Human cytomegalovirus promoting endothelial cell proliferation by targeting regulator of G-protein signaling 5 hypermethylation and downregulation

Interactions between human cytomegalovirus (HCMV) infection and environmental factors can increase susceptibility to essential hypertension (EH). Although endothelial dysfunction is the initial factor of EH, the epigenetic mechanisms through which HCMV infection induces endothelial cell dysfunction are poorly understood. Here, we evaluated whether HCMV regulated endothelial cell function and assessed the underlying mechanisms. Microarray analysis in human umbilical vein endothelial cells (HUVECs) treated with HCMV AD169 strain in the presence of hyperglycemia and hyperlipidemia revealed differential expression of genes involved in hypertension. Further analyses validated that the regulator of G-protein signaling 5 (RGS5) gene was downregulated in infected HUVECs and showed that HCMV infection promoted HUVEC proliferation, whereas hyperglycemia and hyperlipidemia inhibited HUVEC proliferation. Additionally, treatment with decitabine (DAC) and RGS5 reversed the effects of HCMV infection on HUVEC proliferation, but not triggered by hyperglycemia and hyperlipidemia. In summary, upregulation of RGS5 may be a promising treatment for preventing HCMV-induced hypertension.

GO (gene oncology) enrichment analysis of DEGs (differentially expressed genes) and enrichment analysis of pathways. To elucidate endothelial dysfunction-linked biological processes affected by HCMV infection, a GO functional enrichment analysis was conducted on DEGs. The majority of GO terms were related to biological processes, e.g., protein and nucleic acid metabolism, DNA replication, and regulation of oxidative stress ( Fig. 2A). The pathway annotation of differentially expressed genes was performed and the obtained DEGs were all involved in pathway terms using the KEGG database. The DEGs were mainly enriched in pathways, for instance, cell apoptosis, cell cycle, DNA duplication, adhesion factors, and actin cytoskeleton regulation (Fig. 2B).
Two major objectives of the present research were as follows: 1) to investigate the correlation degree between disease and genes, and 2) to clarify genes' role in disease. Thus, according to disease enrichment obtained from our results, the most obvious diseases were colorectal cancer, vascular disease, and cardiac hypertrophy. Therefore, five different databases were selected to evaluate the enrichment of EH (Fig. 2C). The data indicated that several DEGs, including prostaglandin I2 (prostacyclin) synthase (PTGIS), RGS5, selectin E (SELE), endothelin converting enzyme (ECE), and ATPase, Na + /K + transporting, and beta 1 polypeptide (ATP1B1), were highly associated with hypertensive disease (Fig. 2D-G and Table 2). It has been indicated that HCMV infection may be related to hypertension in the Kazakh Chinese population in Xinjiang province (China) through serum angiotensin-converting enzyme (sACE) hypomethylation and 11-d hydroxysteroid dehydrogenase 2(HSD11β2) hypermethylation 30 . Thus, we attempted to further identify the seven candidate genes related to hypertension.
Validation of microarray results using Rt-qpcR(quantitative reverse transcription polymerase chain reaction). To further evaluate significance of the seven identified DEGs, we examined mRNA expressions using RT-qPCR. The mRNA expressions of candidate genes among different groups are revealed by Fig. 3. It was demonstrated that ATP1B1 expressions in uninfected HUVECs treated by HG and ox-LDL were higher than those in the control group, while those were lower than those in HCMV-infected HUVECs treated by HG and ox-LDL (Fig. 3E). Importantly, no significant differences in the expression levels of ACE, ECE, HSD, SELE, and PTGIS were observed in RT-qPCR analysis ( Fig. 3A-D,F). The RGS5 gene expression level was reduced in the HCMV-infected group and in the uninfected group treated by HG and ox-LDL compared to the control; the decreased expression level of RGS5 was more substantial in HCMV-infected HUVECs treated by HG and ox-LDL than that of the other two experimental groups (Fig. 3G). These results above were in agreement with those obtained from the microarray chip analysis. Therefore, RGS5 was selected as a target gene to investigate its role in the function of ECs, involving the interaction of HCMV infection with environmental risk factors.
