SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

Objective

Cutaneous manifestations are the most common clinical features of lupus erythematosus (LE). The aim of this study was to analyze differences in the inflammatory response of keratinocytes from patients with cutaneous LE (CLE) compared with healthy controls.

Methods

Keratinocytes from LE patients and controls were cultured from epidermal stem cells of the hair follicle of anagen head hairs. Functional responses of keratinocytes to cytokine stimulation were determined by flow cytometry and enzyme-linked immunosorbent assay. Biopsy samples of lesional skin were analyzed by immunohistochemistry.

Results

Keratinocytes from CLE patients expressed higher levels of IL-18 receptor on their cell surface in response to tumor necrosis factor α (TNFα) or interferon-γ stimulation. In response to IL-18 stimulation, these cells produced large amounts of TNFα. Of note, in the presence of IL-18, CLE keratinocytes failed to express IL-12. IL-12 has previously been shown to protect keratinocytes from ultraviolet irradiation–induced apoptosis. Keratinocytes from LE patients were more prone to die upon exposure to IL-18, and this increased apoptosis was abrogated by blockade of endogenously produced TNFα as well as by the addition of exogenous IL-12. IL-18 was highly expressed in biopsy samples of lesional skin from CLE patients.

Conclusion

Our results demonstrate an intrinsic difference in the inflammatory response of keratinocytes and indicate an autocrine feedback loop involving TNFα, IL-18, and IL-12 family members. Our results suggest that IL-18 may occupy an important position in the cytokine hierarchy in CLE, indicating the potential benefit of a local agent that blocks IL-18 activity in the treatment of the manifestations of CLE.

Lupus erythematosus (LE) is a chronic multisystem disease for which a number of immunoregulatory abnormalities have been described. The cause of LE remains unknown, but it is believed to be multifactorial, involving genetic and environmental factors. The disease may vary in severity from limited cutaneous lesions to severe systemic disease; however, in all forms of LE, the skin is one of the main target organs.

In cutaneous LE (CLE), epidermal keratinocytes are considered to be target cells of immunologic injury. A high level of apoptotic keratinocytes in lesional skin from CLE patients has been reported (1, 2). In lupus conditions, the noninflammatory clearance of apoptotic cells seems to be impaired (3, 4), and it has been proposed that the increased rate of apoptosis increases the chance of leakage of intracellular antigens that may trigger an autoimmune response. Furthermore, exposure to ultraviolet B (UVB) irradiation and mediators of inflammation, such as tumor necrosis factor α (TNFα), stimulate keratinocytes to translocate cytoplasmic and nuclear antigens, such as Ro/SSA, to the plasma membrane (5–7).

Aside from their role as target cells in immunologic injury, evidence is accumulating that keratinocytes may play an important role in actively regulating and maintaining the pathologic changes in CLE. Major findings regarding pathologic changes in skin include the expression of Fas (CD95), CD54, and class II major histocompatibility complex (MHC) on keratinocytes in lesional skin (1, 8) and the increased epidermal expression of proinflammatory cytokines, such as TNFα, type I interferons, and high mobility group box chromosomal protein 1 (HMGB-1) (9, 10). However, the hierarchical order of proinflammatory mediators reported to be expressed in the lesional skin of CLE patients remains to be determined.

So far, the “intrinsic” role of human tissue-resident cells such as keratinocytes in the maintenance of chronic inflammation has not been deciphered. In this study, we were interested in determining the potential differences in the proinflammatory response of patient-derived keratinocytes. These cells may be fundamental in influencing the micromilieu and leukocyte activity (11–13) in the skin as the target organ for chronic autoimmune inflammation.

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

Cytokines and reagents.

All cytokines used were purified recombinant human preparations. TNFα was purchased from ImmunoTools (Friesoythe, Germany). Interleukin-12 (IL-12), IL-18, and interferon-γ (IFNγ) were purchased from R&D Systems (Wiesbaden, Germany). Infliximab (Remicade), which was used to block TNFα, was obtained from Essex Pharma (Munich, Germany). Staurosporine was purchased from BioVision (Mountain View, CA).

Patients and healthy controls.

Consecutive patients with cutaneous forms of LE (chronic discoid or subacute cutaneous LE) were included in the study. The diagnosis of CLE was confirmed histologically in all patients. Scores on the European Consensus Lupus Activity Measure (14) in these patients were between 0.5 and 6.0. Patients had either no increase in serum antinuclear antibody (ANA) titers (extractable nuclear antigen [ENA]–negative) or had increased ANA titers with predominantly Ro and La ENA autoantibodies. Some patients showed a deficiency of C3 complement. Double-stranded DNA autoantibodies were not found in these patients.

