Inhibition of Protein Geranylgeranylation Causes a Superinduction of Nitric-oxide Synthase-2 by Interleukin-1 (cid:98) in Vascular Smooth Muscle Cells*

Recently, we have designed farnesyltransferase and geranylgeranyltransferase I inhibitors (FTI-277 and GGTI-298) that selectively block protein farnesylation and geranylgeranylation, respectively. In this study, we describe the opposing effects of these inhibitors on in- terleukin-1 (cid:98) (IL-1 (cid:98) )-stimulated induction of nitric-oxide synthase-2 (NOS-2) in rat pulmonary artery smooth mus- cle cells (RPASMC) and rat hepatocytes. Pretreatment of cells with GGTI-298 caused a superinduction of NOS-2 by IL-1 (cid:98) . RPASMC treated with GGTI-298 (10 (cid:109) M ) prior to IL-1 (cid:98) (10 ng/ml) expressed levels of NOS-2 protein five times higher than those exposed to IL-1 (cid:98) alone. This superinduction of NOS-2 protein by pretreatment with GGTI-298 resulted in nitrite concentrations in the medium that were 5-fold higher at 10 ng/ml IL-1 (cid:98) and 10-fold higher at 1 ng/ml IL-1 (cid:98) . Furthermore, NOS-2 mRNA levels in RPASMC were also increased 6- and 14-fold (at 10 and 1 ng/ml IL-1 (cid:98) , respectively) when the cells were pretreated with GGTI-298.

Nitric oxide (NO) is a radical with important physiologic effects that include modulation of vasoreactivity, prevention of thrombosis, and neurotransmission (1). NO is the product of an enzymatic reaction catalyzed by nitric-oxide synthase (NOS). 1 Mammalian cells express three isozymes of NOS, neuronal (NOS-1), endothelial (NOS-3), and inducible (NOS-2) (reviewed in Ref. 2). NOS-1 and NOS-3 are constitutively active, whereas NOS-2 is not present unless the cell is triggered to produce it. The induction of NOS-2 in cells exposed to cytokines and lipopolysaccharides is implicated in the pathophysiology of shock (3).
IL-1␤ is an important cytokine that is released early in inflammation and shock (4). As a single agent, IL-1␤ is capable of inducing NOS-2 at levels comparable to those induced by a mixture of cytokines and lipopolysaccharides (5). IL-1␤ achieves its effects through signaling pathways that have not been fully defined (6), but that appear to include the mitogenactivated protein (MAP) kinase pathway since IL-1␤ activates MAP kinase itself (7)(8)(9)(10), at least in some systems, via the kinase (MAP kinase kinase) immediately preceding it (10 -12). This effect on MAP kinase kinase implicates the small GTPbinding protein Ras since this triggers activation of the pathway by interacting with Raf, a serine/threonine kinase lying immediately upstream of MAP kinase kinase (13). IL-1␤ is also capable of activating stress-activated protein kinase (14,15) and p38 kinase (16), two kinases that occupy positions in pathways distinct from MAP kinase, but that are also potentially under the control of small G proteins such as Ras, Rho, and Rac (17).
