The Rho-associated coiled-coil-containing protein serine/threonine kinases (ROCK, referring collectively to both ROCK1 and ROCK2 isoforms) are downstream effectors of the Rho subfamily GTPase. They regulate cell migration, intercellular adhesion, cell polarity cytokinesis, and cell proliferation by affecting actin cytoskeleton reorganization and microtubule dynamic (Etienne-Manneville & Hall, 2002; Olson & Sahai, 2008; Ridley, 1999). The biological roles and expression patterns of ROCK in mammalian embryo development have been intensively studied by pharmacological disruption of ROCK activity by the specific ROCK inhibitor Y-27632 and by using Real-Time qRT-PCR (Quantitative Reverse Transcription Polymerase Chain Reaction) and immunocytochemistry (ICC) approaches to visualize expression of ROCK genes and proteins (Duan et al., 2014; Kawagishi et al., 2004; Zhang et al., 2014). These studies demonstrated that early embryo developmental competence was affected in a dose/stage-dependent manner when mouse and porcine embryos were exposed to Y-27632 (100 µM in mouse, 5, 10, and 15 µM in porcine). For example, both ROCK1 and ROCK2 were detected during porcine preimplantation embryo development, but ROCK2 was more abundantly expressed in the trophectoderm (TE) of the blastocyst embryos compared to the expression levels of ROCK1. The majority of early cleavage stage embryos failed to progress to the blastocyst stage and remained arrested at the compaction/morula stages after they were exposed to Y-27632 at concentrations higher than 10 µM. Furthermore, porcine and mouse morula embryos treated with progressively higher concentrations of Y-27632 did not develop to the blastocyst stage either.
Recently, it has been reported that pharmacological inhibition of ROCK and RHO by Y-27632 and RHO inhibitor I from the two-cell stage enhanced the inner cell mass (ICM) lineage and suppressed TE lineage establishment through activation of Hippo signaling and disruption of cellular polarity due to suppression of the activity of the LATS kinase in the outer blastomere cells (Kono, Tamashiro & Alarcon, 2014). Furthermore, treatment of blastocysts with Y-27632 disrupted ICM morphology and caused a transient reduction in the blastocyst cavity size and fetal loss after implantation (Laeno, Tamashiro & Alarcon, 2013). In line with these findings, inhibition of ROCK activity by Y-27632 induced defects of position-dependent HIPPO signaling and downregulation of Cdx2 expression in the emerging TE lineage (Cao et al., 2015).
Taken together, these previous studies provided strong evidence that ROCK activity is involved in blastocyst formation by regulating TE and ICM establishment via position-dependent HIPPO signaling. However, it is not known why formation of the fluid-filled cavity is inhibited in early cleaving embryos exposed to the ROCK inhibitor and why the size of the blastocyst cavity transiently decreased in blastocysts treated with Y-27632. It is well documented that blastocyst formation requires water channels as physiological mediators of fluid movement across the TE, a correct Na+∕K+− ATPase-generated trans-TE ion gradient for water accumulation, and a proper assembly of tight junction (TJ) proteins (Watson & Barcroft, 2001; Watson, Natale & Barcroft, 2004). Recently, ROCK has been reported to be down-regulated in TFAP2C-depleted embryos that were arrested in the transition between morula and blastocyst stages. These embryos exhibited defects of paracellular sealing because of the TJ disruption and had a phenotype resembling that of embryos treated with Y-27632 (Cao et al., 2015; Choi et al., 2012). Thus, we hypothesized that inhibition of ROCK activity leads to the impairment of TJ functions. Particularly, our recent finding that CXADR is required for adherens junction (AJ) and TJ assembly during porcine blastocyst formation (Kwon, Kim & Choi, 2016) also supports this hypothesis, because CXADR was reported to be directly associated with ROCK in human tumor cells (Saito et al., 2014). Here, we report that ROCK activity is involved in functional TJ assembly and paracellular sealing during the preimplantation development of parthenogenetic porcine embryos.
Materials and Methods
All reagents were purchased from Sigma-Aldrich (St. Louis, MO, USA) unless stated otherwise.
Collection of porcine oocytes and embryo culture
Porcine-oocyte collection and embryo culture were carried out as described previously (Lee et al., 2015). Briefly, ovaries obtained from a local slaughter house were transported to the laboratory in Dulbecco’s phosphate-buffered saline (DPBS) at 37 °C. Cumulus-oocyte complexes (COCs) aspirated from 3–6 mm follicles were washed three times with 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES)-buffered Tyrode’s medium containing 0.1% (w/v) polyvinyl alcohol (HEPES-TL-PVA). Groups of ∼50 COCs were incubated in the in vitro maturation medium for 44 h at 39 °C and then denuded by pipetting in HEPES-TL containing 1 mg/mL hyaluronidase for 2–3 min. Denuded oocytes were treated with 50 µM calcium ionophore A23187 for 5 min and incubated in porcine zygote medium 3 (PZM3) containing 7.5 mg/mL cytochalasin B for 3 h for parthenogenetic activation. The embryos were washed three times in HEPES-TL-PVA, transferred to PZM3 supplemented with 0.4% (w/v) BSA, and cultured until use at 39 °C in a humidified atmosphere containing 5% of CO2.
