Ki4ney International, Vol. 48 (1995), pp. 344—353 Time-dependent aspects of osmolyte changes in rat kidney, urine, blood and lens with sorbinil and galactose feeding Scarr D. EDMANDS, KARIN S. HUGHS, SE YOUNG LEE, SUZANN D. MEYER, EDWIN Sii, and PAUL H. YANCEY Biology Department, Whitman College, Walla Walla, Washington, USA Time-dependent aspects of osmolyte changes in rat kidney, urine, blood and lens with sorbinil and galactose feeding. Sorbitol plus myo-inositol, betame and glycerophosphorylcholine (GPC) are cellular osmolytes in the mammalian renal medulla. Galactosemia and hyperglycemia can cause excessive levels of galactitol or sorbitol in several organs via aldose reductase (AR) catalysis. AIR inhibitors can reduce these polyols. To examine osmolyte responses to polyol perturbations, male Wistar rats were fed normal diet, the AR inhibitor sorbinil (at 40 mg/kg/d), 25% galactose, or a combination, for 10, 21 and 42 days. All animals at 21 days had higher apparent renal AR activity than at 10 or 42 days, possibly providing resistance to sorbinil. Sorbinil feeding alone tended to increase urinary, plasma and renal urea levels. It reduced AR activity and sorbitol contents in renal inner medulla, though less so at 21 days; other renal osmolytes, especially betaine, were elevated. Galactose feeding caused little change in renal AR activity, and resulted in high galactose and galactitol contents in renal medulla, urine, blood and lens (and higher renal Na contents at 10 days). Renal sorbitol, inositol and GPC decreased, while betaine contents trended higher at all times. Sorbinilgalactose feeding reduced renal AR activities and galactitol contents (again less so at 21 days), urine, blood and lens galactitol, and further reduced renal sorbitol contents. At 10 and 21 days it tended to raise renal betaine more, and restore inositol (but not GPC) contents to control levels. At 42 days it reduced renal and urinary Na and galactose, and decreased renal betaine to control levels. Under most conditions, total phorylcholine (GPC), and betaine (N-trimethylglycine) [3]. Initial studies, which altered renal osmotic conditions through variations in water or salt intake, showed that total renal contents of osmolytes are up- and downregulated [3—7], tracking sodium contents in antidiuretic and diuretic states such that cell volume is probably maintained [1, 5, 7]. Recent studies have focused on regulation of individual osmolytes. Polyols are of particular interest because of unusual conditions which perturb them. First, excesses occur in diabetic hyperglycemia and galactosemia, in which glucose or galactose is converted to largely unmetabolized sorbitol or galactitol [8, 9], respectively, in tissues with aldose reductase (AR; EC 1.1.121), including lens, renal cortex and medulla, and neurons. It is thought that damaging osmotic pressures may result. Also, there is often a depletion of cell myo-inositol contents [8—10]. Second, decreases in polyols result from AR inhibitors, drugs which have been developed to reduce sorbitol or galactitol accumulation. These have been shown to be beneficial in hyperglycemia and galactosemia to some tissues including the kidney, though not consistently [8—14]. However, since sorbitol is used in renal (non-urea) organic osmolyte contents (presumed to be mostly normal renal osmoregulation, we have suggested that AR inhib- intracellular) and Na plus galactose contents (presumed mostly extracellular) changed together such that cell volumes may have been maintained. The exception was 10 days on galactose, where total osmolytes appeared too low. In galactose-fed animals, urine/plasma ratios suggest some renal galactitol efflux, and cellular galactitol probably helps maintain osmotic balance rather than cause swelling. itors could impair this function and perhaps cause osmotic shrinkage [151. Thus it has been hypothesized that polyol changes up or down may be osmotically damaging. However, previous renal studies of this have not been consistent. In vitro, we used PAP cells, a cultured renal line which relies heavily on sorbitol as an osmolyte. These cells in high glucose media had elevated sorbitol contents concomitant with reduced growth, consistent with high sorbitol causing cell swelling. AR inhibitors reduced the sorbitol build-up, which restored growth at low doses, but caused large decreases in both sorbitol and viability at high doses [15]. In vivo studies showed somewhat different patterns. In female Sprague-Dawley disruptive effects on macromolecules that occur with high NaC1 or rats fed 50% galactose for 10 days, Bondy et al [161 reported that, KC1 [1, 2]. Compatible osmolytes in most organisms are polyols, along with high renal galactitol contents, there were large reducmethylamines, and free amino acids [2]; in the mammalian renal tions in AR mRNA, sorbitol, inositol and GPC contents. The medulla, the major ones are sorbitol, myo-inositol, glycerophos- authors proposed that these were adaptive events (such as feedback inhibition) to maintain cell volume in the presence of excess polyols, perhaps by avoiding the damage seen in PAP cells. The changes caused by galactosemia were mostly reversed by the AR Mammalian renal medulla cells must cope with high external salt concentrations produced by the kidney's urine-concentrating mechanism. Like cells of a wide variety of organisms exposed to hypertonicity, renal cells apparently maintain cell volume with "compatible" organic osmolytes, solutes which do not exhibit the Received for publication May 18, 1994 and in revised form January 11, 1995 Accepted for publication February 27, 1995 © 1995 by the International Society of Nephrology inhibitor, sorbinil. As with PAP cells, however, osmotic stress was suggested because total osmolytes appeared lower than needed for volume maintenance [16]. In studying AR inhibition alone in vitro, we found in PAP cells 344 Edmands et al: Sorbinil, galactosemia and osmolytes that AR inhibitors reduced sorbitol contents and, consistent with cell shrinkage, greatly impaired survival of these cells in high-salt (but not normal) media [15]. However, others found that betaine added to culture media provided a substitute osmolyte which restored growth [17]. In vivo, we fed normal male Wistar rats 345 the nearest 0.1 mg, then homogenized in a glass homogenizer in 19 vol cold buffer [10 mivi N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (HEPES), 5 mri EDTA, 2 m ethylene glycolbis(b-aminoethyl ether)-N,N,N',N'-tetraacetic acid [EGTAJ, 2 mM dithiothreitol, pH 7.0). The homogenates were