PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
J Steroid Biochem Mol Biol. Author manuscript; available in PMC 2010 May 1.
Published in final edited form as:
PMCID: PMC2680768
NIHMSID: NIHMS97493

Human Pancreas-Specific Protein Disulfide Isomerase Homolog (PDIp) is an Intracellular Estrogen-Binding Protein that Modulates Estrogen Levels and Actions in Target Cells

Abstract

Earlier studies showed that protein disulfide isomerase (PDI), a well-known protein folding catalyst, can bind estrogens. Whether other PDI homologs can also bind estrogens, and if so, what are the biological functions of this unique property are not known at present and thus are the subject of our present investigation. Here we report that, of the six representative PDI homologs examined (human PDI, PDIp, ERp57, ERp72, PDIA6 and rat PDIr), only the human pancreas-specific PDI homolog (PDIp) had a similar binding affinity for radiolabeled 17β-estradiol (E2) as did PDI, with apparent Kd values of 1.5 ± 0.3 and 1.5 ± 0.2 μM, respectively. However, PDIp and PDI had distinctly different binding preference for several estrogen analogs. Moreover, we found that PDIp could serve as a high-capacity intracellular E2-binding protein and could modulate the intracellular concentrations of E2 in cultured mammalian cells as well as in human pancreatic tissue. The PDIp-bound E2 in a cell could be released following a drop in the extracellular E2 concentrations, and the released E2 could then augment estrogen receptor-mediated transcriptional activity. Notably, the estrogen receptors α and β were also found to be expressed in rodent and human pancreatic tissues where high levels of PDIp were detected. Altogether, these data show that, in addition to its well-documented function as a protein folding catalyst, PDIp can also serve as an effective modulator of the cellular levels and biological actions of endogenous estrogens in certain target sites (such as the pancreas) where estrogen receptors and PDIp are co-present.

Keywords: PDI, PDIp, estrogen receptor, SHBG, pancreas, disulfide bond, estrogen, thyroid hormone

1. INTRODUCTION

The endogenous estrogens exert a wide array of biological functions in various target tissues in females. The hormonal actions of estrogens are not only crucial for the maintenance of normal female physiology such as the development of reproductive organs and secondary sex characteristics, but they also contribute to the induction of tumors in certain target organs (e.g., pituitary, uterus and breast) [13]. Many of the well-known physiological actions of endogenous estrogens are mediated by the estrogen receptor (ER) α and β [4], which are transcription factors that can initiate the expression of target genes. In the past decade or so, enormous advances have been made in delineating ERα- and ERβ-mediated signalling pathways that mediate many of the known biological actions of endogenous estrogens in the body [46]. In addition, the non-genomic actions of ERs have also been suggested to play a role in mediating the estrogen-induced rapid signal transduction [7, 8].

Besides the ER-mediated signaling pathways, other biological processes, such as metabolism, transport and storage of endogenous estrogens, also play important roles in modulating the actions of endogenous estrogens in target tissues. For instance, the interconversion between 17β-estradiol (E2) and estrone catalyzed by 17β-hydroxysteroid dehydrogenase represents one of the most important mechanisms for regulation of estrogenic status in the uterus [9], as well as in other target tissues, including breast cancer [10]. Notably, our earlier study showed that the high levels of the oxidative 17β-hydroxysteroid dehydrogenase (Type II) contained in human breast cancer cells contribute to their selective insensitivity to the anticancer actions of 2-methoxy-E2 (a nonpolar E2 metabolite with strong growth-inhibitory activity) [11]. In addition to the estrogen-specific metabolizing enzymes, another well-documented modulator of global estrogenic status in the body is the sex hormone-binding globulin (SHBG). SHBG is a major E2-binding protein present in blood, and it is estimated that >99% of circulating E2 is tightly bound to SHBG [12], which modulates the bioavailability of circulating E2 and ultimately estrogenic activity in various target tissues [13]. Beside, SHBG was also shown to be involved in mediating estrogen-induced membrane signaling pathways [13, 14].

Protein disulfide isomerase (PDI), also known as P4HB or PDIA1, is a well-known protein folding catalyst and a molecular chaperone [1517]. Earlier studies showed that this protein can bind [3H]E2 in vitro [18, 19], and recently our data confirmed this property [20]. Differing from SHBG, PDI is an intracellular protein. Because this protein is present in liver as well as other tissues at very high levels, it has been suggested that PDI may function as a high-capacity intracellular estrogen-binding protein by serving as an intracellular estrogen reservoir and thus may modulate endogenous estrogen’s biological functions [19]. This hypothesis was supported by our recent finding showing that PDI could effectively accumulate E2 (a representative estrogen tested) inside mammalian cells and augment ER-mediated transcriptional activity of E2 in cells that expressed high levels of PDI [20]. Given that the PDI family is composed of multiple homologs [16], it is of interest to determine whether other members of the PDI family also have E2-binding activity, and if so, what are their potential biological functions in regulating the biological activity of endogenous estrogens.

The pancreas-specific PDI homolog (PDIp) was previously reported to be expressed in pancreatic acinar cells but not in islet cells or a few other tissues examined [2124]. Here we report that, of the six representative PDI homologs examined in the present study, only PDIp has a similar E2-binding activity as does PDI. Moreover, PDIp can effectively modulate the concentrations and hormonal activity of E2 in cells. Since ERα and ERβ were also found to be co-expressed with PDIp in rodent and human pancreatic tissues, collectively these data suggest that PDIp may function as a modulator of the biological functions of endogenous estrogens in this organ.

2. MATERIALS AND METHODS

2.1. Chemicals, reagents, cell lines and tissues

Triiodothyronine (T3), vitamin D3 and 1,25-dihydroxyvitamin D3 [1,25(OH)2D3] were purchased from Sigma-Aldrich (St. Louis, MO). 17β-Estradiol (E2) was purchased from Steraloids Inc. (Newport, RI). [3H]E2 (specific activity of 110 Ci/mmol) was obtained from Perkin Elmer (Waltham, MA). All other chemicals and reagents used in this study were of analytical grade or higher.

