Introduction
A growing body of evidence indicates that type I interferons, such as interferon-α (IFN-α), play a pivotal role in the etiology and pathogenesis of systemic lupus erythematosus (SLE), and single-nucleotide polymorphisms in several key molecules important for IFN-α production and action are associated with SLE [
1,
2]. Moreover, some of these type I IFN pathway polymorphisms have been shown to impact IFN-α levels and responsiveness in SLE patients
in vivo [
3].
Plasmacytoid dendritic cells (pDCs) have been shown to be the major source of IFN-α production in the peripheral blood [
4] and within lymph nodes [
5], and these cells produce IFN-α after stimulation across TLR-7 and/or TLR-9 [
6‐
8]. pDCs have also been implicated as key mediators of pathogenesis in SLE [
9,
10]. However, a number of studies have shown that SLE patients have circulating pDCs that are reduced in number and/or are dysfunctional [
11‐
13]. Since different cell types are known to produce type I IFN in small quantities after microbial challenge [
4], these observations raise the possibility that other circulating cells (for example, those involved in the innate immune system, such as monocytes or myeloid dendritic cells, and/or those expressing TLR-9 and/or TLR-7, such as B cells) are the source of IFN-α production in SLE.
Antimalarial agents, such as quinine, have long been used in the treatment of SLE, first (in 1894) in the context of cutaneous lupus and then, as hydroxychloroquine (HCQ), in the context of SLE [
14‐
17]. In a randomized, double-blind, placebo-controlled study, SLE patients treated with HCQ had fewer disease flares and severe disease exacerbations compared to those receiving a placebo [
18]. Despite some uncertainty regarding the exact mechanism(s) underlying their various effects, the principal mechanism of action of agents such as HCQ relates to their ability to increase the intracytoplasmic pH and to thereby prevent acidification and maturation of endosomes [
19,
20]. IFN-α in SLE patients can be produced by pDCs in response to continuous stimulation by circulating immune complexes [
9] that are internalized by CD32 (FcγRIIA), with subsequent detection of DNA and RNA by endosomal TLR-9 and TLR-7 in pDCs [
10,
21]. HCQ would predictably block TLR-9/7 stimulation [
22,
23] and thus play a beneficial role in the treatment of SLE. Importantly, HCQ has been shown to inhibit the production of IFN-α in pDCs
in vitro, either after induction by DNA-containing immune complexes [
10] or upon stimulation with TLR-9 agonists [
13]. It is not clear, however, whether the same effect occurs
in vivo, that is, in the setting of SLE patients treated with HCQ.
In the current study, we have addressed the above questions directly in a cohort of patients with SLE, treated or not with HCQ, to determine the predominant circulating cell subpopulation capable of IFN-α production and the extent to which such production is inhibited by HCQ in vivo.
Materials and methods
Patients studied
SLE subjects were recruited from the University of California, San Francisco (UCSF) Lupus Genetics Project collection [
1,
24]. From this cohort, we recruited 39 individuals of European ancestry who fulfilled at least four of the American College of Rheumatology criteria for SLE [
25]. Disease activity was assessed using the Systemic Lupus Activity Questionnaire (SLAQ) [
26]. Subjects with the following criteria were excluded: acute infection or vaccination within the prior eight weeks, ongoing treatment with chemotherapy (including cyclophosphamide) or radiotherapy, or active viral hepatitis. Ten healthy gender-matched blood donors served as controls. The University of California, San Francisco Committee on Human Research approved the study and all subjects provided written informed consent.
Preparation of PBMCs
Blood samples obtained from SLE and healthy control subjects were collected into ethylenediaminetetraacetic acid (EDTA) and peripheral blood mononuclear cells (PBMCs) were prepared using a ficoll hypaque gradient. Cells were washed in phosphate buffered saline (PBS) and suspended in RPMI 1640 medium supplemented with penicillin, streptomycin, and L-glutamine.
