Epidermal keratinocytes do not activate peripheral T-cells: interleukin-10 as a possible regulator

Abstract

The immunogenicity of allogeneic cultured human epidermal keratinocytes (cHEKs) has been studied in several models with contradictory results. We studied human T-cell activation in an in vitro assay by incubating, for 4 and 24 hr, cHEK confluent sheets with human peripheral blood mononuclear cells (PBMC); parallel HEK cultures were incubated with interferon (IFN)-γ to induce the expression of major histocompatibility complex (MHC) molecules before their interaction with PBMC. T-cell activation was evaluated by flow cytometry. T cells neither expressed the early and late activation markers CD69 and CD25, respectively, nor proliferated after incubation with the epidermal sheets, despite the IFN-γ-induced expression of MHC and adhesion molecules in cHEKs. Interleukin (IL)-10 was detected in the medium from the co-cultured PBMC and HEK sheets, but not from HEK alone. The results suggest that HEKs are unable to stimulate T lymphocytes through secretion of cytokines that might contribute to the immunosuppressive effect in this in vitro model.

Introduction

The skin allograft is one of the most rapidly rejected tissues (8–11 days after grafting); the rate of this rapid rejection may depend on the vascularity of allografted sites, the number of antigen-presenting cells (APCs) in the donor epidermis, and the migration rate of APCs towards the lymph node that drains the grafted site.¹

Cultured human epidermal keratinocytes (cHEKs) have been successfully applied as autografts to burned patients,^2,3 and as allografts for the temporary coverage of full-thickness burn wounds and to promote faster re-epithelization of partial-thickness burns, donor sites, excised burn scars and leg ulcers;^4–8 the apparent non-rejection of these cultured allografts has been explained by the immunosuppressive effect induced by burn injury.⁹ However, in non-immunocomprised patients they have also been applied to promote wound healing of facial dermabrasion wounds without clinical signs of immune rejection.¹⁰

Skin ulcers caused by venous stasis or diabetes healed after the repeated application of cultured allografts; in some of these patients, more than 15 repeated applications at intervals of 2 weeks were required for complete healing.⁸ In all these cases, there was no evidence of host sensitization. Supporting these observations, cHEK sheets applied onto excisional full-thickness skin wounds in immunocompetent NMR1 mice did not lead to the presence of inflammatory infiltrate or giant cells in the wound bed, which are normally associated with tissue immune rejection.¹¹

Clinical, histological and immunopathological observations in humans and experimental animals suggested that the allografted cHEKs do not induce rejection by the host,^12,13 probably in part because of the absence of APCs in the cHEK sheets^12,14 and the lack of major histocompatibility complex (MHC) class II molecules at the cultured keratinocyte surface.^15,16

HEKs synthesize blood group antigens,¹⁷ and MHC class I¹⁸ and class II antigens after stimulation with interferon (IFN)-γ,¹⁹ suggesting that the cultured allografts would be rejected by the host. Some experiments showed only a slow rejection of the cultured allografts applied to leg ulcers and various surgical skin defects, even after depletion of APCs from the cHEKs.^20–23 Therefore, keratinocytes elicit an alloreactive response, however slow, when either applied to a wound bed as epithelial sheets^20–23 or injected as cell suspensions,²² implying that the APCs are not the only cells in the epidermis that can stimulate the immune system of the host.

As T-lymphocyte activation is one of the early, key events that may initiate allograft rejection,²⁴ we attempted to understand the weak or absent immune reaction elicited by the cHEKs; we explored human T-lymphocyte activation in an in vitro co-cultivation assay with cHEKs. We show that cHEKs were unable to activate and induce proliferation of naïve T-lymphocytes, and that interleukin (IL)-10 release might be part of a mechanism contributing to the lack of T-cell response.

Materials and methods

Materials

Bovine serum albumin (BSA) and phytohaemagglutinin (PHA) were obtained from Sigma Chemical Co. (St Louis, MO). Fetal bovine serum (FBS) was obtained from HyClone Laboratories (Logan, UT), recombinant human IFN-γ from R&D Systems (Minneapolis, MN), and epidermal growth factor (EGF) from Upstate Inc. (Charlottesville, VA). Antibodies raised against cell surface markers were obtained from Immunotech-BioPharmaLink (Warrenale, PA), and anti-intercellular adhesion molecule (ICAM)-1 from AbD Serotec (Raleigh, NC). The mouse immunoglobulin G1 (IgG1) labelled with phycoerythrin (PE) was obtained from BD Biosciences (San Jose, CA). IL-10 was quantified with a Quantikine assay kit (R&D Systems), according to the manufacturer’s instructions.

