The universal importance of iron, its high toxicity, and complex chemistry present a challenge to biological systems in general and to protected compartments in particular. The high mitotic rate and avid mitochondriogenesis of developing male germ cells imply high iron requirements. Yet access to germ cells is tightly regulated by the blood-testis barrier that protects the meiotic and postmeiotic germ cells. To elucidate how iron is supplied to developing male germ cells, we analyzed iron deposition and iron transport proteins in testes of mice with iron overload and with genetic ablation of the iron regulators Hfe and iron regulatory protein 2. Iron accumulated mainly around seminiferous tubules, and only small amounts localized within the seminiferous tubules. The localization and regulation of proteins involved in iron import, storage, and export such as transferrin, transferrin receptor, the divalent metal transporter-1, cytosolic ferritin, and ferroportin strongly support a model of a largely autonomous iron cycle within seminiferous tubules. We show evidence that ferritin secretion from Sertoli cells may play an important role in iron acquisition of primary spermatocytes. During spermatogenic development iron is carried along from primary spermatocytes to spermatids, and from spermatids iron is recycled to the apical compartment of Sertoli cells, which traffic it back to a new generation of spermatocytes. Losses are replenished by the peripheral circulation. Such an internal iron cycle essentially detaches the iron homeostasis within the seminiferous tubule from the periphery and protects developing germ cells from iron fluctuations. This model explains how compartmentalization can optimize cellular and systemic nutrient homeostasis.
- blood-testis barrier
- seminiferous tubule
- Sertoli cells
- ferritin secretion
- iron regulatory protein 2
iron is an essential nutrient that is toxic in excess (18). Therefore, iron transport and iron uptake systems are tightly regulated in all iron-dependent organisms (2, 5). In mammals, iron homeostasis is regulated at systemic and cellular levels through the peptide hormone hepcidin and the iron regulatory proteins (IRP), respectively. A third and very effective way of regulating nutrient homeostasis is by compartmentalization. This strategy of surrounding an organ with a cell barrier is used to protect the most delicate organs as well as cells from pathogen invasion and from peripheral nutritional imbalances and toxic materials (10, 31). Blood-tissue barriers are found in the brain, the retina, the testis, and the epididymis.
In the testis, spermatogenesis starts near the basal membrane of the seminiferous tubule (SFT), where spermatogonia either replicate to give rise to more undifferentiated germ cells or start the differentiation process and gradually move toward the lumen of the SFT. During maturation, germ cells pass through the dynamic tight junctions positioned between neighboring Sertoli cells (SC) from the apical to the luminal compartment of the SFT. There, meiotic and postmeiotic germ cells develop into mature spermatozoa that are finally released from the SC to the SFT lumen (Fig. 1A) (20, 41). SC form a highly polarized epithelial monolayer that functions as a habitat for the developing sperm, supplying it with nutrients and other factors and removing waste such as apoptotic cells and residual bodies that bud off the maturing spermatids during their elongation (44). Further protection and separation of the developing male germ cells from the periphery takes place at the basal membrane of the SFT by peritubular myoid cells (PTM) and due to the fact that most capillaries in the testis are not fenestrated (10, 36, 41).
Spermatogenesis is an iron-dependent process, because developing male germ cells undergo many mitotic divisions with iron needed for DNA synthesis and cell growth, in particular for mitochondriogenesis. In addition, maturing spermatids and spermatozoa are extremely vulnerable to oxidative stress, implying a need for a well-balanced iron homeostasis in the SFT (1). For peripheral iron to reach the developing testicular germ cells, it must pass from the interstitial capillaries across or between the endothelial cells, across the basal membrane of the SFT, and possibly transcellularly across the epithelial SC barrier. It was suggested that the SC barrier is playing a role in iron transport to the developing sperm (48), but the detailed mechanism of iron transport through the blood-testis barrier (BTB) is still unclear. In contrast, iron transport through the epithelial barrier of the small intestine is well characterized. There, iron is transported from the intestinal lumen to the blood stream by the apical divalent metal transporter-1 (DMT1) and the basolateral iron exporter ferroportin. In addition, basolateral transferrin receptor (TfR1) imports iron to epithelial cells from the blood (35).
Iron import to the SFT by transferrin was demonstrated (33, 48, 50), and it was hypothesized that SC could import iron basolaterally, transport it across the BTB, and supply iron to the developing male germ cells (16, 49). Transferrin together with albumin was localized to primary spermatocytes and differentiating spermatids in rat testis (12, 48), suggesting a possible entry of iron to the germ cells through testicular transferrin secreted by SC (45, 52), but the pathway of iron within the SFT is not elucidated. Also, nothing is known so far about the localization of the iron exporter ferroportin that could potentially play a role in SFT iron transport. Ferroportin transcript levels in the SFT are highest in PTM cells and the epithelial SC layer (Table 1; also see http://public.wsu.edu/∼griswold/microarray/). In other epithelial cell layers ferroportin is expressed basolaterally (9, 51), which is a location that extrapolated to SC would exclude it from being an iron supplier to the developing spermatids.
