Abstract
Objective. Childhood-onset systemic lupus erythematosus (cSLE) is usually a more severe and aggressive disease than adult-onset SLE (aSLE), but cellular and subcellular reasons for these differences are not well understood. The present study analyzed Th subsets, STAT1/STAT5 signaling response, and cytokine profiles of cSLE.
Methods. FOXP3+ regulatory (Treg) and effector Th subsets, expression and phosphorylation of STAT1/STAT5 in Th, and cytokine profiles were measured in the peripheral blood of patients with cSLE and healthy controls (HC), using flow cytometry and immunoassay on a biochip.
Results. Significant correlation between expression of the activation marker HLA-DR and decreased Th counts, an increase in the percentage of FOXP3+ Th, and a decrease in the activated Treg (aTreg) subset among them were found in cSLE. In contrast to our previous findings in aSLE, no significant differences in percentages and a significant decrease in the numbers of the naive-resting Treg (rTreg) subset compared to HC were found. The percentages of CD25− cells, possibly reflecting interleukin 2 depletion, were significantly increased in cSLE aTreg, but not in the rTreg subset. Consistent with the results of our previous studies in aSLE, increased expression of STAT1, along with significant correlation between decreased Th counts and their increased basal phosphorylation of STAT5, were also found in cSLE.
Conclusion. Our results suggest that the key difference in Treg homeostasis between cSLE and aSLE is in the rTreg subset. However, perturbed aTreg homeostasis, increased levels of STAT1 protein, and homeostatic STAT5 signaling appear to be intrinsic characteristics of the disease, present in cSLE and aSLE alike.
Systemic lupus erythematosus (SLE) is a chronic multi-system autoimmune inflammatory disease with diverse clinical and laboratory signs1. The exact etiopathogenesis of SLE is not yet fully known2. Pathogenic T lymphocytes from SLE can demonstrate many signs of altered function, which in the end stimulates the autoimmune inflammatory state3. The population of CD4−CD8− double negative T lymphocytes is higher in patients with SLE than in healthy controls (HC) and can induce autoreactive B lymphocytes to produce anti-dsDNA antibodies and higher numbers of inflammatory cytokines4. Changes are also noticeable in individual subpopulations of helper T lymphocytes (Th). It is suspected that disruption of the homeostasis between the effector Th (Teff) and the regulatory Th (Treg) cells can lead to various inflammatory and autoimmune diseases, including SLE5.
According to the major transcription factors, cytokines needed for their differentiation, as well as the characteristic cytokines secreted by the activated cells, Teff can be divided into Th1, Th2, Th17, and others6. Despite the considerable functional specialization of subtypes, there is some plasticity between different Teff, which sometimes allows multifunctional subpopulations to arise. Examples of the latter are Th1Th17, which secrete interferon-γ (IFN-γ), characteristic for Th1, as well as highly inflammatory interleukin (IL)-17, characteristic for Th177.
Treg control the immune response by direct inhibition of Teff, but also by reducing the antigen presentation potential of dendritic cells and altering their release of cytokines. The depletion of Treg can thus contribute to the onset and maintenance of autoimmune diseases in 2 distinct ways. The removal of cells with a suppressor function can lead to unrestrained polyclonal activation of Teff; additionally, systemic expansion and maturation of dendritic cells can increase the likelihood of the presentation of self-antigens8. The total number and activation of Treg is usually reduced in SLE; moreover, an elevated number of FOXP3-expressing cells, which are functionally not suppressive, has been found9,10. Many previous studies have focused also on IL-17, expansion of Th17, and the Treg:Th17 dynamic in the pathogenesis of SLE11,12,13.
Proper functioning of Th depends on their successful signal transduction. Cytokine receptors receive signals of different extracellular molecules and trigger signal transduction through the Janus kinase and STAT pathway14. Canonically it begins with 3 consecutive tyrosine phosphorylations, last being phosphorylation of STAT proteins. The phosphorylation rate of STAT can therefore offer relevant insight into the activation status of the cell. Aberrant cytokine signaling is associated with many immune-mediated disorders, in particular, inflammatory conditions and autoimmune diseases such as SLE15,16.
