Regulation of Reticuloendothelial Iron Transporter MTP1 (Slc11a3) by Inflammation*

Acute and chronic inflammation cause many changes in total body iron metabolism including the sequestration of iron in phagocytic cells of the reticuloendothelial system. This change in iron metabolism contributes to the development of the anemia of inflammation. MTP1, the duodenal enterocyte basolateral iron exporter, is also expressed in the cells of the reticuloendothelial system (RES) and is likely to be involved in iron recycling of these cells. In this study, we use a lipopolysaccharide model of the acute inflammation in the mouse and demonstrate that MTP1 expression in RES cells of the spleen, liver, and bone marrow is down-regulated by inflammation. The down-regulation of splenic expression of MTP1 by inflammation was also observed in a Leishmania donovani model of chronic infection. The response of MTP1 to lipopolysaccharide (LPS) requires signaling through the LPS receptor, Toll-like receptor 4 (TLR4). In mice lacking TLR4, MTP1 expression is not altered in response to LPS. In addition, mice lacking tumor necrosis factor-receptor 1a respond appropriately to LPS with down-regulation of MTP1, despite hyporesponsiveness to tumor necrosis factor-α signaling, suggesting that this cytokine may not be required for the LPS effect. We hypothesize that the iron sequestration in the RES system that accompanies inflammation is because of down-regulation of MTP1.

Iron is an essential nutrient for growth and development of eucaryotes and most prokaryote species. A normal individual will absorb ϳ1 mg of elemental iron a day through the duodenum, to match an equivalent daily physiologic loss. The plasma turnover of iron is ϳ10 -20 mg a day and one source of this pool is iron released from the reticuloendothelial system (RES). 1 Macrophages of the RES release iron from phagocytosed erythrocytes and return it to the circulation for reuse by the erythroid compartment of the body. Difficulties in environmental iron acquisition limit the growth of microorganisms and some of the major virulence factors associated with bacterial infections are genes encoding more efficient means for acquiring iron from the host. One defense against invading microorganisms involves the sequestration of iron in body compartments not readily accessible to these invaders. The acute phase response to infection is characterized by a number of changes in iron metabolism including acute declines in serum iron, increases in the rate of serum iron disappearance, a decline in serum iron turnover, sequestration of the metal in the RES, and a decline in intestinal iron absorption (reviewed in Refs. 1 and 2). Chronic inflammatory states are also characterized by low serum iron levels, RES iron sequestration, and anemia (3,4,5). A number of cytokines including interleukin 1 (IL-1) (6 -11), tumor necrosis factor-␣ (TNF-␣) (11)(12)(13)(14), and IL-6 (16 -18) have been demonstrated to be potentially involved in changes in iron metabolism in animal and human models of acute and chronic inflammation.
During inflammation serum iron levels drop secondary to an increase in the rate of iron clearance from plasma, a decrease in iron mobilization from the RES compartment (the so-called iron exit block) (19 -25), and a decrease in iron absorption (23, 26 -28). The recent identification and characterization of the duodenal epithelial cell basolateral iron exporter solute carrier family 11a member 3 (SLC11a3), also known as IREG1 (29), ferroportin 1 (30), MTP1 (31) as an iron-regulated protein that is also expressed in the RES, has led to the hypothesis that MTP1 may be the protein responsible for iron export from this compartment. There is ample evidence that MTP1 exports iron from cells. Overexpression of MTP1, by transient transfection, in tissue culture cells has been demonstrated to result in depletion of intracellular ferritin and cytosolic iron levels (31) and MTP1 expression in frog oocytes results in measurable iron efflux (29,30). In this paper we report that MTP1 expression in the cells of the RES is regulated by acute inflammation. This inflammation-mediated control of MTP1 expression in the RES may be one component responsible for iron sequestration in the RES in both acute and chronic inflammatory states.
Mice-Unless otherwise indicated C57/BL6 mice aged 8 -12 weeks of either sex were used for the experiments described. LPS, PDTC, and all cytokines were prepared in a 100-l volume of saline and given intraperitoneally. Iron-dextran was given in 50 -100 l of saline in one or two thigh muscles. Turpentine was administered to anesthetized animals as a 50-100-l subcutaneous injection into the back between the scapulas. Unless noted otherwise, a minimum of 4 animals were used per experimental condition. Low iron mouse chow was purchased from Harlan Sprague-Dawley (Indianapolis, IN).
