Upregulated miR-146a expression in peripheral blood mononuclear cells from rheumatoid arthritis patients

Sixteen patients (including two samples from a single patient) who fulfilled the American College of Rheumatology classification criteria for RA were included in the study. Their demographic, clinical, and laboratory characteristics are summarized in Table 1. Four disease control individuals, including one with systemic lupus erythematosus, two with Sjögren's syndrome, and one with systemic sclerosis, were included. Nine healthy donors with no history of autoimmune disease were included as control individuals. This study was approved by the University of Florida Institutional Review Board, and written permission was obtained from all who participated in the study.

Blood samples were collected in EDTA-treated tubes and PBMCs were isolated by standard Ficoll density-gradient centrifugation. PBMCs were washed once in sterile phosphate-buffered saline (PBS) before culture or RNA isolation. Total RNA was isolated from freshly obtained PBMCs using the mirVana miRNA Isolation kit (Ambion, Austin, TX, USA), in accordance with the manufacturer's protocol. RNA concentrations were determined and 10 ng of each RNA sample were used for quantitative real-time RT-PCR (qRT-PCR). miRNA qRT-PCR was performed using the TaqMan MicroRNA Reverse Transcription Kit, TaqMan Universal PCR Master Mix, and TaqMan MicroRNA Assay primers for human miR-146a, miR-155, miR-132, miR-16, and miRNA let-7a (Applied Biosystems, Foster City, CA, USA). mRNA qRT-PCR was performed using the TaqMan High-Capacity cDNA Reverse Transcription Kit, TaqMan Fast PCR Master Mix, and TaqMan mRNA assay primers (Applied Biosystems). All reactions were analyzed using StepOne Real-Time PCR System (Applied Biosystems). The levels of miRNA were normalized to U44 controls, whereas mRNA levels were normalized to 18S RNA. The cycle threshold (Ct) values, corresponding to the PCR cycle number at which fluorescence emission reaches a threshold above baseline emission, were determined and the relative miRNA or mRNA expression was calculated using the 2^-ΔΔCt method [23].

THP-1 human monocytes obtained from American Type Culture Collection (Manassas, VA, USA) were cultured in RPMI 1640 medium with 2 mmol/l L-glutamine, 4.5 g/l glucose, 10 mmol/l HEPES, 1.0 mmol/l sodium pyruvate, 0.05 mmol/l 2-mercaptoethanol, and 10% fetal bovine serum. THP-1 cells were seeded at 5 × 10⁵ cells per well in a six-well plate and treated with 10 ng/ml TNF-α, IFN-γ, IL-12p70, IL-4, IL-10 (BD Biosciences, San Jose, CA, USA), IFN-α, IFN-β (PBL Interferon Source, New Brunswick, NJ, USA), or macrophage colony-stimulating factor (M-CSF; US Biological, Swampscott, MA, USA). Cells were also treated with 25 ng/ml monocyte chemoattractant protein (MCP)-1 (Sigma) in serum-free media. After the designated treatment time had elapsed, cells were harvested and washed once in PBS before analysis.

THP-1 cells were cytospun onto glass slides at 1,000 rpm for 5 minutes. PBMCs were cultured on glass slides at 37°C for 1 hour. Cells were fixed in 3% paraformaldehyde for 10 minutes and permeabilized in 0.5% Triton X-100 for 5 minutes. GWB were detected in THP-1 cells with a human prototype anti-GWB serum [24] used at 1:6,000 dilution, and in PBMCs with rabbit anti-Rck/p54 antibodies used at 1:500 dilution. TNF receptor-associated factor (TRAF)6 and IL-1 receptor-associated kinase (IRAK)-1 were detected using rabbit anti-TRAF6 (1:50; Santa Cruz Biotechnology, Santa Cruz, CA, USA) and rabbit anti-IRAK-1 (1:50; Santa Cruz Biotechnology). Secondary antibodies used were Alexa Fluor 488 goat anti-human IgG or goat anti-rabbit IgG (1:400) from Molecular Probes (Carlsbad, CA, USA). Slides were mounted using Vectashield Mounting Medium with 4',6-diamidino-2-phenylindole (DAPI; VECTOR Laboratories, Burlingame, CA, USA). Fluorescence images were taken with Zeiss Axiovert 200 M microscope and a Zeiss AxioCam MRm camera using the 20× or 40 × 0.75 NA objectives. Color images were assessed using Adobe Photoshop version 7 (Adobe Systems Inc., San Jose, CA, USA). GWB were counted using Cell-Profiler image analysis software [25].

siRNAs targeting TRAF6 and IRAK-1 were transfected into THP-1 cells using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA), in accordance with the manufacturer's instructions. To monitor the transfection efficiency, Cy3-labeled siRNA targeting lamin A/C was transfected into cells in parallel in all transfections, and at least 80% transfection efficiency was achieved. The siRNAs used in this study were all purchased from Applied Biosystems. The sense and antisense strand sequences were as follows: IRAK-1: 5'-GGUUUCGUCACCCAAACAUtt-3' and 5'-AUGUUUGGGUGACGAAACCtg-3'; TRAF6: 5'-GGUUGUUUGCACAAGAUGGtt-3' and 5'-CCAUCUUGUGCAAACAACCtt-3'.

