| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
DEVELOPMENT AND TISSUE PLASTICITY
Departments of Pediatrics and Medicine (1Rehabilitation and 2Neurology), 3Pathology and Molecular Medicine, McMaster University, Hamilton, Ontario, Canada L8N 3Z5; 4School of Health Sciences, Deakin University, Burwood, Victoria, Australia 3125; and 5Department of Biology/Chemistry, Springfield College, Springfield, Massachusetts 01109
Submitted 15 May 2003 ; accepted in final form 12 June 2003
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
dietary supplements; hepatitis; drug toxicity; side effects
There has been some concern regarding the potential for CrM toxicity based on two anecdotal human case reports (21, 34), one animal study in hypertensive rats (8) and the fact that carcinogens can be formed if creatine and sugars are heated to high temperatures (47, 48). In humans, most of the studies that have examined the potential for toxicity have not found evidence of side effects when consumed at "recommended" doses (22, 29, 33, 36, 37, 44). Several recent reviews have concluded that dietary CrM supplementation in humans appears to be relatively safe in the short term; however, they cautioned that the long-term side effects have not been evaluated systematically (19, 44). Encouragingly, more recent studies in humans have not found evidence for CrM-associated toxicity based on blood analysis and side-effect questionnaires in older adults (3, 5), young athletes (29, 44), and in patients with neurological diseases (17, 38, 41, 46).
To date, there have not been any systematic pathology studies in animals treated with CrM, and most of the studies evaluating the efficacy in animal models of disease did not evaluate the potential for side effects (1, 20, 27, 28). Given the widespread use of CrM supplementation among athletes and nonathletes (18), the potential for its use in a variety of disease states (see above), and because there are several ongoing studies in patients with neurological disease, it is important that studies more carefully evaluate the potential for side effects from longer-term CrM supplementation at the tissue and organ level.
One of the theories regarding the potential for toxicity from CrM supplementation is that creatine can increase oxidative stress and potentially form carcinogenic compounds in vitro (47). In contrast, several studies have found that CrM supplementation decreased markers of oxidative stress in animal models of neurodegenerative disease (20, 28) and is a weak, but abundant, antioxidant (24). Other studies have found CrM to have an anti-tumorigenic effect on solid tumor growth in animals (23, 30). If CrM is acting via an increase in oxidative stress, it would be expected that animals with an inherently high level of oxidative stress would show a greater propensity toward CrM-mediated toxicity. The SOD1 G93A transgenic mouse has a mutation in the Cu/Zn-superoxide dismutase gene that results in an increase in oxidative stress (14). As a consequence of the increased oxidative stress, the G93A mouse is a good candidate model to evaluate the potential for enhanced CrM tissue toxicity in addition to the fact that it is a well-studied model of ALS (1, 14, 20).
Our initial goal was to determine whether there was any evidence of histopathological lesions in 22 different organs/tissues obtained from SOD1 G93A mice after 140 days of CrM supplementation. After finding significant hepatic inflammatory lesions in the CrM-treated mice, we sought to determine whether or not these lesions occurred after a lower CrM dose exposure in a nontransgenic murine strain (CD-1) and in a different species (Sprague-Dawley rat). These results indicate that tissues, and results obtained from tissues, of one species may not be representative of other species.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
Study 1. For the SOD1 G93A studies, breeding pairs of transgenic
mice, BbCg-Tg (SOD1-G93A) 1Gur
[PDB]
dl, were obtained from Jackson
Laboratories (Bar Harbor, ME). Off-spring bearing the human transgene were
identified from blood samples (retro-orbital venous sampling), using PCR
amplification of a fragment of the transgene as outlined by Jackson
Laboratories
(www.JAX.org).
Equal numbers of male and female mice were allocated to each group. The
lower-copy-number SOD1-G93A mice (used in the current study) develop
progressive limb weakness at 
180 days of life that progresses to
paralysis within 3-4 wk because of a progressive loss of anterior motor horn
neurons in part as a result of excessive oxidative stress and apoptosis of the
neurons (39).
Study 2. For the longer-term, lower-dose study, CD-1 female mice were obtained from Charles River Laboratory (Wilmington, MA) at 15 wk of age.