Five anagen hairs for use in keratinocyte cultures were plucked from the scalp of CLE patients and healthy volunteers. Punch biopsy (4 mm) of affected skin was performed on patients with CLE (for diagnostic purposes) as well as on patients with active, acutely inflamed atopic dermatitis (for characteristics of the eczema patients, see ref.15). Because of insufficient numbers of keratinocytes in some cases, we could not include keratinocytes from all patients in all experimental settings. However, we describe our findings in 4 patients (3 with subacute cutaneous LE and 1 with chronic discoid LE) in whom all of the key experiments were repeated in order to analyze the stability of the effects that were found.

All study subjects gave their written informed consent. The study was approved by the Ethics Committee of the Hannover Medical School and was conducted according to the principles of the Declaration of Helsinki.

Isolation of messenger RNA (mRNA), reverse transcription (RT), and quantitative real-time polymerase chain reaction (PCR).

RNA was isolated from paraffin-embedded tissue sections (not older than 2 years) using a High Pure RNA Paraffin kit (Roche Molecular Biochemicals, Mannheim, Germany). RT was performed using a First-Strand complementary DNA (cDNA) synthesis kit (MBI Fermentas, St. Leon-Rot, Germany). Quantitative real-time PCR was performed with a LightCycler instrument (Roche Molecular Biochemicals) using a QuantiTect SYBR Green PCR kit (Qiagen, Hilden, Germany). QuantiTect primers from Qiagen were used for IL-18, TNFα, and GAPDH. For analysis, standard curves were created, and targets were quantified using Relative Quantification Software (Roche Molecular Biochemicals) as previously described (16).

DNA isolation and identification of single-nucleotide polymorphisms (SNPs).

DNA was extracted from heparinized blood or cultured keratinocytes (QIAamp DNA Mini kit; Qiagen) obtained from patients and healthy controls. DNA samples were analyzed for the –308 TNFα promoter polymorphism as previously described (17). For detection of the human TNF-308 G/A promoter polymorphism, the following probes and primers were purchased from TIB Molbiol (Berlin, Germany): CCT-GCA-TCC-TGT-CTG-GAA-GTT-A (sense), CTG-CAC-CTT-CTG-TCT-CGG-TTT (antisense), AAC-CCC-GTC-CCC-ATG-CCC-C (sensor; fluorescein labeled), and CAA-AAC-CTA-TTG-CCT-CCA-TTT-CTT-TTG-GGG-AC (anchor; LightCycler Red 640 labeled). The interferon regulatory factor 5 (IRF5) rs2004640 genotype was determined by PCR sequencing–based typing, using the following primers: AAG-TCT-AGG-CCT-AGA-CTG-GG (sense) and TGC-CCA-CTC-CGC-CGC-CTG (antisense). The PCR products were sequenced in both forward and reverse directions using a cycle sequencing kit (BigDye Terminator; Applied Biosystems, Foster City, CA) and an Applied Biosystems model 3730 sequencer. Data were analyzed with the SeqMan II program, version 5.7 (GATC Biotech, Konstanz, Germany).

Keratinocyte isolation and cultivation.

Isolation of outer root sheath cells from plucked human hair follicles was performed as described previously (18). Approximately 5 anagen hairs were plucked from the scalp of healthy volunteers or LE patients. The plucked hairs were cut down to the follicle, and the hair follicles were incubated for 2–24 hours in Hanks' balanced salt solution buffered with 20 mM HEPES (Sigma-Aldrich, Deisenhofen, Germany) and supplemented with 50 μg/ml of gentamicin and 2.5 μg/ml of amphotericin B (PromoCell, Heidelberg, Germany). The hairs were then placed in a 6-well plate that had been covered with a feeder layer of 3T3 fibroblasts (catalog no. ACC 173; DSMZ, Braunschweig, Germany) that had been treated with mitomycin C (Medac, Hamburg, Germany). Within 1–2 weeks, keratinocytes grew out (mainly from the bulge region). The medium (high-glucose Dulbecco's modified Eagle's medium with Ham's F-12 supplemented with 10% fetal calf serum, 1.8 × 10–4M adenine [Sigma-Aldrich], 10–10M cholera toxin [Sigma-Aldrich], 2 × 10–9M T3, 10 ng/ml epidermal growth factor, 4 μg/ml of hydrocortisone, 5 μg/ml of insulin, 5 μg/ml of transferrin, 50 μg/ml of gentamicin, and 2.5 μg/ml of amphotericin B [all components except adenine and cholera toxin were from PromoCell]) was changed every 2–3 days.

