Dose-dependence of endoneurial fluid sodium and chloride accumulation in galactose intoxication

Journal of the Neurological Sciences, 1988, 86:113-124 Elsevier

113

JNS 03022

Dose-dependence of endoneurial fluid sodium and chloride accumulation in galactose intoxication Andrew P. Mizisin 1, Robert R. Myers 2,3, Heidi M. Heckman 2 and Henry C. Powell 1 Departments of IPathology (Neuropathology), 2Anesthesiology, and 3Neurosciences, University of California, San Diego, School of Medicine, and the Veterans Administration Medical Center, San Diego, La Jolla, CA 92093 (U.S.A.) (Received 21 October, 1987) (Revised, received 20 March, 1988) (Accepted 28 March, 1988)

SUMMARY

Endoneurial edema in galactose neuropathy was studied in a colony of Sprague-Dawley rats fed diets containing 0~/o, 10~o, 20~o or 40~o D-galactose for approx. 200 days. Endoneurial fluid was analyzed by X-ray microanalysis for electrolyte concentration, by microgravimetry of whole nerve segments for water content, by measurement of endoneurial fluid pressure and by morphometry in transverse sections of nerve. Galactose intoxication resulted in dose-dependent increases in endoneurial fluid sodium and chloride that were directly associated with increases in nerve water content and endoneurial fluid pressure. The presence of edema and its dose-dependence was also confirmed by morphometric analysis of sciatic nerves at the light microscopic level. The data demonstrate that electrolyte-induced osmotic imbalances in endoneurial fluid are dependent on the amount of galactose ingested and suggest that the doserelated accumulation of sodium and chloride in endoneurial fluid contributes substantially to the pathogenesis of galactose neuropathy.

Key words: Galactose neuropathy; Dose-dependent osmotic imbalance; Endoneurial fluid electrolytes; Edema; Endoneurial fluid pressure

Correspondence to: Andrew P. Mizisin, Ph.D., Department of Pathology (Neuropathology), V-151, University of California, San Diego, La Jolla, CA 92093, U.S.A. Telephone: (619) 543-6608. 0022-510X/88/$03.50 © 1988 Elsevier Science Publishers B.V. (Biomedical Division)

114 INTRODUCTION

Since Van Heyningen (1959) first detected elevated polyol levels in the c ataractous lenses ofgalactose-fed rats, the sorbitol pathway has been implicated in the pathogenesis of complications of diabetic hyperglycemia. Galactose, unlike glucose, can not be completely metabolized by the enzymes of the polyol pathway which results in the rapid accumulation of the polyol, dulcitol, during gatactose intoxication. The work of Kinoshita and colleagues on cataractogenesis (for review see Kinoshita 1965) emphasized the possibility for osmotic imbalance when tissues containing the sorbitol pathway enzymes are exposed to elevated blood-sugar levels. After localization of polyol pathway enzymes in nerve (Gabbay et al. 1966), the osmotic theory of cataractogenesis was extended to this tissue in an attempt to understand the pathogenesis of galactose neuropathy in terms of an osmotic imbalance resulting from sorbitol pathway hyperactivity. However, edema is confined to the extracellular endoneurial space (Myers et al. 1979; Low et al. 1985; Nukada et al. 1986; Griffey et al. 1987) and this, coupled with detection of only micromolar amounts of polyols (Sharma et al. 1976; Willars et al. 1987) and the cellular location of sorbitol pathway enzymes (Gabbay and O'Sullivan 1968; Ludvigson and Sorenson 1980; Kern and Engerman 1982; Chandler and Miller 1986; Chakrabarti et al. 1987), does not support polyol accumulation as the sole explanation for galactose-induced complications in nerve. Recent measurements of endoneurial fluid electrolytes have revealed that sodium and chloride concentrations in rats fed a 40~o D-galactose diet are nearly twice those of normal (Mizisin et al. 1986a). Further studies have associated edema and increased interstitial fluid pressure with endoneurial fluid electrolyte accumulation in galactose neuropathy and demonstrated that this osmotic imbalance is linked to the sorbitol pathway. Evidence for sorbitol pathway involvement includes the restoration of nerve water content, endoneurial fluid pressure and electrolyte concentration to normal levels after treatment of galactose-fed rats with the aldose reductase inhibitor, Statil (Mizisin et al. 1986b). These results suggested that endoneurial fluid electrolyte accumulation accounts for much of the osmotic imbalance produced by gatactose intoxication and that this imbalance is dependent on the level of sorbitol pathway activity. Further support for a link between endoneurial electrolyte accumulation and sorbitol pathway activity was the preliminary observation that electrolyte accumulation appeared to be related, in a dose-dependent manner, to the amount of galactose provided in the diet (Mizisin et al. 1986b). In order to examine this possibility, a colony of age-matched, female Sprague-Dawley rats was divided into 4 groups and fed diets containing either 0 ~o, 10 ~/o, 20 ~o or 40 ~/o D-galactose before measurements of endoneurial fluid electrolytes, nerve water content and interstitial fluid pressure were made. We now report that galactose intoxication results in dose-dependent increases in endoneurial fluid sodium and chloride concentrations, and that this dose-dependent accumulation is associated with dose-dependent increases in nerve water content, endoneurial fluid pressure and edema.

