Genetic variation and alterations of genes involved in NFκB/TNFAIP3- and NLRP3-inflammasome signaling affect susceptibility and outcome of colorectal cancer

(1)

Genetic variation and alterations of genes

involved in NFκB/TNFAIP3- and

NLRP3-inflammasome signaling affect susceptibility

and outcome of colorectal cancer

Jonas Ungerbäck, Dimitri Belenki, Aksa Jawad Ul-Hassan, Mats Fredrikson,

Karin Fransén, Nils Elander, Deepti Verma and Peter Söderkvist

Linköping University Post Print

N.B.: When citing this work, cite the original article.

This is a pre-copy-editing, author-produced PDF of an article accepted for publication in

Carcinogenesis following peer review. The definitive publisher-authenticated version.

Jonas Ungerbäck, Dimitri Belenki, Aksa Jawad Ul-Hassan, Mats Fredrikson, Karin Fransén,

Nils Elander, Deepti Verma and Peter Söderkvist, Genetic variation and alterations of genes

involved in NFκB/TNFAIP3- and NLRP3-inflammasome signaling affect susceptibility and

outcome of colorectal cancer, 2012, Carcinogenesis, (33), 11, 2126-2134.

is available online at

http://dx.doi.org/10.1093/carcin/bgs256

Copyright: Oxford University Press (OUP): Policy B

http://www.oxfordjournals.org/

Postprint available at: Linköping University Electronic Press

(2)

For Peer Review

Genetic variation and alterations of genes involved in NFκB/TNFAIP3- and

NLRP3-inflammasome signaling affect susceptibility and outcome of colorectal cancer

Jonas Ungerbäck1,#,*, Dimitri Belenki1,#, Aksa Jawad ul-Hassan1, Mats Fredrikson2, Karin Fransén3, Nils Elander1, Deepti Verma1, and Peter Söderkvist1

1

Division of Cell Biology, Department of Clinical and Experimental Medicine Faculty of Health Sciences, Linköping University, SE-581 85, Linköping, Sweden

2

Division of Occupational and Environmental Medicine, Department of Clinical and Experimental Medicine, Faculty of Health Sciences, Linköping University, SE-581 85, Linköping, Sweden

3

Department of Clinical Medicine, School of Health and Medical Sciences, Örebro University, SE-701 82, Örebro, Sweden

#

These authors contributed equally to this work.

Key words:

Colorectal cancer, inflammation, inflammasome, polymorphism, TNFAIP3,

*

Corresponding author: jonas.ungerback@liu.se

The authors declare no conflict of interest! 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

at Linkopings Universitet on August 13, 2012

http://carcin.oxfordjournals.org/

(3)

For Peer Review

Abstract

Colorectal tumors are continuously exposed to an inflammatory environment, which together with mitogenic signals sustains several cancer hallmarks. NFκB is a major regulator of inflammation and variation in NFκB-associated genes could potentially be used as biomarkers to identify patients with increased risk of colorectal cancer (CRC) development, and/or a rapidly progressing disease.

In this study 348 CRC cases and 806 randomly selected healthy individuals from southeastern Sweden were examined with regard to seven polymorphisms in NFκB pathway-associated genes. Log-rank-tests and Cox proportional hazard regression analysis examined the association between the polymorphisms and CRC-specific survival while chi-square tests and logistic regression analysis were used to test for associations between the polymorphisms and CRC-susceptibility. Gene expression and loss-of-heterozygosity analyses of TNFAIP3 were carried out in a subset of tumors to assess its role as a tumor suppressor in CRC. Heterozygous and polymorphic TNFAIP3 (rs6920220), heterozygous NLRP3

(Q705K) and polymorphic NFκB -94 ATTG ins/del genotypes were found to be associated with poorer

survival in patients diagnosed with invasive CRC (aHR=5.2, 95% CI 2.5-10.9, P<0.001). TNFAIP3 mRNA levels were significantly decreased in tumors compared to adjacent non-neoplastic mucosa (P<0.0001) and LOH of 6q23.3, (TNFAIP3), was detected in 17% of cases, while only 2.5% of the investigated specimens displayed TNFAIP3 gene mutations. We propose that TNFAIP3 (rs6920220),

NLRP3 (Q705K) and NFκB -94 ATTG ins/del polymorphisms are associated with poor survival in

patients with advanced CRC and may be used as prognostic markers. Experimental results indicate that

TNFAIP3 may act as a tumor suppressor in CRC.

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

(4)

For Peer Review

Introduction

Although for long poorly elucidated, a link between cancer and inflammation was early suspected due to pathological and clinical observations where the conditions seemed to commonly co-exist. Following the last decades of intense research and ever growing possibilities to experimentally investigate the cellular and molecular biology behind both inflammation and cancer, Hanahan and Weinberg recently took the next step and proposed tumor-promoting inflammation as an enabling characteristic of tumors in addition to their previously stated ‘Hallmarks of Cancer’ [1].

Colorectal cancer (CRC) is one of the leading causes of cancer related deaths with about 1.2 million new worldwide cases diagnosed each year [2]. Chronic inflammatory bowel diseases (IBDs), e.g. Crohn’s disease (CD) and ulcerative colitis (UC), strongly increase the risk of colitis-associated cancer (CAC) [3]. However, the molecular mechanisms linking inflammation and sporadic CRC are poorly understood, although it is generally believed that the inflammatory response follows tumor initiation (rather than the opposite) in the sporadic cases [3,4]. The importance of cancer-related inflammation is further emphasized by recent studies which reveal that tumor growth and progression depend on the activation of NFκB, a key controller of the innate immune response, both in the intestinal epithelium as well as in infiltrating inflammatory cells [3-5]. NFκB signaling in epithelial cells is evoked by a variety of signals, including the release of reactive oxygen or nitrogen species (RONS), presentation of antigens to the cell or through binding of cytokines, such as TNFα, IL-6 or IL-1β to their respective membrane receptors (reviewed in Hymowitz [6]). Major activators of canonical NFκB signaling are the Toll like/interleukin-1 receptor family (TLR; IL-1R) and the tumor necrosis factor (TNF) receptor, which are activated upon binding of lipopolysaccharide (TLR4), IL-1 (IL-1R) or TNFα (TNF receptor) [6]. The receptor binding and subsequent pathway signaling lead to activation of the IκB kinase (IKK) complex comprising of IKKγ/NEMO, IKKα and IKKβ, which is followed by the phosphorylation and subsequent ubiquitination and proteasomal degradation of IκBs. Ultimately, liberated NFκB translocates to the nucleus and activates several pro-inflammatory and anti-apoptotic genes through direct interference with the DNA transcription 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

(5)

For Peer Review

machinery [3] [4]. Genetic explanation for aberrant NFκB activation in CRC is sparse, but in two separate studies individuals carrying the deletion allele of the functional -94 ATTG insertion/deletion (ins/del) polymorphism in the promoter region of NFκB were found to carry an increased risk of CRC in Swedish [7] and Danish [8] populations. Polymorphisms in the genes encoding TLRs and ILR ligands were previously associated with the risk of various cancer forms, CRC included [9-14].

NFκB signaling is under control of TNFAIP3 (also known as A20) encoded by the TNFAIP3 gene on chromosome 6q23.3, a 790 amino acid protein with ubiquitin editing properties. It removes ubiquitin moieties via its ovarian tumor (OTU) domain while the C-terminal zinc fingers (ZnF; mainly ZnF4) provide E3 ligase activity [15]. TNFAIP3 disrupts NFκB activation via several mechanisms of which the most well-described consists of a series of ubiquitination and deubiquitination steps of several NFκB-associated proteins which ultimately terminates NFκB activity [6] [15,16]. On the other hand, the expression of TNFAIP3 is regulated by NFκB [17], which in a negative feedback loop manner serves to terminate NFκB activation. TNFAIP3 deficient mice develop severe inflammation in multiple organs, including the intestines where the loss of epithelial integrity is thought to evoke the strong inflammatory response that is observed [18,19]. Similarly, TNFAIP3 expression was decreased in the intestines of CD patients, a disease where the epithelial integrity is thought to be severely damaged [20]. The proposed roles of TNFAIP3 in cancer are diverse and most likely dependent on the cellular context. While the knowledge on this issue is still limited, a few studies report some interesting clinical findings. While undifferentiated head and neck squamous cell carcinomas, nasopharyngeal carcinomas [21], ER negative breast cancer [22] and glioblastoma stem cells [23], show overexpression of TNFAIP3, decreased TNFAIP3 levels have very recently been found in pancreatic cancers [24] and a tumor suppressive role has been proposed in B-cell lymphomas (BCLs) [25,26]. Moreover, several single nucleotide polymorphisms (SNPs), including rs6920220, in TNFAIP3 were connected with increased risk of rheumatoid arthritis [27,28], IBD [29], psoriasis [30] and type 1 diabetes [31], enlightening the importance of the gene in common inflammatory/autoinflammatory disorders.