centrifuged in a microcentrifuge for 15 minutes at 10,000 rpm. Supernatants sorbinil for 10, 21 and 71 days. In the 10- and 71-day treatments, sorbitol contents of inner medullas fell greatly, but in apparent were analyzed kinetically at 37°C for aldose reductase activity with compensation, betaine contents increased such that cell volume 10 mM glyceraldehyde or xylose as substrates in 100 mM Na-PO4 was likely maintained [18, 19]. Therefore betaine and sorbitol may buffer (pH 7.0) with 100 mM NADPH [201. serve as interchangeable compatible osmolytes in vitro and in vivo, eliminating osmotic problems. However, a different pattern was seen in our 21-day treatment: Protein and osmolyte analyses neither sorbitol nor betaine levels were significantly altered, suggesting another type of compensation, such as up-regulation of Frozen kidneys were thawed in a cold room, and inner medullas sorbitol production (not tested for in that study) [18]. Yet a third were immediately removed and weighed to the nearest 0.1 mg, pattern was reported by Bondy et al [16] for rats fed sorbinil for followed by homogenization in a glass homogenizer in 9 vol cold 10 days: sorbitol contents decreased but no other cellular os- 7% perchloric acid. Lenses were treated similarly. The homogemolytes increased enough to compensate, suggesting osmotic nates were centrifuged for 15 minutes at 10,000 rpm (microcenstress under AR inhibition. trifuge); 50 1d aliquots were removed from each homogenate for Thus several different, sometimes conflicting patterns in renal Na analysis (see below). The pellets were dissolved in 1 ml of 1 osmolytes and possible cell stress have been reported under N NaOH, then analyzed for protein with the Bio-Rad assay different times and types of polyol perturbations. To further (Bio-Rad Laboratories, Richmond, CA, USA), using y-globulin elucidate responses of renal osmolytes to such perturbations, we treated similarly (perchioric acid, NaOH) as a standard. The treated Wistar rats with dietary sorbinil and/or galactose for 10, 21 remaining supernatant was neutralized to pH 6.8 to 7.5 with 1 M and 42 days, and measured renal AR activities and, for the first KOH. The samples were passed through Bond Elut C18 cartridges time together, osmolytes in several tissues. We hypothesized that (Varian, Harbor City, CA, USA) and 0.45-mm HV filters (Millipolyol disturbances could cause osmotic problems (from excesses pore, Bedford, MA, USA). Overall tissue dilution was 20-fold. or deficits) or trigger compensatory changes in other renal Peaks were identified using standards treated similarly. osmolytes, with the type and degree of response possibly being Urine and plasma samples were analyzed for osmotic pressure age- or time-dependent. using a vapor pressure osmometer (Wescor, Logan, UT, USA), then treated in the same manner as medullas, starting with Methods perchloric acid treatment (2:1 vol added to plasma). In the 42 day Animals and surgical procedures experiment, whole blood was also treated. Overall urine dilution All animals were male Wistar Rats initially 41 to 44 days old was 15-fold, plasma and blood dilution fivefold. (COBS, Charles River, Wilmington, MA, USA), housed under Sodium in all samples was analyzed by flame photometry standard conditions and given Purina Rodent Chow (5001, Purina Mills, St. Louis, MO, USA) and water ad libitum for 7 to 10 days (Cole-Parmer, Chicago, IL, USA). Organic osmolytes were anabefore the experiments were started. In each of three experiments lyzed using high-performance liquid chromatography (HPLC) as previously described [21] with a Sugarpak-I column (Waters, conducted for 10, 21 and 42 days, four groups of animals were fed Milford, MA, USA) with peak detection by refractive index. The ground chow, with the first group as control (Control 10 days, 21 days, 42 days), the second (Sorbinil 10 days, 21 days, 42 days) fed column separated tissue galactose, myo-inositol, GPC, galactitol, sorbitol, urea, and betaine at 70°C, with a detection limit of about ground chow with sorbinil at 40 mg/kg/day, and the third (Galactose 10 days, 21 days, 42 days) and fourth (Sorbinil + Galactose 10 4 /LM for 50 il injections. Because of unknown peaks interfering with myo-inositol and days, 21 days, 42 days) groups treated as the first and second, respectively, but with 25% galactose added in their chow. All galactose HPLC peaks in urine (but not tissue) samples, inositol animals were given water ad libitum throughout. Water and food was analyzed enzymatically as previously described [22], and the consumption and body wt were determined eveiy one to three carbohydrates were also analyzed by GC-MS (5890/5971A, days; urines were collected for two 24-hour periods near the end Hewlett-Packard, Palo Alto, CA, USA). Thirty-five microliters of the 10 and 42 day experiments. Animals were anesthetized with standards or samples, with 14 jtl of 2.5 m'vi mannitol added as CO2 and killed by cervical dislocation. Urine samples, when internal standard, were dried in 0.1 ml reaction vials in a vacuum present, were taken from the bladder with a syringe; blood over NaOH pellets. To the vials 15 jil acetonitrile and 20 l samples were collected from the renal artely for 10- and 42-day animals. Both were centrifuged for 10 minutes at 12,000 rpm (microcentrifuge). Lenses from 10- and 42-day animals and one kidney from each animal were frozen in liquid nitrogen in vials and stored at —70°C for later analysis. Aldose reductase measurements bis-trimethylsilyl-trifluoroacetamide were added; vials were capped with Teflon inserts and heated at 90°C for two hours [4]. One microliter samples were injected (split injection) onto a 30 m X 0.33 mm non-polar HP-i column (Hewlett-Packard), run at 50°C for two minutes, ramped at 10°C/mm to 170°C, held there for 17 minutes, then ramped at 5°C/mm to 180°C. This separated trimethylsilated galactose, mannitol, sorbitol, galactitol and myoThe second kidney was immediately dissected on a petri plate inositol. Resolution was limited to 30 I.LM. Peaks were identified on ice to obtain the inner medulla. The medulla was weighed to by spectrum library and standards treated similarly. Edmands et a!: Sorbinil galactosemia and osmolytes 346 Table 1. Concentrations of major osmolytes in urines, and protein contents of renal inner medullas, of rats treated in four dietary groups for times indicated Urine solute concentrations (mM SD) Group A. 10 days Control 10 days Sorbinil 10 days Galactose 10 days Sorbinil + galactose 10 days B. 21 days Control 21 days Sorbinil 21 days Galactose 21 days