BIOLOGICAL TRACE ELEMENT RESEARCH 3, 245-256 (1981)
Effect of Form of Iron on Nickel Deprivation in the Rat Liver Content of Copper, Iron, Manganese, and Zinc FORREST H. NIELSEN*
AND TERRENCE R.
SHULER
United States Department of Agriculture, Science and Education Administration, Grand Forks Human Nutrition Research Center, Grand Forks, North Dakota 58202 Received May 27, 1980; Accepted June 10, 1981
Abstract In three fully crossed, factorially arranged, completely randomized experiments, female weanling rats were fed a basal diet (containing about 10 ng of nickel and 2.3 txg of iron/g) supplemented with graded levels of nickel and iron. Iron was supplemented to the diet in experiment 1 at levels of 0, 25, 50, and 100 txg/g as a mixture of 40% FeSO4-nH20 and 60% Fe2(SO4)3"nHzO; in experiment 2 at levels of 0, 12.5, 25, 50, and 100 ~g/g as Fez(SO4)3"nHzO;in experiment 3 at levels of 0, 25, and 50 ixg/g as either the mixture of ferric-ferrous sulfates, or as ferric sulfate only. Nickel as NiCI~.3HzO was supplemented to the diet in experiment 1 at levels of 0, 5, and 50 Ixg/g; in experiment 2 at levels of 0 and 50 Ixg/g; and in experiment 3 at levels of 0 and 5 Ixg/g. Regardless of dietary nickel, rats fed no supplemental iron exhibited depressed iron content and elevated copper, manganese, and zinc contents in the liver. Nickel and iron did not interact to affect iron, manganese, and zinc in liver. Liver copper was inconsistently affected by an interaction between nickel and iron. Nickel deprivation apparently accentuated the elevation of the copper level in livers of severely iron-deficient rats. Experiment 3 showed that the form of dietary iron altered the effect of nickel deprivation on the iron content of the liver. When only ferric sulfate was supplemented to the diet, liver iron content was depressed in nickel-deprived rats. On the @Copyright 1981 by The Humana Press Inc. All fights of any nature whatsoever reserved. 0163--4992/81/0900-024552.40
245
246
NIELSEN AND SHULER
other hand, when the ferric-ferrous mixture was supplemented to the diet, nickel deprivation apparently elevated the iron content in the liver. The findings support the views that (1) parameters that are affected by an interaction between nickel and iron are limited in factorially arranged experiments, and (2) the form and level of dietary iron markedly influence the effect of nickel deprivation in the rat. Index Entries: Nickel, interaction with iron; iron, interaction with nickel; nickel-iron interaction; rat, iron effect on nickel deprivation in; liver, rat, effect of iron on Ni-deprived; liver, rat, Cu,Mn, and Zn in; copper, in rat liver; manganese, in rat liver; zinc, in rat liver.
Introduction Previous studies in our laboratory (1-3) showed that the form of supplemental dietary iron affected the interaction between nickel and iron in the rat. The interaction between nickel and iron affected hematocrit, hemoglobin level, and plasma alkaline phosphatase activity when the dietary iron supplement was ferric sulfate only, but not when the dietary iron supplement was a mixture of ferric and ferrous sulfates. Regardless of the form of dietary iron, nickel and iron did not interact to affect plasma and liver total lipids or phospholipids. The form of supplemental iron also affected the signs of nickel deprivation in the rat (1-3). When dietary iron was supplied as ferric sulfate only, nickel deprivation depressed hematocrit and hemoglobin level, and elevated plasma and liver total lipids. When dietary iron was supplied as a ferric-ferrous sulfate mixture, nickel deprivation depressed hematocrit and hemoglobin level less markedly and depressed, rather than elevated plasma lipids. Liver total lipids were not affected. We, therefore, conducted the following study to ascertain whether the form of dietary iron influenced the effects of dietary nickel and of an interaction between nickel and iron on the contents of copper, iron, manganese, and zinc in liver. We examined those elements because previous studies (1, 4) showed that they were affected by nickel and/or iron nutriture.
Materials and Methods Female weanling Sprague-Dawley rats (Sprague-Dawley, Inc., Madison, WI)* were weighed individually upon arrival and housed three per all-plastic cage measuring 50 x 24 x 16 cm (5) and located inside a mass air flow rack (Lab Products, Garfield, NJ). The rats were assigned to groups of six by the following fully crossed factorial designs: experiment 1, two way, three by four; experiment *Mention of a trademark or proprietary product does not constitute a guarantee or warranty of the product by the US Department of Agriculture, and does not imply its approval to the exclusion of other products that may also be suitable.
IRON EFFECT ON NICKEL DEPRIVATION
247
2, two way, three by five; experiment 3, three way, two by two by three. The levels of iron and nickel supplemented to the basal diet were variables in all experiments. Iron form was the third variable in experiment 3. In experiment 1, the basal diet was supplemented with iron at 0, 25, 50, and 100 Ixg/g and with nickel at 0, 5, and 50 ~g/g. Experiment 2 was identical, except for an additional level of iron, 12.5 p~g/g. In experiment 3, the basal diet was supplemented with iron at 0, 25, and 50 ~g/g and with nickel at 0 and 5 Ixg/g. Nickel was supplemented as NiC12-3HzO ("Ultrapure" grade, Alfa Inorganics, Beverly, MA). Iron was a mixture of 60% Fe2(SOg)3"nH20 and 40% FeSO4"nH20 in experiment 1; was Fez(SOg)3"nH20 in experiment 2. Those forms of iron were used as a treatment variable in experiment 3. The ferric and ferrous sulfates were prepared from iron sponge ("Specpure" grade, Fisher Scientific Co., Chicago, IL) and sulfuric acid (' 'Ultrex" grade, J. T. Baker Chemical Co., Phillipsburg, NJ) by described methods (6). Glucose was added to the ferric, and the ferric-ferrous, sulfate powders to make mixes containing 25 mg Fe/g, as determined by atomic absorption spectrometric methods (1). This was necessary because the waters of hydration of the prepared ferric and ferrous sulfates were not determined. The ferric supplement was ascertained to be 92%, and the ferric-ferrous supplement 60%, in the ferric form by use of the method described by Harris and Kratochvil (7). The iron mixes were added to the diet at the expense of the ground corn. The rats had access to distilled, deionized (Super Q System, Millipore Corp., Bedford, MA) water in plastic cups. Fresh food in plastic cups was provided each day. Plastic equipment and cleaning procedures were described (5, 8, 9). Absorbent paper under the cages caught droppings and was changed every other day. Room temperature was 23~ and lighting was controlled automaticlaly to provide 12 h of light and 12 h of darkness. The diets were mixed 3 days before the start of each experiment, and about biweekly, thereafter. They were stored at - 16~ in tightly capped plastic containers. The basal diet (1) contained about 10 ng of nickel/g and 2.3 ~g ofiron/g on an airdried basis, as determined by atomic absorption spectrometric methods (1). The rats were fed their respective diets for 10 weeks in experiments 1 and 3, and 9 weeks in experiment 2, then decapitated subsequent to ether anesthesia and cardiac exsanguination with a heparin-coated syringe. The liver was immediately removed and the portion not used for lipid analyses (3) was immediately frozen for later trace element anysis. Liver samples were prepared for analyses by a described method (1). Copper, iron, manganese, and zinc were determined by atomic absorption spectrophotometry (10). Because of an accident during the ashing procedure, the liver samples from rats fed 5 Ixg of nickel/g of diet in experiment 2 were lost. Thus, for the statistical analyses, experiment 2 was a two by five factorial design. Data were treated by two-way (experiments 1 and 2), or three-way (experiment 3), analysis of variance and differences between individual treatment means and groups were evaluated for significance by the Scheff6 test with a per experiment rate of 0.05 (11). For comparison, data from a previous study (1) are included in the tables.
