Functional annotation with expression validation identifies novel metastasis-relevant genes from post-GWAS risk loci in sporadic colorectal carcinomas.

BACKGROUND: Colorectal cancer (CRC) is the third highest incidence cancer and is the leading cause of cancer mortality worldwide. Metastasis to distal organ is the major cause of cancer mortality. However, the underlying genetic factors are unclear. This study aimed to identify metastasis-relevant genes and pathways for better management of metastasis-prone patients. METHODS: A case-case genome-wide association study comprising 2677 sporadic Chinese CRC cases (1282 metastasis-positive vs 1395 metastasis-negative) was performed using the Human SNP6 microarray platform and analysed with the correlation/trend test based on the additive model. SNP variants with association testing -log10 p value ≥5 were imported into Functional Mapping and Annotation (FUMA) for functional annotation. RESULTS: Glycolysis was uncovered as the top hallmark gene set. Transcripts from two of the five genes profiled, hematopoietic substrate 1 associated protein X 1 (HAX1) and hyaluronan-mediatedmotility receptor (HMMR), were significantly upregulated in the metastasis-positive tumours. In contrast to disease-risk variants, HAX1 appeared to act synergistically with HMMR in significantly impacting metastasis-free survival. Examining the subtype datasets with FUMA and Ingenuity Pathway Analysis (IPA) identified distinct pathways demonstrating sexual dimorphism in CRC metastasis. CONCLUSIONS: Combining genome-wide association testing with in silico functional annotation and wet-bench validation identified metastasis-relevant genes that could serve as features to develop subtype-specific metastasis-risk signatures for tailored management of patients with stage I-III CRC.

The orphan nuclear receptor NR0B2 could be a novel susceptibility locus associated with microsatellite-stable, APC mutation-negative early-onset colorectal carcinomas with metabolic manifestation.

Colorectal cancer (CRC) is a high incidence cancer and major cause of cancer mortality. Though disease-causing tumor suppressors for major syndromes are well characterized, about 10% of CRC is familial but without mutations in known tumor suppressors. We exhaustively screened 100 polyposis families for APC germline mutations and identified 13, which are APC mutation-negative, microsatellite-stable (MSS), and with undetectable mutation in known tumor suppressors. Whole exome sequencing in three probands uncovered two with germline frameshift NR0B2 mutations, c.293_301delTTGGGTTGGinsAC and c.227delT. Sanger Sequencing identified a third proband with NR0B2 c.157_166delCATCGCACCT frameshift mutation. All three mutations deleted the C-terminus activation/repression domain of NR0B2, thus are loss-of-function mutations. Real-time RT-PCR performed on tumor and matched mucosa of one patient revealed that NR0B2 downstream targets, SMAD3 was derepressed while GLI1 was downregulated in the colonic mucosa compared to healthy controls. Truncated NR0B2 molecule was predicted to have weakened binding with interacting partners SMAD3, GLI1, BCL2, and RXRα, implying perturbation of TGF-ß, Hedgehog, anti-apoptotic and nuclear hormone receptor signaling pathways. Immunostaining also revealed nuclear retention of the most severely truncated NR0B2 molecule compared to the wildtype. Microsatellite and sequencing analysis did not detect loss of wildtype allele in probands' tumors. The patient who acquired somatic KRAS mutation progressed rapidly whist the other two patients manifested with late-onset obesity and diabetes. We propose that haploinsufficiency of NR0B2 is associated with a novel CRC syndrome with metabolic phenotypes.

Genome-wide association study identified copy number variants associated with sporadic colorectal cancer risk.

