Identification of antigenic epitopes on human allergens: studies with HLA transgenic mice.

Environmental factors play an important role in the rise and manifestation of allergic conditions in genetically predisposed subjects. Increased exposure to indoor/outdoor allergens is a significant factor in the development of allergic sensitization and asthma. Recently, strong relationships between the immune response to several highly purified allergens and specific human leukocyte antigen (HLA)-DQ and -DR haplotypes have been reported. The major antigens from clinically important allergens have been cloned and sequenced. However, whether innate structural features of major allergens or peculiar immune recognition of these molecules contribute to the overly robust immune responses is not known. We generated and used transgenic (tg) mice expressing single HLA class II transgene(s) to characterize the allergen epitopes presented by particular HLA class II molecules. Next, we generated in vivo models for asthma in the HLA tg mice by intranasal challenge with allergenic extracts. Furthermore, we used a single epitope to induce an allergic lung inflammation. Our system offers a sophisticated technique for systematically identifying the genetic (individual human class II) and antigenic (individual allergenic epitopes) basis of asthma sensitivity and has important implications for new treatment strategies.

Allergy is the most widespread immunologic disorder in humans. It affects one in four individuals (HayGlass 1995). Today, 50 million Americans suffer from this disease and 15 million of them have asthma (Wasserman 1999). Allergy is defined as a T helper 2 (Th2)-driven hypersensitivity to innocuous antigens (allergens) of complex genetic and environmental origin and is acknowledged as a major risk factor for asthma (Holt et al. 1999). Allergens are derived from different sources. Both indoor allergens (house dust mite, cockroach, fungi, mold, animal-derived substances) and selected outdoor allergens (tree, weed, and grass pollens, fungal spores) have been implicated in causing asthma (Burge and Rogers 2000;Moore et al. 2001;Platts-Mills et al. 2000).
Despite much progress in understanding allergies-the dysfunction of the immune system (Campbell and Weiss 1999)-the molecular and genetic predisposition to allergy has not been identified. Identification of these factors is critical in managing atopic individuals (Cookson 1999). Recently, genome-wide searches and candidate gene approaches have been used to examine the possible involvement of several genes in the development of atopy and asthma. The regions of potential linkage to one or more asthma phenotypes were identified on chromosomes 5q, 6p, 11q, 12q, 13q, and 14q (Blumenthal 2000;Cookson 1999;Ono 2000).
Human leukocyte antigen (HLA) molecules are encoded by highly polymorphic gene families located on chromosome 6p. These molecules play a critical role in the initiation of immune responses to allergens because they bind allergen-derived peptides and present them to T lymphocytes, resulting in T cell activation and proliferation ( Figure 1). CD4 + Th2 cells are at the cellular epicenter of allergic disease. They produce an array of cytokines that directly or indirectly cause an acute and chronic allergic reaction in the airways ( Figure 1). However, CD4 + T-cell response to particular allergen/antigen critically depends on the genotype of HLA II, on the difference in their physicochemical characteristics of the molecular structure, and on binding and presentation of distinct peptides to CD4 + T cells (Klein and Sato 2000;Little and Parham 1999).
The association of HLA haplotypes and short ragweed (SRW) allergy was the first human immune response (Ir) gene to be recognized (Levine et al. 1972), and HLA-DR2 and HLA-DQ6 restriction of immunoglobulin (Ig)E reactions to antigen 5 is well documented (Huang et al. 1991;van Neerven et al. 1996). However, SRW pollens contain 52 different antigens, suggesting that other HLA II molecules can be involved in their recognition (Marsh et al. 1987;Stewart et al. 1996;van Neerven et al. 1996). Furthermore, knowledge of the sets of determinants presented by each HLA molecule is not known, nor is the hierarchy of importance among HLA class II molecules in ragweed allergen presentation. In addition, specific HLA-DR and DQ molecules have been implicated in asthma (Gerbase-DeLima et al. 1997;Hizawa et al. 1998;Malo and Chan-Yeung 2001;Soriano et al. 1997). However, their role and contribution in the development of allergic asthma are difficult to assess in humans because of their genetic heterogeneity. Do particular HLA-DR or HLA-DQ alleles preferentially present peptides that elicit Th2type cytokine production and promote allergic asthmatic conditions? Are there Th1restricted determinants? Do some HLA class II molecules protect against allergic reactions? If so, how is protection from atopy maintained? To address these questions, it is necessary to analyze systematically the variables of the allergen, individual epitopes, and the presenting HLA molecule to a population of T cells. Such controlled in vivo studies are not possible in the human system. HLA class II transgenic (tg) mice provide an excellent model for studying the genetic and molecular basis of allergic response.