115 MATERIALS AND METHODS Forty-eight age-matched Sprague-Dawley rats were randomly assigned to 4 groups of 12 animals, housed in cages with wire bottoms and fed diets containing either 0~o, 10~o, 20~o or 40~o D-galactose for approx. 200 days (Table 1). The galactose was provided in biscuits made of powdered rat chow (Wayne Rodent Blox), D-galactose and/or solka-floc (a non-nutritive fiber used by Purina in rodent chow) mixed together with water and oven-dried. All diets contained 60 ~o powdered rat chow and either 40 ~o solka-floc (0~o galactose diet), 10~o D-galactose and 30~o solka-floc (10~o galactose diet), 20~o D-galactose and 20~o solka-floc (20~o galactose diet) or 40~o D-galactose (40~o galactose diet). The diets and water were available ad libitum. Prior to killing with a 0.5-ml intracardiac injection of Beuthanasia-D, the rats were anesthetized with a mixture (2 ml/kg) of pentobarbital (50 mg/ml), diazepam (5 mg/ml) and saline in volume proportions of 1 : 1 : 2 given intraperitoneally. In 24 animals (6 animals per diet group), the left sciatic nerve was exposed and used in determinations of endoneurial fluid (EF) electrolyte concentrations and the right sciatic nerve for determination of nerve water content. In the remaining 24 animals (again 6 per diet group), the left sciatic nerve was used for endoneurial fluid pressure (EFP) measurements and the right sciatic nerve was removed, processed for histology and used for edema morphometry.

EF electrolytes Energy dispersive X-ray spectrometry was used to quantify EF electrolytes in microdrop samples with a method developed by Quinton (1978) and modified for peripheral nerve by Myers et al. (1983). Using a Teflon-coated glass micropipette, EF was aspirated from a small perineurial incision and transferred immediately to a petri dish containing water-saturated hexadecane. Subsequent transfers of EF, serum and six electrolyte standards to a parlodion-coated nickel grid were made with a Teflon-coated

TABLE 1 BODY WEIGHT AND LENGTH OF DIET All values represent mean + SD (n = 12). Diet ( % galactose)

0 10 20 40

Time on dieta (days)

Weight (g) Initial

Final

263.9 + 28.2 274.1 + 32.6 260.2 + 18.2 263.0 _+23.6

280.3 + 20.7 295.2 + 37.9* 272.2 + 26.4 252.0 + 22.3*

Overall average time on diet 201.9 + 30.7 days (mean _+SD; n = 48). * Difference is significant (P < 0.05).

