|Home | About | Journals | Submit | Contact Us | Français|
In the presence of IL-6, transforming growth factor (TGF)-β1 induces differentiation of T helper (Th) 17 cells in mice. Interleukin (IL)-23, a heterodimeric cytokine composed of IL-23p19 and IL-12/23p40 subunits, stimulates the growth and expansion of Th17 cells, and has been implicated in psoriasis pathogenesis. To study the associations between TGF-β1, the IL-23/Th17 inflammatory pathway, and psoriasis, we investigated inflammatory skin disease in transgenic mice that constitutively overexpress human TGF-β1 in basal keratinocytes (K5.hTGF-β1 transgenic mice); these mice had previously been reported as having a psoriasis-like disease. K5.hTGF-β1 transgenic mice had high levels of TGF-β1 mRNA and protein in both skin and serum. Levels of cytokines involved in IL-23/Th17-mediated inflammation were not elevated in lesional skin compared with those in non-lesional and wild-type skin. It is noteworthy that IL-4 and IgE were markedly elevated in inflamed skin and serum, respectively, of transgenic mice. Monoclonal antibodies (mAbs) specifically directed against IL-23p19 or IL-12/23p40 had no clinical effect on established inflammatory skin disease in K5.hTGF-β1 transgenic mice, whereas the same mAbs were able to block the development of murine experimental autoimmune encephalomyelitis, an IL-23/Th17-mediated disease. In summary, the IL-23/Th17 inflammatory pathway is not responsible for the maintenance of inflammatory skin disease in K5.hTGF-β1 transgenic mice.
Interleukin (IL)-23 is primarily produced by dendritic cells (Oppmann et al., 2000), and is clearly elevated in psoriasis lesions, as indicated by increased levels of mRNA for each of its subunit components, IL-23p19 and IL-12/23p40, in lesional skin as compared with those in non-lesional skin, whereas mRNA levels of IL-12/35p35 are not elevated (Lee et al., 2004; Chamian et al., 2005; Chan et al., 2006; Toichi et al., 2006). Furthermore, an intradermal injection of recombinant IL-23, but not of IL-12, into mice induces acute histological changes reminiscent of human psoriasis (Chan et al., 2006; Zheng et al., 2007). These IL-23-triggered changes are dependent on cytokines produced by T helper (Th) 17 cells (Zheng et al., 2007). Th17 cells are a newly appreciated subset of CD4+ T helper cells, distinct from Th1 and Th2 cells, which are dependent on IL-23 for survival and expansion (Oppmann et al., 2000; Aggarwal et al., 2003). Genetic or antibody targeting of either IL-23 or Th17 cells ameliorates several T-cell-mediated autoimmune disorders in mice, including classic models of rheumatoid arthritis and multiple sclerosis (Cua et al., 2003; Murphy et al., 2003; Langrish et al., 2005; Chen et al., 2006). In contrast, targeting of IL-12 or Th1 cells leads to worsening of these same conditions (Cua et al., 2003; Murphy et al., 2003; Langrish et al., 2005; Chen et al., 2006). Thus, the IL-23/Th17 inflammatory pathway is critical in mediating distinct autoimmune T-cell-mediated inflammatory diseases in mice, and its role in mediating inflammation in human psoriasis appears to be equally important (Fitch et al., 2007).
Although IL-23 stimulates Th17 survival and proliferation, differentiation of murine Th17 cells from naive T-cell precursors is dependent on transforming growth factor (TGF)-β1 in combination with IL-6 (Bettelli et al., 2006; Mangan et al., 2006; Veldhoen et al., 2006). Neutralization of TGF-β1 abrogates differentiation of naive murine CD4+ T cells to Th17 cells in culture, whereas exposure to TGF-β1 and pro-inflammatory cytokines, such as IL-6 induces Th0 cells to become mature Th17 effector cells (Veldhoen et al., 2006). Mice that completely lack TGF-β1 have lower serum levels of IL-17 and produce fewer IL-17+CD4+splenic T cells, whereas mice that overexpress TGF-β1 under the IL-2 promoter produce more antigen-stimulated IL-17 (Mangan et al., 2006). In addition, TGF-β1 expression by T cells is required for the induction of Th17 cells in certain tissues (Li et al., 2007).
