EXPERIMENTAL CELL RESEARCH 232, 246–254 (1997)ARTICLE NO. EX973526

Role of Heme–Hemopexin in Human T-Lymphocyte Proliferation1

Ann Smith,2 Jeffrey D. Eskew, Corina M. Borza, Michael Pendrak, and Richard C. Hunt*

Division of Molecular Biology and Biochemistry, School of Biological Sciences, University of Missouri-Kansas City, 5100 Rockhill Road,Kansas City, Missouri 64110-2499; and *Department of Microbiology and Immunology,

University of South Carolina Medical School, Columbia, South Carolina 29208

mechanism at sites of injury, infection, and inflam-mation. q 1997 Academic PressHeme–hemopexin supports and stimulates prolifer-

ation of human acute T-lymphoblastic (MOLT-3) cells,suggesting the participation of heme in cell growthand division. MOLT-3 cells express approximately INTRODUCTION58,000 hemopexin receptors per cell (apparent Kd 20nM), of which about 20% are on the cell surface. Bind- Deficiencies in iron status are known to impair theing is dose- and temperature-dependent, and growth function of cells of the immune system, especially lym-in serum-free IMDM medium is stimulated by 100– phocytes. Iron is required for the growth and prolifera-1000 nM heme–hemopexin, consistent with the high tion of all mammalian cells, in part because iron depri-affinity of the receptor for hemopexin, and maximal vation arrests DNA synthesis by inhibiting ribonucleo-growth is seen in response to 500 nM complex. Growth tide reductase (ribonucleoside diphosphate reductase,was similar in defined minimal medium supple- E.C. 1.17.4.1.), an iron requiring enzyme catalyzing themented with either low concentrations of heme–he-

first committed step in DNA synthesis [1]. This enzymemopexin or iron-transferrin, and either of these com-contains a unique binuclear ferric iron center whichplexes were about 80% as effective as a serum supple-stabilizes a tyrosyl-free radical essential for enzymement. Heme–hemopexin, but not apo–hemopexin,activity. Iron is transported into cells via receptor-me-reversed the growth inhibition caused by desferriox-diated endocytosis of transferrin, and iron-transferrinamine showing that heme–iron derived from hemeis well known to support the growth of all cells includ-catabolism is used for cell growth. Cobalt-protopor-ing lymphocytes [2, 3]. Moreover, cell surface trans-phyrin (CoPP)–hemopexin, which binds to the recep-ferrin receptors are increased upon activation of lym-tor but is not transported intracellularly [Smith et al.,phocytes [4].(1993) J. Biol. Chem. 268, 7365], also stimulated cell

proliferation in serum-free IMDM but did not ‘‘res- Whether heme (iron-protoporphyrin IX), which cancue’’ the cells from desferrioxamine. Furthermore, maintain hepatoma cell growth in vitro [5], see Discus-CoPP–hemopexin effectively competed for the hemo- sion below], can also support the growth of other cellpexin receptor with heme–hemopexin and dimin- types is not known. Hemopexin is a 60-kDa plasmaished its growth stimulatory effects. In addition, pro- glycoprotein which binds heme with high affinity [Kdtein kinase C (PKC) is translocated to the plasma õ 1 pM; [6]] and forms an important link between hememembrane within 5 min after heme–hemopexin is and iron metabolism by transporting heme, derivedadded to the medium, reaches maximum activity from hemolysis and damaged tissues, via a specific re-within 5–10 min, and declines to unstimulated levels ceptor in vivo to hepatic parenchyma [7]. Heme is ca-by 30 min. Heme–hemopexin and CoPP-hemopexin tabolized by heme oxygenase (HO; [8–11] causes theboth augmented MOLT-3 cell growth stimulated by down-regulation of transferrin receptor mRNA [12],serum. Thus, heme–hemopexin not only functions as and the heme-iron is reutilized [5] or stored by ferritinan iron source for T-cells but occupancy of the hemo- [13]. Hemopexin and transferrin constitute a distinctpexin receptor itself triggers signaling pathway(s) in-

class of endocytic transport systems in which both thevolved in the regulation of cell growth. The stimula-transport protein and the receptor recycle while hemetion of growth of human T-lymphocytes by heme–or iron, respectively, is released intracellularly [14].hemopexin is likely to be a physiologically relevant

Heme was reported to be a mitogen for peripheralblood mononuclear cells (PBMC), and the mitogenic1 A preliminary account of a portion of this work was presented at effect was reportedly macrophage-dependent [15].the ASBMB meeting in Washington DC, 1994.However, unphysiologically high free heme concentra-2 To whom correspondence and reprint requests should be ad-

dressed. Fax: (816)-235-5158. E-mail: annsmith@cctr.umkc.edu. tions were required (50–100 mM; [15]), and in contrast

2460014-4827/97 $25.00Copyright q 1997 by Academic PressAll rights of reproduction in any form reserved.