Modulating the function of these small G proteins has recently become possible with inhibitors of prenyltransferases, the enzymes responsible for attaching prenyl lipids to the proteins as the first and most critical step in post-translational processing (18). The enzymes are farnesyltransferase (FTase) (19,20), which attaches the 15-carbon farnesyl, and geranylgeranyltransferase I (GGTase I) (21,22), which attaches the 20-carbon geranylgeranyl. Among proteins that are farnesylated are Ha-Ras and nuclear lamins; among those that are geranylgeranylated are Rap1A, RhoA, and Rac1. Some proteins that can be both farnesylated and geranylgeranylated include RhoB and Ki-Ras 4B (23). Because farnesylation and/or geranylgeranylation of small G proteins is critical for their function, we (24 -28) and others (29 -34) have designed potent and highly selective inhibitors of FTase. More recently, we have designed GGTase I inhibitors (35)(36)(37). FTase inhibitors have been shown to block the processing of oncogenic Ras, to reverse transformation, to induce the accumulation in the cytoplasm of inactive Ras-Raf complexes, and to inhibit murine and human tumor growth in vivo (24 -34). Furthermore, FTase inhibitors are not toxic in vivo (25,30), are less effective at inhibiting normal cell growth (29,31), and do not inhibit growth factor (epidermal growth factor or platelet-derived growth factor) activation of MAP kinase (37)(38)(39). Much less is known about GGTase I inhibitors, but recently, we have shown that these agents antagonize the signaling of oncogenic Ki-Ras 4B (35), decrease levels of tyrosine phosphorylation of the platelet-derived growth factor receptor (37), and block fibroblasts in the G 1 phase of the cell cycle (36). Very little is known about the effects of FTase and GGTase I inhibitors in cells other than fibroblasts. Recently, Singh et al. (40) showed that an inhibitor of FTase blocked IL-1␤-stimulated induction of NOS-2 and activation of MAP kinase in cardiac myocytes. Here we demonstrate that an inhibitor of GGTase I causes superinduction of IL-1␤-stimulated NOS-2 in smooth muscle cells and hepatocytes.

EXPERIMENTAL PROCEDURES
Cell Culture-The primary culture of rat pulmonary artery smooth muscle cells (RPASMC) has been described in detail previously (5). Briefly, explants of pulmonary artery were obtained by dissection following pentobarbital euthanasia of male Harlan Sprague Dawley rats and placed intimal surface down in plastic flasks containing growth medium (1:1 Dulbecco's modified Eagle's medium/Ham's F-12 medium, 10% fetal bovine serum, and penicillin/streptomycin). Cells were allowed to migrate out of the explants for 5-7 days before removal of the explants and subsequent subpassaging by trypsinization. For the studies reported here, cells were trypsinized, replated at ϳ10 6 cells/75-cm 2 flask, and grown under standard conditions of 37°C, 5% CO 2 , and 100% humidity. Cells were used through the fifth passage.
Hepatocytes were isolated from male Harlan Sprague Dawley rats using the in situ collagenase perfusion technique of Geller et al. (41). After isolation, hepatocytes were plated onto gelatin-coated Petri dishes (Corning Inc.) and grown in Williams' medium E (Life Technologies, Inc.) with the addition of L-arginine (0.50 mM), insulin (1 M), HEPES (15 mM), L-glutamine, penicillin/streptomycin, and 10% low endotoxin fetal calf serum. Cells were incubated in growth medium for 24 h, and then serum-free Williams' medium E with the above additives was used for all experimental conditions.
Induction of NOS-2-This was achieved by treating cells with recombinant human IL-1␤ (R&D Systems, Minneapolis, MN) at the concentrations indicated. After 24 h, the medium was harvested for determination of nitrite concentration, and the cells for assessment of NOS-2 RNA and protein levels.
Treatment of Cells with GGTI-298 and FTI-277-RPASMC were grown to 50% confluence (generally, 24 -48 h after plating). Fresh medium was added containing either FTI-277 or GGTI-298 at the concentrations indicated. The cells were allowed to grow with inhibitors for 24 h. Stimulation of cells with IL-1␤ took place in basal medium (Dulbecco's modified Eagle's medium and 0.1% bovine serum albumin) with the addition of FTI-277 or GGTI-298 where indicated.
Measurement of Nitrite-To quantify the amount of NO released from cells, nitrite, its stable product in aqueous solution, was measured by the Griess reaction. Nitrite concentration was measured by mixing an aliquot of cell supernatant with an equal volume of Griess reagent (1 part 0.1% naphthylethylenediamine dihydrochloride to 1 part 1% sulfanilamide in 5% phosphoric acid) and incubating at room temperature for 10 min. The absorbance at 550 nm was measured, and nitrite concentration was determined using sodium nitrite as a standard.
