Activation of the interferon-γ signaling pathway in systemic lupus erythematosus peripheral blood mononuclear cells

Authors


Abstract

Objective

To investigate interferon-γ (IFNγ) signaling in peripheral blood mononuclear cells (PBMCs) from patients with systemic lupus erythematosus (SLE) by analyzing IFNγ receptor (IFNγR) expression, STAT-1 expression and phosphorylation, and the regulation of IFNγ-inducible genes.

Methods

Fluorocytometry was used to investigate expression of STAT-1, pSTAT-1, CD95, HLA–DR, class I major histocompatibility complex (MHC), IFNγ-inducible 10-kd protein (IP-10), monokine induced by IFNγ (Mig), and IFNγR in PBMCs from SLE patients and healthy individuals. STAT-1 phosphorylation was determined by fluorocytometry and Western blotting after stimulation with IFNα or IFNγ. Quantitative polymerase chain reaction was used to assess messenger RNA (mRNA) expression of the IFNγ-inducible genes IP-10 and Mig shortly after preparation or after stimulation with IFNγ in monocytes.

Results

STAT-1 expression was increased in PBMCs from SLE patients and correlated significantly with disease activity and with the IFN-inducible expression of CD95 and HLA–DR. STAT-1 expression also showed a trend toward association with class I MHC expression. In addition, the expression of other IFNγ-inducible genes, such as IP-10 or Mig, was increased in SLE monocytes. While STAT-1 phosphorylation in SLE PBMCs and PBMCs from healthy individuals was similar after IFNα stimulation, incubation with IFNγ induced STAT-1 phosphorylation only in SLE lymphocytes. Moreover, SLE monocytes showed a considerably higher increase in pSTAT-1 expression upon IFNγ stimulation than monocytes from healthy individuals. Increased responsiveness of SLE monocytes to IFNγ was also confirmed on the mRNA level, where expression of the IFN-inducible, STAT-1–dependent genes IP-10 and Mig was more efficiently increased in SLE cells. However, IFNγR was similarly expressed on SLE lymphocytes and monocytes and those from healthy individuals.

Conclusion

In addition to supporting the role of IFNs in SLE immunopathogenesis in general, the findings of the present study support a role of IFNγ in this disease.

Type I interferons (e.g., IFNα) and type II IFN (IFNγ) have both been implicated in the immunopathogenesis of systemic lupus erythematosus (SLE). IFNα and IFNγ serum levels are increased (1–5), and IFN messenger RNA (mRNA) signatures are expressed in the peripheral blood cells of SLE patients (6–8). IFNα and IFNγ are known to induce SLE flares and drug-induced lupus (9–11). In murine models of the disease, both IFNα and IFNγ may be pathogenetically important (12–15), and, especially, a deficiency in either IFNγ or the IFNγ receptor (IFNγR) totally abates the disease process (16–19).

IFNs act on a variety of cells, including lymphocytes and monocytes (20, 21). The biologic effects of IFNα and IFNγ are mediated via the phosphorylation, and thus the activation, of members of the STAT family (21–25). The binding of IFNγ to its receptor leads to tyrosine phosphorylation of STAT-1 dimers, effected by JAK-1 and JAK-2. Subsequently, serine phosphorylation (26), nuclear translocation, and DNA binding occur, whereupon the activated STAT-1 dimer, termed IFNγ-activated factor, acts as a transcription factor (22, 27). IFNα signal transduction likewise includes STAT-1 phosphorylation but also encompasses phosphorylation of STAT-2 and formation of the IFN-stimulated transcription factor 3 complex, consisting of a heterotrimer of pSTAT-1 and pSTAT-2 with IFN regulatory factor 9. Thus, STAT-1 is a critical component in both IFNα and IFNγ signaling, and phosphorylation at a single tyrosine phosphorylation site is essential for its activation (21–24, 28). In the present study, we have analyzed STAT-1 expression and STAT-1 phosphorylation in SLE peripheral blood lymphocytes and monocytes to evaluate the degree and capacity of IFN signal transduction pathway activation and thus to obtain evidence of the biologic activity of IFNγ.

