Rho GTPase-dependent Signaling Is Required for Macrophage Migration Inhibitory Factor-mediated Expression of Cyclin D1^*

Cell Culture and Drug Treatment

MIF^−/− mice and their wild-type littermates were maintained on a mixed 129Sv × C57Bl/6 background (F3). Mouse embryonic fibroblasts (MEFs) were generated from embryos at day 14.5 and grown in DMEM with 10% FCS and antibiotics (5). NIH 3T3 fibroblasts were maintained in low glucose DMEM with 10% FCS and antibiotics. For recombinant MIF (15) stimulation and inhibitor studies, NIH 3T3 cells were plated at a density of 2 × 10⁵ cells/6-cm plate and allowed to grow for 48–72 h in low glucose DMEM with 10% FCS and antibiotics. Without a medium change, the cells were then treated with either 10 μm ML-7 (Biomol), 10 μm Y27632 (EMD Biosciences), or 50 μm U0126 (Promega). After incubation with the specified inhibitor for 30 min, recombinant MIF (rMIF) was added at 50 ng/ml unless otherwise specified. For transient transfection experiments, NIH 3T3 cells were plated as above, and 24 h later, cells were transfected with the indicated amounts of either pcDNA3.1 GS or pcDNA3.1 GS/MIF using Fugene 6 Transfection Reagent (Roche Applied Science) as described previously (5, 16).

For cell proliferation experiments, 1 × 10⁴ NIH 3T3 fibroblasts were plated in each well of a 96-well plate in 0.5% serum containing low glucose media for 48 h. Without a media change, cells were treated with either 10 μm ML-7, 10 μm Y27632, or 50 μm U0126. After incubation with the specified inhibitor for 30 min, rMIF was added at 50 ng/ml and allowed to sit for 30 h. 0.5 μCi/well [³H]thymidine (PerkinElmer Life Sciences) was added to each well for an additional 4 h, at which time the cells were harvested and the incorporation of [³H]thymidine into DNA was quantified by liquid scintillation counting (Packard Instrument Co.).

For Rho inactivation experiments, pEXV-Rho^N19 or pEXV vector alone were transfected into 1 × 10⁵ NIH 3T3 cells/well of a 6-well plate at 4 μg/well using Fugene 6. rMIF was added to transfected cells 48 h later as indicated.

For adhesion/growth factor induction of Rho activation assays, 2 × 10⁶ MIF^+/+ or MIF^−/− p53^H175 immortalized fibroblasts were serum-starved for 48 h. Cells were lifted and plated onto 10-cm fibronectin-coated dishes (BD Biosciences) in the presence of 10% FCS for the indicated times.

Immunoblotting Studies

Whole cell extracts were prepared from cells after the indicated treatments for the indicated times. Cells first were lysed in ice-cold radioimmune precipitation buffer (containing 1 mm NaVO₄, 2 mM NaF, and a protease inhibitor mixture (Roche Applied Science)). The cells were disrupted by repeated aspiration through a 21-gauge needle. After microcentrifugation at 3000 rpm for 10 min (4 °C), the supernatants were removed, and equal amounts of cellular proteins were fractionated on SDS-polyacrylamide gels (Bio-Rad) and transferred to polyvinylidene difluoride membranes (Millipore Corp.). Immunoblotting was performed with antibodies directed against phospho-ERK MAPK, phospho-MLC (Cell Signaling Technology), cyclin D1, Cdk4, Myc epitope 9E10, HA tag (Santa Cruz Biotechnology, Inc., Santa Cruz, CA), and phospho-FAK (Tyr³⁹⁷) (BIOSOURCE International). Densitometric analysis of Western blots was performed using Scion Image software.

