Nitric oxide synthesis in endothelial cells: evidence for a pathway inducible by TNF-a SANTIAGO LAMAS, THOMAS MICHEL, BARRY M. BRENNER, AND PHILIP A. MARSDEN Renal and Cardiology Divisions, Department of Medicine, Brigham and Women’s Hospital and the Harvard Center for the Study of Kidney Diseases, Harvard Medical School, Boston, Massachusetts 02115

LAMAS,SANTIAGO,THOMASMICHEL,BARRY M. BRENNER, AND PHILIP A. MARSDEN. Nitric oxide synthesis in endothelial cells: evidence for apathway inducible by TNF-a. Am. J. Physiol. 261 (Cell Physiol. 30): C634C641, 1991.-Nitric oxide (NO) release accounts for the biological activity of endotheliumderived relaxing factor. Given that tumor necrosis factor-a (TNF-ac) has been implicated as an important mediator in septic shock, we explored whether TNF-cu enhances L-argininedependent synthesis of NO and L-citrulline in endothelial cells. The release of NO was detected in a coincubation bioassay where measurement of guanosine 3’,5’-cyclic monophosphate (cGMP) production in reporter monolayers, namely glomerular mesangial cells or fetal lung fibroblasts, reflected activation of soluble guanylate cyclase. Reporter monolayer cGMP content was greater in the presence of TNF-a-treated bovine aortic and renal artery endothelial cells than in the presence of vehicletreated endothelial cells. TNF-a-stimulated endothelium-dependent increases in reporter monolayer cGMP content were first evident at 8 h and maximal at 16-24 h. In addition, TNFa-stimulated endothelium-dependent increases in reporter monolayer cGMP content were abrogated by hemoglobin and methylene blue, blunted by N”-nitro-L-arginine and augmented by superoxide dismutase and the calcium agonist bradykinin. These observations suggested that TNF-a! enhanced release of NO. Furthermore, the formation of L-[14C]citrulline from L[ 14C]arginine, as determined by quantitative cation-exchange chromatography and thin-layer chromatography, was enhanced by TNF-cu in a time- and concentration-dependent manner. Thus it is evident that endothelial cells release NO for a prolonged period in response to TNF-cr and transiently when stimulated with calcium agonists. The prolonged release of NO from TNF-a-stimulated endothelial cells may be implicated in the pathogenesis of septic shock. aorta; endothelium; endothelium-derived relaxing factor; cytokines; guanosine 3’,5’-cyclic monophosphate; kidney; mesangial cells; septic shock

TUMORNECROSIS FACTOR-~ (TNF-cr)isconsideredtobe

an important mediator in the pathogenesis of septic shock (4, 38). Animals infused with TNF-a! developed changes in systemic hemodynamics including profound reductions in peripheral vascular resistance and organ dysfunction characteristic of gram-negative septicemia or lipopolysaccharide (LPS) administration (39). The mechanism(s) whereby TNF-ar induces hypotension, however, is incompletely understood. Vascular endothelial cells produce nitric oxide (NO), or a closely related substance, from the guanidino nitrogen(s) of the amino acid L-arginine. Endothelial-derived C634

0363-6143/91

$1.50

Copyright

nitric oxide (EDNO) accounts, in part, for the potent labile vasodilator endothelium-derived relaxing factor (EDRF) (14, 16, 33). EDNO plays a major role in the regulation of local organ blood flow, vasomotor tone, and systemic blood pressure (35) through activation of soluble guanylate cyclase in vascular smooth muscle (26). Basal release of NO from the vascular endothelium modulates vasomotor tone in a variety of species, including humans (40). Endothelium-dependent vasodilators, such as bradykinin and acetylcholine, rapidly but transiently augment the release of EDNO. Agonists that release NO from the endothelial lining of vascular tissues or cultured endothelial cells share in common the ability to elevate the intracellular calcium concentration (25). Agonistinduced release of EDNO is a calcium/calmodulin-dependent process (7). Evidence now suggests constitutive expression of calcium-activated NO synthase in brain (6), nonadrenergic noncholinergic peripheral nerves (8), platelets (34), and the adrenal gland (32). In contrast, nitrogen oxide biosynthesis can be induced in other cell types that do not express NO synthase activity under basal conditions: macrophages (11, 37), Kupffer cells (5), hepatocytes (9), tumor cells (l), glomerular mesangial cells (24)) and vascular smooth muscle cells (3). Activated macrophages express NO synthase activity that is not calcium dependent (22). It has been suggested that the enzymes responsible for NO synthesis may constitute a family of at least two distinct types: type I enzyme being the induced NO synthase of macrophages and type II enzyme the NO synthase constitutively expressed in endothelial and cerebellar cells (31). Therefore, nitrogen oxide biosynthesis is a widely expressed autocrine and paracrine signaling pathway (26, 29). Because of the central role of TNF-cu in the hemodynamic alterations of septic shock, we examined the effect of TNF-cu on EDNO release and L-arginine metabolism in homogeneous populations of cultured endothelial cells derived from bovine aorta and renal artery. We provide evidence that TNF-cu induces NO synthase activity in endothelial cells and that induction of enzymatic activity occurs in cells that exhibit constitutive calcium-regulated release of NO. METHODS Materials. Cell culture media and balanced salt solutions were purchased from GIBCO (Grand Island, NY), low endotoxin defined supplemented bovine calf serum

0 1991 the American

Physiological

Society

Downloaded from www.physiology.org/journal/ajpcell by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on August 23, 2018. Copyright © 1991 American Physiological Society. All rights reserved.

