MicroRNA-21 and Venous Neointimal Hyperplasia of Dialysis Vascular Access

Abstract

Background and objectives

There is increasing evidence that microRNAs (miRNAs) play crucial roles in the regulation of neointima formation. However, the translational evidence of the role of miRNAs in dialysis vascular access is limited.

Design, setting, participants, & measurements

miRNA expression in tissues was assessed by using venous tissues harvested from ten patients on dialysis who received revision or removal surgery, and ten patients who were predialysis and received creation surgery of arteriovenous fistulas served as controls. To extend these findings, 60 patients who received angioplasty of dialysis access were enrolled and the levels of circulating miRNAs were determined before and 2 days after angioplasty. Clinical follow-up was continued monthly for 6 months. The primary outcome of angioplasty cohort was target lesion restenosis within 6 months after angioplasty.

Results

In the surgery cohort, the expressions of miR-21, miR-130a, and miR-221 were upregulated in stenotic tissues, whereas those of miR-133 and miR-145 were downregulated. In situ hybridization revealed similar expression patterns of these miRNAs, localized predominantly in the neointima region. Twenty eight patients in the angioplasty cohort developed restenosis within 6 months. The levels of circulating miR-21, miR-130a, miR-221, miR-133, and miR-145 significantly increased 2 days after angioplasty. Kaplan–Meier plots showed that patients with an increase of miR-21 expression level >0.35 have a higher risk of patency loss (hazard ratio, 4.45; 95% confidence interval, 1.68 to 11.7). In a multivariable analysis, postangioplasty increase of miR-21 expression was independently associated with restenosis (hazard ratio, 1.20; 95% confidence interval, 1.07 to 1.35 per one unit increase of miR-21 expression level; P=0.001).

Conclusions

Certain miRNAs are differentially expressed in the stenotic venous segments of dialysis accesses. An increase in blood miR-21 level with angioplasty is associated with a higher risk of restenosis.

Visual Abstract

Introduction

Dialysis vascular access dysfunction results in hospitalization and morbidity in patients with ESKD and is associated with nearly one tenth of hemodialysis-related cost (1). Majority of hemodialysis vascular access dysfunctions are attributable to intimal hyperplasia of the outflow veins (2). Although angioplasty is effective in restoring function of the accesses, its benefits are attenuated by a high restenotic rate (3–5). Previous studies showed a significant increase in the proliferative activity of vascular smooth muscle cells and myofibroblasts in neointimal lesions (6). Nonetheless, the upstream regulators of the highly proliferative smooth muscle cells and myofibroblasts remain unclear.

MicroRNAs (miRNAs) have emerged as a novel class of gene regulators that work via transcriptional degradation and translational inhibition or activation (7). miRNAs could directly regulate >30% of genes in cells. Thus, miRNAs are crucial regulators in normal development, physiology, and pathogenesis (8). In a variety of vascular diseases, miRNA expression profiles are altered (7). Recent studies showed substantial evidence that certain miRNAs, i.e., miR-21, miR-221/222, miR-130a, miR-133, and miR-145, can regulate the differentiation and proliferation of vascular smooth muscle cells after vessel injury (9–13). These results suggest that these miRNAs may play a critical role in the development of intimal hyperplasia after vessel injury and may be a potential target of intervention against restenosis. However, to our knowledge, there is no translational evidence for these miRNAs in venous stenosis of dialysis vascular accesses in humans.

Thus, our study aimed at investigating the potential role of specific miRNAs in the pathogenesis of venous intimal hyperplasia in patients on hemodialysis (14,15). The expression of these miRNAs was first compared between control vein samples and stenosed vein segments of patients on hemodialysis. The role of miRNAs that were differentially expressed in stenotic tissues was further elucidated by determining the relationship between their circulating levels and the development of restenosis after angioplasty. The results of our study provide new insights of specific miRNAs in the regulation of intimal hyperplasia in dialysis arteriovenous fistulas.