Specific rabbit antibodies against a number of proteins were used in this study, and they were obtained from the following sources: PDI was from Sigma-Aldrich (catalog No. is P7372, dilution of 1:2000 for Western blotting), ERα from Santa Cruz (Santa Cruz, CA, sc-543, 1:200; sc-7207, 1:500), ERβ from Santa Cruz (sc-8974, 1:500) and Invitrogen (Carlsbad, CA, 51-7700, 1:100), GAPDH from Cell Signaling (Danvers, MA, 2118, 1:2000), and calnexin from Santa Cruz (sc-11397, 1:400). The endogenous estrogens contained in fetal bovine serum (FBS) were removed by using dextran-coated charcoal as described previously [25].

The mouse anti-PDIp antiserum (dilution of 1:2500 for Western blotting) was raised in our laboratory. Note that although the antibody against the recombinant human PDIp was raised in BALB/c mice (described in the Supplementary Data section), we found that the antibody could also cross-react with the PDIp proteins present in C57BL/6J mouse pancreatic tissues (Fig. 7) but it had no reactivity with PDI or other proteins (see Fig. S1C, Fig. S2).

FIGURE 7
Co-expression of PDIp with ERα and ERβ in the pancreas of human, rhesus monkey, rat and mouse

All cell lines used in this study were obtained from ATCC (Manassas, VA), and they included the MCF-7 and MDA-MB-435 human breast cancer cells; Mia Paca-2, Bx PC-3 and Capan-2 human pancreatic cancer cells; RIN-m5F rat pancreatic cells; HT22 mouse hippocampal cells; SK-N-SH human neuroblastoma cells; and monkey kidney cos-7 cells. These cell lines were cultured according to supplier’s instructions.

Pancreas tissues of Sprague-Dawley rats and C57BL/6J mice used in this study were collected from untreated control animals. Pancreatic tissues from rhesus monkeys were purchased from US Biomax (Rockville, MD, RhFTS151). Two human normal pancreatic tissue specimens were obtained from the National Disease Research Interchange (catalog No. is 0060960-13 and 0060884-14).

2.2. Construction of plasmids

The vectors carrying the cDNAs encoding each of the six PDI homologs, namely, human PDI, PDIp, ERp57, ERp72, PDIA6 and rat PDIr, were purchased from ATCC (I.M.A.G.E. numbers: 3879411, 3162384, 3542405, 2819726, 3454007 and 7323988, respectively). Primers for each of these genes were designed with a respective pair of the restriction enzyme sites: 5′-NdeI/BamHI-3′ for PDI, 5′-NdeI/XhoI-3′ for PDIp, 5′-SalI/BamHI-3′ for ERp57, 5′-NdeI/BamHI-3′ for Erp72, 5′-NdeI/XhoI-3′ for PDIA6, and 5′-NdeI/BamHI-3′ for PDIr. Each gene was first cloned into the pGEM-T vector (Promega, Madison, WI), and then sub-cloned into the pET-19b vector (Novagen, La Jolla, CA), in which a short sequence of GTCGACGCTGCA (containing SalI site) had been inserted between NdeI and XhoI sites by using the Quickchange Mutation Kit (Stratagene, Cedar Creek, Texas) according to manufacturer’s instructions. The pcDNA3.1 vector for protein expression in mammalian cells was purchased from Invitrogen, and PDIp was cloned into this vector at the sites of 5′-HindIII/XhoI-3′. PCR3.1-ERα and pGL-basic + ERE (estrogen responsive element) + E1b + luciferase reporter gene vector were gifts from Professor Carolyn Smith at Baylor College of Medicine (Houston, TX, USA).

2.3. Protein purification

The recombinant pET-19b vectors carrying each of the PDI homolog cDNAs were transformed into BL-21 (DE3) E. coli host cells, and the respective proteins were over-expressed after induction with IPTG (at 0.2 mM for Erp72 and at 0.1 mM for the other five PDI homologs) at 20°C overnight. Cell pellets were dissolved in 10 mM sodium phosphate buffer (pH 7.4, 0.5 mM EDTA, 1 mM DTT, and 1 mM PMSF), and sonicated for 60 cycles (3 s sonication followed by 15 s break). The supernatants of bacterial cell lysates (after centrifugation at 13,200 rpm for 16 min at 4°C) were incubated with Ni Sepharose High Performance (Amersham, Piscataway, NJ) at 4°C for 1 h, and the unbound proteins were removed by washing three times with 20 mM sodium phosphate buffer (containing 0.5 M NaCl and 20 mM imidazole, pH 7.4). After elution with 0.5 M imidazole (0.5 M NaCl, pH 7.4), fractions containing target proteins (examined by SDS-PAGE analysis) were combined, desalted using a PD-10 column (Amersham), and concentrated in an Amicon Ultra-4 tube (Millipore, Billerica, MA) before stored at −80°C. Protein concentration was determined using protein assay solution (Bio-Rad, Hercules, CA) with bovine serum albumin as standard.

The partially purified PDIp protein was further separated on the ÄKTA FPLC system using ion-exchange chromatography (Mono Q 4.6/100 PE, Amersham). The salt gradient elution was performed from 0 to 0.15 M NaCl (20 mM Tris-HCl, pH 7.5) in 2 column volumes and then gradually to 0.5 M NaCl in 25 column volumes followed by quickly increasing in 3 column volumes to 1.0 M NaCl. PDIp was eluted out with 0.35–0.4 M NaCl (see Fig. S1A, S1B) and the highly-pure PDIp protein was used for raising polyclonal antibodies in female Balb/C mice.

2.4. [3H]E2-binding ability of purified PDI homolog proteins

A number of assays were used to assess the binding ability of PDI homolog proteins for [3H]E2. To determine the [3H]E2-binding activity of these purified proteins (at a final concentration of 0.5 μM) in vitro, each of the proteins was incubated with 13.2 nM [3H]E2 (0.1 μCi per tube) in 10 mM sodium phosphate buffer ( pH 7.4) at room temperature for 2 h and then at 4°C overnight. The samples (100 μL) were centrifuged briefly at 4°C and subject to desalting on a PD-10 column (from Amersham) with 10 mM sodium phosphate buffer (pH 7.4, 0.15 M NaCl). Eluted fractions from 2.3 to 4.3 mL were collected and mixed with 3-volume scintillation cocktail (Fisher Scientific, Pittsburgh, PA) for radioactivity measurement with a Pelkin Beta Counter. For saturation experiments, different concentrations of [3H]E2 were present during the incubation, and the same assay method as described above was used. The nonspecific binding was determined in the presence of 100-fold excess of cold E2 for representative lower concentration points.