TNF-a

ENHANCES

THE

RELEASE

(SCS) from Hyclone Labs (Logan, UT), guanosine 3’,5’cyclic monophosphate (cGMP) radioimmunoassay from Biomedical Technologies (Stoughton, MA), cell culture plates from Costar (Cambridge, MA), Millicell culture inserts from Millipore Products Division (Bedford, MA), and glass cover slips from Bellco Biotechnology (Vineland, NJ). Human recombinant tumor necrosis factor-a (rHuTNF-a, sp act, 9.8 x 106/mg) was a gift of Knoll Pharmaceuticals (Whippany, NJ). [8,5’-3H]cGMP (sp act, 33.3 Ci/mmol), L-[4,5-3H]leucine (sp act, 53 Ci/ mmol), and L-[2,3-3H]arginine (sp act, 40-70 Ci/mmol) were obtained from New England Nuclear (Wilmington, DE), L-[U-14C]arginine (sp act, 305 mCi/mmol) and L[carbamoyl-14C]citrulline (sp act, 50-60 mCi/mmol) were from Amersham (Arlington Heights, IL). Dowex AG 5OWX-8 cation-exchange resin (100-200 mesh) was from Bio-Rad (Richmond, CA), and thin-layer chromatography silica gel 60 plates were from Alltech Associates (Deerfield, IL). L-arginine, N”-nitro-L-arginine (LNNA), LPS (Escherichia coli serotype 0.26:B6, phenol extracted), 3-isobutyl-1-methylxanthine (IBMX), bovine liver superoxide dismutase, bovine red blood cell hemoglobin, methylene blue, L-arginine, L-citrulline, cGMP, lithium chloride, and all other reagents were purchased from Sigma Chemical (St. Louis, MO). Cell isolation and culture. Bovine aortic endothelial cells (BAEC) and bovine renal artery endothelial cells (BRAE) were isolated from the thoracic aorta and main renal artery, respectively, using published methods (25). Primary cultures of endothelial cell clones were initiated on loo-mm tissue culture plates that had been precoated with gelatin (0.2 g/dl), isolated with cloning cylinders, detached with trypsin-EDTA, and passaged at cloning density onto gelatin (0.2 g/dl)-coated loo-mm plates. To obtain homogeneous endothelial cell cultures, single clones were isolated a second time with cloning cylinders to establish subcloned populations of cells. Individual clones were examined for angiotensin I-converting enzyme activity, expression of factor VIII-related antigen, and uniform uptake of fluorescent acetylated low-density lipoprotein (LDL) as described. Fluorescence-activated cell sorting (FACS) of clones labeled with fluorescent acetylated LDL confirmed that endothelial cell clones represented homogeneous populations of cells that were both uniformly labeled and clearly distinguishable from cloned populations of mesangial cells or vascular smooth muscle cells. Endothelial cells were fed every 48 h with RPM1 1640 medium supplemented with L-glutamine (300 mg/l), 15% low endotoxin SCS, 100 U/ml penicillin, and 100 pg/ml streptomycin. Cells were utilized at passages 5-10. Working concentrations of rHuTNF-a and culture medium contained 95% of endothelial cells. Furthermore, LDH activities in culture supernatants were 60.2 t 9.7 and 69.2 t 6.4 U/bin vehicle- and TNF-a-treated cells, respectively (24 h, 100 rig/ml, triplicate determinations; Kodak Ektachem quantitative kinetic determination, Eastman Kodak). Figure 2 demonstrates the time-dependent increase in reporter mesangial cGMP levels following treatment of BAEC with 100 rig/ml rHuTNF-cu. BAEC and GMC were coincubated for the final 1 h. As can be seen, TNFa-induced increases in GMC cGMP levels were first evident at 8 h and maximal at 16-24 h. Although not shown, addition of rHuTNF-a to the coincubation assay for 10 min failed to have a significant effect on GMC cGMP content (n = 4, triplicate determinations).

Downloaded from www.physiology.org/journal/ajpcell by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on August 23, 2018. Copyright © 1991 American Physiological Society. All rights reserved.

TNF-cu

ENHANCES

THE

RELEASE

OF

NO

FROM

l.O-

ENDOTHELIAL

C637

CELLS

1.0 1

;

0.8-

Tii 0 CD z . z E

0.6-

0.4-

z -Q

0.4

;

z (3 0

0.2

0.2-

W

0.6 0 F N

25 (3 0

0.0 0.0

I I

I

I

I

I

I

I

1

0

4

8

12

16

20

24

TIME

(h)

FIG. 2. Time dependence of TNF-cu-stimulated endothelial-dependent increases in GMC-associated cGMP content. IBMX (1 mM) was added 10 min before cGMP extraction. BAEC were treated with 100 rig/ml rHuTNF-cu for the indicated periods before coincubation with GMC for 1 h. Data points represent means t SE of 4 experiments (triplicate determinations). Where error bars are not evident, SE was smaller than data point. * P < 0.05 vs. control (0 h). In the absence of endothelial cells, mesangial cell cGMP content averaged 0.07 t 0.01 pmol/lO” cells.