Materials and Methods

Study Population

Ten patients who underwent surgical revision or removal of vascular access were included in the initial surgery cohort (disease group), enrolled from November 2013 to July 2016. Ten patients who were predialysis and had arteriovenous fistula creation served as the control group. Surgical revision was performed because of recurrent stenoses or failed endovascular therapy; surgical removal was performed because of infection or vascular access aneurysm. The results obtained in the fistula surgery cohort were examined in another cohort of 60 patients who underwent angioplasty for vascular accesses dysfunction, enrolled from August 2016 to November 2016. They were referred on the basis of one or more of the following criteria: (1) clinical signs suggesting vascular access dysfunction, (2) reduction in flow rate of >25% from baseline, (3) total access blood flow rate <500 ml/min as measured by the ultrasound dilution method, or (4) increased venous pressure during dialysis. Patients with thrombosis, arterial lesion, central vein lesions, or failed endovascular therapy were excluded. All eligible patients on hemodialysis in this study received 4 hours of hemodialysis three times a week, using a synthetic dialysis membrane (polyamide or polysulfone). The adequacy of dialysis was accessed monthly by the single-pool Kt/V of urea nitrogen. This study was approved by the institutional review board of National Taiwan University Hospital, Hsinchu Branch (approval no. 103–022-F). This study was conducted in accordance with the Declaration of Helsinki, and written informed consent was obtained from all participants.

Angiography and Angioplasty Procedures

Diagnostic angiography was performed in the morning on a midweek dialysis day, within 4 hours before the upcoming dialysis session. After diagnostic angiography, angioplasty was performed in accordance with the National Kidney Foundation-Dialysis Outcomes Quality Initiative Vascular Access guidelines, i.e., only for patients with clinical dysfunction and a minimum of 50% diameter stenosis (16). The stenosis was treated with standard angioplasty techniques, and high pressure or cutting balloons were used only for resistant stenosis (17). After the angioplasty procedure, only aspirin was prescribed for 3 days. Medications for underlying cardiovascular diseases were continued on the basis of the original indications.

Collection of Specimens, Clinical Information, and Follow-Up of Patients

Blood sampling was scheduled at baseline (after angiography but before angioplasty) and 2 days after angioplasty. Blood samples were drawn directly from the arteriovenous fistulas and the plasma samples were used for miRNA analysis. Baseline data were obtained through a review of medical and dialysis records, angiographic and intervention reports, and interviews with the participants and their referring nephrologists, if needed (18). Study participants of the angioplasty cohort were followed for 6 months. The clinical follow-up included physical examination, dynamic venous pressure monitoring, and transonic examination of access blood flow rate, if available. The follow-up was performed by a coordinating study nurse through telephone contact at monthly intervals. The referring nephrologists were blinded to the participants’ miRNA data. Patients with abnormal clinical or hemodynamic parameters were referred for repeat fistulography and angioplasty, as appropriate. The primary outcome was target lesion restenosis within 6 months.

Quantitative Real-Time PCR for miRNA Expression

Tissue samples were ground and homogenized before RNA extraction. RNA was extracted using TRIzol (Ambion, Austin, TX) according to the manufacturer’s instructions. The RNA extracted would be used by a threshold of RNA integrity number >7. On the basis of previous animal studies, six miRNAs were selected for quantitative real-time PCR quantification (9–13,19–21). For quantitative analysis of miRNA, complementary DNA from 1 μg of total RNA was obtained by the SuperScript II Reverse transcription (Invitrogen) as described above. Quantitative real-time PCR was performed on a LightCycler 1.5 (Roche Diagnostics, Mannheim, Germany). miRNA expression was validated by quantitative stem-loop PCR technology as previously described (22,23). The use of target-specific reverse transcription primers and universal probe library probe #21 (Roche Diagnostics) allowed for the specific detection of mature miRNAs. A final volume of 20 μl consisted of 0.5 μM of each forward and reverse primer, 0.1 μM probe #21, 1× LightCycer TaqMan Master (Roche Diagnostics), and 2.5 μl of complementary DNA. Amplification curves were generated with an initial denaturing step at 95°C for 10 minutes, followed by 45 cycles of 95°C for 5 seconds, 60°C for 10 seconds, and 70°C for 1 second. U6 served as the reference gene for the detection of miRNA expression in tissue. Fluorescence signals were normalized to U6, and the threshold cycle (Ct) was set within the PCR. The target PCR Ct values were standardized by subtracting the U6 Ct value, which provided the Δ(difference)Ct value. The following equation was used to calculate the relative miRNA expression level: (relative quantities =2^{−[ΔCt sample−ΔCt control]}). For the assay of circulating miRNA in plasma, values were normalized to that of cel-miR-39 and were expressed as relative quantities =2^{− (ΔCt sample−ΔCt miR-39)}, as previously described (24).

