Immune sensitization to methylene diphenyl diisocyanate (MDI) resulting from skin exposure: albumin as a carrier protein connecting skin exposure to subsequent respiratory responses

Mouse and bovine albumin, triton X-100, sodium chloride, dithiothreitol (DTT), MDI, protease inhibitor cocktail and Tween 20 were from Sigma (St. Louis, MO). Urea and Tris-HCl were from American Bioanalytical (Natick, MA). Nonidet P40 substitute (Igepal CA-360) was from USB Corporation (Cleveland, OH). Acetone was from J.T. Baker (Phillipsburg, NJ). Ethylenediaminetetraacetic acid (EDTA) and phosphate buffered saline (PBS) were from Gibco (Grand Island, NY). Nunc Maxisorp™ microtiter plates were obtained through VWR International (Bridgeport, NJ). SuperSignal West Femto Maximum Sensitivity enhanced chemiluminescence substrate was obtained through Thermo Fisher Scientific (Rochester, NY). Tetramethylbenzidine (TMB) substrate was from BD Bioscience (San Jose, CA). Streptavidin conjugated alkaline phosphatase and p-nitrophenyl phosphate (pNPP) substrate were from Kirkegaard & Perry Laboratories (Gaithersburg, MD). Peroxidase conjugated rat anti-mouse anti-IgG₁, and anti-IgG_2a were from Pharmingen (San Diego, CA). Protein G Sepharose 4 Fast Flow was from GE Healthcare (Piscataway, NJ). Biotin-labeled rat anti-mouse IgE was from BioSource International, Inc. (Camarillo, CA). Imperial protein stain and rabbit anti-mouse IgG were from Pierce (Rockford, IL). Nitrocellulose and reducing gel electrophoresis buffer were from Bio-Rad (Hurcules, CA). Rabbit anti-tropomyosin, rabbit anti-collagen type 1/α2, and mouse anti-cytokeratin 14 were from Santa Cruz Biotechnology, Inc (Santa Cruz, CA).

Female BALB/c mice, 9 to 12 weeks, from the National Cancer Institute (Frederick, MD), were used in all experiments. The backs of mice were shaved with electric clippers 1 day before exposure to 50 μl of MDI ranging in dose from 0.1%-10% weight/volume (w/v), delivered in a 4:1 acetone/olive oil "vehicle" (approximate surface area 0.5 - 1 cm² on right side). Control mice were identically exposed to 50 μL of an acetone/olive oil mixture without MDI. Mice were anesthetized during the skin exposure, and 20 minutes after application, the exposed area was cleansed with 70% ethanol. Mice were re-exposed a second time 7 days later on the opposite (left) side of their back. Serum of exposed mice was obtained on day 21 and analyzed by ELISA for MDI-specific antibodies, and used as a probe to detect MDI (exposure)-induced antigenic-modification of "self" mouse skin proteins. In some studies MDI skin exposed mice were subsequently exposed to MDI-albumin conjugates via the respiratory tract (see below). A time line of skin/airway exposures and sample acquisition is shown in Figure 1.

Mouse sera samples were analyzed for MDI-specific antibodies using an enzyme-linked immunosorbant assay (ELISA), similar to that our laboratory has recently developed for measuring MDI-specific human antibodies [38]. Microtiter plates were coated with 1 μg/well of mouse albumin conjugated with MDI (see below), or control "mock exposed" mouse albumin, by overnight incubation at 4°C, in 0.1 M carbonate buffer (pH 9.5). Plates were "blocked" with 3% (w/v) bovine serum albumin before murine serum samples were titrated in blocking buffer. Sera were incubated for 1 hour at 25°C, followed by a 1:2000 dilution of peroxidase conjugated rat anti-mouse anti-IgG₁ or anti-IgG_2a. MDI-specific IgE was detected with biotin-labeled secondary rat anti-mouse IgE, followed by streptavidin-conjugated alkaline phosphatase. ELISAs were developed with TMB or p-NPP substrate and optical density (OD) measurements were obtained on a Benchmark microtiter plate reader from Bio-Rad. All samples were tested in triplicate to obtain average values expressed in figures.

MDI-specific IgG data are reported as end-titers; the reciprocal of the highest dilution that yields a positive OD reading, > 3 S.D. units above control serum from unexposed mice. Isocyanate-specific IgE data are represented as a binding ratio, as recommended in previous clinical studies, which is calculated as the (OD of wells coated with MDI-albumin) ÷ (OD of wells coated with control albumin) [39]. Total serum IgE levels were measured as previously described [40].

