Acta physiol. scand. 1975. 93. 269-276 From the Brain Research Laboratory, E-Blocket, University Hospital, Lund, Sweden

Cerebral Blood Flow and Oxygen Consumption in the Rat in Hypoxic Hypoxia BY

H A L L D ~J~HANNSSON R and Bo K. SIESJO Received 9 September 1974

Abstract J~HANNSSON, H. and B. K. SIESJO.Cerebral blood flow and oxygen consumption in the rat in hypoxic hypoxia. Acta physiol. scand. 1975. 93. 269-276. In order t o measure cerebral blood flow (CBF) and cerebral metabolic rate for oxygen (CMRo,) at pronounced degrees of hypoxic hypoxia the Pao, of artificially ventilated and normocapnic rats was reduced to between 47 and 22 mm Hg for 15-25 min with subsequent measurements of CBF, using a ‘“Xenon modification of the Kety and Schmidt technique, and of the arteriovenous difference in oxygen content, the venous blood being obtained from the superior sagittal sinus. When the Pao, was reduced to minimal values of 22 mm Hg CBF increased 4- t o 6-fold, the increase in CBF being unrelated to changes in blood pressure or Paco,. The CMRo, remained unchanged at all levels of hypoxia. It is concluded that the maintenance of a normal, or near-normal, cerebral energy state even at extreme degrees of hypoxic hypoxia depends solely on a homeostatic increase in CBF.

In spite of the fact that brain cells are considered to be markedly sensitive to oxygen deficiency, there are now many studies to show that the energy metabolism of the brain is extraordinarily well maintained even at pronounced degrees of hypoxic hypoxia. Thus, although lactic acid accumulates in the brain at Pa,, values of below about 50 mm Hg, the tissue concentrations of ATP, ADP and AMP remain close to normal even if the Pa,, is reduced to 20-25 mm Hg (Gurdjian et al. 1944, Schmahl et al. 1966, Siesjo and Nilsson 1971, Duffy et al. 1972, MacMillan and Siesjo 1972, Bachelard et al. 1974). The maintenance of a normal, or near-normal, energy state in profound hypoxia indicates that powerful homeostatic mechanisms exist that secure an almost adequate supply of oxygen to the cells. One such mechanism is the increase in cerebral blood flow (CBF), which is clearly observed at Pa,, values of below 50 mm Hg,and may amount to several hundred percent of normal at extreme degrees of hypoxia (Kety and Schmidt 1948, Courtice 1941, McDowall 1966, Cohen et al. 1967, Kogure et al. 1970). Another mechanism was recently proposed by Duffy et al. (1972), who exposed mice to 5 3 % 0, and measured the energy consumption of the brain by means of the “closed box” method of Lowry et a/. (1964). Since Duffy et al. (1972) found a decrease in “energy consumption” in the brain 269

270

H A L L D ~ RJ~HANNSSONAND BO K.

SIESJB

of the hypoxic mice, they postulated that the tissue can protect itself against cellular anoxia by decreasing its energy demands. This attractive hypothesis receives some support from studies of transmission mechanisms in the hypoxic brain. Thus, hypoxia is accompanied by signs of a decreased rate of synthesis of indole and catechol neurotransmitters (Davis and Carlsson 1973, Davis et al. 1973). Furthermore, there are reports of decreased tissue concentrations of the excitatory amino acids aspartate and glutamate and of a n increased concentration of the inhibitory amino acid GABA in hypoxia (Tews et al. 1963, Wood et a/. 1968,Siesjo and Nilsson 1971,Duffy et al. 1972,MacMillan and Siesjo 1972,Bachelard et al. 1974). Previous studies in man have shown that the cerebral metabolic rate for oxygen (CMR,,) remains unaltered at Pa,, values of 35-40 mm Hg (Kety and Schmidt 1948, Cohen et al. 1967). At these degrees of hypoxia there is thus no sign of a decreased consumption of energy. Since there were no previous, quantitative studies of CMR,, at more pronounced degrees of hypoxia, we examined CBF and CMR,, in rats after reducing Pa,, to minimal values of 22 mm Hg. The objective of the study was to evaluate a possible decrease in CMR,, at degrees of hypoxia corresponding to inhalation of 5-8% oxygen. A preliminary report of the findings has been published (Jbhannsson and Siesjo 1973).

