| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Physiology (A.K., S.R., I.H., M.P.), Institute of Biomedicine, and Turku Graduate School of Biomedical Sciences (A.K.), University of Turku, 20520 Turku, Finland; and Institute of Reproductive and Developmental Biology (I.H.), Faculty of Medicine, Imperial College, London W12 ONN, United Kingdom
Address all correspondence and requests for reprints to: Matti Poutanen, Institute of Biomedicine, University of Turku, Kiinamyllynkatu 10, 20520 Turku, Finland. E-mail: matti.poutanen{at}utu.fi.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The wnt glycoproteins are highly conserved, secreted short-distance signaling molecules that bind to receptors of the Frizzeled and lipoprotein receptor-related protein families and induce ß-catenin stabilization (canonical pathway) or act through the Wnt/Ca2+ or Wnt/polarity pathways (11). The intracellular wnt effector, ß-catenin, is mutated or its degradation pathway is altered in many pathological conditions, including several types of cancers (12). ß-Catenin is also found stabilized in human breast cancer (13, 14), but its mutations have not been found in this malignancy. However, deregulated WNT expression and epigenetic silencing of Wnt-pathway inhibitors Wif-1 and Sfrp1 have been found in human breast malignancies (15, 16, 17, 18, 19), and an autocrine mechanism resulting in constitutive activation of the wnt pathway in human breast cancer cells has been identified (20). Ectopic wnt signaling induces the accumulation of mammary progenitor cells and increases susceptibility to tumorigenesis (21).
Wnts are involved in the normal development of the mammary gland (22, 23) and some of them are dependent on ovarian steroids. Wnt expression in the mammary gland is spatiotemporally regulated and altered in the induction of different developmental stages (23, 24). Classically, Wnts encode transforming and nontransforming wnt proteins classified according to their potential to transform mammary epithelial cells in vitro (25). However, the classification does not necessarily correlate with transforming potential in vivo (26).
Studies on Wnt1- and Wnt10b-expressing transgenic (TG) mice have confirmed the role of these effectors as potent growth stimulators for the mammary gland epithelium (27, 28). Wnt4 has been shown to mediate progesterone receptor (Pgr) actions in the mammary gland (29). Wnt2, -5a, -7b and -10b are expressed during ductal development and Wnt4, -5b, and -6 during pregnancy. Expression of the wnt proteins is down-regulated during lactation, whereas Wnt2, 5a, 5b, and 7b are reinduced upon involution (23, 24). The precise nature of hormonal regulation of the Wnts still remains poorly understood. It has been shown that ß-catenin functionally interacts with estrogen receptor-
(Esr1) (30) and that estrogen up-regulates Wnt4 and -5a expression in an Esr1-independent manner in the mouse uterus (31). Recently Wnt7b mRNA was described to be enriched in terminal end buds (TEBs) in the mammary gland, and Wnt5b has been detected in both TEBs and mature ducts (32).
Several TG mouse models with activated Wnt/ß-catenin pathway have been created. These include mouse mammary tumor virus reverse transcriptase-long-terminal repeat-driven ligands Wnt1, Wnt10b (27), and activated ß-catenin mutants (33, 34, 35). These models present with similar mammary gland tumor histopathology, as do the TG mouse models stabilizing ß-catenin by affecting its destruction complex (Apcmin and dominant-negative form of GSK3ß, the dnGSK3ß mice) (36, 37). Yet another mouse model recently described, Epimorphin/syntaxin-2, shares this phenotype (38). One of the most striking features of the wnt pathway tumors is the transdifferentiation of the mammary epithelium into epidermal structures and formation of pilar tumors (36, 39, 40).
We previously described the generation of the TG mouse model overexpressing hCG ß-subunit (hCGß) under the ubiquitin C promoter (41). Dimerization of the TG hCGß subunit with endogenous -subunit in the pituitary gland results in 20- to 30-fold elevated levels of LH/hCG bioactivity in circulation. In short, the mice develop precocious puberty and have elevated levels of estrogen during puberty. This is followed by marked ovarian luteinization, formation of pituitary prolactinomas, and subsequent progressive increase in serum levels of progesterone and prolactin. The hCGß-expressing female mice are obese and infertile and develop mammary gland tumors from the age of 9 months onward, with 90% penetrance at the age of 12 months. All the gross phenotypes of these mice are initially directly or indirectly due to alterations of ovarian function because they all are abolished by ovariectomy at the age of 1–2 months.
In this study, we demonstrate that the endocrinologically induced mammary gland tumors in the hCGß TG mice are histologically similar to those induced by activation of the wnt pathway and demonstrate a link between hCG action and wnt dysregulation in the mammary gland.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
hCG injections
Twelve-week-old FVB/N females were ovariectomized as above, and a pellet releasing estradiol (
0,19 mg/kg/d) and progesterone (P; 4.8 mg/kg·d; Innovative Research of America, Sarasota, FL) or placebo was inserted under the skin. After 10 d, the animals were injected with 20 IU of recombinant hCG sc (provided by A. F. Parlow, National Hormone and Peptide Program, Torrance, CA) or saline for 5 consecutive days and then killed.
Histological and morphological analysis of the mammary gland
For histological studies, the fourth left inguinal mammary gland from the TG females and litter mates or tumor foci from the TG mice were dissected out at different ages between 1 and 12 months of age and fixed overnight in 4% paraformaldehyde. The tissues were dehydrated, embedded in paraffin, and cut into 5-µm-thick sections. Sections were stained with hematoxylin and eosin. For whole-mount preparations, the fourth right inguinal mammary gland from TG and WT litter mates was removed, spread on a glass slide, and fixed overnight with Carnoy’s fixative (acetic acid-ethanol). The slides were then washed with ethanol and distilled water, stained with carmine-alum for 1–2 d, dehydrated in a series of ethanol and xylene, and finally mounted in Permount.
