BRIEF RESEARCH COMMUNICATION Association of the a4 Integrin Subunit Gene (ITGA4) With Autism Catarina Correia,1,2 Ana M. Coutinho,1 Joana Almeida,3 Raquel Lontro,3 Cristina Lobo,3 Teresa S. Miguel,4 Madalena Martins,2 Louise Gallagher,5,6 Judith Conroy,5 Michael Gill,5,6 Guiomar Oliveira,3 and Astrid M. Vicente1,2* 1 Instituto Gulbenkian de Ci^encia, Oeiras, Portugal 2 Instituto Nacional de Saúde Dr. Ricardo Jorge, Lisboa, Portugal 3 Centro de Desenvolvimento da Criança, Hospital Pediatrico de Coimbra, Coimbra, Portugal Direcç~ao Regional de Educaç~ao da Regi~ao Centro, Coimbra, Portugal 4 5 Department of Genetics, Smurfit Institute, Trinity College, Dublin, Ireland 6 Department of Psychiatry, Trinity College, Dublin, Ireland Received 27 March 2008; Accepted 14 January 2009 In the present work, we provide further evidence for the involvement of the integrin alpha-4 precursor gene (ITGA4) in the etiology of autism, by replicating previous findings of a genetic association with autism in various independent populations. The ITGA4 gene maps to the autism linkage region on 2q3133 and is therefore a plausible positional candidate. We tested eight single nucleotide polymorphisms (SNPs) in the ITGA4 gene region for association with autism in a sample of 164 nuclear families. Evidence for association was found for the rs155100 marker (P ¼ 0.019) and for a number of specific marker haplotypes containing this SNP (0.00053 < P < 0.022). a4 integrins are known to play a key role in neuroinflammatory processes, which are hypothesized to contribute to autism. In this study, an association was found between the ITGA4 rs1449263 marker and levels of a serum autoantibody directed to brain tissue, which was previously shown to be significantly more frequent in autistic patients than in age-matched controls in our population. This result suggests that the ITGA4 gene could be involved in a neuroimmune process thought to occur in autistic patients and, together with previous findings, offers a new perspective on the role of integrins in the etiology of autism to which little attention has been paid so far. 2009 Wiley-Liss, Inc. Key words: autism; ITGA4 gene; neuroimmune process A strong genetic component has been postulated for autism, a behavioral syndrome characterized by deficits in social interaction, impaired communication, and restricted and stereotyped behaviors [Folstein and Rosen-Sheidley, 2001]. While multiple genes are likely involved, the results from several genome scans have provided convincing evidence for the presence of an autism susceptibility loci on chromosome 2q31-q33 [Buxbaum et al., 2001; IMGSAC, 2001; Shao et al., 2002; Rabionet et al., 2004; Romano et al., 2005; Lauritsen et al., 2006; Spence et al., 2006]. 2009 Wiley-Liss, Inc. How to Cite this Article: Correia C, Coutinho AM, Almeida J, Lontro R, Lobo C, Miguel TS, Martins M, Gallagher L, Conroy J, Gill M, Oliveira G, Vicente AM. 2009. Association of the a4 Integrin Subunit Gene (ITGA4) With Autism. Am J Med Genet Part B 150B:1147–1151. One possible candidate in this region is the integrin alpha 4 gene (ITGA4), which spreads over 79.7 kb and encodes the integrin alpha-4 precursor. Several independent studies have identified this gene as a potential autism susceptibility locus. Faham et al.  carried out a case/control study, applying multiplexed variation screening (MVS) to scan the genes mapping under the chromosome 2q31-33 linkage peak defined by Buxbaum et al. , using probands from the Autism Genetic Resource Exchange (AGRE) families that were part of the original linkage scan. The results provided evidence for association between autism and one single nucleotide polymorphism (SNP) variant in exon 22 of the ITGA4 gene. Conroy et al.  followed up a previous finding of an Additional Supporting Information may be found in the online version of this article. Grant sponsor: Fundaç~ao Calouste Gulbenkian; Grant sponsor: Fundaç~ao para a Ci^encia e Tecnologia; Grant number: POCTI/39636/ESP/2001. *Correspondence to: A. M. Vicente, Instituto Gulbenkian de Ci^encia, Rua da Quinta Grande, 6, 2781 Oeiras, Portugal; Instituto Nacional de Saúde Dr. Ricardo Jorge, Av. Padre