Introduction

Allergic rhinitis (AR) is an IgE-mediated nasal disorder involving inflammation and hyperactive nasal mucosa, resulting in symptoms of rhinorrhea, sneezing, nasal pruritus, congestion, and aggravation of comorbid asthma [1•]. The prevalence of AR is increasing, currently affecting approximately 10–25 % of the population worldwide [1•]. In recent years, the use of controlled allergen challenge facilities (CACFs) (also referred to as “exposure units” or “challenge chambers”) has contributed to our understanding of the pathophysiology of AR and pharmacological properties, efficacy, and onset of action of new therapies. Treatments investigated using CACF models include antihistamines, leukotriene modifiers, corticosteroids, and immunotherapy [2••]. Herein, we will review advances made in basic and clinical allergy research through the use of CACF models. Studies published in abstract form only are not discussed, due to space constraints.

Brief Historical Context

The first multi-participant CACF was the Vienna Challenge Chamber (VCC; Vienna, Austria), described by Horak and Jäger in 1987 [2••, 3••]. Shortly thereafter, Day and Clark adapted their specially engineered particulate distribution technology (originally used to study urea formaldehyde in 1981), for pollen distribution, creating the Environmental Exposure Unit (EEU) at Kingston General Hospital, Kingston, Ontario, Canada [4–6]. In Europe, Rønborg et al. constructed an allergen exposure facility in Copenhagen, Denmark, and validated it for use with house-dust mite allergen [7•]. The Fraunhofer Institute Environmental Challenge Chamber (ECC) was established in Hannover, Germany in 2003 [8••]. Across North America, facilities have emerged, including the Atlanta Exposure Unit [9–11], the Environmental Exposure Chamber (EEC) [12, 13], the Biogenics Research Chamber [14, 15], and the Allergen BioCube [16]. In Japan, the Environmental Challenge Chamber at Chiba University was built in 2008 [17•], followed by another allergen challenge chamber in Osaka [18]. An Allergen Challenge Theatre has been described in Ottawa, Ontario, but only in abstract form [19]. These facilities were uniquely and independently developed, and employ various technologies to achieve allergen distribution, monitoring, and air quality control/assessment.

Brief Technical Overview and Considerations

CACFs are custom-designed rooms that house study participants in a controlled environment, in which exposure to consistent allergen levels can take place and symptoms can be monitored [2••]. Studies conducted to date have been monocentric, and there are technical differences between units, the most palpable being the method of allergen dispersion and allergen concentration monitoring [2••, 3••].

In the VCC, clean air and a second air circuit carrying allergen-loaded air, enter through the ceiling and are exhausted at floor level [2••, 3••]. Modified Burkard pollen traps are employed to monitor allergen levels [2••, 3••]. The EEU employs a laser-aided pollen dispersion system and directional fans to propel the pollen over the participant seating area [6]. Rotorod samplers, distributed at seven locations, collect pollen at 30-min intervals, which correspond to the participant symptom scoring times [6]. The EEU employs slightly negative pressure to minimize pollen loss to the rest of the hospital, and the ventilation system uses high efficiency filtered outdoor air, with temperature and humidity control, exhausted directly back into the outdoors [6]. The Atlanta Allergen Exposure Unit similarly employs filtered outdoor air and Rotorod samplers [2••, 9].

At the Fraunhofer Institute, HEPA-filtered air enters via swirl inlets and thermal convection from floor heating aids in mixing [8••]. Pollen is introduced via a feeding system and a pneumatic dispersion nozzle located on the floor above the unit [8••]. Pollen concentrations are monitored by a laser particle counter, and every 30 min by two Rotorod samplers [8••]. Pollen in the Chiba environmental challenge chamber is also introduced from reservoirs above the ceiling, where the particles are agitated by fans to fall and drop down through holes into the participant seating area [17•]. Levels are monitored at 56 points using automatic pollen counters, including one on the back of each participants’ chair [17•]. Although the automated counters are not pollen-specific, they have been validated against Durham samplers [20]. Other designs exist; however, the facilities above have published the largest number of manuscripts detailing technical operations. While individual CACFs exhibit differences, it is generally agreed that well-documented, uniform concentration of allergen, selected to be relevant to levels found outdoors, is the most important constant in CACF systems [2••], and a lack of such control would generate concerns when interpreting the study outcomes.

