scholarly journals Evaluation of colorectal cancer subtypes and cell lines using deep learning

2018 ◽  
Author(s):  
Jonathan Ronen ◽  
Sikander Hayat ◽  
Altuna Akalin

ABSTRACTColorectal cancer (CRC) is a common cancer with a high mortality rate and a rising incidence rate in the developed world. The disease shows variable drug response and outcome. Molecular profiling techniques have been used to better understand the variability between tumours as well as cancer models such as cell lines. Drug discovery programs use cell lines as a proxy for human cancers to characterize their molecular makeup and drug response, identify relevant indications and discover biomarkers. In order to maximize the translatability and the clinical relevance of in vitro studies, selection of optimal cancer models is imperative. We have developed a deep learning based method to measure the similarity between CRC tumors and other tumors or disease models such as cancer cell lines. Our method efficiently leverages multi-omics data sets containing copy number alterations, gene expression and point mutations, and learns latent factors that describe the data in lower dimension. These latent factors represent the patterns across gene expression, copy number, and mutational profiles which are clinically relevant and explain the variability of molecular profiles across tumours and cell lines. Using these, we propose a refined colorectal cancer sample classification and provide best-matching cell lines in terms of multi-omics for the different subtypes. These findings are relevant for patient stratification and selection of cell lines for early stage drug discovery pipelines, biomarker discovery, and target identification.

2019 ◽  
Vol 2 (6) ◽  
pp. e201900517 ◽  
Author(s):  
Jonathan Ronen ◽  
Sikander Hayat ◽  
Altuna Akalin

Colorectal cancer (CRC) is a common cancer with a high mortality rate and a rising incidence rate in the developed world. Molecular profiling techniques have been used to better understand the variability between tumors and disease models such as cell lines. To maximize the translatability and clinical relevance of in vitro studies, the selection of optimal cancer models is imperative. We have developed a deep learning–based method to measure the similarity between CRC tumors and disease models such as cancer cell lines. Our method efficiently leverages multiomics data sets containing copy number alterations, gene expression, and point mutations and learns latent factors that describe data in lower dimensions. These latent factors represent the patterns that are clinically relevant and explain the variability of molecular profiles across tumors and cell lines. Using these, we propose refined CRC subtypes and provide best-matching cell lines to different subtypes. These findings are relevant to patient stratification and selection of cell lines for early-stage drug discovery pipelines, biomarker discovery, and target identification.


2016 ◽  
Vol 39 (6) ◽  
pp. 545-558 ◽  
Author(s):  
Elisabetta Bigagli ◽  
Carlotta De Filippo ◽  
Cinzia Castagnini ◽  
Simona Toti ◽  
Francesco Acquadro ◽  
...  

PLoS ONE ◽  
2013 ◽  
Vol 8 (3) ◽  
pp. e59689 ◽  
Author(s):  
Anna C. Luca ◽  
Sabrina Mersch ◽  
René Deenen ◽  
Stephan Schmidt ◽  
Isabelle Messner ◽  
...  

Blood ◽  
2009 ◽  
Vol 114 (22) ◽  
pp. 832-832 ◽  
Author(s):  
Michael A Chapman ◽  
Jean-Philippe Brunet ◽  
Jonathan J Keats ◽  
Angela Baker ◽  
Mazhar Adli ◽  
...  

