Genetic and Pharmacological Inhibition of Rho-associated Kinase II Enhances Adipogenesis*

Rho-associated kinase (ROCK) regulates reorganization of actin cytoskeleton. During adipogenesis, the structure of filamentous actin is converted from long stress fibers to cortical actin, suggesting that the ROCK is involved in adipogenesis. Two ROCK isoforms have been identified: ROCK-I and ROCK-II. However, pharmacological inhibitors of ROCK cannot distinguish two ROCK isoforms. In the present study, we examined the role of ROCK in adipogenesis and actin cytoskeleton using genetic and pharmacological approaches. Y-27632, which inhibits the activity of both ROCK isoforms, enhanced adipogenesis through the up-regulation of adipogenic transcription factors in 3T3-L1 cells. Furthermore, Y-27632 restored inhibition of adipogenesis by lysophosphatidic acid, which activates Rho. Regarding actin cytoskeleton, Y-27632 disrupted stress fibers in 3T3-L1 preadipocytes. Next, we analyzed adipogenesis of mouse embryonic fibroblasts (MEFs) derived from ROCK-I and ROCK-II knock-out mice, respectively. Adipogenesis of ROCK-II (-/-) MEFs was markedly enhanced compared with wild-type MEFs while that of ROCK-I (-/-) MEFs was not. In contrast to pharmacological approaches, no obvious alteration was found in actin cytoskeleton of ROCK-II (-/-) MEFs compared with wild-type MEFs. In 3T3-L1 cells, knockdown of ROCK-II by RNA interference enhanced the expression of adipogenic transcription factors while that of ROCK-I did not. Moreover, Y-27632 inhibited IRS-1 serine phosphorylation and enhanced Akt phosphorylation in 3T3-L1 preadipocytes. Similarly, Akt phosphorylation in ROCK-II (-/-) MEFs was augmented compared with wild-type MEFs. In conclusion, inhibition of ROCK-II, not ROCK-I, enhances adipogenesis accompanied by the up-regulation of adipogenic transcription factors. Augmentation of insulin signaling may contribute to the enhancement of adipogenesis.

regulates various physiological and pathological processes (17)(18)(19)(20), while they cannot refer to the isoform-specific role of ROCK because ROCK inhibitors inhibit the activity of both ROCK-I and ROCK-II. Furthermore, ROCK inhibitors may potentially inhibit other protein kinases such as protein kinase A (12).
Several studies have reported that Rho-ROCK pathway is involved in adipogenesis (21,22). However, the role of ROCK in adipogenesis still remains to be clarified. We show here the effects of inhibition of ROCK on adipogenesis and examine the molecular mechanisms using pharmacological and genetic approaches. Especially, the adipogenic analysis using MEFs derived from ROCK-I or ROCK-II knock-out mice and ROCK-I or ROCK-II siRNA can provide the evidence about the physiological role of ROCK and the isoform-specific role of ROCK in adipogenesis.
Cell Culture, Adipocyte Differentiation, and Oil Red O Staining-3T3-L1 cells (kindly provided by Dr. H. Green and Dr. M. Morikawa, Harvard Medical School, Boston, MA) were cultured and differentiated into adipocytes as described previously (23). Briefly, cells were grown for 2 days post-confluence (referred as day 0) in 10% CS/DMEM. Differentiation was induced with 10% FBS/DMEM containing 0.5 mM IBMX, 0.25 M dexamethasone, and 1 g/ml insulin for 2 days. The cells were then incubated in 10% FBS/DMEM with insulin for 2 days and maintained hereafter with 10% FBS/DMEM to day 8. Pioglitazone was dissolved in Me 2 SO and added to media within 0.1% of volume. The medium was changed every other day. At day 8, the cells were washed with phosphate-buffered saline (PBS) twice, fixed in 3.7% formaldehyde for 1 h and then stained with 0.6% (w/v) Oil Red O solution (60% isopropyl alcohol, 40% water) for 2 h at room temperature. Cells were then FIGURE 1. ROCK expression and actin cytoskeleton during adipogenesis and effect of ROCK inhibitors on adipogenesis. A, equal amounts of proteins harvested from day 0 to day 8 were subjected to Western blot analysis using antibodies against ROCK-I and ROCK-II. B, 3T3-L1 preadipocytes and mature adipocytes were fixed, permeabilized, and stained with Oregon Green phalloidin for F-actin. Bar, 50 m. C, differentiated adipocytes were fixed and stained with Oil Red O at day 8. Macroscopic and microscopic pictures of cells are shown. D, lipid accumulation was assessed by the quantification of A 510 in destained Oil Red O with isopropyl alcohol. Data are expressed as mean Ϯ S.E. from triplicate experiments. **, p Ͻ 0.01 (Student's t test) compared with dexamethasone and IBMX-treated group. Pio, pioglitazone, V, vehicle.
washed with water to remove unbound dye. Stained Oil Red O was eluted with isopropyl alcohol and quantified by measuring the optical absorbance at 510 nm (24).
