Article Text
Abstract
Objective MicroRNAs (miRNAs) regulate the expression of genes involved in immune activation. A study was undertaken to characterise the miRNA signature and identify novel genes involved in the regulation of immune responses in systemic lupus erythematosus (SLE).
Methods The expression of 365 miRNAs in peripheral blood mononuclear cells of patients with SLE and healthy controls was analysed using TaqMan Low Density Arrays. The results were validated by quantitative real-time PCR and potential target genes were identified using prediction analysis software. The effect of miR-21 on T cell function was assessed by transfection with antago-miR-21 or pre-miR-21.
Results A 27-miRNA signature was identified in patients with SLE; 19 miRNAs correlated with disease activity. Eight miRNAs were deregulated specifically in T cells and four miRNAs in B cells. miR-21 was upregulated and strongly correlated with SLE disease activity (r2=0.92). Compared with controls, CD4 T lymphocytes from patients with SLE had higher basal and activation-induced miR-21 expression. Silencing of miR-21 reversed the activated phenotype of T cells from patients with SLE—namely, enhanced proliferation, interleukin 10 production, CD40L expression and their capacity to drive B cell maturation into Ig-secreting CD19+CD38hiIgD−(plasma cells. Overexpression of mMiR-21 in normal T cells led to acquisition of an activated phenotype. Investigation of putative gene- targets showed that PDCD4 (a selective protein translation inhibitor) was suppressed by miR-21 and its expression was decreased in active SLE.
Conclusions miRNAs represent potential biomarkers in SLE as their expression reflects underlying pathogenic processes and correlates with disease activity. Upregulated miR-21 affects PDCD4 expression and regulates aberrant T cell responses in human SLE.
Statistics from Altmetric.com
Introduction
Systemic lupus erythematosus (SLE) is a prototypic autoimmune disease with multiple genetic and environmental factors contributing to its pathogenesis.1 2 Defects in T cell antigen receptor-mediated signalling in SLE have been shown to lead to breakdown of immunological tolerance. Aberrant costimulation and skewed cytokine production by CD4 T lymphocytes in SLE contribute to B cell hyper-responsiveness.3
High-throughput techniques have been used to dissect disease mechanisms and identify novel molecular pathways involved in complex diseases including SLE.4 We and others have previously used cDNA microarrays to identify unique gene signatures in patients with SLE.5,–,7 Although these studies have provided insights into genetic pathways involved in disease pathogenesis, they offer limited information regarding regulatory mechanisms that control gene expression.
MicroRNAs (miRNAs) are, in general, potent negative modulators of genes involved in several cellular processes.8 Compared with other gene regulatory mechanisms such as epigenetic and transcription factors, miRNA-mediated effects occur prior to protein synthesis, thus allowing for the fine tuning of gene expression. Deregulation of miRNA expression has been implicated in the pathogenesis of human diseases and miRNAs are thought to represent novel disease biomarkers and potential therapeutic targets.9
miRNAs regulate the function of both the innate and the adaptive immune system and are involved in various immune pathways.8 10 In mice, deregulation of miRNAs leads to aberrant immune responses and development of systemic autoimmunity.11 12 Altered miRNA expression has been reported in human autoimmune diseases including SLE, rheumatoid arthritis and multiple sclerosis.13,–,16 However, the mechanisms by which these changes promote autoimmunity have not been thoroughly investigated.
In this study we have characterised the miRNA signature of human SLE, reporting a strong correlation of certain miRNAs with SLE disease activity. Importantly, the identified miRNAs are predicted to regulate genes and processes pertinent to the pathogenesis of SLE. Furthermore, in functional studies we show that miR-21, the miRNA with the highest correlation with disease activity, regulates key cellular functions that contribute to aberrant T cell phenotype in patients with SLE and affects PDCD4, a selective protein translation inhibitor of genes involved in immune responses.
Methods
Patients and healthy blood donors
Patients with SLE fulfilling the American College of Rheumatology classification criteria17 were recruited from the Department of Rheumatology, University Hospital of Heraklion (Greece). The SLE Disease Activity Index (SLEDAI) was used to classify patients with active (SLEDAI ≥8) or inactive (SLEDAI <8) disease.18 Peripheral blood samples were obtained from 34 patients (18 with active and 16 with inactive disease) and 25 age- and sex-matched healthy blood donors. All subjects gave written informed consent prior to enrolment in the study, which was approved by the ethics committee of our hospital.