In addition to transferrin and TfR1, other proteins involved in iron transport, such as DMT1 and glycosyl phosphatidylinositol-anchored ceruloplasmin (11, 13, 42, 46), and iron storage, such as cytosolic and mitochondrial ferritin, were found in the testis. DMT1 is suggested to be expressed in a stage-dependent manner in the elongating spermatids and in the cytosol and nuclei of SC of adult rat testis, where it may play a role in iron uptake (13), but the source of this iron has not been identified so far. During their development, elongating spermatids shed residual bodies that contain mitochondria and cytosol and also likely contain cytosolic ferritin (3). SC phagocytose, these residual bodies, and iron may be released from phagocytosed residual bodies to the phagolysosomes and transported across the phagolysosomal membrane by DMT1. In support, the presence of DMT1 on the phagolysosomal membrane of a SC line was demonstrated previously (17). Mitochondrial ferritin was localized to the interstitial Leydig cells and to germinal cells in the SFT, and cytosolic ferritin was found mainly in the interstitium (42). The mRNA analysis of cytosolic ferritin showed that transcript levels of the ferritin H-subunit were significantly higher than L-subunits, and in the cell types analyzed the highest expression was found in PTM and SC, followed by the early spermatocytes and decreasing as spermatocytes mature (Table 1; also see http://public.wsu.edu/∼griswold/microarray/). Cellular localization of glycosyl phosphatidylinositol-anchored ceruloplasmin has never been determined (11, 46).
Here, we investigated the iron supply system to male germ cells using iron-overloaded mice and mice with targeted deletions of Hfe or IRP2, interfering with the systemic iron regulation of hepcidin or with cellular iron regulation, respectively (4, 22, 25, 29). Hfe is the gene mutated in the most prevalent form of hereditary hemochromatosis. Hfe mutations can cause constitutively low levels of the iron-regulating hormone hepcidin and systemic iron overload. In contrast, the targeted deletion of IRP2 causes misregulation of transcripts that contain iron-responsive elements such as TfR1 and the two ferritin subunits and leads to late-onset neurodegenerative disease and anemia. The distribution and regulation of iron transport and storage proteins in the testis give rise to a model for a largely autonomous iron cycle within the SFT that supplies most of the iron needed for spermatogenesis. This autonomous cycle could be a strategy to stabilize the iron pool and to control it within the closed compartment of a target organ.
MATERIALS AND METHODS
All mice were of a C57Bl/6J background. All mouse experiments were approved by the Technion Animal Ethics Committee, Haifa, Israel. IRP2−/− mice were a generous gift from Tracey Rouault (Molecular Medicine Program, National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, MD). Hfe−/− mice were a generous gift from Joanne Levy (deceased) and Nancy Andrews (Pediatrics and Pharmacology & Cancer Biology, Duke University, Durham, NC). Testes of 2-mo-old transgenic mice that express green fluorescent protein (GFP) from the scleraxis (scx) promoter (scx-GFP mice) were a kind gift from Benny Shilo, Weizman Institute, Israel. For the iron overload experiments, 4- to 6-mo-old C57Bl/6J male mice were injected with a total of 100 or 25 mg of Fe in the form of iron dextran (Sigma-Aldrich) for the heavy or moderate iron treatment, respectively. Ten or 2.5 mg of iron daily was injected peritoneally 5 days/wk for 2 wk.
Immunofluorescence and immunohistochemistry.
Testes were fixed in Bouin's solution (Sigma-Aldrich), dehydrated in increasing concentrations of ethanol and xylene, and embedded in paraffin. Sections (5 μm) were cut using a Thermo Microtome and fixed onto Superfrost microscope slides.
Deparaffinized and rehydrated sections were incubated for antigen retrieval with 0.01 M sodium citrate in ddH2O, pH 6.0, for all stainings or 6 M urea in PBS for TfR1 staining, followed by extensive rinses in ddH2O, and blocked with 10% normal goat/donkey serum (Jackson Labroratory) in PBS containing 0.1% bovine serum albumin (BSA; Sigma Aldrich). Sections were incubated overnight at room temperature in a humidified chamber with either affinity-purified rabbit anti-mouse liver ferritin (diluted 1:400), affinity-purified anti-mouse ferritin H-subunit antibody (diluted 1:200, raised against recombinant H-subunit; the plasmid was a kind gift from P. Santambrogio, Milan, Italy), affinity-purified rabbit anti-mouse ferroportin (diluted 1:200), affinity-purified rabbit anti-mouse DMT1 (diluted 1:400), monoclonal mouse anti-human transferrin receptor (diluted 1:200; Zymed), goat anti-mouse transferrin (diluted 1:250; Santa Cruz Biotechnology), or goat polyclonal anti-GFP (diluted 1:500; Abcam). For immunofluoresence (IF), Alexa donkey anti-rabbit 568, donkey anti-goat 488, goat anti-mouse 488, and goat anti-rabbit 568 (Invitrogen) diluted 1:1,000 were used as secondary antibodies and incubated for 1 h. For immunohistochemitry (IHC), slides were rinsed with PBS before being incubated with biotinylated goat anti-rabbit or goat anti-mouse antibody (Vector) diluted 1:500 for 1 h at room temperature. Sections were incubated with Vectastain ABC kit (Vector), following the manufacturer's instructions. The immunohistochemical reaction was visualized using 3,3-diaminobenzidine-tetrahydrochloride (DAB; Sigma-Aldrich). Sections were counterstained with Harris hematoxylin (Sigma-Aldrich), dehydrated, and mounted with Eukitt resin (Sigma-Aldrich). Negative controls were incubated only with secondary antibodies and with only one primary antibody, followed by both secondary antibodies for IF double-staining experiments.