The primary roles of STAT1 are the transmission of IFN signals and the activation of antiviral inflammatory response by excretion of new IFN. It regulates the cytokine production of Th1 and controls the proliferation and apoptosis of other immune cells17. In SLE, as well as in the SLE mice model, basal expression of STAT1 was raised18,19. Phosphorylation of STAT1 was also slightly increased and in positive correlation with the disease duration in mice T lymphocytes, but not in humans with adult-onset SLE (aSLE)20. Higher expression of STAT1 mRNA was reported in lupus nephritis and was correlated with disease activity and IFN-dependent gene expression21,22. The rise in basal expression and phosphorylation of STAT1 can thus be a result of — and a help in maintaining — the inflammatory environment of affected cells.
STAT5 transmits signals of common γ-chain cytokines (such as IL-2), and induces expression of FOXP3, a transcription factor crucial for the maturation of Treg23. STAT5-dependent genes are tissue-specific, but generally STAT5 controls survival, proliferation, differentiation, and regulation of the cell cycle. It plays a central role in maintaining peripheral tolerance through activation of Treg and is thus likely to be involved in the pathogenesis of autoimmune diseases24. Accordingly, the levels of phosphorylated STAT5 (pSTAT5) were elevated in aSLE and were in correlation with their disease activity19. STAT3 and STAT5 can bind to many identical binding sites on the gene for IL-17. STAT5 can competitively block STAT3-dependent transcription of IL-17 and inhibit Th17 polarization, leading to a decrease of inflammation25,26.
Childhood-onset SLE (cSLE) is usually a more severe form of the disease compared to aSLE. It can involve more organs and require more aggressive treatments, increasing the possibility of longterm drug toxicity and disease damage. Systemic manifestations of the disease and development of anti-dsDNA antibodies are also more common in cSLE27. The causes of disease onset at different ages and the clinical differences between cSLE and aSLE are not yet well understood, but factors contributing to the more severe childhood form of the illness are likely to be predominantly genetic predisposition, a higher incidence of acute infections, and immaturity of the immune system and other organ systems28.
In our study, subpopulations of Th, their expression and phosphorylation of STAT1 and STAT5, and plasma cytokine concentrations were analyzed in cSLE. Comparison of our results with findings on aSLE enhances understanding of the similarities and differences between the 2 types of the disease.
MATERIALS AND METHODS
Study subjects
Seventeen cSLE patients with a median age of 18.0 years at enrollment were included in the study. All patients were followed at the Children’s Hospital, University Medical Centre Ljubljana, Slovenia. Disease activity was assessed using the Systemic Lupus Erythematosus Disease Activity Index 2000 (SLEDAI-2K) scoring system29. As a control group, we included 20 healthy adolescents (HC) with a median age of 16.0 years and no history of allergies, acute infections, autoimmune disorders, or medications that could affect the immune system. Demographic, clinical, and laboratory data are presented in Table 1. The study was approved by the National Medical Ethics Committee of the Republic of Slovenia (approval number: 27/11/11); each participant or a legal guardian signed an informed consent form.
Antibodies and sample staining for analysis of lymphocyte subpopulations
Whole venous EDTA blood was aliquoted. Plasma was collected from 1 aliquot and stored for later cytokine analysis. The remaining aliquots were prepared following 3 different protocols. Standard whole-blood staining methodology with premixed multicolor panels (Tube 1–3; Table 2), according to the manufacturer’s instructions, was used when staining solely surface antigens. STAT expression/phosphorylation was studied following the BD Phosflow protocol (Phosflow Lyse/Fix Buffer and Perm Buffer III). Staining of surface and intracellular antigens was in this case done simultaneously (Tube 5; Table 2). Treg were detected after primary staining of surface antigens in whole blood, with subsequent fixation and permeabilization using BD HumanFOXP3 Buffer set and staining of FOXP3 (Tube 4; Table 2).
