点击显示 收起
1 From the Departments of Nutrition (W-IL and BL) and Internal Medicine (BL), University of California, Davis
2 Supported by intramural faculty research grants (BL). 3 Reprints not available. Address reprint requests to B Lönnerdal, Department of Nutrition, University of California, One Shields Avenue, Davis, CA 95616. E-mail: bllonnerdal{at}ucdavis.edu.
ABSTRACT
Background: Mechanisms regulating iron transfer from maternal circulation into milk are yet unknown. Whether intestinal iron transporters, divalent metal transporter 1 (DMT1) and ferroportin 1 (FPN1), are present in the mammary gland and are involved in iron transfer into milk are unknown.
Objective: The objective was to examine DMT1 and FPN1 in rat mammary gland at different stages of lactation and to evaluate the effects of maternal iron status.
Design: Rats were fed either 35 mg Fe (control rats) or 8 mg Fe (low-iron rats) per kg diet for 3 wk and were fed the same diet throughout pregnancy and lactation. Mammary gland DMT1, FPN1, transferrin receptor, and ferritin were examined in control rats on days 1, 5, 10, and 20 of lactation and in low-iron rats on days 10 and 20 of lactation. Tissue and milk iron were measured.
Results: Milk iron, DMT1, and FPN1 decreased throughout lactation. Iron status was compromised in low-iron rats, whereas milk iron was maintained. On day 10 of lactation, mammary gland iron and ferritin were lower in the low-iron rats. DMT1, FPN1, and transferrin receptor values were unchanged; however, a smaller-size DMT1 protein was observed in the low-iron rats. On day 20, transferrin receptor increased in the low-iron rats, whereas mammary gland iron, ferritin, DMT1, and FPN1 were unchanged.
Conclusions: The results show that DMT1 and FPN1 concentrations are higher during early lactation and are possibly involved in iron transfer into milk. Mammary gland regulation of DMT and FPN1 during low iron status appears to be different from that in the intestine.
Key Words: Rats • divalent metal transporter 1 • ferroportin 1 • milk iron • transferrin receptor • ferritin • mammary gland
INTRODUCTION
The important role of iron in the development of the neonate has been shown in several species. Because milk is the only food newborn mammals consume, the capacity of milk to provide adequate iron can be critical. Iron concentrations in both human and rat milk decrease during the course of lactation (1–4). A 30% decrease in the iron concentration of human milk was reported during the first month of lactation (5). The mechanisms responsible for the decrease in milk iron are not clear.
Little correlation has been found between maternal iron intake and milk iron content (4, 6–8). No correlation between iron intake and milk iron has been found, even in women with an iron intake of >200% of the recommended dietary allowance (9). Ethiopian women, who commonly have daily iron intakes of >200 mg, had milk iron concentrations similar to those of Swedish women, whose daily iron intake was 10–15 mg (6). Attempts to increase milk iron through supplementing mothers with iron during lactation have failed (5). Maternal iron reserves also do not seem to play a significant role in milk iron concentrations. Several studies have not found milk iron to be correlated with the number of pregnancies (5, 10, 11), despite the fact that iron stores decrease with an increase in the number of pregnancies (10). Maternal iron deficiency accompanied by a serum iron concentration of 0.34 mg/L compared with a maternal serum iron concentration of 2.35 mg/L in control subjects resulted in comparable milk iron concentrations (12). A large increase in plasma iron produced by administration of iron did not result in an increased milk iron concentration (13). Taken together, most of the evidence suggests that iron uptake into the mammary gland and iron transfer into milk are regulated.
Transferrin receptors (TfRs) have been identified and characterized in the mammary tissues of lactating rats (14, 15) and are believed to transport iron into the mammary tissue via the classic TfR-mediated endocytotic pathway. However, no correlation between milk iron and TfRs was found, which suggests that the control of milk iron lies after iron entry into the mammary gland (16). Recently, 2 intestinal iron transporters—divalent metal transporter 1 (DMT1) and ferroportin1 (FPN1)—were identified (17–21). Whether these iron transporters are present in the mammary gland and are involved in iron transfer from the mammary gland to milk have not been investigated. The purpose of this study, therefore, was to examine at a molecular level the regulation of iron transfer into milk. We examined the expression of DMT1, FPN1, TfR, and ferritin during various stages of lactation and the effects of low maternal iron status on their expression in rat mammary gland.
