| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ORIGINAL RESEARCH COMMUNICATION |
1 From the Institute of Food Research, Norwich Research Park, Norwich, United Kingdom (LJH, JRD, WJH, VJB, JAH, and SJF-T); Rowett Research Institute, Aberdeen, United Kingdom (JHB); Sofia University, Faculty of Chemistry, Sofia, Bulgaria (TIV); and the Fisheries Research Services Marine Laboratory, Aberdeen, United Kingdom (IMD)
2 Supported by the Biotechnological and Biological Sciences Research Council and the European Union (Marie Curie Fellowship, Mass School Programme, contract number HPMT-CT-2000-00140). 3 Address reprint requests to JR Dainty, Institute of Food Research, Colney Lane, Norwich, NR4 7UA, United Kingdom. E-mail: jack.dainty{at}bbsrc.ac.uk.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Objectives: The objectives were to validate a method for estimating endogenous losses of copper, test whether a simple model can predict true absorption from the plasma appearance of labeled copper, and develop a compartmental model for copper metabolism by using stable isotopes.
Design: A stable isotope of copper was intravenously administered to 6 men, and fecal samples were collected for 14 d. Four weeks later the study was repeated, but with an oral dose, and blood samples were collected for 7 d and fecal samples for 14 d.
Results: There was no significant difference (P = 0.48) in the estimated endogenous loss of copper calculated by using either the excreted intravenous dose (
± SD: 32 ± 5%) or the absorbed and excreted oral dose (35 ± 2%). A simple mathematical model fitted to plasma isotope appearance data estimated true absorption to be 8 ± 2% compared with 48–49% measured by fecal monitoring. A more complicated compartmental model predicted that, when newly absorbed copper first enters the blood, 74% is removed by the liver and 99% is bound to ceruloplasmin in the plasma. The exchangeable pool of copper was estimated to be 43 ± 30 mg. Daily endogenous losses were predicted to be 2.4 mg.
Conclusions: The results showed that fecal monitoring is the only method that can reliably measure labeled copper absorption, and it is not necessary to administer an intravenous dose of copper to estimate endogenous losses. The compartmental model provides new insights into human copper metabolism.
Key Words: Copper • stable isotopes • men • compartmental modeling • endogenous losses • true absorption
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
It is generally agreed that the fecal monitoring method is not ideal in terms of volunteer compliance, ease of sample preparation, analysis, or cost. An alternative method used for iron absorption studies (5, 6) involves giving an oral dose and taking serial blood samples over 6 h. The plasma appearance of iron is modeled by a simple technique, from which the quantity of iron absorbed from the oral label can be calculated. The second aim of the present study was to construct a similar model by measuring the appearance of labeled copper in the plasma and to compare the results with fecal monitoring data to see whether the iron technique could be used for copper.
Only one compartmental model of copper metabolism has been developed for humans (7). Although this model incorporates the main physiologic features, it has some experimental drawbacks, such as limited sampling of blood, which means that there is incomplete characterization of the initial metabolism of absorbed copper. Also, much of the earlier work on copper metabolism has focused on animal models, the results of which are assumed to mimic human copper metabolism. Some of the assumptions of the animal models have yet to be validated in humans. Therefore, the third aim of our study was to develop a comprehensive model of human copper metabolism.
|
SUBJECTS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
= 39 y) were recruited to the study. A 10 mL screening blood sample was taken to exclude volunteers whose biochemical and hematologic indexes fell outside the normal range. Other exclusion criteria included use of medications or nutritional supplements and smoking. The aims and procedures of the study were explained to the subjects during a visit to the Human Nutrition Unit at the Institute of Food Research, and written informed consent was obtained. The Norwich District Ethics Committee approved the protocol, and the study was conducted in accordance with the Helsinki Declaration of 1975 as revised in 1983.
Study design
Subjects took part in 2 d of experiments at the Human Nutrition Unit but were otherwise free-living during the course of the study. On the first experimental day, which followed an overnight fast (10 h), the subjects received a 0.5-mg intravenous infusion of a highly enriched copper-65 stable isotope (65Cu = 99.4%). The dose was administered over 60 min via a cannula inserted into a vein in the forearm. The subjects were given a light breakfast 1 h after completion of the infusion. Fecal and urine samples were subsequently collected over the next 14 and 7 d, respectively. No blood samples were taken.
