Abstract Cytokines and endotoxin stimulate inducible NO synthase (iNOS) in different types of cells; however, little is known about regulatory mechanisms. Using the Griess reagent for nitrite levels, Western blots for iNOS protein, Northern blots for iNOS mRNA, and transient transfection studies to monitor transcription, we determined potential mechanisms involved in interleukin-1β stimulation of iNOS in cultured neonatal ventricular myocytes. When myocytes were treated with interleukin-1β (5 ng/mL), nitrite levels increased, and this effect was inhibited 80% by the specific iNOS inhibitor aminoguanidine. Neither interferon gamma nor tumor necrosis factor–α alone stimulated nitrite production. Bacterial endotoxin alone stimulated nitrites and potentiated the effect of interleukin. To determine whether a tyrosine kinase–mediated signaling pathway was involved in interleukin action, we used the inhibitor genistein, which blocked interleukin-stimulated nitrites, iNOS protein, and iNOS mRNA. To determine the effect of activation of protein kinase C, we treated cells with the phorbol ester phorbol 12-myristate 13-acetate (PMA). PMA decreased both interleukin-stimulated nitrites and iNOS protein by 40%. To determine the involvement of cyclic nucleotides, cells were treated with either dibutyryl cAMP or cGMP. cAMP (1 mmol/L) stimulated iNOS mRNA, protein, and nitrite production, whereas cGMP had no effect. To test for a direct effect of interleukin on transcription of the iNOS gene, we transfected the full-length mouse iNOS 5′ regulatory sequences (−1592 to +160) coupled to a luciferase reporter gene (−1592iNOSLuc). Interleukin stimulated luciferase activity 1.8±0.2-fold. To determine whether interleukin also affects iNOS mRNA stability, interleukin-stimulated iNOS mRNA was allowed to decay in the presence of the transcription inhibitor actinomycin D. iNOS mRNA t1/2 (≈1 hour) was not affected by interleukin. Thus, our data suggest that (1) interleukin-1β is the primary cytokine in myocyte iNOS regulation and acts predominantly at the transcriptional level; (2) interleukin stimulation of iNOS mRNA and protein is coupled to a tyrosine kinase–mediated signaling pathway; and (3) protein kinase C and cAMP can modify interleukin signaling by decreasing and increasing iNOS, respectively.
The iNOS was originally characterized in macrophages and shown to be maximally induced by endotoxin (bacterial LPS) and IFN. Since then, iNOS has been characterized in a number of different cell types in the cardiovascular system, including endothelial cells, vascular smooth muscle cells, fibroblasts, and cardiac myocytes.1 2 3 Regulation of iNOS seems to be cell-type specific, with different types of cells responding to different cytokines and using a number of signaling mechanisms.1 PKC has been implicated in the stimulation of iNOS in a murine macrophage cell line4 and rat hepatocytes.5 In contrast, PKC inhibits the induction of iNOS in mesangial cells,6 whereas it is not involved in iNOS regulation in vascular smooth muscle cells.7 Signaling pathways that activate tyrosine kinase4 and protein kinase A8 also modulate the induction of iNOS. At the molecular level, the transcription factors nuclear factor–κB (NFκB) 9 and interferon regulatory factor 110 are involved in the induction of NOS in macrophages.
IL is a major regulator of iNOS in neonatal11 and adult12 cardiac myocytes. Depending on the cell type, IL signaling may involve the sphingomyelin-ceramide pathway, serine-threonine kinases, tyrosine kinases, and the transcription factors NFκB and activator protein 1 (AP1).13 14 15 Because the iNOS isoform in cardiac myocytes may participate in a number of pathophysiological conditions, including myocarditis, contractile dysfunction, allograft rejection, and ischemic injury,1 2 3 it is important to understand the mechanisms of its induction. In this report we further characterize the major cytokine stimulators of iNOS in myocytes, IL signaling, and transcriptional effects.
Preparation of Ventricular Myocytes
This protocol was approved by the Henry Ford Hospital Committee for Care and Use of Experimental Animals. Hearts from 1- to 2-day-old neonatal Sprague-Dawley rats (Charles River, Kalamazoo, Mich) were minced and digested with trypsin and DNAse to prepare primary cultures of ventricular myocytes as described previously.16 Ventricular myocytes were plated at a density of 1×105 cells/cm2 in Dulbecco’s modified Eagle’s medium (Gibco) plus 10% fetal bovine serum (HyClone) and cultured as described previously.16 Medium lacking serum but supplemented with insulin, transferrin, and selenium was added 24 hours before cells were treated with test compounds. Unless otherwise specified, treatments were for 24 hours.