MDI-mouse albumin conjugates used for ELISA and respiratory tract challenge were prepared under the reaction conditions recently defined to yield optimally antigenic MDI-conjugates with human albumin [38]. Mouse albumin in phosphates buffered saline (pH 7.2) at 5 mg/ml was mixed with a freshly prepared solution of 10% (w/v) MDI dissolved in acetone, to achieve a final MDI concentration of 0.1% (w/v). The reaction mixture was rotated end-over-end for 2 hours at room temperature, dialyzed against PBS and (0.2 μM) filtered. "Mock exposed" albumin was identically prepared, using only acetone (1% v/v final concentration) for the 2-hr exposure period. MDI conjugation to mouse albumin was verified based on characteristic shift in electrophoretic mobility, and absorbance at 250 nm, due to MDI's double ring structure [41]. In later experiments, for comparative purposes (with albumin purified from skin exposed to MDI in vivo, see below), we generated MDI-mouse albumin conjugates in vitro with varying levels of MDI/protein molecule, by varying the MDI concentration during conjugation reactions.

Mice were lightly anesthetized with methoxyflurane and exposed to 50 μL of a 2 mg/ml solution of MDI-albumin or control "mock exposed" albumin in PBS by means of an intranasal droplet on days 14, 15, 18, and 19; and sacrificed by means of CO₂ asphyxiation on day 21. Bronchoalveolar lavage (BAL) cell counts and differentials were performed as previously described [40].

Skin samples from MDI exposed mice were Western blotted with serum IgG from autologous mice that had been "skin-only" exposed to MDI, to detect "self" proteins antigenically modified by MDI exposure. Specificity controls included parallel blots with sera from mice exposed to vehicle only, and irrelevant (anti-ovalbumin) hyperimmune sera. Electrophoresis and Western blot were performed as previously described using pre-cast sodium dodecyl sulfate (SDS) acrylamide gels (4-15% gradient) from BioRad, and nitrocellulose membrane [42, 43]. Nitrocellulose strips were incubated for 2 hrs with a 1:100 dilution of sera, washed extensively with PBS containing 0.05% Tween 20, incubated with a 1:2000 dilution of peroxidase conjugated anti-mouse IgG, and developed with enhanced chemiluminescence substrate.

Proteins from MDI exposed mouse skin were purified by a 2-step (isoelectric focusing/electroelution) process, guided by serum IgG from "skin only" exposed autologous mice, to detect antigenic modification. Preparative isoelectric focusing was performed using a Rotofor^® system from Bio-Rad, according to the manufacturers recommendations, to initially separate skin proteins into 20 fractions between pH 3 and 10, with subsequent re-focusing between pH 3 to 6, to increase resolution. Rotofor fractions containing proteins antigenically modified by MDI exposure were further fractionated and analyzed by parallel Western blot/SDS-PAGE, from which they were excised using a Bio-Rad Model 422 Electro-Eluter run at constant current (8-10 mA/glass tube) for 3-5 hrs. Purified proteins were aliquoted and further analyzed for protein sequence (see below) and confirmation of MDI-antigenicity via immunoblot with serum IgG from exposed mice.

Liquid chromatography (LC) followed by tandem mass spectrometry (MS/MS) was performed by the Yale Keck Center on a Thermo Scientific LTQ-Orbitrap XL mass spectrometer, as previously described [44]. Briefly, purified proteins were reduced and carboxamidomethylated, trypsin digested and desalted with a C18 zip-tip column before MS/MS analysis. From uninterrupted MS/MS spectra, MASCOT compatible files (http://​www.​matrixscience.​com/​home.​html) were generated, and searched against the NCBI non-redundant database [45, 46]. For true positive protein identification, the 95% confidence level was set as a threshold within the MASCOT search engine (for protein hits based on randomness search). In addition, the following criteria must also have been met (1) two or more MS/MS spectra match the same protein entry in the database searched, (2) matched peptides were derived from trypsin digestion of the protein, (3) the peptides be murine in origin, and (4) the electrophoretic mobility must agree with the molecular weight. The identity of the purified proteins was further confirmed by Western blots with commercially available polyclonal or monoclonal antibodies (type I collagen, keratin-14, and tropomyosin), using hyperimmune anti-ovalbumin rabbit or mouse serum as a (negative) specificity control.

