MicroRNA profiling differentiates colorectal cancer ...

1 downloads 0 Views 130KB Size Report
Sep 15, 2011 - group of samples were excluded. .... 5p in the mutated KRAS group (q < 0.05; Ta- ble 2). .... tory factor (TMF) is a Golgi-associated protein.
GENES, CHROMOSOMES & CANCER 51:1–9 (2012)

RESEARCH ARTICLES

MicroRNA Profiling Differentiates Colorectal Cancer According to KRAS Status Neda Mosakhani,1 Virinder Kaur Sarhadi,1 Ioana Borze,1 Marja-Liisa Karjalainen-Lindsberg,1 Jari Sundstro¨m,2 ¨ sterlund,4 and Sakari Knuutila1* Raija Ristama¨ki,3 Pia O 1 Department of Pathology,Haartman Institute and HUSLAB,University of Helsinki and Helsinki University Central Hospital,Helsinki, Finland 2 Department of Pathology,University of Turku,Turku University Central Hospital,Turku,Finland 3 Department of Oncology and Radiotherapy,Turku University Hospital,Turku,Finland 4 Department of Oncology,University of Helsinki and Helsinki University Central Hospital,Helsinki,Finland

Recent studies have shown the important role of microRNAs (miRNAs) in a variety of biological processes, and in its ability to distinguish tumors according to their prognostic and predictive properties. To identify miRNA signatures associated with colorectal carcinoma (CRC) and with KRAS status, we studied, using Agilent’s miRNA microarrays, miRNA expression in primary tumors from 55 metastatic CRC patients, including 15 with mutant and 40 with wild-type KRAS. Comparing these with normal colon tissue, we identified 49 miRNAs—including 19 novel miRNAs—significantly deregulated in tumor tissue. The presence of the KRAS mutation was associated with up-regulation of miR-127-3p, miR-92a, and miR-486-3p and down-regulation of miR-378. Increased expression of miR-127-3p and miR-92a in KRAS mutant tumors was significantly confirmed by quantitative reverse transcriptase polymerase chain reaction (qRT-PCR) (P < 0.05). We identified some predicted target genes of differentially expressed miRNAs between mutated and wild-type KRAS, such as RSG3 and TOB1, which are involved in apoptosis and proliferation. Target prediction and pathway analysis suggest a possible role for deregulated miRNAs in nicotinamide adenine dinucleotide phosphate (NADPH) C 2011 Wiley Periodicals, Inc. V regeneration and G protein-coupled receptor signaling pathways.

INTRODUCTION

Colorectal carcinoma (CRC) is one of the most common human malignancies, with a worldwide incidence of 1.2 million new cases in 2007. Recent targeted therapy, especially with antiepidermal growth factor receptor (anti-EGFR) antibodies, has improved the metastatic CRC treatment results (Douillard et al., 2010; Van Cutsem et al., 2011). Only 10–20% of patients with metastatic CRC benefit, however, from antiEGFR treatment (Cunningham et al., 2004; Chung et al., 2005). When the genes which are in the signaling pathway downstream from EGFR, ones such as KRAS and BRAF, are mutated, the EGFR signaling pathway is deleteriously active regardless of the EGFR antibody treatment. Abnormal activation of EGFR signaling pathways can lead to abnormal cell growth, proliferation, and differentiation, and finally to tumor initiation, tumor progression, angiogenesis, and metastasis information (Salomon et al., 1995; Friday and Adjei, 2005). KRAS mutations are predictive of resistance to anti-EGFR treatment in such a way that a successful targeting of the EGFR axis with C 2011 Wiley Periodicals, Inc. V

Cetuximab or Panitumumab depends on the presence of wild-type KRAS (Lievre et al., 2006, 2008; Di Fiore et al., 2007; De Roock et al., 2008; Karapetis et al., 2008; Heinemann et al., 2009). KRAS plays an important role in transduction of extracellular signals from EGFR to downstream effectors involved in cell division, apoptosis, and differentiation of cells, and KRAS mutations occur in 35–45% of all colorectal cancers (Brink et al., 2003; Baldus et al., 2010). Identification of novel targets is therefore necessary for improving therapeutic strategies of CRC patients with mutated KRAS. Additional Supporting Information may be found in the online version of this article. Supported by: Helsinki and Uusimaa Hospital District (HUS EVO), The Academy of Finland, The Sigrid Juse´lius Foundation, The Finnish Cancer Organizations, and The Amgen Ab and Merck Oy. *Correspondence to: Sakari Knuutila, Department of Pathology, Haartman Institute, PO Box 21 (Haartmaninkatu 3), University of Helsinki, Helsinki FI-00014, Finland. E-mail: sakari.knuutila@helsinki.fi Received 19 July 2011; Accepted 8 August 2011 DOI 10.1002/gcc.20925 Published online 15 September 2011 in Wiley Online Library (wileyonlinelibrary.com).