Statistical Analyses

Continuous variables were presented as means±SD for normally distributed variables and median with interquartile range for non-normally distributed variables. There were no missing data that need to be handled. Comparisons of continuous variables were made using t test, the Mann–Whitney U test, and the chi-squared test, with Yates correction and Fisher exact test as appropriate. Univariable linear regression analyses were employed to assess the relationships between miRNA levels and baseline variables. In the angioplasty cohort, the increase in miRNA after angioplasty was defined as the miRNA level 2 days after angioplasty minus the level before angioplasty. Restenosis-free patency was estimated using the Kaplan–Meier method and compared using the log-rank test. Cox proportional hazard regression analysis was used for estimating the relative hazard of restenosis factors. The assumption of proportionality was checked graphically using the log-log plot and was found to be acceptable for the risk factor of interest. All variables with a P value <0.2 in the univariable analysis and biologically relevant factors, including age, sex, diabetes, reference vessel diameter, and use of statin were adjusted in the multivariable regression model. Receiver’s operation curve was used to determine the best cut-off point of miR-21 increase by maximizing both sensitivity and 1-specificity with the Youden index. P values <0.05 were considered significant. All statistical calculations were performed with SPSS v20.0.

An expanded methods section is provided in the online Supplemental Material.

Results

Characteristics of the Patients Who Received Surgery

Venous segments of dialysis vascular accesses were harvested from ten patients on hemodialysis: five (recurrent stenosis, three; failed angioplasty, two) had venous segments with significant stenosis at the time of surgical revision (significant stenosis group), and the other five (vascular access infection, three; vascular access aneurysm, two) had venous segments without significant stenosis (nonsignificant stenosis group). The mean age of the study participants was 70 years, and four were men. The clinical characteristics of the surgery cohort are summarized in Table 1.

Table 1.

Baseline characteristics of the study participants in the surgery and angioplasty cohorts

Factors	Surgery cohort (n=20)			Angioplasty cohort (n=60)
	Control (n=10)	Nonsignificant stenosis (n=5)	Significant stenosis (n=5)	Patent (n=32)	Restenosis (n=28)
Clinical factors
Age, yr	64 (9)	66 (15)	72 (5)	64 (13)	65 (14)
Sex, male	6 (60%)	3 (60%)	1 (20%)	18 (56%)	18 (64%)
Dialysis vintage, mo	NA	60 (12–96)	77 (50–82)	31 (10–108)	36 (14–76)
Kt/V	NA	1.24 (0.15)	1.23 (0.11)	1.37 (0.24)	1.41 (0.21)
Hypertension	8 (80%)	3 (60%)	4 (80%)	23 (72%)	15 (54%)
Diabetes mellitus	5 (50%)	3 (60%)	3 (60%)	9 (28%)	11 (39%)
Dyslipidemia	2 (20%)	0 (0%)	1 (20%)	5 (16%)	2 (7%)
Active smoker	3 (30%)	1 (20%)	1 (20%)	4 (13%)	6 (21%)
Coronary artery disease	6 (60%)	4 (80%)	4 (80%)	8 (25%)	21 (75%)
Antiplatelet	6 (60%)	4 (80%)	4 (80%)	8 (25%)	10 (36%)
ACEI/ARB	2 (20%)	1 (20%)	1 (20%)	3 (9%)	1 (4%)
Statin	2 (20%)	1 (20%)	2 (40%)	4 (13%)	1 (4%)
Access and lesion factors
Shunt age, mo	NA	15 (11–96)	50 (48–82)	24 (11–00)	30 (14–64)
Right arm access	NA	2 (40%)	1 (20%)	10 (31%)	3 (11%)
Upper arm access	NA	0 (0%)	2 (40%)	8 (25%)	10 (36%)
Stenosis before, %	NA	NA	NA	72 (9)	76 (9)
Stenosis after, %	NA	NA	NA	14 (21)	6 (10)
Lesion length, mm	NA	NA	NA	33 (10)	47 (20)
Vessel diameter, mm	NA	NA	NA	6.2 (1.0)	6.7 (0.7)
Outflow vein location	NA	4 (80%)	3 (60%)	22 (69%)	20 (71%)