K5.hTGF-β1 transgenic mice overexpress latent human TGF-β1 in basal keratinocytes and hair follicles under the K5 promoter, which leads to the development of inflammatory skin disease (Li et al., 2004). A recent evaluation of murine models of human psoriasis ranked K5.hTGF-β1 transgenic mice high among the numerous models, because of clinical, histological, and cellular similarities to human psoriasis (Gudjonsson et al., 2007). These mice clinically showed many clinical and cellular hallmarks of psoriasis, including erythematous scaly plaques, epidermal hyperproliferation, hyperkeratosis, acanthosis, and inflammatory cell infiltrates of CD8+ T cells and neutrophils in the epidermis, and CD4+ T cells and macrophages in the dermis (Li et al., 2004). Clinical disease is apparent by 4 months of age in most mice, with consistent Koebnerization of ears at sites of ear tagging (Li et al., 2004). As TGF-β1 induces Th17 differentiation in mice, and given the importance of Th17 cells in human psoriasis, we investigated the role of the IL-23/Th17 inflammatory pathway in the pathogenesis of inflammatory skin disease in K5.hTGF-β1 transgenic mice.
To test the hypothesis that inflammatory skin disease in K5.hTGF-β1 transgenic mice is because of the increased presence of Th17 cells, we measured mRNA levels of several cytokines that are important in the IL-23/Th17 inflammatory pathway, using quantitative real-time reverse transcriptase polymerase chain reaction (qRT-PCR) (Figure 1). Data are expressed in three separate ways: comparison of lesional skin with wild-type (WT) skin (Figure 1a); comparison of clinically normal-appearing (that is, non-lesional) skin with WT skin (Figure 1b); and comparison of lesional skin with non-lesional skin (Figure 1c). As expected, lesional and non-lesional skin contained high levels of hTGF-β1 mRNA. TNF-α, IL-6, and IL-17A mRNAs were also elevated in lesional skin. However, mRNAs specific for the IL-23p19 and IL-12/23p40 subunits of IL-23, as well as for IL-23R, a component of the IL-23 receptor expressed on Th17 cells, were present at similar levels in both lesional and non-lesional skin as those in WT skin. mRNA encoding the Th17 cell-associated transcription factor, RORγt, was present at decreased levels compared with that in WT skin. mRNA encoding IL-22, a key Th17 effector cytokine, was not detected in any skin sample.
We next examined cytokines associated with Th1, Th2, and regulatory T-cell (Treg) responses using qRT-PCR (Figure 1). IL-12/35p35 levels were similar in both lesional and WT skins. It is interesting that levels of IL-4, a Th2 cytokine, were high in transgenic lesional skin compared with those in WT skin. Compared with non-lesional skin, lesional skin had elevated mRNA levels for Treg-associated cytokines, including foxp3, IL-10, and IL-12p35, which is a component of IL-12 and of a newly described and crucial Treg cytokine, IL-35 (Collison et al., 2007).