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

(6)

For Peer Review

Other molecular networks take part in NFκB driven inflammation, such as the recently described NLRP3 inflammasome, a multiprotein complex consisting of NLRP3 (also called CIAS1, NALP3), ASC, CARD8 and procaspase-1. While the exact role of CARD8 in the inflammasome is a debatable issue, its role as an inhibitor of NFκB and caspase-1 is more well-understood [32,33]. Assembly of the NLRP3 inflammasome is triggered by danger associated molecular patterns and results in activation of procaspase-1 to caspase-1, which converts proIL-1β to mature and active IL-1β [34]. Germline mutations in the NLRP3 gene are associated with various inflammatory diseases and familial periodic fever syndromes, e.g. cryopyrin associated periodic syndromes (CAPS). SNPs in the NLRP3 gene and a nonsense polymorphism (C10X) in CARD8 have been associated with susceptibility to CD in men [35], while their role in intestinal malignancy, to our knowledge, has not yet been investigated.

In the present study, enrolling 348 Swedish CRC patients and approximately 800 healthy controls, we reveal that variations in genes being involved in the NFκB/TNFAIP3 and NLRP3 mediated inflammation are associated with the susceptibility and/or outcome of CRC. Further, we suggest that TNFAIP3 is a potential tumor suppressor in CRC, displaying LOH in 17% of the investigated cases. Mutational inactivation is another hypothetical ‘hit’ to the TNFAIP3 gene, but was found to be a rare event in the investigated material. This is, to our knowledge the first study that investigates and identifies a correlation between TNFAIP3 (rs6920220) and NLRP3 (Q705K) SNPs and the survival of CRC patients.

Material and methods

Patient material, DNA and RNA isolation

Clinical information, including age and gender, Dukes’ staging, polypoid/ulcerative phenotype and survival data, was obtained from the Swedish Cancer register for 355 patients with sporadic CRC undergoing surgical resection between 1998 and 2005 at the County Hospital Ryhov, Jönköping and University Hospital, Linköping, Sweden. Normal intestinal mucosal biopsies resected approximately 10 cm from the tumor (n = 235) and/or normal blood samples (n = 223) were collected from 348 of these 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

(7)

For Peer Review

patients. All patients underwent surgery and the tumor tissue was saved, examined and diagnosed by experienced pathologists at respective hospitals. 806 randomly selected healthy individuals within the same geographical area as the CRC patients were collected during the years 1998 to 2000 and used as controls in this study. Tissue and blood samples were immediately snap frozen and stored at -70°C until nucleic acid isolation. Genomic DNA was isolated from the tissue and normal blood samples with the Wizard® Genomic DNA Purification Kit (Promega, Madison, WI, USA) according to the supplier’s recommendations. Total RNA was isolated from intestinal tumor tissue and corresponding non-neoplastic mucosa using IllustraTM triplePrep kit (GE Healthcare, Buckinghamshire, UK) or RNeasy Mini Kit (Qiagen, Hilden, Germany) according to manufacturer’s recommendations. The study was approved by the Research Ethics Committee at the Faculty of Health Sciences, Linköping (Dnr. 98113).

SNP selection and real time PCR TaqMan® SNP genotyping

One aim of the study was to determine and evaluate the role of polymorphisms in genes linked to NFκB pathway mediated inflammation as susceptibility factors for sporadic CRC. The investigated genes, SNPs and their genotyping assays can be found in Table SI. Specifically, NFκB -94 ins/del ATTG promoter polymorphism was determined by using the forward primer:

5’-CCGTGCTGCCTGCGTT-3’, reverse primer:

5’-GCTGGAGCCGGTAGGGAA-3’ as well as probe 1: 5'-VIC-ACCATTGATTGGGCC-MGB-3' and probe 2:

5'-FAM-CGACCATTGGGCC-MGB-3’. Total volume per well was 15 µl comprised of 7.5 µl TaqMan® Universal PCR Master Mix (2x), 1.3 µl (10 µM) forward primer, 1.3 µl (10 µM) reverse primer, 0.3 µl (10 µM) probe 1, 0.3 µl (10 µM) probe 2, 3.3 µl MilliQ dH2O and 1 µl (20 ng/µl) of DNA. TNFAIP3

(rs6920220), NLRP3 (rs10733113), NLRP3 (Q705K), CARD8 (C10X), TLR4 (D299K) and TLR4 (T399I)

(Table SI) were genotyped in CRC patients and healthy controls by using TaqMan® allelic discrimination assay, according to the supplier’s recommendations using a 7500 Fast Real-time qPCR machine (Applied 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

(8)

For Peer Review

Biosystems, Foster City, CA, USA). Genotypes were analyzed using Sequence Detection System (SDS) 1.3.1 (Applied Biosystems).

cDNA synthesis and quantitative real time PCR

125-500 ng total RNA from each sample was reversely transcribed into cDNA with Maxima® First Strand cDNA synthesis Kit for RT-qPCR (Fermentas, St Leon-Rot, Germany) according to supplier’s instructions. The relative mRNA expression of TNFAIP3 was determined by quantitative real-time PCR (7900HT Fast Real-Time PCR System (Applied Biosystems)), which was performed with the TaqMan® Fast Universal PCR Master Mix (Applied Biosystems) and cDNA specific primer/probe mixes for

TNFAIP3 (Hs00234713_m1) (Applied Biosystems). The relative concentrations of TNFAIP3 were

normalized to the expression of the control genes β-glucuronidase (GUSB) (4333767F) and cyclophilin A (PPIA) (4333763T) (Applied Biosystems).

Loss of heterozygosity (LOH) analysis of 6q23.3

Six microsatellite markers, one trinucleotide ins/del polymorphism and one SNP were chosen for LOH mapping of an approximately 2.8 Mb region of chromosome 6q23.3 (including the TNFAIP3 gene). The marker information, primer sequences and PCR conditions are listed in Table SII. The order of these markers was: centromere-D6S292-D6S1009-rs6920220-D6S(1381118)-rs5029940-D6S(1382694)-D6S1587-D6S1569-telomere, where D6S(1381118) and D6S(1382694) are previously unreported dinucleotide repeats (chromosomal positions in hb [hectobases] within brackets) and rs5029940 is an

TNFAIP3 intragenic trinucleotide ins/del polymorphism. LOH analysis of the microsatellite markers and

the trinucleotide repeat were carried out using fluorescently tagged primers and two pools of multiplex-PCR (Table SII). 1 µl of diluted multiplex-PCR product was combined with 0.3 µl of Genescan LIZ 600 size standard (Applied Biosystems) and 12 µl formamide loading buffer (HiDi) (Applied Biosystems) and analyzed on a ABI 3500 Genetic Analyzer (Applied Biosystems). GeneMapper v.4.1 (Applied Biosystems) was used to determine size, height and area for each detected fluorescent peak. Alleles were defined as the two highest peaks within the expected size range and a tumor/normal ratio (T1:T2/N1:N2) 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

(9)

For Peer Review

was used to determine LOH. The ∆∆Ct method was used to determine LOH for the rs6920220 SNP.

Allelic loss of ≥35% (ratio of 0.65 or 1.54) for both peak height and area was scored as LOH for a single marker.