Production and Characterization of HLA Transgenic Mice
We have introduced HLA-DQ8 (HLA-DQB1*0302, HLADQA1*0301) and HLA-DQ6 (HLA-DQB1*0601, HLA-DQA1*0301) genes into H-2Aβ 0 mice (Bradley et al. 1997;Nabozny et al. 1996). H-2Aβ 0 mice have disrupted Aβ gene and therefore could not express H-2Aβ molecule. Because these mice were of the H2 b haplotype and lacked the functional H2-E molecule, no mouse class II molecules were expressed on the cell surface. HLA tg mice express intact HLA molecules on cells of the lymphoid system and respond to several protein, peptide, and parasitic antigens Geluk et al. 1998;Krco et al. 2000;Kudva et al. 2001;Pimtanothai et al. 2000;Raju et al. 2001). HLA-DQ molecules in H-2Aβ 0 mice are expressed on the lymph nodes, spleens, thymus, lungs, and kidneys of transgenic mice but not in liver, pancreas, or brain. Approximately 25-40% of peripheral blood lymphocytes express HLA-DQ molecules.
Using a panel of monoclonal antibodies specific for Aα (clone 7-16-17), Aβ (clone 25-5-16), and Eα/β (clone Y-17) subunits, we detected no cell surface expression of hybrid class II. Thus, the HLA-DQ8 and HLA-DQ6 mice express intact human class II molecules in the absence of endogenous mouse or hybrid molecules.
The specific staining of mouse lungs for HLA-DQ expression was performed and sections were examined using light microscopy. This study demonstrated that the expression of HLA class II molecules in normal lung tissue in tg mice is similar to that in humans ). In addition to the professional APC in lungs, lung alveolar and bronchial epithelial cells and vascular endothelial cells expressed HLA-DQ molecules.
We have generated tg mice expressing HLA-DR2 (DRB1*1 502), DR3 (DRB1*0301), and DR4 (DRB1*0401) (Geluk et al. 1998;Pan et al. 1998Pan et al. , 1999. The DR molecules are also expressed on the cell surface, interact with CD4 + T cells, and positively/negatively select various T-cell receptor-bearing subsets and present antigens. The DR4 mice can present type II collagen and derived peptides and develop mild arthritis (Pan et al. 1998(Pan et al. , 1999. Thus, the DR molecules in tg mice are biologically functional.

Identification of HLA Class II Restricted Epitopes of Allergens in Transgenic Mice
Recent cloning and sequencing of cDNAs that encode major allergens (Gregoire and Chapman 2001;Ledford 1994;Stewart and Thompson 1996) has allowed T-cell epitope mapping in humans using long-term T-cell clones and lines (Bohle et al, 2000;de Silva et al. 2000;Dhillon et al. 1992;Katsuki et al. 1996). However, in vitro cloning procedure could bias toward certain T-cell response (van Neerven, 1996) rendering polyclonal T cells as better representatives of epitope diversity. Furthermore, although studies with T-cell clones from atopic patients might suggest potential epitopes, they do not provide a comprehensive picture of important epitopes for all class II molecules. Conventional mouse models do not fully reflect immune response to allergen seen in humans because major histocompatibility complex (MHC) II molecules are mouse derived. HLA II tg mice provide a powerful model for an exhaustive characterization and identification of the antigenic determinants on allergens.
