Deregulation of hsa-miR-20b expression in TNF-α-induced premature senescence of human pulmonary microvascular endothelial cells

Microvascular Research 114 (2017) 26–33

Contents lists available at ScienceDirect

Microvascular Research journal homepage: www.elsevier.com/locate/ymvre

Deregulation of hsa-miR-20b expression in TNF-α-induced premature senescence of human pulmonary microvascular endothelial cells Pooi-Fong Wong a,⁎, Juliana Jamal a, Kind-Leng Tong a, Eng-Soon Khor a, Chia-Earn Yeap a, Hui-Lan Jong a, Sui-Ting Lee a, Mohd Rais Mustafa a, Sazaly Abubakar b a b

Department of Pharmacology, Faculty of Medicine, University of Malaya, Kuala Lumpur, Malaysia Department of Microbiology, Faculty of Medicine, University of Malaya, Kuala Lumpur, Malaysia

a r t i c l e

i n f o

Article history: Received 13 April 2017 Revised 1 June 2017 Accepted 3 June 2017 Available online 05 June 2017 Keywords: Premature senescence TNF-α hsa-miR-20b Microvascular endothelial dysfunction, retinoblastoma-like 1

a b s t r a c t miRNAs are important regulators of cellular senescence yet the extent of their involvement remains to be investigated. We sought to identify miRNAs that are involved in cytokine-induced premature senescence (CIPS) in endothelial cells. CIPS was established in young human pulmonary microvascular endothelial cells (HMVEC-Ls) following treatment with a sublethal dose (20 ng/ml) of tumor necrosis factor alpha (TNF-α) for 15 days. In parallel, HMVEC-Ls were grown and routinely passaged until the onset of replicative senescence (RS). Differential expression analysis following miRNA microarray profiling revealed an overlapped of eight deregulated miRNAs in both the miRNA profiles of RS and TNF-α-induced premature senescence cells. Amongst the deregulated miRNAs were members of the miR 17–92 cluster which are known regulators of angiogenesis. The role of hsamiR-20b in TNF-α-induced premature senescence, a paralog member of the miR 17–92 cluster, was further investigated. Biotin-labeled hsa-miR-20b captured the enriched transcripts of retinoblastoma-like 1 (RBL1), indicating that RBL1 is a target of hsa-miR-20b. Knockdown of hsa-miR-20b attenuated premature senescence in the TNF-α-treated HMVEC-Ls as evidenced by increased cell proliferation, increased RBL1 mRNA expression level but decreased protein expression of p16INK4a, a cellular senescence marker. These findings provide an early insight into the role of hsa-miR-20b in endothelial senescence. © 2017 Elsevier Inc. All rights reserved.

1. Introduction Cellular senescence is a state of irreversible growth arrest first described following the observation of cells losing their proliferative capacity after a finite number of population doublings (Hayflick, 1965; Hayflick and Moorhead, 1961). Many types of stressors can induce cellular senescence, including oxidative stress, DNA damage, activation of tumor suppressor gene, improper cell contacts or chronic exposure to the pro-inflammatory cytokines as well as telomere shortening from repeated cell division, which is also known as replicative senescence (RS) (Ben-Porath and Weinberg, 2005; Foreman and Tang, 2003; Toussaint et al., 2001). Cytokine-induced premature senescence (CIPS) has been associated with persistent DNA damage response and activation of Abbreviations: RS, replicative senescence; CIPS, cytokine-induced premature senescence; pRB, retinoblastoma protein; TNF-α, tumor necrosis factor alpha; SA-β-gal, senescence-associated beta-galactosidase; HMVEC-Ls, human pulmonary microvascular endothelial cells; PD, population doublings; RBL1, retinoblastoma-like 1; Bi, biotin; HUVECs, human umbilical vein endothelial cells; RTCA, Real Time Cell Analyzer; CI, cell index; SEM, standard error of the mean; ANOVA, one-way analysis of variance. ⁎ Corresponding author at: Department of Pharmacology, Faculty of Medicine, University of Malaya, 50603 Kuala Lumpur, Malaysia. E-mail address: [email protected] (P.-F. Wong).

http://dx.doi.org/10.1016/j.mvr.2017.06.002 0026-2862/© 2017 Elsevier Inc. All rights reserved.

p53-p21CIP1 and p16INK4a retinoblastoma protein (pRB) cell cycle regulatory pathways, which cause growth arrest and senescence (Sasaki et al., 2008). In addition, CIPS is often reinforced by a plethora of senescence-associated secretory phenotypes factors including tumor necrosis factor alpha (TNF-α), interleukins (IL-1α, IL-1β, IL-6), transforming growth factor beta (TGF-β) and chemokines such as vascular endothelial growth factor (VEGF) and matrix metalloproteinases (Kojima et al., 2013). The pulmonary vasculature is lined by a monolayer of endothelial cells with a total surface area of 90 m2 and it is critical for gas exchange in the lung. This monolayer of cells acts as a selective barrier by lying at the interface between blood, airway and lung parenchyma (Goldenberg and Kuebler, 2015). Premature senescence of pulmonary endothelial cells can induce endothelium dysfunction which could increase susceptibility to diseases such as pulmonary sepsis and pulmonary hypertension in elderly patients. Extensive studies have shown that aging pulmonary endothelium is more vulnerable to oxidative stress because of decreased antioxidant activities, reduced bioavailability of nitric oxide which is a vasorelaxant molecule (Rapoport et al., 1983) and insufficient pulmonary cellular repair and regeneration (Jane-Wit and Chun, 2012). In addition, endothelium dysfunction in lung has been reported to be one of the causes of exacerbation of lung inflammatory-

