Copper Deficient Swinedm5migu4zj3pb.cloudfront.net/manuscripts/105000/105891/...Copper-deficient swine failed to absorb dietary iron at the normal rate. Increased amounts of stainable

Iron Metabolism in

Copper - Deficient Swine

G. RiCAiRD LEE, SERGIO NACHT, JoHN N. LUKENS, and G. E. CARTwmGwr

From the Department of Medicine, University of Utah College of Medicine,Salt Lake City, Utah 84112

ABSTRACT The way in which iron is handledby the duodenal mucosa, the reticuloendothelialsystem, the hepatic parenchymal cell, and the nor-moblast was investigated in copper-deficient swine.

Copper-deficient swine failed to absorb dietaryiron at the normal rate. Increased amounts ofstainable iron were observed in fixed sections ofduodenum from such animals. When 59iron wasadministered orally, the mucosa of copper-deficientanimals extracted iron from the duodenal lumen atthe normal rate, but the subsequent transfer toplasma was impaired.

When intramuscular iron supplements weregiven to copper-deficient pigs, increased amountsof iron were found in the reticuloendothelial sys-tem, the hepatic parenchymal cells, and in normo-blasts (sideroblasts). Hypoferremia was observedin the early stages of copper deficiency, even thoughiron stores were normal or increased. When redcells that were damaged by prolonged storage wereadministered, the reticuloendothelial system failedto extract and transfer the erythrocyte iron to theplasma at the normal rate. Administration of cop-per to copper-deficient animals with normal ironstores resulted in a prompt increase in the plasmairon.

The observed abnormalities in iron metabolismare best explained by an impaired ability of theduodenal mucosa, the reticuloendothelial system,and the hepatic parenchymal cell to release iron tothe plasma. It is suggested that copper is essentialto the normal release of iron from these tissues.

Dr. G. Richard Lee is a Markle Scholar in AcademicMedicine.

Received for publication 7 March 1968 and in revisedform 8 May 1968.

This concept is compatible with the suggestionmade by others that the transfer of iron from tis-sues to plasma requires the enzymatic oxidation offerrous iron, and that the plasma copper protein,ceruloplasmin, is the enzyme (ferroxidase) whichcatalyzes the reaction.

Because excessive amounts of iron were foundin normoblasts, it is suggested that an additionaldefect in iron metabolism affects these cells andplays a major role in the development of anemia.As a result of the proposed defect, iron cannot beincorporated into hemoglobin and, instead, ac-cumulates as nonhemoglobin iron.

INTRODUCTION

The severe anemia associated with copper defi-ciency in swine has been characterized as hypo-chromic and microcytic (1, 2). Since anemia ofthis type usually is due to a defect in hemoglo-bin synthesis, a role for copper in the biosynthesisof hemoglobin has been proposed. However,neither the nature of the defect nor the role forcopper has been elucidated.

Defects in hemoglobin synthesis may be classi-fied into three categories: (1) abnormalities in thebiosynthesis of protoporphyrin and heme; (2)abnormalities in the metabolism of iron; and (3)abnormalities in the biosynthesis of globin. In aninvestigation recently reported from this labora-tory (3), the pathways of porphyrin and hemesynthesis in erythrocytes from copper-deficientswine were found to be intact.

Impaired absorption of iron from the gastroin-testinal tract has been observed previously in cop-per-deficient swine along with hypoferremia and

2058 The Journal of Clinical Investigation Volume 47 1968

an increase in the total iron-binding capacity (4).Because of the impaired absorption of iron, suchanimals were actually deficient in iron even thoughlarge amounts of iron were supplied in the diet.However, the parenteral administration of ironfailed either to prevent or to alleviate the anemia.Thus, failure to absorb iron did not appear to bethe only factor involved in the pathogenesis of theanemia.

The purposes of the present investigation were:(1) to gain a clearer understanding of the defectin iron absorption; (2) to search for abnormali-ties in the way iron is handled by other tissues, es-pecially the reticuloendothelial system, the hepaticparenchymal cell, and the normoblast; and (3) todetermine if the defects in iron metabolism con-stituted an adequate explanation for the anemia.Because of the tissue deficiency of iron that occursin copper-deficient swine fed iron by the oralroute, it is not possible to evaluate iron metabolismin tissues other than the duodenal mucosa. Forthis reason, studies were made in pigs given ironintramuscularly in an amount calculated to main-tain the total body iron within normal limits. Theresults are compared with those obtained fromstudies in copper-deficient pigs given iron bymouth, in iron deficient pigs, and in controls.

METHODSThe 82 pigs used in this study were obtained at 3-5 daysof age. They were of mixed breed and were purchasedfrom a single breeder in litters of 6-10 animals.

