing regulators70. The 2D maps could be expanded to include alternate splice forms that have the potential to enhance the protein-protein interaction (PPI) networks and can influence the proteomic and regulatory landscape71,72, inturn impacting the disease networks. With these possibilities, the spectrum and value of 2D Map in the network 13



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

perspective would further increase. Given this complex interplay among miRNA and between miRNA and their targets, it is not surprising that integration of all levels of omics information would generate a more definitive platform for selection of clinical targets.

CONCLUSIONS Our analysis highlights the feasibility of effective integration of transcriptomics and proteomics data on tumors to relate miRNAs and their target mRNA and proteins to connect regulatory and functional contexts. Using 2D maps built this way, we can now look at the combined effect of the changes observed across different omics levels including their regulators acting in concert. Dysregulation of miRNAs is observed in various cancers including gliomas and integration of the miRNA with their mRNA and protein targets combined with the network analysis of the targets, identifies the complex biological relationships between them. The 2D maps thus, a) contribute an added confidence and strength to the discovery panels b) strengthen the network view of the tumors by including the regulators of gene expression and c) help in prioritizing entities to develop clinical assays at three different levels of gene expression - regulatory miRNA or their target mRNA or protein, with at least two of them amenable to detection as circulatory molecules.

ACKNOWLEDGMENTS The work reported here was carried out under the financial support to Ravi Sirdeshmukh from the Indian Council of Medical Research (ICMR), Govt. of India. MKG and SJ are the Ph.D. students registered under Manipal University, Manipal. MKG is also a recipient of Senior Research Fellowship from the Council of Scientific and Industrial Research (CSIR), Govt. of India.

Figure Legends: Figure 1: The histogram showing representative altered microRNAs and their mRNA and protein targets. The differentially expressed proteins from GBM13and the corresponding GBM transcripts from TCGA resource9,18 were mapped against significantly altered microRNAs in GBM9,18. Upregulated microRNAs and their targets are shown in (A) and down regulated microRNAs with their targets, in (B). The targets were derived using prediction program from miRWalk database as discussed in the text. The information on the expression of the microRNAs and their respective mRNA and protein targets was derived 14


Page 14 of 27

Page 15 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


from the differentially expressed transcript and protein datasets discussed in the text. Their inverse relationship is shown in the figure and all the targets are listed in the Supplementary Table 2.

Figure 2: A scheme of 2 - dimensional molecular maps with functional and regulatory linkage.

Figure 3: Cell proliferation associated protein interaction network and 2-dimensional molecular maps. A, represents the cell proliferation interaction network and B represents 2D map of key molecules from the cell proliferation network targeted by altered microRNAs miR-7, miR-128a and miR-218 in GBM. The network was constructed using STRING web resource as described in the text. Arrows represent the up or down regulation trends.

Figure 4: Cell Invasion associated protein interaction network and 2-dimensional molecular maps. A, represents the interaction network consisting of the cell adhesion and migratory molecules involved in cell invasion and B represents 2D map of key miRNAs and their targets with inverse relationship in their expression in GBM. miR-16, miR-155 and miR-21 for cell adhesion and miR-29b, miR-124a and miR-1 for migration are shown. The network was constructed using STRING web resource as described in the text. Up or down regulation trends are represented by arrows.

Figure 5: A scheme for integration of transcriptomic and proteomic data to map molecular networks and 2-dimensional molecular maps.

Supplementary Table Legends: Supplementary Table 1: Overlap of mRNAs and proteins from GBM data. Differentially altered mRNA and proteins observed in GBM were compared using the transcript dataset reported by TCGA group and Dong et.al9,18 and the protein dataset generated by us13. Comparison was done with reference to number of proteins in the dataset (n=710) as mentioned in the ‘Materials and Method’ section. Of the 710 altered proteins, 327 showed an overlap with altered mRNAs and 317 of the overlapping entities also showed concordance in the expression trend in the two empirical datasets. The Table also shows discordant overlapping entities in red. 15



