INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS CONCLUSION REFERENCES

Asthma, which is characterized by variable airflow obstruction and enhanced bronchial responsiveness, forms part of the spectrum of atopic diseases, involving activation of T cells to produce an array of cytokines that orchestrate the synthesis of immunoglobulin E (IgE) and the recruitment of mast cells, basophils, and eosinophils for mediator secretion. Both asthma and atopy have strong genetic components but do not follow simple monogenic patterns of inheritance (1, 2). The genetic analysis of such complex disorders may be confounded by polygenic inheritance (more than one disease gene occurring in the same individual), genetic heterogeneity (different combinations of disease genes occurring in different individuals), and incomplete or age-dependent penetrance. These factors result in a dissociation between genotype and phenotype characteristic of single gene disorders, such that a gene that affects a disease in one population may show little impact in another.

Studies attempting to define the genetic factors underlying asthma and atopy must address the problem of phenotype definition. Atopy is characterized by a persistent IgE-mediated response to common aeroallergens. Conventionally, atopy has been defined by three separate indices: (1) a raised serum total IgE; (2) positive skin-prick tests to common allergens; (3) the presence of allergen-specific IgE antibodies (RAST) in the serum. These can be used either singly or in combination. Asthma can be diagnosed on the basis of characteristic pathological changes found in the airway on bronchial biopsy, including a thickening of the lamina reticulosa of the basement membrane, raised eosinophil and mast cell counts, and thickening of the airway wall (3). However, asthma cannot be defined purely on the basis of such pathological changes, because some of these changes can be found individually in other diseases of the airways (for example, chronic obstructive pulmonary disease), and some individuals without symptoms of asthma can also have these changes in mild form. Altenatively, because asthma is difficult to define clinically, important functional components that reflect aspects of the diseasebronchial hyperresponsiveness (BHR) and IgE statuscan be measured more readily and used as surrogate markers in genetic studies.

The clinical symptoms associated with asthma, wheeze and periodic chest tightness, are usually accompanied by variable airflow obstruction or BHR. The presence of BHR is often indicative of asthma and is moderately correlated with the pathological changes observed in the airway (4). Although the pathophysiology of BHR is poorly understood, in population studies a close relationship between atopy, BHR, and asthma has been demonstrated (5).

Serum IgE levels are closely related to both diagnosed (7) and self-reported asthma (6) and to allergic rhinitis (6). However, while most asthmatics are also atopic (5), most atopics are not asthmatic, implying that additional factorsboth genetic and environmentalare necessary for the clinical expression of asthma.

While recognizing that the syndrome of asthma has a 50- 60% heritability, researchers have no consensus about the contributing genetic mechanisms or the underlying quantitative traits of IgE and BHR. Using segregation analysis, various modes of inheritance have been proposed (8). However, the models derived will be confounded by important environmental influences and there has been little agreement among different studies. It is also possible that some of the heritable components of asthma are independent from those that modulate both BHR and IgE. Thus, interactions between genes and environment involved in the pathogenesis of asthma and atopy are likely to be complex. The application of genetic mapping techniques to large, well-characterized population samples will help us to understand these interactions. Advances in molecular genetics, improved genetic maps, and more powerful methods of analysis have made the mapping of genes involved in complex diseases a practical proposition.

As part of genetic research into asthma, the Southampton group has investigated specific genes of chromosomal regions already implicated in the pathogenesis of asthma or atopy, as well as screening whole chromosomes for the identification of novel genes. We present here the methodologies used and our findings on chromosomes 5, 11, and 12.

	STUDY DESIGN

Population Samples

For any epidemiological study of complex disorders to have sufficient statistical power to detect the underlying genetic factors, it must be based on a large sample size. To exclude false positive results arising by chance (Type 1 errors), it is important to confirm any putative linkages in additional population samples. Therefore, two distinct populations totaling nearly 1,000 individuals have been recruited from the Wessex region of Southern England. The first sample comprised 131 nuclear families with three or more children (n = 685) drawn from General Practice registers in the Southampton area (9). All families were recruited without reference to asthma; these will be referred to as the random sample. The second cohort of families, the multiplex sample, comprised 60 extended families recruited through an asthmatic proband attending Child Health Clinics in Southampton, with at least one other family member affected by asthma. Blood was taken for DNA extraction and for measurement of total serum IgE and specific IgE from 324 members of the multiplex sample and from 666 members of the random sample.