2

MOSAKHANI ET AL.

microRNAs (miRNAs) are small, 19- to 25-nucleotide noncoding RNAs, which negatively regulate gene expression at transcriptional or posttranscriptional level (Bartel, 2004; Lim et al., 2005). miRNAs play an important role in the control of many biological processes such as cellular development, differentiation, proliferation, and apoptosis. Their role has been implicated also in tumor biology such as progression, invasion, and oncogenesis. miRNA profiling can distinguish tumors according to their histopathological, prognostic, and predictive characteristics (Iorio et al., 2005; Mattie et al., 2006; Lebanony et al., 2009). Moreover, compared with expression profiling of mRNA, miRNA profiling proves to be a more accurate method of classifying tumor subtypes (Esquela-Kerscher and Slack, 2006). A number of studies published recently focus on the significance of miRNAs in CRC development, classification, diagnosis, and prognosis (Michael et al., 2003; Bandres et al., 2006; Cummins et al., 2006; Volinia et al., 2006; Monzo et al., 2008; Earle et al., 2010). Little is known, however, about differences in miRNA expression between CRCs with and without the KRAS mutation. We therefore aimed to determine the miRNA signature associated with KRAS status in CRC patients. MATERIALS AND METHODS Patients and Materials

The primary tumor of 60 patients with metastatic CRC, including 30 females and 30 males were investigated in this study. Mean age of these patients was 61  11 years. The site of the primary tumor was the colon in 44 patients and rectum in 16. Formalin-fixed, paraffin-embedded (FFPE) tissue samples were obtained from surgeries on the primary tumors from Helsinki and Turku University Central Hospital patients prior to the start of any treatment. The study was approved by the HUS ethics committee, as no. 173/13/03/02/09. The tumor content of each sample was determined by pathologists at the Helsinki and Turku University Central Hospitals. The control sample was a commercially available colon tissue RNA R Human total RNA, sample (FirstChoiceV Applied Biosystems/Ambion, Austin, TX). Mutational Analysis of KRAS and BRAF

DNA was extracted from tumor tissue samples with the QIAamp DNA FFPE Mini Kit (Qiagen, Genes, Chromosomes & Cancer DOI 10.1002/gcc

Valencia, CA). The TheraScreen KRAS mutation kit (Qiagen, DxS, Manchester, UK) allowed detection of seven KRAS mutations in codons 12 and 13. The BRAF Mutation Test Kit (Qiagen, DxS, Manchester, UK) was used to detect the V600E mutation using a real-time PCR assay. BRAF mutation test was performed only in wildtype KRAS patients, because according to several reports, KRAS and BRAF mutations in CRCs are nearly mutually exclusive (Frattini et al., 2004). Five patients were excluded from further analysis due to BRAF mutations. RNA Extraction

The miRNeasy FFPE Mini Kit (Qiagen, Valencia, CA) was used to extract total RNA, including miRNA, according to the manufacturer’s instructions. The NanoDrop-1000 Spectrophotometer (Thermo Fisher Scientific, Wilmington, DE) allowed quantification of RNA, and Agilent’s Bioanalyzer was used to check the quality of both RNA and miRNA with the RNA 6000 chip for RNA and the small RNA chip for miRNA (Agilent Technologies, Palo Alto, CA). Labeling, Hybridization, Scanning, and Data Processing

We used Agilent’s miRNA Microarray system (V2), containing 723 human and 76 human viral miRNAs (Agilent Technologies, Santa Clara, CA) for miRNA profiling. FFPE samples are a proven source for miRNA profiling (Borze et al., 2011). Labeling and hybridization of RNA samples were performed according to Agilent’s protocol version 2.0, as described (Mosakhani et al., 2010). In brief, 100 ng of total RNA was treated with Calf Intestine Phosphatase for 30 min at 37  C. dimethyl sulfotide (DMSO) (100%) served for denaturation at 100  C for 5 min, after which the samples were immediately transferred to an icewater bath to prevent reannealing. They were then labeled with cyanine3-pCp by incubation with T4 RNA ligase for 2 hr at 16  C. After the labeling reaction, samples were vacuum-dried at medium heat and resuspended in nuclease-free water. Next, they were hybridized to the microarrays in Agilent SureHyb Chambers for 20 hr at 55  C. The microarrays were then washed with washing buffers and scanned with Agilent’s Scanner. The raw data were reprocessed with Agilent’s Feature Extraction Software with default parameters. Details of the miRNA preprocessing