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Supplementary Table 2: Identification of predicted microRNA regulators for mRNAprotein targets altered in GBM. The concordant mRNA and protein entities (n=317) from Supplementary Table 1 were used for the analysis. Using multiple prediction programs in miRWalk database, microRNAs-target pairs were predicted by mapping to altered microRNAs dataset identified in GBM as per criteria described under ‘Materials and Methods’ section. This resulted in identification of 16,253 predicted miRNA-target pairs that covered 312 of 317 differentially expressed mRNA-proteins and 194 of 199 microRNAs used for the analysis. More than 60% of the miRNAs-target pairs (n=9,817) conformed to the inverse trend in their expression and are shown in green in the Table.

Supplementary Table 3: mRNA and protein entities used for the Proliferation interaction network and their microRNA regulators. For the mRNA and protein entities present in the proliferation interaction network shown in Figure 3A, miRNA regulators were identified with reference to experimentally validated target mode in miRWalk database20. Details are discussed in the text under section ‘Functional Network analysis and 2D molecular maps’. Supplementary Table 4: mRNA and protein entities used for the Invasion interaction network and their microRNA regulators. For the mRNA and protein entities present in the invasion interaction network shown in Figure 4A, miRNA regulators were identified as in Supplementary Table 3, with reference to experimentally validated target mode in miRWalk database20. Details of the entities are provided in the text under section ‘Functional Network analysis and 2D molecular maps’.

Supplementary Figure Legends: Supplementary Figure 1: A pipeline for the identification of miRNA and their mRNA and protein targets in GBM data. The details of the analysis are described under ‘Results and discussion’ and given in Supplementary Tables 1 and 2.

Supplementary Figure 2: A pipeline for the identification of miRNA and their mRNA and protein targets in cell proliferation (A) and invasion (B) networks. The details of the

16


Page 16 of 27

Page 17 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


analysis are described under ‘Results and discussion’ and given in Supplementary Tables 3 and 4.

