Differential effects of hypertonic mannitol and glycerol on rat brain metabolism and amino acids

Brain Research, 225 (1981) 143-153

143

Elsevier/North-Holland Biomedical Press

D I F F E R E N T I A L EFFECTS OF H Y P E R T O N I C M A N N I T O L A N D G L Y C E R O L ON RAT BRAIN METABOLISM A N D A M I N O ACIDS

PAK HOO CHAN, EVAN POLLACK and ROBERT A. FISHMAN Brain Edema Clinical Research Center, Department of Neurology, University of California, School of Medicine, San Francisco, Calif. 94143 (U.S.A.)

(Accepted April 2nd, 1981) Key words: hyperosmolality - - brain amino acids - - mannitol - - glycerol - - brain metabolism

SUMMARY Intraperitoneal injections in rats of two different dosages of hypertonic solutions containing mannitol or glycerol caused complex and differential changes in brain amino acids. When plasma osmolalities were elevated to toxic levels of 397-432 mOsm/kg H20, brain sodium was increased, whereas plasma sodium was decreased. Brain potassium was not affected. Brain water decreased significantly, concomitant with elevation of plasma osmolality. Both brain lactic acid and [125I]albumin space rose significantly. Brain amino acids (mostly aliphatic and basic amino acids) as well as GABA and glycine (putative inhibitory neurotransmitters) increased after both mannitol and glycerol. Ammonia was stimulated by mannitol but was unaffected by glycerol. Plasma amino acids, which generally increased after mannitol, were decreased by glycerol. When the plasma osmolalities were elevated only to moderate levels (about 350 mOsm/kg HzO), only glycerol induced a significant increase in brain taurine, aspartic acid, alanine, leucine and lysine. Thus, with moderate hyperosmolality, glycerol has striking effects on brain amino acid metabolism that are not observed with mannitol.

INTRODUCTION Hyperosmolality may be responsible for a severe encephalopathy with coma and focal or generalized seizures. The level of plasma hyperosmolality at which these occur is somewhat variable; however, severe encephalopathy is almost always seen in patients with plasma osmolalities of about 380 mOsm/kg H20 or greater 1. Lesser degrees of hyperosmolality (about 330 mOsm/kg H20) obtained with the intravenous administration of hypertonic mannitol or glycerol, are used therapeutically to acutely

144 reduce brain volume and intracranial pressure. However, the pathophysiology of hyperosmolal encephalopathy is not well understood, nor is it clear whether the therapeutic efficacy of various agents varies with different solutes such as mannitol or glycerol. We demonstrated previously that the elevation of plasma osmolalities to toxic levels (392-446 mOsm/kg H20) by intraperitoneal injection into rats of hypertonic solutions of sodium chloride, sucrose, and glucose caused coma with complex changes in brain metabolism 5. Brain electrolytes, lactic acid, ammonia and amino acids were increased. GABA, glycine, taurine, aspartic acid, glutamine, alanine, threonine and lysine were the most affected amino acids, indicating that the putative neurotransmitter amino acids and their precursors were particularly involved as the brain responded to a hyperosmolal environment. No important differences between the effects of hypertonic sodium chloride, sucrose or glucose were detected. Mannitol and glycerol, the solutes most commonly used in the clinical treatment of brain edema8,13 are employed in the present studies to compare their effects on rat brain amino acid metabolism to the previously published data obtained with hypertonic sucrose, glucose and sodium chloride. The effects of both toxic and more moderate degrees of hyperosmolality are also compared in these experiments. METHODS In vivo models of acute hyperosmolality Sprague-Dawley male rats (purchased from Simonsen, Gilroy, Calif.) weighing 120-150 g, were given 4 i.p. injections at 15 rain intervals. Each injection contained 3.56 ml of 2.25 M (8 mmol) of mannitol, or 2 ml of 4 M (8 mmol) glycerol per 100 g of body weight (the solutions are equiosmolal). These toxic osmotic dosages were the same as previously used 5. Isotonic saline (3.56 ml/100 g body weight, 4 injections) was used for control studies. For the moderate dosage studies, animals received only one injection of either hypertonic solution or the same volume of 0.9 ~ isotonic saline. Animals were decapitated 60 min after the final injection. Decapitated brains were immediately dropped in liquid N2 for chemical analysis. It took 3 sec to freeze the brains in liquid Nz. Blood samples, collected in heparin-coated centrifuge tubes, were centrifuged immediately at 10,000 rpm for 20 rain for plasma preparation. The present studies have compared toxic and moderate levels of hyperosmolality obtained with mannitol or glucose. The moderate levels (about 350 mOsm/kg H20) obtained in the rat are somewhat higher than the therapeutic levels obtained in man, which are usually about 330 mOsm/kg H20 20. With this caveat, these experimental models of toxic and moderate hyperosmolality have analogies with hyperosmolality as seen in human hyperosmolal coma and with osmotherapy respectively. Determination of brain and plasma osmolality, brain water, Na +, K + and lactic acid Methods for determination of brain and plasma osmolality, brain water, Na ÷, and K ÷ and lactate were described previously 5. Brain and plasma osmolalities were determined by measuring the freezing point depression of the respective tissues with an