PATIENTS AND METHODS

Patients and samples.

Peripheral venous blood was obtained from 50 consenting patients with SLE classified according to the revised criteria of the American College of Rheumatology (29, 30), as well as from 30 healthy donors. Disease activity was assessed using the SLE index score (range 0–12, mean ± SD 4.3 ± 3.3) (31) and the SLE Disease Activity Index (range 0–15, mean ± SD 3.6 ± 3.3) (32), both of which are known to correlate well with all other common SLE disease activity parameters (33). Dosages of glucocorticoids and immunomodulatory drugs were recorded from the standardized patient charts. In the case of SLE patients, a complete blood count and erythrocyte sedimentation rate as well as serum levels of creatinine, complement factors (C3 [range 36.8–151 mg/dl, mean ± SD 88.95 ± 25.2 mg/dl], C4, and CH50), and anti–double-stranded DNA (anti-dsDNA) by radioimmunoassay (range 0–1,399 IU/ml, mean ± SD 64.3 ± 203.6 IU/ml) were routinely determined on the same day, as were urinary sediment and urinary protein excretion.

Cell preparation and culture.

Peripheral blood mononuclear cells (PBMCs) were isolated over Lymphocyte Separation Medium (LSM 1077) gradients (PAA Laboratories, Pasching, Austria). Cells were cultured either in pure RPMI 1640 (Invitrogen, Paisley, UK) for short-term experiments (e.g., stimulation with IFNs for 15 minutes) or in complete RPMI (RPMI 1640 plus L-glutamine, penicillin, streptomycin, and HEPES; Invitrogen) with 10% fetal calf serum (heat-inactivated; PAA Laboratories) for long-term kinetic experiments. For some experiments, purified monocytes or T cell populations were prepared from PBMCs, using negative selection by magnetic-activated cell sorting (MACS) (Monocyte Isolation Kit and Pan T-cell Isolation Kit II; both from Miltenyi Biotec, Bergisch Gladbach, Germany). In indicated experiments, recombinant human IFNα (Strathmann Biotech, Hamburg, Germany) or IFNγ (R&D Systems, Minneapolis, MN) was used at a concentration of 100 units/ml (34, 35).

Fluorocytometry of PBMCs.

Indirect intracellular staining of STAT-1 was performed using the Fix&Perm assay (An der Grub Bioresearch, Kaumberg, Austria) and either a monoclonal antibody binding STAT-1 (Cell Signaling Technology, Beverly, MA) or an isotype control antibody (Beckman Coulter, Fullerton, CA) in the first step, both followed by a fluorescein isothiocyanate (FITC)–conjugated secondary antibody (Dako, Glostrup, Denmark) as a second-step reagent (36). Staining for pSTAT-1 was performed by fixing cells in 2% paraformaldehyde (VWR International, Darmstadt, Germany) in phosphate buffered saline (PBS; Invitrogen) for 30 minutes, washing twice, and permeabilizing them in 90% methanol (Sigma-Aldrich, Buchs, Switzerland) in PBS overnight at –20°C, followed by incubation with a phycoerythrin (PE)–conjugated anti–pSTAT-1 monoclonal antibody or a PE-labeled isotype control antibody (both from BD Biosciences, San Jose, CA).

Cell surface staining was done according to standard protocols (37), using the following monoclonal antibodies: FITC-conjugated anti-CD95 (Beckman Coulter), PE-conjugated anti–HLA–DR (BD Biosciences), FITC-conjugated anti–HLA–A/B/C (anti–class I major histocompatibility complex [MHC]) (BD Biosciences), PE-conjugated anti-IFNγRI (BD Biosciences), and FITC-conjugated (Dako) and PE-conjugated (BD Biosciences) isotype control antibodies. Surface staining for INFγRII was performed with a specific unconjugated antibody (Alexis Corporation, Lausen, Switzerland) or a respective isotype antibody (Beckman Coulter) and a FITC-conjugated secondary antibody (Dako). Cells were analyzed by flow cytometry on a FACScan (BD Biosciences) immediately after staining, and individual gates were used for analyses of monocytes and lymphocytes.