TNF-a-stimulated BAEC-dependent increases in reporter monolayer cGMP content were not dependent on the reporter monolayer studied. Rat fetal lung fibroblast (RFL-6) cell-associated cGMP levels were higher in the presence of BAEC than in their absence, an effect which was clearly augmented by rHuTNF-cu (100 rig/ml, 24 h; 0.07 t 0.01, 0.23 t 0.06, and 0.78 t 0.14 pmol/106 RFL6 cells for RFL-6 monolayers and RFL-6 monolayers in the presence of vehicle- and TNF-a-treated BAEC, respectively; n = 3, triplicate determinations; P < 0.05). Consistent with results obtained with BAEC donor monolayers, mesangial cell-associated cGMP levels were greater in the presence of TNF-a-treated BRAE [0.07 t 0.01, 0.11 t 0.01, and 0.30 t 0.01 pmol/106 cells, for control reporter GMC monolayers and reporter monolayers with vehicle- or TNF-a-treated BRAE (100 ng/ ml), respectively; n = 3, triplicate determinations; P c 0.051. Moreover, TNF-a-stimulated increases in mesangial cell-associated cGMP levels were evident when bradykinin was added to the BRAE/GMC coincubation assay (100 nM, 2 min; 0.27 t 0.02 compared with 0.49 t 0.05 pmol/106 cells for bradykinin-stimulated vehicleand TNF-a-treated BRAE, respectively; n = 3, triplicate determinations; P < 0.05). The effects of pharmacological agents on TNF-cu-stimulated endothelial-dependent GMC cGMP accumulation are shown in Fig. 3. Hemoglobin (oxygenated, 10 PM) and methylene blue (10 PM) lowered BAEC-dependent mesangial cell cGMP content. The results shown in Fig. 3 demonstrate inhibition of TNF-a-induced increases on average of 87.4 t 1.7 and 83.6 t 6.4%, respectively (n = 3, triplicate determinations). Hemoglobin and methylene blue also decreased mesangial cell-associated cGMP levels in the presence of bradykinin-stimulated TNF-w treated BAEC (n = 3, triplicate determinations, data not shown). Hemoglobin and methylene blue failed to have a significant effect on mesangial cell cGMP content in

-TNF

tTNF

FIG. 3. Effects of hemoglobin and methylene blue on endothelialdependent GMC-associated cGMP content. BAEC were treated for 24 h with vehicle or rHuTNF-cu (100 rig/ml) before coincubation with GMC for 1 h. The coincubation system was treated for 15 min with vehicle (solid bars), hemoglobin (10 PM; hatched bars), or methylene blue (10 PM; grey bars). IBMX (1 mM) was added 10 min before cGMP extraction. Columns represent means t SE of 3 experiments (triplicate determinations). t P < 0.05 vs rHuTNF-a-treated BAEC. In the absence of endothelial cells, mesangial cell cGMP content averaged 0.03 k 0.02 pmol/106 cells.

the absence of endothelial cells (0.03 t 0.01, 0.03 t 0.01, and 0.04 t 0.01 pmol/106 cells for vehicle-, hemoglobin-, and methylene blue-treated GMC, respectively; n = 3, triplicate determinations). The finding that TNF-a-stimulated BAEC-dependent GMC cGMP accumulation was abrogated by known inhibitors of EDRF action, namely hemoglobin and methylene blue, suggested that TNF-cu enhanced the production of EDNO by BAEC. To further characterize whether TNF-cu enhanced the release of NO it was of interest to determine whether superoxide dismutase (SOD), which is known to prolong the pharmacological half-time ( t112) of EDNO, modulated the effect of TNF-cu on BAECdependent cGMP production in reporter monolayers. Experiments were performed by inserting inverted GMC cover slips in the upper chamber of Millicell culture inserts. Inserts were then placed in wells that contained no donor monolayer or BAEC monolayers, vehicle or rHuTNF-a! treated. The distance between donor and reporter monolayers in this culture insert assay is greater than when cell monolayers are juxtaposed. Because the pharmacological t 1/2 of EDNO is very short, increasing the physical separation of endothelial cells and reporter GMC might be expected to attenuate endothelium-dependent increases in reporter cGMP content. Shown in Fig. 4, GMC cGMP levels were higher is the presence of BAEC than in their absence, but only-in the presence of SOD (200 U/ml, 15 min). Moreover, in the presence of SOD, levels of GMC cGMP were 52.8 + 4.4% greater in the presence of TNF-a-treated BAEC compared with vehicle-treated BAEC. It is unlikely that uptake of extracellular cGMP, released by donor monolayers, mediated the observed increase in reporter monolayer cGMP content. GMC uptake of [3H]cGMP averaged 0.11 and 0.05% of total added radioactivity at 60 and 10 min, respectively. The presence of rHuTNF-cr (100 rig/ml) for 60 min did not modify GMC uptake of [3H]cGMP. The observation that TNF-a-treated BAEC increased the

Downloaded from www.physiology.org/journal/ajpcell by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on August 23, 2018. Copyright © 1991 American Physiological Society. All rights reserved.

C638

TNF-cu

ENHANCES

-SOD

THE

RELEASE

tSOD

FIG. 4. Effect of superoxide dismutase (SOD) on rHuTNF-a-induced increases in GMC-associated cGMP content. BAEC were treated for 24-h with vehicle or rHuTNF-cu (100 rig/ml). In this experiment GMC were grown on cover slips and inserted inverted in the top chamber of Millicell culture inserts. Inserts were incubated for 1 h in the absence or presence of vehicle(stippled bars) or rHuTNF-atreated cells (hatched bars). SOD (200 U/ml) or vehicle were added for 15 min to the coincubation system. IBMX (1 mM) was added 10 min before cGMP extraction. Solid bars represent GMC-associated cGMP levels in the absence of BAEC (reporter). Columns represent means k SE of 3 experiments (duplicate determinations). * P < 0.05 vs. BAEC in the absence of rHuTNF-cu. T P < 0.05 vs. BAEC in the absence of SOD.

