The Effect of Diet and Alcohol on the Development of Folate Deficiency in the Rat ROBERT MCGUFFIN, PAULGOFFAND R. S . HILLMAN Division
of Hematology,
University
of Washington Medical School
(Received 30 December 1974; acceptedfor publication 24 March 1975) SUMMARY. Studies of the rate of depletion of serum and tissue methylated and nonmethylated folates were carried out in rats maintained for long periods on either a folate deficient (sucrose-water/sulphathiazole)diet or a deficient diet plus high alcohol intake. By means of implantation of a feeding gastrostomy tube, it was possible to sustain constant blood ethanol levels of between 50 and 300 mg/dl for 3-4 weeks with relatively normal calorie intake and without death of the animal. Using this animal model, which closely resembles severe alcoholism in man, a very rapid depression in serum 5-methyl tetrahydrofolate was observed similar to that reported in alcoholic man. At the same time, release of folates from liver stores was unimpaired by alcohol ingestion. Liver folate store depletion rates were identical for alcoholic and folate starved animals. The explanation for the sudden alcohol suppression of serum folate levels must, therefore, be sought at a point in the internal metabolic sequences of folate other than the delivery of folate stores to plasma. The behaviour of serum folate activity in nian, in the face of both dietary folate deprivation and ingestion of alcohol, has been the subject of a number of reports (Herbert, 1962, 1963; Herbert et al, 1963; Sullivan & Herbert, 1964; Halsted et a], 1967, 1971, 1972; Eichner & Hillman, 1971, 1973; Eichner et al, 1972, 1974; Paine et al, 1973). Herbert’s (1962) classic study of normal man called attention to the limited nature of ‘folate stores’ and the rapid appearance of megaloblastic erythropoiesis with folate deprivation. Subsequent studies not only have confirmed this observation but also have demonstrated an additive effect of alcohol on the rate of induction of folate deficiency and abnormal erythropoiesis (Herbert et al, 1963 ; Sullivan & Herbert, 1964; Eichner & Hillman, 1971; Eichner et al, 1972, 1974). The latter may in part be related to organ damage in the alcoholic or intestinal malabsorption of folic acid secondary to the local toxicity of the alcohol (Halsted et al, 1967, 1971, 1972; Eichner et al, 1972). However, Eichner & Hillman (1973) and Paine et al(1g73) have shown that serum folate levels fall to deficient levels within hours of oral or intravenous administration of moderate amounts of chemically pure ethanol despite normal dietary and liver folate store patterns. As a part of this work, detailed studies of the effects of alcohol and alcohol metabolites on the Lactobacillus casei assay failed to reveal direct inhibition, raising the question of a separate effect of alcohol on the internal metabolism and storage of folate metabolites. In order to study this phenomenon in more depth, it was felt important to establish an animal model where serum folate levels and tissue stores could be measured while maintaining Correspondence:Dr R. S. Hillman, Division of Hematology, University of Washington Medical School, BB Sciences, University of Washington, Seattle, Washington 98195, U S A .