HcMV infection promoted ec proliferation. Next, CCK-8 assay was undertaken to determine the role of HCMV infection in HUVECs proliferation. As illustrated in Fig. 4A, HCMV infection significantly increased the proliferation of HUVECs compared with the control (P < 0.05). In contrast, the proliferation of HUVECs treated by HG and ox-LDL was reduced (P < 0.05). Moreover, for HCMV-infected HUVECs treated by HG and ox-LDL, the proliferation of ECs was further decreased compared with the HCMV-infected group or the HG and ox-LDL-treated group (P < 0.05). The facts above indicated that HCMV infection promoted the proliferation of ECs, while proliferation of ECs was inhibited by HG and ox-LDL treatment, which enhanced EC proliferation caused by HCMV infection, and that could be blocked by HG and ox-LDL treatment.

HCMV infection significantly increased activity of DNMT in ECs.
To assess the underlying mechanisms, in which the interaction between HCMV infection and environmental risk factors influenced endothelial dysfunction, we initially tested the activity of DNMT in the four treatment groups. The results showed that the activity of DNMT in HCMV-infected or HG and ox-LDL-treated HUVECs was increased compared to the control (P < 0.05). Additionally, the activity of DNMT was decreased in HCMV-infected HUVECs treated by HG and ox-LDL compared with the other two experimental groups (P < 0.05; Fig. 4B). www.nature.com/scientificreports www.nature.com/scientificreports/ increased in both HCMV-infected HUVECs and HUVECs treated by HG and ox-LDL compared to the control (P < 0.05). Additionally, the increase in RGS5 methylation in HCMV-infected HUVECs treated by HG and ox-LDL was further amplified compared with HCMV-infected or HG and ox-LDL-treated cells (Fig. 4C). www.nature.com/scientificreports www.nature.com/scientificreports/ To further indicate whether DNA hypermethylation was associated with RGS5 downregulation, HUVECs were incubated in the presence of decitabine (DAC). It was disclosed that RGS5 mRNA expressions in HCMV-infected or HG and ox-LDL-treated HUVECs were lower compared to the control (P < 0.05). The combined effects of the two groups further decreased RGS5 mRNA expressions (P < 0.05), which is consistent with the above-mentioned results of RGS5 methylation, indicating that HCMV infection and HG and ox-LDL treatment could synergistically reduce RGS5 mRNA expressions. Moreover, the RGS5 mRNA expression in HCMV-infected or HG and ox-LDL-treated HUVECs was increased upon the addition of DAC (P < 0.05; Fig. 4D; the four sets of DAC-free data were identical to HCMV infection-induced EC proliferation was regulated by DNA methylation. In the present research, we focused on the effects of DNA methylation on the regulation of EC dysfunction in the context of interactions between HCMV infection and environmental risk factors using CCK-8 assay. Upon treatment with DAC, the enhanced proliferation of ECs induced by HCMV infection was reversed, whereas HG and ox-LDL treatment did not affect the inhibition of proliferation of ECs (P < 0.05), as depicted by Fig. 4E (the four sets of DAC-free data were identical to Fig. 4A so as to more accurately compare effects of DAC treatment on proliferation of ECs). The expression level of PCNA in HCMV-infected cells was significantly higher (P < 0.05), whereas that level in HUVECs treated with HG and ox-LDL was notably lower compared to the control group (P < 0.05). Additionally, the expression level of PCNA was remarkably lower in HCMV-infected HUVECs treated with HG and ox-LDL than in uninfected HUVECs treated with HG and ox-LDL (P < 0.05). These findings indicated that HCMV infection promoted the expression level of PCNA, whereas HG and ox-LDL inhibited the expression level of PCNA. Notably, the expression level of PCNA was lower in DAC-treated group than that in control group (P < 0.05), iindicating that DNMT inhibitors could reverse the expression level of PCNA. Moreover, the expression level of PCNA in HCMV-infected HUVECs treated with DAC was lower than that in untreated HCMV-infected HUVECs (P < 0.05; Fig. 4F,G). Thus, it could be concluded that DNMT inhibitors could reverse the high expression of PCNA induced by HCMV infection. As previously described, our findings demonstrated that HCMV infection facilitated proliferation of ECs by regulating the RGS5 expression level through methylation.

