Generalized cerebral vasoconstriction induced by intracarotid infusion of angiotensin II in the rabbit

Brain Research, 269 (1983) 91-101 Elsevier

91

Generalized Cerebral Vasoconstriction induced by Intracarotid Infusion of A n g i o t e n s i n II in t h e R a b b i t

A N N E - M A R I E R E Y N I E R - R E B U F F E L , ELISABETH PINARD, PIERRE-FRI~DI~RIC AUBINEAU, PHILIPPE MERIC and J A C Q U E S SEYLAZ

Laboratoire de Physiologie et Physiopathologie Cbrbbrovasculaire, U. 182 LN.S.E.R.M., E.R.A. 361 C.N.R.S., Universit~ Paris VII, 10, avenue de Verdun, 75010Paris (France) (Accepted November 16th, 1982)

Key words: angiotensin II - cerebral blood flow - rabbit - [14C]ethanol tissue sampling technique heat clearance technique - m a s s spectrometry technique - central effect

This study investigated the influence of angiotensin II, perfused into one common carotid artery at a dose of 0.065/tg/kg/min, on the cerebrovascular resistance of the anesthetized rabbit by means of complementary in vivo methods. Heat clearance and mass spectrometry measurements indicated that in the homolateral caudate nucleus angiotensin induced a significant decrease in local blood flow (18.2 +__9%), a fall in pO2 (14.2 ___ 5.3%) and no significant change in pCO2. The [lnc]ethanol tissue sampling technique revealed a significant decrease in flow in all 10 structures sampled in the brain. This decrease was similar in magnitude in both the ipsilateral and the contralateral hemisphere with regard to the site of injection. When expressed in terms of cerebrovascular resistance (CVR) and allowing for a slight increase in blood pressure ( ( 1 0 % ) , these results show that angiotensin II infusion induced an increase in CVR of 18-32%. We conclude that: (1) A unilateral intracarotid infusion of a low dose of angiotensin II induces an increased vascular tone in all cerebral structures. (2) This action, being bilateral, cannot readily be explained by a direct action of angiotensin II on the cerebral vessels in view of the very low recirculating concentration of angiotensin II ( ( 10-9 M). The hypothesis of a cerebral vasomotor influence ofangiotensin II by action on a central structure is discussed. INTRODUCTION

The octapeptide angiotensin II (AII) has been frequently used as a vasopressor agent to study the regulation of cerebral blood flow (CBF). This utilization has been justified during the last decade by the assumption that All does not affect cerebral circulation when administered systemicallyS,8,35:7. More recently, this assumption was challenged by the work of Hardebo et al) 9 who found that systemically administered AII may decrease cerebral venous outflow. Moreover, Gavras and Liang ~6 have demonstrated that an increase in arterial concentration of endogenous All increases the vascular resistance preferentially in some organs, including 'whole brain'. Nevertheless, the use of All in some very recent studies on CBF regulation2,26,32 shows that 0006-8993/83/$03.00 © 1983 Elsevier Science Publishers B.V.

this action of AII on cerebral vasculature still remains controversial. This controversy may stem from two main arguments against a vasoconstrictor action of AII on cerebral vessels: first, the presence and the role of All receptors in cerebral vessels has not yet been clearly demonstrated; second, even if these receptors are present and functional, the existence of the bloodbrain barrier (BBB) may prevent any direct action of AII on cerebrovascular smooth muscle. Concerning the first point, some authors have pointed out a direct vasoconstrictive action of All on pial arteries in vitrog,3L On the contrary, Toda et al.48 have recently demonstrated that All can relax a previously contracted basilar or middle cerebral artery by stimulating All receptors which are associated with a release of 12 profftaglandins. This direct relaxant action of All has also been demonstrated by Webb 5° on

92 other vessels in many species. However, Raichle and G r u b b 4° claim that intraventricularly administered All does not change CBF and direct observations of pial arteries in vivo seem to indicate that these vessels do not react directly to circulating AII, although they react to the drug-induced increase in blood pressure 26,5~. Whatever the direct action of All on cerebrovascular receptors might be in vitro, these last results are consistent with the assertion that All does not cross the endothelial BBB and thereby cannot act on vascular smooth m u s c l e 26. However, All can act on other vascular beds by indirect mechanisms. First, an indirect hypertensive effect is mediated by a facilitatory action of All both on the intracerebral adrenergic pathway 4,13,t4,24,52 and on the sympathetic nervous system ~. The latter effect is mediated via actions on ganglionic transmission 31 and the terminal neurotransmitter releasO 2,2L34. Second, circulating AII may act on some central structures involved in homeostasis of blood volume and pressure: the area postrema, the nucleus tractus solitarius and certain other circum-ventricular structures 22,38,46.In these structures, the BBB is lacking or deficient T M and circulating AII can reach parenchymal receptors. Since Hardebo et al. 19 and Gavras et al. 16 have shown that exogenous or endogenous AII can decrease the total CBF, our aim was to determine whether it could be a direct action of AII on cerebral vessels despite the presence of the BBB, or whether this possible action could be mediated through the indirect mechanisms mentioned previously. To this end we have measured bilaterally in both anterior and posterior regions of the brain the changes of blood flow induced by a monolateral intracarotid infusion of AII. MATERIALS AND METHODS

The complementary approaches used to study the action of angiotensin II on CBF are (1) a quantitative, multiregional tracer method and (2) a continuous method for dynamic measurements of the changes occuring during the administration.