Values were expressed as mean (SD), median (interquartile range), or n (percentage). NA, not applicable; Kt/V, urea clearance; ACEI/ARB, angiotensin-converting enzyme/angiotensinogen receptor blocker.

Expression of miRNAs in Tissues by Quantitative Real-Time PCR

Expression profiles of miRNAs in the stenotic veins of vascular access and control veins were compared by quantitative real-time PCR (Figure 1, Table 2). The expression profiles of some miRNAs were different between the diseased and control groups and between the significant stenosis and nonsignificant stenosis groups. In the diseased group, miR-133 and miR-145 were downregulated and miR-21, miR-130a, miR-221, and miR-222 were upregulated compared with the control group. In the significant stenosis group, miR-21, miR-130a, and miR-221 were upregulated compared with the nonsignificant stenosis group.

Figure 1.

Tissue miRNA levels in venous samples were different among the control veins, non-significant stenotic lesions, and significant stenotic lesions. Expressions of indicated miRNAs in control veins harvested at arteriovenous fistula creation (CT) and diseased veins harvested at surgical revision or removal of dialysis vascular access were examined. The diseased veins were further classified into the nonsignificant stenosis (NS) and significant stenosis (SS). RQ, relative quantities, the relative expression level of miRNAs determined by the cycle number via quantitative real-time PCR, normalized to U6 expression.

Table 2.

miRNA levels in the tissues of control and diseased veins

miRNAs	Control	Disease	P Value	Nonsignificant	Significant
(RQ)	n=10	n=10		n=5	n=5
MiR-21	0.74 (0.27–2.11)	7.29 (3.76–15.2)	<0.001	3.77 (2.80–4.74)	14.9 (10.6–21.2)
MiR-130a	0.47 (0.17–0.99)	1.84 (0.36–6.55)	0.06	0.38 (0.26–0.63)	5.83 (4.21–14.6)
MiR-221	0.89 (1.14–2.63)	14.1 (3.04–60.5)	0.02	3.72 (0.99–6.44)	56.7 (33.8–79.7)
MiR-222	1.53 (1.18–2.56)	10.4 (4.46–24.9)	0.02	5.62 (0.86–10.4)	22.9 (13.6–32.2)
MiR-133	2.57 (2.02–3.02)	0.58 (0.15–1.00)	<0.001	0.91 (0.48–2.24)	0.22 (0.11–0.58)
MiR-145	1.13 (0.93–1.63)	0.09 (0.06–0.80)	0.02	0.75 (0.07–1.94)	0.07 (0.05–0.08)

The miRNA levels in the diseased veins were classified as significant or nonsignificant stenosis. Values are median (interquartile range). miRNA, microRNA; RQ, relative quantities, the relative expression level of miRNAs was determined by the cycle number via quantitative real-time PCR, normalized to U6 expression.

Expression of miRNAs in Tissues by Histologic Examination

The findings of the histologic examination of miRNAs were comparable to those of quantitative real-time PCR. Expressions of miR-21, miR-130a, and miR-221 were found in the diseased veins, but not in the control veins (Figure 2). The expressions of these miRNAs were localized predominantly in the neointima area and were more prominent in the significant stenosis group, as quantified by the immunofluorescence intensity measurements. By contrast, miR-145 expression was prominent in the control veins, less prominent in the nonsignificant stenosis group, and particularly rare in the significant stenosis group.

Figure 2.

