miR-21 induces endothelial progenitor cells proliferation and angiogenesis via targeting FASLG and is a potential prognostic marker in deep venous thrombosis

Male Sprague–Dawley (SD) rats (200–250 g) were purchased from Shanghai Laboratory Animal Center (Shanghai, China) and were kept in specific pathogen-free (SPF) animal rooms. All the procedure was approved by the Institutional Animal Care and Use Committee of the Second Affiliated Hospital of Soochow University. This rat model is well described in previous studies [16]. In brief, the rats were anesthetized by intraperitoneal injection of 7% pentobarbital and underwent midline laparotomy in order to dissect the inferior vena cava (IVC) free from aorta. And then IVC was ligated just below the upper renal vein with a 6-0 Prolene sutures. The posterior venous branches were also tightened. After that, the confluence of iliac vein was clamped with vascular clips for 15 min. Then the incision was closed and the rats were allowed to recover after surgery. Thirty rats were randomly divided into three groups (n = 10 for each group): (A) blank control group received 1 ml EGM-2-MV medium, (B) lentivirus vector group received blank vector and (C) miR-21 overexpression group received pGLV3-H1-Puro-miR-21.

EPCs isolation was performed as previously described [17]. Briefly, male SD rats (200–250 g) were sacrificed and bone marrow was harvest from femurs and tibias. Mononuclear cells were acquired with density gradient centrifugation and cultured in EGM-2-MV (Lonza, MD, USA) medium at 37 °C in a 5% CO₂ incubator. Non-adherent cells were washed 4 days after culture. Then medium was changed every 2 days. EPCs in passage 3–4 were used for subsequent experiments.

Seven milliliters of venous blood were collected from DVT patient from July 2012 to June 2015 in our department. Inclusion criteria consisted mainly age range 30–50 years, suffered from primary acute DVT extending to the high femoral or iliac vein, symptom duration of less than 2 weeks, verified by ultrasound or digital subtraction angiography (DSA), good functional status. The following exclusion criteria were applied: isolated infrapopliteal thrombosis, contraindications to anticoagulation or thrombolytic agents, with malignant tumors, bacterial endocarditis, during pregnancy and declined to provide informed consent. The whole blood was centrifuged at 4 °C, 2000 rpm for 5 min followed by centrifugation at 12,000 rpm for 15 min. The serum samples were portioned in aliquots and stored at − 80 °C. The protocols were approved by the Institutional Review Board of Second Affiliated Hospital of Soochow University. Informed consent was obtained from each participant prior to specimen acquisition.

Small RNAs were extracted from 500 μL of serum using a miR-PARIS kit (AM1556) according to the manufacturer’s instructions. To allow for normalization of sample-to-sample variation in RNA isolation, synthetic Caenorhabditis elegans miRNAcel-miR-21 (purchased as a custom RNA oligo nucleotide from Qiagen) was added (50 pmol/L in a 5 μL total volume) to each denatured sample.

Total RNA of EPCs was extracted with Trizol Reagent (Invitrogen; Carlsbad, CA, USA). Mature miRNA expression analysis was done using the miRNA real-time PCR quantitation kit (Applied Biosystems, Foster City, CA, USA). The expression of miR-21 was carried out using the Applied Biosystems 7500 Real Time PCR System, with U6 as an internal control. mRNA expression analysis was performed using Power SYBR Green (Applied Biosystems, Foster City, CA, USA). PCR primers (forward and reverse) were as follows: rno-mir-21, forward: GCGGCGGTAGCTTATCAGACTG and reverse: ATCCAGTGCAGGGTCCGAGG; U6, forward: GCTTCGGCAGCACATATACTAAAAT and reverse: CGCTTCACGAATTTGCGTGTCAT; GAPDH, forward: CGCATCTTCTTGTGCAGTG and reverse: GAGGGTGCAGCGAACTTTATT.

To regulate the expression of miR-21 in EPCs, miR-21 agomir, antagomir or respective negative control were transfected into cells with Lipofectamine 3000 (Invitrogen; Carlsbad, CA, USA). 72 h after transfection, cells were harvested for subsequent experiments. Transfection efficacy was evaluated by qRT-PCR. The sequence of miR-21 agomir were: 5′-UAGCUUAUCAGACUGAUGUUGA-3′; agomir negative control were: 5′-UUCUCCGAACGUGUCACGUTT-3′; miR-21 antagomir were: 5′-UCAACAUCAGUCUGAUAAGCUA-3′; antagomir negative control were: 5′-CAGUACUUUUGUGUAGUACAA-3′. The target sequence of siRNA against rat FASLG (NM_012908.1) was 5′-GCAGAACUCCGAGAGUCUATT-3′.

A total of 1 × 10⁴ cells of EPCs were seeded to 24-well plates in a final volume of 800 ul medium for assessment of proliferation ability. 72 h after seeding, cell proliferation was evaluated using the Cell Counting Kit-8 (Dojindo, Kumamoto, Japan). All experiments were performed in triplicate.

