Consideration of methods for identifying mite allergens

Using blood serum samples, researchers are able to identify antibodies, in particular IgEs, resulting from reaction to allergen exposure. It is the most common way to identify an allergic response in humans or animals. Early studies used human sera and crossed radioimmunoelectrophoresis to estimate the number of IgE-reactive species in crude mite extracts. CRIE first separates native proteins based on charge, followed by orthogonal electrophoresis into a gel containing sera where antibody protein precipitates form. Subsequent blotting with radiolabeled anti-IgE reveals reactive peaks. When Chinese HDM allergic patients were screened using this technique, 23 allergen-IgE precipitates were detected in D. pteronyssinus and 17 in D. farinae extracts [17].

More recent discovery attempts used additional forms of protein separation (isoelectric focusing (IEF), one dimensional sodium dodecyl sulfate polyacrylamide gel electrophoresis (1D SDS-PAGE), two dimensional SDS-PAGE (2D SDS-PAGE), or chromatography) or DNA-based tools (cDNA libraries or genomic approaches) coupled with IgE-binding to identify potential allergens. These studies along with the allergens of interest are found in Table 1 and are displayed chronologically to reveal how the discovery techniques have evolved.

Ferrandiz et al. [18] used 1D SDS-PAGE and western blotting to detect 13 IgE-reactive components from extracts of the poorly studied mite D. siboney, a species present in Cuba and recently identified in China [19]. This study went on to purify group 1 and 2 homologues from D. siboney by using affinity chromatography with cross-reactive antibodies raised against group one and two proteins from other mite species. Morgan et al. [16] used preparative IEF followed by non-reducing 1D SDS-PAGE and blotting using sera from 16 individual patients to identify 47 potential allergens in the mite E. maynei. The percentage of patients with IgEs reactive to a given allergen varied from 6 to 88%, thus helping to prioritize major and minor reactive species. Morgan et al. went on to identify a group 2 homologue by using cross-reactive monoclonal antibodies raised against other group 2 proteins. These studies serve as examples of how candidate gene approaches can be used to identify species-specific homologues of known mite allergens.

Candidate gene approaches based on expected biochemical activity can also be used to identify potential novel allergens. The major group 1 allergens are serine proteases whose enzymatic activities are proposed to enhance the immune response by affecting the barrier function of the mucosa [20]. Based on previous studies suggesting that there were at least three distinct serine proteases in D. pteronyssinus fecal extracts, King et al. [6] used chromatography to separate and identify the collagenolytic protease Der p 9 based on its enzymatic activity and weak IgE binding. Additionally, Harris et al. [21] used fluorogenic substrates to perform a large-scale screen for active cysteine and serine D. pteronyssinus proteases, but this approach merely resulted in the identification of the known Der p 1 protein indicating that predictions based on proteolytic activity may not be particularly powerful. Alternatively, native and recombinant active proteases may have autolysis activity that interferes with detecting IgE binding [22].

Separation techniques coupled with protein sequencing (a.k.a. proteomics) [23] are capable of identifying a larger range of potential allergens than candidate gene approaches. Le Mao et al. [24] separated D. farinae extracts by IEF, 1D, or 2D SDS-PAGE followed by blotting with individual patient sera to estimate the number of IgE-reactive species to be 15–16, 7, or 12, respectively. Known allergens were identified in this screen, including multiple isoforms of Der f 2 and Der f 3 detected using monoclonal antibodies against these proteins. Homologues of Der p 4 and Der p 5 were tentatively identified based on amylase activity (Der f 4) and pI (Der f 5). For antigen discovery, the authors excised two high molecular weight species and microsequenced them with Edman degradation to identify a protein resembling Der f 14 [25] and a protein resembling the chitinase allergen from prawns. A similar study used IEF, ID, and 2D electrophoresis to catalogue multiple IgE-binding D. pteronyssinus proteins using sera from allergic dogs [26].