Cell isolation
Peripheral blood mononuclear cells (PBMCs) were freshly isolated by Ficoll-Histopaque (Sigma-Aldrich, St Louis, Missouri, USA) density gradient centrifugation of heparinised venous blood. Patients had not taken any medication for SLE during the 24 h prior to blood sampling. CD4 T and CD19 B lymphocytes were obtained by magnetic separation (Miltenyi Biotec, Miltenyi Biotec, Gladbach, Germany).
MiRNA expression and miRNA target prediction analysis
The expression of 365 miRNAs was analysed by TaqMan Low Density Arrays (TLDA human miRNA v1.0, Dana-Farber Molecular Diagnostics Facility, Harvard Medical School, Boston, Massachusetts, USA) using 2 μg total RNA. RNU48 was used to normalise microarray expression results. Potential miRNA gene targets were identified using the miRBase (http://microrna.sanger.ac.uk), PicTar (http://pictar.bio.nyu.edu) and TargetScan version 4.0 (http://www.targetscan.org/index.html) search engines. To optimise the accuracy of prediction, a potential gene target should be predicted by a minimum of two out of three programs and the targeted sequence should be conserved among species.
Real-time PCR and western blot
Microarray results were validated with the mirVana qRT-PCR miRNA Detection Kit and qRT-PCR Primer Sets (Ambion, Austin, Texas, USA). For western blot, cells were lysed with RIPA buffer containing a complete protease inhibitor cocktail. Proteins were separated by 10% sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to nitrocellulose membrane. PDCD4 was detected using a monoclonal antibody (Cell Signaling Technology, Danvers, MA, USA) and protein loading was determined using a monoclonal antibody for β-actin (Santa Cruz, CA, USA). The secondary antibodies horseradish peroxidase (HRP)-conjugated antimouse IgG and antirabbit IgG were from Jackson Immunoresearch Laboratories (West Grove, PA, USA). Enhanced chemiluminescence was from Thermo Scientific (Vantaa, Finland).
Transfection of pre-miRNAs or antisense-miRNAs
The transfection agent siPORT NeoFX agent (Ambion) was used for silencing and overexpression of endogenous miRNAs. Freshly isolated T lymphocytes were seeded in 96-well tissue culture plates (Corning Incorporation, Corning, NY, USA) and mock-transfected or transfected with irrelevant (scramble) miR, antago-miR-21 or pre-miR-21 (all 75 nM) (Ambion).19 The efficiency and optimal dose of the transfection was determined by real-time PCR. Freshly isolated CD4 T lymphocytes were seeded in 96-well tissue culture plates and were transfected with 2 μg control (pEGFP-C1) vector or PDCD4-pEGFP-C1 expression vector (Cloneth Laboratories, USA). Culture supernatants were collected after 48 h to measure interleukin (IL)-10 (Ready-SET-Go, ELISA kit; eBioscience, San Diego, CA, USA).
T cell assays and T/B lymphocytes co-cultures
Full details are given in the online supplement.
Statistical analysis
The GraphPad Prism software was used for statistical analysis and calculation of mean±SEM values. Comparisons were performed using the Mann–Whitney U test (independent samples) and the paired t test (paired samples). miRNA expression was plotted against SLEDAI scores of individual patients and r2 values (coefficient of determination) were obtained from polynomial trend lines. A two-tailed p value <0.05 was considered statistically significant.
Results
Microarray analysis reveals a distinct pattern of miRNA expression in PBMCs of patients with SLE
We analysed the expression of 365 miRNAs in the PBMCs of 34 patients with SLE and 20 healthy individuals. The microarray analysis identified 14 miRNAs significantly downregulated (2.2–18.0-fold) and 13 miRNAs significantly upregulated (3.4–9.2-fold) in patients with active SLE compared with controls (figure 1A). miR-21, miR-25, miR-148b and miR-155 showed the highest induction (6.3–9.2-fold) whereas miR-196a, miR-150 and hsa-let-7a were the most downregulated miRNAs. The microarray results were validated by quantitative real-time PCR (figure 1B). Several of the identified miRNAs are predicted to regulate genes implicated in biological processes pertinent to the pathogenesis of SLE (figure 1C). Four miRNAs (miR-15a, miR-16, miR-21, miR-25) are predicted to regulate genes involved in apoptosis, a process known to be impaired in SLE.20 Upregulated miR-148a and miR-148b regulate DNA methyltransferase 3b (DNMT3b), which may be involved in epigenetic modifications in SLE.21 miR-150 has been identified as a major regulator of B cell development by targeting the transcription factor c-myb,22 and hsa-let-7a regulates Stat3 expression.23 In accordance, an inverse correlation between miR-150 and c-myb, as well as between hsa-let-7a and Stat3, was documented in our SLE cohort (figure 1D). Gene ontology analysis identified miRNA gene networks significantly enriched in patients with SLE involving anti-apoptotic mechanisms, immune responses, DNA methylation and insulin receptor signalling (see table 1 in online supplement). Together these data suggest that miRNAs may regulate major pathogenic pathways in SLE.
