NMR IN BIOMEDICINE. VOL. 5, 48-52 (1992)

Brain Metabolism and Blood Flow in Acute Cerebral Hypoxia Studied by NMR Spectroscopy and Hydrogen Clearance

Kathryn Allen, Albert L. Busza,*t H. Alan Crockard and David G. Gadiant Institute o f Neurology, Queen Squarehndon WClN 3BG. UK and tHunterian Institute, The Royal College of Surgeons of England, 35-43 Lincoln's Inn Ficlds. London WC2A 3PN, UK

We have developed a reliable, reproducible model of hypoxia in the gerbil. 'H and "P NMR spectroscopy demonstrates that cerebral energy metabolism is very resistant to hypoxia. Cerebral blood flow (measured by hydrogen clearance) began to increase when the arterial oxygen tension (p.0,) was reduced to 40-50 mm Hg, and there was no change in phosphorus metabolites or lactate until psOz was below 40 mm Hg. In 50% of the animals lactate increased prior to any change in the phosphorus metabolites or intracellular pH, suggesting that 'H NMR may be more sensitive than 31P NMR at detecting hypoxic or ischaemic changes. The calculated rate of oxygen delivery at a time when phosphorus energy metabolism becomes impaired is similar in both hypoxia and ischaemia (ca 4 mL/100 g/min). We suggest that the critical factor in ischaemia is the reduction in oxygen supply, rather than the accumulation of toxic metabolites, such as lactate.

INTRODUCTION

The combination of high energy consumption and low energy stores renders the brain particularly dependent on a continuous supply of oxygen and nutrients by the blood. Thus conditions such as asphyxia, circulatory arrest and stroke frequently lead to cerebral dysfunc- tion and neuronal death. This may be due either to a problem of nutrient supply to the brain (i.e., oxygen and glucose), or to a failure to remove harmful metabo- lic products such as lactate. However, it is difficult to separate the effects of lack of substrate from the effects of an increase in lactate.

In earlier studies we examined cerebral ischaemia in the gerbil using simultaneous NMR spectroscopy and cerebral blood flow (CBF) measurements. '.' These demonstrated a close link between energy metabolism and CBF, with a flow threshold of ca 20 mL/100 g/min below which there is depletion of high energy phos- phates, increase in lactate and decrease in intracellular pH (pHi) associated with ischaemia. We have also studied the post-ischaemic phase' and shown the time course of recovery of "P metabolites, pH, and lactate.

We have now extended these studies to include cerebral hypoxia, in order to examine separately how decreased oxygen supply alters brain metabolism. A major aim of this study was to isolate the hypoxic component of this complex pathological mechanism.

MATERIALS AND METHODS

Apart from the hypoxia model, much of the experimen- tal methodology has been described elsewhere'.* and so will be dealt with only briefly here. The numbers of

Author to whom correspondence should be addressed. Abbreviations used: BP. blood pressure; CBF, ccrebral blood flow; MABP, mean arterial blood pressure; paOz. arterial oxygen tension; PCr, phasphocreatine.

animals used were as follows: (i) NMR spectroscopy study, n = 17 (of which 10 animals had 'H spectra and CBF measured in addition to "P spectra); (ii) CBF study, n = 10 (included in the NMR spectroscopy study); (iii) blood pressure (BP) and ECG study, n = 7.

NMR spectroscopy

Seventeen adult male gerbils (weight range 55-65 g) were anaesthetized with a halothane/oxygen mixture. Since ventilation of such a small animal inside the NMR magnet is very inefficient, due to the large ventilatory dead space involved, the animals were allowed to breathe spontaneously. Heart and respiratory rates were closely monitored over the course of each experi- ment. Body temperature was maintained at 36.5-37 "C with warm air, and a femoral artery catheter was used to take blood samples without removing the animal from the magnet. Arterial blood gas levels (sample size 50 yL) were measured using a Corning blood gas ana- lyzer (model 178) (Corning Medical, Halstead, UK), and haemoglobin measured using a Sahli haemoglobin- ometer (Gelman Hawksey Ltd., Lancing, England). In 10 of these animals CBF was also measured by hydro- gen clearance .'. '

NMR spectroscopy was performed on a Bruker AM-360 spectrometer, using a vertical 8.5 T magnet and a purpose-built probe-head of outer diameter 7.3 cm. Two concentric surface RF coils were fixed to the skull using epoxy resin. The outer coil, tuned to "P (145.8 MHz), had dimensions of 10 X 8 mm, and the adjacent inner coil, tuned to 'H (360.1 MHz), had dimensions of 9 X 7 mm.

