A Randomized Controlled Trial to Evaluate Inhibition of T-Cell Costimulation in Allergen-induced Airway Inflammation

Abstract

Rationale: T lymphocytes are important in the pathogenesis of allergic asthma. Costimulation through CD28 is critical for optimal activation of T cells, and inhibition of this pathway with CTLA4Ig has been shown to be effective in preventing airway inflammation and hyperresponsiveness in animal models of asthma. Abatacept, a humanized version of CTLA4Ig, has been approved for treatment of rheumatoid arthritis, providing the opportunity to test whether inhibition of costimulation is an effective strategy to treat people with asthma.

Objectives: To determine if 3 months of treatment with abatacept reduced allergen-induced airway inflammation in people with mild atopic asthma.

Methods: Randomized, placebo-controlled, double-blinded study. Bronchoscopically directed segmental allergen challenge was performed on 24 subjects followed by bronchoalveolar lavage 48 hours later. Subjects were randomized 1:1 to receive abatacept or placebo, followed by a second allergen challenge protocol after 3 months of study drug.

Measurements and Main Results: There was no significant reduction in allergen-induced eosinophilic inflammation in the abatacept-treated group compared with placebo (17.71% ± 17.25% vs. 46.39% ± 29.21%; P = 0.26). In addition, we did not detect an effect of abatacept on FEV₁, provocative concentration of methacholine sufficient to induce a 20% decline in FEV₁, or asthma symptoms. Subjects treated with abatacept had an increased percentage of naive and a corresponding decrease in memory CD4⁺ T cells in the blood compared with placebo.

Conclusions: Inhibition of CD28-mediated costimulation with abatacept does not seem to alter the inflammatory response to segmental allergen challenge or clinical measures of asthma symptoms in people with mild atopic asthma.

Clinical trial registered with ClinicalTrials.gov (NCT 00784459).

At a Glance Commentary

Scientific Knowledge on the Subject

Inhibition of CD28-mediated costimulation has been shown to be effective at preventing airway inflammation and hyperresponsiveness in animal models. However, the effectiveness of such a strategy in humans is unknown.

What This Study Adds to the Field

This study in humans with asthma tested the hypothesis that blockade of costimulation would diminish allergen-induced airway inflammation. This study demonstrated that treatment of subjects with mild atopic asthma with abatacept, an inhibitor of CD28-mediated costimulation, had no effect on inflammation induced by segmental allergen challenge, or on clinical measures of asthma symptoms or severity.

Asthma and other allergic diseases are mediated by a complex inflammatory response in which the innate and adaptive immune systems have important roles (1, 2). Exposure to allergen in subjects who are atopic results in IgE-mediated activation of mast cells and basophils with the release of soluble mediators that lead to many of the acute manifestations of allergic disease (3). However, although this may initiate the allergic cascade, these events alone are not sufficient to account for the chronic inflammation characteristic of asthma.

Depending on the combination of signals received, T cells may differentiate down one of several effector lineages. In the presence of IL-4, T cells develop along the Th2 lineage, which are central to allergic inflammation and are capable of secreting IL-4, IL-5, IL-9, and IL-13 (4). These cytokines orchestrate the Th2 inflammatory response, which is characterized by the recruitment of eosinophils. Although the presence and contribution of Th2 cells to airway disease in asthma is well established (5, 6), clinical trials testing agents that block IL-4, IL-5, or IL-13 have had mixed results (reviewed in [7]).

T cells require an ordered series of signals to be fully activated and develop effector function. Binding of the T-cell receptor (TCR) to the antigen–major histocompatibility complex expressed on the surface of an antigen-presenting cell delivers a required but insufficient signal to fully activate the T cell (8). Additional receptors, termed “costimulatory receptors,” deliver essential signals that result in clonal expansion, augmented cytokine secretion, and enhanced cell survival (9). The most critical of these is CD28, which when engaged by its ligands (CD80 or CD86) concomitant with TCR signaling is sufficient to fully activate a resting naive T cell (10). Conversely, inhibition of CD28 prevents T-cell activation in vitro and in vivo (11, 12).