Studies 3 and 4. Young male Sprague-Dawley animals were purchased from Taconic Laboratories (Germantown, NY) at 21 days of age. The rats were Pneumocystis carinii free at the start of the studies and were given a minimum of 1 wk to observe for any illness signs or behavior.
Ethical approval for studies 1, 3, and 4 was obtained from the McMaster University Animal Research Board, and the studies were conducted in accordance with the Canadian Council on Animal Care; the studies were completed at the Central Animal Facility of McMaster University (Hamilton, Canada). Study 2 was completed at Springfield College (Springfield, MA). The procedures followed were approved by the Springfield College Animal Use Review Committee. All studies adhered to the policy outlines in the Guide For the Care and Use of Laboratory Animals published by the National Institutes of Health (NIH Publication No. 80-23, Revised 1996). All efforts were made to minimize the number of animals used and their suffering.
Design
Study 1. The data for the current study were collected during a study comparing the clinical efficacy of CrM compared with riluzole (see below) as therapies for SOD1 G93A mice (see Ref. 39). After being weaned, animals were housed individually with a 12:12-h light-dark cycle. Equal numbers of male and female animals were randomly assigned to one of four groups receiving either standard powdered murine chow (control, n = 16), riluzole (n = 14), CrM (n = 14), or a combination of riluzole and creatine (n = 15). The CrM (Sigma, St. Louis, MO) was mixed with the standard murine chow (Purina) at 2% wt/wt. The riluzole was provided by Rhone Poulenc Rorer, Canada (now Aventis Pharma), and was given in drinking water (100 µg/ml) ad libitum. The combination (riluzole + CrM) included both treatments at the above doses, and treatment for all groups was initiated when the mice were 40 days old for a total of 159 days of treatment. Each of the treatment groups (CrM, riluzole, riluzole + CrM) in this study showed a significant prolongation in the time to onset of clinical symptoms (paralysis) and a lower symptom score at the time of death at 199 days of life compared with the control animals (39). The average lifespan of these high-copy-number animals is between 200 and 230 days with death from paralysis.
All animals were euthanized using an overdose of pentobarbital sodium at 199 days of life (159 days of treatment) and were immediately dissected; tissues were visualized and palpated for evidence of gross pathology. Portions of liver, kidney, adrenal gland, esophagus, duodenum, jejeunum, cecum, colon, urinary bladder, spleen, pancreas, lung, heart, skeletal muscle (red and white gastrocnemius), testes, ovaries, uterus, cervix, thyroid, spinal cord, and brain were immersed in 10% phosphate-buffered formaldehyde (formalin). These samples were dehydrated in increasing concentrations of ethanol and xylene and embedded in paraffin, and 5-µm-thick sections were stained with hematoxylin and eosin and cover slipped.
Study 2. Twenty-four CD-1 mice were randomly separated into three
equal groups of eight animals and housed in these groups. All groups had ad
libitum access to food and water, with the control group receiving no CrM
while the experimental groups received CrM (Pfanstiehl Laboratories, Waukegan,
IL) added to the water supply to achieve a dose of 0.025 g ·
kg-1 · day-1 (very-low-dose group) and 0.05 g
· kg-1 · day-1 (low-dose group). Two
animals from each group were randomly killed at 56 days of age. Water
consumption was monitored weekly, and the doses above represent the mean
intake per animal over the intervention period. The remaining mice were killed
after 300 days of intervention, and the liver, heart, kidneys, and skeletal
muscle (biceps femoris, semitendinosis, semimembranosis, and gracilis) were
excised and processed as outlined above for histological analysis.
Study 3. Equal numbers (n = 20) of male Sprague-Dawley rats were randomly allocated to receive CrM (2% wt/wt; Traco Laboratories, Champaign, IL) or normal rat chow, and each of these groups was further subdivided into groups receiving methylprednisolone (7 mg · kg-1 · wk-1) or a placebo injection starting on day 30 of life. Thus there were four treatment groups [CrM + placebo, CrM + methylprednisolone, normal + placebo, normal + methylprednisolone (all n = 10)], and these treatments were administered over a 50-day period. Because of the risk of immunosupression from the prednisolone, the rats were housed in a barrier room free of rodent infections, and the light-dark cycle was 12:12 h. At day 50, the rats were euthanized with pentobarbital sodium, and the liver was extracted and placed in 10% formalin and processed as above.