Western Blotting for NOS-2-Cell monolayers were lysed in ice-cold buffer containing 50 mM Tris (pH 8.0), 110 mM NaCl, 5 mM EDTA, 1% Triton X-100, and the protease inhibitors antipain, pepstatin, leupeptin, chymostatin, and phenylmethylsulfonyl fluoride, and the lysate was transferred to a conical tube. Protein concentration was determined using the Bradford assay (Bio-Rad). Whole cell lysate was boiled in Laemmli buffer, and 100 g of protein/lane was loaded on a 7.5% SDS-polyacrylamide gel and separated electrophoretically. Proteins were transferred to a nitrocellulose membrane overnight at 90 mA in a Bio-Rad Trans-Blot cell. For immunoblotting, the membrane was blocked with 10% nonfat dry milk in PBS for 1 h. The primary antibody was a polyclonal rabbit anti-murine macrophage NOS-2 (Transduction Laboratories, Lexington, KY) diluted 1:1000 in 1% nonfat dry milk and applied for 2 h. After washing three times in PBS containing 0.1% Tween 20 and 1% nonfat dry milk, the secondary antibody (peroxidaseconjugated goat anti-rabbit IgG, Sigma) was applied at 1:5000 dilution. The blot was washed in PBS containing 0.3% Tween 20 and 1% nonfat dry milk three times over 30 min, and positive immunoreactivity was visualized using enhanced chemiluminescence reagents (ECL, Amersham International, Buckinghamshire, United Kingdom) and by exposure to photographic film.
Determination of Ras/Rap1A Processing-Cell lysates (50 g/lane) were electrophoresed on a 12.5% SDS-polyacrylamide gel and then transferred to nitrocellulose as described above. The membrane was blocked in PBS-T (PBS with 0.1% Tween 20) containing 5% nonfat dry milk. After washing three times in PBS-T, the membrane was probed with Y13-238, an anti-Ras antibody, at 50 g/ml in PBS-T containing 3% nonfat dry milk for 1 h at room temperature. The blot was washed as described above and then incubated with the secondary antibody (peroxidase-conjugated goat anti-rat IgG) at a dilution of 1:1000. Positive immunoreactivity was visualized using ECL. Processing of Rap1A was assessed in the same manner described for Ras except that the membrane was probed with anti-Rap1A antibody (1:1000; Santa Cruz Biotechnology, Santa Cruz, CA), and the positive antibody reactions were visualized using peroxidase-conjugated goat anti-rabbit IgG and ECL.
Isolation of RNA and Northern Analysis-Total RNA was extracted using the method described by Chomczynski and Sacchi (42). Aliquots containing 20 g of total RNA underwent electrophoresis on 1% agarose gel containing 3% formaldehyde. RNA was transferred to a GeneScreen membrane (DuPont NEN) and UV-cross-linked. Membranes were hybridized overnight at 42°C with a cDNA probe to murine macrophage NOS-2 (1-2 ϫ 10 6 cpm/ml) labeled with [ 32 P]dCTP (specific activity of 3000 Ci/mM; DuPont NEN) by random priming (Boehringer Mannheim). The hybridized filters were washed at 53°C and analyzed on a PhosphorImager (Molecular Dynamics, Inc., Sunnyvale, CA).  (36,37). In this study, we treated RPASMC with the methyl ester prodrugs of FTI-276 and GGTI-297, FTI-277 and GGTI-298, which enter the cell more easily (24,36). Fig. 1A shows that treatment of RPASMC with FTI-277 (0 -10 M), as described under "Experimental Procedures," resulted in a dose-dependent inhibition of Ras processing, with no effect on the processing of the geranylgeranylated protein Rap1A. In contrast, treatment of RPASMC with GGTI-298 (0 -10 M) resulted in inhibition of Rap1A processing, with no effect on Ras processing. These results demonstrate that FTI-277 and GGTI-298 are potent inhibitors of FTase and GGTase I in smooth muscle cells and are therefore efficacious in inhibiting cellular processing of farnesylated and geranylgeranylated proteins, respectively.