P.-F. Wong et al. / Microvascular Research 114 (2017) 26–33

related disease such as asthma (Mukhopadhyay et al., 2006; Suarez et al., 2010) and chronic obstructive pulmonary disease (COPD) (Green and Turner, 2017). In situ culture of lung tissue collected from COPD patients showed hallmarks of senescence such as decreased cell-population doublings, early replicative senescence, shorter telomeres, higher levels of p16 and p21 and increased secretion of inflammatory markers at early cell passage compared to those in control subjects (Amsellem et al., 2011). TNF-α is a pleiotropic pro-inflammatory cytokine of the TNF superfamily produced by various cell types and activated macrophages. TNFα exerts multiple biological effects depending on the distinct receptors expressed on the specific cell types (van Horssen et al., 2006). Essentially, overexpression of TNF-α activates intracellular signaling cascades of apoptosis, cell survival, inflammation, immunity and senescence. An earlier study showed that TNF-α-treated human diploid fibroblasts exhibited increased senescence-associated beta-galactosidase (SA-β-gal) activity and irreversible growth arrest which confirmed that TNF-α could induce premature senescence (Dumont et al., 2000). Several other studies further demonstrated that this inflammatory stressor could trigger premature senescence in endothelial cells (Dumont et al., 2000; Zhao et al., 2010; Zhou et al., 2002) by generating reaction oxygen species (Dumont et al., 2000) or by altering mitochondrial functions (Zhou et al., 2002). miRNAs are short (~22 nt), non-coding RNAs (Lee et al., 1993) which regulate gene expression by targeting the 3′-untranslated region (UTR) of mRNA transcripts, causing translation inhibition or degradation of mRNA. Unique alterations in miRNA expression were observed in various pathological conditions including aging-associated diseases (Frenzel et al., 2009). miRNAs have been implicated in modulating senescence in various cell types. For examples, miR-217 and miR-34a were shown to target Sirtuin1 while miR-146a is reported to target NADPH oxidase 4 in the modulation of senescence in endothelial cells (Ito et al., 2010; Menghini et al., 2009; Tabuchi et al., 2012; Vasa-Nicotera et al., 2011; Zhao et al., 2010). Despite extensive investigations onto the role of miRNAs in cellular senescence, the regulation of cellular senescence pathways modulated by miRNAs remains to be further defined. The present study sought to investigate the role of hsamiR-20b in premature senescence induced by chronic exposure to TNF-α. Findings from the present study would contribute new insights into the in-depth mechanisms governing endothelial health and dysfunction. 2. Materials and methods 2.1. Cell culture Human pulmonary microvascular endothelial cells (HMVEC-Ls; Lonza, MD, USA) were maintained in complete microvascular endothelial cell growth medium-2 at 37 °C in a humidified atmosphere of 95% air and 5% CO2. Population doublings (PD) was calculated as described (Cristofalo et al., 2000). Cells were seeded at a constant initial density of 5 × 105 cells per T75 flask and passaged using 0.25% trypsin EDTA every two to three days in order to maintain the cultures in constant log phase. The cells were used at passage 5 (young untreated cells), passage 12 (pre-senescent cells), or passage 19 (senescent cells), in subsequent experiments. HMVEC-Ls were used because they are more delicate in physical structure and thus, more susceptible to the induction of endothelial dysfunction.

27

2.3. Induction of cytokine-induced premature senescence (CIPS) by TNF-α treatment To establish CIPS model, young untreated cells of passage number b 5 were subjected to chronic TNF-α (Sigma-Aldrich, USA) treatment at various concentrations. The cells were seeded and incubated overnight prior to fifteen successive sub-lethal treatments with or without TNF-α at concentrations of 10 ng/ml, 20 ng/ml and 40 ng/ml in complete media. The media were replenished every day for fifteen consecutive days. 2.4. SA-β-galactosidase (SA-β-gal) pH 6.0 staining Senescent cells were detected by fluorescence-based SA-β-gal assay (Debacq-Chainiaux et al., 2009). Young and RS cells at passage b 5 and N20, respectively, were trypsinized and seeded at a sub-confluent level overnight and pre-treated with 50 nM Bafilomycin A1 for 1 h. Similarly, at the end of the TNF-α chronic treatment, cells were trypsinized and subjected to the same procedure prior to incubation with Bafilomycin A1. C12FDG, a fluorogenic substrate for β-galactosidase was then added to a final concentration of 33 μM and further incubated for 2 h. The cells were resuspended in ice-cold PBS with 5% FBS and acquired using FACSCanto II flow cytometer (Becton Dickinson, USA). The C12FDG-fluorescein signal was detected by FITC detector channel. The percentage of senescent positive cells was determined using FlowJo Version 10 (Tree Star, USA). 2.5. Cell cycle analysis Young, RS and CIPS (TNF-α-treated) cells were harvested and stained with propidium iodide using Cycletest plus DNA reagent kit (Becton Dickinson, CA, USA). DNA QC particles (Becton Dickinson) were used to verify the instrument performance and for quality control. 2.6. miRNA microarray Agilent human microRNAs array consisting of 866 human miRNAs and 89 human viral miRNAs, based on Sanger miRNA database release 12.0 was used. Three biological replicates of RS and CIPS were used for microarray analysis. Total RNAs (100 ng) were labeled with Cyanine 3-pCp and hybridized onto Human MicroRNA Array version 3.0 (8X15K) with 40–60 mer oligonucleotides. Microarray data were deposited in the National Center for Biotechnology Information Gene Expression Omnibus (GEO; http://www.ncbi.nlm.nih.gov/geo/) and are accessible through GEO series accession no. GSE45541. 2.7. Microarray data analysis miRNA microarray data was performed using GeneSpring GX version 11.5.1 (Agilent Technologies Inc., CA, USA). Threshold of raw signals were set at 1.0. The normalization algorithm employed was quantile and baseline transformation to median of all samples was performed. Differential miRNA profiles were generated by applying fold change N2.0 and corrected P-value b 0.05. Significance analysis was performed using Student's t-test followed by Benjamini-Hochberg multiple testing corrections. TargetScan program integrated into GeneSpring was used for the prediction of biological targets of the differentially expressed miRNAs. 2.8. qRT-PCR validation of miRNA and mRNA expression levels

