Death Receptor 4 Variants and Colorectal Cancer Risk

Apoptosis or programmed cell destruction is a key regulator of tissue homeostasis. An imbalance between cell death and proliferation may result in tumor formation. The tumor necrosis factor–related apoptosis-inducing ligand (TRAIL) activates the extrinsic apoptotic pathway through the engagement of the proapoptotic death receptor 4 (DR4, TNFRSF10A, TRAILR-1), a member of the tumor necrosis factor receptor superfamily. DR4 consists of two extracellular cysteine-rich, ligand-binding pseudorepeats (50s and 90s loops), one single transmembrane helix as well as the apoptosis-triggering cytoplasmic death domain. Suppression of cell death signaling due to detrimental alterations in DR4 involves a deregulated cell proliferation and predisposes to cancer (1-6).

Immunohistochemical studies have shown extensive expression of DR4 both in sporadic and hereditary colorectal neoplasms together with a higher degree of apoptosis, suggesting neoplastic colorectal cells to be prone to TRAIL-induced apoptosis (5, 6). Moreover, Drosopoulos et al. (7) have reported the induction of cell death in human colon adenocarcinoma cells in a mitogen-activated protein kinase/extracellular signal-regulated kinase kinase–dependent manner by TRAIL. Mutations in DR4, mapping to chromosome 8p21-22, have been described in several human cancers, such as breast, lung, head and neck cancer, and non-Hodgkin's lymphoma (8-10), but have not yet been studied in colorectal cancer. However, frequent loss of heterozygosity on chromosome 8p21-23 has been observed in hepatocellular carcinoma as well as in colorectal cancer, pointing to DR4 as a candidate tumor suppressor gene for both cancers (11).

Recently, we have investigated the effects of DR4 Thr²⁰⁹Arg (626C>G) and Glu²²⁸Ala (683A>C) on familial breast cancer risk (12). Although neither variant showed significance regarding disease risk, the analysis of the rare haplotype 626C-683C resulted in an increased risk of breast cancer (12).

Given these findings, we estimated the colorectal cancer risk associated with the DR4 Thr²⁰⁹Arg and Glu²²⁸Ala variants in a German population-based case-control study.

Colorectal cancer cases and controls were drawn from the German Darmkrebs: Chancen der Verhütung durch Screening (DACHS) study, a large population-based case-control study carried out in the Rhine-Neckar-Odenwald region in the southwest of Germany (13). The cases consisted of 659 unrelated male and female subjects (33-91 years of age; median 68) with incident invasive colorectal cancer diagnosed between January 2003 and March 2005. The median time between diagnosis and ascertainment of the cases was 14 days. Detailed information is given in Table 1. The control group comprised 607 unrelated male and female individuals (34-94 years of age; median 67). None of the controls had a personal history of colorectal cancer. They were randomly selected from lists of population registries and frequency-matched to cases by 5-year age groups, sex, and county of residence. Cases and controls were eligible if they were aged 30 years and above, German-speaking, and mentally and physically capable of participating in an in-person interview of about 1 hour. A detailed description of the study population is given in Table 1. The study was approved by the Ethics Committees of the University of Heidelberg and the State Medical Boards of Baden-Wuerttemberg and Rhineland-Palatinate (Germany).

Details of the data collection procedures are reported elsewhere (13-15). In brief, study subjects were asked to participate in an in-person interview and to give a blood sample or a mouthwash when blood samples were not available. Information on demographic factors, anthropometric measures, medical history, including medication and screening, family history of colorectal cancer, reproductive history, and lifestyle factors were collected by trained interviewers using a standardized questionnaire.

Based on our previous results for familial breast cancer (12), we selected and analyzed the effects of DR4 Thr²⁰⁹Arg (rs4871857) and Glu²²⁸Ala (rs17088993) on colorectal cancer risk. In the single nucleotide polymorphism database, there are five additional coding single nucleotide polymorphisms in DR4: Thr³³Ile (rs20577), Pro¹⁰⁵Arg (rs11986840), His¹⁴¹Arg (rs6557634), Asn²⁹⁷His (rs17088980), and Arg⁴⁴¹Lys (rs2230229). To validate these variants, we sequenced 20 randomly selected breast cancer DNA samples (12). DR4 His¹⁴¹Arg was not analyzed because, according to Fisher et al. (9), it segregated together with Thr²⁰⁹Arg in 96% of their samples (n = 243). None of the investigated variants, except for DR4 Arg⁴⁴¹Lys, could be confirmed. The rare Arg⁴⁴¹Lys was shown not to segregate with the high-risk haplotype 626C-683C and therefore not expected to be disease-causing.

Blinded genotyping was done by TaqMan allelic discrimination. TaqMan primers and probes were provided by the Assay-by-Design service (Applied Biosystems, Foster City, CA) and designed based on the Genbank NT_023666 sequence. Sequences of primers and probes were as follows: Thr²⁰⁹Arg, GGTGGTGAGGAAAGGTCAAG (forward) and ATGGGGTCAGGGCTGATAG (reverse), VIC-TCTCACCCTGTGCTGC, and FAM-TCTCACCCTCTGCTGC; Glu²²⁸Ala, CCCCTGCAGATACGAGGAG (forward) and caggaaaagacaggagtctcg (reverse), VIC-TGACATCGAGTGTGTCC, and FAM-CATCGCGTGTGTCC. Detailed information about sequencing and TaqMan allelic discrimination methods have been described previously (12, 16). To ensure accuracy and to exclude genotyping error, ≥10% of samples were re-genotyped. The duplicates showed a concordance rate of 100%.