Quantitative flow measurements by the [14 C]ethanol tissue sampling method Principle. The basis of this method, an application of the Fick principle, was described by Ekl6f et al. ~° and adapted to the rabbit by Lacombe et al. 27. The tracer, [lnC]ethanol, diffuses into the cerebral tissue at a rate dependent on the local blood flow and the tracer concentration in the arterial blood. Thus, by measuring the final tissue tracer concentration, C~(T), and the arterial tracer concentration, Ca(t), during the measurement period, flow can be calculated in each structure sampled in absolute terms from the equation: C,(T) = Kie-KiT0fr Ca(t)eKi(t) dt where T is the end of the measurement. The tissue:blood partition coefficient, 2~, and K~ are related to flow by the relationship: f~ =XK~. We used a value of 1.14 for X as calculated by Ekl6f et al. 1°. This choice was discussed in a previous study by Lacombe et al. 28. Protocol. CBF measurements were made in 34 rabbits weighing 2.5-2.8 kg. The animals were anesthetized with a short-duration steroid anesthetic (althesin CT 1341, Glaxo, 1 m g / k g / m i n i.v.), paralyzed with gallamine triethiodide (Flaxedil, 10 m g / k g i.v.) and artificially ventilated. Blood pressure (BP) was recorded in a femoral artery and samples of arterial blood were drawn periodically to check pH, paCO2 and paO2- A steady state was established for the measurement period, with paCO2 = 33 ----- 3 mm Hg, pH = 7.386 _ 0.055, paO2 ---- 120 ___ 30 mm Hg and BP = 88 ___ 5 mm Hg (mean _ S.D.). The tracer was injected via a femoral vein at a uniform rate for 35 s, during which arterial blood samples were drawn anaerobically from the second femoral artery. At the end of this measurement period, the animal was decapitated and the head frozen in liquid nitrogen. Dissection of the brain samples was performed at - 1 0 °C to prevent evaporation of the tracer. Since it seemed possible that ihe anesthesia

93 might interfere with the action of angiotensin II (especially central effects, see Introduction), it would have been preferable, in theory, to test the drug on the unanesthetized animal. However, allowing complete dissipation of the anesthesia could lead to interference from immobilization stress29,39. On the basis of pilot experiments, we therefore chose to compare the CBF of the 16 sham-treated animals (perfused with physiological saline) with that of 18 angiotensin-treated animals during the period of anesthesia dissipation, corresponding to 0-20 min after the end of the althesin perfusion. By relating the CBF values obtained to the duration of anesthesia dissipation, we were able to deduce the specific effects ofangiotensin II. Administration of angiotensin II. A saline solution of angiotensin II (Hypertensin, Ciba) was infused unilaterally via a fine needle implanted in the common carotid artery, at a rate of 80/tl/ min corresponding to a dose of 0.065 /~g/kg/ min for 90 s. At this rate of infusion the systemic angiotensin concentration can be estimated at less than 10-9 M. Our aim was to obtain a local concentration sufficient to induce a marked cerebrovascular reaction in the regions irrigated by the internal carotid artery without reaching a systemic concentration high enough to cause a generalized vasoconstriction. Thus any cerebrovascular effect observed could not be attributed to an autoregulatory response to a hypertension induced by systemic angiotensin. In fact, even at the low concentrations used, angiotensin II did induce a slight increase in BP, though in all cases less than 10%. For this reason we have expressed the results in terms of cerebrovascular resistance (BP/local blood flow) rather than flow.

Continuous measurements of local CBF, pO 2 and

pC02 The heat clearance technique for local CBF measurements. This technique is based on the measurement of variations of tissue thermal conductivity, which occur when local blood flow varies. Two thermistor probes are chronically implanted in the tissue, one of which is heated a few tenths of a degree above the tissue tempera-

ture as monitored by the second probe. This temperature increment is maintained constant by a negative feedback system, the changes in the heating current required being directly proportional to the blood flow variations. Further details have been described by Seylaz et al.42. The mass spectrometric technique for local tissue gas partial pressure measurements. Tissue gases are withdrawn continuously under a high vacuum via a chronically implanted cannula and analyzed in a mass spectrometer. The magnitude of the signals obtained is directly proportional to the partial pressure of the gases in the cerebral tissue (N2, CO2 and 02). A technical analysis of the method has been published previously44. Protocol These two techniques were combined in the same animals, so that simultaneous continuous measurements of local blood flow in the caudate nucleus, tissue pO 2 and pCO2 in the same structure, BP, paO2, paCO2 and the electrocorticogram could be made. The animals were prepared as follows. Flow probes and mass spectrometer cannulae were bilaterally implanted in the caudate nuclei of 9 rabbits which were used in an acute experiment 2 weeks later, this delay being allowed for the complete recovery of the cerebral tissue around the probes 7 As in the experiments in which multiregional flow measurements were made, the animals were anesthetized with infused althesin, paralyzed with gallamine triethiodide and artificially ventilated. Blood pressure was measured in one femoral artery, the other being used for continuous measurement of arterial partial pressures of 0 2 and CO2 with a mass spectrometer cannula according to the technique described by Seylaz et al.43. Administration of angiotensin II. Angiotensin II was infused at the same rate and in the same way as in the series of experiments using [14C]ethanol. Several infusions could be made in the same animal with this method, while monitoring all the variables before and during the test. RESULTS