In addition to the use of the desalting method for separation of free and bound E2, the charcoal adsorption method was also used. The samples were mixed with equal volume of a solution containing 1% dextran-coated charcoal plus T70 dextran (0.1%, w/v), incubated at 4°C for 3 min, gently mixed four times during the incubation, and centrifuged at 16,000 g (4°C) for 3 min. The supernatant was mixed with scintillation cocktail for radioactivity determination.

2.5. [3H]E2-binding ability of PDI homolog proteins in E. coli and mammalian cells

We determined the [3H]E2-concentrating effect of E. coli cells that selectively over-expressed each of the PDI homology proteins as a means to assess the E2-binding ability of these proteins in live cells. E. coli cells that over-expressed PDI homologs were cultured in glass tubes and then harvested in eppendorf tubes by centrifugation. After washed with LB medium once, the cells were re-suspended in cell-wall digestion buffer without lysozyme (30 mM Tris-HCl, 1 mM EDTA, 20% sucrose, pH 8.0) and adjusted to the same density (absorbance at 600 nm was 0.1). After that, lysozyme (at a final concentration of 0.3 mg/mL) and [3H]E2 (at increasing concentrations) were added to each tube before incubation at room temperature for 2 h and then at 4°C overnight. The cell wall-digested E. coli cells were centrifuged at 4°C (3000 g, 8 min). Pellets (representing the spheroplasts) were broken down by treatment with 1% SDS and 0.2 M NaOH before mixed with 5-volume scintillation cocktail for radioactivity measurement. The supernatants (representing the periplasmic fracttion) were also subjected to radioactivity measurement.

For studying mammalian cells that over-expressed PDIp, they were incubated in DMEM medium (without FBS) for 1 h, and then incubated with [3H]E2-containing DMEM medium for 2 h to allow [3H]E2 to enter the cells. After the medium was removed, cells were washed once with DMEM medium and digested with a lysis buffer (Promega, E1500) and subjected to radioactivity measurement.

2.6. Preparation of tissue lysates and mammalian cell microsomes

Animal tissues were homogenized in the 20 mM Tris-HCl buffer, pH 7.5, containing 0.15 M NaCl, 0.2% Triton X-100, plus protease inhibitors cocktail (Sigma-Aldrich), and then centrifuged at 4,000 g for 10 min at 4°C. The obtained supernatants were used for Western blotting (see Fig. 7). Crude lysates from human pancreatic tissues were used for [3H]E2-binding assay after removal of PDIp by immunoprecipitation using the mouse anti-PDIp antiserum (the normal mouse serum was used as control).

To prepare microsomal fraction from cos-7 cells, cells were dissolved in 20 mM Tris-HCl buffer, pH 7.5, containing 0.15 M NaCl, plus protease inhibitors cocktail, and were broken down by sonication. The cytosol was obtained after centrifugation of the whole cell lysates at 10,000 g for 16 min at 4°C. Microsomes were prepared by further centrifugation of the cytosolic fraction at 100,000 g for 90 min at 4°C. The whole cell lysates and microsomes were subject to Western blotting.

2.7. Transfection of mammalian cells

The plasmids pcDNA3.1, pcDNA3.1-PDIp and pCR3.1-ERα were used for assaying the [3H]E2-binding in cos-7 cells. Transfection experiments were done in 24-well plates by using lipofectamine 2000 and Opti-MEM I reduced serum medium (both obtained from Invitrogen) according manufacturer’s instructions. At 36 h after transfection, cells were subjected to [3H]E2-binding activity measurement (as described above), protein concentration determination, and western blotting.

The plasmid pGL-basic + ERE + E1b + luciferase reporter gene was used for ERα-mediated transcriptional activity in MCF-7 cells. After transfection, MCF-7 cells were cultured with EMEM medium plus 10%DCC-treated FBS for one day and then cultured with the medium containing E2 released form cos-7 cells for one day. Cell lysates were subjected to luciferase activity measurement (using a kit from Promega, E1500) and protein concentration determination.

3. RESULTS

3.1. Characterization of estrogen-binding activity of the recombinant human PDIp protein

Earlier studies showed that rat PDI has E2-binding activity [19]. Our recent study showed that human PDI can function as an intracellular estrogen-binding protein and can modulate estrogenic status in estrogen target cells that express high levels of PDI [20]. In the present, we were interested in knowing whether other human PDI homologs, such as PDIp, Erp57, Erp72, PDIr and PDIA6 [1517], also have this E2-binding activity. We selectively expressed in E. coli these six representative PDI homologs (each with a catalytic CxxC motif and a non-catalytic thioredoxin-like domain), purified these proteins (Fig. 1A), and then determined their binding affinity for [3H]E2.

FIGURE 1
Characterization of PDIp as an [3H]E2-binding protein

First, we used E. coli cells as a model (refer to our recent study [20]) to probe the binding ability of these PDI homologs for [3H]E2 in live E. coli cells. As shown in Fig. 1B, spheroplasts of E. coli cells that selectively expressed PDIp or PDI exhibited a significantly higher ability to bind [3H]E2 than the spheroplasts that carried an empty vector. However, E. coli cells that expressed any of the other PDI homologs did not show appreciably ability to accumulate [3H]E2. In agreement with this observation, the periplasmic fractions of E. coli cells that expressed PDIp or PDI concomitantly showed a lower radioactivity than the control cells carrying an empty vector or those cells expressing any of the other PDI homologs (data not shown). This initial observation suggested that PDIp (likePDI) in its natural form has the ability to bind [3H]E2, whereas other PDI homologs do not share this biological property.

Then we examined the estrogen-binding activity of these six PDI homologs using purified proteins. Again, we found that only PDIp can bind [3H]E2 in a similar way as PDI, whereas other homologs can not bind [3H]E2 (Fig. 1C). In addition, when a different binding assay (charcoal adsorption) was used, the same observation was made, i.e., the [3H]E2-binding activity was only seen with PDI and PDIp (data not shown). The data from these binding analyses of purified PDI homolog proteins are in agreement with the results described above using E. coli cells (Fig. 1B).

Saturation binding assay showed that the recombinant human PDIp had a similar binding affinity as the recombinant human PDI (Fig. 2A, 2B). The nonlinear regression analysis revealed that the apparent Kd values for PDI and PDIp are 1.5 ± 0.3 μM (R2 = 0.94) and 1.5 ± 0.2 μM (R2 = 0.96), respectively. No significant difference in the binding affinity between the recombinant human PDI (1.5 μM, as determined in this study) and rat PDI (2.1 μM, as determined earlier [19]) was noted, which is in line with the high degree of similarity (93%) between their amino acid sequences.