Advantages of the CACF Model

CACFs offer advantages over traditional in-season multicentre trials when evaluating new AR treatments. Logistical problems are presented by traditional studies that are scheduled to coincide with peak pollen season, as levels can be influenced by weather, highly variable between sites and consecutive seasons, significantly impacting the apparent efficacy of treatment [21]. Furthermore, individual participants’ exposures are affected by lifestyle [22], adherence to scheduled administration of study medication, and symptom recording cannot always be ensured [6, 23]. CACFs were developed to address these variables, while recapitulating many characteristics of an outdoor allergy study [2••, 6, 23]. Although to date, the burden of proof for determining the efficacy of new treatments remains with in-season outpatient trials, expert reports, the European Medicines Agency and the proceedings of an Advisory Committee of the US Food and Drug Administration (FDA) Meeting underscore the utility of CACFs, and perhaps they will contribute to the regulatory determination of efficacy in the future [24–27].

Reproducibility of CACF Findings

Reproducibility of findings between studies is an important consideration in clinical research. If conditions are tightly controlled, repetition of the same protocol on separate occasions should yield the same results. Indeed, both the VCC and the EEU have demonstrated this capability. Consecutive studies in the VCC determined the efficacy of desloratadine on nasal congestion and demonstrated similar results [28, 29]. In the EEU, two studies, separated by approximately 4 years, investigated the effects of cetirizine or loratadine vs. placebo in the treatment of seasonal AR, and yielded almost identical comparative efficacy results [30, 31]. Thus, placebo and medication responses are highly reproducible in CACF studies.

Priming and Other Factors That Affect the AR Symptom Response

Despite the reproducibility of CACF clinical findings, there are known factors that can affect an individual’s symptoms upon exposure. The priming effect is defined as an increase in reactivity of the nasal membrane following repeated exposure to allergen [32••]. This well-recognized effect of repeated allergen exposure in CACFs has become somewhat of an advantage of the model. In CACF clinical trials, priming sessions are often used to increase sensitivity to a specific allergen by inducing mucosal inflammation, facilitating the development of robust symptoms and boosting participant inclusion [6]. In the EEU, priming has been shown to standardize pre-dose symptoms and greatly reduce symptom variability [6, 23]. In the Osaka allergen challenge chamber, a single challenge to Japanese cedar pollen has been proven to induce robust symptoms at the end of pollen season, while three consecutive challenges are required out of season [33].

Important individual factors that affect EEU studies include sensitization to dust mite, dog, or grass, current exposure to dog or cat, and rhinitis-related quality of life [34•]. Thus, in the EEU, perennial allergens have been demonstrated to be confounders under certain conditions. Recently, Jacobs et al. showed that pollen sensitization can conversely become a confounder when perennial allergens are being studied, demonstrating that pollen allergy affects symptom development to house dust mite [35]. Quality of life at study entry also represents an important consideration, affecting response to placebo [36]. As CACF studies, by their nature, induce disease symptoms, participation in CACF trials can, in turn, have effects on an AR participants’ quality of life [37•]. While the above factors, related to the individual research participant, appear to be separate from induced priming, they may collectively reflect the importance of endogenous priming in the participant’s day-to-day life.

Effects of, and Comparison to, the Natural Pollen Season

Due to the effects of natural exposure on participants’ symptom responses, traditionally, CACF studies have been conducted outside of the relevant pollen season. However, scientists from the Fraunhofer Institute demonstrated an identical treatment effect of cetirizine/pseudoephedrine, compared with placebo, both within and outside of the pollen season [38••]. The CACF model also offers better reproducibility and sensitivity/specificity of total nasal symptom score (TNSS), compared to two measures during one natural pollen season [39]. In the Biogenics Research Chamber, concordance between induced allergic symptoms and those experienced during the natural season has been shown [40•]. Multiple allergic sensitizations have also been shown to result in “pre-priming” during the natural season [41]. Therefore, in general, CACF models achieve similar symptoms to those experienced by AR sufferers during the natural pollen season, but pre-priming must be taken into consideration, as they can occasionally affect results.