MATERIALS AND METHODS
Diet
Rats were fed an egg white–based semipurified experimental diet. The composition of the experimental diet is shown in Table 1. The control diet and the low-iron diet differed only in iron content; the control diet contained 35 mg Fe/kg diet, and the low-iron diet contained 8 mg Fe/kg diet as ferrous sulfate. Iron concentrations were verified by atomic absorption spectrophotometry (Model Smith-Heifjie 4000; Thermo Jarrell Ash, Franklin, MA) before the study began. The composition of the mineral and vitamin mixes is shown in Table 1.
View this table:
TABLE 1. Composition of the semipurified diet1
Animals
Female virgin Sprague-Dawley rats (n = 39; body weight: 250 g) were purchased (Charles River, Wilmington, MA) and housed individually in stainless steel hanging cages. Animals were kept in a temperature-, humidity-, and light-controlled room (20 °C, humidity >60%, and a 12-h light-dark cycle) and given access to deionized water ad libitum. Animals were fed a control diet for 1 wk and were then randomly fed either the control diet or the low-iron diet for another 3 wk. Rats were then bred and kept on the same diet throughout pregnancy and lactation. On postnatal day 1, litters were culled to 10. Control rats (n = 6 or 7 per group) on either days 1, 5, 10, or 20 of lactation and low-iron rats (n = 6 or 7 per group) on days 10 or 20 of lactation were separated from their pups for 2 h and were then anesthetized with an intraperitoneal injection of a xylazine-ketamine cocktail (0.25 mL/300 g rat). The cocktail contained 1.6 mg xylazine/kg and 33 mg ketamine/kg. Animals were deeply anesthetized as determined by a lack of reaction to a hard-toe pinch. A dose of oxytocin (10 U/kg) was subcutaneously administered to stimulate milk let-down, and 0.5–1.0 mL milk was collected manually. Dams were always milked between 1200 and 1400 to minimize possible diurnal variations in milk composition (22). After the collection of milk, dams were killed by asphyxiation with carbon dioxide, and blood was collected by cardiac puncture. Samples of mammary glands, livers, and duodenums were dissected. This study was approved by the Animal Research Services at the University of California Davis, which is accredited by the American Association for the Accreditation of Laboratory Animal Care.
Hemoglobin, body weight, and food intake
Whole-blood hemoglobin was analyzed with the cyanomethemoglobin method (Sigma, St Louis). Twenty microliters of each blood sample was mixed with 5 mL Drabkin's solution (0.1% sodium bicarbonate, 0.005% potassium cyanide, and 0.02% potassium ferricyanide) for hemoglobin determination. Food intakes were recorded every other day, and weights were recorded weekly throughout the experiment.
RNA extraction
Mammary glands (100 mg/mL) were dissected and homogenized in TRIzol reagent (Life Technologies, Rockville, MD) immediately after removal from the animals. Homogenates were kept at –80 °C until extraction. Total RNA was extracted from the whole tissues according to the TRIzol protocol.
Real-time quantitative reverse transcriptase–polymerase chain reaction
The relative expressions of DMT1 and FPN1 messenger RNA were determined by real-time semiquantitative reverse transcriptase–polymerase chain reaction (RT-PCR) with the use of TaqMan EZ RT-PCR Core Reagents (Applied Biosystems, Foster City, CA). Primers and probes were designed by using PRIMER EXPRESS software (Applied Biosystems) to span introns to avoid co-amplification of genomic DNA. For quantification of DMT1 cDNA, the following primers were used: 5-GTT TGT CAT GGA GGG ATT CCT-3 and 5-CAT TCA TCC CTG TCA GAT GCT-3, which recognized both the iron-responsive element (IRE) and the non-IRE forms. For quantification of FPN1 cDNA, the following primers were used: 5-GTG CCT CCC AGA TCG CAG-3 and 5-GGG CTG GTT ATA GTA GGA GAC CC-3. The probes for DMT1 (5-AAA ATG GTC GCG CTT TGC CCG A-3) and FPN1 (5-ACC CTT CCG CAC TTT TCG AGA TGG A-3) were 5-labeled with 6-carboxyfluorescein and 3-labeled with 6-carboxytetramethylrhodamine.