FIGURE 2
Comparison of PDIp and PDI in their binding properties to estrogens

Next we compared the binding preference of PDIp and PDI for several biologically-important E2 metabolites or antagonists, and significant differences were observed. As shown in Fig. 2C, whereas PDIp and PDI each had a similar binding affinity for estrone, estriol, and 4-hydroxyl-E2, the latter obviously had a higher binding affinity for 2-methoxyl-E2 and 2-hydroxyl-E2. Interestingly, PDIp, unlike PDI, could not bind tamoxifen. These observations suggest a certain degree of difference in the ligand-binding pockets of PDIp and PDI. Also, both PDIp and PDI had little binding affinity for ICI-182,780, suggesting the presence of steric hindrance caused by the long C-7α side-chain of this synthetic ER antagonist.

Since Erp57 was reported earlier to act as a receptor for the active form of vitamin D3, 1,25(OH)2D3 [26, 27], we also performed the competition binding assay to determine whether D3 and 1,25(OH)2D3 can also bind to PDI and PDIp. We found that these two compounds did not have an appreciable ability to displace the binding of the radioligand [3H]E2 to PDI or PDIp, thus suggesting that they have no binding affinity for PDI and PDIp (Fig. 2C). These observations showed that the PDI family members are rather specific with respect to their ability to bind different biological ligands: while estrogens can bind to PDI/PDIp but not to Erp57, vitamin D3 can only bind to Erp57 but not to PDI and PDIp.

Besides of the estrogen-binding activity, human PDI [28] and recombinant rat PDI [19] were previously shown to have thyroid hormone-binding activity. We observed that the binding of [3H]E2 to PDIp, like its binding to PDI, could be effectively displaced by 3,3′,5-triiodo-L-thyronine (T3) in a concentration-dependent manner (Fig. 2D). The estimated relative binding activity (based on IC50 value) of PDIp for T3 was comparable to its binding activity for PDI, approximately at 4.5 μM, which is similar to its binding activity for rat PDI (4.3 μM) [19]. In addition, we found that PDIp did not have appreciable binding activity for the male sex hormone testosterone (data not shown).

3.2. The [3H]E2-binding activity of the natural PDIp protein present in human pancreas

Since the in-vitro binding activity (data shown in Fig. 1, ,2)2) was based on the recombinant human PDIp selectively-expressed in E. coli cells, next we also determined whether the PDIp present in human pancreas has the same [3H]E2-binding activity. Since earlier studies showed that high levels of PDIp are expressed in human pancreas [2124], we used human pancreas for this purpose. The pancreatic tissues were homogenized in the presence of detergent Triton X-100 that could break cellular membrane compartments (such as endoplasmic reticulum), and the tissue lysates were then incubated with mouse anti-PDIp antiserum (normal mouse serum was used as a control) to remove the PDIp protein by immunoprecipitation (Fig. 3A). Pancreatic tissue lysates after immunoprecipitation of the endogenous PDIp had approximately 40% lower specific binding activity for [3H]E2 compared to tissue lysates without removal of PDIp (Fig. 3B), indicating that PDIp present in human pancreatic tissue lysates could bind [3H]E2. Notably, the remaining [3H]E2-binding activity of the pancreatic tissue lysates after PDIp immunoprecipitation likely was attributable to the residual PDIp and also to the presence of other E2-binding proteins (such as ERα and/or ERβ) (data are described later in Fig. 7).

FIGURE 3
[3H]E2-binding activity of PDIp prepared from human pancreas

3.3. Selectively-expressed PDIp can modulate intracellular E2 concentrations in live cells

Recently we have shown that PDI can help accumulate E2 inside mammalian cells and augment ER-mediated transcriptional activity of E2 in these cells [20]. Given that PDIp is abundantly expressed in pancreas with protein concentrations at up to 0.1% of the total cellular proteins (estimated according to data in Fig. 7), we hypothesized that PDIp might also function as a modulator of intracellular E2 concentrations. To test this hypothesis, first we extended the observation in Fig. 1B by showing that PDIp could increase the total intracellular concentrations of [3H]E2 in E. coli spheroplasts when increasing concentrations of [3H]E2 were present (Fig. 4A). Consistent with this observation, a concomitant decrease in the periplasmic [3H]E2 concentrations was observed with E. coli cells that over-expressed PDIp (data not shown). For comparison, a similar accumulating effect on intracellular [3H]E2 levels was also observed with PDI, whereas Erp57 did not have this property.

FIGURE 4
PDIp modulates intracellular E2 concentrations in E. coli and mammalian cells

Next we set out to determine whether the concentrating effect of PDIp on the intracellular [3H]E2 also occurs in mammalian cells. To choose appropriate cell lines for testing, we first screened a number of cell lines, including Mia Paca-2, BxPC-3, and Capan-2 human pancreatic cancer cells, RIN–m5F rodent pancreatic cancer cells, and several other types of cancer cell lines. Different from PDI that was found to be highly expressed in each of these cell lines (Fig. S1C), PDIp was basically not detected in any of them. The lack of expression of endogenous PDIp in these cancer cell lines prevented us from using the RNA interference to knock down PDIp in these cell lines as a study model. Alternatively, we chose to over-express PDIp in cos-7 cells, which is a commonly-used mammalian cell line with a high transfection efficiency but does not express the endogenous PDIp (right part in Fig. 4B, lane 1). Following the transfection with the pcDNA3.1-PDIp plasmid, high expression of PDIp was detected in cos-7 cells (right part in Fig. 4B, lane 2). Next, we incubated cos-7 cells with [3H]E2 and measured the radioactivity associated with the cells (see Materials and Methods). We observed a strong accumulating effect of the selectively-expressed human PDIp on intracellular [3H]E2 levels (left part in Fig. 4B) when different [3H]E2 concentrations (from 5 to 100 nM) were tested, which agreed with the observations made in E. coli cells.

