Endothelial Cells Inhibit NO Generation by Vascular Smooth Muscle Cells
Role of Transforming Growth Factor-β
Endothelial cell (EC)–released agents are active regulators of vascular smooth muscle cell (VSMC) functions. The first aim of the present work was to analyze the effect of ECs on interleukin-1β (IL-1β)–induced NO production by SMCs. Bovine aortic ECs (BAECs) and BVSMCs in culture were used for the study. IL-1β (0.03 U/L) stimulated nitrite production by BVSMCs. This increase was smaller in the presence of BAECs. This effect was accompanied by reduced expression of inducible NO synthase (iNOS) in BVSMCs coincubated with BAECs, as analyzed by Western blot analysis. The reduction in iNOS protein expression was partially reversed by a polyclonal antibody against transforming growth factor-β (TGF-β). Furthermore, we examined the cytotoxic effect of the NO released from BVSMCs on both BAECs and the BVSMCs themselves. Incubation of BAECs with IL-1β–prestimulated BVSMCs induced EC toxicity, which was partially inhibited by an inhibitor of NO synthesis, Nωnitro-l-arginine methyl ester, or an inhibitor of iNOS expression, dexamethasone. No cytotoxic effect of IL-1β on BVSMCs themselves was detected. ECs modulate iNOS expression in SMCs by mechanisms that include a TGF-β–dependent pathway. The NO released from SMCs exerts cytotoxic effects on the adjacent endothelium without altering the viability of the SMCs.
- Received November 27, 1995.
- Revision received February 26, 1996.
NO is a multifunctional molecule with an important role in the relationships among cells that make up the microvascular environment. NO has been implicated in the regulation of different functions, including blood flow, blood fluidity, SMC relaxation and proliferation (for reviews, see References 1 through 3), platelet and leukocyte aggregation, adhesion to the endothelium,4 5 6 7 and more recently, matrix protein synthesis.8
NO is generated by the metabolic conversion of l-arginine to l-citrulline by the activity of NO-synthesizing enzymes (ie, NOSs). Two major classes of NOS activity have been identified in blood vessels: a constitutive isoform, which is localized in the endothelium under basal conditions, and an inducible isoenzyme, which is mostly absent under “normal” conditions and requires cytokine or endotoxin activation for its expression.9
It has been shown that VSMCs not only respond to NO but also are capable of producing it. Stimulation of VSMCs with specific inflammatory mediators, including IL-1β, induces NO release from these cells by stimulating iNOS expression.10 Several factors activate or inhibit iNOS expression in VSMCs, including platelet-derived growth factor, fibroblast growth factor, insulin-like growth factor I, TGF-β, angiotensin II, and plasmin.11 12 13 14 15 For most of these substances the endothelium is a well-recognized source. Nevertheless, at present there are no reports about the actual role of the endothelium in iNOS expression in VSMCs.
In the past few years several groups have shown that after endothelial denudation in vivo, VSMCs generate NO, which inhibits the contractile response to different agonists.16 Therefore, the first aim of the present study was to investigate the role of normal ECs in regulating the induction of IL-1β–stimulated iNOS activity in VSMCs.
cNOS generates small amounts of NO for short periods of time. In contrast iNOS generates large amounts of NO over several hours.1 3 However, the pathophysiological implications of large amounts of NO released from VSMCs have not been adequately defined. Because large amounts of NO are known to exert toxic effects on many cell types (for reviews see References 1 and 17), in a second set of experiments we examined the possibility that the NO released after BVSMC iNOS stimulation might induce cytotoxic effects in the BVSMCs themselves or adjacent ECs.
IL-1β, TGF-β, L-NAME, dexamethasone, and SNP were purchased from Sigma Chemical Co. SIN-1 was obtained from Alexis Corp, and Na251CrO4 (as a source of 51Cr) was purchased from Amersham. All other chemicals were of the highest commercially available quality from Sigma. The monoclonal antibody against iNOS was purchased from Transduction Laboratories, and the polyclonal antibody against TGF-β was a gift from P. Esbrit (Fundación Jiménez Díaz).
BAECs were obtained and cultured as previously described.18 In brief bovine aortic lumen was filled with 500 mg/L type II collagenase (Sigma) and incubated at 37°C for 20 minutes. BAECs were harvested in RPMI 1640 medium supplemented with 10% fetal calf serum, 5 mmol/L Gln, 2×10−5 U/L penicillin, and 2×10−5 μg/L streptomycin. Cells were seeded into six- or 24-well plates and used after one or three passages. BAECs exhibited the typical “cobblestone” appearance and “stained” positive for von Willebrand factor immunofluorescence.
