BIOLPSYCHIATRY
315
1992;31:315-321
Dietary Calcium Blocks Lithium Toxicity in Hamsters without Affecting Circadian Rhythms Harry Klemfuss, Torsten T. Bauer, Kerry E. Greene, and Daniel F. Kripke
Lithium can be toxic to rodents at plasma concentrations (0.6-1.2 retool~L) that also phase delay circadian rhythms. In hamsters, raising the concentration of calcium in the dietfrom 0.! %-3% reduced weight loss and polydipsia caused by 0.4% lithium carbonate. Calcium ingession did not affect plasma lithium concentration or the phase of the circadian wheel-running rhythm in lithium-treated animals. Calcium ingestion did not alter weight gain, salt or water intake, or circadian rhythms in hamsters not receiving lithium. Dietary calcium supplementation may prevent some toxic effects of lithium without interfering with other central nervous system actions.
Introduction Lithium can cause toxic effects at plasma concentrations within the human therapeutic range. Polyuria, gastrointestinal disturbances, and tremor are common adverse effect,~ ~f lithium treatment in patients taking lithium for affective disorders, as well as in experimental animals, in rats, Olesen et al (1975) demonstrated that plasma lithium concenCations in the range of 0.6-1.2 mmol/L can be produced without apparent toxicity if adequate sodium is available and supplemental potassium is added to a standard rodent diet. We have observed that dietary potassium supplementatio~ in rats (Klemfuss and Kripke 198'7) and hamsters (Klemfuss and Kripke 1989) sufficient to prevent weight loss and mortality caused by lithium treatment did not alter plasma lithium concentration or lithium-induced delay of the circadian rhythm of wheel~rmming. These results support the hypothesis that addition of potassium to the diet of patients taking lithium might increase the therapeutic index of this drag (Schou 1982; Olesen 1984; Cater 1986). Recently it has been reported that a small increase in potassium intake (16-20 mmol/ day) ameliorated lithium-induced tremor and edema in one patient (Cummings et al 1988) and prevented polyuria and subjective thi~t in a sample of four lithium patients (Tripuraneni 1990). The reduction in lithium toxicity seen with potassium supplementation in rodents and in these patients encouraged us to examine whether ingestion of other
From the Veterans Affairs Medical Center (HK, DFK) San Diego, the Department of Psychiatry (HK, TrB, DFK), and Neuroscience (KEG), University of California, San Diego, and Universi~t Bonn (TrB), Bonn, Germany. Address reprint requests to Harry Klemfuss, Research Service V-151, Veterans Affairs Medical Center, 3350 La Jolla Village Drive, San Diego, CA 92161. Received April 18, 1991; revised August 2, 1991. © 1992 Society of Biological PsychiaU7
0006-3223/92/$05.00
316
BIOL PSYCHIATRY 1992;31:315-321
H. Klemfuss et al
cations might differentiz!ly influence toxic and chronobiological actions of lithium in hamsters. Physiological interactions between lithium and calcium might be anticipated based on physicochemical similarities between the two ions (Williams 1973). Lithium interferes with the action of many calcium-dependent enzymes at therapeutic concentrations, and could thereby influence many aspects of peripheral as well as central nervous system functioning (Meltzer 1986, !990; Berridge et al 1989). Chronic lithium treatment has been reported to increase p]as~a c~.lcium ~n l~a~.an.~, a~.~ r,:,.d~,~ (Cmma,~'-,~.~al !98!; McEachron et al 1982; Linder et al 1989). Alterations in plasma and cerebrospinal fluid calcium concentrations have been associated with the expression of affective disorders (Dubovsky and Franks 1983; Carman and Wyatt 1979), and may predict the success of lithium therapy (Carman et al 1974; Linder et al 1989). Thus, ~ateractions with calcium may account for some of lithium's biological actions. The present study was intended to determine whether dietary calcium influences the toxic and the circadian rhythm actions of lithium in the hamster.