In the past years, HLA tg mice have been used to identify epitopes on several autoimmune disease-related antigens such as collagen and cartilage glycoprotein, the autoantigens associated with arthritis (Cope et al. 1999;Krco et al. 1999), glutamic acid decarboxylase and insulin associated with type 1 diabetes Kudva et al. 2001;Raju et al. 1997), nicotinic acetylcholine receptor involved in myasthenia gravis (Raju et al. 2001), RO60 (SS-A) lupus-related antigen (Paisansinsup et al. 2000), myelin basic protein, myelin olygodendrocyte glycoprotein and proteolipid protein implicated in the development of multiple sclerosis (Das et al. 2000;Ito et al. 1996;Madsen et al. 1999;Khare M, David CS. Personal communication). In addition, the use of HLA II tg mice has proven to be a valuable tool for identification of T-cell epitopes on malarial parasites (Pimtanothai et al. 2000) and mycobacterial antigens (Geluk et al. 1998) that could be used to generate a new type of more effective subunit vaccines. Moreover, recent studies of HLA II tg mouse immune responses to tumor-associated proteins allow us to identify epitopes of potential clinical value for design of anticancer vaccines (Chapoval et al. 2001a; Wilson K, David CS. Personal communication).
We tested the ability of HLA-DQ tg mice to respond to house dust mite (Krco et al. 2000;Neeno et al. 1996), ragweed (Chapoval et al. 1998), and cockroach (Papouchado et al. 2000) allergens. The Mini-Monograph | Chapoval and David 246 VOLUME 111 | NUMBER 2 | February 2003 • Environmental Health Perspectives Figure 1. This diagram depicts schematically how the allergic response develops. Allergens are derived from different sources such as cockroaches, ragweed pollens, and house dust mites that are currently being investigated in our laboratory. Antigen-presenting cells (APC) capture, process, and present allergen as an allergen-derived peptide fragment in the contest of specific MHC II (HLA II) molecules. Recognition of this complex by T-cell receptor (TCR) on CD4 + T cells leads to a cytokine (IL-4) mediated clonal expansion of the T helper 2 (Th2) type cells. Several factors bias the development of Th2 type of response (Romagnani 2000). Allergen also interacts with B-cell receptor and is internalized by B cells following antigen (Ag) processing and presentation to the TCR on expanded population of Ag-specific Th2 cells. This together with costimulatory factors and cytokines leads to clonal expansion and differentiation of the B cells involved. B cells switch to Ag-specific antibody (Ab) production, including IgE and IgG1 in mice (IgE and IgG4 in humans, correspondingly). The allergen-specific Ab bind to their receptors displayed on mast cells. Re-exposure to allergen and subsequent crosslinking of Ab molecules leads to mast cell activation, degranulation, and release of mediators. Th2 cells generate IL-5, which increases the production of eosinophils (Eos) in bone marrow, promotes their release into circulation, and regulates their migration into the lung tissue. Activated eosinophils release their toxic granule contents [major basic protein (MBP), eosinophil cationic protein (ECP), eosinophil peroxidase (EPO)] and cause serious tissue damage. Th2 cytokine can directly contribute to the tissue damage or airway remodeling in asthma (Elias et al. 1999) by inducing and promoting a mucus hypersecretion. They also act on fibroblasts (Fb) and favor the subepithelial fibrosis. Asthma and allergy involve a variety of steps and pathways, all of which are influenced by genetics as well as environmental factors.
response of HLA-DQ tg mice to the whole SRW extract was examined. Mice were immunized subcutaneously with SRW in complete Freund adjuvant. Seven days later draining lymph nodes were isolated and mechanically dispersed into a single-cell suspension. Lymph node cells (LNC) were incubated in vitro with various concentrations of SRW (Figure 2). Proliferation was assessed by 3 [H] thymidine incorporation and expressed as the change (∆) in counts per minute (cpm) over the cell cultures without SRW. Stimulation of LNC with SRW resulted in strong T-cell proliferation in a dose-dependent manner giving peak ∆ of 25,824 and 14,312 cpm for HLA-DQ6 and HLA-DQ8 mice, respectively. HLA-DQ6/DQ8/H-2Aβ 0 mice were unresponsive to any dose of SRW at any point in time. Interleukin (IL)-5 and IL-10 were the primary cytokines produced by in vitro challenged LNC of SRW-primed transgenic mice, suggesting that Th2 response was generated in both transgenics (Chapoval et al. 1998). To identify the subpopulation of T cells necessary for responses to SRW in transgenic mice, monoclonal antibodies (mAb) with different specificity were added to the wells containing cultures of LNC with SRW. Proliferation was significantly inhibited or completely eliminated in cultures with either anti-CD4 or anti-DQ mAb. No significant effect was observed in cultures containing mAb specific for mouse CD8, H-2A b , H-2Aβ b , H-2E/β b , or for human MHC class I molecules. Thus, the in vitro response to SRW extract in HLA-DQ6 and HLA-DQ8 transgenic mice is mediated by CD4 + HLA-DQ-restricted T cells.