Methods The experiments were performed on male Wistar rats (320-420 g) which had free access to rat pellets and water until operation. Anesthesia was induced with 2-3 % halothane. When unresponsive to tilting of the jar, the animals were tracheotomized, injected with tubocurarinechloride i.p. (0.5 mg. kg-’) and maintained artificially ventilated o n 70% N,O and 30% 0,. One femoral artery was cannulated for blood pressure recording and the other for sampling of blood. The latter catheter was cut to a length of about 3 cm t o minimize errors due to catheter “smearing”. One femoral vein was cannulated and used for infusion of fresh donor blood during the measurement of CBF (see Norberg and Siesjo 1974). The posterior part of the superior sagittal sinus was exposed by means of a small burr hole for sampling of cerebral venous blood. The body temperature, as measured in the rectum, was kept close to 37°C by means of intermittent heating. The ventilation was adjusted to give a Pam, of about 35 mm Hg. After a steady state period of 30 min the 0,concentration of the insufflated gas mixture was lowered so as to give Pao, values of 2 C 5 0 mm Hg. while maintaining the N,O concentration at 70%. In the control group the 0,concentration was maintained at 30% throughout the experiment. At Pa,* values of 20-25 mm Hg the Paco. fell in spite of constant ventilation. In order to maintain Paco. close t o 35 mm Hg in these animals CO, was added to the gas mixture. In the majority of the experiments the period of hypoxia was 25 min. Fifteen min before the end of the hypoxic (or control) period an identical gas mixture was delivered to the ventilator from a rubber bag that also contained about 10 pCi of 133Xenon (obtained from AB Atomenergi Studsvik Sweden). At the end of the hypoxic (or control) period samples were taken from the artery and from the superior sagittal sinus for measurements of 133Xenon activity and for total oxygen content (To,). The bag containing IS3Xe was then disconnected, the Xenon-free gas mixture was again administered, and repeated samples were taken from artery and vein for measurements of 133Xenonactivity. A new set of arterial and venous samples were taken about 2 min after the beginning of desaturation for To,. During the whole desaturation period fresh donor blood was infused Lo. at a rate sufficient to maintain mean arterial blood pressure above 120 mm Hg. During the 25 min hypoxic period (Pao, < 35 mm Hg) there was a gradual fall of plasma p H and the arterial total 0, content decreased in spite of constant Pa,,. In order to minimize the plasma acidosis six animals were made hypoxic for only 15 min (Pao, 25 mm Hg, or lower) and the arterial Pco, was allowed to fall spontaneously by omitting addition of CO,. In these animals, only three of which met the criteria set up for evaluation (see below), 133Xenonwas administered during the whole hypoxic period of I5 min.

CBF AND CMRo, IN HYPOXIA

271

Arterial PO,, Pco, and pH were measured using microelectrodes operated at 37'C, with due corrections for deviation in temperature. To. was measured in 25 pl samples using the method of Fabel and Lubbers (1964, see Norberg and Siesjo 1974, Borgstrom et al. 1974). 13'Xenon activity in arterial and cerebral venous blood was determined with a scintillation counter (Wallac). CBF was calculated from the desaturation curves for artery and vein, using the trapezoid rule (Kety and Schmidt 1948. For a more detailed account of the CBF and CMRo, methods, see Eklof Pt al. 1973, Norberg and Siesjo 1974). CMRo, was calculated by multiplying CBF with the arteriovenous difference in oxygen content (AVDo,), as measured 2 min after the start of desaturation.