The RT-PCR reactions were carried out according to the manufacturer's protocol on an ABI Prism 7700 Sequence Detector (Applied Biosystems). RT-PCR conditions were 30 min at 60 °C and 10 min at 95 °C followed by 40 cycles of 15 s at 95 °C and 1 min at 60 °C. Analysis of the data were done by using the SEQUENCE DETECTOR 1.7 software. All reactions were performed in duplicate, and the change in Rn and CT were calculated from fluorescence activity data collected during PCR. A no-template control was included in every reaction. The housekeeping gene 18 S ribosomal RNA was used for internal normalization with the use of TaqMan ribosomal RNA control reagent (Applied Biosystems). Identical results were obtained by RT-PCR analysis of 18S in each group; therefore, 18S was selected as the reference for all experiments.
Membrane and soluble-protein preparation
Mammary glands were dissected and snap-frozen in liquid nitrogen once the tissues were removed from the animals. Tissues were then kept at –80 °C until further processing. Frozen mammary glands (300 mg) were then homogenized in 5 mL HEPES-EDTA buffer [20 mmol HEPES/L, pH 7.4; 1 mmol EDTA/L; 250 mmol sucrose/L; and a protease inhibitor mixture containing 4-(2-aminoethyl)benzenesulfonyl fluoride, trans-epoxysuccinyl-L-leucyl-smido(4-guanidino)butane, bestatin, leupeptin, aprotinin, and sodium EDTA (Sigma)]. The homogenate was centrifuged at 500 x g for 5 min at 4 °C. The supernatant fluid was then centrifuged at 100 000 x g for 30 min at 4 °C. The supernatant fluid (soluble protein) was collected for ferritin protein determination, and the crude membrane fraction (pellet) was resuspended in 0.5 mL homogenization buffer and stored at –80 °C for DMT1, FPN1, and TfR protein determination. Protein concentrations of the membrane fractions and the soluble fractions were quantified by the Lowry method (23).
Production of DMT1 and FPN1 antibodies
Peptide fragments of DMT1 (VKPSQSQVLKGMFV) and FPN1 (KQLTSPKDTEPKPLEGTH) were synthesized with an additional cysteine residue for conjugation to keyhole limpet hemocyanin at the C-terminal end (Genemed Synthesis, Inc, South San Francisco). Keyhole limpet hemocyanin–conjugated peptides were injected into New Zealand White rabbits (1 mg peptide/rabbit) for polyclonal antibody production. Peptide sequences were verified by amino acid analysis and mass spectroscopy. DMT1 peptide was synthesized according to the predicted amino acid sequence obtained from the human DMT1 complementary DNA (cDNA) corresponding to amino acids 235–248 of the protein in the presumed fourth external loop between putative transmembrane regions 5 and 6. Because this region is common to both the IRE and the non-IRE forms of DMT1, it will react with both types of DMT1. The amino acid sequence of this region is highly conserved in mammals with little sequence homology to Nramp1. This peptide is in an extracellular region of the protein away from the glycosylation sites and with only 6 amino acids in common with human Nramp1. The FPN1 peptide was synthesized according to the predicted amino acid sequence obtained from the rat FPN1 cDNA corresponding to amino acid 253–270 of the protein in the external loop between putative transmembrane regions 5 and 6. Comparison of both peptide sequences with public sequences databases using the BLAST identified only DMT1 and FPN1 sequences.
Specificity of the antibodies was verified by peptide competition analysis. Briefly, membrane protein from the mammary gland was resolved and transferred as described below. After blocking, the blots were incubated with DMT1 or FPN1 antibodies in the presence or absence of the corresponding peptides for 1 h. Peptides used were in 25-fold excess over the antibodies used. The blots were visualized by using enhanced chemiluminescence after incubation with secondary antibody (Figure 1).