AID ECR 3526 / 6i1f$$$321 04-02-97 10:39:53 ecl

247HEME-HEMOPEXIN STIMULATES T-CELL PROLIFERATION

Cell culture and proliferation assays. MOLT-3 cells were rou-similar heme concentrations (10 to 100 mM) inhibitedtinely cultured in RPMI-1640 supplemented with 10% FBS and 0.5%the growth of mouse spleen lymphocytes [16]. In vivogentamycin. For most experiments the cells were incubated in IMDMthe plasma heme-binding proteins, hemopexin (0.5 to containing insulin (10 mg/ml), gentamycin (0.5%), and b-mercapto-

1.2 g/L, Kd õ 1 pM; [17]) and albumin (35 to 55 g/L, ethanol (10 mM) for 24 h before addition of iron supplements. Beforeeach growth experiment, the cells were washed three times in IMDMKd Çl0 nM; [18]), preclude the existence of significantand then aliquoted (2 1 104 cells/ml) in IMDM containing insulin,amounts of free heme; and hemopexin prevents hemegentamycin, and b-mercaptoethanol with varying concentrations offrom entering cells which lack hemopexin receptors,heme–hemopexin, diferric–transferrin, or desferrioxamine (DF).

such as L7 fibroblasts [19]. Growth studies were carried out using a minimum of quadruplicateMouse hepatoma (Hepa) cells, which express hemo- samples for individual time points from a series of two to four inde-

pendent experiments. Data shown are the mean and S.D. from onepexin receptors on their surface, grow and divide nor-representative experiment. Additional experimental details aremally in a defined medium containing heme-hemopexingiven in the figure legends. Cells incubated in RPMI-1640 containingbut lacking iron salts or iron transferrin [5]. Depriving 10% FBS and gentamycin were used as the control for maximal

these cells of iron by treatment with desferrioxamine growth and had a doubling time for MOLT-3 cells of 26 h. After 24,(DF), a membrane-permeable iron chelator, increases 48, or 72 h incubation at 377C in 5% CO2, viable cell number was

determined using the CellTiter 96 Nonradioactive Proliferation/Cy-hemopexin-mediated heme transport by a mechanismtotoxicity assay (Promega) following the manufacturer’s protocol.independent of protein synthesis, suggesting a link be-This assay is based on the cellular conversion of a tetrazolium salttween hemopexin and iron homeostasis and cell growth into a blue formazan product that is detected by its absorbance at

[5]. Whether nonhepatic cells like lymphocytes can use 570 nm. To improve sensitivity the absorbance at 650 nm was sub-tracted from that at 570 nm. Briefly, 15 ml of assay dye was addedheme–hemopexin as an iron source has not previouslyto 100 ml of cell suspension followed by a 4-h incubation at 377C inbeen addressed. Of particular relevance here is that5% CO2. Then 100 ml of assay solubilization buffer was added tolymphocytes are exposed to heme-hemopexin at sites ofeach sample, and after an overnight incubation at room temperatureinfection and tissue injury. For example, heme released samples were mixed and the absorbance (570 minus 650 nm) re-

from hemoglobin by nanomolar concentrations of hy- corded using a Molecular Devices Thermomax plate reader. Therewas a linear relationship between the measured absorbance anddrogen peroxide, likely to be present during inflamma-MOLT-3 cell number over the range 104 to 106 cells/ml.tion, is rapidly bound by hemopexin, preventing oxida-

Hemopexin binding assays. After rinsing twice in serum-free me-tion of lipoproteins by hemoglobin [20]. Since the irondium, MOLT-3 cells were incubated in Hepes-buffered serum-freesaturation of transferrin is decreased in infection, we medium containing bovine serum albumin (50 mg/ml). Ligand bind-

investigated here whether heme-hemopexin could re- ing was assessed at 37 and 47C for up to 1 h using 125I-hemopexinplace iron–transferrin in supporting the growth of T- complexed with heme using a slight modification of previously pub-