Protein quantification from IF and gels.
Comparative quantification of protein expression levels was done on IF images, immunoblots, or immunoprecipitation (IP) from metabolically labeled cells using Image J (http://rsbweb.nih.gov/ij/) or Adobe Photoshop software. The images were acquired at short exposure times to prevent image saturation. Pixel intensity of multiple areas was quantified, and average and standard deviations were calculated.
The moderately iron-overloaded testes were fixed in 4% paraformaldehyde in PBS, washed with increasing amounts of sucrose, and embedded in OCT. Frozen sections were cut (10 μm) using a Leica cryostate. The heavily iron-overloaded tissues were prepared as for IF. Deparaffinized and rehydrated sections were incubated with peroxidase-blocking solution (3% H2O2 in PBS) for 30 min and washed five times with ddH2O. Slides were stained with Prussian blue solution [1% K4Fe(CN)6 and 1% HCl in ddH2O] for 60 min and washed five times with ddH2O. For enhancement, samples were incubated with DAB solution (0.07% DAB and 0.05% H2O2 in PBS) for 1–6 min and washed again five times with ddH2O.
Microscopy and image processing.
Image visualization was performed on a Nikon Eclipse 50i microscope using an X-cite series 120 microscope light source system. The acquisition software was the NIS-Elements Microscope Imaging Software. Image visualization for colocalizations, SC and TfR1, DMT1, ferritin, and ferroportin was done on an LSM 510 META laser-scanning confocal microscope from Zeiss with a Plan Apochromat 63×/1.4 NA oil DIC lens. The acquisition and analysis software program was the Zeiss LSM 510.
Animals were euthanized, and testes, spleens, and kidneys were quickly extracted and snap-frozen in liquid N2. Tissues were lysed in RIPA lysis buffer [150 mM NaCl, 1.0% NP-40, 0.5% sodium deoxycholate, 0.1% SDS, 5 mM DTT, 0.01 mg/ml leupeptine, and 1 mM 4-(2-aminoethyl)benzenesulfonyl fluoride] containing protease inhibitor cocktail (Complete without EDTA; Roche) for immunoblots and in Triton lysis buffer (10 mM Tris·HCl, pH 7.5, 150 mM NaCl, 1% Triton-X-100, and protease inhibitor cocktail) for IP. Nuclei and debris were removed by centrifugation.
Equal amounts of protein (20–40 μg/lane) were separated by SDS-PAGE on 12% gels and transferred to nitrocellulose membranes. Membranes were blocked with 2% BSA and 3% nonfat milk in Tris-buffered saline plus Tween-20 (TBST; 20 mM Tris, pH 7.6, 137 mM NaCl, and 0.1% Tween-20) at 4°C overnight and incubated for 1–3 h with rabbit anti-L-ferritin antibody diluted 1:5,000 or rabbit anti-mouse ferroportin antibody diluted 1:500 at room temperature. Horseradish peroxidase-conjugated goat anti-rabbit antibody (1:25,000 dilution in 2% BSA in TBST; Abcam) was used as a secondary antibody and added to membranes for 1 h. Blots visualization was achieved using an enhanced chemiluminescence kit (Pierce) according to the manufacturer's instructions.
Detection of germ cell apoptosis by terminal deoxynucleotidyl transferase-mediated deoxyuridinetriphosphate nick end labeling and cleaved caspase-3 IHC.
Testes were fixed in Bouin's fixative and embedded in paraffin (Roti-Plast; Carl Roth, Karlsruhe, Germany). For apoptosis detection, terminal deoxynucleotidyl transferase-mediated deoxyuridinetriphosphate nick end labeling (TUNEL) was performed according to the manufacturer's recommendations by using the ApogTag Plus Peroxidase In Situ Apoptosis Detection Kit (S7101; Chemicon, International). Anti-cleaved caspase-3 immunostaining was assessed using rabbit polyclonal cleaved caspase-3 antibody (Asp175; Cell Signaling Technology, Danvers, MA), as described (43). For each animal, at least two sections (3 μm) of two different areas of each testis were analyzed for labeled cells. The relative amounts of SFTs containing one or more stained cells were determined as described (23).
Ferritin IP from tissues.