Sorbinil + galactose 21 days C. 42 days Control 42 days Sorbinil 42 days Galactose 42 days Sorbinil + galactose 42 days N Sodium 3 Urea Galactose Galactitol 165 31 203 45 178 57 187 56 326 102 ND 285 48" 261 27" ND ND 403 86 3 3 3 5 114 35 111 52 103 36 444 95 271 89 ND ND 320 237 3 169 24 193 40 114 57" 462 111 711 178k 359 83b ND ND 392 178 4 5 4 479 82 526 72 31 5 6 6 91.7 290 121 70 396 99 217 108k 29.5a 379 248 152" 177k Renal inner medulla % protein 6.11 6,38 7.82 6.51 0.99 0.38 0.68a 6.55 6.31 4.4 0.6 6.60 6.20 0.52 0.89 ND ND 7.42 7.02 6.5 3.0 7.28 0.97 0.53 0.97 0.58 24.7 ND 9,4a 3.1 0.5 ND ND 31.3 31.9 17.2a 7.5 1.10 1.10 0.51 7.41 N indicates number of animals; ND, not detected. a Significantly different from Control group (P < 0.01, SNK multiple comparison tests) b Combined galactose treatments had a significant effect (P < 0.005) by 2-way ANOVA "Significantly different from Galactose group (P < 0.01, SNK multiple comparison tests) Table 2. Concentrations of major osmolytes in plasmas of rats treated in four dietary groups for 42 days Plasma solute concentrations (mM N Group Control 42 days 4 Sorbinil 42 days Galactose 42 days Sorbinil + galactose 42 days Blood/plasma ratio" 4 4 4 Sodium 143 4 147 2 148 7 144 6 0.62 0.05 Urea 7.90 su), 42 days Galactose ND ND 0.57 9.31 1.03 7.22 0.96 0.61 0.13 7.22 0.79 5.99 3.94 0.69 1.70 1.80a 0.13 Galactitol ND ND 0.254 QQ47a 0.028 0,016ab 1.50 0.26 N indicates number of animals; ND, not detected. a Significantly different from Control group (P < 0.05, SNK multiple comparison tests) "Significantly different from Galactose group (P < 0.001, SNK multiple comparison tests) C Calculated for individual animals for all groups combined (for sodium, urea) or Galactose and Sorbinil + galactose groups combined (for galactose, galactitol) Statistical methods Statistics were performed by two-way analysis of variance, followed by Student-Newmann-Keuls (SNK) multiple comparison tests, and by Student's two-tailed t-tests, using Statworks (Cricket, Urines of Sorbinil animals tended to have elevated urea levels which were significant at 10 and 42 days (Table 1A, C). Those of Galactose animals had high galactose and galactitol concentra- tions and lower urea concentrations (Table 1), and lower Na Philadelphia, PA, USA) and Instat 2.0 (GraphPad, San Diego, concentrations in the 42 day group (Table 1C). Sorbinil + CA, USA). Sorbinil was provided by Pfizer Central Research Galactose urines showed similar patterns but had greatly reduced (Groton, CT, USA). Other chemicals were purchased from Sigma Chemical (St. Louis, MO, USA) and Boehringer-Mannheim (Indianapolis, IN, USA). Results Live animal, urine, blood and lens measurements Body weight gain was not affected by any experimental condition (data not shown). Galactose-fed (Galactose and Sorbinil + Galactose) animals ate significantly more food (20%), drank more water (average 83 mi/day compared to 47 for the Control plus Sorbinil groups), had greater urine output (measured in 10 and 42 day groups: 40 to 49 ml/day compared to 13 to 19 mllday for Control plus Sorbinil groups), and tended to have larger kidneys (P < 0.05 at 42 days; average 2.2 g compared to 1.8 g for Control plus Sorbinil groups). Urine osmolalities were highly variable (from 800 to 1600 mOsm/kg) and did not differ among groups. levels of galactitol in all cases (Table 1). No methylamines were found in urines; traces (<1 mM) of inositol and sorbitol were found in a few scattered urine samples, but with no significant trends. In plasmas, analyzed at 10 days (data not shown) and 42 days (Table 2), glucose (not shown), Na and urea concentrations did not differ, except for elevated urea in the 42 day Sorbinil group (Table 2). Plasmas of galactose-fed groups at 10 days (not shown) and 42 days (Table 2) had high galactose and some galactitol accumulation, but with galactitol reduced tenfold in the Sorbinil + Galactose group. For 42-day animals, whole blood was also analyzed. Rupturing blood cells had little effect on urea concentrations relative to plasma; however, it reduced concentrations of Na and galactose by 30 to 40%, indicating their primarily extracellular location, and it elevated galactitol concentration by 50%, indicating its primarily intracellular location (blood/plasma ratio, Table 2). 347 Edmands et al: Sorbinil, galactosemia and osmolytes Table 3. Contents of major osmolytes in lenses of rats, treated in four dietary groups for 10 and 42 days Group Galactose Sorbinil + galactose N 5 5 Lens solute contents, 10 days Galactose Galactitol 2.20 l.32a 2.47 0.93a Lens solute contents, 42 days Galactose 187 11.6 24.5 128ab 0.35 0.39 0.29a 0.26a Galactitol 53.7 31.5 11.7a 6.8w' N indicates number of animals. Galactose contents are in mmol/kg wet wt (± so), galactitol in mmol/kg protein (± so; protein content of lenses 5%). Control and Sorbinil groups (not shown) had no detectable levels of these solutes. a different from Control group (P < 0.05, SNK multiple comparison tests) b Significantly Significantly different from Galactose group (P < 0.05, SNK multiple comparison tests) averaged 32 Average urine/plasma ratios (U/P) for galactose (in all galactose-fed groups) and urea (in all groups), compounds mainly concentrated from blood, were statistically identical in 10 and 42 day groups, ranging from 58 to 65. For galactitol the U/P ratios ranged were 104 and 125, respectively, for the 10 and 42 day Galactose groups, suggesting some urinary galactitol arose from the kidney itself (the ratios were around 200 for the Sorbinil + Galactose animals, but as plasma levels were near the HPLC detection limit, these ratios cannot be regarded as accurate). Lenses were examined for at 10 days and 42 days. In galactose-fed groups, galactose and galactitol accumulated, but to considerably lower levels in the 42-day group (Table 3). Sorbinil treatment reduced the galactitol contents in both, but resulted in much less relative decrease in the 42 day case (Table 3). Lens protein contents (averaging 32 5%) did not differ among the groups. Renal protein contents and aldose reductase activities In previous studies we reported osmolyte contents as mmol/kg wet wt [18, 22]. In those studies, tissue protein contents (% of wet wt) did not differ among groups. We now report contents as mmol/kg protein, because of an increase in kidneys of the Galactose group at 10 days (Table 1A). Protein contents did not differ among groups at 21 and 42 days (Table 1B,C). . 90 80 . . 