lished assays [25]. Briefly, the amount of specifically bound hemo-lymphocyte cells using a human acute lymphoblasticpexin was determined from the difference in cell-associated 125I-he-cell line (MOLT-3).mopexin after incubation of the cells in the presence and absence ofnonradiolabeled hemopexin (1 mg/ml). Cells were recovered aftercentrifugation through 20% sucrose in PBS and cell-associated radio-MATERIALS AND METHODSactive ligand measured using an LKB g-counter. The total numberof hemopexin receptors was determined from the plateau value ofMaterials. All metalloporphyrins were obtained from Porphyrinthe specifically cell-associated hemopexin, which was generallyProducts (Logan, UT). Mesoheme (iron-mesoporphyrin IX) was usedreached by 30 to 60 min incubation at 377C. The number of cellin place of heme (iron-protoporphyrin) since mesoheme is more stablesurface receptors was estimated from similar incubations at 47C. Forthan heme and mesoheme–hemopexin complexes are chemically andeach binding experiment, two independent assays were set up usingbiologically equivalent to heme–hemopexin (21, 22). Human iron-either duplicate or triplicate samples. The apparent dissociation con-saturated transferrin and insulin were obtained from Collaborativestant (Kd) for binding was estimated for the high-affinity sites afterBiomedical Products (Bedford, MA). Desferrioxamine was purchaseddetermining Bmax by plotting the amount of specifically bound heme–from Ciba-Geigy Corp. (Summit, NY) and dissolved in PBS before125I-hemopexin at 377C versus concentration of radioactive liganduse. Iscove’s modified Dubecco’s medium (IMDM), RPMI-1640, fetaladded to the cells.bovine serum (FBS), and gentamycin were purchased from Gibco-

Determination of protein kinase C (PKC) activity. PKC assaysBRL (Grand Island, NY). The MOLT-3 cell line was obtained fromwere carried out using a commercial kit (GIBCO-BRL), and the spe-the American Type Culture Collection.cific PKC activity was determined in cytosol and membrane extractsIsolation and characterization of hemopexin and its complexes(ca. 20 mg protein) by measuring kinase activity in the presence andwith heme and heme analogs. Hemopexin was isolated from rabbitabsence of a PKC peptide-specific inhibitor according to the manufac-serum, purified, and characterized as previously described [21, 23].turer’s protocol. PKC activity was increased after purification of theComplexes of hemopexin with mesoheme or cobalt protoporphyrinextracts by ion-exchange chromatography on DE-52 resin. Additional(CoPP) were prepared by mixing 1.1 molar equivalent of tetrapyrroledetails are given in the text and figure legend.with 1 molar equivalent of protein. Unbound ligand was removed by

dialysis at 47C. Full saturation of the hemopexin complexes was con-firmed by absorbance spectroscopy using published procedures [21, 23]. RESULTSTo minimize aggregation, solutions of CoPP or mesoheme in dimethylsulfoxide (DMSO) were prepared fresh each day, and their concentra- Expression of hemopexin receptors on MOLT-3 cells.tions were determined using an extinction coefficient (ArmM01

rcm01)When MOLT-3 cells were incubated with heme–125I-of 180 at 424 nm in 0.1 M NaOH:pyridine:H2O (3:10:17 v/v) for CoPP

and 170 at 394 nm in DMSO for mesoheme [24]. hemopexin complexes at 377C, binding was rapid and

AID ECR 3526 / 6i1f$$$321 04-02-97 10:39:53 ecl

248 SMITH ET AL.

FIG. 1. Expression of hemopexin receptors on MOLT-3 cells. (A) MOLT-3 cells were grown in normal culture medium and then rinsed,and the amount of cell-associated 125I-hemopexin was quantitated after incubation at 377C for the times indicated with heme-125I-hemopexinin the presence or absence of nonradiolabeled heme-hemopexin as described under Materials and Methods. The mean and S.D. are shownfor the data from two independent experiments (s, l) normalized by expressing the specifically bound heme-125I-hemopexin as percentageof B saturation (plateau value). (B) Shown is a nonlinear regression analysis of the amount of heme-125I-hemopexin specifically bound toMOLT-3 cells after incubation at 377C for 1 h with increasing concentrations of radioactive complex. Two independent saturation bindingexperiments (s, l) were used to determine Bmax; and the derived apparent Kd , at 50% Bmax, is 20 nM.

reached a plateau within 30 to 60 min (Fig. 1A). The mercaptoethanol or both did not further stimulate cellgrowth compared with that seen with diferric–trans-average total number of binding sites from two inde-

pendent experiments was 58,000 per cell. By compar- ferrin or heme–hemopexin (data not shown).ing the amount of ligand bound after 60 min at 377C Effect of desferrioxamine (DF) on heme–hemopexinwith that bound on the surface at 47C (1.23 //0 0.58 supported MOLT-3 cell growth. To determine whetherng/106 cells) it was estimated that about one-fifth ofthe receptors are on the cell surface (ca. 11,000 recep-tors per cell). Analysis of binding saturation curves at377C gave a maximum binding, Bmax, of 12.8 fmol/106

cells from which an apparent Kd for hemopexin bindingof approximately 20 nM was derived (Fig. 1B).