BACKGROUND: Multiple single nucleotide polymorphisms (SNPs) have been associated with colorectal cancer (CRC) risk. The role of structural or copy number variants (CNV) in CRC, however, remained unclear. We investigated the role of CNVs in patients with sporadic CRC. METHODS: A genome-wide association study (GWAS) was performed on 1000 Singapore Chinese patients aged 50 years or more with no family history of CRC and 1000 ethnicity-matched, age-matched and gender-matched healthy controls using the Affymetrix SNP 6 platform. After 16 principal component corrections, univariate and multivariate segmentations followed by association testing were performed on 1830 samples that passed quality assurance tests. RESULTS: A rare CNV region (CNVR) at chromosome 14q11 (OR=1.92 (95% CI 1.59 to 2.32), p=2.7e-12) encompassing CHD8, and common CNVR at chromosomes 3q13.12 (OR=1.54 (95% CI 1.33 to 1.77), p=2.9e-9) and 12p12.3 (OR=1.69 (95% CI 1.41 to 2.01), p=2.8e-9) encompassing CD47 and RERG/ARHGDIB, respectively, were significantly associated with CRC risk. CNV loci were validated in an independent replication panel using an optimised copy number assay. Whole-genome expression data in matched tumours of a subset of cases demonstrated that copy number loss at CHD8 was significantly associated with dysregulation of several genes that perturb the Wnt, TP53 and inflammatory pathways. CONCLUSIONS: A rare CNVR at 14q11 encompassing the chromatin modifier CHD8 was significantly associated with sporadic CRC risk. Copy number loss at CHD8 altered expressions of genes implicated in colorectal tumourigenesis.

Amino-terminal p53 mutations lead to expression of apoptosis proficient p47 and prognosticate better survival, but predispose to tumorigenesis.

Whereas most mutations in p53 occur in the DNA-binding domain and lead to its functional inactivation, their relevance in the amino-terminal transactivation domain is unclear. We show here that amino-terminal p53 (ATp53) mutations often result in the abrogation of full-length p53 expression, but concomitantly lead to the expression of the amino-terminally truncated p47 isoform. Using genetically modified cancer cells that only express p47, we demonstrate it to be up-regulated in response to various stimuli, and to contribute to cell death, through its ability to selectively activate a group of apoptotic target genes. Target gene selectivity is influenced by K382 acetylation, which depends on the amino terminus, and is required for recruitment of selective cofactors. Consistently, cancers capable of expressing p47 had a better overall survival. Nonetheless, retention of the apoptotic function appears insufficient for tumor suppression, because these mutations are also found in the germ line and lead to Li-Fraumeni syndrome. These data from ATp53 mutations collectively demonstrate that p53's apoptosis proficiency is dispensable for tumor suppression, but could prognosticate better survival.

Analysis of colorectal cancer glyco-secretome identifies laminin ß-1 (LAMB1) as a potential serological biomarker for colorectal cancer.

The high mortality rate in colorectal cancer is mostly ascribed to metastasis, but the only clinical biomarker available for disease monitoring and prognosis is the carcinoembryonic antigen (CEA). However, the prognostic utility of CEA remains controversial. In an effort to identify novel biomarkers that could be potentially translated for clinical use, we collected the secretomes from the colon adenocarcinoma cell line HCT-116 and its metastatic derivative, E1, using the hollow fiber culture system, and utilized the multilectin affinity chromatography approach to enrich for the secreted glycoproteins (glyco-secretome). The HCT-116 and E1 glyco-secretomes were compared using the label-free quantitative SWATH-MS technology, and a total of 149 glycoproteins were differentially secreted in E1 cells. Among these glycoproteins, laminin ß-1 (LAMB1), a glycoprotein not previously known to be secreted in colorectal cancer cells, was observed to be oversecreted in E1 cells. In addition, we showed that LAMB1 levels were significantly higher in colorectal cancer patient serum samples as compared to healthy controls when measured using ELISA. ROC analyses indicated that LAMB1 performed better than CEA at discriminating between colorectal cancer patients from controls. Moreover, the diagnostic performance was further improved when LAMB1 was used in combination with CEA.

Genome-wide association study identifies a new SMAD7 risk variant associated with colorectal cancer risk in East Asians.