A minimum of 4 wk later, the subjects attended the Human Nutrition Unit for a second time. After an overnight fast, the subjects were given a 3-mg oral dose of highly enriched copper-65 stable isotope (65Cu = 99.7%). Blood samples were collected over 5 h via a cannula inserted into a vein in the forearm. After removal of the cannula, the subjects were given lunch before they returned home. Single blood samples were collected each morning from fasted subjects for the following 4 consecutive days and 7 d postdosing. Each subject made complete fecal and 24-h urine collections for 14 and 7 d, respectively, after dosing. No fecal markers were given but the subjects were made aware of the importance of compliance and the necessity to report any missed samples. The subjects kept a record of all food and beverages consumed for 3 d before and 3 d after the second experimental day. Diary entries were recorded by using household measures and coded with the most appropriate food code selected from UK food-composition tables by using DIET CRUNCHER (Way Down South Software, Dunedin, New Zealand; Internet: www.waydownsouthsoftware.com) nutritional analysis software. Estimated amounts were calculated from values derived from average food portion sizes (8).
Dose preparation and administration
The intravenous doses were prepared by Ipswich Hospital Pharmacy Manufacturing Unit by dissolving 67 mg [65Cu]copper (II) chloride (Trace Sciences International, Richmond Hill, Canada) in 100-mL sterile saline. The doses were portioned into 2-mL sterile glass ampules and stored at 4 °C until used. The doses were tested for sterility by Ipswich Pharmacy Quality Control Department, and copper concentrations were measured by inductively coupled plasma mass spectrometry (ICP-MS). The intravenous doses were administered by a qualified nurse issued with the appropriate approvals by a medical doctor. Before infusion, the 2-mL copper dose (0.25 mg/mL) was mixed with 50 mL sterile normal saline and infused with an Omnifuse pump (Graseby Medical Ltd, Watford, United Kingdom).
Isotopically enriched oral doses of copper chloride were prepared from [65Cu]elemental copper (Chemgas, Boulogne, France) as previously described (6), and the concentration was accurately determined by ICP-MS. The solution was divided into individual doses, which were stored in plastic vials at –20 °C until required. The oral dose was administered in 50-mL water, and the subjects were asked to consume the dose as quickly as possible.
Blood sample collection and analysis
A baseline blood sample (10 mL) was collected on the second experimental day, 15 min before the oral dose was administered; subsequent blood samples (10 mL) were collected at 0, 20, 40, 60, 80, 100, 120, 150, 180, 210, 240, and 300 min. Five additional blood samples were collected 24, 48, 72, 96, and 168 h after dosing. Blood was collected into trace element–free lithium heparin–containing tubes (Sarstedt, Leicester, United Kingdom) mixed gently by inversion, and centrifuged at 15 000 x g for 10 min at room temperature. The supernatant fraction of the plasma was transferred to acid-washed Nalgene cryogenic vials (Nalge Company, Rochester, NY), frozen on dry ice, and stored at –80 °C.
The copper that was not bound to ceruloplasmin, known as directly reacting copper, was extracted from the samples by using a dialysis-chelex method (9) with the following modifications. Human plasma (5 mL) was dialyzed against 100 mL of 175 mmol ammonium phosphate buffer/L (pH 7.0) for 4 h and then 2 x 100 mL of 175 mmol ammonium phosphate buffer/L (pH 7.0) containing 50 mmol histidine/L for 16 h with the use of Spectrapore dialysis tubing (10 mm flat-width cellulose ester membrane, MWCO 5000; Spectrum Europe BV, DG Breda, Netherlands). Copper in the histidine-containing dialysates was extracted by using 2 mL Chelex-100 minicolumns, which were prepared according to the manufacturer’s instructions (Bio-Rad, Hemel Hempstead, United Kingdom). The pH of the dialysates was adjusted to 8.0 with ammonium hydroxide before being applied to the columns. After washing with 1 mol ammonium acetate/L and 18.2 M
water, the divalent metals (including copper) were eluted with the use of 10 mL of 2.5 mol nitric acid/L. Total plasma copper and retentate copper were extracted by adding 0.8 mL water and 0.1 mL concentrated ultrapure nitric acid to a 0.1-mL sample. Precipitates were removed by centrifugation at 10 000 x g for 15 min, and the supernatant fractions were retained. All solutions were then analyzed for copper by atomic absorption spectrometry and for copper isotope ratios by ICP-MS. The methods used to remove the contaminant copper from the reagents used throughout the procedure and to quantify the remaining contaminant copper were described previously (9, 10).