Comparison of the effect of heme-hemopexin and iron-transferrin on cell proliferation. MOLT-3 cells wereincubated in IMDM lacking iron salts or iron–trans-ferrin but supplemented with heme–hemopexin (0.05to 2 mM). Heme–hemopexin supported proliferation ina dose-dependent manner and was stimulatory at con-centrations as low as 0.1 mM (Fig. 2). Cell growth wasincreased by 0.5 mM heme–hemopexin to 120, 160, and300% after 24, 48, and 72 h, respectively, comparedwith cells lacking hemopexin in the medium (Fig. 2).MOLT-3 cells grown in RPMI-1640 containing 10%FBS displayed an average growth of 130% after 24 h,220% after 48 h, and 350% after 72 h. Thus, heme–hemopexin alone was generally 70 to 80% as effectiveas a 10% serum supplement. Growth was also en-hanced to a similar extent by iron-free IMDM supple-

FIG. 2. Dose–response curve for the effects of heme-hemopexinmented with diferric transferrin (0.01–1.0 mM; data on MOLT-3 cell proliferation. MOLT-3 cells (2 1 104 cells/well) werenot shown), and in general heme–hemopexin was as grown in microtiter plates with or without heme-hemopexin (0.05 toefficient as transferrin in supporting cell proliferation 2 mM) and the number of viable cells was determined 72 h later

using the Promega proliferation assay, as described under Materialsin iron-deficient IMDM (Fig. 3). Substantially lessand Methods. The means and standard deviation of triplicate sam-growth was seen when cells were cultured in IMDMples from one representative experiment, from a series of three inde-supplemented with insulin, with or without b-mercap- pendent experiments, are shown. The mean value for cells grown in

toethanol, but lacking either heme–hemopexin or serum-free IMDM was defined as 100% (O.D. 560 minus 650 nm of0.18 //0 0.05) and used to normalize all other values.iron–transferrin; and supplementation with insulin, b-

AID ECR 3526 / 6i1f$$$321 04-02-97 10:39:53 ecl

249HEME-HEMOPEXIN STIMULATES T-CELL PROLIFERATION

tor [26]. Maximal growth (200% control) was observedwith 5 mM CoPP–hemopexin, 80% of that observedwith 0.5 mM heme–hemopexin. When CoPP–hemo-pexin was present in excess (10 mM), the effect ofheme–hemopexin was attenuated, and heme–hemo-pexin (0.5 mM) was no longer able to rescue the cellsfrom the inhibitory effect of DF (Fig. 6). Further in-creases in the ratio of CoPP–hemopexin to heme–hemopexin resulted in a proportional decrease ingrowth, indicating that occupancy of the receptor byexcess CoPP prevents heme–hemopexin entry into thecell and subsequent heme–iron release.

Activation of PKC activity by hemopexin. MOLT-3cells were incubated with heme–hemopexin, and thetime course of the amount of PKC translocated to theplasma membrane was assessed. Within 5 min of expo-sure to hemopexin, there is a significant increase inPKC in the membrane, reaching approximately 50% ofthe total activity (Fig. 7). However, this effect is tran-

FIG. 3. Time course of the effects of heme-hemopexin on MOLT- sient, and levels significantly declined within 10–203 cell proliferation. Three identical 96-well plates were set up for the min and returned to control levels within 30 min.three time points (24, 48, and 72 h). Iron-transferrin, 0.1 mM (m), or

Stimulation of cell growth by hemopexin complexesheme-hemopexin, 0.5 mM (j), was added to 2 1 105 cells/ml in iron-free IMDM. As a control for maximal growth, MOLT-3 cells were in the presence of serum. One important questionincubated in RPMI-1640 containing 10% FBS (.), and as a negative raised by the results presented so far is whether thecontrol the cells were incubated in IMDM alone (l). Shown are the action of hemopexin supports cell growth by supplyingmean and standard deviation of quadruplicate samples from a repre-

a necessary nutrient for the cells, viz. heme and heme–sentative experiment, which was repeated five times. Cell viabilityiron, or whether receptor occupancy induces activationin the serum-free, defined medium (IMDM supplemented with 10

mg/ml insulin, 10 mg/ml transferrin and 10 mM b-mercaptoethanol) of a signal transduction event that stimulates cellaveraged 80% of that of cells cultured in serum-supplemented RPMI-1640. The absorbance for the control data set in this experiment was0.48 //0 0.03.

the heme–iron derived from heme–hemopexin trans-port is used for cell growth and proliferation, MOLT-3cells were incubated in serum-free IMDM containingincreasing concentrations of heme–hemopexin or apo–hemopexin (0.05 to 1 mM) in the presence or absenceof DF (40 mM) for 48 h. Heme–hemopexin, but notapo–hemopexin, efficiently overcame the DF-inducedgrowth inhibition (Fig. 4).