248
NIELSEN AND SHULER
Results All rats fed the diet without iron supplementation developed classic signs of iron deficiency; lethargy, rough coat, pale eyes, incisor depigmentation, and depressed growth. Rats fed 25, 50, or 100 Ixg of iron/g of diet appeared healthy and normal. However, the hematocrit data (2) indicated that rats fed iron at 12.5, 25, and 50 Ixg/g as ferric sulfate only were anemic, and thus iron-deficient. This indicates that rats fed ferric sulfate only consumed the intended relatively unavailable form of iron, whereas rats fed the ferric-ferrous mixture were receiving a substantial portion of their iron as the more available form. Severe iron deficiency markedly elevated the copper content in liver (Table 1). In the previous study (1) and experiment 3, nickel deprivation accentuated this elevation. In experiment 1, the liver copper data from the severely iron-deficient rats were extremely variable (large error mean square). Thus, even though nickel deprivation apparently elevated the copper content of liver, the variability in the severely iron-deficient group probably prevented the statistical demonstration of an effect of either nickel or a nickel-iron interaction. As expecte'd, iron deficiency depressed the iron content in liver (Table 2). The significant interaction between nickel and iron form in experiment 3 showed that the effect of nickel depended upon the form of dietary iron. Nickel deprivation apparently elevated the iron content in livers of rats fed less than 100 fxg of iron as a ferric-ferrous sulfate mixture per gram of diet. On the other hand, nickel deprivation depressed the iron content in livers of rats fed low levels of supplemental ferric sulfate only. The difference between the nickel-deprived and -supplemented groups was significant (P < 0.01) in experiment 1 and approached significance (P < 0.06) in experiment 2. The Scheff6 value of 76 (18 s-test 18 in Table 2) shows that, in experiment 3, when ferric sulfate only was fed, the mean iron content was significantly lower (200 vs 344 txg/g), and when the ferric-ferrous mixture was fed, was slightly (but not significantly by the conservative Scheff6 test) higher (453 vs 404 t~g/g) in nickel-deprived than nickel-supplemented rats. Regardless of the form of dietary iron, nickel and iron did not interact to affect the iron content. In experiment 3, the significant effect of iron form, and of the interaction between iron form and iron, further demonstrated that rats fed ferric sulfate only consumed the intended relatively unavailable form of iron. At the same level of iron supplementation, liver iron content was lower in rats supplemented only with ferric sulfate than in rats supplemented with ferric-ferrous sulfate. The data in Table 3 show that iron deficiency elevated the manganese content in liver of rats. Nickel deprivation apparently elevated the manganese content in liver in experiment 1, but not in ahy other experiment. Regardless of the form of dietary iron, liver manganese content was not affected by an interaction between nickel and iron. Severe iron deficiency elevated the zinc content in liver (Table 4). In experiments 1 and 2, nickel deprivation also appaarently slightly elevated the zinc content in liver. The effect of dietary nickel was most obvious in the severely irondeficient rats. Nickel and iron did not interact to affect the zinc content in liver.
Table 1 E f f e c t s in R a t s o f N i c k e l , I r o n , a n d T h e i r I n t e r a c t i o n o n L i v e r C o p p e r L i v e r c o p p e r , b lxg/g F e 3+ F e 2+ ' F e 3+ Previous Treatmenta Ni, p~g/g
Exp. 1
Exp. 3
Exp. 3
Exp. 2
study (1)
F e , t~g/g
0
0
177
245
172
100
174
0
12.5
--
--
--
78
--
0
25
25
17
28
30
20
0
50
17
16
19
36
17
0
100
16
--
--
73
164
119 .
21
15
5 5
0 12.5
5
25
19
17
5
50
16
16
5
100
16
--
--
--
16
109 --
---
---
109 43
53 --
.
.
.
--
89
16
--
22
16
--
16
.
50 50
0 12.5
50
25
18
--
--
31
17
50 50
50 100
16 15
---
---
26 18
17 17
An~ysis of Variance--P Values Nickel e f f e c t Iron e f f e c t
NS
0.007
NS
0.0002
0.0001
0.0001
0.0001
0.0001
Nickel x iron
NS
0.004
NS
0.0001
Iron f o r m e f f e c t Nickel x iron f o r m
---
---
---
Iron • iron f o r m
--
--
--
Nickel x iron • iron f o r m
--
NS
--
--
1743
1285
922
338
82 44
56 --
Error m e a n s q u a r e
NS NS 0.006
Scheff6 Values C 6 s-test 6 12 s-test 12
124 --
97 51
18 s-test 18 24 s-test 24
45 34
35 26
30 s-test 30
--
36 s-test 36
--
-17
--18 --
21 16 ---
~Amounts o f Ni (nickel chloride) and Fe (either a mixture of 40% ferrous sulfate and 60% ferric sulfate, or ferric sulfate) supplemented to diet. ~Dry weight basis. ~The Scheff6 test (11) is a method for performing multiple comparisons between group means. Means differing by more than the value given are significantly different (P < 0.05). To make a comparison, ascertain the number of animals used to determine a mean. Each treatment contained six animals.
Table 2 E f f e c t s in R a t s o f N i c k e l , I r o n , and T h e i r I n t e r a c t i o n on L i v e r Iron L i v e r iron, b Ixg/g F e 3+ F e 2+, F e 3+ Previous Treatmenta N i , Ixgtg
Exp. 1
Exp. 3
Exp. 3
69 --
Exp. 2
s t u d y (1)
F e , lxg/g
0 0
0 12.5
75 I
65 --
0
25
461
417
114
80
115
0 0
50 100
562 541
878 I
416 --
159 331
354 426
59
68
0
5
12.5
5
25
283
376
301
--
163
5 5
50 100
435 545
776 i
661 __
---
440 395
.
.
.
--
66 --
5
.
61
59 73
59
.
50 50
0 12.5
70 i
---
---
67 83
68 --
50 50
25 50
389 510
-i
---
94 290
227 368
50
100
586
--
--
369
518
Analysis of VafianceIp
Values
Nickel effect
0.01
0.02
0.06
Iron effect
0.0001
0.0001
0.0001
0.0001
NS
N i c k e l x iron
NS
NS
NS
NS
Iron form effect
--
0.0001
--
--
N i c k e l x iron f o r m I r o n x iron f o r m
-i
0.0001 0.0001
---
---
I 8778
0.002 6232
-6734
10,990
N i c k e l x iron x iron f o r m Error mean square
Scheff6 Values c 277
214
222
12 s-test 12
6 s-test 6
i
111
119
18 s-test 18
193
76
--
24 s-test 24
77
57
--
30 s-test 30
__
36 s-test 36
--
I
49 37
--
302 -111 85 ---
'~Amounts of Ni (nickel chloride) and Fe (either a mixture of 40% ferrous sulfate and 60% ferric sulfate, or ferric sulfate) supplemented to diet. t~Dry weight basis. CThe Scheff6 test (11) is a method for performing multiple comparisons between group means. Means differing by more than the value given are significantly different (P < 0.05). To make a comparison, ascertain the number of animals used to determine a mean. Each treatment contained six animals. 250
Table 3 E f f e c t s in Rats o f N i c k e l , I r o n , a n d T h e i r I n t e r a c t i o n o n L i v e r M a n g a n e s e L i v e r m a n g a n e s e , b p,g/g F e 3+ F e z+ , F e 3+ Previous Treatment~ Ni, i.Lg/g
Exp. 1
Exp. 3
Exp. 3
Exp. 2
study (1)
13.0 --
7.6 --
7.8 --
9.0 10.6
10.4 --
9.4 8.9
6.5 6.1
7.7 6.8
10.5 10.5
10.0 7.8
Fe, ~g/g
0 0
0 12.5
0 0
25 50
0
100
8.2
--
--
8.5
7.4
12.2 .
