Integrin α_E(CD103)β₇ influences cellular shape and motility in a ligand-dependent fashion

Positioning and locomotion of leukocytes within tissues provide the basis for the molecular crosstalk with other cells and are prerequisites for a functional immune system. To date, the recruitment cascade of initial endothelial adhesion, activation, firm adhesion, transmigration, and subsequent localization into the connective tissue is one of the best established concepts in leukocyte biology.^1,2  However, many effector functions of immunocytes are exerted within parenchymatous organs, mostly epithelia. It is therefore somewhat surprising that we know relatively little about locomotion and function-determining morphogenesis of lymphocytes within epithelial tissues, such as the epidermis of the skin.³ 

An adhesion receptor that is thought to mediate retention of lymphocytes within epithelial tissues is integrin α_E(CD103)β₇.⁴  First described 2 decades ago as a selective marker for intestinal intraepithelial lymphocytes,⁵  α_E(CD103)β₇ has been implicated in epithelial T-cell retention through binding to E-cadherin.^6,7  Indeed, α_E(CD103)-deficient mice exhibited a reduced number of mucosal intraepithelial T cells.⁸  However, α_E(CD103)β₇ has later been found to be also expressed by some lymphocytes within other epithelia, such as the epidermis of the skin,⁹  where it presumably contributes to recruitment of T cells in inflamed human skin¹⁰  as well as dendritic epidermal T cells (DETCs) in murine skin.¹¹  Thymic DETC precursor cells express integrin α_E(CD103)β₇ before their migration into the periphery, suggesting that α_E(CD103)β₇ is involved in guiding tissue-specific epidermal localization of DETCs.¹²  Expression of α_E(CD103)β₇ has been described on some CD4⁺CD25^+13,14  and CD8^+15,16  regulatory T cells (T_reg). In several experimental models α_E(CD103)β₇ is involved in guiding tissue localization of lymphocyte subsets in inflammatory conditions and/or allograft rejection.^14,17,18  Although α_E(CD103)β₇ is highly expressed by some lymphocytes and CD11^high/MHC-II^high dendritic cells at mucosal and other epithelial sites, its role in immune regulation remains largely elusive. Expression of α_E(CD103)β₇ appears to be associated with cytotoxic activity of CD8⁺ T cells in graft-versus-host disease and allogeneic transplantation.^19,20  Along this line, interactions of α_E(CD103)β₇ with E-cadherin are crucial for lysis of E-cadherin⁺ tumor cells by cytotoxic T cells in some cases.²¹  Moreover, expression of α_E(CD103)β₇ is associated with several important cellular activities such as antigen presentation²²  or stimulation of T_reg²³  and some mucosal CD8⁺ T cells^24,25  by α_E(CD103)β₇⁺ dendritic cells. It is, however, currently unclear whether and to what extent α_E(CD103)β₇ itself contributes directly to such functions.

Overall, a robust body of evidence has accumulated indicating that α_E(CD103)β₇ is involved in tissue-specific retention and/or effector functions of some immune cells. Yet, it is still unclear whether α_E(CD103)β₇ is involved in other cellular functions such as fine-tuning of shape and motility, similar to what has been shown for some other integrins,²⁶  which are presumably crucial for the exertion of effector functions within epithelial compartments. We present here the first experimental data showing that integrin α_E(CD103)β₇ is indeed involved in sculpting the shape as well as stimulating the motility of cells on ligand contact. Thus, integrin α_E(CD103)β₇ also determines the ligand-directed shape and locomotion of cells, a novel function that extends beyond mere adhesion and retention on the substrate.

