Expression of RND Proteins in Human Myometrium

4 downloads 0 Views 517KB Size Report
Mar 22, 2006 - Division of Obstetrics and Gynaecology,3 Henry Wellcome Laboratories for Integrative Neuroscience and. Endocrinology, Clinical Sciences at ...
BIOLOGY OF REPRODUCTION 75, 452–461 (2006) Published online before print 22 March 2006. DOI 10.1095/biolreprod.105.049130

Expression of RND Proteins in Human Myometrium1 J. Lartey,3 A. Gampel,4 J. Pawade,5 H. Mellor,4 and A. Lo´pez Bernal2,3 Division of Obstetrics and Gynaecology,3 Henry Wellcome Laboratories for Integrative Neuroscience and Endocrinology, Clinical Sciences at South Bristol, University of Bristol, Bristol, BS1 3NY, United Kingdom Department of Biochemistry,4 School of Medical Sciences, University of Bristol, Bristol, BS8 1TD, United Kingdom Department of Pathology,5 Bristol Royal Infirmary, University of Bristol, Bristol, BS2 8HW, United Kingdom to the activation of calmodulin-dependent myosin light chain kinase (MYLK) [1]. MYLK phosphorylates the regulatory myosin light chain (MYL), enhancing actin-myosin ATPase activity to cause contraction [2, 3]. However, during agonistinduced contractions, the increase in MYL phosphorylation and tension observed is usually higher than the corresponding increase in [Ca2þ]i. This increase in calcium-independent tension is termed ‘‘Ca2þ sensitization’’ [4, 5]. Dephosphorylation of the phosphorylated MYL by a myosin phosphatase results in relaxation [6]. All myosin phosphatases consist of a protein phosphatase 1 subunit (PP1) and both a large and a small myosin targeting subunit. The large myosin phosphatasetargeting subunit (known as PPP1R12A) regulates the action of PP1[7]. GTP binding proteins of the RHO family are key regulators of the actin cytoskeleton and actin stress fiber formation [8]. RHO GTPases cycle between GTP-bound (active) and GDPbound (inactive) conformations, controlled by three classes of regulators. Guanine nucleotide exchange factors (GEFs) catalyze the exchange of GTP for GDP, thereby activating the GTPase and enabling it to interact with potential effectors. GTPase activating proteins (GAPs) accelerate the hydrolysis of GTP, returning the GTPase to its inactive state [9]. Guanine nucleotide dissociation inhibitors (GDIs) interact with GDPbound GTPase, sequestering it in an inactive complex [10, 11]. The active GTP-bound form of RHOA activates RHOassociated protein kinase (ROCK), which phosphorylates PPP1R12A at Thr 696 [12], inhibiting the phosphatase activity. Moreover, ROCK directly phosphorylates MYL at Ser 19 [13, 14]. Both mechanisms lead to an increase in MYL phosphorylation and Ca2þ independent increase in tension [4, 5]. Thus, stimulation of the RHOA-ROCK pathway sensitizes myometrium to the effects of agonists by the process of Ca2þ sensitization [15, 16]. We and others have demonstrated a high expression of RHOA-associated kinase (ROCK1) in human myometrium [17, 18]. Furthermore, exposure of human uterine smooth muscle cells to agonists like thromboxane induces an increase in ROCK1 [19]. In rat, mouse, and human myometrium, the ROCK inhibitor Y27632 affected oxytocin-, thromboxane-, and carbachol-induced increases in tension and MYL phosphorylation without a reduction in [Ca2þ]i [5, 20, 21; also see Note Added in Proof]. Recently, the use of Y27632 in pregnant rats with lipopolysaccharide and prostaglandin F2alpha induced preterm labor, leading to lower delivery rates [22]. Thus, the tocolytic potential of agents that negatively modulate the RHOA-ROCK pathway may be useful in treating preterm labor. The RND small GTP binding proteins (RND1, RND2 and RND3) are members of the RHO family that lack intrinsic GTPase activity and can inhibit RHOA-stimulated stress fiber formation [23–25]. RND proteins are constitutively active and inhibit the RHOA-ROCK system either by binding to ROCK

ABSTRACT RHO GTPases are key regulators of the actin cytoskeleton and stress fiber formation. In the human uterus, activated RHOA forms a complex with RHO-associated protein kinase (ROCK) which inhibits myosin light chain phosphatase (PPP1R12A), causing a calcium-independent increase in myosin light chain phosphorylation and tension (Ca2 þ sensitization). Recently discovered small GTP binding RND proteins can inhibit RHOA and ROCK interaction to reduce calcium sensitization. Very little is known about the expression of RND proteins in the human uterus. We tested the hypothesis that the uterine quiescence observed during gestation is mediated by an increase in RND protein expression inhibiting RHOA–ROCK-mediated PPP1R12A phosphorylation. Immunohistochemistry and immunoblotting were used to determine RHOA and RND protein expression and localization in nonpregnant, pregnant nonlaboring, and laboring patients at term and patients in spontaneous preterm labor. Changes in protein expression estimated by densitometry between different patient groups were measured. A significant increase of RND2 and RND3 protein expression was observed in pregnant relative to nonpregnant myometrium associated with a loss of PPP1R12A phosphorylation. RND transfected myometrial cells demonstrated a dramatic loss of stress fiber formation and a ‘‘rounding’’ phenotype. RND upregulation in pregnancy may inhibit RHOA– ROCK-mediated increase in calcium sensitization to facilitate the uterine quiescence observed during gestation. female reproductive tract, mechanisms of hormone action, parturition, pregnancy, RHOA, RND proteins, uterine quiescence, uterus