Cell culture

The human epidermal keratinocytes (HEKs, strain HE-125) were isolated from newborn foreskin and cultured as described previously.^3,4 Single-cell suspensions, from second- or fifth-passage keratinocytes, were grown to confluence in medium with 10 ng/ml EGF in the presence of 3T3-feeder layers.^3,4

To study cell surface molecules by flow cytometry, HEKs were grown as above with medium containing 1·3 mm ethyleneglycoltetraacetic acid (EGTA) to reduce Ca²⁺ and decrease cell–cell adhesion, avoiding the use of proteases for cell harvesting and modification of the cell surface. Preconfluent cultures were incubated for 24 hr with medium containing or not containing 20 ng/ml IFN-γ, and then exhaustively washed with phosphate-buffered saline (PBS) to obtain single-cell suspensions by incubation with 0·02% [weight/volume (w/v)] EGTA, at 37°C. Cell suspensions were washed with PBS, pelleted and diluted with 1% (w/v) BSA in PBS to 1 × 10⁵ cells/50 μl, and immunostained.

Isolation and culture of PBMC

Human PBMC were obtained from freshly collected buffy coat fractions from healthy donors. All donors signed a letter of consent agreeing to participate in the protocol, which was approved by the ethics committee of the Institution and the corresponding health authorities, following international regulations. Mononuclear cells were isolated using a Ficoll-Paque gradient at room temperature. The PBMC were collected at the Ficoll-Paque/water interphase, washed twice with PBS and resuspended to 1 × 10⁶ cells/ml in RPMI-1640 medium supplemented with 10% (v/v) fetal bovine serum (FBS), 2 mm glutamine, 100 U/ml penicillin, 100 U/ml streptomycin, and 1 mm pyruvate. For viability determination, PBMC were incubated with anti-CD3-fluorescein isothiocyanate (FITC) (for total T cells) for 30 min, washed and resuspended in 300 μl of PBS containing 1 μg of propidium iodide/1 × 10⁶ cells and analysed by flow cytometry.

Co-cultivation of human PBMC with HEKs, assessment of T-cell proliferation and fluorescence-activated cell sorting (FACS) analysis

Experiments were carried out in duplicate dishes, using PBMC from six different donors. To evaluate T-cell proliferation, 20 × 10⁷ PBMC in 1 ml of PBS were labelled by the addition of 1 μl of a carboxyfluorescein succinimidyl ester (CFSE) stock solution, to yield a 1 μm final concentration of the fluorochrome. PBMC were further incubated for 10 min at 37°, washed with cold PBS and resuspended in RPMI medium at 1 × 10⁶ cells/ml. The labelled cells were co-cultured with cHEKs. As controls, 1 × 10⁶ labelled PBMC/ml were incubated or not with 5 μg/ml PHA for 72 hr at 37°. At the end of the experiments, cells were immunostained with anti-CD3 antibody, washed, resuspended in PBS and analysed by flow cytometry.

To analyse T-cell activation, 1-day postconfluent HEK sheets were incubated with 1 × 10⁶ PBMC for 4 and 24 hr. Thereafter, PBMC were harvested and washed with 0·1% (w/v) BSA in PBS. They were then centrifuged and incubated with anti-CD3, anti-CD25 or anti-CD69 for 30 min at room temperature. Cells were washed twice with 0·1% (w/v) BSA in PBS, and fixed with 3·5% (w/v) formaldehyde in PBS.

To analyse keratinocyte surface markers, 1 × 10⁵ HEKs were incubated with anti-MHC-I-PE, anti-MHC-II-FITC, anti-CD40-PE, anti-CD80-PE, and anti-ICAM-1-FITC. Thereafter, cells were washed and fixed. Flow cytometry was performed with a fluorescence activated cell sorter, FACSCalibur™ (Becton Dickinson, Franklin Lakes, NJ) equipped with a 15-mW argon ion laser. Filter settings were FITC (530 nm), PE (585 nm) and PerCP emitting in the deep red (>650 nm). Cells (10 000 events) were computed in list mode. Data were analysed with gating on CD3 (to estimate total T cells) using the cellquest® research software (Becton Dickinson).

