β5i Subunit Deficiency of the Immunoproteasome Leads to Reduced Th2 Response in OVA Induced Acute Asthma

The immunoproteasome subunit β5i has been shown to play an important role in Th1/Th17 driven models of colitis and arthritis. However, the function of β5i in Th2 dependent diseases remains enigmatic. To study the role of β5i in Th2-driven pathology, β5i knockout (KO) and control mice were tested in different models of experimental allergic asthma. β5i-deficient mice showed reduced OVA/Alum- and subcutaneous/OVA-induced acute asthma with decreased eosinophilia in the bronchoalveolar lavage (BAL), low OVA-specific IgG1 and reduced local and systemic Th2 cytokines. While Th2 cells in the lungs were reduced, Tregs and Th1 cells were not affected. Attenuated asthma in β5i KO mice could not be attributed to defects in OVA uptake or maturation of dendritic cells in the lung. Surprisingly, β5i deficient mice developed HDM asthma which was comparable to control mice. Here, we present novel evidence for the requirement of the β5i immunosubunit to generate a strong Th2 response during OVA- but not HDM-induced acute asthma. The unexpected role of β5i in OVA asthma remains to be clarified.


Introduction
Allergic asthma is a type I hypersensitivity reaction of the upper and lower respiratory tract which affects more than 300 million people worldwide and therefore, represents one of the most common diseases. It is characterized by recurrent episodes of wheezing, breathlessness, chest tightness and coughing [1]. These manifestations are caused by airway hyper-responsiveness, mucus hypersecretion, lung eosinophilia and accumulation of Th2 cells [2,3]. Data suggest that in humans, inflammatory reactions of severe asthma are driven by NF-kB [4]. Similarly, in the rat model of OVA-induced allergic asthma, inhibition of NF-kB activation reduced airway eosinophilia and secretion of IL-1a, eotaxin, TNFa, IL-4, IL-5 and IL-13 [5]. Processing and degradation of many proteins involved in signaling including NF-kB are mediated by proteasomes. The mammalian proteasome is a multicatalytic protease complex which is involved in protein degradation and processing of MHC I class antigens [6]. The 20S proteasome contains three different catalytic activities: caspase-like (b1 subunits), trypsin-like (b2 subunits) and chymotrypsin-like (b5 subunits) [7]. Stimulation with IFN-c or IFN-a induces replacement of constitutive catalytic subunits by the corresponding immunosubunits, b1i, b2i and b5i leading to the formation of immunoproteasomes [8]. Immunoproteasomes show altered cleavage site specificity with improved MHC-I antigen presentation [9,10,11] and increased activation of NF-kB [12,13].
It has been shown that partial proteasome inhibition by bortezomib or b5i subunit deletion markedly attenuates DSS induced colitis [14,15] and selective inhibition of b5i by the inhibitor PR-957 ameliorates development of collagen-induced arthritis and decreases pro-inflammatory IL-6, IL-23 and TNF-a cytokines production by peripheral blood mononuclear cells [16]. Surprisingly, the knowledge on proteasome involvement in Th2driven diseases is scarce. Application of non-specific proteasome inhibitors PS-519 [17] or bortezomib [18] in OVA-induced acute asthma models significantly reduces airway eosinophilia. Isolated eosinophils pretreated with proteasome inhibitor MG-132 show reduced expression of integrin beta-2 (CD18) or stop secreting IL-8 and initiate apoptosis in response to TNF-a [19].
We thus wondered whether specific inhibition of the b5i immunosubunit is an effective strategy for the treatment of acute asthma. Here, we present novel evidence that b5i immunosubunit deletion abrogates OVA-induced acute asthma, as shown by reduced eosinophilia and Th2 responses. However, lack of the b5i immunosubunit did not show any beneficial effects in the clinically relevant HDM asthma model.

