Down-regulation of microRNA-34a* in rheumatoid arthritis synovial fibroblasts promotes apoptosis resistance

Authors


Abstract

Objective

To investigate the expression and effect of the microRNA-34 (miR-34) family on apoptosis in rheumatoid arthritis synovial fibroblasts (RASFs).

Methods

Expression of the miR-34 family in synovial fibroblasts with or without stimulation with Toll-like receptor (TLR) ligands, tumor necrosis factor α (TNFα), interleukin-1β (IL-1β), hypoxia, or 5-azacytidine was analyzed by real-time polymerase chain reaction (PCR). Promoter methylation was studied by combined bisulfite restriction analysis. The effects of overexpression and silencing of miR-34a and miR-34a* on apoptosis were analyzed by annexin V/propidium iodide staining. Production of X-linked inhibitor of apoptosis protein (XIAP) was assessed by real-time PCR and immunohistochemistry analysis. Reporter gene assay was used to study the signaling pathways of miR-34a*.

Results

Basal expression levels of miR-34a* were found to be reduced in synovial fibroblasts from RA patients compared to osteoarthritis patients, whereas levels of miR-34a, miR-34b/b*, and miR-34c/c* did not differ. Neither TNFα, IL-1β, TLR ligands, nor hypoxia altered miR-34a* expression. However, we demonstrated that the promoter of miR-34a/34a* was methylated and showed that transcription of the miR-34a duplex was induced upon treatment with demethylating agents. Enforced expression of miR-34a* led to an increased rate of FasL- and TRAIL-mediated apoptosis in RASFs. Moreover, levels of miR-34a* were highly correlated with expression of XIAP, which was found to be up-regulated in RA synovial cells. Finally, we identified XIAP as a direct target of miR-34a*.

Conclusion

Our data provide evidence of a methylation-specific down-regulation of proapoptotic miR-34a* in RASFs. Decreased expression of miR- 34a* results in up-regulation of its direct target XIAP, thereby contributing to resistance of RASFs to apoptosis.

The pathogenesis of rheumatoid arthritis (RA) is characterized by marked proliferation of the synovial tissue, which invades and destroys periarticular bone and cartilage. The major effector cells of cartilage degradation are synovial fibroblasts (1). RA synovial fibroblasts (RASFs) produce high levels of destructive enzymes and proinflammatory cytokines upon activation (2). A characteristic feature of RASFs is their decreased susceptibility to spontaneous as well as FasL- and TRAIL-mediated apoptosis (3, 4). Whereas activation of the Fas pathway is associated with increased JNK signaling, TRAIL has been reported to act via inhibition of the phosphatidylinositol 3-kinase/Akt molecular pathway (5). Several factors have been demonstrated to inhibit receptor-mediated apoptosis in RASFs, such as small ubiquitin-like modifier 1 (6), FLIP (7), and Mcl-1 (8). However, the mechanisms leading to the characteristic apoptosis resistance of RASFs are still incompletely understood.

MicroRNAs (miRNAs) have emerged as fine-tuning regulators of diverse biologic processes (9, 10). During their biogenesis, miRNA genes are transcribed into primary miRNAs that are processed by Drosha and Dicer to generate miRNA duplexes consisting of a mature strand and a passenger miRNA strand (also called miRNA star [miRNA*] strand) (11). Some miRNAs have been reported to be important regulators of cellular lifespan. Of these, miRNA-34a (miR-34a) has been identified as a modulator of apoptosis. However, existing data are controversial. On the one hand, miR-34a was shown to be an inducer of programmed cell death in colon cancer cell lines and mouse embryonic fibroblasts (12–14), whereas on the other hand, miR-34a protected cells, such as human B lymphocytes and a mouse tubular cell line, against apoptosis (15, 16).

Since alterations of apoptosis are a characteristic finding in rheumatoid synovium, we have analyzed the expression of the miR-34 family in RASFs. MiR- 34a, miR-34b/b*, and miR-34c/c* were not found to be differentially expressed between RASFs and synovial fibroblasts from patients with noninflammatory osteoarthritis (OA). However, miR-34a*, the passenger strand of miR-34a, was significantly down-regulated in synovial fibroblasts from patients with RA compared to those with OA. We analyzed the effects of miR-34a* on apoptotic cell death and identified X-linked inhibitor of apoptosis protein (XIAP) as a direct target of miR-34a*. Finally, we showed that low levels of miR-34a* correlate with high XIAP expression in synovial fibroblasts.