Inhibition of Protein Geranylgeranylation Causes Superinduction of NOS-2 Protein and Nitrite Levels by IL-1␤ in RPASMC and Rat
Hepatocytes-To determine the role of geranylgeranylated proteins in the signal transduction pathways that mediate IL-1␤ induction of NOS-2, we treated RPASMC and rat hepatocytes with the GGTase I inhibitor GGTI-298 prior to treatment with IL-1␤ as described under "Experimental Procedures." Fig. 1B shows that in the absence of inhibitors, IL-1␤ (0 -10 ng/ml) induced a concentration-dependent, but modest, increase in the medium levels of nitrite from 1 M basal levels to 12 M after treatment with 10 ng/ml IL-1␤. Pretreatment with GGTI-298 (10 M) caused a dramatic increase in IL-1␤-induced nitrite formation (from 4 to 52 M). The medium of cells treated with 1 ng/ml IL-1␤ alone accumulated nitrite levels of 4 M, whereas the medium of cells treated with GGTI-298 prior to IL-1␤ accumulated levels of nitrites that were more than 10-fold higher (42 M) (Fig. 1B). In the absence of IL-1␤, GGTI-298 increased basal levels of NOS-2 to levels comparable to those obtained by stimulation of cells with 1 ng/ml IL-1␤ alone. In contrast to the effects of GGTI-298, treatment of cells with FTI-277 inhibited IL-1␤-stimulated nitrite production (Fig. 1B). RPASMC were treated with and without GGTI-298 or FTI-277 and stimulated with IL-1␤. Total RNA was isolated, and the levels of NOS-2 and cyclophilin mRNAs were determined by Northern blot analysis as described under "Experimental Procedures." Lane 1, vehicle; lanes 2 and 3, IL-1␤ (1 and 10 ng/ml, respectively); lanes 4 -6, FTI-277 (10 M) and IL-1␤ (0, 1, and 10 ng/ml, respectively); lanes 7-9, GGTI-298 (10 M) and IL-1␤ (0, 1, and 10 ng/ml, respectively). Data are representation of three independent experiments. The dramatic increase in the level of nitrite brought about by GGTI-298 could be due to direct activation of NOS-2 enzymatic activity or superinduction of NOS-2 protein . To determine the  effects of GGTI-298 on NOS-2 protein levels, RPASMC were  treated with GGTI-298 for 24 h and then stimulated with IL-1␤ for a further 24 h as described under "Experimental Procedures." The cells were harvested, and the lysates were analyzed by Western blotting using a specific anti-NOS-2 antibody. Fig.  2A shows that in the absence of GGTI-298, significant induction of NOS-2 protein levels in RPASMC occurred only at 10 ng/ml IL-1␤, whereas in the presence of inhibitor, NOS-2 was induced at concentrations as low as 1 ng/ml. In RPASMC treated with 10 ng/ml IL-1␤, GGTI-298 enhanced the ability of IL-1␤ to induce NOS-2 protein by 5-fold. A similar effect was demonstrated in hepatocytes (Fig. 2B).