miRNA expression correlates with disease activity in patients with SLE
We next analysed miRNA expression according to SLE disease activity. Nine miRNAs were significantly downregulated and 10 miRNAs were significantly upregulated in active SLE compared with inactive SLE (figure 2A). Expression of miR-21, miR-25, miR-106b and miR-148b showed a significant positive correlation with disease activity and variation in miR expression could explain 84–92% of the SLEDAI variation (r2=0.84–0.92, figure 2B). Notably, miR-196a and miR-379, both downregulated in active SLE, showed a significant inverse association with SLEDAI (r2=0.89 and 0.90, respectively; figure 2C). To characterise the subsets of PBMCs that account for these differences, we repeated the analysis in purified CD4 T and CD19 B cells from a subset of cases and controls. Five miRNAs were significantly downregulated and six miRNAs were significantly upregulated in patients with SLE compared with controls (figure 2D). In B cells, three miRNAs were downregulated and four miRNAs were upregulated in patients with SLE compared with controls (figure 2E). miR-21, miR-25 and miR-106b were upregulated in both T and B lymphocytes from patients with SLE compared with healthy controls; eight miRNAs (let-7a, let-7d, let-7g, miR-148a, miR-148b, miR-324-3p, miR-296, miR-196a) showed altered expression only in T cells from patients with SLE and four miRNAs (miR-15a, miR-16, miR-150, miR-155) only in B cells from patients with SLE (figure 2F).
miR-21 is upregulated and controls aberrant T cell responses in active SLE
Among the upregulated miRNAs, miR-21 exhibited the strongest correlation with disease activity (r2=0.92), being upregulated in both T and B lymphocytes of patients with active SLE (figure 2). miR-21 is upregulated in murine-activated effector and memory T cells and exerts significant cellular proliferative effects.24 25 T cells from patients with SLE show enhanced TCR/CD28-mediated proliferative responses and also provide ‘help’ to B cells for autoantibody production. miR-21 mRNA was 2.3-fold upregulated in freshly isolated CD4 T cells from patients with SLE compared with healthy controls (figure 3A). Anti-CD3/anti-CD28 stimulation induced miR-21 to levels that were significantly higher in patients with SLE than controls (mean 4.0-fold vs 1.6-fold, respectively).
To explore whether aberrant miR-21 expression contributes to T cell hyperactivity in SLE, we stimulated CD4 T cells from patients with SLE with anti-CD3/anti-CD28 monoclonal antibodies and transfected them with antago-miR-21 to silence miR-21 expression (figure 3B). By using CFSE-labelled T cells, we found that silencing of miR-21 reduced T cell proliferation (mean±SEM percentage of undivided T cells on day 6: 48±12% in mock-transfected cells vs 66±9% in antago-miR-21-transfected cells; p<0.05, figure 3C). We also studied T cell IL-10 production and membrane CD40L expression in SLE, both implicated in B cell hyperactivity of SLE. Antago-miR-21 transfection significantly reduced IL-10 production by stimulated CD4 T cells from patients with SLE (612±137 ng/ml in mock-transfected cells vs 347±101 ng/ml in antago-miR-21-transfected cells; p<0.01, figure 3D). miR-21 silencing also suppressed the activation-induced upregulation of membrane CD40L (16.8±1.9% in mock-transfected cells vs 10.1±1.8% in antago-miR-21-transfected cells; p<0.01, figure 3E).
Autoantibody production by B cells in SLE is a T cell-driven response mediated by soluble factors and cell surface interactions. We performed autologous co-cultures of T/B cells from patients with SLE to study the effect of T cell miR-21 silencing in plasma cell generation. In the presence of suboptimal anti-CD3 monoclonal antibody (100 ng/ml), T cells from patients with SLE induced the differentiation of B cells into CD19+CD38hiIgD− plasma cells (figure 3F). Antago-miR-21 transfection significantly reduced the proportion of plasma cells (16.9±2.5% in mock-transfected cells vs 10.5±2.3% in antago-miR-21-transfected cells; figure 3G). Accordingly, the total IgG concentration was reduced in cultures containing antago-miR-21-transfected T cells from patients with SLE (252±72 ng/ml vs 362±63 ng/ml in mock-transfected cells; p<0.05, figure 3H). Together, miR-21 upregulation contributes to the activated T cell phenotype of SLE and the exaggerated T cell-driven B cell differentiation of SLE into Ig-secreting plasma cells.