For in vitro tube formation assay, EPCs were seeded in the presence of EGM-2-MV medium at a density of 2 × 10⁴/mL for 8 h at 37 °C in a 48-well plate coated with Matrigel (R&D Systems, MN, USA). The formation of capillary-like structures was captured under a light microscope. Each experiment was done in triplicate.

The 3′-UTR of FASLG containing the putative miRNA target site(s) was cloned into the SpeI and HindIII sites of the pMIR-REPORT Luciferase vector (Ambion, TX, USA). 293T cells were transfected with firefly luciferase reporter vector, miRNA, and renilla luciferase control vector using lipofectamine 3000. The reporter assays were analyzed by the examination of ratio between firefly and renilla luciferase activities. The experiments were performed in triplicate.

Total proteins extracted from EPCs using RIPA buffer (Sigma-Aldrich, St. Louis, MO, USA) were separated by SDS-polyacrylamide gel and transferred into PVDF membranes. Membranes were blocked with 5% non-fat milk TBST and incubated with primary antibody for FASLG (Abcam, Cambridge, MA, USA), followed by the incubation with appropriate HRP-conjugated secondary antibody. β-actin (Sigma-Aldrich, St. Louis, MO, USA) was then measured as internal control. The densitometry of western blot results was measured using ImageJ software.

The lentiviral expression vector pGLV3-H1-Puro-miR-21 was constructed to stably express miR-21 in EPCs. 293T cells were cotransfected with pGLV3-H1-Puro vector or pGLV3-H1-Puro-miR-21 plasmid using lipofectamine 3000 (Invitrogen; Carlsbad, CA, USA). Then viral particles were harvested from 293T cells and enriched. Finally, viral titers were determined by counting the labeled cells or using qRT-PCR to detect GFP expression. Lentivirus miR-21 or lentivirus vector was used to transfer miRNA into rats. Three days after thrombus formation, rats were injected within the thrombus with different solution. The solution contained 1 × 10⁹ TU/mL lentivirus miR-21 or lentiviral particles. The rats in normal group were injected with 2 mL EGM-2-MV medium.

Seven days after injection, the rats were perfused and IVC segments with thrombus were removed and fixed in 4% paraformaldehyde, embedded in paraffin. All the fixed tissue was sliced at 8-μm intervals. Hematoxylin/eosin staining were done using standard procedures. Images were captured using an inverted microscope. Before the thrombi were weighed, excessive blood on the thrombi was removed by filter paper.

Seven days after injection, IVC venography was acquired with digital subtract angiography (DSA, GE Innova 3100, USA) by injecting contrast media into rat caudal vein or femoral vein to determine the recanalization and resolution of thrombus in vivo. All the acquired images were analyzed using image J software.

Patients attended follow-up visits at 1 month and 3 months after treatment and every 6 months thereafter and were contacted by telephone or e-mail at the 3-month mark between visits. The primary outcome of the study was either a recurrence of venous thromboembolism or PTS. Recurrent DVT was defined as a composite of symptomatic, objectively confirmed deep-vein thrombosis, nonfatal pulmonary embolism, or fatal pulmonary embolism. PTS was defined as patients with suggested symptoms including pain, heaviness, edema, varicose vein, discoloration and or ulcer in affected lower extremity. Villalta score was recorded to assess the severity of PTS during follow-up.

Data are presented as mean ± SEM. Differences among groups were tested by one-way ANOVA. Statistical analyses between two groups were evaluated based on the Student’s two-tailed t- test. Kaplan–Meier method was used to compare the recurrent DVT between patients in different groups with log-rank test. Univariate associations between candidate predictors and recurrence of DVT were examined with 95% confidence interval (CI) by using Cox proportional hazards model. Multivariate Cox regression analysis with backward conditional method was performed to select significant prognostic factors. In all analyses, p < 0.05 was considered significant.

By using endothelial progenitor cells isolated from control (n = 3) and DVT model (n = 3), we found that miR-21 was shown to be significantly down-regulated in DVT model (Fig. 1a). To further verify the downregulation trend of miR-21 in DVT model, EPCs from ten additional DVT model and control were further examined using quantitative RT-PCR. The results revealed that miR-21 had an average of 2.8-fold higher expression level in control than in DVT model (Fig. 1b).

To determine the effect of miR-21 on EPCs function, we tested the proliferation and angiogenesis of EPCs transfected with miR-21 agomir or antagomir. CCK-8 assay showed that miR-21 agomir significantly promoted proliferation of EPCs compared with that in control group, whereas miR-21 antagomir decreased the proliferation of EPCs (Fig. 2a). Furthermore, we test pro-angiogenic effect of miR-21 on EPCs. As is shown in Fig. 2b, c, EPCs transfected with miR-21 agomir exhibited increased angiogenic capability when compared with those transfected with vehicle control, whereas miR-21 antagomir showed decreased angiogenic potential.