INTRODUCTION The human uterus, unlike other types of smooth muscle, has unique qualities which allow it to remain quiescent, despite huge changes in size and tension caused by the growing fetus. Understanding the biochemical events that underlie this prolonged quiescence and lead to the initiation of labor promises more effective treatment of conditions like preterm labor, which remains a major cause of perinatal mortality and morbidity. Uterine smooth muscle contraction is primarily regulated by an increase in intracellular free calcium ([Ca2 þ]i), which leads 1 Supported by a WellBeing of Women (Royal College of Obstetricians and Gynaecologists) Research Training Fellowship to J.L. 2 Correspondence: FAX: 44 117 928 5290; e-mail: [email protected]

Received: 6 November 2005. First decision: 30 November 2005. Accepted: 22 March 2006. Ó 2006 by the Society for the Study of Reproduction, Inc. ISSN: 0006-3363. http://www.biolreprod.org

452

453

RND PROTEINS IN HUMAN MYOMETRIUM TABLE 1. Clinical details of the women whose samples were used in this study.a Patient groups Parameter Patient Age (yr) Gestational Age (wk) Newborn weight (g) APGAR 1 min APGAR 5 min Cervical dilatation (cm) Duration of labor (min)

NP (n ¼ 10)

NIL (n ¼ 10)

40.1 (36–46)*

30.5 (25–37) 38.8 (38–41) 3632 (2980–4320) 8.2 (4–9) 9.3 (6–10) Not determined 0

SL (n ¼ 10) 32.3 39.1 3560 8.1 9.7 6.9 491.0

(23–40) (37–40) (3000–4080) (4–10) (8–10) (4–9) (67–1050)

SPT (n ¼ 9) 30.3 33.1 2254 7.7 9.2 6. 2 428.4

(19–42) (28–36)** (1385–3380)** (4–10) (7–10) (4–10) (50–963)

a

The values are means and ranges; n ¼ number of patients. * P , 0.05–0.01 compared to NIL, SL, and SPT. ** P , 0.01–0.001 compared to NIL and SL.

and preventing it from phosphorylating downstream targets [26] or by increasing the activity of RHO GAPs (ARHGAP), leading to lower levels of GTP bound RHOA [27, 28]. RND1 and RND3 have been described in rat and rabbit myometrium [29, 30], and their levels of expression can be increased by estrogen and progesterone treatment. These increases in RND protein expression are functionally coupled to an increase in the inhibition of agonist- and GTP-induced Ca2þ sensitization during gestation [30]. There is no information regarding the expression of small GTPases other than RHOA in human myometrium. The purpose of this paper is to investigate the presence of RND proteins in human myometrium and to consider the possibility that they might contribute to the uterine quiescence observed during pregnancy. MATERIALS AND METHODS Reagents and Antibodies Primary antibodies and immunizing peptides were purchased as follows: RHOGDI (ARHGDIA) (sc.360), RHOA (sc.418), RND2 (sc.1945) (earlier known as RHO7), ROCK1 (sc.5560) and ROCK2 (sc.5561) from Santa Cruz Biotechnology and alpha-tubulin (TUBA1) mouse monoclonal antibody (T5168) from Sigma. A RND1 rabbit polyclonal antibody (donated by Dr. C. Nobes, University of Bristol) was used for RND1 immunohistochemistry. A RND3 rabbit monoclonal antibody (05–723) (Upstate) was used to detect both RND1 and RND3 [24, 26, 31]. Myosin phosphatase rabbit polyclonal antibody (PRB-457C-200) for total PPP1R12A was purchased from Cambridge Bioscience and anti-phosphoPPP1R12A-Thr 696 rabbit polyclonal antibody (07–251) was purchased from Upstate. Secondary HRP-conjugated antibodies—polyclonal rabbit anti goat (P0160), swine anti rabbit (P0399), and rabbit anti mouse (P0260)—were obtained from DAKO Ltd. pRK5 myc RND2 and RND3 cDNA constructs [24] were obtained from Dr. C. Nobes (University of Bristol). Blotto A was purchased from Santa Cruz Biotechnology, RTU Vectastain Quick Universal Kit (PK-7800) from Vector Laboratories Ltd., Optimax buffer from BioGenex, and Optimem media and Lipofectamine from Gibco BRL. PVDF membrane and Protein Kaleidoscope molecular weight markers were acquired from Bio-Rad Laboratories Ltd. ECL Plus and hyperfilm were purchased from Amersham Biosciences.

Human Myometrial Tissue Collection Myometrial tissue was obtained from nonpregnant (NP) premenopausal women undergoing hysterectomy for benign gynecological disorders and from nonlaboring (NIL) pregnant women at term (37 to 40 wk gestation), undergoing caesarean section with the following indications: maternal request, breech presentation, previous caesarean section, or placenta previa. Myometrial tissue was obtained from the upper border of the uterine incision, taking care to exclude serosal, decidual, or scar tissue. Myometrial tissue was also obtained from women in spontaneous labor at term (SL) and spontaneous idiopathic preterm labor (,37 wk gestation) (SPT) with no evidence of infection. Labor was defined by regular uterine contractions and cervical dilation greater than 4 cm. Patients with underlying medical complications of pregnancy were excluded. The patients’ age, gestational age at delivery, mode of delivery,

newborn weight, cervical dilation, duration of labor, and APGAR scores are outlined in Table 1. NP women were older than women in the pregnant NIL, SL, and SPT groups, but duration of labor cervical dilation and other parameters were similar. As expected, gestational age was shorter and newborn weight less in the SPT group compared to NIL and SL groups. Decidua, placenta, and cord blood samples were also collected for comparisons. Ten patients were recruited into NP, NIL, and SL groups, and nine into the SPT group. Tissues were snap-frozen and stored in liquid nitrogen. The study had the approval of the United Bristol Healthcare Trust Research Ethics Committee. All patients gave informed consent.