TLR stimulation
PBMCs (106) from SLE and control subjects were cultured with the TLR-7 agonist, imiquimod, (5 μg/ml; InvivoGen, San Diego, CA, USA), the TLR-9 agonist, CpG-A ODN 2216 (5 μM; InvivoGen), the TLR-4 agonist, lipopolysaccharide (0.05 μg/ml; InvivoGen), and media for five hours at 37°C in 5% CO2. Brefeldin A (GolgiPlug, BD Pharmingen, San Diego, CA, USA) was added during the final three hours of stimulation to block cytokine secretion.
Flow cytometry
The panels of antibodies used for phenotypic and intracellular cytokine detection are described in Additional file
1, Table S1. Cytokine detection and phenotyping were performed by sequential cell surface and intracellular staining, following the manufacturer's instructions. Potential Fc receptors were blocked by incubating PBMCs with mouse serum prior to the addition of specific mouse anti-human antibodies. Fluorescence activated cell sorting (FACS) analysis was performed on a four-laser BD LSR-II flow cytometer, and data were analyzed using FlowJo software v9-3 (Treestar, Ashland, OR, USA) and transferred into analysis and graphic software including Excel (Windows, Seattle, WA, USA) and/or GraphPad Prism5 (La Jolla, CA, USA). All analyses were carried out without knowledge of the subject's clinical status, including treatment. The strategy used to gate the different subsets of PBMCs is shown in Additional file
1, Figure S1.
Statistical analyses
Exact nonparametric two-tailed tests were used. The Kruskal-Wallis One-Way Analysis of Variance on Ranks (ANOVA) or the Mann-Whitney tests were used to compare continuous variables. The Dunn's multiple comparison test was used for statistical correction of multiple comparisons. The Fisher's exact test was used to compare dichotomous variables. The Spearman rank correlation test was used to determine correlations between variables, with r being the Spearman correlation coefficient. Statistical analysis was performed with GraphPad Prism 5.01 software. P-values of < 0.05 were considered statistically significant.
Discussion
Ongoing IFN-α production in SLE patients appears to play an important pathogenic role in the autoimmune process and pDCs have a pivotal role as the main producers of IFN-α
in vivo [
9,
27]. For this reason, different regimens directed against IFN-α are of high therapeutic benefit in SLE. In this study, we found that pDC production of IFN-α and TNF-α upon TLR-9 or TLR-7 stimulation was markedly reduced in SLE patients treated with HCQ. Although HCQ has been shown to act as a TLR-9 and TLR-7 antagonist
in vitro, these data demonstrate that it also inhibits TLR-9 and TLR-7 stimulation
in vivo. Such inhibition is also specific, in that HCQ treatment has little effect on TLR-4-induced production of TNF-α by monocytes and mDCs. Thus, our data provide new insights regarding the mechanism of action of HCQ in SLE. The reduction of TNF-α TLR-9/7-induced pDC production observed with HCQ is interesting from a therapeutic point of view, even if the involvement of TNF-α in the pathogenesis of SLE remains controversial [
28]. Interestingly, a recent paper has shown that circulating TNF-α and type I IFN levels are correlated in a large cohort of SLE patients [
29]. Finally, similar observations have been made in HCQ-treated HIV-infected "immunologic non-responders" [
28], a condition in which chronic exposure to IFN-α may also lead to immune dysfunction [
30].