The collection and analyses of clinical data with construction of eight variables spanning total IgE; the bronchial response to inhaled histamine; skin-prick tests to allergens; and a recent history of rhinitis, wheezing, and eczema have been described (9). Total serum IgE concentration was measured by an enzyme-linked immunosorbent assay and specific IgE status determined by skin-prick testing to 14 common allergens (Bayer Corporation, Spokane, Washington). The bronchial response to inhaled histamine acid phosphate was measured using the hand-held nebulizer procedure modified from Yan and colleagues (10).

Asthma and Atopy Scores

Asthma can be studied as the principal phenotype or, because of the difficulty in assigning a clinical definition, as surrogate markers (IgE or BHR). Alternatively, the separate variables derived from the clinical and laboratory data can be combined to generate quantitative asthma and atopy scores. An eigenvector plot (Figure 1) illustrates the separation of individual traits into clusters associated with asthma, atopy and wheeze. Principal component and logistic regression analyses of eight traits (IgE, BHR, skin prick, atopy, asthma, wheeze, video questionnaire, migraine) were used to define an asthma and atopy score. Using a physician's diagnosis to define the presence of asthma, the asthma score distributions differ significantly between asthmatics and nonasthmatics from the combined random and multiplex populations (Figure 2).

View larger version (12K): [in this window] [in a new window]   Figure 1.   Eigenvector plot showing principal component regression analysis of data from the combined multiplex and random population samples.

View larger version (21K): [in this window] [in a new window]   Figure 2.   Comparison of the asthma score distributions of asthmatics and nonasthmatics from the combined multiplex and random population samples.

Molecular Approaches to Identifying Disease Genes

For complex disorders, in which the biochemical and physiological basis of a genetic disease will be difficult to define, the genes involved can be identified by two strategies. First, in a candidate-gene approach, individual genes can be directly tested for their involvement in the disease process. If one of the candidate genes is found to be involved, the causative mutations can then be identified without the need for further fine mapping. Genes are selected for investigation because their function is likely to influence disease pathogenesis or from synteny to linked regions identified in animal models. This approach relies upon pre-existing knowledge of the disease process and on the availability of well-characterized candidate genes. Second, in a genome search all genes can be systematically screened using panels of microsatellite DNA markers uniformly distributed throughout the whole genome. The approximate location of a disease gene can be identified by demonstration of co-inheritance of a disease or trait with a particular marker. However, this region could span several cM, and the localization will have to be progressively refined until the gene itself can be isolated.

Both approaches use the same molecular methodshigh throughput polymerase chain reaction (PCR) and semi-automated fluorescence-based genotypingand so can be easily integrated. Research by the Southampton group has utilized both approaches, investigating specific genes or gene clusters (11, 12) as well as screening whole chromosomes with the dinucleotide repeat marker set developed by Dr. John Todd while working on insulin-dependent diabetes mellitus (13) and provided to us by the Medical Research Council.

Statistical Methods

There are several different types of approaches that can be used to map genes in complex disorders, including linkage analysis, association studies and allele-sharing methods (14). There is no universal consensus on the most appropriate method to use, although generally, association studies are best suited to the analysis of candidate genes and linkage analysis best suited to genome screens, particularly with the availability of multipoint analysis. We have used both these approaches as well as developing a nonparametric linkage analysis, i.e., one that is independent of any inheritance model. Nonparametric linkage and association analysis were performed with the NOPAR program (15), incorporating the  statistic (16). This program enables the calculation of a lod score for each locus and allele. Parametric analyses were also performed with COMDS program (9).