Effects of angiotensin H on cerebrovascular resis-

94

tance: multiregional measurements with the f4 C]ethanol tissue sampling technique As explained in Materials and Methods, we used CVR to allow for variations in BP. Nevertheless, regression line analysis of absolute values of CBF during dissipation of anesthesia are given in Table I. We compared the CBF and CVR of angiotensin-infused animals with that of the sham-infused animals. It was first necessary to express CVR in relation to the dissipation of anesthesia. Fig. 1 shows the values of CVR plotted against the time after the end of the anesthesia for one of the structures, the middle cortex, and similar curves could be plotted for 7 other structures (anterior and posterior cortex, caudate nucleus, hypothalamus, thalamus, colliculi, grisea centralis). Linear regression analysis showed a significant negative correlation between the CVR and the time (0-20 min) after the end of the althesin infusion in these 8 (out of 10) structures examined for both the treated and the sham animals (Table II). In the other two structures, the reticular formation and the cerebellum, there was no significant correlation between CVR and anesthesia dissipation, and these results will be dealt with later. Since a significant correlation was obtained for 8 of the structures, the effects of angiotensin infusion could be analyzed by means of a test on

the linked variance, the anesthesia dissipation time being regarded as a concomitant variable 3°. First of all, the slopes of the 3 regression lines for each structure (sham, ipsi- and contralateral sides) were compared to each other. It was found that they were similar for all structures showing that the effects of anesthesia dissipation could be considered identical with or without angiotensin infusion. Second, this condition being fulfilled, the test of linked variance could be performed, showing that (1) the linear regression lines of the ipsi- and contralateral sides are not significantly different, except for the posterior cortex; and (2) the distance between either of these two regression lines and the regression lines for the sham animals is statistically significant. This shows that the angiotensin II infusion caused a statistically significant increase in CVR in these 8 structures, and that this CVR increase was not significantly different between one hemisphere and the other. The amount of the increase in CVR is given by the distance between the regression lines and the increase can be estimated at any time, since these can be considered similar for any given structure. The CVR increases have all been estimated at zero anesthesia dissipation time and their variances computed (Table II, A, B and C). Obviously, the

TABLE I Regression line analysis of regional cerebral bloodflow versus anesthesia dissipation time (t) Homolateral side and heterolateral side indicate the hemisphere sampled with respect to the side of the angiotensin II injection. Homolateral and heterolateral side n -- 18, sham n = 16. Structures:

Totalcortex Anterior cortex Middle cortex Posterior cortex Caudatenucleus Thalamus Hypothalamus Colliculi Griseacentralis Reticular formation Cerebellum*

Sham

Homolateral side

Regressiqn line equation

r

P

y = 1.639t+49.77 y = 1.592t+48.212 y=l.656t+55.36 y--l.481t+50.49 y=1.23 t+33.96 y~l.095t+36.87 y=0.475t+36.14 y=1.32 t+52.84 y=0.475t+35.24 y=0.465t+36.63 y~0.487t+45.88

0.851 <~0.001 0.735 ~0.01 0.788 (0.001 0.763 ~0.001 0.735 (0.01 0.742 ~0.01 0.632 (0.02 0.706 (0.01 0.303 n.s. 0.499 n.s. 0.425 n.s.

Heterolateral side

Regression line equation

r

y---1.995t+38.245 y = 1.861t+36.20 y~-2.219t+40.97 y~-l.827t+36.75 y=l.179t+29.735 y = I.!12t+32.49 y--0.938t+27.06 y=1.66 t+40.99 y--1.239t+23.55 y--0.698t+30.13 y~0.887t+34.66

0.856 ~0.01 0.850 ~0.001 0.837 ~0.001 0.805 ~0.001 0.843 (0.001 0.790 (0.001 0.770 ~0.001 0.926 (0.001 0.786 ~0.001 0.675 ~0.01 0.875 (0.001

* Only one, median sample was taken for the cerebellum.