CD28 has been shown to be essential to the development of allergic airway inflammation in a number of preclinical models (13–15). Mice deficient in CD28 fail to develop airway inflammation hyperresponsiveness (16, 17). In addition, treatment of mice with CTLA4Ig, a soluble inhibitor that interferes with the binding of CD28 to CD80/CD86, also prevented in vivo responses to inhaled allergen (18). Interestingly, CTLA4Ig was effective even if administered only at the time of inhaled challenge and not at sensitization, suggesting a mechanism beyond just prevention of T-cell priming (19). Furthermore, ex vivo treatment of bronchial biopsies obtained from atopic people with asthma with CTLA4Ig reduced allergen stimulated secretion of chemokines important in the recruitment of T cells to the lung (20).

Given the preclinical data demonstrating the importance of CD28-mediated costimulation in allergic airway inflammation, we designed a pilot study to test the efficacy of manipulating this pathway in patients with atopic asthma. This is the first trial reported of which we are aware that directly tests whether blockade of costimulation in humans might be an effective therapy for asthma. Mild atopic people with asthma underwent a segmental allergen challenge (SAC) to determine the baseline inflammatory response. Participants were then treated with the humanized version of CTLA4Ig (abatacept) to block CD28 costimulation or placebo in a randomized, double-blinded fashion for 3 months. This drug has been approved for the treatment of rheumatoid arthritis in adults and juvenile idiopathic arthritis in children (21). After 3 months of treatment participants underwent a second SAC to determine if their baseline inflammatory response was altered. The primary endpoint was recruitment of eosinophils to the lung after allergen challenge. In contrast to data from animal models, our data demonstrate that treatment with abatacept had no effect on allergen-induced airway inflammation.

Methods

Patients

Nonsmoking males and females between 18 and 50 years of age with previously diagnosed mild asthma were enrolled (22, 23). Key inclusion criteria included an FEV₁% of greater than or equal to 70; a provocative concentration of methacholine sufficient to induce a 20% decline in FEV₁ (PC₂₀) of less than or equal to 8 mg/ml (≤16 mg/ml if taking inhaled corticosteroids); a history of atopic symptoms; and a positive skin prick test to cat, ragweed, or dust mite allergen. Individuals were excluded if they had any diagnosed lung disease other than allergic asthma, evidence of an upper or lower respiratory tract infection, or chronic use of oral or inhaled corticosteroids at a dose greater than 440 μg/day of fluticasone. Full inclusion and exclusion criteria can be found in the Methods section in the online supplement.

Study Design

The study was a randomized, double-blind, placebo-controlled parallel group trial (Figure 1A). After screening, eligible patients underwent bronchoscopy with SAC to determine the baseline inflammatory response to instilled allergen. Those that increased the percentage of eosinophils recovered in bronchoalveolar lavage (BAL) fluid by at least 50% 48 hours after allergen challenge were eligible for randomization. Randomization was performed in a 1:1 ratio by the research pharmacist who prepared the study medications but otherwise had no involvement in the trial. Participants received abatacept (Orencia, Bristol-Myers Squibb, New York, NY), 10 mg/kg (intravenously), or placebo for 3 months administered at Days 0, 14, 28, 56, and 84. The dose administered was chosen because it is the approved dose for treatment of rheumatoid arthritis. A second SAC was performed on Day 98 using the same lot and dose of allergen as had been used during the initial SAC. Participants were followed until Day 154. All participants were enrolled at University of Wisconsin, Madison (five participants) or Washington University in St. Louis (19 participants). Institutional review board approval was obtained at each site and each participant provided written informed consent.

Figure 1.