The experimental diet consisted of canned condensedmilk with no supplements other than those described be-low. Details of diet preparation have been reported previ-ously (1). On the basis of the nature and route of theiron and copper supplements they received, the pigs weredivided into five experimental groups: (1) control (oraliron), (2) control (i.m. iron), (3) iron deficient, (4)copper deficient (oral iron), and (5) copper deficient(i.m. iron). Groups 1, 2, and 3 received 0.5 mg of copperas copper sulfate per kg of body wt per day with thediet. Control (oral iron) and copper-deficient (oraliron) swine were given 30 mg of iron as ferrous chlo-ride per kg of body wt per day with the diet. Control(i.m. iron) and copper-deficient (i.m. iron) pigs re-ceived a total of 1.2 or 2.0 g of iron in multiple intra-muscular injections, between 10 and 28 days of age. The

intramuscular iron preparation used was Pigdex 100(colloidal iron oxide stabilized with a low viscosity dex-trin).1 No iron supplements were given to the iron-deficient group. In some experiments, no difference was

1 American Cyanamid Co., Princeton, N. J.

apparent between the two control groups, and they werecombined for statistical purposes.

The routine hematologic methods used in this studyand the methods for determining plasma iron, plasmairon-binding capacity, and bone marrow iron have beenpublished (5). Fixed tissue sections were stained foriron by the method of Gomori (6). Tissue iron concen-tration was determined after wet digestions with sul-furic, nitric, and perchloric acids (3), and the deter-mined value was corrected for tissue hemoglobin con-centration by the method of Andersen and Shoemaker(7).

RESULTS

Effect of intramuscular iron supplements onthe anemia of copper deficiency

The volume of packed red cells and the corpus-cular constants. Values for the volume of packedred cells (VPRC) and the corpuscular constantsduring the development of copper deficiency aredepicted in Fig. 1. Of the 33 copper-deficient pigs,20 were given iron supplements by mouth and 13were given iron intramuscularly. Both types ofiron supplementation were used in the controlgroup, which consisted of 17 pigs.

In the copper-deficient (oral iron) group, a de-crease in the VPRCbecame evident by 6 wk ofage. The VPRCcontinued to decrease in a rapidand linear fashion until the pigs were 10-11 wkof age, at which point some deaths began to occuramong the most anemic animals. These deathsaccount for the apparent tendency of the VPRCto become stable at about 20 ml/100 ml. Microcy-tosis, as measured by the mean corpuscular vol-ume (MCV), was an early development and itsseverity paralleled that of the anemia. Significanthypochromia, as measured by the mean corpus-cular hemoglobin concentration (MCHC), was alater development, appearing at 10-11 wk of age.

In the copper-deficient (i.m. iron) group, ini-tial values for the VPRCwere greater than in thecontrol group, probably because of the residual ef-fects of neonatal iron deficiency in some of thecontrol pigs. Thereafter, anemia progressed at arate similar to that observed in copper-deficient(oral iron) pigs, and by 13 wk of age the anemiawas equally severe in both deficient groups. Theintramuscular administration of iron lessened indegree, but did not prevent the development ofhypochromia and microcytosis; values for thecorpuscular constants in the group receiving in-

Iron Metabolism in Copper-Deficient Swine 2059

45-

'I. 40

N 30

25

Q 50.

45.31

R 30

0 2q

Q 28

27

19

17~

15

3~

- --.0 Conitrol- -.o%* __ *vCcL Del (ti.TZe)

- Ca ef/rN D(opal Ye) G 0

5 6 7 8 9 10 11 12 13 14Age (weeks)

FIGURE 1 Changes in volume of packed red cells(VPRC), mean corpuscular volume (MCV), mean cor-puscular hemoglobin concentration (MCHC), and meancorpuscular hemoglobin (MCH) with development ofcopper deficiency. Hypochromic, microcytic anemia de-veloped in copper-deficient pigs, regardless of the routeby which iron was administered.

tramuscular iron supplements were intermediatebetween those of the control group and of thegroup receiving iron with the diet. At 13 wk ofage, all three groups differed significantly fromone another (P < 0.001) with respect to theMCV, the MCHC, and the mean corpuscular he-moglobin.

Plasma Iron. The plasma iron concentrationand the total plasma iron-binding capacity weredetermined weekly in 12 control, 14 copper-defi-cient (oral iron), and 14 copper-deficient (i.m.iron) pigs (Fig. 2). In copper-deficient (oral iron)pigs, the plasma iron level decreased progressivelyuntil the pigs were about 12 wk of age; thereafter,it tended to increase. In copper-deficient (i.m.

iron) pigs, hypoferremia was observed between theages of 8 and 11 wk. Thereafter, the plasma ironvalues returned to normal. Thus, in both cop-per-deficient groups, three phases could be definedwith respect to the changes in the plasma iron:an initial phase in which the plasma iron was nor-mal; a hypoferremic phase; and a terminal phaseduring which the plasma iron increased. Thesephases differed in degree and duration in the twogroups. Additional data concerning the hypofer-remic phase in the copper-deficient (i.m. iron)group and the terminal phase in both groups willbe presented in later sections.

The total plasma iron-binding capacity was in-creased throughout the entire course in the copper-deficient (oral iron) group. Values observed inthe copper-deficient (i.m. iron) group did not dif-fer significantly from those in the control group.

600

C)400

'i C.300

_-.. 160

11400

0120

100

80,q)a 60

40i 0to20

I-

I-

ca Def (oea IFe) i /

(i. m. Fe) e)

controI% -

>0~~~~

_, I.s _me

cCZtDef (OP-1 e)a

0 'I~~~~at"!

