The integrin αEβ7 is thought to play an important role in the localization of mucosal, but not of cutaneous T lymphocytes. Thus, it was surprising that 89% of adult αE−/− mice on the 129/Sv × BALB/c background developed inflammatory skin lesions without an apparent infectious etiology. Skin inflammation correlated with αE deficiency in mice with a mixed 129/Sv × BALB/c background, but not in mice further backcrossed to BALB/c and housed in a second animal facility. These studies suggested that αE deficiency, in combination with other genetic and/or environmental factors, is involved in lesion development. The lesions were infiltrated by CD4+ T cells and neutrophils, and associated with increased expression of inflammatory cytokines. Furthermore, skin inflammation resulted from transfer of unfractionated αE−/− splenocytes into scid/scid mice, but not from transfer of wild-type splenocytes, suggesting that the lesions resulted from immune dysregulation. We also studied the role of αEβ7 in a murine model of hyperproliferative inflammatory skin disorders that is induced by transfer of minor histocompatibility-mismatched CD4+/CD45RBhigh T cells into scid/scid mice under specific environmental conditions. Under housing conditions that were permissive for lesion development, transfer of αE-deficient CD4+/CD45RBhigh T cells significantly exacerbated the cutaneous lesions as compared with lesions observed in mice reconstituted with wild-type donor cells. These experiments suggested that αE-expressing cells play an important role during the course of cutaneous inflammation. In addition, they suggest that αEβ7 deficiency, in combination with other genetic or environmental factors, is a risk factor for inflammatory skin disease.
The integrin αE (CD103) β7 is expressed by T lymphocytes in or adjacent to mucosal epithelia, including 90% of CD8+ intestinal T cells in or adjacent to the small intestinal epithelium, and 40–50% of the CD4+ T cells within the intestinal lamina propria. It binds to E-cadherin (1, 2, 3) on epithelial cells and may be important in the localization of diffusely distributed T cells within the intestinal epithelium. Integrin αEβ7 also is found on dendritic epidermal T cells, some dendritic cells, some lymphomas and leukemias, and bone marrow-derived mast cells (4, 5, 6, 7, 8, 9, 10) cultured in the presence of TGF-β1, a cytokine that also induced αEβ7 expression on T lymphocytes (1, 11). Like other integrins, the avidity of the interaction between αEβ7 and E-cadherin can be regulated through inside-out signaling in response to T cell stimulation with anti-CD3 (1, 11). In addition, αEβ7 appears to transmit an outside-in signal, as some anti-αE mAbs enhance proliferation of T cells in response to suboptimal concentrations of anti-CD3 or induce the lysis of Fc receptor-bearing targets by αEβ7-expressing T cells (7, 8). Thus, αEβ7 may also function to modulate the response of T lymphocytes to stimulation by epithelial cells, akin to the important accessory role of the integrin CD11a/CD18 (LFA-1) during peripheral T lymphocyte/APC interactions (12, 13, 14).
Consistent with the proposed function of αEβ7 in intraepithelial retention of mucosal T lymphocytes, we recently reported that integrin αE-deficient mice have reduced numbers of T lymphocytes diffusely distributed within the intestinal and vaginal epithelia (3). However, expression of E-cadherin, the known ligand for αEβ7, is not restricted to mucosal epithelial cells, but is also found on epithelial cells in the skin. Furthermore, expression of αEβ7 has been observed in some cutaneous disorders, such as T cell lymphomas, lichen planus, or atopic dermatitis (4, 15, 16, 17). However, it is unclear whether αEβ7 contributes to epidermal localization of T lymphocytes, and a role of integrin αEβ7 in the cutaneous immune system and/or the generation of inflammatory skin disorders has not been described.
In this study, we report the surprising observation that nearly all αE (CD103)-deficient mice spontaneously developed inflammatory ulcerative skin lesions in our initial cohort. This observation prompted further investigations toward a potential function of αEβ7 on cutaneous T cells and/or in the pathogenesis of inflammatory skin disorders. Indeed, we found that αEβ7 significantly influences the generation of hyperproliferative inflammatory skin alterations in mice. These results suggest that the αEβ7 integrin performs a function on T cells localized in the skin and may play a role in hyperproliferative inflammatory skin disorders.