Rho GTPase Activity Assay

The rhotekin binding domain (RBD) and PAK binding domain-GST plasmids were kindly provided by Dr. Keith Burridge (University of North Carolina). The plasmids were transformed into BL-21 Escherichia coli, and GST-PAK binding domain (for Rac1 and Cdc42) and GST-RBD (for Rho) were purified as described (17). For active (GTP-loaded) Rho, Rac, and Cdc42 assays, ~2 × 10⁶ cells/sample were lysed in 10-cm dishes with lysis buffer (50 mm Tris, pH 7.5, 500 mm NaCl, 0.1% SDS, 0.5% deoxycholate, 1% Triton X-100, and 0.5 mm MgCl₂) followed by centrifugation at 14,000 rpm for 10 min at 4 °C. 300–500 μg of lysate/sample was incubated with 60 μl of a 50% slurry of GST-RBD beads or 30 μl of GST-PAK binding domain beads pre-equilibrated with lysis buffer. After rotating for 30 min at 4 °C, the beads were collected by centrifugation, washed once in lysis buffer, and released by boiling in 1× Laemmli sample buffer. Rac1, Cdc42, and RhoA GTP determination and total Rac1, Cdc42, and RhoA from lysates were assessed by Western blot analyses using RhoA, Cdc42 (Santa Cruz Biotechnology), and Rac1 (Upstate Biotechnology, Inc., Lake Placid, NY) antibodies.

Fluorescence Microscopy

5 × 10⁴ cells/ml were plated onto fibronectin-coated (10 μg/ml in phosphate-buffered saline) chamber slides (BD Biosciences) for 2 h, washed, and then fixed with 4% formaldehyde in phosphate-buffered saline. After washing and permeabilization, fluorescein isothiocyanate-labeled phalloidin (Sigma) was added at a final concentration of 1 μg/ml for 1 h. After staining, cells were washed five times with phosphate-buffered saline, fixed with Vectastain (Vector Laboratories), and observed/photographed by fluorescence microscopy at ×20 magnification (Nikon).

Luciferase Promoter Assays

NIH 3T3 or MIF^+/+ and MIF^−/− p53^H175 immortalized cells were plated at 5 × 10⁴ cells/ml onto 24-well dishes. The following day, cells were transiently transfected with 0.4 μg of either serum response factor (SRF)-TA-Luc-, pRb-TA-Luc- (Clontech), or −1745 CD1-Luc- (a generous gift of Dr. Richard Pestell (Georgetown University)) sensitive luciferase promoter constructs using Fugene 6 transfection reagent (Roche Applied Science). 0.1 μg of Renilla pRL-null vector (Promega) was co-transfected with each of the above vectors. In some cases, pK.HA vector alone, pK.HA CPI-17 (18) (generous gifts of Dr. Masumi Eto (University of Virginia)), pEXV alone, or pEXV Rho^V14 (19) was co-transfected along with luciferase plasmids at 1 μg. Luciferase and Renilla luciferase activities were measured by the Dual Luciferase reporter assay system (Promega, Madison, WI) on a TD-20/20 luminometer (Turner Designs). For expression analysis of HA-tagged CPI-17, parallel samples were transfected in duplicate with pK.HA alone or pK.HA CPI-17, and lysates were assessed for HA expression by immunoblotting.

Retroviral Plasmid Constructs and Infections

The p53^H175 (a gift of Dr. Oleksi Petrenko, SUNY Stonybrook, Stonybrook, NY) retroviral expression vector is described elsewhere (20). 1 × 10⁵ fibroblasts were infected with the indicated retroviral supernatants for 24 h, washed, and refed normal media for the indicated times postinfection. The expression of p53^H175 in primary fibroblasts resulted in efficient immortalization of both MIF^+/+ and MIF^−/− cells as evidenced by passaging for greater than 25 times with no evidence of cell senescence.

Exogenously Added Recombinant and Ectopically Expressed MIF Induce Cyclin D1 Expression in Fibroblasts

Our previous studies revealed that both endogenous and exogenously added rMIF induce the sustained activation of ERK MAPK in NIH 3T3 and mouse embryonic fibroblasts (4, 5). We have additionally found that MIF-deficient cells display defective growth factor/adhesion-induced sustained ERK activation, leading to cyclin D1 expression, Rb inactivation, and S phase cell cycle progression (5). Despite these findings, very little is known about how proximal signaling by extracellular MIF promotes the sustained activation of ERK, leading to cyclin D1 expression.