Mutational analyses of TNFAIP3

A detailed description of the mutational analyses of TNFAIP3 can be found in Supplementary Methods. Statistical analysis and figure composition

Standard Chi-square test was used to test genotype frequencies for Hardy-Weinberg-Equilibrium (HWE). Associations between genotype and disease risk were assessed by calculating odds ratios (OR) and corresponding 95% confidence intervals (CI). Linkage disequilibrium (LD) between polymorphisms of the TLR4 gene was examined by pair-wise comparisons of D′-values as well as the correlation coefficient (r2) using HaploView version 4.1 (Broad Institute of MIT and Harvard, Boston, MA). For survival analysis Kaplan-Meier curves were generated and tested for significance by the log-rank test. Overall survival was the time elapsed from diagnosis of CRC to the date of death, while CRC-specific survival was the time elapsed from diagnosis to the date of death from CRC. The median follow-up time after diagnosis was 56 months. Cox proportional hazard modeling was used to estimate hazard ratios and corresponding P-values, associated with the polymorphisms. Crude hazard rations (cHRs) for NFκB -94

ATTG ins/del, TNFAIP3 (rs6920220) or NLRP3 (Q705K) genotypes, gender, age and phenotypes were

determined independently. Adjusted hazard rations (aHR) analyses of NFκB -94 ATTG ins/del, TNFAIP3

(rs6920220) or NLRP3 (Q705K) genotypes on CRC-specific survival of patients in the Dukes’ C+D

group included gender, age and polypoid/ulcerative phenotype and were calculated to investigate simultaneous effects and generalizability of the factors investigated. Group comparisons for the mRNA expression analyses were performed using Mann-Whitney U-tests and Wilcoxon Rank-Sum tests. Error bars display the interquartile range (IQR).

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

(10)

For Peer Review

Statistical analysis was performed with SPSS Statistics 18 (IBM Corporation, Somers, NY, USA), Graphpad Prism 5 (GraphPad Software Inc., San Diego, CA, USA) and Stata 11 (StataCorp LP, Lakeway Drive, Texas, USA). P-values < 0.05 were considered as statistically significant.

Results

NFκB -94 ins/del ATTG, TNFAIP3 (rs6920220) and NLRP3 (Q705K) SNPs are associated with Dukes’ stage related survival

Survival data was available for 348 patients, but normal DNA was only available for 337 of these, which were included in the following analyses. For each SNP, patients were stratified according to their specific genotypes except for TLR4 (D299K) and TLR4 (T399I), which were combined according to their haplotypes: G-T (polymorphic alleles) in one group and A-C, A-T and G-C in the other. No significant differences in CRC-specific survival could be identified (Fig. S1), neither could we detect any differences in overall survival between the different genotypes for the investigated SNPs (Fig. S2).

Exploratory subgroup analyses

Subgroup analysis revealed differences in CRC-specific survival between the NFκB -94 ins/del ATTG,

TNFAIP3 (rs6920220) and NLRP3 (Q705K) genotypes when stratification for Dukes’ stages A+B and

C+D was considered. A Log-rank test showed no significant differences within the A+B group for any of the polymorphisms (Fig. 1 A, B, C). Among the C+D patients the del/del genotype of NFκB -94 ins/del

ATTG led to significantly shorter CRC-specific survival (P = 0.007) (Fig. 1D). Patients homozygous for

the AA genotype of TNFAIP3 (rs6920220) had a significantly shorter CRC-specific survival compared to carriers of the GG genotype (P = 0.001) (Fig. 1E). Individuals with a NLRP3 (Q705K) CA genotype in the C+D group had likewise a significantly shorter CRC-specific survival (P = 0.004) (Fig. 1F). Furthermore, the ulcerative phenotype was found to be significantly associated with poorer CRC-specific survival (cHR = 2.1, 95% CI 1.0-4.3, P = 0.04) (Fig. S3A), but not overall survival (P = 0.2) (Fig. S3B), which is in concordance with a previous study by Nasir et al. [36].

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

(11)

For Peer Review

A multivariate Cox regression analysis was carried out to evaluate the potential prognostic significance of

NFκB -94 ATTG ins/del, TNFAIP3 (rs6920220) and NLRP3 (Q705K) polymorphisms on CRC-specific

survival of patients in the Dukes’ C+D group. The multivariate models included SNP genotypes, gender, age and polypoid/ulcerative phenotype. The NFκB -94 ATTG ins/del, TNFAIP3 (rs6920220) and NLRP3

(Q705K) polymorphisms were statistically significant in these models, with the heterozygous or/and

polymorphic genotypes being significantly associated with poorer CRC-specific survival if diagnosed with Dukes’ stage C+D, but not Dukes’stage A+B (Table II) (Fig. S4A, B). Among 110 Dukes’ C+D patients 35 out of 53 individuals (66%) carries either NFκB -94 ATTG del/del or NLRP3 (Q705K) CA or

TNFAIP3 (rs6920220) GA or AA genotypes, suffered CRC death (Fig. S4A) (opposed to 19 out of 57

(33%) individuals with the remaining genotypes). To more clearly show that carriers of any of the polymorphic variants of these polymorphisms within the Dukes’ C+D group have a poorer survival, we defined a so-called prognostic panel of the polymorphisms with significant genotypes in the multivariate Cox regression models (Table II). The prognostic panel consists of NFκB -94 ATTG ins/del, TNFAIP3

(rs6920220) and NLRP3 (Q705K) and results in cHR of 2.9 (95% CI: 1.7–5.1, P < 0.001, n=110) (Table

II) with a worse prognosis for carriers of the polymorphic and/or heterozygous genotypes. Adjusting for age, gender and polypoid/ulcerative phenotype the panel stays significant (Prognostic Panel: aHR = 5.2, 95% CI 2.5-10.9, P < 0.001, n = 75) (Table II). Altogether, the homozygous polymorphic NFκB -94

ATTG ins/del variant, the heterozygous and homozygous polymorphic TNFAIP3 (rs6920220) variants as

well as heterozygous variant of NLRP3 (Q705K) are associated with poor survival in CRC patients diagnosed with advanced CRC.

NFκB pathway SNPs and colorectal cancer susceptibility

Distribution of the SNP genotypes in the study populations and the odds ratios for developing CRC are summarized in Table III. All the polymorphisms were found to be in HWE in the control population.

TLR4 D299K and TLR4 T399I were found to be in tight LD (D’ = 0.9, r2 = 0.7), with the polymorphic (minor) alleles in phase. Heterozygosity for NFκB -94 ins/del ATTG was significantly associated with 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

(12)

For Peer Review

CRC susceptibility when compared to homozygous wild-type allele (ins/ins) carriers, (OR = 1.6, 95% CI 1.2-2.1, P = 0.003). No other significant associations were identified regarding the investigated polymorphisms and CRC susceptibility, although the NLRP3 (rs10733113) AA genotype showed borderline significance (OR = 2.2, 95 % CI 1.0-5.2, P = 0.07) (Table III). To investigate a potential SNP combinatorial risk increase for the development of CRC, multiple logistic regression analysis was carried out on combinations of polymorphisms in the patient and control material. A moderate, however non-significant risk increase was seen when combining NLRP3 (Q705K) and CARD8 (C10X) for the CA and TA genotypes (OR = 1.7, 95% CI 1.0-2.9, P = 0.08). A similar trend was observed when combining patients with one or more variant alleles for each SNP (OR = 1.6, 95 % CI 1.0-2.6, P = 0.08) (Table SIII). TNFAIP3 gene expression is downregulated in colorectal tumors

To determine whether TNFAIP3 expression levels were significantly altered in colorectal malignancies total RNA was isolated from 57 CRC/non-neoplastic colorectal pairs and additional 12 tumor tissue samples and analyzed by quantitative RT-PCR. Significantly lower TNFAIP3 mRNA levels were found in the tumors as compared to adjacent non-neoplastic intestinal mucosa (Mann-Whitney U-test, P < 0.0001, paired Wilcoxon Rank-Sum test, P < 0.0001) (Fig. 2A; Table SIII). Moreover, TNFAIP3 expression levels in normal intestinal mucosa correlated with TNFAIP3 (rs6920220) genotype status. Homozygous wild-type individuals (GG) displayed an approximately 1.7-fold higher mRNA expression compared to GA+AA individuals (Mann-Whitney U-test, P = 0.016) indicating a gene dose dependent expression.