When sufficient keratinocytes were outgrown, cells were detached using 0.05% trypsin/0.02% EDTA (PAN Biotech, Aidenbach, Germany). Trypsin activity was stopped by the addition of an equal volume of trypsin inhibitor (PAN Biotech). Keratinocytes were then cultured in normal medium (Keratinocyte Growth Medium 2 kit; PromoCell), which was changed every other day. When cells reached 70–80% confluency, they were used for further experiments or were passaged and used between passage 3 and passage 7. When possible, some cells were cryopreserved at passage 3 and thawed at a later time (not exceeding a total of 7 passages). Before stimulation, hydrocortisone and epidermal growth factor were omitted from the medium. At the time the cultures were used in the experiments, they did not contain any contaminating fibroblasts or other cell types, as verified by expression of the epithelial marker cytokeratin (mouse anti-human cytokeratin antibody, clone MNF-116; Dako, Hamburg, Germany) and the fibroblast-specific marker AS02 (CD90; Dianova, Hamburg, Germany).

Flow cytometric analysis of membrane molecules.

Prior to flow cytometric analysis, adherent cells were detached by the addition of 0.025% EDTA for 10 minutes and HyQTase (Perbio, Bonn, Germany) for 10 minutes. Expression of IL-18 receptor α (IL-18Rα) was assessed using phycoerythrin (PE)–labeled mouse anti-human IL-18Rα monoclonal antibody (R&D Systems), as described previously (19). Other fluorescein isothiocyanate (FITC)–labeled or PE-labeled mouse anti-human monoclonal antibodies used were as follows: CD40, Fas (CD95), and CD54 (all from Coulter Immunotech, Hamburg, Germany), TNF receptor type I (TNFRI) and WSX-1 (both from R&D Systems), IL-12Rβ1 and HLA–DR (both from BD Biosciences, Heidelberg, Germany), Toll-like receptor 3 (TLR-3) and TLR-2 (both from Biomol, Hamburg, Germany), and CD91 (BioMac, Leipzig, Germany), lectin-like oxidized low-density lipoprotein receptor 1 (LOX-1; Sanbio, Beutelsbach, Germany), and class I MHC (BioSource, Solingen, Germany). Staining solution 7-aminoactinomycin D (7-AAD) was purchased from BD Biosciences and was used at a final concentration of 2.5 μg/ml. Mouse anti–phospho-histone monoclonal antibody H2A.X (Ser139, clone JBW301) was purchased from Biomol. H2A.X becomes phosphorylated in response to double-stranded DNA breaks that occur during apoptosis.

Stained cells were measured by flow cytometry using a FACSCalibur instrument and were analyzed using CellQuest Pro software (BD Biosciences). In some experiments, FITC-labeled M30 CytoDeath antibody (Roche Molecular Biochemicals) was used to detect apoptosis.

Immunohistochemistry.

Immunohistochemistry was performed as described previously (15). Deparaffinized tissue sections (5 μm thick) were treated with Target Retrieval Solution and, for IL-18 staining, with Peroxidase-Blocking Reagent (both from Dako). The sections were overlaid with the following primary antibodies: mouse anti-human IL-18Rα monoclonal antibody (clone 70625, undiluted; R&D Systems) or rabbit anti-human IL-18 polyclonal antibody (diluted 1:800; MBL, Woburn, MA). Corresponding concentrations of appropriate isotype antibodies were used in the same buffer. After overnight incubation, the sections were overlaid with appropriate biotinylated secondary antibodies (Vector, Peterborough, UK) for 30 minutes at room temperature. A mouse or rabbit alkaline phosphatase Vectastain ABC kit (Vector) was used according to the manufacturer′s instructions.

Measurement of cytokines and chemokines.

Cell-free supernatants from hair-derived keratinocytes were collected 24 hours after stimulation. The following enzyme-linked immunosorbent assay (ELISA) kits were used: CXCL10, CCL20, IL-12p40, and IL-1 receptor antagonist (IL-1Ra) (all from R&D Systems), TNFα (eBioscience, San Diego, CA), and high-sensitivity IL-12p70 (Diaclone, Besançon, France). Supernatants were also used in a Cytometric Bead Array (human Th1/Th2 11-plex kit; Bender MedSystems, Vienna, Austria) covering (among others) the cytokines IL-1β, IL-6, IL-8, TNFα, IL-10, TNFβ, and IL-12p70. In the keratinocyte-derived supernatants, production of IL-12p70, TNFβ, and IL-10 were below the detection limit of this assay.

Statistical analysis.