Experimental Infection-Six-week-old male BALB/c mice were infected intravenously with 1 ϫ 10 6 Leishmania donovani (MHOM/S.D./ 001S-2D) amastigotes in Hanks' balanced salt solution as previously described (32). Age-matched control mice received the same volume of Hanks' balanced salt solution by the same route of inoculation. At 56 days post-infection, mice were killed and the spleen sections were prepared in paraffin. Samples were immunostained with anti-MTP1 antibody as described below.
Immunohistochemistry-Immunohistochemistry was performed on paraffin-embedded mouse organ sections using an affinity purified rabbit anti-MTP1 polyclonal antibody using the Envisionϩ (Dako Corp., Carpenteria, CA) staining kit with Vector VIP or AEC as chromogen (Vector Laboratories, Burlingame, CA) as described previously (31). Immunofluorescence was performed using the affinity purified rabbit anti-MTP1 polyclonal antibody at 4 g/ml, and a rat anti-mouse F4/80 antigen antibody (Serotec, Raleigh, NC) at 1:20. Secondary reagents were Alexa 488-labeled goat anti-rabbit IgG antibody and Alexa 594 goat anti-rat IgG antibody at 1:500 dilution. Sections used for immunofluorescence were fixed with a 50:50 mixture of acetone and methanol for 20 min at Ϫ20°C. Pictures were obtained on color slide film using an Olympus BX-60 fluorescence microscope.
Western Blotting-Liver and spleen lysates were prepared by homogenization of tissue in phosphate-buffered saline supplemented with 0.5% Triton X-100, 5 mM EDTA, 0.1 mg/ml 4-(2-aminoethyl)benzenesulfonyl fluoride, 10 g/ml leupeptin hydrochloride, and 2 g/ml peptostatin A. After Dounce homogenization of tissue pieces, the nuclei and insoluble debris was spun down in a microcentrifuge at 13,000 rpm for 10 min in the cold and the supernatant was removed. For the immunoprecipitation, 15 l of Protein A/G slurry (Calbiochem) was incubated for 1 h with 1 g of affinity purified rabbit anti-MTP1 polyclonal antibody or a similar amount of control normal rabbit IgG in phosphatebuffered saline, 0.5% Triton X-100 and washed extensively. Two hundred micrograms of total protein from the organ lysates was added to the Protein A/G slurry bound to the anti-MTP1 antibody and tumbled for 2 h in the cold. Subsequently, the slurry was washed extensively with phosphate-buffered saline, 0.5% Triton X-100 and then boiled in SDS sample buffer prior to electrophoresis using the Laemmli gel system. The gel was transferred to nitrocellulose and Western blotting was done using the affinity purified rabbit anti-MTP1 polyclonal antibody or control IgG (400 ng/ml antibody concentration). Peptide blocking experiments were done by preincubating the diluted anti-MTP1 antibody with 2 g of specific peptide prior to addition of the antibody to the membrane. Horseradish peroxidase-labeled goat anti-rabbit antibody (Pierce) was used as a secondary reagent at 1:2000 -1:5000 dilution and signal detected on radiographic film using commercially available chemluminescent substrates.
Cell Culture and RT-PCR Analysis-Total splenocytes from C57/Bl6 mice were prepared by manual crushing of the spleen using the back of a 10-ml syringe in 3 ml of Dulbecco's modified Eagle's medium containing 10% fetal bovine serum and antibiotics. The spleen cell suspension was filtered through a nylon mesh and the cells were resuspended at 2-5 million mononuclear cells per milliliter and plated into 12-well plates. An aliquot of the cell suspension was treated with ammonium sulfate to lyse the red blood cells to determine the mononuclear cell number. The cells were incubated in plastic dishes at 37°C for 2 h and then washed four times with complete Dulbecco's modified Eagle's medium to remove nonadherent cells. The plates were then returned to the incubator and 2 h later LPS was added at 5 g/ml concentration. PDTC was added at a concentration of 100 M and a goat anti-rat/ mouse TNF-␣ was used at 1 g/ml. TNF-␣ was added to cultures at 10 ng/ml. After overnight incubation, total RNA was isolated using a commercially available RNA isolation kit (RNA-4-PCR, Ambion. Onefifth of the RNA was transcribed into cDNA using a commercially available kit (Retroscript Kit, Ambion).
