Nutrient Interactions

Manganese, Iron and Lipid Interactions in Rats1 CINDY D. DAVIS, DEFUSE M. NEYANDJ.

L GREGER

Department of Nutritional Sciences, University of Wisconsin, Madison, Wl 53706 (9, 12) and interferes with hemoglobin regeneration in anemic animals (12, 13). Supplemental iron restored hematocrits in pigs fed high levels of manganese (13). Excess manganese may also affect iron metabolism by inhibiting 5-aminolevulinate synthase, and hence heme synthesis (14). Several investigators have reported that manganesedependent Superoxide dismutase (MnSOD) activity (EC 1.15.1.1) in several tissues reflects dietary intake of manganese (15-18). Although MnSOD appears to be a promising indicator of nutritional status in regard to manganese, ingestion of ethanol increases MnSOD ac tivity (19).This suggests that further work is needed to evaluate the effect of oxidative stress, as might be in duced by lipid peroxidation, on MnSOD activity. Also, because iron has been implicated as a catalyst for lipid peroxidation (20), increased concentrations of iron in the tissues of manganese-deficient animals coupled with lower levels of MnSOD activity may contribute to increased lipid peroxidation. Work in the 1950s and 1960s suggested that alter ations in manganese concentrations in in vitro systems affected cholesterol synthesis (21-25). Doisy (26) even noted that manganese deficiency lowered blood choles terol levels in one subject. More recently, Roby et al. (27) reported lower plasma and liver cholesterol levels in manganese-deficient animals. Although KlimisTavantzis et al. (28) and Kawano et al. (29) did not observe depressed plasma cholesterol levels during manganese deficiency, Kawano et al. (29) observed re duced high density lipoprotein (HDL)cholesterol, HDL protein and HDL apo E levels with manganese defi ciency. The purposes of this study were to examine the interactive effects of manganese, iron and lipid on tissue retention of manganese and iron and to determine the functional consequences of the interactions in terms of

ABSTRACT The interactive effects of manganese, iron and lipid on mineral status, manganese-dependent Superoxide dismutase (MnSOD) activity and lipoprotein composition were investigated by feeding weanling rats two levels of manganese (0.4 or 56 Mg Mn/g diet), two levels of iron (29 or 109 ng Fe/g diet) and either 12% high-linoleic acid safflower oil or 12% high-oleic acid safflower oil for 8 wk. Rats fed the manganesedeficient diets had decreased heart MnSOD activity; depressed tibia and kidney manganese concentra tions; lowered plasma and high density lipoprotein (HDL) cholesterol, HDL protein and HDL apo E con centrations; and elevated HDL protein/cholesterol ra tios. Ingestion of supplemental iron slightly decreased heart MnSOD activity and tibia and kidney manganese concentrations but had no effect on hematocrits or on plasma and HDL cholesterol levels. Rats fed the linoleic acid-rich rather than the oleic acid-rich oil had increased heart MnSOD activity but had unchanged plasma and HDL cholesterol levels. The decrease in plasma and HDL cholesterol levels with manganese deficiency appeared not to be a result of increased lipid peroxidation but may have resulted from decreased cholesterol synthesis and/or secretion. J. Nutr. 120: 507-513, 1990. INDEXING KEY WORDS:

•manganese •Superoxide dismutase •high density lipoproteins •rats

•iron

Iron deficiency, especially that induced by an iron-de ficient diet rather than by bleeding, increases manga nese absorption and retention (1-5). This interaction may reflect reduced competition for carriers in the proximal intestine (1) and/or the stimulating effect of erythropoiesis (6). However, the effect of iron supple mentation above requirement levels is less clear. Iron supplementation did not affect tissue manganese con centrations in calves (7), but it decreased liver manga nese concentrations in mice (8),sheep (9)and chickens (10).Rats fed very high levels of iron (>410 ug Fe/g) (11) or injected with iron (4) retained less of oral doses of 54Mnthan did those fed adequate levels of iron. Manganese is also antagonistic to iron. Manganese supplementation depresses tissue iron concentrations

'This work was supported by the College of Agricultural and Life Sciences, University of Wisconsin-Madison, project 2623, and by National Institutes of Health training grant 5T32CA09451.

0022-3166/90 $3.00 ©1990 American Institute of Nutrition. Received 10 April 1989. Accepted 11 December 1989.

507 Downloaded from https://academic.oup.com/jn/article-abstract/120/5/507/4738598 by Washington University, Law School Library user on 21 May 2018

508

DAVIS ET AL.

the activity of MnSOD, the hematocrits and the lipoprotein composition.