Methods In each of two replications carried out in January and June of 1990, 30 adult male Syrian Golden hamsters were purchased from Charles River (Wilmington, MA) at 110-120 g, and were maintained in group cages under a schedule of 10 hr of light at 30 lux and 14 hr of total darkness per day (LD 10:14). Animals were randomly assigned to a standard hamster diet modified to contain 0.1% calcium (TD89347, Teklad, Madison, Wl) or the same diet with 3% calcium (TD89349, in which 75 g calcium carbonate (CaCO3) replaced the same weight of nonnutritive polydextrose filler in each kg of diet). Simultaneously (first replication) or after 2 weeks on the low or high calcium diets (second replication), half of the animals in each group were given diets containing the same calcium concentration plus 0.4% lithium carbonate (Li2CO~) (TD89346, TD89348). Both tap water and 0.4% sodium chloride (NaCi) were available throughout the experiment. Following 2 weeks of lithium treatment, each hamster was transferred to an individual cage containing a running wheel. The same dietary and lighting regimen was continued for another 2 weeks. The number of revolutions of the wheel was recorded in 5-min segments by a computer system that also identified the transition between intervals of low running activity (less than the animal's mean) and intervals of running activity greater than the mean. Thi~ data was used to estimate the average phase of the circadian locomotor activity rhythm during the 2 weeks in the running-wheel cages. Body weight, water intake, and saline intake were recorded every 7 days. A:'ter 2 weeks of LD 10:14 recording, hamsters remained in the running-wheel cages for an additional 2 weeks in constant darkness, with the same diet, and then were anesthetized at a time corresponding to the midpoint of each animal's inactive period. Blood was removed by puncture of the inferior vena cava and immediately heparinized and centrifuged. Subsequent measurement of plasma lithium and calcium by atomic absorption spectrophotometry used added lanthanum to prevent cross reactivity. Statistical significance was tested with analyses of variance and, when appropriate, the NewmanKeuls test for multiple comparisons (SPSS/PC + V2.0), and expressed as mean _ SEM. Some of the data from the first replication have been presented in summary form (Greene et al 1990; Klemfuss and Greene 1991).
Lithium-Calcium Interaction
BIOL PSYCHIATRY
317
1992;31:315-321
i10 ¢O~
Figure 1. Calcium supplementation decreased lithium-induced weight loss. In the present study, hamsters were fed diets containing either 0.1% or 3.0% calcium (Ca) without lithium or similar diets containing 0 . 4 % lithium carbonate (Li). Body weight at the end o f each week is presented as the percentage of each animal's weight prior to lithium treatment (mean _+ SEM). For diets without lithium n = i3 and for diets with lithium n - 16. Data from a study with a sindlar protocol, in which hamsters were fed diets cont~r, ing a standard calcium concentration (0.675%) for 4 weeks, are included for comparison (n =
100 m U
"F: m qO
90
@
13.
80 V 0.1~Ca + Li m 0 . 7 ~ C o + Li Q 3.0~ Co + I.i I
m/group).
I
I
I
I
1
2
3
4
Weeks of Lithium
Results The toxic and circadian rhythm actions of lithium were differentially affected. ;~y the concentration of calcium in the diet. In both replications, hamsters given the low-calcium diet began to lose weight within the first week of lithium treatment, losing over 20% of starting weight over 4 weeks (Figure 1). This change is comparable to weight loss in hamsters fed a standard diet containing 0.675% calcium (data from Klemfuss and Kripke 1989). Animals given the high-calcium diet maintained weight near baseline, and weighed significantly more than animals given lithium and low calcium during the third and fourth weeks of lithium treatment (Table 1). At the conclusion of the second replication, the cage cleaner, who was unaware of diet treatments, noted evidence of diarrhea in five of
Table 1. Summary of Diet Effects°
Diet 0.1% Ca
Body weight (g)
3.0% Ca 0.1% Ca _ Li 3.0% Ca -- Li
136 138 108 119
-+ +-
F df p
10.6 56
315
1992;31:315-321
Dietary Calcium Blocks Lithium Toxicity in Hamsters without Affecting Circadian Rhythms Harry Klemfuss, Torsten T. Bauer, Kerry E. Greene, and Daniel F. Kripke
Lithium can be toxic to rodents at plasma concentrations (0.6-1.2 retool~L) that also phase delay circadian rhythms. In hamsters, raising the concentration of calcium in the dietfrom 0.! %-3% reduced weight loss and polydipsia caused by 0.4% lithium carbonate. Calcium ingession did not affect plasma lithium concentration or the phase of the circadian wheel-running rhythm in lithium-treated animals. Calcium ingestion did not alter weight gain, salt or water intake, or circadian rhythms in hamsters not receiving lithium. Dietary calcium supplementation may prevent some toxic effects of lithium without interfering with other central nervous system actions.