70 60 — 50 40 30 20 10 E0 10 days 21 days 42 days Time of treatment Fig. 1. Aldose reductase activities (units/kg protein) measured in crude renal medulla,y extracts with 10 m glyceraldehyde (G bars) or xylose (X bars) as () substrate. Symbols are: (1J) control; (i) sorbinil; () galactose; galactose + sorbinil. 4P < 0.05 compared to Controls, fP < 0.05 compared to Galactose groups, by ANOVA followed by SNK multiple comparison tests; §P < 0.05 compared to same test at other time points; error bars indicate so. As measured with glyceraldehyde as substrate (G bars in Fig. 1), animals in the 21 day experiment had more apparent renal aldose reductase (AR) activity than older and younger animals in all four among groups), suggesting that mainly aldose reductase was being dietary tests (P < 0.001 for most comparisons; Fig. 1). Sorbinil measured. But they were 31 to 35% of the glyceraldehyde activity feeding greatly reduced AR activity per kg protein in medullary in all 21 day groups (Fig. 1), suggesting the presence of other extracts in both treatment types at all times. Galactose feeding reductases (Discussion). had little effect on AR activity except for a small decrease in 10-day animals (Fig. 1). AR rates were measured in a 400-fold tissue dilution (in the assay cuvette), and probably reflect inhibition by sorbinil in the extract rather than down-regulation of the enzyme in the cells. The inhibition constant for sorbinil is 0.01 M [23]; assuming that the consumed sorbinil (40 mg/kg/day) was evenly distributed in the animal, the extract in the cuvette would have had at least 1 p.M. Assays for AR activity were also run with xylose as substrate in order to distinguish between aldose and other reductases. For example, activity of rat kidney aldehyde reductase with 10 mM Renal osmolytes: Sorbinil or galactose treatment Renal osmolyte patterns were similar, with a few differences, at the three time periods of sorbinil treatment (data in Sorbinil rows, Tables 4 to 6, plotted as percentages of Control values in Fig. 2A). Contents of Na and urea did not change in any case. Renal sorbitol contents were greatly reduced in all cases, although in the 21 day case the amount of reduction was less (Fig. 2A). In apparent compensation, betaine contents in all cases trended upwards (Fig. 2A), and were significant at 10 (Table 4) and 42 days (Table 6). Inositol and GPC also increased significantly at 42 xylose is reportedly 0.7% of the activity with 10 mtvi glyceralde- days. hyde [24], while that of AR from cultured rabbit kidney (PAP) Galactose feeding alone produced similar patterns in most cells is 46% [23]. Xylose-based measurements showed the same renal osmolytes in all three time periods (Galactose rows in patterns as with glyceraldehyde, except that activity was not higher Tables 4 to 6, plotted as percentages of Control values in Fig. 2B). in the 21-day-old groups compared to older and younger groups, Na and urea contents did not change in any case except for as it usually was with glyceraldehyde (X vs. G bars in Fig. 1). Velocities with xylose were 46 to 60% of the glyceraldehyde increased Na at 10 days (Table 4). High levels of galactose activity in the 10 and 42 day experiments (no significant difference increasing the total AR-catalyzed polyoi content (GI + S bars, (Tables 4 to 6) and galactitol (Gl bars in Fig. 2B) were found in all, Edmands et al: Sorbinil, galactosemia and osmolytes 348 Fig. 2. Contents of major osmolytes in renal inner medullas of sorbinil- A U, 0 treated (A), galactose-fed (B) and sorbinil plus galactose-fed (C) rats, using data from Tables 4 to 6 converted to percentages of appropriate Control groups (horizontal line at 100%), with galactitol (Gi bars) plus sorbitol (S 175 () 125 C a, () bars) values compared to Control sorbitol levels. Symbols are: (D) sorbitol + GPC. *P < 0.05 compared to inositol; betaine; galactitol; Controls, tP < 0.05 compared to Galactose groups, by ANOVA followed by SNK multiple comparison tests or (*) by t-test; error bars indicate SD. 150 () 100 75 ,50 0 Fig. 2B). In the 21 day case, galactitol accumulation was signifi- 0 21 days, and decreased sorbitol contents in all (Fig. 2B). In cantly higher than at 10 or 42 days (Tables 4 to 6). There were trends towards decreased inositol and GPC, significant at 10 and 25 0 21 days 10 days 42 days contrast, betaine contents tended to increase, significantly so at 21 and 42 days (Fig. 2B). Time on sorbinil 700 500 300 200 reduced compared to the Galactose group; sorbitol levels were T 1 reduced compared to all other groups. Galactose contents also fell at 42 days (Table 6). GPC contents (Fig. 2C) remained similar to their depressed states seen in untreated galactosemia (Fig. 2B). i %75 0 at 42 days (Table 6). In all cases, galactitol levels were significantly * 175 150 125 100 50 25 Renal osmolytes: Sorbinil plus galactose treatment Sorbinil plus galactose feeding resulted in somewhat different patterns at the three time periods (Sorbinil + Galactose rows in Tables 4 to 6, plotted as percentages of Control values in Fig. 2C). Na and urea contents did not change, except for decreased Na At 10 and 42 days, total galactitol plus sorbitol was reduced to about the level of sorbitol in Control animals, but at 21 days this total was higher (Gl + S bars, Fig. 2C). Other patterns were similar at 10 and 21 days: inositol contents were restored to .1111. 11. 1. __________ 10 days _____ 21 days Time on galactose _________ control levels; betaine contents increased above control levels, tending higher than in Sorbinil or Galactose groups (Tables 4 to 5; Fig. 2C). At 42 days, the depletion in renal Na content was concurrent with betaine reduced significantly relative to the Galactose groups (Fig. 2B), back to control levels (Fig. 2C). 42 days Renal cell-volume maintenance To test whether poiyol perturbations lead to cell volume imbalances, we made the following assumptions about renal cells. First, urea is thought to equilibrate across cell membranes, while Na may be largely extracellular and the normal polyols and C methylamines are likely intracellular [3, 25]. Second, the galactose in renal extracts should be mainly extracellular, given the blood- 225 ' g a) 150 plasma ratios (see above; Table 2) and the high levels in urine (Table 1) compared to renal tissue (Tables 4 to 6). Third, we initially assumed that most of the galactitol is in cells, where it is synthesized (see Discussion). Thus an estimate of cell volume 125 to total methylamines + polyols including galactitol ("total or- 200 175 control may be obtained by comparing contents of Na + galactose ganic osmolytes"). These are plotted as percentages of Controls in Figure 3. Relative osmotic effects of Na + galactose in this figure are underestimated, since cellular osmolytes presumably respond 100 75 o ,' E (I) to increase in extracellular osmotic pressures above isosmotic levels, not above zero. Nevertheless, Figure 3 suggests general 50 trends. 