Iron transport across Caco-2 cell monolayers. Effect of transferrin, lactoferrin and nitric oxide

¢' ~ .~

.

~. •

.~..

BB. l~iochlm~ca d Biophy',ica Acti~ t-~X9 I I=)q¢,i 2¢~1 2tt7

ELSEVIER

etBiochi~& BiophysicaAEta

Iron transport across Caco-2 cell monolayers. Effect of transferrin, lactoferrin and nitric oxide Lourdes Shnchez ~. Maznah Ismail. Foo. Y. Liew. Jeremy H. Brock

"

Department o/hmntet~oh, qv. I '11il('r~it~ ,q Ghls~o~t. W~.~'tt,rn htl~rtnttr~. (;ta~qt,w GI I ,VT..~t othmd. I'K

Received 27 June it)95: re'.iscd 24 October 19t~5" accepted 15 November 1~,,~5

Abstract Differentiated Caco-2 colon carcinoma cell monolayer.~ grown in bicameral chamber,, have been used as an in vitro model to .,i=:',.]3. the effect of different carrier molecule.,, on mucosal iron transport. Tran.,,ter of iron actor.,, the monolayers in the apical-lo-ba.~olateral direction was greater from ferric lactoferrin than front iron citrate, while very little transport occurred from Fe-transferrin. However. a greater proportion of iron was retained by the cells when Fe-citrate was the donor. Caco-2 cell~ expressed tran.~ferrin receptors (n = 1.3 × IO~/cell: K,, = 2 x 10' I/mol}, but binding of lactofcrrin, though substantial in quantity, had an affinity too low to measure. When monolayer~ were incubated with ~''~l-labelled lactoferrin or transferrin ~ome ~:~I-activily was transporled, but almost all was TCA-soluble. ~uggesting that degradation products rather than intact protein were being tran.,,ported. Addition of I0 /.tM S-nitroso-N-acetyl-oj.-penicillamine (SNAP). which produces nitric oxide (NO) in s=~lution, caused a ~,ignificant increa~,e in iron transport from ferric citrate, but not from Fe-lactoferrin or Fe-transferrin. it is concluded that in thi.~ in vitro ~y,~tem lactoferrin but not transferrin enhance~ muco~al iron tran.~port. and that NO may play a regulatory role in iron ab,~orption. l~eywords: Iron absorption: Lactoferrin: Trunr, ferfin: Nitric oxide

nai secretions [4] facilitates iron absorption [5]. Infancy is a

I. Introduction

Iron absorption is essential for the maintenance of iron levels in the body, since excretion mechanisms are poorly regulated. Despite a great deal of work, the factors involved in iron uptake by the intestine have not yet been elucidated, although roles for the iron-binding proteins transferrin and lactoferrin have been proposed. Transferrin has been reported to facilitate the uptake of iron by the inte.,,tinal mucosa in rats [I ], although transferrin receptors appear to be present in the basolateral membrane of duodenal enterocytes rather than on the microvilli [2.3]. This suggests that transferrin may be involved in the transport of iron into these cells for their own metabolic activity. rather than mediating iron uptake at the apical border for subsequent mucosal transport. It has also been suggested that lactoferrin, a closely related iron-binding protein found in milk and other exter-

Corresponding author. jhb I h @el in med.gla.ac.uk.

Fax:

~-44

141

3373 "~t7;

e-mail:

~Pre.~enl addre.~s: Depaftamenlo de Tccnolog~a de Io.,, Alimento.~. Facullad de Veterinariu. Miguel Servel 177.5(X)I3 Zaragozu. Spain. 03(.lq.-4165/q6/$15.(X) ~' 1996 El.',evier Science B.V. All right.', tee,erred SSDI 0 3 0 4 - 4 1 6 5 ( 9 5 1 0 0 1 7 3 - 5

period when iron absorption is of great importance, and it might therefore be expected that during tfiis period iron-binding proteins present in milk could play a role in iron uptake by the intestine, The higher bioavailability of iron in human milk compared with cow's milk [6] has lent support to the idea that lactoferrin promotes iron uptake by the neonate intestine. However. whether lactoferrin actually plays such a role is controversial. While iron bound to lactoferrin was superior to FeSO.= in improving the iron status of anaemic rats [7]. in infants there was no difference in the absorption of inorganic iron and lactoferrinbound iron [81. If lactoferrin does play a role in iron absorption some form of interaction between this protein and the intestinal mucosa must occur. The presence of lactoferrin-binding protein~ has been reported on the brush-border membrane of the rabbit [9]. Rhesus monkey [5], mouse [10] and human fetus [I I]. However. iron release from lactoferrin and uptake by the cells was not demonstrated in these studies and in contrast it has been suggested that lactofertin may serve to control excessive iron absorption, rather than enhance iron uptake [ 12].