a

203.6 + 28.7 211.9 + 24.2 199.3 + 31.2 193.0 + 37.7

116 glass picoliter pipette under hexadecane in order to prevent evaporation. After rinsing in hexane, the grid with the microdrops was flash-evaporated under vacuum. Microanalysis was performed with a scanning electron microscope (Etec Autoscan) equipped with an X-ray detector (Kevex). During each analysis, the accelerating voltage was set at 20 keV and the specimen current at 1.5 nA as measured by a specimen current Faraday trap. A magnification of 200 diameters was used with a reduced raster scan set to just include each microdrop. Nerve water content

Sciatic nerve water content was determined with a microgravimetric method described in detail by Costello et al. (1982). Briefly, a density gradient column was prepared with two nonaqueous fluids of high and low specific gravities (bromobenzene and kerosene respectively). Droplets of salt solutions (potassium sulfate) of known specific gravity were used to calibrate the relationship between depth in the column and specific gravity. The specific gravity of nerve was determined from its position in the column and was in turn related to water content. EFP

Endoneurial fluid pressures were recorded using the servo-nuU micropipette method (Myers et al. 1978), a technique based on changes that occur in the resistivity of the tip of a fluid-filled pipette when it has been inserted through the perineurium into the endoneurial space of an exposed nerve fascicle. Resistivity changes result from positive endoneurial fluid pressure forcing interstitial fluid into the rnicropipette tip and create an error signal sensed by a Wheatstone bridge. The error signal drives a linear motor forcing a reservoir of micropipette fluid back down the pipette shank balancing the Wheatstone bridge and minimizing the error signal. The hydraulic force required to minimize the error signal is equivalent to the EFP of the nerve fascicle. Histology

Sciatic nerves were fixed by immersion in 2.5~o phosphate-buffered glutaraldehyde for 24 h, postfixed in 1 ~o aqueous osmium tetroxide for 2 h and dehydrated in both a graded ethanol series and propylene oxide. After infiltrating with a 1 : 1 mixture of propylene oxide and araldite for 4 h and with 100~o araldite overnight, the nerves were embedded in fresh araldite resin. Thick sections (1 #m) were cut with glass knives and stained with p-phenylenediamine prior to examination with a light microscope. Morphometric analysis of edema was performed on magnified light micrographs of nerve cross sections using a Talos digitizer system linked to a Cromemco computer. Edema was quantified by measuring the linear distance from randomly selected axons to the nearest relevant structure (perineurium, axon or vessel for subperineurial, interstitial and perivascular edema respectively). A total of 100 transverse sections of sciatic nerve (5 transverse sections per nerve x 5 animals per group x 4 groups) were examined to assess subperineurial edema. For measures of perivascular and interstitial edema, only one randomly selected transverse section per animal was examined, All slides were coded so that the linear measurements were made in a blinded fashion.

117 Statistics To insure that the time period over which these data were collected did not introduce a bias, the data were plotted versus animal age or length of diet and the resulting correlation coefficients were compared to the 1 ~o critical value for correlation coefficients. ANOVA was used to analyze the data after which multiple comparisons were made with the Tukey test (Zar 1984).