FIGURE 1.. Verification of the specificity of divalent metal transporter 1 (DMT1) and ferroportin 1 (FPN1) antibodies, determined by peptide competition analysis. Membrane proteins from mammary gland tissue were incubated in the absence or presence of the corresponding peptides. The peptides used were in 25-fold excess over the antibodies used. Mammary gland tissue from the control animals was immunostained on day 10 of lactation. A: Western blot. Sections were incubated with preimmune rabbit serum (1:300, by vol; B) as negative control, DMT1 antibody (1:300, by vol; C), and FPN1 antibody (1:300, by vol; D). Sections were visualized at 100x magnification.
Purification of antibodies
Immunized rabbit sera were purified by using the immunizing peptides. A 2-mL volume of Affi-Gel 10 for DMT1 peptide (pI 9.3) and Affi-Gel 15 for FPN1 peptide (pI 6.3) were transferred to two 1-mL columns (Bio-Rad, Hercules, CA) and washed with coupling buffer (0.1 mol NaHCO3/L, pH 8.3). The immunizing peptides (1 mg/mL) were added to the Affi-Gel columns, which were then incubated overnight on a rotating mixer at 4 °C. Coupling buffer was added to the columns and was collected to check coupling efficiency. The columns were then washed with 10 mL coupling buffer, and 0.1 mol glycine-HCl/L (pH 2.5, elution buffer) was added. The affinity columns were then equilibrated with phosphate-buffered saline (PBS). A 2-mL volume of the antisera was applied to the affinity columns and allowed to run through into collection tubes. The columns were washed with 10 mL PBS Tween containing 0.5 mol NaCl/L. The purified anti-DMT1 immunoglobulin (Ig) G and anti-FPN1 IgG were eluted with elution buffer, and the pH was adjusted to 7 and sodium azide was added.
Western blotting
Western blots were performed to determine DMT1, FPN1, TfR, and ferritin protein concentrations in the rat mammary gland. Protein samples extracted from the same animals were used for all Western blot analysis. Mammary gland membrane protein fractions or the soluble protein fractions (50 µg) were solubilized in Laemmli buffer, boiled for 5 min, and separated by 8% (DMT1, FPN1, and TfR) and 12% (ferritin) sodium dodecyl sulfate–polyacrylamide gel electrophoresis, respectively. Similar loading and transfer of proteins was verified by staining the blots with Ponceau S Red solution (Sigma). Proteins were transferred to nitrocellulose membranes by electroblotting, which were then blocked by 5% nonfat powdered milk in PBS with Tween 20 (PBST) at room temperature for 2 h and at 4 °C overnight. The membranes were then washed in several changes of PBST and incubated for 1 h at room temperature in either affinity-purified DMT1 (1:1000), affinity-purified FPN1 (1:750), mouse anti-human ferritin antibody (1:2500) (Alpha Diagnostic, San Antonio, TX), or mouse anti-rat TfR antibody (1:1000) (Pharmingen, San Diego) in PBST. The membranes were washed again with several changes of PBST, incubated with donkey anti-rabbit Ig, peroxidase-linked species-specific whole antibody (Amersham, Buckinghamshire, United Kingdom) for DMT1 and FPN1, or peroxidase-conjugated rabbit anti-mouse Ig (Dako, Copenhagen) for ferritin and TfR for 1 h at room temperature. Membranes were washed again with several changes of PBST. The immunologically detected proteins were visualized by using enhanced chemiluminescence (Amersham) or Super Signal Femto substrate (Pierce, Rockford, IL). Processed blots were exposed to X-ray film for the optimum exposure time and quantified by using the Chemi-doc Gel Quantification System (Bio-Rad, Hercules, CA).