Intracellular staining of IFNγ-inducible 10-kd protein (IP-10) and monokine induced by IFNγ (Mig; CXCL9) was done using PE-conjugated monoclonal antibodies (both from R&D Systems) according to the manufacturer's protocols. Monocytes were characterized by simultaneous surface staining, using allophycocyanin-conjugated anti-CD14 (Beckman Coulter). Cells were analyzed on a FACSCanto II flow cytometer, using DIVA software (BD Biosciences).

Western blotting.

T cells isolated by MACS were lysed in radioimmunoprecipitation assay lysis buffer containing Complete Mini EDTA-free protease inhibitor cocktail tablets (Roche, Indianapolis, IN) and sodium orthovanadate (Sigma-Aldrich, St. Louis, MO). Protein extracts were separated by electrophoresis on 10% NuPAGE (Invitrogen), followed by electrotransfer onto nitrocellulose membrane. After blocking with nonfat dried milk for 24 hours at 4°C, the membranes were consecutively incubated with a monoclonal antibody against pSTAT-1 (BD Biosciences) and a horseradish peroxidase–conjugated secondary antibody (Dako). Specific bands were detected with the enhanced chemiluminescence (ECL) detection kit on Amersham Hyperfilm ECL (both from GE Healthcare Biosciences, Uppsala, Sweden). Protein expression was quantified using a Fluor-S MultiImager and Quantity One software (both from Bio-Rad, Hercules, CA).

Quantitative real-time polymerase chain reaction (PCR).

For quantitative real-time PCR, the total RNA of purified monocytes from 8 healthy controls and 13 SLE patients was extracted using the RNeasy Mini Kit (Qiagen, Valencia, CA). Fifty nanograms of total RNA was reverse-transcribed using the Sensiscript RT Kit (Qiagen). Two microliters of complementary DNA was used for quantitative PCR, using the following primers: for suppressor of cytokine signaling 1 (SOCS-1), 5′-TGTTGTAGCAGCTTAACTGTATC-3′ (forward) and 5′-AGAGGTAGGAGGTGCGAGT-3′ (reverse) (34); for Mig, 5′-GCACCAACCAAGGGACTATC3′ (forward) and 5′-TCAGTTCCTTCACATCTGCTG-3′ (reverse); and for IP-10, 5′-GCTGTACCTGCATCAGCATT-3′ (forward) and 5′-TTCTTGATGGCCTTCGATTC -3′ (reverse). Real-time PCR was performed on a LightCycler 480 using LightCycler 480 SYBR Green (Roche) following the manufacturer's protocol. Duplicate reactions were run for each sample, and expression of a tested gene was normalized relative to the levels of GAPDH (5′-TGTGATGGTGGGAATGGGTCAG-3′ [forward] and 5′-TTTGATGTCACGCACGATTTCC-3′ [reverse]).

Priming with IFNα or IFNγ prior to IFNγ stimulation.

Magnetically isolated monocytes from 6 healthy donors were cultured in complete RPMI 1640 containing 10% fetal bovine serum for 2 days with macrophage colony-stimulating factor (10 ng/ml; Strathmann Biotech) with or without IFNα or IFNγ (3 units/ml or 30 units/ml). Thereafter, monocytes were restimulated with IFNγ (100 units/ml) for 15 minutes (for pSTAT-1 analysis) or 6 hours (for IP-10 analysis). Phosphorylation of STAT-1 was measured by fluorocytometry as specified above, and mRNA expression of IP-10 was determined by quantitative PCR.

Confocal laser scanning microscopy.

