Skin biopsies (lesional and non-lesional) were collected from patients (n = 15) who had been diagnosed according to CLE criteria (1) and received neither pharmaceutical nor physical therapy. Normal skin tissues (n = 10) were obtained from healthy donors undergoing plastic surgeries. The demographic characteristics of CLE patients and healthy donors were detailed in Table S1 in Supplementary Material. Each biopsy was processed routinely for both paraffin sections and fresh tissue lysates, the latter of which were used for total RNA extraction. This study was carried out in accordance with the recommendations of the guidelines of the Hospital Research Ethics Committee. All subjects gave written informed consent in accordance with the Declaration of Helsinki.

As described previously (17), female MRL/lpr mice (or MRL/MpJ mice) at 10 weeks of age were irradiated with UVB at a dose of 50 mJ/cm². Irradiation was performed on the flank daily and lasted 2 days. There were five mice in each group. Skin biopsies were collected 24 h after last irradiation. The protocol was approved by the Research Ethics Committee.

As described previously (18), paraffin sections were deparaffinized and rehydrated, followed by the addition of dual endogenous enzyme block (DAKO, Glostrup, Denmark). Rabbit anti-Fn14 (clone #EPR3179) or RANTES (catalog #ab9679) IgG (Abcam, Cambridge, MA, USA) and rabbit anti-TWEAK (catalog #sc-5558) or Ro52 (catalog #sc-20960) IgG (Santa Cruz, Dallas, TX, USA) were used as primary antibodies (4 µg/ml, overnight at 4°C). To exclude the possibility of non-specific staining, rabbit polyclonal IgG (catalog #A7016; Beyotime Biotech, Beijing, China) was used as control. Then, the sections were incubated with polymer-horseradish peroxidase-labeled goat anti-rabbit/mouse IgG (catalog #K4065) and 3,3'-diaminobenzidine-chromogen substrate (DAKO) in order. Finally, sections were counterstained with hematoxylin–eosin solution.

The positive staining areas were measured by ImageJ1.61u software (National Institutes of Health, Bethesda, MD, USA) (19). Six to eight viewing fields were randomly selected from epidermal areas within each section. Positive staining areas were expressed as a percentage of the whole field area (% positivity per square millimeter). Such quantitative analysis was performed by two pathologists in a blinded fashion.

Murine keratinocytes (PAM212 cell line) were cultured in Eagle’s minimal essential media supplemented with 10% fetal bovine serum, 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid buffer, l-glutamine, MEM non-essential amino acids, sodium pyruvate, and penicillin/streptomycin (Gibco, Waltham, MA, USA). Before the stimulation assays, keratinocytes were starved in 2% fetal bovine serum-supplemented medium for 24 h. UVB irradiation was done by using a 302-nm UVB instrument (Rayminder, Ottawa Hills, OH, USA). The UVB dose (5.5 mJ/cm²) was chosen based on preliminary experiments (Figure S1 in Supplementary Material). In some experiments, culture media were added with recombinant mouse TWEAK (100 ng/ml, 2 days; R&D Systems, Minneapolis, MN, USA) or bovine serum albumin (BSA; 100 ng/ml, 2 days; Sigma-Aldrich, St. Louis, MO, USA) immediately after UVB irradiation. The inhibitors of JSH-23 (20 µM; Sigma-Aldrich), U0126 (10 µM; Cell signaling, Danvers, MA, USA), and wortmannin (25 µg/ml; Beyotime Biotech) were added 2 h before the UVB and TWEAK treatments.

Total RNA was extracted from fresh tissues or cell cultures by using the PureLink RNA kit (Invitrogen, Grand Island, NY, USA). A commercial cDNA kit (Applied Biosystems, Carlsbad, CA, USA) was used for reverse transcription. qRT-PCR was carried out on the 7900HT Fast PCR system (Applied Biosystems), with SYBR Green Master Mixes (Invitrogen, Grand Island, NY, USA) as fluorescent dye. The expression levels of the objective genes were calculated by using the 2^−ΔCt method (20). The sequences of the primers (AuGCT DNA-SYN Biotech, Beijing, China) are shown in Table S2 in Supplementary Material.