Data were analyzed using the paired or unpaired t-test for normally distributed values. For non-normally distributed values, the Mann-Whitney rank sum test was used. GraphPad Prism for Windows software (version 3.02; GraphPad Software, San Diego, CA) was used to perform the statistical analyses.

RESULTS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

Higher levels of IL-18Rα expression in keratinocytes derived from CLE patients.

We analyzed hair-derived keratinocytes from patients with CLE for their surface receptor expression profile upon stimulation with the proinflammatory cytokines TNFα or IFNγ. In CLE patients, stimulation with TNFα or IFNγ (Figure 1) alone was sufficient to up-regulate IL-18Rα, whereas keratinocytes from healthy donors depended on 2 signals (e.g., IFNγ plus TNFα) to up-regulate IL-18Rα on their cell surface (19). Of note, basal expression of the receptor was significantly higher in keratinocytes from CLE patients compared with those from healthy donors. Whereas some patients showed an extreme response to TNFα (very high patient-to-patient variability), the up-regulation of IL-18Rα in response to IFNγ was more stable.

thumbnail image

Figure 1. Induction of interleukin-18 receptor α (IL-18Rα) on keratinocytes from patients with cutaneous lupus erythematosus (CLE). A, Cultured keratinocytes from CLE patients (LE) or healthy controls (H) were not stimulated (ns) or were stimulated with tumor necrosis factor α (TNFα; 10 ng/ml) or interferon-γ (IFNγ; 20 ng/ml). After overnight incubation, cell surface expression of IL-18Rα, along with the appropriate isotype control, was determined. Results are expressed as the mean fluorescence intensity (MFI) in cells from 8 CLE patients and 9 control subjects. Data are shown as box plots. Each box represents the 25th and 75th percentiles. Lines outside the boxes represent the 10th and the 90th percentiles. Lines inside the boxes represent the median. CLE patient keratinocytes stimulated with IFNγ showed significantly higher IL-18Rα expression than did unstimulated CLE patient keratinocytes (P = 0.04 by paired t-test). ∗ = P < 0.05; ∗∗∗ = P < 0.01 by Mann-Whitney rank sum test. B, Keratinocytes from 4 of the LE patients were analyzed a second time after cells that had been stored in liquid nitrogen after the third passage were thawed. Cells were either not stimulated or were stimulated with IFNγ. Induction of IL-18Rα in cells from these 4 patients was found to be stable over time.

Download figure to PowerPoint

The easy inducibility of IL-18Rα on keratinocytes from CLE patients was stable over time. Figure 1B shows the results of repeated measurements of IFNγ-induced IL-18Rα in keratinocytes from 4 CLE patients after the cells were subjected to a freeze–thaw cycle. A significant difference in IFNγ-induced up-regulation of surface molecules on keratinocytes from CLE patients compared with those from healthy controls was also observed for TLR-2 (mean ± SEM 28.49 ± 5.602 mean fluorescence intensity [MFI] in CLE patients [n = 15] and 13.90 ± 1.175 MFI in healthy controls [n = 10]; P < 0.05 by unpaired t-test) (data not shown).

Other surface molecules examined, including TNFRI, IL-12Rβ1, WSX-1, CD54, Fas (CD95), TLR-3, CD40, class I MHC or class II MHC (HLA–DR), and the scavenger receptors CD91 and LOX-1, were not differentially regulated. These surface molecules were determined by flow cytometric analysis using the same protocol for all experiments.

Greater likelihood of death of CLE patient keratinocytes upon IL-18 stimulation.

To investigate the relevance of IL-18R expression on CLE keratinocytes, we analyzed the functional effects of IL-18Rα. Apoptosis is an important issue in the pathophysiology of LE, and according to the “clearance hypothesis,” the number of apoptotic cells in LE target organs may overwhelm the body's capacity to clear them. We analyzed the number of dead cells (using 7-AAD) or apoptotic cells (using H2A.X) in keratinocytes from CLE patients and healthy subjects. CLE keratinocytes showed a trend toward higher levels of spontaneous apoptosis. Upon incubation with IL-18, a significant dose-dependent increase in apoptosis was detectable in CLE, but not healthy, keratinocytes (Figure 2).