Results The present methods for CBF and CMR,, have now been used in several hundred determinations. At normal flow values the evaluation of the arteriovenous differences in ISSXenon activity and To, presents no difficulties (Fig. 1, unfilled symbols). However, in hypoxia the arteriovenous differences in 133Xenonand To, become so small that accurate determination of CBF and CMR,, is difficult. The values presented here were obtained during the course of one year and represent a selection. The following criteria of selection were applied: (1) The blood pressure remained stable within 10-20 mm Hg during desaturation, and never fall below 110 m Hg. If larger changes in blood pressure occurred, the arteriovenous differences in To,,as measured before and during desaturation, varied unduly ( '0.5 ml. (100 ml-l). (2) It was possible to draw smooth lines through all 13sXenonvalues for both artery and vein (an example is given in Fig. 1, filled symbols). In most of the experiments that did not meet these criteria, and which were therefore excluded from the material, it was possible to obtain approximate CBF and CMR,, values. If these experiments had been included in the material, they would not have changed any of the present conclusions. Fig. 2 shows that the arterial Po,, but not the TOv,values remained stable between the 15th and the 25th rnin of hypoxia. The fall in To, was obviously related to a gradual decrease in plasma pH, which was observed in animals with an arterial Po, of 35 mm Hg, or lower (see Fig. 2). In 3 animals (not shown in the figure) Pa,, was reduced to 20-25 mm Hg for only 15 min, and the PacO,was allowed to fall spontaneously (the values were, 25.2, 24.0 and 31.1, respectively). In these animals the plasma pH remained between 7.26-7.33 and the arterial To, values were 4.35, 4.59 and 4.50 ml.(100 mI)-l, respectively. The individual CBF and CMR,, values are presented in Fig. 3. There were 6 control animals (Pa,, '120 mm Hg) and 19 hypoxic animals with Pa,, values varyinge between 47 and 22 mm Hg. 9 animals had a Pa,, of 25 mm Hg, or lower. Of those, 6 were made hypoxic for 25 rnin (unfilled circles) and 3 for 15 min (filled circles). The figure shows that CBF increased in all hypoxic animals. At the most pronounced degrees of hypoxia (Pa,, 20-25 mm Hg) CBF increased 4-5-fold, with one single value reaching 700 ml. (100 g)-'.min-'. Although 3 of the CMR,, values at Pa,, 20-25 mm Hg (25 min of hypoxia) fell below the lower range of the control CMR,, values, there was no obvious change in CMR,, at any Pa,, range. In order to allow a statistical comparison at different levels of hypoxia the material was divided into the following Pa,, ranges: (1) > 120 mm Hg, (2) 47-37 mm Hg, (3) 33-27 mm Hg, and (4) 25-22 mm Hg. The last Pa,, range was tabulated both by combining the 25 and 15 min animals (n = 9) and by choosing the 25 min animals only (n = 6). Table I gives the

272

H A L L D ~ RJ~HANNSSONAND BO K. S I E S J ~

Fig. 1 . Representative desaturation curves for 13SXenon in normoxia (unfilled symbols) and hypoxia (filled symbols). 0:femoral artery, A : superior sagittal sinus. 0

I

2

3

4

5

6

7

0

9

10 Minutes II

mean arterial blood pressure, body temperature, arterial Po,, P,,,, pH, arterial and venous To:,AVD,,, CBF and CMR,,. Body temperature in the hypoxic groups did not vary significantly from the controls and all animals (except the 15 min animals, see above) had similar P,, values. Mean arterial blood pressure fell significantly in the hypoxic group with Pa,? 20-25 mm Hg and at no Pa,, range was there an increase in blood pressure. Thus, the increase in CBF was unrelated to an increase in blood pressure. Arterial To, fell to reach 20% of normal at the most extreme degrees of hypoxia, and the AVD,, was similarly reduced. However, there was a corresponding increase in CBF and there was thus no change in CMR,,.

Discussion As remarked in the introduction, two previous studies have demonstrated that moderate degrees of hypoxia do not lead to a measurable decrease of CMR,, in man. In the study of Kety and Schmidt (1948) 10%0,was administered to spontaneously breathing subjects and CBF increased to 135% of normal in spite of an associated fall in Pa,,,. Cohen ef al. (1967) exposed artificially ventilated and normocapnic subjects to about 7% O2 and recorded an increase in CBF to 170% of normal. In none of these studies did the CMR,, of the brain deviate from normal. The objective of the present study was to measure quantitatively CBF and CMR,, at more pronounced degrees of hypoxia. This was achieved by reducing Pa,, to minimal values of 22-25 mm Hg in artificially ventilated rats. The degree of hypoxia thus achieved is considerably more severe than in the previous studies on man. This is not only due to the lower Pa,,, but also to the fact that the oxygen dissociation curve of the rat (Sherwood et al. 1950) is shifted to the right (Ph0about 40 mm Hg). In the study of Kety and Schmidt (1948) inhalation of 10% Oa,which lowers Pa,, to about 40 mm Hg, reduced arterial oxygen content from 19 to 16 ml.(100 mI)-l. At a comparable reduction in Pa,, in the rat (see Table I), arterial To,was lowered to 11.7 ml.(100 mI)-l, i.e. to about 5 5 % of normal. At Paor 22-

