Introduction
Systemic lupus erythematosus (SLE) is an autoimmune disease characterized by defective phagocytosis of apoptotic cells (ACs) [
1]. Accumulation and presentation of AC-derived nuclear and membrane autoantigens in lymphoid organs are believed to drive the activation of autoreactive B and T cells, leading to production of antinuclear and antiphospholipoprotein autoantibodies. Immune complexes containing nuclear antigens and antibody-opsonized ACs bind to Toll-like receptors (TLRs) and immunoglobulin G Fc receptors (FcγRs) on innate immune cells, provoking aberrant production of type I interferons α and β (IFN-α/β) and proinflammatory cytokines [
2,
3]. Additionally, noningested ACs undergo secondary necrosis, which fuels ongoing innate inflammation by amplifying TLR activation and oxidative burst [
4,
5].
Clearance of ACs is crucial for resolution of inflammation and maintenance of immune tolerance [
6]. In healthy individuals, discrete populations of phagocytes, called M2c (CD163
+) macrophages, are designated to promptly remove ACs, including activated immune cells undergoing apoptosis [
7‐
9]. Moreover, the physiologic engulfment of ACs is associated with macrophage release of anti-inflammatory cytokines [
6].
The Mer receptor tyrosine kinase (MerTK), which belongs to the family of Tyro3, Axl and MerTK (TAM) receptors (TAMRs), is required for the efficient clearance of ACs exerted by M2c monocytes/macrophages [
9], participates in immune regulation by stimulating interleukin 10 (IL-10) secretion [
9‐
11] and is involved in restoration of tissue homeostasis after inflammatory processes as well as in the maintenance of central and peripheral tolerance [
11‐
14]. Another member of the TAMR family, Axl, is importantly involved in the deactivation of innate immune cells stimulated by TLR agonists and type I IFNs through recruitment of suppressors of cytokine signaling 1 and 3 and the transcriptional repressor Twist [
15‐
17]. Both MerTK and Axl inhibit TLR-induced activation of nuclear factor κB (NF-κB) transcription factors and production of proinflammatory cytokines such as tumor necrosis factor α (TNF-α) and IL-6 [
9‐
11,
15‐
17].
Both the Mer and Axl receptors are susceptible to posttranslational regulation through ectodomain shedding mediated by a disintegrin and metalloprotease domain (ADAM) metallopeptidases [
18‐
20]. In the present study, we measured the soluble (s) ectodomains sMer and sAxl in the circulation of SLE patients and matched healthy individuals with the aim of investigating how these molecules relate to clinical, laboratory and immunological profiles of SLE; how they are related to each other and to the TAMR ligands growth arrest–specific 6 (Gas6) and reduced free Protein S (ProS); and under what immunological conditions they are produced. We found that plasma levels of both sMer and sAxl were related to general aspects of systemic autoimmunity and were associated with hematological and renal involvement. However, sMer and sAxl did not significantly correlate with each other. Compared to sAxl, sMer showed closer relations with specific aspects of SLE immunopathogenesis, such as production of lupus-specific autoantibodies and reduction of free ProS in circulation. Strong correlations with disease activity indices were found for sMer, but not for sAxl. Patients with signs of active SLE showed higher levels of sMer compared to matched controls. Remarkably, sMer levels in SLE patients directly correlated with circulating levels of sCD163, a well-known marker of M2 activation, and sCD163 levels correlated with Systemic Lupus Erythematosus Disease Activity Index (SLEDAI) score. In fact, sMer and sCD163 were found to be released under the same M2c polarizing conditions. Production of sAxl was instead enhanced in the presence of IFN-α or IFN-β, and plasma concentrations of sAxl in SLE patients correlated with increased Gas6 levels. Our data highlight, through the study of sMer and sCD163, a strict relationship between SLE pathogenesis and homeostasis of anti-inflammatory and efferocytic M2c monocytes/macrophages. We also provide indirect proof, through the study of sAxl, that type I IFN stimulation plays a role in the development of systemic autoimmunity but does not seem to be closely related to SLE disease activity. Whether augmented ectodomain shedding of membrane receptors reflects increased turnover and/or activation of the respective pathways or rather contributes to their dysfunction and/or inhibition remains to be clarified.