145 advanced osmometer (Advanced Instruments, Needham Height, Mass.). Plasma osmolality was measured directly. Brain solutes were extracted with boiling distilled water according to the method of Arieff et al. 2. Tissue water content was determined by dessication at 105 °C for 16 h. Brain sodium and potassium were first extracted with 2 N nitric acid for 24 h, and were analyzed with an Eppendorf flame photometer (Brinkman, Westbury, N.Y.). Lactic acid was measured enzymatically. The assay consisted of an aliquot of 0.25 ml of brain tissue homogenate, 13.0 units of lactate dehydrogenase, 1.3 mg of fl-NAD in glycine-hydrazine buffer, pH 9.2 with a final volume of 1.5 ml (enzyme and reagents were obtained from Sigma, St. Louis, Mo.).

Determination of [1251Jbovine serum albumin (BSA) space [125I]BSA (0.828 mCi/mg, 99 % purity, purchased from New England Nuclear, Boston, Mass.) was dialyzed against 50 mM phosphate buffer (pH 7.4) overnight to eliminate the free iodine. Rats were injected with 1/tCi of [125I]BSA i.p. 24 h prior to the experimental manipulation18. After the serial injections, control and experimental animals were decapitated and the hemispheres and blood plasma were counted in a Packard gamma counter. The per cent [125I]BSA space was expressed as: cpm per g of brain cpm per ml of plasma

× 100

The 24 h BSA space is used as a measure of the permeability of brain capillary endothelial cells and of brain extracellular fluid volume.

Amino acid analysis Frozen rat brains were pulverized 10 times with a tissue pulverizer and placed in an acetone/dry-ice bath. About 1 g of the pulverized frozen tissue was homogenized with an ice-cold medium containing 3 N perchloric acid with a tissue-to-medium ratio of 1:10 (v/w) for 1 min. The samples were then centrifuged at 10,000 rpm for 10 min. The supernatant was dried in vacuo, redissolved in double-distilled water and then dried again in vacuo, using Savant Speed Vac concentrator (Savant Instruments, Hicksville, N.Y.). These procedures were repeated at least two times to remove any trace of protein precipitate. The extracted amino acids were dissolved in 0.2 N lithium citrate buffer, pH 2.2 to obtain a concentration of 0.2 g/ml buffer. About 0.1-0.2 ml samples were applied to a Beckman 119 CL amino acid analyzer for physiological fluid analysis. This amino acid preparation method was slightly different from the previously published method5. The present method has achieved a much better resolution between glutamic acid and glutamine and a higher percentage of recovery for phosphoserine, phosphoethanolamine and taurine. The amino acid analyzer was equipped with a single column containing W-3P resin with a buffer and ninhydrin flow rate of 30.5 ml/h. The amino acid calibration standards (obtained from Hamilton, Reno, N.Y.) with the addition of freshly prepared glutamine and tryptophan (2.5 mmol/ml) were used for calibration. The data were processed automatically by a Beckman Model 126 data system. Plasma was deproteinized with an equal volume of 3 % sulfosalicyclic acid (Harleco, Gibbstown, N.J.) and was centrifuged at 10,000 rpm