273

CBF AND CMR,, IN HYPOXIA 25

0

tl

0

I

I

15

25

7.5OOy

hC-+-+

I

I

I

0

15

25

0

15

Minutes

Minutes

I 25 Minutes

Fig. 2. Individual values for Pao,. [OJ,and pH in the rat after 15 and 25 min of hypoxia, respectively. The data are compared to the corresponding values obtained in control rats (means +S.E.).

25 nun Hg the arterial To,was reduced to less than 20% of normal in the animals that were hypoxic for 25 min. Thus, when compared to the previous studies of CBF and CMR,, in hypoxia, the present degree of hypoxic hypoxia must be considered excessive. It should be pointed out, though, that the severity of the hypoxic hypoxia in the present type of experiments is not only given by the reduction in Pa,,, but also by the influence of plasma pH on oxygen dissociation. When normocapnic animals are exposed to very low oxygen

.............. 0.-------------

.... .....

a

Fig. 3. Individual CBF and CMRo, values in the rat after 25 min of exposure to hypoxia. In 3 animals (filled circles) the period of hypoxia was I5 min and Paco, was allowed to fall spontaneously.

0 O

10

20

30

40

50

Pao2 rnrnHg

18 - 155812

O D o

00

120 130 140 150 160

274

H A L L D ~ RJ~HANNSSONAND BO K. SIESJO

TABLE I. Effect of hypoxia on physiological parameters, cerebral blood flow and cerebral metbolic rate in rats. Means k S.E.M.

Num- PaO, ber of animals 6 5 4 6 9

ArtemmHg rial pressure m m Hg 140 f6 40.8 k2.1 30.5 kI.5 23.5 k0.6 23.4 kO.5

Tempera ture "C

PaCO,

pH

[O& [Oal, ml.(100 mI)-l

m m Hg

AVDO,

CBF CMRO~ m1.(100 gamin)-'