(A) Overview of study design. (B) Disposition of participants evaluated for the study. BAL = bronchoalveolar lavage; IV = intravenous; SAC = segmental allergen challenge.

Outcome Measures and Study Endpoints

The primary endpoint was the change in recovery of eosinophils in BAL fluid in response to allergen challenge. Secondary endpoints included changes in total cells recovered into BAL fluid, FEV₁, PC₂₀, changes in asthma control scores as determined by the Asthma Control Questionnaire, changes in exhaled nitric oxide, and changes in serum IgE levels. Exploratory endpoints included changes in specific subsets of inflammatory cells recruited to the lung after allergen challenge as determined by flow cytometry; changes in the cytokine profile of cells isolated from BAL fluid after allergen challenge; and changes in peripheral blood T cells as determined by flow cytometry, proliferation assays, and stimulated cytokine secretion.

Titrated Skin Prick Testing and SAC Protocol

Skin prick testing of cat allergen extract, short ragweed allergen extract (Ambrosia artemisiifolia), and standardized dust mite allergen extracts (Dermatophagoides farinae or Dermatophagoides pteronyssinus) were performed at screening (all glycerinated stocks from Greer Laboratories, Lenoir, NC) and reactivity determined using standard methods (24–26). Titrated skin prick testing was performed using the allergen that induced the strongest positive result to skin prick testing. The most concentrated dilution administered during the titration was a 1:100,000 dilution of the stock allergen. A full description of the titrated skin prick protocol is provided in the online supplement.

Bronchoscopically directed SAC was performed using a modification to the protocol described by Jarjour and coworkers (27) as described in detail in the online supplement. The most dilute solution of allergen that yielded a positive skin test reaction was used to calculate the dose for SAC. A 5-ml solution of the final allergen dilution was instilled bronchoscopically as fully described in the online supplement. At the start of the study, participants were administered a SAC dose 100 times greater than the minimum concentration that provoked a positive result in the titrated skin test. After the first three subjects were enrolled, we modified the protocol to increase the dose to 1,000 times the minimum reactive skin test dose, with a maximum dose being a 1:100 dilution of the stock allergen. This modification was made because the first three subjects had limited eosinophil responses (<10%) in response to the SAC. In all cases, the second SAC was performed at the same allergen dose as the first SAC. All subjects were included in the final data analysis.

Measurement of Lung Function

Spirometry was performed at each visit using standard methods (28). Methacholine bronchoprovocation challenge testing was performed as previously described and the dose required to induce a 20% decrease in FEV₁ (PC₂₀) determined by inhalation of increasing doses of methacholine (0.03125–16 mg/ml) (29).

Analysis of Blood and BAL Fluid

BAL fluid obtained at each bronchoscopy (four bronchoscopies per participant) was analyzed for total cell counts and cell subsets as described briefly below and in detail in the online supplement. Manual differentials were performed on cytospin preparations and immune cell subsets determined by flow cytometry. Peripheral blood was collected before each SAC protocol. Whole blood and cells recovered from BAL fluid were incubated with antibodies directed against the immune cell subsets as indicated in the result tables and described in detail in the online supplement. The markers used to define each cell type are detailed in Table E7 in the online supplement. The cells were then analyzed on a four-color FACSCalibur flow cytometer using CellQuest software (Becton-Dickinson Corporation, Mountainview, CA). Data were further analyzed using Winlist v7 software (Verity Software Corporation, Topsham, ME). Cytokine levels were determined on serum, concentrated BAL fluid, and stimulated culture supernatants by ELISA and T-cell proliferation was determined by tritiated thymidine incorporation as described in the online supplement.

Statistical Analysis

All data were analyzed using SAS version 9.3 (SAS Institute, Cary, NC). Continuous outcomes were presented as the mean ± standard deviation. The comparison of baseline values between placebo and abatacept was done using a two-group t test for continuous outcomes and a chi-square test for categorical outcomes. Comparison of prerandomization, end of treatment intervention, and change from prerandomization to end of treatment values between placebo and abatacept was done using two-group t tests.