Immunohistochemistry Paraffin-embedded tissue sections were dewaxed in xylene, hydrated in alcohol, and incubated (in 0.3% hydrogen peroxide in PBS) for 10 min to block endogenous peroxidase activity. Microwave antigen retrieval in citrate buffer (0.8 mM citric acid, pH 6.0, 82 mM trisodium sulfate) was at 958C for 20 min, followed by blocking with 2.5% horse serum for 30 min to reduce nonspecific binding. The sections were incubated with primary antibody diluted with 10% human serum in PBS for 45 min at room temperature, biotinylated secondary antibody for 20 min, streptavidin peroxidase reagent for 20 min, and diaminobenzidine chromogen solution for 5 min, and then were counterstained in Mayers Haematoxylin for 1 min, dehydrated in alcohol, immersed in xylene, and mounted. The following primary antibodies were used for immunohistochemistry: RHOA mouse monoclonal antibody (1:400 dilution), RND1 rabbit polyclonal antibody (1:100 dilution), RND2 goat polyclonal antibody (1:50 dilution), and RND3 rabbit antibody (1:100 dilution). Prediluted biotinylated pan-specific secondary antibody (Vectastain Quick Universal Kit) was added according to the manufacturer’s instructions. Liver, brain, and placental tissues were used as positive controls for protein expression. The primary antibody was omitted from the staining protocol as a negative control.

Whole Tissue Homogenates Myometrial samples were homogenized using a mechanical homogenizer in ice-cold lysis buffer (100 mM Tris pH 7.5, 5 mM EDTA, 50 mM NaCl, 30 mM sodium pyrophosphate, 50 mM NaF, 1% Triton X-100, 100 lM sodium vanadate, 10 mM PMSF, 10 mM DTT). Proteins were separated from nuclei and non-broken cells by centrifugation (10,000 g for 1 h at 48C). The supernatant was stored at 80 8C until use.

Western Blotting and Densitometry. Homogenate proteins (100 lg) were separated by SDS-PAGE and transferred to PVDF using a semidry blotting system. Membrane-bound proteins were blocked with Blotto A before Western blotting with primary antibody and a compatible secondary antibody. Chemiluminescence reactions were carried out using ECL Plus. Control experiments using immunizing peptide to pre-absorb primary antibody led to loss of the immunoreactive bands. Immunoblots of different patient samples were carried out under conditions in which the intensity of the bands obtained was proportional to protein concentration and ECL reaction. Repeated experiments on the same patient samples performed on different days showed similar results. We used TUBA1 and ARHGDIA as markers of equal protein loading in each sample, but TUBA1 was preferred because its molecular weight was higher and did not interfere with the visualization of RHO and RND GTP binding proteins. All samples were analyzed in duplicate and normalized to their own TUBA1

454

LARTEY ET AL.

FIG. 1. Immunohistochemical localization of RND1, RND2 and RND3 proteins in nonpregnant and pregnant uterine tissue sections. There was increased cytoplasmic peroxidase reaction of the nonpregnant endometrial glands (ESG) relative to endometrial stroma (ES) for RND1 (A) and RND3 (G). This is in comparison to very weak signal for RND2 in both ES and ESG (D). There was a uniform cytoplasmic staining for RND1 in both nonpregnant and pregnant myometrium (B–C). RND2 and RND3 showed a similar pattern of intracellular staining, and both proteins demonstrated more intense myometrial staining in pregnancy (RND2 [E–F]) and (RND3 [H–I]). There was strong staining of lymphoid cells (LN) but no staining of blood cells (BC). Corresponding control reactions with the primary antibody step excluded are presented in J–L. The figures are representative of samples taken from five different women in each patient group. Original magnification 3400; bar ¼ approximately 50 lm.

signal. Densitometry analysis was performed using Image J Program (Wright Cell Imaging Facility, Toronto, Canada).

Human Uterine Smooth Muscle Cell Culture, Transfection, and Immunofluorescence Myometrial cells were isolated using a minced tissue explants method [32] and cultured as described by Phaneuf et al. [33]. Pre-confluent cells were seeded onto cover slips and transfected with eukaryotic expression vectors pRK5 encoding myc-tagged RND2 or RND3 using Lipofectamine according to the supplier’s instructions. Transfected cells were fixed in 4% (wt/vol) paraformaldehyde, permeabilized with 0.2% Triton X-100, and treated with fresh sodium borohydride. The fixed cells were stained for filamentous actin using Alexa 488 labeled phalloidin, and for myc epitope-tagged RND2 and RND3 proteins using an anti-myc antibody (9E10) followed by FITC-labeled secondary antibody. Statistical Analysis. Data were analyzed using the GraphPad Prism Program (Hearne Scientific Software). RND protein signals were not normally distributed and the data were log transformed before analysis by ANOVA with Tukey’s post hoc analysis.

staining was positive only in NIL myometrial fibers and largely absent in NP fibers (Fig. 1 E-F). RND1 immunoreactivity was similar in both NP and NIL sections. There was strong staining of both RND1 and RND3 in nonpregnant endometrial glands, but the lack of a signal for RND2 implied that RND2 may be localized to pregnant myometrial fibers only. RHOA demonstrated an intense cytoplasmic immunoreactivity of myometrial fibers in both nonpregnant and pregnant myometrium (Fig. 2). There was also very strong cytoplasmic staining of NP endometrial glands and the surrounding stroma. Desmin (smooth muscle specific) was used as a myometrial marker. Control experiments also confirmed that RND1 and RND3 proteins were expressed in brain and liver [24], whereas RND2 was highly expressed in the brain [34, 35] (data not presented). Incidentally, there was very high RND protein expression in lymphocytes in both nonpregnant and pregnant tissue sections (Fig. 1). It is likely that RND2 is involved in lymphocyte function by controlling cell motility and adhesion [24].