7.7
8.2 .
--
9.7
7.1
--
10.7
6.4
--
8.8
--
--
--
7.9
10.2 --
---
---
9.2 10.5
11.0 --
5 5
0 12.5
5
25
9.3
6.9
5
50
8.3
5.9
5
100
6.7
.
.
.
50 50
0 12.5
50 50
25 50
7.9 7.6
---
---
9.9 9.0
10.0 8.2
50
100
7.2
--
--
8.2
7.8
Analysis of Variance
P Values
Nickel effect
0.0001
NS
NS
Iron e f f e c t N i c k e l x iron Iron f o r m e f f e c t
0.0001 NS --
0.0001 NS 0.04
0.0001 NS --
0.05 0.0001 NS --
N i c k e l x iron f o r m
--
NS
--
--
Iron x iron f o r m
--
NS
--
--
N i c k e l x iron x iron f o r m
--
NS
--
--
E r r o r m e a n square
1.0
1.3
1.0
0.7
3.0 1.6
2.7 1.4
2.4 --
Scheff6 Valuesc 6 s-test 6 12 s-test 12
2.9 --
18 s-test 18
1.1
1.1
--
0.9
24 s-test 24
0.6
0.8
--
0.7
30 s-test 30 36 s-test 36
---
-0.5
0.6 --
---
aAmounts of Ni (nickel chloride) and Fe (either a mixture of 40% ferrous sulfate and 60% ferric sulfate, or ferric sulfate) supplemented to diet. t'Dry weight basis ~ Scheff6 test (11) is a method for performing multiple comparisons between group means. Means differing by more than the value given are significantly different (P < 0.05). To make a comparison, ascertain the number of animals used to determine a mean. Each treatment contained six animals. 251
Table 4 E f f e c t s in Rats o f N i c k e l , Iron, and T h e i r I n t e r a c t i o n o n L i v e r Z i n c L i v e r z i n c , b ixg/g F e 3+ F e z+, F e 3+ Previous Treatments Ni, txg/g
Exp. 1
Exp. 3
Exp. 3
Exp. 2
s t u d y (1)
F e , p,g/g
0
0
133
108
108
102
110
0
12.5
--
--
--
81
--
0
25
89
80
81
80
91
0
50
97
81
79
83
89
0 5
100 0
89 105
-100
-99
80 --
87 109
5
12.5
5 5
25 50
.
.
5
I00
83
--
--
50
0
93
--
--
93
113 --
91 91
--
.
.
78 81
78 82
. ----
95 88 93
50
12.5
--
--
76
50
25
83
--
--
71
89
50
50
80
--
--
78
94
50
100
79
--
--
72
91
Analysis of V a r i a n c e - - P Values Nickel effect
0.001
NS
0.004
NS
Iron effect N i c k e l x iron
0.0001 NS
0.0001 NS
0.0001 NS
0.0001 NS
Iron form effect
--
NS
--
--
N i c k e l • iron f o r m
--
NS
--
--
I r o n x iron f o r m
--
NS
--
--
N i c k e l x i r o n x iron f o r m
--
NS
--
--
Error mean square
212
84
77
101
Scheff6 V a l u e s c 6 s-test 6 12 s-test 12 18 s-test 18 24 s-test 24 30 s-test 30 36 s-test 36
43 -16 12 ---
25
24
29
13
13
--
9 7
---
11 8
5 --
---
-4
aAmounts o f Ni (nickel chloride) and Fe (either a mixture of 40% ferrous sulfate and 60% ferric sulfate, or ferric sulfate) supplemented to diet. ~ weight basis. ~The Scheff6 test (11) is a method for performing multiple comparisons between group means. Means differing by more than the value given are significantly different (P < 0.05). To make a comparison, ascertain the number of animals used to determine a mean. Each treatment contained six animals. 252
IRON EFFECT ON NICKEL DEPRIVATION
253
Discussion Findings from previous studies suggested that positive evidence of the interaction of nickel with iron would require examination of the parameters thatare affected by borderline iron deficiency (1-3). Data from examination of only the parameters that are affected by severe iron deficiency probably would not confirm the interaction between nickel and iron. Apparently nickel enhances the absorption of iron only when it is present in the diet in less than adequate but not severely inadequate, levels and in a relatively unavailable form. In severe iron deficiency, iron would not be available for interaction, and nickel probably would have little effect on the absorption of dietary iron. Thus, hematocrit and hemoglobin respond to nickelinduced changes in iron absorption, but total lipids and phospholpids in plasma and liver are changed by severe iron deficiency only and are not affected by an interaction between nickel and iron. Because borderline iron deficiency apparently elevated the copper and manganese contents, and depressed the iron content in liver, an interaction between nickel and iron might be expected to affect the copper, iron and, perhaps, manganese content in liver. Failure, in the present study, to find that nickel and iron interacted to affect iron and manganese, and copper consistently, might be explained as follows: In the previous study (1) and experiment 3, the interaction between nickel and iron affected the copper content in liver and the error mean square (Table 1) indicates that the copper data were less variable than those in experiment 1. In the previous study (1) and experiment 3, nickel deprivation mainly accentuated the elevation of the copper level in livers of severely iron-deficient rats. In experiments 1 and 2, the copper values from iron-supplemented rats were quite consistent, but this was probably masked by the extremely variable values from the severely irondeficient rats. Because of this variance, relatively large differences between means would have been required for statistical significance, and the relatively small effects of an interaction between nickel and iron, if present, were obscured. Thus, the data are not unequivocal, but suggest that an interaction between nickel and iron affects the copper content in liver of rats. Statistical demonstration of that interaction, however, would depend upon the experimental design (i. e., use a large number of animals), or the fortuitous consistent elevation of the copper content in livers of severely iron-deficient rats. Experiment 3 was an example of this fortuitous elevation of copper content. The liver copper should have been about equal in the two groups fed no supplemental iron or nickel or fed no supplemental iron and 5 txg of nickel/g of diet. Liver copper was not equal in those two groups. The highest values were usually found in groups designated for the ferric-ferrous mixture treatment. The copper values were consistently lower in the severely iron-deficient rats assigned to the ferric treatment than for rats assigned to the ferric-ferrous treatment. Further evidence that an interaction between nickel and iron affected the liver copper content was found in a previous study (4), in which nickel-deprived rats fed an apparently inadequate amount of iron, as ferric sulfate, exhibited, at age 55 days, elevated copper content in the liver. In rats fed an apparently adequate level of dietary iron as ferric chloride, nickel deprivation did not affect the copper content.