Immunodetection of Langerhans cells

Normal human skin was embedded in Tissue Freezing Medium (Jung, Leica Microsystems, Nussloch, Germany), and frozen at −70° until use. Cryosections (8 μm) in silanized glass slides were acetone-fixed, air-dried, and then hydrated in PBS for 10 min. After blocking, slides were incubated overnight at 4°, with monoclonal antibody (mAb) DCGM4 (IgG1) specific for langerin (CD-207). Immunodetection was performed with FITC-anti-mouse IgG for 45 min at room temperature. The cHEKs were grown onto 18·0 × 18·0-mm glass coverslips, and paraformaldehyde-fixed, confluent cultures were immunostained with the mAb DCGM4 and observed with a Leica high-speed confocal/multiphoton system Model TCS SP2 (Leica Microsystems, Wetzlar, Germany).

Results

Cultured human epidermal keratinocytes do not activate human T lymphocytes

In the epidermis, Langerhans cells (LCs) capture and process native soluble antigens, and then migrate into secondary lymphoid tissue to activate T lymphocytes and take part in allograft rejection.^1,25 Serial cultivation of human keratinocytes from skin biopsies leads to depletion of LCs in the cultures.^23,26 We immunostained the cHEK sheets with an antibody against langerin, a specific marker of LCs. LCs were located at the basal and suprabasal layers of normal human epidermis (Fig. 1a–c), but they were not found in the cHEK sheets (Fig. 1d–f). We obtained similar results by staining for ATPase activity (not shown).

Figure 1.

Immunodetection of Langerhans cells (LCs) in human skin and in cultured human epidermal keratinocyte (cHEK) sheets. Tissues were stained with anti-langerin (green) for the LCs and with propidium iodide (PI; red) for nuclei. (a–c) Normal human skin (bar, 40 μm), (d–f) cHEK sheets, and (g–i) isotype control (negative) for cHEK sheets. For (d–i), the bar represents 80 μm.

CD69 is one of the earliest cell surface markers for T-cell activation. We incubated, for 4 and 24 hr, the cHEK sheets with the human T lymphocytes and followed the presence of CD69 by flow cytometry. PBMC incubated with cHEKs from third- or sixth-transfer cultures showed a similar proportion of CD69 T cells as those incubated without cHEKs (Table 1), whereas incubation with PHA, as a non-specific activator, induced a significant increase in CD69 T cells. When PBMC were incubated with IFN-γ-treated cHEKs, the proportion of CD69⁺ cells was also similar to that of controls (Table 1). Similar results were obtained when CD25 was used as marker for T-lymphocyte activation (Table 1). These results show that cHEKs cannot activate T lymphocytes in this in vitro assay.

Table 1.

Absence of T-lymphocyte activation by cultured human epidermal keratinocytes (cHEKs)

Incubation	T lymphocytes activated (%)
	CD69+		CD25+
	4 hr	24 hr	48 hr	72 hr
Cultured HEKs at third passage
Control PBMC	0·8	2·1	9·1	8·7
PBMC + PHA (5 μg/ml)	14·9	88·8	15·3	21·6
PBMC + cHEK sheets	0·7	1·8	12·8	10·1
PBMC + cHEKs treated with IFN-γ (20 ng/ml)	0·9	1·9	11·9	11·3
Cultured HEKs at sixth passage
Control PBMC	—	4·6	—	—
PBMC + PHA (5 μg/ml)	—	89·0	—	—
PBMC + cHEK sheets	—	5·1	—	—
PBMC + cHEKs treated with IFN-γ (20 ng/ml)	—	5·9	—	—

PBMC, peripheral blood mononuclear cells; IFN-γ, interferon-γ; PHA, phytohaemagglutinin.

Cultured HEKs were incubated with 1 × 10⁶ PBMC for 4 and 24 hr for CD69, and for 48 and 72 hr for CD25. Control cultures consisted of PBMC stimulated or not with 5 μg/ml PHA. Thereafter, PBMC were harvested and immunostained, and T-cell activation was analysed by flow cytometry.

As the outcome of T-cell activation is a proliferative cell response, we also evaluated T-cell proliferation after PBMC labelled with CFSE had been incubated for 24 hr with cHEKs previously treated or not with 20 ng/ml IFN-γ. Proliferation was determined by flow cytometry using the CFSE dilution method.²⁷ About 91% of T lymphocytes incubated with 5·0 μg/ml PHA for 3 days underwent at least one cell division, whereas in PBMC cultured with medium alone only 0·22% of total T cells underwent cell division (Fig. 2a,b). In the PBMC incubated with cHEKs, treated or not with IFN-γ, only about 0·6% of T lymphocytes underwent cell division (Fig. 2c,d).

Figure 2.

