Enhanced Inflammatory Potential of CD4+ T-Cells That Lack Proteasome Immunosubunit Expression, in a T-Cell Transfer-Based Colitis Model

Proteasomes play a fundamental role in intracellular protein degradation and therewith regulate a variety of cellular processes. Exposure of cells to (pro)inflammatory cytokines upregulates the expression of three inducible catalytic proteasome subunits, the immunosubunits, which incorporate into newly assembled proteasome complexes and alter the catalytic activity of the cellular proteasome population. Single gene-deficient mice lacking one of the three immunosubunits are resistant to dextran sulfate sodium (DSS)-induced colitis development and, likewise, inhibition of one single immunosubunit protects mice against the development of DSS-induced colitis. The observed diminished disease susceptibility has been attributed to altered cytokine production and CD4+ T-cell differentiation in the absence of immunosubunits. To further test whether the catalytic activity conferred by immunosubunits plays an essential role in CD4+ T-cell function and to distinguish between the role of immunosubunits in effector T-cells versus inflamed tissue, we used a T-cell transfer-induced colitis model. Naïve wt or immunosubunit-deficient CD4+ T-cells were adoptively transferred into RAG1−/− and immunosubunit-deficient RAG1−/− mice and colitis development was determined six weeks later. While immunosubunit expression in recipient mice had no effect on colitis development, transferred immunosubunit-deficient T- cells were more potent in inducing colitis and produced more proinflammatory IL17 than wt T-cells. Taken together, our data show that modifications in proteasome-mediated proteolysis in T-cells, conferred by lack of immunosubunit incorporation, do not attenuate but enhance CD4+ T-cell-induced inflammation.


Introduction
The immune system senses pathogens through pattern recognition receptors that bind specific pathogen-associated molecular patterns. Ligand binding induces a signaling cascade downstream of the receptor that activates a specific transcriptional program, allowing the immune system to respond efficiently to the invading microorganisms. The proteasome, an abundant cellular protease complex, plays an essential role in those signaling pathways, as the activation of many signaling molecules is regulated by the timely degradation of other molecules in the signaling complex. So depends the activation of the transcription factor NFkB on phosphorylation, ubiquitylation and subsequent proteasome-mediated degradation of its inhibitor IkB [1]. IkB degradation exposes a nuclear localization sequence in NFkB, allowing it to translocate to the nucleus and to initiate the expression of, amongst others, (pro)inflammatory cytokines [1][2][3]. Another function of proteasomes, during infection with intracellular pathogens, is the processing of pathogen-derived antigens into peptides that can be presented by MHC class I molecules on the cell surface, allowing CD8 T-cells to detect and react to the presence of intracellular pathogens (for review see [4]). Thus, proteasome activity plays an essential role at different stages of pathogen-specific immune responses.
Proteasomes consist of a barrel-shaped catalytic core particle, the 20S proteasome, and one or more regulatory particles (for review see [5]). The enzymatic activity of the 20S proteasomes is exerted by three b subunits, located in the inner two rings of the 20S complex, which exhibit caspase-like (b1), trypsin-like (b2) and chymotrypsin-like activity (b5). Exposure of cells to type 1 and type 2 interferons or TNFa induces the expression of three facultative subunits, b1i/LMP2, b2i/MECL-1 and b5i/LMP7, which preferentially incorporate into newly assembled proteasome complexes and thus, when expressed, replace their constitutive homologues in the cellular proteasome population [5]. In addition, in particular cells of the hematopoietic lineage express different quantities of the three facultative subunits and, therefore, often contain so called ''mixed'' proteasomes, containing the constitutive and one or more inducible b subunits [5,6]. Due to altered cleavage preferences, proteasomes containing the facultative subunits (named immunoproteasomes) are more suited to generate high affinity MHC class I ligands than constitutive proteasomes, containing the b1, b2 and b5 subunits [5,7,8]. As a consequence, pathogen-specific CD8 + Tcell responses often target immunoproteasome-generated peptides [5,8,9]. Immunoproteasomes have further been shown to protect cells from interferon-induced oxidative stress, by efficient removal of aggregates of oxydant-damaged, polyubiquitylated unfolded nascent proteins [10,11] and immunoproteasome expression in the peripheral tissues was found to protect against early forms of CD8 + T-cell-mediated autoimmune diseases [7,10]. In specific, we showed that irradiated and BM reconstituted, b2i/MECL-1&b5i/LMP7-deficient recipient mice developed latent forms of CD8 + T-cell-mediated autoimmune diseases, such as insulindependent diabetes mellitus and diabetes insipidus [7]. This could be explained by altered MHC class I antigen processing of tissue antigens in inflamed immunosubunit-deficient tissue, in particular, since adoptive transfer of CD4 + T-cell depleted splenocytes from diseased mice conferred the disease phenotype onto RAG1 2/2 mice that also lacked immunosubunit expression, but not onto RAG1 2/2 mice that expressed the immunosubunits [7]. Alternatively, enhanced death in immunosubunit-deficient tissues following irradiation [10] may have led to presentation of tissue antigens in the draining lymph nodes and thus may have induced priming of auto-reactive CD8 + T-cells.