Materials and Methods
The generation and genotyping of mice deficient for the αE integrin subunit (CD103) have been described previously (3). In reconstitution experiments, C.B-17/lcrTac-scidfDFscid/scid (scid/scid) mice (Taconic Farms, Germantown, NY) of 5–8 wk of age were used as recipients. Mice were kept under specific pathogen-free conditions in microisolator cages. Microbiologic assessment of αE−/− and αE+/+ mice included: Gram stain of inflamed skin and eyelids; bacterial and fungal cultures of blood, inflamed and uninflamed skin, and eyelids; and bacterial cultures of nasopharyngeal washes that revealed growth of resident flora including α-streptococci, Pasteurella aerogenes, Pasteurella multocida, Staphylococcus epidermidis, and Streptococcus acidominimus. Additional evaluation included anal tape test, fecal flotation, and microscopic pelage for metazoan parasites, and serologies for murine viral pathogens including Sendai virus, pneumonia virus of mice, mouse hepatitis virus, minute virus of mice, GD-7 virus, Reo-3 virus, lymphocytic choriomeningitis virus, mouse adenovirus, mouse pox virus (ectromelia virus), K-virus, polyoma virus, epizootic diarrhea of infant mice, mouse CMV, mouse thymic lymphocyte virus, and Hantaan virus, and for procaryotic organisms including cilia-associated respiratory bacillus, Encephalitozoon cuniculi, and Mycoplasma pulmonis.
The following murine Ags were detected by mAbs: CD3ε (500A2; PharMingen, San Diego, CA), CD4 (RM4-5; PharMingen), CD8α (53-6.72; American Type Culture Collection (ATCC), Manassas, VA), CD11b (αM integrin, Mac-1, M1/70; ATCC), CD18 (β2 integrin, 2E6; ATCC), CD25 (IL-2R α-chain, 3C7; PharMingen), CD45R/B220 (RA3-6B2; PharMingen), CD45RB (MB23G2, ATCC; 16A, PharMingen), CD49f (α6 integrin, GoH3; Dianova, Hamburg, Germany), CD54 (ICAM-1, YN1/1.7.4; ATCC), CD103 (αE integrin, M290, P. Kilshaw (Department of Immunology, AFRC Babraham Institute, Cambridge, U.K.) (5), or 2E7, L. Lefrancois (Department of Medicine, University of Connecticut Health Center, Farmington, CT) (8)), β7 (M293; P. Kilshaw (Department of Immunology, AFRC Babraham Institute, Cambridge, U.K.)), CD106 (VCAM-1, M/K-2.7; ATCC), CD32/CD16 (Fc-γII/IIIR, 2.4G2; ATCC), H-2Dd (34-2-12; PharMingen), H-2Kb (AF6-88.5; PharMingen), MHC class II (M5/114.15.2 or N22; ATCC), IFN-γ (XMG1.2; PharMingen), IL-4 (BVD6-24G2; PharMingen), IL-6 (MP5-20F3; PharMingen), GM-CSF (MP1-22E9; PharMingen). In addition, rabbit antisera reactive with murine IL-1α and TNF-α (Genzyme, Cambridge, MA) were used. The following mAbs were used as controls: rat IgG1 (R59-40; PharMingen), rat IgG2a (R35-95; PharMingen), rat IgG2b (SFR3-DR5; ATCC), and hamster IgG (UC8-4B3; PharMingen). FITC anti-CD45RB (mAb 16A) and PE anti-CD4 (mAb RM4-5) were obtained from PharMingen; biotinylated goat anti-hamster, mouse-adsorbed rabbit anti-rat and goat anti-rabbit serum were obtained from Vector Laboratories (Burlingame, CA); and goat anti-rat Microbeads were purchased from Miltenyi Biotec (Auburn, CA).