5 6 7 8 9 10 122S 1JS 4Age (weeAk)

FIGURE 2 Changes in plasma iron (Fe) and total iron-binding capacity (TIBC) with development of copperdeficiency. Hypoferremia and an increase in TIBC wereobserved in copper-deficient (oral iron) pigs. Transienthypoferremia with no change in TIBC was found incopper-deficient (I.M. iron) pigs.

2060 G. R. Lee, S. Nacht, J. N. Lukens, and G. E. Cartwright

TABLE I

Tissue Iron in Copper-Deficient and Control Swine

Copper CopperControls Controls Iron deficient deficient

(oral iron) (i.m. iron) deficient (oral iron) (i.m. iron)

Nos . 6 9 5 14 10Concentration

Liver iron ,ug/g 149 i 8 1738 i 293 18 i 1.5 61 4t 13 1478 + 182

Totalmg/organ 54 i 17 622 +t 65 2.1 :1 0.2 24 i 5.7 632 4- 85

ConcentrationSpleen iron ,g/g 88 + 18 140 i 21 104 + 6 187 + 13.6 676 i 108

Totalmg/organ 2.0 + 0.2 5.5 i 0.7 0.9 t 0.1 6.3 i 0.9 20 + 2.3

Red cell iron, mg, (Total) 661 i: 63 789 ih 59 69 + 7 239 i 15 300 i 30"Total" body iron, mg 715 i 71 1415 i 62 72 i 7 266 + 18 949 ± 105Body weight, kg 16.4 + 2.0 20.3 + 1.6 4.4 i 0.5 14.8 i 0.7 16.8 i 1.2"Total" body iron concentration, 45 -t 3 73 i 7 17 i 1.3 18 + 0.9 59 + 7

mg/kg of body wt)

Values are means 4t SEM.

Duodenal mucosa

Iron absorption. An estimate of total body ironcontent was made in pigs from each of the experi-mental groups. In the animals that received oraliron supplements, this estimate was used as ameasure of the amount of iron absorbed duringthe observation period. Estimated "total" bodyiron was calculated by summing the amounts ofiron in the liver, the spleen, and the circulatingerythrocytes. The assumption was made that arelatively small proportion of the total body ironwould be found outside of these three sites. Theconcentration of iron in the liver and spleen wasdetermined directly on tissues obtained when theanimals were killed (at 12-17 wk of age, exceptin the case of the iron-deficient pigs, who becameanemic more quickly and were, therefore, killedat 6-8 wk of age). The total red cell iron was cal-culated from the blood hemoglobin concentration,the estimated blood volume (8), and the knowniron content of hemoglobin.

The copper-deficient (oral iron) pigs werefound to be severely iron deficient. Their "total"body iron concentration, liver iron content, anderythrocyte iron content were considerably lowerthan in control (oral iron) pigs; in fact, the "to-tal" body iron concentration was of the same or-der of magnitude as the values found in the iron-

deficient group (Table I). When iron supplementswere given intramuscularly, copper-deficient pigsdid not become iron deficient and their "total"body iron concentration was comparable with thatof the control group.

Iron in fixed sections of duodenum. Sectionsof proximal duodenum were obtained at autopsyand stained for iron. In specimens from pigs thathad received iron with the diet, Prussian blue pos-itive material was seen in two locations: in thelamina propria as large granules within macro-phages, and within the columnar epithelial cells asa line of fine granules (Fig. 3). No stainable gran-ules were seen in mucosal specimens from animalsgiven iron supplements exclusively by the intra-muscular route; thus, the visible iron appeared tobe of dietary origin, rather than part of a poolin equilibrium with other body stores.

The amount of iron found in the mucosa wasconsiderably greater in copper-deficient (oral iron)pigs than in controls. The amount of iron in eachspecimen was estimated semiquantitatively on ascale from 0 to 4 +. On this scale, mucosa fromcontrol (oral iron) pigs contained a trace to 1 +stainable iron, whereas specimens from copper-deficient (oral iron) pigs were scored as 3 to 4 +.This observation suggested that the barrier tothe absorption of iron in the copper-deficient ani-

Iron Metabolism in Copper-Deficient Swine 2061

t Blip OK id S'pah

4ni.

P-7~~~~~-4

itM.- 4Pv~~~~~~~~~~~~~~t.W4~~~~~~~-

FIGURE 3 Iron stains of mucosa from a copper-deficient (oral iron) pig (A) and a control (oral iron) pig (B). Inthe copper-deficient animal, large amounts of iron were seen in the lamina propria and within epithelial cells. Sincethe iron is seen more easily in a colored reproduction, the arrows are inserted to indicate the location of epithelialiron.

mals did not lie between the intestinal lumen andthe duodenal mucosa, but at a subsequent site inthe absorption pathway.

Changes in mucosal iron after a test meal. Theabove hypothesis was studied further by measur-ing mucosal radioactivity after the administrationof test meals containing -"iron. These experimentswere performed when the pigs were between 90and 110 days old, at which time the copper-defi-cient pigs were severely anemic (VPRC less than25 ml/100 ml). After a 48 hr fast each pig wasfed, by stomach tube, a test meal consisting of 1mg of iron as ferrous chloride labeled with 5-10,uc of 59Fe. At varying time intervals thereafter,the pigs were killed by exsanguination. The proxi-mal duodenum was excised, opened, and washedwith 0.01 M EDTA in saline. Mucosa was scrapedfree from underlying layers with a glass slide (9),

suspended and washed in EDTA-saline, and cen-trifuged at 1120 g. The volume of packed mucosawas measured in a graduated centrifuge tube.Radioactivity was assayed by crystal scintillationcounting, and mucosal iron content was calculatedfrom the specific activity of the iron in the testmeal.

