Domestic Mite Antigens in Floor and Airborne Dust at Workplaces in Comparison to Living Areas: A New Immunoassay to Assess Personal Airborne Allergen Exposure

Objectives Allergens produced by domestic mites (DM) are among the most common allergic sensitizers and risk factors for asthma. To compare exposure levels between workplaces and living areas a new assay able to measure airborne DM antigen concentrations was developed. Methods At workplaces and in living areas, 213 floor dust samples and 92 personal inhalable dust samples were collected. For sensitive quantification of DM antigens, a new enzyme immunoassay (EIA) based on polyclonal antibodies to Dermatophagoides farinae extract was developed. Reactivity of five house dust mite and four storage mite species was tested. All dust samples were tested with the new EIA and with the Der f 1 and Der p 1-EIAs (Indoor Biotechnologies, UK) which detect major allergens from D. farinae and D. pteronyssinus by monoclonal antibodies. Samples below the detection limit in the DM-EIA were retested in an assay variant with a fluorogenic substrate (DM-FEIA). Results The newly developed DM-EIA detects antigens from all nine tested domestic mite species. It has a lower detection limit of 200 pg/ml of D.farinae protein, compared to 50 pg/ml for the DM-FEIA. DM antigens were detected by DM-EIA/FEIA in all floor dust and 80 (87%) of airborne samples. Der f 1 was found in 133 (62%) floor dust and in only 6 airborne samples, Der p 1 was found in 70 (33%) of floor samples and in one airborne sample. Der f 1 and DM concentrations were highly correlated. DM-antigens were significantly higher in inhalable airborne samples from textile recycling, bed feather filling, feed production, grain storage and cattle stables in comparison to living areas. Conclusions A new sensitive EIA directed at DM antigens was developed. DM antigen quantities were well correlated to Der f 1 values and were measurable in the majority (87%) of airborne dust samples. Some workplaces had significantly higher DM antigen concentrations than living areas.


Introduction
Multicenter studies across Europe have found that house dust mites, among the indoor allergens, yield the highest sensitization frequency [1] and that exposure to mite allergens is an important risk factor for allergic respiratory diseases and asthma [2][3][4][5]. Therefore, many studies on mite allergen concentrations have been performed. In most cases, reservoir dust from floors or mattresses has been analysed by monoclonal antibody-based immunoassays to group I or II allergens from the house dust mites Dermatophagoides pteronyssinus and D.farinae [6][7][8]. However, according to Paufler et al. [9] ''assessing allergens in ambient air would better represent human exposure because inhalation is the main route of uptake, and a close correlation between levels of floor and air antigens has not yet been proved''. Especially for occupational allergen exposure assessment, the method of choice is the sampling of inhalable dust in the breathing zone of the suspected exposed worker and then the determination of allergen levels per cubic meter. Due to extremely low allergen-carrying particle concentrations, airborne exposure assessment of mite allergens is difficult to perform. In several studies, the level of airborne house dust mite allergens in undisturbed conditions could not be determined because the limits of detection of the assays used were inadequate [10]. Therefore, the first step for measuring airborne mite allergen exposure was to develop an assay with sufficient sensitivity.
To assess for allergens besides mite group I or group II allergens, polyclonal antibodies directed at an allergen extract from D. farinae were used for a two-sited EIA amplified by a polyenzyme-conjugate and, in the most sensitive assay variant, by a fluorogenic substrate. The specificity of the new assay was controlled by testing seed, mould, insect and other mite extracts and by comparison with the results of the Der f 1 and Der p 1 assay. In this pilot study, different workplaces with suspected mite exposure were compared with living areas; in total, 213 floor dust samples and 92 personal inhalable dust samples were collected and analysed. Production, purification and biotinylation of antibodies The animal experiments were approved by the local commissioner for animal welfare and notified by the ''Landesamt für Natur, Umwelt und Verbraucherschutz, Nordrhein-Westfalen''. A female New Zealand white rabbit was immunized subcutaneously with 0.5 mg of D. farinae extract emulsified in TiterMax Gold (TiterMaxH USA, Norcross, GA, USA). Two booster injections with the same allergen concentration in TiterMax Gold were carried out at 5-week intervals. The serum was collected 5 weeks after the last injection and stored at 280uC. For antibody purification, the IgG fraction was isolated by affinity chromatography with HiTrap Protein A column and with antigen coupled Nhydroxysuccinimide-ester-activated Sepharose column following the manufacturer's instructions (GE Healthcare, Uppsala, Sweden). Antigen-affinity purified antibodies were biotinylated by mixing with a 30-fold molar excess of biotinoyl-e-aminocaproic acid-N-hydroxysuccinimide ester (Roche, Mannheim, Germany) dissolved in dimethylsulfoxide for 4 h at room temperature.