1. TNF-a-induced endothelial-dependent increases in GMC-associated cGMP are L-arginine dependent

TABLE

Mesangial Cell cGMP, pmol/106 cells

Condition

None Vehicle-treated TNF-a-treated

BAEC BAEC

(-)-L-NNA

(+)-L-NNA

0.04t0.01 0.31t0.04 0.64t0.15

0.04~0.01 0.08~0.02* 0.18t0.04*

Values are means t SE of 3 experiments (triplicate determinations). Bovine aortic endothelial cells (BAEC) were treated for 24 h with vehicle or recombinant tumor necrosis factor-a (100 rig/ml). For these studies the incubation buffer was devoid of L-arginine. The coincubation system was treated with vehicle or N”-nitro-L-arginine (L-NNA; 500 PM) for 15 min. 3-Isobutyl-1-methylxanthine (1 mM) was added 10 min before cGMP extraction. * P < 0.05 vs. BAEC in the absence of L-NNA.

cGMP content of reporter monolayers in Millicell culture inserts only in the presence of SOD is consistent with the hypothesis that TNF- CYaugments the release of EDNO from BAEC. TNF-a-induced BAEC-dependent increases in GMC cGMP levels are inhibited by L-NNA. As shown in Table 1, L-NNA (500 PM, 15 min) decreased mesangial cellassociated cGMP levels in the presence of vehicle- and TNF-a-treated BAEC by 73.4 t 5.6 and 72.6 t 1.4%, respectively. As shown, mesangial cell cGMP content in the absence of endothelial cells was not modified by LNNA. BAEC-dependent GMC cGMP accumulation is augmented by LPS. It was of interest to determine whether treatment of BAEC with the microbial wall product LPS could modulate the cGMP levels in reporter monolayers. LPS treatment of BAEC (10 pg/ml, 24 h) increased mesangial cell cGMP content (0.32 t 0.02, 0.60 t 0.06,

OF

NO

FROM

ENDOTHELIAL

CELLS

and 0.57 t 0.04 pmol/106 cells for vehicle-, LPS-, and TNF-a-treated BAEC, respectively; n = 3, triplicate determinations; P < 0.05 for LPS and TNF-cu treatment vs. vehicle treatment). LPS addition for 1 h had no significant effect on GMC cGMP content in the presence or absence of BAEC (n = 3, triplicate determinations). TNF-LX augments the production of f%‘]citrulline from [14C]arginine. Given that L-arginine serves as the metabolic substrate for NO synthase and that TNF-cw-induced endothelium-dependent increases in mesangial cell cGMP content were L-arginine dependent, we determined the metabolic fate of radioactive L-arginine in BAEC. Treatment of BAEC with TNF-a-augmented the conversion of L- [ 14C]arginine to L- [ 14C]citrulline as determined by quantitative cation-exchange chromatography. Shown in Fig. 5A, after 24-h addition of 100 rig/ml rHuTNF-cu, conversion of L- [ 14C]arginine to L- [‘“Clcitrulline was augmented -3.5-fold above rates of conversion in vehicle-treated cells (n = 3, triplicate determinations). Figure 5 B demonstrates the time dependence of the TNF-cu effect. L-[14C]citrulline formation over a 3-h period was determined in BAEC following treatment with rHuTNF-cw (100 rig/ml) for the indicated times. Although TNF-a-stimulated increases in L- [ 14C]arginine to L-[‘4C]citrulline conversion were first evident at 12 h, a more robust effect was evident at 24 h (n = 3, duplicate determinations). TLC confirmed that 24-h treatment of BAEC with rHuTNF-cu (100 rig/ml) augmented production of L- [ 14C]citrulline from L- [ 14C]arginine (data not shown). Figure 6 shows the concentration dependence of TNF-a-induced increases in L- [ 14C]citrulline formation from L- [ 14C]arginine measured at 24 h. Threshold effects of rHuTNF-a! were observed at l-10 rig/ml. Concentrations of rHuTNF-a required to stimulate half-maximal effects were determined by log-logit analysis and averaged 65.8 t 3.8 rig/ml (n = 3, duplicate determinations). 700

2500 -

z

v)

600

Q) CD o

jj 500

0 7

2

2000

CD 0

400

z 1500 CT

Q

2 a

300

E 3

200

E is

100 0

f .-E

1000

&3 C is

500

0 0

60 TIME

120

(min)

180

0

4

8

12

16

20 24

TIME (h)

FIG. 5. Effect of TNF-cu on L- [ 14C]arginine-L[ r4C]citrulline conversion in BAEC. A: L-[14C]arginine (5 x lo5 cpm/ml) was added for the indicated times to BAEC grown in 12-well plates and treated with vehicle (open circles) or rHuTNF-a (100 rig/ml, 24 h) (closed circles). After extraction with 1 ml of ice-cold 15% trichloroacetic acid, L-[~~C]citrulline was determined by quantitative cation-exchange chromatography. Each point represents means & SE of 3 experiments (triplicate determinations). Where error bars are not evident, SE was smaller than data point. The 2 curves are statistically different (ANOVA, P < 0.05). B: BAEC were grown in 6-well plates and treated for the indicated times with rHuTNF-cu (100 rig/ml). L-[14C]arginine ( lo6 cpm/ml) was added 3 h before termination of experiment. L-[14C]citrulline was determined by quantitative cation-exchange chromatography. Each point represents means k SE of 3 experiments (duplicate determinations).

Downloaded from www.physiology.org/journal/ajpcell by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on August 23, 2018. Copyright © 1991 American Physiological Society. All rights reserved.

TNF-cu

ENHANCES

THE

RELEASE

W 0

2 300 CL 0 zc 200 .3 E 0

100

0

!

m

-

v m ““,

+q

0

m

0.1

TNF

m

. . ‘..‘,

.

1

. .

‘.W..,

.

10

. . . . . ..I

.

100

.