146 for 10 minutes. The samples were dried in vacuo and then dissolved in lithium citrate buffer to obtain a concentration of 1 ml plasma/1 ml buffer. About 0.1-0.2 ml samples were applied to the amino acid analyzer. For the measurement of GABA in control plasma, about five times more plasma was used for analysis. RESULTS Effects o f hyperosmolal solutions on brain and plasrna osmolalities, brain HzO, N a +, K + and lactic acid content Sixty minutes after the i.p. injection of hyperosmolal mannitol or glycerol, the animals were stuporous or prostrate; some had twitching of their limbs, especially the glycerol-treated animals. Table I shows that both plasma and brain osmolalities were increased in mannitol and glycerol-treated animals. These data indicate a mean increase respectively of 99 and 134 mOsm/kg H~O in plasma osmolality concomitant with an increase of 92 and 112 mOsm/kg H20 in brain osmolality. Similarly, brain H20 was reduced by 2 . 2 5 ~ for mannitol and by 2 . 7 7 ~ for glycerol, indicating hypertonic glycerol caused a somewhat greater degree of brain dehydration in response to the higher plasma osmolality obtained with glycerol. Brain Na + was slightly increased by 4 ~ for mannitol- and 6 ~ for glycerol-treated animals. However, brain K + content was not affected by either solute. Plasma Na + was decreased by both mannitol and glycerol. (The hyponatremia reflects in part the movement of sodium into the peritoneal fluid.) These data indicate the opposing effects of hypertonic agents

TABLE I Effects of toxic levels of acute hyperosmolality on plasma osmolality and brain osmolality, Na +, K + and lactate content

Mean S.E.M. are given. Number of animals used in parenthesis. Control

Plasma osmolality (mOsm/kg HzO) Brain osmolality (mOsm/kg tissue H20) Brain H20(~) (mEq/kg dry wt.) Brain Na + (mEq/kg dry wt.) Brain K + (mEq/kg dry wt.) Plasma Na + (mEq/liter) Brain lactate (mmol/kg wet wt.)

Mannitol

Glycerol

297 ± 0.7 (12)

397 ± 2 (12)*

432 4- 6 (t3)*

314 4- 3 (13)

406 4- 3 (10)*

426 4- 6 03)*

79.684- 0.18 (13)

77.43 4- 0.17 (12)*

244 -~ 2(10)

253 ± 4 (10)**

259 :~ 2(7)*

515 4- 10(14)

520 4- 7(10)

524 4- 10(11)

143 4- 1 (8)

126 ± 2 (7)*

137 ± 2 (5)**

2.78 ± 0.19 (10)

3.89 ± 0.15 (8)*

76.9l 4- 0.22 (13)*

6.82 4- 0.70 (13)*

*P < 0.001, **P < 0.05, using Student's t-test for statistical analysis, comparing each experimental group to control. See Methods for dosages used.

147 TABLE II

Effects of toxic levels of acute hyperosmolality on blood-brain barrier permeability Mean ± S.E.M. are given; number of animals used given in parenthesis.

% [125I]Bovine serum albumin space Control (7) Mannitol (5) Glycerol (6)

1.36 ± 0.06 7.82 4- 1.18" 4.94 ± 0.42*

*P < 0.001, using Student's t-test for statistical analysis.

TABLE III

Effects of toxic levels of acute hyperosmolality on free amino acid contents of rat plasma Mean ± S.E.M. is given for control and experimental plasma amino acid concentrations. Number of animals used is given in parenthesis.

Free amino acids (ttmol/liter)

Control (12)

Mannitol (10)

Glycerol (10)

Phosphoserine Taurine Hydroxyproline Aspartic acid Threonine Serine Glutamine Proline Glutamic acid Citrulline Glycine Alanine Valine Cysteine Methionine Isoleucine Leucine Tyrosine Phenylalanine GABA Ornithine Ammonia Lysine Histidine Tryptophan Arginine

6 244 40 35 282 293 952 185 106 26 314 459 149 10 40 82 132 82 65 0.26 56 163 588 80 64 208

16 339 48 34 287 340 690 145 135 52 441 624 197 30 44 103 184 89 97 6 62 219 670 104 76 180