Since cos-7 cells are ER-negative [29], next we determined whether the [3H]E2-concentrating effect of PDIp as observed above (Fig. 4B) could be detected when ERα (a representative ER) was also present. To do so, we co-expressed PDIp and ERα in cos-7 cells and determined their effect on intracellular [3H]E2 levels. We observed that the intracellular [3H]E2 levels were significantly increased following the expression of ERα alone, and it was further increased when ERα and PDIp were co-expressed (Fig. 4C). This observation indicates that PDIp still can accumulate intracellular [3H]E2 in the presence of ERα.

3.4. PDIp can increase intracellular E2 level regardless of its localization in the cytosol or endoplasmic reticulum

The PDI family proteins are known to be associated with the endoplasmic reticulum. However, the subcellular localization of human PDIp appears unclear at present. Alignment of human PDIp amino acid sequences available in the NCBI database (Fig. 5A) showed that the amino acid sequence of PDIp deduced from the genomic DNA sequence (CAI95586.1) theoretically should contain the full-length signal peptide, but the sequence deduced from the brain cDNA sequence (AAH00537.2) actually contains a shorter signal peptide, and PDIp identified for human pancreas (based on its cDNA sequence) does not contain any signal peptide [21]. Therefore, there is a possibility that different forms of PDIp protein may be expressed in different tissues, and the protein may have different intracellular localizations (i.e., cytosol vs. endoplasmic reticulum). With this possibility in mind, we also compared in this study the E2-accumulating effect of PDIp-full (with the full-length signal peptide), PDIp-brain (with a short signal peptide), and PDIp-pancreas (without a signal peptide). The protein sequences of these three forms of PDIp are shown in Fig. 5A. We found that after the cos-7 cells were transfected with a corresponding expression plasmid for PDIp-full, PDIp-brain or PDIp-pancreas protein, each of these proteins was expressed when the whole cell lysates were analyzed using Western blotting (Fig. 5B). When the microsomal fractions prepared from the over-expressed cos-7 cells were analyzed (which also included endoplasmic reticulum membranes), much higher levels of PDIp-full and PDIp-brain proteins were detected whereas the level of PDIp-pancreas was very low (the lower part of Fig. 5B), suggesting that the former two are located in the endoplasmic reticulum whereas PDIp-pancreas is in the cytosol.

FIGURE 5
PDIp accumulates E2 in endoplasmic reticulum of mammalian cells

Next we also determined the accumulation of intracellular [3H]E2 in these over-expressed cos-7 cells. As shown Fig. 5C, the presence of each of the PDIp proteins accumulated more intracellular [3H]E2 than the control cells. The higher [3H]E2-accumulating effect of PDIp-full and PDIp-brain than PDIp-pancreas (Fig. 5C) was apparently due to their relative higher expression levels than the latter (compare lanes 3, 4, 5, 6 vs lane 2, using calnexin as a reference, Fig. 5B, upper part).

3.5. Intracellular PDIp-bound E2 can be released to augment ERα-mediated transcriptional activity

Since PDIp can selectively accumulate E2 in live cells in culture, next we sought to determine whether the intracellular PDIp-bound E2 could be released upon a drop in E2 levels and whether the released E2 could augment its biological activity. The experimental design was summarized in Fig. 6A. Briefly, PDIp was over-expressed in ER-negative cos-7 cells (cultured in 24-well plates) by transfecting the cells with the pcDNA3.1-PDIp plasmid. Thirty-six h later, cells were incubated with 20 nM E2 in 0.3 mL DMEM medium (without FBS) for 2 h to allow E2 to enter the cells. Then the E2-containing medium was removed and cells were washed once with 1 mL EMEM medium. The cells were then incubated with 0.5 mL medium of EMEM + 10% DCC-treated FBS (E2-free) for 2 h to release the intracellular PDIp-bound E2 into the medium. The medium containing the released E2 was used to culture MCF-7 cells (an ER-positive breast cancer line), and these cells were pre-transfected with a vector containing an estrogen responsive element (ERE)-driven luciferase reporter gene that enabled us to assay the ERα-mediated transcriptional activity in these cells.

FIGURE 6
Intracellular PDIp-bound E2 can be released to augment ERα-mediated transcriptional activity A

As shown in Fig. 6B, the medium collected from cos-7 cells pre-incubated with 20 nM E2 induced ERα-mediated transcriptional activity in MCF-7 cells. This effect was enhanced when MCF-7 cells were cultured by using medium collected from PDIp-overexpressing cos-7 cells, reflecting higher E2 concentrations in the medium. Assuming that the cos-7 cells overexpressing PDIp accumulate more intracellular E2 than control cos-7 cells (as observed in Fig. 4B), the former were expected to release more E2 than the latter when they were incubated with E2-free medium. Notably, the medium volume for releasing PDIp-bound E2 in this experiment was over 200 times larger than the volume of the cos-7 cells pellet, and accordingly, there was >200-fold dilution of the intracellular E2. Since the released E2 after >200-fold dilution still showed a detectable increase in ERα-mediated luciferase gene transcription, this observation suggests that the PDIp-rich cells have a rather high E2-binding capacity.

3.6. Detection of ERα and ERβ expression in pancreatic tissues

PDIp is known to be highly expressed in the pancreas of human [21, 24], sheep and rat [22], and a recent study showed that it is also expressed in mouse stomach [30]. To further probe the potential role of PDIp in modulating estrogen actions, we analyzed PDIp and also ERα and ERβ protein levels in pancreatic tissues of human (lane 1), rhesus monkey (lane 2), rat (lane 3), and mouse (lane 5), and also in mouse stomach (lane 4). As shown in Fig. 7, Western blotting showed that PDIp was highly expressed in the pancreas of all species (lanes 1, 2, 3, 5) and was also abundantly expressed in mouse stomach (lane 4). ERα (55 kD), a functional isoform of ERα [8, 31], was expressed in all pancreatic tissues examined and also in mouse stomach. In comparison, the 66-kD ERα (the regular form of ERα) was found to be only expressed in mouse pancreas (lane 5) but barely detectable in human pancreas (lane 1), but not detected in the pancreas of rat and rhesus monkey (lanes 2 and 4). This pattern of expression of ERα was confirmed with another antibody against the N-terminal domain of ERα (data not shown). For ERβ (53 kD), it was detected in all tissues examined (lanes 1–5, Fig. 7), and this observation was also confirmed by using another antibody against the N-terminal domain of ERβ (data not shown). The physiological importance of the species specificity of the 66-kD ERα isoform that was found to be highly expressed in mouse pancreas merits further study.