Nitrite production, an index of NO production, was measured in media samples by the Griess reaction.17 Nanomoles of nitrite were determined by comparison with a standard curve of NaNO2 using an enzyme-linked immunosorbent assay plate reader at 505 nm. NOx from triplicate wells were averaged for each experiment. Controls (untreated cells) were assigned a value of 1, and the values for all treatments were normalized to 1 (x-fold increase versus control). Data were expressed as mean±SE. Differences in mean values among treatment groups were analyzed by a Student’s t test or one-way ANOVA, with pairwise multiple comparisons made by the Student-Newman-Keuls method. A value of P<.05 was considered significant.
Western Blot Analysis
Approximately 3×106 myocytes were lysed in buffer containing 10 mmol/L Tris (pH 7.5), 10 mmol/L EDTA, 0.4% deoxycholate, 1% NP-40, 0.1% sodium dodecyl sulfate (SDS), 1 mmol/L phenylmethylsulfonyl fluoride, and 1× protease inhibitor cocktail (1000X=5 mg/mL each pepstatin A, leupeptin, antipain, and chymostatin). Cellular debris was removed by centrifugation in a microfuge. The protein concentration of the supernatant was determined using Coomassie protein assay reagent (Pierce) with bovine serum albumin as the standard. Fifty micrograms of cytosolic protein was separated out by electrophoresis on an 8% SDS-polyacrylamide gel and then transferred to an Immobilon-P PVDF membrane (Millipore). The membrane was incubated in blocking buffer (Tris-buffered saline [TBS] with 5% nonfat dried milk and 0.2% NP-40) and then with 0.4 mg/mL rabbit polyclonal antibody raised against mouse macrophage iNOS (Santa Cruz Biotechnology) overnight at 4°C. The blot was washed in TBS plus 0.25% sodium deoxycholate, 0.2% NP-40, and 0.1% SDS and incubated with a 1:2000 dilution of the secondary antibody (goat anti-rabbit IgG alkaline phosphatase conjugate; BioRad) in blocking buffer. Colorimetric detection of the reaction product was accomplished with substrate and developing reagents from BioRad according to the manufacturer’s protocol. Laser densitometry was used to quantitate the protein bands.
Northern Blot Analysis
Total RNA was extracted from ventricular myocytes and analyzed for iNOS mRNA as described previously.11 GAPDH mRNA was used to normalize for variations in loading of samples on gels. mRNA levels were quantified by laser densitometry.11
Transient Transfection and Luciferase Assay
Freshly isolated ventricular myocytes were transiently transfected by electroporation as described previously.11 Aliquots of 25 to 50 μg of −1592iNOSLuc18 (containing the 5′ flanking sequences of the mouse iNOS gene extending from −1592 to +160) and 1 μg of RSV-β-galactosidase were transfected per 12×106 cells. Cells (1×106) were plated onto each well of a six-well plate. Luciferase (Luciferase Assay System, Promega) and galactosidase (Galacto-Light chemiluminescent assay; Tropix) activities were measured in an OptoComp 1 luminometer (MGM Instruments) using the manufacturers’ protocols. Luciferase activity was normalized to β-galactosidase or protein and reported as x-fold increase versus untreated controls (controls=1). The increase in response to IL was identical when luciferase activity was normalized either to β-galactosidase or protein. Our previous studies have shown no effect of 24-hour IL treatment on total protein in ventricular myocytes.11 Also, we did not detect any effect of IL on RSV-β-galactosidase activity in these experiments. For each experiment, duplicate aliquots of cell lysates from triplicate wells were assayed. Data were expressed as mean±SE and analyzed for statistical significance as described above.
The cytokines IL, TNF, and IFN were obtained from Bachem. Endotoxin (bacterial LPS), aminoguanidine, cAMP, cGMP, and the phorbol ester (PMA) were obtained from Sigma Chemical Co. Genistein was purchased from Calbiochem. C2 and C16 ceramide were obtained from Fluka and Biomol, respectively, and sphingomyelinase from Sigma. Routine laboratory supplies and chemicals were obtained from Fisher Scientific.