Sampling and extraction of floor dust and inhalable dust
Sampling took place in Germany from mid-2007 until the end of 2011. Floor dust samples were taken in sleeping rooms (n = 56), living and working rooms (n = 20), kitchens (n = 4) and corridors (n = 4) of 60 private homes and in 21 workplaces of 11 different industries (n = 129). Permission for sampling at workplaces was obtained by the employers and involved employees to obtain information on house dust mite allergen concentrations, sampling in own private homes was performed by volunteering colleagues of the authors.
ALK dust collection cassettes (ALK-Abello, Hørsholm, Denmark) with pre-weighed glass fiber filters (Macherey & Nagel, Düren, Germany) were attached on vacuum cleaners, and an area of 1-2 m 2 on floors was vacuumed for 4 minutes. Within the next three days, samples were frozen (220uC) for 24 h to kill mites and subsequently stored at room temperature until extraction.
Personal airborne dust samples were taken at 33 workplaces of 15 different industries (n = 76) and in 16 living areas during housework including vacuuming and bed-making (n = 16) for 1-4 h with GSP-sampling heads for inhalable dust and air samplers at a flow rate of 3.5 l/min in the individuals' breathing zones. All sampling was performed with pre-weighed and coded 3.7 cm polytetrafluoroethylene (PTFE) filters with 1 mm pore size (FALP03700, Millipore). After sampling and conditioning for . Filters with floor dust were extracted in 10-50 ml PBST, and PTFE filters were extracted in 5 ml PBST by mixing. All extracts were centrifuged for 15 min at 3000x g, and the supernatants were stored frozen in aliquots at 280uC until testing in mite assays.

Statistics
The strength of the relationship between EIA measures was analysed using Pearson correlations. All weights or EIA values below the detection limit were replaced by 2/3 of this limit. The Mann-Whitney test was used for the comparison of workplaces with living areas. A pvalue of less than 0.05 was considered statistically significant. All statistical analyses were performed using GraphPad Prism (version 5.03 for Windows; GraphPad Software, San Diego USA, www.graphpad.com).

Results
With the purified polyclonal antibodies directed to antigens of the mite species Dermatophagoides farinae, an EIA was developed with a lower detection limit of 200 pg/ml. To assess the specificity of the assay, extracts from several materials were also tested. No reactivity was observed with seeds and grains or mould extracts (Fig. 1 a, b). Recovery rates of 1-5 ng/ml D. farinae standard protein used as spike in 10 mg/ml mould or 100 mg/ml seed extract was between 86-116% without any extract dependent difference. Whereas extracts from Sitophilus granarius, Tribolium confusum and Ephestia kuehniella showed no reactivity, the Blattella germanica, Periplaneta americana and Tenebrio molitor extracts reacted, although for factors of 3200, 70000 and 60000 lower than D. farinae protein, respectively (Fig. 1c). In addition, extracts from eight different domestic mite species were tested (Fig. 1d). Highest reactivity was obtained with proteins from Dermatophagoides microceras and Blomia tropicalis, and the lowest reactivity was obtained with D. pteronyssinus (a factor of 24 lower than D. farinae protein). Reactivity of the polyclonal rabbit antibodies to different mite extracts was analysed in comparison to IgE binding of a patients' pool serum by immunoblots (Fig. 2). The patients' pool and the rabbit antibodies reacted to proteins from all mite species, although more protein bands were detected by rabbit antibodies than by human IgE. In contrast, IgE-binding to proteins of 15 kDa from L. destructor and G. domesticus showed stronger bands than binding by the rabbit antibodies.
As the newly developed D. farinae assay detects extracts of all other tested Astigmata, it was named ''Domestic Mite EIA (DM-EIA)''. All floor dust and airborne samples were analysed with the DM-EIA and with the Der f 1 and Der p 1 EIAs. Whereas Der f 1 was found in 133 (62%) floor dust and only 6 airborne samples, and Der p 1 was measured in 70 (33%) floor dust and one airborne sample, all floor dust samples and 56 (61%) of the airborne samples were measurable with the DM-EIA. All samples with detectable Der f 1or Der p 1 allergen also contained domestic mite antigens. Sensitivity of the DM-EIA was further improved by using a fluorogenic substrate, resulting in a lower detection limit of 50 pg/ml. A comparison of 100 measurements above the detection limit of DM-EIA and DM-FEIA showed nearly identical values with both assay variants (mean ratio EIA versus FEIA value: 1.160.3). From 36 airborne samples negative in the DM-EIA, 24 were above the detection limit of the DM-FEIA. Thus, 80 (87%) of airborne samples contained measurable amounts of domestic mite  (Fig. 3a). The domestic mite antigen content in double positives was at mean 1446121 times higher than the Der f 1 content. The standards of the Der f 1 EIA and DM-FEIA were also compared in a ''crossassay'' design. The D. farinae standard of the DM-FEIA contained at mean 0.7% Der f 1 while the universal standard of the Der f 1 EIA which contains among other purified single allergens 2.5 mg/ ml Der f 1 yielded a value of 2.4 mg/ml antigen in the DM-FEIA.
Comparing the domestic mite antigen (or Der f 1) values with the Der p 1 values, only a low to moderate correlation was observed (Fig. 3 b, Table 1). The correlation was increased when the sums of Der f 1 and Der p1 were compared to DM-values (Fig. 3 c, Table 1).
In the floor dust samples from living areas, Der f 1 could be measured in 73 of 84 (87%) samples, and Der p 1 in 32 (38%) samples. Differences of Der f 1, Der p 1 and domestic mite antigen concentrations per gram floor dust were not significant between sleeping rooms and living/working rooms, although maximum concentrations were found in sleeping rooms. Comparing floor dust samples from living areas with workplaces, only textile recycling and bed feather cleaning resulted in higher median domestic mite antigen and Der f 1 concentrations ( Table 2). All other workplaces had lower domestic mite and Der f 1 concentrations in dust than living areas. These differences were highly significant in most cases.
Analyzing airborne samples, the comparison between living areas and workplaces showed a somewhat different picture (Table 3). At most workplaces, inhalable domestic mite concentrations were higher than those in living areas. These differences were significant for textile recycling, bed feather filling, feed production, grain storage and cattle stables. Inhalable airborne Der f 1 was measurable during textile recycling (five samples .LOD) and feather bed filling (one sample .LOD) alone (data not shown). Inhalable dust concentrations were significantly higher at nearly all workplaces than in living areas during house work.