The expression of miRNAs in venous tissues demonstrated by in situ hybridization were different among control veins, non-significant stenotic lesions, and significant stenotic lesions. (A) A control cephalic vein from a patient who was predialysis at arteriovenous fistula creation. (B) Nonsignificant stenotic lesions from an infected vascular access. (C) Stenotic lesions from arteriovenous fistula with recurrent stenosis. Smooth muscle cells were stained with α smooth muscle actin (α-SMA) and counterstained with 4′,6-diamidino-2-phenylindole (DAPI). (D) Fluorescence intensity analysis of selected neointima area. CT, control cephalic vein; NS, nonsignificant stenotic lesions; SS, significant stenotic lesions.

Characteristics of Patients Who Received Angioplasty

The study flow of the angioplasty cohort was demonstrated in the Supplemental Figure 1. No patients were lost to follow-up and the mean duration of follow-up was 122 days. During follow-up, 28 patients had restenosis at the same location. The characteristics of the patients with and without target lesion restenosis was summarized in Table 1. Increased medians of relative expression levels of miR-21 (1.72 versus 0.96; P<0.001), miR-130a (2.71 versus 1.26; P<0.001), and miR-221 (2.00 versus 0.99; P<0.001) at 2 days after angioplasty were noted (Figure 3). After stratification into patent and restenosis groups, the restenosis group had a higher postangioplasty level and increases of miR-21 and miR-130a, compared with the patent group (Table 3). Linear regression analysis showed that the increase of miR-21 was positively correlated with lesion length and stenosis severity before angioplasty (Supplemental Table 1).

Figure 3.

The circulating levels of certain miRNAs were elevated at 2 days after angioplasty and were different between the patent and restenosis groups. Expression of selected circulating miRNAs sampled from dialysis vascular access before and after angioplasty, stratified into patent and restenosis groups at 6 months. Whisker plots showed the 25th, 50th, and 75th percentile distributions in each panel; RQ, relative quantities, the relative expression level of miRNAs determined by the cycle number via quantitative real-time PCR, normalized to a control miRNA cel-miR-39 expression.

Table 3.

Circulating miRNA expression levels before and after angioplasty in the angioplasty cohort, according to patent and restenosis groups at 6 months after angioplasty

miRNAs	Patent (n=32)		Restenosis (n=28)		P Value
MiR-21
Before	0.76	(0.23–2.08)	1.06	(0.30–5.45)	0.22
After	1.31	(0.40–4.16)	2.39	(1.11–12.62)	0.02
Δ(after-before)	0.22	(0.12–0.88)	0.91	(0.48–3.53)	<0.001
MiR-130a
Before	0.75	(0.24–2.67)	1.67	(0.65–4.40)	0.09
After	2.10	(1.08–6.32)	4.61	(2.34–7.18)	0.04
Δ(after-before)	1.14	(0.30–2.48)	1.97	(1.32–3.87)	0.04
MiR-221
Before	0.76	(0.14–3.35)	1.42	(0.43–10.31)	0.21
After	1.93	(0.51–4.92)	2.62	(1.49–15.73)	0.96
Δ(after-before)	0.50	(0.20–1.21)	0.82	(0.46–1.48)	0.21
MiR-222
Before	1.24	(0.28–3.37)	1.12	(0.57–1.63)	0.12
After	1.16	(0.35–2.71)	1.15	(0.49–1.75)	0.05
Δ(after-before)	−0.03	(−0.23 to 0.42)	0.16	(−0.44 to 0.79)	0.12
MiR-133
Before	3.23	(0.91–5.43)	3.01	(1.43–5.06)	0.69
After	2.19	(1.31–4.63)	4.47	(1.55–8.58)	0.30
Δ(after-before)	0.25	(−1.26 to 1.90)	2.27	(0.02–4.72)	0.11
MiR-145
Before	1.38	(0.28–4.13)	1.43	(0.75–3.23)	0.79
After	2.40	(0.96–6.16)	2.34	(1.13–5.38)	0.87
Δ(after-before)	1.07	(0.24–2.83)	0.67	(−0.13 to 4.02)	0.61

Data were expressed as median (interquartile range). Unit of miRNA expression levels: relative quantities, the relative expression level of miRNAs determined by the cycle number via quantitative real-time PCR, normalized to a control miRNA cel-miR-39 expression. miRNA, microRNA; Before, value before angioplasty; After, value 2 days after angioplasty; Δ (after-before): value 2 days after angioplasty minus value before angioplasty.