Quantitative PCR was performed using TaqMan polymerase with detection of SYBR Green fluorescence on an ABI Prism 7700 Sequence detector (PE Biosystems, Foster City, CA). MTP1 mRNA levels were normalized using the expression of GAPDH as a housekeeping gene. Relative quantitation of both MTP1 and GAPDH mRNA was based on standard curves prepared from serially diluted mouse mast cell cDNA. The following sense and antisense sequences were employed: mouse MTP1, sense, CTACCATTAGAAGGATTGACCAGCTA, antisense, ACTGGAGAACCAAATGTCATAATCTG; mouse GAPDH, sense, CAT-GGCCTTCCGTGTTCCTA, antisense, TGT CATCATACTTGGCAGGT-TTCT.
Serum Iron Measurements-Iron was reduced and released from serum transferrin by addition of an equal volume of 0.2 N HCl containing 0.05% ascorbic acid. Subsequently the sample was deproteinated by addition of an equal volume trichloroacetic acid (11%). The sample was then centrifuged and iron in the supernatant was measured using a commercially available ferrozine-based Total Iron Measurement Kit (Sigma).

MTP1 Is Expressed in Splenic, Liver, and Bone Marrow
Macrophage Cells-Previous work had indicated that MTP1 is primarily expressed in the Kupffer cells of the liver and red pulp of the spleen and had a distribution similar to that seen with the F4/80, a macrophage-specific cell surface antigen (31). The identification of MTP1 staining cells in the spleen, liver, and bone marrow as macrophages was confirmed by two-color double immunofluorescence staining using antibodies to MTP1 and F4/80 antigen in frozen liver, spleen, and bone marrow cell sections (Fig. 1). In the red pulp of the spleen, there was great overlap between distribution of MTP1 and F4/80 staining cells. Interestingly, there were F4/80 positive cells in the white pulp of the spleen that do not stain with the MTP1 antibody, indicating that not all macrophages express MTP1. In the liver, as reported previously, MTP1 immunostaining was apparent on the surface of hepatocytes and in Kupffer cells. The MTP1 immunofluorescence of the Kupffer cells co-localized to the immunostaining with the F4/80 antigen. In bone marrow cell cytospins, there were many MTP1 positive cells, and immunostaining also co-localized with F4/80 in most of the cells.
Inflammation Results in Down-regulation of MTP1 Expression in the Reticuloendothelial Cells-A murine LPS model of the acute phase reaction was used to examine the connection between MTP1 expression in the RES compartment of the body and inflammation. Experimental mice were treated with 100 g of LPS and changes in MTP1 expression were assessed 16 -18 h later ( Fig. 2A). Immunohistochemical staining using an anti-MTP1 antibody of sections of spleen from LPS-treated mice demonstrated diminished MTP1 staining of spleen macrophages compared with control mice. To study the response of Kupffer cell MTP1 expression to LPS administration, mice were treated with 1-2 mg of iron-dextran to induce Kupffer cell MTP1 expression (31) and 7-10 days later treated with LPS. In iron-treated mice, LPS injection resulted in down-regulation of MTP1 in the Kupffer cells. Additionally, MTP1 was also downregulated in bone marrow cells with LPS treatment. Last, duodenal MTP1 expression was induced with feeding of a low iron diet to mice and subsequent LPS administration to these iron-deficient mice also resulted in down-regulation of duodenal MTP1 expression compared with controls.
The regulation of MTP1 in the spleen in a mouse model of a more chronic infectious disease was examined. The spleens of mice infected with L. donovani for 8 weeks were examined for MTP1 expression by immunohistochemistry using paraffin-embedded tissue and the anti-MTP1 antibody. Infected mice demonstrated diminished MTP1 staining in the spleen in comparison to control mice (Fig. 2B). Tissues from two infected and two control animals were examined. These data demonstrate that MTP1 was also down-regulated in a more chronic inflammatory state induced by an infection.