Copper isotope ratios were measured with a Perkin-Elmer (Norwalk, CT) Elan 6100 DRC ICP-MS instrument (10). Instrumental performance was checked daily with the use of multielement standard solutions. The ratio of cerium oxide to cerium signals was maintained between 2.2% and 3.5% to control the formation of oxides, and the formation of double-charged ions was controlled by keeping the signal for Ba2+/Ba below 3%. Isotope ratios were first corrected for fractionation by reference to the published isotope ratio for 63Cu/65Cu (11). The isotope ratio for a prepared standard was obtained before each measurement and used to derive an isotope ratio correction factor.
Fecal sample preparation and analysis
All equipment used during sample processing was acid-washed before use. Fecal samples were autoclaved, freeze-dried, ground to a fine powder with the use of a mortar and pestle, and subsampled. Samples were prepared for ICP-MS analysis by using a combination of 2 methods (2, 12). Briefly, portions of fecal samples were ashed at 450 °C for 48 h in a muffle furnace (Vulcan 3-550; Jencons Scientific Ltd, United Kingdom). Samples were then prepared for the ion-exchange extraction of copper by taking up 0.2 g fecal ash in a 1:1 mixture of water (Milli-Q; Millipore, Billerica, MA) and concentrated ultrapure nitric acid (Merck Ltd, Lutterworth, United Kingdom), drying on a hotplate, and reashing overnight at 500 °C. The resultant ash was taken up in 2 mL of 6 mol hydrochloric acid/L (Aristar grade; Merck Ltd) and left overnight before being centrifuged at 3000 rpm for 10 min. The supernatant fraction was removed and dried down under hot lamps until only 1 mL solution remained.
Copper was subsequently extracted from the supernatant fraction by ion-exchange chromatography using analytic-grade anion-exchange resin (AG1 x 8, 200-mesh chloride; Biorad Ltd, Hemel Hempstead, United Kingdom). The resin was soaked in deionized water for 
24 h before use, and 2 mL presoaked resin was packed into acid-washed glass columns (1-mL pipette tips; Sarstedt, Nümbrecht, Germany). The columns were connected to a peristaltic pump (Watson Marlow, Falmouth, United Kingdom) via polyethylene tubing (1 mm internal diameter, 2 mm outside diameter) with a flow rate of 1 mL/min. The ion-exchange system was purged of any contaminant copper by washing with 2.5 mol HCl/L (Aristar grade; Merck Ltd) for 10 min. Minerals were eluted from the columns for 1 h with a solution of 2 mol HNO3/L. The resin was regenerated into the chloride form by pumping a 6 mol HCl/L solution through the column for 1 h. Samples were subsequently loaded directly onto the resin, and the columns were flushed with a 6 mol HCL/L solution for 15 min. Copper was eluted with a 2.5 mol HCl/ solution into polytetrafluoroethylene vials in 15-mL fractions. Fractions were dried down under a hot lamp and reconstituted for MC-ICP-MS analysis in 2% (by vol) ultrapure HNO3. Copper totals were measured by atomic absorption spectrometry (model 3300; Perkin-Elmer).
Copper stable isotope ratios were determined by using a multicollector ICP-MS (Micromass "Isoprobe" multicollector inductively coupled plasma mass spectrometer; GVi, Manchester, United Kingdom) combined with a Cetac "Aridus" desolvator (Cetac, Omaha, NE). The samples were analyzed (at a concentration of 1 ppm) with bracketing standards and instrument blanks (13). For correction of instrumental mass bias, a simple linear correction between samples and bracketing standards was applied. Standard Reference Material 976 (copper certified for isotopic composition; National Institutes of Standards and Technology, Gaithersburg, MD) was used as a reference. The internal reproducibility of the measurement was not <0.05% for 63Cu/65Cu ratios. The external reproducibility of the calculated copper isotopic enrichment, which was based on the isotopic variation in the baseline fecal samples, was 0.1%.