Effects of Cytokines and LPS on NO Production
We have previously shown that IL stimulates nitrite production in neonatal ventricular myocytes.11 We extended these studies to include other cytokines and LPS and combinations thereof to determine the major inducers of iNOS in myocytes. Fig 1⇓, top, shows that IL and LPS (both at 5 ng/mL) induced nitrite production 12- and 5.8-fold, respectively, and that the effect of IL was potentiated by LPS (15.8±1.1 versus 12.2±1.5; P<.01 IL plus LPS versus IL alone). These results were confirmed at the protein level by Western blot analysis (Fig 1⇓, bottom). Neither TNF (25 ng/mL) nor IFN (10 ng/mL) stimulated nitrite production (M.C.L., J.R.S., unpublished observations, 1995). Fig 1⇓, top, also shows that the stimulatory effect of IL was decreased 80% by the specific iNOS inhibitor aminoguanidine (100 μmol/L).19 Western blot analysis of iNOS protein indicated that aminoguanidine had no effect on protein levels (Fig 1⇓, bottom).
Possible Signal Transduction Mechanisms and Cross Talk
Since induction of iNOS by LPS in macrophages involves a tyrosine kinase–linked signaling cascade,4 we next studied whether the tyrosine kinase inhibitor genistein (50 and 100 μmol/L) could inhibit the effects of IL in myocytes. Fig 2⇓ shows that genistein inhibited IL-stimulated nitrite production (Fig 2A⇓), iNOS protein (Fig 2B⇓), and iNOS mRNA (Fig 2C⇓), suggesting that the tyrosine kinase–mediated pathway is directly linked to regulation of iNOS gene expression.
We have previously shown that the PKC inhibitor staurosporine has no effect on IL stimulation of nitrites,11 suggesting that PKC is not part of the signaling pathway involved in IL regulation of iNOS. We tested whether direct activation of PKC could modulate IL stimulation of iNOS. Stimulation of PKC with the phorbol ester PMA decreased IL-stimulated nitrite production (Fig 3A⇓) and iNOS protein (Fig 3B⇓) by 40%, but had no effect on iNOS mRNA (Fig 3C⇓). Thus, our data suggest that PKC may affect translation or degradation of iNOS protein.
In vascular smooth muscle cells, cAMP alone increases nitrites and stimulates iNOS mRNA.7 cGMP, an intracellular mediator of NO actions, has also been implicated in potentiating IL stimulation of nitrites.20 Thus, we tested whether cAMP and cGMP mediate IL action or are able to modulate IL stimulation of nitrites. Fig 4⇓ shows that 1 mmol/L cAMP stimulated nitrites 3.4-fold (Fig 4A⇓) and iNOS protein and mRNA (Fig 4B⇓). Treatment with cAMP and IL synergistically stimulated nitrite production (21-fold versus 16-fold stimulation; P<.01). In contrast, cGMP (1 mmol/L) had no effect on nitrite production and iNOS protein levels (M.C.L., J.R.S., unpublished observations, 1995). Thus, cAMP and the pathway it stimulates synergize with the IL signaling pathway.
Both TNF and IL have been shown to activate the sphingomyelinase-ceramide phospholipid signaling pathway.14 To test its involvement in the regulation of iNOS, cells were treated with 10−6 and 10−5 mol/L of C2 and C16 ceramide or 10−5 to 10−1 U/mL sphingomyelinase, concentrations used in studies of EL4 thymoma cells.21 None of these treatments stimulated nitrites (M.C.L., J.R.S., unpublished observations, 1995).
IL Stimulates Transcription of iNOS
To test for a direct effect of IL on transcription of the iNOS gene, we transfected ventricular myocytes with a plasmid containing the full-length mouse iNOS regulatory sequences upstream from a luciferase reporter gene (−1592iNOSLuc). Fig 5⇓ shows that IL stimulated the iNOS promoter 1.8±0.2-fold (n=9; P<.001). A promoterless luciferase vector did not respond to IL (M.C.L., J.R.S., unpublished observations, 1995).
iNOS mRNA t1/2
Since the magnitude of IL activation of the iNOS promoter was not as great as the increases in nitrites and mRNA, we questioned whether IL was stabilizing iNOS mRNA. Because our previous studies showed maximal induction of iNOS mRNA 3 to 24 hours after IL treatment,11 we treated myocytes with IL for 24 hours and then with the transcription inhibitor ActD (5 μg/mL) in the presence or absence of IL. At 0, 1, 3, and 6 hours after addition of ActD, total RNA was extracted, followed by Northern blot analysis. Fig 6A⇓ shows that iNOS mRNA decayed with a t1/2 of approximately 1 hour after the addition of ActD and that this was not affected by the addition of IL. Fig 6B⇓ shows iNOS mRNA levels after the aforementioned treatments. Thus, IL seems to act primarily at the transcriptional level and not through stabilization of iNOS mRNA.