P

Regression line equation

r

P

y = 1.995t+41.846 y = 1.529t+41.22 y=2.038t+43.77 y = 2.127t + 39.80 y-0.752t+33.04 y~0.959t+31.91 y=0.852t+29.41 y--l.348t+46.73 y--1.213t+24.47 y---0.548t+31.86

0.803 ~0.001 0.660 ~0.01 0.813 ~0.001 0.836 (0.001 0.629 (0.01 0.668 ~0.01 0.806 (0.001 0.765 (0.001 0.747 ~0.001 0.613 (0.01

95 CVR

Fig. 2 illustrates the changes in CVR induced in all 10 structures as determined by the two types of statistical analysis described. This multiregional approach shows that the angiotensin

mm Hg/ml.1OOg'1.mn -I Middle Cortex

TABLE II

CVR differences (test of linked variance)for all structures except reticularformation and cerebellum (A) Between the ipsilateral hemisphere and the sham animals. (B) Between the contralateral hemisphere and the sham animals. (C) Between the ipsilateral and the contralateral hemispheres. Section (D) shows CVR values in perfused and sham animals for median structures: reticular formation and cerebellum (Student's t-test), n.s.; not significant. *~ •

~

o

.N,

Angiotensin injected side Contralateral side Sham

Structures

t

6

t'o

rain

ab

Fig. 1. Effect of intracarotid angiotensin II infusion on middle cerebral cortex blood flow under various degrees of dissipation of the anesthesia. Linear regression curves of flow versus time after the end of the althesin infusion for both sham-operated animals and angiotensin II-injected animals. Lower regression line: data of sham-operated animal; y = 0.035 t + 1.6, r = 0.76, P ( 0.001. Upper regression lines: data of unilaterally injected animals, giving the following equations with regard to the side of angiotensin II injection: ipsilateral cortex, y = 0.051 t + 2.2, r - 0.81, P ~ 0.001; contralateral cortex, y = 0.047 t + 2.1, r = 0.85, P ( 0.001. Statistical analysis shows, (1) that the slopes of the 3 regression lines are not significantly different, (2) that the regression lines of the ipsilateral and contralateral cortex are not significantly different.

CVR increases would be statistically similar at any other time of anesthesia dissipation. As regards the two structures for which there was no significant correlation between flow and anesthesia dissipation time (reticular formation, cerebellum), we calculated the mean CVR and the variance using the values at all times, and compared the values o f sham animals with those of angiotensin-treated animals by means of Student's t-test. There was a statistically significant difference (P ~ 0.001) of respectively 21% and 32% between the shams and the treated animal for the reticular formation and the cerebellum.

CVR differences ± S.D. (ram Hg/ml. lOOg-l min

l)

P

(A) Ipsilateral hemisphere compared to sham animal hemisphere Anterior cortex Middle cortex Posterior cortex Total cortex Caudate nucleus Hypothalamus Thalamus Grisea centralis Colliculi

0.601 0.548 0.719 0.485 0.342 0.769 0.538 0.678 0.515

_ _ ± _ _ _ ± ± ±

0.076 0.072 0.089 0.065 0.110 0.104 0.097 0.158 0.056

~0.001 ~0.001 ~0.001 ~0.001 ~0.01 <0.001 ~0.001 ~0.001 ~0.001

(B) Contralateral hemisphere compared to sham animal hemisphere Anterior cortex Middle cortex Posterior cortex Total cortex Caudate nucleus Hypothalamus Thalamus Grisea centralis Colliculi

0.435 0.461 0.484 0.648 0.468 0.542 0.635 0.653 0.322

_ 0.089 _ 0.064 _ 0.073 __. 0.069 _ 0.110 __. 0.090 + 0.128 ___0.185 _ 0.069

<0.001 <0.001 < 0.001 <0.001 ~0.001 ~0.001 ~0.001 ~0.01 ~0.001

(C) lpsilateral hemisphere compared to contralateral hemisphere Anterior cortex Middle cortex Posterior cortex Total cortex Caudate nucleus Hypothalamus Thalamus Grisea centralis Colliculi

0.166 0.086 0.125 0.162 0.125 0.226 0.097 0.025 0.193

_ 0.081 ± 0.070 _ 0.083 _ 0.070 _ 0.090 _ 0.116 _ 0.109 _+ 0.155 ___0.058

n.s. n.s. ~0.01 n.s. n.s. n.s. n.s. n.s. n.s.

(D) C VR of the hind brain structures in perfused and sham animals Perfused Sham Reticular formation 2.674 + 0.4752.197 _+ 0.281 <0.001 Cerebellum 2.264 ± 0.254 1.791 + 0.265 <0.001

96

II infusion induced a generalized increase in CVR of 18-32%. Effects of angiotensin H on local CBF, p02 and pCO2: continuous measurements by heat clearance and mass spectrometry The intracarotid infusion of angiotensin II induced in the ipsilateral caudate nucleus a

CVR

mm Hg/ml.lOOg-lmn

marked decrease in blood flow of 18.2 _ 9.0% (mean ___ S.D., n = 9) and a parallel decrease in pO2 of 14.2 _ 5.3%. Only slight non-significant increases in pCO 2 were observed (Fig. 3). Of the systemic variables (BP, paO2, paCO2), only the BP changed, showing a slight increase of less than 10%. The vascular changes in the caudate nucleus appeared within 2-3 s, the maximum

J

I

M I

J_

Total C t x A n t Ctx Mid Ctx Post Ctx Caud ***

Sham n = 1 6

P ( 0.001 ~

Angiotensin

Hypothal

Thai

Colliculi

Grisea

R.F. Cerebellum

* * P ( 0.01 i n j e c t e d side n=18

C o n t r a l a t e r a l side n - 1 8

Fig. 2. Increases in CVR induced by unilateral intracarotid infusion of angiotensin II at the rate of 0.065 ~g/kg/min in the 10 structures sampled.

