|Home | About | Journals | Submit | Contact Us | Français|
Several studies have documented a substantial genetic component in the aetiology of allergic diseases and a number of atopy susceptibility loci have been suggested. One of these loci is 3q21, at which linkage to multiple atopy phenotypes has been reported. This region harbours the CD86 gene encoding the costimulatory B7.2 protein. The costimulatory system, consisting of receptor proteins, cytokines and associated factors, activates T cells and regulates the immune response upon allergen challenge.
We sequenced the CD86 gene in patients with atopy from 10 families that showed evidence of linkage to 3q21. Identified polymorphisms were analysed in a subsequent family‐based association study of two independent Danish samples, respectively comprising 135 and 100 trios of children with atopy and their parents. Functional analysis of the costimulatory effect on cytokine production was performed in an autologous cell‐based system based on cells expressing CD86 variants.
Two polymorphisms were identified, encoding the amino acid changes Ile179Val and Ala304Thr, respectively. Significant associations were observed between the Ile179Val polymorphism and allergy phenotypes in both samples (eg, asthma, p=4×10−3 in the two samples combined). The undertransmitted (protective) Val179 allele was found to induce higher production of both Th1 and Th2 cytokines than the overtransmitted (risk) Ile179 allele, suggesting a functional impact of the polymorphism.
The CD86 gene, and specifically the Ile179Val polymorphism, may be a novel aetiological factor in the development of asthma and related allergic disorders.
Allergic diseases such as asthma, rhinitis and atopic dermatitis affect approximately one‐fifth of the general population in the Western world and their frequencies seem to be increasing. Asthma alone affects 155 million people in the world and hay fever is even more common.1 Atopic dermatitis at present affects 21% of Danish children.2 Because of their high prevalence and the associated personal morbidity and social costs, the allergic diseases are a major global health problem.
Family, twin and adoption studies have documented a substantial genetic component in the aetiology of allergic diseases.3,4,5 Several genome scans of large numbers of families with atopy have been carried out in the past decade, suggesting a number of atopy loci. Relatively few of the identified loci meet the criteria for genomewide significance. One that does is 3q21, at which significant linkage to atopic dermatitis was reported.6 We and others have reported evidence of linkage to the same region for multiple atopy phenotypes, including asthma, rhinitis and atopy.7,8,9,10 This region harbours the CD86 gene encoding one of two costimulatory B7 proteins required for T cell activation. Antigen‐presenting cells (APCs) express CD86 (B7.2), constituting the ligand for CD28 and cytotoxic T‐lymphocyte‐associated protein 4 (CTLA4) at the T cell surface.11 Binding of CD86 to CD28 initiates a costimulatory signal for T cell activation, proliferation, differentiation and production of a number of cytokines.11 In contrast, binding to CTLA4 negatively regulates T cell proliferation and diminishes the immune response.12,13,14 CTLA4 has higher affinity for CD86 than CD28 molecules and is predominantly expressed upon activation of T cells. T cell responses are categorised as T helper (Th)1 and Th2 according to the cytokines they produce. Th1 cells produce interferon (IFN)‐γ, tumour necrosis factor (TNF)α and interleukins (IL)‐2, IL‐12 and IL‐18, whereas Th2 cells typically produce IL‐4, IL‐5, IL‐9, IL‐10 and IL‐13. It is generally accepted that a Th1/Th2 imbalance is implicated in the development of asthma and atopy,15,16,17,18,19 and a role for the costimulatory B7 molecules in the regulation of the Th1/Th2 cytokine balance has been suggested by several investigators.20,21,22,23,24
The CD86 gene is composed of eight exons and spans >22 kb. Jellis et al found that alternatively spliced cDNAs result from the use of either exon 1 or 2 giving rise to CD86 isoform 1 (GenBank accession number NM_175862) and 2 (GenBank accession number NM_006889), respectively.25 The N‐terminus of isoform 2 is six amino acids shorter than isoform 1. Exon 2 does not contain coding sequences. Exon 3 corresponds to the signal peptide, exon 4 to an IgV‐like domain, exon 5 to an IgC‐like domain and exon 6 to the transmembrane region and part of the cytoplasmic tail. Exons 7 and 8 encode the remainder of the tail. The two CD86 cDNAs with alternate 5′ untranslated sequences may represent constitutively expressed and inducible forms of CD86.25
This study investigated the possible role of CD86 in susceptibility to asthma and related allergic disorders by family‐based association analysis of two independent Danish samples. Furthermore, a substitution polymorphism (Ile179Val) showing association in both samples was subjected to functional analysis in an autologous cell‐based system.
