Polymorphisms in integrin genes and lymphoma risk

Abstract

Immune deficiency is one of the best characterized and strongest known risk factors for non-Hodgkin lymphoma (NHL). We studied the association between single nucleotide polymorphisms (SNPs) in integrin genes that are important components in human innate immunity and the risk of NHL in a population-based case–control study of women in Connecticut, USA. A total of 373 tag SNPs in 33 gene regions were included in the analysis of 448 cases and 525 controls. The ADAM19 rs11466782 SNP was associated with an increased risk of lymphoma (OR, 1.73; 95 % CI, 1.28–2.35; P _additive = 0.0004), and the ICAM3 rs2304240 (OR, 0.67; 95 % CI, 0.52–0.86; P _additive = 0.002) and the PTGDR rs708486 SNPs (OR, 0.75; 95 % CI, 0.63–0.90; P _additive = 0.002) were associated with reduced risk of lymphoma. Two gene regions (ADAM19 (P=0.009) and ICAM3 (P=0.009)) displayed global associations with lymphoma risk at the P<0.01 level. While our results suggest that genetic polymorphisms in integrin genes may play a role in the genesis of lymphoma in women, they should be viewed as exploratory until they are replicated in additional populations.

Introduction

Immune deficiency is one of the best characterized and strongest known risk factors for non-Hodgkin lymphoma (NHL) (Grulich et al, 2007). Certain autoimmune diseases and infections (EBV, HIV), along with weakened immunity have been found to increase lymphoma susceptibility.

Innate immunity is the first barrier of the human body to clear non-specific antigens and interacts with the adaptive immune system during physiological and chronic inflammation (Kabelitz & Medzhitov, 2007). Integrins are important components in innate immunity because they mediate attachment between cells or between a cell and the extracellular matrix. They also play a role in cell communication and transduce cell-cell and cell-matrix information, allowing rapid and flexible responses to changes in the environment (Hynes, 2002).

To test the hypothesis that genetic variation in genes coding integrins may play a role in lymphoma carcinogenesis, we evaluated genetic polymorphisms in 33 integrin genes in a population-based case-control study among women in Connecticut.

Materials and Methods

Details of this population-based case control study are described elsewhere (Morton et al, 2003). Briefly, incident female NHL cases were identified using the Yale Comprehensive Cancer Center’s Rapid Case Ascertainment Shared Resource. Eligible cases included female adults of Connecticut diagnosed with incident NHL (ICD-O: M-9590-9642, 9690-9701, 9740-9750). A population-based control group was enrolled in Connecticut and the controls were frequency matched to cases by age within 5-year groups. From 1995 to 2001, 601 NHL eligible cases (72% of all eligible cases) and 717 qualified controls participated in this study. There were 145 diffuse large B-cell lymphomas (DLBCL), 105 follicular lymphomas (FL), 52 small cell lymphocytic lymphomas/chronic lymphocytic leukemias (SLL/CLL), 29 marginal zone lymphomas (MZBL), and 32 T-cell lymphomas. Blood samples were available from 461 cases and 535 controls.

DNA was extracted from blood samples using phenol-chloroform extraction and genotyped by a GoldenGate Illumina assay. Genes in the same gene region were also genotyped. Gene region was defined as a chromosome region with multiple genes within a range of 200,000 bases upstream or downstream. A total of 973 subjects (448 cases and 525 controls) were successfully genotyped. The assay was initially designed to examine 386 single nucleotide polymorphisms (SNPs) in 33 gene regions (43 genes). However, 13 SNPs were excluded due to assay failure. A final total of 373 tag SNPs in 33 gene regions were included in the analysis (Supplementary Table 1). Approximately 99.7% of assays had a completion rate of ≥95 % for all samples and 99.5% of assays reached ≥95% concordance among quality control samples.

The genotype data were analyzed by comparing the homozygotes and the heterozygotes of the variant with the homozygote of the common allele in an unconditional logistic regression, adjusted for age and race. An additive model was applied by treating the genotypes as values of 0, 1, and 2 in one model in order to test for a linear trend. A gene region-based test, permutation-based resampling method (10,000 permutations), was applied to assess the true statistical significance of the smallest p-trend within each gene region. This method automatically adjusts for the number of tag SNPs tested within that gene region, as well as the underlying linkage disequilibrium pattern. Haplotype analysis was carried out for all gene regions and all contiguous locus subsets in a gene region were examined to identify sub-haplotypes with the strongest omnibus association with lymphoma risk. Given the possibility of false positive results due to multiple comparisons, we evaluated the robustness of our results using the False Discovery Rate (FDR). All data were analyzed with the Statistical Analysis Software, version 9.13 (SAS Institute Inc, 1996) if not specified otherwise.

Results

Selected demographic characteristics of cases and controls were similar with respect to age, race, education, and family history of cancer (Table 1).

Table 1.

Characteristics of cases and controls.

