Wang, Junwen

Abstract In clinical cancer treatment, genomic alterations would often affect the response of patients to anticancer drugs. Studies have shown that molecular features of tumors could be biomarkers predictive of sensitivity or resistance to anticancer agents, but the identification of actionable mutations are often constrained by the incomplete understanding of cancer genomes. Recent progresses of next-generation sequencing technology greatly facilitate the extensive molecular characterization of tumors and promote precision medicine in cancers. More and more clinical studies, cancer cell lines studies, CRISPR screening studies as well as patient-derived model studies were performed to identify potential actionable mutations predictive of drug response, which provide rich resources of molecularly and pharmacologically profiled cancer samples at different levels. Such abundance of data also enables the development of various computational models and algorithms to solve the problem of drug sensitivity prediction, biomarker identification and in silico drug prioritization by the integration of multiomics data. Here, we review the recent development of methods and resources that identifies mutation-dependent effects for cancer treatment in clinical studies, functional genomics studies and computational studies and discuss the remaining gaps and future directions in this area.

Abstract Summary We present MetaMarker, a pipeline for discovering metagenomic biomarkers from whole-metagenome sequencing samples. Different from existing methods, MetaMarker is based on a de novo approach that does not require mapping raw reads to a reference database. We applied MetaMarker on whole-metagenome sequencing of colorectal cancer (CRC) stool samples from France to discover CRC specific metagenomic biomarkers. We showed robustness of the discovered biomarkers by validating in independent samples from Hong Kong, Austria, Germany and Denmark. We further demonstrated these biomarkers could be used to build a machine learning classifier for CRC prediction. Availability and implementation MetaMarker is freely available at https://bitbucket.org/mkoohim/metamarker under GPLv3 license. Supplementary information Supplementary data are available at Bioinformatics online.

Abstract Lung cancer has the highest mortality across all cancers in the world. Epidermal growth factor receptor (EGFR) is commonly mutated in lung adenocarcinomas from Asian non-smoking females. These tumors usually respond to EGFR tyrosine kinase inhibitors (EGFR-TKI), but the outcomes vary in different patients. Though EGFR T790M, amplification of MET and other mechanisms have been shown to confer EGFR-TKI resistance, the molecular mechanisms in around 30% of patients are still unknown. To uncover novel candidate pathways of EGFR-TKI resistance, we analyzed the whole exome mutation profiles of pre-treatment, EGFR T790M-free lung adenocarcinomas with typical activating mutations (L858R, exon 19 deletions). Amongst 19 non-responding (NR) tumors without significant shrinkage after gefitinib or erlotinib treatment, 67 cancer-related genes showed coding region non-synonymous single nucleotide variations (SNV) or insertion-deletions (indels) but they were not detected in 22 responding tumors. Gene set enrichment analysis (GSEA) showed 14 of the 67 genes were involved in the EGFR and WNT signaling network (false discovery rate q=1.14 e-8 for EGFR and q=4.3 e-9 for WNT), including well-known TKI resistance-related genes PIK3CA, PTEN and NF1. Notably, ARID1A, which encodes a subunit of the chromatin remodeling complex, showed nonsense mutations in 2 NR but not responding tumors, which were validated by Sanger sequencing. More than 40% of ARID1A mutations in lung adenocarcinomas recorded in COSMIC were nonsense mutations or frameshift indels. Loss of function mutations of ARID1A have also been reported in various cancers, together suggesting ARID1A might be a tumor suppressor but its involvement in EGFR mutant lung cancers is unknown. To investigate the potential role of ARID1A, its expression was first studied using Kaplan-Meier plotter. Low expression of ARID1A is associated with significantly better overall survival than high expression level in lung adenocarcinoma patients (p=8.5 e-7). To test whether loss of ARID1A could cause TKI resistance, ARID1A was knocked down by shRNA in HCC827 lung cancer cell line harboring EGFR exon19 deletion. MTT cell viability assay using gefitinib showed increase of IC50. Western blot assay further showed ARID1A knockdown increased phosphorylation of AKT at Ser473. Together, the findings suggest loss of ARID1A could lead to EGFR bypass and promote activation of the downstream PI3K/AKT pathway, leading to EGFR TKI resistance. Overall, our findings suggest EGFR-TKI resistance mutations could be detected before TKI drug selection. Resistant gene mutations present in pre-treatment tumors could help to predict the treatment response. More specifically, our study identified a promising gene list involved in EGFR-TKI resistance, and suggested ARID1A loss could be a novel EGFR-TKI resistance mechanism. Citation Format: Xuyuan Gao, Hang Xu, James CM Ho, Oscar SH Chan, Feng Xu, Junwen Wang, Victor HF Lee, Vicky PC Tin, Zhijie Xiao, Siqi Wang, Judy WP Yam, Maria P Wong. EGFR-mutant lung adenocarcinomas mutation profiles reveal ARID1A might be a novel tyrosine kinase resistance pathway [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr LB-196.