Genome-wide association studies (GWAS) of colorectal cancer (CRC) have been conducted primarily in European descendants. In a GWAS conducted in East Asians, we first analyzed approximately 1.7 million single-nucleotide polymorphisms (SNPs) in four studies with 1,773 CRC cases and 2,642 controls. We then selected 66 promising SNPs for replication and genotyped them in three independent studies with 3,612 cases and 3,523 controls. Five SNPs were further evaluated using data from four additional studies including up to 3,290 cases and 4,339 controls. SNP rs7229639 in the SMAD7 gene was found to be associated with CRC risk with an odds ratio (95% confidence interval) associated with the minor allele (A) of 1.22 (1.15-1.29) in the combined analysis of all 11 studies (p = 2.93 × 10(-11) ). SNP rs7229639 is 2,487 bp upstream from rs4939827, a risk variant identified previously in a European-ancestry GWAS in relation to CRC risk. However, these two SNPs are not correlated in East Asians (r(2) = 0.008) nor in Europeans (r(2) = 0.146). The CRC association with rs7229639 remained statistically significant after adjusting for rs4939827 as well as three additional CRC risk variants (rs58920878, rs12953717 and rs4464148) reported previously in this region. SNPs rs7229639 and rs4939827 explained approximately 1% of the familial relative risk of CRC in East Asians. This study identifies a new CRC risk variant in the SMAD7 gene, further highlighting the significant role of this gene in the etiology of CRC.

Targeting the 'Undruggable' Driver Protein, KRAS, in Epithelial Cancers: Current Perspective.

This review summarizes recent development in synthetic drugs and biologics targeting intracellular driver genes in epithelial cancers, focusing on KRAS, and provides a current perspective and potential leads for the field. Compared to biologics, small molecule inhibitors (SMIs) readily penetrate cells, thus being able to target intracellular proteins. However, SMIs frequently suffer from pleiotropic effects, off-target cytotoxicity and invariably elicit resistance. In contrast, biologics are much larger molecules limited by cellular entry, but if this is surmounted, they may have more specific effects and less therapy-induced resistance. Exciting breakthroughs in the past two years include engineering of non-covalent KRAS G12D-specific inhibitor, probody bispecific antibodies, drug-peptide conjugate as MHC-restricted neoantigen to prompt immune response by T-cells, and success in the adoptive cell therapy front in both breast and pancreatic cancers.

KRAS-specific antibody binds to KRAS protein inside colorectal adenocarcinoma cells and inhibits its localization to the plasma membrane.

Colorectal cancer (CRC) is the third highest incidence cancer and a leading cause of cancer mortality worldwide. To date, chemotherapeutic treatment of advanced CRC that has metastasized has a dismayed success rate of less than 30%. Further, most (80%) sporadic CRCs are microsatellite-stable and are refractory to immune checkpoint blockade therapy. KRAS is a gatekeeper gene in colorectal tumorigenesis. Nevertheless, KRAS is 'undruggable' due to its structure. Thus, focus has been diverted to develop small molecule inhibitors for its downstream effector such as ERK/MAPK. Despite intense research efforts for the past few decades, no small molecule inhibitor has been in clinical use for CRC. Antibody targeting KRAS itself is an attractive alternative. We developed a transient ex vivo patient-derived matched mucosa-tumor primary culture to assess whether anti-KRAS antibody can be internalized to bind and inactivate KRAS. We showed that anti-KRAS antibody can enter live mucosa-tumor cells and specifically aggregate KRAS in the cytoplasm, thus hindering its translocation to the inner plasma membrane. The mis-localization of KRAS reduces KRAS dwelling time at the site where it tethers to activate downstream effectors. We previously showed that expression of SOX9 was KRAS-mutation-dependent and possibly a better effector than ERK in CRC. Herein, we showed that anti-KRAS antibody treated tumor cells have less intense SOX9 cytoplasmic and nuclear staining compared to untreated cells. Our results demonstrated that internalized anti-KRAS antibody inhibits KRAS function in tumor. With an efficient intracellular antibody delivery system, this can be further developed as combinatorial therapeutics for CRC and other KRAS-driven cancers.