thumbnail image

Figure 2. Proapoptotic response of keratinocytes from patients with cutaneous lupus erythematosus (CLE) to stimulation with interleukin-18 (IL-18). Keratinocytes from CLE patients (n = 12) or healthy controls (n = 8) were either not stimulated (ns) or were stimulated with recombinant human IL-18 for 24 hours. A, Percentage of 7-aminoactinomycin D (7-AAD)–positive cells in keratinocytes from healthy subjects and patients with CLE, as determined by flow cytometry following stimulation with IL-18 at 1, 10, or 100 ng/ml. Values are the mean and SEM. ∗∗ = P < 0.02 by paired t-test. B, Expression of H2A.X in keratinocytes from 5 healthy subjects and 7 patients with CLE, as determined by flow cytometry following no stimulation or stimulation with 100 ng/ml of IL-18. Values are the mean fluorescence intensity (MFI). ∗ = P < 0.05 by unpaired t-test; ∗∗∗ = P < 0.01 by paired t-test. C, Representative histograms of H2A.X antibody–stained cells from 2 CLE patients and 1 healthy control subject, as determined by flow cytometry following no stimulation or stimulation with 100 ng/ml of IL-18.

Download figure to PowerPoint

We performed repeated experiments with keratinocytes from the 4 patients whose cells were used for the experiments shown in Figure 1B and for repeated experiments. The results showed increased apoptosis, as determined by H2A.X (P = 0.002 by paired t-test) as well as by M30 (P = 0.04 by paired t-test) staining following IL-18 stimulation (data not shown). Preincubation with IFNγ resulted in increased apoptosis in the control medium as well as in IL-18–stimulated cells. The “relative” apoptosis induced by IL-18 remained the same.

Production of higher levels of TNFα upon IL-18 stimulation of CLE patient keratinocytes.

Supernatants from IL-18–stimulated keratinocytes were analyzed by ELISA. Upon stimulation with IL-18, CLE keratinocytes released significantly more TNFα than did healthy control keratinocytes (Figure 3A). Other mediators, such as IL-1Ra, CCL20, and CXCL10 (examined by ELISA), as well as IL-1β, IL-6, and IL-8 (examined by Cytometric Bead Array), did not show significant differences, although a clear trend toward higher IL-6 expression in CLE keratinocytes was observed.

thumbnail image

Figure 3. Interleukin-18 (IL-18) induction of large amounts of tumor necrosis factor α (TNFα) on keratinocytes from patients with cutaneous lupus erythematosus (CLE), but not healthy controls (H). A, Induction of TNFα in keratinocytes from healthy subjects and CLE patients. Cells were not stimulated (ns) or were stimulated overnight with 100 ng/ml of IL-18, and TNFα levels (in pg/ml) in cell-free supernatants were determined by enzyme-linked immunosorbent assay. Bars show the mean. ∗∗∗ = P < 0.01 by unpaired t-test. B, Reduction of IL-18–induced apoptosis in keratinocytes from CLE patients following TNFα blockade with infliximab (ifl), as determined by staining with H2A.X. Keratinocytes from 4 of the LE patients were stimulated with 100 ng/ml of IL-18 for 24 hours in the presence or absence of 10 or 20 μg/ml of infliximab, in medium alone, or in 20 μg/ml of infliximab alone, and H2A.X-positive cells were analyzed by flow cytometry. Staining was calculated relative to that in unstimulated cells, which was set at 100%. Values are the mean and SEM. ∗ = P < 0.05 by unpaired t-test.

Download figure to PowerPoint

Repeated measurements were performed on keratinocytes from the 4 patients whose cells were used for the experiments shown in Figure 1B and for repeated experiments, and the results again showed a high response of cells from CLE patients to stimulation with IL-18 (see bottom of Figure 5C, which is discussed below). Of note, TNFα is known for its instability in culture supernatants, which probably accounts for different absolute levels of this cytokine in the different series of experiments. However, the differences observed between healthy and CLE keratinocytes remained stable.

SNP analysis was performed to examine the frequency of the –308A TNFα allele, which has been suggested to be associated with an increased production of TNFα in CLE patients (17). Patients carrying the A allele (4 AG and 8 GG in the CLE group; 1 AA and 6 GG in the healthy control group) seemed not to show a higher expression of TNFα as compared with patients carrying the GG genotype (data not shown). Furthermore, in this small number of patients analyzed, we found no association between the IRF5 rs2004640 T allele (20) and high producers of TNFα (5 of 7 healthy controls carried the T allele versus 5 of 10 CLE patients).

Abrogation of IL-18–induced apoptosis of LE keratinocytes by TNFα blockade.

Four patients with a positive apoptotic response to IL-18 were selected for TNFα-blocking experiments. All patients showed a dose-dependent decrease in apoptosis, as determined by H2A.X expression (P = 0.01 by unpaired t-test) (Figure 3B) and by M30 expression (P = 0.004 by paired t-test) in repeated experiments using cells from the same patients whose cells were used for the experiments shown in Figure 1B and for repeated experiments, as well as a dose-dependent decrease in cell death (using 7-AAD) in the presence of infliximab, a neutralizing TNFα antibody. These data indicate that IL-18–induced apoptosis is mediated, at least in part, via the release or surface expression of TNFα.