. mlq

1000

(nglml)

FIG. 6. Effect of different concentrations of rHuTNF-a on L-['~C]arginine-L-[14C]citrulline conversion in BAEC. L-[14C]arginine (2.5 x 10' cpm/ml) was added for 3 h to BAEC grown in 6-well plates and treated with rHuTNF-cu for 24 h. After extraction with 1 ml of ice-cold 15% trichloroacetic acid, L-[14C]citrulline was determined by quantitative cation-exchange chromatography. Each point represents means 2 SE of 3 experiments (duplicate determinations). Where error bars are not evident, SE was smaller than data point. * P c 0.05 vs. control.

5000 -cn

1 * t

g 4000 -I w

I 0

OF

NO

FROM

ENDOTHELIAL

CELLS

C639

radioactive L-arginine. Although incorporation of L[ 14C]arginine (1 &i/ml) into TCA-precipitable radioactivity tended to be lower in BAEC treated for 24 h with rHuTNF-cu (100 rig/ml) compared with vehicle-treated BAEC (4.9 t 0.8 x lo5 cpm and 6.0 t 0.4 X lo5 cpm for TNF-a- and vehicle-treated cells, respectively; n = 3, triplicate determinations) the difference failed to reach significance. Similarly, incorporation of L- [ 3H] arginine (1 &i/ml) into TCA-precipitable radioactivity was not significantly affected by 24-h treatment with TNF-a (data not shown). Furthermore, incorporation of the neutral amino acid L- [3H]leucine (1 &i/ml) into TCA precipiate over a l-h period tended to be lower in BAEC treated for 24 h with rHuTNF-a! (100 rig/ml) compared with vehicle-treated cells, although again failed to achieve significance (836 t 35 and 1,057 t 53 cpm for TNF-cu- and vehicle-treated BAEC; 6 determinations). We further evaluated whether TNF-cu modified the uptake of extracellular L-[3H]arginine (0.1 &i/ml) in BAEC. Uptake of L-[3H]arginine over 60 min into vehicle- or TNF-a-treated BAEC (24 h, 100 rig/ml) averaged 1.2 + 0.1 and 1.1 t 0.1 X lo4 cpm, respectively (triplicate determinations). Furthermore, when measured at 5, 10, or 20 min, uptake of L-[3H]arginine into vehicle- or TNFa-treated BAEC was not significantly different (data not shown).

I

I

DISCUSSION

-TNF

+TNF

FIG. 7. Effect of calcium ionophore (A23187) on L-[14C]arginine to L- [ 14C]citrulline conversion in vehicleand TNF-a-treated BAEC. L[14C]arginine (6 x 10’ cpm/ml) was added 3 h before the termination of the experiment to BAEC grown in 6-well plates treated with vehicle (control) or rHuTNF-cu (100 rig/ml, 24 h). A23187 (10 FM; hatched bars) or vehicle (solid bars) were added 15 min before termination of the experiment. After extraction with 1 ml of ice-cold 15% trichloroacetic acid, L- [ “C]citrulline was determined by quantitative cationexchange chromatography. Each column represents means k SE of 9 determinations from 3 separate experiments. * P < 0.05 vs. control; t P c 0.05 vs. vehicle-treated BAEC.

Analysis of variance suggested that maximal responses were observed at 100-250 rig/ml. Figure 7 shows the effect of TNF-a! on BAEC L-[~~C]citrulline formation in the absence and presence of the calcium ionophore A23187 (10 PM, 15 min). L-[~~C]citrulline formation increased in BAEC in response to A23187 consistent with the observation that constitutive NO synthase activity in endothelial cells is calcium/ calmodulin dependent. Basal rates of L- [ “C]citrulline formation were significantly enhanced by rHuTNF-cw (100 rig/ml, 24 h) and augmented in an additive manner by 15 min of A23187. Given that prolonged treatment of BAEC with TNFcyenhanced the conversion of L- [ 14C]arginine to L- [ 14C]citrulline, we further evaluated the metabolic fate of

We developed a coincubation assay that uses cGMP production in reporter monolayers, GMC or RFL-6, as a measure of EDNO action to determine whether TNF-cu modulated basal and agonist-induced release of NO from endothelial cell monolayers. The current study demonstrates that TNF-CY. acts on endothelial cells derived from two different sources to enhance the release of a soluble mediator that stimulates cGMP production in adjacent reporter monolayers. Studies performed to determine the time dependence of the TNF-cu effect demonstrated that this endothelium-dependent effect was first evident at 8 h and maximal at 16-24 h. Moreover, TNF-a-stimulated endothelium-dependent increases in reporter cGMP content were evident following addition of the endotheliumdependent vasodilator bradykinin, suggesting that TNFcy enhanced release of NO in the basal state and in an additive manner to a calcium agonist. TNF-a-stimulated endothelium-dependent increases in reporter monolayer cGMP content were abrogated by methylene blue and hemoglobin, both known to inhibit NO action. Moreover, in the current study, levels of cGMP were higher in reporter GMC menolayers cocultured in Millicell culture inserts with TNF-a-treated BAEC compared with vehicle-treated BAEC, but only in the presence of SOD. The findings that the TNF-cu effect was abrogated by methylene blue and hemoglobin and that SOD and bradykinin potentiated the effect are interpreted to indicate that TNF-cu stimulated EDNO release with consequent stimulation of mesangial cell soluble guanylate cyclase activity. Previous investigations have suggested that TNF-Q treatment for 2 h inhibits the release of EDRF from carotid arteries (2) and isolated perfused hearts (23). The nature of the mechanism by which TNF-a! inhibits

Downloaded from www.physiology.org/journal/ajpcell by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on August 23, 2018. Copyright © 1991 American Physiological Society. All rights reserved.