Abstract Background Illumina paired-end sequencing has been increasingly popular for 16S rRNA gene-based microbiota profiling. It provides higher phylogenetic resolution than single-end reads due to a longer read length. However, the reverse read (R2) often has significant low base quality, and a large proportion of R2s will be discarded after quality control, resulting in a mixture of paired-end and single-end reads. A typical 16S analysis pipeline usually processes either paired-end or single-end reads but not a mixture. Thus, the quantification accuracy and statistical power will be reduced due to the loss of a large amount of reads. As a result, rare taxa may not be detectable with the paired-end approach, or low taxonomic resolution will result in a single-end approach. Results To have both the higher phylogenetic resolution provided by paired-end reads and the higher sequence coverage by single-end reads, we propose a novel OTU-picking pipeline, hybrid-denovo, that can process a hybrid of single-end and paired-end reads. Using high-quality paired-end reads as a gold standard, we show that hybrid-denovo achieved the highest correlation with the gold standard and performed better than the approaches based on paired-end or single-end reads in terms of quantifying the microbial diversity and taxonomic abundances. By applying our method to a rheumatoid arthritis (RA) data set, we demonstrated that hybrid-denovo captured more microbial diversity and identified more RA-associated taxa than a paired-end or single-end approach. Conclusions Hybrid-denovo utilizes both paired-end and single-end 16S sequencing reads and is recommended for 16S rRNA gene targeted paired-end sequencing data.

Abstract Lung cancer is the most lethal malignancy around the world. In Hong Kong, chronic tobacco use is prevalent in male but <20% of female patients smoke. Epidermal growth factor receptor (EGFR) mutations affect 30% and 67% of male and female lung adenocarcinomas, respectively. Treatment responses to tyrosine kinase inhibitors (TKI) are not uniform and all patients eventually develop drug resistance. Mutation screening of immediate pre-treatment tumors is the most ideal for assessing drug targets but in practice, excision specimens of primary or metastatic sites are more efficient in terms of tumor quantity, timeliness and patient acceptance. However, information on the applicability of this approach and correlation with the clinical outcome is limited. For this purpose, we compared the whole exome mutation profiles of 39 excised EGFR mutant cancers treated with first line TKI with the clinical response type, including 15 stage I/II and 24 stage III/IV tumors. They included 16 non-responders (NR) defined clinically as those with stable or progressive disease, and 23 responders (R) defined by partial or complete tumor shrinkage. Considering only the non-synonymous SNV and INDEL mutations of coding regions of known actionable targets and genes listed in both the CGC and COSMIC cancer databases, the tumors harbored 54 mutated genes including 26 recurrently mutated and 28 genes involving only 1 tumor. Excluding EGFR, TP53 was the most commonly mutated gene occurring in 25/39 (64.1%) cases. Most known resistant genes involved in the EGFR and bypass signaling network showed mutations only in the NR group including recurrent mutations of PTEN (3/16), PIK3CA (2/16) and NF1 (2/16), and single case mutations of AKT1, ALK, RAF1 and KDR. Two EGFR network candidates showed mutations in both the NR and R groups, including HGF (2/16 NR, 1/23 R) and ROS1 (1 case in either group). Four post-treatment (post-TKI) tumors of acquired resistance were also analyzed, all of which showed EGFR T790M while no pre-treatment NR or R tumor harbored this mutation. Notably, mutations in the β-catenin pathway were prominent in the NR and post-TKI tumors, including APC (2/16 NR), CTNNB1 (1/16 NR, 1/4 post-TKI) and c-MYC (1/16 NR, 1/4 post-TKI) while they were not detected in the R tumors. Also, nonsense mutations of ARID1A was observed in 2/16 NR but none of the R tumors. In summary, our findings revealed candidate TKI resistant mutations involving the EGFR and bypass signaling networks in pre-treatment excision specimens particularly in non-responding patients. The EGFR T790M was not detectable in pre-treatment samples but was prevalent in post-TKI treated cancers. While this study is limited by its small cohort size, the findings indicate deep sequencing analysis of pre-treatment excision specimens of EGFR-mutant lung adenocarcinomas is warranted for detection of resistant mutations and predicting treatment response. Note: This abstract was not presented at the meeting. Citation Format: Xu-yuan Gao, Hang Xu, James CM Ho, Oscar SH Chan, Feng Xu, Junwen Wang, Victor HF Lee, Vicky PC Tin, Zhijie Xiao, Siqi Wang, Judy WP Yam, Maria P. Wong. Whole-exome sequencing reveals tyrosine kinase-resistant mutations in pretreatment EGFR-mutant lung adenocarcinomas [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr LB-274. doi:10.1158/1538-7445.AM2017-LB-274