Effect of cobalt–protoporphyrin–hemopexin on cellgrowth. The results described above demonstratethat MOLT-3 cells express the hemopexin receptor.CoPP–hemopexin binds to the hemopexin receptor inhepatoma cells, but unlike the heme, CoPP is not takenup by cells [26]. We used this complex to determinewhether stimulation of cell growth by hemopexin takesplace at least in part in response to receptor occupancyalone, which would implicate activation of one or more FIG. 4. Effect of desferrioxamine (DF) on heme-hemopexin-medi-

ated MOLT-3 cell proliferation. Cells (2 1 104 cells/ well) were incu-signaling pathways. As shown in Fig. 5, CoPP–hemo-bated in IMDM supplemented with heme-hemopexin (l), apohemo-pexin (0.5–5 mM) increased T-cell growth in serum-freepexin (.), or heme-hemopexin together with DF (40 mM; j) or apo-IMDM in a dose-dependent manner but on a molar hemopexin and DF (m). The average value for cells grown in IMDM

basis was less effective than heme–hemopexin, consis- was defined as 100% (absorbance of 0.34 //0 0.07) and used to nor-malize all other values.tent with a lack of iron and lower affinity for the recep-

AID ECR 3526 / 6i1f$$$321 04-02-97 10:39:53 ecl

250 SMITH ET AL.

FIG. 5. Effects of CoPP- and heme-hemopexin complexes on MOLT-3 cell growth. MOLT-3 cells were incubated with increasing micromo-lar concentrations of either heme-hemopexin (open bars) or CoPP-hemopexin (hatched bars) as indicated in the figure for 48 h. The meanabsorbance value (O.D. 560 minus 650 nm) for cells incubated in iron-free IMDM was defined as 100% (absorbance of 0.085 //0 0.001) andused to normalize all other values.

growth in the presence of adequate amounts of nutri- hemopexin (apparent Kd 20 nM), with ca. 58,000 hemo-pexin receptors per cell, of which about 20% reside onents. To begin to address this issue, the stimulation of

cell growth by heme–hemopexin or CoPP–hemopexin the cell surface at any one time. This level of expressionof surface receptors (ca. 11,000 per cell) is somewhatin IMDM alone was compared with that of cells growing

in serum (Fig. 8). The presence of 0.25% FBS stimu- lower than those of mouse Hepa cells (35,000 receptors,Kd 17 nM; [5]), human promyelocytic (HL-60) cellslated cell growth, yet heme–hemopexin still aug-

mented growth. Maximal growth rates stimulated by (42,000 receptors, Kd 1 nM; [27]), human polymorpho-nuclear leukocytes (57,000 receptors, Kd 2.5 nM; [28]),1–2 mM heme–hemopexin in 0.25% serum were equiv-

alent to those which occurred in response to IMDM and U937 cells (40,000 receptors, Kd 1 nM [29]) andmore similar to the levels expressed on HeLa cellscontaining 0.5 to 1% serum. A significant increase in

the stimulation of growth by CoPP–hemopexin alone (20,000 receptors, Kd 0.8 nM; [29]). The lowest reportedlevels of hemopexin receptors are on human placentalwas apparent in the presence of 0.1% serum in re-

sponse to increasing concentrations of the complex. cytotrophoblasts (3500–7000, Kd 0.34–0.85 nM; [30])and human erythroleukemic K562 cells [31], whereasEven when cell growth was increased by the serum,the highest reported are on human retinal pigment epi-the cells were stimulated additionally by CoPP–hemo-thelial cells (230,000 per cell, Kd 10 nM; [25]). In thispexin (Fig. 8B). These data support that heme–hemo-highly oxygenated, irradiated barrier tissue of the eye,pexin not only acts to provide nutrient heme–iron buthemopexin protects against damage by heme and reac-also to generate mitogenic signals.tive oxygen species [25, 32].

The present results show that heme–hemopexin sup-DISCUSSIONports proliferation of MOLT-3 cells, can replace iron–

Ligand binding studies showed that MOLT-3 cells transferrin as a growth factor, and can generate mito-genic signals in the presence of serum-stimulated cells.express relatively abundant high affinity receptors for

AID ECR 3526 / 6i1f$$$321 04-02-97 10:39:53 ecl

251HEME-HEMOPEXIN STIMULATES T-CELL PROLIFERATION

FIG. 6. Effects of heme-hemopexin on cell proliferation in the presence of CoPP-hemopexin and desferrioxamine (DF). As indicated inthe figure, MOLT-3 cells were incubated in IMDM alone or containing either 0.5 mM heme-hemopexin (heme-HPX) or 10 mM CoPP-hemopexin(CoPP-HPX) alone, or a mixture of these complexes, in the presence or absence of 40 mM DF for 48 h. As a control for maximal growth,cells were grown in RPMI-1640 supplemented with 10% FBS. The mean value for cells incubated in iron-free IMDM alone was defined as100% (the OD 560–650 nm value for this data set was 0.22 //0 0.016) and used to normalize all other values.