2.2. Induction of replicative senescence (RS) in HMVEC-Ls RS cells were obtained by consecutively passaging the young untreated cells until cell proliferation ceased, approximately at passage 19 (PD ~20). Pre-senescent cells were achieved by routine sub-culturing of the young untreated cells until passage 12 (PD ~13).

miRNA and mRNA were harvested from cells by using miRNeasy Mini Kit (Qiagen, USA) and RNeasy Mini Kit (Qiagen, USA), respectively, following manufacturer's instructions. Total RNA yield was quantitated using NanoDrop 2000c (Thermo Scientific, USA). qRT-PCR was performed to quantitate the expression level of selected miRNAs using

28

P.-F. Wong et al. / Microvascular Research 114 (2017) 26–33

Taqman miRNA assay (Applied Biosystem, USA) with the following ID: hsa-miR-17 (assay ID: 002308), hsa-miR-20a (assay ID: 000580) and hsa-miR-20b (assay ID: 001014) while RNU6B (assay ID: 001093) was used as endogenous reference small RNA for normalization of data. For quantification of retinoblastoma-like 1 (RBL1) mRNA expression level, qRT-PCR was performed using RBL1 (assay ID: 4331182) and Ct values were normalized to those of the endogenous reference control, GAPDH (assay ID: 4332649). Fold-change analysis was performed using the 2−ΔCt method.

2.9. Confirmation of miRNA target by transfection with miRNA mimic and biotin pull-down assay Transfection with biotin (Bi)-labeled miRNA mimic was performed in a hardier endothelial cell line, human umbilical vein endothelial cells (HUVECs) using DharmaFECT 1 (Dharmacon, CO, USA). Cells (2.3 × 105 cells per well) were seeded overnight and then transfected with 100 nM Bi-miR-20b or negative control, C. elegans Bi-cel-miR-67 in six-well plates for 24 h. Biotin pull-down assay was performed as described by Lal et al. (2011). Cells were washed with PBS and resuspended in 0.1 ml lysis buffer containing 50 mM Tris-HCI (pH 8.8), 1% NP-40, 150 mM NaCI, 50 U RNaseOUT (Invitrogen, USA) with protease inhibitor (Roche Diagnostic, Germany) and incubated on ice for 5 min. The cytoplasmic lysate was isolated by centrifugation at 10,000 ×g for 10 min. Streptavidin-coated magnetic beads (Dynabeads® MyOne Streptavidin T1, Invitrogen, USA) at 50 μg beads for each sample were blocked for 2 h at 4 °C in 0.1 ml blocking buffer containing lysis buffer, 1 mg/ml BSA and 1 mg/ml yeast tRNA and washed twice with 0.5 ml lysis buffer. Cytoplasmic lysate was then incubated with the beads and incubated for 4 h at 4 °C. RNA bound to the beads (pull-down RNA) was isolated. For qRT-PCR, mRNA levels were normalized to those of the housekeeping gene HRPT1. The enrichment ratio of the control-normalized pulldown RNA to the control-normalized input levels was then calculated.

2.12. RBL1 gene knockdown study Young HMVEC-Ls (passage 7–9) were transfected with 50 nM of a pool of two RBL1 siRNAs targeting RBL1 transcript variant 1 (NM_002895) and transcript variant 2 (NM_183404) (RiboBio, China), or with the same concentration of non-targeting negative control (NTCsiRNA; Dharmacon, USA) using DharmaFECT #1 transfection reagent (Dharmacon, USA). To determine the efficacy of knockdown, transfected cells were harvested for total mRNA, 24 hour post-transfection to be examined for RBL1 mRNA expression level by qRT-PCR. In a separate experiment, the transfected cells were harvested 48 hour post-transfection for total protein extraction and the expression of p16INK4a was then examined by immunoblotting.

