Antibody conjugates bispecific for intercellular adhesion molecule 1 and allergen prevent migration of allergens through respiratory epithelial cell layers.

To the Editor:

It is becoming increasingly evident that intact epithelial barrier function is important for preventing allergens to reach tissues involved in allergic inflammation. Therefore it is tempting to speculate that strategies that either strengthen epithelial barrier function or prevent allergen from crossing the epithelium might have therapeutic potential.

In this study we produced an antibody conjugate bispecific for intercellular adhesion molecule 1 (ICAM1) and a major respiratory allergen (ie, the major grass pollen allergen Phl p 2) termed P2/ICAM1 to investigate whether such a conjugate can inhibit the migration of allergens through respiratory cell layers and reduce the activation of the inflammatory cells underneath.

ICAM1 was used as a target molecule to anchor allergen-specific antibodies on human bronchial epithelial cells because it is expressed on the surfaces of respiratory epithelial cells, especially under inflammatory conditions, and has a low turnover rate. Furthermore, ICAM1 is a major target molecule for cellular entry of human rhinovirus (HRV) strains, which are implicated in asthma exacerbations, and has been reported to be highly expressed on the respiratory epithelium of allergic patients.

Antibody conjugates were formed through streptavidin-biotin coupling (ie, conjugation) by using different ratios of the individual components (P2/ICAM1: 1:1, 1:0.5, and 1:0.25; see the Methods section in this article's Online Repository at www.jacionline.org). P2/ICAM1, but not a conjugate formed with a Phl p 5–specific antibody, reacted specifically with recombinant Phl p 2 and ICAM1, as shown by means of ELISA (see the Methods section and Figs E1 and E2 in this article's Online Repository at www.jacionline.org). Furthermore, P2/ICAM1 bound specifically to the respiratory epithelial cell line 16HBE14o- and immobilized Phl p 2 on the cells, as shown by using fluorescence-activated cell sorting (FACS) and immunofluorescence microscopy, respectively (Fig 1 and see the Methods section and Fig E3 in this article's Online Repository at www.jacionline.org). Images in Fig 1 show colocalization (Fig 1, A, yellow, merge) of Phl p 2 (red, Alexa Fluor 568) and P2/ICAM1 (green, Alexa Fluor 488) on the cell surface. When either Phl p 2 (Fig 1, B) or Phl p 2–specific rabbit antibodies (Fig 1, C) were omitted, no binding of Alexa Fluor 568 goat anti-rabbit IgG was observed. When P2/ICAM1 was added to the cells, it was detected with Alexa Fluor 488 goat anti-mouse IgG specific for αICAM1 mouse IgG (Fig 1, A-C, green). When P2/ICAM1 was omitted, detection antibodies did not bind (Fig 1, D). No cell staining was found when only secondary antibodies were applied (data not shown). P2/ICAM1 remained bound to 16HBE14o- cell surfaces for up to 72 hours at 37°C (see Fig E4 in this article's Online Repository at www.jacionline.org).

Fig 1

Visualization of P2/ICAM1 and Phl p 2 on 16HBE14o- cells by means of immunofluorescence microscopy. Images (A-D) show cells incubated with different combinations of reactants (left margin and bottom), which were stained with Alexa Fluor 488–labeled anti-mouse antibodies (green, left column) and Alexa Fluor 568–labeled anti-rabbit antibodies (red, middle column) to visualize αICAM1-mouse IgG and Phl p 2, respectively. Nuclei were stained with 4′, 6-Diamidino-2-Phenylinodole, Dihydrochloride (blue), and merged images are shown in the right column. White bars = 20 μm.

Fig E3

Detection of ICAM1 or P2/ICAM1–captured Phl p 2 on respiratory epithelial cells by using flow cytometry. A, ICAM1 was detected on the surfaces of 16HBE14o- cells by using ICAM1–specific antibodies (white graph) or an isotype control (gray graph). FITC, Fluorescein isothiocyanate. B, Cells preincubated with different ratios of P2/ICAM1 (1:1; 1:0.5; and 1:0.25) and Phl p 2 were probed for cell-bound Phl p 2 with specific rabbit antibodies (white graphs) or the isotype control (gray graphs). C, Scatter plots showing percentages of Phl p 2–reactive cells with 3 different ratios of P2/ICAM1. D, Cells incubated with Phl p 2–specific human IgG1 antibodies, followed by Phl p 2 and Phl p 2–specific rabbit antibodies. One representative experiment of 3 independent experiments is shown.

Fig E4

Visualization of P2/ICAM1 and Phl p 2 on 16HBE14o- cells by using immunofluorescence microscopy over a time course of 24 (A), 48 (B), and 72 (C) hours. Images showing cells incubated with different combinations of reactants (left margin and bottom), which were stained with Alexa Fluor 488–labeled anti-mouse antibodies (green, left column) and Alexa Fluor 568–labeled anti-rabbit antibodies (red, middle column) to visualize αICAM1–mouse IgG and Phl p 2, respectively. Nuclei were stained with 4′, 6-Diamidino-2-Phenylinodole, Dihydrochloride (blue), and merged images are shown in the right column. White bars = 20 μm.