Cruz, 1649-016 Lisboa, Portugal. E-mail: email@example.com, firstname.lastname@example.org Published online 3 March 2009 in Wiley InterScience (www.interscience.wiley.com) DOI 10.1002/ajmg.b.30940 1147 1148 AMERICAN JOURNAL OF MEDICAL GENETICS PART B interstitial deletion on chromosome 2q, within the region showing evidence for linkage [Gallagher et al., 2003], and found a significant association with several SNPs and haplotypes within the ITGA4 gene region in two independent populations. Significant association with autism is also reported by Ramoz et al. , who performed a dense marker scan across candidate genes mapping across the 2q24-q33 linkage region in AGRE families, including several markers within ITGA4 and additional genes in this region. Integrins are membrane spanning, non-covalently linked ab heterodimers that act as cell–cell and cell–extracellular matrix (ECM) adhesion receptors throughout the body. The a4 integrin chain dimerizes with either the b1 chain or the b7 chain. The a4b1 integrin, also known as very late antigen4 (VLA4), is the most important member of the beta 1 integrin subfamily of integrins, which are expressed in a variety of different cell types, including neurons, glial cells, meningeal cells, and endothelial cells, subject to regional and developmental regulation [Milner and Campbell, 2002]. The activation of integrins by their extracellular ligands results in a number of changes of integrin properties, such as spatial localization, internalization, ligand affinity, intracellular association with signaling proteins, interaction with the cytoskeleton and transcriptional modulation. In the central nervous system (CNS), integrins mediate adhesive and migratory events in processes as diverse as synaptogenesis, activation of microglia, and stabilization of the endothelium and the blood–brain barrier [Milner and Campbell, 2002], many of which may be disrupted in autism [Folstein and Rosen-Sheidley, 2001]. Integrins are also vital in situations of CNS inflammation, which is thought to occur in autistic individuals. We have previously reported the widespread occurrence of serum autoantibody reactivities to brain tissue in autistic patients [Silva et al., 2004], while autoantibodies to specific brain proteins, such as myelin basic protein (MBP), have been found in sera from children with autism [Singh et al., 1993, 1998; Connolly et al., 1999]. There is also evidence for an active ongoing chronic neuroinflammatory process in multiple areas of the CNS of autistic individuals, with marked activation of microglia and astroglia and increased levels of proinflammatory and anti-inflammatory cytokines [Vargas et al., 2005]. In vitro experiments show that microglial activation is accompanied by increased expression of a4b1 integrin, which is stimulated by inflammatory cytokines [Hailer et al., 1996]. a4 integrin expressed on the surface of lymphocytes binds to their respective endothelial counter-receptor, VCAM1, playing a fundamental role in their adhesion to the vascular endothelium and migration into the CNS. On the other hand, the use of a4b1 antagonists has been shown to reduce inflammatory cellular infiltration in various tissues, haltering the progression of brain inflammatory diseases such as multiple sclerosis (MS) [Von Andrian and Engelhardt, 2003]. Given its chromosomal location within a candidate region for autism and its role in the CNS, ITGA4 is a strong candidate gene for autism susceptibility. We further investigated the involvement of ITGA4 in autism by carrying out a replication study in a population sample of 164 nuclear families with one affected individual. Patients were diagnosed as previously described [Coutinho et al., 2004], and only idiopathic subjects with Developmental Quotient above 50 were included. The Ethical Committee at the HP approved the collection of data and biological specimens from patients for research purposes, and all participants signed an informed consent. Following on the previous association study in an independent population of Irish trios [Conroy et al., 2008], which guided the selection of markers to replicate identified associations and