Abstract Abstract 832 We hypothesized that new therapeutic targets for multiple myeloma (MM) could be discovered through the integrative computational analysis of genomic data. Accordingly, we generated gene expression profiling and copy number data on 250 clinically-annotated MM patient samples. Utilizing an outlier statistical approach, we identified HOXA9 as the top candidate gene for further investigation. HOXA9 expression was particularly high in patients lacking canonical MM chromosomal translocations, and allele-specific expression analysis suggested that this overexpression was mono-allelic. Indeed, focal copy number amplifications at the HOXA locus were observed in some patients. Outlier HOXA9 expression was further validated in both a collection of 52 MM cell lines and 414 primary patient samples previously described. To further verify the aberrant expression of HOXA9 in MM, we performed quantitative RT-PCR, which confirmed expression in all MM patients and cell lines tested, with high-level expression in a subset. To further investigate the mechanism of aberrant HOXA9 expression, we interrogated the pattern of histone modification at the HOXA locus because HOXA gene expression is particularly regulated by such chromatin marks. Accordingly, immunoprecipitation studies showed an aberrantly low level of histone 3 lysine 27 trimethylation marks (H3K27me3) at the HOXA9 locus. H3K27me3 modification is normally associated with silencing of HOXA9 in normal B-cell development. As such, it appears likely that the aberrant expression of HOXA9 in MM is due at least in part to defects in histone modification at this locus. To determine the functional consequences of HOXA9 expression in MM, we performed RNAi-mediated knock-down experiments in MM cell lines. Seven independent HOXA9 shRNAs that diminished HOXA9 expression resulted in growth inhibition of 12/14 MM cell lines tested. Taken together, these experiments indicate that HOXA9 is essential for survival of MM cells, and that the mechanism of HOXA9 expression relates to aberrant histone modification at the HOXA9 locus. The data thus suggest that HOXA9 is an attractive new therapeutic target for MM. Disclosures: No relevant conflicts of interest to declare.


Blood ◽  
2016 ◽  
Vol 128 (22) ◽  
pp. 4095-4095
Author(s):  
Alyssa Bouska ◽  
Chengfeng Bi ◽  
Waseem Lone ◽  
Weiwei Zhang ◽  
Ambreen Kedwaii ◽  
...  

Abstract Burkitt lymphoma (BL) is the most common non-Hodgkin lymphoma (NHL) in children. Although it accounts for only 1-5% of NHL in adults, approximately 60% of the BL cases diagnosed each year in western countries occur in patients >40 years of age. Although adult and pediatric BL cases are indistinguishable by molecular classification, pediatric patients have a significantly better outcome than adults. While translocation of MYC to the immunoglobulin heavy or light chain genes is characteristic of pediatric and adult BL, genetic differences may contribute to the superior clinical outcome of childhood cases. Therefore, we aimed to identify the spectrum of additional genetic abnormalities that occur in adult and pediatric BL. Copy number analysis, gene expression profiling (GEP), and targeted sequencing of ~400 genes known to be mutated in NHLs were performed on a cohort of molecularly defined BL samples. Copy number abnormalities (CNAs) were identified by the Affymetrix 250k NspI SNP array in 73 BL tumors (28 adult, 45 pediatric), and sequencing was performed on 52 BLs (21 adult, 31 pediatric). Pediatric cases had fewer CNAs than adults. The most common focal abnormality identified was a gain on 13q31.3 encompassing MIR17HG. It was more frequent in adult compared to pediatric cases (35% vs 16%, p=0.085) and was associated with increased expression of miR-17~92 cluster members; and among adults, patients with this gain trended towards worse overall survival, though the number of cases with available information was small. Gain of 8q was found in ~20% of adult cases, but in no pediatric cases. Surprisingly, cases with 8q gain had significantly lower MYC mRNA expression (p< 0.001) and lower protein expression. In cases with MYC gain 0/4 cases were positive for MYC protein expression by immunohistochemistry; in contrast,6/10 cases with no MYC gain were positive for MYC expression. This suggests that gain of 8q is driven by another gene in the region. Additional genetic alterations included gains of genomic loci encompassing MCL1 and MDM4 (1q21-24) and losses encompassing RB1, p53 and CDKN2A/CDKN2B. Pathway analysis of genes differentially expressed by CN status showed an enrichment of genes involved in cell cycle regulation, the p53 signaling pathway, and the ubiquitin proteasome pathway. The frequencies of mutations in commonly mutated genes including MYC, ID3, TP53, CCND3, DDX3X, ARID1A, and TCF3 were not significantly different in adult and pediatric BL. However, BCL2, (43%, p<0.001), ZFHX3 (24%, p<0.01), SPTBN5 (20%, p=0.02), RB1 (14%, p=0.06), BTG1 (14%, p=0.06), TCF4 (14%, p=0.06), and TNFRSF14 (14%, p=0.06), were exclusively mutated in adult BL. In contrast, mutations in CDH23 (29% vs 5%, p=0.04) and SMARCA4 (35% vs 19%, p=0.05) were more frequent in pediatric BL. When mutations were placed into oncogenic pathways, mutations in genes regulating the PI3K-AKT pathway did not shown significant differences between adult and pediatric cases. Mutations promoting tonic BCR signaling (TCF3 and ID3) by activation of the PI3K pathway had similar frequencies in the two age groups, however, BCR signaling effectors inducing chronic active NF-kB signaling (CD79A, SYK, MYD88, BCL10, CARD11) were significantly associated with adult BL (adult cases with any mutation: 19% vs 7%). Gene expression studies suggested activated BCR signaling in BL cases with CN gain of miR-17~92. In vitro analysis of miR17~92 in BL cell lines (n=4) showed that functional loss of miR-17 ~ 92 expression using a miRNA sponge led to reduced proliferation. Treatment of BL cell lines with anti-IgM induced BCR activation in a time- and dosage-dependent manner, as estimated by increased phosphorylation of downstream mediators (SYK and BLNK). This activation was reduced upon loss of functional miR-17 ~ 92 expression in cell lines. Since miR-17~92 can directly inhibit proximal negative regulators of BCR signaling, treatment with the FDA-approved BTK inhibitor, Ibrutnib, further inhibited proliferation of BL cell lines carrying the miRNA sponge. The BCR signaling pathway is one example of how unique abnormalities in adult BL can provide possible targets for therapeutic intervention. Disclosures No relevant conflicts of interest to declare.