To determine whether exposure of cells to exogenously added recombinant MIF was sufficient to induce cyclin D1 transcription/translation that is reportedly linked to the sustained activation of ERK (7–9), NIH 3T3 cells were treated with rMIF (15) at various concentrations. As shown in Fig. 1A, cyclin D1 protein expression was dose-dependently induced by treatment with rMIF. To verify this finding and to determine whether forced overexpression of MIF was also capable of inducing cyclin D1 expression, MIF was transiently transfected into NIH 3T3 fibroblasts. As shown in Fig. 1B, a transiently transfected MIF mammalian expression plasmid (16) dose-dependently induced the expression of cyclin D1 (top panel) concordant with increased intra- and extracellular expression of MIF (data not shown).

Because the sustained activation of ERK promotes the expression of cyclin D1 through a transcription-dependent process (7, 8), we then determined whether transfection of MIF was capable of driving the expression of a well characterized cyclin D1 promoter-linked luciferase reporter plasmid (21). NIH 3T3 cells co-transfected with an MIF-expressing plasmid and a luciferase plasmid driven by a 1,745-nucleotide region of the human cyclin D1 promoter (21) revealed that cyclin D1 promoter activity is increased almost 4-fold above the vector control at the highest concentration of MIF plasmid used (Fig. 1B). This finding strongly suggests that the increased expression of cyclin D1 observed with both rMIF and ectopically expressed MIF is due to enhanced transcription from the cyclin D1 promoter.

MIF Activates RhoA, Myosin Light Chain Phosphorylation, and ERK MAPK

Recent studies describe a link between Rho GTPase-dependent activation of ERK MAPK, cyclin D1 expression, and G₁/S phase cell cycle progression in fibroblasts (7). Because we hypothesized that MIF-dependent cyclin D1 expression is linked to its ability to stimulate ERK activity, we examined the possibility that MIF may act to promote ERK activation and cyclin D1 expression by stimulating Rho GTPase activity. As shown in Fig. 2, A and B, recombinant MIF induces Rho GTP accumulation in NIH 3T3 fibroblasts concomitant with its activation of ERK MAPK. Notably, we were unable to detect any effect of modulating levels of MIF on the activation states of either Rac1 or Cdc42 GTPases in mesenchymal cells (data not shown).

Roovers and Assoian (13) report that Rho promotes stress fiber formation and subsequent cyclin D1 expression via a Rho kinase-dependent mechanism. Rho kinase has previously been linked to enhanced myosin light chain phosphorylation and subsequent stress fiber formation by two distinct mechanisms. The first involves a direct Rho kinase-mediated phosphorylation of MLC (22), and the second requires a Rho kinase-dependent inactivating phosphorylation of myosin light chain phosphatase (MLCP) (23). Regardless of the mechanism, enhanced phosphorylation of MLC is thought to lead to stress fiber formation, enhanced integrin clustering, and the subsequent focal adhesion kinase (FAK)-dependent sustained activation of ERK MAPK (13). To determine whether MIF-induced Rho-GTPase activation results in enhanced myosin light chain phosphorylation, immunoblot analysis of MLC using phosphorylation-specific antibodies revealed that MLC phosphorylation increases in direct proportion to the MIF-induced stimulation of GTP-bound Rho (Fig. 2A). These data demonstrate that MIF induces ERK activation concomitant with Rho GTPase activation and enhanced myosin light chain phosphorylation.

Inhibition of Rho, Rho Kinase, Myosin Light Chain Kinase, and MEK1 Prevents the Induction of MAPK, Cyclin D1, and DNA Synthesis by MIF in NIH 3T3 Fibroblasts

To investigate in more detail the precise requirements for this hypothesized pathway of cyclin D1 expression induced by MIF, we next sought to inhibit the key enzymes thought to be involved in this pathway. Because we hypothesize that MIF-dependent induction of cyclin D1 expression in fibroblasts is reliant upon sustained ERK MAPK activation, a well characterized inhibitor of the upstream activator of ERK1/2 was used. As shown in Fig. 3A, pretreatment of NIH 3T3 fibroblasts with the MEK1 inhibitor U0126, prevented the induction of cyclin D1 expression by MIF.