An et al. [27] extended and refined the proteomic approach by first separating D. farinae extracts into IgE-reactive fractions using a gel filtration step followed by a clean-up step to remove contaminants that might interfere with 2D SDS-PAGE. 2D-separated proteins were transferred to membranes and probed for IgE-binding activity using pooled sera from HDM-sensitive asthmatic patients. Seventeen reactive protein spots were recovered and identified using electrospray ionization (ESI) quadrupole time-of-flight (Q-TOF) mass spectrometry and found to represent twelve different protein species, of which four were known D. farinae allergens. Four (Der f 25, Der f 28, Der f 29 and Der f 30) of the eight novel allergens were purified by gel filtration and ion exchange chromatography. Homogeneity of the purified protein samples was inferred by observing a single IgE-band in blots following 1D SDS-PAGE. Using individual patient sera for western blotting and ELISA, An et al. determined that a high proportion of patients (63–76%) had IgEs reactive to the novel potential allergens. A high proportion of patients (60–70%) also had positive skin prick reactions, and the proteins stimulated a response in a basophil activation assay. This study represents a good example of a how modern biochemical methods have improved antigen discovery and the proteins warrant further analysis as part of the confirmation process, particularly in regards to quantitative measurement of specific IgE titers and environmental levels.

Screening of cDNA libraries has also been used to identify both minor and major allergenic components. D. farinae cDNA libraries were screened with HDM-immunized rabbit serum to identify Der f 10 [28] and a high molecular weight reactive protein named Mag3 (Der f 14) [25]. Having a cDNA clone enabled Aki et al. to quickly generate sequence information and produce a recombinant GST-tagged protein for purification and confirmation purposes. Also, once the clone was identified as a tropomyosin homologue, the group was able to use a previously published biochemical purification procedure to isolate the native protein, presumably to a reasonable level of purity. They used dot blots to confirm that approximately 80% of patients had IgEs that recognized native Der f 10 and the major antigens Der f 1 and Der f 2. They used serum from pollen or yeast-sensitive patients as negative controls and followed up the IgE-binding studies with skin prick tests finding that approximately 40% of patients had an immune response to native Der f 10.

More recently, Weghofer et al. screened a D. pteronyssinus cDNA library with pooled serum from asthmatic patients to identify new minor (Der p 21) [29] and major (Der p 23) allergens [30]. Again, the cDNA screening approach allowed for rapid production and purification of a recombinant protein, which the authors used in dot blots to examine serum from three different European populations diagnosed with rhinitis, conjunctivitis, and/or asthma. Approximately 70–87% of all patients exhibited Der p 23-specific binding. Using a chip-based Ig-E binding assay demonstrated that only 58% of patients diagnosed with rhinoconjunctivitis had Der p 23-binding IgEs; whereas, 72% of patients with asthma exhibited Der p 23-binding. Quantitation of IgE levels revealed that Der p 23-specific antibodies were present at similar titers to Der p 1 and Der p 2 antibodies, but some patients only reacted to one of the three proteins. They demonstrated the allergenic activity of Der p 23 using a basophil activation assay, confirmed its presence in mite feces, and were able to detect it at low levels in house dust samples. This study is a strong example of a combined discovery and confirmation approach that yielded exciting results.

Other DNA-based approaches, such as whole genome sequencing or cataloguing expressed sequence tags (ESTs), have been used to identify mite allergens but have lagged behind other methods. This is hardly surprising since until recently less than 1% of the potential genomic information from dust mites was publicly available [31]. A draft sequence now allows up to 95% of genes to be identified and provides a scaffold for DNA fragment assembly [32]. The availability of comprehensive sequence data from mites and mite-associated microorganisms have been exploited by using a candidate gene approach to identify species specific homologues of allergens identified from other sources [31, 32]. Chan et al. also coupled their genomic sequencing efforts to 2D SDS-PAGE followed by immunoblotting to identify twelve new D. farinae proteins with IgE-binding activity. They used their sequence information to rapidly clone and express six of these twelve candidates and found that only one of the recombinant proteins (Der f 24) bound IgE from the majority of HDM-allergic patients without binding in serum from non-allergic or pollen-allergic controls. Der f 24 was found to induce a positive skin prick reaction in 50% of the tested patients.

A bioinformatic approach was used to mine the human genome to identify substrates for the scabies mite protein Sar s 3 in an effort to identify skin proteins that might be affected by scabies infestation [33]. B and T cell epitopes are less well understood than protease sites, however mite sequences may eventually be mined to identify novel mite allergens based on algorithms used to identify these epitopes. Initial models for linear B cell binding motifs had predictive power only slightly better than random, but newer approaches, particularly ones incorporating multiple models, will likely prove more powerful [34]. Lin et al. [35] used such models to define the molecular characteristics of the allergen Der f 29 for the prediction of four peptides comprising B cell epitopes and five peptides comprising T cell epitopes. This led to the identification of a novel subtype of dust mite allergen, Der f 29b.