Study 4. Equal numbers (n = 6) of male Sprague-Dawley rats were randomly allocated to receive CrM in rat chow (2% wt/wt; Traco Laboratories) or normal rat chow over a 365-day period starting on day 30 of life. The rats were housed in pairs with a light-dark cycle of 12:12 h. At day 365, the rats were killed with pentobarbital sodium and were immediately dissected; tissues were visualized and palpated for evidence of gross pathology. Portions of liver, kidney, adrenal gland, esophagus, duodenum, jejeunum, cecum, colon, urinary bladder, spleen, pancreas, lung, heart, skeletal muscle (red and white gastrocnemius), testes, thyroid, spinal cord, and brain were immersed in 10% formalin and processed as above.
A summary of each of the treatment groups is presented in Table 1.
|
The tissue sections collected at the complete necroscopy in study 1 were reviewed blindly by a veterinary pathologist (Kwiecien) and a human pediatric pathologist (Bourgeois) for histological changes. The hepatic sections from studies 1, 2, and 3 were evaluated blindly by the veterinary pathologist (Kwiecien) and a human pediatric pathologist (Bourgeois) using a transmitted white light microscope (Olympus BX-60) to look for histological changes. The pathology was graded on an ordinal scale with 1 = no significant pathological changes (i.e., no more than one area with <10 inflammatory cells); 2 (grade I pathology) = small aggregates of 10 or greater inflammatory cells measuring between 30 and 50 µm with or without associated hepatocellular injury; 3 (grade II pathology) = multifocal aggregates of mixed inflammatory cells measuring between 51 and 100 µm with associated hepatocellular injury; and 4 (grade III pathology) = multifocal aggregates of mixed inflammatory cells measuring >100 µm with associated hepatocellular injury. The vast majority of inflammatory cells were lymphocytes with some histiocytes seen in association with necrotic hepatocytes primarily in grade II and III pathology. The interrater correlation coefficient between the two pathologists using this scale in 59 blinded slides was 0.96. The liver sections for study 2 were also evaluated by one of the authors (Keys) using a slightly different grading system (although they were subsequently reanalyzed using the common grading system described above), and the cell diameters were also determined for the mice treated for 56 days. The test/retest coefficient of variation for the cell diameters was <3%. Finally, all of the histological slides from studies 2 and 4 were examined blindly by one of the pathologists (Bourgeois). The histology results from study 2 were based on the common grading system described. Examples of a normal liver section and the three grades of pathology are presented in Fig. 1.
|
Statistics
The tissue histology was graded according to the ordinal scale above and was analyzed using the Kruskal-Wallis non-parametric statistical test and Dunn's multiple-comparison post hoc test (version 3.03; Prism, San Diego, CA). The data in Tables 1, 2, 3, 4, 5 reflect the number of animals whose tissue fell within a given pathological category compared with the total number of animals. For the cell diameters, a one-way ANOVA was employed with Tukey's post hoc test (version 3.03; Prism). A P value <0.05 was taken to indicate statistical significance.
|
|
|
|
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
Gross changes were not observed in any of the tissues/organs from the SOD1
G93A mice. There were a few minor histopathological changes that were
considered incidental in some of the tissues and organs (i.e., kidney, 3
samples with small collections of lymphocytic aggregates; lung, 4 samples with
scattered lymphocytic aggregates) from the G93A mice [all seen in <5% of
the sections (excluding muscle and spinal cord; see below)], with no
differences between the treatment groups [P = not significant (NS)]
and no sex differences. There was an 50% reduction in the number of motor
neurons in the lumbar region of the G93A mice at 199 days compared with
age-matched nontransgenic mice of the same background, and the skeletal muscle
showed an abnormal number of internalized myonuclei and scattered atrophy,
with no differences between treatment groups (also see Ref.