Methods
Participants
Plasma samples from 107 SLE patients participating in the Oklahoma Cohort for Rheumatic Disease were studied. All patients satisfied at least four of the 1982 revised American Rheumatism Association criteria for SLE [
21]. Forty-five of these patients were matched to healthy controls by age, gender and ethnicity. Clinical and laboratory data were registered into a database which included no personal identifiers. The characteristics of the patients and the controls enrolled are reported in Table
1. Heparinized plasma samples were collected and stored at −70°C immediately after collection. Disease activity was scored using the SLEDAI and the British Isles Lupus Assessment Group (BILAG) index [
22,
23]. Levels of complement fractions C3 and C4 were determined by immunoturbidity. Total 50% hemolytic complement (CH
50) activity was calculated by using a liposome immunoassay. Antinuclear antibodies were detected by indirect immunofluorescence. The
Crithidia luciliae test was used for detection of anti-double-stranded DNA (anti-dsDNA). Antiextractable nuclear antigen (anti-ENA) autoantibodies were measured by Ouchterlony double-immunodiffusion. Prior to participation, all participants gave their informed consent to donate their blood samples. The study was approved by the institutional review boards of the Oklahoma Medical Research Foundation and Temple University.
Table 1
Demographic, clinical and immunological characteristics of the patients
a
Age (years) | 39.6 ± 14.1 | 46.7 ± 15.4 | 45.4 ± 15.9 |
Sex (F:M ratio) | 3.9:1 | 3.1:1 | 3.1:1 |
Ethnicity | | | |
Caucasian (%) | 88.9 | 86.9 | 86.9 |
African (%) | 6.7 | 6.5 | 6.5 |
Asian (%) | 2.2 | 3.7 | 3.7 |
American Indian (%) | 2.2 | 2.8 | 2.8 |
ACR total (number of criteria met) | 5.51 ± 1.69 | 5.52 ± 1.69 | |
Antichromatin Ab | 32.7 | 28.9 | |
Anti-dsDNA Ab (Crithidia luciliae test) (%) | 27.1 | 33.3 | |
Anticardiolipin IgG Ab (%) | 38.3 | 28.9 | |
LAC (%) | 13.1 | 11.1 | |
Anti-Smith Ab (%) | 18.7 | 11.1 | |
Anti-RNP Ab (%) | 20.6 | 17.8 | |
Anti-Ro/SSA (Ro52/60 kDa) Ab (%) | 38.3 (21.5/34.6) | 31.1 (17.8/28.9) | |
Anti-La/SSB (anti-Ro60 + anti-La) Ab (%) | 16.8 (15.0) | 15.6 (13.3) | |
CH50 <40 (U/ml) | 51.4 | 55.6 | |
C4 <16 (mg/dl) | 35.5 | 44.4 | |
SLEDAI score | 5.29 ± 4.20 | 5.02 ± 3.26 | |
Renal involvement (%) | 13.1 | 24.4 | |
(BILAG grades A to C) |
Mucocutaneous involvement (%) | 68.2 | 57.8 | |
Musculoskeletal involvement (%) | 77.6 | 66.7 | |
Cardiovascular/respiratory involvement (%) | 14 | 15.6 | |
Hematological involvement (%) | 37.4 | 31.1 | |
Vasculitis (%) | 53.3 | 53.3 | |
Neurological involvement (%) | 3.7 | 4.4 | |
BILAG total score | 6.17 ± 4.53 | 5.70 ± 3.93 | |
Gas6 plasma levels (ng/ml) | 18.81 ± 8.67 | 17.64 ± 7.10 | 15.89 ± 6.88 |
Free protein S plasma levels (μg/ml) | 6.78 ± 2.36 | 6.37 ± 1.79 | 6.91 ± 1.74 |
Cell cultures
Monocytes from buffy coats of healthy blood donors were isolated with Ficoll-Paque PLUS gradient (GE Healthcare Life Sciences, Pittsburgh, PA, USA) and by magnetic separation using a kit for human monocyte enrichment by negative selection (EasySep cell isolation platform; STEMCELL Technologies, Vancouver, BC, Canada) according to the manufacturer’s instructions. The purity of CD14+ cells was >90% as assessed by flow cytometry. CD14+ cells were cultured for 3 days at 0.8 × 106 cells/ml in 24-well plates containing serum-free X-VIVO 15 medium (Lonza, Walkersville, MD, USA) in the presence or absence of macrophage colony-stimulating factor (M-CSF) (50 ng/ml; PeproTech, Rocky Hill, NJ, USA), granulocyte macrophage colony-stimulating factor (GM-CSF) (100 ng/ml; PeproTech), IL-10 (50 ng/ml; PeproTech), IFN-α (3,000 U/ml; Novus Biologicals, Littleton, CO, USA), IFN-β (3,000 U/ml; PeproTech), IFN-γ (2.5 ng/ml; R&D Systems, Minneapolis, MN, USA), IL-4 (20 ng/ml; Novus Biologicals), IL-17 (100 ng/ml; R&D Systems) or dexamethasone (100 nM; Sigma-Aldrich). When specified, on day 2, cells were coincubated with lipopolysaccharide (LPS) (100 ng/ml; Sigma-Aldrich) for the remaining 24 hours. Cells were then harvested by centrifugation. Supernatants were collected and immediately stored at −20°C for a few days before being tested by enzyme-linked immunosorbent assay (ELISA). Pellets were resuspended in phosphate-buffered saline (PBS) and immediately analyzed by flow cytometry.
Enzyme-linked immunosorbent assay
Plasma concentrations of sAxl, sMer and sCD163 were measured by sandwich ELISA according to standard procedures [
24]. Briefly, 96-well plates were precoated overnight with a capture antibody. Heparinized plasma samples were diluted 1:10 in PBS containing 1% bovine serum albumin (BSA) and applied to precoated plates in duplicate. Serial dilutions of purified recombinant Axl, MerTK or CD163 proteins were used to construct a standard curve. Blank wells were used to hold 1% BSA. For
in vitro studies, cell culture supernatants were not diluted, and blank wells received serum-free X-VIVO 15 medium. Antigens were detected by a secondary biotin-conjugated antibody and horseradish peroxidase–conjugated streptavidin (BioLegend, San Diego, CA, USA). The plate was developed with 3,3′,5,5′-tetramethylbenzidine substrate. The reaction was stopped with 2 N sulfuric acid. Absorbance was detected at 450 nm and read with a reference wavelength set at 570 nm using a VersaMAX ELISA microplate reader (Molecular Devices, Sunnyvale, CA, USA). The optical density for each point was the average of duplicate samples. Concentrations were determined using SoftMax software (Molecular Devices) by applying four-parameter logistic regression to the standard curve. For sAxl quantitation, we used a mouse monoclonal anti-Axl Ab (clone 108724; R&D Systems) for capture, recombinant human Axl (R&D Systems) for the standard curve and a biotinylated goat polyclonal anti-Axl Ab (R&D Systems) for detection. For sMer quantitation, we used the Human Total Mer DuoSet IC (DYC891; R&D Systems) according to the manufacturer’s instructions. For sCD163 quantitation in plasma samples, we used the Human CD163 Quantikine ELISA Kit (DC1630; R&D Systems) according to the manufacturer’s instructions. For sCD163 quantitation in supernatants, we used a mouse monoclonal anti-CD163 Ab (clone EDHu-1; Novus Biologicals) for capture, recombinant human CD163 (R&D Systems) for the standard curve and a biotinylated goat polyclonal anti-CD163 Ab (R&D Systems) for detection.
Flow cytometry
Membrane expression levels of Axl, MerTK and CD163 were measured in cultured monocytes after being washed in buffer containing 2% BSA. Monocytes were gated on the basis of forward and side light scatter and by using a phycoerythrin/cyanin 7 (PE/Cy7)-conjugated anti-CD14 antibody (BioLegend). The following mouse monoclonal antibodies were used for detection: PE-conjugated anti-MerTK (clone 125518; R&D Systems), PE-conjugated anti-Axl (clone 108724; R&D Systems) and allophycocyanin (APC)-conjugated anti-CD163 (clone GHI/61; BioLegend). Expression levels were evaluated using appropriate PE-labeled and APC-labeled isotype controls (BioLegend). Cells were analyzed using a FACSCalibur flow cytometer (BD Biosciences, San Jose, CA, USA) and FlowJo software (Tree Star, Ashland, OR, USA).