292

L Sdm'he: et al. / fliochhnica et BitJphysh'a Act,t 127t9 ( 10961 291-297

A number of other factors can influence iron absorplion. including low molecular weight substances such as ascorbic acid, which has been shown to enhance iron transport across mucosal cell rnonolayers [13]. Although ascorbic acid. through its reducing power, probably enhances the availability of iron to mucosal cells, it may also act by directly influencing intracellular iron metabolism. Another molecule that has recently been shown to influence intracellular iron metabolism is nitric oxide (NO), which is generated enzymatically by a variety of cells through oxidation of arginine [14.15], and may also be produced by bacteria such as Escherichia colt which are components of the gastrointestinal flora [ 16]. Since NO can enhance iron release from cells [I 7] it seems possible that it might also increase transport of iron across mucosal cells, but so far the effect of NO on iron absorption has not been investigated. Iron absorption studies performc:] in vivo have the drawback of dealing with a complex system in which it is difficult it. determine the relative importance of different t%ctors. In vitro cell culture models could overcome this difficulty but attempts to establish differentiated enterocyte celi line~ in culture have not been successful, and differentia:ted maligv:ant colorectal cells, though expressing several brush border enzymes, lack the structural characteristics of differentiated enterocytes [18]. However, the Caco-2 line, derived from a colon carcinoma, is able to differentia,~. spontaneously when grown in standard culture conditions [18]. The differentiated cells polarize, form microvilli ;.l-/ express enzymes associated the duodenal enterocyte brush border such as sucrose-isomaltase, alkaline phosphatase, lactase and aminopepttdase. This cell line thus represents an appropriate model for the study of transport mechanisms related to the intestinal barrier. In this work we have therefore used differentiated Caco-2 cells grown in bicameral chambers as an intestinal cell model to study passage of iron from different sources through monolayers.

2. Materials and methods

2. I. Cell culture Caco-2 cells were provided by Dr, 1. Freshney (Dept. of Oncology. Glasgo,, University) and routinely grown in Dulbecco's modified Eagle's medium (DMEM)(Gibco. Paisley, UK) supplemented with 10CTcfetal bovine serum (Northumbria, Cramlington, UK). ICh non-essential amino-acids (Flow. Rickmansworth, UK). I /.tg/ml bovine insulin (Sigma, Dorset. UK) and antibiotics (100 U / m l Penicillin. I00 p,g/ml Streptomycin; Flow), Cells were normally grown in tissue culture flasks to confluence and seeded into Transwell bicameral chambers (Costar, High Wycombe. UK) of 0.3 # m pore size and 6.5 mm diameter at a density of IO~ cells/era-'. The well contained 200 p,I

