Nitric Oxide Synthesis and Regional Cerebral Blood ... - SAGE Journals

1 downloads 0 Views 421KB Size Report
ticipates in the CBF increases accompanying hyper capnia or ... ter (L-NAME), on the regional CBF (rCBF) re ... tained and a second rCBF determination was made at 10 ..... Faraci FM (1991) Role of endothelium-derived relaxing factor in.
Journal of Cerebral Blood Flow and Metabolism 13:80--87 © 1993 The International Society of Cerebral Blood Flow and Metabolism Published by Raven Press, Ltd., New York

Nitric Oxide Synthesis and Regional Cerebral Blood Flow Responses to Hypercapnia and Hypoxia in the Rat

D. A. Pelligrino, H. M. Koenig, and R. F. Albrecht Department of Anesthesiology, University of Illinois-Chicago, Chicago, Illinois, U.S.A.

(�rCBF in ml 100 g - I min - I ) were 488 versus 57 (CX), 570 versus 60 (SC), 434 versus 72 (BS), and 393 versus 45

Summary: The role of nitric oxide (NO) synthesis in the cerebral hyperemic responses to hypercapnia and hypox­ ia was investigated in anesthetized rats. Regional CBF

(CE), respectively. These differences were all statistically significant (p < 0.05). During hypoxia, when compared to rats not given L-NAME, inhibition of NO synthase activ­ ity resulted in significantly greater (p < 0.05) percentage increases in rCBF (from normoxic baseline values) in most regions. The changes in non-L-NAME- vs. L-NAME-infused rats were 156% versus 262% (CX), 181% versus 309% (SC), and 210% versus 462% (BS), respectively. When the �rCBF values (from normoxic baseline levels) were compared, the changes were greater in the L-NAME group, but the differences were statisti­ cally insignificant. The results of this study indicated that NO synthesis is critically involved in the cerebral hy­ peremic response to hypercapnia but not hypoxia. In fact, the data obtained in the hypoxic groups suggested that reductions in O2 supply may inhibit the NO-generating capacity in the brain. Key Words: Cerebral blood flow­ Endothelium-dependent relaxation-Hypercapnia­ Hypoxia-L-NAME-Nitric oxide-Nitric oxide syn­ thase.

(rCBF) measurements were obtained in the cortex (CX), subcortex (SC), brainstem (BS), and cerebellum (CE) us­ ing radiolabeled microspheres. The rCBF responses to

either hypercapnia (P ac02 = 70-80 mm Hg) or hypoxia (Pa02 = 40--45 mm Hg) were compared in rat groups stud­ ied in the presence and absence of NO synthase inhibition induced via the intravenous infusion of nitro-L-arginine methyl ester (L-NAME, 3 mg kg-1 min- I ). Administra­ tion of L-NAME under normocapnic/normoxic conditions produced a 40-60% reduction in baseline rCBF values, indicating the presence of a NO "tone" in the cerebral vasculature. Infusion of L-NAME resulted in a substan­ tial attenuation, in all regions measured, of the rCBF in­ creases that normally accompany hypercapnia. In com­ paring saline-infused to L-NAME-infused rats, the per­ centage increases in rCBF (from normocapnic baseline values) were 351% versus 166% (CX), 446% versus 199% (SC), 443% versus 206% (BS), and 483% versus 174% (CE), respectively. The rCBF changes from baseline

Vascular tissue, when exposed to certain physi­ ologic stimuli (e,g., activation of specific receptors, shear stress), can exhibit an endothelium­ dependent vasodilation in a wide range of struc­ tures, including the brain, via the synthesis and re­ lease of nitric oxide (NO) or nitrosothiol com­ pounds (Luscher and Vanhoutte, 1990; Myers et

aI., 1990; Moncada et aI., 1991). Furthermore, a continuous NO production appears to occur under resting conditions in many vascular beds, thereby imparting a vasodilatory "tone" (Gardiner et aI., 1990; Faraci, 1991; Persson et aI., 1991; Tanaka et aI., 1991; Pelligrino et aI., 1992). NO is formed from the terminal guanidino nitrogen of L-arginine via the action of NO synthase (NOS) (e.g., Luscher and Vanhoutte, 1990). In the brain, NOS is not only found in vascular endothelium but also in perivas­ cular nerves and astrocytes (Bredt et aI., 1990; Mur­ phy et aI., 1990; Garthwaite, 1991). The constitutive NOS isoform found in the brain is dependent on the presence of calcium, calmodulin, NADPH, and tet­ rahydrobiopterin (Forstermann et aI., 1991; Mayer et aI., 1991). Also, molecular oxygen is an essential substrate in the formation of NO (and citrulline) via NOS (Leone et aI., 1991).

Received February 11, 1992; final revision received May 19, 1992; accepted June 17, 1992. Address correspondence and reprint requests to Dr. D. A. Pelligrino at Dept. of Anesthesiology, Humana Hospital-Michael Reese, 2929 S. Ellis Avenue, Chicago, IL 60616, U.S. A. Abbreviations used: BS, brainstem; CE, cerebellum; CYR, cerebral vascular resistance; CX, cortex; EAA, excitatory amino acid; EDRF, endothelium derived relaxing factor; L-NAME, ni­ tro-L-arginine methyl ester; MANOY A, multivariate analysis of variance; NOS, nitric oxide synthase; PKC, protein kinase C; SC, subcortex.