Yellow fluorescent protein was fused C-terminally to 3 different integrin α_E-constructs. The integrin α_E stop codon was eliminated by site-directed mutagenesis (primer: sense 5′-GTCTGCTCCAAGATCAGCCCCCTGCTTC; antisense 5′-GAAGCAGGGGGCTGATCTTGGAGCAGAC) using the QuikChange II XL Site-Directed Mutagenesis Kit (Stratagene, Amsterdam, The Netherlands). The YFP-gene was amplified by polymerase chain reaction (PCR) using YFP-primers linked with a NotI restriction site (sense 5′-AGGCTGTGAGGCGGCCGCTAATGGTGAGCAAGGGCGAG, antisense 5′-CTCTAGCGTTGCGGCCGCCTACTTGTACAGCTCGTC) and Platinum Pfx DNA Polymerase (Invitrogen, Karlsruhe, Germany). The PCR product as well as the 3-point mutated plasmids were digested with NotI (NEB, Frankfurt/Main, Germany) and ligated with T4 DNA Ligase (Invitrogen). The following PCR conditions were used: mutagenesis 95°C 1 minute, followed by 18 cycles 95°C 50 seconds, 60°C 50 seconds, 68°C 9 minutes; YFP 94°C 2 minutes, followed by 30 cycle, 94°C 15 seconds, 58°C 30 seconds, 68°C 1 minute. All plasmids were sequenced (SEQLAB, Göttingen, Germany).

K562 cells were cultured in RPMI and PAM212 cells were cultured in DMEM (PAA, Pasching, Germany). The Nucleofector Kit V (Amaxa, Cologne, Germany) was used for transient transfections. The transfection efficiency was determined by flow cytometry. For coating, slides were overlayered with 1 μg/mL recombinant murine E-cadherin (R&D Systems, Wiesbaden, Germany) in 2 mM CaCl₂/PBS over night at 4°C. Kollagenreagens Horm (NYCOMED, Linz, Austria) was diluted 1:1 in PBS and coated on cover slips overnight at 37°C. For antibody or cytochalasin D treatment the transfected K562 were incubated for 60 minutes with the 2E7 mAb (20 μg/mL) or cytochalasin D (0.5 μM), respectively. As determined by both trypan blue exclusion and flow cytometry assessing annexin V/propidium iodide staining, cytochalasin D treatment did not affect cell viability under the conditions used in our study (> 95% viable cells).

Cells were cultured for 90 minutes on E-cadherin–coated cover slips. Imaging was performed by serial 1.0 μm Z-steps using a TCS-4D microscope and SCANware software (Leica, Wetzlar, Germany). For coculture experiments, 1.5 × 10⁵ K562 and 4 × 10⁵ PAM212 were coincubated in RPMI. In other experiments, cryostat-cut sections or epidermal sheets were overlayered with transfected cells. Confocal microscopy was performed after 24 hours by serial 0.2 to 0.5 μm Z-steps using a DMI6000 microscope equipped with a TCS Sp5 scanner and LAS AF software (Leica Microsystems, Mannheim, Germany). Morphometric analyses were performed using ImageJ software (version 1.38; National Institutes of Health, Bethesda, MD).

Cells were allowed to attach to E-cadherin for 30 minutes, and were then monitored for 60 to 90 minutes using an Axiovert 200M microscope equipped with a LD Achroplan 20×/0.4 lens (Zeiss, Göttingen, Germany). Images were taken every 150 seconds using a CoolSnap ES camera (Photometrics, Tucson, AZ). Morphometric analyses were performed using MetaMorph Software (version 6.3r2; Visitron Systems, Puchheim, Germany). The following numbers of independent experiments were performed: mock-transfected cells, 7 experiments; wild-type α_E(CD103)/β₇, 3 experiments; α_E-open/β₇, 6 experiments; and α_E-closed/β₇, 4 experiments.

The isolation of animal material was approved by governmental authorities (Regierung von Unterfranken [District Government of Lower Franconia], Germany) and was performed according to institutional guidelines. Mice (C57/BL6 background) were maintained under specific pathogen-free conditions.

Mouse ears were incubated with dispase (Boehringer, Mannheim, Germany) at 6 mg/mL for 1 hour at 37°C. The epidermis was then peeled off and digested again for 5 minutes at 37°C under gentle stirring in dispase (6 mg/mL) and DNase (0.1%). The resulting single-cell suspension was subjected to 2-color flow cytometry using monoclonal antibodies directed against CD3ϵ (PE-conjugated), CD11a, CD18, CD25, CD29, CD44, CD49b, CD49d, CD49e, CD49f, and CD103 (FITC-conjugated; BD Pharmingen, Heidelberg, Germany). Dead cells were excluded by propidium-iodide staining. Cells were analyzed in a FACScan using CellQuest software (Becton Dickinson, Heidelberg, Germany).