Two important findings of our study are that the tyrosine kinase inhibitor genistein prevents IL induction of iNOS mRNA and protein and nitrite production and that the specific iNOS inhibitor aminoguanidine inhibits nitrite production in ventricular myocytes. Moreover, the effect of IL occurs primarily at the transcriptional level.
Of the cytokines tested in our study, only IL and LPS stimulated iNOS synthesis and nitrite production. LPS potentiated the effects of IL. Whether LPS is acting through a signaling mechanism distinct from IL or inducing production of cytokines, as occurs in macrophages,22 was not investigated in our study. Our data are consistent with a recently published study on iNOS expression in adult cardiac myocytes: Either IL or LPS stimulated iNOS mRNA, whereas combined treatment with IL, TNF, and IFN gave maximal stimulation.12
Because IL appeared to be the primary cytokine involved in upregulation of iNOS, we studied its signaling mechanisms, possible second messengers, and cross talk with other intracellular signals. Our data indicate that a tyrosine kinase–mediated signal is required for IL action in neonatal ventricular myocytes. Such a mechanism has recently been reported for cardiac myocytes23 and pancreatic β-cells.24 Since the IL receptor has neither intrinsic tyrosine kinase activity nor has it been shown to couple to a nonreceptor tyrosine kinase,13 25 it is unclear how its effects are transduced and how they result in activation of transcription factors, such as NFκB13 26 and AP1.15 Studies suggest that the effect of IL is transmitted through the sphingomyelinase-ceramide–activated kinase cascade.14 This pathway activates stress-activated protein kinases (SAPKs), also known as jun kinases (JNKs), through tyrosine phosphorylation. Two targets for SAPKs are the transcription factors JUN and ATF-2, which can individually activate target genes or form heterodimers for transactivation of genes at AP1/ATF sites in responsive genes.27 28 Although we were unable to demonstrate the involvement of either sphingomyelinase or ceramide in the induction of iNOS, we have preliminary evidence from Western blot analysis for the presence of JNKs in cardiac myocytes and are exploring whether they are tyrosine-phosphorylated in response to IL (M.C.L., J.R.S., unpublished observations, 1995).
Our data also suggest that other intracellular signals, including those from PKC and cAMP, can modulate IL stimulation of iNOS. Previously we showed that the kinase inhibitor staurosporine did not affect IL stimulation of nitrites, a result confirmed by other investigators.23 29 Thus, PKC is not a component of the IL signaling pathway. On the other hand, our studies suggest that PKC can modulate IL-stimulated NO production. Although we did not determine the mechanism involved in this effect, our Western and Northern blot data suggest that PKC may modulate IL regulation of iNOS at the translational or posttranslational level. Such mechanisms are involved in transforming growth factor–β suppression of NO production in macrophages.30 Two other possible roles of PKC can be excluded. It is unlikely that PKC is directly inactivating iNOS enzyme activity, because the iNOS protein does not have a consensus PKC phosphorylation site. Also, it is unlikely that PKC is downregulating IL receptors, because we have previously shown that phenylephrine, which is known to activate PKC, potentiates IL-stimulated nitrite production by myocytes.11 This potentiation is most likely unrelated to PKC and thus is not contradictory to our present results; instead this potentiation may be related to the high dose of phenylephrine used (50 μmol/L), subsequent activation of β-adrenergic receptors, and increases in cAMP.