of medium in the upper chamt~er and 800/,d in the lower, The polycarbonate membra.rle of the inserts was previously precoated with collagen by adding 50 ~1 of a 2 m g / m l solution of rat tail collagen type I (Boehringer, Mannheim, Germany) in 0. I M acetic acid to each chamber. Excess solution was removed, and the inserts dried inverted under sterile conditions, in order to avoid clumps the cell suspension was passed through a nylon membrane (R. Cadish & Sons, Finchley, UK) before seeding. The cells were maintained at 37°C in an atmosphere of 5% CO,. and 90% relative humidity and the medium changed daily. Confluent cultures of differentiated cells were obtained after 17-20 days and differentiation was checked by measuring transepitheiial electrical resistance (TEER) with an epithelial voltohmeter (World Precision Instruments, New Haven, CT, USA) as described elsewhere [13]. Phenol red exclusion [19] was used as a measure of monolayer integrity. 2.2. Transmission electron microscopy Caco-2 cells grown in Transwell chambers were examined by transmission electron microscopy using standard procedures. Briefly, cells were fixed overnight with 2% glutaraldehyde followed by post-fixation in I% OsO 4 and embedding in Araldite. Sections (2 ~m) were stained with toluidine blue and ~xamined by light microscopy. Ultrathin (90 nm} sections were then cut frt3m suitable areas, mounted on 200-mesh copper grids, stained with uranyl acetate and lead citrate, and examined in a Philips CMI0 electron microscope. 2.3. Binding of lactoferrin and transferrin to Coco2 cells Human lactoferrin was isolated from milk by the method of Johansson [20] and purity checked by SDS-polyacrylamide gel electrophoresis. Transferrin was obtained from Behringwerke (Hounslow, UK). Both proteins were iodinated by the chloramine T method [21] and free '2Sl removed by passage through a Sephadex G-25 column. For the binding studies Caco-2 cells were grown to approximately 80% confluence as adherent cultures in 48-well tissue culture plates. (Preliminary experiments showed that these studies could not be carded out using bicameral chamber cultures due to high non-spe¢.,,, binding of labelled proteins to the collagen-coated membrane.) The cells were washed with serum-free medium and samples of labelled lactoferdn or transferrin at different concentrations were added. To estimate non-specific binding, a 200-fold excess of unlabelled lactoferdn or transferrin was added before the labelled protein. Cells were incubated for 30' at 37°C as incubation at 4°C caused the cells to detach from the plate. Afterwards, the cells were washed three times with PBS containing 2% bovine serum albumin, dissolved in 2% sodium dodecyl sulphate (SDS) and the radioactivity asso~aated with the cells determined in a Compugamma 1232 gamma counter (LKB. Croydon, UK).

293

L 5(mche2 et al. / Biochimico et Biophs ~'ica A c t . 12~(,*~I(Jt,t6~ 2 9 1 - 2 9 7

Binding affinities and number of sites per cell were cldculated by the Scatchard method [22]. the number of cells present being calculated from a standard curve of cell number versus protein content obtained using trypsinised cells in suspension.

2,4. iron uptake and transport .studies Caco-2 monolayers grown in bicameral chambers were used for transport studies when they were differentia'-ed and the monolayer intact, as checked by measuring TEER and phenol red exclusion. After washing both upper and lower chambers with serum-free medium, lactoferri~ )r transferrin solutions were added to the upper chamber. Both proteins were 50% saturated with ['~'~Fe]-citrate (specific activity 10 /xCi//~g, Amersham) and added at a concentration of 50 /xg/ml. Saturation to 100% was then achieved using unlabelled ferric nitrilotriacetate (FeNTA) as described previously [23]. This procedure ensures that the proteins are fully saturated, but avoids the risk of any free 59Fe being present. Transport of iron from Fe-citrate was assessed by adding an equivalent amount of [~'~Fe]citrate to the cultures in the absence of transferrin or 30GO- a 2500'

~

20O0,

.~

1500,

S

f

ID

3. Results

Specific

1000-

3. !. Lactoferrin and transferrin binding to Caco-2

I-

500" Non-specihc

0

~

,

o

i

20

•

_-I

4o

I,,

~;o

do

Jm •

i

~oo

TransfemnadOecl(no) S

b

5

~-

lactoferrin The final concentration of ~'Fe in the samples was 47(1 nM. The lower chamber contained human apotransferrin ( I m g / m l ) as iron acceptor. Serum-free medium wag u,;ed during iron uptake and transport in order to ensure that total iron levels in the medium could be controlled. Use of serum-free medium did not affect the TEER. Medium from the lower compartment was removed for analysis at i, 5 and 23 h of incubation and replaced by fresh medium. At the end of the experiment the TEER was measured, the cells were washed three times with Hank's so!ution and then dis~lved in 2% ( w / v ) SDS. Radioactivity associated with the medium from the upper a,ld lower chambers at different times of incubation, and with the SDS-digest of the cells was determined. The same experimental conditions were u ~ d when transport experiments were carried out with iodinated proteins. Integrity of [1-~Sl]-lactoferrin or transferrin that had pa.s~d across the Caco-2 monolayer wag determined by precipitation of an aliquot with 10c~ ( w / v , final concentration)trichloroacetic acid. The effect of nitric oxide on iron transport was assayed under the same experimental conditions by adding IO~M S-nitroso-N-acetyl-D,L-penicillamine (SNAP) (Schwarz Pharma, Monheim, Germany). which generates nitric oxide in solution, to both compartments. Preliminary experiments showed that this concentration gave optimal responses without loss of cell viability or alteration of the TEER. Iron transport was determined after 48 h,