Statistical analysis
Data are expressed as means ± SD. Comparisons of soluble receptor levels between patients and matched controls or between groups of patients with different laboratory or clinical characteristics were made using the Mann-Whitney U test. Correlations between soluble receptor levels and other continuous laboratory data were analyzed using Spearman’s rank correlation coefficient. Correlations of soluble receptor levels with the weighted scales of SLEDAI and the total BILAG index were made using Pearson’s correlation coefficient. Comparisons in soluble receptor levels among patients with inactive, moderately active or very active SLEDAI scores were made using one-way analysis of variance (ANOVA) with the Newman-Keuls multiple comparison test. For in vitro studies, differences between cell treatment groups were calculated using a paired Student’s t-test or one-way repeated-measures ANOVA (Newman-Keuls post hoc analysis) when more than two treatment groups were compared. Prism software (GraphPad Software, La Jolla, CA, USA) was employed for all analyses and graphing. A P value <0.05 was considered statistically significant.
Discussion
SLE is characterized by impaired macrophage phagocytosis of ACs [
1], delayed and proinflammatory AC clearance [
2,
3] and increased cellular expression of the type I IFN-inducible gene spectrum: the so-called IFN “signature” [
25,
26]. All these events reflect and contribute to aberrant stimulation of innate immunity. The family of the TAMRs acts to impede such events, thereby preventing systemic autoimmunity. In particular, MerTK is key to efficient clearance of early ACs and to macrophage production of anti-inflammatory cytokines [
9‐
14], and Axl is primarily involved in feedback pathways controlling type I IFN-mediated innate immune activation [
15‐
17]. In the present study, we analyzed the levels of sAxl and sMer receptors in the circulation of SLE patients and investigated potential relations with the clinical, laboratory and immunological aspects of the disease.
We found that increased levels of both sMer and sAxl are associated with general traits of systemic immunity, such as antinuclear and antiphospholipid autoantibody positivity. Additionally, both correlated with hematologic and renal involvement. Nevertheless, we found that sMer, but not sAxl, was significantly associated with lupus-specific humoral autoimmune responses, which were characterized by production of anti-dsDNA, anti-Sm, anti-RNP and anti-Ro60 autoantibodies. Remarkably, only sMer showed strong correlations with disease activity indices, such as C3 and C4 reduction, circulating titers of anti-dsDNA and SLEDAI and total BILAG scores. Compared to matched healthy controls, plasma levels of sMer, but not sAxl, were found to be higher in patients with active lupus (SLEDAI score ≥6), active BILAG renal score and anti-dsDNA and anti-Ro60 positivity. The highest values of sMer were observed in patients with very active lupus (SLEDAI score ≥9). Differences between sAxl and sMer also included relations with their ligands, Gas6 and ProS. In particular, sAxl directly correlated with Gas6 levels, whereas sMer correlated with reduced levels of free ProS. Notably, we found that sAxl and sMer were produced by different immune phenotypes of monocytes/macrophages. sAxl release was induced in the presence of either IFN-α or IFN-β, and sMer was released by M2c differentiated cells, similarly to what we observed for sCD163, a well-known marker of M2 activation. In fact, concentrations of sMer in the circulation of lupus patients directly correlated with plasma levels of sCD163, and sCD163, similarly to sMer, significantly correlated with disease activity. Combining type I IFN exposure with M2c polarizing conditions reduced M2c-driven sMer production while increasing IFN-α-induced sAxl release. The prototypical T-helper cytokines IFN-γ, IL-4 and IL-17 did not exert significant influences on either sAxl or sMer production.