We isolated total protein from lesional and non-lesional skin of K5.hTGF-β1 transgenic mice and analyzed tissue and serum samples for the presence of IL-23 and Th17-related cytokines by enzyme-linked immunosorbent assay (ELISA) (Figure 2). Transgenic mice exhibited high levels of total hTGF-β1 protein (Figure 2a). IL-1α protein was detected at similar levels in skin obtained from both K5.hTGF-β1 transgenic mice and WT mice (data not shown). Protein levels of IFN-γ were slightly elevated and IL-4 levels were markedly elevated in lesional skin (Figure 2b and c). As a result of the latter finding, we tested the serum of transgenic mice, and found serum IgE to be very high in K5.hTGF-β1 transgenic mice compared with that in WT mice (Figure 2d). In contrast, IL-17A, IL-22, and IL-23 proteins were not detected in skin of K5.hTGF-β1 transgenic mice or in that of WT mice. WT and transgenic lymph node cultures treated with IL-17-promoting cytokines showed high levels of IL-17A, IL-22, and IL-12/23p40. IL-12/23p40 was also detected in sera of both groups of mice. No differences were observed in IL-17A, IL-22, and IL-12/23p40 protein levels in the lymph node and sera between WT and K5.hTGF-β1 transgenic mice (Figure 2e and h).
To determine the functional role of the IL-23/Th17 axis in mice with advanced skin disease, we neutralized specific components of the pathway with monoclonal antibodies (mAbs), as has been done recently in humans with psoriasis (Krueger et al., 2007; Kimball et al., 2008; Leonardi et al., 2008; Papp et al., 2008). Treated mice were evaluated with a clinical disease-scoring system to determine the response to therapy. Mice with baseline disease scores of two or more were treated for 3 weeks by intraperitoneal injection, with 1 mg of neutralizing mAbs directed against either IL-23p19 or IL-12/23p40, or they were treated with isotype control Abs. Skin disease progressed in a similar manner in isotype control Ab-treated (Figures (Figures3a3a and and4a),4a), anti-IL-23p19 mAb-treated (Figures (Figures3b3b and and4a),4a), and anti-IL-12/23p40 mAb-treated (Figures (Figures3c3c and and4a)4a) K5.hTGF-β1 transgenic mice. It is noteworthy that all histological analyses of lesional skin isolated from K5.hTGF-β1 transgenic mice undergoing treatment were similar, and showed hyperkeratosis with serum crust, acanthosis, and a superficial mixed inflammatory infiltrate consisting of mononuclear cells and scattered eosinophils (Figure 3). Weight loss was similar across all groups tested (Figure 4b). To ensure that mAbs were functional in neutralizing IL-12 and IL-23, treatment with 1 mg of anti-IL-23p19 and anti-IL-12/23p40 mAbs prevented the onset of experimental autoimmune encephalomyelitis (Figure 4c), a disease mediated by IL-23 and Th17 cells (Chen et al., 2006).
The results presented here indicate that the IL-23/Th17 inflammatory pathway is not responsible for maintenance of inflammatory skin disease in K5.hTGF-β1 transgenic mice. As TGF-β1 is a critical component in the development of Th17 cells in mice (Bettelli et al., 2006; Mangan et al., 2006; Veldhoen et al., 2006), we originally hypothesized that overexpression of TGF-β1 by keratinocytes could influence the differentiation of naive T cells in skin-draining lymph nodes, and lead to enhanced development of Th17 cells. These Th17 cells, through expression of skin-homing receptors, CCR6 and CCR4 (Acosta-Rodriguez et al., 2007a, b), would return to the skin and release inflammatory cytokines, such as IL-17A, IL-22, and TNF-α. The net result would be the establishment and maintenance of psoriasiform lesions (Li et al., 2004). Disease would resolve on mAb therapy directed against IL-23, similar to human psoriasis (Krueger et al., 2007; Kimball et al., 2008; Leonardi et al., 2008; Papp et al., 2008). We showed, however, that levels of IL-23 and Th17 cytokines are not elevated in K5.hTGF-β1 transgenic mice (Figures (Figures11 and and2).2). We also showed that mAb targeting of the IL-23/Th17 inflammatory pathway does not affect the maintenance of skin disease in these mice (Figures (Figures33 and and4).4). As we learn more about the role of IL-23 and Th17 cells in the pathogenesis of human psoriasis and other inflammatory skin diseases, it will be critical to investigate this pathway in each of the various murine models of these diseases. We believe that this study will provide the framework to carry out these future studies.