6 ± 0.7 252 ± 32 (a) 54 ± 4* 258 4- 26 201 4- 17" 265 ± 35* 172 4- 14 104 ± 9 29 ± 3 289 ± 9 378 ± 27 106 ± 14 13 4- 2 18 ± 2* 58 ± 7** 89 ± 9* 64 ± 5*** 64 ± 4 5 4- 0.7* 218 ± 23* 70 ± 5* 464 4- 38 65 ± 6 51 ± 35* 72 ± 13"

+ ± ± 44± ± 4± 4444± 4± ± ± 444± ± ± ± 4-

1 25 4 3 25 17 98 27 7 3 5 37 13 1 3 5 9 7 4 0.04 6 6 31 5 5 18

4- 1 * ± 37 4- 4 4- 4 ± 38 4- 44 ± 66* ± 34 ± 17 ± 3" ± 55** 4- 54** ± 13" 4- 4* 4- 2 ± 5* 4- 7* ± 6 4- 4* ± 0.9* ± 6 ± 20*** ± 31 ± 14 4- 6 4- 33

*P < 0.001, **P < 0.02, ***P < 0.05, using Student's t-test for statistical analysis, comparing each experimental group to control. (a) a significant extra peak was co-eluted with bydroxyproline. This new amino acid has not yet been identified.

148 TABLE IV

Effects of toxic levels of hyperosmolality onfree amino acid levels of rat brain Control and experimental brain tissues mean 4- S.E.M. are given; number of animals are presented in parenthesis.

Free amino acids (l~mol/lO0 g brain tissue)

Control (12)

Mannitol (7)

Glycerol (8)

Phosphoserine Phosphoethanolamine Taurine Aspartic acid Threonine Serine Asparagine Glutamine Proline Glutamic acid Citrulline Glycine Alanine Valine Cysteine Cystathionine Methionine lsoleucine Leucine Tyrosine Phenylalanine Ornithine GABA Ammonia Lysine Histidine Tryptophan Arginine

132 ± 10

122 ± 24

135 ± 20

528 780 191 49 61 10 150 7 1010 3 150 52 7 4 4 3 4 5 6 4 18 150 31 20 3 2 8

± 68 ± 58 ± 17 ± 4 4- 3 ± 1 ± l0 ± 0.5 ± 23 4- 0.4 4- 8 4- 3 zk 0.5 4- 0.5 -- 0.5 ± 0.2 4- 0.4 ± 0.3 4- 0.5 -k 0.4 ± 2 ± 9 ± 2 ± 1 4- 0.2 ± 0.2 4- 0.2

910 1147 213 64 67 12 168 10 994 4 226 119 10 7 6 2 6 9 7 6 24 210 68 31 4 2 12

± 125 4- 233 ± 16 4- 7 :k- 7 ± 2 ± 10 4- 1"* 4- 62 4- 0.4 ± 19" 4- 8* 4- 0.5* 4- 1"** 4- 1 4- 0.5 4- 0.6** 4- 0.5* 4- 0.5 4- 0.2* :k 2 4- 8* 4- 9* zk 4* 4- 0.5 4- 0.2 4- 0.5*

624 ± 78 1616 ± 167" 250 4- 6* 61 4- 5 63 ± 5 13 ± 1 171 ± 19 10 zk 1"* 975 ± 47 3 ± 0.2 195 4- 17"* 137 ~- 15" 11 4- 2* 6 4- 1 4 4- 0.5 4 ± 0.3 4 ± 0.6 6± 1 6± 1 5 ± 0.5*** 17 -4- 2 201 4- 18"* 26 5_ 2 32 ± 3* 3 4- 0.4 2 4- 0.2 19 ± 1 *

*P < 0.001, **P < 0.02, ***P < 0.05, using Student's t-test for statistical analysis, comparing each experimental group to control. o n t h e b r a i n a n d p l a s m a N a + c o n t e n t . F u r t h e r m o r e , lactic a c i d levels w e r e significantly i n c r e a s e d by b o t h h y p e r t o n i c solutions. H o w e v e r , h y p e r t o n i c g l y c e r o l i n c r e a s e d lactic a c i d by 145 %, w h e r e a s m a n n i t o l i n c r e a s e d lactic a c i d by o n l y 40 %.