Equal amounts of protein (1 mg) were incubated with protein A-Sepharose beads (Amersham) for 1 h to remove all proteins that bind nonspecifically to the beads. Precleared supernatant was incubated with a 25-μl slurry of protein A-Sepharose beads bound to rabbit anti-mouse liver-ferritin antibody (a kind gift from Dr. A. M. Konijn, Hebrew University of Jerusalem) for 2 h. Beads were washed three times with 1% Triton buffer (20 mM Tris·HCl, pH 8, 13.7 mM NaCl, 1% Triton-X-100, and 10% glycerol) and once with PBS. Immunoprecipitated samples were further analyzed by SDS-PAGE (14% gels) under reducing conditions and visualized by Coomassie staining.
SC line and primary culture.
SC-line MSC-1 was kindly donated by Michael D. Griswold (Washington State University, Pullman, WA). MSC-1 cells were grown as described previously (27). For mice SC primary cultures, a protocol was adapted from Karl and Griswold (19). Twenty-day-old mice were euthanized, and testes were removed, decapsulated in Hanks' basic salt solution (HBS) lacking Mg+2 and Ca+2, and dissociated into tubules by addition of 2.5% trypsin ×10 solution (Biological Industries Israel) and 13.3 mg/ml DNAse I in HBS (Sigma-Aldrich) 500–1,500 μl as needed. Testes were incubated at 37°C with occasional swirling until tubules were separated. After centrifugation at 100 g for 2 min, the supernatant was discarded, and 0.5 ml of soybean trypsin inhibitor ×50 (Biological Industries Israel) was added and allowed to stand for 1 min. Additional HBS was added, and the tube was centrifuged at 200 g for 2 min. HBS wash and centrifugation was repeated twice. PTM were removed by addition of 0.7 mg of collagenase type II in 500–1,000 μl of HBS (Sigma-Aldrich) as needed and DNAse I. Tube was incubated at 37°C with occasional swirling. Every few minutes, aliquots were checked under the microscope until the PTM layers were detached (rugged appearance of periphery of SFT) from SFT, and tubules were broken to small fragments. Three washes with HBS followed by centrifugation at 300 g for 2 min followed. After the supernatant was discarded, tubule fragments were plated in two 24-well plates per mouse, using Dulbecco's modified Eagle's medium-F-12 (1:1) (Biological Industries Israel). Both MSC-1 and SC primary culture were incubated in 5% CO2 and at 37°C.
Inductively coupled plasma-mass spectrometry.
Mouse testes samples were weighted analytically and digested with 70% nitric acid at 60°C for 30 min. Thirty percent H2O2 was added to each digested sample, and tubes were reheated to 60°C for another 30 min. Samples were diluted to 1% nitric acid with ddH2O, filtered, and analyzed in an inductively coupled plasma-mass spectrometry spectrometer, iCAP 6000 Series (Thermo Scientific).
Metabolic labeling and pulse chase experiment.
MSC-1 cells were grown in six-well plates, and the whole experiment was done on adherent cells. Cells were incubated with 100 μM ferric ammonium citrate for 24 h before the metabolic labeling and during the whole experiment, including pulse and chase. Metabolic labeling was done as described previously (30). In short, cells were starved for 30 min in starvation medium [methionine- and cysteine-free DMEM (Sigma-Aldrich) containing 5% dialyzed FCS, 1% glutamine, 1% penicillin-streptomycin, and 10 mM HEPES]. For metabolic S35 labeling of newly synthesized proteins, 0.3 mCi/ml of S35 Protein Labeling Mix (PerkinElmer) was added to the medium. Cells were then incubated for 1 h. To test the effect of molecules secreted from neighboring cells on SC-ferritin secretion, some cells were incubated with a testis cell suspension (TCS). The TCS was prepared freshly from whole testes from which the tunica was removed, and cells were mechanically dispersed by aspirating a testis and 0.5–1 ml of medium through an 18-gauge needle and syringe several times. TCS from one testis was added to eight wells during the starvation and chase period. At indicated time points after the labeling pulse, medium was collected and cells were harvested by scraping. Ferritin was immunoprecipitated from cell lysates and medium, as described above (30).
Iron entry to the SFT is limited.
The BTB plays an essential role in maintaining the environment necessary for the development and maturation of meiotic and postmeiotic male gametes. To assess the iron distribution in the testis, mice were analyzed after two different levels of iron overload, which was induced with peritoneal injections of iron-dextran. Total iron levels were assessed by inductively coupled plasma-mass spectrometry, and a 50-fold increase in total iron levels was found in the testis of iron-overloaded mice, whereas other metals were unaffected by the iron overload (Fig. 1B). Assessment of the distribution of nonheme, ferric iron by Perls staining showed that ferric iron was not distributed evenly in the testis and accumulated mainly in the interstitium (Fig. 1, C and E). In moderately iron-overloaded mice, no iron increase was detected within the SFT by either Perls or DAB-enhanced Perls staining (Fig. 1C). For comparison, in livers of moderately iron-overloaded mice, clearly ferric iron levels were elevated throughout the liver tissue. The Kupffer cells accumulated higher ferric iron concentrations than parenchymal cells of the liver, but no compartmentalization was visible (Fig. 1D). A small increase in iron levels within the SFT of heavily iron-overloaded mice was detectable only with DAB-enhanced Perls staining (Fig. 1E), and within the SFT iron concentration was higher at the periphery. Taken together, a tight barrier that protected the developing germ cells from iron overload was demonstrated at the basal membrane of the SFT.