RESULTS Immunohistochemical Localization of RND and RHOA Proteins in Uterine Tissue Sections We observed positive staining reactions in myometrial fibers using the RND1, RND2, and RND3 antibodies. Interestingly, the observed cytoplasmic immunoreactivity was increased in the pregnant (NIL) relative to the nonpregnant (NP) tissue sections for both RND2 and RND3 (Fig. 1). RND2

RND2 Protein Expression is Confined to Myometrial Tissue RHOA was highly expressed in myometrial, placental, and decidual tissues. Moreover, it was the only GTPase examined that was present in cord blood (Fig. 3). Blood and nonmyometrial tissue contamination may be a potential confounding factor on RHOA GTPase expression in myometrial homogenates. RND1 and RND3 were expressed in myometrial, placental, and decidual samples at 27 and 29 kDa,

455

RND PROTEINS IN HUMAN MYOMETRIUM

FIG. 2. Immunohistochemical localization of RHOA in nonpregnant and pregnant uterine tissue sections. A: Nonpregnant endometrium demonstrates an intense RHOA immunoreactivity in the cytoplasm of both stromal cells (ES) and endometrial glands (ESG). B–C: Nonpregnant and pregnant myometrial fibers (MYO) show an intense cytoplasmic granular staining pattern. D: There was a strong myometrial desmin reaction, associated with negative staining in endometrial gland (ESG) and endometrial stroma (ES) in nonpregnant endometrium. E–F: There was a strong desmin reaction in nonpregnant (dense compact fibers) and pregnant (less compact, hypertrophied fibers) tissue sections. There was no desmin staining of blood cells (BC). G–J: Control sections with primary antibody step omitted. The figures are representative of samples taken from five different women in each patient group. Original magnification 3400; bar ¼ approximately 50 lm.

respectively, but were not present in cord blood. Occasionally (1 in 40 myometrial samples examined), RND1 and RND3 bands were shifted down the gel, probably as a result of posttranslational modification (Fig. 3, first lane). Unlike RND1 and RND3, RND2 was expressed only in pregnant myometrial samples. This suggests that RND2 has a specific myometrial function in pregnancy, although the idea needs further investigation. Increase in RND2 and RND3 Protein Expression in Pregnant versus Nonpregnant Human Myometrium We used immunoblotting to determine the relative expression of RHOA, RND1, RND2, and RND3 proteins in myometrium from four groups of patients. The protein expression of TUBA1 and ARHGDIA used to ensure equal protein loading was constant across all four groups (Table 2 and Fig. 4). The RHOA antibody detected a band at 25 kDa which was present in both nonpregnant and pregnant myometrium, and there were no significant changes in relative expression among the four groups (Table 2 and Fig. 4). The RND2 antibody detected a band at 27 kDa. In contrast to RHOA, pregnancy induced a large increase in RND2 expression in term nonlaboring (NIL, P , 0.001), term in labor (SL, P , 0.001), and preterm laboring (SPT, P , 0.001) samples (Table 2 and Fig. 4). There were no significant differences in RND2 expression among the pregnant NIL, SL, and SPT groups. RND3 and RND1 were detected using a rabbit RND3 antibody which cross-reacts with both proteins, giving a strong band at 27 kDa corresponding to RND1 and a heavier band at 29 kDa corresponding to RND3 (Fig. 5) [26]. The level of RND1 tended to be higher in pregnant compared to nonpregnant myometrial samples, but the difference did not reach statistical significance (P . 0.05) (Table 2 and Fig. 5).

RND3 expression increased significantly in the pregnant NIL (P , 0.001), SL (P , 0.001), and SPT (P , 0.001) samples relative to the NP samples (Table 2 and Fig. 5). There was no change in RND3 expression between the nonlaboring NIL and laboring SL and SPT samples. ROCK1 and ROCK2 Expression is Constant in Nonpregnant and Pregnant Myometrium We further explored the significance of the differences between RHOA and RND protein expression above by

FIG. 3. RHOA, RND1, RND2, and RND3 protein expression in myometrium (MYO); decidua basalis (DB), decidua parietalis (DP), cord blood (PCB), and placenta (PLA). The immunoblotting demonstrates RHOA expression (25 kDa) in all the tissues analyzed and RND1 and RND3 in all but PCB. RND3 band is at 29 kDa and RND1 at 27 kDa. TUBA1 (50 kDa) was used as a marker of equal protein loading. Note that in contrast to the other proteins, RND2 is only expressed in myometrium and not in any of the other tissues analyzed. This experiment shows immunoblots on separate samples from two donors. Another donor was used for the comparison between myometrium and placenta. Similar results were obtained in three independent experiments with samples from other donors.

456

LARTEY ET AL.

FIG. 4. Immunoblotting and densitometry analysis of RND2 and RHOA proteins in human myometrium. Ten independent samples were used for each group of nonpregnant (NP), pregnant not in labor (NIL), spontaneous labor (SL), and nine samples for spontaneous preterm labor (SPT). Each of the samples was analyzed in duplicate, normalized to TUBA1 and plotted on a logarithmic scale within the allotted patient groups. Representative duplicate immunoblots of RHOA and RND2 are presented in A and densitometry graphs in B. To demonstrate equal protein loading, the corresponding TUBA1 and ARHGDIA immunoblots (A) of individual samples in each patient group were analyzed in duplicate and plotted in C. The bars represent means (*p ,0.001).