RESULTS The body weights of animals used in this study were not significantly different at the beginning of the experiment nor were there significant weight gains or losses in each group after approximately 200 days on the diets (Table 1). Between-group differences in final body weight were significant only in animals fed 10~o and 40~o galactose (P < 0.05). Correlations between E F electrolytes, nerve water content, EFP and edema versus animal age or length of diet were not significant (P > > 0.05; r values ranged from 0.10 to 0.24; r 2 values ranged from 0.01 to 0.06). The lack of any significant correlation suggests that no bias due to animal age or length of diet was introduced during the time these data were collected. By 4months, all rats receiving 40~o galactose developed bilateral nuclear cataracts, polydypsia and polyuria. In the 20~o diet group, polydypsia and polyuria were present by 4 months but nuclear cataract development was not noted in any rats until 5 months and was not a consistent feature at death. Nuclear cataracts, polydypsia and polyuria were absent at death in all rats receiving the 0~o and 10~o galactose diets. Galactose intoxication was associated with increases in EF sodium and chloride concentrations that were related, in a dose-dependent manner, to the amount of galactose provided in the diet (Fig. 1). The EF sodium concentrations of animals on the 20 ~o or 40~o galactose diets (245.0+29.5 and 262.7+28.3mEq/1 respectively; mean + SD) was significantly different than sodium concentrations measured in animals receiving the 0~o galactose diet (143.5 + 34.5 mEq/1; P < 0.001). Differences between EF chloride concentrations of animals on 20~o or 40~o diets (148.7 + 17.8 and 167.9 + 14.2 mEq/1 respectively; mean + SD) were also significant when compared to chloride concentrations of animals receiving the diet without galactose (93.5 + 13.5 mEq/1; P < 0.001). However, EF potassium and serum sodium, chloride and potassium concentrations remained nearly constant in all diet groups. The magnitude of increases in nerve water content and EFP were also related to the amount of galactose provided in the diet in a dose-dependent manner (Fig. 2A and B). Differences between nerve water content (64.0 + 1.5 ~o ; mean + SD) of the 0~, galactose diet group and either the 20~o or 40~o group (74.2 + 2.7~o and 75.4 + 4.6~o, respectively) are significant (P < 0.05). Similarly, the EFP of the 0~o galactose group (2.1 + 0.9 cm H20; mean + S D) was significantly different than either the 20 ~o or 40 ~o group (5.1 + 1.0 and 8.4 + 2.2 cm H 2 0 respectively; P < 0.05). Dose-dependent increases in nerve water content and EFP were confirmed by histologic evidence of edema (Fig. 3C and D). Quantitative edema morphometry revealed dose-dependent increases

118

280,

E

Na + 2 1 0 '

p0

uJ

140 18o.

a15

C I" 130

BO

K + 20 ] O

lo

20

4'0

GALACTOSE IN DIET (% by wt)

Fig. l. Endoneurial fluid (O O ) and serum (O . . . . O) sodium, chloride and potassium as a function of galactose diet. Each point represents the mean + SEM of the mean of 6 animals. In points lacking error bars, the SEM is less than the diameter of the point. Serum and endoneurial fluid electrolytes were measured in each animal. Lines were drawn to connect the points at each diet. Asterisks indicate significant difference from EF electrolyte value of 0% galactose diet and serum value of the same diet (P < 0.001). Data were analyzed with 3-way ANOVA after which multiple comparisons were made with the Tukey test.

Q

O

10-

80 ¸

z z~ z 0 a:

~ z

7.5,

75

70-

65 ~

80

f o

:g E

uJ 2.5

~o

2b

GALACTOSE IN DIET (%bywt)

4.0

o

o

~o 20 GALACTOSE IN DIET (%bywt)

;o

Fig. 2. Nerve water content (A)andendoneurialfluidpressure(B) as a function ofgalactose diet, Each point represents the mean + SEM of 6 animals. Lines were drawn to connect the points at each diet. Asterisks indicate significant difference from 0% diet (P < 0.05); Data were analyzed with a l'way ANOVA aPter which multiple comparisons were made with the Tukey test.

119

Fig. 3. Transverse sections of sciatic nerve from rats fed (A) 0%, (B) 10%, (C) 20 % or (D) 40% D-galactose diets. Note the closely packed nerve fibers in (A) and (B) and the presence of perivascular, interstitial and subperineurial edema (arrows) in (C) and (D). Paraphenylenediamine. Original magnification, × 200.

120

40'

=:-~

32

;> E ~

24

,..< --

~=__.

~,,=, .=E ¢

8

== (/) i

0

°"°'"'"

"'°i"

~ ' .....