The TNFAIP3 gene is frequently deleted in colorectal adenocarcinomas, leading to lower gene expression

Further we elucidated potential mechanisms underlying the decreased TNFAIP3 mRNA levels seen in the tumors. TNFAIP3 is frequently lost in follicular lymphoma due to LOH of 6q23.3 [37]. Using eight markers spanning the region 6q23.3, 25 out of 149 CRC tumors (16.8%) showed LOH and of these 25 cases 20 showed continuous loss of the entire analyzed region (2.8 Mb), while five cases showed loss of a 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

(13)

For Peer Review

smaller part (0.42 Mb), however still encompassing TNFAIP3. Fig. 2B displays how the markers a located on the chromosome 6q in relation to TNFAIP3 and also a representative sample showing LOH for five out of eight markers (rs6920220 is left out of Fig. 2B). LOH was associated with decreased TNFAIP3 mRNA expression (Mann-Whitney U-test, P = 0.0033) at least partly explaining the lower expression levels seen in the CRC tumors (Fig. 2C; Table SIII).

The TNFAIP3 gene is rarely mutated in colorectal adenocarcinomas

Using SSCA mutational screening and subsequent DNA sequencing, mutations in TNFAIP3 were detected in two out of 79 colorectal tumors (2.5%). In one case a missense mutation c.305A>G (p.N102S) with germline origin was found (Fig. S5A). The substitution occured one amino acid prior to the catalytic residue of the OTU deubiquitinase domain. The asparagine residue and the adjacent protein sequence are evolutionary conserved, implying a functional importance (Fig. S5B). A second mutation in exon 9 in an unrelated sample uncovered the deletion of one nucleotide: c.2274delC. Sequencing of corresponding DNA from normal intestinal mucosa proved the mutation to be somatic (Fig. S5C). The mutation results in a prolonged protein p.K759Sfs*57X and a complete ablation of the evolutionary conserved ZnF7 domain (Fig. S5D). Interestingly, this mutation was found in a microsatellite instable tumor (data not shown), however none of the two mutated samples showed any signs of LOH.. Multiple sequence alignment (MSA) of TNFAIP3 in silico with ClustalX v2.1 revealed that the OTU and ZnF7 domains are conserved among species. The identified c.305A>G germline mutation is not entered into the NCBI dbSNP database, but has previously been found in patients with B-cell lymphoma [25,38]. We therefore investigated whether TNFAIP3 c.305A>G is a common genetic variant among CRC patients or healthy control individuals. Using PCR-RFLP (Table SIV)) and subsequent gel electrophoresis, 342 CRC cases and 365 control individuals were tested for c.305A>G in their normal DNA (either blood or normal intestinal tissue). The assay was verified using the previously identified mutants as positive controls. No additional mutations were found implying that TNFAIP3 c.305A>G is a rare genetic variant.

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

(14)

For Peer Review

Discussion

It is commonly known that even though acute inflammation is essential for the host’s microbe defense and tissue repair, a chronic inflammatory condition degenerates the tissue, increases the risk of sustained cell and DNA damage and promotes tumor formation and progression. This is clearly evident in patients with IBD, which suffer an increased risk of CRC and especially CAC [3]. A widely accepted link between inflammation and CRC is underlined by studies of the pro-inflammatory NFκB pathway, which promotes cell proliferation, inhibits apoptosis [4] and is found activated in primary and metastatic CRC [39]. An ATTG (ins/del) polymorphism in the NFκB promoter has been associated with increased risk of CRC [7,8]. Consistent with these findings, we found an increased risk of CRC in individuals heterozygous for the ins/del genotype. The deletion allele has been associated with lower gene expression

in vitro [40], which intuitively is associated with a poorer inflammatory response. However, Greten et al.

[41] unexpectedly discovered that NFκB also could function as a negative post-translational regulator of IL-1β secretion through caspase-1 independent but serine protease dependent pathways, possibly in an inflammation stage-dependent manner emphasizing an even more complex role of NFκB in regulating inflammation. In the survival analysis, no association of the NFκB -94 ATTG ins/del polymorphism with CRC-specific or overall survival was found, which is in accordance with a previous study by Riemann et al. [42]. However, stratification for Dukes’ stages A+B and C+D revealed a poorer survival for patients carrying the del/del genotype among the C+D patients (Fig. 1 A, D and Table II). A CRC-specific survival analysis comparing the A+B and C+D groups verifies a poorer survival in the C+D group (P < 0.001) (Fig. S3C) and justifies the crude subgrouping.

The rs6920220 SNP is located in a putative TNFAIP3 regulatory region ~185kb upstream of the gene. Corroborating previous epidemiological and functional studies [27,28,43], we show that the variant allele of the rs6920220 SNP on TNFAIP3 associates with a lower TNFAIP3 expression in non-neoplastic intestinal tissue as compared to the wild-type G-allele. Due to TNFAIP3 inhibitory effects upon NFκB signaling, a hypomorphic TNFAIP3 gene will likely contribute to the NFκB activation seen in primary 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

(15)

For Peer Review

and metastatic CRC [39]. After adjusting for gender, age and polypoid/ulcerative phenotypes, the association with poorer survival remained highly significant, suggesting that TNFAIP3 (rs6920220) may be a novel a prognostic marker in patients with advanced stages of CRC. To our knowledge the findings in this study are the first to suggest a tumor suppressive role for TNFAIP3 in CRC. Previously, Arsenescu

et al. [20] reported decreased TNFAIP3 expression in intestinal tissue from CD patients and here we

suggest that low levels of functional TNFAIP3 in a similar manner may be connected with colorectal malignancy. Hypothetically, loss of TNFAIP3 in intestinal epithelial cells may be an event contributing to a chronic/dysfunctional inflammatory response with chronic uncontrolled NFκB activation, which may foster a tumor. In the present material, TNFAIP3 mRNA levels were found to be significantly lower in tumors compared to the adjacent non-neoplastic mucosa (Fig. 2A) supporting the before mentioned hypothesis

TNFAIP3 is located on chromosome 6q23.3, a region frequently deleted in various cancer forms

[37,44,45] and the suggested tumor suppressor function of TNFAIP3 in B-cell lymphomas and lower

TNFAIP3 expression seen in the colorectal tumors prompted us to do LOH analysis of the 6q23.3 region.

In total, LOH was detected in 25 out of 149 CRC tumors (16.8%), which also was associated with a

significantly lower gene expression (Fig. 2C). Interestingly, loss of 6q23.3 is a relatively frequent event in

sporadic CRC and this chromosomal region contains several additional tumor suppressor candidates and genes involved in the inflammatory response e.g. BCLAF1, MAP3K5, IL-22Rα2 and IFNγR1 that may be

responsible or contributing factors in addition to LOH of TNFAIP3. A possible tumor suppressive role for

TNFAIP3 in CRC is further underlined by the single somatic and germline mutations identified in the

present study. Both identified mutations are located in evolutionary well-conserved domains (Fig. S5B, D), which may affect the function of the protein. The c.2774delC mutation alters the whole ZnF7 domain, which is important for TNFAIP3’s role as an inhibitor of NFκB. The TNFAIP3 N102S mutation was previously reported in three BCLs [25,38] and at least once unambiguously identified as a germline alteration. Together with our data this proposes that TNFAIP3 N102S may be a germline genetic variant 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

(16)

For Peer Review

that predisposes for cancers that are characterized by increased inflammatory NFκB activity. Collectively, our results suggest that an increased NFκB activity, at least partly, could be explained by allelic loss of one copy of TNFAIP3 and/or the low expressing TNFAIP3 (rs6920220) A-allele. In combination with the lower TNFAIP3 expression seen in the tumors, this proposes a tumor suppressive function of TNFAIP3 in CRC.

