Intracellular forms of iron during transferrin iron uptake by mitogen ...

British journal of Hnematologg. 198 6 , 62, 487-494

Intracellular forins of iron during transferrin iron uptake by mitogen-stimulated human lymphocytes A D R I A NBOMFORD,STFPHEN Y O I J N GA N D ROGERW I L L I A MTl7e S Liver Utiit, King's College School of IVIedicinr and Drntistrg, London Received 9 November 1984; accepted for yirblicirtiun 30 iLliiy 1985

SUMM A K Y . Transferrin receptors expressed by mitogen stimulated human lymphocytes mediate the uptake of transferrin iron into haem, ferritin and a

non-haem, non-ferritin component. In spite of different rates of iron uptake by cells from different individuals, the proportional incorporation of 5'Fe into thcsc components was similar, suggesting that there was a n obligatory relationship between the different forms of iron in the cells. By 3 h over 60% of the iron taken up was incorporated into ferritin while less than l(1X was found in tiaem. Initially (10min) non-haem, non-ferritin iron comprised 70'$/,of total iron and this diminished to 30% by 3 h. At 10 min 80%)of iron in the non-haem, nonferritin component was retained by anti-transferrin affinity columns indicating it was transferrin-bound. The proportion retained fell to reach a steady state level of 50% by 60 min. These results indicate that 10-200/, of the iron in the cells was not recognized as transferrin, ferritin or haem iron. The finding that iron incorporation into this unidentified pool reached equilibrium while that into haem and ferritin increased suggests the iron may act as a precursor for functional and storage compounds. Among cells of the immune system there is a relationship between cellular activation or functional status and expression of the receptor for the plasma iron-transport protein transferrin (Hamilton et ul, 1984; Yourig e l crl, 1984a). Stimulation of pcriphcral blood lymphocytes with mitogens or by mixed lymphocyte culture also results in an increase in transfcrrin receptors with up to 50 times morc rcccptors on maximally activated cells than resting cell5 (Larrick & Crewwell, 19 79; Galbraith e t nl, 1980; Bomford et a], 198 3 : Young of nl, 1984b). However, the role of this receptor in the iron metabolism of proliferating cells is not clear since iron uptake rates by cultured mouse terato-carcinoma cells (Karin & blintz. 1981 ) and mitogen-stimulated human lymphocytes (Bomford et al. 1983; Young et a!, 1984-13were low compared to erythroid cells in spite of comparable receptor numbcrs. In addition, in vivo studies suggcst that the iron requirements of non-erythroid cells are very low Correspondence: Dr A. Bomford, The Lwer Unit. King's College School of Medicine and Deninark Hill. London SE5 XRX.

48i

l k r i t i\try,

488

A. Bornford, S. Young and R. Williams

(Fairbanks & Beutler, 1977). In a previous study of mitogen-stimulated human lymphocytes we found that cells from different individuals took up iron from transferrin, albeit at low rates, and that the characteristics of transferrin binding were similar to those of established cell lines (Bornford et al, 1983). In the present report the fate of transferrin iron in mitogenstimulated lymphocytes has been determined to investigate in more detail its distribution among intracellular iron components. MATERIALS AND METHODS The procedures used to harvest peripheral blood mononuclear cells from healthy male volunteers, deplete the preparation of monocytes and stimulate the T lymphocytes with optimal concentrations of phythohaemagglutinin (PHA) (5 pg/ml) have been described in full previously (Young et al, 1984b). Purijication and radio-isotope labelling of transferrin. Transferrin was prepared from pooled human plasma and labelled with '"I and 59Fe (Amersham International) as described previously (Bomford P t a], 1983). Preparation oJantibodies find umnity columns. Antisera to human liver ferritin (Bornford et al, 1977) and to human transferrin (Leibman & Aisen, 1979) were raised in rabbits and examined for specificity by radial immunodiffusion and immunoelectrophoresis (Hudson & Hay, 1980).An IgG fraction of the anti-transferrin antiserum was further purified on DEAEcellulose (Whatman Biochemicals Ltd) and linked to CKBr-Sepharose (Pharmacia) (Hudson & Hay, 1980).The capacity of this affinity gel for transferrin determined with 12'I-transferrin was found to he 0.44 mg/nil of settled gel (Conlon-Hollingshead et nl, 1982). Incubation conditions. Cells (5-10 x l o T )were incubated in 5-10 ml MEM (minimum essential medium) containing 1%w/v BSA (bovine serum albumin) and "T and j9Felabelled transferrin (2.5 pmol 1-') at 37°C under 5% C02/air. At intervals, 1-2 ml of the suspension were added to 3 ml of cold PRS (phosphate buffered saline) (Dulbecco & Vogt, 1 9 54) and the cells washed three times at 4°C. A proportion of the washed cells was counted for "'1- and j9Fe in a gamma counter and the rest kept at 4°C for further analysis. This included treatment with 0.25% (w/v) Pronase (Calbiochem-Behring Corp.) for 3 0 min at 4°C (Karin & Mintz, 1981) followed by washing the cells and counting for l L 5 1 and j8Fe. Incorporation of j g F e into cellular components. Washed cells from the time course experiments were resuspended in cold hypotonic buffer (1 ml 25 mM hepes, pH 7.5, per 5 x lo6cells) and homogenized on ice using 40 strokes in a Dounce homogenizer with a light pestle. The homogenization buffer contained 1%(w/v) sodium deoxycholate and 1%(v/v) Triton X-100, to ensure that all the ferritin protein in the homogenates was accessible to the antibody. Up to 50% of ferritin may be associated with cell membranes and inaccessible to anti-ferritin antibody in the absence of detergent (Sargent & Munro, 1975). The detergents also ensured that transferrin, contained in endocytic-vesicles (van Renswoude et RZ, 1982) became accessible to the anti-transferrin affinity gel. The cell homogenates were centrifuged at 9000 x g (av) for 2 min at 4°C in a Beckman microfuge I3 (Irvine, Scotland). The supernatants which contained 92 f3% (meanf SD, n = 3) of the radioactivity originally associated with the cells were divided and:

lntrncellular Foriris of lron during Trarisferrin lron Uptake

489

( a )Treated with anti-ferritin antiserum (10 p1 to 100 pI supernatant) for 1 h at 3 7 O C and overnight at 4°C after the addition of 2 0 p g of purified human liver ferritin (Romford et al. 1977). The brown precipitates were recovered by centrifugation at 9000 x g (av). washed three times in cold PBS and counted for "Fe. 95* 3y, (mean*SD. n = 3 ) of "Pe-ferritin. prepared as described by Miller & Perkins (1969) was precipitated under these conditions. Correction for non-specific. binding. which varied from 3% to 6% of specific counts, was made by adding non-immune rabbit serum to a parallel incubation. An iininune precipitate was formed by adding goat anti-rabbit IgG and this washed precipitate was counted for "Fe. (b)Acidified (10 p1 1 1 - 6M HCI to 100 pl supernatant) and the haein extracted with ethyl methyl ketone (Teak, 1959). The organic phases from two extractions of each supernatant were pooled and counted for "Fe. Control experiments (data not shown) using a standard haemoglobin solution showed that this technique extracted 98% of haern into the organic phase. (c) 100-400 pl applied to anti-human transferrin affinity columns (gel bed 0.1 5 ml) equilibrated in homogenization buffer. The column was washcd with this buffer until radioactivity in the eluate had fallen to background levels. Column washings and the gel were counted for "'1 and "Fe. '3 7% of a pure '"I. iyFe-transferrin solution was retained by this gel while 95% of "Fe-ferritin and "FeCI1 in MEM, 1%BSA could bc eluted with homogenixation buffer.

RESULTS Cellular uptake and locatinri of ' ' F P Incubation of stimulated cells with 'l'1 arid "Fe labelled diferric transferrin at 3 7°C resulted in the accumulation of 5'l'e which took the form of rapid initial binding of "'Fe, 1 0 l l 0 ~ e dby a steady increase for the next 150 inin (Fig 1).The rates of iron uptake calculated from the slope

MINUTES

Fig 1. Uptake of iron by mitogen stimulated hurnan lyrnphocytes. Cells ( 5 10 x 10:) froin three subjccfs were incubated in MEM contiiiriing 1% ( w / v ) BSA with diferric jYFeand "'I labelled transferriri ( 2 . 5 jrmol 1 - I ) a t 37°C under 57; C02/air.

A. Bornford, S . Young and R. Williams

490

between 30 min and 180 min were 0.78, 0.64 and 0.46 pg/min/106 cells for the three individuals. To confirm that endocytosis of transferrin had occurred and that the 59Fealso resided at an intracellular location we determined the proportion of each label that was resistant to Pronase (Karin & Mintz, 1981j in washed cells from the 30 min sample from experiment 2 in Fig 1. At least 80%ofboth labels was inaccessible and therefore intracellular.

Incorporation of' jYFe into haem andferritin As shown in Fig 2, the major iron-containing compound after 3 h was ferritin, which represented 61% of the iron in the cells. Incorporation was linear after a rapid phase of accumulation during the first 10 min. There was also a slow incorporation of iron into haem which represented 6%of the total iron at the end of the incubation. At the first time point (10 minj, 66% of the iron was not recognized as either ferritin or haem. Iron in this non-haem. non-ferritin compartment increased rapidly for the first 30 min of the iricubatiori and reached a steady-state after 60 min.