2.13. Immunoblotting Cells in 6-well plate were harvested using RIPA lysis buffer (Santa Cruz, USA) that was supplemented with 1 mM orthovanadate, 2 mM PMSF and 1% protease inhibitor. Fifteen micrograms of total protein lysate was separated using 12% SDS gel and then transferred to PVDF membranes prior to blotting with antibodies. Anti-p16 INK4a (#554070) and anti-beta actin (#sc-56459) were purchased from BD Biosciences (USA) and Santa Cruz Biotechnology (USA), respectively.

2.14. Statistical analysis All data are presented as means ± standard error of the mean (SEM). Statistical analysis was performed with Student's t-test or one-way analysis of variance (ANOVA), followed by Bonferroni's multiple comparison post-test using GraphPad Prism software (version 5.0; GraphPad Software Inc., CA, USA), where applicable. P-values b 0.05 were accepted as statistically significant. Statistical significance is expressed as ***, P b 0.001; **, P b 0.01; *, P b 0.05.

2.10. hsa-miR-20b knockdown study Pre-senescent HMVEC-Ls (passage 10–13) were transfected with 100 nM hsa-miR-20b hairpin inhibitor (Dharmacon, USA) for 24 h. Control cells were transfected with the same concentration of miRIDIAN miRNA hairpin inhibitor negative control #2 (Dharmacon, USA). Total RNA enriched with miRNAs was harvested from the transfected cells 24 h post-transfection. Efficacy of the knockdown was determined by comparing the mRNA expression levels of miR20b and RBL1 to those of hairpin inhibitor negative controltransfected cells. Seventy-two hour post-transfection, the transfected cells were harvested for total protein extraction and the expression of p16INK4a , a cellular senescence biomarker was then examined by immunoblotting.

2.11. Real time cell proliferation analysis Cell proliferation was monitored at every 15 min using xCELLigence Real Time Cell Analyzer (RTCA; ACEA Biosciences, USA). Increased cell index (CI) indicated increased in cell proliferation. Cell-free culture media (50 μl) was added into each well and background reading was recorded. Fifty microliter cell suspension containing young and RS cells at cell density of 2.5 × 103 and 3.5 × 103, respectively was added into each well. When cells reached logarithmic growth phase, cells were treated with TNF-α (20 ng/ml) for three consecutive days prior to transfection with hsa-miR-20b hairpin inhibitor for 24 h. Growth curves were normalized to the CI at the last measured time point right before TNF-α treatment. Normalized CI at each time point was analyzed using RTCA Software Version 1.2 algorithm.

3. Results 3.1. Prolonged treatment of young HMVEC-Ls with TNF-α induces premature senescence In the present study, young HMVEC-Ls were exposed to 10, 20 and 40 ng/ml of TNF-α at different duration of successive treatments. As control, young HMVEC-Ls of similar passage were grown in complete growth medium without any addition of TNF-α. The establishment of senescence was confirmed by SA-β-gal expression in the cells and cell cycle analysis. Staining with SA-β-gal, pH 6.0 revealed a significant increase (N52%) of SA-β-gal positive cells in young HMVEC-Ls treated with 10 and 20 ng/ml TNF-α for 15 consecutive treatments over a period of 15 days when compared to the young, untreated cells (Fig. 1A). In subsequent studies, TNF-α at 20 ng/ml was used to induce senescence in HMVEC-Ls. In parallel with an increase in the SA-β-gal expression, young HMVEC-Ls exposed to chronic TNF-α treatment became enlarged and flattened, which is a typical senescent phenotypic morphology, similar to those of the RS cells (Fig. 1B). There were 5.5 ± 1.7% SAβ-gal positive in untreated young cells whereas both RS and chronic TNF-α-treated young untreated cells demonstrated significantly higher SA-β-gal positive cells, which were ~49.4 ± 4.6% and 57.8 ± 10.0%, respectively (Fig. 1C). Furthermore, chronic TNF-α exposure resulted in progressive cell cycle arrest in HMVEC-Ls, similar to the RS cell cycle profile. The TNF-α-treated cell and RS cell cultures contained significantly higher percentages of G0/G1 cell population i.e. ~67% and ~71%, respectively, compared to the untreated young cell culture (~ 54%) (Fig. 1D).

P.-F. Wong et al. / Microvascular Research 114 (2017) 26–33

29

Fig. 1. TNF-α induces premature senescence in HMVEC-Ls. (A) Percentage of SA-β-gal positive cells after chronic TNF-α treatments at various concentrations. (B) Representative HMVECLs images depicting the cells morphology at young (b5 passages), RS (N20 passage) and at the end of the chronic TNF-α treatment. Blue arrows indicate an enlarged and flattened cell morphology. (C) Dot plot (C12FDG bright cells) and the percentage of SA-β-gal positive cells (~1.5 × 105 cells analyzed per group). (D) Cell cycle population in young, RS and TNF-αinduced premature senescent HMVEC-Ls. RS indicates replicative senescent cells. Data are mean ± SEM, n = 3. P-values computed using one-way ANOVA followed by Bonferroni's multiple comparison post-test. ***P b 0.001, **P b 0.01, *P b 0.05 compared to untreated or young control. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