39). In contrast, hepatic
inflammatory changes were observed with a greater frequency in the livers of
the G93A mice in the creatine group compared with the placebo and riluzole
groups (P < 0.01). Similar results were found for the grade
III lesions, with more in the creatine group compared with the placebo
and riluzole groups (P < 0.05). There were no grade II or
III pathological changes observed in the livers from either of the
placebo or riluzole groups (Table
2). When the data from the creatine and creatine + riluzole groups
were combined (i.e., all animals on CrM) compared with the riluzole and
placebo groups (i.e., all animals not exposed to CrM), there were more
grade II and III lesions seen for creatine and creatine +
riluzole compared with the combined non-CrM-treated groups (P <
0.05).
Study 2
There were no remarkable changes seen in the kidney, heart, or skeletal muscle of the mice from any of the CD-1 mice groups. There was a significant increase in the number of inflammatory lesions observed in the liver from the 0.05 g · kg-1 · day-1 supplemented mice compared with the nonsupplemented group, and the results were not different when the histological grading was performed at McMaster University (Bourgeois) or at Springfield College (Keys; Table 3). There was also evidence of mild connective tissue infiltration in most of the CrM-treated mice on both dose levels (data not shown). There was no evidence of inflammation or connective tissue proliferation in the liver sections from the 56-day treated animals. The hepatocytes (mean of 15 cells/animal) from control livers (56 days) averaged 15.2 ± 2.6 (SD) µm in diameter. Hepatocytes from mice maintained on 0.025 g · kg-1 · day-1 averaged 18.5 ± 1.4 µm in diameter (+22.5%, P < 0.05 vs. control). Hepatocytes from mice maintained on 0.05 g · kg-1 · day-1 for 56 days averaged 22.0 ± 1.5 µm in diameter (+45%, P < 0.01 vs. control and P < 0.05 for 0.05 vs. 0.025 g · kg-1 · day-1).
Study 3
There were no gross lesions in the livers of the rats. There were a few grade I and II lesions in some of the livers from the rats (no grade III lesions), with no between-group differences (NS, Table 4).
Study 4
There were some grade I and II lesions seen in the livers of the rats, with no difference between the groups (NS, Table 5). There were a few mild foci of inflammatory cells in the lungs of four control and four CrM-treated animals and two control and one CrM-treated animal, with trivial sclerotic changes seen in the kidneys (NS between groups, data not shown). For all of the remaining tissues, there was no evidence of histopathology in more than one animal and nothing more than minor changes. There was no evidence of neoplasia in any tissue/organ.
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
The hepatic lesions seen in the murine model were strongly associated with the administration of CrM, and there was no independent or interactive effect related to riluzole administration. CrM has been evaluated as a therapeutic modality in a number of murine models of neurological disease, including ALS (SOD1 G93A; see Refs. 1, 20, and 39), Huntington's disease (9, 28), and head trauma (40). In these studies, the dose of CrM was similar or identical to that used in the current series of experiments, yet there were no evaluations of potential hepatic pathology from the CrM administration. The current data are of importance to future evaluations of CrM using murine models, for it is possible that the inflammatory changes in the liver could alter the pharmacokinetics of any coadministered drug that is metabolized by the hepatic route. In addition, alterations in hepatic function could affect glucose and/or urea/ammonia metabolism, or even animal viability if the changes were extensive. Clearly, these hepatic changes must be considered in the design and interpretation of future studies using CrM supplementation in murine models of human medical disorders.
We considered several issues in trying to further understand the observed
hepatic inflammatory changes. Initially, it was felt that the changes were
unique to the SOD1 G93A mice, perhaps as an interactive effect with the known
increase in oxidative stress induced by the mutation
(14). It was considered
plausible that the increased production of free radicals may have contributed
to hepatocellular injury that was exacerbated by CrM. However, given that CrM
supplementation reduced indexes of oxidative stress in animal models of
neurological disease (9,
20,
27,
28), and is an antioxidant
(24), we considered this to be
unlikely. Second, the amount of CrM administered to the G93A mice (2% wt/wt;
human equivalent 70-90 g/day accounting for the higher food intake of
mice) was much higher than doses used in human trials with neurological
disease (17), and the
inflammation could be a dose-dependent toxic effect. Finally, we considered
that the hepatic changes could be unique to the murine model. To answer these
questions, we administered CrM at a lower dose to nontransgenic CD-1 mice and
also to a different species of animal, the Sprague-Dawley rat.