After collagenase digestion to isolate the endothelium, the aortic lumen was scraped to remove contaminating endothelial material. Bovine aortic media tissue was then removed in small strips and transferred to tissue culture flasks containing the RPMI 1640 medium described above. Once the BVSMCs had migrated and grown from the explants, cells were transferred to six- or 24-Transwell inserts (a microporous membrane with a 0.4-μm pore size and coated with collagen type I; Millipore). BVSMCs were used between passages one and three. As described,19 these cells exhibited the typical “hill-and-valley” growth morphology and reacted with anti–α-actin monoclonal antibody (Boehringer Mannheim).
BAEC/BVSMC Coincubation Experiments
A coculture system was prepared by placing the Transwell inserts containing the BVSMCs into the wells of the culture plates containing the BAECs. When required by the experimental protocol, IL-1β (0.03 U/L), a polyclonal antibody against TGF-β, or a nonspecific IgG was added to the BAEC/BVSMC coculture. In this coculture system the medium was shared by both cell types. Therefore, humoral exchange was allowed between them without direct cell-to-cell contact. This coculture technique also permitted the processing of BVSMCs alone. Cytotoxicity assays were performed in 24-well plates, whereas the experiments for nitrite and iNOS protein detection were done in six-well plates.
Measurement of NO Production
NO production by BVSMCs or BAEC/BVSMC cocultures was assessed as nitrite generation. Nitrite contents were measured by using the Griess reagent/reaction as described.20 In brief BVSMCs or BAEC/BVSMC cocultures were incubated with or without IL-1β (0.03 U/L) for 18 hours. When required by the experimental protocol, a rabbit polyclonal antibody against TGF-β (6.7×10−8 mol/L) or a rabbit nonspecific IgG at the same concentration was added to the incubation medium. The supernatants were recovered; after centrifugation (2500 rpm, 10 minutes) nitrite accumulation was measured by mixing equal volumes of supernatant and Griess reagent [1% sulfanilamide and 0.1% N-(1-naphthyl)ethylenediamine dihydrochloride in 2% H3PO4]. Nitrite concentrations were determined at an optical density of 554 nm by comparison with standard solutions of sodium nitrite prepared in the same culture medium.
Determination of iNOS Protein Expression
iNOS protein expression was analyzed in BVSMCs incubated with or without BAECs. iNOS expression was induced by incubation with IL-1β (0.03 U/L) for 18 hours. After incubation BVSMCs were collected and boiled in Laemmli buffer21 containing 2-mercaptoethanol. Proteins obtained from BVSMCs were separated on denaturing SDS/10% polyacrylamide gels (15 μg per lane) and blotted onto nitrocellulose (Immobilon-P, Millipore). The blots were blocked overnight at 4°C with 5% nonfat dry milk in 20 mmol/L Tris HCl, 137 mmol/L NaCl, and 0.1% Tween 20. Western blot analysis was performed with a monoclonal antibody against iNOS. The blots were incubated with the first antibody (1:500) for 1 hour at room temperature, extensively washed, and incubated with the second antibody (horseradish peroxidase–conjugated anti-mouse immunoglobulin antibody) at a dilution of 1:1500 for 1 more hour. Specific iNOS protein was detected by enhanced chemiluminescence (ECL, Amersham) and evaluated by densitometry (Molecular Dynamics). Prestained protein markers (Sigma) were used for molecular mass determinations. The monoclonal antibody specifically recognizes the iNOS isoform (130 kDa) and does not cross-react with endothelial cNOS, since the band of iNOS protein was almost undetectable in a homogenate of cultured BECs but reacted with positive controls obtained from homogenates of lipopolysaccharide-treated mouse macrophages (data not shown).
51Cr Release From BAECs and BVSMCs
We examined whether the NO produced by cytokine-stimulated BVSMCs could either act in a paracrine fashion to damage adjacent BAECs or be autotoxic for BVSMCs. To test the first possibility, BAECs were preincubated for 18 hours in RPMI 1640 containing 10% fetal calf serum and 2×103 μCi/L Na251CrO4. After being washed twice with fresh RPMI 1640, 51Cr-labeled BAECs were coincubated with BVSMCs, which had been previously stimulated with IL-1β (0.03 U/L) for 8 hours. BVSMCs were rinsed twice and cocultured with 51Cr-labeled BAECs for another 48 hours. Therefore, only the BVSMCs were directly stimulated with IL-1β.