REFERENCES 1. Buckner, J. C.; Brown, P. D.; O'Neill, B. P.; Meyer, F. B.; Wetmore, C. J.; Uhm, J. H., Central nervous system tumors. Mayo Clin Proc 2007, 82, (10), 1271-86. 2. Wen, P. Y.; Kesari, S., Malignant gliomas in adults. N Engl J Med 2008, 359, (5), 492-507. 3. Gladson, C. L.; Prayson, R. A.; Liu, W. M., The pathobiology of glioma tumors. Annu Rev Pathol 2010, 5, 33-50. 4. Furnari, F. B.; Fenton, T.; Bachoo, R. M.; Mukasa, A.; Stommel, J. M.; Stegh, A.; Hahn, W. C.; Ligon, K. L.; Louis, D. N.; Brennan, C.; Chin, L.; DePinho, R. A.; Cavenee, W. K., Malignant astrocytic glioma: genetics, biology, and paths to treatment. Genes Dev 2007, 21, (21), 2683-710. 5. Gritsenko, P. G.; Ilina, O.; Friedl, P., Interstitial guidance of cancer invasion. J Pathol 2012, 226, (2), 185-99. 6. Kwiatkowska, A.; Symons, M., Signaling determinants of glioma cell invasion. Adv Exp Med Biol 2013, 986, 121-41. 7. Teodorczyk, M.; Martin-Villalba, A., Sensing invasion: cell surface receptors driving spreading of glioblastoma. J Cell Physiol 2010, 222, (1), 1-10. 8. Brennan, C. W.; Verhaak, R. G.; McKenna, A.; Campos, B.; Noushmehr, H.; Salama, S. R.; Zheng, S.; Chakravarty, D.; Sanborn, J. Z.; Berman, S. H.; Beroukhim, R.; Bernard, B.; Wu, C. J.; Genovese, G.; Shmulevich, I.; Barnholtz-Sloan, J.; Zou, L.; Vegesna, R.; Shukla, S. A.; Ciriello, G.; Yung, W. K.; Zhang, W.; Sougnez, C.; Mikkelsen, T.; Aldape, K.; Bigner, D. D.; Van Meir, E. G.; Prados, M.; Sloan, A.; Black, K. L.; Eschbacher, J.; Finocchiaro, G.; Friedman, W.; Andrews, D. W.; Guha, A.; Iacocca, M.; O'Neill, B. P.; Foltz, G.; Myers, J.; Weisenberger, D. J.; Penny, R.; Kucherlapati, R.; Perou, C. M.; Hayes, D. N.; Gibbs, R.; Marra, M.; Mills, G. B.; Lander, E.; Spellman, P.; Wilson, R.; Sander, C.; Weinstein, J.; Meyerson, M.; Gabriel, S.; Laird, P. W.; Haussler, D.; Getz, G.; Chin, L., The somatic genomic landscape of glioblastoma. Cell 2013, 155, (2), 462-77. 9. Chin, L.; Meyerson, M., Comprehensive genomic characterization defines human glioblastoma genes and core pathways. Nature 2008, 455, (7216), 1061-8. 10. Kalinina, J.; Peng, J.; Ritchie, J. C.; Van Meir, E. G., Proteomics of gliomas: initial biomarker discovery and evolution of technology. Neuro Oncol 2011, 13, (9), 926-42. 11. Verhaak, R. G.; Hoadley, K. A.; Purdom, E.; Wang, V.; Qi, Y.; Wilkerson, M. D.; Miller, C. R.; Ding, L.; Golub, T.; Mesirov, J. P.; Alexe, G.; Lawrence, M.; O'Kelly, M.; Tamayo, P.; Weir, B. A.; Gabriel, S.; Winckler, W.; Gupta, S.; Jakkula, L.; Feiler, H. S.; Hodgson, J. G.; James, C. D.; Sarkaria, J. N.; Brennan, C.; Kahn, A.; Spellman, P. T.; Wilson, R. K.; Speed, T. P.; Gray, J. W.; Meyerson, M.; Getz, G.; Perou, C. M.; Hayes, D. N., Integrated genomic analysis identifies clinically relevant subtypes of glioblastoma characterized by abnormalities in PDGFRA, IDH1, EGFR, and NF1. Cancer Cell 2010, 17, (1), 98-110. 12. Jayaram, S.; Gupta, M. K.; Polisetty, R. V.; Cho, W. C.; Sirdeshmukh, R., Towards developing biomarkers for glioblastoma multiforme: a proteomics view. Expert Rev Proteomics 2014, 11, (5), 621-39. 13. Polisetty, R. V.; Gautam, P.; Sharma, R.; Harsha, H. C.; Nair, S. C.; Gupta, M. K.; Uppin, M. S.; Challa, S.; Puligopu, A. K.; Ankathi, P.; Purohit, A. K.; Chandak, G. R.; 17