In control experiments, we used staurosporine, a known apoptosis inducer, as the control. We did not observe an abrogated or diminished apoptosis response in the presence of neutralizing TNFα antibodies (mean ± SEM 24.1 ± 5.3 MFI for staurosporine-treated cells and 21.9 ± 4.3 MFI for staurosporine plus infliximab–treated cells [n = 5], as determined by flow cytometry using the M30 antibody; P not significant by paired t-test).

High levels of IL-18 expression in lesional skin from CLE patients.

We performed immunohistochemical analysis of lesional skin from CLE patients to determine IL-18 expression. As shown in Figure 4, IL-18 expression was more intense in the epidermis of CLE patients as compared with healthy controls. Staining of IL-18Rα using this method did not show prominent differences between skin from healthy subjects and affected skin from patients with CLE (results not shown).

thumbnail image

Figure 4. Expression of interleukin-18 (IL-18) in biopsy samples of lesional skin from patients with cutaneous lupus erythematosus (CLE) and in biopsy samples of normal skin from healthy subjects. Samples obtained from 5 CLE patients and 2 healthy controls were embedded in paraffin, sectioned, and stained with anti–IL-18 antibody or an equivalent concentration of an isotype control, as described in Patients and Methods. Representative staining of tissue sections from a CLE patient and a healthy subject is shown (original magnification × 100).

Download figure to PowerPoint

In addition, we analyzed the expression of IL-18 mRNA in biopsy samples of lesional skin from CLE patients as compared with lesional skin from patients with actively inflamed, lesional eczema, another type of inflammatory skin disease. Quantitative real-time PCR analysis confirmed a significantly higher expression of IL-18 in lesional skin from CLE patients (n = 7) as compared with lesional skin from eczema patients (n = 3).The mean ± SEM relative expression (normalized to GAPDH) in biopsy samples from patients with CLE was 0.16 ± 0.035 arbitrary units (AU), as compared with 0.017 ± 0.004 AU in biopsy samples from patients with acutely inflamed eczema (data not shown).

It was previously shown that in biopsy samples of normal, noninflamed human skin, IL-18 mRNA is only faintly expressed (21). We therefore compared IL-18 mRNA expression in lesional skin from CLE patients and eczema patients but not from healthy skin. There was no trend toward higher expression of TNFα mRNA in CLE samples compared with eczema samples, but the difference was not significant (data not shown). However, higher TNFα expression in affected skin as compared with unaffected skin of CLE patients has been described in previous studies (9).

Lower production of IL-12 upon IL-18 stimulation of keratinocytes from LE patients.

IL-12 is involved in the polarization of Th1-type immune responses. A number of studies have described a lack of IL-12 in patients with LE. In the context of IL-18 stimulation, we detected a discrete increase in IL-12p40 secretion in keratinocytes from healthy controls that was not detectable in keratinocytes from LE patients (Figure 5A). IL-12p40 seemed to be down-regulated in LE patient keratinocytes, resulting in a significant difference between the 2 groups. Using a high-sensitivity ELISA (detection limit 0.78 pg/ml) to study keratinocytes from the 4 patients whose cells were used for the experiments shown in Figure 1B and for repeated experiments, we were able to show that the results obtained for IL-12p40 were also applicable to IL-12p70 (Figure 5C). Interestingly, the same CLE patients showed a high level of response of TNFα to IL-18 stimulation, whereas keratinocytes from the healthy controls produced IL-12p70, but not TNFα, under the same experimental conditions (Figure 5C).

thumbnail image

Figure 5. Failure of interleukin-18 (IL-18) to induce IL-12 in keratinocytes from patients with cutaneous lupus erythematosus (CLE). A, Levels of IL-12p40 in keratinocytes from CLE patients and healthy controls (H), as determined by enzyme-linked immunosorbent assay (ELISA). Cells were not stimulated (ns) or were stimulated for 24 hours with 100 pg/ml of IL-18. Bars show the mean. ∗ = P < 0.05 by unpaired t-test. B, Apoptosis of keratinocytes from CLE patients stimulated for 24 hours with IL-18 in the presence and absence of IL-12, as determined by H2A.X staining. Recombinant human IL-12 was added to IL-18–stimulated cells from 5 CLE patients that had shown an apoptotic response to IL-18. Staining was calculated relative to that in unstimulated cells, which was set at 100%. Values are the mean and SEM. ∗∗∗ = P < 0.01 by unpaired t-test. C, Production of IL-12p70 and tumor necrosis factor α (TNFα) by keratinocytes from 4 CLE patients whose cells were used for the experiments shown in Figure 1B and for repeated experiments and from 4 healthy control subjects, as determined by high-sensitivity ELISA. Cells were either not stimulated or were stimulated for 24 hours with 100 pg/ml of IL-18. Bars show the mean. ∗ = P < 0.05 by unpaired t-test.