I229 Health
185
I 86
R.McGufin, P.G@and R.S. Hillman
blood ethanol levels similar to those seen in a human alcoholic population. In the past, other workers have used animal models to examine the influence of dietary deprivation on both serum and tissue levels of folate (Briggs, 1959;Grossowicz et a\, 1964;Chanarin et al, 1969). Briggs (1959)reported a successful system for induction of folate deficiency in the mouse but did not study tissue folate levels. Chanarin et al (1969)found that rats maintained on a folate-free diet containing sulphanilamide experienced a 90% depletion of serum folate in 9 weeks, even though it took only 3 weeks to reach a 90% depletion of ‘folate’ in the liver. Grossowicz et nl (1964,using a similar model, demonstrated an even more rapid depletion of ‘tissue folate.’ In both studies, folate measurements were limited to those metabolites measured by L. casei and/or Pediococcus cerevisiae without apparent conjugase treatment. The present study expands upon this approach both as to the identification of specific metabolites, especially the separation of 5-methyl H, PteGlu 1-7(5 methyltetrahydropteroyl glutamates one to seven) from other folates, and the description of the rates of tissue and serum folate depletion with diet deprivation. In addition, a reliable model for investigation of the acute effects of alcohol on blood and tissue folate metabolism is described together with studies confirming the specific rapid effect of ethanol on the serum folate level. MATERIALS AND METHODS Two complete studies were carried out over an 8 month period, using Sprague-Dawley female rats weighing 150-250 g. In the first study, folate deficiency was induced by restricting animals to a liquid diet of 25% sucrose in water to which IOO mg succinylsulphathiazole/~oo ml was added to suppress intestinal production of folates. In addition, coprophagy was prevented by isolation of individual animals in metabolic cages. Control animals were maintained on standard Purina rat lab chow. They demonstrated a 40% gain in liver weight and a 42% increase in total liver folate activity during the 26 days of the study, while sugarwater rats showed a 14% reduction in liver weight and a 95% reduction in liver folate content. At least eight animals on the sucrose-water diet with paired controls were sacrificed at zero time and on days 3, 6,9,13, 18 and 25 of the study. In the second study, a third group of animals was allowed the sucrose-water and succinylaclphathiazole liquid diet ad libitum, while receiving a daily supplement of ethanol by instillation through a surgically placed gastrostoniy tube. Following abdominal incision, a 0.065 inch O D silastic (Dow Corning Corporation) tubing was introduced into the stomach, sewn in place, and then carried subcutaneously to an exit point just behind the head. The tube was kept sealed with fish line except during the instillation of 5-6 ml of the ethanol solution by syringe. The administration of roo ml/kg of 10% ethanol in three or four divided doses resulted in peak blood ethanol level of 200-300 mg/dl, falling to 50-150 nig/dl at 8 h. Gastrostomy tubes were also implanted in sucrose-water control animals. At least seven animals from both the sucrose-water control and ethanol groups plus three regular diet control animals were sacrificed on days I, 3, 6,9,12, 15 and 18 of the study for blood and tissue folate measurements. In both studies, following anaesthesia with sodium pentobarbital, the abdominal wall was incised, portal vein cannulated and 2 ml of blood drawn for serum folate determinations. The liver was then flushed with 40 ml of chilled 0.1 M sodium phosphate buffer, pH 6.0,to
Alcohol and Fohte Deficiency 187 which I g of sodium ascorbateldl had been added. Next, liver and kidneys (first study only) were removed, weighed, and four 10-20 mg tissue samples excised from each. After obtaining an exact weight, each biopsy was placed in a 1-2 m10.1 M phosphate buffer at pH 6.0 with I g/dl sodium ascorbate, homogenized in a ground glass homogenizer and then frozen and thawed four times using dry ice and acetone and a 37°C water bath over a 2 h period. Following this process, the samples were stored at -20°C until the day of assay. At that time they were autoclaved at 10 psi for 10 min to precipitate protein, centrifuged, and the supernatant diluted to 10 ml with phosphate buffer and 0.2-0.5 ml pipetted in duplicate for microbiological assay with both L. cusei and Streptococcus faeculis. Eight measurements
Days
FIG I. Compared to the liver folate levels of control animals (o), liver (0), kidney (A) and serum ( 0 ) levels of 5-methyl H, PteGlu in sugar-water/sulphathiazoleanimals fell rapidly. The rates of fall for tissue and serum levels were virtually identical with a T+of approximately 6 days.