In a subsequent series of experiments, we demonstrated that P2/ICAM1 can prevent the apical-to-basolateral penetration of Phl p 2 but not of the birch pollen allergen Bet v 1, a similarly sized but not related allergen, through a layer of the cultured respiratory epithelial 16HBE14o- cells by using a well-established Transwell culture system (Costar Corning Incorporated, Corning, NY; see Figs E5 and E6 in this article's Online Repository at www.jacionline.org), as described in the Methods section in this article's Online Repository. In the experimental Transwell model allergen concentrations were tested that, according to studies analyzing the release of grass pollen allergens during pollen seasons, might naturally occur on the mucosa. Furthermore, we showed that P2/ICAM1 also reacted specifically with natural grass pollen–derived Phl p 2 (data not shown).

In the Transwell experiments we found that the apical addition of P2/ICAM1 prevented the penetration of Phl p 2 over the full period of analysis (ie, for 72 hours) into the basolateral compartment (see Fig E5, A, gray bars, +) when compared with conditions without addition of P2/ICAM1 (Fig E5 A, gray bars, −). When Phl p 2 and P2/ICAM1 conjugates were omitted, no Phl p 2 signal was detected (see Fig E5, A, no Phl p 2, 72 hours). Additionally, no relevant penetration of P2/ICAM1–Phl p 2 complexes into the basolateral wells was found (Fig E5, B, gray bars, +). The addition of the P2/ICAM1 conjugate alone did not affect the integrity of the epithelial layers when resistance was measured at 24, 48, and 72 hours (data not shown).

Fig E5

Cell-bound P2/ICAM1 conjugates inhibit transepithelial migration of Phl p 2. A, ELISA detection of free Phl p 2 in apical (black bars) and basolateral (gray bars) compartments of cultured 16HBE14o- cell monolayers that had been preincubated with (+) or without (−) P2/ICAM1 conjugates. Concentrations of free Phl p 2 corresponding to OD values (y-axis) were measured at different time points (x-axis: 24, 48, and 72 hours) in wells with or without cells (no cells). In the negative control Phl p 2 was omitted (no Phl p 2). B, Simultaneous detection of P2/ICAM1–Phl p 2 complexes in the same samples as in Fig E5, A. Shown is one of 3 independent experiments producing comparable results.

Finally, we tested whether a reduction of transepithelial allergen migration by P2/ICAM1 has an effect on basophil activation to assess the migration of Phl p 2, capable of activating mast cells or basophils. Results from basophil activation tests (see the Methods section and Fig E7 in this article's Online Repository at www.jacionline.org) performed with samples from 3 independent Transwell experiments with basophils from 3 representative allergic patients are shown in Fig 2. The extent of basophil activation induced with basolateral samples showed a significant and consistent reduction for samples taken from cultures in which P2/ICAM1 had been added to the apical compartments (Fig 2, gray bars, +) compared with those in which P2/ICAM1 conjugates had been omitted (Fig 2, gray bars, −). This effect was observed at each of the analyzed time points. Basophil activation induced by samples taken from the apical compartments (Fig 2, black bars) was comparable and almost identical with that obtained with basolateral samples from cell-free preparations (Fig 2, no cells). Basophil activation was not observed when Phl p 2 was absent from the cultures (Fig 2, no Phl p 2). Concentrations of soluble ICAM1 in 16HBE14o- cell cultures were around 1 ng/mL in apical wells and not detectable in basolateral wells in the experiments and had no influence on allergen-induced basophil activation (data not shown).

Fig E7

Basophils were detected on the basis of side-scatter characteristics and expression of CD203c and analyzed with Paint-a-gate in the FlowJo program. Basophils are defined as CD203c+ cells in all samples. PE, Phycoerythrin; SSC, side scatter.

Fig 2

Inhibition of transepithelial migration of Phl p 2 by P2/ICAM1 conjugates leads to decreased basophil activation with basolateral samples. Samples obtained from apical and basolateral compartments of Transwell cultures were incubated with blood samples from 3 patients allergic to Phl p 2 (A-C), and basophil activation was measured by determining upregulation of the surface marker CD203c on basophils by using flow cytometry. CD203c upregulation expressed as the stimulation index is displayed on the y-axis. Results are means of triplicates, and error bars indicate SDs. *P < .05, **P < .01, and ***P < .001, ANOVA and linear contrasts.

Our experiments thus demonstrate that it is possible to use antibody conjugates bispecific for ICAM1 and a major respiratory allergen to inhibit allergen migration through respiratory cell layers and to reduce the activation of inflammatory cells underneath.

Because increasing numbers of human high-affinity antibodies specific for a large variety of clinically relevant allergens are becoming available, it should be possible to engineer recombinant bispecific antibody constructs for a variety of respiratory allergens that can be combined to protect patients against several different sources for which the major allergens have been identified. The topical application of the conjugates to target organs of allergic inflammation might help prevent allergens from intruding into the tissues and, subsequently, from inducing allergic inflammation and boosting allergen-specific IgE responses. For example, the conjugates can be administered to the nose, eyes, buccal mucosa, and eventually lung to treat respiratory, conjunctival, and oral allergic symptoms.

Taking the finding into consideration that anti–ICAM1 antibodies were shown to inhibit rhinovirus infections and HRV-induced inflammation, one might also speculate that antibody conjugates that are bispecific for ICAM1 and allergens might have a synergistic protective effect in allergic patients with HRV-induced asthma exacerbations.