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Pandey, A.; Sirdeshmukh, R., LC-MS/MS analysis of differentially expressed glioblastoma membrane proteome reveals altered calcium signaling and other protein groups of regulatory functions. Mol Cell Proteomics 2012, 11, (6), M111 013565. 14. Wilhelm, M.; Schlegl, J.; Hahne, H.; Moghaddas Gholami, A.; Lieberenz, M.; Savitski, M. M.; Ziegler, E.; Butzmann, L.; Gessulat, S.; Marx, H.; Mathieson, T.; Lemeer, S.; Schnatbaum, K.; Reimer, U.; Wenschuh, H.; Mollenhauer, M.; Slotta-Huspenina, J.; Boese, J. H.; Bantscheff, M.; Gerstmair, A.; Faerber, F.; Kuster, B., Mass-spectrometry-based draft of the human proteome. Nature 2014, 509, (7502), 582-7. 15. Santarpia, L.; Nicoloso, M.; Calin, G. A., MicroRNAs: a complex regulatory network drives the acquisition of malignant cell phenotype. Endocr Relat Cancer 2010, 17, (1), F5175. 16. Sumazin, P.; Yang, X.; Chiu, H. S.; Chung, W. J.; Iyer, A.; Llobet-Navas, D.; Rajbhandari, P.; Bansal, M.; Guarnieri, P.; Silva, J.; Califano, A., An extensive microRNAmediated network of RNA-RNA interactions regulates established oncogenic pathways in glioblastoma. Cell 2011, 147, (2), 370-81. 17. Li, Z.; Rana, T. M., Therapeutic targeting of microRNAs: current status and future challenges. Nat Rev Drug Discov 2014, 13, (8), 622-38. 18. Dong, H.; Luo, L.; Hong, S.; Siu, H.; Xiao, Y.; Jin, L.; Chen, R.; Xiong, M., Integrated analysis of mutations, miRNA and mRNA expression in glioblastoma. BMC Syst Biol 2010, 4, 163. 19. Ovaska, K.; Laakso, M.; Haapa-Paananen, S.; Louhimo, R.; Chen, P.; Aittomaki, V.; Valo, E.; Nunez-Fontarnau, J.; Rantanen, V.; Karinen, S.; Nousiainen, K.; LahesmaaKorpinen, A. M.; Miettinen, M.; Saarinen, L.; Kohonen, P.; Wu, J.; Westermarck, J.; Hautaniemi, S., Large-scale data integration framework provides a comprehensive view on glioblastoma multiforme. Genome Med 2010, 2, (9), 65. 20. Dweep, H.; Gretz, N., miRWalk2.0: a comprehensive atlas of microRNA-target interactions. Nat Methods 2015, 12, (8), 697. 21. Lyustikman, Y.; Momota, H.; Pao, W.; Holland, E. C., Constitutive activation of Raf1 induces glioma formation in mice. Neoplasia 2008, 10, (5), 501-10. 22. Hermanson, M.; Funa, K.; Hartman, M.; Claesson-Welsh, L.; Heldin, C. H.; Westermark, B.; Nister, M., Platelet-derived growth factor and its receptors in human glioma tissue: expression of messenger RNA and protein suggests the presence of autocrine and paracrine loops. Cancer Res 1992, 52, (11), 3213-9. 23. Rollbrocker, B.; Waha, A.; Louis, D. N.; Wiestler, O. D.; von Deimling, A., Amplification of the cyclin-dependent kinase 4 (CDK4) gene is associated with high cdk4 protein levels in glioblastoma multiforme. Acta Neuropathol 1996, 92, (1), 70-4. 24. Weber, G. L.; Parat, M. O.; Binder, Z. A.; Gallia, G. L.; Riggins, G. J., Abrogation of PIK3CA or PIK3R1 reduces proliferation, migration, and invasion in glioblastoma multiforme cells. Oncotarget 2011, 2, (11), 833-49. 25. Heroux, M. S.; Chesnik, M. A.; Halligan, B. D.; Al-Gizawiy, M.; Connelly, J. M.; Mueller, W. M.; Rand, S. D.; Cochran, E. J.; LaViolette, P. S.; Malkin, M. G.; Schmainda, K. M.; Mirza, S. P., Comprehensive characterization of glioblastoma tumor tissues for biomarker identification using mass spectrometry-based label-free quantitative proteomics. Physiol Genomics 2014, 46, (13), 467-81. 26. LeBrun, D. G.; Li, M., MicroRNAs in Glioblastoma Multiforme: Profiling Studies and Therapeutic Impacts. Molecular and Cellular Pharmacology 2011, 3, (3). 27. Halatsch, M. E.; Schmidt, U.; Behnke-Mursch, J.; Unterberg, A.; Wirtz, C. R., Epidermal growth factor receptor inhibition for the treatment of glioblastoma multiforme and other malignant brain tumours. Cancer Treat Rev 2006, 32, (2), 74-89. 18