Western blotting of liver and spleen samples from LPStreated and control mice confirmed the decreased tissue expression of MTP1 secondary to LPS administration (Fig. 3). The down-regulation of MTP1 expression by LPS was apparent in Western blots of immunoprecipitated MTP1 from spleen (Fig. 3, left panel) and liver lysates (Fig. 3, right panel). In most, but not all experiments at least two or more distinct bands were apparent in the immunoblots of spleen and liver lysates. The first was a smear occurring at 60,000 -65,000 and the second was of higher molecular weight appearing at 110,000 -130,000. The appearance of the 60,000 -65,000 and the higher molecular weight bands were specific to immunoprecipitation with the anti-MTP1 antibody and blotting with the anti-MTP1 antibody. The peptide used to purify the anti-MTP1 antiserum blocked the appearance of the bands. The higher molecular weight bands were present in most experiments and the ratio of these bands to the 60,000 -65,000-band varied from experiment to experiment. The 60 -65-kDa-band is most likely MTP1. Preliminary data indicates that the larger bands do not represent either glycosylation or ubiquitination intermediates of MTP1. Anti-ubiquitin antibodies do not recognize the higher molecular weight bands by Western blotting and treatment with N-linked carbohydrate endoglycosidases does not alter the mobility of these higher molecular weight bands (data not shown). The larger bands probably represent a fraction of MTP1 that is modified, aggregated, or exhibits an aberrant migration.
In LPS-treated mice, hypoferraemia was induced more rapidly than changes in splenic MTP1 protein expression. Serum iron was diminished as early as 2 h after LPS administration (Fig. 4A); whereas, there was no change in MTP1 expression at this time point (Fig. 4B). Diminished MTP1 expression was noted at 6 h after LPS administration and this declined persisted as long as 72 h (data not shown). These data indicate that the early hypoferraemia because of LPS administration precedes MTP1 protein down-regulation and is not likely to be caused by changes in MTP1 expression. Other mechanisms, such as an increased clearance of blood iron may be more important in the development of the initial hypoferraemia.
Toll-like Receptor 4 Signaling Is Required for MTP1 Downregulation by LPS-Mice lacking a functional LPS receptor, the Toll-like receptor 4 (TLR4), such as the C3H/HeJ mouse strain are resistant to the lethal effects of LPS and fail to mount an acute phase response to LPS challenge and do not demonstrate a decline in serum iron (33,34). Although C3H/HeJ mice are resistant to induction of the acute phase response with LPS, they are known to be sensitive to other inflammatory mediators. We examined the response of splenic MTP1 to LPS in the C3H/HeJ mouse strain (Fig. 5). LPS-mediated down-regulation of MTP1 in the spleen was abrogated in the C3H/HeJ compared with the LPS-sensitive C3H/HeJ-FeB strain. A turpentine injection model of acute inflammation (23,27,35) was used to determine whether MTP1 down-regulation was specific for TLR4 or could be achieved by other stimuli in the C3H/HeJ mice. Turpentine injection of C3H/HeJ mice results in sterile inflammation and the response has been associated with a decline in serum iron and induction of the acute phase response (33,34). C3H/HeJ mice treated with turpentine demonstrated marked down-regulation of MTP1 expression in the spleen compared with untreated controls. These results indicate that inflammatory stimuli other than LPS can induce down-regulation of MTP1 in the spleen by a non-TLR4 dependent mechanism.

. Down-regulation of MTP1 in mouse spleen, liver, and duodenum is induced by LPS administration and chronic infection.
A, immunohistochemistry using the anti-MTP1 antibody was performed as outlined above on paraffin-embedded tissue sections of spleen, liver, bone marrow, and duodenum of mice treated with 100 g of LPS 18 h prior to sacrifice. Mice used for liver immunohistochemistry staining were treated with an intramuscular injection of 1-2 mg of iron-dextran 1 week prior to LPS treatment. Mice used for duodenal sections were given an iron-free diet for the month preceding LPS injection. B, paraffin-embedded spleens of mice infected with L. donovani and control animals were analyzed for MTP1 content by immunohistochemistry using an anti-MTP1 antibody.