80

NITRIC OXIDE AND CEREBRAL VASODILATION

It is unclear whether enhanced NO release par­ ticipates in the CBF increases accompanying hyper­ capnia or hypoxia. There is evidence from in vitro experiments of an endothelium-dependent compo­ nent to the cerebral hyperemic response to hypoxia, suggestive of an enhanced NO activity (Pearce et al., 1989, 1990). On the other hand, one might sus­ pect a suppression of NO release during hypoxia, due to the O2 requirement for NO synthesis (Leone et al. , 1991; Rengasamy and Johns, 1991). Under hypercapnic conditions, recent reports indicated that in the presence of NOS inhibition, a blunting of the cortical CBF increase occurred, suggesting that NO release, at least in part, normally contributes to the CBF increase (Iadecola, 1992; Wang et al. , 1992). However, evidence for this effect in other brain regions is lacking. In the present study, we evaluated the effects of NOS inhibition, via intra­ venous administration of nitro-L-arginine methyl es­ ter (L-NAME), on the regional CBF (rCBF) re­ sponses to hypoxia and hypercapnia. We recently demonstrated that intravenous L-NAME could completely block the profound cerebral hyperemic response accompanying administration of a musca­ rinic receptor agonist (Pelligrino et al. , 1992). Acti­ vation of muscarinic receptors is an accepted model for vascular relaxation mediated by endothelium­ derived NO (e.g., Luscher and Vanhoutte, 1990). Thus, the use of the L-NAME in this manner per­ mitted us to evaluate whether NO release is in­ volved in the rCBF increases induced by hypoxia or hypercapnia. METHODS

The study protocol was approved by the Institutional Animal Care and Use Committee. Male Sprague-Dawley rats (350-450 g, n 23) were used. For study, anesthesia was induced with halothane. The rats were then para­ lyzed with curare (1 mg kg-I i.v.), tracheotomized, and connected to a rodent respirator. The anesthesia used during the remaining surgery was 0.7% halothane170% NzO/30% Oz. Catheters were implanted in both femoral arteries and veins. An additional catheter was advanced into the left ventricle, following insertion into the right brachial artery. This catheter, employed in the injection of radiolabeled microspheres, was tapered over the final 2-3 cm from the tip so as to not occlude the right common carotid artery. Following surgery, the halothane was dis­ continued and wound sites were infiltrated with bupiva­ caine. The 70% NzO/30% 0z was maintained and an i.v. infusion of fentanyl was initiated (10 ILg kg-I loading dose) and continued throughout the study (25 ILg kg-I h -I ) . The experiments were started at -1 h post­ halothane. The arterial pressure pulse was recorded con­ tinuously during the study. The rectal temperature was servocontrolled at 3rc. The experimental groups were as follows: hypercapnia (PaCOZ 7�0 mm Hg), with (n 7) and without (n 4) i.v. L-NAME; hypoxia (PaOZ 40-45 mm Hg), with (n =

=

=

=

=

81

5) and without (n 4) i.v. L-NAME; and a 20 min i.v. L-NAME time control (n 3). Hypercapnia was intro­ duced by adding COz to the inspiratory gas mixture (giv­ ing a concentration of 8%). For hypoxia, the inspiratory 0z concentration was reduced to 13% and Nz was added (at 17%). The L-NAME was infused at 50 ILl min-I to provide a dose of 3 mg kg - 1 min-I. The rCBF was mea­ sured via radiolabeled microspheres CS7Co, 113Sn, and 85Sr) in the cortex (CX), subcortex (SC), brainstem (BS), and cerebellum (CE). Details of the IT :lOdology were provided in an earlier publication fn . our laboratory (Pelligrino and Albrecht, 1991). In all experiments, a control rCBF measurement was made at 10 min following the initiation of an i.v. saline infusion (50 ILl min -I). In experiments where i. v. L-NAME was not given, the saline infusion was main­ tained and a second rCBF determination was made at 10 min of hypercapnia or hypoxia. Methoxamine was in­ fused (0.1-0.2 mg kg-I min-I) during hypoxia to prevent the mean arterial pressure (MAP) from falling below 90 mm Hg. The total dose of methoxamine given was similar when comparing experiments with and without L-NAME. Preliminary experiments confirmed that this dose of methoxamine had no influence on rCBF in a MAP range of 90-150 mm Hg. When L-NAME was used, a second rCBF measurement, following control, was obtained at 10 min of L-NAME (under normoxic/normocapnic condi­ tions). Hypercapnia or hypoxia was then introduced, while maintaining the L-NAME infusion, and a third rCBF determination was made 10 min later. In three rats, an L-NAME time control was performed, in which rCBF measurements were made (in normoxic/normocapnic rats) at 10 min of i.v. saline and at 10 and 20 min of i.v. L-NAME. At the end of the experiment, the rats were killed with an overdose of halothane. The brains were then removed and dissected into the right and left CX, right and left SC (contains tissue primarily from the striatum, thalamus, and hippocampus), BS, and CEo Brain tissue samples were then weighed and placed (along with the blood with­ drawn during and following microsphere injections-see Hoffman et aI., 1983; Pelligrino and Albrecht, 1991) in a gamma counter (Nuclear Data, Inc.). In this study, as in previous studies from our laboratory (Pelligrino and Al­ brecht, 1991; Pelligrino et aI., 1992), we found no side-to­ side differences in rCBF in the CX or SC regardless of experimental manipulation. Thus, all rCBF values in these two regions are reported as the average of the two sides. Partial pressures for arterial CO2 and 02' along with arterial pH, were measured using an IL 1303 blood gas/ pH analyzer. Cerebral vascular resistance (CVR) was es­ timated from the MAP divided by the rCBF. Statistical comparisons of rCBF values between experimental groups were made employing a multivariate analysis of variance (MANOVA, Systat). Analyses of results within experimental groups (e.g., comparisons to control or L-NAME alone) were performed using a repeated­ measures two-way ANOVA, along with a post hoc C ma­ trix multiple comparisons analysis (Systat). Statistical significance was taken at the p < 0.05 level. =