Preparation of Primary Mouse Embryonic Fibroblasts and Induction of Adipogenesis-The generation of ROCK-I and ROCK-II knock-out mice was previously reported (25,26). ROCK-II knock-out mice of a mixed background 129/Sv and C57BL/6N genetic background were backcrossed to the C57BL/6N strain for more than 8 generations. Primary MEFs were harvested from 13.5 d.p.c. embryos of each genotype as previously described (27). Differentiation of MEFs was induced with 0.5 mM IBMX, 0.25 M dexamethasone and 10 g/ml insulin for 6 days and maintained with 10% FBS/DMEM to day 8.
Small Interfering RNA (siRNA)-Synthetic siRNAs were purchased from Invitrogen (Carlsbad, CA). siRNAs were delivered into 3T3-L1 cells by a pulse of electroporation with Nucleofector (Amaxa, Gaithersburg, MD). After the electroporation, cells were immediately mixed with the fresh medium for 10 min and reseeded on plates for further examinations. The cells were differentiated 12 h after the electroporation. Negative control siRNA was used as the control of experiments. The targeted nucleotide sequences of the siRNAs were ATTGGTGCTTGT-CAGTTAGGCGTGC (ROCK-I) and ATAATTACTCATAAGGT-TGACGAGG (ROCK-II).
Preparation of Total Cell Lysates and Western Blot Analysis-Cells were washed twice with ice-cold PBS and harvested in a lysis buffer (40 mM HEPES, 10 mM EDTA, 100 mM NaF, 10 mM sodium pyrophosphate, 1 mM Na 3 VO 4 , 50 M okadaic acid, 1% (v/v) Nonidet P-40, 1 mM phenylmethylsulfonyl fluoride, and 0.1 mg/ml aprotinin at pH 7.5). For the examination of transcription factors, cells were harvested in a lysis buffer (60 mM Tris-HCl and 1% SDS at pH 6.8). Lysates were heated at 100°C for 10 min. After centrifugation, supernatants were normalized for protein concentration via the Bradford method and subjected to immunoblotting. For Western blot analysis, proteins were subjected to SDS-PAGE and electroblotted to polyvinylidene fluoride membranes (Polyscreen, NEN Life Science Products, Inc., Boston, MA). Transferred membranes were blocked with Block Ace (Yukijirushi Nyugyo, Sapporo, Japan) and then incubated with the primary antibodies. After washing with PBS, membranes were reacted with secondary antibodies and developed with ECL plus as instructed by the manufacturer. The signal on the blot was detected and quantified with Lumino-Image Analyser LAS-1000 System (Fuji Photo Film Co., Tokyo, Japan).
Statistical Analysis-Data are presented as means Ϯ S.E. Student's t test was used to compare with the control. Differences were accepted as significant at p Ͻ 0.05 level.

ROCK Expression and Actin Cytoskeleton during Adipogenesis and Effect of ROCK Inhibitors on Adipogenesis-
We detected protein expression of ROCK in 3T3-L1 cells by Western blot analysis (Fig. 1A). In 3T3-L1 preadipocytes (day 0) and mature adipocytes (day 8), both ROCK-I and ROCK-II were expressed. The expression levels of two ROCK isoforms did not significantly change during adipogenesis. Then, we observed actin cytoskeleton by phalloidin staining at preadipocytes and mature adipocytes. During adipogenesis, the structure of filamentous actin was converted from long stress fiber (preadipocyte) to cortical actin (mature adipocyte), suggesting that Rho-ROCK pathway is involved in adipogenesis (Fig. 1B). To examine the effects of ROCK inhibitors on adipogenesis, 3T3-L1 cells were exposed to 1 M, 10 M, 30 M Y-27632, 10 M fasudil, 1 M pioglitazone, and 1 g/ml insulin, respectively in the presence of dexamethasone and IBMX. Y-27632, fasudil, or pioglitazone was added to media from day 0 to day 8. Insulin was added to media from the start of differentiation (day 0) to day 4. At day 8, we evaluated the lipid accumulation by Oil Red O staining (Fig. 1, C and D). Y-27632 enhanced lipid accumulation in a dose-dependent manner. Y-27632 or fasudil-treated dishes accumulated lipid droplets markedly compared with vehicle-treated dishes, indicating that ROCK inhibitors enhance adipogenesis. Consistent with previous reports (28 -30), insulin and pioglitazone enhanced lipid accumulation, respectively. Then, to determine when the ROCK inhibitor principally affects adipogenesis, 30 M Y-27632 was added in various periods of 3T3-L1 adipogenesis stimulated by dexamethasone and IBMX (Fig. 2, A and B). The exposure of 30 M Y-27632 during days 0 -4 fully enhanced lipid accumulation, while during days 4 -8, Y-27632 did not enhance it at all. In cases of days 0 -2 and 2-4, Y-27632 partially enhanced lipid accumulation. Treatment with 10 M or 30 M Y-27632 did not significantly alter total cell number during adipogenesis compared with vehicle treatment (Fig. 2C).