Unlike the effects of the phorbol ester PMA, cAMP by itself was able to stimulate iNOS mRNA, protein, and nitrite production. Moreover, it synergized with IL. Such effects have previously been reported in cultured vascular smooth muscle8 and mesangial31 cells. In the latter case, cAMP increased the t1/2 of IL-stimulated iNOS mRNA. The mechanism by which cAMP stimulates nitrite production was not addressed in our study. Since steady-state levels of iNOS mRNA were increased by cAMP, one could postulate either a direct effect on transcription or an indirect effect occurring via stimulation of an intermediate regulatory factor or a posttranscriptional effect. If cAMP directly stimulates transcription, then the DNA sequence and transcriptional regulation of the rat gene must be different from those of the mouse macrophage gene, which does not contain binding sites for cAMP-stimulated transcription factors.20
We used two approaches, transient transfection of the iNOS promoter and mRNA stability studies, to show that IL regulates iNOS primarily by a transcriptional mechanism. After transfection of −1592iNOSLuc into ventricular myocytes, stimulation with IL activated the luciferase reporter gene 1.8-fold. In a mouse macrophage cell line, −1592iNOSLuc was activated 150-fold by LPS and IFN.20 Such a discrepancy could be attributable to species-specific differences in regulatory elements in the mouse and rat iNOS genes, or it could reflect different transcriptional mechanisms for IL versus LPS plus IFN. It is also possible that other regions of the iNOS gene, in addition to the −1592 to +160 region used in these studies, contain key regulatory elements. From a methodological point of view, it could also be that traces of LPS in our plasmid preparations desensitized iNOS gene expression to subsequent cytokine stimulation, as suggested by Bogdan et al.32
A second line of evidence indicating a transcriptional effect of IL was our study on mRNA t1/2. Our data suggest that IL did not increase mRNA stability. After treatment with IL for 24 hours, the addition of ActD resulted in rapid decay of iNOS mRNA (t1/2=1 hour). This was not significantly changed by combined treatment with IL and ActD. However, if IL stabilization of mRNA required transcription, then our studies would not have detected a change in mRNA t1/2. Nonetheless, our determination of t1/2 was consistent with studies in which a number of different cell types and inducers of iNOS mRNA were used, including mesangial cells (t1/2=1 hour),30 vascular smooth muscle cells (t1/2=2 hours),8 adult cardiac myocytes (t1/2=4 hours),12 and mouse macrophages (t1/2=3 hours).33
In summary, our studies indicate that IL, acting through a tyrosine kinase–mediated pathway, is the primary cytokine in myocyte iNOS regulation and exerts its effect primarily at the transcriptional level. The IL signaling pathway can be modified by cross talk from other pathways, including those that activate PKC and cAMP.
Selected Abbreviations and Acronyms
|iNOS||=||inducible NO synthase|
|PKC||=||protein kinase C|
|PMA||=||phorbol 12-myristate 13-acetate|
|TNF||=||tumor necrosis factor–α|
This work was supported by grants HL-28982 and HL-03188 from the National Institutes of Health. We thank Dr Charles J. Lowenstein for kindly providing the −1592iNOSLuc plasmid.
Morris SM Jr, Billiar TR. New insights into the regulation of inducible nitric oxide synthesis. Am J Physiol. 1994;266:E829-E839.
Dinerman JL, Lowenstein CJ, Snyder SH. Molecular mechanisms of nitric oxide regulation: potential relevance to cardiovascular disease. Circ Res. 1993;73:217-222.
Hortelano S, Genaro AM, Bosca L. Phorbol esters induce nitric oxide synthase activity in rat hepatocytes: antagonism with the induction elicited by lipopolysaccharide. J Biol Chem. 1992;267:24937-24940.
Muhl H, Pfeilschifter J. Possible role of protein kinase C-ε isoenzyme in inhibition of interleukin 1β induction of nitric oxide synthase in rat renal mesangial cells. Biochem J. 1994;303:607-612.
Kanno K, Hirata Y, Imai T, Marumo F. Induction of nitric oxide synthase gene by interleukin in vascular smooth muscle cells. Hypertension. 1993;22:34-39.
Imai T, Hirata Y, Kanno K, Marumo F. Induction of nitric oxide synthase by cyclic AMP in rat vascular smooth muscle cells. J Clin Invest. 1994;93:543-549.
Xie Q-W, Kashiwabara Y, Nathan C. Role of transcription factor NFκB/rel in induction of nitric oxide synthase. J Biol Chem. 1994;269:4705-4708.
Kamijo R, Harada H, Matsuyama T, Bosland M, Gerecitano J, Shapiro D, Le J, Koh SI, Kimura T, Green SJ, Mak TW, Taniguchi T, Vilcek J. Requirement for transcription factor IRF-1 in NO synthase induction in macrophages. Science. 1994;263:1612-1615.