Low concentrations of heme–hemopexin are effective, transferrin as a growth factor indicates that heme–consistent with the high affinity of hemopexin for its hemopexin functions as a nutrient iron source. This isreceptor. The dose–response curve of growth to heme– supported by the results of experiments with DF, whichhemopexin is bell-shaped, as seen with mitogens, in- chelates iron from both extra- and intracellularcluding concanavalin A [33], and no growth stimulation sources. Intracellularly, DF prevents both the forma-occurs when the defined medium is supplemented with tion of the iron center in newly synthesized apoR2 sub-higher concentrations of heme–hemopexin (generally units of ribonucleotide reductase as well as regenera-greater than 5 mM). Importantly, these higher amounts tion of active enzyme after iron loss [34]. Heme–hemo-of heme–hemopexin are not toxic to the cells, since pexin reverses the inhibitory effect of DF, showing thatthere is no loss of cells or cell viability. As shown else- heme–iron from heme catabolism (like iron derivedwhere using mouse hepatoma cells,3 the arrested state from iron–transferrin) is used for cell growth, and pro-induced by high concentrations of heme–hemopexin is viding evidence that heme catabolism supplies iron forreversible, and replacement of the experimental me- ribonucleotide reductase. This was further supporteddium with serum-supplemented culture medium by the inability of CoPP–hemopexin to rescue the cellsallows the cells to resume growth. growing in serum-free IMDM from the inhibitory effect

The ability of heme–hemopexin to replace iron– of DF. Furthermore, by competing with heme–hemo-pexin for the receptor, a 20-fold excess of CoPP–hemo-

3 Morales, P., Eskew, J., and Smith, A., manuscript in preparation. pexin was no longer able to rescue the cells from the

AID ECR 3526 / 6i1f$$$321 04-02-97 10:39:53 ecl

252 SMITH ET AL.

mitogenic signal in cells with sufficient nutrients tosupport growth.

The mechanism whereby CoPP–hemopexin stimu-lates MOLT-3 cell proliferation appears to be due tohemopexin receptor occupancy, which has been pro-posed to activate one or more signalling pathways in-volving the generation of superoxide and hydrogen per-oxide and an H7-sensitive process likely to be activa-tion of PKC [35]. As a result CoPP–hemopexinincreases transcription of the metallothionein-1 geneindependently of metalloporphyrin uptake [26]. PKC is

FIG. 7. Time course of the activation of protein kinase C by heme-hemopexin in MOLT-3 cells. The cells were incubated with 10 mMheme-hemopexin for 0, 5, 10, 20, 30, and 60 min in serum-free IMDM.Specific PKC activity was determined in cytosol and detergent ex-tracts of membranes as described under Materials and Methods inthe presence and absence of a peptide inhibitor and normalized toprotein content. The data shown are for membrane-associated spe-cific PKC activity (dpm/mg) as a percentage of the total cellular-specific PKC activity (dpm/mg cytosolic extract plus dpm/mg mem-brane extract).

DF-induced growth inhibition. These data suggest thatan extracellular iron source as well as a means forits uptake are mandatory for cell growth as well asactivation of a signaling pathway by hemopexin recep-tor occupancy.

It is important to distinguish between the stimula-tion of cell growth by supplying a necessary nutrientand the generation of a signal transduction event thatstimulates cell growth in the presence of adequateamounts of nutrients. The data presented here showingthat heme–hemopexin augments the stimulation ofcell growth in response to serum supports that heme– FIG. 8. Effects of heme-hemopexin or CoPP-hemopexin on cell

growth in the presence of serum. MOLT-3 cells (2 1 104 cells/well)hemopexin can act as a growth factor. In work to bewere grown with or without heme-hemopexin (0.25 to 2 mM; A) orpresented elsewhere,3 heme–hemopexin stimulatesCoPP-hemopexin (0.25 to 2 mM; B) in IMDM alone (open bars) or in[3H]thymidine incorporation into DNA in the presence IMDM containing 0.1% (v/v) serum (fine hatchmarks) or 0.25% serum

of serum in mouse hepatoma (Hepa) cells. Interest- (coarse hatchmarks). The number of viable cells was determined 48 hlater using the Promega proliferation assay. The means and standardingly, increasing concentrations of CoPP–hemopexindeviation of quadruplicate samples from one representative experi-also stimulated cell growth in the presence of serum,ment are shown. Each individual growth experimental series wasand the extent of stimulation was dependent on the repeated twice. The average value for cells grown in serum-free

complex concentration. In sum, these observations sup- IMDM alone was defined as 100% (560–650 nm absorbance of 0.238//0 0.012) and used to normalize all other values.port that hemopexin receptor occupancy generates a

AID ECR 3526 / 6i1f$$$321 04-02-97 10:39:53 ecl

253HEME-HEMOPEXIN STIMULATES T-CELL PROLIFERATION

We thank L. Khalifah for her technical help with the isolation andknown to play a pivotal role in the control of cell growthcharacterization of hemopexin used in this work. Dr. J. J. Thompsonand proliferation including T-cell activation [36], and(Louisiana State University Medical Center, New Orleans, LA) washeme–hemopexin is shown here to activate rapidly the M.S. thesis adviser of Dr. K.-T. Chan and both are thanked for