Since our previous work demonstrated that GWB can be used as biomarkers for miRNA activity (Pauley KM, unpublished data), we began to examine a variety of cytokines and chemokines for their ability to stimulate miRNA activity in human monocytic THP-1 cells. THP-1 cells were treated with 10 ng/ml TNF-α, IFN-α, IFN-β, IFN-γ, IL-12p70, M-CSF, IL-4, IL-10, or 25 ng/ml MCP-1 for 4 hours. Indirect immunofluorescence (IIF) was performed using a human anti-GWB serum to detect GWB in the cells. As shown in Figure 1, the proinflammatory cytokines/chemokines TNF-α, IFN-α, IFN-β, IFN-γ, and MCP-1 resulted in a significant increase in the number of GWB per cell compared with untreated cells cultured in parallel (P < 0.0001, as determined using one-way analysis of variance). However, IL-12p70, M-CSF, IL-4, and IL-10 had no significant effect on the number of GWB. TNF-α elicited the strongest response in THP-1 cells, with fourfold increase in the average number of GWB per cell (Figure 1a,c). These experiments were repeated at least three times, with reproducible results each time.

Next, we decided to examine the effect of TNF-α stimulation on human PBMC GWB. GWB staining using human PBMCs from a healthy donor, after 4 hours stimulation with TNF-α (1 ng/ml), is shown. Similar to THP-1 cells, the number of GWB per cell increased 3.5-fold after TNF-α stimulation of PBMCs (Figure 1b,c; P < 0.0001, as determined by Mann Whitney test). These data indicated that THP-1 cells may be suitable substitutes for human PBMCs in some of the subsequent experiments.

In Figure 1 we showed that TNF-α is a potent inducer of GWB and therefore miRNA activity. Our preliminary studies and work from other investigators have confirmed that TNF-α stimulation induces the expression of certain miRNAs, including miR-146a and miR-155 [20, 21]. Based on these data and the important role played by TNF-α in RA pathogenesis and therapies, we began to investigate the expression levels of miRNA in RA patients as compared with those in healthy and disease control individuals. PBMCs were obtained from patients (n = 17 RA patients and n = 4 disease control individuals) and healthy donors (n = 9) and isolated by Ficoll density-gradient centrifugation. Initially, RA PBMCs were monitored by IIF for GWB; however, we did not observe an increased number of GWB in RA compared with healthy control individuals (not shown). This discrepancy could be due to limited sensitivity in the quantitation of GWB.

As shown in Figure 2a, the average relative expression levels of miR-146a, miR-155, miR-132, and miR-16 were 2.6-, 1.8-, 2.0-, and 1.9-fold, respectively, higher for RA patients than for healthy control individuals (P < 0.01 for miR-146a and P < 0.05 for miR-155, miR-132, and miR-16, as determined by one-way analysis of variance). The expression of miRNA let-7a was not significantly different between RA patients and healthy control individuals (Figure 2a). Disease control miRNA expression resembled that in healthy control individuals.

To examine the relationship between RA disease activity and miRNA expression levels, patients were classified into inactive/remission and active patients, based on C-reactive protein (CRP) and erythrocyte sedimentation rate (ESR) values. Three patients with normal CRP and ESR (Table 1) were classified as inactive, whereas eight patients with elevated CRP and/or ESR (Table 1) were classified as active (Figure 2b). Those patients with incomplete or no available data were omitted (Table 1). miRNA expression levels were compared between the groups. Interestingly, high miR-146a and miR-16 expression levels appeared to correlate with active disease, whereas low expression level correlated with inactive disease (Figure 2b; P < 0.05, as determined by t-test). These data indicates that miR-146a and miR-16 expression levels may be a useful marker of RA disease activity. Further studies involving a larger patient cohort are needed to determine fully whether monitoring miRNA expression as a marker for disease activity can improve upon CRP or ESR measurements.

Figure 2c shows the miRNA expression levels in two samples from a single RA patient collected over a 2-month interval, during which time this patient's CRP and ESR values increased despite methotrexate treatment. The miRNA levels of this patient were largely unchanged over the 2-month interval, remaining elevated compared with those in healthy control individuals. This indicates that the elevated miRNA expression in this patient may reflect the patient's lack of improvement, as indicated by the increased CRP and ESR values. In this patient, miR-146a, miR-155, and miR-132 expression levels were stable over this time period, whereas the expression levels of miR-16 and miRNA let-7a increased by approximately 3.5-fold and 2.4-fold, respectively. A larger patient population must be examined in order to determine whether miRNA expression levels may be indicative of treatment efficacy.

To further analyze the increased miRNA expression exhibited by these RA patients, we compared miR-146a, miR-155, and miR-132 expression levels with patient clinical and demographic data (Table 1) and found no significant trends or correlations between high expression levels and age, race, or medications. Patients receiving no medications at the time of miRNA analysis exhibited the same trend toward elevated miRNA expression, indicating that treatment with medications is not responsible for the increased miRNA expression in RA patients.