Association of miRNAs with Restenosis

Univariable Cox proportional hazard regression analysis revealed that hypertension, coronary artery disease, lesion length (Supplemental Table 2), baseline miR-221, postangioplasty miR-21 and miR-221, and postangioplasty increase of miR-21 were risk factors for restenosis (Table 4). After adjustment for age, sex, hypertension, diabetes, statin use, coronary artery disease history, access side, vessel diameter, and lesion length and severity, only postangioplasty increases of miR-21 remained an independent risk of restenosis (adjusted hazard ratio, 1.20; 95% confidence interval, 1.07 to 1.35, per one unit increase of miR-21 expression level; P=0.001) (Table 4). We classified the patients by lesion length and degree of stenosis to eliminate the effect of lesion burden. Similar associations were observed between increase in miR-21 and restenosis-free patency (Supplemental Figure 2). On the basis of the receiver operating characteristic analysis, an increase of miR-21 expression level >0.35 was the best cut-off value for patency loss (Figure 4). Kaplan–Meier plots showed that the group with miR-21 expression level increase >0.35 had a higher risk of patency loss than the group with miR-21 expression level increase <0.35 (hazard ratio, 4.45; 95% confidence interval, 1.68 to 11.7; P=0.001) (Figure 4).

Table 4.

Associations of miRNAs and changes in miRNAs after angioplasty with time to restenosis

miRNA	Univariate		Multivariate
	HR (95% CI)	P Value	HR (95% CI)	P Value
MiR-21
Before	1.01 (0.99–1.01)	0.07	1.01 (1.01–1.08)	0.81
After	1.01 (1.00–1.01)	0.004	1.00 (0.93–1.05)	0.99
Δ(after-before)	1.19 (1.10–1.29)	<0.001	1.20 (1.07–1.35)	0.001
MiR-130a
Before	1.01 (0.95–1.07)	0.70	1.03 (0.96–1.11)	0.45
After	1.01 (0.98–1.04)	0.52	1.01 (0.98–1.05)	0.40
Δ(after-before)	1.02 (0.97–1.06)	0.48	1.02 (0.97–1.07)	0.46
MiR-221
Before	1.07 (1.03–1.11)	<0.001	1.01 (0.96–1.06)	0.79
After	1.06 (1.03–1.09)	0.001	1.01 (0.97–1.06)	0.58
Δ(after-before)	1.03 (0.94–1.13)	0.54	1.04 (0.94–1.14)	0.51
MiR-222
Before	0.85 (0.65–1.11)	0.24	0.80 (0.57–1.13)	0.21
After	0.90 (0.73–1.11)	0.32	0.81 (0.59–1.11)	0.19
Δ(after-before)	0.98 (0.77–1.26)	0.89	0.89 (0.66–1.22)	0.48
MiR-145
Before	0.92 (0.64–1.33)	0.67	0.70 (0.36–1.36)	0.19
After	0.96 (0.80–1.14)	0.63	0.76 (0.52–1.10)	0.16
Δ(after-before)	0.97 (0.80–1.17)	0.76	0.83 (0.56–1.23)	0.36
MiR-133
Before	1.06 (0.94–1.19)	0.38	1.09 (0.87–1.38)	0.45
After	1.02 (0.99–1.04)	0.22	1.03 (0.98–1.09)	0.21
Δ(after-before)	1.02 (0.99–1.05)	0.20	1.05 (0.98–1.12)	0.19

Cox proportional hazard regression analysis was used to estimate relative hazard for restenosis factors. Multivariable analysis was adjusted by age, sex, hypertension, diabetes, statin use, coronary artery disease history, access side, vessel diameter, and lesion length and severity. All miRNAs in the first column were entered into the multivariable analysis separately. The hazard ratio was estimated per one unit increase of miRNA expression level (relative quantities, the relative expression level of miRNAs determined by the cycle number via quantitative real-time PCR, normalized to a control miRNA cel-miR-39 expression). miRNA, microRNA; HR, hazard ratio; 95% CI, 95% confidence interval; Before, before angioplasty; after, 2 days after angioplasty; Δ(after-before): value 2 days after angioplasty minus value before angioplasty.