METHODS Experimental design. The basic design of this study was a 2 x 2 x 2 factorial. The diets contained low (LMn) or adequate (AMn) levels of manganese (by analysis, 0.4 or 56 ug Mn/g diet), adequate (A Fe)or high (H Fe) levels of iron (by analysis, 29 or 109 ug Fe/g diet) and 12% high-linoleic acid safflower oil [byanalysis, 73% 18:2(n6)] (L)or 12% high-oleic acid safflower oil [by analysis, 70% 18:l(n-9)] (O). Hence, the dietary treatments were as follows: LMnAFeL, LMnHFeL, AMnAFeL, AMnHFeL, LMnAFeO, LMnHFeO, AMnAFeO and AMnHFeO. Six rats were fed each of these eight diets ad libitum for 8 wk. In addition, six rats were pair-fed (treatment PFL) the AMnHFeL diet to the level consumed by the rats fed the LMnHFeL diet, and six rats were pair-fed (treatment PFO) the AMnHFeO diet to the level consumed by the rats fed the LMnHFeO diet. Animals and diets. Weanling male Sprague-Dawley rats (HarÃ-anSprague-Dawley, Madison, WI)were housed individually in stainless steel, wire-bottomed cages. The rats' initial average weight was 67 ±l g (mean ± SEM). The formulation of the purified diets was consistent with the AIN-76 diet (30). The diet contained 20% casein (Teklad Test Diets, Madison, WI), 12% safflower oil [either high in linoleic acid (U.S.Biochemical, Cleve land, OH) or high in oleic acid (California Fats and Oils, Richmond, CA)],5% cellulose (Teklad Test Diets), 3.5% AIN-76 mineral mixture without manganese or iron, 1% AIN-76 vitamin mix (Teklad Test Diets), 0.3% DLmethionine (Teklad Test Diets), 0.2% choline bitartrate (Teklad Test Diets), 0.02% BHT (Sigma Chemical, St. Louis, MO), 50% sucrose (Kohl's, Madison, WI) and 7.98% cornstarch (Pocahontas Foods, Richmond, VA). Manganese carbonate (J. T. Baker Chemical, Phillipsburg, NJ) and ferric citrate (J. T. Baker Chemical) were added to the levels described above. Sample collection. Rats were fasted overnight before being anesthetized with CO2 and killed by exsanguination. Blood was collected by cardiac puncture into sy ringes containing 1 mg EDTA and 0.1 mg gentamicin sulfate/mL blood. Hematocrits were determined. Kid neys, tibiae, spleens and 1-g samples of liver were cleansed of adhering material, weighed and frozen in acid-soaked plastic containers. Hearts and 1-g samples of liver were washed with cold 0.9% NaCl, weighed and placed in liquid nitrogen. Mineral analyses. Samples of diet, liver, spleen and reference materials were dried, heated at 450°Cin a muffle furnace and analyzed for iron and copper by atomic absorption spectroscopy (Model 372, PerkinElmer, Norwalk, CT). Downloaded from https://academic.oup.com/jn/article-abstract/120/5/507/4738598 by Washington University, Law School Library user on 21 May 2018

Kidneys, tibiae, diet and reference materials were analyzed for manganese content using an atomic absorp tion spectrophotometer with a graphite furnace atom izer (Hitachi Model 170-70 polarized Zeeman, Tokyo, Japan) by a standard additions technique (31).Each sam ple was injected three times with standards containing 0, 3 and 5 ug Mn/L. Samples of bovine liver (#1557a) or spinach (#1570) standard reference material obtained from the Office of Standard Reference Materials (Gaithersburg, MD) were analyzed with each set of samples. Liver (n = 6) and spinach (n = 6) samples were determined to contain 103% and 106%, respectively, of their certified values for manganese. Enzyme analyses. Hearts were homogenized with a Polytron in a 0.05 mol/L phosphate buffer, pH 7.2, with 0.25 mol/L sucrose for 2 min. The homogenates were centrifuged at 1000 x g for 30 min at 4°C. The pellet was discarded. Manganese-dependent Superoxide dismutase (MnSOD) activity of the supernatant was determined in 50 mmol/L TAPS (A/-tris[hydroxymethyl]methyl-3aminopropanesulfonic acid), 0.1 mmol/L DTPA (diethylenetriaminepentaacetic acid) and 1 mmol/L KCN, at pH 8.2. Pyrogallol in 0.1 mol/L HCl was added to the reaction mixture (so that the final concentration of pyrogallol was 0.2 mmol/L), and the change in absorbance at 420 nm was recorded. The addition of KCN to the reaction mixture slowed the reaction as noted by Marklund and Marklund (32),but the response to vari ous levels of enzyme remained linear. One unit of MnSOD activity was defined as the amount of enzyme needed to obtain 50% inhibition of pyrogallol auto-oxi dation (32). Lipoprotein isolation and lipid analyses. Plasma was obtained from blood by centrifugation at 4°C for 20 min at 1200 x g. Lipoproteins were fractionated from 2-mL plasma samples containing 0.45 mg DTNB [5,5'dithiobis(2-nitrobenzoic acid)] and 0.3 mg PMSF (phenylmethylsulfonylfluoride) by sequential ultracentrifugation, according to the following densities: very low density lipoprotein (VLDL)(d < 1.006 g/mL), low density lipoproteins (LDL) (d = 1.006-1.050 g/mL) and HDL (d = 1.050-1.196 g/mL) (33, 34). Gradient gel electrophoresis was used to determine the proportion of HDLi (diameter 12.2-17 nm) and HDL2(diameter 8-12.1 nm) subpopulations (33, 34). Fatty acid composition of the dietary oils was deter mined by gas chromatography (35, 36). Plasma choles terol concentrations were determined enzymatically (37).Concentrations of protein (38)and cholesterol (37) were determined in VLDL, LDL and HDL fractions. Phospholipid (39) and triglycérides(Sigma Triglycéride Kit No. 336, Sigma Chemical) were determined in HDL. Lipids were extracted from 1-g hepatic samples with chloroform-methanol (35) and assayed for cholesterol (40) and triglycéride (41)content. Apolipoprotein concentrations in plasma and HDL were determined by the Laurell electroimmunoassay