The second possibility was studied by preincubating BVSMCs in RPMI 1640 containing 10% fetal calf serum and 2×103 μCi/L Na251CrO4 for 18 hours. The BVSMCs were rinsed twice with fresh medium and incubated for an additional 48 hours in RPMI 1640 with or without IL-1β. To exclude possible interference from an NO-independent cytotoxicity induced by IL-1β, similar experiments were performed with dexamethasone (10−6 mol/L), an inhibitor of iNOS expression, or the l-arginine competitive analogue L-NAME (10−4 mol/L). Dexamethasone was added and washed concurrently with IL-1β. L-NAME was added during BAEC/BVSMC coincubation. In other experiments, we examined whether a polyclonal anti–TGF-β antibody modified the EC toxicity of the IL-1β–stimulated BVSMCs. Supernatant fractions were counted for radioactivity at different times after centrifugation (1800 rpm for 10 minutes). 51Cr release at each time was calculated with respect to the total 51Cr content.
Results are expressed as mean±SEM. Unless otherwise stated, each value corresponds to a minimum of six experiments done in triplicate. Comparisons were done by either ANOVA or paired or unpaired Student's t test when appropriate. Bonferroni's correction for multiple comparisons was used to determine the level of significance of the probability value.
IL-1β–Stimulated Nitrite Production by BVSMCs With or Without BAECs
Incubation of confluent BVSMC cultures with IL-1β induced the accumulation of nitrite in the culture medium (Fig 1⇓). To determine the specificity of the nitrite determination as a measure of NO production by BVSMCs, L-NAME (10−4 mol/L) or dexamethasone (10−6 mol/L) was added to the incubation medium. Both L-NAME and dexamethasone significantly reduced nitrite production by BVSMCs (86±2% and 63±4%, respectively; n=4, P<.05). IL-1β–stimulated nitrite release was significantly greater in BVSMCs alone than in BAEC/BVSMC cocultures (Fig 1⇓).
To elucidate the mechanisms by which BAECs reduced the efficacy of IL-1β to stimulate nitrite release from BVSMCs, a polyclonal antibody against TGF-β was used. The hypothesis was based on recent data that vascular SMCs and ECs cooperate in producing the biologically active form of TGF-β,22 a potent blocker of IL-1β–dependent iNOS induction in SMCs.13 Addition of the anti–TGF-β antibody to BAEC/BVSMC cocultures partially reversed the inhibition by BAECs of IL-1β–induced nitrite release from BVSMCs (Fig 1⇑). This effect was not observed when a nonspecific IgG was used. Addition of the TGF-β antibody alone to BVSMCs had no effect on IL-1–stimulated nitrite production (data not shown). The efficacy of the polyclonal antibody to inhibit TGF-β activity was confirmed by its ability to prevent the inhibition of nitrite production evoked by addition of exogenous TGF-β (4×10−10 mol/L) to the IL-1β–activated BVSMCs (Table⇓).
iNOS Protein Expression
Modifications of the levels of iNOS protein expression were detected by Western blotting. Under baseline conditions BVSMCs expressed minimal amounts of iNOS protein, which increased markedly in the presence of IL-1β (Fig 2⇓). Coincubation of BVSMC monolayers with BAECs decreased IL-1β–induced iNOS expression (Fig 2⇓). Addition of a polyclonal antibody against TGF-β partially reversed the inhibition by the endothelium (Fig 2⇓). This effect was not observed by incubation with a nonspecific IgG (data not shown).
Measurement of BVSMC and BAEC Cytotoxicity
Treatment of 51Cr-labeled BVSMCs with IL-1β (0.03 U/L) for 48 hours did not affect 51Cr release from BVSMCs with respect to untreated BVSMCs (Fig 3a⇓). This observation indicates that levels of IL-1β that induced iNOS expression and activity in BVSMCs did not induce cytotoxicity. iNOS activity has been reported to be toxic to ECs in an autocrine manner.23 To test whether the NO produced by stimulated BVSMCs was toxic to adjacent BAECs, both IL-1β–stimulated and unstimulated BVSMCs plated on 24-Transwell inserts were placed over a monolayer of 51Cr-labeled BAECs for 48 hours. In these assays BVSMCs only were activated by the cytokine during 8 hours (see “Methods”). Baseline 51Cr release from BAECs was not significantly changed by the presence or absence of unstimulated BVSMCs (Fig 3b⇓). However, when BAECs were exposed to IL-1β–treated BVSMCs, 51Cr release from the BAECs was significantly enhanced, indicating that IL-1β–stimulated BVSMCs caused a time-dependent cytotoxicity (Fig 3b⇓).