For the 31P spectra, R F pulses were applied every 700 ms as described previously,2 and blocks of 400 scans were accumulated. Where appropriate the data were added together to give a better SIN ratio. For data processing of the "P spectra the convolution difference technique' was used, with line broadenings of 40 and

0952-3480/92/010048-05 $05.00 0 1992 by John Wiley & Sons, Ltd

Received 22 July 1991 Accepted (revised) 27 September 1991

NMR AND CBF IN HYPOXTA 49

400 Hz, to suppress the broad underlying signal from bone and phospholipids. The spectra were analysed by calculating peak areas, using squared paper, and the Pi/(PCr + Pi) ratio was calculated.

'H spectra were acquired in 10 of the animals, and were accumulated using two interleaved pulse sequences, with the water signal on resonance and optimization on lactate:

133 i-t-'18Oo'-t-acquire (A)

133i -r-2662-t-acquire (B)

where t = 68 ms and '180"' is a hard refocusing pulse as described previously.'

The addition of the spectra acquired under these conditions provides a simple method of monitoring the lactate signal at 1.32 ppm, while suppressing the water and fat signals.' Typically 64 or 128 scans were accumu- lated at interpulse intervals of 1.9s. The data were processed using a trapezoidal filter covering the first 12.5 ms of the free induction decay, followed by expo- nential multiplication equivalent to a line broadening of 5 Hz. Lactate concentrations and pHi were determined as described by Gadian et al.'

Cerebral blood flow

In those animals where CBF was recorded (n = 10) the scalp was removed and four burr holes made in the skull (two over each parietal region). Platinum elec- trodes were inserted to a depth of 1 mm in the cortex and fixed in position with cyanoacrylate adhesive. A reference silver/silver chloride electrode was placed subcutaneously in the lower flank; the platinum elec- trodes were polarized to +400 mV. CBF was measured by hydrogen clearance, after the method of Crockard et a1.: and as described by Gadian et a1.2

After being placed vertically in the magnet, the animals were switched from halothane/oxygen to halothane/air. Control CBF, spectra and blood gases were taken. The amount of oxyen in the inspired air was then progressively decreased by the proportional addition of nitrogen to the gas mixture. Each time it was lowered, CBF, blood gases (i.e., pa02, paC02 and pH) and interleaved sets of 31P and 'H spectra were acquired.

BP and ECG study

Due to the difficulty involved in measuring BP with the animal inside the NMR magnet, a separate study was performed (n = 7) under identical conditions but out- side the magnet. The animals were anaesthetized and body temperature maintained as before. An arterial catheter allowed access to blood samples and measure- ment of BP. Hypoxia was produced as previously described, and BP continuously monitored.

ECG was monitored in this study and in the NMR/CBF studies using two unipolar electrodes, one over the right atrium and the other over the apex of the left ventricle. A conventional cardiometer (Cardiorater model CR5, Cardiac Recorders Ltd., London, UK), fitted with an amplifier to detect the potentials pro- duced by the gerbil heart, was used.

Calculation of oxygen delivery

Oxygen delivery (in mL/100 g/min) was calculated as oxygen content (determined from our own oxygen saturation curve for the gerbil, and assuming a mean haemoglobin content of 12.4 g/100 mL) multiplied by the CBF. Since there was no significant change in blood pH throughout these experiments, no correction for Bohr shift was necessary (see Results).

Statistical analysis was carried out using Student's t- test for unpaired data.