Informed consent was obtained from each participant prior to inclusion in the study. Local ethics committee approval was obtained in all regions where families were recruited.
In total, 235 families (964 individuals) were recruited from two studies on atopy. In the first study (sample A), 135 nuclear families with asthmatic sibpairs were enrolled primarily from patient records in four clinical centres in the eastern, western and central areas of Denmark.26,27 Inclusion criteria were recurrent cough, wheezing and dyspnoea and a positive metacholine challenge using the methods described by Yan et al.28 Both affected siblings were aged 15–45 years. Skin‐prick tests using 10 common aeroallergens and lung function measurements including FEV1 (forced expiratory volume in 1 second) and FVC (forced vital capacity) were performed on all participants. Blood samples were analysed for allergen‐specific IgE (CAP Phadiatop; Pharmacia Upjohn, Copenhagen, Denmark); a positive test was defined as IgE 0.35 kUA/l for one or more of the allergens. Hay fever and eczema was self‐reported in the questionnaires.
The second study (sample B) included a total of 424 individuals from 100 Danish atopic sibpair families enrolled in the Danish Inheritance of Type I Allergy (ITA) project. Recruitment, examination and clinical information have been described previously.7,8 In brief, patients were recruited through four paediatric and one adult outpatient allergy clinic in Western Denmark. Inclusion criteria were a minimum of two atopic siblings with doctor‐diagnosed asthma, hayfever or atopic dermatitis and reported effect of appropriate medication. Mean age among the offspring was 10.8 years. All subjects were clinically examined and questionnaire tested by the same doctor. Atopic disease was diagnosed by this doctor according to standard criteria in consensus with a second doctor evaluating the questionnaires only. Blood was drawn for serum measurement of specific IgE to 11 common allergens (CAP RAST FEIA test; Pharmacia Upjohn); a positive test was defined as above.
For each phenotype, only one affected offspring from each family (in both samples A and B) was selected for genotype analysis. Where there were 2 affected offspring in a family, the oldest child was selected for the analysis. The family structures and phenotypes of the two family samples are shown in table 11.
From all family members in both sample A and B, genomic DNA was isolated from blood samples according to standard procedures. Exon 1–8 of CD86, all intron/exon boundaries of the two isoforms and approximately 1 kb of the promoter region (of isoform 1) were screened for sequence variations in 10 unrelated patients from families (from sample B) showing evidence of linkage to 3q21.8 Direct sequencing was performed using a commercial kit (Big Dye Terminator Cycle Sequencing Kit; Applied Biosystems, Foster City, CA, USA) and an automated analyser (ABI‐310 Genetic Analyzer; Applied Biosystems). Genotyping of parents and offspring was performed by single base extension technology (SNaP‐shot Kit; Applied Biosystems). Standard PCR conditions were used for both sequencing and genotyping. Primer sequences are available upon request.
Pairwise linkage disequilibrium was calculated using Simwalk229 and the GOLD software package (http://www.sph.umich.edu/csg/abecasis/GOLD).30 For association analysis, the transmission disequilibrium test (TDT) was used.31 As only one affected child per family was analysed, the method is a valid test for association even in the presence of linkage, which is the case for both sample A10 and sample B.7,8 The allergy phenotypes investigated are strongly correlated (table 11)) and the two SNPs tested are in (moderate) linkage disequilibrium with each other. Consequently, the presented p‐values were not corrected for multiple testing. However, if analysis of the two SNPs were considered independent, most of the associations reported remain significant after Bonferroni correction.