The primary analysis was the change between prerandomization and postrandomization BAL fluid eosinophilia. This was done using an analysis of covariance, where the postrandomization outcome was compared between placebo and abatacept groups after adjusting for the prerandomization outcome. A Wilcoxon nonparametric two-group analysis was performed on the differences between prerandomization and postrandomization BAL fluid cytokine level outcomes of IL-4, IL-10, IL-17A, and tumor necrosis factor-α.

Secondary analyses were done for BAL % eosinophils and total cells. For % eosinophils, analyses were performed to compare the level of change in pre- to post-SAC values between placebo and abatacept. This was done using an analysis of covariance, where the post-SAC outcome was compared between placebo and abatacept groups after adjusting for the preallergen BAL % eosinophil. This analysis was done for both SAC1 and SAC2 separately. For total BAL cells, matched pair t tests were done to test for change between SAC1 and SAC2 for placebo and abatacept groups separately. The same approach was used to test for change in total BAL cells between preallergen and postallergen outcomes within each SAC and treatment group separately.

A two-tailed P value less than 0.05 was considered significant.

Results

A total of 70 participants were screened with 24 randomized; two withdrew after the first infusion of study medication because of an anaphylactoid reaction, both of whom were randomized to abatacept (Figure 1B). The remaining 22 participants, 11 in each arm, completed the entire study protocol. Overall, there were no significant differences in the baseline characteristics of the participants (Table 1). The results of the titrated skin testing for participants randomized to abatacept or placebo are shown in Table E8. These are the concentrations used to determine the allergen challenge dose used for the SAC protocol.

TABLE 1.

BASELINE CHARACTERISTICS OF PARTICIPANTS ENROLLED INTO THE TRIAL

	Abatacept (n = 13)	Placebo (n = 11)	P Value
Age (range)	35 ± 9.8 (20–47)	28 ± 7.8 (19–42)	0.08
Sex			0.34
Female	7 (54%)	8 (73%)
Male	6 (46%)	3 (27%)
Race			0.13
White	8 (62%)	5 (45%)
African American	5 (38%)	3 (28%)
Mixed	0	3 (27%)
Inhaled corticosteroids (number receiving)	3	2	0.77
FEV1, L	3.19 ± 0.92	3.05 ± 0.81	0.71
FEV1, % predicted	89.1 ± 12.3	87.4 ± 14.9	0.81
PC20, mg/ml	5.00 ± 10.52	1.46 ± 2.12	0.29
IgE, mg/ml	313.2 ± 339	312 ± 302	0.79
Allergen			0.20
Cat	7 (54%)	3 (27%)
Ragweed	6 (46%)	5 (45%)
Dermatophagoides farina	0	2 (18%)
Dermatophagoides pternyssinus	0	1 (9%)

Definition of abbreviations: PC₂₀ = provocative concentration of methacholine sufficient to induce a 20% decline in FEV₁.

For FEV₁, PC₂₀, and IgE, the data are presented as mean ± standard deviation.

Primary Outcome

The prespecified primary endpoint was the change in percentage of eosinophils recovered from BAL fluid after allergen challenge between the abatacept and placebo groups. We hypothesized that treatment with abatacept would blunt the response to allergen resulting in a decrease in eosinophils recovered in BAL fluid. However, there was no significant difference in the recovery of eosinophils in BAL fluid in participants treated with abatacept compared with placebo (change in eosinophils post-SAC: 17.53% ± 17.27% vs. 44.84% ± 28.16%; P = 0.26) (Figure 2). When assessed using the change in the absolute number of eosinophils recovered instead of percent eosinophils, no difference in response was observed (P = 0.32). The analysis remained not significant when adjusted for race, age, sex, or allergen. Overall, participants randomized to abatacept had a less robust inflammatory response, as determined by total cell counts and cell differential analysis of the postallergen BAL fluid (Figure 3; see Figure E1). However, this diminished response was apparent before and after randomization, and when adjusted for the prerandomization response to allergen challenge, the post randomization inflammatory response to allergen did not differ between treatment arms. In addition, there was no difference in the cellular composition of the preallergen BAL fluid sample between abatacept- or placebo-treated participants at either SAC1 or SAC2 time points (Figure 3; see Figure E3 and data not shown). Thus, our data suggest that abatacept does not alter the inflammatory response in the lung to SAC.