97 being attained 30-60 s after the beginning of the infusion. Since there was no change in paO2 (the animal being artificially ventilated), the decrease in pO2 can presumably be attributed to the fall in flow, because an increase in the local metabolism would have been expected to induce a significant rise in local pCO2.

B.P.

DISCUSSION

The present results show that intracarotid infusion of angiotensin II can induce a significant decrease in CBF and therefore an increase in CVR in many cerebral structures, and that these modifications are not significantly modified, in

150 i iiii i _

--

_-

II_

"

ill I

I ll-

I II

] 5o 35

PaCO

2

Pa 0 2

......

.

.

mmHg

.

.

.

.

15 25O 200 150

.

m

P CO 2

-10 po 2

.

:

-25 +25

i

|

t

! anRiotensin

7o

II i-¢

(0.0 6 5pg/kg/m n)

'lmn

'

Fig. 3. Simultaneous and continuous recordings of the effect on local CBF and related variables of a one-minute unilateral intracarotid injection of 0.065 ~tg/kg/min of angiotensin II in a lightly anesthetized rabbit. Notice the small increase in arterial blood pressure (BP), the absence of effect on the electrocorticogram (E.Co.G) and on the arterial partial pressures of O 2 and CO2, the decrease of caudate pO2 (the latency of the changes is due to the response time of the sampling cannula), and the distinct 20% decrease in CBF.

98 absolute terms, by the anesthetic used. The mean blood flow measured under anesthesia, i.e., the value of y at t = 0 in the equation presented in Table I and the mean blood flow measured after 20 min dissipation of anesthesia, i.e., the value of y at t = 20, are comparable to that determined by other authors with other tracer techniques in the same animaF 8. Since blood flow in all structures was significantly reduced despite the slight rise in blood pressure which generally occurred, it is clear that the autoregulatory response was not the only factor involved in the reaction observed. This induced vasoconstriction was of comparable magnitude both in the cerebral structures directly dependent on the carotid artery for their blood supply, and in the contralateral and posterior structures, which, as we discovered in pilot experiments with tritiated water infused in the same way as angiotensin II, only received recirculating concentrations of the infused peptide. The infused tracer reached the homolateral cortical structures and the hypothalamus directly in all cases, and the homolateral caudate nucleus in about 80% of the experiments. These data are compatible with the arteriographic mapping of the rabbit brain 23 and were recently confirmed by Scremin et al.41 who found that an internal carotid artery supplies mainly the homolateral cortex and subcortical structures except for the thalamus and the posterior portion of the nucleus caudatus. It also seems extremely unlikely that locally occurring vasoconstriction (or vasospasm), by rerouting the blood in the circle of Willis, could have caused the unilaterally infused angiotensin to circulate through the brain. Considering first the side-to-side distribution of angiotensin II, the fact that flow in the two hemispheres was similar (Table I) excludes the possibility that angiotensin II-infused blood could have been rerouted to the heterolateral hemisphere (unless pressure in the heterolateral internal carotid artery was much lower than in the homolateral artery). Regarding angiotensin II distribution between the fore- and hindbrain, it would be necessary for carotid blood to flow backwards through the circle of Willis at least to the superior and inferior cerebellar arteries in order for angiotensin II to

act directly on the posterior structures. This again seems impossible since cerebrovascular resistance was increased more in the hindbrain than in the forebrain (Fig. 2). Consequently, it appears that although different cerebrovascular beds received very different concentrations of angiotensin II, they constricted to approximately the same degree. Two hypotheses may by considered in attempting to explain this observation. (1) The recirculating concentration of angiotensin II was equal to or greater than the concentration which causes a maximal contraction by a direct effect on cerebral vessels. (2) The angiotensin, either at the directly infused concentration or at the recirculating concentration, acted by an indirect mechanism on the whole cerebral vasculature. As the estimated recirculating concentration of angiotensin was less than 10-9 M, assuming no catabolism or uptake, the first hypothesis does not seem entirely compatible with the following observations. In vitro and in vivo experiments have shown a distinct effect of angiotensin II at concentrations 10 or 100 times higher. In vitro, Edvinsson et al. 9 observed a very slight effect of angiotensin II on cat middle cerebral arteries at a concentration of 10- 9 M, the maximal contraction being obtained at 10- 7 M. Furthermore, a marked effect on the electrical activity of rabbit middle cerebral arteries was elicited by concentrations of 5 × 10-6 M by Lusamvuku et al.33. In vivo, Hardebo et a119 obtained a significant decrease of rat CBF with angiotensin II at concentrations in the internal carotid artery greater than 10-8 and 10- 7 M, angiotensin II being infused bilaterally in these experiments. It therefore seems unlikely that the low recirculating concentration could induce a maximal contraction in cerebral vessels in our experiments. Furthermore, it is well known that peptides such as angiotensin II do not cross the blood-brain barrier except in certain highly localized areas of the brain. The permeable areas described in the literature46 are periventricular structures such as the area postrema, the subfornical organ and the organum vasculosum lamina terminalis. Consequently, although we cannct exclude a direct ac-