Introduction Lithium can cause toxic effects at plasma concentrations within the human therapeutic range. Polyuria, gastrointestinal disturbances, and tremor are common adverse effect,~ ~f lithium treatment in patients taking lithium for affective disorders, as well as in experimental animals, in rats, Olesen et al (1975) demonstrated that plasma lithium concenCations in the range of 0.6-1.2 mmol/L can be produced without apparent toxicity if adequate sodium is available and supplemental potassium is added to a standard rodent diet. We have observed that dietary potassium supplementatio~ in rats (Klemfuss and Kripke 198'7) and hamsters (Klemfuss and Kripke 1989) sufficient to prevent weight loss and mortality caused by lithium treatment did not alter plasma lithium concentration or lithium-induced delay of the circadian rhythm of wheel~rmming. These results support the hypothesis that addition of potassium to the diet of patients taking lithium might increase the therapeutic index of this drag (Schou 1982; Olesen 1984; Cater 1986). Recently it has been reported that a small increase in potassium intake (16-20 mmol/ day) ameliorated lithium-induced tremor and edema in one patient (Cummings et al 1988) and prevented polyuria and subjective thi~t in a sample of four lithium patients (Tripuraneni 1990). The reduction in lithium toxicity seen with potassium supplementation in rodents and in these patients encouraged us to examine whether ingestion of other
From the Veterans Affairs Medical Center (HK, DFK) San Diego, the Department of Psychiatry (HK, TrB, DFK), and Neuroscience (KEG), University of California, San Diego, and Universi~t Bonn (TrB), Bonn, Germany. Address reprint requests to Harry Klemfuss, Research Service V-151, Veterans Affairs Medical Center, 3350 La Jolla Village Drive, San Diego, CA 92161. Received April 18, 1991; revised August 2, 1991. © 1992 Society of Biological PsychiaU7
0006-3223/92/$05.00
316
BIOL PSYCHIATRY 1992;31:315-321
H. Klemfuss et al
cations might differentiz!ly influence toxic and chronobiological actions of lithium in hamsters. Physiological interactions between lithium and calcium might be anticipated based on physicochemical similarities between the two ions (Williams 1973). Lithium interferes with the action of many calcium-dependent enzymes at therapeutic concentrations, and could thereby influence many aspects of peripheral as well as central nervous system functioning (Meltzer 1986, !990; Berridge et al 1989). Chronic lithium treatment has been reported to increase p]as~a c~.lcium ~n l~a~.an.~, a~.~ r,:,.d~,~ (Cmma,~'-,~.~al !98!; McEachron et al 1982; Linder et al 1989). Alterations in plasma and cerebrospinal fluid calcium concentrations have been associated with the expression of affective disorders (Dubovsky and Franks 1983; Carman and Wyatt 1979), and may predict the success of lithium therapy (Carman et al 1974; Linder et al 1989). Thus, ~ateractions with calcium may account for some of lithium's biological actions. The present study was intended to determine whether dietary calcium influences the toxic and the circadian rhythm actions of lithium in the hamster.