Iron acquisition in the SFT.
Within the SFT, spermatogonia undergo multiple cell divisions, and spermatocytes go through meiosis and extensive mitochondriogenesis that depend on sufficient iron supply for DNA production and cell growth. To elucidate how early spermatocytes meet their iron needs, we performed IHC and IF studies for transferrin, TfR1, and DMT1. TfR1 internalizes transferrin to endosomes from where iron is usually transported to the cytosol by the divalent metal transporter (DMT1). Transferrin was detected in the interstitium and in early to midspermatocytes at the periphery of the SFT (Fig. 2A). In addition, strong transferrin staining was also visible in SC, where testicular transferrin is synthesized (Fig. 2A, arrowheads). The highest expression of TfR1 was also detected in the interstitial space and the periphery of the SFT in spermatogonia and in early to midprimary spermatocytes (Fig. 2D). Because TfR1 was reported earlier on SC (32), we further analyzed which cell types in the SFT express TfR1. To differentiate between SC and developing germ cells, IF colocalization and quantification analyses were carried out in scx-GFP mice (37) and various iron metabolism proteins, including TfR1, using the Zeiss LSM 510 analysis software program. In the testis, scx is expressed only in the SC with the protein found in cytoplasm, and therefore, this construct provides a SC marker. Little TfR1 expression was detected in SC (Fig. 2E, white signal), whereas spermatocytes located close to the basal membrane of the SFT exhibited stronger TfR1 staining (asterisks in Fig. 2E). DMT1 was detected in the same location as TfR1 (Fig. 2F), although DMT1 levels in early spermatocytes are low. Taken together, within the SFT, the highest expression of TfR1 is found on early to midprimary spermatocytes. The colocalization of TfR1 with transferrin and DMT1 suggested an active role for TfR1 in transferrin-dependent iron import to these cells.
DMT1 is polarized to the adluminal area of the SFT.
In the final steps of spermatid maturation into spermatozoa, the number of mitochondria per spermatid is reduced from several thousands to <100 (15), and mitochondria and much spermatid cytosol are budded off the spermatids as residual bodies that are mostly phagocytosed by SC (3). To determine whether DMT1 may play a role in the iron import from residual bodies to SC, DMT1 IHC and IF were performed (Fig. 3). In concert with the findings of Griffin et al. (13), we detected DMT1 in elongating spermatids and SC. The appearance of DMT1 location in different stages of the spermatogenic cycle is changing significantly (Fig. 3A). DMT1 expression level remained constant throughout the steps of elongation until the mature spermatozoa were released in late stage VIII (step 16 spermatids). In stage IX (step 9) no pronounced DMT1 expression was detected, whereas in stage XI (step 11) clear DMT1 expression was visible, indicating that DMT1 expression begins then. IF analysis of DMT1 enabled us to refine the characterization of its cellular and subcellular localization. The DMT1 staining showed typical SC staining spanning the radius of the tubule (Fig. 3B, arrowheads), but DMT1 staining often did not reach the basal membrane. To further analyze this apparent polarization of DMT1 in SC, testes of scx-GFP mice were stained for DMT1, and colocalization of DMT1 with GFP (white signal in Fig. 3C) was quantified. DMT1 expression within the SC was polarized to the luminal compartment (Fig. 3C). DMT1 was located on or near the SC plasma membrane adjacent to spermatids (Fig. 3C, inset) and within the SC cytosol. In summary, DMT1 accumulation in the luminal compartment of the SFT supported our hypothesis that it may play a role in iron transport during the final steps of germ cell maturation.
Compartmentalized and polarized distribution of cytosolic ferritin.