were performed on each organ with a minimum of seven animals at each time point. Serum folate determinations were performed in duplicate. L. cusei and S . fueculis assays were performed according to the aseptic technique described by Herbert (1966). In view of the described tissue preparation technique for activation of liver conjugase, the two microbiological assays would be expected to measure the major portion of liver and kidney folate stores, all methylated and non-methylated monoglutamates, with partial response to di-glutamates and no response to any residual polyglutamates of greater size (Bennet et al, 1964; Bird et a!, 1965; Tamura el al, 1972). Total active folate in liver (L. cusei) at zero time and for control animals ranged from 75 to 95 pg (7.5-9.5 &g), a range of values which would reflect measurement of better than 80% of the reported total folate content of liver as studied by Thenen et ul (1973) in rats and Corrocher et a1 (1972) in the guinea-pig. Data were analysed assuming S. faeculis activity as representative of non-methylated folates (glutamate derivatives of H4 Pte, 10 formyl H4 Pte, 5 formimino H4 Pte, 5-10 methylene H4 Pte and Pte), while 5-methyl folate levels were determined as the difference between L.
R. McGufin, P. G o f a n d R. S . Hillman
188
cusei and S.faeculis values. Significance was tested using the Wilcoxon Signed Ranks statistical method in order to adjust for differences in original serum folate values. All measurements of blood ethanol levels were performed by a modification of the alcohol dehydrogenase method (Eichner & Hillman, 1973).
RESULTS When animals were placed on the sucrose-water/sulphathiazolediet, both serum and tissue levels of 5-methyl H, Pte Glu and its polyglutamate derivatives (in the case of liver predominantly penta- and hepta-glutamates) fell rapidly (Fig I). A 50% reduction in folate
100
80 0
-a
> 0
._ 60 .-01 L 0 r
E 40
z
L
a" 20
0
4
12
8 Days
FIG2
16
20
0
4
B
12
16
20
Doys
FIG 3
FIG2. The rate of fall of 5-methyl HLPteGlu, levels in the livers of alcoholic rats (0) was identical to the rates of fall observed in sugar-water animals (A) from study I , and sucrose-water control animals ( 0 )from study 2. Normal diet controls are shown as open squares. FIG 3. Serum 5-methyl HIPteGlu levels of alcoholic animals ( 0 )fell more rapidly then serum levels from study I sucrose-water rats(0) and study 2 sucrose-water control animals(a).(At day 3, P < 0.005, at day 6, P = 0.05). Normal diet controls are shown as open squares.
content was appreciated in liver, kidney and serum by the sixth day. Subsequently, the rate of depletion slowed and was not complete untiI the zoth-24th day. Measurements of nonmethylated folates (S.faeculis activity) in liver and kidney revealed a similar decline with 50% reduction in the first 6 days. In some animals a rise in S.faecalis values then occurred on or about the ninth day, followed by a second decline, though never below 50% of the initial value. The majority of animals, however, demonstrated comparable reductions in both methylated and non-methylated folates. While serum levels of S.faecalis activity appeared to follow a similar pattern, the extremely low values appreciated in serum made accurate measurement difficult. In the second study, normal and sucrose-water control animals were compared to animals receiving sucrose-water plus divided doses of ethanol sufficient to sustain a continuous blood
Alcohol and Folate Deficiency
189
ethanol level between 50 and 300 mgldl. The rats maintained on sucrose-water plus ethanol showed no difference in the rate of decline of liver 5-methyl H, PteGlu, (n = 1-7) as compared to sucrose-water control rats (Fig 2). Similarly, there was no appreciable difference between the two groups in the behaviour of liver non-methylated folate levels. However, the ethanol group did show a more rapid decline of their serum 5-methyl H, PteGlu levels (Fig 3). By the third day, serum values had fallen below 30% of the original value in the ethanol group, while the sugar-water diet group maintained levels greater than 60% of the Starved
Alcohol Day A
0
w
3
6
A T
T
6.316.3
1*9/31
L T
L
T
I-6/ 2.9
1.4/2*2
FIG 4. The relative sizes of non-methylated liver folate stores (heavy shading) and methylated (5methyl H4PteGlu,) liver stores (light shading) are compared for alcohol-treated animals and folatestarved, sucrose-water rats, study 2. The rate of shrinkage of these storage pools is virtually identical for the two groups of animals, despite major differencesin serum values of 5-methyl H4PteGlu at day 3 , 6 and g (central numbers).