=

=

RESULTS

The values for arterial Po2, Pco2, and pH and MAP in the four major experimental groups, obJ Cereb Blood Flow Metab, Vol. 13, No.1, 1993

D. A. PELL/GRIND ET AL.

82

TABLE 1. Arterial blood variables

Group

C

Hypercapnia (n = 4) Hypercapnia + L-NAME (n = 7) Hypoxia (n = 4) Hypoxia + L-NAME (n = 5)

124 ±22

L-NAME

130 ±8 110 ±18

134 ±32

127 ±22

138 ±16

HC/HX

C

L-NAME

136 ±22

37. 8 ±2. 4

142

39. 0 ±2. 1 39. 6 ±1. 0

35. 6 ± L oa

37. 0 ±2. 5

35. 3 ±2. 2

±29 44. 3 ±2. 4a 43. 3 ±5. 1a. b

pH

Pco2 (mm Hg)

P02 (mm Hg)

HC/HX

MAP (mm Hg)

L-NAME

C

77. 6 ±2. 8a

7.41 ±0. 04

74. 8 ±4. 8a. b 39. 1 ±1. 0

7.36 ±0.05 7.38 ±0.02

7.34 ±0.05

7.37 ±0.07

7.34 ±0. 07

35.7 ±2. 7

HC/HX

C

L-NAME

7.14 ±0.02a

137 ±22

7.02 ±0.05a 7.36 ±0.04

137 ±11 136

7.16 ±0.09a. b

128 ±12 161 ± 8a

139 ±13b 110 ± loa

157 ± 9a

114 ±13a. b

�'! 134 ±7

HC/HX

All values are means ± SD. C, control; HC, hypercapnia; HX, hypoxia. a p < 0.05 vs. C; bp < 0.05 vs. L-NAME.

tained at the time of each rCBF measurement, are summarized in Table 1. In both hypercapnic groups, as one would expect, inhalation of 8% CO2 produced significant pH reductions from control, but equivalent elevations in Pe02. In the rats given L-NAME, a slightly lower Pe02 value (vs. saline­ infused control) was measured at 10 min of infusion, immediately prior to induction of hypercapnia. At the same time, a significant 24 mm Hg increase in MAP from control was seen. The MAP increase was reversed during hypercapnia. Neither of the hypoxia-exposed groups showed variations in Pe02 at any time point during the experiments. Equal ar­ terial P02 values were measured in both groups fol­ lowing exposure to an inspired O2 level of 13%. During hypoxia, the MAP values in these groups were similar but lower than control by 20-26 mm Hg. Administration of L-NAME prior to hypoxic induction, as observed in the corresponding hyper­ capnic group, produced a significant rise in MAP from control (LlMAP 24 mm Hg). The arterial pH was significantly reduced from control during com­ bined L-NAME / hypoxia. In the absence of L-NAME, no fall in pH during hypoxia was ob­ served.

(CE). Following the onset of hypercapnia, a se­ verely attenuated rCBF response, compared to non-L-NAME-infused rats, was observed. In fact, in only two regions (SC, BS) was the change from the normocapnic/L-NAME-infused level found to be statistically significant (Fig. O. When expressed as a percentage of the normocapn�c/L-NAME flow values (baseline-Table 2), the rCBF values during hypercapnia/L-NAME were in the range of 166206%, and when expressed as LlrCBF from base­ line, the values ranged from 45 to 72 ml 100 g-l min - 1 (Table 2). These rCBF changes were signif­ icantly reduced from the changes that normally oc­ cur in hypercapnia (see Table 2 and above). Re­ gional CVR values are presented in Fig. 2. As one

� �

I

's

=

Hypercapnia and cerebral hemodynamics

In the absence of L-NAME, hypercapnia resulted in substantial increases in rCBF, from normocapnic levels, in all regions (Fig. O. As a percentage of the normocapnic (i.e. , baseline) values (Table 2), these increases ranged from 351 to 483% in the four re­ gions analyzed. The range of rCBF changes (LlrCBF) was 393-570 ml 100 g - 1 min - 1 (Table 2). Infusion of L-NAME, prior to induction of hyper­ capnia, resulted in significant reductions in rCBF, from saline-infused control levels, in all regions (Fig. O. As a percentage of control, rCBF de­ creased to 62% (CX), 56% (SC), 55% (BS), and 53% J Cereb Blood Flow Metab. Vol. 13, No.1, 1993

.::

I

bl) 0 0

S ....., ""

iJ:1 U I-.

BOO

Control

..

.

700

m

L-NAME

600

c=J

Hypercapnia

.

.

500 400 300 200 100 0

.

*

#

.



.

SC

BS

CE

��W�

CX

L-NAIlE

+ Hypercapnia

CX

SC

BS

Hypercapnia

I

CE alone

FIG. 1. Regional cerebral blood flow (rCBF) values in the cortex (CX), subcortex (SC), brainstem (BS), and cerebellum (CE) during hypercapnia, with (left) and without (right) intra­ venous infusion of the NO synthase inhibitor, L-NAME. All rCBF values during hypercapnia + L-NAME were signifi­ cantly lower than those measured during hypercapnia alone. Additional information is provided in the text. All values are means ± SD. *p < 0.05 versus control (normocapnic saline infusion); #p < 0.05 versus L-NAME (normocapnic).