Metabotyping of human colorectal cancer using two-dimensional gas chromatography mass spectrometry.

Colorectal cancer (CRC) is the fourth most common cause of death from cancer in the world. The limitations of the currently available methods and biomarkers for CRC management highlight the necessity of finding novel markers. Metabonomics can be used to search for potential markers that can provide molecular insight into human CRC. The emergence of two-dimensional gas chromatography time of flight mass spectrometry (GC × GC/TOFMS) has comprehensively enhanced the metabolic space coverage of conventional GC/MS. In this study, a GC × GC/TOFMS was developed for the tissue-based global metabonomic profiling of CRC. A Pegasus GC × GC/TOFMS (Leco Corp., St. Joseph, MI, USA) system comprising an Agilent 7890 GC and Pegasus IV TOFMS was used for this purpose. An Agilent DB-1 (30 m × 250 µm × 0.25 µm) fused silica capillary column and a Restek Rxi®-17 (1 m × 100 µm × 0.10 µm) fused silica capillary column were used as the primary and secondary columns, respectively. The method was applied for global metabonomic profiling of matched CRC and normal tissues (n = 63) obtained from 31 CRC patients during surgery. An attempt was also made to compare GC × GC/TOFMS with GC/MS and NMR in similar application. The results showed that the metabotype associated with CRC is distinct from that of normal tissue and led to the identification of chemically diverse marker metabolites. Metabolic pathway mapping suggested deregulation of various biochemical processes such as glycolysis, Krebs cycle, osmoregulation, steroid biosynthesis, eicosanoid biosynthesis, bile acid biosynthesis, lipid, amino acid and nucleotide metabolism.

Metabolic profiling in colorectal cancer reveals signature metabolic shifts during tumorigenesis.

Colorectal cancer (CRC) arises as the consequence of progressive changes from normal epithelial cells through polyp to tumor, and thus is an useful model for studying metabolic shift. In the present study, we studied the metabolomic profiles using high analyte specific gas chromatography/mass spectrometry (GC/MS) and liquid chromatography tandem mass spectrometry (LC/MS/MS) to attain a systems-level view of the shift in metabolism in cells progressing along the path to CRC. Colonic tissues including tumor, polyps and adjacent matched normal mucosa from 26 patients with sporadic CRC from freshly isolated resections were used for this study. The metabolic profiles were obtained using GC/MS and LC/MS/MS. Our data suggest there was a distinct profile change of a wide range of metabolites from mucosa to tumor tissues. Various amino acids and lipids in the polyps and tumors were elevated, suggesting higher energy needs for increased cellular proliferation. In contrast, significant depletion of glucose and inositol in polyps revealed that glycolysis may be critical in early tumorigenesis. In addition, the accumulation of hypoxanthine and xanthine, and the decrease of uric acid concentration, suggest that the purine biosynthesis pathway could have been substituted by the salvage pathway in CRC. Further, there was a step-wise reduction of deoxycholic acid concentration from mucosa to tumors. It appears that to gain a growth advantage, cancer cells may adopt alternate metabolic pathways in tumorigenesis and this flexibility allows them to adapt and thrive in harsh environment.

KRAS mutation-independent downregulation of MAPK/PI3K signaling in colorectal cancer.