Statistical significance was determined using ANOVA with a block design for pooled data from more than one experiment. Antibody data, calculated through 2-fold dilutions, were log(2) transformed for analysis.

The capacity of MDI skin exposure to induce an MDI-specific antibody response was evaluated through ELISA analysis of sera from mice exposed to MDI diluted in acetone, at varying concentrations ranging from 0.1-10% weight/volume (w/v). We found that skin exposure to ≥ 1% MDI resulted in the development of high serum levels of MDI-specific antibodies. As shown in Figure 2, the end titers for MDI-specific antibody reached >1:100,000 and >1:30,000 for IgG₁ and IgG_2a subclasses respectively. MDI-specific IgE and total IgE serum levels were also elevated, up to 6-fold above control levels. The IgG and IgE induced by MDI skin exposure did not bind to unexposed proteins, or other reactive chemical "haptens" such as DNCB or adipoyl chloride (not shown).

Mice initially exposed to MDI via the skin, were subsequently exposed via the respiratory, to a water soluble derivative of MDI (mouse albumin conjugates), in an adaptation of our murine HDI asthma model [22]. In the present experiments, mice that received only vehicle (acetone/olive oil) skin exposure, exhibited no change in bronchoalveolar lavage (BAL) cell numbers or differentials, when (airway) challenged with MDI-albumin conjugates. However, mice with previous (≥1%) MDI skin exposure developed significant airway inflammatory responses to respiratory challenge. The observed increase in total cell numbers of BAL samples (obtained 48 hours post exposure) was primarily due to increases in eosinophils and lymphocytes (Figure 3). Thus, respiratory tract exposure, to concentrations of MDI (albumin conjugates) that normally do not evoke cellular inflammation, causes pathologic changes (increased number of airway cells with Th2-profile) in mice previously exposed to MDI via the skin.

The initial MDI (skin) exposure dose was found to have a strong affect on the level of airway inflammation subsequently induced by respiratory tract challenge. The largest degree of airway inflammation was observed in mice initially (skin) exposed to MDI at a 1% (w/v) concentration, with more limited, albeit significant, inflammation in mice that had been skin exposed to 10% (w/v). The reason for the paradoxically limited airway inflammation in mice (skin) exposed to the highest test dose of MDI (10% w/v) remains unclear; however, analogous findings have been reported in HDI exposed mice [22]. A similar (non-linear dose-response) phenomenon is well-described for contact sensitization to many other reactive chemicals, e.g. formaldehyde, picryl chloride, DNCB [47].

In mice with prior MDI skin exposure, subsequent respiratory tract exposure to MDI-albumin conjugates was found to boost MDI-immune sensitization, based on levels of MDI-specific serum IgG and IgE. As shown in Figure 4, statistically significant increases were detectable among Th2-associated subclasses/isotypes, IgG₁ and IgE, but not in the Th1-associated subclass, IgG_2a. Thus, in mice previously exposed to MDI via the skin, subsequent respiratory tract exposure to MDI (albumin conjugates) further boosts MDI immune sensitivity.

As shown in Figure 5A, detergent extracts from 1% MDI exposed skin contained a single antigenically-modified protein, specifically recognized by antibodies from autologous MDI skin (only) exposed mice, but not control mouse sera. The "MDI antigen" was purified from exposed skin by a 2-step process (Figure 5B, and 6A), and identified as albumin through LC-MS/MS analysis (see Additional file 1). The antigenically modified albumin from exposed skin exhibited biophysical properties consistent with MDI conjugation, when compared with albumin purified from vehicle-only exposed skin, or MDI-mouse albumin conjugates prepared in vitro; specifically, alterations in electrophoretic migration and change in absorbance at 250 nm (Figure 6A&6B).

Additional "MDI antigens", specifically recognized by antibodies from MDI skin (only) exposed autologous mice, but not control mouse sera, were detectable in urea extracts from skin exposed to the highest test dose of MDI (10%), as shown (Figure 7A). Among these antigenically-modified proteins, the most prominent, based on recognition by serum IgG from skin exposed autologous mice, were purified through elecrophoretic fractionation methods, and identified by LC-MS/MS as pro-collagen type 1/α2, keratin 14, and tropomyosin (see Additional file 1). Their (MDI) antigenicity and identity were further confirmed by Western blot with autologous serum IgG from skin exposed mice (Figure 7B) and commercially available protein-specific (collagen, keratin, tropomysosin) antibodies (not shown).