Page 18 of 27

Page 19 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


28. Brennan, C.; Momota, H.; Hambardzumyan, D.; Ozawa, T.; Tandon, A.; Pedraza, A.; Holland, E., Glioblastoma subclasses can be defined by activity among signal transduction pathways and associated genomic alterations. PLoS One 2009, 4, (11), e7752. 29. Gupta, M. K.; Jayaram, S.; Madugundu, A. K.; Chavan, S.; Advani, J.; Pandey, A.; Thongboonkerd, V.; Sirdeshmukh, R., Chromosome-centric human proteome project: deciphering proteins associated with glioma and neurodegenerative disorders on chromosome 12. J Proteome Res 2014, 13, (7), 3178-90. 30. Jayaram, S.; Gupta, M. K.; Shivakumar, B. M.; Ghatge, M.; Sharma, A.; Vangala, R. K.; Sirdeshmukh, R., Insights from Chromosome-Centric Mapping of Disease-Associated Genes: Chromosome 12 Perspective. J Proteome Res 2015. 31. Lakka, S. S.; Gondi, C. S.; Yanamandra, N.; Olivero, W. C.; Dinh, D. H.; Gujrati, M.; Rao, J. S., Inhibition of cathepsin B and MMP-9 gene expression in glioblastoma cell line via RNA interference reduces tumor cell invasion, tumor growth and angiogenesis. Oncogene 2004, 23, (27), 4681-9. 32. Mathew, L. K.; Skuli, N.; Mucaj, V.; Lee, S. S.; Zinn, P. O.; Sathyan, P.; Imtiyaz, H. Z.; Zhang, Z.; Davuluri, R. V.; Rao, S.; Venneti, S.; Lal, P.; Lathia, J. D.; Rich, J. N.; Keith, B.; Minn, A. J.; Simon, M. C., miR-218 opposes a critical RTK-HIF pathway in mesenchymal glioblastoma. Proc Natl Acad Sci U S A 2014, 111, (1), 291-6. 33. Mathew, L. K.; Huangyang, P.; Mucaj, V.; Lee, S. S.; Skuli, N.; Eisinger-Mathason, T. S.; Biju, K.; Li, B.; Venneti, S.; Lal, P.; Lathia, J. D.; Rich, J. N.; Keith, B.; Simon, M. C., Feedback circuitry between miR-218 repression and RTK activation in glioblastoma. Sci Signal 2015, 8, (375), ra42. 34. Jun, G. J.; Zhong, G. G.; Ming, Z. S., miR-218 inhibits the proliferation of glioma U87 cells through the inactivation of the CDK6/cyclin D1/p21 pathway. Oncol Lett 2015, 9, (6), 2743-2749. 35. Venkataraman, S.; Birks, D. K.; Balakrishnan, I.; Alimova, I.; Harris, P. S.; Patel, P. R.; Handler, M. H.; Dubuc, A.; Taylor, M. D.; Foreman, N. K.; Vibhakar, R., MicroRNA 218 acts as a tumor suppressor by targeting multiple cancer phenotype-associated genes in medulloblastoma. J Biol Chem 2013, 288, (3), 1918-28. 36. Song, L.; Huang, Q.; Chen, K.; Liu, L.; Lin, C.; Dai, T.; Yu, C.; Wu, Z.; Li, J., miR218 inhibits the invasive ability of glioma cells by direct downregulation of IKK-beta. Biochem Biophys Res Commun 2010, 402, (1), 135-40. 37. Kefas, B.; Godlewski, J.; Comeau, L.; Li, Y.; Abounader, R.; Hawkinson, M.; Lee, J.; Fine, H.; Chiocca, E. A.; Lawler, S.; Purow, B., microRNA-7 inhibits the epidermal growth factor receptor and the Akt pathway and is down-regulated in glioblastoma. Cancer Res 2008, 68, (10), 3566-72. 38. Papagiannakopoulos, T.; Friedmann-Morvinski, D.; Neveu, P.; Dugas, J. C.; Gill, R. M.; Huillard, E.; Liu, C.; Zong, H.; Rowitch, D. H.; Barres, B. A.; Verma, I. M.; Kosik, K. S., Pro-neural miR-128 is a glioma tumor suppressor that targets mitogenic kinases. Oncogene 2012, 31, (15), 1884-95. 39. Zhong, J.; Paul, A.; Kellie, S. J.; O'Neill, G. M., Mesenchymal migration as a therapeutic target in glioblastoma. J Oncol 2010, 2010, 430142. 40. Ortensi, B.; Setti, M.; Osti, D.; Pelicci, G., Cancer stem cell contribution to glioblastoma invasiveness. Stem Cell Res Ther 2013, 4, (1), 18. 41. Rath, B. H.; Fair, J. M.; Jamal, M.; Camphausen, K.; Tofilon, P. J., Astrocytes enhance the invasion potential of glioblastoma stem-like cells. PLoS One 2013, 8, (1), e54752.