Cytosolic ferritin is the main iron storage protein and consists of 24 subunits of two types, L and H. Both subunits are translationally regulated by iron. Under low-iron conditions, IRPs bind to the 5′-untranslated region and inhibit protein translation, and in iron overload ferritin translation is high (39). In mice with targeted deletion of IRP2 (IRP2−/−), ferritin regulation is disrupted, and ferritin expression is highly uncontrolled (22, 29). Also in the testis, ferritin protein level is regulated accordingly, and testes of IRP2−/− mice or mice with iron overload have high levels of ferritin (Fig. 4A). Nevertheless, IRP2−/− mice were fertile, and viable sperm were found in the epididymides. In the SFT, no difference in the number of apoptotic cells was found between wild-type (WT) and IRP2−/− mice, which was tested by the TUNEL assay and IHC of cleaved caspase-3 (n = 5 of each genotype; data not shown). However, the ferritin level in testes of Hfe−/− mice was not significantly elevated. Expression level of ferritin was low in testis and kidney compared with spleen in WT mice, and both ferritin subunits, H and L, were expressed (Fig. 4B). IRP2 deficiency enhanced the expression of both ferritin subunits in the testis, kidney, and spleen, and in iron overload the L-subunit was more induced than the H-subunit, leading to a higher L/H ratio. Ferritin protein expression was reported in the testis mainly in interstitial Leydig cells (42). To identify the detailed distribution of cytosolic ferritin, we localized it by IHC and IF in untreated WT, iron-overloaded, and IRP2−/− mice. In agreement with the earlier findings, cytosolic ferritin was found in the interstitium (Fig. 4, C–I). In addition, we detected cytosolic ferritin within the SFT in SC and in primary spermatocytes and some spermatogonia (Fig. 4, C–I). The same distribution pattern was detected with the L- and H-subunit-specific antibodies (compare Fig. 4, C and G), indicating that most ferritin 24-mers consist of both subunits. Colocalization with TfR1 demonstrated that cytosolic ferritin was expressed in the same primary spermatocytes as TfR1 (data not shown). In IRP2−/− mice, SC-ferritin levels were significantly elevated, whereas spermatocyte ferritin was less affected (compare Fig. 4, D and E), suggesting that ferritin synthesis takes place mainly in SC. To further characterize the ferritin localization in the SFT, scx-GFP testes were used. Colocalization of cytosolic ferritin with the SC-specific scx-GFP showed that expression levels of cytosolic ferritin in the cytoplasm of SC decreased from the basal to the luminal compartment (Fig. 4I), and quantification of cytosolic ferritin in the luminal and basal compartments of SC verified this observation [colocalization in white; luminal domain: 0.020 ± 0.002 arbitrary units (AU); basal domain: 9.3 ± 1.1 AU; Fig. 4I]. Cytosolic ferritin accumulation around the primary spermatocytes suggested a transfer of ferritin from the SC to the primary spermatocytes and implied that SC secrete ferritin. IP of ferritin from the metabolically labeled SC line MSC-1 and its conditioned medium demonstrated that these cells secrete ferritin (Fig. 4J). The predominant subunit expressed in these cells is the ferritin H-subunit. To elucidate whether the ferritin secretion from SC can be triggered by the SC environment, the MSC-1 cells were incubated with a cell suspension of dispersed testis cells (TCS). These experiments indicated that ferritin secretion was induced by the presence of testicular cells (+TCS) during the starvation and chase period of a pulse chase experiment (Fig. 4, J–L).
Taken together, our data show that most peripheral iron remains stored in ferritin in the interstitial space and that ferritin within the SFT is synthesized mainly in SC that are capable of secreting ferritin, which may be taken up by primary spermatocytes.
Localization and regulation of ferroportin in the testis.
We showed that the SFT is well protected from iron overload, and only a little peripheral iron crossed the basal membrane. In addition, within the SFT, DMT1 was present and likely mediated both transferrin/TfR1-dependent and -independent iron uptake systems. The polarized expression of DMT1 in the luminal compartment of SC suggested that much iron is taken up by the SC in that area, and the polarized accumulation of ferritin in the basolateral compartment of SC supported the notion of iron export from SC through ferritin. We wondered what role the iron exporter ferroportin might play in the SFT iron transport. To test this, ferroportin was localized by IF in testes of adult mice (Fig. 5, A and B) and in murine primary SC cultures (Fig. 5C). Expression of ferroportin was seen clearly in the PTM cells, which line the SFT, and on the muscular wall of larger blood vessels (Fig. 5, A and B). IF localization of ferroportin in testis of scx-GFP mice suggested that ferroportin may be expressed sparsely on the basal membrane of adult SC (Fig. 5B). This was further supported by IF analyses of ferroportin in primary SC cultures in which the membrane of SC was clearly stained (Fig. 5C). The possible presence of ferroportin expression within the membrane of SC suggested a role of ferroportin in exporting iron from the SFT. But more prominently it seems to play a role in iron transport between the SFT and interstitial blood vessels.
To examine further the regulation of ferroportin expression in the testis, protein levels were analyzed in testes from control iron-overloaded Hfe−/− and IRP2−/− mice by Western blot and compared with ferroportin levels in the spleens of these same animals. As expected, ferroportin protein level decreased in the spleens of iron-overloaded mice and increased in the spleens of Hfe−/− mice compared with WT control mice, confirming ferroportin regulation by the peptide hormone hepcidin in the periphery. This pattern of regulation was not found in the testes, where ferroportin expression levels did not change (Fig. 5, D and E). Because Western blots were performed on whole testis lysates and therefore represented an average protein level of a mixture of cell types, we tested whether ferroportin localization was shifted under the different conditions. IF on testes of iron-overloaded Hfe−/− and IRP2−/− mice demonstrated a similar expression pattern of ferroportin on PTM and blood vessel walls in the different genotypes (Fig. 5F). Taken together, ferroportin was located mostly outside of the SFT and demonstrated a testis-specific regulation.
Many iron transport and storage proteins have been found in the testis, but their detailed localization and regulation have not been studied to date. We explored the mechanism and regulation of iron transport and storage in the testis to reveal how compartmentalization by tissue barriers can protect delicate and vital cells from environmental stress. On the basis of our findings, we propose a new model of a largely autonomous internal iron cycle within the SFT that is supported by iron import and export from and to the periphery (Fig. 6).