3.2. miRNA expression profile Comparison of miRNA profiles between RS and young untreated cells revealed that 15 miRNAs were significantly differentially expressed, in which, seven miRNAs were found to be upregulated while eight miRNAs were downregulated in RS (Fig. 2A). Members of the miRNA 17–92 cluster including miR-17, -18a, -19b, -20a and -92a were notably downregulated. Sixty miRNAs were significantly differentially expressed in the miRNAs profile of cells treated chronically with 20 ng/ml TNF-α versus young untreated cells, with 34 upregulated and 26 downregulated miRNAs (Fig. 2B). miR-17, miR-18a, miR-19b, miR-20a and miR-92a were also downregulated in the TNF-α-treated cells. We further identified eight significantly deregulated miRNAs (both 5p and 3p arms of miR-20a, miR-126, miR-1226, miR-20a, miR20b, miR-92a, miR-17, miR-19b, miR-1308) common to both RS and

chronic TNF-α-treated cells (Fig. 2C). qRT-PCR results corroborated with microarray expression, whereby hsa-miR-17 and -20a were downregulated in RS and TNF-α-treated cells when compared to young untreated cells (Fig. 2D). However, hsa-miR-20b was instead found upregulated with qRT-PCR analysis and this discrepancy could be due to the inherent differences between the two methods in detecting miRNAs. For miRNAs that are highly homologous to other members in the same family, detection by PCR is reported to be more sensitive than conventional microarray hybridization (Git et al., 2010). 3.3. RBL1 as a target gene of hsa-miR-20b TargetScan analysis predicted RBL1 transcript variant 1 (NM_002895) as a target of hsa-miR-17, hsa-miR-20a and hsa-miR20b. Homology regions between hsa-miR-20b and 3′UTR of RBL1

30

P.-F. Wong et al. / Microvascular Research 114 (2017) 26–33

Fig. 2. miRNA expression profile of TNF-α-induced premature senescent HMVEC-Ls. Differential expression of miRNAs in (A) RS and (B) TNF-α-induced premature senescent HMVEC-Ls compared to young control cells. (C) Venn diagram of differentially expressed miRNAs. Fold change: negative values indicate down-regulation, positive values indicate up-regulation; log2 fold change cut-off: 2.0; P-value computation by Student's t-test; Multiple testing correction: Benjamini Hochberg False Discovery Rate; Corrected P-value cut-off: 0.05. *Indicates less predominantly expressed miRNA (from the opposite arm of the precursor). (D) qRT-PCR validation of selected miRNAs expression (~5.6 × 104 cells analyzed per group). a.u. indicates arbitrary units and RS indicates replicative senescent cells. Data are mean ± SEM, n = 3. P-values were computed using one-way ANOVA followed by Boferroni's multiple comparison post-test. ###P b 0.001, ##P b 0.01, #P b 0.05 compared to young control.

transcript variant 1 (NM_002895) and 2 (NM_183404) were identified by RNAhybrid, which start at position 1792 of RBL1 mRNA (Fig. 3A). To confirm whether RBL1 is indeed a target of hsa-miR-20b, we performed biotin pull-down assay following transfection with biotinylated hsamiR-20b. Bi-hsa-miR-20b pull-down mRNAs were significantly enriched (2-fold) with RBL1 transcripts compared to C. elegans Bi-celmiR-67 pull-down mRNAs (Fig. 3B). The RBL1 mRNA expression level was also stably downregulated in chronic TNF-α-treated cells after 15 days of consecutive treatment with 20 ng/ml TNF-α, similar to that of the replicative senescent cells, as compared to young untreated cells (Fig. 3C). To further confirm whether hsa-miR-20b modulates the expression of RBL1, cells were transfected with hsa-miR-20b hairpin inhibitor to knockdown the expression of hsa-miR-20b. Following transfection, the expression of hsa-miR-20b was downregulated by 50% (Fig. 3D. i), together with the upregulation of the mRNA expression of RBL1 (Fig. 3D. ii). These results suggests that chronic exposure of HMVEC-Ls to TNF-α is necessary to induce stable upregulation of hsamiR-20b, leading to the downregulation of its target gene, RBL1.

3.4. Knockdown of hsa-miR-20b improves cell proliferation and decreases p16INK4a protein expression level To study the effects of hsa-miR-20b in TNF-α-treated HMVEC-Ls growth profile, pre-senescent cells (passage 10–13) were treated with 20 ng/ml TNF-α for three consecutive days with daily replenishment of fresh medium followed by hsa-miR-20b knockdown. Cell growth was monitored using RTCA system. The hsa-miR-20b knocked-down cells showed significant increased in cell proliferation (P b 0.01, Student's t-test), as indicated by elevated normalized cell index, when compared to cells transfected with hairpin inhibitor negative control (Fig. 4A). To ascertain the role of hsa-miR-20b in the development of endothelial senescence phenotype, the effect of the knockdown of hsamiR-20b on a cellular senescence biomarker, p16INK4a was investigated. Following the knockdown of hsa-miR-20b, significant reduction of p16INK4a protein expression was observed (Fig. 4B) suggesting that the decreased in the senescence phenotype was partly associated with improved cell growth profile.