MATERIALS AND METHODS

Patient samples and tissue preparation.

Synovial tissue specimens were obtained from RA and OA patients at the Department of Orthopedic Surgery, Schulthess Clinic. All patients provided informed consent. All RA patients fulfilled the American College of Rheumatology criteria for classification of the disease (17). For cell culture, tissue samples were digested and synovial fibroblasts were grown as previously described (10). Cultures of synovial fibroblasts were subjected to experimental procedures at passages 4–9.

Stimulation assays.

Synovial fibroblasts were plated in 12-well plates (5 × 104 cells/well) in 1 ml supplemented Dulbecco's modified Eagle's medium and stimulated for 2, 8, or 24 hours with 10 ng/ml tumor necrosis factor α (TNFα), 1 ng/ml interleukin-1β (IL-1β) (both from R&D Systems), 300 ng/ml bacterial lipopeptide palmitoyl-3-cysteine-serine-lysine-4, 10 μg/ml poly(I-C) (both from InvivoGen), or 10 ng/ml lipopolysaccharide (LPS; from Escherichia coli J5) (List Biological Laboratories). For methylation studies, 5-azacytidine (5-azaC; Sigma-Aldrich) was added to the cell culture medium at a concentration of 0.1 μM or 1 μM per day for 7 days. For induction of hypoxia, fibroblasts were exposed to a humidified atmosphere containing 1% O2 (hypoxia) or 20% O2 (normoxia) (volume/volume) for 2, 6, or 20 hours.

RNA isolation, reverse transcription, and real-time polymerase chain reaction (PCR).

For miRNA analysis, RNA was isolated from fibroblasts with the mirVana Isolation Kit (Applied Biosystems). Total RNA was reverse transcribed and analyzed by real-time PCR using miR-specific TaqMan primers (Applied Biosystems). Expression of let-7a was tested as an endogenous control for relative quantification. For analysis of messenger RNA (mRNA) expression, isolated RNA was digested with DNase and reverse transcribed. Single-reporter real-time PCR was performed using the ABI Prism 7700 Sequence Detection system. XIAP mRNA was quantified by real-time PCR with SYBR Green. The primer sequences used for XIAP were as follows: forward 5′-AAT-TGG-GAA-CCT-TGT-GAT-CG-3′, reverse 5′-AGG-AAA-GTG-TCG-CCT-GTG-TT-3′. Eukaryotic 18S ribosomal RNA was measured using a predeveloped primer/probe system (Applied Biosystems), as an endogenous control. Results were analyzed using the ΔCt method for relative quantification, as described previously (18).

Transfection experiments.

Synovial fibroblasts were transfected with 100 nM of synthetic precursor molecules (pre-miR) or inhibitors (anti-miR) of miR-34a and miR-34a*, or negative controls for pre-miR and anti-miR (all from Applied Biosystems), using Lipofectamine 2000 reagent according to the protocol of the manufacturer (Gibco Invitrogen). Transfected synovial fibroblasts were incubated at 37°C for 72 hours for analysis of basal expression or for 48 hours when subjected to further experiments. Successful transfection was confirmed by real-time PCR using miR-specific TaqMan primers as described above.

Fluorescence-activated cell sorting.

Synovial fibroblasts were transfected with pre- and anti-miRs as described above. After 48 hours, medium was changed and cells were left untreated or stimulated for 24 hours with 200 ng/ml FasL (Alexis) or 20 ng/ml TRAIL (R&D Systems). Cells were detached with accutase (PAA Laboratories) and stained with fluorescein isothiocyanate–conjugated annexin V and propidium iodide using an Annexin-V-Fluos Staining Kit (Roche). Cells were subsequently analyzed on a FACSCalibur flow cytometer. Data were processed using CellQuest software (BD Biosciences).

Immunohistochemistry analysis.

Tissues were fixed in paraformaldehyde and embedded in paraffin. Synovial tissue sections were deparaffinized and pretreated with 10 mM citrate buffer. After blocking of endogenous peroxidase and nonspecific binding, slides were incubated with mouse-anti-human XIAP antibodies (1:200) (clone 117318; R&D Systems) or mouse IgG1 isotype control (Jackson ImmunoResearch) overnight at 4°C. Sections were incubated with biotinylated goat-anti-mouse antibodies (Jackson ImmunoResearch) followed by incubation with horseradish peroxidase–conjugated streptavidin complex (ABC kit; Vector). XIAP-positive cells were visualized using aminoethylcarbazole. All sections were counterstained with hematoxylin.