Discussion
House dust mites provide the predominant inhalant allergens in many parts of the world, but geographic patterns for the common species Dermatophagoides pteronyssinus and D. farinae differ. Europe has mixed populations, with a trend of higher D. pteronyssinus allergen quantities in Western Europe near the coast and dominating D. farinae allergens in continental regions with cold winters. [8]. D. farinae populations increase with relative humidity of 50%, D. pteronyssinus increases at humidity over 60% [13]; thus, the most prevalent house dust mite in Germany varies from year to year. Low winter temperatures seem to reduce the major allergen Der p 1 from D. pteronyssinus rather than levels of Der f 1 from D. farinae [8]. In our study, 133 of the 213 floor dust samples contained Der f 1, whereas Der p 1 was detected in only 70 of these samples. One reason might have been the relative cold winters in Germany in 2008 to 2011. Thus, in our study, the house dust mite D. farinae was clearly the dominant species and therefore is the focus of the analysis presented here.
Using an allergen extract from D. farinae for immunization, a new sandwich EIA was developed with increased detection rate but lower specificity in comparison to the monoclonal antibody based Der f 1 assay from Indoor Biotechnologies. The EIA detects many antigens from D. farinae and from other Astigmata in the following order: D. microceras, Blomia tropicalis, Euroglyphus maynei, Tyrophagus putrescentiae, Acarus siro, Glycyphagus domesticus, Lepidoglyphus destructor and D. pteronyssinus. A second immunization with a D. farinae preparation from Allergon had been performed (data not shown). This D. farinae preparation contained 88% of the antigen and 38% of the Der f 1 content in comparison to the first immunogen. With the new antibodies again all other mite extracts reacted in nearly the same order. Thus, the reactivity pattern does not seem to be arbitrary but may be caused by the amount and similarities of immunodominant mite proteins which do not necessarily reflect the evolutionary relatedness of the species in all cases. Taking this broad detection pattern into account, the newly developed assay was named ''domestic mite EIA'', as, according to Spieksma [14], ''domestic mites' are free-living mites found regularly in the home environment and belonging to two ecologically defined groups: house-dust mite and storage mites.'' Whereas high concentrations of seeds and mould proteins did not influence the domestic mite EIA, very high concentrations of cockroach (Blattella germanica and Periplaneta americana) and mealworm (Tenebrio molitor) protein reacted, although for orders of magnitude less than domestic mites.
However, human IgE antibodies also have a broad detection pattern. Typically, most mite sensitized persons have IgE antibodies to both common house dust mites [15][16][17] and, to a lesser extent, to several storage mites [18,19], and also in some cases to cockroaches [20] partially due to cross-reactivity of antibodies. The pool serum from ten house dust mite sensitized patients contained IgE antibodies to all tested domestic mite species and detected protein bands of all mite extracts in the IgE immunoblot. As the rabbit serum IgG immunoblot, the human IgE immunoblot showed many more mite allergens than group I (,30 kDa in SDS-PAGE under reducing conditions) or group II (,14 kDa) allergens. As the newly developed DM assay detects antigens of many domestic mite species comparable to the human IgE antibodies of domestic mite sensitized patients, its less specific detection profile in comparison to monoclonal antibody-based assays might not be a disadvantage for the assessment of health relevant exposure. Although the rabbit antibodies detect more proteins than the human IgE antibodies, and clearly more proteins than monoclonal antibodies to major allergens, for exposure assessment this problem can be handled as long as the measured antigen amount is proportional to the allergen amount.