In the future, the most powerful discovery techniques will likely combine the genomic and proteomic approaches discussed above, i.e., using protein separation techniques such as 2D gel electrophoresis to identify IgE-reactive components followed by protein sequencing (Edmann Degradation) and now, more commonly, tandem mass spectrometry (MS/MS) (with various peptide fragmentation (e.g., trypsinization), ionization (e.g., ESI, matrix assisted laser desorption), and mass detection techniques (e.g., time of flight (TOF)) in combination with deep RNA or DNA sequencing. Briefly, the differing MS/MS applications generate ionized protein fragments that are categorized by mass and charge and then subjected to a further round of fragmentation and categorization. The resulting mass information of the given fragments is used to identify proteins predicted via the RNA and/or DNA sequencing strategies. These approaches allow for the comprehensive and high throughput identification of multiple IgE-binding proteins from mite body or feces extracts. Recently, this technique was used successfully to identify new IgE binding proteins in both D. pteronyssinus and D. farinae [36] and could be considered a model approach for use in additional mite species.

As discussed above, there are a wide range of techniques available to discover new potential allergens, and these discovery methods have both general and specific pitfalls. For all screening strategies, it is important to consider the source of serum. Animal models are useful because they can provide a non-limiting supply of reagents and can provide antibodies for later purification purposes. For this reason, early screening methods often used serum from rabbits immunized with whole mite extracts [25, 28] or monoclonal antibodies derived from HDM-sensitized mice to identify potential allergens, which were then confirmed using patient sera. However, this approach can be limiting since animals and humans respond to different mite proteins. For example, the major group 1 and group 2 allergens in humans are not major allergens in dogs [37].

When considering pooling patient sera for screening purposes, it should be noted that patients with different allergic presentations and/or age may exhibit different patterns of IgE reactivity. Of note, allergic symptoms may represent a time course with AD appearing and subsiding in childhood, with appearance of allergic rhinitis and possibly asthma later in life [38]. IgE-binding may also follow a time course.

Geographic differences should also be considered when comparing results of previous studies and when considering new screening efforts. Patients from different geographic areas are exposed to allergens from different mite species and from other invertebrates, and this influences their IgE-binding profiles. Additionally, dust mites, even from the same species, may have regional geographic polymorphisms. Yi et al. [15] reported differences in the IgE-binding patterns from B. tropicalis extracts derived from Columbia and Singapore, suggesting possible differences in mite subpopulations. Additionally, naturally occurring variations in IgE binding sites have been identified in Blo t 5 [39].

A particular confounding factor for identifying dust mite causal allergens is the presence of IgEs generated by allergic responses to other invertebrate (i.e., helminth, cockroach, or prawn). This can impact the screening process as allergic patients with low titers of HDM-specific IgEs but high titers of non-HDM-related IgEs can yield false positive binding. This is particularly true for allergens with shared epitopes (i.e., glycosylation) [40]. This issue has raised concerns over whether the group 10 proteins are bona fide HDM allergens, as it is known that tropomyosin has IgE-reactive glycosylation sites [41]. Aki et al. [28], who discovered Der f 10, noted that the IgE-binding of recombinant Der f 10 was 25× less than the IgE-binding of the native protein, which could indicate differences in glycosylation as proteins synthesized in E. coli lack such modifications. A later study in the same geographic area detected very low titers of recombinant Der f 10-reactive IgEs in asthmatic patients and little activity in a bronchoprovocation test, which indicates that Der f 10 may play little role in respiratory HDM allergies [42]. In contrast, a recent study using 2D SDS-PAGE followed by immunoblotting with sera from patients with allergic rhinitis detected native Der f 10 as a major Ig-E binding species [43]. In this study, 75% of patients had IgEs reactive to Der f 10 while only 50% reacted to Der f 2. The varying results obtained for Der f 10 could be due to differences in the native and recombinant forms of the protein or differing patient populations. For example, Banerjee et al. [4] found that only 10% of asthmatic patients but 67% of patients with atopic dermatitis had recombinant Der p 10-reactive IgEs. Alternatively, patients with high reactivity to group 10 proteins may have been sensitized to tropomyosin homologues from other species [11].