Characteristic	Cases (%)(n=448)	Controls (%)(n=525)	P a
Age (years)
<50	89 (20%)	99 (19%)	0.17
50–69	218 (49%)	231 (44%)
70+	141 (31%)	195 (37%)
Race
Caucasian	432 (96%)	494 (94%)	0.09
Other	16 (4%)	31 (6%)
Education
<12 years	58 (13)	63 (12)	0.46
12 years–Some college	337 (75)	386 (74)
College graduate or higher	53 (12)	76 (14)
Family history b
No	93 (21%)	129 (25%)	0.16
Any cancer	355 (79%)	396 (75%)

^aPearson χ² test;

^bFamily history of cancer in first degree relatives.

The genotype frequencies for cases and controls, and the effect of these SNPs and gene regions on lymphoma risk are provided in Supplementary Table 2. Among the 373 tag SNPs in 33 gene regions, three SNPs were statistically significant at the level of 0.005 in the additive model (Table 2). ADAM19 rs11466782 was associated with increased risk of lymphoma (OR, 1.73; 95 % CI, 1.28–2.35; P _additive = 0.0004), and ICAM3 rs2304240 (OR, 0.67; 95 % CI, 0.52–0.86; P _additive = 0.002) and PTGDR rs708486 (OR, 0.75; 95 % CI, 0.63–0.90; P _additive = 0.002) were associated with reduced risk of lymphoma. The association of ADAM19 rs11466782 with NHL risk was also significant for several NHL subtypes, especially for FL (OR _trend, 2.20; 95 % CI, 1.37–3.52; P, 0.001).

Table 2.

Logistic regression analysis of selected SNPs on the risk of NHL among women in Connecticut, adjusted for age and race.

	Controls	%	Cases	%	OR	95% CI	P
ADAM19 rs11466782 (IVS14-152 A>G)
AA	445	85	341	76	1
AG	77	15	101	23	1.76	1.26–2.45	0.0009
GG	3	1	6	1	2.65	0.66–10.71	0.17
AG/GG	80	15	107	24	1.79	1.29–2.48	0.0004
Trend					1.73	1.28–2.35	0.0004
ICAM3 rs2304240 (Ex2-35 G>A)
GG	349	66	335	75	1
AG	154	29	107	24	0.73	0.55–0.98	0.04
AA	22	4	6	1	0.29	0.11–0.72	0.01
AG/AA	176	34	113	25	0.68	0.51–0.90	0.01
Trend					0.67	0.52–0.86	0.002
PTGDR rs708486 (IVS1-478 A>G)
AA	163	31	168	38	1
AG	252	48	215	48	0.80	0.60–1.06	0.12
GG	109	21	65	15	0.55	0.38–0.81	0.002
AG/GG	361	69	280	63	0.72	0.55–0.95	0.02
Trend					0.75	0.63–0.90	0.002

Two gene regions (ADAM19 (P=0.009) and ICAM3 (P=0.009)) displayed global association with lymphoma risk at the P<0.01 level. The smallest FDR value (0.16) was obtained for ADAM19 rs11466782. Haplotype analyses did not provide additional information beyond results from individual SNP analyses.

Discussion

We found that several SNPs in genes involved in integrins were associated with lymphoma risk in women. This is the first report that a polymorphism in ADAM19 is associated with the risk of lymphoma.

The protein coded by ADAM19 is a cell surface glycoprotein with several functional domains and works as an endopeptidase/metalloprotease that cleaves extracellular matrix proteins and sheds growth factors and cytokines. It plays essential roles in embryo implantation, cardiovascular morphogenesis, neurogenesis, and other developmental processes (Qi et al, 2009). ADAM19 is upregulated in human brain tumors and is also over-expressed by many human cancerous cell lines including cancers of the colon, ovary, lung, and brain (Qi et al, 2009). Abnormally high expression of this gene is also linked to inflammation and fibrosis of the lung and kidney. ADAM19 may participate in the pathology of cancers through several mechanisms by promoting cell growth and invasion (Mochizuki & Okada, 2007). The ADAM19 rs11466782 is in Intron 14 (ivs14 -152 T>C), close to exon 15, which is a conserved cysteine-rich domain that regulates the protein's metalloprotease activity.

ICAM3 is a transmembrane glycoprotein and expressed by all leucocytes. It functions not only as an adhesion molecule, but also as a potent signalling molecule for immune response. The ICAM3 was found to be overexpressed in lymphoid neoplasms (Beltran & Boissier, 2009; Cordell et al, 1994). The polymorphism ICAM3 rs2304240 is located at one end of a conserved intercellular adhesion molecule domain and close to the other important conserved immunoglobulin domain.

In summary, we found that genetic polymorphisms in a number of integrin genes were associated with lymphoma risk in women. This suggests that integrins, especially ADAM19, may play a role in the pathogenesis of lymphoma. While we attempted to account for possible spurious findings due to multiple comparisons by evaluating False Discovery Rates of the minP results, it is still possible that our analyses may have led to false positive or false negative findings; therefore, these results need to be replicated.