To the best of our knowledge, herein we describe for the first time sMer as a biomarker of M2c activation, in parallel with sCD163. We confirmed the correlation between SLEDAI scores and plasma levels of sMer reported by Wu
et al.[
28] and Recarte-Pelz
et al.[
29]. We also have shown a direct correlation of sMer with sCD163 levels and a significant correlation between SLEDAI and sCD163 levels. Our data strongly suggest a strict relation between SLE activity and M2c homeostasis, in agreement with recent data from Nakayama
et al. showing sCD163 associations with anti-dsDNA positivity and leukopenia [
30]. Similarities between sMer and sCD163, with regard to their expression patterns and their associations in SLE, are consistent with the fact that their respective membrane receptors MerTK and CD163 are both upregulated on the surface of regulatory M2c monocytes/macrophages [
9]. Both are cleaved by the same metalloproteinase, ADAM-17 [
20,
27], in contrast to sAxl, which is cleaved by ADAM-10 [
18]. Both MerTK and CD163 serve to trigger IL-10 release from M2c cells [
9,
31], and both protect macrophages from oxidative stress and subsequent apoptosis induced by hydrogen peroxide, oxidized lipoproteins or iron-containing heme [
32‐
34].
The biological significance of sMer and sCD163 in SLE can be construed as due to at least two mechanisms. Correlations of sMer and sCD163 with SLE activity may indicate a compensatory increase in M2c activation and turnover of monocytes and/or macrophages, with the aim of promoting efferocytosis and immune regulation in response to the still poorly defined inflammatory triggers and to the increased rates of apoptosis. Alternatively, excess ectodomain shedding of MerTK and CD163 by ADAM-17 may account for a functional impairment of M2c monocytes/macrophages and could itself contribute to chronic inflammation, defective clearance of early ACs and autoimmunity. It is known, in fact, that TAMR-knockout mice develop hyperreactive immune responses and severe lymphoproliferation [
11,
35]. In particular, disrupted MerTK expression is associated with a SLE-like syndrome in mice [
36], and gene polymorphisms of MerTK and Gas6 are associated with clinical manifestations in SLE patients [
37,
38]. Besides gene defects and polymorphisms, posttranslation inhibition of these molecular pathways through ectodomain shedding may affect efferocytosis and regulatory responses [
19], thus favoring accumulation of AC-derived autoantigens. ADAM metalloproteinases are, in fact, activated upon multiple conditions, including infections, oxidative stress and paracrine signals [
18‐
20,
27,
39‐
42]. Of note, ADAM-17 is also known to cleave and inhibit the membrane receptor for M-CSF [
43], which is needed for complete M2c differentiation [
9]. Besides its crucial role in promoting macrophage release of major proinflammatory mediators, including TNF-α and IL-6 [
40], ADAM-17 may thus exert its proinflammatory effects by interfering with differentiation and activity of regulatory M2c macrophages. From this perspective, impeding ectodomain shedding by the use of safe and selective ADAM inhibitors might help to restore macrophage homeostasis in SLE [
44].
Cleavage of Axl into sAxl may in turn alter the homeostatic mechanisms regulating TLR-mediated activation [
15,
16], thus resulting in exaggerated production of IFN-α in response to AC-derived autoantigens. Excess activation of TLR/IFN pathways may ultimately lead to dendritic cell maturation, presentation of autoantigens to autoreactive T cells, chronic B-cell activation, oligoclonal expansion of plasmablasts and production of autoantibodies [
26]. In addition, both sMer and sAxl are able to sequester the ligand Gas6 [
18,
19,
45], thus interfering with membrane TAMR-induced regulatory signaling. Contrary to Ekman
et al.[
46], however, we could not confirm a significant association between SLEDAI scores and plasma levels of sAxl. Similarly, Recarte-Pelz
et al.[
29] failed to find such an association. The discrepancy might be due to differences between patient populations or to the use of different detection reagents. The same ELISA kit (R&D Systems) was used by Wu
et al.[
28], Recarte-Pelz
et al.[
29] and our laboratory for detection of sMer in SLE patients. For sAxl, instead, Ekman
et al.[
46] used an ELISA type developed in their laboratory, whereas we and Recarte-Pelz e
t al.[
29] used the same commercially available anti-Axl detection antibody (R&D Systems). The weaker association with SLE activity of sAxl compared to sMer suggests a more indirect role of sAxl in SLE pathogenesis. Whereas the cleavage of MerTK may be critical for the accumulation of AC-derived autoantigens and production of pathogenic lupus-specific autoantibodies, the cleavage of Axl could be more generally related to uninhibited TLR activation and production of IFN-α/β and other proinflammatory cytokines. Consistent with this view, sAxl, but not sMer, was found to be increased in nonautoimmune inflammatory diseases such as critical limb ischemia [
28,
47].