Methods In each of two replications carried out in January and June of 1990, 30 adult male Syrian Golden hamsters were purchased from Charles River (Wilmington, MA) at 110-120 g, and were maintained in group cages under a schedule of 10 hr of light at 30 lux and 14 hr of total darkness per day (LD 10:14). Animals were randomly assigned to a standard hamster diet modified to contain 0.1% calcium (TD89347, Teklad, Madison, Wl) or the same diet with 3% calcium (TD89349, in which 75 g calcium carbonate (CaCO3) replaced the same weight of nonnutritive polydextrose filler in each kg of diet). Simultaneously (first replication) or after 2 weeks on the low or high calcium diets (second replication), half of the animals in each group were given diets containing the same calcium concentration plus 0.4% lithium carbonate (Li2CO~) (TD89346, TD89348). Both tap water and 0.4% sodium chloride (NaCi) were available throughout the experiment. Following 2 weeks of lithium treatment, each hamster was transferred to an individual cage containing a running wheel. The same dietary and lighting regimen was continued for another 2 weeks. The number of revolutions of the wheel was recorded in 5-min segments by a computer system that also identified the transition between intervals of low running activity (less than the animal's mean) and intervals of running activity greater than the mean. Thi~ data was used to estimate the average phase of the circadian locomotor activity rhythm during the 2 weeks in the running-wheel cages. Body weight, water intake, and saline intake were recorded every 7 days. A:'ter 2 weeks of LD 10:14 recording, hamsters remained in the running-wheel cages for an additional 2 weeks in constant darkness, with the same diet, and then were anesthetized at a time corresponding to the midpoint of each animal's inactive period. Blood was removed by puncture of the inferior vena cava and immediately heparinized and centrifuged. Subsequent measurement of plasma lithium and calcium by atomic absorption spectrophotometry used added lanthanum to prevent cross reactivity. Statistical significance was tested with analyses of variance and, when appropriate, the NewmanKeuls test for multiple comparisons (SPSS/PC + V2.0), and expressed as mean _ SEM. Some of the data from the first replication have been presented in summary form (Greene et al 1990; Klemfuss and Greene 1991).
Lithium-Calcium Interaction
BIOL PSYCHIATRY
317
1992;31:315-321
i10 ¢O~
Figure 1. Calcium supplementation decreased lithium-induced weight loss. In the present study, hamsters were fed diets containing either 0.1% or 3.0% calcium (Ca) without lithium or similar diets containing 0 . 4 % lithium carbonate (Li). Body weight at the end o f each week is presented as the percentage of each animal's weight prior to lithium treatment (mean _+ SEM). For diets without lithium n = i3 and for diets with lithium n - 16. Data from a study with a sindlar protocol, in which hamsters were fed diets cont~r, ing a standard calcium concentration (0.675%) for 4 weeks, are included for comparison (n =
100 m U
"F: m qO
90
@
13.
80 V 0.1~Ca + Li m 0 . 7 ~ C o + Li Q 3.0~ Co + I.i I
m/group).
I
I
I
I
1
2
3
4
Weeks of Lithium
Results The toxic and circadian rhythm actions of lithium were differentially affected. ;~y the concentration of calcium in the diet. In both replications, hamsters given the low-calcium diet began to lose weight within the first week of lithium treatment, losing over 20% of starting weight over 4 weeks (Figure 1). This change is comparable to weight loss in hamsters fed a standard diet containing 0.675% calcium (data from Klemfuss and Kripke 1989). Animals given the high-calcium diet maintained weight near baseline, and weighed significantly more than animals given lithium and low calcium during the third and fourth weeks of lithium treatment (Table 1). At the conclusion of the second replication, the cage cleaner, who was unaware of diet treatments, noted evidence of diarrhea in five of
Table 1. Summary of Diet Effects°
Diet 0.1% Ca
Body weight (g)
3.0% Ca 0.1% Ca _ Li 3.0% Ca -- Li
136 138 108 119
-+ +-
F df p
10.6 56