99 tion on hypothetical endothelial receptors, a direct action on smooth muscular receptors of cerebral vessels seems very unlikely. What is more, Mann et al.37 pointed out that the brain receptors they characterized for angiotensin II are probably not located in the vascular tissue but within the parenchyma. This finding corroborates the work of Bennet and Snidera who found that the distribution of angiotensin II receptors in the brain does not correlate with the density of vascularisation. For these reasons, we believe the second hypothesis offers a better possible explanation of our observation of a similar vasoconstriction in all parts of the brain. It is in fact well known that angiotensin II influences systemic vascular resistance not only by a direct effect on the vascular wall, but also by various interactions with the sympathetic nervous system, as we mentioned in the introduction. Moreover, angiotensin II can modify blood pressure and cardiac activity by acting on the hypothalamic periventricular organs in which the blood-brain barrier is deficient. It has been shown that electrical stimulation of these angiotensin-sensitive structures in the rat produced small changes in arterial pressure concomitant with profound shifts in blood flow of various organs '5. Destruction of the region antero-ventral to the third ventricle (AV3V) diminishes the pressor effect of intracarotid angiotensin II in the rat ~5, and Simpson 46 has proposed that these brain structures 'contain elements critical for normal cardiovascular and body fluid control' and that the animals in which this area had been destroyed 'are deficient in their response to angiotensin II delivered either systemically or via the ventricles'. As already mentioned, it is probable that in our experiments this region of the hypothalamus received relatively high concentrations of angiotensin II, so that it is perfectly feasible that this area, which has close connections with another angiotensin-sensitive structure, the subfornical organ, and with the preoptic nuclei, is directly or indirectly responsible for the generalized cerebral vasoconstriction we have observed. Such a hypothetical central action of angiotensin II could be mediated via nervous or hu-

moral pathways which remains to be elucidated. The most obvious nervous pathway, the sympathetic nervous system, does not seem to be involved in this phenomenon, since experiments in 5 animals (Reynier-Rebuffel, unpublished observations) did not show any modification of the cerebrovascular effects of intracarotid infusion of angiotensin II after unilateral post-ganglionic section of the cervical sympathetic chain. Consequently, if one considers the relatively slow development of the cerebrovascular response (30-60 s), another possibility could be an indirect action of angiotensin II through a centrally mediated release of humoral vasoactive substances. Vasopressin seems to be the most likely candidate: this peptide may constrict cerebral vessels in vitro at low concentration6,18. It has also been shown that angiotensin II induces vasopressin release in pressor amounts 2°,25,49and that this is at least implicated in the slow component of the centrally mediated pressor response to angiotensin II. In conclusion, whatever the mechanism involved in the phenomenon observed here, whether or not a central action of angiotensin II is concerned, our results indicate that circulating angiotensin II may interfere with CBF regulation, and should not be considered an inert agent in regard to pressure/flow curves in studies on autoregulation. Furthermore, such an indirect, centrally mediated action of angiotensin II on cerebrovascular tone could well conciliate the previous observation 19 of a decrease of CBF induced by circulating angiotensin II with the notion of a generally impervious barrier to this peptide within the brain. ACKNOWLEDGEMENTS

Thanks are due to Mrs. M. C. Miller, Mrs. S. Damato and Mr. J. L. Corr6ze for skillful technical assistance. We are grateful to Mrs. J. Leizerovici for typing the manuscript. This work was supported by grants from the Centre National de la Recherche Scientifique (E.R.A. 361) and the Institut National de la Sant6 et de la Recherche M6dicale (U. 182).