Histopathologic analysis and in vivo 5-bromo-2′-deoxyuridine (BrdU)3 labeling
For histochemical analysis, tissue samples were fixed, dehydrated, and embedded in JB-4 plastic resin according to the manufacturer’s instructions (Polysciences, Warrington, PA). Sections (3 μm) were stained with hematoxylin-eosin or for chloroacetate esterase reactivity, as described elsewhere (18, 19). Samples were evaluated by an investigator blinded to the mouse genotype. For immunohistochemistry, 5-μm acetone-fixed cryostat-cut sections were stained by indirect immunoperoxidase using an avidin-biotin complex (ABC) kit (Vector Laboratories) using 3-amino-9-ethylcarbazole (Aldrich, Milwaukee, WI). To detect proliferating cells, mice were injected i.p. with 5 mg BrdU in 500 μl PBS at both 9 and 6 h before sacrifice. JB-4 plastic-embedded sections were immersed in 0.03% H2O2 in methanol and denatured with 0.4% pepsin (Sigma, St. Louis, MO) in 0.1 N HCl and 0.8 N HCl. Sections then were stained with anti-BrdU (Becton Dickinson, Hamburg, Germany) and the ABC immunoperoxidase method. All immunohistochemical stainings were performed using tissue sections from at least three mice.
Cell purification and reconstitution of scid/scid mice
PBMCs from αE+/+ and αE−/− mice were isolated by density gradient centrifugation and H-2 typed, and mice homozygous for the H-2Dd haplotype were used as donors. For adoptive transfer experiments, splenocytes were isolated and passed over a glass wool column to remove debris and to enrich for T cells. To enrich for CD4+ or CD8+ T lymphocytes, negative selection was performed with anti-CD4 or anti-CD8 mAbs using a magnetic cell separation system according to the manufacturer’s instructions (Miltenyi Biotec, Cologne, Germany). The resulting populations were ∼85% pure. In addition, contamination of the CD4+ population with CD8+ cells or of the CD8+ population with CD4+ cells was <1% in each case, as determined by FACS. C.B-17/IcrTac-scidfDFscid/scid recipient mice were injected i.v. with 1.7–2 × 107 of unfractionated splenocytes or with 2.2 × 106 CD4+- or CD8+-enriched T cells. The ear thicknesses of the recipient scid/scid mice were monitored over 6 wk, and tissues were harvested for histopathologic analysis at the end of the observation period.
For generation of psoriasiform skin lesions by reconstitution of scid/scid mice with CD4+/CD45RBhigh T lymphocytes (18), splenocytes from αE+/+ or αE−/− donors that were MHC matched, but minor histocompatibility Ag mismatched, were purified as described previously (19, 20, 21). Briefly, splenocytes were incubated with 20 μg each of anti-B220, anti-αM integrin, anti-CD8α, and anti-MHC class II mAbs/107 cells, followed by 20 μl goat anti-rat microbeads/107 cells and stained cells were depleted using a magnetic cell separation system separation column (type CS; Miltenyi Biotec). The CD4+-enriched population then was incubated for 30 min with 7.5 μg PE-conjugated rat anti-CD4, and 12.5 μg cells FITC-conjugated anti-CD45RB/107 cells. Using a FACSvantage (Becton Dickinson), the 40–45% of cells expressing the highest levels of CD45RB (naive T cells) (22) were selected from the CD4+ population (hereafter referred to as CD4+/CD45RBhigh T cells). scid/scid recipient mice were injected with 2–2.5 × 105 cells each. Recipient mice started to develop environmentally modulated psoriasiform skin lesions within 3–4 wk (19, 23). Mice were weighed at weekly intervals. To evaluate the overall skin inflammation, the ear thicknesses were monitored using a skin thickness gauge (Oditest; Dyer, Lancaster, PA), and a clinical score for the psoriasis-like skin disorder was applied as described previously (19). Of note, mice developed lesions in this model when they were fed mouse breeder chow with 9% fat, kept in Pine Soft Wood Chips (100% virgin), and under static air in microisolator cages with weekly airflow. However, lesions did not develop if they were fed Prolab RMP3000, kept in Bed-o-cob bedding with continuous air flow in the same animal facility, after transfer was performed at the same time (data not shown).
Statistical significance was assessed by the paired two-tailed Student t test, and p < 0.05 was considered to be significant.