Combined bisulfite restriction analysis of the miR-34a/34a* promoter.

Genomic DNA was prepared from RASFs and OASFs using a QIAamp DNA blood Mini kit and bisulfite modified using an EpiTect bisulfite kit (both from Qiagen). PCR amplification was performed using HotStart and AmpliTaq Gold (both from Applied Biosystems). Primers were designed for the CpG area upstream of the miR-34a/34a* promoter, as follows: hemi-forward 5′-GGG-GAT-TGT-AGT-GTT-AGT-TTT-TTT-3′, forward 5′-TGT-TGG-TTT-GTT-TTT-TGG-ATT-TTA-3′, reverse 5′-AAA-AAA-TCA-ACA-CTT-CCC-TAA-AAA-AA-3′. The PCR product was Taq I digested for 5 hours and analyzed on an agarose gel.

Cloning of XIAP 3′-untranslated region (3′-UTR) expression plasmids.

The 6,800-bp 3′-UTR of XIAP was cloned into pGL3control vector (Promega) in 3 overlapping parts. For PCR, human genomic DNA (Promega) and the following primers were used: XIAP 3′-UTR part 1 (1–2,299 bp) forward 5′-GTA-ATT-CTA-GAT-CTA-ACT-CTA- TAG-TAG-GCA-TGT-TA-3′, reverse 5′-TAA-ATT-TAA-AGT-CTA-GAC-TAT-ACA-GAC-CAA-ATT-C-3′; XIAP 3′-UTR part 2 (2,249–4,652 bp) forward 5′-AGC-TCG- CTA-GCC-TGC-CAC-TTA-GTT-TGG-TTA-TAT-AG-3′, reverse 5′-CCG-ACG-CTA-GCC-ACA-TTG-TGT-TAA-CTG-TAT-GAG-TC-3′; XIAP 3′-UTR part 3 (4,573–6,790 bp) forward 5′-TCG-AGT-CTA-GAG-AGC-TTT-CTA-AGA-GAA-GCA-ATT-GG-3′, reverse 5′-CCG-ACT-CTA-GAA-ATT-TTA-AAG-AAT-AGT-ATT-TTA-3′. The PCR products were gel purified, digested with Xba I (parts 1 and 3) or Nhe I (part 2), and inserted into the pGL3control plasmid via the Xba I restriction site downstream of the luciferase gene. The constructs were analyzed for correctness by sequencing.

Reporter gene assay.

HEK 293 cells were cotransfected with the 3′-UTR constructs, the pRL_SV40 vector (Promega) as internal control, and synthetic pre–miR-34a* or pre-control oligonucleotide, using Lipofectamine 2000. Firefly luciferase activity was measured using a Dual-Luciferase Reporter Assay system (Promega) and normalized to the activity of Renilla luciferase.

Statistical analysis.

Mean ± SEM values were calculated. Paired and unpaired t-tests were used as appropriate for statistical evaluation of the data (with GraphPad Prism 5.0). P values less than 0.05 were considered significant.

RESULTS

Reduced constitutive expression levels of miR-34a* in RASFs.

Since the miR-34 family has been found in several studies to influence apoptotic cell death, we performed expression analysis of miR-34a, miR-34b, and miR-34c and of the corresponding passenger strands in synovial fibroblasts. Constitutive expression levels differed between individual miRs, which may be attributable to their varying chromosomal locations. MiR-34a is present on chromosome 1, while miR-34b and miR-34c are located on chromosome 11. The expression of miR-34a, as well as that of miR-34b and miR-34c, was not found to be significantly altered in RASFs compared to OASFs. No differences in expression of the passenger strands of miR-34b and miR-34c were found between synovial fibroblasts from RA patients and those from OA patients. However, constitutive expression levels of the passenger strand miR-34a* differed significantly between RASFs and OASFs (Figure 1). Levels of miR-34a* were down-regulated by 44% in RASFs (mean ± SEM ΔCt 5.96 ± 0.51) as compared to OASFs (5.13 ± 0.71) (P < 0.001).

Figure 1.