Overlapping peptides spanning the entire amino acid sequence of antigen 5 from SRW were used to identify immunodominant and cryptic determinants in HLA-DQ6 and HLA-DQ8 tg mice (Chapoval et al. 1998). To identify immunodominant determinants, mice were immunized with individual overlapping peptide and LNC were challenged in vitro with relevant peptide. To identify cryptic epitopes, mice were immunized with crude allergenic SRW extract and LNC were challenged in vitro with either SRW or individual peptide representing antigen 5. SRWimmunized HLA-DQ6 mice respond to peptide 11-30 of Amb a 5, whereas HLA-DQ8 mice strongly recognize peptide 1-20 (Table 1). We concluded that a naturally processed epitope for HLA-DQ6 molecule resides within residue 10-20 or residue 18-30, whereas for HLA-DQ8 molecule it is residue 1-10 or 8-20. When immunized with peptides, HLA-DQ6-restricted T-cell responses were detected to two determinants (residues 1-20 and 11-30) on Amb a 5. In contrast, LNC of HLA-DQ8 mice recognized three Amb a 5 determinants (residues 1-20, 11-30, and 21-40). Therefore, there is at least one HLA-DQ6-restricted determinant within the region 1-30 of antigen 5, and there are at least two HLA-DQ8-restricted determinants within residue 1-40. We are performing truncation analysis to precisely localize critical residues (Krco et al. 2000). Control H-2Aβ 0 mice failed to show a proliferative response to any of the peptides. The mAb inhibition studies demonstrated that the immune response to individual allergen peptide in tg mice depended on CD4 + T cells and was HLA-DQ restricted (Chapoval et al. 1998). With knowledge of T-cell epitopes on allergens, we are exploring new strategies for immune intervention in allergic diseases by specifically interfering with the interaction among APC, allergen, and T lymphocyte.

HLA Transgenic Mice as Model for Allergen-Induced Asthma
The conventional murine models of antigeninduced experimental asthma have been extraordinarily helpful in addressing selected pathogenetic issues (Krinzman et al. 1996;Kung et al. 1994;Renz et al. 1992). Generation and use of transgenic and knockout mice for the multitude of mouse cytokines and their receptors (Borchers et al. 2001;Cohn et al. 2001;McMillan et al. 2002;Wang et al. 2001;Zhu et al. 2001), cells (Chapoval et al. 2001b;Corry et al. 1998;MacLean et al. 1999;Zuany-Amorim et al. 1998), chemokine receptors (Lukacs et al. 2001;Schuh et al. 2002), costimulatory (Gonzalo et al. 2001;Jember et al. 2001;Mehlhop et al. 2000), adhesion (Fiscus et al. 2001;Wolyniec et al. 1998), and signaling (Das et al. 2001) molecules demonstrate new evidence for the role and significance of the immune system in the mechanisms of allergic tissue inflammation and airway hyperreactivity.
To study the role and contribution of specific HLA class II molecules in the pathogenetic mechanism in atopic diseases and asthma, H-2Aβ 0 , HLA-DQ6, and HLA-DQ8 mice were actively sensitized and later challenged intranasally with SRW (Chapoval et al. , 2002. In a sensitization phase, SRW was applied intraperitoneally on day 0 and day 7 with aluminum hydroxide as an adjuvant. Seven days after the sensitization phase, airway inflammation was induced by application of dialyzed extract into the respiratory tract two times at 6 hr apart. Vehicle-sensitized control mice were injected with an equal volume of aluminum hydroxide suspension in phosphate-buffered saline (PBS), and challenged intranasally with PBS alone. The development of disease was evaluated both histologically and physiologically at 48 hr after challenge. After immunizations and challenge, HLA-DQ6 and HLA-DQ8 tg mice had an increase in cell accumulation in the BAL (Figure 3) and perivascular and peribronchial inflammatory infiltrates in the lung tissue (Figure 4). The severe inflammatory pulmonary reaction involved predominantly eosinophils (Figures  3 and 4). The in vivo response to SRW was mediated by HLA-DQ-restricted CD4 + T cells . H-2Aβ 0 mice did not show any histopathologic features of Mini-Monograph | HLA transgenic mice as models for allergic diseases and epitope recognition Environmental Health Perspectives • VOLUME 111 | NUMBER 2 | February 2003    (Figures 3 and 4) and the T-cell response to subsequent rechallenge with SRW was absent in these mice. HLA-DQ6 mice were treated with anti-DQ mAb by intraperitoneal injections 24 hr before, simultaneous with, and after each SRW application. This treatment significantly lowered the number of BAL total cells, eosinophils, and lymphocytes recovered from allergen-sensitized HLA-DQ6 mice, suggesting an important role for HLA molecules in the development of allergic inflammation in our model (Figure 3).