Urine sample preparation and analysis
Urine samples (20 mL) were weighed into acid-washed crucibles and gently evaporated on a hot plate until the volume was reduced to 1–3 mL; 1 mL of the concentrated urine was transferred to a 5-mL acid-washed plastic tube (Sarstedt, Leicester, United Kingdom) and gently mixed with 1 mL 2% ultrapure nitric acid. Total copper concentrations were measured by atomic absorption spectrometry (model 3300; Perkin-Elmer).
Kinetic data analysis
Estimation of endogenous losses
In a previous publication we described a method to estimate endogenous losses of copper from a labeled oral dose, which involved the simultaneous administration of the rare earth marker holmium (3). That article showed that copper and holmium share identical excretory patterns; therefore, it was deemed unnecessary in the present study to give holimum as a marker to indicate when all the unabsorbed copper had been excreted. This point comes when the mole fraction of labeled copper in the feces falls below 0.02 (2% of the total copper in the feces). After this time, all of the subsequent stool samples contain only labeled copper that has been absorbed and then excreted. By plotting a straight line through these subsequent points in a graph of mole fraction of labeled copper in feces versus time and extrapolating back to the time of label administration (t = 0), the mole fraction of all labeled copper absorbed and then excreted can be estimated. The process for converting this into mass of copper is detailed in the article (3). The procedure for estimating the loss of the intravenous dose into the feces is simple; it is the cumulative appearance of the intravenous label over the period of fecal collection.
Simple model for predicting copper absorption
A previous publication (5) details the mathematical approach taken for this simple model. In summary, it is assumed that a certain fraction of the copper from the oral dose is absorbed at a constant rate over a clearly defined absorptive time and appears in the plasma, where it is eliminated according to a rate constant.
Compartmental model for investigating copper metabolism
Data from plasma and feces were analyzed by using the SAAMII (SAAM Institute Inc, Seattle, WA) program (14) and the compartmental model (Figure 1
).
|
Data fitting.
Two parameters, the apparent volume of distribution of the accessible compartments (V) and the fractional transfer rate from compartments 1 to 7 (k7,1), were held constant for all volunteers. Both of these parameters were difficult to estimate from our data; therefore, V was fixed at 5 L according to information received from a personal communication (N Lowe, 2003) and k7,1 set to 10.0 d–1. The other parameters were given initial estimates consistent with published data on human copper metabolism. During the fitting process, the parameters are allowed to vary until a minimum of the objective function is reached. The software then returns the mean and the SD of the parameters. The 3 data sets (plasma copper concentrations from the oral dose, copper from the oral dose that appeared in the feces, and copper from the intravenous dose that appeared in the feces) were combined within SAAMII. The fecal data were entered as a cumulative total. It is assumed that within the time frame of this experiment, none of the oral or intravenous isotope would have been incorporated into bone and subsequently released and returned to the accessible compartment. The final model structure was arrived at by a process of trial and error but with the guiding principle that it must contain the fewest compartments to adequately describe the data (Principle of Parsimony). The final model parameters are nonuniquely identifiable, which means that they have more than one but a finite number of solutions. Several compartmental structures were attempted based on known physiology and metabolism, but the final structure (Figure 1
) was chosen because it seemed to capture most of the known features of copper metabolism with the minimum SD on the parameters. The mass of copper in each compartment was estimated by using SAAMII in "system" mode with an exogenous input into the gut compartment equivalent to the calculated daily intake from the diet.
Statistics
All data are expressed as means ± SDs. Student’s t test was used to test for differences between parameters of interest. Differences were considered significant if the P value was < 0.05.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The oral and intravenous isotope excretion and oral absorption estimates are summarized in Table 1. The quantity of intravenous copper isotope excreted (32 ± 5%) was not significantly different (P = 0.27) from the estimate of excretion of the absorbed oral copper label (35 ± 2%). The mean apparent absorption from the oral dose was 33 ± 3%. When a correction was made to the apparent absorption for the quantity of oral dose absorbed and then excreted, the mean true absorption was estimated to be 49 ± 4%. This was not significantly different (P = 0.48) from the true absorption (48 ± 5%) when the correction was made by using the estimation of the quantity of the intravenous dose excreted. True absorption as calculated by the simple model was 8 ± 2%, which was significantly different (P < 0.01) from that estimated by fecal monitoring.