After informed consent was obtained, a cutaneous biopsy specimen was taken from a patient with metastatic malignant melanoma. From this biopsy, T cell and melanoma cell lines were established.32,33 Both cell lines were maintained in Dulbecco's modified Eagle's medium (Life Technologies Inc., Gaithersburg, Maryland, USA), with 10% fetal calf serum (Life Technologies Inc.), 100 U/ml penicillin and 0.1 mg/ml streptomycin (both Leo Pharma, Ballerup, Denmark), in 5% CO2. In addition, 100 pM IL‐2 and 500 pM IL‐4 were added to the T cell medium.
Total RNA from human T cells expressing the most frequent CD86 allele, Ile179, was purified (Aurum total RNA kit; Bio‐Rad, Hercules, California, USA) and cDNA was synthesised using Omniscript RT‐kit (Qiagen, Valencia, California, USA) according to the manufacturers' instructions. Jellis et al have previously suggested that translation of both cDNAs (isoforms 1 and 2) begins at the methionine codon in exon 3.25 Based on this, full‐length, isoform 2 human CD86 cDNA was amplified using a forward primer located in the exon 1–3 boundary and a reverse primer in the 3′ end of exon 8 and cloned into pcDNA3.1 (Invitrogen Corp., Carlsbad, California, USA). The CD86 Ile179Val mutation was introduced into the CD86‐Ile179 clone using site‐directed mutagenesis (QuickChange; Stratagen. La Jolla, California, USA). To achieve stable expression in melanoma cells, CD86 variants were inserted into the pBabePuro retroviral vector.34 All vectors were sequenced.
pBabePuro vectors containing the CD86 variants were packaged in packaged PG13 cells (American Type Culture Collection), and supernatants were used to transduce human melanoma cells using methods described previously.35 Transduced melanoma cells were selected with 5 μg/μl puromycine (Sigma Chemical Co., Minneapolis, Minnesota, USA). Expression of Ile179 or Val179 CD86 receptor proteins was confirmed by sequencing of reverse‐transcribed cDNA from the transduced melanoma cells (data not shown).
For flow cytometry analysis, transduced melanoma cells were harvested by incubation in phosphate‐buffered saline (PBS) containing 0.6 mmol/l EDTA followed by incubation in PBS containing 0.6 mmol/l EDTA and 0.005% trypsin (v/v) (Life Technologies Inc). Approximately 2×105 cells were stained with either 10 μl primary anti‐CD86 monoclonal antibody (Becton Dickinso, Franklin Lakes, NJ, USA) or 10 μl R‐phycoerythrin (R‐PE)‐conjugated anti‐HLA‐class I (Dako, Glostrup, Denmark) for 30 min at 4°C in 50 μl PBS containing 5% human serum (obtained from the local blood bank). For visualisation of CD86, 1 μl secondary fluorescein isothiocyanate (FITC)‐conjugated Alexa Fluor goat‐anti‐mouse antibody (Molecular Probes, Breda, The Netherlands) in 50 μl PBS containing 5% human serum was added. Following incubation, the cells were washed in PBS and resuspended in PBS containing 5% human serum. Approximately 2×104 live cells were analysed on a FACS Calibur flow cytometer with CELLQuest V.3.3 software (Becton Dickinson). Dead cells were excluded by propidium iodide staining.
Cellular localisation and confocal microscopy was carried out as described.36 The melanoma cells were fixed in neutral buffered formalin (Mallinckrodt Baker, Deventer, The Netherlands) for 15 min at room temperature, labelled with anti‐CD86 monoclonal antibody (Becton Dickinson) and visualised using FITC‐conjugated Alexa Fluor goat‐anti‐mouse antibody (Molecular Probes). Nuclear DNA was counterstained with Hoechst 33258 (1 μg/ml). Labelled cells were analysed by confocal microscopy analysis (Leica TCS microscope; Leica, Heerbrugg, Switzerland).