Harding P, Carretero OA, LaPointe MC. Effects of interleukin-1β and nitric oxide on cardiac myocytes. Hypertension. 1995;25:421-430.
Balligand J-L, Ungureanu-Longrois D, Simmons WW, Pimental D, Malinski TA, Kapturczak M, Taha Z, Lowenstein CJ, Davidoff AJ, Kelly RA, Smith TW, Michel T. Cytokine-inducible nitric oxide synthase (iNOS) expression in cardiac myocytes. J Biol Chem. 1994;269:27580-27588.
Muegge K, Vila M, Gusella GL, Musso T, Herrlich P, Stein B, Durum SK. Interleukin 1 induction of the c-jun promoter. Proc Natl Acad Sci U S A. 1993;90:7054-7058.
Ding AH, Nathan CF, Stuehr DJ. Release of reactive nitrogen intermediates and reactive oxygen intermediates from mouse peritoneal macrophages: comparison of activating cytokines and evidence for independent production. J Immunol. 1988;141:2407-2412.
Lowenstein CJ, Alley EW, Raval P, Snowman AM, Snyder SH, Russell SW, Murphy WJ. Macrophage nitric oxide synthase gene: two upstream regions mediate induction by interferon-γ and lipopolysaccharide. Proc Natl Acad Sci U S A. 1993;90:9730-9734.
Griffiths MJD, Messent M, MacAllister RJ, Evans TW. Aminoguanidine selectively inhibits inducible nitric oxide synthase. Br J Pharmacol. 1993;109:963-968.
Inoue T, Fukuo K, Nakahashi T, Hata S, Morimoto S, Ogihara T. cGMP upregulates nitric oxide synthase expression in vascular smooth muscle cells. Hypertension. 1995;25:711-714.
Mathias S, Younes A, Kan C-C, Orlow I, Joseph C, Kolesnick RN. Activation of the sphingomyelin signaling pathway in intact EL4 cells and in a cell-free system by IL-1β. Science. 1993;259:519-522.
Ungureanu-Longrois D, Balligand J-L, Simmons WW, Okada I, Kobzik L, Lowenstein CJ, Kunkel SL, Michel T, Kelly RA, Smith TW. Induction of nitric oxide synthase activity by cytokines in ventricular myocytes is necessary but not sufficient to decrease contractile responsiveness to β-adrenergic agonists. Circ Res. 1995;77:494-502.
Palmer JN, Hartogensis WE, Patten M, Fortuin FD, Long CS. Interleukin-1β induces cardiac myocyte growth but inhibits cardiac fibroblast proliferation in culture. J Clin Invest. 1995;95:2555-2564.
Corbett JA, Kwon G, Misko TP, Rodi CP, McDaniel ML. Tyrosine kinase involvement in IL-1β-induced expression of iNOS by β-cells purified from islets of Langerhans. Am J Physiol. 1994;267:C48-C54.
Taniguchi T. Cytokine signaling through nonreceptor protein tyrosine kinases. Science. 1995;268:251-255.
Tsujino M, Hirata Y, Imai T, Kanno K, Eguchi S, Ito H, Marumo F. Induction of nitric oxide synthase gene by interleukin-1β in cultured rat cardiocytes. Circulation. 1994;90:375-383.
Vodovotz Y, Bogdan C, Paik J, Xie Q-W, Nathan C. Mechanisms of suppression of macrophage nitric oxide release by transforming growth factor β. J Exp Med. 1993;178:605-613.
Kunz D, Muhl H, Walker G, Pfeilschifter J. Two distinct signaling pathways trigger the expression of inducible nitric oxide synthase in rat renal mesangial cells. Proc Natl Acad Sci U S A. 1994;91:5387-5391.
Bogdan C, Vodovotz Y, Paik J, Xie Q-W, Nathan C. Traces of bacterial lipopolysaccharide suppress IFN-γ-induced nitric oxide synthase gene expression in primary mouse macrophages. J Immunol. 1993;151:301-309.
Lorsbach R, Murphy W, Lowenstein C, Snyder S, Russel S. Expression of the nitric oxide synthase gene in mouse macrophages activated for tumor cell killing. J Biol Chem. 1993;268:1908-1913.