KRAS is a gatekeeper gene in human colorectal tumorigenesis. KRAS is 'undruggable'; hence, efforts have been diverted to inhibit downstream RAF/MEK/ERK and PI3K/Akt signaling. Nevertheless, none of these inhibitors has progressed to clinical use despite extensive trials. We examined levels of phospho-ERK1/2(T202/Y204) and phospho-Akt1/2/3(S473) in human colorectal tumor compared to matched mucosa with semi-quantitative near-infrared western blot and confocal fluorescence immunohistochemistry imaging. Surprisingly, 75.5% (25/33) of tumors had lower or equivalent phospho-ERK1/2 and 96.9% (31/32) of tumors had lower phospho-Akt1/2/3 compared to matched mucosa, irrespective of KRAS mutation status. In contrast, we discovered KRAS-dependent SOX9 upregulation in 28 of the 31 (90.3%) tumors. These observations were substantiated by analysis of the public domain transcriptomics The Cancer Genome Atlas (TCGA) and NCBI Gene Expression Omnibus (GEO) datasets and proteomics Clinical Proteomic Tumor Analysis Consortium (CPTAC) dataset. These data suggest that RAF/MEK/ERK and PI3K/Akt signaling are unlikely to be activated in most human colorectal cancer.

Ultra-pressure liquid chromatography/tandem mass spectrometry targeted profiling of arachidonic acid and eicosanoids in human colorectal cancer.

Cumulative evidence shows that eicosanoids such as prostaglandins, leukotrienes, thromboxanes and hydroxy eicosatetraenoic acids play an important role in associating inflammation with human colorectal cancer (CRC). In this study an ultra-pressure liquid chromatography/tandem mass spectrometry (UPLC/MS/MS) method was developed and validated for the targeted profiling of eight relevant eicosanoids and the major metabolic precursor, arachidonic acid (AA), in human colon. Multiple reaction monitoring (MRM) experiments were performed in negative electrospray ionization mode. The metabolites were separated using a C(18) column consisting of 1.7 µm ethylene-bridged hybrid particles (100 × 2.1 mm i.d.) and gradient elution (50 to 95% of solvent B) with a mobile phase comprising water (0.1% formic acid) [solvent A] and acetonitrile (0.1% formic acid) [solvent B] at a flow rate of 0.4 mL/min. The analysis time for each sample was 5.5 min. Our UPLC/MS/MS method demonstrated satisfactory validation results in terms of selectivity, sensitivity, matrix effect, linearity, extraction efficiency, intra- and inter-day precision, accuracy and autosampler stability. The method was applied for the clinical profiling of matched pairs of cancerous and normal colon mucosae obtained from eight colorectal cancer patients. Endogenous levels of AA and selected eicosanoids such as prostaglandin E(2) (PGE(2)), prostacyclin (PGI(2)) [assayed as its stable hydrolytic product 6-keto-prostaglandin(1α) (6-k PGF(1α))] and 12-hydroxy-5Z,8Z,10E,14Z-eicosatetraenoic acid (12-HETE) were found to be significantly different (p <0.05; paired t-test) between cancerous and normal mucosae.

Advances in colorectal cancer genomics and transcriptomics drive early detection and prevention.

Colorectal carcinoma (CRC) is a high incidence cancer and leading cause of cancer mortality worldwide. The advances in genomics and transcriptomics in the past decades have improved the detection and prevention of CRC in familial CRC syndromes. Nevertheless, the ultimate goal of personalized medicine for sporadic CRC is still not within reach due no less to the difficulty in integrating population disparity and clinical data to combat what essentially is a very heterogenous disease. This minireview highlights the achievement of the past decades and present possible direction in the hope of early detection and metastasis prevention for reducing CRC-associated morbidity and mortality.

Enhancer-derived long non-coding RNAs CCAT1 and CCAT2 at rs6983267 has limited predictability for early stage colorectal carcinoma metastasis.