10 GALACTOSE

,

,

20

40

IN D I E T

(%

bywt)

Fig. 4. Subperineurial(I ~),perivascular (O . . . . O) and interstitial (& .......... &)edema as a function of galactose diet. Edema was quantified by measuring the linear distance from randomly selected axons to the nearest relevant structure (perineurium, axon or vessel for subperineurial, interstitial or perivascular edema respectively). Each point represents the mean + SEM of 5 animals. In points lacking error bars, the SEM is less than the diameter of the point. Lines were drawn to connect the points at each diet. Asterisks indicate significant differences from 0% galactose diet (P < 0.05). Data were analyzed with 2-way ANOVA after which multiple comparisons were made with Tukey test.

in subperineurial edema; dose-dependent increases of similar magnitude were not evident in measures o f perivascular and interstitial edema (Fig. 4).

DISCUSSION The data presented above demonstrate that the magnitude o f the endoneurial osmotic imbalance associated with galactose neuropathy is dependent on the amount of galactose ingested. In peripheral nerve, galactose intoxication results in a dosedependent accumulation of sodium and chloride in endoneurial fluid (Fig, I) that is associated with dose-dependent increases in nerve water content (Fig. 2A), endoneurial fluid pressure (Fig. 2B) and edema (Fig. 4). The influence of E F sodium or chloride concentration on nerve water content is further emphasized when nerve water content is expressed as a function of electrolyte concentration (Fig. 5). These data, in conjunction with the magnitude o f sodium and chloride accumulation (nearly 200 mosmol/t increase on a 40~o galactose diet over 0~o diet; Fig. 1) and extracellular location of edema (Fig. 3C and D ; Myers et al. 1979; Low et al. 1985; N u k a d a et al. 1986; Griffey et al. 1987), suggest that these electrolytes make a substantial contribution to the osmotic imbalance characteristic of galactose neuropathy. Furthermore, the magnitude of polyol accumulation (nearly 10 mosmol/1 increase in total nerve polyhydric substances on a 4070 galactose diet over control; S h a r m a et al. 1976), cellular location o f

121 85 ¸

?~

80

7=

LU z 0 0

75

70

z

65

60

loo

1so

200

2~o

300

EF S O D I U M ( m E q / I )

Fig. 5. Nervewater content as a function ofendoneurial fluid sodium concentration. Each point represents the nerve water content of the right sciatic nerve plotted against endoneurial fluid sodium concentration of the left sciatic nerve of an individual animal. The line is the least-square fit to the data. Correlation coefficient = 0.77; coefficientof determination = 0.60. Plot of an animal's nerve water content as function of the sum of endoneurial fluid sodium and chloride concentrations (plot not shown) is similar (r = 0.77; r 2 = 0.60).

sorbitol pathway enzymes (Gabbay and O'Sullivan 1968; Ludvigson and Sorenson 1980; Kern and Engerman 1982; Chandler and Miller 1986; Chakrabarti et al. 1987) and relatively impermeant nature ofpolyols (LeFevre and Davies 1950; Wick and Drury 1951) argue against an appreciable polyol contribution to the osmotic defects resulting from galactose intoxication. The dose-dependent response of EF sodium and chloride concentration (Fig. 1), nerve water content (Fig. 2A) and subperineurial edema (Fig. 4) to galactose intoxication is characterized by an abrupt increase between values associated with the 10~o and 20~o galactose diets. This abrupt increase, coupled with the lack of substantial increases in dose response between 0 ~o and 10 ~o and between 20 ~o and 40 ~o, suggests that, in this study, the limit of galactose intoxication tolerated by sciatic nerve without appreciable endoneurial osmotic imbalance is within this range. In this regard it is important to note that the dose response of EFP (Fig. 2B) is linear over the same range of galactose diets and does not mirror the dose response of EF sodium and chloride (Fig. 1) or nerve water content (Fig. 2A). However, EFP is a function of perineurial compliance (Myers et al. 1980; Low 1981, 1984). The relationship between changes in endoneurial volume and EFP is sigrnoidal, with initial increases in endoneurial volume associated with only small increments in EFP as potential space is occupied by edema. Further increases in edema shift the relationship of endoneurial volume and E F P to the steep slope of the compliance curve where small increments in volume are associated with larger increments in pressure. Indeed, plotting average E F P values (Fig. 2B) versus average nerve water content (Fig. 2A) results in a curvilinear relationship (plot not shown) similar to the initial steep rise in the perineurial compliance curve. Furthermore, it is possible that marked increases in EF electrolytes and nerve water content between the 20~o and 40~o galactose diets are blunted by increases in E F P which would