Download figure to PowerPoint

IL-12 protection of keratinocytes from IL-18–induced apoptosis.

It was previously shown that IL-12 could protect keratinocytes from UV irradiation–induced damage/death (22). Therefore, we investigated the potential of IL-12 to protect keratinocytes from IL-18–induced apoptosis. In keratinocytes from LE patients, which showed an apoptotic response to IL-18 stimulation, the addition of exogenous IL-12 was able to protect them from cell death (Figure 5B). This finding held true when M30 staining was used to examine apoptosis protection of IL-18–stimulated keratinocytes from the 4 patients whose cells were used for the experiments shown in Figure 1B and for repeated experiments (P = 0.045 by paired t-test). Staurosporine-induced apoptosis could not be modified by the addition of exogenous IL-12, as analyzed by M30 staining (mean ± SEM 24.1 ± 5.3 MFI for staurosporine-treated cells and 22.6 ± 3.8 MFI for staurosporine plus IL-12–treated cells [n = 5]; P not significant by paired t-test).

DISCUSSION

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

Tissue-resident cells, such as synovial fibroblasts (23), endothelial cells, and epithelial cells, are being increasingly recognized as active cells in the pathophysiology of various autoimmune diseases. The results presented in this study add new aspects to the notion that keratinocytes are involved in the inflammatory process of cutaneous lesions in patients with LE.

We showed that keratinocytes from patients with CLE display an intrinsically different pattern of immunologic response to IL-18. We found a higher level of apoptosis induction by IL-18 in keratinocytes from CLE patients, but not from normal controls, and we found that this was mediated by high levels of TNFα production and diminished levels of IL-12 release. Thus, IL-18 seems to regulate TNFα as well as IL-12p40. These data support the in vivo data derived from studies of the murine system (24, 25) and indicate that IL-18 may occupy an important position in the proinflammatory cytokine cascade in lupus conditions.

IL-18 is a proinflammatory cytokine (26), and it has also been demonstrated to mediate immune cell infiltration into tissues (27). We have previously shown that primary human keratinocytes respond directly to IL-18, with a prominent effect on the production of CXCL10 (19).

Our results comparing the expression of IL-18 protein in lesional skin from CLE patients with that in skin from healthy controls, as well as the expression of IL-18 mRNA in lesional skin from CLE patients with that in lesional skin from eczema patients, another inflammatory skin disease, indicate high levels of IL-18 expression in CLE skin lesions. A number of studies have found hints of a pathogenetic role of IL-18 in systemic lupus erythematosus (SLE) patients or in murine models of lupus. Increased serum levels have been reported in LE patients (28–30), and high serum levels of IL-18 seem to be correlated with disease activity (31, 32). Amerio et al (32) proposed that the role of IL-18 in the pathogenesis of SLE might be important through its apoptosis-mediating properties. An up-regulation of IL-18, which was also expressed by tubular epithelial cells, has been reported (33) and recognized as a hallmark of lupus nephritis (34). Interestingly, MRL/lpr mice showed clear benefit with regard to glomerulonephritis, renal damage, and mortality rates from the targeting of IL-18 with cDNA vaccination (24), whereas administration of IL-18 to MRL/lpr mice resulted in accelerated proteinuria, glomerulonephritis, vasculitis, raised levels of proinflammatory cytokines, and the development of a butterfly facial rash (25). Thus, there may be important parallels in the pathogenesis of skin and renal inflammation. In both conditions, IL-18 may play a prominent role in the cytokine hierarchy in the local inflammatory micromilieu at epithelial “interfaces.” However, more data are clearly needed to prove this hypothesis.

The most prominent effect of IL-18 on keratinocytes from CLE patients that was identified in this study was the increased production of TNFα. From our data, we cannot conclude whether the increased TNFα production in CLE keratinocytes is due to increased sensitivity to IL-18 (and thus dysregulation at the level of IL-18 “signaling”) or altered regulation of TNFα (e.g., at the promoter level). An IL-18–mediated increase in TNFα has been described in relation to synovial macrophages (35), human peripheral blood mononuclear cells (36), and in murine trinitrobenzene sulfonic acid–induced colitis (37). Sullivan et al (38) evaluated chromatin at the TNFα locus within the monocyte population of SLE patients as compared with healthy controls and found more highly acetylated histones at the TNFα locus in monocytes from SLE patients. An increased transcriptional competence of TNFα could also play a role in IL-18–stimulated CLE keratinocytes.