•

4,

0 o.oo

" NN,.. x o.;~

o.;,-

o.;3

Binding studies (Fig. ia,b) revealed that there was specific binding of transferrin to Caco-2 cells, Scatchard analysis (Fig. i b) showed a single binding site with an affinity constant of 1.5. IOs l / m o l and 1.3-10 s binding sites per ceil. Although lactoferrin bound to Caco--2 cells, the binding was non-saturable and the Scatchard plot obtained did not show the existence of a receptor for lactoferrin with measurable affinity (Fig. 2). it seems unlikely that this result was due to denaturation of lactoferrin during iodination, as we have previously demonstrated specific binding to monocytes of iodinated lactoferrin prepared in this way [23].

4. Transport stmlimi o.h

" o.;s

Bound (pmoles) Fig. I. Tr,msfcrrin binding to Caco-2 cells, a. Binding curves. Ceils were incubated with I00, 250, 400, 600, 800 and 1000 ng [l:-~l]-transferrin per well in the presence or absence of 200-fold excess of unlabeled transfer-

rin for 30" at 37°C. b. Scatchard plot of me data shown in Fig. In.

4. I. Iron transport and

uptake

Transmission electron microscopy revealed that Caco-2 monolayers consisted of differentiated cells (Fig. 3). These were polarised with a typical asymmetric morphology, and

L S/m+'h,': e! ~11./ Bi,,'l+imi,'el ,,t flir+phy.~'i,'a Arla 12,~9 ~IU~/~)2W-2~J7

2t~4

3000t a

2000~ Total 1500

Specific 1ooo ,,,J

5OO Non-Specific

&

I

o

,

,

",'0

o

~

•

60

,;o

e'o

Lactofemn aUded (rig) 0.4"

b

% ,e-,. ~

0.2

OA

0.0

•

0.00

!

0,05

•

l 0.10

•

"i

0.15

t 0.20

•

0.25

Bound {pmoles)

Fie. 2. Lactol+crrinhindir'g to Caco-2 cells, a. Binding curves. Cells were incuhalcd with ~:~l-lacl~)ferrina~ described for transfcrrin in Fig. la. b. Scatchard plot of the dala shown in [.ig. 2a.

glycogen deposits and brush border microvilli were pre.sent. Integrity of the, monolayer was c o , fin;ned by the presence of tight junctions in the apical zone, and interdigitation and desmosomes along the cells. Iron transpori experiments were carried out when the monolayers reached the maximal TEER. about 230 / t c m : . and the passage of phenol red was < 2 . 5 ~ per hour. Despite the lack of specific binding of lactoferrin to Cat:o-2. iron bound to this protein was transported across the monolayer more readi:y than iron bound to citrate ( P < 0.0i ). while almost no lransport of transferrin-bound iron occurred (Fig. 4). in all cases the amount transported formed only a small proportion ( < 3%) of the total iron added, a finding that does. however, mirror the limited absorption of dietary iron that occurs in vivo. The amount of iron associated with the cells (Table I) was however significantly greater ( P < O.OI ) for citrate than for transf e n i n or lactoferrin. Integrity of the monolayer after the transport experimenl,+ was checked by measuring the TEER.

Fig. 3. Transmission electron mierograph of a monolayer of Caco-2 cells after 16 days of culture in a bicameral chamber (x5000). (a) brush border microvilli; (b) junctional complexes.

4.2. Transport of lactoferrin and transferrin When the passage of lactoferrin or tran~ferdn proteins across Caco-2 monolayers was studied, it was found that 3.8% and 4,7% of the 12"+!label originally associated with lactoferrin and transferrin respectively had traversed the monolayer in 23 h (Fig. 5), At this time the amount of

Fe-lactoferrin

Fe.civate

.o

Fe-~'ansferrin o

0

0

-

i

i

5

~0

!

i

15

20

25

Time (hours)

Fig, 4. Iron tran~porl across Caco-2 mono|aycrs in hicamera| chambeTs. Iron (470 nM) was added Io the upper chamber as ~'~Fe-citrate, '~Felac(ofcrrin or ~'~Fe-tran.-.i'crrin.Values represent the percentage of total radioactivity tran.~ferredto tbe lower chamber (mean + SD of six or eight replical¢~ from three diflerent experimcnls).