122 represent a hydrostatic force opposing movement of electrolytes and water into the endoneurium. The sorbitol pathway enzymes, aldose reductase and sorbitol dehydrogenase, are found in insulin-independent tissues such as nerve, eye and kidney where tissue sugar levels reflect blood-sugar levels (Gabbay 1973). Because the activity of these enzymes is highly substrate-concentration dependent, hyperglycemia accelerates sorbitol pathway activity (Gabbay 1973). While the notion linking the osmotic imbalance in galactose neuropathy to sorbitol pathway activity is not new (Gabbay and Snider t972), the suggestion that EF sodium and chloride accumulation is dependent on polyol pathway activity is novel and supported by the dose-dependent responses to galactose intoxication reported here. Further evidence is derived from a study in which treatment of galactose-intoxicated rats with the aldose reductase inhibitor, Statil, restored EF sodium and chloride concentrations to normal levels while also ameliorating edema and elevated EFP (Mizisin et al. 1986b). These observations emphasize that, in addition to generating organic osmolytes, the sorbitol pathway in galactose neuropathy is a metabolic process that drives endoneurial fluid sodium and chloride accumulation and parallel increases in nerve water content and EFP. The accumulation of EF sodium and chloride in rats fed 20 °/o and 40. .... ~, galactose is in marked contrast to normal serum sodium and chloride levels recorded in these animals (Fig. 1, Table 1 ; Mizisin et al. 1986a) and suggests movement against agradient maintained by the blood-nerve barrier. How EF sodium and chloride accumulation occurs is not known but it appears to be dependent on the amount of galactose ingested (Fig. 1) and on increased sorbitol pathway activity (Mizisin et al. 1986b). These observations and the fact that polyol pathway metabolism is substrate-concentration dependent suggest that electrolyte accumulation is linked to the movement of sugar into the endoneurium. While cotransport of monosaccharides and sodium is known to occur in the small intestine (Frizzell et al. 1973) and kidney (for review see Ullrich 1976), current evidence suggests that transcellular movement of monosaccharides across the blood-nerve barrier into nerve is mediated by facilitated transport and that there is no facilitated or active ion transport into the endoneurium (for review see Rechthand and Rapoport 1987). Although paracellular movement of nonelectrolytes and electrolytes across the blood-nerve barrier has been demonstrated (Weerasuriya et al. 1980; MacKenzie et al. 1987), in the absence of precise knowledge of all the forces involved it is not possible at this time to adequately describe a mechanism by which movement of galactose into nerve is linked to sodium and chloride accumulation. However, the recent report of increased Na/K-ATPase activity in nerve from galactose-intoxicated rats (Lambourne et al. 1987) raises the interesting possibility that this altered enzyme activity might participate in the endoneurial electrolyte accumulation reported here. While most animal models of galactose toxicity are not entirely comparable to human transferase or galactokinase deficiency galactosemia, the presence of dulcitol in the tissues and urine emphasizes the importance of the sorbitol pathway as an alternative means of galactose metabolism in both animal and human disorders (for review see Segal 1983). Without considering whether galactose is provided as such in the diet (most animal models of galactose toxicity) or derived from lactose-containing milk (human

123 galactosemia or marsupial models of galactose toxicity), the observations in this study suggest that the osmotic imbalance seen in nerve following galactose intoxication is linked to the amount ofgalactose available for reduction to dulcitol by aldose reductase. Furthermore, these observations predict sodium and chloride accumulations in other tissues, such as the brain and retina, that contain the sorbitol pathway and are protected by blood-tissue barriers and suggest that the brain edema characteristic of human galactosemia in infancy (Quan-Ma et al. 1966) may be associated with similar electrolyte accumulations.