Ears were removed and depilated. The split skin was floated for 20 minutes on 0.5 M NH₄SCN and the epidermal sheet was peeled off. After fixation in acetone for 20 minutes at −20°C and washing in phosphate-buffered saline (PBS), the epidermal sheets were soaked for 90 minutes with antimurine CD3ϵ (Becton Dickinson) followed by anti-hamster fluorescein isothiocyanate FITC (Vector Laboratories, Burlingame, CA) for 60 minutes. The sheets were analyzed using an Axioscop 2 microscope equipped with a Plan-Neofluar 40×/0.75 lens (Zeiss), and images were recorded using an AxioCam HRc (Zeiss). Morphometric analysis was performed using AxioVision AC software (Zeiss).

Data are displayed as mean (± SD or ± SEM as indicated), P values were determined using the 2-tailed t test, and P values less than .05 (confidence interval [CI] of 95%) were considered statistically significant. All statistical analyses were 2-sided.

Given that little is known about the role of α_E(CD103)β₇ for the motility of leukocytes, we addressed this issue by transfecting cells with the murine receptor, in which the α_E(CD103) chain was locked in different functional conformations by point mutations.²⁷  Unfortunately, transfection of several populations of primary murine lymphocytes under various conditions did not result in the expression of functional heterodimers, whereas control proteins were readily expressed (data not shown). In addition, Jurkat cells showed constitutive expression of the human β₇ integrin subunit, thus precluding functional studies of the murine heterodimer (data not shown). However, when the erythroleukemia cell line K562 was double-transfected with the wild-type α_E(CD103) and β₇ integrin subunits, surface expression and binding to E-cadherin were readily detected. The overall migration (determined by single-cell-tracking) on immobilized murine E-cadherin was similar in all transfectant lines expressing α_E(CD103)β₇ (mean95 μm/h, ± 18), while mock-transfected cells migrated significantly longer distances (average 159 μm/h ± 37, P < .001 compared with either of the α_E(CD103)β₇ transfectants). However, some remarkable differences became apparent regarding the motility of cells transfected with the different α_E(CD103) species at their site of attachment: Time-lapse microscopy demonstrated that expression of wild-type (α_E-WT) or a point-mutated α_E(CD103) species locked in a constitutively active conformation (α_E-open) resulted in a significantly increased ability of the cells to move in an amoeboid fashion on E-cadherin–coated surfaces and to extend cellular protrusions and filopodia in a highly dynamic manner. Integrin α_E-open/β₇ and α_E-WT/β₇ transfected K562 cells showed markedly more pronounced changes in morphology as compared with a constitutively inactive species (α_E-closed/β₇)or mock (pcDNA3.1) transfected cells (Figure 1A; Videos S1,S2, available on the Blood website; see the Supplemental Materials link at the top of the online article). Quantitative analyses demonstrated that the number of cells showing amoeboid motility was significantly increased by 50% (± 19) in cultures transfected with integrin α_E-WT/β₇ (Figure 1B) as compared with mock transfectants (P < .04; Figure 1B). When the cells were transfected with integrin α_E-open/β₇, the number of motile cells within the cultures was even increased by 85% (± 27, P < .001 compared with controls; Figure 1B). In addition, the magnitude of cellular movement and the size of the protrusions formed by the cells appeared to be markedly greater in transfectants expressing functionally active α_E(CD103)β₇ species as compared with the controls (visualized in Videos S1,S2). Transfection of integrin α_E-closed/β₇ did not result in enhanced motility of the transfectants (P = .5 compared with mock-transfected controls; Figure 1B). Culture of the cells on uncoated plastic surfaces, on collagen-coated surfaces, in the presence of α_E(CD103)-directed antibodies, or cytochalasin D completely abrogated the increases in amoeboid movement (Figure 1B). Thus, the observed gains of motility were strictly dependent on interactions of functional α_E(CD103)β₇ with E-cadherin and appeared to be relayed through the actin-based cytoskeleton.