P.-F. Wong et al. / Microvascular Research 114 (2017) 26–33

31

4. Discussion

Fig. 3. Gene target of hsa-miR-20b. (A) Homology regions of hsa-miR-20b and 3′UTR of RBL1. (B) Enrichment of RBL1 transcripts in biotin-hsa-miR-20b (Bi-hsa-miR-20b) pulldown mRNA in HUVECs. (C) mRNA expressions of RBL1 in young, RS and TNF-αinduced premature senescent HMVEC-Ls (~5.6 × 104 cells analyzed per group). (D) (i) hsa-miR-20b and (ii) RBL1 mRNA expressions after TNF-α treatment followed by knockdown of hsa-miR-20b (~ 5.6 × 104 cells analyzed per group) in HMVEC-Ls. a.u. indicates arbitrary units; RS indicates replicative senescent cells. Data are mean + SEM, n = 3. P-values were computed using Student's t-test or one-way ANOVA followed by Bonferoni's multiple comparison post-test. ***P b 0.001, **P b 0.01, *P b 0.05 compared to young untreated cells or hairpin inhibitor negative controls.

3.5. Knockdown of RBL1 increases p16INK4a protein expression level As increased RBL1 mRNA expressions were observed in both RS and TNF-α-treated HMVEC-Ls, we further investigated the association of RBL1 expression and p16INK4a with cellular senescence. Briefly, RBL1 gene knockdown was performed in young HMVEC-Ls as the level of RBL1 was found to be higher in the young cells compared to RS (Fig. 3C). Following the RBL1 gene knockdown, a significant decrease of RBL1 mRNA expression level was observed in young HMVEC-Ls when compared to NTCsiRNA control (Fig. 4C. i). Concurrently, there was also an increase in the protein expression of p16INK4a in the young HMVEC-Ls (Fig. 4C. ii). This indicates that absence or lower expression of RBL1 could trigger endothelial cell senescence.

TNF-α engages various intracellular signaling pathways in promoting apoptosis, cell survival and inflammation (Bradley, 2008). Previous studies have established premature senescence models by sublethal treatment with TNF-α across various different cells. Beyne-Rauzy et al. induced senescence in leukemic KG1 cells by 20 ng/ml TNF-α treatment for 15 days whereas premature senescence in HUVECs was successfully established by 10 ng/ml TNF-α treatment for 10 days (Beyne-Rauzy et al., 2004; Khan et al., 2017). Here, we showed that similar chronic exposure to sublethal concentrations of TNF-α could also induce premature senescence in human microvascular endothelial cells. Identification of deregulated miRNAs common to both TNF-α-induced premature senescence and RS is of interest as they could be miRNAs that potentially modulate premature senescence via intracellular pathways engaged by TNF-α. Members of the miR 17–92 cluster are known to regulate pathways of cellular senescence and aging (Grillari et al., 2010), whereby miR-17, -19b, -20a and 106a are down-regulated in aged human tissues (Hackl et al., 2010). Our results also concurred with these earlier studies whereby hsa-miR-17 and -20a, in particular, were down-regulated in both RS and TNF-α-treated cells. Interestingly, hsamiR-20b which shares near identical sequences with hsa-miR-17 and hsa-miR-20a was instead found up-regulated in cells chronically treated with TNF-α and in RS cells. hsa-miR-20b belongs to the miR-106a-363 cluster in chromosome X, a paralogous cluster of miR-17-92 and miR106b-25 clusters of the large miR-17 family (Tanzer and Stadler, 2004). It promotes tumor growth and survival (Lei et al., 2009; Li et al., 2013) in various cancers (Li et al., 2012; Xue et al., 2015) and given that it is also an anti-angiomir (Borges et al., 2016), its unexplored effect on endothelial health was investigated here. Chronic TNF-α treatment (up to 15 days) was required to sustain hsa-miR-20b up-regulation, consistent with the observation of senescence phenotype at day 15. However, earlier studies have shown that hsa-miR-20b was downregulated in human dermal fibroblasts and trabecular meshwork cells in H2O2-induced senescence (Li et al., 2009) and UVB-induced senescence in human fibroblasts (Greussing et al., 2013). These discrepancies could be attributed to the use of different cell models and stressors to induce premature senescence. Moreover, McCall et al. (2011) have shown that miRNAs expression could differed significantly across different endothelial cell types, especially for hsa-miR-20b where its expression was found highest in HUVEC, moderate in human pulmonary microendothelial cells and lower in human pulmonary artery endothelial cells (McCall et al., 2011). In addition, TNF-α also engages various signaling pathways that could have modulated the expression of hsamiR-20b differently. Various senescence-inducing stressors engage p53 or pRB tumor suppressor pathways to halt cell proliferation during premature cellular senescence (Campisi and d'Adda di Fagagna, 2007). RBL1 (p107), a member of the small RB family, was identified as one of the gene targets of hsa-miR-20b. Increased mRNA expression of RBL1 with the knockdown of hsa-miR-20b further supports the modulatory role of hsamiR-20b on RBL1. Knockdown of hsa-miR-20b has not only increased cell growth but has also resulted in a decrease in the expression of p16INK4a, a cellular senescence marker. Moreover, the lower expression of RBL1 as a consequence of increased hsa-miR-20b expression also resulted in an increase in the expression of p16INK4a, as demonstrated in the RBL1 siRNA knockdown study. These strongly suggest that both hsa-miR-20b and RBL1 are involved in the modulation of endothelial cell proliferation and the development of the senescence phenotype. The involvement of RBL1 in premature cellular senescence is plausible as RBL1 is a homolog of RB protein, a key regulator of entry into cell division that acts by repressing E2F activity during cell cycle progression. It is shown that RBL1 is required for irradiation-induced premature senescence when there is a loss of pRB function (Lehmann et al., 2008). Similarly, RBL2 (p130), another member of RB family, is also involved in senescence induction where it is recruited to the cell cycle regulatory