Similar with TNFAIP3 (rs6920220), we find an association between heterozygous NLRP3 (Q705K) and poor patient survival in advanced tumors (Dukes’ C+D) and to our knowledge this is the first study that reports an association between SNPs in the NLRP3 gene and CRC patient survival. These results, together with the lack of association of these SNPs with CRC susceptibility or patient survival in the Dukes’ A and B group, support a model where these inflammatory signals rather contribute to late stage sporadic colorectal tumorigenesis than early as often is the case in CAC [3]. The Q705K SNP is quite prevalent in the healthy population and has been associated with increased risk of RA and CD [35,46,47] and most recently with malignant melanoma [48]. In a recently published study by our group using the monocytic cell line THP-1, we have demonstrated that the Q705K in NLRP3 leads to an increased production of IL-1β and IL-18 [49]. This gain of function polymorphism, in combination with other triggering factors might contribute to an increased predisposition towards inflammatory conditions. Studies of CAC affected mice with inducible NLRP3 deficiency reveal a protective role for the NLRP3 inflammasome against colitis and CAC [50,51]. This protective role has been attributed to NFκB and NLRP3 inflammasome regulated cytokine IL-18 production, which plays a pivotal role in the repair of ulcerated epithelium during the acute inflammatory phase and has also been proposed to have anti-tumorigenic effects [51]. These results suggest that in the acute inflammatory phase during the early formation of intestinal adenomas, the NLRP3 inflammasome may have a protective role, but if overactivated it may contribute to chronic inflammation and in more advanced sporadic CRC cases, lead to poorer survival. Both overactivation of the NLRP3 inflammasome and downregulation or genetic alterations in TNFAIP3, may contribute to increased IL-1β and subsequent IL-6 levels and a poorer outcome for the CRC patient. 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

(17)

For Peer Review

As an adjuvant to standard cancer treatment, IL-1β or IL-6 blocking therapy has been proposed to alleviate inflammation and tumor progression in advanced CRC. We acknowledge that the significance of the observed associations for susceptibility and survival may not survive stringent correction for multiple testing. However, due to the explorative nature of this study, we purposely did not correct for multiple comparisons so not to eliminate potentially important results. Therefore the data should be interpreted with some caution and we would welcome independent replications of this study as well as further functional analyses of the polymorphisms herein enrolled.

In conclusion, we here suggest that the anti-inflammatory gene TNFAIP3 may act as a novel tumor suppressor in CRC and further, that a panel of SNPs in genes being involved in inflammation pathways,

i.e. rs6920220 in TNFAIP3, Q705K in NLRP3 (rs35829419) and NFκB -94 ATTG ins/del (rs28362491)

polymorphism could be used as prognostic markers to identify patients at high risk of rapid CRC progression and poor survival. Our results highlight the crucial role of the investigated inflammation pathways in advanced/terminal stages of sporadic CRC rather than in early stages. Speculatively, analyses of inflammation-associated genes may aid in identifying patients for future interleukin- and/or NFκB-modulating therapies.

Acknowledgment

The authors would like to sincerely thank Piiha-Lotta Jerevall for help with the survival analyses as well as Jenny Welander, Annette Molbaek and Åsa Schippert for technical and laboratory assistance.

Grant support

This work was funded by the Swedish Research Council. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

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

(18)

For Peer Review

References

1. Hanahan, D., et al. (2011) Hallmarks of cancer: the next generation. Cell, 144, 646-674. 2. Tenesa, A., et al. (2009) New insights into the aetiology of colorectal cancer from

genome-wide association studies. Nat Rev Genet, 10, 353-358.

3. Terzic, J., et al. (2010) Inflammation and colon cancer. Gastroenterology, 138, 2101-2114 e2105.

4. Vaiopoulos, A.G., et al. (2010) Colon carcinogenesis: Learning from NF-kappaB and AP-1. Int

J Biochem Cell Biol, 42, 1061-1065.

5. Greten, F.R., et al. (2004) IKKbeta links inflammation and tumorigenesis in a mouse model of colitis-associated cancer. Cell, 118, 285-296.

6. Hymowitz, S.G., et al. (2010) A20: from ubiquitin editing to tumour suppression. Nat Rev

Cancer, 10, 332-341.

7. Lewander, A., et al. (2007) Polymorphism in the promoter region of the NFKB1 gene increases the risk of sporadic colorectal cancer in Swedish but not in Chinese populations.

Scand J Gastroenterol, 42, 1332-1338.

8. Andersen, V., et al. (2010) Polymorphisms in NFkB, PXR, LXR and risk of colorectal cancer in a prospective study of Danes. BMC Cancer, 10, 484.

9. Kutikhin, A.G. (2011) Impact of Toll-like receptor 4 polymorphisms on risk of cancer. Hum

Immunol, 72, 193-206.

10. Ito, H., et al. (2007) Interleukin-1beta gene in esophageal, gastric and colorectal carcinomas.

Oncol Rep, 18, 473-481.

11. Sehouli, J., et al. (2002) Polymorphism of IL-1 receptor antagonist gene: role in cancer.

Anticancer Res, 22, 3421-3424.

12. Boraska Jelavic, T., et al. (2006) Microsatelite GT polymorphism in the toll-like receptor 2 is associated with colorectal cancer. Clin Genet, 70, 156-160.

13. Rakoff-Nahoum, S., et al. (2009) Toll-like receptors and cancer. Nat Rev Cancer, 9, 57-63. 14. Eyking, A., et al. (2011) Toll-like receptor 4 variant D299G induces features of neoplastic

progression in Caco-2 intestinal cells and is associated with advanced human colon cancer.

Gastroenterology, 141, 2154-2165.

15. Verstrepen, L., et al. (2010) Expression, biological activities and mechanisms of action of A20 (TNFAIP3). Biochem Pharmacol, 80, 2009-2020.

16. Vereecke, L., et al. (2009) The ubiquitin-editing enzyme A20 (TNFAIP3) is a central regulator of immunopathology. Trends Immunol, 30, 383-391.

17. Krikos, A., et al. (1992) Transcriptional activation of the tumor necrosis factor alpha-inducible zinc finger protein, A20, is mediated by kappa B elements. J Biol Chem, 267, 17971-17976.

18. Lee, E.G., et al. (2000) Failure to regulate TNF-induced NF-kappaB and cell death responses in A20-deficient mice. Science, 289, 2350-2354.

19. Kolodziej, L.E., et al. (2011) TNFAIP3 Maintains Intestinal Barrier Function and Supports Epithelial Cell Tight Junctions. PLoS One, 6, e26352.

20. Arsenescu, R., et al. (2008) Signature biomarkers in Crohn's disease: toward a molecular classification. Mucosal Immunol, 1, 399-411.

21. Codd, J.D., et al. (1999) A20 RNA expression is associated with undifferentiated nasopharyngeal carcinoma and poorly differentiated head and neck squamous cell carcinoma. J Pathol, 187, 549-555.

22. Vendrell, J.A., et al. (2007) A20/TNFAIP3, a new estrogen-regulated gene that confers tamoxifen resistance in breast cancer cells. Oncogene, 26, 4656-4667.

23. Hjelmeland, A.B., et al. (2010) Targeting A20 decreases glioma stem cell survival and tumor growth. PLoS Biol, 8, e1000319.

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

(19)

For Peer Review

24. Wang, Q., et al. (2012) Expression of A20 is reduced in pancreatic cancer tissues. J Mol Histol. 25. Honma, K., et al. (2009) TNFAIP3/A20 functions as a novel tumor suppressor gene in several

subtypes of non-Hodgkin lymphomas. Blood, 114, 2467-2475.

26. Malynn, B.A., et al. (2009) A20 takes on tumors: tumor suppression by an ubiquitin-editing enzyme. J Exp Med, 206, 977-980.

27. Orozco, G., et al. (2009) Combined effects of three independent SNPs greatly increase the risk estimate for RA at 6q23. Hum Mol Genet, 18, 2693-2699.

28. Thomson, W., et al. (2007) Rheumatoid arthritis association at 6q23. Nat Genet, 39, 1431-1433.

29. (2007) Genome-wide association study of 14,000 cases of seven common diseases and 3,000 shared controls. Nature, 447, 661-678.

30. Nair, R.P., et al. (2009) Genome-wide scan reveals association of psoriasis with IL-23 and NF-kappaB pathways. Nat Genet, 41, 199-204.

31. Fung, E.Y., et al. (2009) Analysis of 17 autoimmune disease-associated variants in type 1 diabetes identifies 6q23/TNFAIP3 as a susceptibility locus. Genes Immun, 10, 188-191. 32. Razmara, M., et al. (2002) CARD-8 protein, a new CARD family member that regulates

caspase-1 activation and apoptosis. J Biol Chem, 277, 13952-13958.