cover genetic variability throughout the gene, we tested eight SNPs at the ITGA4 gene: rs1449263 upstream of the gene, rs3770136 and rs1449260 in intron 2, rs155100 in intron 9, rs3770116 in intron 15, rs3770112 in intron 17, rs2305581 in intron 20, and rs3770105 in intron 22. All SNPs were genotyped using the competitive allele specific PCR system (KASPar). Markers were tested for Hardy–Weinberg equilibrium and call rates were above 90% (Supplementary Table I). TABLE I. Association Analysis of ITGA4 Individual Markers With Autism SNP rs1449263 rs3770136 rs1449260 rs155100 rs3770116 rs3770112 rs2305581 rs3770105 Allele A G A G A G A T A C C T A G A G Transmitted 56 59 25 31 45 39 60 37 20 18 37 30 20 20 40 29 Not transmitted 59 56 31 25 39 45 37 60 18 20 30 37 20 20 29 40 x2 0.0783 P-value 0.779657 0.6441 0.422269 0.4289 0.512543 5.5059 0.018991 0.1053 0.745563 0.7327 0.392056 0 1 1.7611 0.184544 CORREIA ET AL. 1149 FIG. 1. Haploview-generated LD map of the eight SNPs within ITGA4 gene in the Portuguese population. The numbers within the boxes indicate the D0 statistic values between corresponding two SNPs. Black shading indicates strong LD (no number means a score of 1), gray shading indicates uninformative, and white shading indicates strong evidence for recombination. Two-, three-, and four-marker haplotypes significantly associated with autism. Horizontal bars represent significant over transmitted (black lines) and under transmitted ITGA4 marker haplotypes (gray lines). The extended transmission disequilibrium test (ETDT) was used to examine the association of individual markers with autism [Sham and Curtis, 1995]. Haplotypes were analyzed using TRANSMIT [Clayton and Jones, 1999]. Pair-wise linkage disequilibrium between markers was calculated using the Haploview program. Other statistical analyses were performed using the SPSS package. All results are presented uncorrected for multiple testing. Association results for all markers are shown in Table I. A significant transmission disequilibrium of alleles at the rs155100 marker was found (c2 ¼ 5.5059; P ¼ 0.019), with preferential transmission of allele A. Evidence of association with autism was also found for multiple marker haplotypes involving this SNP (Fig. 1, Supplementary Table II). Intermarker linkage disequilibrium (LD) was calculated and we found that the associated marker rs155100 is located in a region with low levels of LD between two regions with high levels of LD (Fig. 1). The fact that only the rs155100 marker was associated with autism may therefore be explained by the LD structure of this gene in our population. Considering that all risk haplotypes identified include the A allele of marker rs155100, the association of these haplotypes is therefore likely driven by this marker. Although SNP rs155100, located in intron 9, per se is not likely to have a functional consequence, it may be in LD with a nearby functional variant. Notably, exon 9 encodes a FG-GAP domain important in ligand binding (http://pfam. sanger.ac.uk). Our results replicate previous studies reporting the association of the ITGA4 gene with autism in various population groups [Faham et al., 2005; Conroy et al., 2008; Ramoz et al., 2008], thus providing additional evidence for association of the ITGA4 gene with autism in an independent cohort. Different SNPs were, however, associated in the different studies, including our own. For instance, in the study by Conroy et al. SNP and haplotype associations were identified in the Irish population that were not replicated in the AGRE and/or Vanderbilt samples, although in some cases the combined sample provided evidence of association. On the other hand, specific SNPs were associated in the Vanderbilt but not in the Irish or AGRE populations. Faham et al. only found an associated SNP in exon 22, while a number of different markers in introns 13 and 15–28 were significantly associated in the study by Ramoz et al., even though these two studies may possibly have an important overlap in terms of probands tested, as both used AGRE families, and thus would be expected to identify associations in the same gene regions. The observation of