original value (P
of Hematology,
University
of Washington Medical School
(Received 30 December 1974; acceptedfor publication 24 March 1975) SUMMARY. Studies of the rate of depletion of serum and tissue methylated and nonmethylated folates were carried out in rats maintained for long periods on either a folate deficient (sucrose-water/sulphathiazole)diet or a deficient diet plus high alcohol intake. By means of implantation of a feeding gastrostomy tube, it was possible to sustain constant blood ethanol levels of between 50 and 300 mg/dl for 3-4 weeks with relatively normal calorie intake and without death of the animal. Using this animal model, which closely resembles severe alcoholism in man, a very rapid depression in serum 5-methyl tetrahydrofolate was observed similar to that reported in alcoholic man. At the same time, release of folates from liver stores was unimpaired by alcohol ingestion. Liver folate store depletion rates were identical for alcoholic and folate starved animals. The explanation for the sudden alcohol suppression of serum folate levels must, therefore, be sought at a point in the internal metabolic sequences of folate other than the delivery of folate stores to plasma. The behaviour of serum folate activity in nian, in the face of both dietary folate deprivation and ingestion of alcohol, has been the subject of a number of reports (Herbert, 1962, 1963; Herbert et al, 1963; Sullivan & Herbert, 1964; Halsted et a], 1967, 1971, 1972; Eichner & Hillman, 1971, 1973; Eichner et al, 1972, 1974; Paine et al, 1973). Herbert’s (1962) classic study of normal man called attention to the limited nature of ‘folate stores’ and the rapid appearance of megaloblastic erythropoiesis with folate deprivation. Subsequent studies not only have confirmed this observation but also have demonstrated an additive effect of alcohol on the rate of induction of folate deficiency and abnormal erythropoiesis (Herbert et al, 1963 ; Sullivan & Herbert, 1964; Eichner & Hillman, 1971; Eichner et al, 1972, 1974). The latter may in part be related to organ damage in the alcoholic or intestinal malabsorption of folic acid secondary to the local toxicity of the alcohol (Halsted et al, 1967, 1971, 1972; Eichner et al, 1972). However, Eichner & Hillman (1973) and Paine et al(1g73) have shown that serum folate levels fall to deficient levels within hours of oral or intravenous administration of moderate amounts of chemically pure ethanol despite normal dietary and liver folate store patterns. As a part of this work, detailed studies of the effects of alcohol and alcohol metabolites on the Lactobacillus casei assay failed to reveal direct inhibition, raising the question of a separate effect of alcohol on the internal metabolism and storage of folate metabolites. In order to study this phenomenon in more depth, it was felt important to establish an animal model where serum folate levels and tissue stores could be measured while maintaining Correspondence:Dr R. S. Hillman, Division of Hematology, University of Washington Medical School, BB Sciences, University of Washington, Seattle, Washington 98195, U S A .