Up-regulation of long non-coding RNAs (lncRNAs), colon-cancer associated transcript (CCAT) 1 and 2, was associated with worse prognosis in colorectal cancer (CRC). Nevertheless, their role in predicting metastasis in early-stage CRC is unclear. We measured the expression of CCAT1, CCAT2 and their oncotarget, c-Myc, in 150 matched mucosa-tumour samples of early-stage microsatellite-stable Chinese CRC patients with definitive metastasis status by multiplex real-time RT-PCR assay. Expression of CCAT1, CCAT2 and c-Myc were significantly up-regulated in the tumours compared to matched mucosa (p < 0.0001). The expression of c-Myc in the tumours was significantly correlated to time to metastasis [hazard ratio = 1.47 (1.10-1.97)] and the risk genotype (GG) of rs6983267, located within CCAT2. Expression of c-Myc and CCAT2 in the tumour were also significantly up-regulated in metastasis-positive compared to metastasis-negative patients (p = 0.009 and p = 0.04 respectively). Nevertheless, integrating the expression of CCAT1 and CCAT2 by the Random Forest classifier did not improve the predictive values of ColoMet19, the mRNA-based predictor for metastasis previously developed on the same series of tumours. The role of these two lncRNAs is probably mitigated via their oncotarget, c-Myc, which was not ranked high enough previously to be included in ColoMet19.

Germline bone morphogenesis protein receptor 1A mutation causes colorectal tumorigenesis in hereditary mixed polyposis syndrome.

OBJECTIVES: Hereditary mixed polyposis syndrome (HMPS) is characterized by polyps of mixed adenomatous/hyperplastic/atypical juvenile histology that are autosomal dominantly inherited and that eventually lead to colorectal cancer (CRC). Although CRC with adenomatous polyps is initiated by inactivating adenomatous polyposis coli (APC), the initiating event of CRC with mixed polyps remains unclear. We aimed to identify the underlying germline defect in HMPS. METHODS: We screened for bone morphogenesis protein receptor 1A (BMPR1A) mutation by exonic sequencing, reverse-transcriptase polymerase chain reaction (PCR) followed by cDNA sequencing, and multiplex ligation-dependent probe amplification (MLPA) analysis in eight Singapore Chinese HMPS families. RESULTS: Germline BMPR1A defects were found in four (50%) families. In two families, it is shown to co-segregate with the disease phenotype in all affected members over three generations, indicating that it is the disease-causing mutation. CRC incidence is 75%. The most defining characteristic is the presence of mixed hyperplastic-adenomatous polyps. Juvenile polyps are rarely reported, and if present, are usually of mixed components. Detailed histology of the polyps from one patient over 11 years distinguishes HMPS from juvenile polyposis syndrome (JPS). We report further the first cases of Wilms' tumor and papillary thyroid carcinoma associated with BMPR1A germline defect. CONCLUSIONS: Germline BMPR1A defect is the disease-causing mutation in 50% of the HMPS families. If patients present with mixed morphology polyps in the large bowel that are autosomal dominantly inherited and corresponding absence of upper gastrointestinal abnormalities, the gene to begin mutation screening should be BMPR1A rather than APC.

A susceptibility gene set for early onset colorectal cancer that integrates diverse signaling pathways: implication for tumorigenesis.

PURPOSE: The causative genes for autosomal dominantly inherited familial adenomatous polyposis (FAP) and hereditary nonpolyposis colorectal cancer have been well characterized. There is, however, another 10% to 15% of early onset colorectal cancers (CRC) in which the genetic components are unclear. In this study, we used microarray technology to systematically search for differentially expressed genes in early onset CRC. EXPERIMENTAL DESIGN: Young patients with non-FAP or non-hereditary nonpolyposis colorectal cancer, and healthy controls were age- (

Human colorectal cancer initiation is bidirectional, and cell growth, metabolic genes and transporter genes are early drivers of tumorigenesis.