100 REFERENCES 1 Barnes, K. L., Ferrario, C. M. and Conomy, J. P., Comparison of the hemodynamic changes produced by electrical stimulation of the area postrema and nucleus tractus solitari in the dog, Circulat. Res., 45 (1979) 136-143. 2 Beausang-Linder, M., Effect of sympathetic stimulation on cerebral and ocular blood flow, Acta physiol, scand, 144 (1982) 212224. 3 Bennett, J. P., Jr. and Snyder, S. H., Angiotensin II binding to mammalian brain membranes, J. biol. Chem., 251 (1976) 7423- 7430. 4 Bickerton, R. K. and Buckley, J. P., Evidence for a central mechanism in angiotensin-induced hypertension, Proc. Soc. exp. Biol. Med., 106 (1961)834-836. 5 Bill, A. and Linder, J., Sympathetic control of cerebral blood flow in acute arterial hypertension, Acta physiol. scand., 96 (1976) 114-121. 6 Conde, M. V., Gomez, B. and Lluch, S., Contractile response of isolated human middle artery to vasoactive agents, J. cereb. Blood Flow Metabol., 1 Suppl 1 (1981) 350-351. 7 Edvinsson, L., Nielsen, K. C., Owman, Ch. and West, K. A., Alterations in intracranial pressure blood brain barrier, and brain edema after sub-chronic implantation of a cannula into the brain of conscious animals, A cta physiol. scand., 82 (1971) 527-531. 8 Edvinsson, L., Owman, Ch. and Siesjr, B. K., Physiological role ofcerebrovascular sympathetic nerves in the autoregulation of cerebral blood flow, Brain Research, 117 (1976) 519-523. 9 Edvinsson, L., Hardebo, J. E. and Owman, Ch., Effects of angiotensin II on cerebral blood vessels, A cta physiol. scand, 105 (1979) 381-383. 10 Eklrf, B., Lassen, N. A., Nilsson, L., Norberg, K., Siesjr, B. K. and Torlof, P., Regional cerebral blood flow in the rat measured by the tissue sampling technique; a critical evaluation using four indicators Cl4-antipyrine, C 14ethanol, H3-water and xenon 133, A cta physiol, scand., 91 (1974) 1-10. 11 Falcon, J. C. II, Phillips, M. I., Hoffman, W. E. and Brody, M. J., Effect of intraventricular angiotensin II mediated by the sympathetic nervous system, A mer. J. Physiol., 235 (1978) 392-399. 12 Feldberg, W. and Levis, G. P., The action 0fpeptides on the adrenal medulla. Release of adrenaline by bradykinin and angiotensin, J. Physiol. (Lond.), 171 (1964) 98108. 13 Ferrario, C. M., Dickinson, C. J. and McCubbin, J. W., Central vasomotor stimulation by angiotensin, Clin. ScL, 39 (1970) 239-245. 14 Ferrario, C. M., Gildenberg, P. L. and Mc Cubbin, J. W., Cardiovascular effects of angiotensin mediated by the central nervous system, Circular. Res., 30 (1972) 252262. 15 Finck, C. D., Haywood, J. R., Bryan, W. J., Packwood, W. and Brody, M. J., Central site for pressor activity of blood-borne angiotensin in rat, Fed. Proc., 38 (1979) 1233. 16 Gavras, H. and Liang, C., Acute renovascular hypertension in conscious dogs, Circular. Res., 47 (1980) 356-365. 17 Gildenberg, P. L., Ferrario, C. M. and McCubbin, J. W., Two sites of cardiovascular action of angiotensin II in the brain of the dog, Clin. Sci., 44 (1973) 417 420. 18 Henko, J., Hardebo, J. E. and Owman, Ch., Effect ofva-

19

20

21 22 23 24 25 26

27

28 29

30 31 32 33

34 35

36

rious neuropeptides on cerebral blood vessels, J. cereb. Blood Flow Metabol., l Suppl 1 (1981)346-347. Hardebo, J. E., Nilsson, B. and Owman, Ch., Circulating angiotensin decreases cerebral blood flow relevance to studies on the upper limit of autoregulation and blood brain barrier function, Acta neurol, scand., 60 Suppl 72 (1979) 604- 605. Hoffman, W. E., Phillips, M. I., Schmid, P. G., Falcon, J. C. and Weet, J. F., Antidiuretic hormone release and the pressor response to central angiotensin II and cholinergic stimulation, Neuro#harmacology, 16 (1977) 463-472. Hugues, J. and Roth, R. H., Evidence that angiotensin enhances transmitter release during sympathetic nerve stimulation, Brit. J. Pharmacol., 41 (1971) 239-255. lshibashi, S. and Nicolaidis, S., Hypertension induced by electrical stimulation of the subfornical organ (SFO), Brain Research, 6 (1980) 135-139. Jeppsson, P. G. and Olin, T., CerebralAngiography in the Rabbit, CWK Gleerup, Lund, 1960, pp. 13-26. Joy, M. D. and Lowe, R. D., Evidence that the area postrema mediates the central cardiovascular response to angiotensin II, Nature (Lond.), 228 (1970) 1303-1304. Keil, L. C., Summy-Long, J. and Severs, W. B., Release of vasopressin by angiotensin II, Endocrinology, 96 (1975) 1063-1065. Kontos, H. A., Wei, E. P., Dietrich, W. D., Navari, R. M., Povlishock, J. T., Ghatak, N. R., Ellis, E. F. and Patterson, J. L., Jr., Mechanism of cerebral arteriolar abnormalities after acute hypertension, Amer. J. Physiol., 240 (1981)511-527. Lacombe, P., Meric, Ph., Reynier-Rebuffel, A. M. and Seylaz, J., Critical evaluation of cerebral blood flow measurements made with 14C-ethanol, Med. biol. eng. Comput., 17(1979)602 618. Lacombe, P., Meric, P. and Seylaz, J., Validity of cerebral blood flow measurements obtained with quantitative tracer techniques, Brain Res. Rev., 2 (1980) 105-169. Lacombe, P., Reynier-Rebuffel, A. M. and Seylaz, J., Evidence of an adrenal factor responsible for the cerebral vasodilation occurring during immobilization stress, J. cereb. BloodFlow Metabol., 1 Suppl 1 (1981)492-493. Lellouch, J. and Lazar, P., Mbthodes Statistiques en Exp@rimentation Biologique, Flammarion Mrdecine-Sciences, Paris, 1974, pp. 107-114. Lewis, G. P. and Reit, E., Further studies on the action of peptides on the superior cervical ganglion and suprarenal medulla, Brit. J. Pharmacol., 26 (1966) 444-460. Linder, J., Cerebral and ocular blood flow during a 2blockade: evidence for a modulated sympathetic response, A ctaphysiol, scand, 113 (1981) 511- 517. Lusamvuku, N. A. T., Sercombe, R., Aubineau, P. F. and Seylaz, J., Correlated electrical and mechanical responses of isolated rabbit pial arteries to some vasoactive drugs, Stroke, I0 (1979) 722732. Mac Cubbin, H. W. and Page, I. H., Renal pressor system and neurogenic control of arterial pressure, Circular. Res., 12(1963)553 559. MacKenzie, E. T., Strandgaard, S., Graham, D. I., Jones, J. V. and Harper, A. M., Effects of acutely induced hypertension in cats on pial arteriolar caliber, local cerebral blood flow, and the blood-brain barrier, Circulat. Res., 39 (1976) 33-41. Mangiapane, M. L. and Simpson, J. B., Subfornical organ: forebrain site of pressor and dipsogenic action of