Immunohistochemistry
Mammary glands of the control animals on day 10 of lactation were fixed in cacodylic acid–formaldehyde buffer and were then dehydrated and embedded in paraffin. Sections (4 µm) were cut and mounted on glass slides. The endogenous peroxidase activity of the tissues was blocked by using 0.3% H2O2 in methanol for 30 min. The slides were then rinsed in PBS, and nonspecific binding of primary antibody was avoided by blocking sections with 10% normal goat serum (Vector Laboratories Inc, Burlingame, CA). To detect DMT1 and FPN1, slides were incubated for 1 h at room temperature with either affinity-purified DMT1 antibody (1:300) or affinity-purified FPN1 antibody (1:300) diluted in blocking solution. After being washed in PBS, slides were incubated with biotinylated goat anti-rabbit IgG secondary antibody (Vector Laboratories Inc) for 1 h at room temperature. Slides were again washed in PBS and incubated for another 50 min with an avidin-biotin complex (ABC kit; Vector Laboratories Inc) with alkaline phosphatase as a reporter enzyme. Staining was visualized with the use of 3-diaminobenzidine tetrahydrochloride substrate working solution. The sections were counterstained with hematoxylin and rehydrated in ethanol and xylene, and the slides were mounted.
Tissues and milk iron analysis
Mammary glands, livers, and small intestines were dissected from the animals. Mammary glands were rinsed, and small intestines were perfused with fresh isotonic saline. Tissues and milk were wet-ashed in 16 mol HNO3/L for 7 d (24). Samples were then diluted with an appropriate amount of distilled water and analyzed for iron by atomic absorption spectrophotometry (model Smith-Heifjie 4000; Thermo Jarrell Ash, Franklin, MA).
Statistical analysis
Values are expressed as means ± SEMs. Statistical analysis was performed by using PRISM GRAPHPAD version 3.02 (GraphPad Software, San Diego). Student's unpaired t test was used to compare 2 data sets, and one-way analysis of variance (ANOVA) was used to compare multiple data sets. When the P value obtained from ANOVA was significant, Tukey's test was applied to test for differences among groups. Significance was considered to be P < 0.05. When the variances within the group differed significantly (P < 0.05) according to Bartlett's test, data were log transformed before being tested by one-way ANOVA.
RESULTS
Food intake and body weight
Food intake and body weight of the control and low-iron rats were not significantly different throughout the experiment (data not shown).
Hemoglobin, tissue, and milk iron concentration throughout lactation
Hemoglobin was significantly higher day 20 of lactation than on day 1, and liver iron was significantly higher day 10 and day 20 of lactation than on day 1 in the control subjects (Table 2). Mammary gland iron concentrations remained constant throughout lactation, whereas the milk iron concentration was significantly higher on day 1 and decreased during the course of lactation (Table 2).
View this table:
TABLE 2. Hemoglobin and milk and tissue iron concentrations in the control and low-iron rats by day of lactation1
Mammary gland DMT1 and FPN1 throughout lactation
The gene expression of both DMT1 and FPN1 was significantly higher during early lactation and declined throughout lactation (Figure 2). DMT1 and FPN1 protein concentrations also decreased throughout lactation. Western blot analysis showed bands at 144, 60, and 54 kDa for DMT1, with the 60-kDa band having the strongest intensity. Both the 144- and 60-kDa bands decreased on day 10 and day 20 of lactation, whereas the 54-kDa band remained constant throughout lactation. For FPN1, a single band with a molecular mass of 140 kDa was observed; the FPN1 protein concentration was also higher during early lactation and decreased throughout lactation.
FIGURE 2.. A: Mean (±SEM) relative gene expression of mammary gland divalent metal transporter 1 (DMT1) and ferroportin 1 (FPN1) in the control rats on days 1, 5, 10, and 20 of lactation, determined by real-time reverse transcriptase–polymerase chain reaction. n = 6 or 7 per group. Means with different lowercase letters are significantly different, P < 0.05 (one-way ANOVA and Tukey's test). B: Mean (±SEM) DMT1 and FPN1 protein concentrations in mammary gland tissue from control rats on days 1, 5, 10, and 20 of lactation by Western blot analysis. n = 3 per group. Means with different lowercase letters are significantly different, P < 0.05 (one-way ANOVA and Tukey's test); the comparison was made within the same molecular weight class.