33. Fontalba, A., et al. (2007) Deficiency of the NF-kappaB inhibitor caspase activating and recruitment domain 8 in patients with rheumatoid arthritis is associated with disease severity. J Immunol, 179, 4867-4873.

34. Agostini, L., et al. (2004) NALP3 forms an IL-1beta-processing inflammasome with increased activity in Muckle-Wells autoinflammatory disorder. Immunity, 20, 319-325.

35. Schoultz, I., et al. (2009) Combined polymorphisms in genes encoding the inflammasome components NALP3 and CARD8 confer susceptibility to Crohn's disease in Swedish men. Am

J Gastroenterol, 104, 1180-1188.

36. Nasir, A., et al. (2004) Flat and polypoid adenocarcinomas of the colorectum: A comparative histomorphologic analysis of 47 cases. Hum Pathol, 35, 604-611.

37. Ross, C.W., et al. (2007) Comprehensive analysis of copy number and allele status identifies multiple chromosome defects underlying follicular lymphoma pathogenesis. Clin Cancer Res, 13_{, 4777-4785.}

38. Dong, G., et al. (2011) A20, ABIN-1/2, and CARD11 mutations and their prognostic value in gastrointestinal diffuse large B-cell lymphoma. Clin Cancer Res, 17, 1440-1451.

39. Puvvada, S.D., et al. (2010) NF-kB and Bcl-3 activation are prognostic in metastatic colorectal cancer. Oncology, 78, 181-188.

40. Karban, A.S., et al. (2004) Functional annotation of a novel NFKB1 promoter polymorphism that increases risk for ulcerative colitis. Hum Mol Genet, 13, 35-45.

41. Greten, F.R., et al. (2007) NF-kappaB is a negative regulator of IL-1beta secretion as revealed by genetic and pharmacological inhibition of IKKbeta. Cell, 130, 918-931.

42. Riemann, K., et al. (2006) No association of the NFKB1 insertion/deletion promoter polymorphism with survival in colorectal and renal cell carcinoma as well as disease progression in B-cell chronic lymphocytic leukemia. Pharmacogenet Genomics, 16, 783-788. 43. Elsby, L.M., et al. (2010) Functional evaluation of TNFAIP3 (A20) in rheumatoid arthritis.

Clin Exp Rheumatol, 28, 708-714.

44. Uribe, P., et al. (2009) Allelotyping, microsatellite instability, and BRAF mutation analyses in common and atypical melanocytic nevi and primary cutaneous melanomas. Am J

Dermatopathol, 31, 354-363.

45. Bragadottir, G., et al. (1995) Loss of heterozygosity at chromosome 6q correlates with tumor progression and patient survival. Int J Oncol, 7, 871-876.

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

(20)

For Peer Review

46. Villani, A.C., et al. (2009) Common variants in the NLRP3 region contribute to Crohn's disease susceptibility. Nat Genet, 41, 71-76.

47. Kastbom, A., et al. (2008) Genetic variation in proteins of the cryopyrin inflammasome influences susceptibility and severity of rheumatoid arthritis (the Swedish TIRA project).

Rheumatology (Oxford), 47, 415-417.

48. Verma, D., et al. (2012) Inflammasome polymorphisms confer susceptibility to sporadic malignant melanoma. Pigment Cell Melanoma Res.

49. Verma, D., et al. (2012) The Q705K Polymorphism in NLRP3 Is a Gain-of-Function Alteration Leading to Excessive Interleukin-1beta and IL-18 Production. PLoS One, 7, e34977.

50. Allen, I.C., et al. (2010) The NLRP3 inflammasome functions as a negative regulator of tumorigenesis during colitis-associated cancer. J Exp Med, 207, 1045-1056.

51. Zaki, M.H., et al. (2010) IL-18 production downstream of the Nlrp3 inflammasome confers protection against colorectal tumor formation. J Immunol, 185, 4912-4920.

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

(21)

For Peer Review

Tables

Table I: Distribution of CRC patient characteristics All patients in the study 355

Patients with survival data 348

Variable Subcategory Missing cases

n (%) n (%) n (%) n (%) Age Median 72.5 Mean 71 Range 25-94 Gender Female 167 (46.9) 4 (1.1) Male 185 (52.0)

Survival Alive 170 (48.9) Death cause No data 32 (9.2) Deceased 178 (51.1) CRC specific 94 (27.0) Other death causes 52 (14.9) Staging Dukes’ A 45 (12.6) 172 (48.3) 65 (18.3) B 127 (35.7) C 107 (30.1) 119 (33.4) D 11 (3.1) Adenoma 7 (2.0) Phenotype Polypoid 83 (23.3) 138 (38.8) Ulcerative 121 (34.0) Mixed 14 (3.9) Location Colon 190 (53.4) 27 (7.6) Rectum 139 (39.0) 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

(22)

For Peer Review

Table II: Cox regression modeling of Dukes’ C+D stratified CRC-specific survival. Analyzed with SPSS Statistics 18 Enter method.

Factor Patients (n) (Alive/Deceased) cHR (95% CI) P-value Patients (n) (Alive/Deceased) aHR (95% CI) * P-value

NFκB -94 ATTG ins/del model

NFκB -94 ATTG ins/del ins/ins 35 (22/13) 23 (15/8) ins/del 62 (31/31) 1.5 (0.8-2.9) 0.2 43 (20/23 1.9 (0.8-4.3) 0.1 del/del 13 (3/10) 2.9 (1.3-6.7) 0.01 9 (2/7) 3.3 (1.2-9.3) 0.03 Gender F 52 (27/25) 37 (20/17) M 58 (29/29) 1.1 (0.6-1.8) 0.8 38 (17/21) 1.4 (0.7-2.8) 0.3 Age 110 (56/54) 1.0 (1.0-1.03) 0.6 75 (37/38) 1.0 (1.0-1.02) 0.4 Phenotype P 29 (18/11) 29 (18/11) U 46 (19/27) 2.1 (1.0-4.3) 0.04 46 (19/27) 2.1 (1.0-4.3) 0.04 TNFAIP3 (rs6920220) model TNFAIP3 (rs6920220) GG 71 (40/31) 44 (27/17) GA 34 (15/19) 1.7 (0.9-3.0) 0.06 28 (10/18) 3.6 (1.8-7.3) <0.001 AA 5 (1/4) 4.8 (1.7-13.7) 0.004 3 (0/3) 21.4 (4.8-94.6) < 0.001 Gender F 52 (27/25) 37 (20/17) M 58 (29/29) 1.1 (0.6-1.8) 0.8 38 (17/21) 1.8 (0.9-3.5) 0.1 Age 110 (56/54) 1.0 (0.98-1.0) 0.6 75 (37/38) 0.98 (0.95-1.0) 0.2 Phenotype P 29 (18/11) 29 (18/11) U 46 (19/27) 2.1 (1.0-4.3) 0.04 46 (19/27) 2.5 (1.2-5.3) 0.02 NLRP3 (Q705K) model NLRP3 (Q705K) CC 100 (55/45) 70 (36/34) CA 10 (1/9) 2.7 (1.3-5.6) 0.006 5 (1/4) 3.2 (1.0-10.1) 0.05 Gender F 52 (27/25) 37 (20/17) M 58 (29/29) 1.1 (0.6-1.8) 0.8 38 (17/21) 1.5 (0.8-2.9) 0.2 Age 110 (56/54) 1.0 (0.98-1.0) 0.6 75 (37/38) 1.0 (0.97-1.02) 0.5 Phenotype P 29 (18/11) 29 (18/11) U 46 (19/27) 2.1 (1.0-4.3) 0.04 46 (19/27) 2.6 (1.2-5.7) 0.02

Modeling of combined genotypes§

Prognostic panel§ 110 (56/54) 2.9 (1.7-5.1) <0.001 75 (37/38) 5.2 (2.5-10.9) <0.001 §

The Prognostic panel consists of the polymorphic variant NFκB -94 ATTG ins/del, heterozygous and polymorphic variant TNFAIP3 (rs6920220) and heterozygous NLRP3 (Q705K).

*_{P-values were calculated with adjustment for age, gender and polypoid/ulcerative phenotype in a}

multivariate Cox regression analysis. 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

(23)

For Peer Review

Table III: Genotype distribution of the different polymorphisms in a CRC patient and normal control population investigated in the study and their association with CRC susceptibility.