Decreased constitutive expression of microRNA-34a* (miR-34a*) in rheumatoid arthritis synovial fibroblasts (RASFs). Basal expression levels of miR-34a* (n = 8–10), miR-34a (n = 18–20), miR-34b* (n = 5), miR-34b (n = 6–7), miR-34c* (n = 5), and miR-34c (n = 8–10) in RASFs or osteoarthritis (OA) synovial fibroblasts are shown. Data are presented as box plots, where the boxes represent the 25th to 75th percentiles, the lines within the boxes represent the median, and the lines outside the boxes represent the minimum and maximum values. ∗ = P < 0.001 by unpaired t-test.

Methylation-specific inactivation of miR-34a*.

To investigate the regulation of miR-34a* expression, we stimulated RASFs with either TNFα, IL-1β, or the Toll-like receptor (TLR) ligands bacterial lipopeptide (TLR-2), poly(I-C) (TLR-3), and LPS (TLR-4). Real-time PCR analysis showed no changes in miR-34a* expression in RASFs (Figure 2A) or OASFs (data not shown) after stimulation for 24 hours. Time course analysis of TNFα-treated cells revealed that this cytokine also had no influence on miR-34a* expression after 2-hour or 8-hour stimulation (data not shown). During the establishment of RA, hypoxia occurs in the synovium (19). To assess whether changes in the oxygen supply could alter the expression of miR-34a*, we incubated RASFs under hypoxic conditions for 2, 6, or 20 hours. Subsequent real-time PCR showed reduced miR-34a* levels after 20 hours of hypoxia; however, the reduction did not reach statistical significance (Figure 2B).

Figure 2.

Epigenetic inactivation of miR-34a and miR-34a* in RASFs. A, Expression of miR-34a* after stimulation for 24 hours with tumor necrosis factor α (TNFα; n = 7), bacterial lipopeptide (BLP; n = 4), poly(I-C) (pIC; n = 5), lipopolysaccharide (LPS; n = 4), or interleukin-1β (IL-1; n = 4) was measured by real-time polymerase chain reaction (PCR). Expression levels relative to those in untreated RASFs (set at 1 [dotted line]) were calculated. Data are presented as box plots, where the boxes represent the 25th to 75th percentiles, the lines within the boxes represent the median, and the lines outside the boxes represent the minimum and maximum values. B, RASFs (n = 6) were cultured for 20 hours under conditions of normoxia (20% O2) or hypoxia (1% O2), and expression of miR-34a* was measured by real-time PCR. Each pair of symbols represents RASFs from an individual patient. C, RASFs (n = 6) were stimulated with 0.1 μM or 1 μM 5-azacytidine (5-aza), and expression of miR-34a* and miR-34a was measured by PCR before treatment and after 1 week of treatment. Each set of symbols represents RASFs from an individual patient. ∗ = P < 0.05 by paired t-test. D, Bisulfite-modified DNA isolated from RASFs (n = 3) was PCR amplified using primers recognizing the CpG island upstream of the miR-34a/34a* promoter. The PCR products were left untreated (U) or were digested (D) using Taq I; Taq I recognizes the sequence TCGA and therefore cleaves only when the CpG sequence has been preserved by a methylated cytosine during bisulfite conversion. DNA was visualized by agarose gel electrophoresis. The restricted bands demonstrate that the promoter of miR-34a/34a* is methylated in these cells. See Figure 1 for other definitions.

Since the mature miR-34a strand was previously shown to be regulated by DNA methylation (14, 20) and RASFs display alterations in their DNA methylation pattern (21), we investigated the influence of DNA-demethylating agents on the expression of miR-34a and its passenger strand. RASFs were treated for 1 week with 5-azaC at 0.1 μM or 1 μM. MiR-34a and miR-34a* were induced (34% and 70%, respectively) upon demethylation with 1 μM 5-azaC, but statistical significance was reached only for miR-34a* (P < 0.05) (Figure 2C). In fact, combined bisulfite restriction analysis showed that the CpG island upstream of the promoter of miR-34a/34a* is methylated in RASFs (Figure 2D), indicating that expression of miR-34a and its passenger strand is under epigenetic regulation.

Proapoptotic effect of miR-34a* in RASFs.