254
NIELSEN AND SHULER
55 days, elevated copper content in the liver. In rats fed an apparently adequate level of dietary iron as ferric chloride, nickel deprivation did not affect the copper content. In previous studies (1,2), when the iron supplement was ferric sulfate only, the interaction between nickel and iron affected hematocrit and hemoglobin. These parameters increased and reached a plateau as dietary iron increased. At the low levels of iron supplementaion, nickel deprivation depressed hematocrit and hemoglobin levels. At high levels of iron supplementaion, where these parameters had reached a plateau, nickel deprivation had no effect. Thus, an interaction between nickel and iron was shown because nickel affected hematocrit and hemoglobin differently at different levels of iron supplementation. In the present study, the iron and manganese contents of the liver reacted differently to increasing dietary iron. As the dietary iron increased, generally liver iron content increased and manganese content decreased--no plateau was reached. There was no point at which the effect of dietary nickel was precluded, which probably explains the absence of interaction between nickel and iron. Although the interaction between nickel and iron was not statistically significant, the form of dietary iron did alter the effect of dietary nickel on the iron content of the liver. When only ferric sulfate was supplemented to the diet, liver iron was significantly lower in nickel-deprived than in nickel-supplemented rats in experiment 3; the difference approached significance (P < 0.06) in experiment 2. In contrast, when the dietary iron was a ferric-ferrous sulfate mixture, the liver content of iron was higher in nickel-deprived than in nickel-supplemented rats; the difference was significant (P < 0.01) in experiment 1. Similar findings were published (4). In those studies, at age 35 and 55 days, nickel-deprived rats fed 60 p~g iron/g diet as ferric sulfate exhibited a depressed concentration of iron in the liver. At age 30 days, the iron concentration in livers from rats fed 30 p~g iron/g diet as ferric chloride was higher in nickel-deprived than in nickel-supplemented rats. Also, at age 55 days, the iron concentration tended to be higher in the nickeldeprived than in nickel-supplemented rats. Apparently the form and level of iron fed to rats influenced the direction of the change in liver iron concentration that was associated with nickel deprivation. The changes in manganese content in liver caused by various treatments were small in the present study and probably of little physiological meaning. Perhaps any effect on the manganese content of the liver was indirectly caused by changes in liver iron content induced by dietary nickel and iron. Apparently the iron content of the liver has a major influence on its manganese content. The findings in the present study and in previous studies (1, 4) suggest that as iron increases in the liver, manganese decreases. Schnegg and Kirchgessner (12-14) found that nickel deprivation depressed the total copper, iron, and zinc contents in liver of rats. The initial impression is that their findings contrast with those of our present and previous (4) studies, which showed that nickel deprivation apparently elevated the iron (ferric-ferrous mixture fed), copper and zinc concentration, on a per gram basis, in livers of rats, espe-
IRON EFFECT ON NICKEL DEPRIVATION
255
cially if they were iron-deficient. However, if the data of Schnegg and Kirchgessner (12) is recalculated to give the ~g copper, iron, or zinc/g liver, much of the divergence in findings is eliminated. When their data is expressed on the per gram basis, the liver content of copper and zinc was as high, or higher, in nickeldeprived as in nickel-supplemented rats. Still, the liver content of iron on the per gram basis was generally lower in nickel-deprived than nickel-supplemented rats. In our studies, this was observed in nickel-deprived rats fed ferric sulfate only. We did not determine the total weight of the dry liver in these experiments because a portion of the fresh liver was used for lipid analysis. Thus, the total copper, iron, and zinc contents of the liver could not be accurately determined. However, we made a rough estimate of total contents (data not presented) by using the total fresh weight, and the dry weight concentrations of copper, iron, and zinc, of liver. These calculations did not change any of our major conclusions stated in this discussion. On the whole liver, in contrast to the per gram liver, basis, we found that: (1) manganese was not elevated in rats fed no supplemental iron, (2) manganese was not elevated in nickel-deprived rats in experiment 1, and (3) the elevation in the zinc content was less marked in severely iron-deficient rats. Thus, on the whole liver basis, our findings apparently contrast with those of Schnegg and Kirchgessner (12-14). Perhaps part of this divergence can be explained by the difference in experimental design. In other words, factors such as severity of nickel deprivation, form or level of dietary iron, or age of animals influenced the direction and magnitude of the change in total copper, iron, and zinc contents in the livers of nickeldeprived rats. This might be expected because these factors were shown to influence signs of nickel deprivation (3, 4, 15). The findings in the present and previous studies (1-3) indicate that the parameters that are affected by an interaction between nickel and iron are limited in factorially arranged experiments. These parameters reach a plateau when dietary iron is adequate, and vary directly with dietary iron when it is inadequate. Parameters that did not reach a plateau with the iron spplements used, or were affected by severe iron deficiency only, are not suitable to ascertain a nickel-iron interaction. Furthermore, the findings show that the form of dietary iron markedly influences the interaction between nickel and iron, and the effect of nickel deprivation, in the rat.
Acknowledgment The authors thank Barry Shull and LuAnn Johnson for their help with the statistical analysis of the data.
References 1. F. H. Nielsen, T. R. Shuler, T. J. Zimmerman, M. E. Collings and E. O. Uthus, Biol. Trace Element Res. 1, 325 (1979). 2. F. H. Nielsen, J. Nutr. 110, 965 (1980).
256 3. 4. 5. 6.
7. 8. 9. 10. 11. 12. 13. 14.
15.
NIELSEN AND SHULER
F. H. Nielsen, Biol. Trace Element Res. 2, 199 (1980). F. H. Nielsen and T. R. Shuler, Biol. Trace Element Res. 1, 337 (1979). F. H. Nielsen and B. Bailey, Lab. Anim. Sci. 29, 502 (1979). E. Pietsch, G. Blinoff-Achapkin, H. Gruss, A. Kotowski, M. Du Made, and G. Nachod, Gmelins Handbuch Der Anorganischen Chemie, Verlag Chemie G.m.b.H., Berlin, 1957, pp. 395,440. W. E. Harris and B. Kratochvil, Chemical Separation and Measurements, Saunders, Philadelphia, 1974, pp. 124-132. F. H. Nielsen, D. R. Myron, S. H. Givand, and D. A. Ollerich, J. Nutr. 105, 1607 (1975). F. H. Nielsen, D. R. Myron, S. H. Givand, T. J. Zimmerman, and D. A. Ollerich, J. Nutr. 105, 1620 (1975). Analytical Methods for Atomic Absorption Spectrophotometry, Perkin-Elmer Corp., Norfolk, CT (1976). H. Scheff6, The Analysis of Variance, Wiley, New York, pp. 90-137 (1959). A. Schnegg and M. Kirchgessner, Arch. Tierern?ihrung 26, 543 (1976). A. Schnegg and M. Kirchgessner, Internat. Z. Vit. Ern. Forschung 46, 96 (1976). A. Schnegg and M. Kirchgessner, in Trace Element Metabolism in Man and Animals-3 (M. Kirchgessner, ed.), Technical University of Munich, FreisingWeihenstephen, 1978, pp.236-243. F. H. Nielsen, T. J. Zirnmerman, M. E. Collings, and D. R. Myron, J. Nutr. 109, 1623 (1979).