Since TGF-β was involved in the BAEC inhibition of iNOS expression in BVSMCs, we determined the effect of a polyclonal anti–TGF-β antibody on EC cytotoxicity of IL-1β–treated BVSMCs. The anti–TGF-β antibody was added during coincubation. When BAECs were coincubated with IL-1β–treated BVSMCs with the anti–TGF-β antibody, 51Cr release from the BAECs increased slightly but significantly (Fig 3b⇑). This significant increase was observed for as long as 12 hours after coincubation. This effect was not observed when a nonspecific IgG was used (data not shown).
The question was then raised whether an 8-hour incubation with IL-1β was sufficient to induce iNOS expression and to produce large amounts of NO by BVSMCs. Stimulation of BVSMCs with IL-1β for 8 hours was sufficient to increase nitrite production (Fig 4a⇓) and to induce iNOS expression (Fig 4b⇓). In addition, we tested nitrite accumulation after a 24-hour coculture by following the same protocol described previously for the cytotoxicity experiments. Despite the fact that IL-1β was removed from BVSMCs, these cells released high amounts of NO (Fig 4c⇓). When BAECs were coincubated with IL-1β–treated BVSMCs, we observed a slight and significant decrease in nitrite accumulation (Fig 4c⇓). This inhibition of nitrite accumulation was less than that observed when IL-1β was added simultaneously with BAECs (Figs 1 and 4c⇑⇓). Therefore, although the presence of endothelium decreased nitrite production by IL-1β–treated BVSMCs, this decrease was not sufficient to avoid the endothelial cytotoxicity of NO.
Once we observed an increase in NO production after 8 hours of IL-1β stimulation of BVSMCs, the next set of experiments was performed to demonstrate that NO was the diffusible mediator produced by IL-1β–stimulated BVSMCs, which subsequently induced cytotoxicity in adjacent BAECs. For this purpose iNOS expression and activity were inhibited by dexamethasone and L-NAME, respectively. Dexamethasone (10−6 mol/L) was added to the BVSMCs simultaneously with IL-1β, whereas L-NAME (10−4 mol/L) was added during coincubation. BAEC cytotoxicity induced by IL-1β–stimulated BVSMCs was partially inhibited by dexamethasone (10−6 mol/L) and L-NAME (10−4 mol/L) (by 56±5% and 68±4%, respectively; n=4, P<.05). In this regard we also observed that the increase in nitrite production in cocultures of BAECs and IL-1β–stimulated BVSMCs was blocked by L-NAME (Fig 4c⇑).
Finally, we tested the cytotoxic capacity of exogenous NO on BAECs by incubating these cells with the NO-generating compounds SIN-1 and SNP. Both SIN-1 (10−3 mol/L) and SNP (10−3 mol/L) markedly induced 51Cr release from BAECs in a time-dependent manner (Fig 5⇓). SIN-1 induced a higher 51Cr release from BAECs, which was correlated with a higher nitrite accumulation than that observed with SNP (Fig 5⇓).
The present study provides new evidence about the way that the two major constituents of the vascular wall, ie, ECs and SMCs, interact. The main findings of our investigation demonstrate that BAECs produce an inhibitory effect on both IL-1β–stimulated NO production and iNOS expression in adjacent BVSMCs. Furthermore, the capability of BVSMCs to injure adjacent BAECs through the production of NO as a cytotoxic agent has also been demonstrated.
To clarify the mechanisms of the endothelium-dependent inhibition of iNOS induction, we examined the role of TGF-β as a putative candidate for mediating this effect. The rationale for this approach was based on two types of evidence: first, the cooperative role of endothelial and vascular SMCs in producing active TGF-β22 and second, the fact that addition of exogenous, active TGF-β is capable of inhibiting iNOS expression in both macrophages and SMCs.13 24 Our results support the concept that the endothelium regulates iNOS expression in SMCs by a TGF-β–dependent mechanism, because addition of a polyclonal antibody against TGF-β enhanced both nitrite production and iNOS protein expression induced by IL-1β in BAEC/BVSMC cocultures.