To investigate the role of miR-34a and miR-34a* in apoptosis of RASFs, we overexpressed each miR using synthetic pre-miRs. Successful overexpression was confirmed by real-time PCR. Basal cell death was not significantly changed by transfection with pre-miRs (data not shown). Interestingly, we observed opposite effects of miR-34a and of its passenger strand on FasL-mediated apoptosis (Figure 3A). Overexpression of miR-34a reduced the number of FasL-stimulated annexin V–positive cells compared to control transfected RASFs by 20% (P < 0.05), whereas cells transfected with pre–miR-34a* were more susceptible to FasL-induced apoptosis, as demonstrated by an increase in the proportion of annexin V–positive cells from 25% to 40% (P < 0.05). Consistent with this, overexpression of miR-34a* further promoted TRAIL-induced apoptosis, with the proportion of annexin V–positive cells increasing from 22% to 30% (P < 0.05) (Figure 3B). Transfection of miR-34a did not affect levels of TRAIL-mediated apoptosis. Thus, the results suggest that the constitutively low expression levels of miR-34a* in RASFs result in decreased apoptosis.

Figure 3.

Induction of apoptotic cell death in RASFs by miR-34a*. A and B, RASFs (n = 5–6) were transfected with synthetic precursor miRs to overexpress miR-34a (pre-34a), miR-34a* (pre-34a*), or nonspecific control miR (pre-ctr). Forty-eight hours after transfection, cells were stimulated with 200 ng/ml FasL (A) or 20 ng/ml TRAIL (B) for 24 hours and annexin V binding was analyzed by flow cytometry. C, RASFs (n = 4) were treated with inhibitors of miR-34a* (anti-34a*) or nonspecific control miR (anti-ctr). Forty-eight hours after transfection, cells were stimulated with 20 ng/ml TRAIL for 24 hours and annexin V binding was analyzed. D, OASFs (n = 8) were transfected with pre-34a* to overexpress miR-34a* or with pre-ctr as a transfection control. Forty-eight hours after transfection, cells were stimulated with 20 ng/ml TRAIL for 24 hours and annexin V binding was analyzed. Values are the mean ± SEM. ∗ = P < 0.05 by paired t-test. See Figure 1 for other definitions.

In accordance with the data obtained with pre–miR-34a*, silencing of miR-34a* in RASFs resulted in reduced levels of apoptosis, verified by a decrease in the proportion of annexin V–positive cells from 19% to 17% (P < 0.05) (Figure 3C). When OASFs overexpressed miR-34a*, apoptosis was further increased from 13% to 20%, similar to the results obtained with RASFs (P < 0.05) (Figure 3D).

XIAP mediates the proapoptotic effects of miR-34a*.

Based on target prediction algorithms (http://diana.cslab.ece.ntua.gr/), we identified XIAP as a potential target gene for miR-34a*. To investigate for a possible regulatory role of miR-34a* on the expression of XIAP, we transfected RASFs with pre–miR-34a* or a nonspecific pre-miR control and performed silencing experiments using anti-miR-34a* or nonspecific control anti-miR. When RASFs were transfected with pre–miR-34a*, levels of XIAP were significantly reduced in comparison to those in RASFs transfected with control pre-miR–transfected cells (by 36%; P < 0.05), whereas silencing of miR-34a* (anti-34a*) increased expression of XIAP in comparison to that in control anti-miR–transfected RASFs (by 90%; P < 0.05) (Figure 4A). In addition, we studied the effect of the mature miR-34a strand on XIAP. Neither overexpression nor silencing of miR-34a altered production of XIAP (Figure 4A). Our findings suggest that miR-34a* exerts its proapoptotic effect via XIAP.

Figure 4.

Modulation of X-linked inhibitor of apoptosis protein (XIAP) expression by microRNA-34* (miR-34a*). A, Rheumatoid arthritis synovial fibroblasts (RASFs) were transfected with precursor miRs (pre; n = 9–13) or miR inhibitors (anti; n = 5) specific for miR-34a and miR-34a*, respectively, or pre-miR and anti-miR negative controls. Expression of XIAP mRNA was analyzed by real-time polymerase chain reaction using XIAP-specific primers. Results were normalized to 18S ribosomal RNA and expression levels relative to those in nontransfected RASFs (set at 1 [dotted line]) were calculated. ∗ = P < 0.05 versus nontransfected RASFs, by paired t-test. B, Basal expression of XIAP in RASFs and in osteoarthritis synovial fibroblasts (OASFs) was measured (n = 4 per group). ∗ = P < 0.05 by unpaired t-test. Data in A and B are presented as box plots, where the boxes represent the 25th to 75th percentiles, the lines within the boxes represent the median, and the lines outside the boxes represent the minimum and maximum values. C, The correlation between basal levels of XIAP and constitutive expression of miR-34a* in synovial fibroblasts (n = 11) was determined by Spearman's nonparametric correlation test. Dashed lines represent the 95% confidence interval.