As is known to all that miRNAs cause mRNA cleavage or translational repression by forming imperfect base pairing with the 3′UTR of target genes. We searched multiple database including TargetScan, miRDB and Miranda, the results suggested that FASLG might be a potential target of miR-21, which contained a miR-21 binding site in 3′UTR region (Fig. 3a). To confirm that miR-21 could bind FASLG mRNA 3′UTR region, we constructed a luciferase reporter vector which contained this region. The results revealed that luciferase activity significantly decreased in the presence of miR-21. However, miR-21 could not affect luciferase activity if 3′UTR FASLG was muted (Fig. 3b). To further address this question, we applied miR-21 agomir and antagomir to determine the expression of FASLG protein levels. As is shown in Fig. 3c, FASLG protein level is significantly lower than that in the group of EPC transfected with miR-21 agomir, while EPCs transfected with miR-21 antagomir upregulated FASLG protein expression. Moreover, decreased FASLG protein level was observed when we employed FASLG siRNA in EPCs (Fig. 3d).

To test the effect of FASLG on EPCs function, FASLG siRNA was applied to downregulate the expression of FASLG. EPCs transfected with FASLG siRNA revealed increased cell proliferation and angiogenesis when compared to that of control group. However, miR-21 antagomir reversed this effect when co-transfected with FASLG siRNA in EPCs (Fig. 4a–c). Furthermore, we performed the rescue experiment and found the similar results when co-transfected with miR-21 agomir and FASLG (Fig. 4d–f). Taken together, these results demonstrated that miR-21 regulated function of EPCs via targeting FASLG.

To verify whether miR-21 could affect the prognosis of DVT in vivo, a rat model of DVT was established. Cell culture medium, lentivirus miR-21 or lentivirus vector was injected 3 days after thrombus formation. Compared with blank control and vehicle control groups, the rats injected with lentivirus miR-21 showed less thrombosis in IVC segment (Fig. 5a). Moreover, in the lentivirus miR-21 groups, the weight of thrombus was obviously lower than that in other groups (Fig. 5b). Besides, we investigated the thrombus resolution in vivo by DSA. The results showed that significant increase of recanalization ratio in rats received lentivirus miR-21 compared to rats received lentivirus vector (Fig. 5c). Collectively, these results confirmed that miR-21 contributed to thrombus resolution in vivo.

Since miR-21 regulated EPCs biological function and promoted thrombus resolution, we then investigated miR-21 expression level in DVT patients by using RT-PCR. A total of 146 patients with DVT were referred to our department from July 2012 to June 2015. In the initial phase, 63 patients were excluded based on the exclusion criteria. In addition, 11 patients refused to provide blood sample were also excluded. Thus, we collected specimen from 72 patients. 18 patients were geographically unable to return for follow-up visits and 10 patients were confirmed with invalid Villalta score. Another four cases were excluded for case–control purpose. Finally, 40 patients were recruited in this study (Additional file 1: Figure S1). 40 patients were divided into two groups with low or high serum miR-21 using the median expression level of all 72 cases as the cut-off point. The follow‑up duration was 2 year. The patient demographic data are shown in Table 1. Kaplan–Meier analysis revealed that patients with low level of serum miR-21 had unfavorable trends of recurrent DVT. There was significant difference on recurrence-free rate between high level and low level group using the log-rank test (94.7% vs 64.6%, p = 0.019), respectively (Fig. 6a). PTS is a complication of DVT, which affect 23–60% of patients in the 2 years following DVT of the leg. PTS was assessed using the Villalta scale (Table 2), consisting of five patient‑rated leg symptoms and six clinician-rated physical signs. The results showed that patients with high miR-21 expression had lower Villalta score at 2 years after treatment (Fig. 6b). Accordingly, PTS was considered present if the Villalta score was > 4 in the leg ipsilateral to DVT and was categorized as mild if the score was 5–9, moderate if the score was 10–14 and severe if the score was > 14 or an ulcer was present. Figure 6c were representative images for PTS patients using digital subtraction angiography (DSA). We evaluate the number of patients with different severity of PTS categorized by miR-21 expression level. Our data showed that patients with more severe PTS have lower miR-21 expression level while patients with mild PTS have higher miR-21 expression level (Fig. 6d).

To further investigate independent risk factors for recurrence of DVT. Six candidate predictors including age, gender, BMI (BMI ≥ 30 or BMI < 30), categories of DVT (unprovoked or secondary to risk factors), duration of anticoagulation and miR-21 expression (low level or high level) were selected based on existing retrospective researches and a priori hypotheses. Univariate Cox proportional hazard regression analyses were performed and found that miR-21 expression level was significantly associated with recurrence of DVT. The other significant prognostic factors in univariate analyses including age, category of DVT, duration of anticoagulation and miR-21 expression (Table 3). Moreover, multivariate analysis indicated that miR-21 expression, age and duration of anticoagulation demonstrated significant correlation with recurrence (Table 3). Taken together, these results showed that miR-21 expression level can serve as an independent predictor of recurrence of DVT.