association with different SNPs in different populations is not unexpected, and may have several explanations. Many of the tested SNPs are not likely to be functional polymorphisms, but may be in LD with a nearby functional variant; 1150 it is therefore very plausible that variable underlying patterns of LD lead to different SNPs being associated with autism in populations with different origins and likely variable ethnic constitution. Other factors may also contribute to the discrepancies in associated markers between populations. For instance, the SNPs tested in the different populations did not necessarily overlap. In our study, only one marker significantly associated in the work by Ramoz et al.  was tested, whereas the associated SNP in our population, rs155100, was only tested in the study of Conroy et al. Four markers were common between the Ramoz et al. study and the Irish population, of which only one was nominally associated in both studies. Study design also led to discrepancies: while the Ramoz et al. and Faham et al. studies possibly shared a proportion of the sample, they have a very different design, one searching for sequence variants in exons that might be associated with autism and the other testing linkage and association with non-functional SNPs. The discrepancies among populations regarding the associated SNPs may therefore be explained by different patterns of LD in the populations tested and/or by variable study designs with little SNP overlap. We have not attempted any correction for multiple testing of either SNPs or haplotypes, as there is no consensus regarding the most appropriate method for this purpose. The application of the most common correction methods, such as the Bonferroni correction, may result in an excess of type II errors, especially in relatively small samples, and therefore the replication, in several independent populations, of a genetic association is considered the best evidence of a true association. The present study is a replication of a previously reported association with this gene in multiple independent populations. It thus reinforces the suggestion of an involvement of the ITGA4 gene in autism susceptibility, indicating that this region should be further explored. One of the described pathologic functions of a4 integrins in humans is the recruitment of circulating activated T-cells, monocytes, and macrophages to the CNS [Von Andrian and Engelhardt, 2003]. Besides mediating leukocyte adhesion, a4b1 has been implicated in costimulatory signals for T-cell proliferation [Nojima et al., 1990; Sato et al., 1995] and in the differentiation of T cells and B cells through their interaction with fibronectin [Williams et al., 1991; Salomon et al., 1994]. In agreement with the hypothesis of a chronic neuroinflammatory process in autism, we have previously shown the widespread occurrence of autoantibodies to brain tissue in autistic individuals, suggestive of an immune dysregulation in autism [Silva et al., 2004]. More specifically, we defined patterns of autoantibody repertoires directed to brain tissue that were characteristic of subsets of autistic patients, and identified one specific autoantibody, to an as yet unidentified brain antigen, which strongly discriminated between patients and controls sera (P ¼ 0.00046, Mann–Whitney U-test). Autoantibody reactivities were assessed by a quantitative immunoblotting technique defining a quantitative population distribution of optical densities for each specific autoantibody reactivity. Given the role of a4 integrins in neuroimmune processes, we tested the association of ITGA4 with the occurrence and strength of this specific autoantibody reactivity in our patients, using a non-parametric analysis of variance (Kruskal–Wallis test). We found a significant association of the ITGA4 promoter marker rs1449263 with the optical density distri- AMERICAN JOURNAL OF MEDICAL GENETICS PART B bution of this autoantibody reactivity (c2 ¼ 9.58 (2 df); P ¼ 0.008). Interestingly, this marker is associated with susceptibility to MS [O’Doherty et al., 2007], with the MS risk allele (rs1449263 C) also more frequent in patients with higher levels of