3

miRNA IN KRAS MUTATED COLORECTAL CANCER

protocol are provided in the manufacturer’s reference guide for the Agilent Feature Extraction tool. Statistical Analysis

The statistical analysis of microarray data was carried out with the GeneSpring GX Analysis Software v11.0.2 (Agilent). Data were preprocessed by log 2 transformation, and normalization between all arrays was done by the 75th percentile method. The miRNAs not detectable in any samples or controls were excluded from analysis. miRNAs with a ratio of total gene signal/total gene error < 3 were considered as unexpressed. miRNAs not expressed in at least 100% of one group of samples were excluded. Significance of differential expression between two groups of samples was estimated by t-test for those miRNAs with at least a 1.5-fold reduced or increased mean expression level between the two groups. Only those microRNAs with an adjusted P value (q value) < 0.05 (Benjamini correction for multiple testing) were considered as significantly changed. Verification of Microarray Results with Quantitative reverse transcriptase polymerase chain reaction (RT-PCR)

Microarray expression profiles of the selected miRNAs miR-127-3p, miR-92a, and miR-378 were verified by RT-PCR. These miRNAs were selected, because they showed significant differences (fold change and/or low P value) in expression between KRAS mutant and wild-type tumors. Reverse transcription of RNA was performed with the miScript Reverse Transcription Kit (Qiagen, Valencia, CA), according to manufacturer’s guidelines. RT-PCR was performed on a Light-cycler, software v.3.5 (Roche Applied Science, Mannheim, Germany) by the SYBR Green miScript PCR system (Qiagen, Valencia, CA). Each reaction was performed in a 20-lL volume with 5 ng template cDNA. The primers for amplification of selected miRNAs and U6 were purchased from Qiagen. Cycling conditions for RT-PCR using capillary cyclers consisted of an initial incubation at 95  C for 15 min, followed by 40 cycles of 94  C for 15 sec, 55  C for 20 sec, and 72  C for 20 sec. PCR for each RNA sample was performed in duplicate, with a negative control (no template of cDNA) included in every run. Melting curve analysis was also performed to check for nonspecific

amplification. The snRNA U6 primer assay (Qiagen) served as a control for normalization. The relative quantification (RQ) for each miRNA, compared with U6 was calculated using equation 2DDCt. Expression data of miRNAs are represented as fold change (fold change ¼ log 2 RQ). Student’s t-test was used to evaluate statistically significant differences in miRNA expression between the two groups. P values less than 0.05 were considered statistically significant. Target Prediction and Pathway Analysis for Differentially Expressed miRNAs

We used the miRBase target prediction database (http://microrna.sanger.ac.uk), TargetScan (http://www.targetscan.org), miRanda (www. microRNA.org), mirTarget2 (http://mirdb.org/ miRDB), Tarbase (http://diana.cslab.ece.ntua.gr/ tarbase), and PICTAR (http://pictar.bio.nyu.edu) to search for predicted mRNA targets of the differentially expressed miRNAs. To reduce the number of false positives, only those mRNA targets were considered that were predicted by at least four of six programs. Predicted mRNA targets, of the differentially expressed miRNAs, were screened by chipster v1.4.7 (http:// chipster.csc.fi/) for known interactions and involvement in biological networks, by the hypergeometric test in the ConsensusPathDB (CPDB). The CPDB database integrates pathway and other functional interaction resources like Reactome, HumanCyc, BioCarta, Integrating network objects with hierarchies (INOH), and Pathway Interaction Database and provides a list of significant pathways (with P < 0.05) as output. RESULTS Mutational Profiling of KRAS and BRAF

We identified 15 KRAS mutated tumors and 45 wild-type KRAS tumors, including 5 with the BRAF mutation. We carried out miRNA analyses on 55 samples including 15 mutated KRAS and 40 wild-type KRAS and BRAF samples. Five patients with the BRAF mutation were not included in the miRNA analysis. Mutated KRAS samples included 12 females and 3 males, mean age 66  8 years. The site of the primary tumor was the colon in 10 patients and rectum in 5. Wild-type KRAS and BRAF samples came from 15 females and 25 males, mean age 60  11 years. The tumor location in 30 patients was the colon and in 10 patients the rectum. Tumor Genes, Chromosomes & Cancer DOI 10.1002/gcc

4

MOSAKHANI ET AL.