All reagents, except for anti–CD3-PerCP, pSTAT5A-AlexaFluor647 (antibodies-online GmbH), anti-CXCR3-APC (BioLegend Inc.), and anti-CD161-FITC (eBioscience Inc.), were acquired from BD Biosciences. Cells were analyzed with the FACSCantoII Flow Cytometer, equipped with blue and red lasers, running FACSDiva software (both BD Biosciences). Digital data were analyzed using FlowJo software (Tree Star Inc.).
Flow cytometric analysis of lymphocyte subpopulations
Percentages of Th, cytotoxic lymphocyte T (Tc), CD4−DC8− T lymphocytes, HLA-DR+ T lymphocytes were analyzed after gating on CD3+CD45+ lymphocytes. Absolute cell counts were calculated using a dual-platform approach with panleukogating after measuring leukocytes on the Beckman Coulter AcT8 Hematology Analyzer (BD Biosciences). Analysis of all lymphocyte subpopulations and measurement of STAT expression/phosphorylation were conducted after gating on the CD3+CD4+ cell population. The exact antibodies combination used to identify each subset can be found in Table 2. Gating strategies were the same as previously described30. Median fluorescent intensity (MFI) was used to measure expression and phosphorylation of STAT proteins.
Samples were obtained and studied individually; standard calibration beads (BD Biosciences) to set the forward, side scatter, and photomultiplier voltage were used for consistency before each experiment. MFI was normalized by subtraction of MFI values of CD3−CD4− cells in each sample (∆MFI).
Plasma cytokine array on a biochip
All plasma samples were stored at −80°C and analyzed in a single experiment. Cytokine concentrations were analyzed using a kit with predesigned multiplex cytokine immunoassay on a biochip with pre-applied spatially discrete test regions, allowing for simultaneous determination of 10 cytokines [IL-1α, IL-1ß, IL-2, IL-4, IL-6, IL-8, IL-10, tumor necrosis factor–α (TNF-α), IFN-γ, monocyte chemoattractant protein 1 (MCP-1)] in a single sample at a single timepoint (Cytokine & Growth Factors Array I) with the Evidence Investigator immunoanalyser (both Randox), according to the manufacturer’s instructions.
Statistical analysis
The Mann-Whitney U test was used to test differences between 2 groups and the Kruskal-Wallis among 3 groups. For within-group comparisons, the Wilcoxon matched-pairs signed-rank test was used. Correlations between experimental results were examined using Spearman rank test. A value of p < 0.05 was considered significant in all statistical tests. Statistical data analysis was performed using the GraphPad Prism software (GraphPad Software Inc.).
RESULTS
Severe lymphopenia leads to lower numbers of lymphocyte subsets in cSLE, but percentages of Th lymphocytes do not differ significantly
Lymphocyte number and their function are strongly implicated in the pathogenesis of SLE31,32. As well, significant lymphopenia was found in our patients with cSLE (p < 0.0001) compared to HC. The number of analyzed subsets including T lymphocytes, CD4−DC8− T lymphocytes, Th, and Tc was significantly lower in cSLE (first 3 subsets: p < 0.0001, Tc: p < 0.05; Figure 1A, Table 3). Lymphocytes were decreased also relatively in cSLE (% of leukocytes, p < 0.01), as were T lymphocytes (% of lymphocytes, p < 0.01; Figure 1B). While percentages of Th did not differ between groups, we found decrease within CD4−DC8− T lymphocytes and Tc in cSLE (p < 0.01 and p < 0.05; Figure 1B, Table 3).
Treg and Teff subsets show signs of perturbed homeostasis leading to decline in Treg function
Cells undergoing homeostatic lymphopenia-induced proliferation were shown to develop an activated phenotype33. To determine whether this effect influenced T lymphocyte reconstitution in cSLE Th depletion, we examined the expression of the marker of activation HLA-DR. Significant negative correlation was found between Th counts and the percentage of HLA-DR+ T lymphocytes from cSLE, but not HC (Figure 2A).
Because Treg reconstitution in cSLE could also be characterized by a shift in the activated/effector status of these cells, we performed Treg quantification as proposed by Miyara, et al34, differentiating between CD45RA−FOXP3high activated-effector Treg (aTreg) cells, CD45RA+FOXP3low naive-resting Treg (rTreg), and the CD45RA−FOXP3low activated effector T lymphocytes [abbreviated as non-Treg, even though they represent only a small and distinct subset of the conventional T lymphocyte (Tcon) fraction] subset among FOXP3-expressing cells.