In hematopoetic cells, expression of the immunosubunits supports the production of many cytokines. [12][13][14]. Based on the latter observations, immunoproteasomes have been proposed to drive inflammatory processes, an assumption that is supported further by the anti-inflammatory effects of proteasome subunitspecific inhibitors [14]. The role of immunoproteasomes in inflammatory diseases has been investigated in further detail in different mouse models of colitis [15][16][17]. These studies showed that inflammatory disease was accompanied by high levels of immunosubunit expression in the inflamed tissue, with cellular infiltration and expression of inflammatory cytokines [15,16]. Inhibition of activity of the proteasome subunit LMP7/b5i or of both LMP7/b5i and b5 largely prevented DSS-induced colitis, and was associated with reduced quantities of inflammatory cytokines [14][15][16]. Notably, a deficiency of b5i/LMP7, in mice lacking the gene encoding this subunit, protected from DSSinduced colitis [15,16]. This was attributed to diminished processing of the transcription factor NFkB, which initiates cytokine transcription, in hematopoietic cells by Schmidt et al. [15]. Groettrup and coworkers [17], however, reported a reduced differentiation of ''proinflammatory'' Th1 and Th17 cells and enhanced differentiation of regulatory T-cells (Tregs) in b5i/ LMP7 2/2 mice, suggesting that altered T-cell differentiation may explain for the ameliorating effects of b5i/LMP7 deficiency in development of DSS-induced colitis. The last authors also showed that LPS or anti-CD3 stimulated spleen cells of wt and b5i/ LMP7 2/2 mice produce similar amounts of IL-6, IFNy and IL-17 [14,16], arguing against a direct effect of LMP7 on the production of those cytokines. Taken together, these data establish an important regulatory role for the immunosubunit b5i/LMP7 in inflammatory disease, but the underlying molecular mechanism remains poorly understood. In addition, the protective effects of b5i/LMP7 deficiency in inflammatory disease are remarkable in light of the observed importance of immunoproteasomes in clearing potentially toxic protein aggregates in cytokine-exposed cells [10].
To further investigate the role of proteasome immunosubunits in autoimmune, inflammatory disease, we analyzed T-cell-induced colitis development in RAG1 gene-deficient mice compared to RAG1&b2i/MECL-1&b5i/LMP7-deficient mice, in a T-cell transfer model. Using this model, we were able to distinguish between disease modifying effects of immunosubunit expression in inflamed tissue and in effector T-cells, respectively.
Enhanced Induced (i)Treg Differentiation and Elevated Levels of IL17 in Colitic Mice that Received b2i/MECL-1&b5i/LMP7-deficient CD4 + T-cells b5i/LMP7-deficient mice are resistant to development of DSSinduced colitis, due to reduced NFkB activation or to altered CD4 + T-cell differentiation, which both may dampen the production of (pro)inflammatory cytokines [15,17]. Intracellular staining of FoxP3 in spleen-derived CD4 + T-cells of mice with Tcell transfer-induced colitis ( Figure 2D, E) showed enhanced numbers and percentages of FoxP3 pos cells, i.e. iTregs, in mice that On the other hand, deficiency of b2i/MECL-1&b5i/LMP7 in recipient mice had no effect on iTreg differentiation ( Figure 2D, E).