Quantification of transcript levels
Isolation of mRNA from porcine blastocysts (ten samplings per each biological replicate) was carried out by using the Dynabeads mRNA Direct Kit (Dynal ASA, Oslo, Norway). First-strand cDNA was synthesized with the Superscript Reverse Transcriptase Enzyme (Invitrogen, Grand Island, NY, USA). Quantitative real-time RT-PCR (qRT-PCR) was carried out using cDNA synthesized with the Superscript Reverse Transcriptase Enzyme (Invitrogen, Grand Island, NY, USA) on a DNA Engine Opticon 2 Fluorescence Detection System (MJ Research, Waltham, MA, USA) with the DyNAmo SYBR Green qPCR Kit (Finnzymes Oy, Espoo, Finland). Transcripts of porcine CXADR, OCLN, and TJP1 genes were amplified using specific primer pairs and conditions (Kwon, Kim & Choi, 2016). Relative quantification of gene expression was performed by the 2−ΔΔCt method from three technical and biological replicates for the control and ROCK inhibitor-treated groups. GAPDH was used as the internal control in all experiments.
The procedure was carried out as described previously (Lee et al., 2015). Porcine preimplantation embryos were washed in DPBS that contained polyvinyl alcohol (1 mg/mL) and fixed for 20 min in 3.7% (w/v) paraformaldehyde dissolved in DPBS. Thereafter, embryos were permeabilized and blocked in DPBS that contained 0.5% (v/v) Triton X-100 and 5% donkey serum at room temperature for 1 h. Embryos were then incubated with a rabbit polyclonal anti-CXADR antibody (Sigma), a mouse monoclonal anti-TJP1 antibody (ZO-1; Zymed, San Francisco, CA, USA), or a mouse monoclonal anti-OCLN antibody (Zymed) in the blocking solution (DPBS with 5% donkey serum) overnight at 4 °C, followed by incubation with Alexa Fluor 488- and 594-conjugated antibodies (Molecular Probes, Eugene, OR, USA) as secondary antibodies. The embryos were then mounted in Vectashield containing 4′,6-diamidino-2-phenylindole (DAPI; Vector Laboratories, Burlingame, CA, USA) and analyzed by a laser-scanning confocal system with a krypton-argon ion laser under a Leica DM IRB inverted microscope (DM IRB; Leica, Wetzlar, Germany).
Treatment of embryos with the ROCK inhibitor Y-27632
To examine the effects of the ROCK inhibitor Y-27632 on embryo development and polarized distribution of AJ and TJ proteins, late four-cell and morula embryos were incubated in 30 µl droplets supplemented with Y-27632 at final concentrations of 10, 20, and 100 µM for 24 h at 39 °C in a humidified atmosphere containing 5% of CO2. Since the stock solution of Y-27632 was prepared in dimethyl sulfoxide (DMSO), control embryos were treated with the equivalent amount of DMSO and cultured until use.
TJ permeability assay using FITC-dextran uptake
To assess the effects of inhibition of ROCK activity on TJ permeability, DMSO vehicle control embryos and blastocysts derived from morula embryos treated with 10 µM and 20 µM of Y-27632 were incubated in the culture medium containing 1 mg/mL 40-kDa (fluorescein isothiocyanate) FITC-dextran for 30 min at 37 °C. Then, they were immediately washed several times in 50-µl droplets of PZM5 to remove FITC-dextran on the surface of embryos, placed into a clean droplet of PZM5, and visualized under an inverted fluorescence microscope (TE2000U; Nikon, Tokyo, Japan).
Quantitative development data as well as results of protein and gene expression experiments were subjected to the analysis of variance (ANOVA), whereas data from the tight junction permeability assay were analyzed by the chi-square test using the Statistical Analysis System software (Statistical Analysis System, Inc., Cary, NC, USA). The data are presented as the mean ± standard error of the mean (SEM). Differences with P values <0.05 were considered statistically significant unless stated otherwise.
Blastocyst development of porcine embryos treated with the specific ROCK inhibitor Y-27632
To evaluate the effects of ROCK inhibition on blastocyst development, we treated the four-cell embryos with 0 (control), 10, 20, and 100 µM Y-27632 for 24 h. We observed that exposure to the inhibitor led to significantly lower blastocyst development rates (39.1, 20.7, 10.0, and 0%, respectively) (Fig. 1A). The effects were significantly greater at higher Y-27632 concentrations (20 and 100 µM). The embryos that failed to develop to the blastocyst stage were arrested at the pre-compaction stage in presence of 10 and 20 µM Y-27632 (Figs. 1B and 2A).