Normal CD4 T cells transfected with pre-miR-21 acquire an activated phenotype
We next addressed whether miR-21 overexpression could render normal T cells hyperactive. Purified CD4 T cells from healthy controls were stimulated with anti-CD3/anti-CD28 monoclonal antibody and transfected with pre-miR-21 to induce miR-21 expression (figure 3B). Compared with mock-transfected or irrelevant miR-transfected cells, pre-miR-21-transfected T cells had enhanced proliferation (mean±SEM percentage of divided T cells on day 6: 31±3% in scramble-transfected vs 46±1% in pre-miR-21-transfected cells; p<0.05, figure 4A). In accordance with previous studies,26 normal T cells expressed low levels of membrane CD40L; pre-miR-21 transfection caused a small but consistent increase in CD40L expression (figure 4B,C). Moreover, pre-miR-21 transfection caused a significant induction in activation-induced IL-10 secretion (567±0.4 ng/ml vs 255±4 ng/ml in scramble-transfected cells; p<0.001, figure 4D). Finally, miR-21 overexpression in normal CD4 T cells enhanced their capacity to promote autologous CD19 B cell differentiation into CD19 CD38hi IgD− plasma cells (figure 4E), and this effect was associated with increased IgG production (figure 4F). Together, these results suggest that miR-21 overexpression can induce normal T cells to acquire an activated ‘lupus-like’ phenotype.
PDCD4 levels correlate inversely with miR-21 and are decreased in patients with active SLE
MicroRNAs inhibit target gene expression usually at the post-transcriptional level. Previous studies have identified PDCD4, a protein translation inhibitor involved in immune responses,27 as a target gene of miR-21. We found diminished PDCD4 mRNA and protein expression in patients with active SLE compared with controls (figure 5A). To further examine the inverse relationship between miR-21 and PDCD4, we prospectively studied two patients with active SLE until they reached remission. Upon remission, miR-21 mRNA levels were decreased whereas PDCD4 protein levels were significantly increased (figure 5B). Moreover, anti-CD3/anti-CD28 stimulation of T cells resulted in significant suppression of PDCD4 expression (figure 5C). To directly demonstrate that miR-21 negatively regulates PDCD4, normal T cells were transfected with mock or pre-miR-21 to induce miR-21 levels, and PDCD4 protein levels were assessed after 48 h. Pre-miR-21-transfected T cells had reduced PDCD4 compared with mock-transfected cells (figure 5D). In accordance, PDCD4 overexpression significantly reduced IL-10 production (495±23 ng/ml in empty vector vs 312±9 ng/ml in mock-transfected cells, p<0.001) by anti-CD3/anti-CD28-stimulated normal CD4 T cells (figure 5E). Collectively, these results corroborate previous studies showing that miR-21 negatively regulates PDCD4 and imply that the effects of miR-21 in T cells of patients with active SLE may, at least in part, be due to diminished PDCD4 expression.
Discussion
In this study we characterised the miRNA signature of human SLE in PBMCs and in isolated T and B lymphocytes and found a strong correlation of certain miRNAs with SLE disease activity. The identified miRNAs are predicted to regulate genes and processes pertinent to the pathogenesis of SLE, such as DNA methylation, apoptosis and proliferation. miR-21, the miRNA with the highest correlation with disease activity, was found to regulate key cellular functions that contribute to aberrant T cell phenotype in SLE. Finally, the paper reports an inverse correlation between miR-21 and its putative gene target, PDCD4, a selective protein translation inhibitor of genes involved in immune responses.
Our analysis revealed several differentially expressed miRNAs in active versus inactive SLE. Among these genes, miR-21, miR-25, miR-106b (expressed by both T and B lymphocytes) and miR-148b (expressed by T cells) correlated strongly with SLEDAI. The longitudinal analysis of two patients who entered remission showed a significant decrease in miR-21 (figure 5B), suggesting the potential use of miRNAs as disease biomarkers. To this end, miRNAs are attractive as potential biomarkers in SLE because their expression pattern reflects the underlying pathophysiological processes correlating with disease activity, and they can be detected in a variety of tissues.28 Unsupervised clustering of additional cases and controls and longitudinal studies in large patient cohorts would be required to define the use of miRNAs as disease biomarkers.