One puzzling and previously reported result is the overall diminished IFN-α production by pDCs in SLE patients compared to controls [
11‐
13]. It has been hypothesized that the reduced ability of pDCs to produce IFN-α may be a consequence of the redistribution of the efficient pDCs into tissues [
31] or of the exhaustion of circulating pDCs as a consequence of a high level of stimulation by persistent endogenous IFN-α inducing factors [
13]. Although pDCs are probably migrating to tissues and producing IFN- α there, we did not observe any difference in the percentage of circulating pDCs between SLE patients and healthy controls. We did not either observe any association with IFN-α production by
in vitro stimulated pDCs and SLE disease activity. Kwok
et al. [
13] previously showed that repeated stimulation of pDC with TLR-9 ligand decreased levels of IFN-α production. These authors concluded that the persistent presence of endogenous IFN-α-inducing factors induced TLR tolerance in pDCs of SLE patients. We have now demonstrated that, at least in some patients, such "tolerance" is strongly correlated with the
exogenous administration of HCQ. In fact, most of the SLE patients in previous studies who showed decreased IFN-α production were receiving HCQ [
11‐
13]. Our comprehensive study of all subsets of PBMCs potentially targeted by TLR-9/7 ligand also excludes the possibility that other circulating mononuclear cells (besides pDCs) may produce IFN-α. It would have been of interest to determine IFN-α levels and the absolute number of pDCs in the circulation, but such measurements are themselves only uncertain surrogates of the IFN-α and pDC levels in tissues, where most of the pathology of SLE plays out. Although conflicting data may result from the use of different methods to isolate and culture the pDCs between studies, our results support a key role for HCQ treatment in explaining the impaired functional ability of pDCs in SLE.
Low serum HCQ levels have been reported to be a marker for and a predictor of SLE exacerbations [
32], leading to the question of a possible relationship between blood HCQ concentrations and efficacy. We did not observe any significant differences in IFN-α/TNF-α production by TLR-7/9-stimulated pDCs between subjects receiving 400 mg/day and 200 mg/day. We do not know, however, the extent to which each patient was compliant with treatment and serum HCQ levels were not routinely measured.
Lupus activity is commonly measured by valid and reliable disease activity physician-rated scores, such as the Systemic Lupus Activity Measure (SLAM) [
33], the Systemic Lupus Erythematosus Disease Activity Index (SLEDAI) [
34], or the British Isles Lupus Assessment Group (BILAG) [
35]. All assess cumulative disease activity, including not only parameters that are immunological (for example, the level of antibodies against dsDNA and complement) but also clinical and biological. Practical considerations limited our ability to obtain all of the biological and immunological parameters required to calculate the SLEDAI or BILAG scores.
The SLAQ is a patient-reported assessment of subjective disease activity in SLE that has been shown to correlate strongly (
r = 0.62) with the SLAM-omitting laboratory items [
26] and is considered to be the best measure of self-reported disease activity in the field. It also has the advantage of being the most cost effective way to track disease activity. Using a cutoff score of 13 points on the SLAQ results in a negative predictive value of 74% for clinically important disease activity [
26] defined by a SLAM-omitting laboratory items score ≥3 points. In our study, 30 (77%) subjects had a SLAQ score that was less than 13 and the mean SLAQ score was 10.4. We did not observe an association between TLR-9/7 induced IFN-α production by pDCs and SLE disease activity. However, most of the patients studied here had low levels of disease activity and, because of the number of subjects included, our study may have lacked power to fully address this question.
Acknowledgements
We would like to thank Ms. Ruby Harrison for organizing the recruitment of study patients.
This research was supported by Assistance Publique Hôpitaux de Paris and Monahan Foundation (KS); and by a Mary Kirkland Scholar Award, the Alliance for Lupus Research, the Harvey V. Berneking Living Trust, and the following NIH grants: P60-AR-053308 and R01 AR052300 (LAC), and OD000329 and R01AI40312 (to JMM, a recipient of the NIH Director's Pioneer Award Program, part of the NIH Roadmap for Medical Research, through grant DPI OD00329).
Competing interests
The authors do not have conflicts of interest related to this study.
Authors' contributions
KS conceived the project. JMM and KS designed and interpreted the experiments, and wrote the manuscript. KS conducted the experiments. LAC selected appropriate subjects for analysis, provided samples and clinical data, and helped to analyze data and to write the paper. All authors have read and approved the manuscript for publication.