83

NITRIC OXIDE AND CEREBRAL VASODILATION

TABLE 2. reEF values expressed as percentage of baselinea and as change (A) from baseline flows Region Cerebellum

Brainstem

Subcortex

Cortex Group

%



%



%



%



Hypercapnia Hypercapnia + L-NAME Hypoxia Hypoxia + L-NAME

351 ± 51

488 ± 58

446 ±101

570 ±49

443 ± 122

434 ± 152

483 ± 169

393 ± 229

166 ± 85b 156 ± 12

57 ± 94b 109 ± 20

199 ±87 b 181 ±16

60 ±64b 136 ± 44

206 ± 96b 210 ± 31

72 ± 67b 178 ± 42

174 ± 75b 188 ± 44

45 ± 53b 145 ± 57

262 ± 49c

157 ± 65

309 ± 5Y

155 ± 41

462 ± 208c

180 ±67

2S

:!:

93

176 ± 68

All values are means ± SD. Units for MCBF are ml 100 g - I min - I. a Baseline flows were obtained during saline (hypercapnia or hypoxia alone groups) or L-NAME infusions under normocapnicl normoxic conditions (see the text). b p < 0.05 vs. hypercapnia; 'jJ < 0. 05 vs. hypoxia.

would expect from the flow data, hypercapnia alone resulted in significant reductions in CVR in the dif­ ferent regions. With respect to control, infusion of L-NAME produced approximately twofold in­ creases in regional CVR values. When hypercapnia was then introduced, regional CVR values fell to levels not significantly different from control but, with the exception of the CX, significantly lower than values obtained during normocapnic L-NAME infusions. During hypercapnia, the regional CVR values were significantly (four- to fivefold) greater in rats infused with L-NAME compared to rats not given L-NAME. Hypoxia and cerebral hemodynamics

When L-NAME was not infused, hypoxia re­ sulted in significant rCBF increases, from normoxic control, in all of the regions analyzed (Fig. 3). These changes, expressed as a percentage of control, ranged from 156 to 210% (Table 2). The �rCBF val­ ues were in the range of 109 to 178 mll00 g-I min-I (Table 2). Infusion of L-NAME under normoxic conditions produced significant reductions in rCBF



s::

's

OG

g

...

I

5.0

4.0

.. L-NAME o Hypercapnia

3.5 3.0

E

1.5

OIl :I:

! p::

Ei

I

_ Conlrol

S

OIl a

6

1.0

0.5

CX

SC

L-NAME +

BS

CE

Hypercapnia

CX

SC

BS

CE

Hypercapnia alone

FIG. 2. Regional cerebrovascular resistance (CVR) values calculated for the groups represented in Fig. 1. Values are means ±SO. 'p < 0.05 versus control; #p < 0.05 versus the hypercapnic CVR value in the L-NAME-treated group.

Control

L-NAME

Hypoxia *

200

100

O�

L-JlLIIOU--"'OiU..'-.... >W..

_ !§Il§II D

400

300 I



2.0

0.0

=

_�

4.5

2.5

from control in all regions, similar to the changes represented in Fig. 1 (hypercapnia experiments). Thus, rCBF decreased to 53% in the CX, 50% in the SC, 42% in the BS, and 49% in the CEo When hyp­ oxia was introduced, in spite of the lower baseline (i.e., normoxic L-NAME infusion) rCBF levels, the flows in all regions increased to values not statisti­ cally different from those measured during hypoxia without added L-NAME. When expressed as a per­ centage of the baseline values (Table 2), the rCBF reached levels of 262-462%. With the exception of the CE, these values were significantly higher than those seen during hypoxia in the absence of L-NAME (see Table 2 and above). On the other hand, expressing the hypoxic flows as a change from baseline (6.rCBF) produced values (range 150-180 ml 100 g-I min-I) during L-NAME infu­ sions that were larger but not significantly different from those measured in the absence of L-NAME (see Table 2 and above). Regional CVR values are presented in Fig. 4. Hypoxia alone reduced CVR by �50% in the regions studied. Infusion of L-NAME



CX

L..



SC

BS

CE

L-NAME + Hypoxia

CX

SC

BS

CE

Hypoxia alone

(left) and without (right) intravenous L-NAME. Values are means ±SO. See the text for more information. *p < 0.05 versus control (normoxic saline infusion); #p < 0.05 versus normoxic L-NAME infusion. FIG. 3. Regional CBF values during hypoxia with

J Cereb Blood Flow Metab, Vol.13, No.1, 1993

D. A. PELL/GRINO ET AL.

84

produced a greater than or equal to twofold increase in regional CVR values. With hypoxia, these values fell by 70-80% to levels significantly lower than c?ntrol in the SC, BS, and CE, but not significantly different from values obtained in the hypoxia alone group. Investigations of the effects of L-NAME on rCBF responses to hypercapnia or hypoxia used a 20 min L-NAME infusion (i.e. , 10 min of L-NAME alone followed by 10 min of L-NAME plus hypercapnia or hypoxia). It was, therefore, of some importance to ?etermine the influence on rCBF of L-NAME by itself over the time interval of 10-20 min. Figure 5 shows the rCBF values measured prior to introduc­ tion of L-NAME (10 min saline infusion) and at 10 and 20 min of an i. v. infusion of L-NAME. Consis­ tent with the data summarized in Figs. 1 and 3, rCBF values in the four regions were reduced to 50-60% of control at 10 min. When going from 10 to 20 min, a slight further drop in rCBF (to 45-55% of control) was seen in three of the regions (CX, SC, and BS), but these changes were not statistically signi�cant. In these animals, arterial P02 and Peo2 remamed constant and within normal limits throughout the experiments and MAP was in­ creased by 22 ± 5 mm Hg at 10 min and 19 ± 4 mm Hg at 20 min following L-NAME administration from an initial (saline infusion) value of 132 ± 7 mm Hg. DISCUSSION

The results of this study indicate that NO synthe­ sis is critical to the cerebral vasodilatory response to hypercapnia. On the other hand, NO synthesis �n� release does not appear to contribute to hypox­ ta-mduced rCBF increases. In fact, inhibition of NOS prior to hypoxic induction may actually en-

's

oil

0 0 -

i

S

'"' :r:

S S

� ;> u

5.0

_

4.5

� D

4.0 3.5

Control

L-NAME

Hypoxia

3.0

2.5

2.0

1.5

1.0

0.5

0.0

cx

SC

BS

CE

L-NAME + Hypoxia

0

·s

I

CX

SC

BS

CE

Hypoxia alone

FIG. 4. Regional CVR values calculated for the groups rep­ resented in Fig. 3. Values are means ::t SO. *p < 0.05 versus control.