RESULTS

NMR spectroscopy

Figure 1(A) shows the relationship between the P,/(phosphocreatine(PCr) + Pi) ratio (calculated from the "P NMR spectra) and pa02. The closed data points represent values calculated from those animals in which the lactate was also measured (n = lo), while the open symbols represent values from animals where only 3'P spectra were acquired. Figure 1(B) shows the concen- tration of lactate (calculated from the 'H spectra) as a function of paOz. The solid (and broken) lines in Figs 1(A) and (B) indicate the means (+2SD) of all data points above 50mm Hg (before any changes were evident in the NMR spectra). The mean P,/(PCr+P,) was 0.18 k 0.07 (2 SD) and the mean lactate concentra- tion was 1.20f0.68 (2 SD) mmol/kg. In 50% (5/10) of those animals where both 'H and 31P NMR spectra were collected, lactate increased before any changes occurred in the 3'P spectra. As an example, Fig. 2 shows the 'H and 3'P NMR spectra from one such animal where the blood pa02 was measured at 23 mm Hg. These spectra show the presence of high concentra- tions of lactate, in the absence of any significant changes in the "P spectrum. In the remaining 50% of the animals where both 'H and 31P spectra were collect- ed, changes in the proton spectra were always accom- panied by changes in the phosphorus spectra.

Changes in pH, were calculated from the chemical shift of the P, peak. The control (pre-hypoxia) pHi was 7.17 f 0.08 (SD). There was no change in the pHi until the paOz was reduced below 25mm Hg, when it fell steeply as the paOz decreased further (Fig. 1(A), inset).

CBF measurements

Data from the CBF measurements are shown in Fig. 3. In common with the spectroscopy data, the CBF remained remarkably stable (with the exception of two outlying points, both from the same animal) as the paOz was progressively reduced to 40 mm Hg. The CBF then began to increase fairly rapidly, but before any change in the energy metabolites was evident. Figure 4 shows the Pi/(PCr+Pi) ratio as a function of calculated oxy- gen delivery. It can be seen that the ratio remains low until the oxygen delivery reaches ca 4 mL/100 glmin.

BP and ECG study

BP and blood gas analysis results are shown in Fig. 5. There was no significant change in mean arterial blood pressure (MABP) until the paOz was reduced below 45mm Hg (Fig. 5(A), closed symbols). The MABP

SO

0 1.1 -

1.1 - 0 PHi '

6.0 -

K . ALLEN E T A L .

-+ ** . .& a,., ... . **:

0.9

0.8

0.7

0.6 - n- +, 0.5 0 II

0.4 >- n -

0.3

0.2

0.1

0.0

8.0 -

7.0 - - c

6.0 - Y

1

A

I "1

- 5.0 -

E a 4.0 - I

c

200 -

160 - - c I e .- E

0 I P.02 (mm Hgl

W 0 %

b

0 20 40 60 80 100 120 140

P,02 (mm Hg)

9-0 iB

2.0 .... 1

0

mm

0 . 0 1 . I . I I 1 . l . 1 I , I 1

0 20 40 60 80 100 120 140

P,02 (mm Hg) Figure 1. (A) Relationship of the P,I(PCr+P,) ratio with pa02. The data are from 17 animals. Open symbols represent data from animals where only 31P NMR spectroscopy was performed; closed symbols represent data from animals where CBF was measured and 'H spectra acquired in addition to 31P spectra (n= 10). Inset: lntracellular pH (derived from the 31P spectra) plotted as a function of pa02. (B) Relationship of lactate concentration with paOz (n= 10). These data are from the same 10 animals as depicted by the closed symbols in (A), and were acquired by alternating the 'H and 31P collections. In some cases, it was not possible to determine the P, (and therefore pH,) andlor lactate concentrations from the NMR spectra due to the inadequate S/N ratio of these signals at high paOz values. In both (A) and (B), the solid lines represent the mean of all data where the paOz was 50 m m Hg or greater; the broken lines represent 2 SD either side of the mean.

then fell sharply and significantly (p = O.OOS), and conti- nued falling steadily as the paOz was lowered further. Also shown in Fig. 5(A) are the data for paCOz, determined by blood gas analysis (open symbols). There is an obvious and steady decline in blood CO? as the animals hyperventilate.