Immunostaining of the mammary gland showed that DMT1 protein was seen in the mammary secretory epithelial cells lining the alveoli lumen but not in myoepithelial cells (contractile epithelial cells), which surrounded the secretory epithelial cells or adipocytes. The immunostaining showed that the protein was localized intracellularly in the secretory epithelial cell. FPN1 was also seen in the intracellular part of the mammary secretory epithelial cells (Figure 1).
Mammary gland TfR and ferritin protein throughout lactation
Both TfR and ferritin protein concentrations were highest on day 1 and lowest on day 20 (Figure 3).
FIGURE 3.. Mean (±SEM) transferrin receptor (TfR) and ferritin protein concentrations in mammary gland tissue from control rats on days 1, 5, 10, and 20 of lactation, determined by Western blots analysis. n = 3 or 4 per group. Means with different lowercase letters are significantly different, P < 0.05 (one-way ANOVA and Tukey's test).
Effect of maternal iron status on hemoglobin, tissue, and milk iron
At midlactation (day 10), the iron concentrations in the small intestines, livers, and mammary glands of the low-iron rats were significantly lower than those in the controls. Hemoglobin tended to be lower in the low-iron rats, but not significantly so, and there was no difference in milk iron (Table 2). At late lactation (day 20), small intestine and liver iron concentrations were significantly lower in the low-iron rats. Hemoglobin also tended to be lower in the low-iron rats, but not significantly so, and there was no significant difference in milk or mammary gland iron (Table 2).
Effect of maternal iron status on mammary gland DMT1 and FPN1
There was no significant difference in DMT1 gene expression between the control and the low-iron rats on day 10. On day 20, DMT1 gene expression was significantly higher in the low-iron rats than in the controls (Figure 4). The DMT1 protein concentration was not significantly different between the control and the low-iron rats on day 10 or day 20. However, besides the 60- and 54-kDa DMT1 proteins, an additional 39-kDa band was observed in the low-iron rats on day 10. FPN1 gene expression and the protein concentration were not significantly different between the low-iron rats and the controls on both days 10 and 20.
FIGURE 4.. Mean (±SEM) relative divalent metal transporter 1 (DMT1) and ferroportin 1 (FPN1) gene expression and protein concentrations in mammary gland tissue of control and low-iron rats on days 10 and 20 of lactation, determined by real-time reverse transcriptase–polymerase chain reaction and Western blot analysis, respectively. n = 6 or 7 per group. *Significantly different from the control rats, P < 0.05 (Student's unpaired t test).
Effect of maternal iron status on TfR and ferritin protein concentrations
The TfR protein concentration was not significantly different between the low-iron rats and the controls on day 10 and was 81% higher in the low-iron rats than in the controls on day 20 (Figure 5). The ferritin protein concentration was 48% lower in the low-iron rats than in the controls on day 10 and was not significantly different from controls on day 20 (Figure 5).
FIGURE 5.. Mean (±SEM) transferrin receptor (TfR) and ferritin protein concentrations in mammary gland tissue of control and low-iron rats on days 10 and 20 of lactation, determined by Western blots analysis. n = 6 per group. *Significantly different from the control rats, P < 0.0 (Student's unpaired t test).
DISCUSSION
Little is known about the mechanisms and regulation of iron transfer from the mammary gland into milk. Milk iron concentrations decrease throughout lactation (1–4) and do not seem to be affected by maternal iron intake and maternal iron status (4–13). In the current study, we showed that the iron transporters DMT1 and FPN1 are present in the rat mammary gland. Their expression is higher during early lactation and decreases throughout lactation. The decrease in DMT1 and FPN1 correlates with the normal decline in milk iron during lactation. The requirements for iron by the mammary gland can be higher during early lactation. Glandular development in rats begins in pregnancy and continues until day 5 of lactation (25). Both higher milk iron concentrations and higher requirements of iron by the gland may contribute to the need for expressing more DMT1 and FPN1 during early lactation.