4. DISCUSSION

The results of our present study showed that PDIp, a pancreas-specific PDI homolog, has a similar binding activity for E2 as does PDI (Fig. 1, ,2),2), and it can function as an effective intracellular E2-binding protein and can modulate the intracellular levels of E2 in live cells (Fig. 4B, 4C, ,5C).5C). This finding suggests a new biological function of PDIp in vivo beyond as a protein-folding catalyst, and it also aids in our understanding of its modulation of estrogen’s biological functions in the pancreas, where detectable expression levels of ERα and ERβ were observed (Fig. 7).

Our finding that PDIp can selectively accumulate E2 in cultured cells suggests that that PDIp may be able to accumulate E2 in target tissues (like pancreas) that express high levels of this protein. This possibility is supported by the observation that the endogenous PDIp present in human pancreas can accumulate E2, which accounts for ~40% of total E2 bound to various cellular proteins (Fig. 3B). Moreover, the PDIp-bound intracellular E2 can be released upon a drop in the E2 levels in the extracellular compartment and the released E2 can augment ERα-mediated transcriptional activity (Fig. 6B). Based on these observations, it is suggested that the PDIp-bound E2 in pancreas may help buffer the oscillations of the levels and biological activity of endogenous estrogens in the pancreas. Since PDIp expression was also detected in some other tissues (such as stomach and uterus; shown in Fig. S2), it is likely that the estrogen-accumulating effect of PDIp may also be present in these tissues.

There is an emerging body of evidence suggesting that estrogens are associated with certain physiological functions of the pancreas. It was observed earlier that estrogens can be directly biosynthesized in canine and human pancreas [32, 33] and they can affect pancreatic zymogen production in rats [34, 35]. In line with this observation, ERα and/or ERβ were found to be expressed in pancreas in our present study (Fig. 7) and also in studies by others [36, 37]. Therefore, PDIp, as an intracellular estrogen-binding protein, may affect the physiological functions of estrogens in pancreas. In addition, estrogenic status has been suggested to be negatively associated with the progression of certain pancreatic cancers [3742], which have a higher prevalence in males than in females based on both human and animal studies [38, 4345]. Given that PDIp is expressed in pancreas at very high levels, our study may provide an important missing link for the long-standing clinical speculation that estrogenic status and actions in pancreas are one of the protective factors in the development of pancreatic cancer in humans [39]. Also, it is of interest to suggest that an unknown E2-binding protein that was detected earlier in rat pancreatic acinar cells [46, 47] most likely is PDIp in light of the observations made in our present study.

In addition, we also observed that PDIp can bind a number of endogenous E2 metabolites (Fig. 2C). This observation appears to be in line with an earlier study showing that the peptide binding activity of PDIp was inhibited by various estrogen analogs [48]. Some of these E2 metabolites not only have important biological functions in the classic ER-mediated signaling pathways [2], but they can also exert other unique biological functions that are not shared by their parent hormone E2. For instance, 2-methoxy-E2 has strong apoptotic and antiangiogenic activities [49, 50]; 4-hydroxy-E2 has unique biological activity and may have its specific intracellular binding protein [51, 52]; and estriol has a uniquely strong immunosuppressive effect [53]. The binding activity of PDIp for these bioactive estrogen metabolites may favorably increase their levels in the pancreas, and thus may enhance their biological actions in this organ. Interestingly, the results of our present study showed that PDIp has no binding affinity for tamoxifen (a widely-used ER modulator for breast cancer treatment), which is very different from PDI. This observation may be explained according to a previous study which showed that PDIp could not bind ligands with no hydroxyl group [48], since tamoxifen also does not have any hydroxyl group. It appears that the formation of hydrogen bond(s) between the hydroxyl group(s) of a ligand and PDIp is crucial for the binding interactions. PDIp, like PDI, also does not have significant binding activity for ICI-182,780 (a pure ER antagonist that has hydroxyl groups), and this observation may suggest that the binding pockets of PDIp and PDI cannot accommodate the long C-7α side-chain of this antiestrogen. Computational molecular modeling and mutagenesis studies are currently underway in our laboratory to characterize the ligand-binding pockets of PDI and PDIp, which are expected to provide structural information on their differential binding affinity for various ligands.

In summary, of the six representative PDI homologs examined in the present study, only the pancreas-specific PDI homolog (PDIp) was found to have a similar E2-binding activity as has PDI. Moreover, PDIp can effectively modulate the concentrations and hormonal activity of E2 in cells. Since ERα and ERβ were found to be co-expressed with PDIp in rodent and human pancreatic tissues, these data suggest that PDIp can also serve as an effective modulator of the intracellular levels and biological actions of endogenous estrogens in certain target sites (such as the pancreas) where the ERs and PDIp are co-present.

Acknowledgments

We would like to thank Dr. Jian Ding for assistance with the production of the mouse anti-PDIp antiserum, Ms. Pan Wang for mammalian cell culture, and the NDRI for supplying human pancreatic tissues. We appreciate Professor Carolyn Smith at Baylor College of Medicine for the gift of PCR3.1-ERα and pGL-basic-ERE-luciferase plasmids. This study was supported, in part, by a grant from the NIH (CA97109). Some of the analytical and imaging instruments employed in this study are part of the COBRE core facility that is supported by the NIH Grant P20RR021940 from the National Center for Research Resources.

Abbreviations

PDI
protein disulfide isomerase
PDIp
pancreas-specific PDI
ERα and ERβ
estrogen receptor α and β respectively
E2
17β-estrodial
T3
3,3′,5-triiodo-L-thyronine