Genotype-specific odds ratios (OR), 95% asymptotic confidence intervals (95% CI), and Ps were computed by unconditional logistic regression using a tool offered by the Institute of Human Genetics, Technical University Munich (Munich, Germany).⁷

Adjustment for sex and age and the two-sided Cochran-Armitage test for trend were done using the Statistical Analysis System software (version 9.1; SAS Institute, Inc., Cary, NC). The Cochran-Armitage statistic was used to test for trend in genetic ORs with increasing stage. We used the software package haplo.stats to test for association of indirectly deduced haplotypes (17, 18). It does the joint modeling of observed multipoint single nucleotide polymorphism genotype and phenotype, using a generalized linear model framework to test for haplotype-trait association and to calculate ORs and 95% CIs. The most common haplotype was chosen as the reference group. To account for the uncertainty of the haplotype estimation, each haplotype pair, consistent with the genotype of an individual and weighted by its estimated probability, is used to model the individual's phenotype. To support our findings, we derived empirical Ps for haplotype-trait association under an additive model, using the score statistic implemented in haplo.score, based on 20,000 simulations.

Genotype frequencies for both polymorphisms analyzed were in agreement with Hardy-Weinberg expectations in controls (P = 0.48 for Thr²⁰⁹Arg and P = 0.55 for Glu²²⁸Ala, respectively). Although there were no significant differences in genotype frequencies between colorectal cancer cases and controls for DR4 Glu²²⁸Ala (683A>C; OR, 1.19; 95% CI, 0.94-1.50; Table 2), heterozygous carriers of DR4 Thr²⁰⁹Arg (626C>G) showed a significant association with a decreased colorectal cancer risk (OR, 0.73; 95% CI, 0.54-0.97; Table 2). The risk, however, did not decrease with the number of variant alleles, which may point to a finding by chance. Adjustment for sex and age made no relevant difference to the findings. The allele frequencies for both single nucleotide polymorphisms among cases and controls were consistent with the allele frequencies reported in Caucasian populations (9, 12). Stratification by sex and age revealed a significant association of DR4 Thr²⁰⁹Arg with a decreased colorectal cancer risk for male heterozygotes (OR, 0.68; 95% CI, 0.46-0.99; Table 2). Because the association was independent of allele dosis (Table 2), it may be a chance finding. DR4 Glu²²⁸Ala showed an increased colorectal cancer risk for females in an allele dose-dependent manner (OR, 1.52; 95% CI, 1.04-2.22 and OR, 1.95; 95% CI, 0.80-4.77; P_trend = 0.01; Table 2) as well as a decreased risk for Arg²⁰⁹ carriers ≥65 years (OR, 0.65; 95% CI, 0.46-0.92; Table 2), the equivalent of an increased risk for homozygote carriers of the common C allele (Thr²⁰⁹). Analysis by subsite showed the same effect on rectal cancer risk (OR, 0.67; 95% CI, 0.48-0.95; Table 3). Moreover, the risk associated with the DR4 Glu²²⁸Ala variant increased with advanced colorectal cancer stages, revealing a significant risk for Ala²²⁸ carriers (OR, 1.40; 95% CI, 1.04-1.86; P_trend = 0.04; Table 3), which may be a sign of a greater tumor aggressiveness in Ala²²⁸ (683C) carriers.

When analyzing haplotypes, the results from logistic regression pointed to a haplotype effect (global P = 0.07). The combination of DR4 Thr²⁰⁹Arg and Glu²²⁸Ala led to four distinct haplotypes (Table 4). In the generalized linear model framework, 626C-683A, 626G-683A, and 626G-683C haplotypes showed no association with colorectal cancer. The rare haplotype 626C-683C was found to be more common in cases than in controls, resulting in an increased colorectal cancer risk (OR, 2.37; 95% CI, 0.98-5.76; P = 0.06; Table 4). The score statistic yielded a significant empirical P of 0.03, based on 20,000 simulations, for 626C-683C. The identification of the 626C-683C haplotype as risk factor for colorectal cancer is in line with our recently published findings on familial breast cancer risk (626C-683C: OR, 3.52; ref. 12). Because both DR4 Thr²⁰⁹Arg and Glu²²⁸Ala reside next to the DR4 TRAIL-binding ectodomain, the observed association may be due either to the joint effects of Thr²⁰⁹ and Ala²²⁸ by altering ligand-binding and thus apoptotic signaling or to an unknown causative variant in strong linkage disequilibrium with the haplotype.

In summary, we detected a significant association of the DR4 Thr²⁰⁹Arg variant with a decreased colorectal cancer risk for heterozygotes and Arg²⁰⁹ carriers ≥65 years of age. Female Ala²²⁸ variant carriers exhibited a significantly enhanced risk dependent on allele dose. Subsite analysis revealed a protective effect of Arg²⁰⁹ on rectal cancer risk as well as an increased risk for advanced colorectal cancer stages. In addition, the DR4 626C-683C haplotype conferred a 2.4-fold colorectal cancer risk, suggesting its relevance in human cancer.

Our results on DR4 variants and colorectal cancer risk are explorative. Given the modest stratum-specific Ps for association, they may be artifacts due to multiple testing and need confirmation in larger studies.

We thank the clinicians who supported the study, all patients and control individuals who participated in the study, Belinda-Su Kaspereit for technical assistance, and Kerstin Wagner for critical comments on the article.