509

MM, FEAND LIPID INTERACTIONS

technique (42, 34). Antisera and standardized plasma were provided by Paul Roheim (Program Project Grant HL 25596, National Heart, Lung and Blood Institute). Statìstica!analyses. Standard errors of the means (SEM)were calculated as the square root of the quantity [the error mean square divided by 6 (the n per treat ment)]. The data were analyzed by a three-way ANOVA (diet manganese, diet iron and type of fat) using an SAS general linear model program (43).To include all treat ment groups, including the pair-fed rats (treatments PFL and PFO), in the three-way ANOVA, data from rats fed similar diets (i.e., treatments PFL and AMnHFeL and treatments PFO and AMnHFeO) were combined. When the average response to high iron or high manganese is reported in the text, the means were also calculated on that basis—that is, treatments PFL and AMnHFeL and treatments PFO and AMnHFeO are combined. WallerDuncan K ratio i tests were also applied to the data. For this test the data for pair-fed rats were not combined with those for rats fed similar diets. RESULTS The dietary treatments did not affect the weight gain of rats,- the average weight of the rats at the end of the study was 349 ±9 g (mean ±SEM). Tissue mineral levels. Average concentrations of manganese in tibiae and kidneys of manganese-deficient rats were reduced to 38% (286 vs. 750 ug/g) and 37% (404 vs. 1086 ug/g), respectively, of those in rats fed adequate manganese,- the effect was significant (Table 1). Ingestion of high amounts of dietary iron was also found by three-way ANOVA to significantly lower tis sue manganese concentrations. This effect of iron was not apparent when Waller-Duncan K ratio î tests were used to compare means. The type of oil had no effect. Differences in manganese intake did not affect the concentrations of iron in spleens and livers (Table 2). Rats fed the manganese-deficient diets had significantly depressed hematocrits by analysis with the three-way ANOVA, but the practical significance of such small changes in hematocrits is questionable. Spleen iron concentrations reflected iron intakes, but liver iron concentrations and hematocrits did not. On average, liver iron concentrations were 13% greater in rats fed the linoleic acid-rich oil rather than the oleic acid-rich oil; this difference was statistically signifi cant. However, there was an inconsistency: Rats fed AMnHFeL had lower liver iron levels than those fed AMnHFeO. Spleen and liver copper concentrations were unaffected by changes in dietary manganese or iron (data not shown). Manganese-dependent Superoxide dismutase (MnSOD). As determined by the three-way ANOVA, all three dietary factors (manganese, iron and the type of oil) significantly affected heart MnSOD activity (Table 1).Heart MnSOD activity was 40% lower among Downloaded from https://academic.oup.com/jn/article-abstract/120/5/507/4738598 by Washington University, Law School Library user on 21 May 2018

TABLE 1

Indices of manganese status in rats fed various levels of manganese and iron with high-Iinoleic acid or high-oleic acid safflower oil1

Diet2

Tibia Mn

Kidney Mn

Heart Mn Superoxide dismutase u/g wet wt

ng/g wet wt

LMnAFeLLMnHFeLAMnAFeLAMnHFeLPFLLMnAFeOLMnHFeOAMnAFeOAMnHF

SEM

34

47

Statistical effectdetermined byanalysis ofvariance:MnFeOil0.00010.01NS30.00010.05NS0.00010.0050.0001

'Values are means (n - 6). Means in columns without common superscript letters are significantly different as determined by WallerDuncan K ratio t tests. 2Diet notation is as follows: LMn - low level of manganese, AMn - adequate manganese, AFe = adequate iron and HFe - high level of iron. Final L or O refers to safflower oil rich in either linoleic or oleic acid, respectively. For the PFL diet, rats were pair-fed the AMnHFeL diet to the level consumed by the rats fed the LMnHFeL diet; for the PFO diet, rats were pair-fed the AMnHFeO diet to the level consumed by the rats fed the LMnHFeO diet. 3NS - not significant. No interactions were significant.