0 10 days 21 days 42 days Time on sorbinil + galactose Despite reduced renal sorbitol contents in all Sorbinil animals, the same total organic osmolytes (Fig. 3, left) were seen at 10 and 21 days, due to increased betaine (though not significant at 21 days) (Fig. 2A). The simultaneous increase in GPC and inositol at 349 Edmands et al: Sorbinil, galactosemia and osmolytes Table 4. Contents of major osmolytes in renal inner medullas of rats, treated in four groups for 10 days Renal medullary osmolyte content mmollkg wet wt Group N Control 10 days Sorbinil 10 days Galactose 10 days Sorbinil + galactose 10 days 5 5 6 Sodium 199 8 204 14 234 15 6 14' 196 Significance (p)d Sorbinil Galactose Galactose Urea Galactitol ND ND 27.4 16.4' 19.3 13.2' 149 63 164 58 197 44 172 21 ND ND 201 29' 64.9 18.8" NS NS NS NS <0.001 <0.001 <0.001 0.004 0.035 0.002 Interaction SD <0.001 NS mmollkg protein Sorbitol Inositol 58.3 16.1 12.3 3.8a 6.7a 28.5 5.2 4.1" <0.001 <0.001 0.008 SD 303 18 312 38 207 278 43' 22C <0.001 0.006 0.039 Betaine GPC 117 34 166 27" 161 37 275 42 0.003 0.005 NS <0.001 NS 217 40'' NS 263 13 179 22' 202 26' N indicates number of animals; ND, not detected. a Significantly different from Control group (P < 0.05, SNK multiple comparison tests; or b by two-tailed Student's t-test) 'Significantly different from Galactose group (P < 0.05, SNK multiple comparison tests) d Two-way analysis of variance; NS = not significant (P> 0.05) 150• * * * * 140 130 2 t U 120 C) 110 C 100 a, 90 Fig. 3. Contents of Na (Na bars) plus Galactose (Ga bars) and the total of GPC, 80 betaine, sorbito4 inositol, and galactitol (GI bars) ("total oanic osmolytes") in renal inner medullas of non-control rats of Tables 4 to 6. The shaded bars indicate total organic osmolytes. Totals were calculated by summing osmolytes for each individual animal before statistical analysis, then converted to percentages of appropriate control group (horizontal line at 70 60 50 10 days 21 days 42days Sorbirill - 10 days 21 days 42 days - lOdays 21 days 42 days Galactose Sorbinil + galactose 100%). 'P < 0.05 compared to Controls, tP < 0.05 compared to Galactose groups, by ANOVA followed by SNK multiple comparison tests; error bars indicate SD. 42 days (Fig. 2A) yielded a higher average total, but not signifi- However, while some work has suggested this may happen in renal cantly compared to controls (Fig. 3). Galactose animals probably systems, other has shown that compensating processes may occur had increased extracellular renal osmotic pressures due to galac- (see Introduction). Our results indicate that the kidney generally tose. At 10 days this correlated with higher renal Na (and does compensate for polyol disturbances; in particular, betaine protein) contents, but the higher external solute level (Na + appears to be used consistently to replace osmolytes which have galactose) was not matched by an increase in total organic been reduced, confirming findings of other studies [16—191. Howosmolytes (Fig. 3, middle 10 days), where high galactitol and ever, using renal, blood and urinary analyses done together for the betaine contents were offset by decreased sorbitol, GPC and first time, we suggest in contrast to others (such as [161) that inositol (Fig. 2B). At 21 and 42 days, however, organic osmolyte contents apparently matched external solutes (Fig. 3 middle). In Sorbinil + Galactose animals, total organic osmolytes appeared to match Na + galactose contents at 10 and 42 days even though Na and galactose were reduced in the latter (Fig. 3, right). At 21 days, galactitol might not cause cell swelling, and that not all osmolyte responses are directly adaptive for cell volume regulation. Finally, we propose that the apparent inconsistencies among previous studies may be due to age of animals used—because of developmental changes in AR—or to length of treatment. the variability in galactose was too great to estimate osmotic balance. Discussion Previously it has been proposed that polyol perturbations could cause osmotic problems in tissues with aldose reductase (AR). Responses to high galactose and galactitol Galactose feeding, as expected, caused a large increase in galactitol in the renal medulla (Fig. 2B). Potentially this could cause osmotic swelling, thought to be a source of damage from Edmands et al: Sorbinil, galactosemia and osmolytes 350 Table 5. Contents of major osmolytes in renal inner medullas of rats, treated in four groups for 21 days Renal medullaiy osinolyte content Group Control 21 days Sorbinil 21 days Galactose 21 days Sorbinil + galactose 21 days Significance (P)° Sorbinil Galactose Interaction N 5 5 4 6 mmol/kg wet wt Sodium Galactose 174 17 184 15 199 8 187 21 NS NS NS ND ND 28.3 30.3 8.9a 15.5° NS <0.001 NS so Urea Galactitol 168 72 194 14 144 61 140 15 ND ND NS 0.07 NS 308 53° 76.2 11•6ab 0.001 <0.001 0.001 mmollkg protein Sorbitol Inositol 49.5 15.4 29.8 21.6 7•3 6.0° 8.2° 1,2ab <0.001 <0.001 NS 194 29 199 25 SD Betaine 96.6 27.9 188 25b 117 22 137 27° 162 21° 0.03 <0.008 0.04 <0.001 NS 133 16° 0.05 GPC 228 15 243 28 165 27° 184 37° NS <0.001 NS N indicates number of animals; ND, not detected. a different from Control group (P < 0.05, SNK multiple comparison tests) b Significantly Significantly different from Galactose group (P < 0.05, SNK multiple comparison tests) Two-way analysis of variance; NS, not significant (P> 0.05) polyol accumulation in diabetic hyperglycemia [8] and galac- treatment (they did not measure AR activity). However, in our tosemia [9]. However, our results argue against such damage in study, while there was a small (26%) reduction in AR activity in the inner kidney, as there are three possible effects that may medullary extracts at 10 days of galactose feeding, there were no reductions at later times (Fig. 1). These results are not necessarily prevent it: (1.) Our urine and blood data suggest that galactose is elevated contradictory. Wu et a! [28] recently reported that in rats on 21 extracellularly in the medulla, such that increases in cellular days of galactose feeding, there was little correlation in kidney and osmolytes might be necessary to balance this. Galactitol in renal other organs between AR activity and mRNA, with the former cells might thus become a useful osmolyte (Fig. 3 middle, 21 and