The established T cell line was evaluated by flow cytometry for the expression of the markers T cell receptor (TCR), CD28, CD4 and CD8. In parallel experiments, approximately 2×104 T cells were stained with either 20 μl FITC‐conjugated mouse anti‐human TCR‐α/β‐1 (IgG1, κ) antibodies (Becton Dickinson), 10 μl FITC‐conjugated mouse anti‐human CD4 (IgG1, κ) antibodies (Dako), 20 μl FITC‐conjugated mouse anti‐human CD8 (IgG1, κ) antibodies (Becton Dickinson) or 10 μl primary anti‐rat CD28 antibodies (Serotech, Raleigh, North Carolina, USA), followed by 1 μl secondary FITC‐conjugated Alexa Fluor goat‐anti‐mouse antibody (Molecular Probes) in 50 μl PBS containing 5% human serum. Anti‐human CD16 (IgG1, κ) antibodies (Becton Dickinson) were used as isotype control for TCR, CD4 and CD8 staining.
Approximately 0.4×106 washed T cells were incubated with 0.2×106, 0.1×106 or 0.05×106 autologous, transduced melanoma cells. The cells were incubated for 2, 4 or 24 hours at 37°C. Following incubation medium was collected from the respective wells and centrifuged for 15 seconds at 10000 g. The supernatant (free of cytokine‐containing T cells) were analysed for various cytokines (Becton Dickinson human Th1/Th2 CBA kit and a Beckman Coulter FC500 flow cytometer) according to the manufacturer's instructions. Samples were then immediately analysed on the cytophotometer.
To identify potential susceptibility variants, we screened CD86 (figure 11)) for sequence variation in 10 unrelated atopic patients from families that showed evidence of linkage to 3q21 in a previous linkage study.8 Two SNPs were identified, encoding the amino acid changes Ile179Val in exon 5 (rs2681417) and Ala304Thr (rs1129055) in exon 8 (numbering according to CD86 isoform 2), respectively. No other sequence variations were found in the ~3.8 kb region surveyed.
In a subsequent family‐based association study of two independent Danish samples, comprising 135 (sample A) and 100 (sample B) families with members with atopy (table 11),), significant associations were observed between the Ile179Val polymorphism and most of the allergy phenotypes investigated (table 22).). The most frequent allele encoding Ile179 was overtransmitted to the affected offspring in both samples (whereas the Val179 allele was conversely undertransmitted). The results were most significant in sample A, showing association in all phenotypes except rhinitis. In sample B, there was a general trend towards Ile179 overtransmission, but significant association was observed only for asthma. In the two samples combined, the association was strongest for asthma (p=4×10−3) but also strong for atopy (positive RAST) and in patients with both positive RAST and asthma or atopic dermatitis. The Ala304Thr polymorphism showed no evidence of association, and haplotype analysis of the two SNPs merely reflected the association of Ile179Val separately (results not shown).
The two SNPs were found to be in Hardy‐Weinberg equilibrium in both samples and showed moderate pairwise linkage disequilibrium (LD) with a D′ of 0.70 and δ2 of 0.02 (p<0.0001).
To examine whether a variation in CD86 has any impact on the stimulation of T cells, it is necessary to avoid the costimulatory signal being masked by an alloresponse. Because CD86 present on melanoma cells can elicit effective costimulation of T cells,37,38,39 genetically modified melanoma cells can be used to investigate the influence of CD86 polymorphisms on T cell activation. We therefore used autologous T cells and CD86 expressing melanoma cells cultured from the same patient to determine whether the Ile179Val variation in CD86 may influence T cell activation.
Human CD86 Ile179 and Val179 were expressed in melanoma cells after retroviral transduction as described in Methods. Analysis of the expression levels of the two CD86 variants by means of flow cytometry using a CD86 antibody showed almost identical amounts of CD86 protein on the surface (figure 2A2A).). To further investigate the subcellular localisation of the two CD86 variants, labelling with anti‐CD86 followed by confocal microscropy of transduced melanoma cells was performed (figure 2C,D2C,D).). In all cases, both variants displayed a typical surface‐staining pattern and no other subcellular structures were labelled. No surface staining was observed in untransduced cells (data not shown). These data corroborate our findings that transduced CD86 Ile179 and Val179 are transported to the cell surface. To characterise the melanoma cells further, the amount of human leucocyte antigen (HLA) classes I and II was investigated by flow cytometry. We found comparable amounts of HLA class I in melanoma cells expressing CD86 Ile179 and Val179, and no expression of HLA class II.