CACF models are also capable of recapitulating different phenotypic responses to allergen exposure observed in the “real world”. Recently, it has been shown that participants’ AR responses can be clearly phenotyped into isolated early phase responders (EPR), dual responders (early and late phase), as well as an intermediate protracted EPR phenotype, using the EEU [42]. Significant differences were evident between EPR, protracted EPR, and dual responders beginning 3 and 8 h after a 3 h ragweed pollen exposure, respectively [42]. This study demonstrated that the protocols and technology employed in CACF models allow for the phenotyping of participants based on AR symptoms during and after pollen exposure [42].

Oral Antihistamines

Over the past decade, CACF models have contributed extensively to our understanding of the efficacy, onset of action, and duration of action of various antihistamines. Traditionally, antihistamines have been thought of as H1 receptor antagonists; however, it has recently been demonstrated that they may act as inverse agonists, stabilizing inactive forms of the H1 and H2 receptors [43–45]. Whether this is essential or clinically important for antihistamines has not yet been clarified [46]. The second-generation antihistamines terfenadine, astemizole, cetirizine, and loratadine and their efficacy against AR symptoms were investigated in the EEU [47•]. Through these studies, cetirizine and terfenadine continuously ranked higher than loratadine and astemizole in terms of efficacy and onset of action [47•]. This was followed by a study with increased participant numbers that confirmed that cetirizine significantly reduces symptoms compared to both loratadine and placebo, with an onset of action of 60 min [31]. In the VCC, levocetirizine was compared to loratadine in seasonal AR (challenge to grass pollen) and perennial AR (challenge to house dust mite) [48]. This study demonstrated superiority of levocetirizine in improving seasonal AR symptoms, and a trend towards the same  in perennial AR [48]. Levocetirizine was compared to desloratadine in the EEU, revealing that while both were effective, levocetirizine produced a greater improvement in symptom scores during sessions on two consecutive days [49]. Levocetirizine was also examined in the VCC, with consistent results, revealing a longer duration of action than fexofenadine [50]. Fexofenadine was compared to placebo in the EEU and provided clinically important relief with onset of action at 60 min [51]. Cetirizine and fexofenadine were further examined in the EEU, demonstrating a longer duration of action of cetirizine [52, 53•]. Comparative studies continued in the VCC as well, bilastine and cetirizine were found to exhibit longer durations of action, compared to fexofenadine [54•].

Prophylactic experimental designs have also been employed. In the Chiba environmental challenge chamber, participants were exposed to Japanese cedar pollen after a single administration of levocetirizine, levocetirizine for 8 days, or placebo [55•]. Symptoms were lower in both treated groups, compared to placebo, but prophylactic treatment was not superior to single treatment [55•]. This study illustrated an important feature of the fundamental efficacy characteristics of antihistamines. Namely, that there is no added benefit to taking antihistamines continuously, unless exposure to allergen is continuous. Finally, newer antihistamines, with additional pharmacologic properties, have also been examined using CACF models. Rupatadine, a second generation antihistamine and platelet-activating factor antagonist was investigated for prophylactic utility in the VCC [56]. Subjective symptoms and mean secretion weights were significantly lower [56], demonstrating the efficacy of rupatadine, compared to placebo.

Intranasal Antihistamine and/or Steroid Formulations

Intranasal antihistamines, steroids, and combination therapies, and the importance of route of administration, through comparison to oral treatments, have been examined in CACF studies. Recently, the EEU was used to examine azelastine nasal spray vs. orally administered cetirizine, loratadine, or placebo [57]. Azelastine nasal spray significantly reduced nasal symptoms, compared with placebo, and exhibited a faster onset of action compared to oral treatments [57]. An assessment of the onset and duration of action of olopatadine nasal spray was conducted in the EEC by Patel et al. [58]. Olopatadine was significantly more effective than placebo at all time-points starting at 90 min and continuing over 12 h [58]. Azelastine nasal spray was compared to desloratadine tablets in the VCC [59]. While desloratadine significantly improved symptom scores, azelastine nasal spray was superior [59]. Azelastine nasal spray exhibited an onset of action of 15 min, consistent with the onset of action determined in the EEU [59]. Thus, CACF studies have demonstrated consistent results across sites, both in terms of the improvement over systemic treatment, and the onset of action.