|
. The exchangeable pool size was estimated to be 43 ± 30 mg, with most of the copper being located in compartment 6 (84%). True absorption was estimated to be 49 ± 4% [true absorption = (k2,1 + k3,1)/(k2,1 + k3,1 + k7,1)]. Plasma copper was present in 2 compartments, the vast majority (99%) being in compartment 5; the remaining 1% was found in compartment 2. Most of the rate constants (ki,j) that were fitted by the model had a CV (CV = 100 x SD/ki,j) <25%, although k1,3, k1,6, and k5,6 had greater uncertainty (CV: 
50% on average). Rate constant k6,5 for subjects 4 and 6 was fixed because of fitting problems and has no uncertainty associated with it. The same applies to k4,3 for subject 5. The mean model fit to the average of the 6 subjects’ experimental data is shown in Figure 2. The first peak concentration (Cmax = 0.028 ± 0.008 µg/mL) in the labeled plasma occurred 90 min after the dose before steadily decaying and then peaking again (Cmax = 0.021 ± 0.005 µg/mL) 2–3 d later. An example of the distribution of labeled copper between components of human plasma in subject 4 is shown in Figure 3. The directly reacting copper was the first to peak, at 80 min, followed by the ceruloplasmin-bound copper 3 d later.
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The attempt to quantify copper absorption with the use of a simple mathematical model (single compartment) was unsuccessful. True absorption estimated from the simple model was only 8%, whereas it was 48–49% when measured by fecal monitoring. This difference can probably be explained by the large first-pass effect in the liver that newly absorbed copper undergoes on passing through the portal vein. This effect is not incorporated into our simple model.
The more complicated compartmental model (Figure 1) was developed by using the well known Principle of Parsimony, which states that one should not increase, beyond what is necessary, the number of entities required to explain anything. Therefore, this model had the minimum number of parameters to produce the observed behavior of the experimental data. The parameters are nonuniquely identifiable, which means that different starting estimates for the parameters will result in the program optimizer converging to different estimates for the parameters. This requires the starting estimates to be based on sound physiologic principles, otherwise the final estimates will be meaningless even though the optimizer may converge to what appears to be a good fit. Convergence also depends on how much noise or uncertainty is in the data, and, unfortunately, it was necessary to fix rate constant k (6, 5) for subjects 4 and 6 and k (4, 3) for subject 5 to fit the remaining parameters. This is an indication that the model suffers from some a posteriori identifiability issues when the data are noisy. The uncertainty on most of the parameters was acceptable for modeling of this type (15). The experimental data that was simultaneously fitted to the model was the fecal appearance of the oral and intravenous isotopes and the plasma appearance of the oral isotope in the form of total labeled copper concentration. Originally, we planned to use the ceruloplasmin-bound labeled copper plasma concentration and the directly reacting labeled copper plasma concentration as model inputs rather than the total labeled copper plasma concentration. Unfortunately, the labeled ceruloplasmin-bound and directly reacting copper plasma concentrations proved to be very noisy data sets (see Figure 3 for an example), so only the total labeled copper concentration was used.
One of the main findings from our compartmental model (Figure 1) was the necessity of including a pathway (k3,1) from the gut to the liver. This so-called first-pass effect accounts for the surprisingly small quantity of newly absorbed copper that appears in the plasma and explains why the simple model of predicting copper absorption on the basis of plasma appearance data does not work. The ratio of k3,1 to k3,1 + k2,1 would indicate that 74% of absorbed copper is removed by the liver on the first pass before undergoing a delayed entry into the plasma bound to ceruloplasmin (80%) or excreted back into the gastrointestinal tract (20%). The other copper excretory path back into the gut in our model is via compartment 6, which we speculate as containing muscle and other soft tissue. This excretory pathway is consistent with known copper physiology, which suggests that a large flux of copper (2.0–2.8 mg/d) is excreted via salivary, gastric, duodenal, and pancreatic routes (16). The estimated copper flux from our model is 1.4 mg/d via these routes and from biliary excretion (k1,3 x mass of compartment 3) is 1.0 mg/d. which, again, is lower than the value (ie, 2.5 mg/d) reported by Linder et al (16). Our estimates of copper pool sizes are also lower than those previously reported. It is thought that a 70-kg man contains 110 mg Cu (17), but our model has estimated this to be 43 mg. One explanation for this low value may be that the time frame of our blood sampling was a relatively short (7 d) period; therefore, we would not have seen any exchange of the labeled copper with skeletal copper. This seems likely because it is known from calcium kinetic studies that even 14 d of plasma sampling does not allow detection of exchange or equilibrium of labeled calcium with the native element in bone matrix (18). It is thought that almost one-half of the copper stored in the body is contained in the skeleton, so our estimate of total body copper (43 mg) almost certainly excludes any copper in this region.