32

P.-F. Wong et al. / Microvascular Research 114 (2017) 26–33

Fig. 4. Effects of hsa-miR-20b knockdown and RBL1 knockdown on the cell growth and protein expression of p16INK4a in HMVEC-Ls. (A) Real-time growth profile of cells transfected with hsa-miR-20b hairpin inhibitor (~3000 cells analyzed per group). (B) Protein expressions of p16INK4a after knockdown of hsa-miR-20b is shown as fold-change and in a representative blot (~5.6 × 104 cells analyzed per group). (C) (i) RBL1 mRNA expression after the knockdown of RBL1 (~5.6 × 104 cells analyzed per group). (ii) Protein expression of p16INK4a protein is presented as fold change and in a representative blot after transfection with RBL1 siRNA (~5.6 × 104 cells analyzed per group). a.u. indicates arbitrary units. Data are mean ± SEM, n = 3. P-values were computed using Student's t-test. ***P b 0.001, **P b 0.01, *P b 0.05 compared to hairpin inhibitor negative control or NTCsiRNA-transfected cells.

genes and results in senescent-like phenotypes, as demonstrated in doxorubicin-treated breast cancer cells (Jackson and Pereira-Smith, 2006). They further showed that RBL1 compensates for the loss of RBL2 introduced by RBL2 knockdown by increasing its own expression level at the cell cycle promoter genes. In summary, our findings showed that chronic exposure to TNFα causes premature senescence of endothelial cells with the involvement of hsa-miR-20b and its target gene, RBL1. Endothelial cell proliferation and senescence could be mediated at least in part via hsa-miR-20b and RBL1 interaction but this will require further studies to delineate the exact pathway involved as well as in vivo investigation of their functions. Findings from our study thus add to the current limited information on the biological role of hsa-miR-20b and RBL1 and warrant further investigations on how this miRNA acts in tandem with other miRNAs in endothelial senescence.

Competing interests The authors have no competing interests associated with the manuscript. Funding information This study was supported by Fundamental research Grant Scheme (FP042/2010A), University Malaya Research Grant (RG097/09HTM) and Postgraduate Research Fund (PV044/2011B). Author contribution statement WPF participated in experimental design, data interpretation and manuscript preparation. YCE, JJ, JHL, LST, TKL and KES performed experiments. MRM and SAB are advisors of the project and provided critical

P.-F. Wong et al. / Microvascular Research 114 (2017) 26–33

review of the manuscript. All authors have read and approved the final manuscript. References Amsellem, V., et al., 2011. Telomere dysfunction causes sustained inflammation in chronic obstructive pulmonary disease. Am. J. Respir. Crit. Care Med. 184, 1358–1366. Ben-Porath, I., Weinberg, R.A., 2005. The signals and pathways activating cellular senescence. Int. J. Biochem. Cell Biol. 37, 961–976. Beyne-Rauzy, O., et al., 2004. Tumor necrosis factor alpha induces senescence and chromosomal instability in human leukemic cells. Oncogene 23, 7507–7516. Borges, N.M., et al., 2016. Angiomirs expression profiling in diffuse large B-cell lymphoma. Oncotarget 7, 4806–4816. Bradley, J.R., 2008. TNF-mediated inflammatory disease. J. Pathol. 214, 149–160. Campisi, J., d'Adda di Fagagna, F., 2007. Cellular senescence: when bad things happen to good cells. Nat. Rev. Mol. Cell Biol. 8, 729–740. Cristofalo, V.J., et al., 2000. Use of the fibroblast model in the study of cellular senescence. In: Barnett, Y.A., Barnett, C.R. (Eds.), Aging Methods and Protocols. Humana Press, Totowa, NJ, pp. 23–52. Debacq-Chainiaux, F., et al., 2009. Protocols to detect senescence-associated beta-galactosidase (SA-betagal) activity, a biomarker of senescent cells in culture and in vivo. Nat. Protoc. 4, 1798–1806. Dumont, P., et al., 2000. Appearance of biomarkers of in vitro ageing after successive stimulation of WI-38 fibroblasts with IL-1alpha and TNF-alpha: senescence associated beta-galactosidase activity and morphotype transition. J. Anat. 197 (Pt. 4), 529–537. Foreman, K.E., Tang, J., 2003. Molecular mechanisms of replicative senescence in endothelial cells. Exp. Gerontol. 38, 1251–1257. Frenzel, L., et al., 2009. A new mode of TNF-[alpha] inhibition by microRNA. J. Immunol. 182 (Meeting Abstract Supplement). Git, A., et al., 2010. Systematic comparison of microarray profiling, real-time PCR, and next-generation sequencing technologies for measuring differential microRNA expression. RNA 16, 991–1006. Goldenberg, N.M., Kuebler, W.M., 2015. Endothelial cell regulation of pulmonary vascular tone, inflammation, and coagulation. Compr. Physiol. 5, 531–559. Green, C.E., Turner, A.M., 2017. The role of the endothelium in asthma and chronic obstructive pulmonary disease (COPD). Respir. Res. 18, 20. Greussing, R., et al., 2013. Identification of microRNA-mRNA functional interactions in UVB-induced senescence of human diploid fibroblasts. BMC Genomics 14, 224. Grillari, J., et al., 2010. miR-17-92 cluster: ups and downs in cancer and aging. Biogerontology 11, 501–506. Hackl, M., et al., 2010. miR-17, miR-19b, miR-20a, and miR-106a are down-regulated in human aging. Aging Cell 9, 291–296. Hayflick, L., 1965. The limited in vitro lifetime of human diploid cell strains. Exp. Cell Res. 37, 614–636. Hayflick, L., Moorhead, P.S., 1961. The serial cultivation of human diploid cell strains. Exp. Cell Res. 25, 585–621. van Horssen, R., et al., 2006. TNF-alpha in cancer treatment: molecular insights, antitumor effects, and clinical utility. Oncologist 11, 397–408. Ito, T., et al., 2010. MicroRNA-34a regulation of endothelial senescence. Biochem. Biophys. Res. Commun. 398, 735–740. Jackson, J.G., Pereira-Smith, O.M., 2006. Primary and compensatory roles for RB family members at cell cycle gene promoters that are deacetylated and downregulated in doxorubicin-induced senescence of breast cancer cells. Mol. Cell. Biol. 26, 2501–2510.