Analysis of cytokine expression in TCR-activated splenic T-cells by RT-PCR demonstrated enhanced quantities of IL17 mRNA in recipients of b2i/MECL-1&b5i/LMP7-deficient CD4 + T-cells ( Figure 3A), while expression levels of IFNy, IL10 and TNFa did not differ between the different mouse groups ( Figure 3B-D). Thus, both iTregs and IL17-producing cells were more frequent in spleens of colitic mice that had been transferred with b2i/MECL-1&b5i/LMP7-deficient CD4 + T-cells than in mice that had received wt CD4 + T-cells.

Discussion
Previous studies have shown that proteasome immunosubunits may modify host susceptibility to autoimmune and inflammatory diseases at multiple levels, i.e. by altering MHC class I antigen processing in inflamed tissues, by modulating the ability of cells to cope with cell stress and by modulating signal transduction and transcriptional programs in cells, leading to modified T-cell differentiation [7,10,15,17]. We have used a murine model of T-  cell transfer-induced colitis to further examine the role of the proteasome immunosubunits in inflammatory diseases. We show that naïve CD4 + T-cells of mice that lack the two immunosubunits b2i/MECL-1&b5i/LMP7 cause colitis upon transfer into RAG1 2/2 mice, which is more severe than colitis induced by transfer of naïve wt CD4 + T-cells. Enhanced colitis scores were associated with enhanced numbers of total splenic CD4 + T-cells, enhanced frequencies and numbers of iTregs and enhanced IL17 production. Previous studies in humans and DSS-treated mice showed that not only immune cells but also inflamed colon tissue expressed the proteasome immunosubunits [16,21]. In our experiments, the severity of T-cell transfer-induced colitis did not differ between Rag1 2/2 mice and RAG1 2/2 mice that had been backcrossed onto an b2i/MECL-1&b5i/LMP7-deficient background (Figure 1), arguing against a role for immunosubunit expression in the inflamed gut in colitis development. This is consistent with prior observations, which showed that transplantation of wt bone marrow rendered b5i/LMP7-deficient mice susceptible to development of DSS-induced colitis [15]. Thus, in the T-cell transfer model, b2i/MECL-1&b5i/LMP7-deficiency in transferred, naïve CD4 + T-cells enhances colitis development in RAG1 2/2 mice, whereas immunosubunit-deficiency in the targeted gut does not have any effect on colitis development.
Studies by different groups have demonstrated that proteasomes play an important role in cytokine production and that modulation of proteasome subunit composition may alter cytokine production. For example, dendritic cells of LMP2-deficient mice produce less IFNa, IL-6, IL-1b and TNFa following influenza virus infection than wt dendritic cells [12]. This defect was attributed to the formation of mixed proteasomes, containing the immunosubunits b2i/MECL-1 and b5i/LMP7 and the constitutive subunit b1/d, which may diminish transcription of these cytokines by delaying NFkB processing in those cells. A diminished transcription of the same cytokines, IL17 and IL23 was detected also during DSSinduced colitis in b5i/LMP7-deficient, in b1i/LMP2 deficientand in b2i/MECL-1-deficient mice, as well as in mice treated with selective inhibitors targeting the b5i/LMP7 and b5 subunits or the b5i/LMP7 subunit only [15,16]. This was accompanied by a reduced activation of NFkB, in b5i/LMP7-deficient cells [15] and reduced Th1 and Th17 but enhanced Treg cell differentiation of CD4 + T-cells in the lamina propria of b5i LMP7-deficient/ inhibitor-treated mice with mild forms of DSS-induced colitis [15][16][17]. Following T-cell transfer into RAG2 2/2 mice, Kalim et al. [17] further observed a diminished Th1 and Th17 and enhanced Treg cell differentiation in recipients of b5i/LMP7-deficient CD4 + T-cells, suggesting that differences in T-cell differentiation explain for the relative resistance of inhibitor-treated or b5i/LMP7deficient mice to development of DSS-induced colitis.