CrM was administered to nontransgenic (CD-1) mice at doses that were lower
than those used for the rats, and hepatitis changes were found that were
similar to those seen in the G93A mice. These results suggest that exogenous
CrM results in hepatic inflammatory changes in otherwise healthy mice and at a
dose (relative to body weight) that is similar to that habitually consumed in
food by humans (1-1.5 g/day). It is likely that the CrM per se and not an
impurity accounted for the observations, for the administered CrM was obtained
from three separate chemical supply sources (Sigma, Pfanstiehl, and Traco)
with stringent quality control at the source, including HPLC analysis for
impurities. In addition, it would be very improbable that an identical
impurity could exist from two separate companies that were used in the mice
studies where inflammation was seen from two separate university vivariums.
Given that the animals from all groups were caged in the same room, the
probability that murine hepatitis virus (MHV) or some other infectious agent
selectively affected only those animals on CrM is statistically improbable. A
few of the SOD1 G93A mice tested positive for antibodies against the MHV using
an ELISA test; however, the numbers of MHV-positive mice were identical
between the groups (<2/group). In addition, the detection of anti-MHV
antibodies is rarely coincident with liver pathology in wild-type mice
(Kwiecien, unpublished observations, 1997-2003). Although MHV was not tested
in the CD-1 mice, the statistical probability that identical findings were
observed merely by random chance in a different mouse species with the
experiments being conducted in a separate vivarium was extremely low
(P < 0.01). Many of the CD-1 mice tested positive for
Helicobacter bili in the intestines (PCR-based assay), and a few were
positive in livers; however, this was not differentially found in those on CrM
treatment. Although CrM treatment could have contributed to formation of
inflammatory foci, presumably because of periodical bacterial showering via
the portal venous system, this potential effect, however, was restricted to
mice, both the transgenic and wild type, and was not observed in rats. Without
an extensive microbiological evaluation, we cannot exclude the possibility
that CrM may render the murine liver more sensitive to microrganisms that
would otherwise be nonpathogenic.
It is important to reference the doses of CrM used in the current series of
experiments in animals to a human perspective. For longer-term consumption,
several studies have employed doses of between 3 and 10 g CrM/day in humans
(15,
17,
38,
41). For a 70-kg human, this
amounts to between 0.043 and 0.142 g · kg-1 ·
day-1, which is in the range of the dose used in the CD-1 mice
(0.05 g · kg-1 · day-1), and is less than
the estimated intake for the SOD1 G93A studies. Assuming that a 70-kg
sedentary human consumed a 2,000-kcal mixed diet of carbohydrate (55%), fat
(30%), and protein (15%), and CrM were given at 2% wt/wt, the CrM intake would
amount to 8 g/day or 0.114 g · kg-1 ·
day-1. These results clearly indicate that the amount of CrM
provided to the CD1 mice in the current study was not supraphysiological and
was well within the dose range that is being used in longer-term human
clinical trials with CrM (15,
17,
38,
41); the dose in the rats and
G93A mice was also within this range using a human 2% wt/wt dose. If one
corrects for the higher food intake in mice and rats relative to their total
body mass, a 2% wt/wt diet is 8-12 times higher than the dose used in
longer-term human clinical trials with CrM
(15,
17,
38,
41). Although we did not
measure creatine content in the livers in the current study, a recent study
has found that total creatine content in livers of rats and mice ranged from 5
to 8 µmol/g wet wt and showed a large increase (260-500%) in response to
oral CrM supplementation (1.4-2.8% wt/wt diet) within 4-8 wk
(16).