In control animals, mucosal radioactivity in-creased to a maximum value 2-3 hr after the testmeal, and then decreased rapidly, so that littleradioactivity remained in the mucosa 8 and 24 hrafter the meal (Fig. 4). In iron-deficient pigs, thepattern was similar to that seen in controls, exceptthat the maximum value was reached somewhatmore quickly, and was of lesser magnitude.

The pattern observed in copper-deficient ani-mals was quite different. The magnitude of theinitial peak was greater, and the subsequent de-

2062 G. R. Lee, S. Nacht, J. N. Lukens, and G. E. Cartwright

crease was considerably less, leaving much largeramounts of test meal iron in the mucosa after 8and 24 hr. Mucosal radioactivity at 24 hr wasabout half as great in copper-deficient (i.m. iron)pigs as it was in copper-deficient (oral iron) pigs.

Subcellular distribution of mucosal radioiron.To define more clearly the nature of the iron re-tained by the mucosa, a crude subcellular fraction-ation was performed. Mucosa was homogenizedwith high-frequency sound.2 A "sedimentable"fraction was recovered by centrifuging the ho-mogenate at 9000 g for 30 min. The supernatantsolution was heated at 750C for 5 min, and theprecipitate was harvested and designated "heat-labile proteins." Perchloric acid (final concentra-tion, 2%) or ammonium sulfate (60%o saturation)was added to the supernatant solution, and theprecipitate so obtained was designated "ferritin."All radioactivity in this fraction was excluded bySephadex G-200 gel and could be sedimented at90,000 g in a preparative ultracentrifuge. The su-pernatant solution which remained after "ferritin"precipitation was termed the "nonprotein" frac-tion. Radioactivity in each of the four fractionswas assayed.

The distribution of radioactivity among thesemucosal fractions was determined in one animalfrom each group at 2 hr after the test meal, and in

2 Sonifer, Branson Instruments, Inc., Stamford, Conn.

10o 1I 14 16 1a 20 22 24.27t.e (hocafw)

FIGURE 4 Mucosal iron after a test meal containing59iron. Note the excessive retention of iron at 22 hr inboth copper-deficient groups.

one from each group at 24 hr (Table II). The"ferritin" fraction contained most of the radioac-tivity, both at 2 and at 24 hr, in all pigs exceptthose in the iron-deficient group. In these, mostof the radioactivity was in the nonprotein frac-tion at 2 hr and in the sedimentable fraction at 24hr.

In copper-deficient pigs, an increased amount ofmucosal radioiron was found in all four subcellularfractions at 24 hr; most notably in the ferritinfraction.

TABLE I ISubcellular Distribution of Mucosal Iron* After a Test AMeal

Subcellular fraction

Group Time after test meal Sedimentable Heat-labile Ferritin NonProtein

hrControl 2 18.3 5.5 - 81.9 4.2

24 4.8 1.7 6.2 1.6

Copper deficient 2 26 7.9 143.0 3.6oral iron

24 39.8 15.2 97.4 6.3

Copper deficient 2 5.9 3.2 33.1 10.8i.m. iron

24 19.0 6.9 42.7 5.3

Iron deficient 2 12.3 4.6 7.7 27.7

24 3.0 0.8 1.6 1.3

* In mumoles of Fe/ml of mucosa.

Iron Metabolism in Copper-Deficient Swine 2063

The reticuloendothelial system

Splenic and hepatic iron. When iron deficiencydevelops in an otherwise normal animal, iron inthe liver and spleen is depleted before anemia de-velops. In copper-deficient (oral iron) animals, onthe other hand, a relatively large amount of ironwas retained by these organs (Table I). The cop-per-deficient (oral iron) group and the iron-defi-cient group were equally iron deficient as mea-sured by "total" body iron concentration. Nev-ertheless, the liver iron concentration in copper-deficient (oral iron) pigs was more than threetimes as great as in iron-deficient pigs. Thesplenic iron concentration of copper-deficient(oral iron) pigs was also significantly greaterthan that of iron-deficient pigs. In fact, copper-deficient pigs retained more iron in their spleensthan did controls. Still greater iron retention bythe spleen occurred in the copper-deficient groupthat received intramuscular iron supplements.These findings suggested that in copper-deficientpigs, the reticuloendothelial system, like the du-odenal mucosa, retains iron.

Hypoferremia with normal iron stores.Changes in the plasma iron with time in a typical

6 8 10 X2 14 16A?e (weels)

FIGURE 5 Changes in plasma iron in a single copper-deficient (I.M. iron) pig with the development of copperdeficiency. A 2 wk period of hypoferremia was observedfollowed by an increase in plasma iron to hyperferremiclevels.