Effect of Form of Iron on Nickel Deprivation in the Rat Liver Content of Copper, Iron, Manganese, and Zinc FORREST H. NIELSEN*
AND TERRENCE R.
SHULER
United States Department of Agriculture, Science and Education Administration, Grand Forks Human Nutrition Research Center, Grand Forks, North Dakota 58202 Received May 27, 1980; Accepted June 10, 1981
Abstract In three fully crossed, factorially arranged, completely randomized experiments, female weanling rats were fed a basal diet (containing about 10 ng of nickel and 2.3 txg of iron/g) supplemented with graded levels of nickel and iron. Iron was supplemented to the diet in experiment 1 at levels of 0, 25, 50, and 100 txg/g as a mixture of 40% FeSO4-nH20 and 60% Fe2(SO4)3"nHzO; in experiment 2 at levels of 0, 12.5, 25, 50, and 100 ~g/g as Fez(SO4)3"nHzO;in experiment 3 at levels of 0, 25, and 50 ixg/g as either the mixture of ferric-ferrous sulfates, or as ferric sulfate only. Nickel as NiCI~.3HzO was supplemented to the diet in experiment 1 at levels of 0, 5, and 50 Ixg/g; in experiment 2 at levels of 0 and 50 Ixg/g; and in experiment 3 at levels of 0 and 5 Ixg/g. Regardless of dietary nickel, rats fed no supplemental iron exhibited depressed iron content and elevated copper, manganese, and zinc contents in the liver. Nickel and iron did not interact to affect iron, manganese, and zinc in liver. Liver copper was inconsistently affected by an interaction between nickel and iron. Nickel deprivation apparently accentuated the elevation of the copper level in livers of severely iron-deficient rats. Experiment 3 showed that the form of dietary iron altered the effect of nickel deprivation on the iron content of the liver. When only ferric sulfate was supplemented to the diet, liver iron content was depressed in nickel-deprived rats. On the @Copyright 1981 by The Humana Press Inc. All fights of any nature whatsoever reserved. 0163--4992/81/0900-024552.40
245
246
NIELSEN AND SHULER
other hand, when the ferric-ferrous mixture was supplemented to the diet, nickel deprivation apparently elevated the iron content in the liver. The findings support the views that (1) parameters that are affected by an interaction between nickel and iron are limited in factorially arranged experiments, and (2) the form and level of dietary iron markedly influence the effect of nickel deprivation in the rat. Index Entries: Nickel, interaction with iron; iron, interaction with nickel; nickel-iron interaction; rat, iron effect on nickel deprivation in; liver, rat, effect of iron on Ni-deprived; liver, rat, Cu,Mn, and Zn in; copper, in rat liver; manganese, in rat liver; zinc, in rat liver.
Introduction Previous studies in our laboratory (1-3) showed that the form of supplemental dietary iron affected the interaction between nickel and iron in the rat. The interaction between nickel and iron affected hematocrit, hemoglobin level, and plasma alkaline phosphatase activity when the dietary iron supplement was ferric sulfate only, but not when the dietary iron supplement was a mixture of ferric and ferrous sulfates. Regardless of the form of dietary iron, nickel and iron did not interact to affect plasma and liver total lipids or phospholipids. The form of supplemental iron also affected the signs of nickel deprivation in the rat (1-3). When dietary iron was supplied as ferric sulfate only, nickel deprivation depressed hematocrit and hemoglobin level, and elevated plasma and liver total lipids. When dietary iron was supplied as a ferric-ferrous sulfate mixture, nickel deprivation depressed hematocrit and hemoglobin level less markedly and depressed, rather than elevated plasma lipids. Liver total lipids were not affected. We, therefore, conducted the following study to ascertain whether the form of dietary iron influenced the effects of dietary nickel and of an interaction between nickel and iron on the contents of copper, iron, manganese, and zinc in liver. We examined those elements because previous studies (1, 4) showed that they were affected by nickel and/or iron nutriture.
Materials and Methods Female weanling Sprague-Dawley rats (Sprague-Dawley, Inc., Madison, WI)* were weighed individually upon arrival and housed three per all-plastic cage measuring 50 x 24 x 16 cm (5) and located inside a mass air flow rack (Lab Products, Garfield, NJ). The rats were assigned to groups of six by the following fully crossed factorial designs: experiment 1, two way, three by four; experiment *Mention of a trademark or proprietary product does not constitute a guarantee or warranty of the product by the US Department of Agriculture, and does not imply its approval to the exclusion of other products that may also be suitable.
IRON EFFECT ON NICKEL DEPRIVATION
247
2, two way, three by five; experiment 3, three way, two by two by three. The levels of iron and nickel supplemented to the basal diet were variables in all experiments. Iron form was the third variable in experiment 3. In experiment 1, the basal diet was supplemented with iron at 0, 25, 50, and 100 Ixg/g and with nickel at 0, 5, and 50 ~g/g. Experiment 2 was identical, except for an additional level of iron, 12.5 p~g/g. In experiment 3, the basal diet was supplemented with iron at 0, 25, and 50 ~g/g and with nickel at 0 and 5 Ixg/g. Nickel was supplemented as NiC12-3HzO ("Ultrapure" grade, Alfa Inorganics, Beverly, MA). Iron was a mixture of 60% Fe2(SOg)3"nH20 and 40% FeSO4"nH20 in experiment 1; was Fez(SOg)3"nH20 in experiment 2. Those forms of iron were used as a treatment variable in experiment 3. The ferric and ferrous sulfates were prepared from iron sponge ("Specpure" grade, Fisher Scientific Co., Chicago, IL) and sulfuric acid (' 'Ultrex" grade, J. T. Baker Chemical Co., Phillipsburg, NJ) by described methods (6). Glucose was added to the ferric, and the ferric-ferrous, sulfate powders to make mixes containing 25 mg Fe/g, as determined by atomic absorption spectrometric methods (1). This was necessary because the waters of hydration of the prepared ferric and ferrous sulfates were not determined. The ferric supplement was ascertained to be 92%, and the ferric-ferrous supplement 60%, in the ferric form by use of the method described by Harris and Kratochvil (7). The iron mixes were added to the diet at the expense of the ground corn. The rats had access to distilled, deionized (Super Q System, Millipore Corp., Bedford, MA) water in plastic cups. Fresh food in plastic cups was provided each day. Plastic equipment and cleaning procedures were described (5, 8, 9). Absorbent paper under the cages caught droppings and was changed every other day. Room temperature was 23~ and lighting was controlled automaticlaly to provide 12 h of light and 12 h of darkness. The diets were mixed 3 days before the start of each experiment, and about biweekly, thereafter. They were stored at - 16~ in tightly capped plastic containers. The basal diet (1) contained about 10 ng of nickel/g and 2.3 ~g ofiron/g on an airdried basis, as determined by atomic absorption spectrometric methods (1). The rats were fed their respective diets for 10 weeks in experiments 1 and 3, and 9 weeks in experiment 2, then decapitated subsequent to ether anesthesia and cardiac exsanguination with a heparin-coated syringe. The liver was immediately removed and the portion not used for lipid analyses (3) was immediately frozen for later trace element anysis. Liver samples were prepared for analyses by a described method (1). Copper, iron, manganese, and zinc were determined by atomic absorption spectrophotometry (10). Because of an accident during the ashing procedure, the liver samples from rats fed 5 Ixg of nickel/g of diet in experiment 2 were lost. Thus, for the statistical analyses, experiment 2 was a two by five factorial design. Data were treated by two-way (experiments 1 and 2), or three-way (experiment 3), analysis of variance and differences between individual treatment means and groups were evaluated for significance by the Scheff6 test with a per experiment rate of 0.05 (11). For comparison, data from a previous study (1) are included in the tables.