Among lupus-specific autoantibodies, sMer levels showed the strongest association with anti-Ro60 antibodies, particularly in the absence of a concomitant anti-La positivity. No association was found with anti-Ro52 antibodies. In fact, serum positivity for anti-Ro60 best discriminated patients with significantly higher levels of sMer compared to matched healthy controls. It is noteworthy that Ro60 is translocated to the cell surface of ACs during early apoptosis independently of La and Ro52 [
48]. Autoantibodies against surface-exposed Ro epitopes are specific for a subset of SLE patients showing positivity of anti-Ro60 without anti-La, whereas double-positivity of anti-Ro60 and anti-La is consequent to intermolecular spreading from Ro to La, in which antigens are exposed on late ACs or released from necrotic ACs [
48]. Anti-Ro52 antibodies are instead more prevalent in conditions other than SLE, such as primary Sjögren’s syndrome and idiopathic inflammatory myopathies [
49]. Because production of anti-Ro60 antibodies represents a lupus-specific humoral autoimmune response against early ACs [
48] and MerTK is specifically required for M2c macrophage phagocytosis of early ACs [
9], the strong association that we found between anti-Ro60 and sMer in SLE patients might reflect a compensatory increase in M2c activation of monocytes and/or macrophages to enhance the clearance of early ACs by MerTK. Alternatively, the accumulation of early ACs fostering anti-Ro60 production might be itself a consequence of excess ectodomain shedding of MerTK, which would interfere with the clearance efficiency of M2c cells. The latter hypothesis suggests a putative role for the cleavage of MerTK in SLE pathogenesis, at least in a subgroup of anti-Ro60-positive patients.
Our data pertaining to the relation of sAxl and sMer to Gas6 and ProS levels are consistent with previous data on receptor-ligand binding affinity. Correlation between sAxl and Gas6 is in fact consistent with the tenfold higher binding affinity of Gas6 to Axl than to MerTK [
50], as well as with the previous finding that Gas6 is primarily complexed with sAxl in human blood [
45]. Correlation between sMer and reduced free ProS levels is consistent with the fact that ProS binds to MerTK [
50], whereas no connection between ProS and Axl has been demonstrated to date. ProS serves as the main bridging molecule between phosphatidylserine on ACs and MerTK on the surface of human monocytes and/or macrophages [
8]; however, whether sMer binds to ProS remains to be established. ProS needs to oligomerize to bind to Mer by interacting with other molecules of ProS on phosphatidylserine-containing surfaces [
51]. Intriguingly, ProS is also able to bind to microparticles, besides ACs [
52]. Plasma microparticles and/or circulating ACs may therefore serve as a scaffold for ProS oligomerization in circulation. In SLE, levels of microparticles increase with disease activity [
53], whereas levels of free ProS decrease with disease activity [
24]. It is tempting to speculate that, in active SLE patients, ProS may bind to microparticles, thus provoking reduction in free ProS levels, ProS oligomerization and potential formation of ProS-sMer complexes. In support of this view, it has been shown that HIV-infected patients also show reduced free ProS levels, and this reduction has been related to ProS binding to circulating microparticles [
54]. Further investigation is needed to address this hypothesis.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
GZ contributed to the conception and design of the study; performed experiments; collected, analyzed and interpreted data; and wrote the manuscript. JG and LMD performed experiments and collected data. JTM provided plasma samples and clinical characterization of the Oklahoma Cohort for Rheumatic Diseases. PLC conceived, designed and coordinated the study; interpreted data; and critically revised the manuscript. All authors approved the final version of the manuscript.