Effect of Y-27632 on the Expression of Adipogenic Genes-Because ROCK inhibitors enhanced lipid accumulation, we examined the expression of adipogenic transcription factors C/EBP␤ and C/EBP␦ at day 2, and C/EBP␣ and PPAR␥ at day 4 by Western blot analysis (Fig. 3, A and B). In the presence of dexamethasone and IBMX, Y-27632 with a concentration of 30 M augmented the expression of C/EBP␤ (ϳ1.8-fold increase), C/EBP␣ (ϳ6.5-fold increase) and PPAR␥ (ϳ5.6-fold increase) while it suppressed that of C/EBP␦ (ϳ40% decrease) compared with the non-treated group. 1 g/ml insulin did not affect the expression of C/EBP␤ and C/EBP␦ at day 2 while it enhanced that of C/EBP␣ and PPAR␥ at day 4. 1 M pioglitazone augmented the expression of C/EBP␤ and C/EBP␣ while it suppressed those of C/EBP␦ and PPAR␥. The effects of insulin and pioglitazone on the expression of adipogenic trasnscription factors were compatible with previous reports (29,31). Furthermore, we examined protein expression of adipogenic genes during the course of differentiation; day 0, day 1, day 2, day 4, and day 8 (Fig. 3C). Compared with the vehicle treatment A, equal amounts of proteins harvested at day 4 were subjected to Western blot analysis using antibodies against PPAR␥, C/EBP␣, and ␤-actin. B, equal amounts of proteins harvested at day 2 were subjected to Western blot analysis using antibodies against C/EBP␤, C/EBP␦, and ␤-actin. C, equal amounts of proteins harvested at day 0, 1, 2, 4, and 8 were subjected to Western blot analysis using antibodies against PPAR␥, C/EBP␣, C/EBP␤, C/EBP␦, aP2, adiponectin, and ␤-actin. ␤-Actin proteins were monitored as hallmarks during adipocyte differentiation. Pio, pioglitazone. All data are representatives of at least three independent experiments. OCTOBER 5, 2007 • VOLUME 282 • NUMBER 40 group, 30 M Y-27632 enhanced the expression of adipogenic transcription factors: C/EBP␤ (ϳ1.7-fold increase, day 1), PPAR␥ (ϳ5.2-fold increase, day 4), and C/EBP␣ (ϳ11.7-fold increase, day 4) and adipogenic late markers: aP2 (ϳ8.9-fold increase, day 4) and adiponectin (ϳ8.1-fold increase, day 4). Also in this analysis, Y-27632 suppressed the expression of C/EBP␦ (ϳ31% decrease, day 2).

Role of ROCK-II in Adipogenesis
Effect of LPA and Y-27632 on Actin Cytoskeleton and Adipogenesis-LPA activates small GTPase Rho via specific G protein-coupled receptors: LPA [1][2][3][4] (32). It has been reported that LPA induces the stress fiber formation in Swiss 3T3 fibroblast, which is inhibited by Y-27632 (33,34). Therefore, we examined the regulation of actin cytoskeleton by LPA and Y-27632 in 3T3-L1 preadipocytes (Fig. 4A). We stimulated serum-starved 3T3-L1 preadipocytes with or without 10 M LPA after the pretreatment of 30 M Y-27632 or non-treatment and assessed the effects on actin cytoskeleton by staining cells with phalloidin. LPA addition to serumstarved 3T3-L1 cells spread the cells and induced stress fibers while this change was inhibited by Y-27632. In Y-27632-treated cells without LPA addition, we could not observe stress fibers or thick actin rim in the cell periphery.