To visualize the formation of cellular protrusions and filopodia in relation to α_E(CD103)β₇ expression, α_E-WT, constitutively active (α_E-open) and constitutively inactive (α_E-closed) species, respectively, were fused to YFP and transfected into K562 cells. The cells were cotransfected with the unlabeled β₇ subunit. As readily detectable by fluorescence microscopy and flow cytometry, all of the transfectants showed robust expression of the integrin α_E-YFP/β₇ fusion proteins (data not shown). The expected functional properties were readily exhibited by the transfectants: Using wild-type (α_E-WT/YFP/β₇) or the point-mutated species locked in a constitutively active conformation (α_E-open/YFP) resulted in significantly increased cell adhesion to E-cadherin (P < .001 in both cases as compared with mock-transfected cells), while expression of a constitutively inactive species (α_E-closed/YFP) led to significantly weaker binding (P < .01 compared with the control; Figure 2A), indicating the suitability of the YFP fusion proteins for further functional experiments. To further ascertain that fusion of YFP did not alter the function of α_E(CD103)β₇, additional control experiments were performed in which α_E(CD103) was fused to YFP using either 5 amino acid or 21 amino acid spacer sequences, both yielding essentially the same results in functional adhesion experiments (data not shown).

Serial stacks of the K562 cells analyzed by confocal microscopy demonstrated that numerous filopodia were extended by the cells in the contact zone to E-cadherin–coated surfaces. As demonstrated by α_E(CD103) specific antibody-mediated abrogation of filopodia formation, this function was strictly dependent on α_E(CD103)β₇ (Figure 2B). Strikingly, the number and length of the filopodia clearly depended on the conformational state of the α_E-YFP fusion protein transfected. Transfection of α_E-open/YFP resulted in formation of the largest filopodia, whereas transfection of α_E-WT/YFP led to an intermediate size, and cells expressing integrin α_E-closed/YFP exhibited the shortest filopodia (Figure 2C). Statistical analyses verified the significant difference (P < .02 comparing the ratios cell body/filopodia between α_E-open/YFP/β₇- and α_E-closed/YFP/β₇-transfected K562 cells). Control experiments using YFP-transfected cells on E-cadherin (Figure 2C) or integrin α_E(CD103)/YFP/β₇-transfected cells on collagen (not shown) did not provoke the formation of filopodia, further indicating that the filopodia of the integrin α_E(CD103)β₇-transfected cells were dependent on the specific interaction between α_E(CD103)β₇ and its ligand, E-cadherin.

Confirming the functional properties in an experimental system approaching the natural situation within the skin, the formation of filopodia induced by functionally active α_E(CD103)β₇ was also observed when K562 transfectants were cocultured with PAM212, an E-cadherin-expressing murine keratinocyte line.²⁸  In contrast, K562 cells expressing α_E-closed/YFP/β₇ showed only few and short filopodia, and K562 cells transfected with YFP alone did not show formation of filopodia when cocultured with PAM212 keratinocytes (Figure 2D).

To more closely mimic the in vivo situation where immune cells interact with epidermal cells organized in a polarized and stratified epithelium, we performed 2 complementary series of experiments: First, K562 cells expressing α_E(CD103)/YFP constructs and the wild-type β₇ subunit were seeded on cryostat-cut sections of murine skin. In the second series of experiments, epidermal sheets were prepared from mouse ears, and the sheets were placed upside-down into plastic dishes. The α_E(CD103)β₇ transfectants were then plated onto the exposed undersurface of the epidermal sheets. Indeed, in both types of experiments some of the transfectant cells expressing active α_E(CD103)β₇, but not the controls, formed conspicuous filopodia and protrusions extending clearly toward the epidermal keratinocytes, thus generating a striking dendritic cell-like morphology on intact epidermis (examples depicted in Figure 3A,B).