rats fed the diets with low rather than adequate manga nese (93 vs. 153 u/g wet wt, respectively) and 20% higher in rats fed diets with linoleic acid-rich rather than oleic acid-rich oil (137 vs. 109 u/g wet wt, respec tively). The rats fed high rather than adequate levels of iron were found by three-way ANOVA to have reduced MnSOD activity (115 vs. 130 u/g, respectively) overall. The effect was apparent only among rats that were also fed low levels of manganese when the Waller-Duncan K ratio i tests were applied. Lipid and lipoprotein content. Plasma cholesterol levels were significantly lower in animals fed low rather than adequate manganese (Table 3). Plasma and liver cholesterol concentrations were not affected by the level of dietary iron or type of dietary oil. Plasma triglyc éride, liver triglycéride and liver phospholipid concen trations were unaffected by any dietary alterations (data not shown).

510

DAVIS ET AL.

TABLE 2

TABLES

Indices of iron status in rats fed various levels of manganese and iron with high-linoleic acid or high-oleic acid safflower oil1

Plasma and liver cholesterol and apolipoprotein concentrations in rats fed various levels of manganese and iron with high-linoleic acid or higb-oleic acid safflower oil

Diet2HematocritSpleen iron%

ironLiver

Diet2LMnAFeLLMnHFeLAMnAFeLAMnHFeLPFLLMnAFeOLMnHFeOAMn Bmg/100 wtLMnAFeLLMnHFeLAMnAFeLAMnHFeLPFLLMnAFeOLMnHFeOAMnAFeOAMnHFeOPFOS Hg/gwet wt3.183.122.922.583.493.313.522.733.284.100.43NS3NSNSNS wet ml151114151317181916192NSNS

EMStatistical

effectas determinedby analysisof asdeterminedby variance:MnFeOil49ab48b51'50-"50*49*50*51*51*51'0.70.005NSNS323C377**359*489'419*°346C388^322C455*424*°33NS30.001NS65.6*69.2*72 analysisof variance:MnFeOUMn 'Values are means |n - 6). Means in columns without common superscript letters are significantly different as determined by WallerDuncan K ratio i tests. 2SeeTable 1 footnote or text section on experimental design for explanation of diet notation. 3NS - not significant. No interactions were significant.

Plasma apo B and apo A-JTVlevels were fairly insensi tive to the dietary treatments. Dietary manganese and the type of oil had interactive effects on plasma apo A-IV concentrations. Rats fed the deficient rather than the adequate levels of manganese had depressed levels of HDL protein and HDL cholesterol (Table 4). This suggested a general reduction in HDL concentrations in plasma. But the larger HDL protein/cholesterol ratios noted among manganese-deficient rats fed the oleic acid-rich oil in dicated a reduction in the proportion of HDL choles terol. The manganese-deficient rats, especially those fed the oleic rich-rich oil, tended to have lower HDL apo E but constant HDL apo AI concentrations. These obser vations are consistent because apo E-enriched HDL particles are larger and contain greater proportions of cholesterol than do apo AI-enriched HDL particles (36). HDL protein, cholesterol and the ratio of protein/cho lesterol were not affected by the level of dietary iron or the type of dietary oil. Plasma HDL triglycérides,phospholipids LDL protein and cholesterol, and VLDL pro tein levels were not affected by any dietary alterations (data not shown). Downloaded from https://academic.oup.com/jn/article-abstract/120/5/507/4738598 by Washington University, Law School Library user on 21 May 2018

xoilLivercholesterolmg/g 'Values are means (n - 6). Means in columns without common superscript letters are significantly different as determined by WallerDuncan K ratio t test. 2SeeTable 1 footnote or text section on experimental design for explanation of diet notation. 3NS = not significant. No other interactions were significant.

According to three-way ANO VA, consumption of the different oils significantly affected HDL apo E and AI levels. Comparison of means by the Waller-Duncan K ratio i test suggested that consumption of linoleic acidrich rather than oleic acid-rich oil tended to result in high levels of HDL apo E and lower levels of HDL apo AI. The rats fed the linoleic acid-rich oil also utilized feed significantly better (34.0 vs. 32.5%, p < 0.005), had significantly smaller livers (2.77 vs. 2.86 g/100 g body wt, p < 0.05) and had significantly lower plasma VLDL cholesterol levels (3.85 vs. 5.73 mg/100 mL, p < 0.05) than rats fed oleic acid-rich oil (data not shown in tables). Together, these data suggest that the saturation level of the oil fed had subtle effects on the liver.