Polymorphisms CRC n (%) Controls n (%) OR (95% CI) P-value NFκB -94 ATTG ins/del ins/ins 114 (33.1) 256 (41.2) ins/del 187 (54.4) 270 (43.4) 1.6 (1.2-2.1) 0.003 del/del 43 (12.5) 96 (15.4) 1.0 (0.7-1.5) 1.0 ins/del+del/del 230 366 1.4 (1.1-1.9) 0.02 TNFAIP3 (rs6920220) GG 217 (63.1) 516 (64.0) GA 111 (32.3) 255 (31.6) 1.0 (0.8-1.4) 0.8 AA 16 (4.7) 35 (4.3) 1.1 (0.6-2.0) 0.8 GA+AA 126 288 1.0 (0.8-1.4) 0.8 NLRP3 (rs10733113) GG 260 (75.8) 575 (72.5) GA 72 (21.0) 207 (26.1) 0.8 (0.6-1.0) 0.1 AA 11 (3.2) 11 (1.4) 2.2 (1.0-5.2) 0.07 GA+AA 83 218 0.84 (0.6-1.1) 0.3 NLRP3 (Q705K) CC 296 (86.3) 693 (87.4) CA 47 (13.7) 94 (11.9) 1.2 (0.8-1.7) 0.4 AA 0 (0) 6 (0.8) 0.2 (0.01-3.2) 0.2 CA+AA 47 100 1.2 (0.8-1.7) 0.4 CARD8 (C10X) TT 131 (38.1) 332 (41.4) TA 166 (48.3) 365 (45.6) 1.2 (0.9-1.5) 0.3 AA 47 (13.7) 104 (13.0) 1.1 (0.8-1.7) 0.5 TA+AA 213 469 1.2 (0.9-1.59 0.3 TLR4 (D299K) AA 314 (91.3) 742 (92.1) AG 28 (8.1) 56 (6.9) 1.2 (0.7-1.9) 0.5 GG 1 (0.3) 1 (0.1) 2.3 (0.2-38) 0.5 AG+AA 29 57 1.2 (0.7-1.9) 0.5 TLR4 (T399I) CC 316 (91.9) 729 (91.1) CT 28 (8.1) 71 (8.9) 0.9 (0.6-1.4) 0.7 TT 0 (0) 0 (0) N/A N/A 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

(24)

For Peer Review

TLR4 (D299K+T399I) A-C, A-T, G-C G-T 319 (92.7) 25 (7.3) 744 (92.3) 55 (6.8) 1.1 (0.6-1.7) 0.8 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

(25)

For Peer Review

Figure legends:

Figure 1

Kaplan-Meier curves of NFκB -94 ATTG ins/del, TNFAIP3 (rs6920220), NLRP3 (Q705K) genotypes and CRC-specific survival stratified for Dukes’ stages A+B and C+D. Dukes’ stage A+B: A) NFκB -94 ATTG

ins/del B) TNFAIP3 (rs6920220) C) NLRP3 (Q705K). Dukes’ stage C+D: D) NFκB -94 ATTG ins/del E) TNFAIP3 (rs6920220) F) NLRP3 (Q705K). Log-rank P-values compared to the wild-type genotype are

displayed. Figure 2

TNFAIP3 expression is downregulated as a consequence of frequent LOH in colorectal tumors. (A)

Relative mRNA expression of TNFAIP3 (ntumor = 69, nnormal = 57). Tumor (black bar) median relative

expression = 0.67, IQR = 0.45-1.27 compared to adjacent non-neoplastic mucosa (white bar) median relative expression = 1.26, IQR = 0.90-2.0 (Mann-Whitney U-test, P < 0.0001, paired Wilcoxon Rank-Sum test, P < 0.0001). (B) A schematic representation of 6q23.3 displaying the positions of the eight markers in relation to TNFAIP3 and LOH analysis of 6q23.3 in one representative tumor, L32. The D6S1009, rs5029940, D6S(1382694) and D6S1569 markers show LOH while D6S292, D6S(1381118) and D6S1587 are uninformative in this case. (C) Relative TNFAIP3 mRNA expression in tumor samples with no 6q23.3 loss (n = 54, white bar), median relative expression = 0.90, IQR = 0.53-1.38, compared to tumor samples with 6q23.3 loss (n = 13, black bar), median relative expression = 0.45, IQR = 0.19-0.55 (Mann-Whitney U-test, P = 0.0033). Error bars = IQR.

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

(26)

For Peer Review

Figure 1. 168x134mm (300 x 300 DPI) 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

(27)

For Peer Review

Figure 2. 201x250mm (300 x 300 DPI) 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)

For Peer Review

Supplementary Information

Supplementary Methods

Mutational analysis of TNFAIP3

Primer sequences and PCR conditions for sequencing of TNFAIP3 are presented in Table SV.

32

P-nucleotide radiolabeling of PCR amplicons for single-strand conformation analysis (SSCA) was

performed in a secondary PCR reaction in which 1 µl of the primary PCR product was amplified for 15

cycles in the presence of 1.25 µCi of [α-

32

P]-ATP (3000 Ci/mmol) (Perkin-Elmer, Boston, MA, USA).

The samples were diluted 1:20 with denaturing loading solution (0.1% SDS, 14.9 mM EDTA, 49%

formamide, 0.05 bromphenol blue and 0.05% xylene cyanol FF) denaturated at 95°C for 3 min,

immediately cooled on ice and applied to a 6% polyacrylamide gel containing 10% glycerol or onto a

MDE

TM

gel (Mutation Detection Enhancement; BMA products, Rockland, ME, USA). The fragments

were separated at 6 W constant power at room temperature (RT) for 16-18 h, transferred to 3MM

chromatography paper and autoradiographed for 24-72 h at -70°C or 72 h at RT. Shifted fragments were

excised, reamplified and purified for sequencing analysis. Sequencing was performed on an ABI 3500

Genetic Analyzer (Applied Biosystems) using BigDye® Terminator v1.1 Cycle Sequencing Kit for

labeling samples. All procedures were performed according to the manufacturers’ protocols. The obtained

sequences were aligned with a BLAST tool to a genomic Ensembl reference sequence (TNFAIP3:

ENSG00000118503). Mutations were confirmed by a second PCR reaction and mutation analysis and

their somatic origin verified by sequencing of the corresponding normal DNA.

Supplementary Tables

Supplementary Table SI: SNPs investigated in the study.

Gene

SNP

rs-number Genotyping

method

Genotyping assay

ID

TNFAIP3

G>A

rs6920220

TaqMan

C__29431952_10

c.305A>G (N102S)

N/A

PCR-RFLP

N/A

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

(29)

For Peer Review

NFκB

-94 ins/del ATTG

rs28362491 TaqMan

NLRP3

14348 G>A

(ENST00000336119)

rs10733113 TaqMan

C_30713847_10

c.2107C>A(Q705K)

rs35829419 TaqMan

C__25648615_10

CARD8

c.30T>A (C10X)

rs2043211

TaqMan

C__11708080_1

TLR4

c.896A>G (D299G)

rs4986790

TaqMan

C_11722238_20

c.1196C>T (T399I)

rs4986791

TaqMan

C_11722237_20

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

(30)

For Peer Review

Supplementary Table SII: LOH markers, primers and PCR conditions

Marker Chrom. position Forward primer (5'-3') Reverse primer (5'-3') Type of Marker (Sense strand repeat)

D6S292 136.3 Mb FAM-5' agaactaaagttgcctgttc 5' acaagacacaatctcagtgg Dinucleotiderepeat (GT) D6S1009 137.3 Mb FAM-5' gagtgacaaactatggtccc 5' ctaggcaacagagtgaaaga Tetranucleotide repeat (CTAT) rs6920220 138.0 Mb ABI TaqMan assay ABI TaqMan assay G>A SNP

D6S(1381118) 138.11 Mb FAM-5' caaatgatgggtaatagtcgc 5' ttctgggatagtttgtggg Dinucleotiderepeat (CA) rs5029940 138.19 Mb FAM-5' ttgctgggtcttacatgcag 5' cccaccatggagctctgtta DIP (CCT/-)