Effects o f hyperosmolality on blood-brain barrier permeability to [12~I]BSA T h e c h a n g e s in b l o o d - b r a i n b a r r i e r (BBB) p e r m e a b i l i t y to [125I]BSA i n d u c e d by t o x i c doses o f m a n n i t o l a n d g l y c e r o l are s h o w n in T a b l e II. B o t h h y p e r t o n i c m a n n i t o l a n d g l y c e r o l i n d u c e d a m a j o r i n c r e a s e in 125I-BSA space. H y p e r t o n i c m a n n i t o l i n d u c e d a five-fold i n c r e a s e w h e r e a s g l y c e r o l h a d a lesser effect o n 125I-BSA space, c a u s i n g o n l y a 263 % increase. T h e s e d a t a i n d i c a t e t h a t e q u i m o l a r d o s e s o f h y p e r t o n i c

149 glycerol induced a smaller degree of BBB permeability change than mannitol, although the glycerol-treated animals had an even higher plasma osmolality. Effects o f hyperosmolality on plasma amino acids When plasma osmolalities were elevated to 397-432 mOsm/kg HzO by hypertonic solutions of mannitol or glycerol, complex and divergent changes were observed in the free amino acids of plasma as shown in Table III. Hypertonic mannitol, like sodium chloride, sucrose and glucose, caused a generalized increase in the plasma amino acid content. Aliphatic branched amino acids, glycine, alanine, valine, isoleucine and leucine were stimulated by 2 0 - 4 0 ~ . Other amino acids, including GABA, taurine, phosphoserine, cysteine and phenylalanine, were also increased. A m m o n i a was enhanced by 34 %, whereas glutamine was decreased by 27 %. On the other hand, hypertonic glycerol generally decreased the plasma free amino acids. Serine, glutamine, methionine, isoleucine, leucine, tyrosine, ammonia and arginine were greatly reduced. A m o n g these, ammonia, arginine and glutamine were reduced 57 %, 63 % and 72 % respectively. However, G A B A and ornithine were significantly increased in plasma. Effects o f toxic levels o f hyperosmolality on brain amino acids When plasma osmolality was raised to 392 and 432 mOsm/kg H 2 0 respectively by hypertonic solutions of mannitol and glycerol, a general increment in the free amino acids of brain was observed. Table IV shows that both mannitol and glycerol induced a significant increase in proline (43 %), alanine (129 and 165 %), valine (43 and 57 %), phenylalanine (25 and 50%), lysine (55 and 60%) and arginine (50 and 138%). TABLE V Effects of moderate levelsof hyperosmolality on major brain amino acids Mean d: is given for control and experimental brain amino acids. Each group consisted of 6 animals. Amino acid

Mannitol Glycerol ( ~ of control)

Taurine Aspartic acid Threonine Serine Glutamine Glutamic acid Glycine Alanine Leucine GABA Ammonia Lysine

95 ± 9 86 i 10 95 d- 14 77 d- 7 123 ± 15 87 4. 5 105 + 10 99 4- 13 120 ± 40 85 ± 8 80 i: 6 112+ 12

137 + 4** 216 -- 24* 122 ± 16 120 4- 13 109 4. 9 106 ± 8 128 4. 7 139 4. 7* 180 ± 10" 1104- 8 126 + 7 184± 8"

*P < 0.001, **P < 0.02, using Student's t-test for statistical analysis, comparing each experimental group to control.

150 Hypertonic mannitol and glycerol also increased GABA and glycine, the putative neurotransmitters, by 34 and 40 ~ and by 30 and 51 ~ respectively. On the other hand, phosphoethanolamine, cysteine, isoleucine and leucine were increased by 72 ~, 75 ~, 50 ~ and 80 ~, respectively by mannitol. Glycerol did not have an effect on the latter amino acids. Ammonia levels were increased two-fold by mannitol, but were unaffected by glycerol. It is noteworthy that glutamine levels were not affected by either solute. Aspartic acid and taurine, the putative neurotransmitter amino acids, were increased by hypertonic glycerol by 31 ~ and 107 ~ respectively, but mannitol had no such effect.