this autoantibody reactivity in our population of autistic individuals. However, this marker was not associated with autism and was not in LD with the associated marker. While we cannot fully explain this inconsistency, we believe it may mostly be related to methods of analysis and population issues, possibly reflecting a variation in statistical power to detect association, rather than diverse genetic effects. In fact, association with autism was tested using a family-based approach, examining allelic transmissions, while association with autoantibodies reactivities was tested using a non-parametric analysis of variance, which tests the effect of genotypes rather than alleles in the patient population. In addition, there is a small percentage of the population sample (8.6%) that is not overlapping in the two analyses. The upregulation of integrin a4, among other inflammation molecules, in conjunction with a well-documented autoantigen upregulation, represents a signature of enhanced immune cell activation and costimulation in MS, while a4 integrin antagonists are used for MS treatment because they prevent the migration of autoimmune cells to the CNS [Iglesias et al., 2004]. Together with other observations in autism and MS, our results lead us to hypothesize that a chronic neuroinflammatory process occurs in autism involving both cell-mediated and humoral autoimmune responses. This process requires the recruitment of inflammatory cells to the brain, which is known to be influenced by variants of the ITGA4 gene, while the association of specific autoreactivities with an ITGA4 marker can be explained by the parallel occurrence of the humoral and cellular processes. In summary, the present study provides supporting evidence for an association between the ITGA4 gene and autism and, taken together with previous work, suggests a plausible role for this integrin in a neuroinflammatory process occurring in some autistic patients. ACKNOWLEDGMENTS This work was supported by the Fundaç~ao Calouste Gulbenkian and by the Fundaç~ao para a Ci^encia e Tecnologia (POCTI/39636/ ESP/2001). C. Correia and A.M. Coutinho are supported by fellowships from the Fundaç~ao para a Ci^encia e Tecnologia (SFRH/BD/ 16907/2004 and SFRH/BD/3145/2000). The genotyping was performed by KBiosciences. REFERENCES Buxbaum JD, Silverman JM, Smith CJ, Kilifarski M, Reichert J, Hollander E, Lawlor BA, Fitzgerald M, Greenberg DA, Davis KL. 2001. Evidence for a susceptibility gene for autism on chromosome 2 and for genetic heterogeneity. Am J Hum Genet 68:1514–1520. Clayton D, Jones H. 1999. Transmission/disequilibrium tests for extended marker haplotypes. Am J Hum Genet 65:1161–1169. Connolly AM, Chez MG, Pestronk A, Arnold ST, Mehta S, Deuel RK. 1999. Serum autoantibodies to brain in Landau-Kleffner variant, autism, and other neurologic disorders. J Pediatr 134:607–613. CORREIA ET AL. Conroy J, Cochrane L, Anney RJ, Sutcliffe JS, Carthy P, Dunlop A, Mullarkey M, O’Hici B, Green AJ, Ennis S, Gill M, Gallagher L. 2008. Fine mapping and association studies in a candidate region for autism on chromosome 2q31-q32. Am J Med Genet Part B Neuropsychiatr Genet. Coutinho AM, Oliveira G, Morgadinho T, Fesel C, Macedo TR, Bento C, Marques C, Ataide A, Miguel T, Borges L, Vicente AM. 2004. Variants of the serotonin transporter gene (SLC6A4) significantly contribute to hyperserotonemia in autism. Mol Psychiatry 9:264–271. Faham M, Zheng J, Moorhead M, Fakhrai-Rad H, Namsaraev E, Wong K, Wang Z, Chow SG, Lee L, Suyenaga K, Reichert J, Boudreau A, Eberle J, Bruckner C, Jain M, Karlin-Neumann G, Jones HB, Willis TD, Buxbaum JD, Davis RW. 2005. Multiplexed variation scanning for 1,000 amplicons in hundreds of patients using mismatch repair detection (MRD) on tag arrays. Proc Natl Acad Sci USA 102:14717–14722. Folstein SE, Rosen-Sheidley B. 2001. Genetics of autism: Complex aetiology for a heterogeneous disorder. Nat Rev Genet 2:943–955. Gallagher L, Becker K, Kearney G, Dunlop A, Stallings R, Green A, Fitzgerald M, Gill M. 2003. Brief report: A case of autism associated with del(2)(q32.1q32.2) or (q32.2q32.3). J Autism Dev Disord 33: 105–108. Hailer NP, Jarhult JD, Nitsch R. 1996. Resting microglial cells in vitro: Analysis of morphology and adhesion molecule expression in organotypic hippocampal slice cultures. Glia 18:319–331. Iglesias AH, Camelo S, Hwang D, Villanueva R, Stephanopoulos G, Dangond F. 2004. Microarray detection of E2F pathway activation and other targets in multiple sclerosis peripheral blood mononuclear cells. J Neuroimmunol 150:163–177. IMGSAC. 