Although the concentration of TGF-β antibody was enough to reverse the inhibitory effect of large amounts of exogenous TGF-β (4×10−10 mol/L) on IL-1β–stimulated nitrite production by SMCs, it only partially prevented the endothelium-dependent inhibition on IL-1β–stimulated NO release and iNOS expression by BVSMCs. This finding suggests that TGF-β is not the only agent by which the endothelium modulates NO release from SMCs. In this regard different authors have reported that angiotensin II, platelet-derived growth factor, and insulin-like growth factor I inhibit IL-1β–activated iNOS induction in vascular SMCs.11 12 14 Further studies are needed to clarify whether other inhibitors released from the endothelium might be implicated in the above-described effects.
In contrast to our results, Durante et al25 described that TGF-β released from activated platelets did not modify NO production by IL-1β–treated VSMCs. The differences between our experimental conditions and those of Durante et al are, however, significant. Platelets and other cells release TGF-β in a latent, biologically inactive form. TGF-β released from ECs and SMCs, although also secreted in a latent form, is transformed to the active form during coculture of these two cell types, under conditions similar to those in the present experiments.22
In the second part of this study we asked whether the NO released from BVSMCs either had an autotoxic role on vascular SMCs or could act in a paracrine fashion to damage adjacent ECs. The capacity for NO generation by SMCs could be important in different pathophysiological situations. After local endothelial injury, circulating blood cells are recruited to the site of injury, where they are activated and then release inflammatory mediators, including IL-1β. These inflammatory mediators can further stimulate iNOS expression and NO release from vascular SMCs.
Our results demonstrate that IL-1β–induced NO production by BVSMCs does not affect SMC viability. On the other hand, cytotoxicity experiments on BAECs showed that induction of iNOS protein in BVSMCs stimulated 51Cr release from BAECs. Interestingly, BAECs strongly inhibited NO production by BVSMCs when the coculture was performed at the same time as the addition of IL-1β. Eight hours after IL-1β stimulation of BVSMCs, BAECs still prevented NO generation, although this inhibition was insufficient to block its cytotoxic effect. In this regard we observed that the polyclonal antibody against TGF-β only slightly enhanced 51Cr release from BAECs for as long as 12 hours after cell coincubation. Perrella et al13 demonstrated that TGF-β may inhibit IL-1β–stimulated iNOS transcription in SMCs when administered either before or no longer than 6 hours after IL-1β is added. In our experiments when IL-1β–prestimulated BVSMCs were cocultured with BAECs, BVSMCs had already expressed iNOS protein, which could explain the modest inhibition of nitrite production by BAECs under these conditions.
The finding that both nitrite production by BVSMCs and BAEC death were reduced by incubation of BVSMCs with either dexamethasone, an inhibitor of iNOS expression, or L-NAME, an antagonist of the l-arginine/NO pathway, strongly suggests that the mediator by which IL-1β–activated BVSMCs damaged BAECs was indeed NO. In this regard inhibition of nitrite production by L-NAME was correlated with the inhibition of 51Cr release. To further support this suggestion, the NO donor compounds SIN-1 and SNP stimulated 51Cr release from ECs.
Palmer et al23 showed that ECs are damaged by the NO synthesized by iNOS expression within ECs. Our present findings provide further insight into the results of Palmer et al and demonstrate that the damaging effect of NO may be the by-product of a previously undescribed interaction between the two main cell types of the vessel wall. Therefore, under certain conditions, vascular smooth muscle can be the source of severe injury to the endothelium.
At sites of endothelial injury, eg, in the case of balloon angioplasty or atherosclerotic plaque erosion, the effect of locally released cytokines would not be counterbalanced by the presence of ECs. Therefore, local release of large amounts of NO might occur, which could potentially damage the surrounding ECs or interfere with the mobility of cells participating in reendothelialization.
In conclusion, the presence of endothelium negatively modulates iNOS expression in SMCs, at least in part through a TGF-β–dependent mechanism. The NO released from these SMCs exerts cytotoxic effects on the adjacent endothelium without negatively affecting the viability of the SMCs themselves. Since NO is also likely to limit adhesion and activation of blood cells at sites of endothelial injury (and thus the formation of the thrombus), further studies are needed to determine whether large amounts of NO generated by SMCs in the absence of endothelium have beneficial or detrimental effects in situations of intimal vascular lesions.