Cytospin preparations of FITC-conjugated anti–STAT-1–stained cells were performed and analyzed on an Axiovert 100M confocal laser scanning microscope (Zeiss, Wetzlar, Germany) (38).

Statistical analysis.

All group results are expressed as the mean ± SD. Unpaired Student's t-tests and paired t-tests were used for comparing groups and paired samples, provided that the data followed Gaussian distributions. If data were not normally distributed, Mann-Whitney tests were used instead of t-tests, and Wilcoxon matched pairs nonparametric tests were used instead of the paired t-test. Pearson and Spearman correlation coefficients were calculated for investigating possible associations between variables that were or were not normally distributed. P values less than 0.05 were considered significant.

RESULTS

Increased levels of STAT-1 and tyrosine-phosphorylated STAT-1 (pSTAT-1) in SLE PBMCs.

Since levels of both IFNα and IFNγ are elevated in SLE sera (1–4), and since STAT-1 protein expression is under control of IFNs (21, 39), we first investigated whether levels of STAT-1 and its activated form (pSTAT-1) are increased in SLE PBMCs. Indeed, as assessed by fluorocytometry, levels of STAT-1 protein were significantly increased in SLE PBMCs compared with levels in healthy cells (in lymphocytes, mean ± SD mean fluorescence intensity [MFI] 16.2 ± 13.2 versus 5.3 ± 1.9 [P < 0.0001]; in monocytes, mean ± SD MFI 17.8 ± 12.6 versus 7.4 ± 3.3 [P < 0.0001]) (Figure 1A). Laser scanning microscopy confirmed the increased amounts of STAT-1 expression in SLE patients (Figure 1B). Moreover, when analyzed immediately after purification from peripheral blood, SLE lymphocytes and monocytes contained significantly more pSTAT-1 protein than healthy cells (mean ± SD MFI 1.64 ± 0.36 versus 1.37 ± 0.2 [P = 0.0001] and 4.53 ± 1.79 versus 3.35 ± 0.92 [P = 0.0005], respectively) (Figure 1A), indicating in vivo preactivation of STAT-1 protein in SLE cells. When potential connections with the patients' medication were analyzed, no association was observed between immunosuppressive medication or prednisolone dose and STAT-1 or pSTAT-1 expression in SLE.

Figure 1.

Increased STAT-1 and pSTAT-1 expression in peripheral blood mononuclear cells (PBMCs) from patients with systemic lupus erythematosus (SLE). A, Both STAT-1 expression and pSTAT-1 expression were increased in SLE monocytes and lymphocytes compared with monocytes and lymphocytes from healthy individuals (HC), as measured by the mean fluorescence intensity (MFI). The horizontal lines indicate the means. B, Confocal laser scanning microscopy also revealed increased STAT-1 protein expression in SLE PBMCs compared with PBMCs from healthy individuals (original magnification × 630).

Greater efficiency of SLE PBMCs in IFNγ signaling.

We next addressed the question of whether STAT-1 protein would be effectively phosphorylated in SLE PBMCs despite its overexpression in the nonphosphorylated and phosphorylated forms. In preliminary experiments, maximal STAT-1 phosphorylation was observed after 15 minutes of stimulation with either IFNα or IFNγ; therefore, healthy and SLE PBMCs were stimulated for 15 minutes. Despite starting from a somewhat higher level, IFNα-induced STAT-1 phosphorylation in SLE cells was similar to that seen in cells from healthy individuals. In lymphocytes from healthy individuals, the MFI increased from 1.39 ± 0.16 to 1.84 ± 0.38, while in SLE lymphocytes, the MFI increased from 1.56 ± 0.25 to 1.85 ± 0.41. In monocytes from healthy individuals, the MFI increased from 3.47 ± 0.91 to 5.01 ± 1.56, while in SLE monocytes, the MFI increased from 4.09 ± 1.21 to 5.16 ± 1.74.