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

1. Nilsson S, Makela S, Treuter E, Tujague M, Thomsen J, Andersson G, Enmark E, Pettersson K, Warner M, Gustafsson JA. Mechanisms of estrogen action. Physiol Rev. 2001;81:1535–1565. [PubMed]
2. Zhu BT, Conney AH. Functional role of estrogen metabolism in target cells: review and perspectives. Carcinogenesis. 1998;19:1–27. [PubMed]
3. Wu H, Chen Y, Liang J, Shi B, Wu G, Zhang Y, Wang D, Li R, Yi X, Zhang H, Sun L, Shang Y. Hypomethylation-linked activation of PAX2 mediates tamoxifen-stimulated endometrial carcinogenesis. Nature. 2005;438:981–987. [PubMed]
4. Ciocca DR, Roig LM. Estrogen receptors in human nontarget tissues: biological and clinical implications. Endocr Rev. 1995;16:35–62. [PubMed]
5. Reid G, Denger S, Kos M, Gannon F. Human estrogen receptor-alpha: regulation by synthesis, modification and degradation. Cell Mol Life Sci. 2002;59:821–831. [PubMed]
6. Zwijsen RM, Wientjens E, Klompmaker R, van der Sman J, Bernards R, Michalides RJ. CDK-independent activation of estrogen receptor by cyclin D1. Cell. 1997;88:405–415. [PubMed]
7. Razandi M, Pedram A, Merchenthaler I, Greene GL, Levin ER. Plasma membrane estrogen receptors exist and functions as dimers. Mol Endocrinol. 2004;18:2854–2865. [PubMed]
8. Li L, Haynes MP, Bender JR. Plasma membrane localization and function of the estrogen receptor alpha variant (ER46) in human endothelial cells. Proc Natl Acad Sci U S A. 2003;100:4807–4812. [PubMed]
9. Tseng L, Gurpide E. Changes in the in vitro metabolism of estradiol by human endometrium during the menstrual cycle. Am J Obstet Gynecol. 1972;114:1002–1008. [PubMed]
10. Purohit A, Tutill HJ, Day JM, Chander SK, Lawrence HR, Allan GM, Fischer DS, Vicker N, Newman SP, Potter BV, Reed MJ. The regulation and inhibition of 17beta-hydroxysteroid dehydrogenase in breast cancer. Mol Cell Endocrinol. 2006;248:199–203. [PubMed]
11. Liu ZJ, Lee WJ, Zhu BT. Selective insensitivity of ZR-75-1 human breast cancer cells to 2-methoxyestradiol: evidence for type II 17beta-hydroxysteroid dehydrogenase as the underlying cause. Cancer Res. 2005;65:5802–5811. [PubMed]
12. Zhang Y, Graubard BI, Klebanoff MA, Ronckers C, Stanczyk FZ, Longnecker MP, McGlynn KA. Maternal hormone levels among populations at high and low risk of testicular germ cell cancer. Br J Cancer. 2005;92:1787–1793. [PMC free article] [PubMed]
13. Fortunati N, Catalano MG. Sex hormone-binding globulin (SHBG) and estradiol cross-talk in breast cancer cells. Horm Metab Res. 2006;38:236–240. [PubMed]
14. Fortunati N, Fissore F, Fazzari A, Becchis M, Comba A, Catalano MG, Berta L, Frairia R. Sex steroid binding protein exerts a negative control on estradiol action in MCF-7 cells (human breast cancer) through cyclic adenosine 3′,5′-monophosphate and protein kinase A. Endocrinology. 1996;137:686–692. [PubMed]
15. Lumb RA, Bulleid NJ. Is protein disulfide isomerase a redox-dependent molecular chaperone? Embo J. 2002;21:6763–6770. [PubMed]
16. Ellgaard L, Ruddock LW. The human protein disulphide isomerase family: substrate interactions and functional properties. EMBO Rep. 2005;6:28–32. [PubMed]
17. Gruber CW, Cemazar M, Heras B, Martin JL, Craik DJ. Protein disulfide isomerase: the structure of oxidative folding. Trends Biochem Sci. 2006;31:455–464. [PubMed]
18. Tsibris JC, Hunt LT, Ballejo G, Barker WC, Toney LJ, Spellacy WN. Selective inhibition of protein disulfide isomerase by estrogens. J Biol Chem. 1989;264:13967–13970. [PubMed]
19. Primm TP, Gilbert HF. Hormone binding by protein disulfide isomerase, a high capacity hormone reservoir of the endoplasmic reticulum. J Biol Chem. 2001;276:281–286. [PubMed]
20. Fu X, Wang P, Zhu BT. Protein disulfide isomerase is a multifunctional regulator of estrogenic status in target cells. J Steroid Biochem Mol Biol. 2008;112:127–137. [PubMed]
21. Desilva MG, Lu J, Donadel G, Modi WS, Xie H, Notkins AL, Lan MS. Characterization and chromosomal localization of a new protein disulfide isomerase, PDIp, highly expressed in human pancreas. DNA Cell Biol. 1996;15:9–16. [PubMed]
22. Klappa P, Stromer T, Zimmermann R, Ruddock LW, Freedman RB. A pancreas-specific glycosylated protein disulphide-isomerase binds to misfolded proteins and peptides with an interaction inhibited by oestrogens. Eur J Biochem. 1998;254:63–69. [PubMed]
23. Dias-Gunasekara S, Gubbens J, van Lith M, Dunne C, Williams JA, Kataky R, Scoones D, Lapthorn A, Bulleid NJ, Benham AM. Tissue-specific expression and dimerization of the endoplasmic reticulum oxidoreductase Ero1beta. J Biol Chem. 2005;280:33066–33075. [PubMed]
24. Desilva MG, Notkins AL, Lan MS. Molecular characterization of a pancreas-specific protein disulfide isomerase, PDIp. DNA Cell Biol. 1997;16:269–274. [PubMed]
25. Fu X, Zhu BT. Human pancreas-specific protein disulfide isomerase homolog (PDIp) is redox-regulated through formation of an inter-subunit disulfide bond. Arch Biochem Biophys. 2009 doi: 10.1016/j.abb.2008.12.021. [PubMed] [Cross Ref]
26. Nemere I, Farach-Carson MC, Rohe B, Sterling TM, Norman AW, Boyan BD, Safford SE. Ribozyme knockdown functionally links a 1,25(OH)2D3 membrane binding protein (1,25D3-MARRS) and phosphate uptake in intestinal cells. Proc Natl Acad Sci U S A. 2004;101:7392–7397. [PubMed]
27. Nemere I, Safford SE, Rohe B, DeSouza MM, Farach-Carson MC. Identification and characterization of 1,25D3-membrane-associated rapid response, steroid (1,25D3-MARRS) binding protein. J Steroid Biochem Mol Biol. 2004;89–90:281–285. [PubMed]