DISCUSSION MnSOD as an indicator of manganese status. Our observation that rats fed deficient rather than adequate levels of manganese had significantly less MnSOD ac-

511

MN, FE AND LIPID INTERACTIONS TABLE4 HDL composition of rats fed various levels of manganese and iron with high-linoleic acid or high-oleic acid safflower oil1 HDL protein

Diet2

HDL cholesterol

Protein/ cholesterol

mg/100 ml plasma

mg/mg

HDL ApoE

HDL ApoAI

mg/100 ml HDL

HDL23

HDLi3 % staining

intensity

jgabcd2.46abc2.13ed218abcd1.88d2.58"*2.60a2.14"1.99d2.13a128.4d58.4ab58.5'"54.6abc61.6a39. LMAFeLLMHFeLAMAFeLAMHFeLPFLLMAFeOLMHFeOAMAFeOAMHFeOPFO86.6e92.6*107.

6**"120.4*99.

7b60.1*62.7a58.6*51. 4^96.3°*96.4ede131.3a113.2**102.0bcde40.4ede38.2e51.2b53.7*50.2*37.6e38.9*62.0a58.7*489bcd2 lcd32.9a51.4abc429bcd26. S"1*63.9*73.2a68.7a66.2a"119.9e48.1a23.0e21.6e44.6ab22.0e25.

ld60.9*48.3be54.9*°40.9e59.0abc53.

SEM

7.9

3.8

5.9

0.12

4.7

5.3

Statistical effect as determined by analysis of variance: MnFeOilMnxFeFex

oil0.0001NSNSNSNS0.0001NSNSNSNS0.0001NSNSNSNS0.05NS0.050.050.005NS4NS0.05NSNSNS0.05NSNS0.05NS0.05NSNS0.05 'Values are means (n - 6). Means in columns without common superscript letters are significantly different as determined by Waller-Duncan K ratio t tests. 2See Table 1 footnote or text section on experimental design for explanation of diet notation. 3HDLi = diameter 12.2-17.0 nm, HDL2 = diameter 8.0-12.1 nm. "NS = not significant. No other interactions were significant.

tivity in their hearts confirms the observations of others (15, 18, 44). These data suggest that MnSOD activity might be used as an index of nutritional status in regard to manganese. However, other factors also influence the activity of this enzyme. Investigators have observed that conditions that increase production of Superoxide radi cals, such as exposure to hyperbaric oxygen (45), ozone (46), or ethanol (19), increase levels of MnSOD activity. Moreover, MnSOD activity in various tissue does not respond uniformly to environmental factors (44). In this study, rats fed the polyunsaturated rather than the monounsaturated oil were exposed to increased peroxidative stress and had significantly elevated MnSOD activity. Fat accounted for about 26% of energy in these diets. Americans participating in the U.S. Department of Agriculture Nationwide Food Consumption Survey consumed 41% of their energy from fat (47). Thus, the peroxidative stress induced by dietary fat was not ex treme. In this study, manganese-deficient rats fed supple mental iron had decreased heart MnSOD activity, ac cording to statistical analysis by three-way ANOVA. There are several possible mechanisms for this iron-re lated decrease in MnSOD activity. One is increased peroxidative stress. Zidenberg-Cherr et al. (48) impli cated increased tissue iron as the mechanism for Downloaded from https://academic.oup.com/jn/article-abstract/120/5/507/4738598 by Washington University, Law School Library user on 21 May 2018

increased lipid peroxidation in ethanol-fed, manganesedeficient rats. Nevertheless, increased peroxidative damage due to iron accumulation in tissues does not appear likely in this study because MnSOD activity was decreased in animals fed high levels of dietary iron. Furthermore, iron concentrations in livers and spleens were not elevated in manganese-deficient rats. A second, more likely, explanation is that ingestion of supplemental iron depressed manganese absorption and, accordingly, tissue manganese levels. Thus, less manganese was available for incorporation into MnSOD. More work is needed to evaluate the mecha nisms underlying the effect of iron on MnSOD activity. Our observation has practical implications. The two levels of iron that we fed were adequate (29 [ig/g diet) and high (109 ug/g diet) but not excessive. The average woman (23-34 years old) in the United States consumes 11.0 mg of iron daily (61% of RDA for iron) (47). Many (25.7%) American women consume supplements regu larly, and 90% of the supplements contain iron (49). Although many supplements contain 18 mg iron daily, some, particularly those for pregnant women, contain > 60 mg iron. Thus, many women increase their iron intake more than fourfold with supplements, and there fore the dietary levels of iron fed in this study are not extreme.

512

DAVIS ET AL.