measuring the expression of their effector proteins ROCK1 and ROCK2 in the four patient groups. ROCK1 and ROCK2 antibodies detected bands at 150 and 160 kDa, respectively. There were no significant differences in ROCK1 and ROCK2 expression across the four patient groups (Table 2 and Fig. 6). Upregulation of RND2 and RND3 in Pregnancy is Associated with a Loss of Phosphorylation of the Myosin Phosphatase Regulatory Subunit PPP1R12A The myosin phosphatase holoenzyme is composed of three subunits: a 38 kDa catalytic protein phosphatase subunit (PP1c); a large 110–130 kDa regulatory myosin binding subunit (PPP1R12A), the target site of ROCK inhibitory phosphorylation; and a small 20 kDa subunit [16, 36]. RND proteins inhibit RHOA-mediated activation of ROCK by binding to a site which prevents ROCK-RHOA interaction. As a result, there is a reduction in ROCK-mediated phosphorylation of PPP1R12A [37]. We compared RHOA and RND protein expression with the phosphorylation state of PPP1R12A by using a phosphospecific antibody to Thr 696, the phosphorylation site of ROCK. Total nonphosphorylated myosin phosphatase was visualized using a PPP1R12A antibody. The gestation-related upregulation of expression of RND2 and RND3 was associated

with a loss of PPP1R12A phosphorylation in the pregnant samples (Fig. 7), indicating an increase in the activity of myosin phosphatase in pregnancy. This is in contrast to a constant level of expression of nonphosphorylated PPP1R12A in both nonpregnant and pregnant samples. This loss of PPP1R12A phosphorylation occurred despite the constant high level of ROCK expression in pregnant myometrial samples (Fig. 7), [17], [18]. By contrast, we noted an increase in phosphorylation of PPP1R14A, a PP1 inhibitory subunit, in pregnant relative to nonpregnant samples. PPP1R14A (also known as CPI-17) is another potential mediator of Ca2þ sensitization whose activity is dependent on protein kinase C (PRKCB) phosphorylation [16, 38]. Phosphorylated PPP1R14A inhibits myosin phosphatase by binding to its catalytic PP1c subunit [39, 40]. This finding supports data previously reported by Ozaki et al. in human myometrium [41] which links a pregnancy-related increase in PPP1R14A (CPI-17) with enhanced capacity for PRKCB-induced contraction and MYL phosphorylation. RND2 and RND3 Cause Loss of Actin Stress Filaments and ‘‘Rounding’’ of Human Myometrial Cells We examined the biological effects of RND proteins on the phenotype of cultured myometrial cells. Serum-starved cells

457

RND PROTEINS IN HUMAN MYOMETRIUM TABLE 2. Densitometric analysis of RHOA, RND1, RND2, RND3, ROCK1, and ROCK2 protein expression in NP, NIL, SL, and SPT.a Patient groups Proteins RHOA RND1 RND2 RND3 ROCK1 ROCK2 ARHGDIAb TUBA1

NP (n ¼ 10) 118 14 2 4 164 161 180 186

(60–191) (1–102) (1–6) (2–9) (98–203) (132–192) (172–185) (176–192)

NIL (n ¼ 10) 83 42 33 30 162 148 180 192

(22–173) (12–89) (6–76)* (11–43)* (103–205) (100–183) (175–190) (177–207)

SL (n ¼ 10) 71 48 29 39 161 148 178 190

(7–181) (24–126) (8–47)* (18–151)* (103–207) (103–190) (171–190) (170–202)

SPT (n ¼ 9) 87 65 31 43 169 157 183 194

(19–184) (23–107) (9–63)* (4–128)* (139–206) (127–195) (166–192) (182–209)

a Individual patient samples were analyzed in duplicate and expressed as arbitrary densitometry units. The values presented are geometric means and ranges; n ¼ number of samples in each group. b ARHGDIA was measured in five patients in each group. * P , 0.001 compared to NP.

transfected with pRK5 vectors expressing myc-tagged RND2 and RND3 had fewer stress fibers and retracted to produce a rounded phenotype (Fig. 8). More than 70% of transfected myometrial cells with either RND2 or RND3 were rounded and demonstrated fewer stress fibers compared to only 10% of RND-negative cells (Fig. 8, panel B). Transfection efficiency of RND2 and RND3 in duplicate coverslips was 15%–20% per experiment. No differences in phenotype were observed on transfection of RHOA. The results are representative of three different experiments. DISCUSSION RND proteins are members of the RHO family which lack intrinsic GTPase activity and were first characterized by their ability to cause cell rounding by inhibiting the formation of RHOA-dependent actin stress fibers [24]. RND proteins are constitutively active and are regulated by their level of expression [24, 37]. We propose that the balance of expression of RHOA and RND proteins is likely to determine the level of

activity of ROCK and thus Ca2þ sensitization in human myometrium. To the best of our knowledge, this is the first demonstration of RND1, RND2, and RND3 proteins in human myometrium and an upregulation in the expression of RND2 and RND3 proteins in pregnancy. This is consistent with previous reports demonstrating similar pregnancy-induced increases of RND1 in rat [29] and RND3 in rabbit myometrium [30] and an increase in RND1, RND2, and RND3 mRNA in rat and human myometrium [42]. Interestingly, the increase in RND2 expression in pregnant NIL samples is also sustained in the laboring SL and SPT samples. RND2 was the only RND related protein that was expressed in myometrium but not in other intrauterine tissues examined. However, very little is known about the function of RND2 in the uterus. RND2 is primarily expressed in the testis [24, 35], neuronal cells, and hepatic cells [34]. In the testis, RND2 binds to a GTPase activating protein, namely RACGAP1, required for cytokinesis in male germ cells [35]. In neuronal cells, the binding of constitutively active RND2 to rapostlin, an RND2 specific FIG. 5. RND1 and RND3 protein expression in human myometrium. Immunoblotting for RND1 and RND3 was carried out using a rabbit RND3 antibody which recognizes RND3 at 29 kDa and crossreacts with RND1 at 27 kDa. Representative duplicate immunoblots are presented in A and densitometry data of RND1 and RND3 normalized to TUBA1 are presented in B on a logarithmic scale. The bars represent means (*P ,0.001).

458

LARTEY ET AL.