28. Kitagawa S, Obata T, Hasumura S, Pastan I, Cheng SY. A cellular 3,3′,5-triiodo-L-thyronine binding protein from a human carcinoma cell line. Purification and characterization. J Biol Chem. 1987;262:3903–3908. [PubMed]
29. Ikeda M, Inoue S, Muramatsu M, Minatogawa Y. Characterization and identification of a steroid receptor-binding protein, SRB-RGS. Biol Pharm Bull. 2007;30:1056–1564. [PubMed]
30. Niki S, Oshikawa K, Mouri Y, Hirota F, Matsushima A, Yano M, Han H, Bando Y, Izumi K, Matsumoto M, Nakayama KI, Kuroda N, Matsumoto M. Alteration of intra-pancreatic target-organ specificity by abrogation of Aire in NOD mice. J Clin Invest. 2006;116:1292–1301. [PubMed]
31. Couse JF, Curtis SW, Washburn TF, Lindzey J, Golding TS, Lubahn DB, Smithies O, Korach KS. Analysis of transcription and estrogen insensitivity in the female mouse after targeted disruption of the estrogen receptor gene. Mol Endocrinol. 1995;9:1441–1454. [PubMed]
32. Iqbal MJ, Greenway B, Wilkinson ML, Johnson PJ, Williams R. Sex-steroid enzymes, aromatase and 5 alpha-reductase in the pancreas: a comparison of normal adult, foetal and malignant tissue. Clin Sci (Lond) 1983;65:71–75. [PubMed]
33. Fernandez-del Castillo C, Diaz-Sanchez V, Varela-Fascinetto G, Altamirano A, Odor-Morales A, Lopez-Medrano RM, Robles-Diaz G. Testosterone biotransformation by the isolated perfused canine pancreas. Pancreas. 1991;6:104–111. [PubMed]
34. Beaudoin AR, Grondin G, St-Jean P, Vachereau A, Cabana C, Grossman A. Steroids and the secretory function of the exocrine pancreas. Endocrinology. 1986;119:2106–2117. [PubMed]
35. Beaudoin AR, St-Jean P, Proulx J, Grondin G. Influence of steroids on the exocrine pancreas: presence of laminated bodies in the acinar lumen following castration and adrenalectomy. Pancreas. 1989;4:219–227. [PubMed]
36. Satake M, Sawai H, Go VL, Satake K, Reber HA, Hines OJ, Eibl G. Estrogen receptors in pancreatic tumors. Pancreas. 2006;33:119–127. [PubMed]
37. Singh S, Baker PR, Poulsom R, Wright NA, Sheppard MC, Langman MJ, Neoptolemos JP. Expression of oestrogen receptor and oestrogen-inducible genes in pancreatic cancer. Br J Surg. 1997;84:1085–1089. [PubMed]
38. Longnecker DS, Sumi C. Effects of sex steroid hormones on pancreatic cancer in the rat. Int J Pancreatol. 1990;7:159–165. [PubMed]
39. Morales-Miranda A, Robles-Díaz G, Díaz-Sánchez V. Steroid hormones and pancreas: a new paradigm. Rev Invest Clin. 2007:124–129. [PubMed]
40. Sumi C, Brinck-Johnsen T, Longnecker DS. Inhibition of a transplantable pancreatic carcinoma by castration and estradiol administration in rats. Cancer Res. 1989;49:6687–6692. [PubMed]
41. Sumi C, Longnecker DS, Roebuck BD, Brinck-Johnsen T. Inhibitory effects of estrogen and castration on the early stage of pancreatic carcinogenesis in Fischer rats treated with azaserine. Cancer Res. 1989;49:2332–2336. [PubMed]
42. Bourhis J, Lacaine F, Augusti M, Huguier M. Protective effect of oestrogen in pancreatic cancer. Lancet. 1987;2:977. [PubMed]
43. Lo AC, Soliman AS, El-Ghawalby N, Abdel-Wahab M, Fathy O, Khaled HM, Omar S, Hamilton SR, Greenson JK, Abbruzzese JL. Lifestyle, occupational, and reproductive factors in relation to pancreatic cancer risk. Pancreas. 2007;35:120–129. [PubMed]
44. Lowenfels AB, Maisonneuve P. Risk factors for pancreatic cancer. J Cell Biochem. 2005;95:649–656. [PubMed]
45. Parkin DM, Bray F, Ferlay J, Pisani P. Global cancer statistics, 2002. CA Cancer J Clin. 2005;55:74–108. [PubMed]
46. Grossman A, Oppenheim J, Grondin G, St Jean P, Beaudoin AR. Immunocytochemical localization of the [3H]estradiol-binding protein in rat pancreatic acinar cells. Endocrinology. 1989;124:2857–2866. [PubMed]
47. Beaudoin AR, Grondin G, St Jean P, Pettengill O, Longnecker DS, Grossman A. Marked differences in immunocytological localization of [3H]estradiol-binding protein in rat pancreatic acinar tumor cells compared to normal acinar cells. Endocrinology. 1991;128:1617–1622. [PubMed]
48. Klappa P, Freedman RB, Langenbuch M, Lan MS, Robinson GK, Ruddock LW. The pancreas-specific protein disulphide-isomerase PDIp interacts with a hydroxyaryl group in ligands. Biochem J. 2001;354:553–559. [PubMed]
49. D’Amato RJ, Lin CM, Flynn E, Folkman J, Hamel E. 2-Methoxyestradiol, an endogenous mammalian metabolite, inhibits tubulin polymerization by interacting at the colchicine site. Proc Natl Acad Sci U S A. 1994;91:3964–3968. [PubMed]
50. Fotsis T, Zhang Y, Pepper MS, Adlercreutz H, Montesano R, Nawroth PP, Schweigerer L. The endogenous oestrogen metabolite 2-methoxyoestradiol inhibits angiogenesis and suppresses tumour growth. Nature. 1994;368:237–239. [PubMed]
51. Markides CS, Liehr JG. Specific binding of 4-hydroxyestradiol to mouse uterine protein: evidence of a physiological role for 4-hydroxyestradiol. J Endocrinol. 2005;185:235–242. [PubMed]
52. Das SK, Taylor JA, Korach KS, Paria BC, Dey SK, Lubahn DB. Estrogenic responses in estrogen receptor-alpha deficient mice reveal a distinct estrogen signaling pathway. Proc Natl Acad Sci U S A. 1997;94:12786–12791. [PubMed]
53. Kotani M, Fujii H, Tsuchiya H, Matsuno K, Ekino S, Harada S. Effects of estrogen on the lymphoid regeneration and immune response in irradiated and marrow-reconstituted mice. Acta Anat (Basel) 1979;105:298–308. [PubMed]