We next used an alternative method to demonstrate the involvement of geranylgeranylated proteins in the regulation of NOS-2 induction by IL-1␤. Recently, we have shown that treatment of NIH-3T3 cells with lovastatin inhibited the processing of both farnesylated and geranylgeranylated proteins, with a much more pronounced effect on geranylgeranylated proteins, but that treatment of these cells with a combination of lovastatin and either geranylgeraniol or farnesol restored protein geranylgeranylation or farnesylation, respectively (43). Applying this strategy, we observed that RPASMC treated with lovastatin alone were much more responsive to IL-1␤ in inducing NOS-2 (Fig. 3). This lovastatin-mediated superinduction of IL-1␤-stimulated NOS-2 was reversed by geranylgeraniol, but not by farnesol. Lovastatin inhibition of the processing of Ras and of the geranylgeranylated protein Rap1A in RPASMC was reversed with farnesol and geranylgeraniol, respectively (Fig.  3). The above results confirmed that inhibition of protein geranylgeranylation is responsible for the hypersensitivity of RPASMC to IL-1␤ induction of NOS-2. The results also suggest that inhibition of protein farnesylation is not involved in the superinduction of NOS-2 by IL-1␤.
FTI-277 Blocks IL-1␤ Induction of NOS-2 in RPASMC and Rat Hepatocytes-The above results demonstrated that inhibition of protein geranylgeranylation causes a superinduction of NOS-2 by IL-1␤. We next determined the consequences of inhibiting protein farnesylation. RPASMC were pretreated for 24 h with FTI-277 prior to a 24-h treatment with various concentrations of IL-1␤ (0 -10 ng/ml). The cells were then harvested, and the levels of NOS-2 protein as well as the cumulative levels of nitrites in the cell-conditioned medium were determined as described under "Experimental Procedures." Fig.  4A shows that pretreatment of RPASMC with FTI-277 (10 M) blocked IL-1␤ induction of NOS-2 protein. Hepatocytes were treated similarly, but with slightly higher IL-1␤ levels (0 -50 ng/ml). Fig. 4B shows a similar effect in the hepatocytes, although the effect was not as complete.
The data presented in this report clearly demonstrate that IL-1␤ induction of NOS-2 in RPASMC and rat hepatocytes is under the tight control of farnesylated and geranylgeranylated proteins. Whereas inhibition of protein farnesylation blocked NOS-2 expression, inhibition of protein geranylgeranylation had the opposite effect of dramatically enhancing NOS-2 expression. The results suggest that a farnesylated protein mediates IL-1␤ induction of NOS-2, whereas a geranylgeranylated protein represses this induction. A potential candidate for mediating IL-1␤ induction of NOS-2 is the Ras protein. This suggestion is consistent with the recent report by Singh et al. (40) that showed that IL-1␤ activates the MAP kinase pathway and induces NOS-2 in cardiac myocytes and that this can be blocked by the FTase inhibitor BZA-5B. Therefore, Ras proteins that are farnesylated and are known to activate the MAP kinase pathway are good candidates for mediating IL-1␤ induction of NOS-2. Among geranylgeranylated proteins that can serve to antagonize IL-1␤ induction of NOS-2 are the small G proteins Rho and Rac. Although the involvement of these proteins in growth factor signaling is now well documented, their role in cytokine signaling is less clear. It is plausible that cytokines such as IL-1␤ could activate both the Ras/MAP kinase (7-10) and the Rho/Rac/stress-activated protein kinase (14,15) pathways. Integration of the MAP kinase and stressactivated protein kinase pathways in smooth muscle cells and hepatocytes would result in the net outcome of induction of NOS-2 expression. Inhibition of the stress-activated protein kinase pathway that depends on the geranylgeranylated RhoA and Rac proteins would tip the equilibrium in the direction of releasing the repressing pathways and superinducing NOS-2. On the other hand, inhibition of the MAP kinase pathways that depend on farnesylated proteins such as Ras would block NOS-2 induction. Regardless of the mechanism by which inhibition of protein geranylgeranylation causes superinduction of NOS-2, the consequences of this novel finding are of great therapeutic potential. We are presently evaluating the potential of GGTI-298 to reverse in animal models intimal hyperplasia associated with restenosis and atherosclerosis. Our ultimate goal is to prevent local hyperplasia that compromises the success of angioplasty and surgical bypass for obstructive vascular lesions.