Effects of moderate levels of hyperosmolafity on brain amino acids In addition to the toxic levels of osmolalities that were used, lower levels of hyperosmolality (i.e. plasma osmolalities of about 350 mOsm/kg H20) closer to that obtained with osmotherapy were used to study their effects on brain amino acids. Animals received only one injection of either mannitol or glycerol (8 mmol/100 g body weight) and were killed 1 h later. At this time, the animals were alert and able to move freely. The plasma osmolalities rose from 294 ± 4 (n ~ 6; ± S.E.M.) to 344 A- 5 (n : 6) for mannitol and to 346 % 5 (n : 6) for glycerol, respectively. The brain osmolalities were 358 4- 3 (n : 6) and 358 ~ 7 (n ~ 6) and the H20 content was 77.8 ± 0.3 % and 78.3 ~ 0.3 %, respectively, for mannitol- or glycerol-treated animals. Table V shows that moderate levels of hyperosmolality obtained with mannitol, unlike the toxic levels, did not cause any change in free amino acid levels in rat brain. However, equiosmolal glycerol caused a significant increase in alanine (39 %), leucine (80 %) and lysine (84%). Moreover, the putative neurotransmitter amino acids, taurine and aspartic acid, were increased 37 % and 116/o°/ respectively. The major amino acids, including the putative neurotransmitters glutamic acid, GABA and glycine, were not affected by glycerol. DISCUSSION Our previous studies of the metabolic effects in brain of acute plasma hyperosmolality obtained with sodium chloride, sucrose or glucose, showed the induction of complex changes in brain amino acid metabolism associated with the formation of 'idiogenic osmoles '5. The latter indicate the presence in brain of a mechanism which increases intracellular osmolality to protect against a volume change. The production of idiogenic osmoles may be derived from the dissociation of lipid-bound potassium as suggested by Katzman and Leiderman15. The increase in amino acid concentration observed was insufficient to account for most of the idiogenic osmoles, but might contribute to the mechanism of hyperosmolal encephalopathy. In the present study, GABA and glycine (inhibitory neurotransmitters), alanine, valine, leucine, isoleucine (aliphatic branched amino acids) and ammonia in both brain and plasma, were significantly increased by hypertonic mannitol. The increases of these amino acids in both plasma and brain were very similar to the previously reported effects induced by hyperosmolal sodium, sucrose or glucose. Hypertonic

151 mannitol caused a significant increase in plasma taurine, whereas brain taurine was not affected. This change was also seen with sucrose and glucose. Furthermore, the brain levels of phosphoethanolamine, proline, phenylalanine, ornithine, lysine and arginine were increased in brain without the concomitant increment in plasma. On the other hand, hypertonic glycerol, like sodium chloride and sugars, enhanced brain alanine and valine, the non-essential amino acids. The neurotransmitters aspartic acid, GABA, taurine and glycine, were also stimulated by toxic levels of glycerol. However, unlike other hypertonic solutions, glycerol caused a decrease in most plasma free amino acids, with the exception of aspartic acid and GABA, which were increased by glycerol. When the plasma osmolalities were moderately elevated to about 350 mOsm/kg H20 with mannitol or glycerol, (closer to therapeutic levels used for the treatment of cerebral edemaS, 10,14) differential effects of these two agents on brain amino acids were observed. Hypertonic mannitol did not affect the free amino acid levels in either brain or plasma. However, hypertonic glycerol caused an increase in brain of aspartic acid, taurine, alanine, leucine, and lysine. These data indicate that the levels of 3 major groups of amino acids, including aliphatic, basic and neurotransmitter, were stimulated by glycerol. It has been shown that brain cells can utilize glycerol as a metabolic substrate in vivoz2. The role of glycerol oxidation in the differential response of the various amino acids requires further study. These studies have shown that only toxic doses of mannitol caused a 2-fold increase in brain ammonia, unlike glycerol which had no effect. In addition, glutamine levels were unaffected. These data differ from our earlier work with glucose, sucrose and sodium chloride, where both ammonia and glutamine levels were increased. This led us to suggest that hyperosmolal encephalopathy might be related to intracellular ammonia intoxication. This hypothesis is negated by the current studies. The higher brain ammonia levels noted in the earlier work may have been due to less rapid tissue freezing. In the present studies, rapid brain freezing was achieved in less than 3 sec. The mechanisms involved in the differential increases of brain amino acids induced by both toxic and moderate levels of hypertonic mannitol compared to glycerol are intriguing and complicated. The present studies demonstrate that toxic doses of hypertonic solutions cause an increase of the brain albumin space (Table II) at 24 h. This is considered to reflect both an increase in the permeability of brain capillary endothelial cells which comprise the blood-brain barrier to macromolecules as well as an increase in brain extracellular fluid volume. It is not possible to separate the relative importance of these two effects. The disruption of BBB may allow the increased diffusion of plasma amino acids, perhaps through the openings of tight junctions of brain capillary endothelial cells19,25 and/or by increased micropinocytosis12. The transport of plasma amino acids to the brain is limited by a stereospecific carriermediated system17. However, alterations in the carrier mechanism do not adequately account for the increment of brain amino acids, since glutamine, threonine and serine, the major amino acids in plasma (Table III), were not increased significantly in brain after either hypertonic solutions. GABA, on the other hand, which constitutes only a minute amount (0.25 #mol/liter) of plasma amino acids, was increased profoundly in