Overexpression of RGS5 reversed EC proliferation triggered by HCMV infection.
We also assessed the effects of overexpression of RGS5 on proliferation of ECs. Green fluorescence was observed in both the RGS5-transfected and GFP-positive control groups, while that was not observed in the untransfected group ( Fig. 5A), demonstrating that the adenovirus transfection was successful, with a transfection rate of more than 90%. Additionally, RT-qPCR analysis further confirmed the successful construction and transfection of the RGS5-overexpression vector (P < 0.05; Fig. 5B).
Cell proliferation was measured using CCK-8 assay, and it was uncovered that proliferation of ECs in the RGS5-overexpression group was lower than that in the vehicle group (P < 0.05). Notably, proliferation of ECs in HCMV-infected cells overexpressing RGS5 was lower compared to HCMV-infected cells without overexpression of RGS5 (P < 0.05; Fig. 5C). To further elucidate the effects of RGS5 on proliferation of ECs induced by HCMV infection, the expression level of PCNA was measured after overexpression of RGS5. As depicted in Fig. 5D,E, overexpression of RGS5 decreased the expression level of PCNA protein, suggesting that proliferation of ECs could be induced by HCMV infection via hypermethylation and downregulation of RGS5.

Discussion
In our previous research, it was indicated that HCMV infection was associated with hypertension in Kazakh and Han Chinese populations, and interactions with environmental factors, e.g. HG and high fat, could increase susceptibility to hypertension 11 . Our study first found that RGS5, a negative regulator of G protein-coupled receptors (GPCRs), was a target gene for dysfunction of ECs caused by HCMV infection, while that was not triggered by HG or ox-LDL. Furthermore, it was demonstrated that HCMV infection downregulated RGS5 by DNA hypermethylation, resulting in increased proliferation of ECs. The evidence can be described as follows: First, HCMV  www.nature.com/scientificreports www.nature.com/scientificreports/ infection or hyperglycemia and hyperlipidemia have the same downregulation effect on RGS5 caused by DNA hypermethylation, and they have synergistic influence, while opposite effects on the proliferation of ECs were noted. Second, methyltransferase inhibitor could only reverse the enhanced proliferation of ECs mediated by HCMV infection, whereas accelerate the inhibitory influences of hyperglycemia and hyperlipidemia on the proliferation of ECs. In addition, the overexpression of RGS5 reversed the regulatory effect of HCMV on proliferation of ECs, while did not affect the regulation of HG and high fat on proliferation of ECs.
In the current research, it was revealed that there were several DEGs in the HCMV-infected group compared to the control. Among these DEGs, more than 2000 genes were upregulated, and 3275 genes were downregulated. The FBJ murine osteosarcoma viral oncogene homolog gene showed the highest upregulation with a FC of up to 57, indicating that the mentioned gene was remarkably influenced by HCMV infection. However, HUVECs pretreated with HG and ox-LDL showed less induction of DEGs, with 371 upregulated genes and 545 downregulated genes, suggesting that the transcriptional features of cells were altered under different environmental conditions. When the relationships between DEGs in the HCMV-infected group and HG-and ox-LDL-treated group were compared, we found that the DEAD box helicase 3, Y-linked gene was the main DEG with a FC of almost 100. This gene was found to be involved in ATPase metabolism, intracellular electron conversion, and other biological processes. Thus, these data indicated that ECs induced innate immune responses to prevent infection by viruses. Additionally, HCMV infection resulted in substantial numbers of DGEs, and the majority of them were downregulated. This indicated that the reproductive mechanism of viruses, e.g. by borrowing the replication machinery of the host cell to force replication of the genetic material of the virus itself, led to the downregulation of the majority of DEGs in the ECs.
In enrichment analysis, DEGs were mainly involved in apoptosis, cell cycle, DNA replication, adhesion factors, and actin cytoskeleton regulation. According to the results of GO and functional enrichment analysis, the DEGs induced by infection were mainly enriched in protein metabolism, nucleic acid metabolism, and regulation of oxidative stress. These results demonstrated that HCMV infection induced oxidative stress in ECs, leading to tissue damage. A previous study uncovered that HCMV infection increased the production of reactive oxygen species (ROS) in ECs, leading to cellular oxidative stress. Upon generation of ROS, inflammatory cell infiltration could be increased by enhancing cell permeability, thereby influencing cell proliferation, activating multiple signal transduction pathways, mediating migration and differentiation of monocytes and secretion of cytokines by macrophages, as well as promoting the occurrence and development of hypertension 32 . This is consistent with our previous results that patients with EH, particularly those with HCMV infection, exhibited a remarkable increase in 8-hydroxy-2-deoxyguanosine levels, which may have contributed to development of EH in the Kazakh Chinese population 30 . Further screening of DEGs based on disease enrichment and previous studies showed that only the expression level of RGS5 was altered in RT-qPCR analysis. Thus, we concentrated on this gene in the subsequent analyses.
Moreover, RGS5 is an important modulator of signal transduction in the cardiovascular system and is highly expressed in the aorta, heart, skeletal muscle in smooth muscle cells (SMCs), and ECs in cardiovascular tissues [33][34][35] . Moreover, RGS5 plays substantial roles in remodeling blood vessels and regulating blood pressure [36][37][38] . A study demonstrated that RGS5 may be a candidate target gene for the treatment of human and animal www.nature.com/scientificreports www.nature.com/scientificreports/ hypertension. Indeed, RGS5 genetic polymorphisms have been shown to be associated with human hypertension 39 . Additionally, the expression level of RGS5 in the vascular smooth muscle and intima was markedly reduced in spontaneously hypertensive rats and adrenocorticotropic hormone-and Ang-II-induced hypertensive rats, while the blood pressure level was increased in RGS5-knockout mice 37,40,41 . Moreover, RGS5 might also be involved in the regulation of blood pressure by modulating Na + transport in renal epithelial cells, G renal epit AngII type 1 receptor signaling, protein kinase C, mitogen-activated protein kinase (MAPK)/extracellular signal-regulated kinase (ERK), RhoA kinase activation-mediated increases in vascular resistance and vascular remodeling, and downregulation of peroxisome proliferator-activated receptor β-induced oxidative stress and inflammatory responses 37,[42][43][44] . However, it is still elusive whether RGS5 can induce dysfunction of ECs in the context of HCMV infection. It has been reported that HCMV infection promotes proliferation of ECs 45 . Similarly, our study found that HCMV infection stimulated the proliferation of ECs, whereas HG and ox-LDL inhibited proliferation of ECs. In addition, the interactions between these factors further reduced proliferation of ECs. ECs grown under hyperglycemic conditions have been shown to exhibit decreased proliferation and fibrinolytic potential, in addition to increased apoptosis 46,47 . Additionally, a high concentration of ox-LDL decreases migration and proliferation of ECs 48 . Li et al. found that hyperglycemia and hyperlipidemia enhanced the expression levels of ox-LDL receptors (e.g., LOX-1) in ECs [49][50][51] ; this can induce apoptosis, suggesting that the effects of HG and lipids on inhibiting proliferation of ECs may be related to the induction of apoptosis of ECs.
It is noteworthy that HCMV infection induces development of a series of cardiovascular diseases 52 , and ECs can serve as host cells for HCMV infection 53 . However, the molecular mechanisms, in which HCMV infection regulates proliferation of ECs, leading to endothelial dysfunction and causing vascular diseases, have remained www.nature.com/scientificreports www.nature.com/scientificreports/ obscure yet. Our study indicated that the activity of DNMT and DNA methylation levels were increased in ECs after HCMV infection or HG and ox-LDL treatment. Importantly, global DNA hypermethylation and gene-specific methylation have been implicated in cardiovascular diseases 54,55 . Besides, HCMV infection has been reported to inhibit DNA methylation in host cells, which in turn affects cell susceptibility to HCMV infection 31 . HCMV infection decreases the methylation of the human aquaporin-1 gene in the submandibular gland of mice, resulting in the occurrence of salivary gland inflammation 56 . As reported previously, HCMV infection induces sACE hypomethylation and HSD11β-2 hypermethylation in peripheral blood mononuclear cells, which could be involved in the occurrence of EH in the Xinjiang Kazakh population 57 . Notably, in the present research, we found that DAC reversed the low expression level of RGS5 caused by HCMV infection. Furthermore, DAC reversed proliferation of ECs induced by HCMV infection, while further aggravated the inhibition of ECs by hyperglycemia and hyperlipidemia. Based on these findings, we suggest that HMCV infection may induce proliferation of ECs by reducing the expression level of RGS5 through DNA methylation.
Animal experiments have confirmed that mRNA levels of RGS5 decrease in vascular SMCs in atherosclerotic lesions 58 , and loss of RGS5 results in profound hypertension 37 . Downregulation of RGS5 has been identified in bypass graft neo-intima, atherosclerotic arteries, and hypertension 42,59,60 . Consistent with these findings, we found that RGS5 overexpression reversed the proliferation of ECs induced by HCMV infection, whereas did not influence the effects of HG or ox-LDL on proliferation of ECs. Several previous studies reported that ox-LDL suppressed cell proliferation though inhibiting expression levels of basic fibroblast growth factor-5 (bFGF-5) or www.nature.com/scientificreports www.nature.com/scientificreports/ nuclear translocation of cell-cycle proteins 61 . Ox-LDL also suppresses vascular endothelial growth factor-induced EC migration via its inhibitory effect on the AKT/endothelial nitric oxide synthase pathway 62 . In addition, HG regulates bFGF-2 to influence proliferation of ECs through changes in permeability of ECs 63 . The mechanisms, in which hyperglycemia and hyperlipidemia induce proliferation of ECs, are complex and may involve various metabolic abnormalities; however, a low expression level of RGS5 owing to hypermethylation may not be one of these mechanisms. Overall, our data demonstrated that HCMV infection could reduce the expression level of RGS5 by promoting hypermethylation, thereby increasing the proliferation of ECs.
The current research contains several limitations. First, further studies need to be conducted to verify our in vitro results. Indeed, because HCMV infection is a complicated process involving the interactions of virus, host, and host cell factor-1, further analyses are essential. Additionally, to facilitate the application of these findings to human disease, further studies in human and animal models are required, and the exact mechanisms and possible therapeutic applications should be meticulously evaluated. Second, we are entering a new era of understanding how genomes interact with the environment to influence the pathogenesis of diseases. A new evidence showed that epigenetic and genetic factors are essential for regulation and maintenance of blood pressure, and there are complex interactions among genetic and environmental factors, thereby influencing the risk of EH. In the current experiments, we, for the first time, observed that the mechanisms of DNA methylation in the occurrence and development of EC dysfunction were related to hypertension. HCMV infection can induce differential expressions of host genes, and lead to changes in DNA methylation levels in the host. The results of the current research demonstrated that HCMV infection promoted proliferation of ECs by downregulating the expression level of RGS5 via DNA hypermethylation. To our knowledge, we first demonstrated that interactions among HCMV infection, environmental risk factors, and DNA methylation may lead to EC dysfunction, and thus, participate in the occurrence and development of EH.
In summary, in this study, we provided evidence demonstrating that RGS5 was an important negative regulator of the proliferation of ECs following HCMV infection. Mechanistically, we showed that DNA hypermethylation promoted the proliferation of ECs by RGS5 downregulation. Our findings provided a basis for exploring methods for CMV treatment that could also prevent EH, either directly with antivirals or by modulation of RGS5 induced by the virus. The main toxicity of systemic ganciclovir is neutropenia, and although foscarnet is an effective drug for overcoming this limitation of ganciclovir, its main side effects include nephrotoxicity and electrolyte imbalance 64 . Thus, it is urgent to find an effective vaccine and new antiviral intervention strategies to mitigate the deficiencies and toxicities of existing antiviral drugs 65,66 . Accordingly, RGS5 may represent a promising therapeutic target to prevent EH induced by HCMV infection.