A wide variety of intranasal corticosteroids have been approved for treatment of AR, and many have been examined in CACF models. In 1996, the first evaluation of a nasal corticosteroid was performed in the EEU, and revealed a 10 h onset of action of triamcinolone acetonide, which was substantially shorter than expected, as it was previously thought that nasal corticosteroids required several days of administration for efficacy [60••]. A 7 h onset of action was subsequently determined for budesonide in the EEU [61], and a 6 h onset of action for ciclesonide (200mcg) was determined in Mississauga, Ontario [62]. Another assessment of ciclesonide by the same group found an onset of action of 1 h [63], a sizeable difference that is uncharacteristic of a CACF model study.

Another intranasal steroid, fluticasone furoate, was studied at the VCC [64]. Participants were exposed to grass pollen following 8 days of treatment, demonstrating significant reduction in symptoms compared to placebo [64]. Mometasone was compared to azelastine and placebo using a CACF model [12]. Azelastine showed greater reduction in symptoms, with a rapid onset of action starting at 15 min [12]. A subsequent study using 200 mcg mometasone furoate showed a significant difference between ragweed allergic participants who received the treatment and those who were administered the placebo once daily for 8 days [65]. The onset of action was estimated at 6 h, and its duration of action was found to exceed 24 h [65].

CACFs were also used to compare the efficacy of the intranasal anti-histamine olopatadine to mometasone (50 mcg) in ragweed allergic participants [66]. Olopatadine had an earlier onset of action at 30 min compared to 2.5 h [66]. Olopatadine demonstrated a reduction in TNSS for the duration of the study, up to 12 h post administration of the single dose [66]. At Chiba University, cedar pollen allergic individuals were randomized to receive either fexofenadine or mometasone for 7 days to evaluate prophylaxis [67]. Although TNSS in both groups were not different following 3 h of pollen exposure, a difference was present on days 8 to 11 from the start of treatment, suggesting that intranasal steroids had a prolonged anti-inflammatory effect, compared to the antihistamine [67].

A study comparing the solubilized form with the suspended form of budesonide, in combination with azelastine, found all active treatments demonstrated significant decreases in TNSS, compared to placebo, but the solubilized forms offered greater relief and demonstrated a faster onset of action [68]. Thus, CACF models have been instrumental in demonstrating the efficacy of intranasal steroids, their relative efficacy compared to medications targeting other pathways, determining their onset of action, and exploring characteristics of different formulations.

Decongestants and Combination Therapies

While antihistamines are consistently effective against a range of AR symptoms, they are notably lacking in efficacy against congestion, one of the most bothersome symptoms of AR [69, 70, 71••]. As congestion may be mediated by dilatation of venous capacitance vessels, sympathomimetic agents such as pseudoephedrine are effective decongestants [72–74]. Thus, combination treatments have been developed, including formulations of pseudoephedrine plus antihistamines. The congestion-specific efficacy of an oral formulation of cetirizine/pseudoephedrine was compared to budesonide nasal spray in the VCC [75]. Rhinomanometry demonstrated superior efficacy of the combination therapy for nasal congestion [75]. A further study employing the VCC compared the decongestant properties of phenylephrine and pseudoephedrine [76•]. They found that phenylephrine was not significantly different from placebo in terms of participant-scored nasal congestion, whereas pseudoephedrine was [76•]. At approximately the same time, phenylephrine was investigated in the EEU and compared to loratadine-montelukast and placebo [77]. Similarly, phenylephrine did not significantly reduce symptoms, relative to placebo, while loratadine-montelukast was effective [77]. A further CACF study compared the effect sizes of antihistamine vs. pseudoephedrine alone and in combination, relative to placebo [78•]. Participants underwent four 6 h pollen exposures at the Fraunhofer Institute, with administration of drug after 2 h, in a double-blind, four-way crossover [78•]. Nasal obstruction was significantly lower after treatment with cetirizine/pseudoephedrine, compared to either treatment alone or placebo [78•]. Thus, CACF models played a significant role in establishing the efficacy of combination therapies and contributed to the discussion surrounding the level of efficacy of phenylephrine.