CrIPCr

( A ) ATP

00

0 0

0

Lactate I,, I ' " ' I ' " ' I ' ~ ' ~ 1 ' ~ ' ~ I I " " 1

3.5 3.0 2.5 2.0 1.5 1.0 0.5

PPm

I ~ I ~ I ~ I

10 0 -10 -20

PPm 0 0 20 40 60 80 100 120

Figure 2. 'H and 31P spectra acquired when the pa02 was 23 m m Hg. For the 'H spectra, (A) is the spectrum obtained using sequence (A) and (B) is the spectrum obtained using sequence (B) (see text). Summing pulse sequences (A) and (B) suppresses any unwanted fat signals and reveals the presence of lactate at 1.32 ppm.

P,02 (mm Hg)

Figure 3. Relationship of CBF with pa02. The data are from 10 animals.

NMR AND CBF IN HYPOXIA 51

Figure 5(B) shows that blood pH remained very stable throughout the experiments. The small dips in the curve at paOz = 18 and 55 mm Hg are not significant (p > 0.05).

ECG was unaffected until the pa02 was reduced to ca 25-30 mm Hg, whereupon small alterations in the pat- tern were observed (data not shown). This disruption in the normal rhythm was much more dramatic when the paOz was reduced to 20-25mm Hg (25-50% satu- ration), when severe dysrhythmias, such as ventricular tachycardia, were observed.

DISCUSSION

Since NMR enables us to monitor brain energy meta- bolites non-invasively and repeatedly within a single animal, we have been able in these studies to correlate changes in cerebral energy status with a number of physiological variables as the brain is made increasingly hypoxic.

On decreasing the paOz, CBF begins to increase at a pa02 of ca 40-50 mm Hg. This observation is supported by previous work where an increase in CBF was reported when pa02 was reduced below 50mm Hg.7 The increase in CBF has been suggested by some workers to be elicited by lactic acidosis,* although others have found that changes in blood flow precede alterations in lactate and pHi .9 In our studies, the CBF increased slightly earlier than the increase in lactate in two of the animals (one of which contributed to the outlying points in Fig. 3), i.e., at a higher p a o r . Thus although the lactic acidosis certainly plays an important role in the increased CBF, it is possible other factors

I a 0.5 1 +,

0

0

* * 0.2 OS3 1 .om

o*,/ , I , ; , ;, , 0 .o 0 2 4 6 8 1 0 1 2 1 4

Oxygen delivery (rnl 100 g'' min") Figure 4. The P,/(PCr + Pi) ratio plotted as a function of oxygen delivery. Oxygen delivery was calculated as described in the text.

100 -

90 - 80 -

0 . - 70 - - 3 I

A

6 0 -

E 5 O l -1

0 s 0 10 20 30 40 50 60 70 80 90 100

P.02 (mm Hg) 'I 7.6

7.4

m

7.2 4

0 10 20 30 40 50 60 70 80 90 100

P,02 (mm Hg) Figure 5. MABP (closed symbols k SEM), paCO, (open symbols- * SEMI and blood pH (+SEMI derived from blood gas analysis plotted in relation to pa02. Each point on the abscissa is the mean +SEM of measurements lying in the range 11-20,21-30, 31-40 mm Hg, etc. These measurements were made in a separate group of seven animals. The numbers in parentheses represent the number of data points contributing to each mean.

may be operating. For example, hypoxia, particularly systemic hypoxia, is a significant stress factor, leading to increased systemic catecholamine production. This affects the heart by increasing heart rate and stroke work'" and increasing cardiac output. The BP is nor- mally raised, as a result of a combination of increased peripheral resistance and increased cardiac output. The increase in BP observed as an agonal event in some of these animals may be due to increased catecholamine production.