L. S{lm'he: et al, / Biorhimi<'a et Biophvsica Acta 12,~9 r 1996~ 291-297 Table I Iron uptake into Caco-2 cells Fe-citrate Fe-lactofetrin Fe-transferrin

7.55 + 1.5 I 2.69 -I-0,56 3.45 5- 0.67

Values are expressed as percentage of the total iron added to the upper

chamber, determined in iron transport experiments across Caco-2 monolayers in bicameral chambers. Assayswere performed 23 h after addition of the iron source. Values are the mean-l- SD of six or eight replicates from three different experiments.

each protein associated with the cells, as det,:rmined by cell-bound 1251 activity, was 1.84% for lacttferr!,~ and 0.45% for transferrin. However, for both pro eins only about 20% of the radioactivity that had traversed the monolayer was TCA-precipitable, indicating that most of the protein was degraded when crossing the cells. To eliminate the possibility of the protein being degraded by proteases released at the apical surface of the ceils prior to transport, the integrity of lactoferrin and transfem'n in the upper chamber was also checked by T e A precipitation. No change occurred over the incubation pedoa, all the protein 6

I=

o

- - "

i

I

'

5

'~

i

•

10

q

•

15

i

"

2C

~

25

Time [hours)

~" t=

il °

"5

3-

P"

1-

Total

•

0

• 0

;

1'0

1'5

20

15

Time (hours] Fig. 5. Passage of I"Sl-lactoferrin(a) or i:Sl-transferrin (b) across Caco-2 monolayers in bicameral chambers. Proteins (100% iron-saturated) were added at a concentration of 50 /,¢g/ml to the upper chamber. Values represent the percentage of total radioactivity transferred to the lower chamber (mean ± S D of six replicates from three different experiments).

295

in the upper chamber being intact both at the beginning and at the end of the incubation period, it should be noted that the amount of intact protein traversing the membrane was only about 0.5% of that added to the upper chamber, indicating that "leakage" of proteins across the monolayers was minimal. 4.3. Effect t+f nitri," oxide Nitric oxide generated by the addition of 10/.tM SNAP had no effect on transport of iron from lactoferrin or transferrin, the figures being 108.4 +_ 8.4 and 99.8 :t: 2.4% (n = 6) of control cultures without SNAP respectively. However, it caused a significant ( P < 0.01) increase to 143.6 ___0.4% of the control value in iron transport from ferric citrate.

5. Discussion

Differentiated Caco-2 cells provide an in vitro model of the intestinal barrier for the study transport of various substances [24,7'i]. Although there may be some differences between differentiated Caco-2 cells and duodenal enterocytes the Caco--2 cell monolayers in bicameral chambers used in this study had an enterocyte-like appearance and were differentiated, showing polarization and expression of features characteristic of small intestinal cells such as microvini in the brush border and tight junctions between the ,.ells. Therefore despite these reservations, we consider that Caco-2 cells provide a useful in vitro model for studying factors regulating iron absorption, and indeed have been used for this purpose by otheJs [13,19,26], Caco2 ceils expressed transferrin receptors as previously reported by Halleux and Schneider [19]. We found 1.3-IO s binding sites per cell with a K, of 1,5-I0 s I/mot, these figures being in the same range as those found for other cells [27], although since the study had to be perforrned at 37°C (due to detachment of the cells at 4°C) the number of binding sites probably represents both surface and intracellular receptors, Although lactoferrin binding to Caco-2 occurred, it was not possible to detect a receptor with measurable affinity using Scatchard analysis. These results partly contradict some reports in which apparently specific lactoferrin binding proteins have been isolated from the brush border of several species, such as monkey [5], mouse [I0], rabbit [9] and human fetus [ i I]. Specific binding of lactoferrin has also been reported on differentiated HT-29cI.19A colon carcinoma cells [28]. However, none of these putative receptors is as well characterised as the transferrin receptor, and the properties of these lactoferrin-binding molecules varies according to species. Nevertheless, it is possible that although Caco-2 cells ex~ress many features of the enterocytes, they might lack a component with more