ACKNOWLEDGEMENTS

Supported in part by NS 14162, NS 18715, the Veterans Administration Research Service and a Juvenile Diabetes Foundation Postdoctoral Fellowship to A. P. Mizisin. The technical assistance of Ms. L. Woodward is gratefully acknowledged. We thank Dr. Michael W. Kalichman for helpful discussions about experimental design.

REFERENCES Chandler, C.E. and L.J. Miller (1986) Studies of aldose reductase using neuronal cell culture, Metabolism, 35: 71-77. Chakrabarti, S., A.A.F. Sima, T. Nakajima, S. Yagihashi and D.A. Greene (1987) Aldose reductase in the BB rat: isolation, immunological identification and localization in the retina and peripheral nerve, Diabetologia, 30: 244-251. Costello, M.L., H.C. Powell and R.R. Myers (1982) Microgravimetric analysis of nerve edema, Muscle Nerve, 5: 261-264. Frizzell, R.A., H.N. Nellans and G.C. Schultz (1973) Effects of sugars and amino acids on sodium and potassium influx in the rabbit ileum, J. Clin. Invest., 52: 215-217. Gabbay, K. H. (1973) The sorbitol pathway and complications in diabetes, N. Engl. J. Med., 288:831-836. Gabbay, K.H. and J.B. O'Sullivan (1968) The sorbitol pathway - - enzyme localization and content in normal and diabetic nerve and cord, Diabetes, 17: 239-243. Gabbay, K.H. and J.J. Snider (1972) Nerve conduction defect in galactose-fed rats, Diabetes, 21(5): 295-300. Gabbay, K.H., L.O. Merola and R.A. Field (1966) Sorbitol pathway: presence in nerve and cord with substrate accumulation in diabetes, Science, 151: 209-210. Griffey, R. H, R.P. Eaton, C. Gasparovic and W. Silmer (1987) Galactose neuropathy. Structural changes evaluated by nuclear magnetic resonance spectroscopy, Diabetes, 36: 776-778. Kern, T. S. and R.L. Engerman (1982) Immunohistochemical distribution of aldose reductase, Histochem. J., 14: 507-515. Kinoshita, J.H. (1965) Cataracts in galactosemia - - The Jonas Friedenwald Memorial lecture, Invest. Ophthalmol., 4: 786-799. Lambourne, J. E., D. R. Tomlinson, A.M. Brown and G.B. Willars (1987) Opposite effects of diabetes and galactosaemia on adenosine triphosphatase activity in rat nervous tissue, Diabetologia, 30: 360-362. LeFevre, P. G. and R. I. Davies (1951) Active transport into human erythrocyte: Evidence from comparative kinetics and competition among monosaccharides, J. Gen. Physiol., 34:515-524. Low, P. A. (1981 ) In vitro study of acute elevations of endoneurial pressure in mammalian peripheral nerve sheath, Exp. NeuroL, 74: 160-169. Low, P. A. (1984) Endoneurial fluid pressure and microenvironment of nerve. In: P.J. Dyck, P. K. Thomas, E.H. Lambert and R. Bunge (Eds.), Peripheral Neuropathy, Vol. 1, W.B. Saunders, Philadelphia, pp. 599-617.