often showing little change despite large changes in the latter. They suggested that translation, activation and/or turnover may be 42 days). (2.) Galactitol effiux, suggested by the presence of this polyol in more important than mRNA levels in determining tissue AR urine (Table 1), may reduce cell volume changes if there is excess activities [28J. We suggest that sorbitol reduction in our Galactose galactitol. Urinary galactitol could be concentrated from blood animals is due more to AR-substrate competition than to feed(from other organs), but since our U/P ratios for galactitol were back inhibition, especially since AR has a higher affinity for twice those for urea or galactose, at least some is probably is from galactose than for glucose [9]. Second, while renal GPC levels fell in Galactose animals, they the kidney itself. Efflux could be due to membrane damage from cell swelling, or through a polyol release channel. The former is remained depressed in Sorbinil + Galactose groups (Fig. 2B, C). unlikely given that other osmolytes were not detected in urines. Therefore, GPC does not appear to be regulated in response to The detection limit of HPLC was 60 LM for undiluted samples; polyol changes directly. Instead, GPC contents may be altered to with levels of some osmolytes as high as galactitol in the kidney track urea contents, which tended to decrease in all galactoseafed (Tables 4 to 6), some should have been detectable in urines (Table groups. This is not suprising given the diuresis galactose feeding 1) had there been nonspecific membrane leakage. A specific apical membrane sorbitol permease has been reported in cultured renal (PAP) cells, which allows for fast release of sorbitol during diuresis [26], though work on primary cultures of inner medulla cells suggests a basolateral sorbitol channel [27]. induced (increased animal drinking and urinary output) in this and other studies [9, 10]. GPC has been hypothesized to act as a macromolecule-stabilizing osmolyte to counteract the perturbing effects of urea [1, 29]. Several studies have shown that renal GPC contents co-vary with urea levels in kidney and urine [1, 5, 22, 29, If some renal galactitol is extracellular, it would alter the 30]. A similar correlation (P = 0.02) was found in our data solute-total bars in Figure 3 somewhat. This would make the between urinary urea and renal GPC. For example, renal, plasma and urinary urea levels were highest in 42-day Sorbinil kidneys, lowest in 21-day Sorbinil + Galactose kidneys (Tables 1, 2, 5, 6); these groups had the highest and near-lowest levels of renal GPC (Tables 5, 6). Third, inositol downregulation appears linked to AR-polyol 2B). Bondy et al [16], who found the same pattern in galactose-fed accumulation, since sorbinil-galactose feeding both reduced gaSprague-Dawley rats, suggested these changes were specific ad- lactitol content and restored inositol contents to control levels aptations to reduce excess osmotic pressure. However, we suggest (Fig. 2C). Inositol depletion has been seen in many tissues that that they are not necessarily mechanisms for cell volume regula- accumulate sorbitol in hyperglycemia, or galactitol in galaction, but could be side effects of galactosemia, for the following tosemia, though depletion is not consistent [31]; e.g., inositol (and reasons. sorbitol) contents increased in one study in outer medullas of First, for sorbitol, Bondy et a! [16] concluded that galactitol streptozotocin-diabetic rat kidneys [11]. Most workers have proaccumulation exerts feedback inhibition on AR gene transcrip- posed depletion to be a nonadaptive side effect, harmful to cell tion, since they found 60% reduction in papillary AR mRNA after functions [31, 32]. Nevertheless, in organs such as kidney it could 10 days of galactosemia, an effect largely reversed by sorbinil be adaptive if inositol is reduced as an osmolyte specifically to apparent imbalance at 10 days on galactose (Fig. 3 middle) even worse, but conclusions about other galactose and the sorbinilgalactose groups probably would not change. (3.) Finally, in the presence of high galactitol, other renal osmolytes (sorbitol, inositol and GPC) generally decreased (Fig. 351 Edmands et al: Sorbinil, galactosemia and osmolytes Table 6. Contents of major osmolytes in renal inner medullas of rats, treated in four groups for 42 days Group N Control 42 days Sorbinil 42 days Galactose 42 days Sorbinil + galactose 42 days 5 5 5 5 Significance (P)' Sorbinil Galactose Interaction mmollkg wet wt Sodium Galactose 169 9 177 10 174 18 138 12 0.023 0,024 0.005 ND ND 24.7 5.0 9.8a 30ab <0.001 <0.001 0.001 SD Renal medullaiy osmolyte content mmolikg protein Urea Galactitol Sorbitol Inositol 202 69 245 67 215 36 159 51k' NS NS 0.05 ND ND 216 59 50.8 l1.1' <0.001 <0.001 <0.001 50.0 21.6 23.2 17.0 9,4a 3,4a 223 20 272 32 3•3 1ab 192 23 188 25 <0.001 <0.001 NS <0.001 0.07 0.03 sD Betaine 87.8 13.5 116 25 117 17 GPC 254 62 314 44 228 37 83.6 122b 182 17 NS NS 0.003 NS <0.001 0.008 N indicates number of animals; ND, not detected. a Significantly different from Control group (P < 0.05, SNK multiple comparison tests) bC Significantly different from Galactose group (P < 0.05, SNK multiple comparison tests) Two-way analysis of variance; NS, not significant (P> 0.05) offset excessive intracellular poiyols [16]. But from our data, renal galactitol accumulation does not appear to cause intracellular osmolytes to be excessive (Fig. 3, middle), thus reduced inositol may in fact be non-adaptive in terms of volume regulation. Finally, the increase in betaine contents in galactose feeding (Fig. 2B) is more consistent with cells adapting to shrinkage; it does not seem particularly adaptive to downregulate 3 osmolytes other osmolyte (including betaine) changed; total osmolytes decreased, suggesting osniotic stress. We also saw no significant changes in other osmolytes in our 21 day results (Fig. 2A). However, the increase in average betaine, though not significant by itself (Table 5; Fig. 2A), replaced the lost sorbitol exactly, keeping total organic osmolytes constant (Fig. 3, left). Time-dependent patterns and up-regulate one (betaine) simply for cell volume maintenance. Thus we conclude that galactitol accumulation does not Finally, there are some patterns that relate to animal age or cause cell swelling, but that it and galactose diuresis lead to length of experimental treatment. First, effects of galactose feeddepletion of other osmolytes, with betaine then up-regulated to ing changed with time. After 10 days, galactosemia appears to dehydrate the