The relevance of our findings for properties of naturally occurring lymphocytes in vivo can be deduced from our further experiments in which we analyzed wild-type and α_E(CD103)-deficient mice.⁸  To address the question whether α_E(CD103)β₇ affected shape and morphology of natural cells in vivo, we investigated murine DETCs, a CD3⁺ cell population of the γδ T-cell lineage that constitutively expresses α_E(CD103)β₇ under normal conditions.¹²  As expected,¹¹  the number of DETCs was reduced in α_E(CD103)-deficient mice as compared with the wild-type controls (littermates, C57/BL6 genetic background). However, there were conspicuous morphologic differences between DETCs within the epidermis of wild-type mice and their counterparts in the skin of α_E(CD103)-deficient mice.

When whole mounts of epidermal sheets were stained by a CD3-directed monoclonal antibody, it became apparent that both number and size of dendrites per cell in α_E(CD103)-deficient mice were significantly reduced as compared with wild-type animals (Figure 4A). Extensive quantitative morphometric analyses revealed that the dendrites of DETCs from integrin α_E-deficient mice spanned an average area of 325 μm² (± 64.6), as compared with 495 μm² (± 91.9) in wild-type littermates, which was larger by 52.3% (P < .001; Figure 3B). In addition, the average number of dendrites per DETC was reduced significantly in α_E(CD103)-deficient mice (P < .001). DETCs isolated from the epidermis of wild-type mice showed a significantly higher proportion of cells forming protrusions/filopodia when plated on E-cadherin as compared with DETCs isolated from α_E(CD103)-deficient mice (P < .003). Again, the formation of such protrusions was abrogated by antibodies directed against α_E(CD103), β₇, or E-cadherin (Figure S1). Given that flow cytometric analysis of single-cell suspensions of murine epidermis showed no compensatory alterations in other adhesion molecules including the integrin chains α_L(CD11a), β₂(CD18), β₁(CD29), α₂(CD49b), α₄(CD49d), α₅(CD49e) and α₆(CD49f) on DETCs of α_E(CD103)-deficient mice (data not shown), the morphologic differences between DETCs of α_E(CD103)β₇-deficient mice and DETCs of wild-type mice can most likely be attributed to the expression status of α_E(CD103)β₇.

Our results indicate that expression of functional α_E(CD103)β₇ integrin is important for some cells to adapt and sculpt their shape on encountering E-cadherin, a ligand that is constitutively expressed within many epithelial tissues, in a fashion that allows close contact with ligand-bearing structures. This is the first example of such a function of a lymphocyte adhesion receptor that acts primarily within parenchymatous tissues, where many target cells for immune reactions are located and where effector functions of immigrating immune cells are exerted. In contrast, the cascade of lymphocyte extravasation and locomotion within the connective tissue is well established.^1,2  It is therefore conceivable that this novel functional property of α_E(CD103)β₇ plays a role for the intraepithelial functions of α_E(CD103)β₇-bearing cells in vivo.

First described as a marker for intestinal intraepithelial lymphocytes,^4-7,29  integrin α_E(CD103)β₇ has been an enigmatic and tantalizing heterodimeric receptor. Although α_E(CD103)β₇ is expressed at high levels by CD8⁺ mucosal and epidermal T cells, it is also found on small subsets of T lymphocytes elsewhere. Approximately 30% of splenic CD4⁺CD25⁺ regulatory T cells express α_E(CD103)β₇, apparently independent of T-cell activation.^14,30,31  The receptor has been implicated in guiding tissue localization of lymphocyte subsets under inflammatory conditions^10,14,18  and/or allograft rejection.^17,19,32,33  Expression of α_E(CD103)β₇ appears to be associated with cytotoxic activity of CD8⁺ T cells.^19,20,34,35  Along this line, interactions of α_E(CD103)β₇ with E-cadherin appear to be crucial for lysis of autologous E-cadherin⁺ tumor cells by cytotoxic T cells in some cases.²¹  Expression of α_E(CD103)β₇ is associated with several important cellular activities of mucosal dendritic antigen presenting cells, such as antigen presentation^13,22,36  or stimulation of T_reg²³  and some mucosal CD8⁺ T cells^24,25  by α_E(CD103)β₇⁺ dendritic cells. Overall, a robust body of circumstantial evidence suggests an association of α_E(CD103)β₇ expression with important immune functions. It is nevertheless currently unclear whether and to what extent α_E(CD103)β₇ itself contributes directly to such cellular functions, so the precise role of α_E(CD103)β₇ in immune regulation still remains largely elusive. However, it appears reasonable to assume that the proper exertion of all of the above-mentioned immunologic functions attributed to α_E(CD103)β₇-expressing cells requires that the interacting immune cells adapt their shape according to preformed tissue structures that need to come in close contact to each other. Based on our results it is now conceivable that α_E(CD103)β₇ contributes to this process in some cell types.