124 Low, P. A., H. Nukada, J.D. Schmelzer, R.R. Tuck and P.J. Dyck (1985) Endoneurial oxygen tension and radial tomography in nerve edema, Brain Res., 341: 147-154. Ludvigson, M.A. and R.L. Sorenson (1980) Immunohistoehemical localization of aldose reductase. I. Enzyme purification and antibody preparation - - localization in peripheral nerve, artei'y and testis, Diabetes, 29: 438-449. MacKenzie, M. L., M. N. Ghabriel and G. Allt (1987) The blood-nerve barrier: an in vivo lanthanum tracer study, J. Anat., 154: 27-37. Mizisin, A. P., R. R. Myers and H.C. PoweU (1986a) Endoneurial sodium accumulation in galactosemic rat nerves, Muscle Nerve, 9: 440-444. Mizisin, A. P., H. C. Powell and R.R. Myers (1986b) Edema and increased endoneurial sodium in galactose neuropathy. Reversal with an aldose reduetase inhibitor, J. Neurol. Sci., 74: 35-43. Myers, R.R., H.C. Powell, M.L. Costello, P.W. Lampert and B.J. Zweifach (1978) Endoneurial fluid pressure: direct measurement with micropipettes, Brain Res., 48: 510-515. Myers, R.R., M.L. Costello and H.C. Powell (1979) Increased endoneurial fluid pressure in galactose neuropathy, Muscle Nerve, 2: 299-303. Myers, R. R., H. C. Powell, H. M. Shapiro, M. L. Costello and P.W. Lampert (1980) Changes in endoneurial fluid pressure, permeability, and peripheral nerve ultrastructure in experimental lead neuropathy, Ann. Neurol., 8: 392-401. Myers, R. R., H. M. Heckman and H.C. Powell (1983) Endoneurial fluid is hypertonic. Results of microanalysis and its significance in neuropathy, J. Neuropath. Exp. Neurol., 42: 217-224. Nukada, H., P.J. Dyck, P.A. Low, A.C. Lais and M.F. Sparks (1986) Axonal caliber and neurofilaments are proportionately decreased in galactose neuropathy, J. Neuropath. Exp. Neurol., 45: 140-150. Quan-Ma, R., H.J. Wells, W.W. Wells, F.E. Sherman and T.J. Egan (1966) Galactitol in the tissues of ~ galactosemic child, Am. J. Dis. Child., 112: 477-478. Quinton, P.M. (1978) Uttramicroanalysis of biological fluids with energy dispersive x-ray spectrometry, Micron., 9: 57-69. Rechthand, E. and S.I. Rapoport (1987) Regulation of the mieroenvironment of peripheral nerve: role of the blood-nerve barrier, Prog. Neurobiol., 28: 303-343. Segal, S. (1983) Disorders of galactose metabolism. In: J. B. Stanbury, J. B. Wyngaarden, D. S. Fredrickson, J.L. Goldstein and M.S. Brown (Eds.), Metabolic Basis of Inherited Disease, McGraw-Hill Book Company, New York, pp. 167-191. Sharma, A. K., P. K. Thomas and R. W. R. Baker (1976) Peripheral nerve abnormalities related to galactose administration in rats, J. Neurol. Neurosurg. Psychiat., 39: 794-802. Ulrich, K.N. (1976) Renal tubular mechanisms of organic solute transport, Kidney Int., 9: 134-148. Van Heyningen, R. (1959) Formation of polyols by the lens of the rat with 'sugar' cataract, Nature, 184: 194-195. Weerasuriya, A., S.I. Rapoport and R. E. Taylor (1980) Ionic permeabilities of the frog perineurium, Brain Res., 191: 405-415. Wick, A. N. and D. R. Drury (1951) Action of insulin on the permeability of cells to sorbitol, Am. J~ Physiol., 166: 421-423. Willars, G.B., J.E. Lambourne and D.R. Tomlinson (1987) Does galactose feeding provide a valid model of consequences of exaggerated polyol-pathway flux in peripheral nerve in experimental diabetes? Diabetes, 36: 1425-1431. Z ar, J.H. (1984)Biostatistical Analysis. Prentice Hall, Englewood Cliffs, NJ.