However, we did not observe a correlation between increased TNFα secretion and the known –308 A polymorphism in the TNFα promoter. It has been reported that LE (39), subacute cutaneous LE (17), as well as other autoimmune diseases are associated with that polymorphism. This SNP has been associated with enhanced TNFα production, particularly after UVB irradiation (17). However, the issue remains a subject of controversy, since other groups of investigators, such as Popovic et al (9), did not find higher expression of TNFα in either affected or unaffected skin from CLE patients carrying the A allele as compared with patients carrying the GG genotype. It is clear that many patients with LE do not carry the –308 A polymorphism in the TNFα promoter, indicating the involvement of other regulatory factors. Based on recently published studies (20, 40), we also analyzed the IRF5 rs2004640 genotype, but again, we found no association between the IRF5 rs2004640 T allele and high levels of TNFα production. Of note, with regard to the functional significance of IRF5, Kozyrev et al (40) suggested that there may be other functional polymorphisms in IRF5 that are yet to be identified. Probably an interaction between different SNPs may be required to lead to functionally relevant alterations in immune responses.

Our results showed a high level of TNFα expression along with a low level of IL-12 expression upon stimulation of CLE keratinocytes with IL-18. As for the role of IL-12 in lupus conditions, there are some conflicting data. However, a number of studies showed reduced production of IL-12 by blood-derived cells from patients with SLE (28, 41–45). So far, no studies analyzing the expression of IL-12 in the skin of patients with LE inflammation have been published. Importantly, Werth et al (46) demonstrated that IL-12 completely blocks UV irradiation–induced secretion of TNFα from skin keratinocytes and fibroblasts. Thus, the low expression of IL-12 observed in our studies could contribute to the increased TNFα production seen in response to IL-18 in CLE keratinocytes. IL-12 was previously found to inhibit gene transcription and IL-10 release from UVB-irradiated keratinocytes (47) as well as UVB irradiation–induced apoptosis and DNA damage (22). Werth et al (46) suggested that IL-12 might play an important role in healthy individuals by decreasing TNFα-mediated apoptosis of keratinocytes, thereby diminishing one source of self antigen.

Dysregulation of programmed cell death and impaired removal of apoptotic cells have been discussed as pathogenetic factors in LE. According to the “clearance hypothesis,” the number of apoptotic cells in LE target organs may overwhelm the body's capacity to clear them. The role of IL-18 in apoptosis is a subject of controversy in the literature. While Schwarz et al (48) found IL-18 to be a protective cytokine, other investigators found IL-18 to be a proapoptotic mediator (49) in human endothelial cells. Like many other effects of IL-18, the predominant outcome may depend on the cell type, the species examined, and the surrounding milieu.

Apart from IL-18, other cytokines, such as CXCL10, type I and type II IFNs, HMGB-1, and TNFα (9, 13), have been described as effector molecules in CLE. Our data support the idea that IL-18, acting as a proximal regulatory cytokine, may control the micromilieu response pattern in the skin compartment. Therefore, the findings of this study support the notion that targeting the bioactivity of IL-18 may be a promising immunopharmacologic intervention in the treatment of LE conditions (involving the kidney or skin inflammation) (50). IL-18 binding protein (51) is one of the compounds that might prove clinically useful in LE therapy.

In conclusion, our findings suggest that IL-18 may play an important role in triggering inflammation in CLE, promoting a cytokine imbalance toward a high TNFα response and a low IL-12 response, thus providing a proapoptotic microenvironment for keratinocytes. Since the keratinocytes used in our study were not derived from lesional skin, the different response pattern to IL-18 must be based upon intrinsic differences in CLE keratinocytes as compared with healthy keratinocytes.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

Dr. Wittmann had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study design. Wittmann.

Acquisition of data. Wang, Drenker, Eiz-Vesper.

Analysis and interpretation of data. Wang, Werfel, Wittmann.

Manuscript preparation. Wang, Eiz-Vesper, Werfel, Wittmann.

Statistical analysis. Wang, Wittmann.

Acknowledgements

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

We thank Christina Hartmann for excellent technical assistance, Torsten Witte for critical discussion, and Mario Kratz for reading the manuscript. We thank our patients for their support.

REFERENCES

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES