The Effects of Knockdown of Rho-Associated Kinase 1 and Zipper-Interacting Protein Kinase on Gene Expression and Function in Cultured Human Arterial Smooth Muscle Cells

Rho-associated kinase (ROCK) and zipper-interacting protein kinase (ZIPK) have been implicated in diverse physiological functions. ROCK1 phosphorylates and activates ZIPK suggesting that at least some of these physiological functions may require both enzymes. To test the hypothesis that sequential activation of ROCK1 and ZIPK is commonly involved in regulatory pathways, we utilized siRNA to knock down ROCK1 and ZIPK in cultured human arterial smooth muscle cells (SMC). Microarray analysis using a whole-transcript expression chip identified changes in gene expression induced by ROCK1 and ZIPK knockdown. ROCK1 knockdown affected the expression of 553 genes, while ZIPK knockdown affected the expression of 390 genes. A high incidence of regulation of transcription regulator genes was observed in both knockdowns. Other affected groups included transporters, kinases, peptidases, transmembrane and G protein-coupled receptors, growth factors, phosphatases and ion channels. Only 76 differentially expressed genes were common to ROCK1 and ZIPK knockdown. Ingenuity Pathway Analysis identified five pathways shared between the two knockdowns. We focused on cytokine signaling pathways since ROCK1 knockdown up-regulated 5 and down-regulated 4 cytokine genes, in contrast to ZIPK knockdown, which affected the expression of only two cytokine genes (both down-regulated). IL-6 gene expression and secretion of IL-6 protein were up-regulated by ROCK1 knockdown, whereas ZIPK knockdown reduced IL-6 mRNA expression and IL-6 protein secretion and increased ROCK1 protein expression, suggesting that ROCK1 may inhibit IL-6 secretion. IL-1β mRNA and protein levels were increased in response to ROCK1 knockdown. Differences in the effects of ROCK1 and ZIPK knockdown on cell cycle regulatory genes suggested that ROCK1 and ZIPK regulate the cell cycle by different mechanisms. ROCK1, but not ZIPK knockdown reduced the viability and inhibited proliferation of vascular SMC. We conclude that ROCK1 and ZIPK have diverse, but predominantly distinct regulatory functions in vascular SMC and that ROCK1-mediated activation of ZIPK is not involved in most of these functions.

that ROCK1 and ZIPK have diverse, but predominantly distinct regulatory functions in vascular SMC and that ROCK1-mediated activation of ZIPK is not involved in most of these functions. cells [31][32][33][34][35][36]. In both smooth muscle and non-muscle cells and tissues, ZIPK was found to phosphorylate LC 20 at both Ser19 and Thr18 in a Ca 2+ -independent manner [4,18,19]. Phosphorylation or activation of myosin is considered to contribute to the reorganization of actin filaments [37], cell motility [4] and cell death [38] in both smooth muscle and non-muscle cells. The impairment of these processes has been shown to be involved in several pathological conditions, such as hypertension, stroke, vasospasm, atherosclerosis, heart failure, pulmonary hypertension and various cancers [39][40][41][42][43][44].
Many aspects of regulation via ROCK and ZIPK pathways remain unclear. For example, while ROCK1 has been shown to phosphorylate and activate ZIPK [24], it is unclear if this linear relationship is common in the regulation of physiological processes by these two kinases. To test this hypothesis, we utilized a siRNA approach in combination with microarray analysis using a whole-transcript expression chip to identify changes in gene expression profiles induced by ROCK1 and ZIPK knockdown. Cytokine signaling and cell cycle regulatory pathways were particularly affected, which prompted functional assays to test the involvement of these kinases in cytokine production, cell viability and cell proliferation. To our knowledge, this is the first such gene expression study to be performed and leads to the identification of novel roles for these protein kinases and the conclusion that direct activation of ZIPK by ROCK1 is not a universal mechanism in signal transduction.

Materials
All chemicals were reagent grade unless indicated otherwise.

Primary human arterial smooth muscle cell culture
Human coronary artery smooth muscle cells (CASMC: CC-2583) and human umbilical artery smooth muscle cells (UASMC: CC-2579) were purchased from Lonza and cultured in smooth muscle growth medium (Lonza SmGM-2: CC-3182) containing 5% fetal bovine serum (FBS), growth factors (hEGF, insulin and hFGF-β), gentamicin and amphotericin-B at 37°C in a humidified incubator with 5% CO 2 . These SMC maintain their morphological and phenotypic characteristics for up to 10 passages; therefore, cells were used for experiments within 10 passages.

siRNA transfection
Human ROCK1 siRNA (purchased from Santa Cruz Biotechnology; sc-29473) consists of three target-specific siRNAs (19-25 nucleotides). Human ZIPK siRNA (purchased from Ambion; part number 4390824, siRNA ID #s559, Lot #ASO0UMMU), has the following sequence: sense 5 0 -GAGGAGUACUUCAGCAACAtt-3 0 , and antisense 5 0 -UGUUGCUGAA-GUACUCCUCgt-3 0 . Negative control siRNA (Ambion negative control #2, part #AM4613) contains sequences that do not target any gene product. Cultured human CASMC and UASMC were transfected with siRNAs targeting ROCK1, ZIPK or negative control using HiPerFect Reagent (Qiagen) according to the manufacturer's protocol. Briefly, on the day of transfection cells were seeded into a 4-well plate at a density of 6 x 10 4 cells per well in 100 μL of complete SmGM and incubated under normal growth conditions. siRNAs were diluted in 100 μL of medium without serum to give a final concentration of 20 nM. After mixing by vortexing, HiPerFect Reagent (3 μL) was added to the diluted siRNA and mixed by vortexing for 10 s. The HiPerFect/siRNA complexes were added to the cells after incubation for 10 min at room temperature. The cells were incubated at 37°C in a CO 2 incubator for another 4 h and a further 300 μL of complete culture medium were added. The transfected cells were incubated under normal growth conditions and the medium was changed at 30 h after transfection. Cells were cultured for another 18 h and both cells and culture supernatants were harvested at 48 h after transfection. Cells were used for RNA isolation or western blotting, and supernatants were used for ELISA to determine interleukin-6 (IL-6) and other cytokine levels (see below). To determine transfection efficiencies, CASMC and UASMC were co-transfected with FITClabeled control siRNA and either ROCK1-or ZIPK-targeted or control siRNA, nuclei were labeled with DAPI and transfected cells were detected by fluorescence confocal microscopy. In each case, > 90% of the cells were fluorescently labeled.

RNA isolation and cDNA microarrays
Total RNA was isolated from CASMC using the RNeasy Micro kit (Qiagen) according to the manufacturer's instructions. The isolated RNA was quantified by UV absorption with a Nano-Drop 1000 (NanoDrop Technologies, Inc.) and the quality verified with an Agilent RNA 6000 NanoChip on a 2100 Bioanalyzer (Agilent Technologies). Samples with an RNA integrity number (RIN) ! 8 were considered suitable for use in microarrays. A total of 250 ng of RNA in each sample with RIN > 9 was reverse transcribed, amplified, labeled with the Ambion WT Express kit (Ambion) and hybridized to Human Gene 1.0 ST arrays (Affymetrix) at 45°C for 16 h. The arrays were washed and stained on an Affymetrix GeneChip Fluidics-450 following the manufacturer's protocol and scanned on an Affymetrix GeneChip Scanner 3000 7G System. The raw data sets for array comparisons have been deposited in the Gene Expression Omnibus website: www.ncbi.nlm.nih.gov/geo/ (accession number: GSE56819). Of the 28,868 genes represented on the GeneChip Human Gene 1.0ST array, genes for which the log 2-transformed signal intensities differed between the negative control and ROCK1-or ZIPK-knockdown groups by more than 0.585 (1.5-fold change without adjustment) were identified as genes whose expression was altered significantly by the kinase knockdown.

Cell viability assay
Cell viability after kinase knockdown was evaluated with the MTT assay. 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphen yltetrazolium bromide (MTT) was purchased from Sigma-Aldrich. After transfection with the siRNA/HiPerFect complexes for 30 h, the medium was replaced with fresh complete medium and the cells were cultured for a further 18 h. An amount of MTT (5 mg/ml) solution in phosphate-buffered saline equal to 10% of the culture volume was added to each well and the cells were further incubated for 4 h to form purple MTT formazan crystals. The MTT-containing medium was removed and 200 μL of MTT solvent (0.1 N HCl in anhydrous isopropanol) was added to dissolve MTT crystals. After the cells were gently agitated for 5 min, the absorbance at 540 nm was measured with a BioRad iMark Microplate Reader.

Cell proliferation assay
UASMC were transfected with siRNA to ZIPK or ROCK1 or with negative control siRNA as described above for 30 h. The medium was replaced with fresh complete medium (containing 5% FBS) and the cells were cultured for a further 18 h. 5-bromo-2-deoxyuridine (BrDU; 10 μM) was then added to the culture medium and the cells were cultured for an additional 1 h. The cells on coverslips were fixed with 10% formalin in neutral buffer (Sigma) and BrDU was detected with Alexa Fluor 488-conjugated anti-BrDU antibody (Invitrogen).

Quantitative real-time PCR (qRT-PCR)
Single-stranded cDNA was synthesized with random hexamer primers using the Superscript II reverse transcriptase according to the manufacturer's instructions (Invitrogen). Levels of mRNA expression were quantified by RT-PCR using a BioRad iCycler MyiQ detection system (BioRad) with pre-synthesized primers (Applied Biosystems) for ZIPK (Hs00154676_m1), IL-1β (Hs00174097_m1) and IL-6 (Hs00174131_m1). Reactions, which were performed in a MicroAmp Optical 96-well reaction plate, contained: 1x Master Mix, 200 μM each primer, and 100 μM probe in a volume of 25 μL. PCR conditions were 50°C for 2 min, 95°C for 10 min, 40 cycles of 95°C for 15 min/60°C for 1 min. All measurements were performed in duplicate. Expression of gene transcripts was normalized to the housekeeping gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH) [45]. Relative expression levels of the genes of interest were determined using the 2 -ΔΔC T method [46]. To identify genes significantly altered by ROCK1 or ZIPK knockdown, Student's t-test was performed for each assay group. To adjust for multiple hypothesis testing, the resulting p values were used to obtain the false discovery rates (FDR) using the q value method. Gene expression levels were considered significantly different with FDRs of < 5% (i.e. q values 0.05).

Western blotting
In order to categorize biological functions related to gene expression altered by kinase knockdown in our microarray analysis, Ingenuity Pathway Analysis (IPA, Ingenuity Systems, www.ingenuity.com) was used. The differentially expressed genes (DEGs) identified through the statistical analysis described above (fold-change 1.5 and q 0.05) in both knockdown assays were uploaded into IPA analysis.
Otherwise, statistically significant differences were determined by Student's unpaired t-test at a nominal significance of p < 0.05 and (where applicable) a fold-change in expression > 1.5. ANOVA followed by Bonferroni's post hoc test was used for multiple comparisons.

ROCK1 and ZIPK knockdown
The efficiency of siRNA-mediated knockdown of ROCK1 and ZIPK in CASMC was evaluated at both mRNA and protein levels by microarray analysis and western blotting, respectively. ROCK1-targeted siRNA decreased the ROCK1 mRNA level~12-fold without significantly affecting ZIPK mRNA expression, while ZIPK-targeted siRNA reduced the ZIPK mRNA level >2.5fold without significantly affecting the ROCK1 mRNA level ( Table 1). The reduction in ZIPK mRNA in cells transfected with ZIPK siRNA was confirmed by qRT-PCR (Fig. A in S1 File). At the protein level, ROCK1-targeted siRNA reduced ROCK1 protein expression by~80% and ZIPK-targeted siRNA reduced ZIPK protein expression by~50% ( Fig. 1 and Table 2). ROCK1 siRNA had no effect on ZIPK protein expression, but ZIPK siRNA treatment was accompanied by a small increase (~30%) in ROCK1 protein expression ( Table 2). We also examined the effect of ROCK1 knockdown on ROCK2 expression and found that ROCK1 knockdown (by 80% at the protein level) decreased ROCK2 protein expression to 65.2 ± 9.1% of control (n = 6).
Very similar effects were seen in the case of UASMC, in which the efficiency of siRNAmediated knockdown was evaluated at the protein level by western blotting. ROCK1-targeted siRNA again reduced ROCK1 protein expression by~80% and ZIPK-targeted siRNA reduced ZIPK protein expression by~50% (Table A in S1 File). ROCK1 siRNA again had no effect on ZIPK protein expression, but ZIPK siRNA treatment was accompanied by a small increase (~30%) in ROCK1 protein expression (Table A in S1 File).

Changes in gene expression profiles induced by ROCK1 and ZIPK knockdown
Three independent microarray assays were performed on CASMC 48 h after transfection with control, ROCK1 or ZIPK siRNAs using the GeneChip Human Gene 1.0ST array (28,868 genes represented). After normalization through the preprocessing module, a 3D Principal Components Analysis Plot showed clear separation of the data for ROCK1 / ZIPK knockdown from  Table 2.
doi:10.1371/journal.pone.0116969.g001 the negative control. Differentially expressed genes (DEGs) were identified on the basis of standard and advanced quality control metrics (FC ! 1.5 and q values 0.05); complete lists of these genes are provided in Tables B and C in S1 File for ROCK1 knockdown and ZIPK knockdown, respectively. ROCK1 knockdown affected the expression of 553 genes (355 downregulated and 198 up-regulated), while ZIPK knockdown affected the expression of 390 genes (219 down-regulated and 171 up-regulated). The gene expression profiles of the most affected genes are shown as heat maps in Fig. 2A (ROCK1 knockdown) and B (ZIPK knockdown). The DEGs were divided into different functional groups (iReport from http://www. ingenuity.com) ( Table 3). A high incidence of up-and down-regulation of transcription regulator genes was observed in both ROCK1 and ZIPK knockdowns. Other markedly affected groups included transporters, kinases, peptidases, transmembrane and G protein-coupled receptors, with multiple genes encoding growth factors, phosphatases and ion channels also being significantly affected. In each of these subgroups, more genes were down-regulated than up-regulated, in contrast to the sum total of affected genes. In the case of the cytokine subgroup, ROCK1 knockdown resulted in up-regulation of 5 genes and down-regulation of 4 genes, in contrast to ZIPK knockdown, which affected the expression of only two genes (both down-regulated).
To gain insight into potential overlapping roles of ROCK1 and ZIPK in vascular SMC, we selected the DEGs that were common to ROCK1 and ZIPK knockdown. There were 76 such DEGs, of which 41 were down-regulated (italics in Table 4) and 26 up-regulated (bold in Table 4) by both treatments, while the other 9 genes were differentially up/down-regulated (bold italics in Table 4). It is clear that many more genes were uniquely affected by ROCK1 or ZIPK knockdown.

Analysis of biological pathways and processes regulated by ROCK1 and ZIPK
Ingenuity Pathway Analysis (IPA; Ingenuity Systems, www.ingenuity.com) was used to facilitate the identification of biological themes in the microarray data. IPA shifts the emphasis from the evaluation of single genes to an evaluation of molecular pathways, networks and biological functions. 72 canonical pathways with p values < 0.05 were affected by ROCK1 knockdown (Table 5) and 27 by ZIPK knockdown (Table 6). Most interestingly, ROCK1 knockdown affected several cytokine signaling pathways. For example, IL-6 gene expression was upregulated by ROCK1 knockdown (2.24-fold; p = 0.015) according to the microarray analysis (Table 5), and this was confirmed by qRT-PCR ( Fig. 3A) and quantification of IL-6 protein levels in the medium of CASMC and UASMC transfected with ROCK1 siRNA (Fig. 3B). ZIPK knockdown, on the other hand, reduced IL-6 mRNA expression (Fig. 3A) and IL-6 protein secretion (Fig. 3B). qRT-PCR also demonstrated no significant changes in expression level of the housekeeping gene GAPDH as a result of ROCK1 or ZIPK knockdown (data not shown). These results suggest that ROCK1 may inhibit IL-6 secretion. It has been reported that IL-6 promotes tumor cell migration and invasion, and that ROCK1 may be involved in this process [47]. Furthermore, since ZIPK knockdown results in increased ROCK1 protein expression ( Fig. 1, Table 2, and Table A in S1 File), the observed decrease in IL-6 secretion following transfection with ZIPK siRNA may be due to increased ROCK1 expression. This possibility was supported by the effects of combined knockdown of ROCK1 and ZIPK. The combination of ZIPK-siRNA and ROCK1-siRNA reduced ZIPK protein levels to 64.9% of control and ROCK1 levels to 64.1% of control. The combined knockdown of ZIPK and ROCK1 had an intermediate effect on IL-6 secretion compared with knockdown of ZIPK and ROCK1 individually, i.e. IL-6 increased by 47.8 ± 13.4% (n = 9) in the double-knockdown compared with a decrease in IL-6 secretion in ZIPK-knockdown cells to 59.8 ± 5.5% of control siRNA-treated cells (n = 9) and an increase in IL-6 secretion by 143.6 ± 11.4% (n = 9) in ROCK1-knockdown cells. These results support the conclusion that some genes are regulated by both kinases and that upregulation of ROCK1 partially compensates for reduced ZIPK expression. IL-1β mRNA (Fig. 3A) and protein levels ( Fig. 4) were increased~2-fold in response to ROCK1 knockdown. ZIPK knockdown slightly reduced IL-1β mRNA expression compared to control (Fig. 3A), but the apparent reduction in IL-1β protein expression was not statistically significant ( Fig. 4). We also quantified the levels of several other cytokines in the culture medium of control cells and cells transfected with siRNA to ROCK1 or ZIPK. As shown above for IL-6, MCP1 (CCL2) secretion was increased in cells transfected with siRNA to ROCK1 and decreased in cells transfected with siRNA to ZIPK (Table D in S1 File).
Five common pathways were affected by both ROCK1 and ZIPK knockdown (highlighted in bold in Tables 5 and 6) and they are mainly involved in cell cycle regulation. The Cell Cycle: G1/S Checkpoint Regulation pathway contains the largest number of genes whose expression is affected by both ROCK1 and ZIPK knockdown: ROCK1 knockdown resulted in 8 DEGs (4 up-regulated) while ZIPK knockdown resulted in 10 DEGs (8 up-regulated). Three upregulated genes (HDAC1, CDC25A, and CCNE2) were shared by both ROCK1 and ZIPK knockdowns and all others were differentially up/down-regulated, indicating that ROCK1 and ZIPK regulate the cell cycle by different mechanisms. These differences may account for our observation that ROCK1, but not ZIPK knockdown, significantly reduced the viability of vascular SMC from both coronary and umbilical arteries: the viability of both cell types was reduced (by~25% in the case of CASMC and~29% in the case of UASMC) as a result of ROCK1 knockdown (Fig. 5). ZIPK knockdown, on the other hand, had no significant impact on cell viability when compared to the negative controls.
The effects of ROCK1 and ZIPK knockdown on UASMC proliferation were investigated by measuring BrDU incorporation 48 h after siRNA transfection. ZIPK knockdown had no significant effect on the proportion of BrDU-positive cells compared with cells transfected with control siRNA (Fig. 6). On the other hand, ROCK1 knockdown reduced the proportion of BrDU-positive cells by~50% (from 40.2 ± 1.6% to 21.9 ± 2.2% of the total cell population). These results support an important role for ROCK1 in vascular SMC proliferation, as  previously proposed [48], and suggest that ZIPK does not affect proliferation. It is important to note that, in the experiments of Fig. 6, fields were selected to contain approximately the same number of cells although, consistent with the MTT assay (Fig. 5), ROCK1 knockdown reduced overall cell viability. Furthermore, we cannot rule out the possibility that the lack of effect of ZIPK knockdown on viability and proliferation may be due to the fact that~50% of cellular ZIPK is retained. Apart from classifying individual genes into pathway categories, IPA also predicts corresponding biological processes that are significantly associated with identified DEGs. The main processes associated with the most significantly altered genes in ROCK1-knockdown CASMC were cell migration and movement, apoptosis, tissue development, smooth muscle cell proliferation, cell differentiation and protein phosphorylation ( Table 7). The main processes associated with the most significantly altered genes in ZIPK-knockdown CASMC were cell cycle, cell death and viability, necrosis, DNA metabolism and replication (Table 8). Four biological processes (highlighted in bold in Tables 7 and 8) were shared by the two groups: necrosis, apoptosis, proliferation of cells, and cell death of tumor cell lines. Processes unique to ROCK1 knockdown include cell migration and movement, and smooth muscle development, and those unique to ZIPK knockdown include cell cycle and cell viability.
IPA was also utilized to identify DEGs that connect to other DEGs in biological processes or pathways. The top 10 DEGs with the most connections in the ROCK1-and ZIPK-knockdown datasets are listed in Table 9. Among these genes, HDAC1 was the only gene affected by knockdown of both ROCK1 and ZIPK, and was the top gene in the ZIPK group (with 14 neighbors) and fourth in the ROCK1-knockdown group (with 21 neighbors). MYC was the top gene in the ROCK1-knockdown group, which also contains three other transcription regulators (PPARG and VDR, in addition to HDAC1), all of which were up-regulated. Interestingly, upregulated IL-6 is the number 2 gene in the ROCK1-knockdown group, and it has the most connections in this group at 28.
In comparison to ROCK1 knockdown, all eight proteins encoded by the up-regulated DEGs in the ZIPK-knockdown group are localized to the nucleus and related to transcriptional    regulation. Four of these eight genes belong to a nuclear enzyme family (minichromosome maintenance complex component: MCM 3, 6 and 7, together with CDK6) and are involved in control of the chromosomal replication pathway in the cell cycle. These genes are essential for eukaryotic DNA replication. This pathway was not observed in the ROCK1-knockdown group, which may provide another explanation for the decreased viability of ROCK1-knockdown SMCs (Fig. 5). This information is consistent with the pathway and biological function categories discussed above, and indicates again that both kinases are involved in regulation of the cell cycle and cell proliferation, but via different mechanisms. The majority of the DEGs (9 of 10) with the most connections were not shared by the two groups, indicating unique biological functions.

Analysis of diseases associated with ROCK1 and ZIPK knockdown
In order to explore the pathological effects of ROCK1 and ZIPK expression in cells, DEGs were categorized in accordance with their association with different diseases. 127 different diseases were associated with DEGs in the ROCK1-knockdown group and 149 with the ZIPKknockdown group. The top 25 pathological conditions from each group are presented in Tables 10 (ROCK1 knockdown) and 11 (ZIPK knockdown), based on their p values. Disease associations common to ROCK1 and ZIPK knockdown are depicted in bold in Tables 10 and 11. 18 of the 25 diseases in the ROCK1-knockdown group are cancer/tumor-related, 5 are immune-or autoimmune-related, and 2 are seizure-related. These observations are consistent with other ROCK1-knockdown microarray information, such as the higher up-regulation rate of 56% of affected genes in the cytokine sub-group, in contrast to only 36% up-regulation for all DEGs ( Table 3). The higher protein levels of three pro-inflammatory cytokines (IL-6, IL-1β and MCP1 (CCL2)) in ROCK1-knockdown cells (Fig.s 3B and 4, and Table D in S1 File) provide further support for a role for ROCK1 in inflammation and immune-related processes. Almost all of the top 25 diseases in the ZIPK-knockdown group are cancer/tumor-related (the sole exception being mosaic variegated aneuploidy). Finally, we investigated the effect of ROCK1 and ZIPK knockdown on phosphorylation of a well-known substrate, LC 20 [19,28]. The level of expression of LC 20 was unaffected by knockdown of either ROCK1 or ZIPK: LC 20 was quantified in UASMC transfected with scrambled siRNA, ROCK1-or ZIPK-siRNA by western blotting with GAPDH as loading control. The levels of expression of LC 20 in ZIPK-and ROCK1-knockdown cells relative to control cells were 95.1 ± 2.8% and 97.6 ± 3.0%, respectively (n = 14). Phosphorylation of LC 20 at Thr18 and Ser19 [19,28] was then examined by western blotting with a phosphospecific antibody that   (Table 12 and Fig. B in S1 File).

Discussion
Given that ROCK1 and ZIPK mediate important cellular functions and have been implicated in many pathological conditions, mainly vascular diseases such as hypertension, stroke, vasospasm, atherosclerosis, heart failure and pulmonary hypertension, as well as cancer, they are potentially useful therapeutic targets. Many compounds have been developed as inhibitors of ROCK or ZIPK and have been proposed to have therapeutic benefits for multiple diseases [10,49]. However, these compounds usually lack specificity because the ATP-binding sites of protein kinases, with which most chemical inhibitors interact, are quite conserved, which limits their therapeutic promise and usefulness as tools to investigate signaling pathways involving these kinases. As an alternative strategy, RNA interference (RNAi) is being widely utilized to deplete specific proteins from cells or tissues.
Targeting of gene transcripts using RNAi also provides a powerful approach to the analysis of gene function and to assess the role of specific proteins in particular cell functions. The use of microarray analysis to study gene expression patterns and profile changes is a powerful approach to identification of the biological functions of specific gene products. In this study, we have utilized RNAi and microarray analysis to further understand ROCK1 and ZIPK functions. This is the first attempt of its kind aimed at studying global gene expression changes resulting from kinase knockdown in order to identify unique genes regulated by ROCK1 and ZIPK, to explore novel functions of the two kinases and to test the hypothesis that ROCK1 commonly lies upstream of ZIPK. In this manuscript, we present the significant gene expression pattern variations and the analysis of the most relevant biological pathways and processes in ROCK1and ZIPK-knockdown SMC to understand their potential functions. We confirmed 8 of the up-or down-regulated genes using qPCR, western blotting and/or ELISA: IL-1β and IL-6 (Figs. 3 and 4), MCP1 (CCL2) (Table D in S1 File), COX2, vEGF, ICAM, SOCS3 and ILK (data not shown). Our gene expression analysis clearly distinguished the effects of ROCK1 from ZIPK knockdown in CASMC. Numerous genes showed altered expression following kinase knockdown and several classes of gene functions were identified. It is interesting to note that both ROCK1 and ZIPK are involved in regulation of the cell cycle. Cell cycle: G1/S Checkpoint regulation was the principal pathway (based on the p value) affected by both knockdowns. However, the functional implications of the significant DEGs included in the pathway are unclear. As verified by the MTT test, ZIPK knockdown did not affect cell viability, whereas ROCK1 knockdown inhibited cell viability significantly. Gene expression changes may be due to a direct effect on transcriptional regulation or an indirect effect on the cytoskeleton. We did not observe any cell shape change by light microscopy in either ZIPK-or ROCK1-knockdown SMC. Furthermore, staining of the cytoskeleton with Alexa Fluor 388 phalloidin in kinaseknockdown cells did not reveal any effect on the structure of the actin cytoskeleton (data not shown), suggesting direct effects of ZIPK and ROCK1 knockdown on transcriptional regulation.
Accumulating evidence suggests that ROCK1 plays an important role in the regulation of cell proliferation and survival. ROCK is activated during the cell cycle at both the G1/S-phase and during mitosis. Over-expression of constitutively-active ROCK1 increases cell proliferation [50]. In contrast, inhibition of ROCK activity delays cytokinesis [51]. The G1-phase cell cycle checkpoint is the critical juncture at which DNA integrity is assessed and determines whether a cell commits to the cell cycle or reverts to the quiescent G0-phase. ROCK expression in 3T3 fibroblasts stimulates their progression into S-phase [52] while inhibition of ROCK activity in gastric cells impairs the G1/S-phase transition [53]. This is also supported by studies of cardiomyocytes showing that ROCK inhibition reduces the expression of G1/S-phase transition-dependent proteins [54]. A recent report indicated that ROCK1 knockdown decreased PDGF-BB-generated vascular SMC proliferation by inhibiting the expression of proliferating cell nuclear antigen (PCNA) and cyclin D1 [48]. Indeed, in our study ROCK1 knockdown significantly decreased cell proliferation and the expression of both CCND2 and 3 (cyclin D2 and 3; FC for CCND2 is-1.47), although cyclin D1 was unchanged. Over-expression of ZIPK causes cell death via both caspase-dependent apoptosis and caspase-independent autophagy [23,55,56]. Cells that over-express ZIPK typically show signs of rounding, membrane blebbing, DNA fragmentation and detachment from the extracellular matrix. We observed a high incidence of altered DEGs associated with the cell cycle and nuclear location in ZIPK-knockdown SMC. Further study of these DEGs associated with ZIPKknockdown may provide valuable insights into the mechanisms of cell death induced by ZIPK over-expression.
Inflammation is involved in most chronic diseases and the inflammatory response is regulated by diverse cytokines, chemokines, growth factors and cell adhesion molecules. These mediators are produced by multiple cell types, including mucosal epithelium, smooth muscle cells, glial cells and various leukocytes. Smooth muscle cells act as both sources and targets of these mediators, establishing an intricate, dynamic balance that maintains the inflammatory  response [57,58]. Intestinal inflammation in human and animals has been widely shown to be associated with changes in contractility [57,59] and IL-1β was found to inhibit both initial and sustained contraction and LC 20 phosphorylation stimulated by acetylcholine in the presence of an M2 receptor antagonist [60]. Inflammation is observed to various degrees in a diverse group of vascular diseases, including atherosclerosis and myocardial infarction. IL-1β and IL-6 are the two major pro-inflammatory cytokines involved in regulation of the inflammatory process and exhibit increased levels in these pathological situations. Epidemiologically, vascular inflammation is detected by enhanced circulating IL-6 produced by activated inflammatory cells within the vessel wall. The immune response following inflammation directly induces apoptosis and indirectly increases cell death by signaling cellular death programs [61,62], although the precise mechanisms are unknown. Interestingly, in this study we observed that cytokine gene expression, specifically of IL-6, IL-1β and MCP1 (CCL2), was differently affected by ROCK1 and ZIPK knockdown in CASMC, i.e., ROCK1 knockdown increased while ZIPK knockdown decreased their expression. As noted earlier, ROCK1 knockdown by~80% reduced ROCK2 protein expression by~35%, making it difficult to distinguish between the effects of ROCK1 and ROCK2. However, knockdown of ROCK2 at the protein level by~55% had no effect on IL-6 expression suggesting that ROCK1 is likely the isoform involved in regulation of IL-6 expression.  MYC  29  IL6, CEP57, STAT1, PIM2, MYO9A, BTRC, FBXW7, CAMK2D, BCR, ANGPTL4, STK38, IQGAP3, VDR,  RAD21, RPN1, PIM1, PLOD3, CHCHD3, IGF2BP2, CDC25A, SQRDL, TAF12, CYR61, KDM5A, ADM,  SNW1, CDKN2B, HDAC1, NEK11   IL6  28  ABCG2, CCL11, SOCS3, FCGR1A, STAT1, PIM2, F2RL1, PTGS2, RARRES3, FCGR1B, CEBPD, VDR,  BMP6, PTHLH, NNMT, MYC, PIM1, LBP, TNFSF13B, CDC25A, PTX3, IGFBP3, PPARG, GHR, IRF1,  ADM, STX2, MAPK9   STAT1  24  IL6, USP18, PARP9, SOCS3, IFNE, FCGR1A, TEK, TOM1L1, BTRC, PAWR, ANGPTL4, VDR, CCR1,  SCGB3A2, MYC, LBP, STAT4, IGFBP3, AGTR1, PPARG, GHR, IRF1, IFNAR1, HDAC1   HDAC1  Although microarray analysis provides a detailed transcription profile, it has limitations associated with the data analysis. Some genes with higher variations (in both knockdown and microarray assays) are easily excluded by p values because of the low number of replicates (commonly, n = 3). IL-1β is a good example: its differential expression was verified by both western blotting at the protein level and by qRT-PCR at the mRNA level, but it was excluded from the array dataset because of a high p value. We used FC > 1.5 and q value < 0.05 as the criteria to identify DEGs, but some genes excluded by these criteria were in fact significantly altered by kinase knockdown. For example, IL-6 (FC = -1.26) and ROCK1 (FC = -1.006) were found to be down-and up-regulated, respectively, in ZIPK-knockdown cells at the protein level. RNAi also has its limitations, which must be taken into account when interpreting the effects of down-regulation of specific target genes by siRNA. For example, as shown in this study, the extents of down-regulation of ROCK1 (~80%) and ZIPK (~50%) at the protein level were quite different. ZIPK down-regulation affected ROCK1 expression, which was increased by~30%, whereas ROCK1 down-regulation had no effect on ZIPK expression. Furthermore, ROCK1 down-regulation (by~80%) reduced the level of ROCK2 expression (by~35%).
LC 20 has been identified as a substrate of both ROCK1 and ZIPK, at least in vitro, with phosphorylation occurring at Thr18 and Ser19 [19,28]. We showed that down-regulation of ROCK1 or ZIPK reduced LC 20 phosphorylation (Table 12 and Fig. B in S1 File). This could be explained by direct phosphorylation of LC 20 by both kinases or activation of ROCK1 leading to activation of ZIPK, which phosphorylates LC 20 . In summary, this study has identified sets of genes whose expression is altered by knockdown of ROCK1 or ZIPK in vascular SMC, as well as signaling pathways, biological processes and diseases involving each of these kinases. At this point we cannot conclude whether these Values indicate the mean ± SEM (n = 20). * , **significantly different from control (*p < 0.001; **p < 0.01).
Representative western blots are shown in Fig. B in S1 File.
doi:10.1371/journal.pone.0116969.t012 gene expression changes represent global effects of ROCK1 and ZIPK knockdown caused by key upstream regulators, or if they are genes that are directly targeted by the respective kinases. In addition, certain compensatory changes may occur over time, as opposed to primary effects of gene knockdown. Further studies focused on these genes may provide valuable insights into the physiological and pathophysiological roles of ROCK1 and ZIPK in vascular SMC. For example, the opposing effects of ROCK1 and ZIPK knockdown on the expression of IL-6, IL-1β and MCP1 (CCL2) suggest that ZIPK may be a more suitable therapeutic target in the case of chronic inflammatory diseases. Finally, the fact that, while knockdown of ROCK1 or ZIPK has a number of similar effects, the differences are far greater, suggesting that ROCK1mediated activation of ZIPK may be involved in only a subset of signaling pathways involving these kinases.
Supporting Information S1 File. Fig. A, Quantification of ZIPK mRNA levels in CASMC transfected with ZIPK or ROCK1 siRNAs. CASMCs were transfected with siRNA to ZIPK or ROCK1 or with negative control siRNA. Cells were lysed 48 h later for qRT-PCR to quantify ZIPK mRNA levels. Values represent means ± SEM (n = 9). Ã significantly different from control (p < 0.001) . Fig. B, Effect of ROCK1 and ZIPK knockdown on myosin phosphorylation. Representative western blots of control, ZIPK-and ROCK1-knockdown UASMC with anti-2P-LC 20 (antibody recognizing LC 20 only when phosphorylated at Thr18 and Ser19) and GAPDH as loading control. See Table 12 for cumulative quantitative data. Table A, ROCK1 and ZIPK knockdown in UASMC at the protein level. UASMC were transfected with siRNAs targeting ROCK1 or ZIPK or with negative control siRNA as described in the Materials and Methods section. The efficiency of knockdown at the protein level was determined by western blotting. ROCK1 and ZIPK signals were normalized to GAPDH. Values are expressed relative to levels in control cells (means ± SD, n = 11 for ROCK1 knockdown and n = 8 for ZIPK knockdown). Ã significantly different from control (p < 0.001). Table B, Genes whose expression is altered by ROCK1 knockdown. Table C, Genes whose expression is altered by ZIPK knockdown. Table D, Effects of ZIPK and ROCK1 knockdown on cytokine secretion. A Human Custom Multi-Analyte ELISArray kit (CELISA-CMEH0590A) was purchased from Qiagen. The indicated cytokines were assayed in the medium of CASMC transfected with control siRNA (Control) and CASMC transfected with siRNA to ZIPK (ZIPK knockdown) or ROCK1 (ROCK1 knockdown). Positive controls provided with the kit verified the viability of the assay for each cytokine. Negative controls indicated that, of the 6 cytokines listed, IL-1α, MCP1 and GROα were secreted at detectable levels. Values indicate absorbance at 450 nm ± S.E.M. (n = 4). Ã p < 0.05 compared to Control. (PDF)