Adipogenesis of ROCK-II (Ϫ/Ϫ) and ROCK-I (Ϫ/Ϫ) MEFs-To
Effect of Knockdown of ROCK on Expression of Adipogenic Transcription Factors-To determine whether inhibition of ROCK-II enhances adipogenesis in 3T3-L1 cells, we examined the effect of ROCK-I or ROCK-II knockdown on 3T3-L1 adipogenesis using siRNA technique. We introduced siRNA targeted for ROCK-I or ROCK-II into 3T3-L1 preadipocytes. 3T3-L1 cells were differentiated 12 h after introduction of siRNA. We confirmed the knockdown of ROCK-I or ROCK-II expression by Western blot analysis at day 2. Compared with the negative control group, introduction of ROCK-I and ROCK-II siRNA resulted in ϳ96 and ϳ86% decrease, respectively (Fig. 7, A and B). There was no obvious compensatory increase of the expression of the other ROCK isoform. At day 3, we examined the expression of adipogenic transcription factors by quantitative RT-PCR. ROCK-II siRNA enhanced the expression of PPAR␥ and C/EBP␣ compared with the control (ϳ2.1fold and ϳ1.8-fold increase, respectively, **, p Ͻ 0.01) (Fig. 7C). ROCK-I siRNA did not enhance the expression of PPAR␥ and C/EBP␣ significantly (Fig. 7D).

DISCUSSION
In the present study, we show that inhibition of ROCK enhances adipogenesis. It is impossible to determine the isoform-specific role of ROCK by treatment of ROCK inhibitors. In addition to pharmacological approaches, we examined the adipogenesis of both ROCK-I (Ϫ/Ϫ) MEFs and ROCK-II (Ϫ/Ϫ) MEFs. Furthermore, we examined the effects of knock-down of ROCK-I or ROCK-II on 3T3-L1 adipogenesis by RNA interference. To the best of our knowledge, this is the first study demonstrating that inhibition of ROCK-II enhances adipogenesis and insulin-induced Akt phosphorylation. Our results suggest that ROCK-II plays an important role in adipogenesis and insulin signaling.
We examined the ROCK expression in 3T3-L1 cells by Western blot analysis. Both ROCK isoforms are expressed during the course of adipogenesis. A previous study showed that ROCK-II was expressed in 3T3-L1 mature adipocytes (36) while the expression of ROCK-I has not been reported.
Next, we examined the effects of ROCK inhibitors on lipid accumulation in 3T3-L1 cells. We found that two ROCK inhibitors, Y-27632 and fasudil, enhance lipid accumulation, respectively. These findings suggest that inhibition of ROCK enhances adipogenesis. We also assessed when Y-27632 affects adipogenesis. Supplementation with Y-27632 during days 0 -4 fully enhances adipogenesis. These results let us speculate that inhibition of ROCK enhances the expression of adipogenic transcription factors. As expected, Y-27632 enhances the expression of adipogenic transcription factors, C/EBP␤, C/EBP␣, and PPAR␥ in the presence of dexamethasone and IBMX, which indicates that inhibition of ROCK enhances adipogenesis through the up-regulation of adipogenic transcription factors. Following this, Y-27632 enhances the expression of adipogenic late markers, such as aP2 and adiponectin.
We assessed the effects of LPA and Y-27632 on adipogenesis and actin cytoskeleton. Because the Rho-ROCK pathway regulates actin cytoskeleton, we examined the effects of LPA and Y-27632 on actin cytoskeleton with phalloidin staining. We showed that stress fiber formation is augmented by LPA and that it is inhibited by Y-27632 in 3T3-L1 preadipocytes. These results provide the evidence that LPA activates the Rho-ROCK pathway and Y-27632 inhibits the pathway. Then, we showed that LPA inhibits adipogenesis in 3T3-L1 cells consistent with a previous report in 3T3-F442A cells. Furthermore, we showed that Y-27632 restores its inhibition in a dose-dependent manner. These results indicate that the Rho-ROCK pathway plays an important role in adipogenesis.