The role of stem cells in the development of solid tumors remains controversial. In colorectal cancers (CRC), this is complicated by the conflicting "top-down" or "bottom-up" hypotheses of cancer initiation. We profiled the expressions of genes from the top (T) and bottom (B) crypt fractions of normal-appearing human colonic mucosa (M) at least 20 cm away from the tumor as a baseline and compared this to the genes of matched mucosa adjacent to tumors (MT) in twenty-three sporadic CRC patients. In thirteen patients, the genetic distance (M-MT) between the B fractions is smaller than the distance between the T fractions, indicating that the expressions diverge further in the top fractions (B < T). In the remaining patients, the reverse effect is observed (B > T). Assuming that a greater genetic divergence in the top or bottom fractions indicates that position as the initiation site, it is thus equally likely that human CRC initiates from 'top-down' via de-differentiated colonocytes or 'bottom-up' via dysregulated intestinal stem cells. Dysregulated genes that persist until tumor stage are not limited to tumor suppressors or oncogenes but include metabolic and transporter genes such as CA7, PHLPP2, and AQP8.

Chromosome 19q13 disruption alters expressions of CYP2A7, MIA and MIA-RAB4B lncRNA and contributes to FAP-like phenotype in APC mutation-negative familial colorectal cancer patients.

Familial adenomatous polyposis (FAP) is an autosomal-dominantly inherited form of colorectal cancer (CRC) caused by mutation in the adenomatous polyposis coli (APC) gene. Our ability to exhaustively screen for APC mutations identify microsatellite-stable and APC-mutation negative familial CRC patients, enabling us to search for novel genes. We performed genome-wide scan on two affected siblings of one family and 88 ethnicity- and gender-matched healthy controls to identify deletions shared by the siblings. Combined loss of heterozygosity, copy number and allelic-specific copy number analysis uncovered 5 shared deletions. Long-range polymerase chain reaction (PCR) confirmed chromosome 19q13 deletion, which was subsequently found in one other family. The 32 kb deleted region harbors the CYP2A7 gene and was enriched with enhancer, repressor and insulator sites. The wildtype allele was lost in the polyps of the proband. Further, real-time RT-PCR assays showed that expressions of MIA and MIA-RAB4B located 35 kb upstream of the deletion, were up-regulated in the polyps compared to the matched mucosa of the proband. MIA-RAB4B, the read-through long non-coding RNA (lncRNA), RAB4B, PIM2 and TAOK1 share common binding site of a microRNA, miR-24, in their 3'UTRs. PIM2 and TAOK1, two target oncogenes of miR-24, were co-ordinately up-regulated with MIA-RAB4B in the polyps, suggesting that MIA-RAB4B could function as competitive endogenous RNA to titrate miR-24 away from its other targets. The data suggest that the 19.13 deletion disrupted chromatin boundary, leading to altered expression of several genes and lncRNA, could contribute to colorectal cancer via novel genetic and epigenetic mechanisms.

A formalin-fixed paraffin-embedded (FFPE)-based prognostic signature to predict metastasis in clinically low risk stage I/II microsatellite stable colorectal cancer.

Approximately 20% early-stage (I/II) colorectal cancer (CRC) patients develop metastases despite curative surgery. We aim to develop a formalin-fixed and paraffin-embedded (FFPE)-based predictor of metastases in early-stage, clinically-defined low risk, microsatellite-stable (MSS) CRC patients. We considered genome-wide mRNA and miRNA expression and mutation status of 20 genes assayed in 150 fresh-frozen tumours with known metastasis status. We selected 193 genes for further analysis using NanoString nCounter arrays on corresponding FFPE tumours. Neither mutation status nor miRNA expression improved the estimated prediction. The final predictor, ColoMet19, based on the top 19 genes' mRNA levels trained by Random Forest machine-learning strategy, had an estimated positive-predictive-value (PPV) of 0.66. We tested ColoMet19 on an independent test-set of 131 tumours and obtained a population-adjusted PPV of 0.67 indicating that early-stage CRC patients who tested positive have a 67% risk of developing metastases, substantially higher than the metastasis risk of 40% for node-positive (Stage III) patients who are generally treated with chemotherapy. Predicted-positive patients also had poorer metastasis-free survival (hazard ratios [HR] = 1.92, design-set; HR = 2.05, test-set). Thus, early-stage CRC patients who test positive may be considered for adjuvant therapy after surgery.