Combinations of antihistamines with sympathomimetic agents have been proven effective against congestion; however, due to side effects such as insomnia and hypertension, they are contraindicated in those with cardiovascular problems [73, 74, 79, 80]. H3 receptor antagonists as decongestants were suggested as potentially safer alternatives. At the Fraunhofer Institute, dual H1/H3 receptor antagonists significantly attenuated nasal symptoms and blockage, relative to placebo, but were not superior to cetirizine [81]. Barchuk et al. investigated a specific H3 receptor antagonist, compared to placebo, and pseudoephedrine as an active control [82]. Hourly minimal cross-sectional area measurements revealed less of a decrease in nasal patency for the H3 receptor antagonist versus placebo, which was of borderline significance (p = 0.06) [82]. The same measure for pseudoephedrine was statistically significant [82]. However, the H3 receptor antagonist showed superior efficacy compared to pseudoephedrine when the baseline-adjusted area under the curve of participant-scored congestion was used [82]. Recently, a placebo-controlled crossover study of a specific H3 receptor antagonist with or without fexofenadine was conducted in the EEU and compared to pseudoephedrine/fexofenadine as an active control [80]. Although the combination of H3 antagonist/fexofenadine significantly reduced TNSS, relative to placebo, with an onset of action of 60 min, the treatment was not superior to pseudoephedrine/fexofenadine [80]. Additionally, an elevated incidence of adverse events was noted, potentially related to the role of H3 receptors in the central-nervous system [80, 83, 84]. Further development of H3 receptor antagonists with lesser penetrance into the central nervous system may exhibit a more favorable safety profile, or even increased efficacy against allergic rhinitis symptoms [80].

Antileukotrienes

CACF models have similarly provided key information regarding the efficacy and pharmacodynamics of antileukotrienes [85]. Montelukast was compared to levocetirizine in the EEU with ragweed allergic participants. Although both medications reduced symptoms, levocetirizine was significantly more effective 24 h after the first dose and 4.5 h after the second dose [86•]. A similar study found participants treated with levocetirizine experienced significantly lower symptom scores compared to placebo and montelukast [87]. The onset of action of levocetirizine was estimated at 2.5 h, whereas montelukast could not be evaluated since it did not achieve significant efficacy [87].

Combination therapies involving antileukotrienes have also been evaluated using CACF models. The onset of action of an oral combination of loratadine/montelukast was evaluated in the EEU [88]. This study revealed an onset of action of 75 min against total symptom scores, nasal congestion, and peak nasal inspiratory flow [88]. A follow-up study found loratadine/montelukast was superior to phenylephrine and placebo in reducing nasal congestion, total nasal symptoms, and non-nasal symptoms [77]. Another study at the EEU determined the onset of action of the same dose of loratadine/montelukast to be 75 min [88]. A similar study conducted at the VCC also investigated a combination tablet of loratadine/montelukast [89]. They found the onset of action to be 105 min, and significantly improved nasal congestion symptoms and rhinomanometry compared to placebo [89]. Thus, CACF models have been instrumental in determining efficacy and pharmacodynamic features of antileukotrienes, and a close agreement has been demonstrated between facilities.