I229 Health
185
I 86
R.McGufin, P.G@and R.S. Hillman
blood ethanol levels similar to those seen in a human alcoholic population. In the past, other workers have used animal models to examine the influence of dietary deprivation on both serum and tissue levels of folate (Briggs, 1959;Grossowicz et a\, 1964;Chanarin et al, 1969). Briggs (1959)reported a successful system for induction of folate deficiency in the mouse but did not study tissue folate levels. Chanarin et al (1969)found that rats maintained on a folate-free diet containing sulphanilamide experienced a 90% depletion of serum folate in 9 weeks, even though it took only 3 weeks to reach a 90% depletion of ‘folate’ in the liver. Grossowicz et nl (1964,using a similar model, demonstrated an even more rapid depletion of ‘tissue folate.’ In both studies, folate measurements were limited to those metabolites measured by L. casei and/or Pediococcus cerevisiae without apparent conjugase treatment. The present study expands upon this approach both as to the identification of specific metabolites, especially the separation of 5-methyl H, PteGlu 1-7(5 methyltetrahydropteroyl glutamates one to seven) from other folates, and the description of the rates of tissue and serum folate depletion with diet deprivation. In addition, a reliable model for investigation of the acute effects of alcohol on blood and tissue folate metabolism is described together with studies confirming the specific rapid effect of ethanol on the serum folate level. MATERIALS AND METHODS Two complete studies were carried out over an 8 month period, using Sprague-Dawley female rats weighing 150-250 g. In the first study, folate deficiency was induced by restricting animals to a liquid diet of 25% sucrose in water to which IOO mg succinylsulphathiazole/~oo ml was added to suppress intestinal production of folates. In addition, coprophagy was prevented by isolation of individual animals in metabolic cages. Control animals were maintained on standard Purina rat lab chow. They demonstrated a 40% gain in liver weight and a 42% increase in total liver folate activity during the 26 days of the study, while sugarwater rats showed a 14% reduction in liver weight and a 95% reduction in liver folate content. At least eight animals on the sucrose-water diet with paired controls were sacrificed at zero time and on days 3, 6,9,13, 18 and 25 of the study. In the second study, a third group of animals was allowed the sucrose-water and succinylaclphathiazole liquid diet ad libitum, while receiving a daily supplement of ethanol by instillation through a surgically placed gastrostoniy tube. Following abdominal incision, a 0.065 inch O D silastic (Dow Corning Corporation) tubing was introduced into the stomach, sewn in place, and then carried subcutaneously to an exit point just behind the head. The tube was kept sealed with fish line except during the instillation of 5-6 ml of the ethanol solution by syringe. The administration of roo ml/kg of 10% ethanol in three or four divided doses resulted in peak blood ethanol level of 200-300 mg/dl, falling to 50-150 nig/dl at 8 h. Gastrostomy tubes were also implanted in sucrose-water control animals. At least seven animals from both the sucrose-water control and ethanol groups plus three regular diet control animals were sacrificed on days I, 3, 6,9,12, 15 and 18 of the study for blood and tissue folate measurements. In both studies, following anaesthesia with sodium pentobarbital, the abdominal wall was incised, portal vein cannulated and 2 ml of blood drawn for serum folate determinations. The liver was then flushed with 40 ml of chilled 0.1 M sodium phosphate buffer, pH 6.0,to
Alcohol and Fohte Deficiency 187 which I g of sodium ascorbateldl had been added. Next, liver and kidneys (first study only) were removed, weighed, and four 10-20 mg tissue samples excised from each. After obtaining an exact weight, each biopsy was placed in a 1-2 m10.1 M phosphate buffer at pH 6.0 with I g/dl sodium ascorbate, homogenized in a ground glass homogenizer and then frozen and thawed four times using dry ice and acetone and a 37°C water bath over a 2 h period. Following this process, the samples were stored at -20°C until the day of assay. At that time they were autoclaved at 10 psi for 10 min to precipitate protein, centrifuged, and the supernatant diluted to 10 ml with phosphate buffer and 0.2-0.5 ml pipetted in duplicate for microbiological assay with both L. cusei and Streptococcus faeculis. Eight measurements
Days
FIG I. Compared to the liver folate levels of control animals (o), liver (0), kidney (A) and serum ( 0 ) levels of 5-methyl H, PteGlu in sugar-water/sulphathiazoleanimals fell rapidly. The rates of fall for tissue and serum levels were virtually identical with a T+of approximately 6 days.