D6S(1382694) 138.27 Mb FAM-5' ctcacagactatcggacatac 5' aagagcagcaactctttcg Double-Dinucleotiderepeat (CA…CT) D6S1587 138.42 Mb FAM-5' aggcagactgtccatcttta 5' acagaaacgcagcatctatt Dinucleotiderepeat (GT)

D6S1569 139.05 Mb FAM-5' agtatgcaagaaactgctagat 5' tatgtgtgtagtactgcttctga Dinucleotiderepeat (GT)

Multiplex-PCR

conditions: PCR-program

0.5 UDreamTaqDNA polymerase 95 °C for 3 min 2,0 mM Mg2+; 0,2 mM dNTPs 95 °C for 0:30 min 15 µl reactions 58 °C for 0:30 min

72 °C for 0.30 min (Pool1) or 0:40 min (Pool2) 72 °C for 5 min

POOL 1 Primer conc. (µM)

D6S292 0.4

D6S1009 0.2

D6S1587 0.2

POOL 2 Primer conc. (µM) D6S(1381118) 0.3 rs5029940 0.2 D6S(1382694) 0.4 D6S1569 0.5 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

(31)

For Peer Review

Supplementary Table SIII: TNFAIP3 Gene Expression and LOH

Patient ID

Genotype

rs6920220 LOH

Relative TNFIP3 expression (normal tissue)

Relative TNFIP3 expression (tumor tissue) L9 GG No LOH 1,18 1,70 L13 GG No LOH 5,48 1,39 L23 GG No LOH 1,98 0,27 L28 GG No LOH 1,06 0,15 L31 GG No LOH 1,91 1,14 L33 GG No LOH 3,23 1,75 L35 GG No LOH 1,02 0,41 L36 GG No LOH 2,64 0,46 L37 GG No LOH 0,99 1,17 L38 GG No LOH 0,62 0,55 L39 GG No LOH NA 0,61 L42 GG No LOH 1,28 48,42 L44 GG No LOH 2,28 1,66 L49 GG No LOH 4,37 0,97 L51 GG No LOH 0,70 3,48 L61 GG No LOH NA 7,31 L67 GG No LOH 1,91 1,09 L83 GG No LOH 1,53 1,38 J28 GG No LOH 0,94 0,53 J37 GG No LOH NA 0,66 J43 GG No LOH NA 0,37 J69 GG No LOH 30,31 34,60 J95 GG No LOH 2,29 0,38 J113 GG No LOH 4,02 0,56 J114 GG No LOH 1,35 0,67 J115 GG No LOH 2,03 0,36 J134 GG No LOH 0,88 0,85 J109 GG NA 2,53 0,46 L8 GA No LOH 1,89 0,21 L22 GA No LOH 2,07 0,88 L24 GA No LOH 0,90 0,93 L25 GA No LOH 0,79 0,52 L40 GA No LOH 1,66 1,20 L48 GA No LOH 0,31 0,85 L50 GA No LOH 1,11 0,65 L56 GA No LOH 2,90 0,57 J6 GA No LOH 0,41 1,08 J7 GA No LOH NA 1,04 J11 GA No LOH 1,06 1,71 J32 GA No LOH NA 0,97 J33 GA No LOH 0,60 0,60 J35 GA No LOH 1,08 0,95 J42 GA No LOH NA 0,28 J47 GA No LOH 1,93 1,37 J85 GA No LOH 1,26 0,93 J110 GA No LOH 1,30 0,27 J112 GA No LOH 1,48 1,02 L1 AA No LOH NA 3,49 L26 AA No LOH 0,99 1,34 L68 AA No LOH 0,93 1,80 L75 AA No LOH 0,79 0,57 J2 AA No LOH 2,21 1,57 J54 AA No LOH 1,78 0,80 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

(32)

For Peer Review

J64 AA No LOH 1,16 NA J100 AA No LOH 2,32 0,25 J132 AA No LOH 1,08 0,48 L46 GG LOH 0,56 0,56 J40 GG LOH NA 1,69 J45 GG LOH NA 0,32 L32 GA LOH 0,82 0,19 L41 GA LOH 0,82 0,53 L79 GA LOH 1,22 0,47 J20 GA LOH NA 2,54 J39 GA LOH 0,91 0,33 J66 GA LOH 1,14 0,19 J108 GA LOH 0,74 0,16 J122 GA LOH 1,77 0,45 L54 AA LOH 1,61 0,10 J119 AA LOH 0,71 0,47 J8 GG NA NA 0,75 J17 AA NA NA 0,34 Genotype rs6920220 Relative TNFIP3 expression (tumor tissue) (mean/media n/cases) Relative TNFIP3

expression (Normal tissue) (mean/median/cases) GG 4.18 / 0.85 / 273.22 / 1.91 / 23 GA 0.84 / 0.93 / 191.30 / 1.18 / 16 AA 1.29 / 0.80 / 8 1.41 / 1.12 / 8 GG 0.86 / 0.56 / 3 0.56 / 0.56 / 1 GA 0.61 / 0.39 / 8 1.06 / 0.91 / 7 AA 0.29 / 0.29 / 2 1.16 / 1.16 / 2 no LOH LOH 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

(33)

For Peer Review

Supplementary Table SIV: PCR-primers and conditions for SSCA and PCR-RFLP analyzes of TNFAIP3 Gene-exon-segment Application Forward primer (5'-3') Reverse primer (5'-3') product

size / bp TA / °C

TNFAIP3-2a PCR/SSCA tgcaggcagctatagaggagt ggatgatctcccgaaactga 282 55

TNFAIP3-2b PCR/SSCA tgcaccgatacacactggaaatgtt gcgggagctatcacccaggc 216 55

TNFAIP3-3 PCR/SSCA ttgctgggtcttacatgcag cccaccatggagctctgtta 299 55

TNFAIP3-3* PCR/SSCA gtcctcaggtgacggcaat cccaccatggagctctgtta 245 55

TNFAIP3-4 PCR/SSCA agggagtacaggatacattcaagc aaggctgaaagcatttaagtacaga 245 55

TNFAIP3-5 PCR/SSCA acctaagggcctcattttcc aagcaaaaaggaaaaccctga 277 55

TNFAIP3-5* PCR/SSCA gggcctcattttcctttctc ttttctcaccaggcccact 213 _{55, 57.5}# TNFAIP3-6 PCR/SSCA tgagatctacttacctatggccttg cagatgacacaggagagagctg 288 55 TNFAIP3-7a PCR/SSCA tgtaaaatcttgtgtgtgattttgtg tgggcgtttcacattttaca 246 55

TNFAIP3-7b PCR/SSCA ccagaatcccatggaacc tcatggcagtggtctcac 323 58

TNFAIP3-7c PCR/SSCA acagcacccagccctttt gcaagtactgcagatcccatt 231 55

TNFAIP3-7d PCR/SSCA ctgcctccaggatgttac gcttttctgcacttgctcgt 261 58

TNFAIP3-7e PCR/SSCA gtctcaagctgcacggactc aggaacaaaaccccttctgg 194 55

TNFAIP3-8 PCR/SSCA gacccctatgtggtactaactagcat ggagggaaagctcctcca 250 55

TNFAIP3-9a PCR/SSCA tgtgctctccctaagaaatgtg tgggatgctgacactccat 201 55

TNFAIP3-9b PCR/SSCA ctcctgcaagaacatcct tagcaccatgatgactgac 282 58

c.305 A>G (N102S) § _PCR-RFLP _{ttgctgggtcttacatgcag} _{cccaccatggagctctgtta} ₂₉₉ ₅₅ PCR/SSCA primers coverage of 2915 bp of which 2373 bp are coding.

PCR conditions for PCR/SSCA: Denaturation step at 94-95°C for 2-3 min.

Cycles were of denaturation at 94-95°C fro 20-30 sec, followed by annealing at according temperature for 20-30 sec, elongation step at 72°C for 30-40 sec. Final extension was at 72°C for 5min.

* These primer(s) were used in order to avoid common polymorphisms that could mask real mutational mobility shifts. #_{Annealing temperature during SSCA labelling for 15 cycles}

§ _{PCR product subsequently digested with MunI (Fermentas, St. Leon-Rot, Germany) at 37°C for 16 h.}

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