The SFT iron cycle.
Male germ cells from early developmental stages to pachytene spermatocytes express TfR1 and ferritin highly, indicating active iron uptake from the surrounding and its storage (Figs. 2 and 4). Developing spermatocytes move from the SFT periphery across the BTB toward the lumen of the SFT, and iron stored in ferritin in these cells is used gradually. The function of this accumulated iron is unknown, but it can be speculated that it is used for the meiotic cell divisions that these cells undergo and for mitochondriogenesis. The number of mitochondria increase progressively from the primary spermatocyte to the early acrosomal spermatid stage (8, 15), a process requiring iron for mitochondrial iron sulfur cluster and heme synthesis. Therefore, it makes sense that the ferritin concentration decreases toward the SFT lumen (Fig. 4). Iron uptake by mitochondria was shown recently to be essential for sperm development in Drosophila (28). During the final steps of spermatid maturation by elongating spermatids, many mitochondria are shed into residual bodies that are phagocytosed by SC. At this stage the number of mitochondria per spermatid is reduced from several thousands to less than 100 (15), which implies that large amounts of iron are transferred from elongating spermatids to the apical area of SC. The appearance of DMT1 in elongating spermatids correlates with the mitochondrial segregation between midpiece and residual bodies in these cells (21), but a role for DMT1 in that location is not clear. We found DMT1 expressed in the apical pole of the SC, but not restricted to the apical membrane, strongly supporting the notion that DMT1 is involved in iron uptake from phagocytosed residual bodies. (Figs. 3 and 6). Because DMT1 is not a soluble protein, we hypothesize that it is bound to endo- and phagolysosomes. Thus the localization of DMT1 suggests that it plays a TfR1-independent role in iron transport within elongating spermatids and the luminal compartment of SC. In addition, the low levels found throughout the SFT likely support DMT1 involvement in transferrin/TfR1-mediated iron import.
The iron transferred to the luminal compartment of SC can be stored in ferritin, and we demonstrated that SC contain ferritin. Interestingly, the cytosolic ferritin distribution in SC is not uniform, and ferritin accumulates near the basolateral domain around primary spermatocytes (Fig. 4I). SC and macrophages have many similarities that include common markers, and both cell types function as phagocytes and nurse cells to spermatocytes or erythropoietic cells, respectively. Macrophages secrete ferritin that is taken up by human erythroid precursors (24) and are likely a source for serum ferritin through a nonclassical secretion pathway (6). We demonstrated that the SC cell line expressed and secreted ferritin in a regulated manner. Therefore, it is conceivable that SC in vivo secrete ferritin that is subsequently taken up by the next generation of spermatocytes (Fig. 4, J–L). In addition, we observed pronounced misregulation and overexpression of cytosolic ferritin in SC of IRP2−/− mice. The microarray data also show that mRNA levels of both ferritin subunits are high mainly in SC (Table 1; also see http://public.wsu.edu/∼griswold/microarray/). Taken together, these data suggest that much ferritin synthesis takes place in SC and that other cell types high in ferritin, such as the early spermatocytes, may not only synthesize but also acquire ferritin through receptor-mediated endocytosis. Scara-5, a ferritin receptor, is expressed in the testis, but the cell type being expressed it is not yet known (26). Recycling iron from mature sperm back to the primary spermatocyte through the SC completes an autonomous cycle of iron transport within the SFT (Fig. 6).
Peripheral support for the SFT cycle.
The small obligatory iron loss by mature sperm that is leaving the testis through the SFT fluid to the epididymis has to be replaced from the periphery. In contrast to earlier findings (33, 38, 48, 49), very little TfR1 was found on the SC basolateral membrane (Fig. 2E), but the high TfR1 expression on primary spermatocytes suggested that these cells take up iron from testicular transferrin within the SFT. Ferroportin, the iron exporter that is lining the blood vessels and the SFT (Fig. 5G), may facilitate bidirectional iron transport across the interstitium, where semipermeable barriers were reported at the interstitial capillaries and the PTM cells near the basal membrane of the SFT (36). Transferrin that is secreted by SC can participate further in this interstitial iron transport (12, 33). Finally, SC may export excess iron through ferroportin to the interstitial space (Fig. 5B). Ferroportin expression on the basal membrane of SC is low, but the fact that ferroportin does colocalize with scx sporadically (Fig. 5B), that testes of WWv mice and SC primary cultures do contain ferroportin mRNA (Table 1; also see http://public.wsu.edu/∼griswold/microarray/), and that ferroportin protein can be detected easily in SC primary cultures (Fig. 5C) strongly supports that some ferroportin is located basally on SC. In comparison, in the retinal pigment epithelial cell layer, which functions similarly to SC as a cellular barrier, ferroportin is located basolaterally, and no ferroportin is found in the apical cilia (14). Therefore, we suggest that ferroportin also plays a role in iron export from the SFT. In summary, iron transport to and from the SFT basal membrane through the interstitium is likely supported by ferroportin and testicular transferrin. Iron may enter the SFT with the transferrin-TfR1 system through primary spermatocytes and exit the SFT through ferroportin expressed on the basolateral membrane of SC. However, we suggest that iron uptake from the periphery is a minor contribution to the total iron cycle of the SFT.