Next, we treated the four-cell and morula embryos with 20 µM Y-27632 for 24 h and examined the stage-dependent effects of the ROCK inhibitor on morphological responses such as compaction and cavitation at 120 h.p.a (hours post activation). The majority of the four-cell embryos did not develop beyond the compacted/morula stage and remained at the four-cell stage as a result of treatment (Figs. 2A and 2B). Furthermore, while 49.1% of control morula reached the blastocyst stage with cavity and expansion, only 9.2% of morula treated with the ROCK inhibitor developed to the blastocyst (Fig. 2B). At 120 h.p.a, we found that the total cell number in embryos treated with 20 µM Y-27632 was lower than that in control embryos (16.25 ± 2.1 and 29.7 ± 4.4, respectively) (Fig. 2C).
We also investigated the subcellular localization of the CXADR protein, which has been reported to be associated with ROCK activity, to determine whether ROCK affects polarized distribution of adhesion and tight junction proteins. In embryos arrested at the precompaction stage by their exposure to 100 µM Y-27632 for 24 h, the fluorescent signals were dispersed in the cytoplasm and very weak signals were detected at 48 h after the treatment, but control embryos developed to the morula stage and showed continuous thin lines at the apical region of cells (Fig. 2D). Embryos treated with the inhibitor at the morula stage for 24 h showed relatively visible continuous lines along the cell–cell boundaries compared to those in the four-cell treatment group but less distinct than those in control blastocysts at 120 h.p.a (Fig. 2D). These results suggested that ROCK activity affects cellular polarization of adhesion and tight junction proteins.
Inhibition of AJ and TJ assembly in embryos by Y-27632
Based on our findings and previous reports regarding the role of ROCK in blastocyst formation (Choi et al., 2012; Kawagishi et al., 2004; Zhang et al., 2014), we speculated that inhibition of the cavitation onset in Y-27632-treated embryos is caused by mislocalization or AJ and TJ proteins and impaired TJ assembly. Thus, we examined localization and expression levels of CXADR, OCLN, TJP1, and CDH1 mRNA and protein molecules, which are essential for the integrity of the AJ-TJ complex, by using ICC and qRT-PCR at the blastocyst stage. In embryos exposed to 20 µM Y-27632, we observed that the CXADR and TJP1 proteins were barely detected at the apical regions of the cell–cell boundaries. In addition, while the OCLN protein was expressed in continuous lines at the apical edge of cell–cell contacts, its total expression in the treated blastocyst was very weak (Fig. 3A). The relative levels of CXADR, OCLN, TJP1, and CDH1 mRNA transcripts were also significantly lower in embryos from the Y-27632 treatment group (P < 0.05, Fig. 3B).
These findings led us to hypothesize that integrity of the TJ complex was disrupted by mislocalization and lower expression levels of tight junction proteins in presence of the ROCK inhibitor. To examine the functional barrier properties of the tight junction assembly, we treated morula embryos with 10 µM and 20 µM Y-27632 for 24 h and cultured them until the blastocyst stage. At 144 h.p.a, we exposed blastocysts, including control ones treated with DMSO, to 40-kDa FITC-dextran. We found a significant difference in the tight junction permeability in blastocysts from control and inhibitor-treated groups (P < 0.05), but no difference between the two treatment groups (Table 1). This observation demonstrated that inhibition of ROCK activity led to impairment of tight junction assembly during porcine blastocyst cavitation.
|DMSO||10 μM||20 μM|
|FITC dextran Positive||3||7||10|
|FITC dextran Negative||59||25||21|
Y-27632 treatment at the morula stage
Previous studies reported that inhibition of ROCK activity by its specific inhibitor Y-27632 prevented blastocyst cavitation, tight junction biogenesis, TE and ICM formation, establishment of the intracellular polarity in the form of the apical-basal axis, and affected developmental competency in a dose- or a stage-dependent manner in mouse (Cao et al., 2015; Duan et al., 2014; Kawagishi et al., 2004; Kono, Tamashiro & Alarcon, 2014). The defects found in embryos treated with the ROCK inhibitor were attributed to the loss of position-dependent HIPPO signaling that is mediated via the RHO-ROCK signaling (Kono, Tamashiro & Alarcon, 2014). A number of studies have shown that ion gradient, adherens junctions, tight junctions, and water channels are properly expressed and localized on the outer cells of embryos and the TE to form blastocyst (reviewed in Choi et al., 2012; Watson & Barcroft, 2001; Watson, Natale & Barcroft, 2004). However, the specific cellular events and molecules involved in developmental perturbations in early embryos induced by ROCK inhibition were not defined. Here, focusing on tight junction assembly, we investigated how inhibition of ROCK activity affected preimplantation development of porcine embryos.