The NMR results show that the energy metabolism of the brain is resistant to a lowering of the arterial paOz. The brain clearly has a considerable metabolic reserve, which sustains it until paOz approaches 40 mm Hg. At this point adaptation occurs, with an increase in CBF which enables the brain to function until p a 0 2 approaches 20 mm Hg.

In 50% of the animals in this study, the increase in lactate that we observed occurred earlier (i.e., at a higher paOz) than the decline in high energy phosphates or pHi . It is generally assumed that the principal factor contributing to intracellular acidosis in the brain in hypoxia (and ischaemia) is the generation of lactic acid by anaerobic glycolysis," and there is much evidence that lactic acid formation is an important factor in causing neuronal damage."-13 Glucose from the circula- tion is the main substrate for anaerobic glycolysis in the

52 K. ALLEN ETAL.

non-starving animal, and since systemic glucose meta- bolism is affected by circulating catecholamines, there is a possibility that the early rise in lactate is of systemic origin, occurring before the 3'P cerebral metabolites are affected. However, we consider this is unlikely since the blood-brain barrier is relatively impermeable to circulating 1a~ta te . l~ Also, the blood pH remained unaffected even when the p ,Oz was below 20 mm Hg. In addition, raising the concentration of lactic acid in blood causes little apparent uptake into the brain and CSF. " Therefore we conclude that the lactate observed in our studies is generated intracellularly by anaerobic glycolysis. In addition, it is unlikely that the lactate signal that we observe in the absence of any changes in the phosphorus spectra is originating from a more vulnerable region of the brain than is being detected by the phosphorus surface coil. The combination of a 'H surface coil that was smaller than the corresponding 31P surface coil, together with the use of a spin-echo sequence, would ensure that the effective tissue volume interrogated during 'H NMR spectroscopy was some- what smaller than, and included within, that interro- gated during 31P spectroscopy. Our observation that the lactate can increase while the "P spectra remain unchanged therefore suggests that glycolysis can be controlled independently of the concentrations of the phosphorus metabolites, and also that there is a mecha- nism, possibly Na+/H+ exchange, for maintaining normal intracellular pH." It also suggests that 'H NMR spectroscopy may be more sensitive than 3'P at detect- ing hypoxia-induced changes in the brain.

At pa02 values of 20-25 mm Hg, energy metabolism became impaired, as indicated by the increase in the ratio P,/(PCr+P,). The rate of oxygen delivery under these conditions is ca 4 mL/100 g/min, which is similar

to that at which energy failure occurs in ischaemia.' This suggests that the critical factor in ischaemia is delivery of oxygen rather than, for example, removal of waste products such as lactate.

The rapid CBF changes during ventricular tachycar- dia demonstrate the labile situation during stress, and illustrate that in such situations steady-state flow con- ditions may not prevail, a point of importance in some clinical studies. Some of the flow changes noted in Fig. 3 may be due to the above factors, rather than the accumulation of lactate. In fact, in our studies we consider that the death of the animal is due not to the effects of hypoxia on the brain but rather on the myocardium, as reflected by the dramatic changes in the ECG pattern. This has important clinical impli- cations, because if the heart could be protected from hypoxia, the brain might tolerate an even lower paOz.

CONCLUSIONS

The brain is very resistance to hypoxia, and energy metabolism is not adversely affected until paOz is reduced below 40mm Hg. The fact that half the ani- mals showed an increse in lactate prior to any change in phosphorus metabolites suggests that 'H NMR may be more sensitive than 31P NMR at detecting hypoxic damage. The results suggest that reduction in oxygen supply rather than the accumulation of toxic metabo- lites may be the critical factor in ischaemia.

Acknowledgements We thank Action Research, the Rank Foundation, Picker International and the Medical Research Council for their support. We are grateful to Dr E. Proctor for his help and advice.

REFERENCES

1. Crockard, H. A., Gadian, D. G., Frackowiak, R. S. J., Proctor,

2.

3.

4.

5.

6.

7.