In contrast, IFNγ signaling was more effective in phosphorylating STAT-1 in SLE PBMCs than in PBMCs from healthy individuals. As shown in Figure 2A, lymphocytes from many SLE patients phosphorylated STAT-1 in response to IFNγ (MFI from 1.56 ± 0.25 to 1.72 ± 0.37; P < 0.002), but this was not seen in lymphocytes from healthy controls (MFI from 1.39 ± 0.16 to 1.39 ± 0.18; P not significant). This finding was supported by Western blotting (Figure 2B). Also, monocytes from SLE patients showed a considerably higher increase in pSTAT-1 expression upon IFNγ stimulation than monocytes from healthy individuals (P = 0.004, by comparing the slopes of the best-fit lines with a sum-of-squares F test) (Figure 2A). The difference in IFNγ signaling was not due to increased expression of IFNγ receptors on SLE cells, since IFNγRI and IFNγRII were similarly expressed on SLE and control lymphocytes (MFI 10.9 ± 3.1 versus 11.3 ± 2.7 and 4.6 ± 0.9 versus 4.3 ± 0.9, respectively) and on SLE and control monocytes (MFI 52.2 ± 15.5 versus 48.8 ± 14 and 13.9 ± 5.2 versus 11.9 ± 5, respectively), and similar percentages of PBMCs carried either receptor (Figure 3). However, STAT-1 protein expression in SLE monocytes as measured ex vivo by flow cytometry correlated significantly with the increase in pSTAT-1 expression after IFNγ stimulation (r = 0.53, P = 0.0005), suggesting either that pSTAT-1 expression was merely dependent on the amount of the available substrate, or that sensitization to IFNγ signaling might be mediated by the elevated levels of STAT-1 protein.

Figure 2.

Greater responsiveness of SLE PBMCs to interferon-γ (IFNγ). A, PBMCs were treated with IFNγ for 15 minutes and were analyzed on a fluorocytometer immediately after staining for pSTAT-1. Separate gates were set for monocytes and lymphocytes. IFNγ induced STAT-1 phosphorylation in SLE lymphocytes but not in lymphocytes from healthy individuals and had a less pronounced effect in monocytes from healthy individuals than in SLE monocytes. B, Western blot analysis of purified T cells revealed that T cells from some SLE patients (SLE 1) were more responsive to IFNγ than were T cells from healthy individuals. The level of pSTAT-1 expression was quantified using a Fluor-S MultiImager and Quantity One software. The pSTAT-1:actin ratios are shown in parentheses in the top panel and in the bar graph in the bottom panel. NS = not significant (see Figure 1 for other definitions).

Figure 3.

Interferon-γ (IFNγ) receptor expression. Horizontal lines indicate the means. There was no difference in IFNγ receptor expression between SLE monocytes or lymphocytes and monocytes or lymphocytes from healthy individuals. NS = not significant (see Figure 1 for other definitions).

Gene expression analyses.

Next, we were interested in whether the increased sensitivity to IFNγ would affect IFNγ-induced gene expression in SLE monocytes. To this end, the expression of genes such as IP-10, SOCS-1, or Mig was assessed in monocytes from healthy donors or SLE patients by quantitative PCR immediately after isolation or after 6 hours of incubation with IFNγ (22, 40). Immediately after isolation, the fold expression relative to GAPDH of both IP-10 (6.14 ± 1.18 versus 1.46 ± 0.26; P < 0.002) and Mig (9.25 ± 1.5 versus 3.97 ± 0.88; P < 0.01) was found to be up-regulated ex vivo in SLE monocytes compared with control monocytes (Figure 4A). Moreover, when analyzed by fluorocytometry immediately after preparation, CD14+ cells from SLE patients showed increased IP-10 and Mig (CXCL9) protein expression compared with those from healthy individuals (for IP-10, median fluorescence intensity 1,605 ± 359 versus 1,381 ± 382 [P = 0.08]; for Mig, median fluorescence intensity 1,609 ± 300 versus 1,310 ± 311 [P < 0.02]) (Figure 4B). Stimulation of SLE monocytes with IFNγ resulted in higher levels of IP-10 mRNA than in monocytes from healthy individuals (989 ± 717 versus 479 ± 263; P < 0.05). Also, levels of Mig mRNA were significantly higher in SLE monocytes after IFNγ stimulation (79,387 ± 58,750 versus 32,466 ± 20,777; P < 0.05) (Figure 4C). Our findings of increased levels of IP-10 and Mig mRNA are consistent with reports of increased plasma concentrations of IP-10 and Mig in SLE patients (41, 42). No differences in mRNA expression of the negative regulator of the IFN signaling pathway, SOCS-1 (43), were found between SLE monocytes and monocytes from healthy individuals upon incubation with IFNγ (data not shown).