Proliferation of T-lymphocytes after 3 days of incubation with cultured human epidermal keratinocyte (cHEK) sheets. Peripheral blood mononuclear cells (PBMC) labelled with carboxyfluorescein succinimidyl ester (CFSE) were co-cultivated with HEKs for 3 days. PBMC were then stained with anti-CD3 and analysed by flow cytometry. M1 corresponds to the proliferative fraction of total T lymphocytes. (a) PBMC stimulated with phytohaemagglutinin (PHA). (b) Non-stimulated PBMC incubated with medium alone. (c) PBMC incubated with cHEK sheets. (d) PBMC incubated with interferon (IFN)-γ-treated cHEK sheets.

It has been reported that, depending on the presence or absence of co-stimulatory molecules, activation of T lymphocytes can also lead to apoptosis,^28,29 and apoptotic deletion of alloreactive T cells also mediates tolerance to allografts.²⁹ We examined, using propidium iodide (PI) staining and flow cytometry, the viability of PBMC after their interaction for 24 hr with cHEKs previously treated or not with IFN-γ. In all conditions, the proportion of PI-positive, or dead T lymphocytes was about 1%, or less, of the total cell number. All these results suggested that T lymphocytes were not activated and did not become apoptotic after in vitro incubation with cHEKs.

Keratinocyte expression of MHC class I and class II and co-stimulatory molecules

A possible explanation of the lack of T-lymphocyte activation after their interaction with HEKs would be the absence, or presence only at low levels, of MHC antigens or co-stimulatory molecules at the keratinocyte surface. We examined the presence of MHC antigens by flow cytometry in single-keratinocyte suspensions. As the cHEKs do not show MHC class II molecules at their surface,^15,16 unless they are stimulated with IFN-γ,¹⁹ and as MHC class II antigens are strongly associated with immune rejection, we cultured the HEKs under low-Ca²⁺ conditions, and 1 day before harvesting from preconfluent cultures they were incubated with or without 20 ng/ml IFN-γ. Keratinocytes had moderate levels of MHC class I, which increased fivefold after incubation with IFN-γ (Fig. 3a), whereas MHC class II molecules were not detectable unless the keratinocytes were stimulated with IFN-γ, leading to an up-regulation of MHC class II (Fig. 3b). These results suggested that cultured keratinocytes still have the capability to sensitize T cells under inflammatory conditions in vivo.

Figure 3.

Expression of surface molecules on cultured human epidermal keratinocytes (cHEKs). Subconfluent cHEKs were treated or not with 20 ng/ml interferon (IFN)-γ. One day later, cells were harvested as a single-cell suspension and stained for (a) major histocompatibility complex (MHC) class I, (b) MHC class II, (c) CD80, (d) intercellular adhesion molecule (ICAM)-1, (e) CD40 or (f) lymphocyte function-associated antigen (LFA)-1.

As the inability of HEKs to activate T cells could not be explained by a lack of MHC antigens, we examined the levels of co-stimulatory molecules such as CD80 (B7-1), ICAM-1, CD40 and lymphocyte function-associated antigen (LFA)-1. CD80 and LFA-1 were absent from the surface of the keratinocytes even after their stimulation with IFN-γ (Fig. 3c,f). ICAM-1 was not detectable on HEKs, but its levels were increased after stimulation with IFN-γ (Fig. 3d). In contrast, CD40 was found in HEKs treated or not with IFN-γ (Fig. 3e).

Interaction of PBMC with HEKs leads to IL-10 release into culture medium

In addition to the low levels of MHC class II and the absence of co-stimulatory molecules on the keratinocyte surface, another mechanism that might explain the lack of T-lymphocyte activation would be soluble mediators that regulate the immune system. Antigen-specific T-cell suppression by IL-10, a known suppressive cytokine of T-cell proliferation and cytokine production, is essential in peripheral tolerance to allergens, autoantigens, and transplantation and tumour antigens.³⁰ We examined the release of IL-10 to the culture medium after interaction of T lymphocytes with cHEKs. The PBMC were incubated with post-confluent HEKs treated or not with 20 ng/ml IFN-γ for 5 days; the IL-10 concentration was then determined in the culture medium. Controls were culture medium collected from cHEKs treated or not with IFN-γ, and from naïve T lymphocytes alone. Interaction of PBMC with cHEKs treated or not with IFN-γ led to IL-10 release to the culture medium; in four different experiments, the mean IL-10 levels were 1·4 to 1·6-fold higher than those found in medium collected from naïve T-cell cultures (P < 0·005) (Fig. 4). In contrast, HEKs treated or not with IFN-γ, when cultivated alone, produced IL-10 at levels similar to those found in medium from cultured PBMC (Fig. 4).

Figure 4.