FIG. 3. Immunoblot analysis of MTP1 protein expression in LPS-treated mouse spleen and liver. Representative immunoprecipitation followed by immunoblotting using an anti-MTP1 antibody of spleen (left panel) and liver lysates (right panel) prepared from control and LPS-treated mice was done as described. The band on the gel at 45-50 kDa is IgG heavy chain. Data shown is for two mice for each condition.
spleens was assessed in C57/Bl6 mice administered TNF-␣. Intraperitoneal injection of mice with 1 g of rTNF-␣ had no effect on MTP1 expression assessed by IHC on spleen sections using an anti-MTP1 antibody (Fig. 6). In addition, the response of MTP1 in the spleen to LPS was assessed in B6.129-Tnfrsf1a tm1Mak mice, which lack expression of TNF-␣ receptor type 1a and are hyporesponsive to TNF-␣ stimulation (36). The mice that were treated with LPS and a neutralizing anti-TNF-␣ antibody responded to LPS by down-regulation of splenic MTP1 indicating that TNF-␣ stimulation is not required for this effect (Fig. 6). Furthermore, administration of PDTC, a well characterized inhibitor of NF-B activation and TNF-␣ production (37), to mice prior to treatment with LPS did not alter the change in MTP1 expression induced by LPS (data not shown).
MTP1 mRNA in Adherent Spleen Cell Fractions Is Down-regulated by LPS in Vitro-To determine the mechanism of regulation of MTP1 expression by LPS, spleen and liver MTP1 mRNA expression was examined by real-time RT-PCR of total RNA from animals treated with LPS and untreated controls. There was a 2-3-fold down-regulation of MTP1-specific PCR product in spleen and liver samples from LPS-treated animals compared with control mice (Fig. 7A). To better determine whether there may be a change in macrophage-specific MTP1 mRNA secondary to LPS treatment (as opposed to total spleen MTP1 mRNA), splenic macrophages were enriched by adherence to plastic and these cells were treated with LPS, PDTC, or rTNF-␣. Total RNA was isolated from these cells and MTP1 and GAPDH mRNA levels were assessed using RT-PCR. In vitro LPS treatment resulted in down-regulation of MTP1 mRNA expression relative to GAPDH expression in the splenic adherent cells (Fig. 7B). The addition of PDTC, an inhibitor of NF-〉, did not abrogate the effect of LPS. In addition, direct addition of TNF-␣ to the adherent cells did not result in downregulation of MTP1 mRNA. The data indicate that LPS results in down-regulation of MTP1 mRNA in adherent mouse spleen cells and that this action of LPS is probably independent of NF-〉 activation and TNF-␣ synthesis.

DISCUSSION
MTP1 is a metal transporter that exports iron from the cytosol to the outside of cells and was initially identified as the duodenal epithelial basolateral iron transporter (29 -31). MTP1 has also been demonstrated to export iron when expressed in tissue culture cells and Xenopus oocytes. In addition, there is genetic evidence that MTP1 is involved in iron export from the yolk sac of zebrafish embryos. The recent identification of MTP1 mutation leading to hemochromatosis in man adds further weight to the hypothesis that MTP1 is involved in iron homeostasis (38,39). RES cells are responsible for the recycling of iron from the breakdown of heme from senescent erythrocytes and MTP1 has been hypothesized to be the key iron exporter in these cells. Supporting this hypothesis is the observation that MTP1 is expressed in the RES macrophages of the spleen (29 -31), Kupffer cells (29 -31), bone marrow, and FIG. 4. MTP1 down-regulation by LPS is delayed relative to development of hypoferraemia. Mice were treated with LPS for the times indicated and serum iron levels (A) and Western blotting (B) of splenic MTP1 protein levels were assayed as indicated previously. Data shown is for three mice for each time period.

FIG. 5. Spleen expression of MTP1 in C3H/HeJ mice is downregulated turpentine but not by LPS.
Paraffin-embedded spleen sections from C3H/HeJ-FeB and C3H/HeJ mice administered 100 g of LPS and C3H/HeJ mice administered turpentine, and untreated controls were immunostained with anti-MTP1 antibody (4 g/ml).

FIG. 6. LPS-mediated MTP1 regulation is independent of TNF-␣ signaling.
A, C57/Bl6 mice were injected with TNF-␣ (1 g/ mouse), LPS (100 g/mouse), or untreated and 18 h later spleens were collected and embedded in paraffin. B, B6.129-Tnfrsf1a tm1Mak mice were treated with LPS (50 g/mouse) and control IgG (75 g/mouse) or LPS and anti-TNF-␣ antibody (75 g/mouse) or left untreated and 18 h later spleens were collected and embedded in paraffin. Tissues were assayed for MTP1 expression using immunohistochemistry with the anti-MTP1 antibody (4 g/ml). lymph node histiocytes. 2 Although there has been no direct demonstration of MTP1-mediated iron export from RES cells, the ability of MTP1 to export iron from other cell types and its involvement in iron export from the duodenal epithelial cells and the zebrafish yolk sac support the supposition that MTP1 may export iron derived from senescent erythrocytes from the RES cells and back into the blood for reutilization by the erythroid compartment.