To analyze expression of XIAP in the synovium, we performed immunohistochemistry analysis of RA synovial tissue sections. In accordance with previously reported results (22), we found that XIAP was abundantly expressed throughout the entire RA synovial tissue, whereas in synovial tissue specimens from patients with OA, XIAP was found mainly at perivascular areas (Figure 5). The results of the immunohistochemistry analysis were confirmed by quantitative real-time PCR analysis of synovial fibroblasts (Figure 4B). Constitutive levels of XIAP were 2.1-fold higher in RASFs compared to OASFs, with mean ± SEM ΔCt values of 12.54 ± 1.67 and 11.45 ± 0.57, respectively (P < 0.05). Furthermore, basal expression of XIAP mRNA was highly correlated with constitutive miR-34a* production (Spearman's r = 0.8, P < 0.0047) (Figure 4C). These data suggest that constitutive down-regulation of miR-34a* in RASFs is responsible for the up-regulation of XIAP.

Figure 5.

XIAP expression in synovial tissue from patients with RA and OA. Synovial tissue specimens were stained for XIAP (red) and counterstained with hematoxylin (blue). A representative RA specimen and a representative OA specimen with positive XIAP staining, as well as a representative RA specimen stained with mouse IgG1 as a negative control, are shown. Boxed areas in the upper panels are shown at higher magnification in the lower panels. See Figure 4 for definitions.

XIAP is a direct target of miR-34a*.

To test whether XIAP is a direct target of miR-34a*, we performed a reporter gene assay. Due to its long sequence, the 3′-UTR of XIAP was divided into 3 overlapping parts and each fragment was cloned into the pGL3 control vector (Figures 6A and B). Cotransfection of HEK 293 cells with the XIAP 3′-UTR construct 1 and pre–miR-34a* yielded a mean ± SEM of 45 ± 8% less relative luciferase activity as compared to control transfected cells (P < 0.002) (Figure 6C). When HEK 293 cells were transfected with XIAP 3′-UTR constructs 2 or 3, no reduction in relative luciferase activity was observed after transfection with pre–miR-34a*. These findings indicate a direct regulatory interaction of miR-34a* with the apoptosis inhibitor XIAP via binding to the first part of the XIAP 3′-UTR.

Figure 6.

Evidence that miR-34a* directly targets XIAP. A, Due to its size, the 3′-untranslated region (3′-UTR) of XIAP was divided into 3 overlapping parts: XIAP 1, XIAP 2, and XIAP 3. Putative miR-34a* seed matches were identified using Diana Lab algorithms (http://diana.cslab.ece.ntua.gr/), and their positions inside the XIAP 3′-UTR are indicated. B, Each part of the 3′-UTR of XIAP was cloned downstream of the luciferase gene, which is under the control of the constitutively active SV40 promoter. C, Reporter gene assay was performed using HEK 293 cells (n = 4) cotransfected with luciferase constructs containing the XIAP 3′-UTR (XIAP 1, 2, or 3) downstream of the luciferase gene and pre–miR-34a* or negative pre-control (pre-ctr), and luciferase activity relative to that obtained with pre-ctr (set at 1) was determined. Values are the mean ± SEM. ∗ = P < 0.002 by paired t-test. See Figure 4 for other definitions.

DISCUSSION

The prolonged cellular lifespan is a fundamental characteristic of activated synovial fibroblasts, contributing to chronic inflammation and joint destruction in RA. In the current study we focused on the expression and function of miR-34, a family of microRNAs that is known to influence apoptotic cell death (12, 15, 23–25). Whereas miR-34a, miR-34b/b*, and miR-34c/c* did not show any significant differences in their expression levels, we demonstrated significant down-regulation of the passenger strand miR-34a* in RASFs compared to OASFs.

Although microRNAs have been the subject of much research in recent years, little is yet known about the relevance of miR passenger strands. Some years ago, it was proposed that regulation of the expression and accumulation of miR/miR* duplexes represents a tissue-dependent mechanism (26). There is evidence of active sorting of both miR and miR* species into regulatory complexes, suggesting that miR* strands may have a biologic role (27). For instance, the passenger strand of miR-17 was shown to directly target the p50 subunit of NF-κB in a human cell line for malignant pleural mesothelioma (28). In addition, miR-30* was able to bind and suppress the translation of target genes, indicating that both the mature and the passenger strands are functional (26). Of relevance to the present results, the passenger strand miR-34* has been found to be localized to argonaute 1, a protein that is crucial for miRs to exert their biologic functions (27, 29).