Figure 4.

Increase of circulating miRNA-21 level >0.35 at 2 days after angioplasty was associated with patency loss at 6 months. (A) Receiver operating characteristic (ROC) curve of postangioplasty miRNA increase for restenosis. (B) Kaplan–Meier analyses showing the proportion of patients without restenosis. Patients are divided according to the best cut-off value of miRNA-21 increase (ΔMIR-21=0.35 RQ) obtained from the ROC curves. ΔMIR-21 indicates MIR-21 level 2 days after angioplasty minus MIR-21 level before angioplasty. AUC, area under curve; PTA, percutaneous transluminal angioplasty; RQ, relative quantities, the relative expression level of miRNAs determined by the cycle number via quantitative real-time PCR, normalized to a control miRNA cel-miR-39 expression.

Discussion

This study demonstrated that the expression of certain miRNAs is upregulated in venous intimal hyperplasia in dialysis vascular accesses. Increases in the concentration of these miRNAs in the peripheral circulation were found after venous stenosis angioplasty. An increase in blood miR-21 level with angioplasty was associated with a higher risk of restenosis, independent of clinical and anatomic factors. Our study, using tissue specimens from patients on dialysis, may be the first study of its kind to show that neointima is characterized by an aberrant expression of certain miRNAs related to smooth muscle cell proliferation, and some of these miRNAs may be associated with the development of venous intimal hyperplasia in dialysis vascular access.

There is increasing evidence that certain miRNAs are involved in a variety of vascular pathologic processes in arterial diseases (7). In animal studies, miR-21, miR-145, and miR-221/222 have been shown to be involved in the neointimal growth of the carotid artery after injury by balloon dilation (7,9–11). Nevertheless, tissue-specific expression is an important characteristic of miRNAs, and studies investigating the role of miRNAs in venous tissues are rare. Venous stenosis of dialysis vascular access is histologically characterized by intimal thickening, mainly due to the proliferation of smooth muscle cells and myofibroblasts and deposition of extracellular matrix, and hence is different from atherosclerotic lesions. In a porcine vein graft model, Cao et al. (25) found that six miRNAs were downregulated and ten miRNAs were upregulated in vein grafts. Using quantitative real-time PCR, McDonald et al. (26) revealed an upregulation of miR-21 levels in mouse, pig, and human models of vein graft neointima formation. Similar to our study, Lv et al. (27) investigated the expression of miRNAs in stenotic arteriovenous fistulas of patients on hemodialysis. The discrepancy between our findings and those of Lv et al. may be attributable to the different methods used to choose target miRNAs. Lv et al. (27) used a hybridization-based miRNA array to screen differentially expressed miRNAs in stenotic venous tissues, which was different from the target miRNAs evaluated by the quantitative real-time PCR method used in our study. This hybridization-based platform used high-throughput technology to analyze hundreds to thousands of miRNAs in one assay, but it may be considered to have lower dynamic range and specificity than the quantitative real-time PCR method (28). In addition, some relevant miRNAs might be missed and more miRNAs with vague biologic significance may be selected out. In this study, the miRNAs evaluated were selected in accordance with the published biologic significances in animal or human cell studies. The expression of these miRNAs was verified by the quantitative real-time PCR method, a gold standard for miRNA detection, and further substantiated by the in situ hybridization method, which demonstrated an anatomic and severity correlation with the neointima area.