Spontaneous inflammatory autoimmune skin lesions in αE-deficient mice
To evaluate the impact of αE deficiency on the overall development and life span of mice, wild-type (αE+/+), heterozygous (αE+/−),and homozygous-deficient (αE−/−) mice on the 129/Sv × BALB/c background were monitored at weekly intervals until senescence. Surprisingly, spontaneous inflammatory skin lesions were observed in this cohort of αE−/− mice at ∼6 mo after birth, and almost all animals >9 mo old were affected (16/18; 88.9%). Heterozygous mice were affected with an intermediate frequency (4/30; 13.3%), consistent with their expression of reduced αEβ7 levels on T lymphocytes (3), and wild-type mice were not affected (0/8; 0.0%) (Fig, 1a). The inflammatory skin lesions were most commonly seen on the eyelids, but were also observed on trunk and tail, and were manifested by chronic inflammation resulting in ulceration in some cases (Fig. 1 a). The course of these lesions was variable, as some lesions spontaneously resolved over time, whereas others worsened. Spontaneous skin lesions developed independently in mice kept in two different animal facilities. When the αE-deficient mice were further backcrossed to the BALB/c strain and housed in the second animal facility, skin lesions also occurred in some wild-type mice and were not clearly correlated with αE genotype. Thus, modifying factors besides αEβ7 contributed to the development of the skin lesions. Indeed, one such factor may be environment.
Histologic analysis revealed normal skin in wild-type mice (Fig. 1,b, left panel), but subclinical skin lesions in 100% of αE−/− mice (10/10 mice, 13–25 mo of age (mean, 18 mo), including two mice without macroscopic lesions; Fig. 1,b, middle and right panels). Within the subclinical lesions, a dense mixed leukocytic infiltrate was noted, composed of lymphocytes and granulocytes (Fig. 1, b and c). Although few CD8+ T cells were found in these lesions, there were abundant CD4+ T cells both within the dermal and the epidermal compartment, some of which expressed the activation marker CD25 (high-affinity IL-2R α-chain) (Fig. 2,a). Epidermal changes included marked acanthosis (increase of viable cell layers) and moderate ortho-hyperkeratosis (thickening of the stratum corneum) (Fig. 1,b), de novo expression of ICAM-1 and MHC class II by keratinocytes, and suprabasal expression of integrin α6 (Fig. 2,b). In addition, there was an increase in number and size of dermal blood vessels (Figs. 1,b and 2b). In clinically apparent lesions, ulcers were surrounded by a marked reactive epidermal hyperplasia and contained a dense infiltrate of lymphocytes and neutrophils (Fig. 1,c). In comparison with uninvolved skin, expression of IFN-γ, GM-CSF, IL-6, IL-4 (Fig. 2 c), IL-1α, and TNF-α (data not shown) was increased dramatically in the skin lesions of αE−/− mice, suggesting a mixed Th1/Th2 response.
Studies were performed to evaluate for a possible infectious organism that might be important in the pathogenesis of these lesions. In this analysis, Gram stains did not reveal an increase in bacteria in lesions with an intact epidermis from αE−/− mice as compared with the colonization observed in normal-appearing skin from either αE+/+ or αE−/− mice. In addition, bacterial and fungal blood cultures were sterile, and the culture of inflamed skin or eyelids and of nasopharyngeal washes revealed only commensual organisms, which were similar in αE+/+ and αE−/− mice. Finally, viral and bacterial serology were negative in αE+/+ and αE−/− mice, which have been demonstrated to have normal levels of Igs and Ig isotypes (3). Thus, there was no apparent infectious pathogen to account for these lesions, leading us to consider whether the lesions were autoimmune in pathogenesis.