Results
Reduction of the asthmatic phenotype in b5i KO mice in experimental OVA/Alum model To investigate the role of b5i subunit of the immunoproteasome, C57BL/6N (Wt) and b5i KO mice were sensitized and challenged either with PBS/OVA or OVA/OVA as shown in Figure 1A. BAL fluid was collected and differential cell counting was performed as described in the materials and methods section. In comparison to Wt mice, b5i KO mice demonstrated reduced recruitment of inflammatory cells as shown by the lower number of eosinophils, lymphocytes and neutrophils in the BAL ( Figure 1B). Further, OVA specific IgG1 titers (Figure1C) and the Th2 cytokines IL-4 and IL-13 produced by BALs were significantly decreased in b5i KO mice as compared to Wt controls ( Figure 1D). Additionally, lung Th2 and Treg cells profile of the Wt and b5i KO mice was assessed by flow cytometry from the homogenized lungs. We found that b5i KO mice had decreased Th2 cells percentage (0.860.35% vs 0.460.05%, +P,0.05) in comparison to the Wt during the asthma (Figure 2, Figure S1). Despite increased percentage of lung Tregs during asthma, no difference between the Wt and b5i KO was demonstrated ( Figure 2, Figure S2). Similarly, no significant differences in Th1 lung cells content was observed, although a slight decrease in Th1 cells in the Wt asthma group was detectable (data not shown). To further investigate whether the observed reduction of pulmonary Th2 responses in b5i KO mice is systemic or restricted to the lung, splenocytes from OVA-asthma induced Wt and b5i KO mice were in-vitro restimulated with OVA and cytokines were measured in the culture supernatants. Unexpectedly, splenocytes of OVA sensitized and challenged b5i KO mice produced less IL-4, IL-5 and IL-13 as compared to Wt mice ( Figure 3).

Effects of b5i KO during OVA acute asthma are independent of Alum
It is known that Alum affects dendritic cell activation and consequent antigen presentation [20]. To exclude any influence of alum on the asthmatic phenotype, we employed our wellestablished, Alum-free subcutaneous OVA-induced asthma model ( Figure 4A) [21]. Similarly to OVA/Alum asthma, b5i KO mice demonstrated a significant reduction in the allergic inflammation as compared to Wt mice. The frequency of eosinophils, lymphocytes and neutrophils in the BAL was significantly diminished in the mutant mice ( Figure 4B). Analysis of lung Tcells revealed comparable reduction of Th2 cells in b5i KO mice ( Figure 4C), while Tregs were equally increased in both the Wtand b5i KO mice ( Figure 4C), which is comparable to the OVA/ Alum asthma ( Figure 2).

Reduced FITC-OVA uptake by lung inflammatory dendritic cells (iDCs) in b5i KO
The b5i subunit of the immunoproteasome plays an important role in NF-kB activation [12,13] and thus promotes maturation of dendritic cells [22]. To investigate whether reduced Th2 response during OVA/Alum asthma in b5i KO mice is due to impairment in DC maturation, mice were challenged intranasally with a FITC-OVA conjugate and FITC-OVA positive iDCs (F4/ 802CD11c+CD11b+Ly6C+B2202, gating strategy in Figure S3) were analyzed on the following day. Compared to WT mice, iDCs from b5i KO mice revealed significantly decreased FITC-OVA uptake and reduced maturation as shown by low expression of MHC II molecules ( Figure 5). Similar analysis of conventional lung DCs (cDCs) in b5i KO mice revealed no differences between WT mice (data not shown). Further analysis of the FITC-OVA iDCs migration to the mediastinal lymph nodes showed no difference in the distribution between the Wt and b5i KO mice (data not shown), suggesting that an intrinsic defect of iDCs in b5i KO mice might contribute to the observed phenotype.