19



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

42. Sliwa, M.; Markovic, D.; Gabrusiewicz, K.; Synowitz, M.; Glass, R.; Zawadzka, M.; Wesolowska, A.; Kettenmann, H.; Kaminska, B., The invasion promoting effect of microglia on glioblastoma cells is inhibited by cyclosporin A. Brain 2007, 130, (Pt 2), 476-89. 43. Cuddapah, V. A.; Robel, S.; Watkins, S.; Sontheimer, H., A neurocentric perspective on glioma invasion. Nat Rev Neurosci 2014, 15, (7), 455-65. 44. Rebres, R. A.; Kajihara, K.; Brown, E. J., Novel CD47-dependent intercellular adhesion modulates cell migration. J Cell Physiol 2005, 205, (2), 182-93. 45. Kim, H.; Hwang, J. S.; Lee, B.; Hong, J.; Lee, S., Newly Identified CancerAssociated Role of Human Neuronal Growth Regulator 1 (NEGR1). J Cancer 2014, 5, (7), 598-608. 46. Boulter, E.; Garcia-Mata, R.; Guilluy, C.; Dubash, A.; Rossi, G.; Brennwald, P. J.; Burridge, K., Regulation of Rho GTPase crosstalk, degradation and activity by RhoGDI1. Nat Cell Biol 2010, 12, (5), 477-83. 47. Reyes, S. B.; Narayanan, A. S.; Lee, H. S.; Tchaicha, J. H.; Aldape, K. D.; Lang, F. F.; Tolias, K. F.; McCarty, J. H., alphavbeta8 integrin interacts with RhoGDI1 to regulate Rac1 and Cdc42 activation and drive glioblastoma cell invasion. Mol Biol Cell 2013, 24, (4), 474-82. 48. Konduri, S.; Lakka, S. S.; Tasiou, A.; Yanamandra, N.; Gondi, C. S.; Dinh, D. H.; Olivero, W. C.; Gujrati, M.; Rao, J. S., Elevated levels of cathepsin B in human glioblastoma cell lines. Int J Oncol 2001, 19, (3), 519-24. 49. Sun, C.; Wang, Q.; Zhou, H.; Yu, S.; Simard, A. R.; Kang, C.; Li, Y.; Kong, Y.; An, T.; Wen, Y.; Shi, F.; Hao, J., Antisense MMP-9 RNA inhibits malignant glioma cell growth in vitro and in vivo. Neurosci Bull 2013, 29, (1), 83-93. 50. Veeravalli, K. K.; Rao, J. S., MMP-9 and uPAR regulated glioma cell migration. Cell Adh Migr 2012, 6, (6), 509-12. 51. Weber, G. F.; Bjerke, M. A.; DeSimone, D. W., Integrins and cadherins join forces to form adhesive networks. J Cell Sci 2011, 124, (Pt 8), 1183-93. 52. Wolfenson, H.; Lavelin, I.; Geiger, B., Dynamic regulation of the structure and functions of integrin adhesions. Dev Cell 2013, 24, (5), 447-58. 53. Zaidel-Bar, R.; Itzkovitz, S.; Ma'ayan, A.; Iyengar, R.; Geiger, B., Functional atlas of the integrin adhesome. Nat Cell Biol 2007, 9, (8), 858-67. 54. D'Urso, P. I.; D'Urso, O. F.; Storelli, C.; Mallardo, M.; Gianfreda, C. D.; Montinaro, A.; Cimmino, A.; Pietro, C.; Marsigliante, S., miR-155 is up-regulated in primary and secondary glioblastoma and promotes tumour growth by inhibiting GABA receptors. Int J Oncol 2012, 41, (1), 228-34. 55. Gabriely, G.; Wurdinger, T.; Kesari, S.; Esau, C. C.; Burchard, J.; Linsley, P. S.; Krichevsky, A. M., MicroRNA 21 promotes glioma invasion by targeting matrix metalloproteinase regulators. Mol Cell Biol 2008, 28, (17), 5369-80. 56. Fowler, A.; Thomson, D.; Giles, K.; Maleki, S.; Mreich, E.; Wheeler, H.; Leedman, P.; Biggs, M.; Cook, R.; Little, N.; Robinson, B.; McDonald, K., miR-124a is frequently down-regulated in glioblastoma and is involved in migration and invasion. Eur J Cancer 2011, 47, (6), 953-63. 57. Bronisz, A.; Wang, Y.; Nowicki, M. O.; Peruzzi, P.; Ansari, K. I.; Ogawa, D.; Balaj, L.; De Rienzo, G.; Mineo, M.; Nakano, I.; Ostrowski, M. C.; Hochberg, F.; Weissleder, R.; Lawler, S. E.; Chiocca, E. A.; Godlewski, J., Extracellular vesicles modulate the glioblastoma microenvironment via a tumor suppression signaling network directed by miR-1. Cancer Res 2014, 74, (3), 738-50. 58. Ivanov, A. A.; Khuri, F. R.; Fu, H., Targeting protein-protein interactions as an anticancer strategy. Trends Pharmacol Sci 2013, 34, (7), 393-400. 20