Recently, two reports [26, 27] showed increased miR-146 and miR-155 expression levels in RA synovial tissue and fibroblasts. Stanczyk and coworkers [27] reported a fourfold increase in miR-146a expression and a twofold increase in miR-155 expression in RA synovial fibroblasts compared with osteoarthritis synovial fibroblasts. They also demonstrated that miR-155 expression can repress the induction of matrix metalloproteinases 3 and 1, indicating that miR-155 may be involved in modulating the destructive properties of RA synovial fibroblasts. However, in that report, miR-155 expression from RA PBMCs was not significantly different from that in control PBMCs. This discrepancy could be due to differences in experimental techniques or patient populations. Nakasa and colleagues [26] also reported an approximately fourfold increase in miR-146a expression in RA synovial tissue. Our data demonstrate that RA patient PBMCs exhibit elevated miRNA expression in a similar manner to RA synovial tissue, with a 2.6-fold increase in miR-146a expression and a 1.8-fold increase in miR-155 expression. Because of the invasiveness involved in collecting samples, monitoring miRNA expression in RA synovial tissue is, in most cases, limited to extremely severe disease in patients undergoing joint surgery or replacement. Because blood collection is not invasive, this allows for easy sample collection over time, which is a distinct advantage when monitoring disease activity and treatment efficacy.

The expression levels of miR-146a, miR-155, miR-132, and miR-16 were 2.8-, 1.6-, 4.2-, and 3.4-fold higher, respectively, in monocytes than in lymphocytes (Figure 3a). Let-7a expression was similar between monocytes and lymphocytes (not shown). Figure 3b shows the average expression (miR-146a, miR-155, miR-132, and miR-16 combined) for the monocyte and lymphocyte populations of two RA patients (P < 0.02, as determined by Mann-Whitney test). These findings suggest that monocytes/macrophages contribute to the increased miRNA expression observed in RA patients more than lymphocytes, but further studies must be performed to confirm this observation.

Because most of the RA patients exhibited increased expression of miR-146a compared with healthy and disease control individuals, we decided to examine the expression of two confirmed targets of miR-146a, namely TRAF6 and IRAK-1 [21]. TRAF6 and IRAK-1 mRNA expression levels were analyzed by qRT-PCR (Figure 4a,b). RA patients exhibited increased miR-146a production compared with control individuals, and we therefore expected to observe decreased TRAF6 and/or IRAK-1 expression in RA patients as compared with control individuals. However, TRAF6 and IRAK-1 mRNA expression levels were very similar between RA patients and control individuals, and overall the mRNA levels of TRAF6 and IRAK-1 did not exhibit the same degree of variability between patients that we observed with miRNA expression. This may indicate that TRAF6 and IRAK-1 transcripts are under other levels of control. To confirm this discrepancy, we analyzed TRAF6 and IRAK-1 protein levels by IIF in one healthy control individual and one RA patient whose miR-146a level was increased (Figure 2a). PBMCs were processed for IIF as previously described and were stained for TRAF6 and IRAK-1. Image J software was used to quantify the relative level of fluorescence for at least 20 cells. As shown in Figure 4c, there was no significant difference in TRAF6 or IRAK-1 protein levels between the RA patient and healthy control individual, which is consistent with the mRNA analysis.

It is interesting to speculate that this lack of regulation of TRAF6/IRAK-1 by miR-146a could play a role in RA pathogenesis, especially because it has been reported that inhibition of IRAK-1 using antisense oligonucleotides results in decreased LPS-induced cytokine production [28], and our preliminary data have shown that transfection of miR-146a into THP-1 monocytes results in knockdown of TRAF6 and IRAK-1 expression and inflammatory cytokine production (Pauley KM and coworkers, unpublished data). To investigate this possibility further, we transfected siRNA targeting TRAF6 and/or IRAK-1 into THP-1 cells. The knockdown efficiency was determined by analyzing TRAF6 and IRAK-1 mRNA levels by qRT-PCR, and at least 80% and 60% knockdown was achieved for TRAF6 and IRAK-1, respectively (Figure 5a). Two days after transfection, knockdown and control cells were treated with 1 μg/ml LPS for 24 hours. Culture supernatants were collected and cytokines/chemokines were quantitatively detected using a human cytokine multiplex assay. TNF-α production was drastically reduced in the TRAF6 and/or IRAK-1 deficient cells compared with mock transfected cells (Figure 5b), whereas MCP-1 production was not affected by TRAF6 or IRAK-1 knockdown (Figure 5c).

These data demonstrate that TRAF6 and IRAK-1 are required for the production of TNF-α in THP-1 cells. Taken together, it is reasonable to hypothesize that the absence of TRAF6/IRAK-1 regulation by miR-146a in RA patients could contribute to the prolonged production of TNF-α that many of these patients exhibit. Furthermore, it would be interesting to investigate the expression patterns of miR-146a, TRAF6, and IRAK-1 in RA patients who are responsive to anti-TNF-α therapy versus those who are not responsive. Clearly, further studies are needed to elucidate the role played by miR-146a regulation in RA pathogenesis and the mechanism by which TRAF6/IRAK-1 escape miR-146a regulation.