.1 X 3 4Mimbe (Aouas)

FIGURE 6 The effect of an infusion of damaged red cellson plasma iron. Values for the mean + SEM are depictedat each point. The infusion induced relatively little changein the copper-deficient pigs.

copper-deficient (i.m. iron) animal are depictedin Fig. 5. A 2 wk period of hypoferremia was ob-served, followed by a rise to above normal levels.This pattern was also observed in other animals inthis group, but the age at which the hypoferremiaoccurred varied. Thus, the calculation of meanvalues at a single age, as was done in constructingFig. 2, tended to obscure the finding. In Fig. 2,hypoferremia was clearly evident only at week 10.

To document this phenomenon more clearly,the lowest plasma iron out of a series of weeklydeterminations over a 3 month period was re-corded for each pig. The mean "lowest" plasmairon value in 11 control pigs was 117 ± 7.3 jug/100 ml (mean + SEM). The mean "lowest" plasmairon value in 11 copper-deficient (i.m. iron) pigswas 62 + 7.3 ug/100 ml. The values are signifi-cantly different from one another (P < 0.01).

It should be noted that the hypoferremia incopper-deficient (i.m. iron) pigs was not a con-sequence of iron deficiency, since the total bodyiron concentration was normal and the iron stor-age sites were laden with iron in such animals(Table I). Thus, the ability to transfer iron fromreticuloendothelial storage sites to plasma seemedto be impaired in copper-deficient animals.

Effect of an infusion of damaged erythrocyteson plasma iron. As another measure of the ca-pacity of the reticuloendothelial system to processand release iron, the changes in plasma iron thatfollowed the administration of damaged red cells

2064 G. R. Lee, S. Nacht, J. N. Lukens, and G. E. Cartwright

were observed. Pig erythrocytes were damagedby storage at 4VC for a minimum of 3 wk, underaseptic conditions, in plasma containing heparin.An intravenous infusion that contained 1 mg ofiron as damaged erythrocytes per kg of body wtwas administered to each of five control and eightcopper-deficient pigs. The plasma iron was mea-sured before the infusion and at intervals of 1, 2, 3,and 4.5 hr thereafter. In control animals, theplasma iron concentration increased markedly andreached a peak value at about 3 hr. In copper-de-ficient animals, on the other hand, the curves wererelatively flat (Fig. 6).

Plasma iron after administration of copper. Toobtain further evidence for an effect of copper onthe release of iron from reticuloendothelial cells,the changes in plasma iron that followed the ad-ministration of copper to copper-deficient pigswere studied. In animals with anemia due to spe-cific nutritional deficiencies, the plasma iron usu-ally decreases when the deficient nutrient is sup-plied. By contrast, when copper (0.1 mg/kg of

6 i

O-t ~ ~ ~

:,.

4

Of1ti IraAn

..a,..s I v * toot- W~~~~~l

'N 500

'40

0 300o

A 200P1

too-

0 1 2 3 4 s 6Xoutr.g after .iecfxbn

FIGURE 7 The effect of the administration of copper on

the plasma iron in a copper-deficient pig.

body wt) was administered intravenously to cop-per-deficient (i.m. iron) pigs, the plasma ironconcentration promptly increased and continued to*os ''.4

t. ijot- in \ ' -2'S e "lo . .S; ' RevAm

FIGURE 8 Iron stains of liver sections from a copper-deficient (I.M. iron) pig (A) and a control pig (B).Excessive parenchymal cell iron was observed in the copper-deficient animals.

Iron Metabolism in Copper-Deficient Swine

I-

41-

2065

increase until the plasma iron-binding capacitywas saturated (Fig. 7). Thus, copper administra-tion appeared to remove an obstruction to the flowof iron from reticuloendothelial cells to plasma.

The hepatic parenchymal cell

The distribution of iron in hepatic cells was as-sessed in sections of liver stained by the Prussianblue technique. Sections of liver from five controland five copper-deficient animals were evaluated.Animals in both groups had received intramuscu-lar iron supplements, and the concentration ofliver iron was not significantly different in the twogroups. In control animals, almost all stainableiron was confined to the Kupffer cells. A trace ofhepatic parenchymal cell iron was seen in onlyone of the control animals. In contrast, all of thelivers from copper-deficient animals containedparenchymal cell iron in addition to Kupffer celliron. In three of these liver specimens, Kupfferand parenchymal cell iron was about equal, whereasin two, parenchymal cell iron exceeded that in theKupffer cells. Representative photomicrographsfrom one pig in each group are shown in Fig. 8.

Normoblast iron

Toward the end of the observation period,plasma iron values in copper-deficient (i.m. iron)pigs increased to normal or above normal valueswithout a coincident increase in the VPRC(Figs.2 and 5). This phenomenon was not restricted topigs receiving iron intramuscularly, but was alsofound in certain of the copper-deficient (oral iron)

TABLE IIISideroblasts in Bone Marrow of Copper-

Deficient and Control Swine

Group No. Sideroblasts

(%)Controls

oral iron 6 35.8 : 6.9

Controlsi.m. iron 6 36 i 6.9

Copper Deficientoral iron 9 14.8 i 6.5

Copper Deficienti.m. iron 8 76.6 i 8.3

Values are means :1: SEM.

pigs, despite the fact that they were deficient iniron as well as copper. The plasma iron values atthe end of the 12-17 wk observation period in 20copper-deficient (oral iron) pigs appeared to fallinto two, widely separated groups. In 13 of thepigs, the plasma iron value was "low," rangingfrom 13 to 53 jug/100 ml. In the remaining sevenanimals, the plasma iron value was "high," rang-ing from 132 to 247 jug/100 ml. The reticulocytecount in the "low iron" group was 7.9 ± 1.0%(mean + SEM), whereas in the "high iron" group,

the reticulocyte count was 1.9 ± 1.0%o. It appeared,therefore, that the terminal increase in serum ironmight be related to decreased erythropoiesis, andthat factors other than the supply of iron limitedthe response of the marrow.