2%

L. Sdnche: et .I. /Bi.chimica <,t Bwphysica Acta 128q ¢19tJ6J 2~1-297

specific lactofcffin-hinding prope.ie,~ II ,~hould be noted that ligand-binding studies were cawTied out on cells grown in tissue culture plates ralher than bicameral chambers, which might possibly have had an effect on the interaction of Coco2 cells with ]actoferin or transferdn. The pattern of iron uptake and transport across the monolayer~ varied according to the carrier, From the data in Fig, 4 and Table I it can be calculated that total iron uptake tie transported + retained) amounted to 9.1c~o 5.2c~ and 3,8~'~ of the added dose for Fe-citrate, Fe-lactoferrin and Fe-tmnslerrin respectively, Of this, the proportion transported out of the cells to the lower compartment represented 48c~ for lactoferrin, 16c~ for citrate and only 8 ~ for transfen in. TIzus despite the presence of transferrin receptors on Coco-? cells, transport of iron bound to trunsferrin was much lower than from lactoferrin or citrate. This suggests that transferrin receptors, which mediate iron uptake by receptor-mediated endocytosis [27] are primarily involved in uptake of iron for the cell's own metabolic use rather than in iron absorption. Iron uptake from citrate, which occurs by a different mechanism [29], was much higher than from transferrin or lactoferdn, but the proportion of ~his actually transporled across the monolayers, although greater than from transferdn, was considerably less than from lactoferrin. The way in which lactoferdn-iron enters the cells and is transported is unclear: it may be that lactaferrin, through its ability to bind non-specifically in relatively large amounts to the cell membrane, can target iron to the cell membrane where it is released and taken up in some other form. Release of iron from lactoferrin at the cell membrane has been reported in monocytes [23]. Alternatively lactoferdn may enter the cell by adsorptive endc,cytosis and release its iron following lysosomal degradation, as occurs in hepatocytes [30]. The amount of degraded iactoferrin transported across the monolayer (Fig. 5a) was comparable to the amount of iron transported (Fig. 4). which is consistent with the latter mechanism. These findings differ from those of Mikogami et al [28] who found that the amount of iron transported across differentiated HT-29 cell monolayers corresponded more closely to the amount of intact rather than degraded lactoferrin transported. As with the present study, however, most of the transported lactoferrin was degraded. The majority of transferrin transported across the monolayers was also degraded, suggesting that the small amount of transferrinbound iron transported across the cells probably arose from a minor proportion of transferdn intemalised by a route other than the classical receptor-mediated endocytosis. h is known that inflammatory stimuli can modulate iron absorption [31] but tile mechanisms involved are unclear. One possible candidate is nitric oxide, which is released in increased amounts by various cells during inflammation [17]. in the present work nitric oxide significantly increased iron transport from Fe-citrate. but not from lactolerrin or transterrin. NO has recently been shown to influ-

ence intracetlular iron metabolism [14,15] and is now thought to be the agent responsible for increased release of iron from tumour cells in the presence of activated macropha~es [ 17]. An ability of NO to increase the rate of transport through the Coco2 monolayers is consistent with these previous reports, though it is less obvious why an effect was only seen when Fe-citrate was used as the iron source. Possibly in this case the rate-limiting step is release from some intracellular transport compartment susceptible to the effect of NO, whereas with iron presented as lactoferdn or transferdn release from the carder protein is the rate-limiting step. Alternatively, the greater retention of iron derived from ferric citrate may provide a more substantial pool of intracellular iron susceptible to the action of NO. Further studies would be required to evaluate these proposals. Nevertheless, the fact that intestinal bacteria such as E. c o l t can produce NO [16] suggests that microorganisms may have a role in regulation of iron absorption in vivo via the production of NO.

Acknowledgements We thank Jane Hair for carrying out the electron microscopy. LS was supported by a scholarship from the Ministerio de Educaci6n y Ciencias, Spain, and MI by a scholarship from the Government of Malaysia,