were performed on each organ with a minimum of seven animals at each time point. Serum folate determinations were performed in duplicate. L. cusei and S . fueculis assays were performed according to the aseptic technique described by Herbert (1966). In view of the described tissue preparation technique for activation of liver conjugase, the two microbiological assays would be expected to measure the major portion of liver and kidney folate stores, all methylated and non-methylated monoglutamates, with partial response to di-glutamates and no response to any residual polyglutamates of greater size (Bennet et al, 1964; Bird et a!, 1965; Tamura el al, 1972). Total active folate in liver (L. cusei) at zero time and for control animals ranged from 75 to 95 pg (7.5-9.5 &g), a range of values which would reflect measurement of better than 80% of the reported total folate content of liver as studied by Thenen et ul (1973) in rats and Corrocher et a1 (1972) in the guinea-pig. Data were analysed assuming S. faeculis activity as representative of non-methylated folates (glutamate derivatives of H4 Pte, 10 formyl H4 Pte, 5 formimino H4 Pte, 5-10 methylene H4 Pte and Pte), while 5-methyl folate levels were determined as the difference between L.
R. McGufin, P. G o f a n d R. S . Hillman
188
cusei and S.faeculis values. Significance was tested using the Wilcoxon Signed Ranks statistical method in order to adjust for differences in original serum folate values. All measurements of blood ethanol levels were performed by a modification of the alcohol dehydrogenase method (Eichner & Hillman, 1973).
RESULTS When animals were placed on the sucrose-water/sulphathiazolediet, both serum and tissue levels of 5-methyl H, Pte Glu and its polyglutamate derivatives (in the case of liver predominantly penta- and hepta-glutamates) fell rapidly (Fig I). A 50% reduction in folate
100
80 0
-a
> 0
._ 60 .-01 L 0 r
E 40
z
L
a" 20
0
4
12
8 Days
FIG2
16
20
0
4
B
12
16
20
Doys
FIG 3
FIG2. The rate of fall of 5-methyl HLPteGlu, levels in the livers of alcoholic rats (0) was identical to the rates of fall observed in sugar-water animals (A) from study I , and sucrose-water control animals ( 0 )from study 2. Normal diet controls are shown as open squares. FIG 3. Serum 5-methyl HIPteGlu levels of alcoholic animals ( 0 )fell more rapidly then serum levels from study I sucrose-water rats(0) and study 2 sucrose-water control animals(a).(At day 3, P < 0.005, at day 6, P = 0.05). Normal diet controls are shown as open squares.
content was appreciated in liver, kidney and serum by the sixth day. Subsequently, the rate of depletion slowed and was not complete untiI the zoth-24th day. Measurements of nonmethylated folates (S.faeculis activity) in liver and kidney revealed a similar decline with 50% reduction in the first 6 days. In some animals a rise in S.faecalis values then occurred on or about the ninth day, followed by a second decline, though never below 50% of the initial value. The majority of animals, however, demonstrated comparable reductions in both methylated and non-methylated folates. While serum levels of S.faecalis activity appeared to follow a similar pattern, the extremely low values appreciated in serum made accurate measurement difficult. In the second study, normal and sucrose-water control animals were compared to animals receiving sucrose-water plus divided doses of ethanol sufficient to sustain a continuous blood
Alcohol and Folate Deficiency
189
ethanol level between 50 and 300 mgldl. The rats maintained on sucrose-water plus ethanol showed no difference in the rate of decline of liver 5-methyl H, PteGlu, (n = 1-7) as compared to sucrose-water control rats (Fig 2). Similarly, there was no appreciable difference between the two groups in the behaviour of liver non-methylated folate levels. However, the ethanol group did show a more rapid decline of their serum 5-methyl H, PteGlu levels (Fig 3). By the third day, serum values had fallen below 30% of the original value in the ethanol group, while the sugar-water diet group maintained levels greater than 60% of the Starved
Alcohol Day A
0
w
3
6
A T
T
6.316.3
1*9/31
L T
L
T
I-6/ 2.9
1.4/2*2
FIG 4. The relative sizes of non-methylated liver folate stores (heavy shading) and methylated (5methyl H4PteGlu,) liver stores (light shading) are compared for alcohol-treated animals and folatestarved, sucrose-water rats, study 2. The rate of shrinkage of these storage pools is virtually identical for the two groups of animals, despite major differencesin serum values of 5-methyl H4PteGlu at day 3 , 6 and g (central numbers).
original value (P