Page 20 of 27

Page 21 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


59. Wu, G.; Feng, X.; Stein, L., A human functional protein interaction network and its application to cancer data analysis. Genome Biol 2010, 11, (5), R53. 60. Thevenin, A.; Ein-Dor, L.; Ozery-Flato, M.; Shamir, R., Functional gene groups are concentrated within chromosomes, among chromosomes and in the nuclear space of the human genome. Nucleic Acids Res 2014, 42, (15), 9854-61. 61. Franceschini, A.; Szklarczyk, D.; Frankild, S.; Kuhn, M.; Simonovic, M.; Roth, A.; Lin, J.; Minguez, P.; Bork, P.; von Mering, C.; Jensen, L. J., STRING v9.1: protein-protein interaction networks, with increased coverage and integration. Nucleic Acids Res 2013, 41, (Database issue), D808-15. 62. Vehlow, A.; Cordes, N., Invasion as target for therapy of glioblastoma multiforme. Biochim Biophys Acta 2013, 1836, (2), 236-44. 63. Chamberlain, M. C.; Cloughsey, T.; Reardon, D. A.; Wen, P. Y., A novel treatment for glioblastoma: integrin inhibition. Expert Rev Neurother 2012, 12, (4), 421-35. 64. Carter, A., Integrins as target: first phase III trial launches, but questions remain. J Natl Cancer Inst 2010, 102, (10), 675-7. 65. Scaringi, C.; Minniti, G.; Caporello, P.; Enrici, R. M., Integrin inhibitor cilengitide for the treatment of glioblastoma: a brief overview of current clinical results. Anticancer Res 2012, 32, (10), 4213-23. 66. Schittenhelm, J.; Schwab, E. I.; Sperveslage, J.; Tatagiba, M.; Meyermann, R.; Fend, F.; Goodman, S. L.; Sipos, B., Longitudinal expression analysis of alphav integrins in human gliomas reveals upregulation of integrin alphavbeta3 as a negative prognostic factor. J Neuropathol Exp Neurol 2013, 72, (3), 194-210. 67. Schaffner, F.; Ray, A. M.; Dontenwill, M., Integrin alpha5beta1, the Fibronectin Receptor, as a Pertinent Therapeutic Target in Solid Tumors. Cancers (Basel) 2013, 5, (1), 27-47. 68. Tang, R.; Li, L.; Zhu, D.; Hou, D.; Cao, T.; Gu, H.; Zhang, J.; Chen, J.; Zhang, C. Y.; Zen, K., Mouse miRNA-709 directly regulates miRNA-15a/16-1 biogenesis at the posttranscriptional level in the nucleus: evidence for a microRNA hierarchy system. Cell Res 2012, 22, (3), 504-15. 69. Liang, W. C.; Wang, Y.; Xiao, L. J.; Wang, Y. B.; Fu, W. M.; Wang, W. M.; Jiang, H. Q.; Qi, W.; Wan, D. C.; Zhang, J. F.; Waye, M. M., Identification of miRNAs that specifically target tumor suppressive KLF6-FL rather than oncogenic KLF6-SV1 isoform. RNA Biol 2014, 11, (7), 845-54. 70. Makeyev, E. V.; Zhang, J.; Carrasco, M. A.; Maniatis, T., The MicroRNA miR-124 promotes neuronal differentiation by triggering brain-specific alternative pre-mRNA splicing. Mol Cell 2007, 27, (3), 435-48. 71. Resch, A.; Xing, Y.; Modrek, B.; Gorlick, M.; Riley, R.; Lee, C., Assessing the impact of alternative splicing on domain interactions in the human proteome. J Proteome Res 2004, 3, (1), 76-83. 72. Ellis, J. D.; Barrios-Rodiles, M.; Colak, R.; Irimia, M.; Kim, T.; Calarco, J. A.; Wang, X.; Pan, Q.; O'Hanlon, D.; Kim, P. M.; Wrana, J. L.; Blencowe, B. J., Tissue-specific alternative splicing remodels protein-protein interaction networks. Mol Cell 2012, 46, (6), 884-92.