Selected Abbreviations and Acronyms
|L-NAME||=||Nωnitro-l-arginine methyl ester|
|SMC(s)||=||smooth muscle cell(s)|
|TGF-β||=||transforming growth factor-β|
This work was supported by grants from Fundacíon Ramon Areces, Fundacíon de Estudios Cardiovasculares, Laboratorios Uriach SA, and Lacer SA. Dr J.R. Mosquera is a fellow of Fundacíon Renal. L. Sánchez de Miguel, T. de Frutos, and M. Montón are fellows of Fundacíon Conchita Rabago. The authors thank Concepcíon San Martín and Liselotte Gulliksen for editorial assistance.
López-Farré A, Caramelo C, Esteban A, Alberola ML, Millás I, Montón M, Casado S. Effects of aspirin on platelet-neutrophil: role of nitric oxide and endothelin-1. Circulation.. 1995;91:2080-2088.
Kubes P, Suzuki M, Granger DN. Nitric oxide: an endogenous modulator of leukocyte adhesion. Proc Natl Acad Sci U S A.. 1991;88:587-590.
Scott-Burden T, Schini VB, Elizondo E, Junquero DC, Vanhoutte PM. Platelet-derived growth factor suppresses and fibroblast growth factor enhances cytokine-induced production of nitric oxide by cultured smooth muscle cells. Circ Res.. 1992;71:1088-1100.
Schini VB, Schray-Utz B, Busse R, Vanhoutte PM. Insulinlike growth factor I inhibits induction of nitric oxide synthase in vascular smooth muscle cells. Circ Res.. 1994;74:24-32.
Perrella MA, Yoshigumi M, Fen A, Tsai JC, Hiesh CM, Kourembanas S, Lee ME. Transforming growth factor-β1 but not dexamethasone, down-regulates nitric oxide synthase mRNA after its induction by interleukin-1β in rat smooth muscle cells. J Biol Chem.. 1994;269:14595-14600.
Nakayama I, Kawahara Y, Tsuda T, Okuda M, Yokoyama M. Angiotensin II inhibits cytokine-stimulated inducible nitric oxide synthase expression in vascular smooth muscle cells. J Biol Chem.. 1994;269:11628-11633.
Durante W, Schini VB, Catovsky S, Kroll MN, Vanhoutte PM, Schafer AI. Plasmin potentiates induction of nitric oxide synthesis by interleukin-1β in vascular smooth muscle cells. Am J Physiol.. 1993;264:H617-H624.
Joly GA, Schini VB, Vanhoutte PM. Balloon injury and interleukin-1β induce nitric oxide synthase activity in rat carotid arteries. Circ Res.. 1992;71:331-338.
Wright CE, Rees DD, Moncada S. Protective and pathological roles of nitric oxide in endotoxic shock. Cardiovasc Res.. 1992;26:48-57.
López-Farré A, Riesco A, Espinosa G, Digiuni E, Cernadas MR, Alvarez V, Montón M, Rivas F, Gallego MJ, Egido J, Casado S, Caramelo C. Effect of endothelin-1 on neutrophil adhesion to endothelial cells and perfused heart. Circulation.. 1993;88:1166-1171.
Caramelo C, Riesco A, Outeiriño J, Millas I, Blum G, Monzú B, Casado V, Sanchez L, Mosquera JR, Casado S, López-Farré A. Effects of nitric oxide on red blood cells: changes in erythrocyte resistance to hypotonic hemolysis and potassium efflux by experimental maneuvers that decrease nitric oxide. Biochem Biophys Res Commun.. 1994;199:447-454.
Sato Y, Okada F, Abe M, Seguchi T, Kuwano M, Sato S, Furuya A, Hanai N, Tamaoki T. The mechanism for the activation of latent TGF-β during co-culture of endothelial cells and smooth muscle cells: cell-type specific targeting of latent TGF-β to smooth muscle cells. J Cell Biol.. 1993;123:1249-1254.
Ding A, Nathan CF, Graycar J, Denynck R, Stuehr DJ, Srimal S. Macrophage deactivating factor and transforming growth factors β1, -2 and -3 inhibit induction of macrophage nitrogen oxide synthesis by interferon-gamma. J Immunol.. 1990;145:940-944.
Durante W, Schini VB, Kroll MH, Catovsky S, Scott-Burden T, White JG, Vanhoutte PM, Schafer AI. Platelets inhibit the induction of nitric oxide synthesis by interleukin-1β in vascular smooth muscle cells. Blood.. 1994;83:1831-1838.