Figure 2.

The percentage of eosinophils recovered in the bronchoalveolar lavage (BAL) fluid from bronchoscopy performed before and after challenge at the start of the study (before randomization) and again at the end of the intervention period (after randomization). Values shown are the mean ± standard deviation, although for clarity of presentation, the overlapping segments of the error bars have been omitted from the figure. SAC = segmental allergen challenge.

Figure 3.

Bronchoalveolar lavage (BAL) cell counts and cellular differential analysis from participants randomized to either placebo (top) or abatacept (bottom) for the preintervention and postintervention segmental allergen challenge (SAC). (Left) Total number of cells recovered in the BAL. (Right) Change in the percentage of each cell type recovered in the BAL after allergen challenge, calculated by subtracting the preallergen value from the postallergen value (postallergen % cell type − preallergen % cell type). P values were calculated by matched paired t test.

Secondary Outcomes

We assessed a number of clinical endpoints to determine if treatment with abatacept affected lung function or asthma symptoms (Table 2). There was no difference in FEV₁ at baseline, nor did treatment with abatacept change the FEV₁ when compared with placebo (P = 0.12). To assess airway hyperresponsiveness, methacholine challenge tests were performed before initiation of study drug and again before the second SAC protocol. Although participants randomized to abatacept were less reactive at baseline, as determined by PC₂₀, compared with those randomized to placebo, this difference was not statistically significant (P = 0.19). Treatment with abatacept did not alter the PC₂₀, suggesting no effect on airway hyperresponsiveness. We also assessed asthma control using the Asthma Control Questionnaire. Both study arms were similar at baseline and did not change over the course of the trial. No difference in serum IgE or exhaled nitric oxide between groups or in response to treatment with abatacept was observed (data not shown). Thus, abatacept did not significantly affect asthma symptoms or indicators of control airway inflammation.

TABLE 2.

CLINICAL ENDPOINT DATA

	Abatacept (n = 11)	Placebo (n = 11)	P Value
FEV1, L
Before randomization (V4)	3.01 ± 0.84	3.03 ± 0.86	0.97
End of treatment intervention (V10)	2.89 ± 0.83	2.99 ± 0.88	0.79
Change (V10 − V4)	−0.12 ± 0.17	−0.04 ± 0.27	0.40
FEV1, % predicted
Before randomization (V4)	85.7 ± 14.9	85.5 ± 10.5	0.97
End of treatment intervention (V10)	82.0 ± 14.6	85.5 ± 9.60	0.51
Change (V10 − V4)	−3.70 ± 5.0	0 ± 5.60	0.12
PC20, mg/ml
Before randomization (V4)	5.87 ± 11.41	1.46 ± 2.12	0.19
End of treatment (V9)	7.06 ± 12.33	1.45 ± 1.80	0.17
Change (V9 − V4)	0.65 ± 4.9	0.09 ± 1.89	0.92
Asthma Control Questionnaire
Before randomization (V4)	0.87 ± 0.50	1.03 ± 0.58	0.51
End of treatment (V9)	0.77 ± 0.61	0.95 ± 0.84	0.57
Change (V9 − V4)	−0.10 ± 0.62	−0.08 ± 0.78	0.94

Definition of abbreviation: PC₂₀ = provocative concentration of methacholine sufficient to induce a 20% decline in FEV₁; V = visit.