Materials and Methods
cell culture and viral titers. Human umbilical cords (from patients with normal body weight) were obtained after full-term normal deliveries under protocols approved by the Ethics Committee of Obstetrics and Gynecology of an affiliated hospital of Shihezi University, and informed consent was obtained from all the patients. Human umbilical vein endothelial cells (HUVECs) were isolated and grown in endothelial cell medium (ECM; Life Technologies, Carlsbad, CA, America). HUVECs were cultured in ECM supplemented with fetal bovine serum (FBS; 10%) (HyClone, Logan, UT, America), streptomycin (0.1 mg/mL) and penicillin (100 IU/ mL). Human embryonic lung fibroblast cells (MRC-5 cells) were purchased from Wuhan Institute of Virology (Wuhan, China), and were maintained in Dulbecco's modified Eagle's medium (DMEM; Life Technologies, Carlsbad, CA, America) supplemented with FBS (10%), penicillin (100 IU/mL), and streptomycin (0.1 mg/mL). All cells were cultured at 37 °C in presence of 5% CO 2 .
The laboratory-adapted strain of HCMV AD169 obtained from Wuhan Institute of Virology (Wuhan, China) was propagated in confluent monolayers of MRC-5 cells in DMEM supplemented with FBS (10%), penicillin (100 IU/mL), and streptomycin. Supernatants of the infected MRC-5 cells displaying 90-100% cytopathic effects were collected, and centrifuged, which were then stored under −80 °C until the next analysis. Plaque assays were performed to determine Viral stock titers, which were calculated as 10 5 PFU/ml.

HcMV infection. HUVECs were plated in 12-well plates and cultured in ECM until cells reached 70-80%
confluency. Cells were then infected with HCMV (multiplicity of infection = 1, 10 5 PFU/ml) for 2 h, incubated for the indicated number of days post-infection, and then, washed, and refed with fresh ECM. To evaluate HCMV infectivity, immunofluorescence analysis was performed to examine infection rates of HCMV in HUVECs. After fixation with formaldehyde solution (4%), the cells were washed with PBS (phosphate-buffered saline) for three times, and subjected to indirect immunofluorescence analysis using anti-HCMV pp65 (Boster Corp., Wuhan, China), monoclonal antibodies, and fluorescein isothiocyanate-labeled sheep anti-rabbit IgG antibodies (dilution, 1:400; Boster Corp., Wuhan, China). The nuclei were stained with DAPI (4′6′-diamidino-2-phenylindole). The protein expression was observed under a fluorescence microscope.

RNA extraction and gene expression microarray analysis. TRIzol reagent (provided by Invitrogen,
Carlsbad, CA, America), chloroform, and isopropanol were used to extract total RNA from HCMV-infected or uninfected HUVECs according to the manufacturer manual. The extracted RNA was dissolved in 50 μL RNase-free H 2 O. The qualities and purities of the RNA preparations were determined by 1% agarose gel electrophoresis, and the quantities were determined by using an ND-1000 spectrophotometer (NanoDrop Technologies LLC, Wilmington, DE, America).
Microarray analysis was undertaken using RNA samples from HCMV-infected HUVECs, HUVECs treated with high glucose (HG; 4500 mg/L), and oxidized low-density lipoprotein (ox-LDL; 100 mg/L) for 48 h, as well as HCMV-infected HUVECs treated with HG and ox-LDL; uninfected cells were processed in parallel as a control. Microarray analysis was conducted by Beijing Biotech Co. Ltd. (Beijing, China). Four biological samples were added onto an Affymetrix PrimeView Human Gene Expression Array platform (Affymetrix Technologies, Inc., Santa Clara, CA, America), containing 49395 specific probes. RNA linear amplification was carried out according to the manufacturer manual. An Affymetrix Gene Chip Scanner 3000 (Affymetrix, CA, America) was employed to scan microarray slides after microarray hybridization. The raw data were converted logarithmically first, and the original data were standardized and analyzed using Affymetrix ® GeneChip ® Command Console ® Software software. DEGs between the infected and control groups were recognized as genes with a FC (fold-change) > ±2

Validation of microarray results by quantitative reverse transcription polymerase chain reaction (Rt-qpcR).