FIG. 6. ROCK1 and ROCK2 protein expression in pregnant human myometrium. Immunoblotting (A) and densitometry data (B) for RHOA effectors ROCK1 and ROCK2, using antibodies which recognize ROCK1 at 150 kDa and ROCK2 at 160 kDa. Duplicate samples in each group were normalized to TUBA1 and plotted on a logarithmic scale (B). The bars represent means.

effector, induces neurite branching by interacting with neural Wiskott-Aldrich syndrome protein [43]. Moreover RND2 binds vacuolar proteins involved in vesicular trafficking [44]. The functional significance of the pregnancy-related increase in RND3 protein expression is not known either. Similar increases in RND3 expression in pregnant rabbit myometrium, described by Cario-Toumaniantz et al. [30], were associated with the restoration of RND3-inhibited Ca2 þ sensitization by farnesyl-transferase inhibitors at midpregnancy but not in late pregnancy [30]. RND3 binds to ROCK and inhibits ROCK phosphorylation of PPP1R12A [26]. This loss of ROCK-mediated inhibition of myosin phosphatase activity via PPP1R12A is consistent with the hypothesis that FIG. 7. Pregnancy-induced changes in the phosphorylation of myosin phosphatase regulatory subunit PPP1R12A in human myometrium. The tissue homogenates were resolved by SDS PAGE and the proteins identified using immunoblotting. The molecular weights of the proteins are indicated in the left in kilodaltons and their identities on the right. Three myometrial samples each from different patients from NP and NIL groups are shown. The results are representative of six experiments with tissue from different donors.

upregulation of RND proteins mediates the relative uterine quiescence observed during gestation. RND3 itself is a substrate for ROCK which phosphorylates it and improves its half life [37]. Further research is necessary to investigate RND3-ROCK interaction in pregnant myometrium. This is also the first demonstration of RND2 and RND3 causing a loss of stress fibers and the induction of a ‘‘rounding’’ phenotype in cultured human myometrial cells. This may be a myometrial-specific effect given that RND2 does not inhibit stress fiber formation in 3T3 fibroblasts [24, 35]. The mechanism of this inhibition requires further investigation, but it is interesting to note that RND2 induces the reorganization of F-actin fibers, directly opposing the effects

RND PROTEINS IN HUMAN MYOMETRIUM

459

FIG. 8. Effects of RND2 and RND3 on actin filament assembly in human myometrial cells. The cells were transfected with pRK5 vectors expressing RND2 and RND3. The permeabilized cells were stained for filamentous actin, myc-tagged RND2 and RND3 (A). Note the dramatic rounding effect and loss of stress fiber formation produced by RND2 and RND3 in myometrial cells (A). B: Cell counts of RND positive (n ¼ 80) and RND negative (n ¼ 200) myometrial cells with an altered phenotype. This result is representative of three different experiments. The bars represent means 6 SEM.

of RHOA, which is a negative regulator of neurite branching [43]. This is a precedent of RND2 inhibiting RHO-mediated effects but it is worthwhile noting that these interactions may be tissue specific. In contrast to RND2 and RND3, RND1 expression was similar in nonpregnant and pregnant myometrial samples. This suggests that RND2 and RND3 rather than RND1 are likely to be the negative regulators of RHOA-mediated Ca2 þ sensitization in human myometrium. Our data also show no significant change in RHOA and ROCK expression in relation to pregnancy or labor, and the overall pattern of expression is in agreement with other reports [18; also see Note Added in Proof]. This is the first report of RHOA, ROCK, and RND protein expression in human myometrium directly comparing NIL, SL, and SPT tissues. The data show that the onset of term or preterm labor is not associated with changes in RHOA or ROCK expression in myometrium. This would imply that the increase in uterine contractility characteristic of labor is not likely to be regulated at the level of RHOA/ ROCK protein expression. We have not measured RHOA activation directly, but it is thought that only a small proportion of available RHOA needs to be in the GTP-bound form to activate ROCK [45]. Moreover, it is possible that there are changes in agonistinduced Ca2 þ sensitization, for example in response to

oxytocin stimulation [46] with the onset of term or preterm labor. Such changes would only be detected by simultaneous measurement of [Ca2þ]i and tension in response to contractile agonists in myometrial samples obtained before and after the onset of labor. Moreover, since RND2 and RND3 are constitutively active, the fact that their level of expression does not change in term and preterm labor suggests that their role is likely to be related to pregnancy maintenance rather than to the mechanism of parturition. Since PPP1R12A is a target for ROCK, measurement of PPP1R12A phosphorylation is a good indication of RHOA/ ROCK activation. We suggest that the increase in RND2 and RND3 expression relative to RHOA during gestation may be functionally linked in vivo to enhancement of myosin phosphatase activity during pregnancy as suggested by the loss of phosphorylated PPP1R12A in our pregnant samples. Moreover, additional modulation of myosin phosphatase PP1c activity by the PRKCB pathway in pregnancy is suggested by the increased levels of phosphorylated PPP1R14A in pregnant myometrium. There is a possibility that the activities of RND2 and RND3 are regulated by mechanisms other than expression, such as subcellular localization and phosphorylation levels [23, 24, 37]. RND1 and RND3 are upregulated in response to growth factors and sex steroids (estrogen and progesterone)

460

LARTEY ET AL.

[28, 37, 47]. This may explain the increase in expression observed during gestation. We propose that Ca2 þ sensitization in pregnant myometrium is determined by a complex balance of the activity of differing mechanisms like ROCK and PPP1R14A inhibition of myosin phosphatase, secondary to the relative activity of RHOA and the expression levels of RND GTP binding proteins. This balance might be altered at the onset of labor by mechanisms that need to be investigated. Other possible targets for the RHO family of GTPases need to be explored in human myometrium [16, 45]. In conclusion, we have demonstrated a pregnancy-related upregulation of RND2 and RND3 in human myometrium and the induction of a rounding phenotype and loss of stress fibers in cultured human myometrial cells. The upregulation of myometrial RND protein expression is likely to be linked to an increase of myosin phosphatase activity that enhances uterine relaxation during pregnancy. ACKNOWLEDGMENTS We want to thank Dr. C. Nobes for experimental suggestions and for comments during the preparation of this manuscript. This research is integrated in the SAFE Network of Excellence.