248
NIELSEN AND SHULER
Results All rats fed the diet without iron supplementation developed classic signs of iron deficiency; lethargy, rough coat, pale eyes, incisor depigmentation, and depressed growth. Rats fed 25, 50, or 100 Ixg of iron/g of diet appeared healthy and normal. However, the hematocrit data (2) indicated that rats fed iron at 12.5, 25, and 50 Ixg/g as ferric sulfate only were anemic, and thus iron-deficient. This indicates that rats fed ferric sulfate only consumed the intended relatively unavailable form of iron, whereas rats fed the ferric-ferrous mixture were receiving a substantial portion of their iron as the more available form. Severe iron deficiency markedly elevated the copper content in liver (Table 1). In the previous study (1) and experiment 3, nickel deprivation accentuated this elevation. In experiment 1, the liver copper data from the severely iron-deficient rats were extremely variable (large error mean square). Thus, even though nickel deprivation apparently elevated the copper content of liver, the variability in the severely iron-deficient group probably prevented the statistical demonstration of an effect of either nickel or a nickel-iron interaction. As expecte'd, iron deficiency depressed the iron content in liver (Table 2). The significant interaction between nickel and iron form in experiment 3 showed that the effect of nickel depended upon the form of dietary iron. Nickel deprivation apparently elevated the iron content in livers of rats fed less than 100 fxg of iron as a ferric-ferrous sulfate mixture per gram of diet. On the other hand, nickel deprivation depressed the iron content in livers of rats fed low levels of supplemental ferric sulfate only. The difference between the nickel-deprived and -supplemented groups was significant (P < 0.01) in experiment 1 and approached significance (P < 0.06) in experiment 2. The Scheff6 value of 76 (18 s-test 18 in Table 2) shows that, in experiment 3, when ferric sulfate only was fed, the mean iron content was significantly lower (200 vs 344 txg/g), and when the ferric-ferrous mixture was fed, was slightly (but not significantly by the conservative Scheff6 test) higher (453 vs 404 t~g/g) in nickel-deprived than nickel-supplemented rats. Regardless of the form of dietary iron, nickel and iron did not interact to affect the iron content. In experiment 3, the significant effect of iron form, and of the interaction between iron form and iron, further demonstrated that rats fed ferric sulfate only consumed the intended relatively unavailable form of iron. At the same level of iron supplementation, liver iron content was lower in rats supplemented only with ferric sulfate than in rats supplemented with ferric-ferrous sulfate. The data in Table 3 show that iron deficiency elevated the manganese content in liver of rats. Nickel deprivation apparently elevated the manganese content in liver in experiment 1, but not in ahy other experiment. Regardless of the form of dietary iron, liver manganese content was not affected by an interaction between nickel and iron. Severe iron deficiency elevated the zinc content in liver (Table 4). In experiments 1 and 2, nickel deprivation also appaarently slightly elevated the zinc content in liver. The effect of dietary nickel was most obvious in the severely irondeficient rats. Nickel and iron did not interact to affect the zinc content in liver.
Table 1 E f f e c t s in R a t s o f N i c k e l , I r o n , a n d T h e i r I n t e r a c t i o n o n L i v e r C o p p e r L i v e r c o p p e r , b lxg/g F e 3+ F e 2+ ' F e 3+ Previous Treatmenta Ni, p~g/g
Exp. 1
Exp. 3
Exp. 3
Exp. 2
study (1)
F e , t~g/g
0
0
177
245
172
100
174
0
12.5
--
--
--
78
--
0
25
25
17
28
30
20
0
50
17
16
19
36
17
0
100
16
--
--
73
164
119 .
21
15
5 5
0 12.5
5
25
19
17
5
50
16
16
5
100
16
--
--
--
16
109 --
---
---
109 43
53 --
.
.
.
--
89
16
--
22
16
--
16
.
50 50
0 12.5
50
25
18
--
--
31
17
50 50
50 100
16 15
---
---
26 18
17 17
An~ysis of Variance--P Values Nickel e f f e c t Iron e f f e c t
NS
0.007
NS
0.0002
0.0001
0.0001
0.0001
0.0001
Nickel x iron
NS
0.004
NS
0.0001
Iron f o r m e f f e c t Nickel x iron f o r m
---
---
---
Iron • iron f o r m
--
--
--
Nickel x iron • iron f o r m
--
NS
--
--
1743
1285
922
338
82 44
56 --
Error m e a n s q u a r e
NS NS 0.006
Scheff6 Values C 6 s-test 6 12 s-test 12
124 --
97 51
18 s-test 18 24 s-test 24
45 34
35 26
30 s-test 30
--
36 s-test 36
--
-17
--18 --
21 16 ---
~Amounts o f Ni (nickel chloride) and Fe (either a mixture of 40% ferrous sulfate and 60% ferric sulfate, or ferric sulfate) supplemented to diet. ~Dry weight basis. ~The Scheff6 test (11) is a method for performing multiple comparisons between group means. Means differing by more than the value given are significantly different (P < 0.05). To make a comparison, ascertain the number of animals used to determine a mean. Each treatment contained six animals.
Table 2 E f f e c t s in R a t s o f N i c k e l , I r o n , and T h e i r I n t e r a c t i o n on L i v e r Iron L i v e r iron, b Ixg/g F e 3+ F e 2+, F e 3+ Previous Treatmenta N i , Ixgtg
Exp. 1
Exp. 3
Exp. 3
69 --
Exp. 2
s t u d y (1)
F e , lxg/g
0 0
0 12.5
75 I
65 --
0
25
461
417
114
80
115
0 0
50 100
562 541
878 I
416 --
159 331
354 426
59
68
0
5
12.5
5
25
283
376
301
--
163
5 5
50 100
435 545
776 i
661 __
---
440 395
.
.
.
--
66 --
5
.
61
59 73
59
.
50 50
0 12.5
70 i
---
---
67 83
68 --
50 50
25 50
389 510
-i
---
94 290
227 368
50
100
586
--
--
369
518
Analysis of VafianceIp
Values
Nickel effect
0.01
0.02
0.06
Iron effect
0.0001
0.0001
0.0001
0.0001
NS
N i c k e l x iron
NS
NS
NS
NS
Iron form effect
--
0.0001
--
--
N i c k e l x iron f o r m I r o n x iron f o r m
-i
0.0001 0.0001
---
---
I 8778
0.002 6232
-6734
10,990
N i c k e l x iron x iron f o r m Error mean square
Scheff6 Values c 277
214
222
12 s-test 12
6 s-test 6
i
111
119
18 s-test 18
193
76
--
24 s-test 24
77
57
--
30 s-test 30
__
36 s-test 36
--
I
49 37
--
302 -111 85 ---
'~Amounts of Ni (nickel chloride) and Fe (either a mixture of 40% ferrous sulfate and 60% ferric sulfate, or ferric sulfate) supplemented to diet. t~Dry weight basis. CThe Scheff6 test (11) is a method for performing multiple comparisons between group means. Means differing by more than the value given are significantly different (P < 0.05). To make a comparison, ascertain the number of animals used to determine a mean. Each treatment contained six animals. 250
Table 3 E f f e c t s in Rats o f N i c k e l , I r o n , a n d T h e i r I n t e r a c t i o n o n L i v e r M a n g a n e s e L i v e r m a n g a n e s e , b p,g/g F e 3+ F e z+ , F e 3+ Previous Treatment~ Ni, i.Lg/g
Exp. 1
Exp. 3
Exp. 3
Exp. 2
study (1)
13.0 --
7.6 --
7.8 --
9.0 10.6
10.4 --
9.4 8.9
6.5 6.1
7.7 6.8
10.5 10.5
10.0 7.8
Fe, ~g/g
0 0
0 12.5
0 0
25 50
0
100
8.2
--
--
8.5
7.4
12.2 .