One example of a naturally occurring cell type of the T-cell lineage, DETCs, has been investigated in this study using α_E(CD103)β₇-deficient mice and their wild-type counterparts. Supporting the hypothesis that α_E(CD103)β₇ facilitates the formation of cellular protrusions and filopodia “squeezing” between epidermal keratinocytes, DETCs of α_E(CD103)-deficient mice showed significantly fewer and shorter dendrites as compared with wild-type DETCs. Forming dendrites and taking an appropriate shape are presumably important prerequisites for proper contact formation of DETCs with a maximum number of neighboring keratinocytes and other epidermal cells and, therefore, may play a role for at least some of the functions attributed to DETCs including maintenance of epidermal integrity, epithelial defense, wound healing, regression of skin tumors, or suppression of cutaneous graft-versus-host disease.^37-40  It is therefore conceivable that α_E(CD103)β₇ is indirectly involved in such functions through facilitating the proper shape of DETCs within the epidermis. The hypothetical concern that the observed functional differences between wild-type and α_E(CD103)β₇-deficient DETCs might result from compensatory changes in other adhesion molecules appears unlikely, because no such changes could be detected in several candidate molecules. New transfection strategies, such as RNA electroporation into primary lymphocyte populations,⁴¹  might help to further investigate such issues in the future.

How the α_E(CD103)β₇ integrin influences the cellular shape and formation of filopodia is not entirely clear yet, but likely involves signaling interactions with components of the cytoskeleton, as evidenced by the abrogation in the presence of cytochalasin D. Such signaling events have been demonstrated for some other integrins including α_IIbβ₃ or α₁β₁.^26,42  Likewise, it is possible that adaptor molecules contribute to the intracellular transmission of “outside-in” signals generated through the α_E(CD103)β₇ integrin, as has also been demonstrated for other integrins such as α₄β₁ or α_Lβ₂.^43-45  In any case, the morphologic differences in a defined population of naturally occurring T cells strongly suggest that α_E(CD103)β₇ influences the cellular shape not just in transfection-based artificial systems, but also in vivo, where the situation might be more complex and might involve a plethora of other adhesive interactions and/or mediators.

This work was supported by a Rudolf Virchow Award and a research grant from the Deutsche Forschungsgemeinschaft (M.P.S.). G.H., M.J.L., and M.P.S. are members of the Rudolf Virchow Center.

Contribution: S.S. generated and validated YFP-coupled constructs, performed transfections, video microscopy, animal experiments, and most of the functional experiments, and drafted the manuscript; H.S. generated YFP-coupled constructs and performed transfections; M.F., G.H., and M.J.L. critically contributed to the confocal microscopy, set up experimental conditions, contributed to the images containing photomicrographs, and made significant intellectual input to the manuscript; P.K. generated point-mutated α_E-constructs and provided significant intellectual input to the manuscript; M.P.S. conceived and planned the study, contributed to the morphometric analyses, coculture studies, and preparation of epidermal sheets, and wrote the manuscript. All authors critically read and approved the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Michael P. Schön, Department of Dermatology and Venereology, Georg August University Göttingen, von Siebold Str 3, 37075 Göttingen, Germany, e-mail: michael.schoen@med.uni-goettingen.de.