Effect of manganese status on lipid metabolism. Lower plasma cholesterol levels in manganese-deficient animals are associated with decreased cholesterol car ried in the HDL fraction. Our results are consistent with those of Kawano et al. (29) in that manganese-deficient rats had lower HDL cholesterol and HDL apo E levels. However, our results differed from those of Kawano et al. in that our manganese-deficient rats had lower plasma cholesterol concentrations. This is probably be cause our rats had a more severe manganese deficiency, as noted by changes in kidney manganese concentra tions. One possible mechanism for the lower plasma cho lesterol levels in manganese-deficient animals is that manganese deficiency resulted in increased lipid peroxidation of the endoplasmic reticulum, which is the site of lipoprotein synthesis. Bell and Hurley (50) ob served swollen and irregular endoplasmic reticulum in tissues of manganese-deficient mice. Consumption of diets high in iron or polyunsaturated fat might be ex pected to increase the amount of lipid peroxidation and potential damage to the endoplasmic reticulum (51). However, in our study, high intakes of iron and polyun saturated fat did not further depress plasma or HDL cholesterol levels in manganese-deficient animals. Thus, this hypothesis seems unlikely. Furthermore, Paynter (44) noted that mitochondrial peroxidation did not occur in livers of manganese-deficient rats. A second possible mechanism for the lowered plasma cholesterol concentrations in manganese-deficient animals is that manganese was needed for cholesterol synthesis. We measured the activity of ß-hydroxy-ßmethylglutaryl (HMG) CoA synthase, an enzyme in volved in cholesterol biosynthesis, which is expected to parallel the activity of HMG CoA reducÃ-ase,the ratelimiting enzyme in cholesterol biosynthesis. We found that the dietary treatments did not affect the activity of this enzyme. However, manganese is believed to be a cofactor for mevalonate kinase (23) and farnesyl pyrophosphate synthase (25),enzymes involved in later steps of cholesterol synthesis. Further work with modern techniques is needed to evaluate whether manganese is indeed a cofactor of these enzymes. A final possible mechanism for the lowered plasma cholesterol concentrations in manganese-deficient ani mals is that manganese deficiency impaired hepatic cholesterol secretion. Liver cholesterol concentrations tended to be greater in the manganese-deficient animals than in rats fed adequate manganese ad libitum (3.25 vs. 2.88 mg/g). Bell and Hurley (50)found that lipid droplets accumulated in the livers of manganese-deficient mice. Furthermore, the converse has been observed. Hambidge et al. (52)noted that blockage of cholesterol secre tion by the liver in patients resulted in elevated plasma manganese levels. The significance of the effect of poor manganese status on cholesterol levels of humans cannot be deter Downloaded from https://academic.oup.com/jn/article-abstract/120/5/507/4738598 by Washington University, Law School Library user on 21 May 2018

mined from this study. Unlike humans, rats transport more cholesterol in the HDL fraction than in the LDL fraction. Both Doisy (26) and Friedman et al. (53) ob served lower plasma cholesterol levels in manganesedeficient subjects. However, Friedman et al. (53) observed that cholesterol levels did not respond to man ganese repletion, and HDL cholesterol levels were unaf fected. Interpretation of these data is complicated by the fact that the metabolic diet provided during the study of Friedman et al. (53) was a purified formula that con tained less cholesterol (135 mg/d) than was typical. Further work is needed to evaluate the practical signif icance of manganese depletion on cholesterol and lipo protein metabolism in humans.

LITERATURE CITED 1. THOMSON,A. B. R., OLANTUNBOSUN, D. & VALBERG, L. S. (1971) Interrelation of intestinal transport system for manganese and iron. /. Lab. Clin. Med. 78: 642-655. 2. FLANAGAN, P. R., HABT, J. & VALBERG, L. S. (1980) Comparative effects of iron deficiency induced by bleeding and a low iron diet on the intestinal absorptive interactions of iron, cobalt, manga nese, zinc, lead and cadmium. /. Nutr. 110: 1754-1763. 3. POLLACK, S., GEORGE,J. N., REBA,R. C., KAUFMAN, R. M. & CROSBY, W. H. (1965) The absorption of nonferrous metals in iron defi ciency. /. Clin. Invest. 44: 1470-1473. 4. DŒZ-EWALD, M., WEINTRAUB, L. R. & CROSBY,W. H. (1968) In terrelationship of iron and manganese metabolism. Proc. Soc. Exp. Biol. Med. 129: 448-451. 5. KING, B. D., LASSITER, J. W., NEATHERY, M. W., MILLER,W. J. & GENTRY,R. P. (1980) Effect of lactose, copper and iron on man ganese retention and tissue distribution in rats fed dextrose-casein diets. ]. Alum. Sci. 50: 452-458. 6. ZŒCK, A. (1976) Influence of erythropoiesis stimulation on 54Mn distribution in rats. Acta Physiol. Pol. 27: 287-292. 7. Ho, S. Y., MILLER,W. ]., GENTRY,R. P., NEATHERY,M. W. &. BLACKMAN, D. M. (1984) Effects of high but nontoxic dietary manganese and iron on their metabolism by calves. /. Dairy Sci. 67: 1489-1495. 8. HURLEY,L. S., KEEN,C. L. & LÖNNERDAL, B. (1983) Aspects of trace element interactions during development. Fed. Proc. 42: 1735-1739. 9. IVAN,M. & HIDROGLOU,M. (1980) Effect of dietary manganese on growth and manganese metabolism in sheep. /. Dairy Sci. 63: 385-390. 10. BLACK, J.R., AMMERMAN, C. B., HENRY,P. R. & MILES,R. D. (1984) Effect of dietary manganese and age on tissue trace mineral composition of broiler-type chicks as a bioassay of manganese sources. Poultry Sci. 64: 688-693. 11. GRUDEN,N. (1986) The effect of iron dose on manganese absorp tion in neonatal and weanling rats. Nutr. Rep. Int. 34: 21-27. 12. HARTMAN,R. H., MATRONE,G. & WISE,G. H. (1955) Effects of high dietary manganese on hemoglobin formation. /. Nutr. 57: 429-^39. 13. MATRONE, G., HARTMAN, R. H. & CLAWSON, A. J. (1959) Studies of a manganese-iron antagonism in the nutrition of rabbits and baby pigs. /. Nufr. 67: 309-^317. 14. MAINES,M. D. (1980) Regional distribution of the enzymes of haem biosynthesis and the inhibition of 5-aminolevulinate syn thase by manganese in the rat brain. Biochem. f. 190: 315^21. 15. ZIDENBERG-CHERR, S., KEEN,C. L., LÒNNERDAL, B. & HURLEY, L. S. (1983) Superoxide dismutase activity and lipid peroxidation in the rat: developmental correlations affected by manganese défi-