C640

TNF-a)

ENHANCES

THE

RELEASE

EDRF release and/or action was not evaluated. These previous studies did not address the long-term response to TNF-cu. Although we found levels of cGMP in reporter GMC monolayers to be lower after treatment of the coincubation assay with TNF-cu for 60 min, the results failed to achieve significance (Fig. 2). EDNO is synthesized from the amino acid L-arginine. L-arginine analogues, such as L-NNA, are potent inhibitors of NO synthase (7,17, 32). In the current study we have demonstrated that L-NNA inhibited TNF-a-stimulated endothelium-dependent increases in reporter GMC cGMP content. It was therefore of interest to determine the metabolic fate of extracellular L-arginine in TNF-ar- and vehicle-treated BAEC. Although TNF-cu failed to have a major effect on the uptake of extracellular L-[3H]arginine in BAEC and further failed to significantly modulate the incorporation of L-[~H]- or L-[~~C]arginine into nascent BAEC protein, it had a profound effect on the conversion of L-[ 14C]arginine to L-[~~C] citrulline as determined by quantitative cation-exchange chromatography and TLC. Induction of L- [ 14C]arginine to L-[‘4C]citrulline conversion by TNF-ar exhibited a time dependency qualitatively similar to TNF-a!-stimulated increases in NO release. Hecker et al. (15) reported that BAEC did not metabolize L-[14C]arginine to L-[~~C] citrulline in the absence of calcium ionophore. In the current study, calcium ionophore enhanced arginine to citrulline conversion to the same degree in vehicle- and TNF-a-treated cells. Therefore, the observations in the current study, that TNF-a-treated BAEC produce significant amounts of L-[14C]citrulline in the absence of calcium ionophore, that vehicle-treated BAEC produce significant amounts of L- [ 14C]citrulline in the presence of calcium ionophore, and that L-[14C]citrulline production in the presence of both TNF-a! and calcium ionophore is additive, suggest that TNF-a-induced and Ca2+activated NO synthase activity in BAEC are independent enzymatic processes. Evidence presented in this study suggests that the calcium-sensitive enzyme is not induced. However, further studies are necessary to characterize the biochemistry of cofactor requirements and cellular localization of the NO synthase isoform induced by TNF-cu. Nonetheless, the most reasonable interpretation of our observations is that TNF-ar induces NO synthase activity in BAEC and that the NO synthase activity induced most likely represents the Ca2+-calmodulin-independent NO synthase suggested to be present in the endothelial cells (13, 30). Thus it is evident that endothelial cell release of NO can be rapidly stimulated in response to calcium agonists, such as bradykinin, and also induced over several hours by a protein mediator of the effector phase of host defense, namely TNF-cu, and a microbial wall product, namely LPS. Recent studies reported that mouse brain endothelial cells can release nitrate and nitrite, oxidation products of NO, in response to interferon-y combined with LPS, TNF-ar or interleukin-lp (20). In the absence of interferon-y, cytokines and microbial wall products were without effect. Furthermore, these studies did not indicate whether these endothelial cells release NO in a calcium-regulated fashion. The findings in the current study, that TNF-a! alone induces NO generation in endothelial cells and that cytokine-induced and Ca2+-acti-

OF

NO

FROM

ENDOTHELIAL

CELLS

vated NO generation exists in the same cell type, are novel in this regard. That endothelial cell clones used in the current study were cloned, subcloned, and then characterized by FACS analysis following labeling of cells with fluorescent acetylated LDL excludes the possibility that cultures were contaminated with other cell types. These studies demonstrate that TNF-cu and LPS induce, in a time-dependent manner, a biosynthetic pathway for the prolonged production of NO in endothelial cells. Moreover, TNF-cu induces NO synthase activity in cells that constitutively express the capacity to respond to calcium agonists with rapid NO release and citrulline formation. Induction of NO synthase activity in response to TNF-cw was first evident at 8-12 h. Concentrationresponse relationships for TNF-a-induced citrulline formation in BAEC demonstrated that l-10 rig/ml rHuTNF-a elicited a threshold response. Of interest, these concentrations are only l- to loo-fold greater than serum levels of TNF-a observed in humans with severe sepsis (10). Given that circulating levels of cytokines may not accurately reflect local tissue concentrations, paracrine release may result in much greater tissue concentrations of cytokines (12,41). Recent observations suggest that the loss of vascular responsiveness induced by endotoxin or TNF-cu can be reversed by L-arginine analogues known to inhibit NO synthase (19, 21) and that the vascular endothelium contributes to the impairment of vascular contractility (28). In this regard, previous studies from this laboratory have demonstrated that TNF-cu activates soluble guanylate cyclase in mesangial and vascular smooth muscle cells (3,24). In these cell types induction of NO synthase activity is evident at 8 h and maximal at 18-24 h. It is therefore possible that cytokine-induced formation of NO in vascular tissues is both endothelium dependent and independent. Although these studies do not address whether TNF-a! induces endothelial NO synthase activity in vivo, it is possible that endothelial-derived NO contributes to the systemic hypotension that is characteristic of septic shock. NOTEADDEDINPROOF

The observation that endothelial cells express different forms of nitric oxide synthase was recently reported. (M. W. Radomski, R. M. J. Palmer, and S. Moncada. Proc Nutl. Ad. Sci. USA 87: 10043-10047,199O).

We thank Knoll Pharmaceuticals (Whippany, NJ) for supplying human recombinant tumor necrosis factor-a. We are grateful to Elizabeth VanDeCarr for excellent technical assistance and Drs. C. Carpenter, C. Serhan, and J. Loscalzo for helpful comments and suggestions. S. Lamas is a fellow of the Fulbright Foundation-Ministerio de Education y Ciencia, Spain and supported in part by the Spanish Society of Nephrology. T. Michel is a recipient of a Clinician Scientist Award from the American Heart Association. P. A. Marsden is a Medical Research Council of Canada fellowship recipient. This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-35930. Portions of this work were presented in abstract form at the 1990 Annual Meeting of the American Society of Nephrology, Washington, DC. Address for reprint requests: P. A. Marsden, Rm. 7360, Med. Sci.