Immunotherapy

Traditional and newer forms of immunotherapy have been tested for efficacy in CACF models. In 1996, the efficacy of two years of traditional ragweed immunotherapy was demonstrated in the EEU, and symptom scores of recipients were significantly reduced compared to those of non-treated controls after 45 min of pollen exposure [90••]. More recently, CACFs have been employed to contribute to the development of novel immunotherapy medications in phase 2 and 3 clinical trials [91]. In one phase 2b study, participants were treated with four weekly subcutaneous injections of short ragweed pollen allergoid [92]. The vaccine was found to significantly improve symptoms and rhinoconjunctivitis-related quality of life, relative to placebo [92]. In a similar study of an intradermally administered synthetic T-cell epitope vaccine for cat allergy, participants were challenged with cat allergen in a CACF prior to treatment, 18–22 weeks from start of study, and again at 50–54 weeks [93]. Efficacy was demonstrated during the 18–22 week and 1 year follow-up exposures [93]. A sub-lingual immunotherapy study conducted at the VCC evaluated a grass pollen tablet, and participants demonstrated a significant improvement of symptom scores, relative to placebo, during pollen challenges beginning at the first month, and maintained to 2 and 4 months [94]. Thus, CACF models have been useful for determining the efficacy, onset and mechanisms of action of new immunotherapy treatments over the past decade.

Basic Science and Novel/Experimental Therapies

Finally, as CACF models provide a sensitive, specific, and reproducible methodology for allergen challenge, they are also useful for the investigation of basic mechanisms of AR and novel therapies. Researchers at the VCC examined circulating blood cell dynamics in an animal model and in human allergic participants exposed to grass pollen in the same study [95]. A significant drop in erythrocyte counts were found both in the murine model and human participants [95]. This represents a novel basic mechanism that may contribute to the development of AR symptoms. However, the authors did not find agreement in the effects on peripheral leukocyte counts in mice and humans [95].

Many novel and experimental therapies that elucidate basic mechanisms of AR have been investigated in CACF models. Researchers employed the VCC to examine the role of the chemoattractant receptor homologous molecule (CRTH2), which mediates activation of Th2 cells, eosinophils and basophils in response to prostaglandin D(2) [96]. A CRTH2 antagonist or placebo was administered twice daily for 8 days, and participants were exposed to grass pollen on the 2nd and 8th days of treatment [96]. A crossover arm demonstrated a significant effect on the 2nd and 8th days, with some persistence of effects despite the 3-week washout period [96]. Scientists at the Fraunhofer Institute conducted a follow-up study with a different CRTH2 antagonist administered orally at three doses (50, 200, and 400 mg twice daily) [97]. Fluticasone propionate nasal spray (200 μg once daily) and oral montelukast (10 mg once daily) were both employed as active controls and all treatments or placebo were administered for 2 weeks prior to orchard grass exposure [97]. Mean TNSS was significantly reduced versus placebo with 200 mg of CRTH2 antagonist, montelukast and fluticasone propionate, relative to placebo [97]. These studies obtained similar results across sites despite the use of different CRTH2 antagonists, and collectively suggest that CRTH2 antagonists may represent novel AR therapies.

Another novel pathway that has recently been explored is the ion channel transient receptor potential vanilloid 1 (TRPV1) [98]. An intranasal formulation of TRPV1 antagonist was examined in the VCC after 8 days of treatment, 8 days of treatment in combination with fluticasone, fluticasone alone, or placebo [98]. A single 4-h allergen exposure was carried out on day 8, revealing no differences in mean TNSS between TRPV1 antagonist alone and placebo, or between TRPV1 antagonist plus fluticasone and fluticasone alone [98]. This study provides an example of how CACF models can help target effective molecules for further development and bring to light cases where despite promising preclinical data, there is a lack of translation to clinical efficacy, before more costly studies are undertaken.

Conclusions

Controlled allergen challenge facilities provide a unique model system, which overcomes certain challenges evident in standard phase 3 efficacy trials by controlling additional study variables. Symptoms generated in CACF studies compare well to the natural season, demonstrate consistent priming characteristics and the ability to phenotype AR participants. Over the past decade, CACF studies have made important contributions to our understanding of the efficacy, onset of action, and other pharmacodynamic characteristics of AR treatments such as antihistamines, antileukotrienes, immunotherapy, and nasal steroids. CACF models are also well suited to basic science research, owing to their sensitive, specific, and reproducible methodology, and will continue to contribute substantially to the medical literature as newer therapies for the treatment of allergic rhinitis emerge.