Th2 response is not restored by transfer of normal dendritic cells into b5i KO mice
To study whether Th2 driven responses can be rescued in b5i KO mice, we attempted to restore iDCs deficiencies of b5i KO mice by adoptive transfers of in vitro generated Wt BMDCs ( Figure 6A). Although transfer of BMDCs resulted in increased numbers of iDCs in the lung of b5i KO and control mice ( Figure  S4A), Wt but not b5i KO mice showed increased allergic inflammation upon OVA-challenge ( Figure 6B, C). In order to compare the immune reactivity of OVA-asthma mice with or without transfer of DCs, ratios for differential BAL cell counts, Th2 cells and Th2 cytokines produced by OVA reactivated splenocytes (Figure 7) of Wt OVA to KO OVA and Wt DCs OVA to KO DCs OVA groups were calculated. We obtained the following ratios for lymphocytes (5 and 2.3), eosinophils (5 and 4), neutrophils (4 and 1), Th2 cells (2 and 3.5), IL-4 (8 and 10), and IL-13 (3 and 3). In both groups, similar ratios for eosinophils, Th2 cells, and IL-4 and IL-13 suggest that adoptive transfer of BMDCs was not able to complement the Th2 defect in b5i KO mice. Nevertheless, transfer of BMDCs into b5i KO mice increased BAL lymphocytes and completely restored BAL neutrophils numbers to Wt levels ( Figure 6B), which corresponds with the increased frequency of lung Th1 cells ( Figure S4B).

Naïve CD4+ T cells from b5i KO mice differentiate into Th2 cells in vitro
It is well recognized that the b5i subunit of the immunoproteasome plays an important role in Th1 [14,15] and Th17 [16] responses. However, no information is available regarding its involvement in Th2 cell development. Hence, we analyzed Th2 differentiation of naïve CD4+ T cells from b5i KO mice in order to understand its function in Th2 transcriptional program and cytokine production. Antigen-independent differentiation of b5i KO naïve CD4+ cells revealed that Th2 cells from b5i KO mice produced normal amounts of IL-4, IL-5 and IL-13 ( Figure 8, Figure S5), suggesting intact Th2 differentiation.
Th2 response reduction during acute asthma in b5i KO mice is OVA specific We wanted to understand if the observed reduction of Th2 responses was OVA specific or rather a general trait of b5i KO mice. Thus, Wt and b5i KO mice were intranasally challenged with HDM ( Figure 9A) followed by differential BAL cell counts and T-cells analysis. HDM application caused robust allergic inflammation as characterized by significant increase in BALs lymphocytes, eosinophils and neutrophils ( Figure 9B). Similarly to OVA induced asthma, HDM challenge led to significantly elevated levels of lung Th2 and Treg cells ( Figure 9C). However, no substantial difference in response to HDM was observed between Wt and b5i KO mice. These results ruled out that b5i KO mice have an inherited inability to mount an efficient Th2 response.