MM, FE AND LIPID INTERACTIONS ciency. /. Nutr. 113: 2498-2504. 16. ZIDENBERG-CHERR, S., KEEN,C. L., CASEY,S. M. & HURLEY,L. S. (1985) Developmental change affected by Mn deficiency: Mn-superoxide dismutase, CuZn-superoxide dismutase, Mn, Cu, Fe and Zn in mouse tissues. Biol. Trace, Elem. Res. 7: 209-222. 17. DEROSA,G., KEEN,C. L., LEACH,R. M. &. HURLEY,L. S. (1980) Regulation of Superoxide dismutase activity by dietary manga nese. /. Nutr. 110: 795-804. 18. PAYTNER,D. I. (1980) Changes in activity of the manganese Superoxide dismutase enzyme in tissues of the rat with changes in dietary manganese. /. Nutr. 110: 437-447. 19. KEEN,C. L., TAMURA,T., LÔNNERDAL, B., HURLEY,L. S. & HALSTED, C. H. (1985) Changes in hepatic Superoxide dismutase activity in alcoholic monkeys. Am. f. Clin. Nutr. 41: 929-932. 20. DELLARD, C. ]., GOVEMO,V. C. & TAPPEL,A. C. (1983) Relative antioxidant effectiveness of cc-tocopherol and y-tocopherol in iron-loaded rats. /. Nutr. 113: 2266-2273. 21. CURRAN,G. (1954) Effect of certain transition group elements on hepatic synthesis of cholesterol in the rat. /. Biol. Chem. 210: 765-770. 22. CURRAN,G. L. & AZARNOFF,D. L. (1961) Effect of certain tran sition elements on cholesterol biosynthesis. Fed. Proc. 20(suppl 10): 109-111. 23. AMDUR,B., RILLING,H. & BLOCK,K. (1957) The enzymatic con version of mevalonic acid to squalene. /. Chem. Soc. 79: 26462647. 24. TCHEN,T. (1957) On the formation of a phosphorylated deriva tive of mevalonic acid. /. Am. Chem. Soc. 79: 6344-6345. 25. BENEDICT,C., KETT,J. & PORTER,J. (1965) Properties of farnesyl pyrophosphate synthetase of pig liver. Arch. Biochem. Biophys. 110: 611-621. 26. DOISY, E. A. (1973) Effect of deficiency in manganese upon plasma levels of clotting proteins and cholesterol in man. In: Trace Element Metabolism in Animals-2. (Hoekstra, W. G., Suttie, J. W., Ganther, H. E., Mertz, W., eds.), pp. 668-670, University Park Press, Baltimore. 27. ROBY,M. J., VANN,K., FREELAND-GRAVES, J. & SHOREY,R. (1982) Plasma and liver cholesterol in the manganese deficient rat. Fed. Proc. 41: 786 (abs.). 28. KLIMIS-TAVANTZIS, D. J., LEACH,R. &. KRIS-ETHERTON, P. (1983) The effect of dietary manganese deficiency on cholesterol and lipid metabolism in the Wistar rat and in the genetically hypercholesterolemic RICO rat. /. Nutr. 113: 328-336. 29. KAWANO,J., NEY,D. M., KEEN,C. L. & SCHNEEMAN, B. O. (1987) Altered high density lipoprotein composition in manganese-defi cient Sprague-Dawley and Wistar rats. /. Nutr. 117: 902-906. 30. AMERICANINSTITUTEOFNUTRITION(1977) Report of the ad hoc committee on standards for nutritional studies. /. Nutr. 107: 1340-1348. 31. GREGER,J. L, BULA,E. N. & GUM, E. T. (1985) Mineral metabo lism of rats fed moderate levels of various aluminum compounds for short periods of time. /. Nutr. 115: 1708-1716. 32. MARKLUND,S. & MARKLUND,G. (1974) Involvement of the superoxide anión in the autooxidation of pyrogallol and a conve nient assay for Superoxide dismutase. Eur. /. Biochem. 47: 469-474. 33. NEY,D. M., LEFEVRE, M. & SCHNEEMAN, B. O. (1986) Alteration in lipoprotein composition with intravenous compared to intragastric fat feeding in the rat. /. Nutr. 116: 2106-2120. 34. NEY, D. M., LASEKAN, J. B. & SHINNICK,F. L. (1988) Soluble oat fiber tends to normalize lipoprotein composition in cholesterolfed rats. /. Nutr. 118: 1455-1462.