Figure 4.

Expression of interferon-γ (IFNγ)–inducible 10-kd protein (IP-10) and monokine induced by IFNγ (Mig; CXCL9) in SLE monocytes and monocytes from healthy individuals. A and B, Expression of IP-10 and Mig was increased in SLE monocytes, as assessed by quantitative polymerase chain reaction (PCR) (A) and fluorocytometry (B). C, After stimulation with IFNγ for 6 hours, purified monocytes from SLE patients contained significantly more IP-10 or Mig mRNA, as assessed by quantitative PCR, indicating increased sensitivity of SLE monocytes to IFNγ. The fold expression of the indicated genes relative to their expression in unstimulated cultures is shown. Values in A and C are the mean and SD; the horizontal lines in B indicate the means. See Figure 1 for other definitions.

Response to IFN priming.

So far, our experiments revealed increased responsiveness of SLE monocytes to 100 units/ml IFNγ, as assessed by STAT-1 phosphorylation and the expression of IFN-inducible genes such as IP-10 and Mig. Since prestimulation of monocytes with low doses of IFNs was shown to result in higher cellular responsiveness to IFNγ (34, 44), we next sought to determine whether priming of monocytes from healthy individuals with either IFNα or IFNγ would increase IFNγ responsiveness to mimic our findings in SLE cells. Thus, monocytes from 6 healthy donors were primed for 48 hours with or without IFNα or IFNγ (3 units/ml or 30 units/ml) and restimulated with 100 units/ml for 15 minutes, when the increase in pSTAT-1 expression was determined by fluorocytometry. While prestimulation with 30 units/ml IFNγ led to increased responsiveness upon restimulation with IFNγ as compared with medium control (pSTAT-1 MFI 20.5 ± 10.9 versus 10.0 ± 5.0; P = 0.008), priming with either 3 units/ml IFNγ or 3 units/ml IFNα for 48 hours did not influence responsiveness to IFNγ. Prestimulation with 30 units/ml IFNα even decreased the sensitivity to IFNγ, when compared with culture in medium (pSTAT-1 MFI 7.1 ± 4.9 versus 10.0 ± 5.0; P < 0.02).

In contrast, when the same experiment was performed, but increased IP-10 mRNA by quantitative PCR was used as the outcome parameter after 6 hours of restimulation, prestimulation with either 30 units/ml IFNα or 30 units/ml IFNγ resulted in a decreased response to IFNγ (not shown). This was probably due to direct IFN effects on transcription (27).

Correlation of STAT-1 levels with expression of other IFN-inducible proteins.