In some cases, cross-reactivity has been used as a tool to identify new dust mite allergens. Binder et al. [44] screened a cDNA library to identify allergens from the Indian meal moth isolating a clone encoding for an arginine kinase (termed Plo i 1). They generated recombinant Plo i 1 in E. coli and confirmed its allergenicity using basophil activation and skin prick assays. Intriguingly, they found that their recombinant protein could compete away IgE-binding to related proteins in multiple species, including dust mites. This suggests that arginine kinases are pan-allergens, which in dust mites represent the minor group 20 proteins. This cross-species candidate gene approach is not uncommon in the allergen discovery phase [31] and is often in screens from little analyzed species used as the first step to identify proteins resembling known allergens [45].

For protein-based screens, the source of extracts should be carefully considered as well as any confounding effects based on the chosen separation conditions. Of note, there are differences in the number of allergic components identified in whole mite extracts versus feces-specific extracts [46]. This can limit sensitivity for detecting even known allergens. Choopong et al. [43] separated D. farina proteins from whole mite body extracts and detected little Der f 1 because, as the authors claim, this protein is enriched in feces. If the goal of a study is to identify allergens specific for a given condition, appropriate patients and extracts should be used. For example, for asthmatics, it may be more appropriate to consider allergens specifically present in feces since this is believed to be the inhaled component. Additionally, techniques such as IEF, which separates proteins based on charge, and SDS-PAGE, which denatures and separates proteins based on size, have resolutions within specific windows of pI and molecular weight that depend on the chosen conditions. Also, conditions which denature proteins may reduce antibody binding. For example, multiple monoclonal antibodies raised against Der f 1 detected proteins from IEF but not from 1D or 2D SDS-PAGE-separated samples, leading Le Mao et al. [24] to conclude that the antibodies recognize the native protein but not heat denatured forms present in the 1D and 2D PAGE applications. For comprehensive detection of protein species (a.k.a. proteomics) multiple separation techniques should be attempted (see [47]), with the understanding that the number of reactive species can be overestimated due to the presence of multiple isoforms, aggregates, or break-down products of a single expressed protein or underestimated due to separation conditions that affect protein structure and antibody binding.

The majority of papers in Table 2 describe the use of histidine-tagged recombinant proteins synthesized in E coli and purified using Ni++ affinity chromatography. GST-tagged Der f 6 [5] and Blo t 3 [48] were purified with glutathione affinity chromatography. The rarity of this approach likely stems from the fact that the GST-tag is bulkier and more likely to interfere with the activity of the purified protein, although the tag can be removed with a standard enzymatic technique [5].

There are few reports of the use of eukaryotic expression systems, e.g. Pichia pastoris or Spodoptera frugiperda (the baculoviral system), to synthesize dust mite proteins. The touted benefits of these systems include the production of proteins with proper post-translational modifications and proper folding. Misfolding in E. coli can lead to the production of insoluble proteins requiring a biochemical refolding step to restore full IgE-biding and/or enzymatic activity. Olsson et al. [49] reported that recombinant Lep d 2 expressed in E. coli or in S. frugiperda had IgE- binding and basophil activation activities similar to the native protein. Bordas-Le Floch et al. [50] found that recombinant Der p 2 expressed in E. coli did have structural differences detected by circular dichroism but had Ig-E binding and basophil stimulation activities comparable to those of the native or recombinant Der p 2 expressed in P. pastoris. In contrast, Der p 1 or Der f 1 expressed in E coli had significantly less IgE-binding activity when compared to the native form [51]. Hence it is important to consider the structural and post-translational requirements of individual proteins when considering expression systems. For newly discovered allergens, the activity of native and recombinant proteins should be compared.

For isolating native proteins, the most commonly used technique is affinity chromatography using antibodies. Recombinant proteins can be used as antigens to generate specific antibodies which can then be used to isolate the native form [25]. Antibodies can also be generated by immunization with cDNA encoding for the desired target protein [52]. Alternatively, monoclonal antibodies derived by immunizing mice with crude or fractionated mite extracts can serve as tools for purification of potential new native allergens [53]. Additionally, for purification of allergen homologues, cross-reactive antibodies identified in one species can be applied to a new species [18]. In the absence of specific antibodies, alternative chromatography techniques, including ion exchange, gel filtration and hydrophobic exchange, have been used to fractionate whole mite or feces extracts [6, 27, 30, 50]. Fractions were then tested for IgE-binding activity and/or desired enzymatic activity [6] and assayed for homogeneity.