2002 ◽  
Vol 1 (1) ◽  
pp. 26-34 ◽  
Author(s):  
Matthias Futschik ◽  
Aaron Jeffs ◽  
Sharon Pattison ◽  
Nikola Kasabov ◽  
Michael Sullivan ◽  
...  

2013 ◽  
Vol 31 (15_suppl) ◽  
pp. e14544-e14544
Author(s):  
Eva Budinska ◽  
Jenny Wilding ◽  
Vlad Calin Popovici ◽  
Edoardo Missiaglia ◽  
Arnaud Roth ◽  
...  

e14544 Background: We identified CRC gene expression subtypes (ASCO 2012, #3511), which associate with established parameters of outcome as well as relevant biological motifs. We now substantiate their biological and potentially clinical significance by linking them with cell line data and drug sensitivity, primarily attempting to identify models for the poor prognosis subtypes Mesenchymal and CIMP-H like (characterized by EMT/stroma and immune-associated gene modules, respectively). Methods: We analyzed gene expression profiles of 35 publicly available cell lines with sensitivity data for 82 drug compounds, and our 94 cell lines with data on sensitivity for 7 compounds and colony morphology. As in vitro, stromal and immune-associated genes loose their relevance, we trained a new classifier based on genes expressed in both systems, which identifies the subtypes in both tissue and cell cultures. Cell line subtypes were validated by comparing their enrichment for molecular markers with that of our CRC subtypes. Drug sensitivity was assessed by linking original subtypes with 92 drug response signatures (MsigDB) via gene set enrichment analysis, and by screening drug sensitivity of cell line panels against our subtypes (Kruskal-Wallis test). Results: Of the cell lines 70% could be assigned to a subtype with a probability as high as 0.95. The cell line subtypes were significantly associated with their KRAS, BRAF and MSI status and corresponded to our CRC subtypes. Interestingly, the cell lines which in matrigel created a network of undifferentiated cells were assigned to the Mesenchymal subtype. Drug response studies revealed potential sensitivity of subtypes to multiple compounds, in addition to what could be predicted based on their mutational profile (e.g. sensitivity of the CIMP-H subtype to Dasatinib, p<0.01). Conclusions: Our data support the biological and potentially clinical significance of the CRC subtypes in their association with cell line models, including results of drug sensitivity analysis. Our subtypes might not only have prognostic value but might also be predictive for response to drugs. Subtyping cell lines further substantiates their significance as relevant model for functional studies.


Sign in / Sign up

Export Citation Format

Share Document