To confirm the results obtained by analysis of mRNA expression profiles, quantitative reverse transcription polymerase chain reaction was employed for determination of the expression level of dysregulated mRNAs. 3 µg total RNA which was extracted from each treatment group was reversely transcribed into cDNA (Tiangen Biotech, Shanghai, China), according to the manufacturer manual. In addition, RT-qPCR was carried out in a reaction system (25-μL) containing reverse primer, forward primer, SYBR Green/Fluorescein qPCR Master Mix and cDNA. The relative gene expression level was normalized using β-actin as the internal reference. 7300 Real-Time PCR system (provided by Applied Biosystems, Singapore) was employed. The SDS 2.0.1 software (provided by Applied Biosystems, Foster City, CA, USA) was used to calculate cycle threshold (Ct) values, which were normalized to β-Actin according to the 2 −ΔΔCt method. The sequences of the sense and antisense primers used for amplification are listed in Table 3.

Establishment of a recombinant adenovirus for regulating overexpression of RGS5 (G-protein signaling 5).
We constructed recombinant adenoviruses which expressed RGS5 using the AdMEX adenoviral vector system. Both the shuttle vector (pHBdTrack-CMV) and backbone plasmid (pBHGlox [delta] E1, 3cre) for this system were provided by Shanghai Biotech Co. Ltd. (Shanghai, China). This pHBdTrack-CMV vector carries two separate CMV promoters which drive the expression of RGS5 and the GFP (green fluorescent protein). Targeted cells were inoculated into a 24-well plate (1 × 10 5 cells/well). Cells were infected with the adenovirus carrying the target gene or an equal titer of control virus, and duplicate wells were infected for each group. After infection for 8 h, the virus was removed and replaced with a fresh medium. After 48 or 72 h, the level of fluorescence was measured by using a fluorescence microscope. Overexpression of RGS5 was confirmed by RT-qPCR as well.
Western blot analysis. We extracted proteins from all experimental samples, which were separated by electrophoresis on 10% or 12% SDS-PAGE (sodium dodecyl sulfate-polyacrylamide gel electrophoresis). Proteins were then transferred onto PVDF (polyvinylidene difluoride) membranes. The membranes were incubated at four centigrade degree overnight with antibodies directed against the following proteins after blocking with nonfat dry milk (5%) dissolved in TBS-T for 2 h at room temperature: proliferating cell nuclear antigen (PCNA; dilution, 1:400; Boster Corp., Wuhan, China) and β-actin (dilution, 1:1000; Boster Corp., Wuhan, China). After washing, the membranes were incubated with horseradish peroxidase (HRP)-conjugated secondary antibodies (dilution, 1:25000; Boster Corp., Wuhan, China) for 2 h under room temperature. After TBS-T washing for three times, 10 min each time, an enhanced chemiluminescence system (Pierce Company, Waltham, MA, USA) was used to visualize the proteins. Eventually, expression levels were quantified by normalization to β-actin using Image J software (National Institutes of Health, Bethesda, MD, America).

Statistical analysis.
Mean ± standard deviation (SD) was used to describe the data. One-or two-way analysis of variance (ANOVA) was employed for intergroup comparison. The significant difference between groups was evaluated using two-tailed unpaired Student's t-test. P < 0.05 indicated significant difference. SPSS 20.0 software (IBM Corp., Armonk, NY, USA) was employed for statistical analysis.

Data availability
The analyzed data sets generated during the study are available from the corresponding author on reasonable request.