143 2

36.8 k0.2

k0.l

7.382 k0.009

22.18 k0.72

12.92 k0.65

9.25 k0.71

k

114 6

10.31 20.34

133

36.7 k0.2

38.4 1.7

7.257 k0.061

11.74 k1.41

7.06 k0.99

4.68 k0.73

235 k21

10.33 i0.36

37.5 k1.4

7.146 k0.020

5.54 k0.39

3.25 k0.50

2.43 kO.14

439 k28

10.55 k0.31

2 5

37.1 k0.l 36.5 k0.2

35.4 k0.7

7.150 k0.021

4.08 k0.33

2.06 k0.31

2.02 k0.06

516 f41

10.36 k0.71

127 3

36.5 k0.2

32.5

7.185 k0.029

4.66 L0.44

2.18 50.21

2.03 k0.04

528 +28

10.65 kO.48

k

f 3 124 +I1 128

37.3

k 1.6

concentrations for 2&30 min, an excessive plasma acidosis develops, aggravating the hypoxia. Furthermore, at any given artcrial oxygen content a fall in blood pressure adversely affects the oxygenation of the brain. If spontaneously breathing animals are exposed to 5 yo 02,there is a fall in Pa,, to about 20 mm Hg (Lewis et al. 1974), but since the animals hyperventilate vigorously, most of them show a plasma alkalosis that would promote oxygenation of hemoglobin at the existing Pa,,. There are thus reasons to believe that the degrees of hypoxic hypoxia studied presently is ar least as severe as that encountered when spontaneously breathing rats are exposed to 5 % 0,. Available evidence indicates that the Kety and Schmidt technique (1948), as applied to the rat, quantitatively measures C B F in supratentorial cerebral (cortical) regions (Eklof et al. 1973, Norberg and Siesjo 1974). Of particular importance are the facts that the tissues drained by the superior sagittal sinus do not contain slowly perfused tissue masses, and that extracerebral contamination of venous blood can be considered insignificant (see Norberg and Siesjo 1974). In view of these facts, there are no a priori reasons to suspect that systematic errors should be expected in high flow situations, like hypoxia. Our data show an unchanged CMR,, at moderate degrees of hypoxia (see Table I) and in hypercapnia (Eklof et al. 1973), quite in agreement with previous studies (Kety and Schmidt 1948, Cohen et al. 1967). The difficulties encountered in measuring CBF and CMR,, at extreme degrees of hypoxia are therefore technical and due to the low arteriovenous differences in To, and lS3Xenonactivity. Due to these difficulties conclusions should be drawn with some caution when minor differences in CMR,, between the groups are considered. The present results unequivocally show that CMR,, is maintained close to normal even at extreme degrees of hypoxic hypoxia. The data cannot exclude the possibility that a minor reduction in CMR,, may occur in some animals that have Pa,, reduced to about 20 mm Hg for 25 min and that have an excessive plasma acidosis, and therefore also a very marked lowering of arterial To,. However, although such a reduction would be com-

CBF AND CMRO, IN HYPOXIA

275

patible with results demonstrating a minor change in cerebral energy state at Pa,, values of about 20 mm Hg (MacMillan et al. 1974), the present results fail to give clear proof of its existance. At any rate, it can be concluded that there is no measurable reduction in CMR,, before the energy state of the tissue is affected. Furthermore, the present techniques fail to corroborate a reduction in CMR,, of the magnitude reported by Duffy et al. (1972). These authors calculated a fall in CMR,, by 40% when using the rate of disappearance of glucose and glycogen, and a 15% reduction when using instead the rate of increase in lactate. It has subsequently been demonstrated that a reduction in body temperature of the order occurring in the experiments of Duffy et al. (1972) (34°C) lowers CMR,, by 15-20% (Hagerdal et al. 1974). Thus, the fall in body temperature can explain all of the reduction in CMR,,, derived from lactate data, and a large part of that calculated from the figures for glucose and glycogen. We therefore conclude that any theory that attempts to explain energy homeostasis in the brain in hypoxic hypoxia must take into account that not only the energy state, but also CMR,,, are maintained close to normal even at extreme degrees of hypoxia. This study was supported by grants from the Swedish Medical Research Council (Projects No. 14X-263 and 14X-2179), from the Swedish Bank Tercentenary Fund, from U.S.PHS Grant No. 5 R01 NS0783806 from NIH, and from Magnus Bergvalls Stiftelse.