To determine if such was the case, sideroblastcounts were performed on bone marrow (TableIII). A marked increase in the proportion ofsideroblasts was observed in the bone marrowsfrom copper-deficient (i.m. iron) pigs. Not onlywas the proportion of sideroblasts increased, butthe amount of stainable iron in each sideroblastalso appeared greater than in control animals. Insideroblasts from control animals, only one to threesmall granules were seen, whereas in the copper-deficient (i.m. iron) pigs, over half of the sidero-blasts contained more than six granules. More-over, individual iron granules appeared larger.No tendency was observed for the granules toform rings about the nucleus.

DISCUSSION

In these studies, as in others reported previously(1, 2), severe hypochromic, microcytic anemiadeveloped in copper-deficient swine. The intra-muscular administration of iron did not affect therate at which anemia developed, nor did it preventhypochromia and microcytosis.

Abnormalities of iron metabolism were observedin the duodenal mucosa, the reticuloendothelialsystem, the hepatic parenchymal cells, and in thenormoblasts of copper-deficient swine. The firstthree of these tissues have in common the abilityto do three things with iron. First, each can extractiron from some source; second, each can store ironto a variable extent; and third, each can releaseiron to the plasma. The abnormalities in iron me-tabolism that were observed in copper-deficient

2066 G. R. Lee, S. Nacht, J. N. Lukens, and G. E. Cartwright

swine can best be explained by impaired ability ofthese three tissues to release iron.

Duodenal mucosa. On the basis of studies per-formed in vitro with rat duodenum, Manis andSchachter have suggested that iron absorptionoccurs in two distinct, energy-dependent steps(10). In step one, iron is extracted from the in-

testinal lumen by mucosa; in step two, mucosaliron is released into the blood stream. Our findingswith radioactive test meals (Fig. 4) may be in-terpreted in terms of this hypothesis. The amountof dietary iron in the mucosa at any one time rep-resents the difference between the iron extracted instep one and that released in step two. Thus, incontrol animals, the rate of transfer in step oneexceeded that in step two for about 2-3 hr. andmucosal radioiron accumulated. Thereafter thetransfer rate in step two exceeded step one, andthe mucosal iron decreased. In copper-deficientpigs, the uptake of iron from duodenal lumen bymucosa appeared to be normal or increased. Onthe other hand, most of the iron accepted by themucosa remained there over the 24 hr observationperiod. Thus, iron taken up by mucosa was notsubsequently released at the normal rate; step twowas impaired.

The rate of iron transfer in step one appeared tobe related to the status of the iron stores, sincemucosal uptake was more rapid in iron-deficientthan in control animals. This hypothesis may alsoaccount for the difference between the two copper-deficient groups. In copper-deficient (oral iron)pigs, step one was increased as a consequence ofdepleted iron stores; thus, a higher final plateauwas reached (Fig. 4).

The expected end result of impaired mucosaliron release (step two) would be excessive mu-cosal iron storage. Excessive storage was, in fact,demonstrated (Fig. 3 and Table IT), a particu-larly significant finding in the presence of a totalbody deficiency of iron.

Most of the iron retained by copper-deficientmucosa was probably in the form of ferritin (TableIII). The fractionation method used allowed therecovery of a compound which was soluble, pre-cipitable by reagents that precipitate protein, andresistant to heat denaturation. Studies with Sepha-dex gel exclusion and ultracentrifugation indicatedthat the compound wras of high molecular weight.These properties, while not entirely specific for

ferritin, are not characteristic of any other knownbiological iron compounds. The accumulation offerritin in copper-deficient mucosa does not neces-sarily imply that copper is essential to ferritiniron release. Ferritin may not be a direct inter-mediate in the absorptive pathway, but simply acompound formed to protect the cell from damageby ionic iron during periods in which step oneexceeds step two. It is interesting that in iron-deficient pigs, at a time when both step one andstep two were increased (2 hr), most of the ironwas present in the nonprotein fraction of mucosalhomogenates (Table II). It seems probable thatthe active transport form of iron across the cellwall is a nonprotein compound, possibly ioniciron itself.

Reticitloendothelial tissues. Reticuloendothelial(R-E) cells acquire iron by phagocytosis of senes-cent erythrocytes. The R-E system maintains theprincipal storage pool of iron, and constitutes themajor source of supply of iron to the plasma. Un-der most circumstances, the plasma iron concen-tration reflects the balance between the rate atwhich iron is released from R-E cells and the rateat which iron is removed by tissues, especially theerythroid cells of the bone marrow. In copper-deficient pigs, the rate at which iron is removedfrom the plasma (the plasma iron turnover rate)is normal or nearly normal (11). Therefore, thehypoferremia with normal iron stores found incopper-deficient (i.m. iron) pigs implies that re-lease of iron from R-E cells is defective.