NOTE ADDED IN PROOF Recent publications by Riley et al. show similar results in ROCK expression in the mouse myometrium [48] as well as human [49].

15. 16.

17.

18.

19.

20.

21.

22.

23. 24.

REFERENCES 1. Kamm KE, Stull JT. The function of myosin and myosin light chain kinase phosphorylation in smooth muscle. Annu Rev Pharmacol Toxicol 1985; 25:593–620. 2. Word RA, Stull JT, Casey ML, Kamm KE. Contractile elements and myosin light chain phosphorylation in myometrial tissue from nonpregnant and pregnant women. J Clin Invest 1993; 92:29–37. 3. Gallagher PJ, Herring BP, Griffin SA, Stull JT. Molecular characterization of a mammalian smooth muscle myosin light chain kinase [published erratum appears in J Biol Chem 1992 May 5;267(13):9450]. J Biol Chem 1991; 266:23936–23944. 4. Somlyo AP, Somlyo AV. Signal transduction by G-proteins, Rho-kinase and protein phosphatase to smooth muscle and non-muscle myosin II. J Physiol (Lond) 2000; 522:177–185. 5. Tahara M, Morishige K-i, Sawada K, Ikebuchi Y, Kawagishi R, Tasaka K, Murata Y. RhoA/Rho-kinase cascade is involved in oxytocin-induced rat uterine contraction. Endocrinology 2002; 143:920–929. 6. Ikebe M, Hartshorne DJ, Elzinga M. Phosphorylation of the 20,000-dalton light chain of smooth muscle myosin by the calcium-activated, phospholipid-dependent protein kinase. Phosphorylation sites and effects of phosphorylation. J Biol Chem 1987; 262:9569–9573. 7. Ceulemans H, Bollen M. Functional diversity of protein phosphatase-1, a cellular economizer and reset button. Physiol Rev 2004; 84:1–39. 8. Hall A. Rho GTPases and the actin cytoskeleton. Science 1998; 279:509– 514. 9. Van Aelst L, D’Souza-Schorey C. Rho GTPases and signaling networks. Genes Dev 1997; 11:2295–2322. 10. Hakoshima T, Shimizu T, Maesaki R. Structural basis of the Rho GTPase signaling. J Biochem (Tokyo) 2003; 134:327–331. 11. Ueda T, Kikuchi A, Ohga N, Yamamoto J, Takai Y. Purification and characterization from bovine brain cytosol of a novel regulatory protein inhibiting the dissociation of GDP from and the subsequent binding of GTP to RhoB p20, a ras p21-like GTP-binding protein. J Biol Chem 1990; 265:9373–9380. 12. Kimura K, Ito M, Amano M, Chihara K, Fukata Y, Nakafuku M, Yamamori B, Feng J, Nakano T, Okawa K, Iwamatsu A. K. K. Regulation of myosin phosphatase by Rho and Rho-associated kinase (Rho-kinase). Science 1996; 273:245–248. 13. Amano M, Ito M, Kimura K, Fukata Y, Chihara K, Nakano T, Matsuura Y, Kaibuchi K. Phosphorylation and activation of myosin by Rhoassociated kinase (Rho-kinase). J Biol Chem 1996; 271:20246–20249. 14. Kureishi Y, Kobayashi S, Amano M, Kimura K, Kanaide H, Nakano T, Kaibuchi K, Ito M. Rho-associated kinase directly induces smooth muscle

25.

26. 27.

28.

29.

30.

31. 32.

33.

34.

35.

36.

37.