In line with lymphopenia, numbers of aTreg and rTreg were significantly lower in cSLE than HC (p < 0.001 and p < 0.05, respectively; Table 3), which was reflected in the significant decrease of the total number of Treg (aTreg + rTreg) in cSLE (p < 0.01). However, in contrast to the absolute numbers, while percentages of aTreg were significantly decreased among FOXP3+ Th from cSLE and no significant difference was found in the rTreg, the percentage of non-Treg was significantly increased in cSLE (p < 0.001) compared to HC (Figure 2B).
Because data suggest that an acquired insufficiency of IL-2 in SLE accounts for the reduced expression of CD25 in FOXP3+ Treg and the inverse increase in the proportions of the CD25− subset35, we also analyzed CD25− cells in the rTreg, non-Treg, and aTreg subsets from cSLE and HC. While the highest percentage of CD25− cells was found in the non-Treg from both cSLE and HC, it was significantly increased only in aTreg from cSLE compared to HC (Figure 2C).
Percentages of CD25− cells among all FOXP3+ Th, as well as the percentage of FOXP3+ cells among Th, increased significantly in cSLE compared to HC (Figures 2D–E).
The percentage of FOXP3+ cells among Th was also significantly increased after in vitro treatment of purified Th from healthy donors with IL-7 (Figure 2F). Therefore, if diminished in vivo availability of IL-2 accounts for the increase in the CD25− subset from SLE, the other homeostatic STAT5 signaling cytokine IL-7 could be responsible for the increase in the frequency of FOXP3+ cells among Th.
In contrast to FOXP3+ Th, the percentage of CD25+ cells among FOXP3− Tcon was higher in cSLE, but the difference was not statistically significant (p = 0.10; Figure 2G).
Analysis of cell numbers showed a significant decrease in the number of Tcon, Th1, and Th1Th17-like cells in cSLE (p < 0.001, p < 0.05, and p < 0.01, respectively). Percentages of the latter effector Th populations, Th2, and Th17-like cells were, however, not significantly different compared to HC (Table 3).
Increased expression of STAT1 and homeostatic IL-7–dependent STAT5 activation are present in cSLE
We found significantly increased STAT1 expression in Th from cSLE (Figure 3A).
Because significant increases in Th, but not CD3− lymphocyte levels of basal pSTAT5, were observed previously by our research group in aSLE, the increase in Th pSTAT5 (ΔMFI and %) was also examined in cSLE36. Although no statistical differences in basal pSTAT5 between groups were found (Figure 3B), higher pSTAT5 ΔMFI was significantly correlated to lower CD4 counts in cSLE, but not HC (Figure 3C).
Finally, the increase in pSTAT5 levels in Th was dependent on homeostatic cytokine IL-7, because incubation of whole-blood samples from SLE with neutralizing anti–IL-7, but not anti–IL-2 antibodies, resulted in significant reduction of pSTAT5 ΔMFI values (Figure 3D).
Concentrations of plasma cytokines do not differ significantly between the groups
A comparison of the number of cytokines IL-1α, IL-1β, IL-2, IL-4, IL-6, IL-8, IL-10, TNF-α, IFN-γ, and MCP-1 in plasma samples found no statistically significant differences between groups (Supplementary Table 1, available from the authors on request). There was also no statistically significant correlation between SLEDAI-2K of patients at the time of the study and any of the analyzed laboratory markers.
DISCUSSION
A more aggressive clinical course usually occurs with cSLE compared to aSLE, but pathogenetic reasons for this difference are not entirely understood. Our previous studies have identified significant disruption of immune cell homeostasis and changes of cytokine signaling in aSLE36,37. In this study, we determined subpopulations of Treg and Teff, cytokine profile, and basal STAT1/STAT5 expression and phosphorylation in a single-center cohort of cSLE.