The autologous T cells isolated from the same biopsy specimen as the melanoma cells were cultured as described above. A functional T cell normally expresses TCR, CD28, CD4 and/or CD8 proteins on the cell surface, which are known to play an important role in the activation and development of the T cell response.11 To investigate whether these proteins were present at the surface of the isolated T cells, immunostaining using antibodies against TCR, CD28, CD4 and CD8 was performed and aliquots of stained T cells were analysed by flow cytometry. As shown in figure 33,, the T cells were found to collectively express TCR, CD28, CD4 and CD8 on the surface.
To examine the positional effect of the Ile179Val substitution on costimulation, washed T cells (free of IL‐2 and IL‐4) were incubated with melanoma cells expressing either CD86 Ile179 or CD86 Val179. Flow cytometry analysis of the cell‐culture media after 2, 4 and 24 hours of incubation revealed that melanoma cells expressing the Val179 variant induced a higher production of both Th1 cytokines (IL‐2, TNFα, IFNγ) and Th2 cytokines (IL‐4, IL‐5) than melanoma cells expressing the Ile179 variant (figure 44,, A–C and D–E, respectively). No difference between the two variants was observed for the production of IL‐10 (predominantly an anti‐inflammatory cytokine, typically produced by either Th2 or T regulatory (Tr) cells). As the melanoma cells do not express HLA‐class II, the cytokine production in the autologous system may thus derive from CD8+ T cells.40 However, the possibility that CD4+ T cells also contribute to the cytokine production cannot be excluded, as it is known that IFNγ produced early during the antigen response can induce HLA‐class II expression on melanoma cells.41
The presented data show that the T cells of the autologous system are capable of producing a wide range of cytokines upon stimulation by CD86‐expressing melanoma cells and indicate that the two CD86 variants differentially affect the cytokine production.
Using family‐based association analysis of two independent Danish samples and functional analysis in an autologous cell‐based system, we obtained results implicating the Ile179Val substitution polymorphism, located in the costimulatory protein CD86, with the aetiology of allergic disorders.
In this study, the allele encoding Ile179 was overtransmitted to the affected offspring, suggesting that this variant confers increased susceptibility to asthma and atopy, whereas Val179, which was undertransmitted, conversely may act as a protective allele. The association was observed in two independent Danish samples, most prominently in sample A, whereas significant replication was observed only for the asthma phenotype in sample B. The less significant association in sample B may partly be explained by demographic and clinical differences between the two samples: all the individuals in sample A were >15 years old, whereas in sample B the probands were mainly children. Other inclusion criteria and diagnostics were also not completely identical. Thus, even though the diagnoses were the same, the underlying risk alleles might not be identical or of equal importance in the two samples. Even though there was a general trend toward Ile179 overtransmission in all phenotypes in sample B, and the analysis of the combined samples showed association for all the phenotypes investigated, the results regarding the “non‐asthma” phenotypes in particular should be replicated in additional populations.
The data provided by the HapMap project (http://www.hapmap.org) show that most of the CD86 gene (from the proximal promoter region to exon 3) resides within a haplotype block of around 37 kb, and the remainder of the gene is located in smaller distinct blocks. Specifically, the Ile179Val polymorphism is located in a small block encompassing exon 5 and 6, and only very limited LD seems to extend from this block to the surrounding regions. This implies that it is not likely that the association reported is due to LD to an SNP residing outside this region. This is in accordance with the observation that no evidence of association was seen for the Ala304Thr in exon 8. However, it is possible that other CD86 susceptibility variants, such as rare coding variants or intronic SNPs, might exist, which were not targeted in our screening procedure. Moreover, the region showing positive linkage is broad, and the maximum peaks vary between studies and between the specific atopy phenotypes investigated,7,8,9,10 indicating that there might be two or more susceptibility genes located in this region. In particular, the maximum rhinitis peak recently reported by our group is located closer to the centromere at 3q13.31,10 and rhinitis was the phenotype showing the least evidence of association in the present study, again suggesting that there might be another susceptibility gene in a more centromeric location in the region.