siRNA transfection was done as described previously (11). In brief, keratinocytes reaching 50% confluence were incubated with a mixture of target (or control) siRNA and Lipofectamine 2000 reagent (Life Technologies, Carlsbad, CA, USA) for 24 h. The sequences of Fn14 siRNA were CCCAUACUAAGGAACUGCATT (sense) and UGCAGUUCCUUAGUAUGGGTC (antisense). The sequences of TNF receptor associated factor 2 (TRAF2) siRNA were GGUACACUAUGAGGUCUGCTT (sense) and GCAGACCUCAUAGUGUACCTC (antisense). The sequences of Ro52 siRNA were GCCUAUGAGUAUCGAAUGUTT (sense) and ACAUUCGAUACUCAUAGGCTC (antisense). The target genes were verified by qRT-PCR in total RNA extractions, showing an efficiency of >80%.

Cell cultures were extracted for protein lysates. After electrophoresis and transfer onto membranes, the samples were incubated with rabbit anti-Ro52 IgG (2 µg/ml; Abcam). Similarly, rabbit anti-TNF receptor type 1 (TNFR1; catalog #13377), TNFR2 (catalog #3727), cellular inhibitor of apoptosis protein 1 (cIAP1; catalo #4952), TNF receptor-associated factor 2 (TRAF2; catalog #4712), extracellular signal-regulated kinase (ERK; catalog #4695), p-ERK (catalog #4370), Akt (catalog #9272), or p-Akt (catalog #5012) IgG (Cell signaling) and rabbit anti-nuclear factor-kappa B (NF-κB) p50 (clone #E381), NF-κB p65 (catalog #ab16502), or β-actin (clone #SP124) (Abcam) were used as primary antibodies (2 µg/ml). Next, biotinylated goat anti-rabbit IgG (catalog #A0277; 2 µg/ml; Beyotime Biotech) was added. Horseradish peroxidase-streptavidin and an ECL kit (Thermo Scientific, Waltham, MA, USA) were used for color development. The ImageJ1.61u software was used for quantitation of the band intensities, which were then normalized to the β-actin values accordingly.

Flow cytometry was carried out by using phycoerythrin-conjugated IgG targeting TNFR1 (catalog #sc-12746) or TNFR2 (catalog #sc-12750) (Santa Cruz). In brief, keratinocytes were digested with 0.25% trypsin–EDTA solution and then collected from dishes. Cells were blocked with 2% BSA in phosphate buffer saline for 1 h at 4°C, followed by twice washing. Primary antibodies (2 µg/ml) were added for 2 h at 4°C. After washing again, the stained cells were detected by using an LSRII instrument (BD Biosciences, San Jose, CA, USA). All data were analyzed with the FlowJo 7.6.1 software (Tree Star, Ashland, OR, USA).

Apoptotic cells were detected by flow cytometry (16). Keratinocytes were transfected with control or Ro52 siRNA, followed by UVB irradiation, TWEAK stimulation, or their combination. Cells were harvested and resuspended in annexin binding buffer. Annexin-phycoerythrin and 7-aminoactinomycin D (BD Biosciences) were added to samples in order. Flow cytometry was performed immediately. In this assay, early apoptotic cells were annexin positive but 7-aminoactinomycin D negative, while late apoptotic cells were positive for both of them.

Immunofluorescence was performed as reported previously (21). Briefly, keratinocytes were grown on a glass-bottom culture dish (MatTek, MA, USA), and incubated with fluorescein isothiocyanate-labeled antibodies targeting Ro52 (clone #sc-25351), TNFR1 (clone #sc-12746), or TNFR2 (clone #sc-12750) (2 µg/ml; Santa Cruz). After routine washing, cells were observed under a digital confocal microscopy (Leica Co., Wetzlar, Germany). Immunofluorescence was also carried out with frozen sections. Goat anti-Ro52 IgG (catalog #sc-21362; Santa Cruz) and rabbit anti-Fn14 IgG (clone #EPR3179; Abcam) were primary antibodies (2 µg/ml). The Alexa 488-conjugated donkey anti-goat IgG (catalog #ab150129) and Alexa 594-conjugated donkey anti-rabbit IgG (catalog #ab150076) were used as secondary antibodies (2 µg/ml; Abcam). Finally, sections were counterstained with 4′,6-diamidino-2-phenylindole before observation under digital confocal microscopy.