7.7
8.2 .
--
9.7
7.1
--
10.7
6.4
--
8.8
--
--
--
7.9
10.2 --
---
---
9.2 10.5
11.0 --
5 5
0 12.5
5
25
9.3
6.9
5
50
8.3
5.9
5
100
6.7
.
.
.
50 50
0 12.5
50 50
25 50
7.9 7.6
---
---
9.9 9.0
10.0 8.2
50
100
7.2
--
--
8.2
7.8
Analysis of Variance
P Values
Nickel effect
0.0001
NS
NS
Iron e f f e c t N i c k e l x iron Iron f o r m e f f e c t
0.0001 NS --
0.0001 NS 0.04
0.0001 NS --
0.05 0.0001 NS --
N i c k e l x iron f o r m
--
NS
--
--
Iron x iron f o r m
--
NS
--
--
N i c k e l x iron x iron f o r m
--
NS
--
--
E r r o r m e a n square
1.0
1.3
1.0
0.7
3.0 1.6
2.7 1.4
2.4 --
Scheff6 Valuesc 6 s-test 6 12 s-test 12
2.9 --
18 s-test 18
1.1
1.1
--
0.9
24 s-test 24
0.6
0.8
--
0.7
30 s-test 30 36 s-test 36
---
-0.5
0.6 --
---
aAmounts of Ni (nickel chloride) and Fe (either a mixture of 40% ferrous sulfate and 60% ferric sulfate, or ferric sulfate) supplemented to diet. t'Dry weight basis ~ Scheff6 test (11) is a method for performing multiple comparisons between group means. Means differing by more than the value given are significantly different (P < 0.05). To make a comparison, ascertain the number of animals used to determine a mean. Each treatment contained six animals. 251
Table 4 E f f e c t s in Rats o f N i c k e l , Iron, and T h e i r I n t e r a c t i o n o n L i v e r Z i n c L i v e r z i n c , b ixg/g F e 3+ F e z+, F e 3+ Previous Treatments Ni, txg/g
Exp. 1
Exp. 3
Exp. 3
Exp. 2
s t u d y (1)
F e , p,g/g
0
0
133
108
108
102
110
0
12.5
--
--
--
81
--
0
25
89
80
81
80
91
0
50
97
81
79
83
89
0 5
100 0
89 105
-100
-99
80 --
87 109
5
12.5
5 5
25 50
.
.
5
I00
83
--
--
50
0
93
--
--
93
113 --
91 91
--
.
.
78 81
78 82
. ----
95 88 93
50
12.5
--
--
76
50
25
83
--
--
71
89
50
50
80
--
--
78
94
50
100
79
--
--
72
91
Analysis of V a r i a n c e - - P Values Nickel effect
0.001
NS
0.004
NS
Iron effect N i c k e l x iron
0.0001 NS
0.0001 NS
0.0001 NS
0.0001 NS
Iron form effect
--
NS
--
--
N i c k e l • iron f o r m
--
NS
--
--
I r o n x iron f o r m
--
NS
--
--
N i c k e l x i r o n x iron f o r m
--
NS
--
--
Error mean square
212
84
77
101
Scheff6 V a l u e s c 6 s-test 6 12 s-test 12 18 s-test 18 24 s-test 24 30 s-test 30 36 s-test 36
43 -16 12 ---
25
24
29
13
13
--
9 7
---
11 8
5 --
---
-4
aAmounts o f Ni (nickel chloride) and Fe (either a mixture of 40% ferrous sulfate and 60% ferric sulfate, or ferric sulfate) supplemented to diet. ~ weight basis. ~The Scheff6 test (11) is a method for performing multiple comparisons between group means. Means differing by more than the value given are significantly different (P < 0.05). To make a comparison, ascertain the number of animals used to determine a mean. Each treatment contained six animals. 252
IRON EFFECT ON NICKEL DEPRIVATION
253
Discussion Findings from previous studies suggested that positive evidence of the interaction of nickel with iron would require examination of the parameters thatare affected by borderline iron deficiency (1-3). Data from examination of only the parameters that are affected by severe iron deficiency probably would not confirm the interaction between nickel and iron. Apparently nickel enhances the absorption of iron only when it is present in the diet in less than adequate but not severely inadequate, levels and in a relatively unavailable form. In severe iron deficiency, iron would not be available for interaction, and nickel probably would have little effect on the absorption of dietary iron. Thus, hematocrit and hemoglobin respond to nickelinduced changes in iron absorption, but total lipids and phospholpids in plasma and liver are changed by severe iron deficiency only and are not affected by an interaction between nickel and iron. Because borderline iron deficiency apparently elevated the copper and manganese contents, and depressed the iron content in liver, an interaction between nickel and iron might be expected to affect the copper, iron and, perhaps, manganese content in liver. Failure, in the present study, to find that nickel and iron interacted to affect iron and manganese, and copper consistently, might be explained as follows: In the previous study (1) and experiment 3, the interaction between nickel and iron affected the copper content in liver and the error mean square (Table 1) indicates that the copper data were less variable than those in experiment 1. In the previous study (1) and experiment 3, nickel deprivation mainly accentuated the elevation of the copper level in livers of severely iron-deficient rats. In experiments 1 and 2, the copper values from iron-supplemented rats were quite consistent, but this was probably masked by the extremely variable values from the severely irondeficient rats. Because of this variance, relatively large differences between means would have been required for statistical significance, and the relatively small effects of an interaction between nickel and iron, if present, were obscured. Thus, the data are not unequivocal, but suggest that an interaction between nickel and iron affects the copper content in liver of rats. Statistical demonstration of that interaction, however, would depend upon the experimental design (i. e., use a large number of animals), or the fortuitous consistent elevation of the copper content in livers of severely iron-deficient rats. Experiment 3 was an example of this fortuitous elevation of copper content. The liver copper should have been about equal in the two groups fed no supplemental iron or nickel or fed no supplemental iron and 5 txg of nickel/g of diet. Liver copper was not equal in those two groups. The highest values were usually found in groups designated for the ferric-ferrous mixture treatment. The copper values were consistently lower in the severely iron-deficient rats assigned to the ferric treatment than for rats assigned to the ferric-ferrous treatment. Further evidence that an interaction between nickel and iron affected the liver copper content was found in a previous study (4), in which nickel-deprived rats fed an apparently inadequate amount of iron, as ferric sulfate, exhibited, at age 55 days, elevated copper content in the liver. In rats fed an apparently adequate level of dietary iron as ferric chloride, nickel deprivation did not affect the copper content.