kidney as indicated by elevated Na (Table 4 and compensate. Fig. 3) and protein contents (Table 1A), but sorbitol, inositol and Responses to polyol reduction GPC contents decreased in the presence of galactitol such that In all sorbinil-fed animals, renal betaine contents generally osmolyte total appeared deficient (Fig. 3 middle). This could increased (Tables 4 to 6, ANOVA values; Figs. 2A, C). Previously suggest cell shrinkage, surprising given that excessive polyol we found the same trend for 10- and 71-day sorbinil treatments of normal Wistar rats [18, 19]. Renal betaine content also rose 85% in sorbinil-galactose-fed Sprague-Dawley rats [16], and also rose in our galactose treatment (Fig. 2B) as discussed above. Betaine has been seen to increase in renal medullas when urea, GPC and inositol contents were reduced by salt loading or low-protein diets [29, 30]. Thus renal cells may be programmed to use betaine as a substitute osmolyte for volume regulation when other osmolytes build-up is generally thought to cause cell swelling. However, the kidney appears to adapt later to this, with Na contents lower and total cell osmolytes higher at 21 and 42 days than at 10 days (Fig. 3 middle). Lenticular galactose and galactitol accumulation was much reduced at 42 days compared to 10 days (Table 3). Repeated increases and decreases with time of galactose feeding have been reported in rat lenses, but the reasons are unclear [35]. are reduced. As a methylamine, however, betaine has different Second, in sorbinil feeding alone for 42 days, inositol and GPC properties (such as urea counteraction) than polyols [33], so that also increased with betaine (Table 6 and Fig. 2A; the apparently this substitution may not be without problems. higher total of cell osmolytes was not significantly so; Fig. 3). We The mechanism for this process is unknown. Betaine accumu- only found increased betaine in 71 days sorbinil feeding [19], lation is stimulated in renal cells in vitro by external hypertonicity suggesting the 42 day pattern is a temporary effect. Sorbinil-only through a Na/Cl-dependent membrane transporter [34]. This feeding tended to cause higher urea levels in urine, plasma and could be stimulated when galactose increases external osmotic kidneys (Tables 1, 2, 5, 6), with the 42 day group having the pressure and/or when other cell osmolytes are suppressed. In our highest levels; this may have triggered the increased GPC (see 42-day Sorbinil + Galactose group, betaine did not rise despite above). This effect of sorbinil suggests increased protein cataboreduced sorbitol and GPC (Fig. 2C). However, since renal Na lism, but we can find no reports of similar effects in the literature, and galactose contents fell (Table 6; Fig. 3), betaine would not be and have no explanation for it. needed for cell volume maintenance, again suggesting that it is Third, in sorbinil-galactose feeding, by 42 days there were used as a compensatory osmolyte specifically in response to cell reduced Na and galactose contents in kidneys and Na in urine shrinkage. (Tables 1C, 6; Fig. 3 right), suggesting a reduced glomerular Betaine compensation is not universal, however. Bondy et al filtration rate (GFR). Total cell osmolytes decreased in apparent [16], using female Sprague-Dawley rats fed sorbinil for 10 days, compensation (Fig. 3). Bank et a! [10] found that, while galactose found papillary sorbitol contents were greatly reduced, but no feeding increased GFR in rats, 10 to 14 days sorbinil + galactose 352 Edmands et al: Sorbinil, galactosemia and osmolytes feeding (but not sorbinil alone) reduced it below that of controls. The mechanism for this was not determined, but a stimulation of vascular smooth muscle by sorbinil was suggested [10]. Finally, there may developmental changes in AR-type activities. Animals at 21 days of testing were chronologically about 73 days old, about at the age of sexual maturation. Compared to younger or older animals (10 day and 42 day groups), they had higher AR activities as measured by glyceraldehyde (Fig. 1), more galactitol accumulation (Fig. 2B; Tables 4 to 6), and less of a reduction in polyols under sorbinil treatments (Figs. 2A, C). Because xylosebased activity did not differ among groups (Fig. 1), some of the apparent AR activity at 21 days may be attributable to aldehyde reductase (A1R) (Results), or perhaps to a separate "high Kin" aldose reductase recently reported in dog renal medulla [36]. Both enzymes can use galactose as a substrate (though not nearly as well as does aldose reductase), and both are less sensitive to sorbinil [36, 37]. This might explain the greater resistance to 2. YANCEY PH, CIitK ME, HAND SC, BowLus RD, SOMERO GN: Living with water stress: Evolution of osmolyte systems. Science 217:1214— 1222, 1982 3. GARCIA-PEREZ A, BURG MB: Renal medullary organic osmolytes. Physiol Rev 71:1081—1114, 1991 4. BAGNASCO 5, BALABAN R, FALEs H, YANG Y-M, BURG M: Predom- inant osmotically active organic solutes in rat and rabbit renal medullas. J Biol Chem 26 1:5872—5877, 1986 5. COwLEY BD JR, FERRARIS JD, CARPER D, BURG MB: In vivo osmotic regulation of aldose reductase mRNA, protein, and sorbitol content in rat renal medulla. Am J Physiol 258:F154—F161, 1990 6. GULLANS SR, BLUMENFELD JD, BALscIn JA, KALETA M, BRENNER RM, HEILIG CW, HEBERT SC: Accumulation of major organic osmolytes in rat renal medulla in dehydration. Am J Physiol 255:F626— F634, 1988 7. YANCEY PH, BURG MB: Distributions of major organic osmolytes in rabbit kidneys in diuresis and antidiuresis. Am J Physiol 257:F602— F607, 1989 8. BURG MB, KADOR PF: Sorbitol, osmoregulation, and the complications of diabetes. J Clin Invest 81:635—640, 1988 9. KINOSHITA JH: Cataracts in galactosemia. Invest Ophthalmol 4:486— 499, 1965 10. BANK N, Coco M, AYNEDJIAN HS: Galactose feeding causes glomerular hyperperfusion: Prevention by aldose reductase inhibition. Am J sorbinil in our 21 day animals, analogous to previous work [37] showing that AR-specific inhibitors reduce polyols in the renal medulla (where AR dominates) more so than in the cortex (where Physiol 256:F994—F999, 1989 both AR and AiR occur). This might also explain our previous results with rats on sorbinil for 21 days (chronologically about 77 11. SCI-IMOLKE M, SCHLEICHER E, GUDER WG: Renal sorbitol, myoinositol and glycerophosphorylcholine in streptozotocin-diabetic rats. days old) which showed no drop in renal sorbitol contents [19]. Eur J Clin Chem Clin Biochem 30:607—614, 1992 Other studies have shown age-related changes in AR. Newborn 12. ENGERMAN RL, KERN TS: Aldose reductase inhibition fails to prevent Sprague-Dawley rats have been have little AR mRNA, but develop high levels by 12 days [38]. AR activity may continue to fall after maturation: in our previous study of rats on sorbinil for 71 days (final age 125 days), medullary AR activities with glyceraldehyde were 34.1 4.5 units in control, 9.4 1.5 units in sorbinil animals [19] (compare values to Fig. 1). Thus there may be lifelong developmental changes in polyol metabolism, with animal age possibly important to the effectiveness of aldose reductase inhibitors. In conclusion, the mammalian renal inner medulla generally appears to be able to alter its total contents of organic osmolytes when subjected to disturbances by galactosemia and aldose reductase treatment. However, the patterns of osmolyte changes are not the same at all animal ages or lengths of treatment. Acknowledgments retinopathy in diabetic and galactosemic dogs. Diabetes 42:820—825, 1993 13. MEYER WR, DOYLE MB, GRIFO IA, LIPETZ KJ, OATES PJ, DECHER- NEY AH, DIAMOND MP: Aldose reductase inhibition prevents galac- tose-induced ovarian dysfunction in the Sprague-Dawley rat. Am J Obstet Gynecol 167:1837—1843, 1992 14. TsAI SC, BURKANIS TG: Aldose reductase inhibitors: An update. Ann Pharinacother 27:751—754, 1993 15. YANCEY PH, BURG MB, BAGNASCO SM: Effects of NaCl, glucose and aldose reductase inhibitors on cloning efficiency of renal cells. Am J Physiol 258:C156—C163, 1990 16. BONDY C, COWLEY BD JR. LIGI-rrMAN SL, KADOR PF: Feedback inhibition of aldose reductase gene expression in rat renal medulla: Galactitol accumulation reduces enzyme messenger RNA levels and depletes cellular inositol content. J Clin Invest 86:1103—1108, 1990 17. MORIYAMA T, GARCIA-PEREZ A, Oi.SoN AD, BURG MB: Intracellular betaine substitutes for sorbitol in protecting renal medullary cells from hypertonicity. Am J Physiol 260:F494—F497, 1991 18. YANCEY PH, HANER RG, FREUDENBERGER T: Effects of an aldose reductase inhibitor on organic osmotic effectors in rat renal medulla. Sorbinil was kindly provided by Pfizer Central Research. This research was supported by funds from the Howard Hughes Medical Institute, M.J. Am J Physiol 259:F733—F738, 1990 19. EDMANDS S, YANCEY PH: Effects on rat renal osmolytes of extended technical assistance. 103C:499—502, 1992 20. UCHIDA S. GARCIA-PEREZ A, Muiu'ny H, BURG MB: The signal for induction of aldose reductase in renal medullary cells by high external NaCI. Am J Physiol 256:C614—C620, 1989 Murdock Charitable Trust and Research Corporation, and a Whitman College Sally Ann Abshire Award. We thank Douglas Peterson for Reprint requests to: Dr. PR. Yancey, Biology Department, Whitman College, Walla Walla, Washington 99362 USA. Appendix treatment with an aldose reductase inhibitor. Comp Biochem Physiol 21. WOLFF SD, YANCEY PH, STANTON TS, BALABAN RS: A simple HPLC method for quantitating the major organic solutes of the renal medulla. Am J Physiol 256:F954—F956, 1989 22. YANCEY PH: Osmotic effectors in kidneys of xeric and mesic rodents: Corticomedullaiy distributions and changes with water availability. J Comp Physiol 158B:369—380, 1988 Abbreviations are: GPC, glycerophosphoryicholine; AR, aldose reductase; AiR, aldehyde reductase; SNK, Student-Newmann-Keuls; MDCK, Madin-Darby canine kidney. 23. BEDFORD JJ, BAGNASCO SM, KADOR PF, HARRIS HW JR, BURG MB: References 24. TERUBAYASHI H, SATO S, NISHIMURA C, KADOR PF, KINOSHITA JS: 1. YANCEY PH: Compatible and counteracting aspects of organic osmolytes in mammalian kidney cells in vivo and in vitro, in Water and mt 36:843—851, 1989 25. BECK FX, SCHMOLKE M, GUDER WG, DORGE A, THURAU K: Os- Life: A Comparative Analysis of Water Relationships at the Organismic, Cellular, and Molecular Levels, edited by GN SossaRo, CB OSMOND, CL Bous, Berlin, Springer-Verlag, 1992, pp 19—32 Characterization and purification of a mammalian osmoregulatory protein, aldose reductase, induced in renal medullary cells by high extracellular NaCI. J Biol Chem 262:14255—14259, 1987 Localization of aldose and aldehyde reductase in the kidney. Kidney molytes in renal medulla during rapid changes in papillary tonicity. Am J Physiol 262:F849—F856, 1992 26. GARTY H, FURLONG TI, Ews DE, SPRING KR: Sorbitol permease: Edmands et al: Sorbinil, galactosemia and osmolytes 353 An apical membrane transporter in cultured renal papillary epithelial feeding provide a valid model of consequences of exaggerated polyolpathway flux in peripheral nerve in experimental diabetes? Diabetes SARnw' E: Hypotonicity-evoked release of organic osmolytes from distal renal cells: Systems, signals and sidedness. Renal Physiol Bio- 33. YANCEY PH, BURG MB: Counteracting effects of urea and betaine on colony-forming efficiency of mammalian cells in culture. Am I Physiol 258:R198—R204, 1990 34. YAMAUCHI A, UCHIDA S, KWON HM, PRESTON AS, ROBEY RB, GARcIA-PEREz A, BURG MB, HANDLER JS: Cloning of a Na- and cells. Am J Physiol 260:F650—F656, 1991 27. KINNE RKH, CZEKAY R-P, GRUNEWALD JM, MOOREN FC, Kiachem 16:66—78, 1993 28. Wu RR, LYONS PA, WANG A, SAINSEURY AJ, CHUNG S, PALMER TN: Effects of galactose feeding on aldose reductase gene expression. I Gun Invest 92:155—159, 1993 29. PETERSON DP, MURPHY KM, URsIN0 R, STREETER K, YANcEY PH: Effects of dietary protein and salt on rat renal osmolytes: Co-variation in urea and glyeerophosphorylcholine contents. Am I Physiol 263: F594—F600, 1992 30. HEIUG CW, STROMSKI ME, Guu.r's SR: Methylamine and polyol responses to salt loading in renal medulla. Am I Physiol 257:F1117— F1123, 1989 31. OLGEMÔLLER B, SCHWAABE S, SCHLEICFIER ED, GERBITZ KD: Up- regulation of myo-inositol transport compensates for competitive inhibition by glucose. An explanation for the inositol paradox? Diabetes 42:1119—1125, 1993 32. WIuRs GB, LAMBOURNE JE, TOMLINSON DR: Does galactose 36:1425—1431, 1987 C1-dependent betaine transporter that is regulated by hypertonicity. I Biol Chem 267:649—652, 1992 35. UNAKAR NJ, TsuI JY, JOHNSON MJ: Effect of aldose reductase inhibitors on lenticular dulcitol level in galactose fed rats. I Ocular Pharm 8:199—212, 1992 36. OHTA M, TANIMOTO T, TAN.& A: Localization, isolation and prop- erties of three NADPH-dependent aldehyde reducing enzymes from dog kidney. Biochim Biophys Acta 1078:395—403, 1991 37. SATO S: Rat kidney aldose reductase and aldehyde reductase and polyol production in rat kidney. Am I Physiol 263:F799—F805, 1992 38. SCHWARTZ GJ, ZAvlLowrrz BJ, RADICE AD, GARCIA-PEREZ A, SANDS JM: Maturation of aldose reductase expression in the neonatal rat inner medulla. I Cliii Invest 90:1275—1283, 1992