In vitro chemotaxis of macrophages was determined as described previously (18), but with some modifications. Briefly, keratinocytes were grown in six-well plates and stimulated with TWEAK or BSA (10 µg/ml, 2 days). Murine macrophages (J774.2 clone, 1 × 10⁶/ml) were transferred to an inner insert with an 8-µm pore membrane (Corning Inc., Corning, NY, USA), which allowed macrophage migration. The inner insert was placed in an outer insert (pore size = 0.4 µm) that selectively blocked the penetration of TWEAK (27.22 kDa) but allowed culture nutrients and most other cytokines, including chemokine C–C motif ligand (CCL) 17 and CCL22 (<20 kDa). The transwell system was incubated at 37°C for 2 h before the counting of crystal violet–positive cells on the bottom side of membranes.

J774.2 cells were also cultured alone in dishes, followed by collection of supernatants. The concentrations of interferon-α were determined in the supernatants. Keratinocytes received 2-day stimulation of J774.2 supernatant or recombinant mouse interferon-α (PBL Assay Science, Piscataway, NJ, USA), which had identical concentrations of interferon-α (7 pg/ml) in culture media. Cell lysates were processed for extraction of total mRNA or proteins.

The supernatants of the keratinocyte cultures were harvested for the detection of CCL17 and CCL22. Sandwich ELISAs were done according to the instructions for the two commercial kits (R&D Systems). The actual values were calculated from the standard curves generated by using calibrator diluents. Similarly, interferon-α in the J774.2 supernatants was determined by an ELISA kit (Elabscience Biotech, Wuhan, China).

All data were expressed as means ± SEM. The Stata 10.0 software package (StataCorp, College Station, TX, USA) was used for statistical analysis. Analysis of variance was used for the comparison of more than two groups. In comparing two groups, a two-tailed Student’s t-test was used for statistical differences. Chi-square test was used for comparing the ratios in different groups. Differences were considered significant at p < 0.05.

By immunohistochemistry, both TWEAK and Fn14 were detected in paraffin sections of human subjects, with strong staining observed in lesional skin but not in non-lesional or normal controls (Figure 1A). Quantification of positivity percentage per square millimeter showed significantly increased expression of TWEAK and Fn14 in lesional skin (Figure 1B). There were no differences in TWEAK or Fn14 expression between non-lesional and normal controls (Figures 1A,B) or between the clinical subtypes of CLE (Figure S2 in Supplementary Material). The mRNA expression of TWEAK and Fn14 was further determined in fresh tissues, showing higher levels in lesional skin when compared with normal controls (Figure 1C). Accordingly, the mRNA expression levels of RANTES, monocyte chemoattractant protein 1 (MCP-1), and interferon gamma-induced protein 10 (IP-10) increased in lesional samples (Figure 1D). Under confocal microscopy, the Fn14 and Ro52 expression showed high overlap in upper epidermis of lesional skin (Figure 1E). Moreover, both immunohistochemistry and qRT-PCR confirmed more Ro52 and RANTES expression in tissue samples of patients with CLE (Figure S3 in Supplementary Material).

Previous studies had demonstrated that UVB irradiation increases Fn14 expression in lesional skin of MRL/lpr mice (16, 17). In this study, we determined TWEAK expression in skin lesions of MRL/lpr mice. By immunohistochemistry, TWEAK expression was stronger in MRL/lpr mice when compared with MRL/MpJ mice, and UVB irradiation even promoted TWEAK expression in MRL/lpr mice (Figure S4 in Supplementary Material).