Lymphopenia is frequent in all SLE, but more likely to appear in cSLE, with T lymphocytes, especially Th, more affected than B lymphocytes28,32. Our results showed the same trend of pronounced lymphopenia in cSLE, including decreased numbers of T lymphocytes, Th, and even the CD4− DC8− T lymphocytes subpopulation (Figure 1A), which was shown to be expanded in aSLE and to produce IL-17 and IFN-γ in vivo4. In addition, percentages of CD4−DC8− T lymphocytes, but not Th among T lymphocytes, were significantly decreased in cSLE compared to HC (Figure 1B).
Lymphopenia with associated compensatory homeostatic proliferation in response to IL-7 can release autoreactive Th from inhibitory networks38. T lymphocyte recovery was driven by homeostatic proliferation also in patients with multiple sclerosis treated with the lymphocyte-depleting monoclonal antibody alemtuzumab, and their T lymphocytes showed evidence of chronic activation33. Consistent with that, we found significant correlation between decreased Th numbers in cSLE and expression of the activation marker HLA-DR (Figure 2A). However, in contrast to our findings on cSLE and aSLE (previous study36,37), increased aTreg were reported after alemtuzumab therapy33.
The Treg population should ideally be defined in a way that reliably excludes effector cells, and FOXP3+CD127low cells are advocated to accurately identify real Treg39. However, stimulation of purified Th with IL-7 significantly decreased the difference in CD127 (IL-7R) expression between FOXP3+ and FOXP3− subsets, while it increased percentage of FOXP3+ cells among Th. Downregulation of CD127 is therefore not entirely specific for Treg either (Supplementary Figures 1–2, available from the authors on request). We performed identification and analysis of Treg by using antibodies against CD4, CD25, FOXP3, and CD45RA antigens (Table 2) and used activation status of the FOXP3+ cells — expression of CD45RA — as the distinguishing marker for Treg34. Increased percentages of FOXP3+ Th were found in patients with cSLE (Figure 2E), but were not reflected in increased percentages of both Treg subsets: aTreg, which actively perform a suppressive function, and rTreg, which upon activation differentiate in aTreg. However, the non-Treg subset was significantly increased among Th from cSLE (Figure 2B). The same subset, which we found to be increased after stimulation with IL-7 in vitro before, was increased also in our aSLE36,40.
Systemic reduction of IL-2 levels in the early stages of the disease promoted Tcon hyperactivity and accelerated disease progression in a mouse model, highlighting the importance of the Treg-IL-2 axis41. As shown by von Spee-Mayer, et al, in the aSLE population, deficiency of IL-2 can lead to a disease activity–related decrease in the expression of CD25 on Treg35. In this case, distinction between cells may be difficult; they are in reality just “exhausted” Treg owing to IL-2 deprivation, and FOXP3-expressing “false” Treg, which actively excrete IFN-γ and IL-17. Indeed, in cSLE, CD25− cells were enriched in non-Treg and were significantly increased among FOXP3+ Th from cSLE compared to HC (Figure 2C). IL-2 levels in plasma from cSLE patients were, in our study, not decreased compared to HC (Supplementary Table 1, available from the authors on request). However, because of the short half-life of IL-2, measurement of plasma IL-2 is probably not an adequate method to detect shortage of IL-2 in vivo35. It has recently been suggested that in lymphoid tissues FOXP3 expression is maintained in Treg by STAT5-signaling cytokines, such as IL-2 and IL-7. These signals are lost during recirculation in the bloodstream, resulting in decay of FOXP3 in many Treg cells42. Therefore, increased levels of IL-7, which were described previously in our aSLE study and were associated with Th depletion36, may be responsible for the increase in FOXP3+ cells (Figure 2F) and the non-Treg subset among Th. However, IL-7–dependent basal pSTAT5 levels were in our aSLE study36 not significantly increased in the most suppressive aTreg subset34, which was relatively decreased in the pool of FOXP3+ Th from our patients with cSLE (Figure 2B). In addition, the percentage of CD25− cells, possibly also reflecting in vivo IL-2 deficiency, was significantly increased in the aTreg from cSLE (Figure 2C). As for FOXP3, CD25 expression is also regulated by the STAT543. An increased Tcon/aTreg pSTAT5 ratio, found previously in aSLE36, could explain a higher percentage of CD25− cells among aTreg and CD25+ cells among Tcon also in cSLE (Figure 2C and Figure 2G).