J Cereb Blood Flow Metab, Vol.13, No.1, 1993

Cortex

..

Brainslem

0

200

'" Cerebellum

150

-

100

r... (:Q U ....



250

tlD 0 0

S .......

L-NAME and rCBF over time

'2



I

Subcortex

50 0

o

10

Time (min)

20

FIG. 5. .Regional CBF values measured, in normocapnic/ normoxlc rats, during saline infusion (rCBF values at 0 min) and at 10 and 20 min of intravenous L-NAME. Values are means::t SO.

hance the hyperemic response, although the data are somewhat equivocal in this regard. The question of which mechanisms are responsi­ ble for hypercapnia- or hypoxia-induced increases in CBF was addressed in many reports, too numer­ ous to cite here (e.g. , see review by Busija and Heistad, 1984). In spite of such wide attention in the literature, there remains a large measure of uncer­ tainty with regard to factors contributing to the va­ sodilatory process and factors that may modify or attenuate this process. Much of this diversity in ex­ perimental findings may be attributable to differ­ ences related to animal species, methodology (e.g. , in vivo vs. in vitro preparations), or animal age (i. e. , adult vs. neonate). In the rat, vasodilator pros­ tanoid production reportedly plays a large role in the hypercapnic response but no role in the hypoxic response (Sakabe and Siesj6, 1979). During hypercapnia, there appears to be little question that a decrease in brain extracellular pH is critical to the cerebral vasorelaxation response (e. g. , Lassen, 1968; Kontos et ai. , 1977). Thus, any additional mechanisms one might propose (e. g. , prostanoid release and, in the present case, NO re­ lease) as participating in the cerebral hyperemic re­ sponse to hypercapnia must in some way require the presence of an increased H+ ion concentration. The simplest explanation would be that increases in [H+] enhance NOS activity. In vitro activity of brain NOS was shown to be positively correlated with [H+], but only at pH values> 7. 0 (Heinzel et ai., 1992). However, with relevance to hypercapnic states, no changes in NOS activity were seen in the 6.5-7.0 pH range. Perhaps increased H+ levels af­ 2 fect intracellular Ca + availability. Calcium is re­ quired for activation of the constitutive NOS iso­ form (Forstermann et ai., 1991) and in the activation of the enzyme initiating the prostanoid synthesis cascade, phospholipase A2 (see Siesj6, 1981). Inter-

NITRIC OXIDE AND CEREBRAL VASODILATION

estingly, Cortey et ai. (1991), in a preliminary study, reported that inhibition of protein kinase C (PKC) blocked the relaxation of pial arterioles induced by hypercapnia. These results, coupled with the fact 2 that the principal isoforms of PKC are Ca + depen­ dent (Nishizuka, 1988), lend additional support to 2 the idea that altered Ca + availability may play a role in hypercapnic vasodilation. The present results of an attenuated CBF re­ sponse to hypercapnia following NOS inhibition confirms the findings of others (Iadecola, 1992; Wang et aI. , 1992). These authors employed CBF measurement techniques in rats (laser-Doppler flowmetry or intracarotid xenon-133) that assess primarily cortical blood flow. Our findings demon­ strated that enhanced NO production is not only a component of the hypercapnic CBF response in the cortex, but also plays a major role in the hyperemic responses of subcortical and hindbrain structures as well. We recently reported a normal CBF response to hypercapnia in chronically hyperglycemic diabetic rats (Pelligrino and Albrecht, 1991), a putative model of cerebral vascular endothelial damage (Moore et aI., 1985) to the extent that endothelium­ and receptor-dependent (e.g. muscarinic, puriner­ gic P2) vasodilatory responses are profoundly im­ paired (Mayhan, 1989; Mayhan et at. , 1991; Pelli­ grino et aI., 1992). These findings may suggest that hypercapnia induces NO synthesis in nonendothe­ lial structures and/or via a process different from the receptor-dependent mechanisms through which agonists like acetylcholine and ADP act. The source of the L-NAME-sensitive NO forma­ tion in hypercapnia is not likely to be found in vas­ cular smooth muscle cells in that the intracellular 2 Ca + concentration needed to activate NOS would be inhibitory to guanylyl cyclase (Knowles et aI. , 1989). This would block the increase in cyclic GMP essential to the vasorelaxation mechanism. Perivas­ cular astrocytes and neurons remain as potential sources of NO during hypercapnia (see the Intro­ duction), although the paucity of relevant informa­ tion in the literature renders any further discussion far too speCUlative. Vasodilator nerves to vascular smooth muscle do, however, represent a viable op­ tion. Toda and Okamura (1990, 1991) reported that the vasorelaxant response to transmural electrical stimulation in canine cerebral arterial strips, both endothelium intact and denuded, could be blocked in the presence of an NOS inhibitor, suggesting that nonendothelial NO plays a role in transmitting the vasodilatory signal. Interestingly, in an earlier study from the same laboratory, employing the same experimental preparation, it was found that