Chronic and acute inflammation are well characterized conditions in which iron metabolism is altered. These changes in iron metabolism are characterized by a drop in serum iron, an increase in the rate of plasma iron disappearance, a decline in the rate of plasma iron turnover, RES cell iron sequestration, and hyperferritinemia (reviewed Refs. [1][2][3][4][5]. In vivo studies of animal models of acute inflammation secondary to LPS or turpentine administration have demonstrated several mechanisms for the decline in serum iron. LPS and turpentine induce an accelerated clearance of iron from blood, thought to be because of an increase in transferrin-dependent uptake of iron by hepatocytes and other cells (13, 14, 23, 24, 40 -42). Release of lactoferrin by inflammatory cells may also be a contributing factor to this acute decline (43), but this hypothesis has been challenged more recently (7,9). Other changes in iron metabolism because of inflammatory stimuli include a decline in RES and hepatocyte cell iron turnover; and the decrease in RES cell iron turnover results in an accumulation of iron in the RES compartment (23-25, 40 -42, 44). The mechanism for the RES iron sequestration is not known, although an inflammationinduced increase in ferritin synthesis has been hypothesized to play a role (35). A decrease in intestinal iron absorption as a consequence of LPS or turpentine administration has also been demonstrated suggesting another possible mechanism for the more chronic effects of inflammation on iron metabolism (23, 26 -28).
The data presented here demonstrate that acute LPS administration to mice results in down-regulation of MTP1 expression in cells of the RES in the liver, marrow, and spleen and in duodenal epithelial cells. Western blotting of total liver and spleen lysates confirmed the decline in MTP1 protein expression. Furthermore, in an infectious model of chronic visceral leishmaniases, MTP1 was also down-regulated in the infected spleen. The effects of LPS on MTP1 expression are specific for the RES cells and a significant change in staining of blood vessels or kidney cells in response to acute LPS administration was not observed (data not shown). These observations support a hypothesis that the RES cell iron exit block of acute and/or chronic inflammation result from down-regulation of MTP1. The down-regulation of MTP1 in the spleen by LPS was not apparent at 2 h after injection, whereas hypoferraemia was present at this time. Other reports have also pointed to a fast onset of hypoferraemia with LPS (43,45). These observations make it unlikely that the LPS-mediated MTP1 down-regulation in the RES is responsible for the initial hypoferraemia observed with LPS. The initial hypoferraemia probably result from an increased rate of transferrin-mediated uptake of blood iron. It is more likely that MTP1-mediated RES cell iron exit block may serve to maintain the hypoferraemia rather than initiate it.
Many of the effects on iron metabolism of LPS are known to be secondary to LPS-induced synthesis of known proinflammatory mediators such as the interleukins and TNF-␣. Although it is difficult to draw many conclusions from the literature because of differences in animal models used, doses of cytokines, and the modes of administration among others, numerous reports indicate administration of IL-6 (17), IL-1 (6,7,14), and TNF-␣ (11,14,15) in humans and animals result in rapid drops in serum iron but this decline is relatively short-lived compared with that observed with LPS. Chronic administration of TNF-␣ (11,15) or IL-1 (10, 15) results in more prolonged hypoferraemia. In these studies, the major effects of IL-6 (17), IL-1 (7, 15), 2 D. J. Haile, unpublished observation.

FIG. 7.