Induction of inflammatory skin lesions in scid/scid mice through adoptive transfer of αE-deficient T lymphocytes
To further evaluate the cutaneous lesions resulting from αE deficiency, adoptive lymphocyte transfer studies were performed into MHC-matched scid/scid mice. In these studies, recipient mice that had been reconstituted with 1.7–2 × 107 αE+/+ splenic leukocytes appeared to be normal and healthy and did not develop cutaneous lesions during the 6-wk observation period. In contrast, 100% of mice reconstituted with splenocytes from αE−/− donors developed inflammatory skin lesions associated with a 50–60% increase in ear thickness (0.438 mm (n = 3, SD = 0.019) in recipients of αE-deficient donor cells vs 0.285 mm (n = 3, SD = 0.005) in recipients of wild-type cells, p = 0.007). This was especially apparent on the ears (data not shown), feet, and surrounding the eyes (Fig. 3). These lesions were characterized by an infiltration of leukocytes, including CD3+/CD4+ T cells (which were presumed to be donor cells, since scid/scid recipients have very few CD3+ T cells) and neutrophils. Epidermal changes included acanthosis and hyperkeratosis (Fig. 3,b) and ulcer formation in ∼30% of the recipients (one of three animals in each of two independent experiments). Similar to the spontaneous lesions in αE-deficient mice, very few CD8+ T cells were present within the inflammatory lesions in scid/scid recipients reconstituted with αE-deficient splenocytes (Fig. 3,b), suggesting a primary pathogenic role of CD4+ αE-deficient T cells. The inflammatory skin lesions in scid/scid mice first were observed 2 wk after adoptive transfer and persisted during the 6-wk observation period (Fig. 3 c). Although ulcers did not develop consistently and the skin was more diffusely involved, the cutaneous inflammation in mice reconstituted with αE−/− splenocytes was reminiscent of the spontaneous lesions in aging αE−/− mice. Thus, transfer of spleen-derived leukocytes from αE-deficient mice into MHC-matched scid/scid mice initiated skin lesions in immunodeficient recipients appeared to be induced by CD4+, but not CD8+ T cells. This hypothesis was supported by an additional adoptive transfer in which inflammatory skin alterations were observed similarly in mice reconstituted with αE-deficient unfractionated splenocytes, purified T cells, or purified CD4+ T cells, but not in recipients of CD8+ T cells (n = 3 in each group, data not shown).
Integrin αEβ7 is involved in the generation of murine hyperproliferative inflammatory skin lesions
To further test the hypothesis that αEβ7 is involved in the pathogenesis of hyperproliferative inflammatory skin alterations, a T cell-mediated mouse model of a psoriasiform skin disorder was used (18). In this system, transfer of MHC-matched but minor histocompatibility-mismatched αE+/+ CD4+/CD45RBhigh donor cells induced psoriasiform skin inflammation in scid/scid mice in addition to the intestinal inflammation that is known to follow transfer of syngeneic CD4+/CD45RBhigh donor cells (19, 20). This inflammatory skin disease model is environmentally influenced. Within the murine psoriasiform lesions, epidermal T cells expressed αEβ7 (Fig. 4,a). To determine whether the αEβ7 integrin was involved in the pathogenesis of the murine hyperproliferative inflammatory skin disorder, the development of these lesions was compared in scid/scid mice reconstituted with CD4+/CD45RBhigh T cells derived from wild-type or αE-deficient mice. In both cases, abundant CD3+ donor T cells infiltrated the skin of the recipients. As expected, there was prominent expression of αEβ7 on epidermal T cells in recipients of wild-type CD4+/CD45RBhigh donor cells, but not in recipients of αE-deficient donor cells (Fig. 4,a). Interestingly, the psoriasiform skin inflammation was worsened significantly when αE−/− donor cells were used, as assessed by both clinical score and ear thickness (Fig. 4, b and c). In addition, in vivo BrdU incorporation in epidermal keratinocytes was 44% higher in recipients of αE−/− donor cells as compared with recipients of wild-type cells, indicating stronger hyperproliferation (288 BrdU+ cells/mm (±20.89) vs 200.1 (±23.03), n = 3, Fig. 4 d). These findings suggest that αE deficiency on the donor cells worsened the disorder induced by CD4+/CD45RBhigh T cells. However, it is also possible that αE deficiency induced skin changes that were superimposed upon the psoriasiform alterations that result from T cell transfer.