85

hypercapnia-induced relaxation was unaffected by , endothelium removal (Toda et aI. , 1989). In addition to indicating that the endothelium does not partici­ pate in the hypercapnic response, in support of our suggestion above, these authors also concluded that hypercapnia acts directly on the smooth muscle. In light of subsequent reports (Toda and Okamura, 1990, 1991), the possibility exists that hypercapnia may act by increasing NO production in the nona­ drenergic, noncholinergic vasodilator nerves asso­ ciated with this preparation, rather than via a direct action on smooth muscle cells. Recent findings sug­ gest that these nerves may arise from the spheno­ palatine but not the trigeminal ganglion (Nozaki et aI. , 1992; see also Sakas et aI. , 1989). Present results showed that NOS inhibition did not diminish but rather may have augmented the rCBF increases associated with hypoxia. However, the conclusion of a heightened vasodilatory re­ sponse was dependent on the manner in which this was expressed (see the Results section). Neverthe­ less, these findings at a minimum speak against any role for NO/endothelium-dependent relaxation fac­ tor (EDRF) release in mediating hypoxia-induced cerebral vasodilation, as some investigators have suggested (Pearce et aI. , 1989, 1990; Kozniewska et aI. , 1992). In the study by Kozniewska et aI. (1992), the absolute hypoxic CBF value was significantly lower in the group of rats treated with an NOS in­ hibitor, but the normoxic CBF was also reduced by NOS inhibition. This had the net effect that the rel­ ative hypoxia-induced increase in CBF was nearly the same in treated compared to untreated rats (2vs. 4. 3-fold). Thus, any conclusion of a role for NO production in mediating the CBF response to hyp­ oxia would be questionable. The studies by Pearce et at. (1989, 1990) were conducted in vitro using large cerebral arteries (basilar and internal carotid) harvested from rabbits. Hypoxic vasodilation was attenuated following endothelial denudation or gua­ nylyl cyclase inhibition via methylene blue. Yet, when methylene blue was administered in vivo, the hypoxia-induced increase in CBF was unaffected (Pearce et aI., 1990). One might speculate that the failure to attenuate hypoxic CBF increases in vivo by methylene blue, and by NOS inhibition (present report), is related to an inability of hypoxia to induce NO release in the vicinity of smaller cerebral vessels. This is unlikely in that hypoxia is capable of inducing a profound vasodilatory response in brain microvessels (Morii et at. , 1987), and there is evidence of an active va­ sodilatory mechanism in these small vessels that is sensitive to NOS inhibitors and methylene blue (Kontos et at., 1988; Marshall et aI., 1988; Busija et J Cereb Blood Flow Metab, Vol.13, No.1, 1993

D. A. FELL/GRING ET AL.

86

al., 1990; Rosenblum et al. , 1990; Faraci, 1991; Ha­ berl et al., 1991). The possibility also exists that the factors mediating hypoxia-induced vasodilation in the brain are not the same in rabbits and rats. Fi­ nally, previous studies, employing adenosine recep­ tor antagonists (Hoffman et al. , 1984; Morii et al. , 1987), demonstrated that the CBF increasel vasodilation accompanying moderate hypoxia (i.e. , P02 40-50 mm Hg) in rats was almost entirely mediated by adenosine release (see also Winn et al. , 1981). Thus, in moderate hypoxia in the rat, there appears to be little or no room remaining for an NO/EDRF component. With increasing hypoxic se­ verity, the relative role of adenosine decreases (Hoffman et al. , 1984), but other factors such as increased tissue H+, extracellular K+, or excitato­ ry amino acid (EAA) release could make up the difference (Busija and Heistad, 1984; Phillis and Walter, 1989; Hamilton et al. , 1992). The last may be of some relevance to the present discussion in that EAAs can promote neuronal NO production and release (Garthwaite, 1991) (see below). The explanation for an enhanced CBF response following L-NAME administration may be related to the need for an adequate oxygen supply to sus­ tain NOS-mediated NO production (see Leone et al., 1991). In support of this, Rengasamy and Johns (1991) reported that lowering the O2 tension mark­ edly reduced NOS activity in brain homogenates, an effect that was readily reversed by normoxia. If one combines this O2 requirement for NOS activa­ tion with the presence of a NO-related vasodilatory "tone" in the brain, as the present and other studies have shown (Faraci, 1991; Tanaka et al. , 1991; Pel­ ligrino et al., 1992), a hypoxia-induced attenuation of NO synthesis may act (in the absence of inhibi­ tors) to limit the hypoxia-associated cerebral hy­ peremic response. Then, if one inhibits NOS prior to hypoxia, thereby removing this "brake" on the CBF increase, one might see an augmented CBF response. We might add that this process, in rats, may be localized to large vessels, in that larger ar­ teries, rather than microvessels, appear to be the site of basal NO/EDRF production (Faraci, 1991). Consistent with this reasoning and in support of our findings, an enhancement of hypoxia-induced va­ sorelaxation in the face of EDRF antagonism was observed in sheep middle cerebral artery rings (Klaas and Wadsworth, 1989). Nevertheless, there are experimental findings that appear to be incon­ sistent with the concept of hypoxia producing a di­ rect limitation on NO availability. In the study by Kozniewska et al. (1992), the percentage increase from baseline CBF during hypoxia was similar in treated relative to untreated rats, whereas we found =

J Cereb Blood Flow Metab, Vol.13, No.1, 1993

much greater relative increases in treated animals. This occurred despite the fact that the hypoxia was more severe in the Kozniewska et al. study (Pa02 33 mm Hg). One explanation for this apparent in­ consistency can be derived from the observation that hypoxia can prolong the half-life of nitrosothi­ ols (Mulsch et al. , 1991). Thus, the actual amount of NO/nitrosothiol available during hypoxia may be a function of two opposing processes-diminished synthesis but reduced inactivation, with the latter perhaps having a greater relative influence at a Pao2 of 33 mm Hg than at the current study level of -44 mm Hg. Another possibility is that more severe hypoxic challenges could lead to release of EAAs (see above). Therefore, one might speculate that NOS inhibition in more severe hypoxia may remove EAA-stimulated perivascular neuronal NO produc­ tion and also oppose the effects of diminished syn­ thesis, These issues are not going to be resolved here and need to be addressed in future investiga­ tions. In summary, our study demonstrated that hyper­ capnia in rats induces an increase in CBF via a mechanism that is critically dependent on NO syn­ thesis. On the other hand, hypoxia increases CBF through a process that does not require NO. In fact, results suggested, although not conclusively, that hypoxia may suppress cerebral NO production. =

Acknowledgment: The authors thank Anthony Sharp for expert technical assistance. This work was supported by Michael Reese MRIC.