Discussion
The present study focused on the role of b5i proteasome immunosubunit in experimental acute asthma. Currently, there is no information available on the role of this subunit in Th2 driven pathologies, while numerous reports have shown that b5i plays a role in Th1-and Th17-driven experimental inflammation such as DSS-induced colitis [14,15], Crohns disease and ulcerative colitis [12,23] and collagen-induced arthritis [16]. Because non-specific proteasome inhibition by Bortezomib and PS-519 has been shown to significantly reduce airway eosinophilia [17,18] in the models of acute asthma. Therefore, we hypothesized that b5i can also be involved in asthma development and first investigated its expression in healthy and asthmatic lungs of mice. Various methods including proteasome activity assays, two-dimensional gel electrophoresis and western blots confirmed strong expression of the b5i immunosubunit in the lung which however was not altered between naïve and asthmatic lungs (data not shown).
Induction of OVA/Alum asthma in b5i KO mice revealed impaired Th2 responses characterized by reduced eosinophilia, decreased numbers of pulmonary Th2 cells and OVA-specific IgG1 titers. In fact, this finding demonstrates for the first time the impact of the b5i immunosubunit on eosinophil recruitment directly into the lung. Notably, blood eosinophils of non-asthmatic b5i KO mice were comparable to normal control mice. In agreement with previous data from the colitis model [14], b5i KO mice also exhibited reduced amounts of neutrophils in the BAL during the asthma. Interestingly, the frequency of Th1 cells and Tregs in the lung of b5i KO mice was similar to Wt controls suggesting that they are not involved in the down regulation of the Th2 response. Histological analysis of goblet cell numbers and mucus secretion surprisingly revealed no differences between b5i mutant and Wt mice (data not shown).
Next, we wondered whether the reduced Th2 response is airway specific or rather a general trait of these mice. Thus, splenocytes of b5i KO and control mice were in vitro restimulated with OVA and production of IL-4, IL-5 and IL-13 was analyzed. Data revealed that cytokine secretion was reduced in b5i KO, indicating a systemic defect in OVA sensitization. There is substantial evidence that inhibition of the b5i proteasome subunit ameliorates autoimmune pathologies mainly via reduction of Th1 and Th17 cytokines and/or blocking the humoral antibody response [14,15,16]. Yet, the precise mechanisms and pathways by which b5i may affect these immunological processes is not completely understood. Possibly, the b5i subunit mediates its broad regulatory effects in the mucosal systems via activation of NF-kB in adaptive and innate immune cells.
Dendritic cells require NF-kB signaling for responding to exogenous stimuli and subsequent maturation [24,25]. Further, the adjuvant effect of Alum is mediated via interaction with dendritic cells [20]. Therefore, we wanted to exclude any adjuvant mediated effect by employing the Alum independent, subcutaneous OVA asthma model. Similarly, to the OVA/Alum asthma, similar results were achieved with the alum-free protocol, demonstrating that Alum has no effect on reduced Th2 responses in b5i KO mice. While the OVA model of allergic inflammation shows several limitations with respect to the human disease [26], we were interested whether a similar phenotype also occurs in the clinically more relevant HDM asthma model. To our surprise, Th2 responses of HDM acute asthma were similar between Wt and b5i KO mice. To further substantiate this, we tested in vitro Th2 cell differentiation of naïve CD4+ T cells from b5i KO mice and found normal production of IL-4, IL-5 and IL-13 by these cells. We thus concluded that b5i KO mice are capable of mounting normal Th2 response to the different set of allergens, thus demonstrating that these mice do not show an inherited inability to mount normal Th2 responses which is in accordance with a recent publication by Kalim et. al. [27]. It has to be noted that there are major differences concerning the antigenicity between OVA and HDM. While OVA does not cause allergy without prior subcutaneous or intraperitoneal immunization [21], HDM is allergenic by itself. HDM is a complex mixture of Dermatophagoides pteronyssinus antigens which stimulates bronchial epithelial cells to produce eosinophil-attracting/Th2 cytokines such as GM-CSF, TSLP, IL-25 and IL-33, which in turn prime subsequent allergic inflammation [28]. In contrast to HDM, challenging mice with OVA without prior sensitization develops tolerance [29,30]. Multiple sensitization steps with OVA/Alum induce formation of memory T and B-cells, which, after OVA aerosol challenges provoke asthma. We speculate that a possible failure to develop an efficient memory T and B-cells response towards OVA in b5i KO mice could partially explain the observed phenotype.
Immunoproteasome and b5i are constitutively expressed in lymphoid tissues and especially in immature dendritic cells, where they activate NF-kB and degrade byproducts of inflammation induced cellular oxidative stress [22,31,32]. It has been shown that intratracheal instillation of naïve mice with OVA pulsed BMDCs is sufficient to induce asthma after OVA aerosol challenge [33] and that depletion of lung CD11c+ dendritic cells during OVA challenge abrogates allergic inflammation [34]. In line with this, we assumed that b5i KO mice have defects in the inflammatory dendritic cells (iDCs) compartment. To this end, we investigated FITC-OVA uptake and maturation of lung iDCs from b5i KO mice under control and asthmatic conditions. Indeed, b5i KO iDCs showed reduced FITC-OVA uptake and maturation but normal migration into the mediastinal lymph nodes. Complementation of these defects by adoptive transfers of in vitro generated Wt BMDCs into b5i KO mice led to significantly increased allergic inflammation in Wt but not b5i KO mice, although the transfer equally increased iDCs, BAL neutrophils and Th1 cells in the lungs of both groups. The impaired antigen uptake and maturation of iDCs from b5i KO might thus be responsible for the low numbers of BAL neutrophils during OVA asthma but not for the decreased Th2 responses. However, neutrophil influx plays an important role in chronic allergic asthma and acute exacerbations of COPD [35], which makes b5i subunit a possible target for pharmacological intervention in these pathologies.
In summary, we show that the b5i immunosubunit is necessary for mounting a strong Th2 response in the OVA asthma model. Deficiency of this subunit results in a systemically impaired Th2 response, independent of b5i expression on DCs. The observed phenotype is OVA specific as no attenuation is detected in response to HDM allergens. b5i deficient naïve CD4+ T cells differentiate into normal IL-4, IL-5 and slightly elevated IL-13 producing Th2 cells, suggesting intact Th2 specific transcriptional program. Despite defects of lung iDCs from b5i KO mice in the uptake of antigen and maturation, they are not responsible for the reduced Th2 phenotype during Ova asthma. Further molecular and immunological studies have to be performed in order to fully elucidate the mechanisms of the observed b5i phenotype.