Given that the expression not only of STAT-1 but also that of CD95, HLA–DR, and class I MHC is inducible by IFN, these surface molecules were also investigated (20, 21). We therefore used fluorocytometry to investigate the surface expression of class I MHC molecules (HLA–A/B/C), of HLA–DR, and of CD95 as well as their possible association with STAT-1 expression. As expected (4, 7, 45, 46), comparing SLE lymphocytes with those from healthy individuals, we observed increased expression of class I MHC and HLA–DR (MFI 554 ± 139 versus 403 ± 118 [P = 0.001] and 9.0 ± 3.8 versus 6.8 ± 1.9 [P < 0.01], respectively). Similarly, the MFI of CD95 on SLE lymphocytes was significantly increased compared with that on lymphocytes from healthy individuals (11.4 ± 4.1 versus 8.7 ± 2.2; P < 0.01). Likewise, the MFI of CD95 and class I MHC on SLE monocytes was significantly increased compared with that on monocytes from healthy individuals (29.9 ± 6.8 versus 24.1 ± 5.1 [P = 0.0008] and 811.5 ± 260.7 versus 641.6 ± 162 [P < 0.05], respectively).

STAT-1 protein expression in SLE lymphocytes as measured by MFI correlated with the MFI of CD95 (r = 0.45, P < 0.05) as well as with the MFI of HLA–DR (r = 0.51, P < 0.02) and showed a trend toward association with class I MHC expression (r = 0.44, 0.05 < P < 0.1). Similar associations were observed for monocytic STAT-1 expression with the MFIs of both CD95 (r = 0.53, P < 0.01) and class I MHC (r = 0.68, P < 0.01). These data suggested the possibility of a partly common induction mechanism for these molecules in SLE.

Correlation of STAT-1 expression in PBMCs with SLE disease activity.

If IFN production is a consequence of active disease, then the signaling molecules should likewise be associated with SLE activity. Indeed, STAT-1 expression in SLE lymphocytes, as estimated by the MFI, correlated significantly with SLE disease activity as measured by the SLE index score (r = 0.65, P < 0.0001) (Figure 5). The SLE index score was not normally distributed, resulting in a significant association but with a lower r value when correlation was tested with the nonparametric Spearman's rank test (r = 0.37, P < 0.01). Moreover, we observed an association with anti-dsDNA antibodies (r = 0.5, P < 0.0005), which are a well-known marker of SLE disease activity (47), as well as a negative correlation with C3c (r = –0.4, P < 0.005), which is known to correlate negatively with disease activity (48) (Figure 5). Similar results were obtained for STAT-1 expression in monocytes (Figure 5). No association was observed between the ex vivo amount of pSTAT-1 in PBMCs and the SLE index score or other parameters of SLE disease activity.

Figure 5.

Correlation of STAT-1 expression with SLE disease activity. STAT-1 expression in SLE monocytes or lymphocytes, as estimated by the MFI, correlated significantly with SLE disease activity as assessed by the SLE index score (SIS). C3c, a well-known marker of SLE disease activity, correlated negatively with STAT-1 expression. See Figure 1 for other definitions.

DISCUSSION

For many years, studies showing elevated levels of IFNs have implicated both type I and type II IFNs in the immunopathogenesis of human SLE (1–5). These data have more recently been supplemented by respective clinical observations (9–11) and by gene expression data (6–8).

Although a prominent role of IFNγ as a prime mediator of disease pathogenesis has been clearly demonstrated in murine models of SLE (15), recent gene array data favor type I IFNs in human SLE (6–8). In addition to supporting the role of IFNs in general in SLE immunopathogenesis, the present study provides some additional evidence for a role of IFNγ in this disease.

PBMCs from SLE patients contained significantly more STAT-1 protein than those from healthy individuals, revealing that signaling via IFNs is intact in SLE, and that the increase of IFNs has a functional effect in vivo. Moreover, STAT-1 protein contents correlated with clinical disease activity and also with serologic markers associated with disease activity, such as anti-dsDNA autoantibodies and C3c. This relationship between STAT-1 and disease activity in SLE would fit the well-established correlation of IFNα and SLE disease activity. Moreover, STAT-1 protein levels also correlated with the expression of CD95 and HLA class I and class II (HLA–DR) molecules, all of which are likewise known to be inducible by both types of IFNs (20, 21). This and the in vitro effects of IFNs observed here further support the notion of a role for IFNs in SLE pathophysiology.