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS CONCLUSION REFERENCES

Chromosome 5

The 5q31-q33 region of the genome contains several candidate genes implicated in the pathogenesis of asthma and in the regulation of IgE, including the interleukin-4 (IL-4) cytokine cluster (IL-3, IL-4, IL-5, IL-9, IL-13, and the granulocyte/macrophage colony-stimulating factor) and the ₂-adrenergic receptor (ADRB2). This makes chromosome 5q a very attractive candidate region, and recently evidence has been reported for linkage to both total serum IgE (17, 18) and to BHR (19). A schematic map of the region is shown in Figure 3, with the approximate location of the candidate genes indicated.

View larger version (39K): [in this window] [in a new window]   Figure 3.   Schematic map of the region of chromosome 5q showing positive evidence for linkage. Markers genotyped by the Southampton group are indicated by arrows. IL = interleukin; IRF-1 = interferon regulatory factor 1; CDC25 = cell division cycle 25; CSF2 = colony-stimulating factor 2; EGR1 = early growth factor response 1; ADRB2 = 2 adrenoreceptor; FGF-A = acidic fibroblast growth factor; GRL1 = lymphocyte-specific glucocorticoid receptor 1; PDGFR = platelet-derived growth factor receptor; CSF1R = colony-stimulating factor 1 receptor; and IL12B =  chain of interleukin 12.

Evidence of linkage to total serum IgE, centered around the IL-4 locus, was first reported in 11 large Amish pedigrees (17). The linkage evidence was strongest when subjects with high specific IgE were excluded from the analysis, suggesting that specific IgE responsiveness is a confounding factor in the analysis of the genetics of total serum IgE. Using a similar phenotype, a second group also reported linkage to total IgE in 92 Dutch families (18) as well as to a phenotype based on BHR (19). The strongest evidence for both phenotypes in the Dutch population was more distal than the IL-4 locus, centered around the ADRB2 gene. This makes ADRB2 a promising candidate, particularly as coding variants identified within the gene have been associated with hyperresponsiveness (20) and total serum IgE (21). There is also evidence that these variants could have functional significance (22).

In our populations, we have genotyped polymorphic markers within the IL-4 and IL-9 genes as well as five anonymous markers (at loci with no known biological function) that span the 5q31-q33 region in the random and multiplex samples. These are indicated by arrows in Figure 3. Although positive linkage has been demonstrated to several of these markers in other studies (17, 18, 19), we were unable to identify significant linkage to any of the seven markers typed (23). We did, however, find strong allelic association with log total IgE at the IL-9 locus (p < 0.003) in the random sample (12). The association could not be repeated in the multiplex sample, but the positive linkage to other markers in the region and the biological plausibility of the candidate genes within the cytokine cluster suggest that this may not be a Type 1 error. For confirmation, the association will need to be tested in additional populations.

To date, none of the studies on chromosome 5q have had sufficient power to produce an accurate localization of the disease gene(s), and it is highly likely that more than one important locus exists in the region. Polymorphisms within the promoter regions of several candidate genes have now been identified and there is evidence for (24) and against (25) an association between a nucleotide substitution in the IL-4 promoter and total serum IgE.

Chromosome 11

Cookson and colleagues first reported genetic linkage of atopy to chromosome 11q13 in extended (26) and nuclear families (27, 28) with the "atopy gene" preferentially active in maternally-derived alleles (29). The gene encoding the beta subunit of the high-affinity IgE receptor (FcRI-) was proposed as the candidate gene (30). Determination of its coding sequence identified two amino acid substitutions within exon 6 designated Leu 181 and Leu 183, which were highly predictive of atopy when inherited maternally (31). An additional substitution, E237G, has now been identified in exon 7 (Figure 4), which is also associated with asthma but with no maternal effect (32). While some independent studies support a role for chromosome 11q in the etiology of atopy or BHR, most independent groups have been unable to replicate the original linkage results (33).

View larger version (40K): [in this window] [in a new window]   Figure 4.   Amino acid polymorphisms within the  chain of the tetrameric FcRI receptor. Transmembrane domains are boxed. Reported amino acid substitutions are colored black.