To account for any potential influence of drugs on miRNA expression, all medication for SLE was held for 24 h prior to blood sampling. This time interval is 3–4 times greater than the half-life of most drugs (including glucocorticoids); however, the possibility of a drug effect cannot be ruled out completely. Nevertheless, the strong correlation between the expression of certain miRNAs and disease activity suggests that aberrant miRNA expression is predominantly disease-driven.
Our findings agree with those of Pan et al29 who found increased miR-21 and miR-148b levels in T cells from patients with active SLE. In contrast, we did not observe the previously reported downregulation of miR-146a in PBMCs from patients with SLE, which correlates with activation of the type I interferon (IFN) pathway.16 This is in agreement with previous work from our laboratory showing no activation of the type I IFN pathway in cDNA microarrays from our SLE cohort,7 and highlights the effect of genetic background in the miRNA profile and SLE phenotype.
Having defined the miRNA signature of human SLE, we next sought to identify novel molecules involved in the pathogenesis of SLE. We selected miR-21 as this was one of the most upregulated miRNAs and displayed the strongest correlation with disease activity. Previous work has reported that miR-21 is overexpressed in several types of malignancies and may contribute to carcinogenesis by promoting cell proliferation.24 30 miR-21 is also upregulated upon stimulation with lipopolysaccharide27 and within inflammatory milieu,31 32 indicating that this miRNA may represent an inflammatory marker.
T cells from patients with SLE had increased miR-21 expression and silencing of miR-21 reversed their aberrant phenotype, including enhanced activation-induced proliferation, IL-10 production and surface CD40L expression. This is consistent with the observation that miR-21 is upregulated in murine effector/memory versus naive T cells.25 In SLE, T cells promote B cell maturation, isotype switch and production of high-affinity pathogenic autoantibodies, and this effect is mediated by both cell contact interactions and soluble factors.33 Inhibition of miR-21 in T cells from patients with SLE significantly reduced their capacity to drive differentiation of B cells from patients with SLE into immunoglobulin-secreting plasma cells. Together, miR-21 may regulate multiple pathways that contribute to T and B cell abnormalities in SLE with important therapeutic implications. Of note, miR-21 was also found to be upregulated in B cells from patients with SLE, and this is currently under investigation.
miR-21 regulates multiple gene targets including PDCD4,34 35 a selective protein translation inhibitor of genes involved in immune responses.27 PDCD4 levels were significantly decreased in active SLE and were restored upon disease remission. Accordingly, activated T cells from patients with SLE upregulated miR-21 and suppressed PDCD4. PDCD4 has been shown to regulate IL-10 production, indicating that the effect of miR-21 on IL-10 production by T cells in SLE may be through inhibition of PDCD4.27 36 Moreover, PDCD4 inhibits AP-1,37 which is an important transcription factor for CD40L,38 also shown to be regulated by miR-21. These findings suggest that the effects of miR-21 on T cells in SLE may, at least in part, be mediated through inhibition of PDCD4. In support of this, we found that PDCD4 overexpression significantly reduced IL-10 production by normal T cells. The beneficial effect of translation-inhibiting drugs such as rapamycin in murine and human SLE T cells39 makes PDCD4 a plausible target for treatment.
In summary, we have provided evidence for altered expression of miRNAs in patients with active SLE, underscoring the importance of this novel class of genes in regulation of immune responses and pathogenesis of autoimmunity. miR-21 correlated strongly with SLE activity and regulated several functions that contribute to the aberrant phenotype of T cells in SLE. By fine-tuning gene regulation, miRNAs can maintain the balance between immune activation and tolerance, and could be exploited as novel therapeutic targets for SLE.
Acknowledgments
The authors acknowledge the help of Eva Choustoulaki, Prodromos Sidiropoulos, Amalia Raptopoulou, Aryro Repa and Eleni Kteniadaki in sample collection, and Christianna Choulaki, Irene Kyrmizi, Eleni Krasoudaki, Melina Kavoussanaki and Marriana Ioannou for technical assistance. The authors also wish to thank Themis Allisafi, Elena Kontaki and Dimitra Vyrla.
References
Supplementary materials
Web Only Data
Files in this Data Supplement:
Footnotes
ES, GB, AI and DTB contributed equally
-
Funding This work is supported by the Hellenic Society of Rheumatology, the Pancretan Health Association, the Hellenic Ministry of Education, Hellenic Republic and the European Union (EPEAEK Fund and Sixth Framework Programme AutoCure program).
-
Competing interests None.
-
Patient consent Obtained.
-
Ethics approval This study was conducted with the approval of the University Hospital of Heraklion, Crete, Greece.
-
Provenance and peer review Not commissioned; externally peer reviewed.