Once a potential allergen has been isolated, the majority of studies go on to test for IgE-reactivity in individual patient sera drawn from a specific study population. The percentage of patients who exhibit IgE-binding provides a crude measure of whether a protein is a major, mid-tier, or minor allergen. Unfortunately, this aspect of allergen testing has a great deal of variability in terms of techniques used and outcomes reported. Dot blots, western blots, and ELISA have all been used to measure the prevalence of IgE response to a given allergen, but most data in the literature is qualitative and difficult to compare between groups. Allergic patients often exhibit a wide range of total IgE levels which represents antibodies derived from multiple sensitizing proteins found in the environment, and this is worth considering when performing assays with binary outcomes (a.k.a. binding versus non-binding). Quantitative solid phase assays such as ImmunoCAP [54] are easier to compare between groups and can provide a more comprehensive picture of how relevant a given allergen will be for an individual patient. Naturally, the accuracy of these tests (both qualitative and quantitative) are highly dependent on the quality of the purified proteins. Standardization of techniques used for isolation, verification of purity, and quantification of IgE binding should help reduce some of the variation in the field.

Additionally, it should be noted that simply binding IgEs does not indicate that a protein is an allergen. The ultimate test of allergenicity is when a protein can elicit an immune response. Skin prick tests are considered the gold standard for demonstrating sensitivity to a given allergen [54]. Both positive (histamine) and negative (diluent) controls are necessary and a wheal size > 3 mm larger than the negative control is considered a positive result. In general, there is a good concordance between IgE-binding and skin prick responses [55]. However, An et al. [27] reported that 100% of patients had IgE-reactivity to Der f 24, but only 50% were positive in the skin prick test. The other most common used test for allergenicity, is the basophil activation assay where peripheral blood basophils isolated from allergic patients are tested for upregulation of CD203c when challenged with a purified potential allergen.

Skin prick and basophil activation assays are currently the most frequently used tools to verify an allergic response. However, there are additional models relevant for airway allergies. Animal models have been developed which recapitulate aspects of respiratory allergies in humans. Bordas-Le Floch et al. [50] examined T cells isolated by bronchial lavage from Der p 2-sensitized mice demonstrating that recombinant Der p 2 could stimulate cytokine release. They also used an in vivo mouse model of asthma to test the effectiveness of Der p 2 protein immunotherapy. Sublingual treatment of Derp-2 sensitized mice with recombinant Der p 2 significantly reduced airway hyperreactivity as measured by whole body plethysmography. Airway challenge models have also been used in clinical testing. Minami et al. [42] used a bronchoprovocation test to demonstrate that there was a strong correlation between Der p 1 and Der p 2-specific IgE levels and airway response to an HDM-challenge. There are additional clinical assays for allergen testing relevant to rhinitis, conjunctivitis, or other allergic responses [56].

Confirming an allergic response under laboratory conditions is an important step towards defining an allergen. However, ultimately, the protein should be able to elicit a response under natural conditions, including the concentrations found in the home or work environments. Also, understanding these local concentrations can help lead to the identification of high risk areas and the development of strategies to mitigate exposure.

The growing availability of isolated allergens offers an opportunity for developing assays (particularly ELISA-based assays) that could be used to quantify allergen levels in dust. Yasueda et al. [57] developed a fluorometric ELISA for detecting the major allergen Der p1/Der f1 and used this technique to assay the levels of this allergen on skin and bedding finding values ranging from 1.1 to 354 ng/m². A highly sensitive assay was necessary to detect such low levels, and similar or even greater sensitivity may be required for the detection of other putative allergenic mite proteins. Additionally, airborne concentrations of these allergens, which are likely the most relevant form when considering the development of respiratory symptoms, are even more difficult to detect. It should also be noted that studies using environmental sampling typically report considerable variation even when testing different sites in the same room [58], making it difficult to calculate the precise dose experienced by a given patient. This limits our understanding of how environmental exposure contributes to sensitization and development of symptoms and is a major limitation when attempting to definitively demonstrate the allergenicity of a given protein.