contraction through myosin light chain phosphorylation. J Biol Chem 1997; 272:12257–12260. Mackay DJG, Hall A. Rho GTPases. J. Biol. Chem. 1998; 273:20685– 20688. Somlyo AP, Somlyo AV. Ca2 þ sensitivity of smooth muscle and nonmuscle myosin II: modulated by G proteins, kinases, and myosin phosphatase. Physiol Rev 2003; 83:1325–1358. Moore F, Da Silva C, Wilde JI, Smarason A, Watson SP, Lopez Bernal A. Up-regulation of p21- and RhoA-aactivated protein kinases in human pregnant myometrium. Biochem Biophys Res Commun 2000; 269:322– 326. Friel AM, Curley M, Ravikumar N, Smith TJ, Morrison JJ. Rho A/Rho kinase mRNA and protein levels in human myometrium during pregnancy and labor. J Soc Gynecol Investig 2005; 12:20–27. Moore F, Lopez Bernal A. Chronic exposure to TXA2 increases expression of ROCKI in human myometrial cells. Prostaglandins Other Lipid Mediat 2003; 71:23–32. Oh JH, You SK, Hwang MK, Ahn DS, Kwon SC, Taggart MJ, Lee YH. Inhibition of Rho-associated kinase reduces MLC20 phosphorylation and contractility of intact myometrium and attenuates agonist-induced Ca2 þ sensitization of force of permeabilized rat myometrium. J Vet Med Sci 2003; 65:43–50. Kupittayanant S, Burdyga T, Wray S. The effects of inhibiting Rhoassociated kinase with Y-27632 on force and intracellular calcium in human myometrium. Pflugers Archiv 2001; 443:112–114. Tahara M, Kawagishi R, Sawada K, Morishige K, Sakata M, Tasaka K, Murata Y. Tocolytic effect of a Rho-kinase inhibitor in a mouse model of lipopolysaccharide-induced preterm delivery. Am J Obstet Gynecol 2005; 192:903–908. Foster R, Hu KQ, Lu Y, Nolan KM, Thissen J, Settleman J. Identification of a novel human Rho protein with unusual properties: GTPase deficiency and in vivo farnesylation. Mol Cell Biol 1996; 16:2689–2699. Nobes CD, Lauritzen I, Mattei M-G, Paris S, Hall A, Chardin P. A new member of the Rho family, Rnd1, promotes disassembly of actin filament structures and loss of cell adhesion. J Cell Biol 1998; 141:187–197. Guasch RM, Scambler P, Jones GE, Ridley AJ. RhoE regulates actin cytoskeleton organization and cell migration. Mol Cell Bio. 1998; 18: 4761–4771. Riento K, Guasch RM, Garg R, Jin B, Ridley AJ. RhoE binds to ROCK I and inhibits downstream signaling. Mol Cell Biol 2003; 23:4219–4229. Wennerberg K, Forget M-A, Ellerbroek SM, Arthur WT, Burridge K, Settleman J, Der CJ, Hansen SH. Rnd proteins function as RhoA antagonists by activating p190 RhoGAP. Current Biology 2003; 13:1106– 1115. Loirand G, Cario-Toumaniantz C, Chardin P, Pacaud P. The Rho-related protein Rnd1 inhibits Ca2 þ sensitization of rat smooth muscle. J Physiol (Lond) 1999; 516:825–834. Kim YS, Kim B, Karaki H, Hori M, Ozaki H. Up-regulation of Rnd1 during pregnancy serves as a negative-feedback control for Ca2 þ sensitization of contractile elements in rat myometrium. Biochem Biophys Res Commun 2003; 311:972–978. Cario-Toumaniantz C, Reillaudoux G, Sauzeau V, Heutte F, Vaillant N, Finet M, Chardin P, Loirand G, Pacaud P. Modulation of RhoA-Rho kinase-mediated Ca2 þ sensitization of rabbit myometrium during pregnancy – role of Rnd3. J Physiol (Lond) 2003; 552:403–413. Solutions UCS.Certificate of Analysis Anti- RHoE/Rnd3 clone 4 catalog # 05–723. Lot #24565. Cavaille F, Fournier T, Dallot E, Dhellemes C, Ferre F. Myosin heavy chain isoform expression in human myometrium: presence of an embryonic nonmuscle isoform in leiomyomas and in cultured cells. Cell Motil Cytoskeleton 1995; 30:183–193. Phaneuf S, Europe-Finner GN, Carrasco MP, Hamilton CH, Lopez Bernal A. Oxytocin signalling in human myometrium. Adv Exp Med Biol 1995; 395:453–467. Nishi M, Takeshima H, Houtani T, Nakagawara K-i, Noda T, Sugimoto T. RhoN, a novel small GTP-binding protein expressed predominantly in neurons and hepatic stellate cells. Molecular Brain Research 1999; 67:74– 81. Naud N, Toure A, Liu J, Pineau C, Morin L, Dorseuil O, Escalier D, Chardin P, Gacon G. Rho family GTPase Rnd2 interacts and co-localizes with MgcRacGAP in male germ cells. Biochem J 2003; 372:105–112. Hartshorne DJ, Ito M, Erdodi F. Myosin light chain phosphatase: subunit composition, interactions and regulation. J Muscle Res Cell Motil 1998; 19:325–341. Riento K, Totty N, Villalonga P, Garg R, Guasch R, Ridley AJ. RhoE function is regulated by ROCK I-mediated phosphorylation. EMBO J. 2005; 24:1170–1180.

RND PROTEINS IN HUMAN MYOMETRIUM 38. Somlyo AV. New roads leading to Ca2 þ sensitization. Circ Res 2002; 91: 83–84. 39. Kitazawa T, Eto M, Woodsome TP, Khalequzzaman M. Phosphorylation of the myosin phosphatase targeting subunit and CPI-17 during Ca2 þ sensitization in rabbit smooth muscle. J Physiol (Lond) 2003; 546:879– 889. 40. Eto M, Senba S, Morita F, Yazawa M. Molecular cloning of a novel phosphorylation-dependent inhibitory protein of protein phosphatase-1 (CPI17) in smooth muscle: its specific localization in smooth muscle. FEBS Letters 1997; 410:356–360. 41. Ozaki H, Yasuda K, Kim Y-S, Egawa M, Kanzaki H, Nakazawa H, Hori M, Seto M, Karaki H. Possible role of the protein kinase C/CPI-17 pathway in the augmented contraction of human myometrium after gestation. Br J Pharmacol 2003; 140:1303–1312. 42. Kim YS, Hori M, Yasuda K, Ozaki H. Differences in the gestational pattern of mRNA expression of the Rnd family in rat and human myometria. Comp Biochem Physiol A Mol Integr Physiol 2005; 142:410– 415. 43. Fujita H, Katoh H, Ishikawa Y, Mori K, Negishi M. Rapostlin is a novel

44. 45. 46. 47. 48. 49.

461

effector of Rnd2 GTPase inducing neurite branching. J Biol Chem 2002; 277:45428–45434. Tanaka H, Fujita H, Katoh H, Mori K, M N. Vps4-A (vacuolar protein sorting 4-A) is a binding partner for a novel Rho family GTPase, Rnd2. Biochem J 2002; 365:349–353. Dvorsky R, Blumenstein L, Vetter IR, Ahmadian MR. Structural insights into the interaction of ROCKI with the switch regions of RhoA. J Biol Chem 2004; 279:7098–7104. Woodcock NA, Taylor CW, Thornton S. Effect of an oxytocin receptor antagonist and Rho kinase inhibitor on the [Ca þþ]i sensitivity of human myometrium. Am J Obstet Gynecol 2004; 190:222–228. Villalonga P, Guasch RM, Riento K, Ridley AJ. RhoE inhibits cell cycle progression and Ras-induced transformation. Mol Cell Biol 2004; 24: 7829–7840. Riley M, Wu X, Baker PN, Taggart MJ. Gestational-dependent changes in the expression of signal transduction and contractile filament-associated proteins in mouse myometrium. J Soc Gynecol Investig 2005; 12:e33–43. Riley M, Baker PN, Tribe RM, Taggart MJ. Expression of scaffolding, signalling and contractile-filament proteins in human myometria: effects of pregnancy and labour. J Cell Mol Med 2005; 9:122–134.