If the association between STAT-1 and disease activity and the association between STAT-1 and the expression of additional IFN-inducible proteins were, in fact, directly influenced by IFNs, one would expect to find IFN-induced STAT-1 phosphorylation intact. Indeed, in response to IFNα, SLE PBMCs rapidly phosphorylated STAT-1, although this activation did not exceed that observed in PBMCs from healthy individuals. Interestingly, however, IFNγ signaling was much more effective in inducing STAT-1 phosphorylation in SLE monocytes than in monocytes from healthy individuals. Likewise, while lymphocytes from healthy individuals largely failed to phosphorylate STAT-1 after stimulation with IFNγ, SLE lymphocytes were responsive to IFNγ. This finding of increased responsiveness of SLE cells to IFNγ, as assessed by STAT-1 tyrosine phosphorylation, was confirmed by quantitative PCR data. Therefore, following IFNγ stimulation, mRNA levels of IFN-inducible and STAT-1–dependent genes, such as IP-10 or Mig, were more efficiently increased in monocytes from SLE patients than in those from healthy individuals. These data indicate that not only STAT-1 phosphorylation after IFNγ stimulation but also IFNγ-induced gene expression are increased in SLE.

In this context, it is of interest that priming with low concentrations of IFNγ or IFNα in vitro was reported to result in enhanced cellular responsiveness to IFNγ (34, 44, 49). One mechanism underlying IFN priming appeared to involve increased levels of STAT-1 and increased phosphorylation of this protein. This is totally in accordance with our findings in SLE, in which STAT-1 expression was increased, and SLE PBMCs showed increased responsiveness to IFNγ compared with PBMCs from healthy individuals. The idea of IFN priming in SLE is also consistent with the increased in vivo expression of Mig and IP-10, both of which are preferentially regulated by IFNs (8), and with the finding that STAT-1 expression correlates with that of other IFN-inducible proteins, such as CD95 or class I MHC.

To mimic our findings in SLE monocytes, we tested such a priming process in in vitro experiments. However, we found increased STAT-1 phosphorylation only after pretreatment with high-dose IFNγ, while neither low-dose IFNγ nor any dose of IFNα had such an effect. While not fully explained, this could in part be due to small experimental differences such as different cell origin and preparation or lower amounts of IFNγ on restimulation (34, 49).

However, several mechanisms limit the extent and duration of IFN signaling, including loss of expression of IFNγR in T cells, attenuation of signaling by phosphatases, inhibition of JAK-1 and JAK-2 by SOCS-1, and suppression of STAT-1 DNA binding by protein inhibitor of activated STAT (43, 50). Therefore, differences in the expression or function of any of these negative regulatory mechanisms could influence our findings of different IFNγ responsiveness between SLE cells and those from healthy individuals.

Thus, while we had expected to find a situation consistent with IFNα stimulation only, based on recent reports of mRNA array or quantitative PCR studies that did not detect an IFNγ signature (6, 8), our results may suggest some chronic IFNγ stimulation in SLE in addition to IFNα stimulation. This is in accordance with findings in murine SLE, in which IFNγ apparently plays an essential role (16–19), and with new data on human SLE (51).

Taken together, our data provide evidence that both type I IFN signaling and type II IFN signaling are intact in PBMCs from patients with SLE, and they suggest constant IFN influence in active disease. While not excluding a major influence of type I IFNs only, the data better fit a complex model in which both IFNs play a role.

AUTHOR CONTRIBUTIONS

Dr. Karonitsch 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. Smolen, Aringer.

Acquisition of data. Karonitsch, Feierl, C. W. Steiner, Dalwigk, Korb, Binder, Rapp, Scheinecker.

Analysis and interpretation of data. Karonitsch, C. W. Steiner, Binder, G. Steiner, Smolen, Aringer.

Manuscript preparation. Karonitsch, C. W. Steiner, Smolen, Aringer.

Statistical analysis. Karonitsch, Aringer.

Ancillary