The mechanisms regulating the decline in mammary gland DMT1 and FPN1 throughout lactation are not known. Although both mRNA and protein concentrations decrease over time, their decreases are not parallel, ie, there is a significant decrease in DMT1 mRNA from day 1 to day 5, whereas no significant difference was found in protein concentration from day 1 to day 5 (Figure 2). Both DMT1 and FPN1 possess an IRE (17, 18, 20, 21) and may be regulated by the IRE and iron regulatory protein (IRP) system (IRE/IRP). However, the unchanged mammary gland iron concentration throughout lactation and the decline in mammary gland DMT1 and FPN1 during lactation implies that the IRE/IRP system may not play a key role in regulating the normal decline of DMT1 and FPN1. Whether the IRE or the non-IRE DMT1 is the predominant form present in the mammary gland needs to be investigated. Examining the 5-splice variants of DMT1 (1A and 1B) may provide more insight into the regulation of this transporter in the mammary gland. Although a change in the rate of gene transcription cannot be excluded, it is possible that mRNA turnover is changed. Both DMT1 and FPN1 contain adenylate uridylate–rich elements (AREs) in their 3-untranslated region, and these motifs have been shown to enhance mRNA turnover (26). The newly compiled ARE database (ARED) shows that ARE mRNAs represent as much as 5–8% of human genes and encode functionally diverse proteins that are involved in many transient biological processes, such as cell growth and differentiation, transcriptional and translational control, hematopoiesis, and nutrient transport (27).
Hormonal changes during lactation may also regulate the decline in DMT1 and FPN1. Three different sizes of DMT1 protein were observed, with the 60-kDa band having the strongest intensity. The 60-kDa band is in agreement with the predicted size of DMT1 on the basis of its amino acid sequence. The 144- and 54-kDa DMT1 forms may be posttranslationally modified. Indeed, different sizes of DMT1 and FPN1 have been reported previously (28–33). The predicted size of FPN1 is 62 kDa from the amino acid sequence, and what we observed was 140 kDa. It is possible that FPN1 is modified and aggregated. Other researchers have reported 90–110 kDa for FPN1 in human bronchial epithelial cells (34), 60–65 kDa and 110–130 kDa for mouse liver and spleen (35), 80 and 110 kDa for rat brain (36), and 66 kDa for mouse and rat duodenum (17). The reason for the different sizes of FPN1 found in different tissues is not clear, but several investigators have suggested that the different sizes may be due to posttranslational modification, primarily because DMT1 glycosylation has been previously reported in the mouse.
Immunostaining showed that both DMT1 and FPN1 are localized intracellularly in the mammary epithelial cells. This led us to believe that DMT1 may be the iron transporter in the endosomal membrane and that it transports iron out of the endosome to the cytoplasm of the mammary epithelial cells. The intracellular localization of FPN1 suggests the possibility that FPN1 may be involved in the intracellular trafficking of iron between the cytosol and organelles. Apart from being expressed on the basal surface of duodenal enterocytes (21) and placental syncytiotrophoblasts (18), FPN1 is also expressed in the cytoplasm of Kupffer cells (17).
Maternal low iron intake resulted in decreased iron status at both mid- and late lactation, yet milk iron was maintained at the same concentration as in the control rats. This finding implies that there is regulation in the mammary gland that increases the efflux of iron into the milk of the low-iron rats. In most human studies, maternal iron status and maternal iron intake do not appear to affect milk iron concentrations, whereas milk iron can be affected in some animal studies (22, 37, 38). However, the level of iron supplementation used in those studies were very high, such as 240 µg/g Fe (22) or 2500 µg/g Fe (37), and the level of iron supplementation used in deficiency studies was very low (5 µg/g) (38). In some animal studies, repeated bleeding was used to induce iron deficiency (1). The level of dietary iron used in this study was intentionally not as low. It is possible that the extent of iron deficiency and excess needed to affect milk iron is quite pronounced and that such conditions are unlikely to occur in humans.