PKC in MOLT-3 cells. Recently, we showed that heme– permission to cite data from this thesis. This research is supportedin part by grants from the U.S. Public Health Service, National Insti-hemopexin activates PKC and stimulates its transloca-tutes of Health (DK 37463), the University of Missouri Researchtion to the plasma membrane in mouse hepatomaBoard and a post-doctoral fellowship (to M.P.) from the Scientific(Hepa) cells [37].Education Partnership of Hoechst–Marion–Roussel Inc. (formerly

It is likely that quite complex interrelationships exist Marion–Merrell–Dow, Kansas City, MO).between the hemopexin and transferrin transport sys-tems that are mediated by both surface and intracellu-

REFERENCESlar events. Heme–hemopexin causes a rapid down-reg-ulation of transferrin receptor mRNA in Hepa cells [12]

1. Reichard, P. (1988) Annu. Rev. Biochem. 57, 349.and also decreases the expression of transferrin recep-2. Brock, J. H. (1981) Immunology 43, 387–392.tors on the surface of HeLa and U937 cells [29]. This3. Brock, J. H., Mainou-Fowler, T., and Webster, L. M. (1986) Im-suggests that in vivo during trauma, inflammation or

munology 57, 105–110.hemolysis cells which express both hemopexin and

4. Larric, J. W., and Creswell, P. (1979) J. Supramol. Struct. 11,transferrin receptors when exposed to heme–hemo- 579–585.pexin will not only use the heme–iron to replace trans- 5. Smith, A., and Ledford, B. E. (1988) Biochem. J. 256, 941–950.ferrin-delivered iron but that intracellular events 6. Kipping, D., Barth, J., and Petter, O. (1972) Dermatol. Mo-following endocytosis of heme–hemopexin will down- natsschr. 158, 873–877.regulate transferrin receptors. Such intracellular regu- 7. Smith, A., and Morgan, W. T. (1981) J. Biol. Chem. 256, 10902–lation contrasts with non-transferrin-mediated iron 10909.uptake by liver and spleen cells, which takes place in 8. Kikuchi, G., and Yoshida, T. (1976) Ann. Clin. Res. 8, 10–17.congenital atransferrinemia [38] and by phytohemag- 9. Noguchi, M., Yoshida, T., and Kikuchi, G. (1983) J. Biochem.

(Tokyo) 93, 1027–1036.glutinin-stimulated lymphocytes [39], which does notappear to be regulated and can be toxic. 10. Brown, S. B., Houghton, J. D., and Wilks, A. (1990) in Biosyn-

thesis of Heme and Chlorophylls (Dailey, H. A., Ed.), pp. 543–In conclusion, the evidence presented here and else-575, McGraw-Hill, New York.where [37] is consistent with a mechanism for hemo-

11. Wilks, A., and Ortiz de Montellano, P. (1993) J. Biol. Chem.pexin stimulation of the growth of MOLT-3 cells, in the268, 22357–22362.absence of transferrin, that involves both a promoting

12. Alam, J., and Smith, A. (1989) J. Biol. Chem. 264, 17637–signal from receptor occupancy, i.e., PKC activation,17640.

and heme–hemopexin as a source of nutrient iron. It13. Davies, D. M., Smith, A., Muller-Eberhard, U., and Morgan,

has been proposed that the transferrin receptor confers W. T. (1979) Biochem. Biophys. Res. Commun. 91, 1504–1511.not only a signal for lymphocyte proliferation [40] but 14. Smith, A., and Hunt, R. C. (1990) Eur. J. Cell Biol. 53, 234–is also required for the obligatory iron uptake. Heme 245.is released from hemoglobin by nanomolar concentra- 15. Stenzel, K. H., Rubin, A. L., and Novogrodsky, A. (1981) J. Im-tions of hydrogen peroxide likely to be present during munol. 127, 2469–2473.inflammation and is rapidly bound by hemopexin [20]. 16. Stepien, H., Kunert-Radeck, J., Stanisz, A., Zerek-Melen, G.,

and Pawlikowski, M. (1991) Biochem. Biophys. Res. Commun.Since heme–hemopexin can replace diferric–trans-174, 313.ferrin as a growth factor for MOLT-3 cells as well as

17. Hrkal, Z., Vodrazka, Z., and Kalousek, I. (1974) Eur. J. Bio-for normal human PBMC [41], we propose that utiliza-chem. 43, 73–78.tion of heme–hemopexin complexes at sites of injury,

18. Xanthoudakis, S., Graham, M., Wang, F., Pan, E. Y.-C., andinfection, or inflammation is likely to be a physiologi-Curran, T. (1992) EMBO J. 11, 3323–3334.cally relevant mechanism enabling T-lymphocytes to