As mentioned above, one pathway by which Rho could promote signaling to ERK, leading to cyclin D1 expression, is through the inactivation of MLCP by Rho kinase (13, 24). The net result of MLCP inactivation by Rho kinase would be an enhancement of steady state MLC phosphorylation that is dependent upon myosin light chain kinase activity (13). We therefore tested whether inhibitors of myosin light chain kinase could prevent MIF-mediated FAK and ERK activation, leading to cyclin D1 expression. As depicted in Fig. 3B, pretreatment of cells with the myosin light chain kinase inhibitor ML-7, blocked the induction/activation of these proposed MIF effectors, suggesting a strict requirement for MLC-dependent stress fiber formation and subsequent FAK activation in carrying out MIF signaling to cyclin D1 expression.

To verify the importance of Rho kinase as a mediator of MIF-stimulated signaling to focal adhesion formation, sustained MAPK activation, and cyclin D1 expression, the Rho kinase inhibitor, Y27632, was tested for its ability to block MIF signaling events. As shown in Fig. 3C, inhibition of Rho kinase blocks MIF-induced MLC phosphorylation, FAK phosphorylation, and sustained ERK activation and, to a slightly lesser extent, inhibits the relative -fold induction of cyclin D1 induced by MIF. Finally, to verify that RhoA itself was required for MIF-mediated signaling to cyclin D1 expression, forced overexpression of Rho^N19 was found to very effectively block the ability of exogenously added rMIF to induce MLC, FAK, and ERK MAPK phosphorylation and cyclin D1 expression. Together, these findings indicate that Rho, Rho kinase, MLC kinase, and ERK MAPK are all part of an MIF-induced signaling program that serves to enhance cyclin D1 expression in mesenchymal cell types.

Because we hypothesized that this Rho-dependent signaling pathway leading to cyclin D1 expression is responsible for the previously observed induction of mesenchymal cell proliferation initiated by MIF, we next sought to determine whether blockade of the individual components of this signaling cascade could prevent MIF-induced S phase cell cycle progression. As shown in Fig. 3E, inhibition of Rho kinase, myosin light chain kinase, or MEK1 all prevent G₁-S phase progression induced by rMIF.

MIF-deficient Mouse Embryonic Fibroblasts Are Defective in Rho GTPase Activation

Given the fact that exogenously added rMIF activates Rho, leading to the sustained activation of ERK and cyclin D1 expression (Fig. 3), we would predict that cells deficient in MIF would possess less active Rho. To test this, quiescent p53^H175 immortalized MIF^+/+ and MIF^−/− MEFs (5) were plated onto fibronectin-coated dishes in the presence of growth factors for varying times. Welsh et al. (7) previously reported that under these conditions, Rho is responsible for the sustained activation of ERK that had previously been shown to promote mid-G₁ phase cyclin D1 expression (6, 13). As shown in Fig. 4, A and B, immortalized MIF^−/− fibroblasts are strongly impaired in adhesion/growth factor-induced Rho activation when compared with the levels of Rho activity in MIF^+/+ cells. Similar to our earlier findings with ERK (5), the requirement for MIF in modulating adhesion/growth factor-dependent Rho activation appears to be highest during the sustained phase (i.e. 1–4.5 h) of cell activation as opposed to the acute phase (i.e. 0.2 h). This finding is also consistent with our prior observation that MIF is required for maximal cyclin D1 expression in response to adhesion/growth factor stimulation (5).

Active RhoA GTPase, Stress Fiber Formation, Phosphorylated ERK, and Cyclin D1 Are Defective in MIF-deficient Fibroblasts

To investigate in greater detail the relative requirements for endogenous MIF in maintaining Rho activity and downstream effector activation in murine embryonic fibroblasts, we next determined the extent of stress fiber formation in the absence and presence of MIF. Normally cycling, asynchronous MIF^+/+ and MIF^−/− p53^H175-immortalized MEFs were plated onto fibronectin and assessed for their relative abilities to form stress fibers. As expected, MIF-deficient cells display significantly fewer and less organized stress fiber formation than their wild type counterparts (Fig. 5A). Importantly, and consistent with this result, we also find that Rho activity is substantially reduced in normally cycling, asynchronous MIF^−/− cells, and this decreased Rho activity correlates closely with decreased cyclin D1 expression and sustained ERK activity in these cells (Fig. 5B). Together, these findings suggest that both induced (Fig. 4) and steady state (Fig. 5) Rho activity and downstream signaling to cyclin D1 are, at least in part, dependent upon functional MIF.