33

Jane-Wit, D., Chun, H.J., 2012. Mechanisms of dysfunction in senescent pulmonary endothelium. J. Gerontol. A Biol. Sci. Med. Sci. 67, 236–241. Khan, S.Y., et al., 2017. Premature senescence of endothelial cells upon chronic exposure to TNFα can be prevented by N-acetylcysteine and plumericin. Sci. Rep. 7, 39501. Kojima, H., et al., 2013. IL-6-STAT3 signaling and premature senescence. JAK-STAT. 2, p. e25763. Lal, A., et al., 2011. Capture of microRNA–bound mRNAs identifies the tumor suppressor miR-34a as a regulator of growth factor signaling. PLoS Genet. 7, e1002363. Lee, R.C., et al., 1993. The C. elegans heterochronic gene lin-4 encodes small RNAs with antisense complementarity to lin-14. Cell 75, 843–854. Lehmann, B.D., et al., 2008. Distinct roles for p107 and p130 in Rb-independent cellular senescence. Cell Cycle 7, 1262–1268. Lei, Z., et al., 2009. Regulation of HIF-1alpha and VEGF by miR-20b tunes tumor cells to adapt to the alteration of oxygen concentration. PLoS One 4, e7629. Li, G., et al., 2009. Alterations in microRNA expression in stress-induced cellular senescence. Mech. Ageing Dev. 130, 731–741. Li, J.Y., et al., 2012. Differential distribution of miR-20a and miR-20b may underlie metastatic heterogeneity of breast cancers. Asian Pac. J. Cancer Prev. 13, 1901–1906. Li, D., et al., 2013. Crucial role for early growth response-1 in the transcriptional regulation of miR-20b in breast cancer. Oncotarget 4, 1373–1387. McCall, M.N., et al., 2011. MicroRNA profiling of diverse endothelial cell types. BMC Med. Genet. 4, 78. Menghini, R., et al., 2009. MicroRNA 217 modulates endothelial cell senescence via silent information regulator 1. Circulation 120, 1524–1532. Mukhopadhyay, S., et al., 2006. Role of TNFalpha in pulmonary pathophysiology. Respir. Res. 7, 125. Rapoport, R.M., et al., 1983. Endothelium-dependent relaxation in rat aorta may be mediated through cyclic GMP-dependent protein phosphorylation. Nature 306, 174–176. Sasaki, M., et al., 2008. Proinflammatory cytokine-induced cellular senescence of biliary epithelial cells is mediated via oxidative stress and activation of ATM pathway: a culture study. Free Radic. Res. 42, 625–632. Suarez, Y., et al., 2010. Cutting edge: TNF-induced microRNAs regulate TNF-induced expression of E-selectin and intercellular adhesion molecule-1 on human endothelial cells: feedback control of inflammation. J. Immunol. 184, 21–25. Tabuchi, T., et al., 2012. MicroRNA-34a regulates the longevity-associated protein SIRT1 in coronary artery disease: effect of statins on SIRT1 and microRNA-34a expression. Clin. Sci. (Lond.) 123, 161–171. Tanzer, A., Stadler, P.F., 2004. Molecular evolution of a microRNA cluster. J. Mol. Biol. 339, 327–335. Toussaint, O., et al., 2001. Stress-induced Premature Senescence (SIPS). eLS. John Wiley & Sons, Ltd. Vasa-Nicotera, M., et al., 2011. miR-146a is modulated in human endothelial cell with aging. Atherosclerosis 217, 326–330. Xue, T.M., et al., 2015. miR-20b overexpression is predictive of poor prognosis in gastric cancer. Onco. Targets. Ther. 8, 1871–1876. Zhao, T., et al., 2010. MicroRNA-34a induces endothelial progenitor cell senescence and impedes its angiogenesis via suppressing silent information regulator 1. Am. J. Physiol. Endocrinol. Metab. 299, E110–E116. Zhou, J.J., et al., 2002. Human endothelial senescence can be induced by TNF-alpha. Chin. Sci. Bull. 47, 118–122.