There are some limitations to our study. The K5.hTGF-β1 transgenic mice showed clinical, histological, and cytokine expression features that were different than those originally described by Li et al. (2004). More specifically, this original publication described many characteristics of human psoriasis, yet our findings suggest that these mice have features that resemble human atopic dermatitis more so than human psoriasis (for example, crusted skin lesions, scattered tissue eosinophils, high lesional IL-4 expression, and high-serum IgE levels). It is unclear why we found such different results compared with the results arrived at by Li et al., but this may be because of changes in transgene expression that could occur after years of breeding of K5.hTGF-β1 transgenic mice. Although we have definitively shown that the IL-23/Th17 inflammatory pathway is not involved in the inflammatory skin disease observed in these transgenic mice, the precise mechanisms underlying pathogenesis in these mice remain unresolved. For example, the contributions of bacterial colonization and scratching in exacerbating skin lesions in these mice may be studied further, and would be particularly relevant in further understanding the pathogenesis of atopic dermatitis.
TGF-β1 expression can paradoxically activate and suppress the immune function (Wan and Flavell, 2007). Although the combination of TGF-β1 and IL-6 can direct the differentiation of murine Th17 cells, TGF-β1 can also inhibit the pathogenic function of these cells, even in the presence of IL-23 (McGeachy et al., 2007). It is possible that TGF-β1, in the setting of K5.hTGF-β1 transgenic mice, does not drive Th17 cells toward an inflammatory phenotype, but restricts their pro-inflammatory activities. Furthermore, TGF-β1, in a non-inflammatory tissue milieu, can induce Treg differentiation from naive T cells as well as the production of the Treg transcription factor, foxp3 (McGeachy and Cua, 2008). Our data indicate that mRNA levels of Treg-associated factors, foxp3, IL-12/35p35, and IL-10, are elevated in lesional skin than in non-lesional skin. It is also noteworthy that the protein subunit encoded by IL-12/35p35 can bind with the Epstein–Barr virus-induced 3 protein, or EBI3, to create the anti-inflammatory Treg-associated cytokine known as IL-35 (Collison et al., 2007). Although foxp3, IL-12/35p35, and IL-10 mRNA are elevated in skin of K5.hTGF-β1 transgenic mice, it is still not known whether Treg populations play a key pathogenic role in these mice.
TGF-β1 has an unclear role in the pathogenesis of human psoriasis. In one study of 60 patients with psoriasis and 38 without psoriasis, serum TGF-β1 was increased in psoriatics than in controls, and higher TGF-β1 correlated with higher clinical disease scores (Nockowski et al., 2004). Another study showed a correlation of disease severity with concentrations of TGF-β1 in psoriatic scales (Flisiak et al., 2002). Several other studies, however, have shown a negative or absent correlation of expression of TGF-β1 or its receptors in psoriasis (Leivo et al., 1998; Doi et al., 2003). These published studies are insufficient to determine definitely whether there is a positive or negative correlation between TGF-β1 and psoriasis. Indeed, recent studies showed that TGF-β1 is not involved in the differentiation of human Th17 cells (unlike in the development of murine Th17 cells)(Acosta-Rodriguez et al., 2007a, b), and that there is no association between TGF-β1 polymorphisms and psoriasis susceptibility (Baran et al., 2007). Taken together, our current study, as well as these recent studies in humans, suggests that TGF-β1 does not play a major role in the pathogenesis of psoriasis.
All breeding and experiments were undertaken with review and approval from the Portland Veterans Affairs Medical Center Institutional Animal Care and Use Committee. K5.TGF-β1 transgenic mice were a kind gift from Dr Xiao-Jing Wang, University of Colorado (Li et al., 2004). C57BL/6 mice were purchased from Jackson Laboratories (Bar Harbor, ME). To ensure the proper genetic identification of mice, DNA was isolated from tails of mice at 3 weeks of age using a DNeasy Blood and Tissue Kit (Qiagen, Valencia, CA), according to the manufacturer’s instructions. PuReTaq Ready-To-Go PCR beads (GE Healthcare, Piscataway, NJ) were used for DNA amplification, following the manufacturer’s instructions. Primer sequences and product sizes were as follows: K5.hTGF-β1 forward, 5′-GCGTCTGCTGAGGCTCAAGTT-3′ and reverse 5′-ACCTCGGCGGCCGGTAG-3′, 271 bp. The annealing temperature for K5.hTGFβ1 primers was 54°C.