Our findings were supported by previous experimental studies. In a rat carotid artery injury model, multiple miRNAs were dysregulated at neointimal lesions. The time series of their expression is consistent with the development of neointima. miR-21 was one of the most upregulated miRNAs in the vascular wall after balloon injury. Moreover, inhibition in miR-21 expression significantly decreased neointimal formation after angioplasty (11). In animal models of in-stent restenosis, genetic ablation of miR-21 attenuated neointimal formation via increased anti-inflammatory M2 macrophage levels coupled with an impaired sensitivity of smooth muscle cells. These results could explain how miR-21 promotes vascular remodeling. For ethical considerations, tissues from different time points to demonstrate the temporal correlation between miRNA expression and neointima development were not obtained in our study. Nevertheless, these miRNAs were found predominantly in the neointimal layer of stenotic lesions. The anatomic correlation was consistent with that found in animal models. Furthermore, the expression of miRNAs not only was different between normal veins and stenotic fistulas, but also paralleled stenosis severity. This dose effect on expression provided additional support to the roles of miRNAs in neointima formation. In addition, previous animal studies showed that miR-221/222 and miR-130a could enhance neointimal growth after balloon injury, and miR-145 is also a modulator of smooth muscle cell function and suppresses neointima formation (9). Our study demonstrated a significant downregulation of miR-145 in neointimal tissues of dialysis accesses, which is in agreement with the findings of animal studies.

In this study, the function of the investigated miRNAs was not evaluated at cellular levels. A number of targets responsible for miR-21–mediated effects on smooth muscle cell proliferation have been proposed in previous studies. In animal models, PTEN and Bcl-2 have been shown to be involved in miR-21–mediated smooth muscle cell proliferation and apoptosis (11). miR-221 has been demonstrated to enhance smooth muscle cell proliferation via the mediation of PDGF-BB (10). miR-145 targets a network of transcriptional factors, including Elk-1, KLF-4, and myocardin, to promote differentiation and repress proliferation of smooth muscle cells (9). These targets provide evidence supporting the mechanistic link between the regulation of these miRNAs and neointima formation after balloon injury.

Because of the small sample size and cross-sectional design of our study, we attempted to extend our findings by examining circulating miRNAs in a prospective cohort. Although miRNA expression is typically tissue specific, it could be detected not only intracellularly but also in body fluids (29). Circulating miRNAs may come from passive release after cell lysis or active release by exosomes, microvesicles, and apoptotic bodies that were shed from disease cells (7). Although baseline circulating miRNAs did not correlate with stenosis severity, a significant upregulation of circulating miRNAs was observed after angioplasty. The change in miRNA levels was associated with restenosis rate at 6 months, even after correction for clinical and anatomic factors. The increased miRNAs were measured 2 days after angioplasty and therefore probably came mainly from a vascular response to intervention, rather than the preexisting miRNAs in the stenotic lesions. The disrupted endoluminal surface after intervention may facilitate the release of miRNAs into the circulation. Our findings suggest that the increase of circulating miRNAs may parallel their change in tissues and could be a surrogate of the proliferative activity in the disrupted lesions.

Some limitations should be addressed. First, the number of patients in the surgery cohort was small and our findings should be considered pilot in nature. Second, nonstenosed venous segment in dialysis vascular access could not be obtained for ethical reasons. Consequently, the control vein segments were harvested from patients who were predialysis and undergoing vascular access creation. Third, the venous segment in the nonsignificant stenosis group was obtained from infected or aneurysmal accesses. Fourth, the protein or gene expression profiles in tissues were not investigated. Further studies are needed to identify the mechanisms of the interaction between miRNAs and their targets. Fifth, the discrepancy in the degree of miRNA upregulation between in situ hybridization and quantitative real-time PCR methods was found. It may be attributable to the variations in contamination by normal venous tissues at harvesting stenotic lesions of different severity. Finally, this is a single-center study in the Han population and more evidence is needed before the generalization to all patient populations.

Our study revealed that stenotic venous segments of dialysis vascular accesses are characterized by a unique expression of miRNAs, which in turn suggests that the differentially expressed miRNAs may have relevance in the development of venous intimal hyperplasia. Some of these miRNAs could be detected in the circulation. In particular, an increase in miR-21 after angioplasty is independently associated with the development of restenosis. On the basis of these findings, miR-21 increase may serve as a biomarker to identify high-risk patients of postangioplasty restenosis. It will be helpful in therapeutic or surveillance planning, such as aggressive monitoring, miR-21 modulating intervention, or early surgical revision.

Disclosures

None.