To show the effect of TWEAK/Fn14 activation on Ro52 expression, murine PAM212 keratinocytes were treated with UVB irradiation or plus TWEAK stimulation. We found that the mRNA levels of Fn14, RANTES, MCP-1, and IP-10 increased significantly upon UVB irradiation, which was further augmented by TWEAK (Figure 2). Moreover, TWEAK enhanced the Ro52 mRNA expression in keratinocytes, regardless of whether they received UVB irradiation or not (Figure 3A). Western blotting analysis further confirmed the increase in Ro52 levels in both the cultural supernatants and cell lysates (Figures 3B–D). To exclude the effect of signals other than those of Fn14, keratinocytes were transfected with Fn14 siRNA, followed by UVB irradiation or plus TWEAK stimulation. Western blotting showed that the enhancement effect of TWEAK on Ro52 expression was abrogated by Fn14 siRNA transfection in UVB-irradiated keratinocytes (Figures 3E,F). Furthermore, immunofluorescence verified such decrease in both Ro52 and Fn14 expression induced by Fn14 siRNA (Figures 3G,H).

To further reveal whether Ro52 participates in TWEAK-induced photosensitization, murine keratinocytes were transfected with Ro52 siRNA before TWEAK stimulation or UVB irradiation. The results showed that transfection of Ro52 siRNA significantly reduced cell apoptosis induced by TWEAK stimulation, UVB irradiation or their combination (Figure S5 in Supplementary Material).

To understand the downstream signals in the TWEAK/Fn14 regulation of Ro52 expression, keratinocytes were pretreated with specific inhibitors of the NF-κB (JSH-23), mitogen-activated protein kinases (MAPK)/ERK (U0126), and PI3K/Akt (wortmannin) pathways, respectively. Then, these cells received UVB plus TWEAK treatments. The results showed that JSH-23 and wortmannin decreased the Ro52 mRNA levels, whereas U0126 promoted Ro52 expression (Figure 4A). Accordingly, Western blotting analysis showed similar trends in keratinocytes pretreated with these inhibitors (Figures 4B,C). The expression levels of NF-κB, ERK, and Akt proteins were also determined, showing that these specific inhibitors significantly decreased their expression accordingly (Figure S6 in Supplementary Material). To further confirm the effect of these inhibitors, the mRNA levels of RANTES, MCP-1, and IP-10 were determined in keratinocytes (Figure 5). JSH-23 was found to decrease the production of RANTES and MCP-1. U0126 and wortmannin had different effects on the two cytokines, whereas IP-10 was not affected by any of the three inhibitors.

Dermal infiltration of macrophages contributes to the development of CLE through the production of interferon-α (2), which may induce the upregulation and nuclear translocation of Ro52 in cells (22). Hence, we studied the effect of TWEAK/Fn14 activation on macrophage chemotaxis. The results showed that the addition of TWEAK to keratinocyte cultures (not contacting macrophages) enhanced macrophage migration (Figure 6A). Moreover, compared with blank and BSA groups, these keratinocytes expressed more CCL17 and CCL22 on TWEAK stimulation (Figures 6B,C).

The effect of macrophage-derived interferon-α on Ro52 expression in keratinocytes was also analyzed. The results showed that the supernatants of J774.2 cells had high levels of interferon-α (4.65–9.53 ng/ml). The keratinocytes stimulated with J774.2 supernatant or recombinant interferon-α expressed more Ro52 at mRNA and protein levels (Figure S7 in Supplementary Material).

The effect of TWEAK on the regulation of cell fate depends on TNFR (12, 23). Thus, the expression of TNFR subtypes was determined in keratinocytes irradiated with UVB. Interestingly, UVB irradiation promoted the mRNA expression of TNFR1 but inhibits that of TNFR2 in these cells (Figure 7A). Western blotting confirmed such effect of UVB irradiation on TNFR subtypes in cell lysates (Figure 7B). Moreover, both flow cytometry (Figures 7C,D) and immunofluorescence (Figure 7E) verified the shift in TNFR expression profile in these cells.

In these keratinocytes, TWEAK increased the mRNA levels of cIAP1 and TRAF2 but showed no effect on the mRNA expression of NF-κB-inducing kinase (NIK) (Figures 8A,B). Moreover, TWEAK reduced the protein expression of cIAP1 and TRAF2 in a time-dependent manner (0–24 h) (Figures 8C,D). Furthermore, the transfection of TRAF2, but not of control siRNA, decreased the Ro52 protein level in keratinocytes treated with UVB irradiation plus TWEAK stimulation (Figures 8E,F).