0

.

c

b

b

60

I

120

180

MINUTES

Fig 2 . Incorporation of iron into haem, ferritin and non-haem, non-ferritin components in stimulated lymphocytes. Cells from experiment 2 in Fig 1 were analysed for j9Feincorporation into haem ( 0 )and ferritin (M). The 5'Fe in the non-haem. non-ferritin fraction (A)was calculated by dillereme (total cell iyFe- [haem 59Fe+ferritin5yFe]).

Although different amounts of i9Fe accumulated in cells from different individuals the proportional incorporation of 5yFe into the haem, ferritin and non-haern, non-ferritin fractions was similar (Fig 3j. When expressed in this way, iron incorporation in haem and ferritin increased with time while that in the non-haem, non-ferritin fraction fell. This was because the absolute amounts of iron identified as haem and ferritin increased while the incorporation into the non-haem, non-ferritin fraction reached a steady state (Fig 2 ) .

Affinity chromatography of cell supernatants The nature of the non-haem, non-ferritin iron was investigated by passing cell supernatants

lntracellular Forms of lrori during Transfurin Trori Uptakc

49 1

Lo

LL

0 I-

&

25

z X

-€

T

Y!

I

0

1

120

60

I

180

MINUTES

Fig 3. Summary of 59Feincorporation into hacsm. ferritin and non-haem. non-ferritin components in stimulated lymphocytes from three individuals. Cells from three individuals described in Fig 1 were fractionated as described in Fig 2. Incorporation of 5yFeinto haem (0'1,ferritin (N)and non-haem. nonfcrritin (A)components was expressed as a percentage of total cellular "Fe. Mean results&SD for the three individuals are given.

over anti-transferrin affinity columns. 82 f 5% (nieanf SD, n = 3 ) of the 12'I-label bound to these columns, confirming that the transferrin was immunologically intact and undegraded. Early in the incubation (10min) over 50% of the "Fe in the supernatants was retained by the columns (Pig 4) indicating that the iron was transferrin-bound. As the incubation progressed, the proportion of 5'Fe retained fell to less than 2 5% at 2 h. These figures are not corrected for the diferric transferrin bound to the cell membrane (that removed by pronase) which at the 30 min time point accounted for approximately 20% of cell bound transferrin. To correct for this, surface bound transferrin was stripped from the cells with pronase before they were homogenized. There was approximately 20% less transferrin-bound 5yFein these supernatants during the first 3 0 niin of the incubation (Fig 4). Since the distribution of iron between haem, ferritin and non-haem, non-ferritin components was known (Pig 3 ) it was possible to calculate the contribution that transferrin associated 5yFemade to the non-haem, non-ferritin components. We use the mean values for non-haern, non-ferritin iron in the cells from the three subjects shown in Fig 3 . Transferrin bound 5'Fe represented nearly 80% of non-haem, non-ferritin iron in the cell at 10 min but fell to reach a steady-state level of approximately 50% by 75 min (Fig 4. inset). DISCUSSION The distribution of iron in mitogen stimulated lymphocytes was not influenced by rates of iron uptake since j9Fe was incorporated into the intracellular components in a proportionally similar manner in cells from three individuals (Fig 31. This suggests thcre was an obligatory relationship between these different forms of iron.

492

A. Bornford, S. Young nnd R. Williams

MINUTES

0

0

60

120

MINUTES

Fig 4.Contribution of transferrin-bounds9Feto total cellular 5’Fe and (inset)the non-haem. non-ferritin component in stimulated lymphocytes. The percentage contribution made by transferrin bound jYFeto total cellular s9Fefrom washed cells ( 0 )and cells from which surface diferric transferrin was removed with pronase ( 0 )is shown. The contribution (inset)made to the non-haem, non-ferritin component was calculated by using the mean percentage values for non-haem, non-ferritin iron shown in Fig 3 .