To determine the isoform-specific role of ROCK and to entirely exclude the nonspecific effects of ROCK inhibitors, we examined adipogenesis of ROCK-I (Ϫ/Ϫ) and ROCK-II (Ϫ/Ϫ) MEFs. We demonstrated that targeted disruption of ROCK-II enhances adipogenesis and that, in contrast, that of ROCK-I does not enhance adipogenesis. Furthermore, Y-27632 enhances lipid accumulation in ROCK-I (Ϫ/Ϫ) MEFs and does not enhance lipid accumulation in ROCK-II (Ϫ/Ϫ) MEFs. It suggests that inhibition of endogenous ROCK-II enhances adipogenesis. In 3T3-L1 cells, we examined the effect of knockdown of ROCK-I or ROCK-II on adipogenesis using siRNA. We showed that ROCK-II siRNA enhances the expression of adipogenic transcription factors while ROCK-I siRNA does not enhance it. Taken together, our results from the pharmacological and genetic approaches indicate that inhibition of ROCK, especially ROCK-II, enhances adipogenesis.
We investigated the mechanisms of the enhancement of adipogenesis by inhibition of ROCK. We demonstrated that Y-27632 inhibits IRS-1 serine phosphorylation and markedly enhances Akt phosphorylation upon the stimulation of insulin in 3T3-L1 preadipocytes. Previously, it has been reported that insulin/IGF-1 signaling is required for adipogenesis. MEFs derived from IRS-1 (Ϫ/Ϫ) mice do not have enough differentiation properties (27). IRS-1 and IRS-3 double knock-out mice exhibit lipodystrophy (38). Constitutively active Akt causes spontaneous differentiation in 3T3-L1 cells (39). Furthermore, we also demonstrated that an inhibitor of PI3-kinase, wortmannin, blocks the enhancement of adipogenesis induced by Y-27632. Therefore, augmentation of insulin signaling may contribute to the enhancement of preadipocyte differentiation. Concerning MAPK signaling, we showed that Y-27632 slightly enhances p38MAPK phosphorylation and slightly inhibits ERK1/2 phosphorylation in this cell line. Regarding ERK1/2 and p38MAPK, a great deal of evidence about the molecular link between MAPK and adipogenesis has been accumulating (40 -45). However, their roles in adipogenesis are still controversial. In addition to pharmacological approaches, we showed that in ROCK-II (Ϫ/Ϫ) MEFs and even in ROCK-II (ϩ/Ϫ) MEFs, insulin-induced Akt phosphorylation is enhanced compared with that in wild-type MEFs while in ROCK-I MEFs, there is no significant difference in insulin-induced Akt phosphorylation between each genotype. It provides the direct evidence that loss of ROCK-II enhances insulin-induced Akt phosphorylation.
Regarding the ROCK and insulin signaling, we showed that a ROCK inhibitor reduces IRS-1 serine phosphorylation and increases Akt phosphorylation in 3T3-L1 preadipocytes. It is known that ROCK directly associates with IRS-1 and phosphorylates its serine residues (35)(36)(37). In addition, previous studies have reported that IRS-1 serine phosphorylation negatively regulates insulin signaling in vascular smooth muscle cells (37) and NIH3T3 fibroblasts (46). However, recently, it has been reported that ROCK phosphorylates IRS-1 serine residues and positively regulates insulin signaling in 3T3-L1 mature adipocytes (36). The discrepancy between these reports shows that IRS-1 serine phosphorylation regulates insulin signaling differently depending on cell types. Therefore, there could be a difference in the regulation of insulin signaling by IRS-1 serine phosphorylation between preadipocytes and mature adipocytes.
Adipogenic analysis of both ROCK-I (Ϫ/Ϫ) and ROCK-II (Ϫ/Ϫ) MEFs revealed that loss of ROCK-II, not ROCK-I, enhances adipogenesis. The isoform-specific role of ROCK has been still obscure both in vitro and in vivo. As previously reported, loss of ROCK-I results in the eyelid open at birth (EOB) and omphalocele, while loss of ROCK-II results in placental dysfunction, intrauterine growth retardation, and fetal death. However, recently it has been reported that ROCK-II knock-out mice backcrossed in the C57BL6/N background exhibit not only the placental phenotype but also EOB and omphalocele (47). Concerning actin cytoskeletal regulation, it suggests that ROCK-I and ROCK-II cooperatively regulate actin bundle assembly required for eyelid and ventral body wall closure. In this context, our findings that inhibition of ROCK-II, not ROCK-I, enhances adipogenesis is a novel isoform-specific role of ROCK.
In summary, the present study first provides the robust and direct evidence that inhibition of ROCK-II enhances adipogenesis accompanied by the up-regulation of adipogenic transcription factors and that inhibition of ROCK-II enhances insulin-induced Akt phosphorylation. These results indicate that inhibition of ROCK-II enhances adipogenesis, at least in part, via augmentation of insulin signaling. Finally, these findings raise the possibility that ROCK-II is a novel target for the treatment of obesity, insulin resistance, and type 2 diabetes.