Our previous study revealed a significant increase of CD25−FOXP3+ Th in healthy children, compared to healthy adolescents and adults, suggesting that occurrence of this subpopulation could be influenced by the immaturity of the immune system30. The same could be true for the rTreg subset, which showed (in contrast to our previous findings on aSLE36,37) decreased numbers and did not significantly increase among FOXP3+ Th in cSLE compared to HC (Figure 2C).
To our knowledge, numbers and percentages of peripheral blood circulating Th17 were previously not directly investigated in cSLE, but a comparison of numbers of plasma cytokines, associated with Th1, Th2, and Th17, showed strong evidence of crucial involvement of IL-17 in aSLE pathogenesis11,12,13,44. Unlike in aSLE, we found no significant differences in percentages of Th17-like cells in cSLE compared to HC (Table 3). However, our analysis of plasma cytokine concentrations indirectly indicated onset of functional changes in these cells. Despite the significant decrease in numbers of Th1 and Th1Th17-like cells and CD4−DC8− T lymphocytes in cSLE (Table 3), the plasma concentrations of cytokines secreted by these cells did not differ from those in the HC (Supplementary Table 1, available from the authors on request). The cSLE cells were, therefore, probably hyperactive in their response to the stimulus. Lower numbers of Teff subsets could also be a consequence of their recruitment to the site of inflammation, out of the periphery. The low SLEDAI-2K score (Table 1) of cSLE patients without extensive actively inflamed sites at the time of the study and normal plasma cytokine levels, however, make this theory less probable. Our previous study of the age-dependent dynamics of Th subpopulations in healthy subjects suggests that the numbers of Teff are likely to increase in adulthood, owing to the aging process itself30.
On the other hand, STAT1 and STAT5 have both been shown to be capable of suppressing Th17 responses26,45. Consistent with results of our study on aSLE37, an increased expression of STAT1 (Figure 3A) and homeostatic IL-7–dependent STAT5 activation in Th (Figure 3C–3D) were also found in cSLE despite the low disease activity scores of the patients at the time of the study (Table 1), which could be the reason for the absence of expected Th17 expansion. Th depletion was in cSLE associated with higher pSTAT5 levels (Figure 3C), which were shown to confer a worse prognosis in aSLE36. The increased levels of STAT1 expression do not necessarily reflect rapid changes in the clinical activity of the disease, but probably demonstrate an “interferon signature,” and may be associated with increased sensitivity to the new inflammatory signals20,46,47.
Our previous study on aSLE showed that among Th, aTreg are the most sensitive to IFN-α stimulation. They exhibited the highest IFN-α–induced pSTAT1 response combined with decreased proliferation assessed by Ki-67 expression37. IFN-γ, which signals mainly through STAT1, was not increased in plasma from cSLE (Supplementary Table 1, available from the authors on request), but the concentration of IFN-α was not measured in our study. Elevated levels of IFN-α, characteristic for SLE, could inhibit the proliferation of aTreg in cSLE. This could assist with a break of peripheral tolerance and lead to further increase in STAT1 expression and maintenance of an inflammatory condition. Further study of cSLE should perform detailed analysis of STAT1/STAT5 expression and phosphorylation in different subpopulations of FOXP3− Tcon and FOXP3+ Th.
According to our results, decreased numbers of the rTreg subset and lack of significant expansion of Teff subsets compared to HC could, therefore, be interpreted as the key difference in Treg/Teff homeostasis between cSLE and aSLE, while perturbed aTreg homeostasis, higher expression of STAT1 and homeostatic IL-7–dependent basal STAT5 activation, associated with Th depletion, appear to be present in both the childhood- and adult-onset disease.
Footnotes
This work was partially supported by the Slovenian Research Agency grant (grant number L7-8274) and the University Medical Centre Ljubljana research grant (grant number 20180093).
- Accepted for publication June 20, 2019.