In agreement with Kalim et al [17], we observe an enhanced differentiation of naïve immunosubunit-deficient CD4 + T-cells into iTregs in adoptively transferred RAG1 2/2 mice. However, we also find that these iTregs fail to protect the transferred mice against colitis development. Thus, recipients of b2i/MECL-1&b5i/LMP7-deficient T-cells develop colitis six weeks after Tcell transfer, with even higher histological scores and higher quantities of IL17a in the spleens than detected in recipients of wt CD4 + T-cells, despite enhanced numbers and frequencies of iTregs. On the other hand, FACS-purified Tregs derived from b2i/MECL-1&b5i/LMP7-deficient, immunocompetent mice suppress CD4 + T-cell-induced colitis upon cotransfer (not shown), indicating that b2i/MECL-1&b5i/LMP7-deficiency does not disable Treg function. Interestingly, in a recent publication, Basler et al. [22] showed that both b5i/LMP7-deficient and wt mice develop experimental autoimmune encephalitis (EAE) following immunization with the MOG 35-55 peptide. Disease intensity did not significantly differ between mouse groups in this study, but was found to be more intense in b5i/LMP7-deficient mice by Seifert et al. [10], using the same model. As Th17 cells play a central role in the onset and maintenance of EAE, these studies indicate that an exchange of b5i/LMP7 for b5 in T-cells does not diminish T cell-induced inflammatory disease. Moreover, specific inhibition of b5i/LMP7 ameliorated peptide-immunization-induced EAE in wt mice, while specific inhibition of the constitutive proteasome subunit b5 ameliorated EAE in LMP7-deficient mice [22], indicating that the enzymatic activity of the b5/b5i proteasome active site subunit is critical for CD4 + T-cell function. Taken together, we conclude that the exchange of b5i/LMP7 for b5 in T cells does not abolish but may modulate CD4 + T-cell mediated disease, with aggravating effects on inflammation in case of T-cell transfer-induced colitis.
While our data show that immunosubunit-deficiency does not ameliorate T-cell transfer-induced colitis, immunosubunit-deficient mice are protected from DSS-induced colitis [15][16][17]. Although unlikely, it cannot be ruled out that this discrepancy between effects in the two models may be explained by the use of different immunosubunit deficient mice in these studies. While resistance to DSS-induced colitis was observed in single immunosubunit-deficient mice, lacking either b1i/LMP2, b2i/MECL-1 or b5i/LMP7 [15][16][17], we used b2i/MECL-1&b5i/LMP7-deficient T-cells to determine the effects of immunosubunit deficiency in development of T-cell transfer colitis. Thus we cannot exclude that specific proteasome subtypes [23,24], present in single immunosubunit-deficient cells but lacking in double immunosubunitdeficient cells, or the other way around, are responsible for the discrepancy between effects of immunosubunit deficiency in colitis induction in the two models. A more likely explanation however lies in the involvement of different immune effector cells in DSSinduced compared to T-cell transfer-induced colitis and perhaps the differing roles of NFkB activation in these cell types. In DSSinduced colitis, inflammation is driven primarily by leukocytes of the innate immune system, such as monocyte-derived macrophages, which are dependent on NFkB activation for cytokine expression. Thus, delayed NFkB processing may modify the cytokine environment in DSS-treated mice and thereby prevent the onset of lead to modest inflammation at most. In T-cell transfer-induced colitis, inflammation is mediated by activated CD4 + T-cells. In T cells, NFkB activation plays a role only during priming, since CD28-mediated costimulatory signals are dependent on the NFkB pathway but once primed, T-cells secrete cytokines mainly upon antigen-induced TCR stimulation. As discussed above, b5i/LMP7-deficient T cells are activated in inflammatory disease models (as well as infection models) and secrete cytokines upon TCR engagement with quantities comparable to those secreted by wt T cells [14,16], thus a possibly delayed NFkB activation does not impair their function.