Ethics statement
All experiments with animals were performed corresponding to international and German guidelines on animal research and ethics. The used protocols were approved by Regierungsprä sidium Gießen, Hessen, Germany. Permit number: MR 20/6 Nr.71/ 2011.

Animals
Female and male 8-10 weeks old C57BL/6 mice were purchased from Charles River Laboratories, Germany. Female and male b5i KO mice on C57BL/6N background were obtained from the Max Planck Institute for Infection Biology (Berlin, Germany). Both genders were used in the same ratio in all the experiments.

Mouse models of acute allergic airway inflammation
The OVA/Alum-induced allergic airway inflammation model was induced as described previously [21], except that mice were sensitized at days 0, 14 and 21. Control mice were sensitized with 1.5 mg Al(OH) 3   The adjuvant free, subcutaneous allergic airway inflammation model was performed as described previously [21]. House dust mite (HDM) allergic asthma was induced by weekly i.n. challenges of slightly anesthetized mice (i.p. injection of Ketamine/Rompun (Ketamin-Inresa; Inresa Arzneimittel GmbH, Freiburg, Germany)/Rompun 2% (Bayer, Leverkusen, Germany) with 100 mg HDM extract (GREER, Lenoir, USA) in 50 mL PBS on days 0, 7, 14 and 21. One day after last OVA or two days after last HDM challenges mice were sacrificed and blood, bronchoalveolar lavage fluids (BAL), lungs and spleens were collected for further analysis.

Bronchoalveolar lavage
BAL was performed by using 1 mL ice-cold PBS containing 16 protease inhibitor cocktail (Roche, Manheim, Germany). BAL cells analysis was performed as previously described [21]. BAL fluids were centrifuged and the supernatants were stored at -20 uC until cytokine were measured. The measurement of cytokines IL-4, IL-5 and IL-13 was performed by Cytometric Bead Array (BD, Heidelberg, Germany) according to manufacturer's protocol.

Measurement of OVA-specific antibodies levels
OVA-specific IgG1 antibody titers were measured in the serum as described previously [21].

Spleen cells restimulation and cytokine analysis
Spleens were harvested and single cell suspensions were prepared. Erythrocytes were lysed and the cell concentration was adjusted to 2610 6 cells/mL by RPMI-1640 supplemented with 10% FCS, 5 mM glutamine, 100 mg/ml streptomycin, 100 U/ml penicillin. Next, 200 mL of the suspension were transferred into a 24-well plate and restimulated with 50 mg/mL OVA (grade VI) for 72 h (37uC, 5% CO 2 ). Then plates were centrifuged (350 g, 10 min), supernatants were collected and stored at 220uC until the cytokines measurement. IL-4, IL-5 and IL-13 cytokines quantification was performed by Cytometric Bead Array (BD, Heidelberg, Germany) according to the manufacturer's protocol.