REFERENCES Bredt OS, Hwang PM, Snyder SH (1990) Localization of nitric oxide synthase indicating a neural role for nitric oxide. Na­

ture (Lon d) 347:768-780 Busija OW, Heistad 00 (1984) Factors involved in the physio­ logical regulation of the cerebral circulation. Rev Physiol

Biochem Pharmacol 101:162-211 Busija OW, Leffler CW, Wagerle LC (1990) Mono-L-arginine­ containing compounds dilate piglet pial arterioles via an en­ dothelium-derived relaxing factor-like substance. eirc Res

67:1374-1380 Cortey A, Kim SJ, Wagerle LC (1991) Effect of protein kinase C inhibitor, H-7, on cerebral microvascular responses to CO2 in newborn pigs [Abstract]. FASEB J 5:A374 Faraci FM (1991) Role of endothelium-derived relaxing factor in cerebral circulation: Large arteries vs microcirculation. Am

J PhysioI261:HI038-HI042 Forstermann U, Schmidt HHHW, Pollock JS, Sheng H, Mitchell JA, Warner TO, Nakane M, Murad F (1991) Isoforms of nitric oxide synthase. Characterization and purification from different cell types. Biochem Pharmacol 42:1849-1857 Gardiner SM, Compton AM, Bennett T, Palmer RM, Moncada S (1990) Control of regional blood flow by endothelium­ derived nitric oxide. Hyperten sion 15:486-492 Garthwaite J (1991) Glutamate, nitric oxide and cell-cell signal­ ling in the nervous system. Trends Neurosci 14:60-67 Haberl RL, Oecker PJ, Einhaupl KM (1991) Angiotensin degra­ dation products mediate endothelium-dependent dilation of rabbit brain arterioles. eirc Res 68:1621-1627

NITRIC OXIDE AND CEREBRAL VASODILATION

Hamilton K, Bunegin L, Albin MS (1992) The effect of ketamine hydrochloride (KH) on cerebral blood flow (CBF) during acute hypoxia in the rat [Abstract]. Anesth Analg 74:S128 Heinzel B, John M, Klatt P, Bohme E, Mayer B (1992) Ca2+/ calmodulin-dependent formation of hydrogen peroxide by brain nitric oxide synthase. Biochem J 281:627'-{'30 Hoffman WE, Miletich DJ, Albrecht RF, Anderson S (1983) Regional cerebral blood flow measurement in rats with radio­ active microspheres. Life Sci 33:1075-1080 Hoffman WE, Albrecht RF, Miletich DJ (1984) The role of aden­ osine in CBF increases during hypoxia in young vs aged rats.

Stroke 15:124-129 Iadecola C (1992) Does nitric oxide mediate the increases in cerebral blood flow elicited by hypercapnia? Proc Natl Acad Sci USA 89:3913-3916 Klaas M, Wadsworth R (1989) Contraction followed by relax­ ation in response to hypoxia in the sheep isolated middle cerebral artery. Eur J Pharmacol 168:187-192 Knowles RG, Palacios M, Palmer RM, Moncada S (1989) For­ mation of nitric oxide from L-arginine in the central nervous system: A transduction mechanism for stimulation of the soluble guanylate cyclase. Proc Natl Acad Sci USA 86:

5159-5162 Kontos HA, Raper AJ, Patterson JL (1977) Analysis of vasoac­ tivity of local pH, PC02 and bicarbonate on pial vessels.

Stroke 8:358-360

Kontos HA, Wei EP, Marshall JJ (1988) In vivo bioassay of endothelium-derived relaxing factor. Am J Physiol 255:

H1259-H1262 Kozniewska E, Oseka M, Stys T (1992) Effects of endothelium­ derived nitric oxide on cerebral circulation during normoxia and hypoxia in the rat. J Cereb Blood Flow Metab 12:311-

317 Lassen NA (1968) Brain extracellular pH: The main factor con­ trolling cerebral blood flow. Scand J Clin Lab Invest 22:247-

251 Leone AM, Palmer RMJ, Knowles RG, Francis PL, Ashton DS, Moncada S (1991) Constitutive and inducible nitric oxide synthases incorporate molecular oxygen into both nitric ox­ ide and citrulline. J Bioi Chern 266:23790-23795 Luscher TF, Vanhoutte PM (1990) The Endothelium: Modulator of Cardiovascular Function. Boca Raton, CRC Press Marshall 11, Wei EP, Kontos HA (1988) Independent blockade of cerebral vasodilation from acetylcholine and nitric oxide.

Am J Physiol 255:H847-H854 Mayer B, John M, Heinzel B, Werner ER, Wachter H, Schultz G, Bohme E (1991) Brain nitric oxide synthase is a biopterin­ and flavin-containing multi-functional oxido-reductase.