References BACHELARD, H. S., L. D. LEWIS, U. P O N T ~and N B. K. SIEsJO, Mechanisms activating glycolysis in the brain in arterial hypoxia. J. Neurochem. 1974. 22. 395-401. BORGSTROM, L., M. HAGERDAL, L. D. LEwis and U. P O N T ~ N Polarographic , determination of total oxygen content in small blood samples. Scand. J. d i n . Lab. Inoesr. 1974. 34. 375-380. T. C. SMITH, M. RElVlCH and H. WOLLMAN, Effects of hypoxia and norCOHEN,P. J., S. C. ALEXANDER, mocarbia on cerebral blood flow and metabolism in conscious man. J. appl. Physiol. 1967. 23. 183-189. COURTICE, F. C., The effect of oxygen lack on the cerebral circulation. J . Physiol. (Lond.) 1941. 100. 19821 1. DAvrs, J. N. and A. CARLSSON, The effect of hypoxia on monoamine synthesis, levels and metabolism in rat brain. J . Neurochern. 1973. 21. 783-790. V. MACMILLAN and B. K. sIESJ6, Brain tryptophan hydroxylation: Dependence DAVIS, J. N., A. CARLSSON, on arterial oxygen tension. Science 1973. 182. 72-74. DUFFY,T. E., S. R. NELSONand 0. H. LOWRY,Cerebral carbohydrate metabolism during acute hypoxia and recovery. J . Neurochem. 1972. 19. 959-977. EKLOF,B., N. A. LASSEN,L. NILSSON, K. NORBERG and B. K. SIESJO,Blood flow and metabolic rate for oxygen in the cerebral cortex of the rat. Arta physiol. scand. 1973. 88. 587-589. FABEL, H. and D. W. LUBBERS,Eine schnelle Mikromethode zur serienmassigen Bestimmung der 0,konzentration im Blut. Pfliig. Arch. ges. Physiol. 1964. 281. 32-33. E. S.,W. E. STONEand J. W. WEBSTER, Cerebral metabolism in hypoxia. Arch. Neurol. Psychiat. GURDJIAN, (Chic.) 1944. 51. 472-477. H ~ E R D A M., L , J. HARP,L. NlLssoN and B. K. SIESJO,The effect of induced hypothermia upon oxygen consumption in the rat brain. J. Neurochem. In press. J6HANNSSON. H. and €3. K. SIESJO, Blood flow and oxygen consumption of rat brain in profound hypoxia. Acta physiol. scand. 1974. 90. 281-282. KETY,S. S. and C. F. SCHMIDT, The effects of altered arterial tensions of carbon dioxide and oxygen o n cerebral oxygen consumption of normal young men. J. clin. Inuesr. 1948. 27. 484-492. K., P. SCHEINBERG, 0. M. REINMUTH, M. FUJISHIMA and R. BUSTO,Mechanisms of cerebral vasoKOGURE, dilatation in hypoxia. J. appl. Physiol. 1970. 29. 223-229.

276

H A L L D ~ RJ~HANNSSONAND BO K. S I E S J ~

LEWIS,L. D.,U. PONTEN,and B. K. SiesJ6, Arterial acid-base changes in unanaesthetized rats in acute hypoxia. Resp. Physiol. 1973. 19. 312-321. LOWRY,0. H., J. V. PASSONNEAU, F. X. HASSELBERGER and D.W. SCHULZ,Effect of ischemia on known substrates and cofactors of the glycolytic pathway in brain. J. biol. Chem. 1964. 239. 18-30. MCDOWALL, D. G., Interrelationships between blood oxygen tension and cerebral blood flow. In: Oxygen Measurements in Blood and Tissues, ed. by J. P. Payne and D. W. Hill, Churchill London, 1966. 205214.

MACMILLAN. V. and B. K.SIESI~, Brain energy metabolism in hypoxemia. Scand. J. din. Lab. Invest. 1972. 30. 127-136.

MACMILLAN. V., SALPORD, L. G. and B. K. SIESl6, Metabolic state and blood flow in rat cerebral cortex, cerebellum and brainstem in hypoxic hypoxia. Acra physiol. scand. 1974. 92. 103-1 13. NORBERG, K. and B. K. SlESl6, Quantitative measurement of blood flow and oxygen consumption in the rat brain. Actaphysiol. scand. 1974. 91. 154-164. SCHMAHL, F. W., E. BETZ,E. DETTINQER and H. J. HOHORST, Energiestoffwechsel der Grosshirnrinde und Elektroencephalogram bei Sauerstoffmangel. Pfliig. Arch. ges. Physiol. 1966. 292. 46-59. SHERWOOD, E. J., B. G. MACGRAITH and H.H. SCULTHORPE, Pathological processes in disease. 1. Blood of the albino rat: Approximate physico-chemical description. Amer. J. trop. Med. Parasit 1950. 44. 168-1 86.

SIESJ~, B. K. and L. NILSSON. The influence of arterial hypoxemia upon labile phosphates and upon extracellular and intracellular lactate and pyruvate concentration in the rat brain. Scand. J. clin. Lab. Inuesr. 1971. 27. 83-96.

TEWS, J. K., S. H. CARTER, P. D. ROAand W. E. SToNe, Free amino acids and related compounds in dog brain: post-mortem and anoxic changes, effects of ammonium chloride infusion, and levels during seizures induced by picrotoxin and by pentylenetetrazol. J. Neurochem. 1963. 10. 64-653. WOOD,J. D.,V. J. WATSON and A. J. DUCKER, The effect of hypoxia on brain y-aminobutyric acid levels. J. Neurochem. 1968. IS. 603-608.