Given that the CrM supplementation was clearly associated with inflammatory
changes in the liver, it was important to determine whether a similar process
occurred in a different species of animals. In contrast to the inflammatory
changes seen in both strains of mice, there was no evidence of hepatic
inflammation in the Sprague-Dawley rat exposed to an identical dose (2% wt/wt)
but shorter duration (50 vs. 159 days) of CrM administration. To determine
whether duration was a factor, we exposed Sprague-Dawley rats to the same dose
(2% wt/wt) for 1 yr (just under 1/2 of the average lifespan) and did not find
any evidence of differential hepatitis in the rats. Given the higher relative
amount of food eaten by a rat compared with a human, the 2% wt/wt CrM dose for
the rats represented 1.3 g · kg-1 ·
day-1, which was far in excess of that used in any human study;
still, there was no evidence of CrM inducing hepatic inflammation. It is
difficult to extrapolate the current findings to humans; however, there have
been no substantiated reports of alterations in liver function tests or
transaminases in intermediate and longer-term human trials with CrM
supplementation (5,
17,
22,
36,
37). The lack of any
CrM-induced nephritis (renal inflammation) in any of the species studied
herein also supports the fact that most human studies have not found evidence
for renal dysfunction with CrM
(5,
17,
22,
29,
33,
36,
37), and only two
unsubstantiated case reports exist
(21,
34). Finally, CrM
supplementation appears to increase renal disease progression in hypertensive
rats (Han:SPRD; see Ref. 8),
again showing a species and strain difference.
The above results indicate a species difference in the susceptibility to the hepatotoxic effects from CrM administration. One factor that could explain some of this difference, although by an unknown mechanism, is the fact that mice are predominantly herbivores and rats are more omnivorous (in the wild). As a result, mice may not be accustomed to the delivery of high exogenous creatine concentrations via the portal vein. There does not appear to be any difference in the basal hepatic total or phosphocreatine content between rats and mice, and the magnitude of the increase in response to CrM supplementation is similar for these species (16). As mentioned above, a microorganism that is unique to mice may also be a coinitiator or copromoter of the hepatic inflammatory changes; however, we cannot support or refute this hypothesis at this time. Similarly, the lack of any renal pathology in the rat or two mice strains, in contrast with the previous report of renal pathology in the Han:SPRD rats (8), suggests that an unknown factor between and even within species can influence the toxicity of a drug. Irrespective of the mechanism, the mice appeared to have developed hepatitis while on CrM.
The histological findings in the murine CrM-supplemented livers were identical to those of chronic hepatitis caused by persistent viral or bacterial infection, certain hepatotoxins, or autoimmune hepatitis. It is possible that there is some other infectious agent indigenous to mice and not rats that somehow conferred a differential sensitivity to CrM supplementation; however, common agents such as MHV and H. bili did not provide the answer. An autoimmune process was considered unlikely, for there were scattered histopathological findings in the rats that were not suppressed by concomitant methylprednisolone treatment. To rule this out in the mice, further studies with immunosuppressive agents and CrM given separately and in combination would help. A plausible explanation for the inflammatory changes could be the significant swelling that was seen at 56 days of age in the hepatocytes of the CrM-treated CD-1 mice. This swelling may have either represented the osmotic effects of the large increase in total creatine that occurs with CrM supplementation in mice (16), or it was a histopathological feature of a toxic stress. When hepatocytes are first challenged by pathological conditions, they typically exhibit a pale-staining or vacuolated appearance caused by fluid accumulation and swelling of membrane-bound organelles within the cells (6). Similar signs of chronic hepatitis and cirrhosis can be caused by long-term exposure to toxic levels of ethanol; however, the chronic hepatitis observed in the CrM-supplemented mice differed from ethanol-induced hepatitis in several ways. First, the inflammatory infiltrate consists mostly of lymphocytes instead of neutrophils as in alcoholic hepatitis (2). Second, there was no evidence of steatosis in the CrM-treated mice in contrast to that seen with ethanol hepatitis (10). Finally, alcoholic hepatitis and cell swelling are associated with ATP depletion and oxidative stress (6, 10), both of which would be unlikely with CrM supplementation (1, 4, 20).
In summary, CrM supplementation for one-third to one-half of the lifespan of CD-1 mice at doses that are just over those habitually consumed by humans induced hepatitis in most animals. Furthermore, a higher CrM dose given to transgenic mice that exhibit high levels of oxidative stress (SOD1 G93A) also caused hepatitis, yet did not induce any inflammation or other histological abnormality examined in any of the other body tissues and organs. CrM supplementation to rats, in supraphysiological doses for almost one-half their average lifespan, did not differentially induce hepatic inflammatory changes in the many tissues and organs that were sampled, including the liver.
|
DISCLOSURES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  R. B. Kreider Species-specific responses to creatine supplementation Am J Physiol Regulatory Integrative Comp Physiol, October 1, 2003; 285(4): R725 - R726. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