On day 10 of lactation, mammary gland iron was lower in the low-iron rats, which was accompanied by a decrease in mammary gland ferritin protein. However, we observed no increase in DMT1 or FPN1 gene and protein expression, as would be expected in the small intestine during iron deficiency. Other investigators also reported no change in FPN1 expression in the placenta of iron-deficient rats (29), which suggests different regulatory mechanisms than in the intestine; such mechanisms remain to be characterized. In addition, no significant change in transferrin receptor protein concentration was found. Interestingly, a 39-kDa DMT1 protein was observed in the low-iron rats, which may be triggered by the lower mammary gland iron. The function of this smaller-size DMT1 protein is not clear, but we speculate that it may be responsible for the increase in iron efflux into the milk, thereby maintaining milk iron concentrations in the low-iron rats. This 39-kDa DMT1 protein may act as an iron chaperone; however, further studies are needed to confirm its role in increasing iron efflux into the milk during low iron status.
During late lactation, milk output decreases as the rat pups begin to wean. Mammary gland iron on day 20 was not significantly different between the low-iron and the control rats. This was accompanied by unchanged ferritin protein concentrations. The maintained mammary gland iron in the low-iron rats is most likely due to 1) an increase in transferrin receptor in the low-iron rats, which, in turn, increases iron uptake into the mammary gland and hence maintains mammary gland iron, and 2) a lower milk output during late lactation and hence less iron efflux into the milk. FPN1 was similar in the low-iron rats and the controls. DMT1 gene expression increased significantly but the protein concentration was unchanged in the low-iron rats, possibly because of increased protein turnover and degradation. Even though there was an increase in DMT1 gene expression, this increase was very minor when compared with the level of DMT1 during early lactation. With no difference in mammary gland iron between the low-iron dams and the controls on day 20 of lactation, the 39-kDa DMT1 was not observed in the low-iron dams. The lower mammary gland iron of the low-iron dams on day 10 of lactation may be necessary to trigger the presence of this 39-kDa DMT1 protein. Studies that examine the specific subcellular localization of these transporters in mammary epithelial cells are needed to provide more information on the precise role of these transporters.
A likely scenario for the metabolic handling of iron in the mammary gland is as follows: iron bound to transferrin in the circulation is taken up by the transferrin receptor at the basal side of the mammary epithelial cell. The transferrin-transferrin receptor is endocytosed into an endosome, where iron is released from transferrin at low pH (39, 40). Iron is then transported out from the endosome to the cytoplasm via DMT1. The intracellular localization of DMT1 is consistent with its roles in endosomal ferrous iron transport. Once transported out of the endosome, ferrous iron may be oxidized by a ferroxidase such as ceruloplasmin (41, 42) or hephaestin, which allows iron to be incorporated into ferritin or bound to iron-transport proteins in ferric form. Iron may then be utilized by the oxidative phosphorylation cytochrome system because lactation is an energy-expensive process. Iron in the cytoplasm can also be transported into the Golgi and secretory vesicles by unknown mechanisms. FPN1, which is found to be intracellularly localized, may be involved in this intracellular trafficking of iron between the cytosol and organelles. Iron can then bind to iron-binding proteins, such as casein, lactoferrin and transferrin, which are present in the Golgi and are secreted by exocytosis. Iron in the cytoplasm may also be incorporated into iron-containing enzymes, such as xanthine oxidase, which are secreted with the milk fat globule.
In summary, we showed that both DMT1 and FPN1 are present in rat mammary epithelial cells and that their expression decreases throughout the course of lactation. The presence of DMT1 and FPN1 points to an involvement of these transporters in the transfer of iron from the mammary gland into milk. The response of DMT1 and FPN1 in the mammary gland to low iron status is different from that in the small intestine, which indicates a difference in the regulation of these transporters between tissues. Mammary gland DMT1 and FPN1 may not simply be up-regulated to increase iron efflux into milk during iron deficiency. Other factors or proteins may be involved in the regulation of iron transfer to milk during iron deficiency.
ACKNOWLEDGMENTS
W-IL and BL planned the research and W-IL conducted the experiments. Neither of the authors had a conflict of interest.
REFERENCES