Downloaded from www.physiology.org/journal/ajpcell by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on August 23, 2018. Copyright © 1991 American Physiological Society. All rights reserved.

TNF-a Bldg., Univ. of Toronto, lA8, Canada. Received

31 January

1 King

1991;

ENHANCES

College

accepted

Circle,

in final

THE Toronto,

form

29 May

RELEASE Ontario

M5S

1991.

REFERENCES 1. AMBER, I. J., J. B. J. HIBBS, R. R. TAINTOR, AND Z. VAVRIN. Cytokines induce an L-arginine-dependent effector system in nonmacrophage cells. J. Leukocyte BioZ. 44: 58-65, 1988. AND A. M. LEFER. Anti-EDRF effect of 2. AOKI, N., M. SIEGFRIED, tumor necrosis factor in isolated, perfused cat carotid arteries. Am. J. Physiol. 256 (Heart Circ. Physiol. 25): H1509-H1512, 1989. 3. BEASLEY, D., J. H. SCHWARTZ, AND B. M. BRENNER. Interleukin1 induces prolonged L-arginine-dependent cyclic guanosine monophosphate and nitrite production in rat vascular smooth muscle cells. J. CZin. Inuest. 87: 602-608, 1991. 4. BEUTLER, B., I. MILSARK, AND A. C~RAMI. Passive immunization against cachectin/tumor necrosis factor protects mice from lethal effect of endotoxin. Science Wash. DC 229: 869-871, 1985. 5. BILLIAR, T. R., R. D. CURRAN, D. J. STUEHR, M. A. WEST, B. G. BENTZ, AND R. L. SIMMONS. An L-arginine-dependent mechanism mediates Kupffer cell inhibition of hepatocyte protein synthesis in vitro. J. Exp. Med. 169: 1467-1472, 1989. 6. BREDT, D. S., AND S. H. SNYDER. Nitric oxide mediates glutamatelinked enhancement of cGMP levels in the cerebellum. Proc. N&Z. Acad. Sci. USA 86: 9030-9033,1989. 7. BREDT, D. S., AND S. H. SNYDER. Isolation of nitric oxide synthetase, a calmodulin-requiring enzyme. Proc. N&Z. Acud. Sci. USA 87: 682-685,199O. 8. BULT, H., G. E. BOECKSTAENS, P. A. PELCKMANS, J. A. JORDAENS, Y. M. VAN MAERCKE, AND A. G. HERMAN. Nitric oxide as an inhibitory non-adrenergic non-cholinergic neurotransmitter. Nuture Lond. 345: 346-347,199O. 9. CURRAN, R. D., T. R. BILLIAR, D. J. STUEHR, K. HOFFMAN, AND R. L. SIMMONS. An L-arginine-dependent mechanism mediates Kupffer cell inhibition of hepatocyte protein synthesis in vitro. J. Exp. Med. 170: 1769-1774,1989. 10. DAMAS, P., A. REUTER, P. GYSEN, J. DEMONTY, M. LAMY, AND P. FRANCHIMONT. Tumor necrosis factor and interleukin-1 serum levels during severe sepsis in humans. Crit. Cure Med. 17: 975-978, 1989. 11. DRAPIER, J.-C., J. WIETZERBIN, AND J. R. T. HIBBS. Interferon-y and tumor necrosis factor induce the L-arginine-dependent cytotoxic effector mechanism in murine macrophages. Eur. J. Immunol. 18: 1587-1592,1988. 12. DUNCOMBE, A. S., AND M. K. BRENNER. Is circulating tumor necrosis factor bioactive? N. EngZ. J. Med. 319: 1227-1228, 1988. 13. F~RSTERMANN, U., J. S. POLLOCK, H. H. H. W. SCHMIDT, M. HELLER, AND F. MURAD. Calmodulin-dependent endotheliumderived relaxing-factor/nitric oxide synthase activity is present in the particulate and cytosolic fractions of bovine aortic endothelial cells. Proc. NutZ. Acud. Sci. USA 88: 1788-1792, 1991. 14. FURCHGOTT, R. F., AND J. V. ZAWADZKI. The obligatory role of endothelial cells in the relaxation of arterial smooth muscle by acetylcholine. Nature Lond. 288: 373-376, 1980. 15. HECKER, M., W. C. SESSA, H. J. HARRIS, E. E. ANGG;~RD, AND J. R. VANE. The metabolism of L-arginine and its significance for the biosynthesis of endothelium-derived relaxing factor: cultured endothelial cells recycle L-citrulline to L-arginine. Proc. NutZ. Acud. Sci. USA 87: 8612-8616,199O. 16. IGNARRO, L. J. Biosynthesis and metabolism of endotheliumderived nitric oxide. Annu. Rev. Phurmucol. ToxicoZ. 30: 535-560, 1990. 17. ISHII, K., B. CHANG, J. F. KERWIN, Z. J. HUANG, AND F. MURAD. N-w-nitro-L-arginine: a potent inhibitor of endothelium-derived relaxing factor formation. Eur. J. Phurmucol. 176: 219-223, 1990. R., D. J. STUEHR, AND M. A. MARLETTA. Macrophage 18. IYENGAR, synthesis of nitrite, nitrate, and N-nitrosamines: precursors and role of the respiratory burst. Proc. NutZ. Acud. Sci. USA 84: 63696373,1987. 19. JULOU-SCHAEFFER, G., G. A. GRAY, I. FLEMING, C. SCHOTT, J. R. PARRATT, AND J.-C. STOCLET. Loss of vascular responsiveness induced by endotoxin involves L-arginine pathway. Am. J. Physiol.