First, we demonstrated that treatment with Y-27632 resulted in the failure of early porcine embryo development in a dosage-dependent manner. In line with previous results obtained in the mouse model (Duan et al., 2014), treatment with Y-27632 (100 µM) prevented four-cell embryos from undergoing compaction and also inhibited the morula-to-blastocyst transformation during porcine preimplantation development (summarized in Fig. 4A). A possible explanation of defective compaction and impaired blastocyst formation is that the two ROCK isoforms, ROCK1 and ROCK2, have different stage-specific expression patterns and subcellular localizations during preimplantation development and, therefore, can have different biological activities (Zhang et al., 2014). ROCK1 is predominantly detected in the cytoplasm and localized mainly at the cell–cell boundaries at the four-cell to morula stages. However, ROCK1 is barely detected in the blastocyst, while ROCK 2 is present in both the cytoplasm and the nucleus with predominant expression in the TE nuclei (Zhang et al., 2014) (see schematic representation of expression and localization of ROCK1 and ROCK2 in porcine embryos in Fig. 4). Considering the spatial and temporal expression patterns of ROCK1 and ROCK2 and the stage-dependent impairment of embryo development by ROCK inhibition, we suggest that the arrest of embryos at pre-compaction is associated with inhibition of ROCK1 rather than ROCK2, because ROCK1 is involved in cell adhesion via regulating actomyosin contraction and association with E-cadherin complexes (Shi et al., 2013; Smith et al., 2012).
Next, we examined subcellular localization of CXADR, which was recently reported to be required for adherens and tight junction formation during porcine blastocyst development (Kwon, Kim & Choi, 2016) and shown to directly associated with ROCK1 and ROCK2 in human carcinoma cells (Saito et al., 2014). The observed disruption of localization of the CXADR protein and reduced expression of the CXADR gene indicated that clustering of junctional proteins complex and its functional barrier can be mediated by CXAD.R. Thus, we examined single-pass membrane protein (CXADR), four-pass membrane domain (OCLN), and cytoplasmic plaque protein (TJP1), and adherens protein (CDH1), found a significant decrease in the protein and mRNA expression levels, and demonstrated increased paracelluar permeability in the ROCK inhibitor treated blastocysts (Table 1). The deficient barrier function in blastocysts was likely caused by the disruption of tight junction assembly and deregulation of gene expression on the morula-to-blastocyst developmental stage. A growing body of evidence has demonstrated that adherens and tight junctions are both linked intracellularly to F-actin via adaptor proteins such as TJP1 and OCLN, which are phosphorylated by ROCK (Aijaz, Balda & Matter, 2006; Balda et al., 1996; Gopalakrishnan et al., 1998; Hartsock & Nelson, 2008; Hirase et al., 2001; Ohnishi et al., 2004; Sahai & Marshall, 2002; Wittchen, Haskins & Stevenson, 1999; Yamamoto et al., 2008). In addition, nuclear ROCK2 has been reported to be involved in transcriptional regulation of the trophoblast lineage and co-localized with the p300 acetyltransferase (Carey et al., 2013; Rodriguez et al., 2004; Tanaka et al., 2006). Converging lines of evidence suggest that inhibition of cytoplasmic ROCK1 activity in morula can adversely affect the interaction of cytoskeleton F-actin with TJP1 and phosphorylation of the carboxy-terminal domain of OCLN, which is important for paracellular sealing of the TE during the transition stage (Fig. 4B). Furthermore, inhibition of nuclear ROCK2 activity may lead to reduced expression of the CXADR, OCLN, TJP, and CDH1 genes via transcriptional regulation associated with p300 activity (Fig. 4B).
In summary, we demonstrated that inhibition of ROCK activity impairs early porcine embryo developmental steps such as compaction and cavitation. We suggest that observed defective embryo development and lower gene expression of AJ/TJ associated genes in presence of the specific ROCK inhibitor are caused by specific temporo-spatial inhibition of distinct ROCK1 and ROCK2 functions in porcine embryos. In this context, it will be important to establish the molecular mechanisms of differential subcellular targeting of ROCK isoforms during intracellular polarization and adherens/tight junction biogenesis throughout early embryo development.
Fig1A raw data
The blastocyst development rates of the four-cell stage embryos exposed to different concentrations of Y-27632 (control 0, treatment 10, 20, and 100 µM).
Fig2A raw data
The fractions of embryos treated with 20 µM Y-27632 at the four-cell or morula stages that developed to the morula and blastocyst stages, respectively, at 96 and 120 h.p.a.
Fig2C raw data
Total cell number of embryos at 120 h.p.a after the treatment of morula embryos with 20 µM Y-27632.