All mice were monitored from birth to determine the extent, if any, of skin or systemic disease. K5.hTGF-β1 transgenic mice developed a visible skin disease between 2 and 4 months of age, with disease being somewhat variable among individual animals. A skin-scoring system was instituted to standardize all experiments involving cytokine levels, treatment of established disease, and genetic crosses. Scores were assessed as follows: 0.5 = altered hair growth; 1 = thinning of ventral and dorsal hair and mild scaling on the upper back; 2 = <50% areas with hair loss, widespread scaling, and Koebner’s phenomenon near ear tags; 3≥50% areas with hair loss and erythematous scaly plaques; and 4≥50% areas with hair loss, erythematous scaly plaques, retracted eyelids, and dehydration. All animals that were used for measurement of cytokine levels were between 4 and 6 months of age. The mice were euthanized at the end of the experiments by CO2, on >10% weight loss, or if showing obvious dehydration. The mice were housed in specific pathogen-free facilities and given food and water ad libitum.
Clinically abnormal (that is, lesional) and clinically normal (that is, non-lesional) skins were obtained from C57BL/6 and K5.hTGF-β1 transgenic mice, snap frozen in liquid nitrogen, placed into Trizol (Sigma, St louis, MO), and homogenized with a mechanical rotor for 30 seconds. RNA was isolated according to standard Trizol protocol, and was followed by further purification using an RNeasy kit (Qiagen). cDNA was prepared from 1 μg of total RNA by reverse transcription using iScript (Bio-Rad, Hercules, CA). Five reactions were pooled and diluted 5 times with water to establish the same pool of cDNA for all qRT-PCR experiments. qRT-PCR was performed with TaqMan primers and fluorescent probes for GAPDH, hTGF-β1, IL-23p19, IL-12/23p40, IL-12/35p35, IL-23R, IL-6, IL-17A, IL-22, IFN-γ, RORγt, CCR6, IL-4, foxp3, and IL-10 (Applied Biosystems, Carlsbad, CA) on the MyiQ system (Bio-Rad). Relative quantification by the delta–delta Ct method was carried out in Excel, using GAPDH as the housekeeping gene. To generate logarithmic graphs, 100 was added to all observations and the log of each was calculated. WT results were fixed at two (log 100). Gains relative to WT were shown as greater than two and losses were shown as less than two.
Sera, lesional skin, and non-lesional skin were obtained from C57BL/6 and K5.hTGF-β1 transgenic mice. Skin was snap frozen in liquid nitrogen, pulverized when immersed in liquid nitrogen, and subsequently re-suspended in a protein lysis buffer (10 mm Tris-HCl, pH 7.5; 0.5 mm EDTA-Na2; 0.5 mm EGTA; 1% Triton X-100; 0.5 mm PMSF; and Protease Inhibitor Mix, diluted from 100× stock (Sigma)). The samples underwent three freeze–thaw cycles and were spun at full speed for 15 minutes to fractionate the sample and remove protein. Human TGF-β1 and murine IL-17A, IL-22, IL-1α, TGF-β1, and IL-12/23p40 ELISA kits (R&D Systems, Minneapolis, MN), the IL-23 p19/p40, IFN-γ, and IL-4 ELISA kits (eBiosciences, San Diego, CA), and the IgE ELISA kit (Biolegend, San Diego, CA) were used according to the manufacturer’s instructions.