OF

NO

FROM

ENDOTHELIAL

CELLS

C641

259 (Heart Circ. Physiol. 28): H1038-H1043, 1990. 20. KILBOURN, R. G., AND P. BELLONI. Endothelial cell production of nitrogen oxides in response to interferon-gamma in combination with tumor necrosis factor, interleukin-1 or endotoxin. J. NutZ. Cancer Inst. 82: 772-776,199O. 21. KILBOURN, R. G., S. S. GROSS, A. JUBRAN, J. ADAMS, 0. W. GRIFFETH, R. LEVI, AND R. F. LODATO. N-methyl-L-arginine inhibits tumor necrosis factor-induced hypotension: implications for the involvement of nitric oxide. Proc. NutZ. Acud. Sci. USA 87: 3629-3632,199O. 22. KWON, N. S., C. F. NATHAN, AND D. J. STUEHR. Reduced biopterin as a cofactor in the generation of nitrogen oxides by murine macrophages. J. BioZ. Chem. 264: 20496-20501,1989. 23. LEFER, A. M., P. TSAO, N. AOKI, AND J. M. A. PALLADINO. Mediation of cardioprotection by transforming growth factor-p. Science Wash. DC 249: 61-64, 1990. 24. MARSDEN, P. A., AND B. J. BALLERMANN. Tumor necrosis factor activates soluble guanylate cyclase in bovine glomerular mesangial cells via an L-arginine-dependent mechanism. J. Exp. Med. 172: 1843-1852,199O. 25. MARSDEN, P. A., T. A. BROCK, AND B. J. BALLERMANN. Glomerular endothelial cells respond to caclium-mobilizing agonists with release of EDRF. Am. J. Physiol. 258 (Renal FZuid Electrolyte Physiol. 27): F1295-F1303, 1990. 26. MARSDEN, P. A., M. S. GOLIGORSKY, AND B. M. BRENNER. Endothelial cell biology in relation to current concepts of vessel wall structure and function. J. Am. Sot. Nephrol. 1: 931-948, 1991. 27. MARTIN, E. R., B. M. BRENNER, AND B. J. BALLERMANN. Heterogeneity of cell surface endothelin receptors. J. BioZ. Chem. 265: 14044-14049,199o. 28. MCKENNA, T., F. MARTIN, B. CHERNOW, AND F. BRIGLIA. Vascular endothelium contributes to decreased aortic contractility in experimental sepsis. Circ. Shock 19: 267-273, 1986. 29. MONCADA, S., R. M. J. PALMER, AND A. E. HIGGS. The discovery of nitric oxide as the endogenous nitrovasodilator. Hypertension Dallas 12: 365-372,1988. 30. MULSCH, A., E. BASSENGE, AND R. BUSSE. Nitric oxide synthesis in endothelial cytosol: evidence for a calcium-dependent and a calcium-independent mechanism. Nuunyn-Schmiedelberg’s Arch. Phurmucol. 340: 767-770, 1989. 31. NATHAN, C. F., AND D. J. STUEHR. Does endothelium-derived nitric oxide have a role in cytokine-induced hypotension? J. NutZ. Cancer Inst. 82: 726-728,199O. 32. PALACIOS, M., R. G. KNOWLES, R. M. J. PALMER, AND S. MONCADA. Nitric oxide from L-arginine stimulates the soluble guanylate cyclase in adrenal glands. Biochem. Biophys. Res. Commun. 165: 802-809,1989. 33. PALMER, R. M. J., A. G. FERRIGE, AND S. MONCADA. Nitric oxide release accounts for the biological activity of endothelium-derived relaxing factor. Nature Lond. 327: 524-526, 1987. 34. RADOMSKI, M. W., R. M. J. PALMER, AND S. MONCADA. An Larginine/nitric oxide pathway present in human platelets regulates aggregation. Proc. NutZ. Acud. Sci. USA 87: 5193-5197, 1990. 35. REES, D. D., R. M. J. PALMER, AND S. MONCADA. Role of endothelium-derived nitric oxide in the regulation of blood pressure. Proc. NutZ. Acud. Sci. USA 86: 3375-3378, 1989. 36. RODRIGUEZ, H. J., AND J. T. YATES. Studies on amino acid incorporation in isolated toad bladder epithelial cells. Seasonal changes in protein synthesis. Biochim. Biophys. Actu 596: 64-80, 1980. 37. STUEHR, D. J., AND C. F. NATHAN. Nitric oxide: a macrophage product responsible for cytostasis and respiratory inhibition in tumor target cells. J. Exp. Med. 169: 1543-1555, 1989. 38. TRACEY, K. J., Y. FONG, D. G. HESSE, K. R. MANOGUE, A. T. LEE, G. C. KUO, S. F. LOWRY, AND A. CERAMI. Anti-cachectin/ TNF monoclonal antibodies prevent septic shock during lethal bacteraemia. Nature Lond. 330: 662-664, 1987. AND A. CERAMI. Cachectin/tumour 39. TRACEY, K. J., H. VLASSARA, necrosis factor. Luncet 2: 1122-1126,1989. 40. VALLANCE, P., J. COLLIER, AND S. MONCADA. Effects of endothelium-derived nitric oxide on peripheral arteriolar tone in man. Luncet 2: 997-1000, 1989. 41. WILMORE, D. Is circulating tumor necrosis factor bioactive? N. EngZ. J. Med. 318: 1227-1228,1988.

Downloaded from www.physiology.org/journal/ajpcell by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on August 23, 2018. Copyright © 1991 American Physiological Society. All rights reserved.