The three‐dimensional structure of the CD86 protein has not been resolved. To investigate the structural consequences of the Ile179Val substitution, we therefore turned to the known structure of the homologous CD80.42 However, it is difficult to predict the effect of Ile179Val on the structure/function relationships, as comparison of the amino acid sequences of CD86 and CD80 revealed low homology in the region harbouring the Ile179Val codon. On the other hand, as CD86 Ile179 and Val179 proteins were expressed in equal amounts on the cell surface (figure 22),), we hypothesise that the Ile179Val SNP may influence the interaction with CD28 or CTLA‐4 by affecting the conformation or assembly of CD86 rather than protein synthesis, trafficking or membrane embedding.
It is known that Th1 cells secrete IFNγ, TNFα, IL‐2, IL‐12 and IL‐18. These cytokines help to clear invading bacteria and viruses by activating macrophages and cytotoxic T cells, and Th1 cells may dampen allergic inflammation.11 In contrast, Th2 cells produce IL‐4, IL‐5, IL‐9 and IL‐13, which promote allergic inflammation and stimulate B cells to generate antibodies.43 The two subgroups of T cells develop in response to various stimuli, including cytokines. Interestingly, a Th1/Th2 counter‐regulation between these groups has recently been described, with each cell population able to influence the development induced by the other.19 Although the possible role of costimulatory B7 molecules in the regulation of Th1/Th2 development is not clear, accumulating evidence suggest a role of CD86 in regulation of the Th1/Th2 cytokine balance in vivo.20,21,22,23,24,44 Furthermore, it has been found that the Th1/Th2 regulation seems to depend on the nature of the APC, emphasising the role of CD86 for Th2 priming by both macrophages and dendritic cells.45 The second finding of this study is that the production of both Th1 and Th2 cytokines triggered by melanoma cells expressing the overtransmitted CD86 Ile179 variant was lower than that of melanoma cells expressing the undertransmitted Val179 variant (figure 44).). Having established that the clear difference in the evoked cytokine production between the two variants is mediated by equal expression of CD86 Ile179 and Val179, we suggest that the Ile179Val polymorphism has a functional effect on cytokine production. However, from this study we cannot predict whether the alleles predominantly lead to a Th1 or Th2 polarised response in vivo. As the Th1/Th2 balance is a pivotal factor in the development and aetiology of asthma and atopy, we hypothesise that the increased disease risk associated with the Ile179 variant and the protective effect of the Val179 variant might be mediated through a differential effect on the Th1/Th2 balance.
In conclusion, we suggest that the CD86 gene and specifically the Ile179Val polymorphism may influence susceptibility to asthma and related allergic disorders.
The work was supported by grants from the Hørslev Foundation, The Lundbeck Foundation, The Danish Lung Association, Aarhus University Research Foundation, The Augustinus Foundation and The Danish Medical Research Council. We thank Tina F. Hindkjær and Hanne Jacobsen for excellent laboratory work. Professor Poul Nissen is thanked for fruitful scientific discussions regarding the structural considerations.
APC - antigen‐presenting cell
CTLA4 - cytotoxic T‐lymphocyte‐associated protein 4
FEV1 - forced expiratory volume in 1 second
FITC - fluorescein isothiocyanate
HLA - human leucocyte antigen
IFN - interferon
IL - interleukin
LD - linkage disequilibrium
PBS - phosphate‐buffered saline
RAST - radioallergosorbent test
TCR - T cell receptor
TDT - transmission disequilibrium test
Th - T helper
TNF - tumour necrosis factor
UTR - untranslated region
Competing interests: None declared.