Genotyping of three anonymous markers in the region produced no evidence for linkage (11), although only one marker, D11S480, was close to the putative atopy locus. A schematic map of the region is shown in Figure 5, with the markers and polymorphisms studied in Southampton indicated by arrows. Two strong allelic associations were observed (12) both in the random sample, between D11S527 and BHR (p < 0.0003) and between D11S534 and log IgE (p = 0.007). The association between D11S527 and BHR could not be validated in the multiplex sample and multiple association tests, particularly with anonymous markers, can produce positive results by Type 1 errors.

View larger version (18K): [in this window] [in a new window]   Figure 5.   Schematic map of the 11q13 region with details of the FcRI- gene. Dinucleotide repeat markers and specific mutations investigated in the Southampton populations are indicated by arrows.

An intronic marker within the FcRI- gene gave weak evidence for linkage to log IgE (p = 0.05) in the multiplex sample. This result could not be repeated in the random sample, and analysis of the combined sample failed to reach significance (23). Using two large population samples ascertained by separate criteria and employing different analytic methodologies overcomes the limitations apparent in many of the other negative linkage reports.

The Leu 181 substitution has been reported by the Oxford group at a frequency of 15% in an unselected UK population (31). However, using direct cycle sequencing of PCR-amplified fragments encompassing exon 6 from all 120 parents of the multiplex families and from both parents of 90 random families, the wild-type nucleotide sequence was found in every case (23). This is entirely consistent with every other independent investigation, which among them, using a variety of different methodologies, have failed to identify a single Leu 181 or Leu 183 carrier (33) (Table 1). If these substitutions do exist, then it must be at a very low frequency and without any clinical relevance to asthma.

The substitution of glutamic acid for glycine at residue 237 (E237G) occurs at frequencies of 5.3% and 6% in unselected Australian (32) and Japanese (34) populations. In both cases the presence of E237G is strongly associated with asthma (p = 0.005), and in the Australian population it is also associated with BHR (p = 0.0009). The substitution occurs at a frequency of around 3.5% in both our random and multiplex UK populations (32). Analysis of these data showed no evidence of any association to either asthma or atopy. There was significant linkage to both the asthma score (p = 0.009 in both samples) and the atopy score (p = 0.0001 in the multiplex sample), but no significant linkage to log IgE (32). This suggests that the E237G substitution could be important, although it is unclear whether it is a causal mutation or simply in linkage disequilibrium with another mutation.

The tetrameric FcRI receptor complex is central to the process of IgE-dependent allergic inflammation (35), and the subunits (₂) are likely candidate genes for asthma susceptibility. The  chain of the complex contains an amino acid motif similar to the immunoreceptor tyrosine-based activator motif (ITAM), which is implicated in antigen-mediated cell activation. Although the  chain has no autonomous ability for signal generation, it amplifies by fivefold to sevenfold the signal strength mediated by the dimer (36). Alterations in the wild-type sequence of a gene can have biological and clinical significance, but clearly it is important to attribute functions to FcRI in the expression of elevated IgE levels and asthma if polymorphisms within the  chain are to be clinically relevant. Human mast cells and basophils express high levels of FcRI involved in allergen-specific signaling of these cells to release preformed and newly generated mediators involved in the immediate hypersensitivity response. The IgE signaling also releases preformed cytokines and initiates transcription for cytokines including IL-4, IL-5, and IL-13, whose genes are encoded in the IL-4 gene cluster on chromosome 5q (37, 38). Thus, cytokines derived from activated mast cells might be involved in the isotype switching of B cells for IgE synthesis involving ligation of CD40.

Human monocytes and dendritic cells also express FcRI. In asthma, the total number of dendritic cells and the proportion expressing the  chain of FcRI are increased in the bronchial mucosa (39) and in the skin of patients with eczema (40). The subunits of FcRI are under the control of cytokines and can be differentially expressed (40). In monocytes, engagement of IgE by allergen increases allergen uptake. Because the E237G polymorphism of FcRI- is adjacent to an ITAM site, it is able to interact with the two  chains and influence the efficiency of IgE cell signaling involving lyn and syc. Thus, it is possible that the E237G polymorphism could influence both mast cell cytokine secretion and dendritic cell function as a professional antigen-presenting cell.