19. Alam, J., and Smith, A. (1992) J. Biol. Chem. 267, 16379–proliferate and in certain instances to compete for 16384.heme–iron which some human pathogens use for

20. Miller, Y. I., Smith, A., Morgan, W. T., and Shaklai, N. (1996)growth [42–45]. When peripheral nerves are injured, Biochemistry 35, 13112–13117.e.g., by axotomy of the sciatic nerve, hemopexin is one 21. Smith, A., and Morgan, W. T. (1984) J. Biol. Chem. 259, 12049–of only a few proteins which are induced [46, 47], sug- 12053.gesting that it acts as a trophic factor for nerve and 22. Smith, A., and Morgan, W. T. (1979) Biochem. J. 182, 47–54.possibly also muscle cells. The studies here showing 23. Smith, A. (1985) Biochem. J. 231, 663–669.that heme–hemopexin can replace iron–transferrin 24. Yonetani, T., Yamamoto, H., and Woodrow, G. V. (1974) J. Biol.thus provide evidence that hemopexin is potentially Chem. 249, 682–690.important in the regulation of cell growth in both nor- 25. Hunt, R. C., Hunt, D. M., Gaur, N. K., and Smith, A. (1996) J.

Cell Physiol. 168, 71–80.mal and cancerous cells.

AID ECR 3526 / 6i1f$$$321 04-02-97 10:39:53 ecl

254 SMITH ET AL.

26. Smith, A., Alam, J., Escriba, P. V., and Morgan, W. T. (1993) 36. Berry, N., Ase, N., Kishimoto, A., and Nishizuka, Y. (1990) Proc.Natl. Acad. Sci. USA 87, 2294–2298.J. Biol. Chem. 268, 7365–7371.

37. Pendrak, M., Khalifah, L., Ren, Y., and Smith, A. (1994) FASEB27. Taketani, S., Kohno, H., and Tokunaga, R. (1987) J. Biol. Chem.J. 8, A1392. [Abstract 775]262, 4639–4643.

38. Hamill, R. L., Woods, J. C., and Cook, B. A. (1991) Am. J. Clin.28. Okazaki, H., Taketani, S., Kohno, H., Tokunaga, R., and Koba-Pathol. 96, 215–218.yashi, Y. (1989) Cell Struct. Funct. 14, 129–140.

39. Hamazaki, S., and Glass, J. (1992) Exp. Hematol. 20, 436–441.29. Taketani, S., Kohno, H., Sawamura, T., and Tokunaga, R. 40. Keyna, U., Nusslein, I., Rohwer, P., Kalden, J. R., and Manger,

(1990) J. Biol. Chem. 265, 13981–13985. B. (1991) Cell. Immunol. 132, 411–422.30. Van Dijk, J. P., Kroos, M. J., Starreveld, J. S., Van Eijk, H. G., 41. Chan, K.-T. (1989) M.S. Thesis, pp. 1–66, Louisiana State Uni-

Tang, S.-P., Song, D.-Y., and Muller-Eberhard, U. (1995) Bio- versity Medical Center, New Orleans, LA.chem. J. 307, 669–672. 42. Lee, B. C. (1992) Infect. Immun. 60, 810–816.

31. Taketani, S., Kohno, H., and Tokunaga, R. (1986) Biochem. Int. 43. Pickett, C. L., Auffenberg, T., Pesci, E. C., Sheen, V. L., and13, 307–312. Jusuf, S. S. D. (1992) Infect. Immun. 60, 3872–3877.

44. Henderson, D. P., and Payne, S. M. (1993) Mol. Microbio. 7,32. Hunt, R. C., Handy, I., and Smith, A. (1996) J. Cell Physiol.461–469.168, 81–86.

45. Smith, A. (1990) in Biosynthesis of Heme and Chlorophylls33. Di Sabato, G., Hall, J. M., and Thompson, L. (1987) Methods (Dailey, H. A., Jr., Ed.), pp. 435–489, McGraw Hill, New York.Enzymol. 150, 3–17.46. Swerts, J.-P., Soula, C., Sagot, Y., Guinardy, M.-J., Guillimot,

34. Nyholm, S., Mann, G. J., Johansson, A. G., Bergeron, R. J., J.-C., Ferrar, P., Duprat, A.-M., and Cochard, P. (1992) J. Biol.Graslund, A., and Thelander, L. (1993) J. Biol. Chem. 268, Chem. 267, 10596–10600.26200–26205. 47. Madore, N., Sagot, Y., Guinardy, M. J., Cochard, P., and Swerts,

J.-P. (1994) J. Neurosci. Res. 39, 186–194.35. Ren, Y., and Smith, A. (1995) J. Biol. Chem. 270, 23988–23995.

Received May 28, 1996Revised version received February 5, 1997

AID ECR 3526 / 6i1f$$$321 04-02-97 10:39:53 ecl