152 dehydrated brain tissues which reflects changes in brain metabolism. Thus, the changes in brain amino acids induced by hyperosmolality appear to be a result of changes in cerebral metabolism and possibly of aminotransferase activity. In addition, both toxic and moderate doses of glycerol induced a significant increase in cerebral taurine content. The observation that the increase in taurine was among the largest of the changes in cerebral amino acids is of special interest, in light of the proposal that taurine plays an important role in osmotic regulation in mammalian brain 2z. A second possible mechanism might be that the increment of brain amino acids was derived from an endogenous source. In studies of brain cortex in vitro, it has been demonstrated that brain cells produce and accumulate high concentrations of amino acids, particularly neurotransmitters, in response to an hyperosmotic environment 9, 11. Shank and Baxter have shown in the toad that an hyperosmolal environment increases brain amino acids 21. Lockwood 16 in a more chronic model of sodium chloride-induced hyperosmolality reported a two-fold increase in brain glutamine after 48 h, but the source of the glutamine was not known. It was demonstrated earlier that hyperosmolality affects glial cell function and neurotransmitter metabolism3,4, 6. Furthermore, the stress of hyperosmolality on brain cells might cause the breakdown of endogenous proteins or glutamyl and aspartyl peptides. These processes might cause the increase of certain amino acids. A third possible mechanism for the increment of brain amino acids is that the transport mechanism for the removal or entry of brain amino acids by the choroid plexus was altered by hypertonicity, DiMattio et al. 7 and Hochwald et al. la have shown in cat that acute serum hyperosmolality obtained with sucrose or glucose reduced CSF formation by the choroid plexus. The plexus has specific bidirectional transport systems for various amino acids 24. Whether hyperosmolality has specific effects on amino acid transport across the choroid plexus or brain capillary endothelial cells requires further study. However, the possibility that a combination of the above 3 mechanisms might be involved in the increment of brain amino acids cannot be excluded. There are limited implications of this work for clinicians using either hypertonic mannitol or glycerol for the reduction of intracranial pressures. It is clear that the effects of hyperosmolality on the brain vary with different solutions. Equiosmolal doses of hypertonic solutions have different effects on brain metabolism in the rat, and it appears likely that the human brain would also respond differently to a metabolized solute like glycerol compared to metabolically inert mannitol. The complex changes in brain amino acid levels, particularly in the neurotransmitter amino acids, induced with acute hyperosmolality may be relevant to the changes in brain excitability such as seizures, myoclonus and depression of consciousness, that characterize patients with hyperosmolal encephalopathy. ACKNOWLEDGEMENT This work was supported by N.I.H. Grant NS-14543.