------.------

21



Page 22 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 Figure 1: The histogram showing representative altered microRNAs and their mRNA and protein 54 55 targets. The differentially expressed proteins from GBM13 and the corresponding GBM transcripts from 56 TCGA resource9,18 were mapped against significantly altered microRNAs in GBM9,18. Upregulated 57 microRNAs and their targets are shown in (A) and down regulated microRNAs with their targets, in (B). The 58 targets were derived using prediction program from miRWalk database as discussed in the text. The 59 information on the expression of the microRNAs and their respective mRNA and protein targets was 60 ACS Paragonand Plusprotein Environment derived from the differentially expressed transcript datasets discussed in the text. Their inverse

relationship is shown in the Figure and all the targets are listed in the Supplementary Table 2.

Page 23 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42


Figure 2: A scheme of 2 - dimensional molecular maps with functional and regulatory linkage.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42

Page 24 of 27

Figure 3: Cell proliferation associated protein interaction network and 2-dimensional molecular maps. A, represents the cell proliferation interaction network and B represents 2D map of key molecules from the cell proliferation network targeted by altered microRNAs miR-7, miR-128a and miR-218 in GBM. The network was constructed using STRING web resource as described in the text. Arrows represent the up or down regulation trends. ACS Paragon Plus Environment

Page 25 of 27


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 Figure 4: Cell Invasion associated protein interaction network and 2-dimensional molecular maps. A, 35 represents the interaction network consisting of the cell adhesion and migratory molecules involved in cell 36 37 invasion and B represents 2D map of key miRNAs and their targets with inverse relationship in their 38 expression in GBM. miR-16, miR-155 and miR-21 for cell adhesion and miR-29b, miR-124a and miR-1 for 39 STRING web resource as described in the text. Up ACS Paragonusing Plus Environment 40 migration are shown. The network was constructed 41 or down regulation trends are represented by arrows. 42


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 Figure 5: A scheme for integration of 41 and 2-dimensional molecular maps. 42

Page 26 of 27

ACS Paragon Plusand Environment transcriptomic proteomic data to map molecular networks

Page 27 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42


2D molecular map

TOC_Graphical Abstract ACS Paragon Plus Environment

Transcriptomic and Proteomic Data Integration and ... - ACS Publications

Recommend Documents