In this study, we found that aging αE-deficient mice developed inflammatory skin lesions, apparently due to an aberrant immune reaction, as transfer of T cells from αE−/− mice induced a similar phenotype in scid/scid recipients. Integrin αEβ7 expression has been observed on some epidermal T cell malignancies (4, 15, 24) and on T lymphocytes in some benign skin conditions (16, 17). However, it is not generally thought to play an important role in the localization or function of cutaneous T cells. Thus, our observation was surprising and suggests a broader role of the αEβ7 integrin than had been previously appreciated. As E-cadherin, the known αEβ7 counterreceptor (1, 2), is constitutively expressed within the epidermis (25), it is likely that αEβ7 expressed on cutaneous T cells binds to its ligand there. This interaction may serve to enhance the immigration or retention of αEβ7+ cutaneous T cells in the epidermis. This would be consistent with the expression of αEβ7 on a greater proportion of the epidermal T cells than of the dermal T cells in cutaneous lesions. The adhesive functions of αEβ7 on cutaneous T cells are likely to be mediated in the context of other adhesive molecules. These might include the binding of β1 integrins, LFA-1, or as yet unknown molecules on cutaneous T cells with ICAM-1 or the newly identified lymphocyte endothelial epithelial cell adhesion molecule (LEEP-CAM) on keratinocytes (26, 27, 28, 29, 30). In addition, it has been suggested that there is another ligand for αEβ7 on keratinocytes that has not yet been identified at the molecular level, which could be important in keratinocyte interactions with αEβ7+ leukocytes (31).
The αEβ7–E-cadherin interaction could also result in outside-in signaling in response to ligand binding by either αEβ7 on T cells or E-cadherin on epithelial cells. Specifically, a role for αEβ7 in signaling has been indicated by evidence that anti-αEβ7 mAbs enhance the proliferation of T lymphocytes in response to suboptimal concentrations of anti-CD3 (7) and can induce the lysis of target cells in vitro (6, 8). In addition, E-cadherin is known to transmit signals that are important in regulating epithelial cell functions (32, 33, 34). Thus, intracellular signals induced by αEβ7/E-cadherin binding in either T cells or epidermal keratinocytes could modulate the cytokine responses or differentiation, proliferation, or survival of either cell type. Consistent with this possibility, the presence of αEβ7+ T lymphocytes adjacent to acinar epithelial cells has been found to correlate with epithelial cell apoptosis in tissue samples derived from Sjögren’s syndrome patients (35).
Interestingly, integrin αE deficiency also worsened murine psoriasis-like skin lesions induced by CD4+/CD45RBhigh T lymphocytes in environmental conditions that were permissive for lesion development. This was in contrast to the observation that αE deficiency alleviated intestinal inflammation that occurred concomitantly in these mice (M.P.S., J.P.D., and C.M.P., unpublished observations). There are several potential mechanisms by which αE deficiency could modify the phenotypic outcome of partial immune reconstitution in immunodeficient mice. First, the lack of αE expression might prevent some T cell subsets from localizing to the intestinal epithelium. These cells might then be shunted to the skin, where they would induce inflammation due to their production of inflammatory cytokines (36). In addition, αE deficiency could result in the loss of a regulatory signal transmitted through the αEβ7 integrin that modulates T cell responses. Such signaling might modify the response of cutaneous T cells to stimulation, and thereby result in T cell dysregulation in the skin under conditions in which the T cells become activated.
Overall, our results demonstrate a novel and unexpected function of the αEβ7 integrin within the cutaneous immune system. These studies also suggest that αE deficiency may be a genetic factor predisposing to inflammatory skin disease under some environmental conditions and/or in the context of other genetic factors, based upon the presence of inflammatory skin lesions in αE−/− mice. This possibility is supported by the observation that hyperproliferative inflammatory skin lesions in a murine model of psoriasis were exacerbated when the transferred T cells lacked αE expression. Additional genetic studies will be required to determine whether αE deficiency is associated with an increased risk of inflammatory skin disease in humans, as it appears to be in mice, and to identify the role of αEβ7 in cutaneous T lymphocyte responses during antigenic or autoimmune responses.
We thank E. Meluleni, J. Connolly, X. Hu, and R. Kubitza for excellent technical assistance.
This work was supported by grants from the Deutsche Forschungsgemeinschaft (Scho 565/2-1) and the Forschungskommission of the Heinrich-Heine-University (to M.P.S.), a National Research Service institutional training grant from the National Institutes of Health (to J.P.D.), a grant from the Deutsche Forschungsgemeinschaft (to M.S.); and grants from the National Institutes of Health (DK52978) and the Crohn’s and Colitis Foundation, a Cancer Research Institute Investigator Award, and a Pilot and Feasibility grant from the Harvard Skin Disease Research Center (to C.M.P.).
Abbreviation used in this paper: BrdU, 5-bromo-2′-deoxyuridine; ABC, avidin-biotin complex.