TABLE 1. Significantly Under- and Over-Expressed miRNAs in CRCs Compared with Normal Colon Tissue

Under-expressed

Over-expressed

miRNA

q value

P value

Fold change

hsa-miR-1 hsa-miR-195 hsa-miR-143 hsa-miR-363 hsa-miR-215 hsa-miR-31 hsa-miR-497 hsa-miR-28-5p hsa-miR-194 hsa-miR-30c hsa-miR-133b hsa-miR-139-5pa hsa-miR-192 hsa-miR-27b hsa-miR-30b hsa-miR-31*a hsa-miR-218 hsa-miR-24-1*a hsa-miR-378* hsa-miR-30a*a hsa-miR-582-5pa hsa-miR-192*a hsa-miR-30a hsa-miR-145*a hsa-miR-30e*a hsa-miR-9* hsa-miR-143*a hsa-miR-490-3pa hsa-miR-362-3pa hsa-miR-598a hsa-miR-10b hsa-miR-137 hsa-miR-122 hsa-miR-590-5pa hsa-miR-7-1*a hsa-miR-144* hsa-miR-133a hsa-miR-378 hsa-miR-365 hsa-miR-494 hsa-miR-513ba hsa-miR-500 hsa-miR-892ba hsa-miR-513a-5pa hsa-miR-513ca hsa-miR-21*a

0.03 0.03 0.04 0.02 0.02 0.03 0.05 0.05 0.03 0.03 0.01 0.01 0.04 0.05 0.05 0.02 0.03 0.02 0.01 0.02 0.02 0.04 0.01 0.04 0.04 0.01 0.05 0.02 0.04 0.01 0.05 0.02 0.03 0.02 0.02 0.04 0.04 0.02 0.04 0.01 0.01 0.01 0.02 0.02 0.05 0.04

0.002 0.003 0.005 7.8E4 0.001 0.002 0.007 0.008 0.002 0.002 1.6E4 1.4E4 0.005 0.009 0.009 0.001 0.003 9.01E4 6.1E5 8.5E4 0.001 0.004 2.2E4 0.005 0.006 7.9E5 0.008 0.001 0.004 1.9E4 0.008 0.001 0.003 0.001 0.002 0.006 0.005 0.001 0.006 1.3E4 2.8E4 7.3E5 0.001 8.6E4 0.009 0.005

60.2 35.7 26.7 25.4 22.3 19.0 17.2 15.9 15.7 15.2 14.5 14.1 13.8 12.6 12.6 12.0 10.9 10.6 10.4 9.9 9.6 8.8 8.4 8.3 8.1 7.9 7.9 6.8 6.6 6.4 6.2 5.9 5.7 5.6 4.6 4.3 4.1 4.0 3.2 25.2 15.6 12.7 12.5 10.3 7.8 4.7

a

Novel miRNAs in CRC.

content cells in both groups were similar (tumorcell content in the mutated type was 65  19%, and in the wild type, 59  20%). miRNA Expression Profile of Normal Colon and Tumor Tissue

Our analysis showed 46 miRNAs with significantly different expression in CRC samples from that of normal colon tissue (q < 0.05; Table 1). Of Genes, Chromosomes & Cancer DOI 10.1002/gcc

these, 39 miRNAs were significantly underexpressed in all CRC samples, with miR-122 being expressed exclusively in normal colon tissue. The remaining seven miRNAs (miR-494, miR-500, miR-513b, miR-892b, miR-513a-5p, miR-513c, and miR-21*) were significantly over-expressed in CRC samples. miR-1, miR-195, and miR-143 had the highest fold change among under-expressed miRNAs, and miR-494 had the highest fold change among over-expressed miRNAs.

miRNA IN KRAS MUTATED COLORECTAL CANCER

5

TABLE 2. Over- and Under-Expressed miRNAs in KRAS Mutated Group Versus Wild-Type KRAS Group miRNA hsa-miR-92a hsa-miR-127-3p hsa-miR-486-5p hsa-miR-378

P valuea (