Interleukin (IL)-10 release into culture medium after co-cultivation of peripheral blood mononuclear cells (PBMC) with cultured human epidermal keratinocyte (cHEK) sheets. After a 5-day incubation of PBMC with cHEKs, culture media were collected and the IL-10 concentration was determined. The mean of four different experiments is shown. The significance of the data was determined using the paired Student’s t-test (P < 0·005).

These results suggested that IL-10 could mediate the lack of T-lymphocyte activation after their interaction with HEKs; however, in experiments designed to block IL-10 activity with an anti-IL-10 neutralizing antibody, we did not observe the activation of PBMC (results not shown). Together, the results suggest that the immunomodulatory activity of the cultured keratinocytes does not seem be attributable only to a single cytokine, but might involve the combined participation of immunosuppressive cytokines, probably including IL-10.

Discussion

There exist several studies wherein the use of cultured allografts induced immune rejection,^21,22 whereas in others rejection was not detected.^13,16 As activation of T lymphocytes is one of the early, key events for initiation of immune tissue rejection by a host, we used an in vitro assay to determine whether cHEKs sensitize and activate naïve T lymphocytes. As we and others have shown that cHEKs lack LCs, this assay allowed us to examine human T-lymphocyte activation by the cHEK sheets without the interference of APCs normally found in the epidermis, such as the LCs.

Naïve T lymphocytes were not activated by their interaction with allogeneic cHEKs, as early activation markers such as CD69, and late activation markers such as CD25, were not induced. This was also confirmed by our data showing that lymphocytes co-cultivated with the HEKs did not proliferate or undergo apoptosis; both events are a clear sign of T-cell activation. As the lack of T-cell activation might also be explained by low levels of MHC class II and/or of co-stimulatory molecules in the cHEKs, we stimulated, by four- to sixfold, the expression of the keratinocyte MHC class I and class II complex by incubation with IFN-γ, a well-known inducer of MHC molecules;^19,31 these conditions did not induce the early and late activation markers CD69 and CD25, or stimulate lymphocyte proliferation.

The absence of some co-stimulatory molecules in cultured keratinocytes might also contribute to the lack of T-cell activation: CD80 and LFA-1 were not detected at the HEK surface even after stimulation with IFN-γ. CD80 is required for optimal T-lymphocyte activation,³² and its absence contributes to the inability of resting keratinocytes to present antigen to T cells, favouring the induction of anergy;³³ in transgenic mice, the targeted expression of CD80 in keratinocytes produced an abnormal immune response in the skin.^34,35 Also, our results suggest that ICAM-1 and CD40 did not contribute to the inability of keratinocytes to sensitize lymphocytes, as they were found in IFN-γ-treated keratinocytes which did not activate T cells.

Another mechanism that might explain the lack of stimulation of T lymphocytes by the allogeneic cHEKs is the production of soluble factors with immunosuppressive activity. Among the cytokines reported to be immunosuppressive, transforming growth factor (TGF)-β inhibits T-cell activation and proliferation,^36,37 whereas IL-10 blocks synthesis of pro-inflammatory cytokines.³⁸ Moreover, the proliferative response of T cells to certain antigens is inhibited by IL-10.³⁹ As we found that IL-10 released into the culture medium increased about 1·6-fold after the interaction between the cHEKs and the human PBMC, we suggest that part of the possible immunosuppressive mechanism preventing an immune response to allogeneic cHEKs might involve the participation of IL-10 as a regulator of the host’s T-lymphocyte populations. However, blocking only IL-10 activity with an anti-IL-10 neutralizing antibody did not lead to activation of PBMC (results not shown), suggesting that the in vitro immunosuppressive effect of the cultured keratinocytes is not attributable only to a single cytokine, but might involve the combined participation of several immunosuppressive cytokines, probably including IL-10. We previously reported that cHEKs contain TGF-β isoforms 1, 2 and 3,⁴⁰ raising the possibility that the immunomodulatory activity of the cultured keratinocytes might combine the effects of the two cytokines, and maybe others.

In conclusion, active suppression involves the utilization of multiple mechanisms. There is accumulated evidence of successful application of cultured allogeneic human epidermal sheets without any sign of immune sensitization of the host, or slow rejection at most. Our results suggest that such success may depend on a combination of secreted immunosuppressive cytokines such as IL-10 and TGF-β, the lack of a key cell surface molecule with co-stimulatory activity during antigen presentation (CD80), and the absence of LCs. Accumulation of data that substantially support and extend our results might open the way to the development of new applications of cultured keratinocytes, taking advantage of the possible temporary lack of response of the immune system to this cell type.