254
NIELSEN AND SHULER
55 days, elevated copper content in the liver. In rats fed an apparently adequate level of dietary iron as ferric chloride, nickel deprivation did not affect the copper content. In previous studies (1,2), when the iron supplement was ferric sulfate only, the interaction between nickel and iron affected hematocrit and hemoglobin. These parameters increased and reached a plateau as dietary iron increased. At the low levels of iron supplementaion, nickel deprivation depressed hematocrit and hemoglobin levels. At high levels of iron supplementaion, where these parameters had reached a plateau, nickel deprivation had no effect. Thus, an interaction between nickel and iron was shown because nickel affected hematocrit and hemoglobin differently at different levels of iron supplementation. In the present study, the iron and manganese contents of the liver reacted differently to increasing dietary iron. As the dietary iron increased, generally liver iron content increased and manganese content decreased--no plateau was reached. There was no point at which the effect of dietary nickel was precluded, which probably explains the absence of interaction between nickel and iron. Although the interaction between nickel and iron was not statistically significant, the form of dietary iron did alter the effect of dietary nickel on the iron content of the liver. When only ferric sulfate was supplemented to the diet, liver iron was significantly lower in nickel-deprived than in nickel-supplemented rats in experiment 3; the difference approached significance (P < 0.06) in experiment 2. In contrast, when the dietary iron was a ferric-ferrous sulfate mixture, the liver content of iron was higher in nickel-deprived than in nickel-supplemented rats; the difference was significant (P < 0.01) in experiment 1. Similar findings were published (4). In those studies, at age 35 and 55 days, nickel-deprived rats fed 60 p~g iron/g diet as ferric sulfate exhibited a depressed concentration of iron in the liver. At age 30 days, the iron concentration in livers from rats fed 30 p~g iron/g diet as ferric chloride was higher in nickel-deprived than in nickel-supplemented rats. Also, at age 55 days, the iron concentration tended to be higher in the nickeldeprived than in nickel-supplemented rats. Apparently the form and level of iron fed to rats influenced the direction of the change in liver iron concentration that was associated with nickel deprivation. The changes in manganese content in liver caused by various treatments were small in the present study and probably of little physiological meaning. Perhaps any effect on the manganese content of the liver was indirectly caused by changes in liver iron content induced by dietary nickel and iron. Apparently the iron content of the liver has a major influence on its manganese content. The findings in the present study and in previous studies (1, 4) suggest that as iron increases in the liver, manganese decreases. Schnegg and Kirchgessner (12-14) found that nickel deprivation depressed the total copper, iron, and zinc contents in liver of rats. The initial impression is that their findings contrast with those of our present and previous (4) studies, which showed that nickel deprivation apparently elevated the iron (ferric-ferrous mixture fed), copper and zinc concentration, on a per gram basis, in livers of rats, espe-
IRON EFFECT ON NICKEL DEPRIVATION
255
cially if they were iron-deficient. However, if the data of Schnegg and Kirchgessner (12) is recalculated to give the ~g copper, iron, or zinc/g liver, much of the divergence in findings is eliminated. When their data is expressed on the per gram basis, the liver content of copper and zinc was as high, or higher, in nickeldeprived as in nickel-supplemented rats. Still, the liver content of iron on the per gram basis was generally lower in nickel-deprived than nickel-supplemented rats. In our studies, this was observed in nickel-deprived rats fed ferric sulfate only. We did not determine the total weight of the dry liver in these experiments because a portion of the fresh liver was used for lipid analysis. Thus, the total copper, iron, and zinc contents of the liver could not be accurately determined. However, we made a rough estimate of total contents (data not presented) by using the total fresh weight, and the dry weight concentrations of copper, iron, and zinc, of liver. These calculations did not change any of our major conclusions stated in this discussion. On the whole liver, in contrast to the per gram liver, basis, we found that: (1) manganese was not elevated in rats fed no supplemental iron, (2) manganese was not elevated in nickel-deprived rats in experiment 1, and (3) the elevation in the zinc content was less marked in severely iron-deficient rats. Thus, on the whole liver basis, our findings apparently contrast with those of Schnegg and Kirchgessner (12-14). Perhaps part of this divergence can be explained by the difference in experimental design. In other words, factors such as severity of nickel deprivation, form or level of dietary iron, or age of animals influenced the direction and magnitude of the change in total copper, iron, and zinc contents in the livers of nickeldeprived rats. This might be expected because these factors were shown to influence signs of nickel deprivation (3, 4, 15). The findings in the present and previous studies (1-3) indicate that the parameters that are affected by an interaction between nickel and iron are limited in factorially arranged experiments. These parameters reach a plateau when dietary iron is adequate, and vary directly with dietary iron when it is inadequate. Parameters that did not reach a plateau with the iron spplements used, or were affected by severe iron deficiency only, are not suitable to ascertain a nickel-iron interaction. Furthermore, the findings show that the form of dietary iron markedly influences the interaction between nickel and iron, and the effect of nickel deprivation, in the rat.
Acknowledgment The authors thank Barry Shull and LuAnn Johnson for their help with the statistical analysis of the data.
References 1. F. H. Nielsen, T. R. Shuler, T. J. Zimmerman, M. E. Collings and E. O. Uthus, Biol. Trace Element Res. 1, 325 (1979). 2. F. H. Nielsen, J. Nutr. 110, 965 (1980).
256 3. 4. 5. 6.
7. 8. 9. 10. 11. 12. 13. 14.
15.
NIELSEN AND SHULER
F. H. Nielsen, Biol. Trace Element Res. 2, 199 (1980). F. H. Nielsen and T. R. Shuler, Biol. Trace Element Res. 1, 337 (1979). F. H. Nielsen and B. Bailey, Lab. Anim. Sci. 29, 502 (1979). E. Pietsch, G. Blinoff-Achapkin, H. Gruss, A. Kotowski, M. Du Made, and G. Nachod, Gmelins Handbuch Der Anorganischen Chemie, Verlag Chemie G.m.b.H., Berlin, 1957, pp. 395,440. W. E. Harris and B. Kratochvil, Chemical Separation and Measurements, Saunders, Philadelphia, 1974, pp. 124-132. F. H. Nielsen, D. R. Myron, S. H. Givand, and D. A. Ollerich, J. Nutr. 105, 1607 (1975). F. H. Nielsen, D. R. Myron, S. H. Givand, T. J. Zimmerman, and D. A. Ollerich, J. Nutr. 105, 1620 (1975). Analytical Methods for Atomic Absorption Spectrophotometry, Perkin-Elmer Corp., Norfolk, CT (1976). H. Scheff6, The Analysis of Variance, Wiley, New York, pp. 90-137 (1959). A. Schnegg and M. Kirchgessner, Arch. Tierern?ihrung 26, 543 (1976). A. Schnegg and M. Kirchgessner, Internat. Z. Vit. Ern. Forschung 46, 96 (1976). A. Schnegg and M. Kirchgessner, in Trace Element Metabolism in Man and Animals-3 (M. Kirchgessner, ed.), Technical University of Munich, FreisingWeihenstephen, 1978, pp.236-243. F. H. Nielsen, T. J. Zirnmerman, M. E. Collings, and D. R. Myron, J. Nutr. 109, 1623 (1979).