Compartmentalization as a strategy for protection.
The systemic iron overload applied caused heavy iron accumulation in peripheral tissues and in the testicular interstitium but caused only little iron overload within the SFT, which was detectable only by Perls-DAB stain but not by Perls stain alone (Fig. 1, C and E). This indicated that the SFT basal membrane together with PTM cells provided a strong first protective barrier (36) that protects the developing germ cells of all developmental stages from peripheral iron fluctuations. Thus, spermatogonia and preleptotene spermatocytes that have not entered the BTB are also protected from iron overload. The location of this iron barrier outside the BTB further supports our hypothesis of an internal SFT iron cycle (Fig. 1E). A secondary protection level may exist at the BTB, since ferric iron within the SFT is higher at the periphery, where early spermatocytes are located outside of the BTB.
An additional protective mechanism was found at the level of ferroportin. Mice that are heavily iron overloaded showed unchanged testicular ferroportin expression levels, whereas their splenic ferroportin is downregulated as expected (Fig. 5, D and E). The molecular basis for this testis-specific regulation of ferroportin is currently under investigation. It implies that in contrast to other tissues such as the intestine and the spleen, where iron circulation through ferroportin is decreased during iron overload, the testis can continue to export iron under these conditions and thus further protect the developing male germ cells from iron excess.
As mentioned earlier, in the retina, the outer blood-retina barrier consists of the retinal pigment epithelial cell layer (7, 47), a cell type that, similarly to SC, functions as a polarized cellular barrier. Retinal pigment epithelial cells are involved in nutrient transport, and phagocytose membrane-bound particles shed from the neighboring cone and rod cells. Ferroportin in genetically iron-overloaded retina is elevated and, as in the testis, does not reflect the regulation seen in spleen and intestinal epithelium of iron-overloaded animals (14).
Recycling and compartmentalization as a strategy to maximize tissue independence.
In systemic iron homeostasis, most iron is recycled between red blood cells and macrophages, and only small regular iron losses are replaced by dietary iron. Similarly, we suggest that most iron in the SFT is recycled between developing germ cells and SC, and small losses caused by mature sperm taking some iron along when moving on to the epididymis are replaced by iron uptake by primary spermatocytes from the periphery. This compartmentalized iron cycle has the advantage to detach male germ cell development from peripheral iron fluctuations, and SFT iron levels can be kept low and constant to protect the developing germ cells from iron-induced oxidative damage while continuous iron supply is secured.
Another well-established epithelial cell barrier is the choroid plexus, which is part of the blood-brain barrier. Recently, high representation of mRNAs of all proteins needed for iron uptake was demonstrated, and some of these transcripts were also verified to be present at the protein level (40), suggesting that this barrier contributes to the regulation of brain iron homeostasis. Our findings support a new model of an internal iron cycle within a gated compartment in which most of the iron is recycled, and only small amounts of iron are lost and replaced to maintain homeostasis. This is an optimal system to protect cells in closed compartments from damaging nutrient fluctuations and is an example for a biological strategy to control the flow of resources. We suggest that such internal cycles may exist in other compartmentalized tissues such as the retina and the brain.
This work was financially supported by a grant from the State of Lower Saxony and the Volkswagen Foundation, Hannover, Germany, to S. Schubert and E. G. Meyron-Holtz and benefited from core services and support from the Larry I. Lokey Center for Life Science and Engineering at the Technion.
No conflicts of interest, financial or otherwise, are declared by the authors.
Y.L.-B., A.W., and B.M. performed the experiments; Y.L.-B., L.A.C., A.M., and E.G.M.-H. analyzed the data; Y.L.-B., S.S., A.M., and E.G.M.-H. interpreted the results of the experiments; Y.L.-B. and L.A.C. prepared the figures; Y.L.-B., L.A.C., S.S., A.M., and E.G.M.-H. edited and revised the manuscript; Y.L.-B., L.A.C., A.W., B.M., S.S., A.M., and E.G.M.-H. approved the final version of the manuscript; E.G.M.-H. did the conception and design of the research; E.G.M.-H. drafted the manuscript.
We express our special thanks to Michael D. Griswold, Alice Karl, and the whole Griswold Laboratory, Pullman, Washington State University, for hosting Y. Leichtmann-Bardoogo and E. G. Meyron-Holtz and introducing us to the testis world. Ryan Evanoff and Chris Small from the Griswold Laboratory were also very helpful with the MSC-1 cell line and the analysis of the microarray data. We also thank the European Molecular Biology Organization short-term fellowship that funded this trip for Y. Leichtmann-Bardoogo, Benny Shilo and Shay Rotkopf for giving us the scx-GFP mouse testes, Benjamin Podbilewicz, Wing-Hang Tong, and Abraham M. Konijn for fruitful discussions and critical reading of the manuscript, Mahmoud Huleihel and his laboratory for technical support, and Gallit Assouline for graphical assistance.
- Copyright © 2012 the American Physiological Society