Evidence for this hypothesis was obtained byadministering a "load" of iron in the form of dam-aged red cells. This test was used by Noyes, Both-well, and Finch to evaluate "R-E iron block" inthe anemia of infection (12). These workers re-ported that in normal human subjects the plasmairon level increased in a predictable manner aftersuch an infusion, apparently because the iron fromthe infused cells was extracted by R-E tissues andpromptly released. In the present studies, thechanges in serum iron after infusion of damagedred cells into normal pigs resembled those whichwere observed in normal human subjects. On theother hand, little change in the plasma iron oc-curred in copper-deficient pigs after the infusion.This observation is consistent with the postulateddefect in release of iron from R-E cells.

A prompt and steady increase in plasma iron

Iron Metabolism in Copper-Deficient Swine 2067

occurred when copper was administered (Fig. 7).A similar response was observed in copper-de-ficient rats by Marston and Allen (13). Thesestudies may be contrasted with observations madein cetrain other nutritional deficiencies, such asvitamin B0 deficiency in swine (14), or in per-nicious anemia in humans (15). In such disorders,the plasma iron decreases rapidly when the de-ficient nutrient is supplied, because the normo-blasts withdraw iron from the plasma more rap-idly than iron is introduced into the plasma. Theincreasing plasma iron observed in treated copper-deficient pigs suggests that iron flowed into plasmamore rapidly than it flowed out. Since a hemato-poietic response ensues in such animals (1), itseems improbable that the change which occurredon copper administration was a decrease in plasmaclearance. Instead, an increase in flow of iron toplasma appears likely. The most reasonable inter-pretation of the phenomenon is that it reflects thecorrection of the defect in R-E iron release. Theiron could not have come from the intestine, sincethe animals studied had not received oral ironsupplements; however, the hepatic perenchymalcell is a possible alternative source.

The expected end result of defective iron releasewould be accumulation of iron by R-E tissues.Indeed, R-E iron storage was increased in copper-deficient (oral iron) pigs by comparison with iron-deficient pigs with similar amounts of total bodyiron.

It appears, therefore, that the R-E cells in cop-per-deficient animals have an abnormality that issimilar to the one found in mucosa. Iron releaseis impaired, consequently iron storage is increased.

The hepatic parenchymal cells. Hepatic cellsreceive iron directly from the plasma and returnit to the same pool. No direct measurements weremade of the rate of iron uptake or release by he-patic cells in copper-deficient pigs. However, ex-cessive storage of iron by these cells was apparent(Table I, Fig. 8). This finding implies that atsome point, iron uptake exceeded iron release. Itseems likely that hepatic cells, like duodenal mu-cosa and R-E cells, have an impaired ability torelease iron and that the excessive iron storageoccurred for that reason. However, the possibilitythat the increased storage resulted from increaseduptake of iron cannot be excluded.

The role of copper in iron metabolism. The

failure of three different tissues to release ironnormally in copper-deficient animals suggests thatcopper is essential to a biochemical reaction thatresults in the transfer of iron from cells to plasmaand which is common to all three tissues. Al-though the nature of the proposed copper depend-ent reaction is unknown, a reasonable hypothesismay be offered.

Iron is incorporated into apotransferrin only inthe ferric state. Osaki, Johnson, and Frieden (16)have demonstrated that ceruloplasmin functionsin vitro as a ferroxidase, and they have suggestedthat this enzyme may have a biological role in theincorporation of iron into transferrin accordingto the following reaction:

CFe++ -> Fe+++ + ApoTf > Tf

in which C refers to ceruloplasmin, ApoTf refersto apotransferrin, and Tf refers to transferrin.

A deficiency of ceruloplasmin is one of theearliest manifestations of copper deficiency (1).If duodenal mucosal cells, reticuloendothelial cells,and hepatic parenchymal cells transfer iron totheir surfaces in the form of ferrous iron, this ironwould not be oxidized to the ferric state, in theabsence of ceruloplasmin, at a sufficiently rapidrate to allow for the normal movement of ironfrom these cells to the plasma. Such a hypothesisis quite compatible with the observation reportedherein that the release of iron from these cells toplasma is impaired. A possible objection to thishypothesis is that patients with Wilson's diseaseand a deficiency of ceruloplasmin are not knownto have abnormalities in iron metabolism.

Pathogenesis of the anemia. It is tempting topostulate that the anemia of copper deficiency isthe consequence of an inadequate supply of ironto marrow, brought about by impaired release ofiron from mucosal cells, reticuloendothelial cells,and hepatic parenchymal cells. However, this ex-planation seems unlikely since the hypoferremiaobserved in copper-deficient (i.m. iron) pigs wasof brief duration and bore no clear temporal rela-tion to the development of anemia (Fig. 5). Fur-thermore, late in the course of the deficiency, whenthe reticulocyte count and the volume of packedred cells continued to decline, the plasma ironlevel was frequently normal or increased. Finally,in conditions in which erythropoiesis is limited bythe supply of iron available to the bone marrow,

2068 G. R. Lee, S. Nacht, J. N. Lukens, and G. E. Cartwright

stainable normoblast iron is markedly reduced(17). By contrast, some normoblast iron was seenin copper deficient pigs, even in the group thatreceived oral iron supplements. Moreover, normo-blast iron was not only present, but markedly in-creased in the group that received intramusculariron supplements. These observations would ap-pear to exclude the possibility that iron-limitederythropoiesis is an adequate explanation for theanemia.