153 REFERENCES 1 Arieff, A. I., Guisado, R. and Lazarowitz, V. C., Pathophysiology of hyperosmolar states. In T. E. Andreoli, J. J. Grantham and F. C. Rector, Jr. (Eds.), Disturbances in Body Fluid Osmolality, American Physiological Society, Bethesda, Md., 1977, pp. 227-250. 2 Arieff, A. I., Kleeman, C. R., Keushkerian, A. and Bogdoyan, H., Brain tissue osmolality: method of determination and variations in hyper- en hypo-osmolar states, J. Lab. clin. Med., 79 (1972) 334-343. 3 Chan, P. H. and Fishman, R. A., Effects of hyperosmolality on release of neurotransmitter amino acids from rat brain slices, J. Neurochem., 29 (1977) 179-181. 4 Chan, P. H., Wong, Y. P. and Fishman, R. A., Hyperosmolality induced GABA release from rat brain slices: studies of calcium dependency and sources of release, J. Neurochem., 30 (1978) 1363-1368. 5 Chan, P. H. and Fishman, R. A., Elevation of rat brain amino acids, ammonia and idiogenic osmoles induced by hyperosmolality, Brain Research, 161 (1979) 293-301. 6 Chan, P. H., Fishman, R. A., Lee, J. and Candelise, L., Effects of excitatory neurotransmitter amino acids on swelling of rat brain cortical slices, J. Neurochem., 33 (1979) 1309-1315. 7 DiMattio, J., Hochwald, G. M., Malhan, C. and Wald, A., Effects of changes in serum osmolality on bulk flow of fluid into cerebral ventricles and on brain water content, Pfliigers Arch., 359 (1975) 253-264. 8 Feig, P. U. and McCurdy, D. K., The hypertonic state, New EngL J. Med., 297 (1977) 1444 1454. 9 Fishman, R. A., Cell volume, pumps and neurologic function: brain's adaptation to osmotic stress, Res. Publ. Ass. nerv. ment. Dis., 53 (1974) 159-172. 10 Fishman, R. A., Brain edema. In R. A. Fishman (Ed.), Cerebrospinal Fluid in Diseases of the Nervous System, W. B. Saunders, Philadelphia, 1980, pp. 107-128. 11 Fishman, R. A., Reiner, M. and Chan, P. H., Metabolic changes associated with iso-osmotic regulation in brain cortex slices, J. Neurochem., 28 (1977) 1061-1067. 12 Hardebo, J. E. and Owman, C., Barrier mechanisms for neurotransmitter monoamines and their precursors at the blood-brain interface, Ann. Neurol. (Chic.), 8 (1980) 1-11. 13 Hochwald, G. M., Wald, A. and Malhan, C., The sink action of cerebrospinal fluid volume flow. Effect on brain water content, Arch. NeuroL (Chic.), 33 (1976) 339-344. 14 Katzman, R., Clasen, R., Klatzo, I., Meyer, J. S., Pappius, H. M. and Waltz, A. G., Brain edema in stroke, Stroke, 8 (1977) 509-540. 15 Katzman, R. and Leiderman, P. H., Brain potassium exchange in normal adult and immature rats, Amer. J. Physiol., 175 (1953) 263-270. 16 Lockwood, A. H., Adaptation to hyperosmolality in the rat, Brain Research, 200 (1980) 216-219. 17 Oldendorf, W. H., Szabo, J., Amino acid assignment to one of three blood-brain barrier amino acid carriers, Amer. J. Physiol., 230 (1976) 94-98. 18 Prioleau, G. R., Fishman, R. A. and Chan, P. H., Induction of brain edema by fatty acids in vivo, Ann. Neurol. (Chic.), 6 (1979) 156. 19 Rapoport, S. I., Blood-brain Barrier in Physiology and Medicine, New York, Raven Press, 1976. 20 Rottenberg, D. A., Hurwitz, B. J. and Posner, J. B., The effect of oral glycerol on intraventricular pressure in man, Neurology, 27 (1977) 600-608. 21 Shank, R. P. and Baxter, C. V., Metabolism of glucose, amino acids, and some related metabolites in the brain of toads (Bufo boreas) adapted to fresh water or hyperosmotic environments, J. Neurochem., 21 (1973) 301-313. 22 Sloviter, H. A., Shimkin, P. and Suhara, K., Glycerol as a substrate for brain metabolism, Nature (Land.), 210 (1966) 1334-1336. 23 Thurston, J. H., Hauhart, R. E. and Dirgo, J. A., Taurine: a role in osmotic regulation of mammalian brain and possible clinical significance, Life Sci., 26 (1980) 1561-1568. 24 Wright, E. M., Transport processes in the formation of the cerebral spinal fluid. Rev. PhysioL Biochem. PharmacoL, 83 (1978) 1-34. 25 Zanchin, G., Sershen, H. and Lajtha, A., The effect of hyperosmolal urea on the transport of amino acids into rat brain, Exp. Neurol., 51 (1976) 292-303.