FEV₁, PC₂₀, and Asthma Control Questionnaire scores are shown for the 11 participants in each arm of the study who completed the trial. The two subjects who withdrew after the first infusion of study drug are not included in the analysis. Data are presented as the mean ± standard deviation.

Safety

The study protocol was well tolerated with six adverse events reported in those receiving abatacept and 21 in those receiving placebo (see Table E1). Among those receiving abatacept two participants experienced an anaphylactoid reaction during the first infusion of study drug, prompting a change in the protocol to premedicate all participants with cetirizine and acetaminophen before study drug administration. No further reactions were observed. According to the investigators brochure for abatacept, acute infusion-related reactions occur between 0.1% and 1% of patients, with anaphylactoid reactions occurring less frequently. Premedication is not routinely used in clinical practice, raising the possibility that in individuals who are atopic, there may be a higher incidence of infusion reactions. However, our study was not designed to detect this. The remaining adverse events were mild in severity and occurred primarily in the placebo group. Many of the adverse events that were reported are common in an asthmatic population, and although they occurred mostly in the placebo group, are of such low number that no conclusions can be drawn based on these data.

Exploratory Outcomes

To determine if abatacept altered the nature of the inflammatory response to SAC, we performed a comprehensive analysis of the BAL fluid, analyzing cellular composition by flow cytometry (see Table E2) and cytokine content by ELISA (see Table E3). Overall, the change in most cell subsets was similar between abatacept- and placebo-treated participants (see Table E3). Although some of the differences may be statistically significant, the magnitude of the difference between abatacept and placebo are relatively small and confounded by baseline differences observed for the placebo arm. In addition, although we detected higher levels of IL-2, IL-10, and IL-6 in the BAL fluid of the placebo-treated group, these are driven by a few outliers. Thus, these data are unlikely to indicate a meaningful difference between groups. Therefore, we conclude that abatacept had little biologic effect on the local inflammatory response to allergen challenge.

We assessed whether treatment with abatacept affected the percentages of specific immune cell subsets in the peripheral blood (Table 3). Blood was collected at two time points, before allergen challenge at the start of the study treatment period (SAC1) and again at the end (SAC2). The objective of this sample collection and analysis was to determine the effect of abatacept on the circulating immune cell profile. After treatment with abatacept, there was an increase in naive CD4⁺ T cells in the blood accompanied by a decrease in memory CD4⁺ T cells. No change was observed in placebo-treated participants, suggesting this was an effect specific to abatacept treatment. We did not observe any other significant changes in the distribution of cell types. Expression of the T-cell activation marker CD25 was not affected by abatacept treatment, suggesting no effect on the baseline activation status of the circulating T cells. T-cell function was assessed by stimulating peripheral blood mononuclear cells in vitro with either α-CD3 alone, in combination with α-CD28 antibody, or with CTLA4Ig to block any endogenous CD80- or CD86-mediated costimulation (see Table E4). There were no differences in proliferation in response to α-CD3 or α-CD3/α-CD28 stimulation or when stimulated with α-CD3 in the presence of CTLA4Ig. However, the CD80/CD86-dependent proliferation approached statistical significance (P = 0.08) in those treated with abatacept compared with placebo. Cytokines were measured from α-CD3 and α-CD3/CD28 antibody-stimulated cultures at 48 hours. No differences were detected with the exception of IL-17A, which increased in the placebo-treated group but remained unchanged in abatacept-treated over the course of the study intervention.

TABLE 3.