We propose, therefore, that an additional de-fect plays a role in the pathogenesis of the anemiaof copper deficiency, and that this defect lies withinthe normoblast itself. Both the terminal rise in theplasma iron values and the increase in stainablenormoblast iron could be explained by such a de-fect. Previous investigations failed to find evidencethat the defect affected porphyrin or heme biosyn-thesis (3). In view of the defective metabolism ofiron found in other tissues, it is possible that thedefect in the normoblast also affects a biochemicalreaction involving iron. If this is the case, it wouldbe necessary to implicate a step involving intra-cellular iron metabolism rather than one connectedwith the transportation of iron across a cell wall.Such a defect might result in the accumulation ofiron in a form that could not be incorporated intohemoglobin.

ACKNOWLEDGMENTSThe authors would like to thank Dr. Wayne H. Linken-heimer of the American Cyanamid Company for supply-ing the intramuscular iron preparation used in thesestudies and Dr. Fenimore T. Johnson of the UpjohnCompany for supplying heparin for use as an anticoagu-lant. The technical assistance of Miss Jacqueline Thomas,Mrs. Alice Tustison, and Mrs. Barbara Halliday isgratefully acknowledged.

This investigation was supported in part by Researchgrant AM-04489 and Training grant AM-5098 from theNational Institute of Arthritis and Metabolic Diseases,National Institutes of Health, Bethesda, Md.

REFERENCES1. Lahey, M. E., C. J. Gubler, M. S. Chase, G. E. Cart-

wright, and M. M. Wintrobe. 1952. Studies on copper

metabolism. II. Hematologic manifestations of copperdeficiency in swine. Blood. 7: 1053.

2. Cartwright, G. E., C. J. Gubler, J. A. Bush, andM. M. Wintrobe. 1956. Studies on copper metabolism.XVII. Further observations on the anemia of copperdeficiency in swine. Blood. 11: 143.

3. Lee, G. R., G. E. Cartwright, and M. M. Wintrobe.1968. Heme biosynthesis in copper deficient swine.Proc. Soc. Exptl. Biol. Med. 127: 977.

4. Gubler, C. J., M. E. Lahey, M. S. Chase, G. E. Cart-wright, and M. M. Wintrobe. 1952. Studies on coppermetabolism. III. The metabolism of iron in copperdeficient swine. Blood. 7: 1075.

5. Cartwright, G. E. 1968. Diagnostic Laboratory H-ema-tology. Grune and Stratton, New York. 4th edition.

6. Gom6ri, G. 1936. Microtechnical demonstration ofiron. A criticism of its methods. Am. J. Pathol. 12:655.

7. Andersen, D., and W. C. Shoemaker. 1965. Methodfor tissue hemoglobin analysis. Clin. Chem. 11: 422.

8. Bush, J. A., W. N. Jensen, G. E. Cartwright, andM. M. Wintrobe. 1955. Blood volume studies in nor-mal and anemic swine. Am. J. Physiol. 181: 9.

9. Brown, E. B., and M. L. Rother. 1963. Studies onthe mechanism of iron absorption. I. Iron uptake bythe normal rat. J. Lab. Clin. Med. 62: 357.

10. Manis, J. G., and D. Schachter. 1962. Active trans-port of iron by intestine: features of the two-stepmechanism. Am. J. Physiol. 203: 73.

11. Bush, J. A., W. N. Jensen, J. W. Athens, H. Ashen-brucker, G. E. Cartwright, and M. M. Wintrobe.1956. Studies on copper metabolism. XIX. Thekinetics of iron metabolism and erythrocyte lifespanin copper deficient swine. J. Ezptl. Med. 103: 701.

12. Noyes, W. D., T. H. Bothwell, and C. A. Finch. 1960.The role of the reticulo-endothelial cell in ironmetabolism. Brit. J. Haematol. 6: 43.

13. Marston, H. R., and S. H. Allen. 1967. Function ofcopper in the metabolism of iron. Nature. 215: 645.

14. Cartwright, G. E., and M. M. Wintrobe. 1948. Studieson free erythrocyte protoporphyrin, plasma copper,and plasma iron in normal and in pyridoxine-deficientswine. J. Biol. Chem. 172: 557.

15. Hawkins, C. F. 1955. Value of serum iron levels inassessing effect of haematinics in the macrocyticanaemias. Brit. Med. J. 1: 383.

16. Osaki, S., D. A. Johnson, and E. Frieden. 1965. Thepossible significance of the ferrous oxidase activityof ceruloplasmin in normal human serum. J. Biol.Chem. 241: 2746.

17. Bainton, D. F., and C. A. Finch. 1964. The diagnosisof iron deficiency anemia. Am. J. Med. 37: 62.

Iron Metabolism in Copper-Deficient Swine 2069