Recently, our group reported alterations in the DNA methylation status in synovial cells from patients with RA versus those with OA (21). In the present study, we demonstrated that the expression of miR-34a and its passenger strand was increased upon demethylation, although statistical significance was reached only for miR-34a*. Consistent with our data, miR-34a expression has been found to be silenced in various cancers due to aberrant CpG methylation (14, 23). In contrast, cytokines such as TNFα and IL-1β, as well as TLR ligands, did not influence the expression of miR-34a*.

Whereas impaired apoptosis of synovial fibroblasts is important for the pathogenesis of RA, little is known about the miR-dependent regulation of apoptosis. The present findings suggest that both miR-34a and miR-34a* are involved in the regulation of apoptotic pathways. We present evidence here that miR-34a* promotes apoptosis in both FasL- and TRAIL-stimulated RASFs, whereas overexpression of the mature strand miR-34a protects cells against FasL-mediated apoptosis but has no effect on TRAIL-induced cell death.

Conflicting data on the effect of miR-34a on apoptosis have been published. In the colon cancer cell line HCT116, miR-34a was found to promote apoptosis (12). Similarly, miR-34a was found to suppress sirtuin 1, ultimately leading to p53-dependent apoptosis in cancer cells (24). However, a recent study demonstrated antiapoptotic effects of miR-34a in B lymphoid cells (15). Consistent with this finding, up-regulation of miR-34a was positively associated with longer survival of lung cancer cells (23). Concerning the passenger strand miR-34a*, no data on its effect on apoptosis have been published previously. In this study, we demonstrated opposite effects on apoptotic cell death of miR-34a and miR-34a*, with the passenger strand increasing FasL- and TRAIL-induced apoptosis of RASFs. A similar conflicting outcome was observed with miR-155 and its corresponding passenger strand in terms of modulation of interferon-α/β expression (30).

Furthermore, we found that the proapoptotic effects of miR-34a* are mediated upon targeting of XIAP. XIAP is a member of the IAP family of apoptosis inhibitors that block apoptosis by direct binding to caspases, leading to inhibition of the execution phase of apoptosis (31). It was recently shown that XIAP is implicated in innate immunity by mediating signaling of nucleotide-binding oligomerization domain–containing proteins 1 and 2 (32, 33). Consistent with our findings, higher expression levels of XIAP were found in synovial tissue from patients with active RA compared with normal subjects (22) and, in addition, XIAP was reported to be overexpressed in synovial fluid from patients with juvenile idiopathic arthritis (34). Induction of XIAP was observed upon stimulation with hematopoietic cytokines, e.g., granulocyte–macrophage colony-stimulating factor, granulocyte colony-stimulating factor, and stem cell factor, in a human leukemia cell line (35).

Herein we present new evidence of a regulatory role of miR-34a* in the expression of XIAP. Basal expression levels of XIAP correlated with constitutive miR-34a* production, and overexpression/inhibition of miR-34a* modulated XIAP production in RASFs. Using reporter gene assays, we showed that miR-34a* directly targets XIAP via binding to its 3′-UTR region. To our knowledge, this is the first report of an effect of the passenger strand miR-34a* on apoptosis. Our data indicate a functional effect of miR* species, which should be taken into account when analyzing the role of miRs in the pathogenesis of RA.

In conclusion, our results indicate that miR-34a* has a proapoptotic role that is mediated by modulation of the expression of XIAP. As miR-34a* is down-regulated in RASFs, unopposed expression of XIAP may contribute importantly to the apoptosis resistance of these cells.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Kyburz had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Niederer, Ospelt, R. E. Gay, Detmar, S. Gay, Jüngel, Kyburz.

Acquisition of data. Niederer, Trenkmann, Karouzakis, Kolling, Jüngel.

Analysis and interpretation of data. Niederer, Ospelt, Karouzakis, Neidhart, Stanczyk, S. Gay, Jüngel, Kyburz.

Acknowledgements

We thank Ioannis Pandis for valuable suggestions and Maria Comazzi, Peter Künzler, and Ferenc Pataky for excellent assistance with immunohistochemistry and cell culture.

Ancillary