ANALYSIS OF PERIPHERAL BLOOD CELL SUBSETS

	Before Randomization (SAC 1)		After Randomization (SAC 2)
Cell Type %	Placebo	Abatacept	Placebo	Abatacept	P Value
CD4+ T cell	47.3 ± 6.6	44.2 ± 7.3	45.5 ± 15.1	45.3 ± 10.7
Naive	48.8 ± 14.2	37.0 ± 10.3	49.7 ± 14.5	44.2 ± 10.4	0.041
Memory	46.2 ± 13.5	54.7 ± 10.3	45.2 ± 14.1	47.3 ± 11.0	0.013
CD28+CD25+	14.7 ± 8.5	13.9 ± 4.2	15.3 ± 6.1	15.4 ± 5.0
CD28+CD25−	83.9 ± 8.6	83.5 ± 5.3	83.3 ± 6.9	81.6 ± 7.9
CD8+ T cell	26.0 ± 3.7	32.0 ± 6.1	23.7 ± 8.5	28.1 ± 6.1
Naive	71.9 ± 11.4	69.1 ± 12.5	71.8 ± 8.9	69.6 ± 11.9
Memory	21.6 ± 9.5	22.5 ± 9.0	22.0 ± 8.0	22.9 ± 9.8
CD28+CD25+	0.7 ± 0.8	0.7 ± 0.6	0.7 ± 0.7	1.1 ± 0.6	0.08
CD28+CD25−	57.5 ± 17.0	53.0 ± 13.5	61.3 ± 14.2	55.3 ± 9.4
B cell	17.6 ± 5.1	15.8 ± 6.0	15.7 ± 5.5	17.4 ± 6.8
Nk cell	15.7 ± 12.2	19.9 ± 13.3	18.7 ± 16.4	16.1 ± 9.5
Treg	3.2 ± 3.1	3.7 ± 1.5	3.9 ± 2.2	3.1 ± 1.9
MDSC (Lin1−)	17.8 ± 16.7	28.3 ± 20.3	13.5 ± 14.2	13.3 ± 16.3
Myeloid DC (Lin1−)	2.4 ± 2.1	3.0 ± 3.4	3.1 ± 2.8	1.9 ± 1.2
Plasmacytoid DC (Lin1−)	1.4 ± 1.2	1.3 ± 1.0	1.3 ± 1.3	1.1 ± 0.9

Definition of abbreviations: DC = dendritic cell; MDSC = myeloid-derived suppressor cell; SAC = segmental allergen challenge.

Peripheral blood was collected before bronchoscopy at the prerandomization allergen challenge time point and again before the second SAC at the end of the study intervention period (after randomization) and analyzed by flow cytometry for specific cell subsets. Shown are the mean ± standard deviation of each subset as a percentage of the lymphocyte gate. DC and MDSC are shown as a percentage of the lineage cocktail negative gate. P values were calculated comparing whether the postrandomization values were different between placebo and abatacept after correcting for differences present at the prerandomization time point.

Discussion

Data from clinical and animal studies support the central role of T lymphocytes in the pathogenesis of allergic asthma. Therefore, interventions that target T-cell function have been attractive candidates for development as new therapeutics. Many of these agents have been directed at individual T cell–derived cytokines, with clinical trials yielding mixed results (7, 30, 31). Given that multiple effector molecules are involved in the pathogenesis of asthmatic inflammation, a more broad inhibition of T-cell responses to antigen might potentially be an effective therapeutic strategy.

Optimal T-cell expansion and function requires the delivery of costimulatory signals provided by CD28. CTLA4Ig is a soluble fusion protein that binds the ligands for CD28 and is thought to work primarily by preventing receptor engagement and downstream signaling by CD28 (12, 32). The humanized version of this protein, abatacept, was approved for use in the treatment of rheumatoid arthritis and juvenile idiopathic arthritis in 2005 (21). Trials of abatacept have demonstrated efficacy in autoimmune diseases including psoriasis, psoriatic arthritis, systemic lupus erythematosus, and being effective in reducing β-cell loss in early diabetes mellitus (21, 33–36). A modified version of the drug, belatacept, was approved in 2011 to prevent kidney transplant rejection (37).