The sampled compartments (2 and 5) represent the plasma. It was decided to fix the volume of distribution of this space to 5000 mL based on unpublished data (N Lowe, 2003). This value was the mean of 8 subjects who each received an intravenous dose of labeled copper and then had 12 blood samples taken over 90 min, which allowed an estimate to be made of the apparent volume of distribution by a standard pharmacokinetic method (19). The ratio of the volume of distribution to the total copper contained in compartments 2 and 5 represents the predicted plasma concentration of copper. The model predicted that the distribution of copper in the plasma was 99% in compartment 5 (identified as ceruloplasmin-bound copper) and 1% in compartment 2 (identified as directly reacting copper). This compares with 94% ceruloplasmin-bound copper as measured by Cartwright and Wintrobe (20), 97% measured by Buckley et al (21), and 96% measured by Inagaki et al (22).
The half-life (t1/2) of ceruloplasmin copper in this study was estimated to be 27 d, which compares well with the value of 20 d that was found in 10 "normal" control subjects in the study by Lyon et al (23). The t1/2 of the directly reacting copper was estimated to be 60 min, which is higher than the values reported by Buckley et al (21) in 2 subjects after an intravenous infusion (t1/2 = 11 min) and estimated indirectly by Scott and Turnlund (7) in 5 subjects after an oral dose (t1/2 = 26 min) during a metabolic period when the subjects were consuming 1.68 mg Cu/d. Given the small sample size of these studies, there is still some uncertainty as to the true population mean of t1/2 for directly reacting copper. Interestingly, the indirect estimate from the Scott and Turnlund model after the intravenous dose of labeled copper (t1/2 = 41 min for the directly reacting copper) suggests that it depends on the route of administration. We cannot comment on this on the basis of our study because no blood samples were taken after the administration of the intravenous dose.
In conclusion, the previously published method (3) for estimating endogenous losses of copper from an oral dose has been shown to be consistent with that from an intravenous dose. The use of a simple model for estimating the absorption of an oral dose of copper from its appearance in plasma does not warrant further investigation. Fecal monitoring remains the only feasible method for estimating copper absorption. A more complicated compartmental model showed that newly absorbed copper undergoes a substantial first-pass effect in the liver, where 74% of the copper is initially removed from the circulation. The exchangeable pool of copper in the body was estimated to be 43 mg; this value is well below the total body copper concentration of 110 mg reported from autopsy data but consistent with the fact that most copper in the body cannot be quantified by using stable isotopes and compartmental modeling because it is contained in very slowly exchanging bone.
|
ACKNOWLEDGMENTS |
|---|
We thank the Human Nutrition Unit staff for their technical assistance with the human study, Nicky Lowe for allowing the use of unpublished data in the compartmental modeling, and all the volunteers who contributed to the successful outcome of the study.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  L. J Harvey, J. R Dainty, W. J Hollands, V. J Bull, J. A Hoogewerff, R. J Foxall, L. McAnena, J. Strain, and S. J Fairweather-Tait Effect of high-dose iron supplements on fractional zinc absorption and status in pregnant women Am. J. Clinical Nutrition, January 1, 2007; 85(1): 131 - 136. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. S Watson and T. D. Lyon Use of mathematical modeling to study copper metabolism in humans Am. J. Clinical Nutrition, December 1, 2005; 82(6): 1357 - 1357. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
, Total labeled copper; +, labeled copper bound to ceruloplasmin; 
, labeled copper bound to albumin.