Downloaded from https://academic.oup.com/jn/article-abstract/120/5/507/4738598 by Washington University, Law School Library user on 21 May 2018

513

35. FOLCH,J., LEES,M. & SLOANE-STANLEY, G. M. (1957) A simple method for the isolation and purification of total lipids from animal tissue. /. Biol. Chem. 226: 497-509. 36. NEY, D. M., LASEKAN, J. B. & KIM, J. (1989) Relative effects of dietary oleic and linoleic-rich oils on plasma lipoprotein compo sition in rats. f. Nutr. 119: 857-863. 37. ALLAIN,C. C., POON,L. S., CHEN,C. S. G., RICHMOND,W. & Fu, P. C. (1974) Enzymatic determination of total serum cholesterol. Clin. Chem. 20: 470-475. 38. PETERSON,G. L. (1977) Simplification of protein assay method of Lowry et al. which is more generally applicable. Anal. Biochem. 83: 346-356. 39. ROUSNER,G. (1970) Two dimensional thin layer Chromato graphie separation of polar lipids and determination of phospholipids by phosphorous analysis of spots. Lipids 5: 494-496. 40. CARLSON,S. E. & GOLDFARB,S. (1977) A sensitive enzymatic method for the determination of free and esterified tissue choles terol. Clin. Chim. Acta 79: 575-582. 41. FLETCHER,M. J. (1968) A colorimetrie method for estimating serum triglycérides.Clin. Chim. Acta 22: 393-397. 42. DORY,L. & ROHEIM,P. S. (1981) Rat plasma lipoproteins and apolipoproteins in experimental hypothyroidism. /. Lipid Res. 22: 287-296. 43. SAS INSTITUTEINC. (1985) SAS®User's Guide: Statistics, Ver sion 5 Edition, SAS (Statistical Analysis System) Institute Inc., Cary, NC. 44. PAYNTER,D. I. (1980) The role of dietary copper, manganese, selenium, and vitamin E in lipid peroxidation in tissues of the rat. Biol. Trace Elem. Res. 2: 121-135. 45. STEVENS,J. B. & AUTOR,A. P. (1980) Proposed mechanism for neonatal rat tolerance to normobaric hyperoxia. Fed. Proc. 39: 3138^143. 46. DUBICK,M. A. & KEEN,C. L. (1983) Tissue trace elements and lung Superoxide dismutase activity in mice exposed to ozone. Toxicol. Lett. 17:355-360. 47. SCIENCE ANDEDUCATION ADMINISTRATION (1980) Food and nutri ent intakes of individuals in 1 day in the United States, Spring 1977, U.S. Department of Agriculture, Washington, D.C., pp. 75. 48. ZIDENBERG-CHERR, S., HURLEY,L. A., LÖNNERDAL, B. & KEEN,C. L. 11985) Manganese deficiency: effects on susceptibility to ethanol toxicity in rats. /. Nutr. 115: 460-467. 49. LOOKER,A., SEMPOS,C. T, JOHNSON,C. & YETLEY,E. A. (1988) Vitamin-mineral supplement use: Association with dietary in take and iron status of adults. /. Am. Diet. Assoc. 48: 808-814. 50. BELL,L. T. &. HURLEY,L. S. (1973) Ultrastructural effects of manganese deficiency in liver, heart, kidney and pancreas of mice. Lab. Invest. 29: 723-736. 51. DOUGHERTY, J. J., CROFT,W. A. &.HOEKSTRA, W. G. (1981) Effects of ferrous chloride and iron dextran on lipid peroxidation in vivo on vitamin E and selenium adequate and deficient rats. /. Nutr. Ill: 1784-1796. 52. HAMBIDGE, K. M., SOKOL,R. J., FIDANZA,S. J. & GOODALL,M. A. (1989) Plasma manganese concentrations in infants and children receiving parenteral nutrition. /. Parenter. Enter. Nutr. 13: 168171. 53. FRIEDMAN, B. J., FREELAND-GRAVES, J. H., BALES,C. W., BEHMARDI, F. S., KUTSCHKE, R. L., WILLIS,R. A., CROSBY,J. B., TRICKETT,P. C. & HOUSTAN,S.D. (1987) Manganese balance and clinical obser vations in young men fed a manganese-deficient diet. /. Nutr. 117: 133-143.