Given the numerous animal studies that demonstrated CTLA4Ig could potently inhibit allergic airway inflammation (13–15, 17, 19, 38), we tested the efficacy of abatacept in preventing inflammation in response to allergen challenge in humans with mild atopic asthma. In contrast to the results from animal models, we found that administration of abatacept had no effect on the inflammatory response in the lung, nor did we observe a clinical benefit as measured by bronchial hyperreactivity, lung function, or asthma control. One possible explanation for the lack of efficacy is that the drug did not exert its intended biologic effect, perhaps because of an inadequate dosing regimen. However, the dose used was the same as that used for several treatment trials of inflammatory diseases (33–36, 39, 40) and is the approved dose for rheumatoid arthritis (http://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=0836c6ac-ee37-5640-2fed-a3185a0b16eb). An additional possibility is that the small sample size studied may have precluded our ability to detect a treatment effect. A post hoc analysis of the data demonstrated that the study was 80% powered to detect a between-group difference of 29.4% in the % BAL eosinophils at a 0.05 level of significance. Thus, although we may have missed a lesser effect, it seems unlikely that a smaller reduction in eosinophil recruitment would be clinically meaningful.

CD28 costimulation markedly enhances T-cell proliferation and cytokine secretion in vitro, and blockade with CTLA4Ig diminishes both responses (11, 12, 41, 42). Therefore, we predicted that participants who had received abatacept would have reduced cytokine levels in BAL fluid, and that T-cell function would be impaired when peripheral blood mononuclear cells were ex vivo stimulated. However, neither effect was observed suggesting that in vivo treatment with abatacept did not substantially modulate these aspects of T-cell function. Although most of our exploratory endpoints were negative, we did observe a change in the percentage of memory and naive CD4⁺ T cells detected in the blood. In addition, we observed a dampening of CD80/CD86-dependent proliferative responses. Although no assay has been developed for either clinical or research use to determine whether abatacept has exerted its intended biologic effect in vivo, these data would suggest that abatacept did in fact exert a clinically detectable effect.

An additional consideration is recent data that suggest alternative mechanisms of action for abatacept. CTLA4Ig binding to CD80/CD86 on dendritic cells has been shown to induce indoleamine 2,3-dioxygenase, an immunoregulatory enzyme that can inhibit T-cell function through a number of pathways, including alteration of dendritic cell function and promotion of Treg development (43–46). However, we did not detect changes in Treg number in the BAL or blood, although these data do not exclude the possibility of changes in Treg function. Recently we demonstrated in a murine model that CTLA4Ig can inhibit airway inflammation independent of CD28 by activation of nitric oxide synthase in macrophages (38). This pathway is independent of indoleamine 2,3-dioxygenase and distinct from the conventional mechanism by which abatacept blocks engagement of CD28 on the T cell by CD80/CD86 expressed on the antigen-presenting cell. The parameters we measured would not detect alterations in this pathway.

There are limitations to this study. The SAC model provides a very strong stimulus that activates numerous effector pathways of the innate and adaptive immune system. All of the participants were atopic and the T-cell responses would be expected to be predominantly memory responses, which are less dependent on CD28-mediated costimulation than activation of naive T cells (47–49). Thus, the intensity of the provoked inflammatory stimulus may have overridden any requirement for T-cell costimulation. Patients with asthma may in reality experience a more chronic low level of exposure and airway inflammation, punctuated by intermittent severe exacerbations. It is certainly possible that long-term treatment with abatacept may modulate the underlying chronic inflammatory response favorably in a more severely affected asthma population, perhaps resulting in decreased exacerbations and improving asthma control. Such an effect would not have been detected in the present study but the negligible differences after 3 months of therapy suggest this is unlikely.

In conclusion, the current study demonstrated that treatment with abatacept did not significantly alter the inflammatory response to SAC in people with mild atopic asthma. Thus, modulation of CD28-mediated costimulation is unlikely to be an effective strategy in manipulating the T-cell response for therapeutic benefit in allergic inflammation.