Chromosome 12

While candidate gene approaches investigate genes of known function, genome screens, such as that recently completed in Oxford (41), can identify novel regions of the genome that could also be important in asthma pathogenesis. Linkage to chromosome 12 was identified by analysis of a chromosome 12- specific marker set (13), together with additional 12q markers taken from published Genethon sequences. These were genotyped on the multiplex sample, and single locus nonparamet-ric analysis revealed linkage evidence to several markers (42). The positive linkages are shown in Table 2.

A schematic representation of chromosome 12q is shown in Figure 6, with the markers genotyped by the Southampton group appearing inside shaded boxes. Linkage to five of the markers was localized to a distal region of 12q spanning around 35 cMbetween the markers D12S366 and D12S357with the strongest linkage evidence around D12S97. Linkage to this marker has also been replicated in the random sample to the atopy score with p = 0.006 (42). Analysis of these data also identified two allelic associations in the same region: D12S366 with atopy (p = 0.002) and D12S78 with log IgE (p = 0.001) (42).

View larger version (14K): [in this window] [in a new window]   Figure 6.   Schematic representation of chromosome 12q. Markers genotyped on the Southampton populations appear inside shaded boxes, including intronic markers within the IFN- and IRF-1 genes. The remaining markers are those used by Barnes and colleagues (44). Candidate genes are shown as boxes. The location of the markers and the approximate genetic distances between them in cM are based on a YAC contig map of chromosome 12 (45).

Guidelines for interpreting lod scores generated from the dissection of complex traits have recently been proposed (43). Although the linkage values for chromosome 12 would be classed as significant by this convention (p < 0.05), they fall short of being highly significant (p < 0.001). However, the linkage has been replicated in a second population, which is a strict criterion to be fulfilled for confirmed linkages (43). There are two further reasons to think that these linkages are genuine. Firstly, chromosome 12q contains several candidate genes including interferon- (IFN), a mast cell growth factor (MGF), and insulin-like growth factor (IGF-1), the constitutive form of the nitric oxide synthase gene (NOS1) and the  subunit of the nuclear factor Y involved in transcription of human leukocyte antigen genes (NFYB). Secondly, there is independent evidence for linkage to chromosome 12q (44), from a genome screen in two populations: 29 extended Afro-Caribbean families recruited through an asthmatic proband and 12 Caucasian Amish families selected on the basis of a detectable IgE antibody to a common inhalant allergen in at least one child, as previously described (17). This genome screen identified many new putative linked regions, but only chromosome 12q was positive in more than one of the ethnic groups studied. Positive linkage and association values spanned most of 12q, but the strongest association (p = 0.006 to total IgE) and the maximum linkage value from a multipoint analysis were to the marker D12S379. This marker is a considerable distance from the linked region predicted from the Southampton data, and considerable work remains to characterize this region further.

	CONCLUSIONS

TOP ABSTRACT INTRODUCTION RESULTS CONCLUSION REFERENCES

Considerable success has been achieved using linkage analysis to map single-gene disorders onto the human genome. The application of these mapping techniques complex multifactorial disorders has not yet yielded the same level of success, and we are still a long way from understanding the genetic factors underlying asthma and atopy and how these interact with the environment. Genome screens are now identifying novel regions of interest, such as chromosome 12q, but from studies completed so far it is clear that claims for linkage or association must be repeated in separate population samples with adequate power to be validated. Emphasis also needs to be focused on a clearer definition of asthma and associated surrogate markers. Finally, the application of simple analytical statistical models to complex disorders such as asthma is fraught with difficulty. The development of new nonparametric multipoint linkage tests, which do not rely on a knowledge of the pattern of inheritance of a specific disorder and enable a more precise location of the gene in relation to others that may influence the expression of the disease, herald a new era of greater precision in the genetics of asthma and atopy.