Taken together, we conclude that immunosubunit deficiency may influence disease processes in different ways, depending on the etiology of disease. Our data argue against any direct attenuating effects on inflammatory CD4 + T-cells but, in contrary, show that b2i/MECL-1&b5i/LMP7-deficiency enhances CD4 + T-cell-mediated inflammation in T-cell transfer-induced colitis.

Ethics Statement
All animal experiments were carried out in strict accordance with the Dutch Animal Experimentation Act and EU directives 86/609/CEE and 2010/63/EU related to the protection of vertebrate animals used for experimental or other scientific purposes. The experimental protocols were approved by the Committee on Animal Experiments of the University of Utrecht (DEC 2012.II.02.029) and performed in the Central Laboratory Animal Research Facility of the University of Utrecht, which has AAALAC (Association for Assessment and Accreditation of Laboratory Animal Care) accreditation.

T-cell Transfer Colitis Model
To induce colitis, CD4 + CD45RB high cells were FASC-purified from the spleens of B6 wt and B6 b2i/MECL-1 2/2 b5i/LMP7 2/2 donor mice, using a BD Influx TM (BD Biosciences). A total of 4610 5 CD4 + CD45RB high cells were injected intraperitoneally (i.p.) in 200 ml PBS into B6 RAG-1 2/2 or B6 RAG-1 2/2 b5i/LMP7 2/ 2 b2i/MECL-1 2/2 recipient mice. Mice were monitored for 6 weeks for clinical signs and body weight and then sacrificed. Colons were flushed with 10% formalin to remove feces, fixed and stored in 70% ethanol for H&E histopathology. Colitis development was evaluated by histology by two independent experts in a blinded fashion, with minor modifications from the protocol described in [20]. In brief, inflammatory infiltrates, depletion of goblet cells and epithelial hyperplasia were scored as specified in Table 1. Scores for specific colon sections represent the sum of scores for these different criteria. The overall histological score per mouse colon is the average of individual scores for the proximal, mid and distal colon segments.

Flow Cytometry and T-cell Stimulation
Spleens were passed through a 70 mm cell strainer to prepare single cell suspensions, and erythrocytes were lysed by treatment with ammonium chloride. Cells for T-cell transfer were stained with anti-CD4 mAb (RM4-5, eBioScience) in the presence of Fcblock, enriched using a BD Influx TM (BD Biosciences) and then were stained with anti-CD45RB (16A, BD Pharmingen) and anti-CD25 (PC61, eBioscience) mAbs and resorted to obtain a CD4 + CD45RB high CD25 2 population that was used for induction of colitis. To analyze spleens of T-cell transferred mice, total cell counts in single cell suspensions were determined and cell samples then were stained for CD4 cell surface and intracellular FoxP3 expression (clone FJK-16S, eBioscience) according to manufacturer's instructions. Cells were measured on a FACSCantoII (BD Biosciences) and analyzed with FlowJo software (Tree Star). To determine cytokine expression, splenocytes were stimulated with or without anti-CD3 mAb (145.2C11, BD Pharmingen, 2 mg/ml) for 4 h, pelleted and then stored at 220uC in TRIzol reagent (Invitrogen, Breda, NL) prior to RNA purification.

Quantitative PCR
RNA was extracted from TRIzol stored samples as instructed (Invitrogen, Breda, NL) and quantified using a Nanodrop ND-1000 (Thermo Scientific, Etten-Leur, NL). 1 mg of total RNA was reverse transcribed into cDNA synthesis using an iScript cDNA Synthesis Kit (Bio-Rad Laboratories, Veenendaal, NL). TaqManH Gene Expression Assays (Applied Biosystems, Austin, TX, USA) were used for qRT-PCR amplification and expression of the following genes was assessed: IL17a, IFNy, TNFa and IL10. Relative expression of target genes was calculated using the Pfaffl method with 18S rRNA as a reference gene.