E., Allen, K., Williams, S. R. and Ross Russell, R. W. Acute cerebral ischaemia: concurrent changes in cerebral blood flow, energy metabolites, pH, and lactate measured with hydrogen clearance and 31P and 'H NMR spectroscopy. II. Changes during ischaemia. J. Cereb. Blood Now Metab. 7,

Gadian, D. G., Frackowiak, R. S. J., Crockard, H. A., Proctor, E., Allen, K., Williams, S. R. and Ross Russell, R. W. Acute cerebral ischaemia: concurrent changes in cerebral blood flow, energy metabolites, pH, and lactate measured with hydrogen clearance and "P and 'H NMR spectroscopy. 1. Methodology. J. Cereb. Blood Flow Metab. 7, 199-206 (1987). Allen, K., Busza, A. L., Crockard, H. A., Frackowiak, R. S. J., Gadian, D. G., Proctor, E., Ross Russell, R. W. and Williams, S. R. Acute cerebral ischaemia: concurrent changes in cerebral blood flow, energy metabolites, pH and lactate measured with hydrogen clearance and 31P and 'H nuclear magnetic resonance spectroscopy. Ill. Changes following ischaemia. J. Cereb. Blood Flow Metab. 8, 816-821 (1988). Crockard, H. A., lannotti, F., Hunstock, A. T., Smith, R. D., Harris, R. J. and Symon, L. Cerebral blood flow and edema following carotid occlusion in the gerbil. Stroke 11,494-498 (1 980). Campbell, I. D., Dobson, C. M., Williams, R. J. P. and Xavier, A. V. Resolution enhancement of protein NMR spectra using the difference between a broadened and a normal spec- trum. J. Magn. Reson. 11, 172-181 (1973). Williams, S. R., Gadian, D. G. and Proctor, E. A method for lactate detection in vivo by spectral editing without the need for double irradiation. J. Magn. Reson. 66, 562-567 (1986). Kogure, K., Scheinberg, O., Reinmuth, M., Fujishima, M. and

394-402 ( 1 987).

Busto, R. Mechanisms of cerebral vasodilation in hypoxia. J.

8. Johanssonn, H. and Siesjo, B. K. Cerebral blood flow and oxygen consumption in the rat in hypoxic hypoxia. Acta Physiol. Scand. 939, 269-276 (1975).

9. Nilsson, B., Norberg. K., Nordstrom, C.-H. and Siejso, B. K. Influence of hypoxia and hypercapnia on CBF in rats, in International Symposium on Cerebral Blood Flow and Metabolism, ed by M. Harper, B. Jennett, D. Miller and J. Rowan, pp. 19-23. Churchill Livingstone, Edinburgh (1975).

10. Ganong, W. F. The autonomic nervous system, in Review of Medical Physiology, pp. 183-188. Appleton & Lange, Norwalk (1987).

11. Siesjo, B. K. Cerebral circulation and metabolism. J. Neurosurg. 60, 883-908 (1984).

12. Myers, R. E. and Yagamuchi, S. Nervous system effects of cardiac arrest in monkeys. Arch. Neurol. 34, 65-74 (1977).

13. Rehncrona, S. G., Rosen, 1. and Siesjo, B. K. Excessive cellular acidosis: an important mechanism of neuronal damage in the brain. Acta Physiol. Scand. 110, 435-437 ( 1980).

14. Hawkins, R. Cerebral energy metabolism, in Cerebral Energy Metabolism and Metabolic Encephalopathy, Ed. by D. W. McCandless, pp. 3-17. Plenum Press, New York (1985).

15. Weyne, J. and Leusen, I. Lactate and CSF in relation to brain and blood, in Fluid Environment of the Brain, Ed. by H. F. Cserr, J. D. Fenstermacher and V. Fenci, pp. 255-276. Academic Press. New York (7975)

16. Mahnensmith, R. L. and Aronson, P. S. The plasma mem- brane sodium-hydrogen exchanger and its role in physiolo- gical and pathophysiological processes. Circ. Res. 56, 773- 788 (1985).

Appl. Physiol. 29, 223-229 (1970).