References [! ] Huebers, H.A.. Huebers, E., Csiba, E., Rummel. W. and Finch, C.A. (1983) Blood 61,283-290+ [2l Banerjee. D.. Flanagan. P.R., Clueu. J, and Valberg, L.S. (1986) Gastroenterology 9 i, 861-869. [3] Parmley, R.T.. Barton, J.C. and Conrad. M,E. (1985) Br. J. Haematol. 60, 81-89. [4] Masson, P.L. and Heremans, J.F. (1971)Comp. Biochem. Physiol. 39B. 119-129. [5] Davidson, L.A. and L~innerdal, B. (1985) in Proteins of Iron Storage and Transport (Spik, G.. Montreuil, J., Crichton, R.R. and Mazurier, J., eds.), pp. 275-278, Elsevier Science Publishers, Amsterdam, The Netherlands. [6] McMillan,J,, Landaw. S. arid Oski. F. (1976) Pediatrics 58. 686,-691. [7] Kawakami, H.. Hiratsuka, M. and Shun'ichi. D. (1988) Agric, Biol, Chem. 52, 903-909. [8] Fairweathcr-Tait, SJ.. Balmer. S.E., Scott, P.H. and and Minski, M.J. (1987)Pediatr. Res. 22, 651-654. [9] Mazurier. J.. Montredil, J. and Spik. G. (1985) Biochim Biophys, Acta 821,453-460, [10] Hu, W.L.. Sawatzki, G., Montrcuil. J. and Spik, J. (1988) Biochem. J. 249. 435-441. Ill] Kawakami, H. and Ltinnerdal+ B. (1991) Am. J. Physiol, 261, G84 I-G846, [12] Brock, J,H (1980) Arch. Dis, Child. 55, 4|7-421. [13] AIvarez-Hernfindez,X,, Nichols, G.M. and Glass, J. (1991) Biochim. Biophys. Acta 1070. 205-208. [14] Drapicr. J.C., Hiding, H., Wielzerbin, J., Kaldy, P. and Kiihn, L+C. (1993) EMBO J. i2, 3043-3649. [15] Weiss, G., Goossen, B,, Doppler, W., Fuchs, D., Pantopoulos, K.. Wemer-Fclmcyer. G., Wachter. H. and Hentzc, M.W. (1993) EMBO J. 12. 3651-3657.

1,, ~ie;m'he: et ul, / lliach~mica et Biophysica At'to 12,~tl ~1996~ 2 9 1 - 2 9 7

[16] Britta'.n, T,, Blackmore, R.. Greenwood, C. and Thomson, A.J. (1992) Eur. J. Bits:hem, 209. 793-802. [17] Henry. Y.. Lepoivre, M.. Drapier. J-C.. Ducrocq. C., Bcmcher. J.-L. and Guissani. A, (1993) FASEB J. 7. 1124-tl34. [18] Pinto. M., Robine-Leon. S., Appay. M.D., Kedinget', M., Triadou. N., Dussaulx. E., Lacroix. B., Simon-Assmann P. Haffen. K., Fogh, J. av.d Zweibaum, A. (1983) Biol, Cell 47. 323-330. [19] Halleux. C. and Schneider. YJ. (1991) In Vitro Ceil, Dev. Biol. 27A, 293-302. [20] Johansson. B. (1960) Acta Chem. Scand. [4. 5]0-513. [21] Hunter, W,M. and Greenwood, F.C. (1962) Nature [94, 495-496. [22] Scatchard, G (1949) Ann. N.Y. Acad. Sci. 51. 660-672. [23] Ismai} M. and Brock. J,H. (1993)£ Biol, Chem. 268, 21618-21625. [24] Burlon. P.S.. Conradi. R.A., I-[iigers. A.R. and Ho. N.F.H (1:)93) Biochem. Biophys. Res. Commun. 190, 760-766.

297

[25] Souba, W.W. and Copeland, E.M, (1992] Ann, Surg. 215, 6g6-706, [26] Alwwez-Hernandez. X.. M~uJigosky. S,R,. Stewart. B. and Gl~s, J. {1~4) J. Nutr. 12~. 1574-1580. [27] Klausner. R D , A~hwell. G., Van Renswoud¢, ]., Harford. J.B. and Bridge,~, KR. (Iq83) Proc. Natl. Acad. Sci USA, 80. 2263-2266, [28] Mik~gami. T., Heyman. M., Spik, G, and Dcsjeux. l-P, (19~) Amer J, Physiol. 267. G308-G315. [29] Sturrock. A., Alexander. J.. Lamb. J,. Craven. C,M, and Kaplan. J. (I~)0) J, Biol. Chem. 265. 3130-3145. [30] McAhee. D.D. and Esben~n, K. (1991)j. Biol. Chem. 266. 2362423631. [3I] Skiknc, B. and Bayncs. R.D. (1994) in Iron Metabolism in Healtb and Disease (Brock. J.H., Halliday, J.W.. Pippard, M.J. and Powell, L.W., eds.}, pp. 151-187. W.B, Saunders. London. U.K.