Abstract Oncogenic role of p21-activated protein 4 (PAK4) in relation to cell growth and migration through various signalling cascades has been well studied in multiple cancer types. Such role might be contributed largely by cytoplasmic PAK4 as most of the known PAK4 signalling events occur in the cytoplasm. However, recent studies have demonstrated that PAK4 is capable of shuttling between nucleus and cytoplasm. The precise function of nuclear PAK4 (nPAK4) remains substantially unexplored. We have previously shown that PAK4 can act as a transcription factor. Herein, we aim to identify nPAK4-regulated genes by delineating the PAK4-binding landscape in ovarian cancer genome and mapping the PAK4 transcriptome in ovarian cancer cells. PAK4-enriched genomic regions were identified by chromatin immunoprecipitation with a PAK4 antibody followed by next-generation sequencing (ChIP-seq) in OVCAR3 cells. PAK4 transcriptome data was generated using RNA extracted from SKOV3 cells stably expressing PAK4 in the Affymetrix GeneChIP® Human Transcriptome Array 2.0 system. Gene lists generated from both ChIP-seq and microarray dataset were subjected to pathway enrichment analysis by Partek® PathwayTM. Among the 20 most significantly enriched pathways from microarray studies, several pathways in which PAK4 has no known reported function by far have been identified. Notably, DNA repair pathways that are known to influence drug chemosensitivity in ovarian cancer, such as the nucleotide excision repair (p=0.008), homologous recombination (p=0.012), mismatch repair (p=0.018) and DNA replication (p=0.018) pathways were significantly enriched. In concordance with the findings, the nucleotide excision repair pathway was also significantly enriched (p<0.006) in the gene list derived from promoter-restricted PAK4 binding sites in the ovarian cancer genome. Among the many candidates that were deregulated by PAK4, cyclin-dependent kinase 7 (CDK7), a core component in the CAK subcomplex crucial to nucleotide excision repair pathway, was further validated. The transcription start site (TSS) of cyclin-dependent kinase 7 (CDK7) was bound by PAK4. CDK7 mRNA expression was upregulated upon PAK4 overexpression. Quantitative-PCR showed downregulated CDK7 mRNA expression upon siRNA-mediated knockdown of PAK4 in OVCAR3 and TUOS3. This study points to the possible role of PAK4 in regulating gene transcription as PAK4 is mapped to several core promoter/TSS regions in the ovarian cancer genome. Our findings also shed light into the potential role of PAK4 in regulating drug chemosensitivity via its transcriptional control on genes implicated in DNA repair pathways. We postulate CDK7 as a direct transcriptional target of PAK4 and thereby contributes to chemoresistance in ovarian cancer cells. Citation Format: Ivy Tsz-Lo Wong, Oscar Gee-Wan Wong, Yiming Qin, Junwen Wang, Annie Nga-Yin Cheung. Genome-wide profiling of PAK4 DNA-binding sites and transcriptome reveals its potential transcriptional control on DNA repair-related genes in ovarian cancer cells [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 2419. doi:10.1158/1538-7445.AM2017-2419

Tumor-initiating cells (TIC) are dynamic cancer cell subsets that display enhanced tumor functions and resilience to treatment but the mechanism of TIC induction or maintenance in lung cancer is not fully understood. In this study, we show the calcium pathway transcription factor NFATc2 is a novel regulator of lung TIC phenotypes, including tumorspheres, cell motility, tumorigenesis, as well as in vitro and in vivo responses to chemotherapy and targeted therapy. In human lung cancers, high NFATc2 expression predicted poor tumor differentiation, adverse recurrence-free and cancer-specific overall survivals. Mechanistic investigations identified NFATc2 response elements in the 3’ enhancer region of SOX2, and NFATc2/SOX2 coupling upregulates ALDH1A1 by binding to its 5’ enhancer. Through this axis, oxidative stress induced by cancer drug treatment is attenuated, leading to increased resistance in a mutation-independent manner. Targeting this axis provides a novel approach for the long-term treatment of lung cancer through TIC elimination.

BackgroundFunctional genomics employs dozens of OMICs technologies to explore the functions of DNA, RNA and protein regulators in gene regulation processes. Despite each of these technologies being powerful tools on their own, like the parable of blind men and an elephant, any one single technology has a limited ability to depict the complex regulatory system. Integrative OMICS approaches have emerged and become an important area in biology and medicine. It provides a precise and effective way to study gene regulations.ResultsThis article reviews current popular OMICs technologies, OMICs data integration strategies, and bioinformatics tools used for multi‐dimensional data integration. We highlight the advantages of these methods, particularly in elucidating molecular basis of biological regulatory mechanisms.ConclusionsTo better understand the complexity of biological processes, we need powerful bioinformatics tools to integrate these OMICs data. Integrating multi‐dimensional OMICs data will generate novel insights into system‐level gene regulations and serves as a foundation for further hypothesis‐driven research.