It has been firmly established that allergic inflammatory responses depend on the presence of Th2 cells (Krug and Frew 1997). As such, we examined BAL and spleen MNC cultures for cytokine expression. SRW treatment induced local lung IL-5 and IL-13 production in HLA-DQ mice with undetectable levels of interferon-γ (Chapoval et al. , 2002 indicative for Th2 effector function. Similar to our observations in BAL, direct analysis of cytokine synthesis in spleen MNC cultures of HLA-DQ6 mice also demonstrated the production of Th2 cytokines (Chapoval et al. 2002). Thus, the immune system of HLA-DQ tg mice exhibits Th2 cell effector activity in local lymphoid and target tissues. None of Th2 cytokines was detected in BAL and culture supernatant preparations obtained from H-2Aβ 0 control animals, suggesting that their inability to receive CD4 + T-cell stimulation from MHC II molecules leads to a complete failure to express Th2 type cytokines and allergen-specific Ab  or to develop allergic eosinophilic airway inflammation.
Human asthma is characterized and its severity is diagnosed by heightened airway reactivity to varying concentrations of nonspecific bronchoconstrictors or spasmogens such  . Cytospin analysis of BAL cells from SRW-sensitized and -challenged HLA-DQ tg mice. BAL fluids were collected at 48 hr after intranasal challenge, total leukocyte numbers were determined with standard hemocytometer, and differential cell counts were made using Giemsa-stained cytospin slides. Total BAL cell numbers: H-2Aβ 0 , 674 x 10 3 ; HLA-DQ6, 1,460 x 10 3 ; HLA-DQ8, 1,048 x 10 3 ; HLA-DQ6, anti-DQ, 556 x 10 3 , correspondingly. RT, relaxation time; Pef, peak expiratory flow; Pif, peak inspiratory flow. Airway reactivity was measured 48 hr after the last intranasal challenge by recording respiratory pressure curves by whole body plethysmography (Buxco Electronics Inc., Sharon, CT). After establishing a stable baseline of breathing in the plethysmograph, we exposed mice for 3 min to nebulized PBS and subsequently increasing concentrations of nebulized methacholine in PBS. After each nebulization, recordings were taken for 3 min. Values of PenH (index of airway obstruction) were calculated by BioSystem XA software with the following equation: PenH = (Te/RT-1) x (Pef/Pif). as methacholine. To demonstrate these tendencies in our tg mice, we used a noninvasive barometric plethysmography, which allowed us to measure lung physiology in conscious, unrestrained mice (Hamelmann et al. 1997). When we challenged PBS-treated or SRWtreated mice with increasing concentrations of methacholine in a whole-body plethysmograph ( Figure 5), we did not find an airway hyperreactivity in MHC II knockout mice. Allergen challenge induced strong hyperreactivity in HLA-DQ6 mice, but hyperreactivity was lower for HLA-DQ8 mice. The presence of specific HLA-DQ molecules was required for the induction of heightened airway reactivity in vivo. Overall, transgenic expression of two different HLA-DQ molecules in MHC II knockout mice produced different effects on several parameters of experimental allergen-induced asthma (levels of BAL eosinophilia, pulmonary inflammation, AHR, BAL cytokine, and protein concentrations). Thus, HLA class II transgenic mice offer the opportunity to investigate the role of distinct HLA molecules in allergen sensitivity and may be useful for devising new immune based therapies for allergic asthma.