Splenic and liver MTP1 mRNA levels decline in response to LPS treatment. A, MTP1 and GAPDH mRNA levels were assayed in total RNA from spleen and liver tissue of control or LPS-treated mice with MTP1 and GAPDH-specific primers using real time RT-PCR as indicated above. Data shown is compiled from four LPS and four control animals from two independent experiments. Ratios of MTP1 mRNA to GAPDH mRNA were calculated and the MTP1/GAPDH mRNA ratio of one control per experiment was set arbitrarily to 100%. Other values are the MTP1/GAPDH mRNA ratios normalized to this one control. B, adherent mouse splenocytes were isolated as indicated above and treated overnight with LPS (5 g/ml) or LPS and PDTC (100 M) or TNF-␣ (10 ng/ml). Total RNA from the treated and control cells was subjected to RT-PCR analysis using MTP1-specific and GAPDH housekeeping gene primers. Ratios of MTP1 mRNA to GAPDH mRNA are calculated and the MTP1/GAPDH mRNA ratio of one control per experiment was set arbitrarily to 100%. Other values are the MTP1/GAPDH mRNA ratios normalized to this one control. The data for the control and LPS animals is from 6 mouse spleens used in three separate experiments. PDTC and TNF-␣ data represent duplicates of an experiment in which adherent cells were derived from two pooled mouse spleens. * indicates results that have calculated p Ͻ 0.05 for differences. and TNF-␣ (15) on iron metabolism appear to be an increase in the rate of plasma iron disappearance and an increase in hepatocyte uptake of transferrin-iron (17) but there is no consistent data on changes in RES iron mobilization to these cytokines (7,15).
Treatment of mice with TNF-␣ did not result in an appreciable decline in MTP1 expression in mouse spleen at 24 h. Furthermore, mice lacking TNF-␣ receptor type 1a and pretreated with neutralizing anti-TNF-␣ antibodies responded to LPS with decreased MTP1 expression indicating that TNFR1a and possibly TNF-␣ signaling is not required for the LPS effect. Moreover, while in vitro treatment of mouse adherent splenocytes directly with LPS resulted in down-regulation of MTP1 mRNA, addition of TNF-␣ had no effect. Furthermore, PDTC did not block the in vitro effect of LPS on MTP1 mRNA downregulation. The lack of antagonism between LPS and PDTC in vitro and in vivo indicates that NF-B activation may not be required for MTP1 down-regulation. Signaling by TLR4 through the c-Jun NH 2 -terminal kinase, p38 MAPK , or c-Jun NH 2 -terminal kinase-dependent pathways may be more important in regulation of MTP1 by LPS.
This lack of effect of TNF-␣ on splenic MTP1 protein expression in vivo and the lack of effect of in vitro TNF-␣ on adherent splenocyte MTP1 mRNA production is not inconsistent with the observations that this cytokine induces hypoferraemia when administered in vivo. First, in our in vivo experiments, the cytokine was administered as a single injection and it has been characterized that the decline in serum iron induced by these cytokines is not long-lived and is usually reversed within 24 h. In addition, RES iron sequestration is only one of several mechanisms responsible for the hypoferraemia of inflammation. TNF-␣ may induce increased uptake of iron from blood without altering the RES iron sequestration. Therefore it would not be unexpected that serum iron may be altered but MTP1 expression in the RES may not change in response to TNF-␣.
A clue to the molecular signaling mechanism for MTP1 down-regulation by LPS comes for the study of the C3H/HeJ mouse, which lacks a functional TLR4 receptor. The LPS receptor, TLR4, is required to mount an acute phase response including hypoferraemia to LPS. Similarly, we do not observe a down-regulation of MTP1 in response to LPS in the C3H/HeJ mice. This finding indicates that the down-regulation of MTP1 by LPS requires signaling through TLR4. Despite resistance to LPS, the C3H/HeJ mice are responsive to other mediators of inflammation and we observed that treatment of C3H/HeJ mice with turpentine for a period of 24 h resulted in downregulation of MTP1 expression in the spleen. Therefore, there are both TLR4 and non-TLR4-dependent mechanisms for the down-regulation of MTP1 by inflammatory stimuli.
The anemia of chronic diseases is the cause of much morbidity in patients with a variety of illnesses. In addition to direct effects on erythropoiesis, chronic inflammation also perturbs normal body iron metabolism and results in RES cell iron sequestration. Abnormalities in iron kinetics have been observed in many disease states including cancers, rheumatologic conditions, chronic renal failure, and acute and chronic infectious diseases. This work is significant because the data suggest that strategies aimed at abrogating the down-regulation of MTP1 by inflammation may improve the supply of iron to the erythron and led to improvements in anemia in these patients. Furthermore, new evidence suggests that chronic inflammation may play a role in the pathogenesis of a variety of disease states such as chronic renal failure and heart disease. Downregulation of MTP1 by chronic inflammation may result in iron sequestration in tissues and predispose to iron-dependent oxidative stresses.