In agreement with previous reports. using cultured fibroblasts (Octave et al, 1981) and established cell lines such as Chang (White et aJ, 1976) or human erythroleukaemia cells (Klausner ef aJ, 1983). at the end of this incubation, the majority of iron taken up from transferrin was sequestered in ferritin. A small proportion of the iron, 10f2% (mean fSD, n= 3 ) of the total, was extracted into acidic butanone which represented the iron in haem containing enzymes. The other major intracellular component, which was not recognized as haem or ferritin iron, constituted 70% oftotal radioactive cell iron after 1 0 min incubation and by 90 rnin had decreased to 50%. Since the amount of iron in this non-haem, non-ferritin fraction reached steady-state after 60 min (Fig 2) its proportional contribution to total iron continued to diminish (Fig 3 ) . White et UI(19 76), who also noted the presence of a large non-haem, nonferritin pool of iron in Chang cells, were unable to identify transferrin bound iron in cell homogenates based on the absence of iron eluting from Sephadex G-200 chromatography gel at the elution volume of transferrin protein. However, it has since been recognized that transferrin remains bound to its receptor during endocytosis (Enns et al, 1983) and the molecular size of the transferrin-receptor complex is in excess of 200 000 daltons. This complex would appear in the exclusion volume of the chromatography gel used by White et al (1976) and be indistinguishable from ferritin bound iron. We identified transfcrrin bound iron by affinity chromatography and found that it contributed approximately 50% to total cell iron at 10 min or 30%when a correction was made by removing surface bound diferric transferrin with protease. The results in Fig 4 (inset) suggest that initially a high proportion (80%)of non-haem. non-ferritin iron was recognized as transferrin bound and that with time a steady-state was established between this and the remaining iron in the compartment. It

lntrucelluliir Forms oj Iron during Transjerrin Iron Uptake

49 3

should be stressed that the remaining iron, which we calculate may be as much as 1 5% of cell iron at steady-state, was not positively identified but n7as that not bound by transferrin, ferritin or haem. This estimate depends on the efficiency of the various procedures used to identify protein-bound iron and in the case of the immuno-affinity chromatography method used to recognize transferrin, binding was not completely efficient as judged by the retention of l r 5 1label in cell homogenates. While these difficulties may lead to uncertainties about the size of an unidentified pool of iron there is little doubt that a small proportion of the iron in a cell is dialysable and available for binding by iron chelating agents or apotransferrin added to cell supernatants (White rt nl, 1976). Although there is little information on the nature and location of this intracellular pool (Jacobs, 19 7 7 ) ,the results of the present study showing that the amount of iron in this pool reaches an equilibrium while incorporation into ferritin and haem increased provides support for the concept that the iron may act as a precursor for functional and storage compounds (Lynch et al. 1974).

ACKNOWLEDGMENT

This work was supported in part by grants from the Mason Medical Research Foundation and the Medical Research Council. REFERENCES ROMFORD, A,, LIS, Y., MCFARLANE, I.G. CL WILLIAMS. R. (1977) Variation in the distribution of

logy (ed. by W. J. Williams), pp. 168-177. McGraw-Hill, New York. Iwo human heart ferritin species. Isoferritin GALBRAITH, R.M.. WERNER.P., ARNAIJD, P. 8r profile and subunit composition in normal and GALRRAITH,G.M.P. ( 19 8 0 ) Transferrin binding iron overload subjects. Biochemical /oitrnal. to peripheral blood lymphocytes activated by 167, 309-312. phytohemagglntinin involves a specific recepBOME.OHI>, A., YOUNG, S.P., NOURI-AHIA, K. & tor: ligand interaction. /ournu1 of Clinical InvestiWILLIAMS, R. (1983) llptake and release of gation, 66, 1135-1143. transferrin and iron by rnitogen stimulated HAMILTON, T.A.. WEIEL,J.E. & ADAMS.D.0. ( 1984) human lymphocytes. British /ourno/ of HuertiuExpression of the transferrin receptor in murine lology, 55. 93-101. peritoneal macrophages is modulated in the CONL~N-HOLLINCSHEAD. C., BOMFORI). A. & MUNRO. different stages of activation. Journal of ImmuH.N. (1982) Rapid one-step purification of rrology, 132, 2285-2290. ferritin from cell-free translation systems. Ana- HIJDSON.I, b; HAY. F.C. (1980) Prarticul Irrirriulytical Biochemistry, 120, 2 3 5-242. noloyy, 2nd edn, pp. L O 3-22 5. Blackwell ScienDULBECCO. R. & VOGT. M. (19 54) Plaque formation tific Publications, Oxford. and isolation of pure lines with poliomyelitis JACOBS, A. ( 1 Y 77) Low molecular weight intracelviruses. Iournal of Experimental Medicine, 99, M a r iron transport compounds. Blood. 50, 167-1 99. 4 3 3-4 3 9. ENNS, C.A., LARRICK, J.W.. SLOMALAINEN, H.. KARIN, M. & MINTZ,B. (1951) Receptor rnrdiakd SCHRODER, J. & SUSMAN. H.H. ( 195 3 ) Co-migraendocytosis of transferrin in developmentally tion and internalisation of trarisferrin and its totipotent mouse terato-carcinoma stem cells. receptor on K562 cells. Journal of Cell Bioloqly. loiirnal of Biulogicul Chemistry, 256, 324597,579-55s. FAIRBANKS, V.1,.8.BEUTLER. E. ( 19 77) In: H(lenl