2001. A genome wide screen for autism: Strong evidence for linkage to chromosomes 2q, 7q, and 16p. Am J Hum Genet 69:570–581. Lauritsen MB, Als TD, Dahl HA, Flint TJ, Wang AG, Vang M, Kruse TA, Ewald H, Mors O. 2006. A genome-wide search for alleles and haplotypes associated with autism and related pervasive developmental disorders on the Faroe Islands. Mol Psychiatry 11:37–46. Milner R, Campbell IL. 2002. The integrin family of cell adhesion molecules has multiple functions within the CNS. J Neurosci Res 69:286–291. Nojima Y, Humphries MJ, Mould AP, Komoriya A, Yamada KM, Schlossman SF, Morimoto C. 1990. VLA-4 mediates CD3-dependent CD4þ T cell activation via the CS1 alternatively spliced domain of fibronectin. J Exp Med 172:1185–1192. O’Doherty C, Roos IM, Antiguedad A, Aransay AM, Hillert J, Vandenbroeck K. 2007. ITGA4 polymorphisms and susceptibility to multiple sclerosis. J Neuroimmunol 189:151–157. Rabionet R, Jaworski JM, Ashley-Koch AE, Martin ER, Sutcliffe JS, Haines JL, Delong GR, Abramson RK, Wright HH, Cuccaro ML, Gilbert JR, Pericak-Vance MA. 2004. Analysis of the autism chromosome 2 linkage region: GAD1 and other candidate genes. Neurosci Lett 372:209–214. 1151 Ramoz N, Cai G, Reichert JG, Silverman JM, Buxbaum JD. 2008. An analysis of candidate autism loci on chromosome 2q24-q33: Evidence for association to the STK39 gene. Am J Med Genet Part B Neuropsychiatr Genet. Romano V, Cali F, Seidita G, Mirisola M, D’Anna RP, Gambino G, Schinocca P, Romano S, Ayala GF, Canziani F, De Leo G, Elia M. 2005. Suggestive evidence for association of D2S2188 marker (2q31.1) with autism in 143 Sicilian (Italian) TRIO families. Psychiatr Genet 15: 149–150. Salomon DR, Mojcik CF, Chang AC, Wadsworth S, Adams DH, Coligan JE, Shevach EM. 1994. Constitutive activation of integrin alpha 4 beta 1 defines a unique stage of human thymocyte development. J Exp Med 179:1573–1584. Sato T, Tachibana K, Nojima Y, D’Avirro N, Morimoto C. 1995. Role of the VLA-4 molecule in T cell costimulation. Identification of the tyrosine phosphorylation pattern induced by the ligation of VLA-4. J Immunol 155:2938–2947. Sham PC, Curtis D. 1995. An extended transmission/disequilibrium test (TDT) for multi-allele marker loci. Ann Hum Genet 59:323– 336. Shao Y, Wolpert CM, Raiford KL, Menold MM, Donnelly SL, Ravan SA, Bass MP, McClain C, von Wendt L, Vance JM, Abramson RH, Wright HH, Ashley-Koch A, Gilbert JR, DeLong RG, Cuccaro ML, PericakVance MA. 2002. Genomic screen and follow-up analysis for autistic disorder. Am J Med Genet 114:99–105. Silva SC, Correia C, Fesel C, Barreto M, Coutinho AM, Marques C, Miguel TS, Ataide A, Bento C, Borges L, Oliveira G, Vicente AM. 2004. Autoantibody repertoires to brain tissue in autism nuclear families. J Neuroimmunol 152:176–182. Singh VK, Warren RP, Odell JD, Warren WL, Cole P. 1993. Antibodies to myelin basic protein in children with autistic behavior. Brain Behav Immun 7:97–103. Singh VK, Lin SX, Yang VC. 1998. Serological association of measles virus and human herpesvirus-6 with brain autoantibodies in autism. Clin Immunol Immunopathol 89:105–108. Spence SJ, Cantor RM, Chung L, Kim S, Geschwind DH, Alarcon M. 2006. Stratification based on language-related endophenotypes in autism: Attempt to replicate reported linkage. Am J Med Genet Part B Neuropsychiatr Genet 141B:591–598. Vargas DL, Nascimbene C, Krishnan C, Zimmerman AW, Pardo CA. 2005. Neuroglial activation and neuroinflammation in the brain of patients with autism. Ann Neurol 57:67–81. Von Andrian UH, Engelhardt B. 2003. Alpha4 integrins as therapeutic targets in autoimmune disease. N Engl J Med 348:68–72. Williams DA, Rios M, Stephens C, Patel VP. 1991. Fibronectin and VLA-4 in haematopoietic stem cell-microenvironment interactions. Nature 352:438–441.