FEBS Lett 288:187-191 Mayhan WG (1989) Impairment of endothelium-dependent dila­ tion of cerebral arterioles during diabetes mellitus. Am J

PhysioI256:H621-H625 Mayhan WG, Simmons LK, Sharpe GM (1991) Mechanism of impaired responses of cerebral arterioles during diabetes mellitus. Am J Physiol 260:H319-H326 Moncada S, Palmer RM, Higgs EA (1991) Nitric oxide: Physiol­ ogy, pathophysiology, and pharmacology. Pharmacol Rev

43:109--142 Moore SA, Bohlen HG, Miller BG, Evan AP (1985) Cellular and vessel wall morphology of cerebral cortex arterioles after short-term diabetes in adult rats. Blood Vessels 22:265-277 Morii S, Ngai AC, Ko KR, Winn HR (1987) Role of adenosine in regulation of cerebral blood flow: Effects of theophylline during normoxia and hypoxia. Am J PhysioI253:HI65-HI75

87

(1990) Vasorelaxant properties of the endothelium-derived relaxing factor more closely resemble S-nitrosocysteine than nitric oxide. Nature (Lond) 345:161-163 Nishizuka Y (1988) The molecular heterogeneity of protein ki­ nase C and its implications for cellular regulation. Nature

(Lond) 334:661'-{'65 Nozaki K, Moskowitz MA, Maynard KI, Koketsu N, Dawson TM, Bredt DS, Snyder SH (1992) Origin and distribution of nitric oxide synthase-immunoreactive nerve fibers in rat ce­ rebral arteries [Abstract]. Stroke 23:154 Pearce WJ, Ashwal S, Cuevas J (1989) Direct effects of graded hypoxia on intact and denuded rabbit cranial arteries. Am J

PhysioI257:H824-H833 Pearce WJ, Reynier Rebuffel AM, Lee J, Aubineau P, Ignarro L, Seylaz J (1990) Effects of methylene blue on hypoxic cere­ bral vasodilatation in the rabbit. J Pharmacol Exp Ther 254:

616-t>25 Pelligrino DA, Albrecht RF (1991) Chronic hyperglycemic dia­ betes in the rat is associated with a selective impairment of cerebral vasodilatory responses. J Cereb Blood Flow Metab

II : 667'-{'77

Pelligrino DA, Miletich DJ, Albrecht RF (1992) Diminished mus­ carinic receptor-mediated cerebral blood flow response in the streptozotocin-treated rat. Am J PhysioI262:E447-E454 Persson MG, Wiklund NP, Hedqvist P, Gustafsson LE (1991) Microvascular effects of local or systemic inhibition of en­ dogenous endothelium-derived relaxing factor (nitric oxide) production. J Cardiovasc Pharmacol 17 (Suppl 3) :

SI69-SI72 Phillis JW, Walter GA (1989) Hypoxia/hypotension evoked re­ lease of glutamate and aspartate from the rat cerebral cortex.

Neurosci Lett 106:147-151 Rengasamy A, Johns RA (1991) Characterization of endothe­ lium-derived relaxing factor/nitric oxide synthase from bo­ vine cerebellum and mechanism of modulation by high and low oxygen tensions. J Pharmacol Exp Ther 259:310-316 Rosenblum WI, Nishimura H, Nelson GH (1990) Endothelium­ dependent L-Arg- and L-NMMA-sensitive mechanisms reg­ ulate tone of brain microvessels. Am J Physiol 259: HI 396--H140I Sakabe T, Siesjo BK (1979) The effect of indomethacin on the blood flow-metabolism couple in the brain under normal, hypercapnic and hypoxic conditions. Acta Physiol Scand

107:283-284 Sakas DE, Ogilvy flow in during

Moskowitz MA, Wei EP, Kontos HA, Kano M, CS (1989) Trigeminovascular fibers increase blood cortical gray matter by axon reflex-like mechanisms acute severe hypertension or seizures. Proc Natl

Acad Sci USA 86:1401-1405 Siesj6 BK (1981) Cell damage in the brain: A speculative syn­ thesis. J Cereb Blood Flow Metab 1:155-185 Tanaka K, Gotoh F, Gomi S, Takashima S, Mihara B, Shirai T, Nogawa S, Nagata E (1991) Inhibition of nitric oxide syn­ thesis induces a significant reduction in local cerebral blood flow in the rat. Neurosci Lett 127:129-132 Toda N, Hatano Y, Mori K (1989) Mechanisms underlying re­ sponse to hypercapnia and bicarbonate of isolated dog cere­ bral arteries. Am J PhysioI257:HI41-HI46 Toda N, Okamura T (1990) Possible role of nitric oxide in trans­ mitting information from vasodilator nerve to cerebroarterial muscle. Biochem Biophys Res Commun 170:308-313

Mulsch A, Mordvintcev P, Vanin AF, Busse R (1991) The potent vasodilating and guanylyl cyclase activating dinitrosyl­

Toda N, Okamura T (1991) Suppression by NG-monomethyl-L­ arginine of cerebroarterial responses to nonadrenergic, non­ cholinergic vasodilator nerve stimulation. J Cardiovasc Pharmacol 17 (suppl 3):S234-S237

iron(II) complex is stored in a protein-bound form in vascu­ lar tissue and is released by thiols. FEBS Lett 294:252-256 Murphy S, Minor RL Jr, Welk G, Harrison DG (1990) Evidence for an astrocyte-derived vasorelaxing factor with properties similar to nitric oxide. J Neurochem 55:349-351 Myers PR, Minor RL Jr, Guerra R Jr, Bates IN, Harrison DG

Winn HR, Rubio R, Berne RM (1981) Brain adenosine concen­ tration during hypoxia in rats. Am J Physiol 241:H235-H242

Wang Q, Paulsen OB, Lassen NA (1992) Effect of nitric oxide blockade by �-nitro-L-arginine on cerebral blood flow re­ sponse to changes in carbon dioxide tension. J Cereb Blood

Flow Metab 12:947-953

J Cereb Blood Flow Metab. Vol.13, No.1, 1993