101 angiotensin II,Amer. J. Physiol., 239 (1980) 382-389. 37 Mann, J. F. E., Schiller, P. W., Schiffrin, E. L., Boucher, R. and Genest, J., Brain receptor binding and central actions of angiotensin analogs in rats, Amer. J. Physiol., 24 (1981) 124~ 129. 38 Miselis, R. R. and Fluharty, S. J., Local cerebral blood flow in the anterior circum ventricular organs: a test of the vasomotor hypothesis for drinking to angiotensin II. In: XXVII International Congress of Physiological Sciences, Budapest, Hungary, 14 (1980) 586. 39 Pinard, E. and Seylaz, J., Cerebral vasodilatory mechanisms involved in a physiological reaction related to immobilization stress, J. cereb. Blood Flow Metabol., 1 Suppl 1 (1981) 333-334. 40 Raichle, M. and Grubb, R., Jr., Centrally-administered angiotensin I1 increases brain water permeability, Acta neurol, scand., (supp172) 60 (1979) 82-83. 41 Scremin, O. U., Sonnenschein, R. R. and Rubinstein, E. H., Cerebrovascular anatomy and blood flow measurements in the rabbit, J. cereb. Blood Flow Metabol., 2 (1982) 55-66. 42 Seylaz, J., Aubineau, P. F., Correze, J. L. and Mamo, H., Techniques for continuous measurement of local cerebral blood flow, PaO2, PaCO2 and blood pressure in the non anesthetized animal, Pillagers Arch. ges. PhysioL, 340 (1973) 175--180. 43 Seylaz, J., Pinard, E., Correze, J. L., Aubineau, P. F. and Mamo, H., Quantitative continuous measurement of blood gas tensions by mass spectrometry, J. appl. PhysioL, 37 (1974) 937-941. 44 Seylaz, J. and Pinard, E., Continuous measurement of

gas partial pressures in intracerebral tissue, J. appl. Physiol., 44 (1978) 528-533. 45 Simpson, J. B., Epstein, A. N. and Camardo, J. S., The localization of dipsogenic receptors for angiotensin I1 in the subfornical organ, J. cornp. Physiol. Psychol., 92 (1978) 581-608. 46 Simpson, J. B., The circum-ventricular organs and the central actions of angiotensin, Progr. Neuroendocrinol., 32 (1981) 248-256. 47 Strandgaard, S., Jones, J. V., MacKenzie, E. T. and Harper, A. M., Upper limit of cerebral blood flow autoregulation in experimental renovascular hypertension in the baboon, Circular, Res., 37 (1975) 164--167. 48 Toda, N. and Miyazaki, M., Angiotensin-induced relaxation in isolated dog renal and cerebral arteries, Amer. J. Physiol., 240 (! 981) 247-254. 49 Uhlick, E., Weber, P. and Gr/3schel-Stewart, U., Angiotensin stimulated AVP release in humans, KlinWochschr., 53 (1975) 177-180. 50 Webb, R. C., Angiotensin II-induced relaxation of vascular smooth muscle, Blood Vessels, 19 (1982) 165-176. 51 Wei, E. P., Kontos, H. A. and Patterson, J. L., Vasoconstrictor effect of angiotensin on pial arteries, Stroke, 9 (1978) 487-489. 52 Yu, R. and Dickinson, C. J., Neurogenic effects of angiotensin, Lancet, 2 (1965) 1276-1277. 53 Yu, R. and Dickinson, C. J., The progressive pressor response to angiotensin in the rabbit. The role of the sympathetic nervous system, Arch. Int. Pharmacodyn., 191 (1971)24-36.