Bone marrow derived dendritic cells (BMDCs) preparation
BMDCs preparation and growth was performed as described in [36] with modifications. Bone marrow was isolated from femur and tibia and a single cell suspension was prepared in dendritic cell incubation medium (DCIM; RPMI-1640 (with Hepes; PAA, Pasching, Austria) supplemented with 10% FBS (Life Technologies GmbH, Darmstadt, Germany), 0,5% b-Mercaptoethanol, 1% Penicillin, 1% Streptomycin and 1% L-Glutamin (Life Technologies GmbH, Darmstadt, Germany)). Cell pellets were resuspended in erythrocytes lysis buffer (8 g/l NH 4 Cl; 1 g/l KHCO 3 ; 37,2 mg/l EDTA) and incubated at room temperature for 3-5 mins. Then, cells were washed, centrifuged and resuspended in DCIM. About 4610 6 cells were plated in Nunclon TM D Surface Petri dishes (Nalge Nunc International, Rochester, USA) in 10 mL DCIM supplemented with 10% GM-CSF (supernatant of murine hybridoma cell line 66310). Freshly prepared DCIM was added to the cell suspension at days 3 (supplemented with 10% GM-CSF) and 6 (supplemented with 10% GM-CSF and 300 mg/ml OVA). As control, one batch of DCs was prepared without any OVA stimulation. On day 7 cells were washed twice with PBS, and injected i.p. in PBS at amount of 0.8610 6 cells per mouse.

T-cells in vitro differentiation
Naive CD4 + CD62L hi cells were isolated from spleens and lymph nodes of WT and b5i KO mice by magnetic associated cell sorting (MACS, Miltenyi Biotec, Bergisch-Gladbach, Germany). Isolated cells were differentiated into Th2 cells in the presence of plate-bound anti-CD3 (5 mg/ml, clone 145-2C11), soluble anti-CD28 (1 mg/ml, clone 37.51), anti-IFN-c antibodies (XMG1.2) 5 mg/ml and recombinant murine IL-4 20 ng/mL (PeproTech, Rocky Hill, USA). After 3 days of incubation, cells were transferred into new plates containing IL-2 100 U/ml and incubated further for the next 3 days. On the last incubation day cells were restimulated and analyzed by Flow Cytometry as described above.

Statistical analysis
Data were expressed as mean 6 standard error of mean (SEM). One-way ANOVA statistical analysis was performed followed by Tukey's HSD test using GraphPad Prism 5 Software. P-values of P,0.05 were considered significant. Figure S1 Representative lung Th2 cells staining of Wt and b5i KO mice during OVA/Alum acute asthma. The percentages of IL5+ IL13+ of CD4+ T cells were measured by flow cytometry from the homogenized lungs. The presented plots correspond to the data points of Th2 cells in Figure 2. The gating strategy was determined by comparing the stained samples to the negative and isotype controls. (TIF) Figure S2 Representative lung Treg cells staining of Wt and b5i KO mice during OVA/Alum acute asthma. The percentages of CD25+ FoxP3+ of CD4+ T cells were measured by flow cytometry from the homogenized lungs. The presented plots correspond to the data points of Treg cells in Figure 2. The gating strategy was determined by comparing the stained samples to the negative and isotype controls. (TIF) Figure S3 Gating strategy for identification of lung inflammatory dendritic cells (iDCs). Macrophages were excluded based on the F4/80 staining. F4/80 negative cells were further gated to exclude CD11c negative cells. CD11c positive cells were gated to exclude CD11b negative cells. Double positive CD11c+ CD11b+ population was separated into two groups based on the expression of Ly6C inflammatory monocyte precursor marker. In order to eliminate possible B-cells contamination triple positive CD11c+ CD11b+ Ly6C+ population was further gated to remove B220 positive cells. CD11c+ CD11b+ Ly6C+ B2202iDCs were analyzed on the presence of the uptaken OVA-FITC with the subsequent MHC II surface expression. The presented strategy corresponds to the data points of iDCs in Figure 6 and Figure S4A. The gating strategy was determined by comparing the stained samples to the negative and isotype controls. The percentages of IL24+, IL25+ and IL213+ of CD4+ T cells were measured by flow cytometry from in vitro differentiated Th2 cells. The presented plots correspond to the data points in Figure 9. The gating strategy was determined by comparing the stained samples to the negative and isotype controls. (TIF)