Rescue of Defective Cyclin D1 Transcription in MIF^−/− Cells by Rho GTPase and MLC Phosphatase Inhibition

In order to establish a functional link between decreased Rho activity and defective cyclin D1 expression observed in MIF^−/− cells, we next sought to determine whether restoration of Rho-dependent signal transduction could rescue MIF-stimulated cyclin D1 transcription and Rb inactivation (5). Rho^V14, a constitutively active mutant of Rho, was co-transfected into p53^H175 immortalized MIF^−/− cells along with a SRF, cyclin D1, or Rb-sensitive luciferase reporter plasmid. SRF is a well characterized downstream transcriptional effector of Rho activity (25, 26), and, in line with the decrease in active Rho observed in MIF^−/− fibroblasts (Figs. 4 and 5), these cells contain nearly 75% less active SRF than MIF^+/+ cells (Fig. 6A). Importantly, reconstitution of constitutively active Rho restores defective SRF activity in MIF^−/− cells well above and beyond that of the MIF^+/+ cells.

Consistent with our hypothesis that MIF-mediated Rho activity promotes steady state cyclin D1 expression and Rb inactivation in rodent fibroblasts, transcription from both the cyclin D1 promoter and associated Rb inactivation are both significantly reduced in MIF-deficient cells versus MIF-containing cells (Fig. 6A). Importantly, both of these defects are overcome by the reintroduction of constitutively active Rho into MIF^−/− cells (Fig. 6A). Combined, these data suggest that the observed decrease in Rho activity associated with the loss of MIF is at least partially responsible for the defective cyclin D1 transcription and Rb inactivation also found in these cells (Fig. 5) (5).

To further investigate whether Rho-dependent enhancement of myosin light chain phosphorylation leading to cyclin D1 expression is part of an effector pathway utilized by MIF to promote cyclin D1 transcription, we attempted to recapitulate MLCP inhibition by CPI-17. CPI-17 is an endogenous inhibitor of MLCP and is also reported to be a substrate of Rho kinase (27–29). When phosphorylated at Thr³⁸ by Rho kinase or a variety of other kinases, the inhibitory activity of CPI-17 toward MLCP is increased nearly 1000-fold (28).

To rescue the defective Rho-dependent MLCP inhibition in immortalized MIF^−/− fibroblasts, we overexpressed CPI-17 or a vector control along with a cyclin D1 promoter-driven luciferase construct. CPI-17 overexpression in MIF^+/+ p53^H175-expressing cells had only a very modest effect on transcription from the cyclin D1 promoter, suggesting that the requirement for MLCP inhibition on cyclin D1 promoter activity in MIF-containing cells was nearly maximal at the time of analysis (Fig. 6B). In contrast, however, transcription from the cyclin D1 promoter in MIF^−/− cells was only half that of MIF^+/+ cells (Fig. 6, A and B), and restoration of MLC phosphorylation (not shown) through MLCP inhibition by CPI-17 fully restored cyclin D1 transcription from these cells (Fig. 6B). It is important to note that several other enzymes, in addition to Rho kinase, are capable of stimulating the activity of CPI-17 (30–32). Because we anticipate that Rho kinase activity is reduced in MIF-deficient cells, the activation of CPI-17 is probably induced by one of these other enzymes in an MIF-deficient setting. Together, these data imply that both Rho activity and myosin light chain phosphorylation are important effectors of MIF-dependent cyclin D1 expression in fibroblasts.

In summary, these results provide a mechanistic link between MIF and cyclin D1 expression in rodent fibroblasts. This pathway includes the MIF-dependent activation of Rho, Rho kinase, MLC phosphorylation, FAK, and ERK MAPK (Fig. 7). Because Rho activation impinges upon several parallel effector pathways (14), it is likely that MIF-dependent modulation of steady state Rho activity leads to phenotypic changes other than those examined in this study. Further investigation is warranted to evaluate the relative role of MIF in the modulation of these and other signaling intermediates.