Lymph nodes from WT, K5.hTGF-β1 transgenic, or IL-17A knockout mice were harvested and cultured for 3 days in RPMI-1640, 5% FBS, 2mm l-glutamine, and 100 U ml−1 penicillin/100 μg ml−1 streptomycin. Cells were cultured alone or in the presence of IL-17-promoting cytokines or antibodies, including hTGF-β1, IL-6, anti-IFN-γ, and recombinant murine IL-23 (Li et al., 2007). The cells were plated at a density of 5 × 10−6 cells per ml in six-well plates. Then, 150 μl aliquots of cell culture supernatant, to use as protein assay controls, were taken at 24-hour intervals, beginning at 12 hours. The culture medium was supplemented with 1 ml new medium on day 3.
Neutralizing mAbs directed against murine IL-23p19 (CNTO 6163) and IL-12/23p40 (CNTO 3913), as well as murine IgG isotype control antibody (CNTO 1322), were obtained from Centocor R&D (Radnor, PA). Generation and in vitro and in vivo characterization CNTO 3913 and CNTO 1322 have been described elsewhere (Held et al., 2008). CNTO 6163 is a rat (v)/mouse(c) IgG2aκ chimeric antibody developed by Centocor using variable regions from the neutralizing rat anti-mouse IL-23p19 antibody, CNTO 209. Functional variable region genes were identified from the hybridoma cell line and the heavy-chain and light-chain variable-region genes were cloned into murine IgG2a and Kappa expression vectors, respectively. Both single-gene vectors were combined into a single double-gene vector, which was transfected into GS-CHO cells to generate a stable cell line expressing the CNTO 6163 antibody.
A concentration of 1 mg of mAb per animal was used. The animals were paired by disease score and birth month and placed into the anti-IL-23p19, anti-IL-12/IL-23p40, or isotype control Ab treatment groups. Intraperitoneal (i.p.) injections were given once weekly for 3 weeks. Weight changes were monitored weekly, and clinical skin disease was scored between days 0 and 21. The animals were killed on day 21, and lesional and non-lesional skin samples were prepared for routine histology.
Female SJL mice (8–12 weeks old) were immunized subcutaneously with PLP139-151 peptide (Genscript, Piscataway, NJ) emulsified in CFA, supplemented with 2 mg ml−1 Mycobacterium tuberculosis (Difco, San Jose, CA). The mice (n = 3 per group) were injected i.p. with 1 mg anti-IL-23p19 mAb, anti-IL-12/23p40 mAb, or isotype control IgG at days −1 and 6. Clinical disease was assessed daily using a scoring system as follows: 0 = normal; 0.5 = limp tail; 1 = hind limb weakness; 1.5 = inability to right>; 2 = walks with difficulty; 2.5 = hind limb paresis; 3 = complete hind limb paralysis; 3.5 = hind limb paralysis with front limb weakness; 4 = hind limb paralysis with front limb paresis; and 5 = moribund/death. Disease was scored daily until day 15.
Statistical analyses for qRT-PCR results were carried out using a two-tailed t-test, with four to nine degrees of freedom and 95% confidence intervals for graphing. Statistical analyses for protein results comparing WT with transgenic samples were carried out using a one-way analysis of variance with Tukey’s post test, except for the serum IgE results, which were compared using an unpaired t-test.
This work was supported by a Veterans Affairs Merit Award (A.B.) and US National Institutes of Health Grant nos. R21 AR054495-01A1 (A.B.) and T32 CA106195-04 (E.L.F.). We greatly appreciate receiving the K5.hTGF-β1 transgenic mice from Dr Xiao-Jing Wang at the University of Colorado. We also thank Michael Lasarev for help with statistical analysis of the RT-PCR data and Jill M. Carton for helping produce the Centocor monoclonal antibodies used in this study.
CONFLICT OF INTEREST
Dr A Blauvelt is a scientific consultant and clinical investigator for Centocor and Abbott. Jacqueline Benson and Wei Gao are employees of Centocor.