Therefore, the concentrations of the major allergens Der f 1 and Der p1 were compared to the DM-antigen values. Despite the differences in specificities between the DM-EIA and the Der f 1 EIA, the measurement values were highly correlated (Pearson r = 0.97, Pearson of log-transformed floor dust values r = 0.90). Whereas the correlation of DM-antigen concentrations to Der p 1 in floor dust samples was only moderate and similar to the correlation between Der f 1 and Der p 1 (Pearson of logtransformed values r = 0.58 and 0.56, respectively), they were also well correlated to the sums of Der f 1 and Der p 1 (Pearson of logtransformed values r = 0.83). We conclude that the main allergen Der f 1 is a relatively constant proportion of total domestic mite antigens that ranged from 0.3 to 2.2% for 90% of the double positive samples. In line was also the mean factor of 144 between the Der f 1 and DM-antigen content of samples and the amount of 0.7% Der f 1 in the D. farinae standard of the DM-EIA.
Proportionality of DM antigen and Der f 1 values was true for both floor dust and airborne dust, with the limitation that only six airborne samples had detectable levels of Der f 1. Therefore, the DM-EIA and, especially, the DM-FEIA, with their much higher number and rate of measurable airborne samples (80 positives, 87%), should be valuable tools for quantification of airborne domestic mite antigens and thus, estimation of allergen exposure.
For floor dust samples, the DM-EIA had sufficient sensitivity; all samples were above the detection limit. Most workplaces had significantly lower DM concentrations per gram of dust than living areas, with the exception of textile recycling, bed feather cleaning and tailoring. Nevertheless, only floor dust from textile recycling plants had significantly higher DM concentrations in dust than living areas. As expected from the high correlation between the DM and Der f 1 values, the floor dust results are affirmed by the Der f 1 measurements. Again, only textile recycling had significantly higher values, whereas nearly all other workplaces had significantly lower allergen concentrations than living areas. The latter results are in line with studies from Italy in which Der f 1 and Der p 1 allergens were detected with low frequencies in schools and workplaces and with high frequency in homes [21]. However, inhalable airborne mite antigen concentrations displayed a different pattern. At most workplaces, a higher exposure occurred than in living areas. These differences were significant for textile recycling, bed feather filling, feed production, grain storage and cattle stables. One reason for higher inhalable airborne antigen concentrations may lie in greater air disturbances at workplaces producing significantly higher concentrations of airborne dust compared to living areas. Especially during feather filling of bed inlets using an air blower, dust and mite antigen exposure was significantly increased, although antigen concentrations in floor dust were lower than in living areas. Using the DM-FEIA, 87% of the airborne samples contained measurable domestic mite antigen concentrations, whereas only 6 samples (6.5%) contained Der f 1, and one sample Der p 1 (1.1%). Low concentrations of group I and group II mite allergens in airborne samples have been described previously and were explained by the relatively larger sizes of mite allergen-carrying particles when compared to the ones carrying the main cat allergen Fel d 1 [22]. The authors stated that particles with mite allergens only become airborne during disturbance and fall within 15 minutes. Quantification of single airborne mite allergens without vigorous air disturbances needed hitherto even with amplified or radioactive assay variants sampling times .24 hours [23][24][25]. The newly developed DM-assays with lower specificity enabled mite antigen quantification in personal dust samples using sampling times of 2-4 h during usual activities in household and at workplace.

Conclusions
Overall, the newly developed DM-EIA and DM-FEIA are valuable tools to measure airborne domestic mite antigen exposure. The results of our first measurements of personal inhalable antigen exposure indicate that at some workplaces, such as textile recycling, higher inhalable domestic mite antigen concentrations occur than during house work. Further studies on domestic mite antigen exposure (e.g., in daycare centers) are planned.