Immunohistochemistry
Five-micrometer-thick paraffin sections were mounted on slides, deparaffinized, and rehydrated in a series of xylene and ethanol. Antigen retrieval was performed by boiling the slides in antigen unmasking solution (Vector Laboratories, Inc., Burlingame, CA) for 15 min in a microwave oven (850 W) and keeping them for 20 min in hot solution. After washing with PBS + 0.1% Tween 20, the slides were treated with 1% H2O2 for 15 min. The sections were incubated in PBS + 3% BSA overnight at + 4 C with the antibodies against one of the following antigens: 1) estrogen receptor 1, dilution 1:200 (clone 1D5, Dako A/S, Glostrup, Denmark); 2) progesterone receptor, dilution 1:400 (Dako, A0098); 3) -smooth muscle actin, dilution 1:1000 (clone 1A4, Sigma, St. Louis, MO); 4 and 5) keratin 5 and 6, dilution 1:200 (BabCo, Richmond, CA); and 6) ß-catenin, dilution 1:100 (Transduction Laboratories Inc., Lexington, KY). As the secondary antibody, Dako EnVision antimouse or antirabbit secondary for primary antibodies 1–3 were used at room temperature for 30 min, followed by visualization with 3'3'-diaminobenzidine (Dako). Finally, the slides were stained with Gill’s hematoxylin followed by dehydration and mounting. For antibodies 4–6 (Molecular Probes, Eugene, OR), AlexaFluor 594 antimouse and antirabbit secondary antibodies were used (1:400 dilution at 37 C for 1 h) followed by staining with 4',6'-diamino-2-phenylindole for 15 min. Fluorescent slides were mounted in DakoCytomation fluorescent mounting medium. For fluorescent pictures, pseudocolors were created by Adobe Photoshop CS (San Jose, CA).
RNA isolation and quantitative (q) RT-PCR
The fourth mammary gland or mammary gland tumor was excised and snap frozen in liquid nitrogen. Total RNA was extracted with a lipid tissue minikit (QIAGEN, Valencia, CA) and treated with amplification grade DNase I (Invitrogen, Carlsbad, CA). For cDNA synthesis and subsequent qRT-PCR, the SYBR Green DyNAmo HS qRT-PCR kit (Finnzymes, Espoo, Finland) was used. For each qPCR, an aliquot of cDNA (diluted 1:20) was used except for the housekeeping genes in which 1:40 dilution was used. qRT-PCR analysis was performed using the DNA Engine Opticon system (MJ Research, Inc., Waltham, MA) with continuous fluorescent detection. Samples and standards were run in triplicates. The two housekeeping genes (L19 and cyclophilin) were analyzed to normalize the results between the samples. All primers used (Table 1
) were intron spanning. The presence of a single, right-sized PCR product was confirmed by running the samples on 2% agarose gels.
|
Statistical analysis
SigmaStat 3.1 (SPSS Inc., Chicago, IL) was used for the t test or the Mann-Whitney rank sum test. The limit of statistical significance was set at P < 0.05. The results are presented as mean ± SEM.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
), increased lateral side branching and budding was observed in mammary gland ducts of the TG females from the age of 1 month as a response to precocious hormonal activity (Fig. 1b). Budding was evident from primary and secondary ducts and hyperbranching was detected (Fig. 1, c and d). At 3 months of age, the TG mammary gland structurally resembled mid- to late-pregnancy mammary glands with apparent alveolar development.
|
, a and b). The lobuloalveolar structures increasingly replaced the fat pad, but the organization of alveoli remained irregular with secretion trapped inside the glands. In contrast to diffuse alveolar hyperplasia, the first atypical focal lesions with no secretory activity were present from month 3 onward. These low- and high-grade foci of mammary intraepithelial neoplasias (MINs) presented with several layers of epithelium, filling the lumen with hyperplastic cells and loss of cell polarity (Fig. 2c). These masses of nonlactating epithelium became more abundant over time, and progressive keratinization was seen histologically in these areas at and beyond 5 months of age (Fig. 2d). The TG mammary glands expressed mRNA of the milk protein ß-casein, but the secretory activity of the tumors was suppressed. Relative mRNA expression of ß-casein is shown in Table 2.
|
|
), pilar (37%) (Fig. 2g), glandular (19%) (Fig. 2f), and acinar (7%) were the main subtypes in 27 independent tumor samples studied, but most often tumor areas presented with a combinations of these subtypes together with myoepithelial differentiation and abundant stroma. The tumor growth-induced inflammatory responses and pushing margins were observed at the growing borders of the tumors (Fig. 2g). Aggressive growth was apparent with rapidly growing tumor mass, and necrosis was frequently found in the center of the tumors. Keratinizing nodules preceded by ghost cell formation were often present in the tumors and precursor lesions (Fig. 2g). Based on histological criteria, solid tumors typical to ErbB/Ras TG pathways were never observed.
The hCGß mammary glands present with abnormal, inconsistent, and scattered myoepithelium layer
-Smooth muscle actin (SMA) was used as the myoepithelial marker. It is normally expressed in the monolayer of the myoepithelium lining the ducts and alveoli, as shown in the mammary glands of WT mice (Fig. 3a). In TG mice, continuous -SMA staining was detected around the secretory alveoli (Fig. 3c), but staining in the hyperplastic areas was either totally absent (Fig. 3b), increased, or inconsistently scattered (Fig. 3e). -SMA staining also often lined the aberrant hyperplastic ductal structures (Fig. 3d). Myoepithelial differentiation was consistently found in tumor areas, which is a typical feature of tumors of the wnt pathway (Fig. 3f).
|
). From the age of 3 months onward, Esr1 could not be detected by immunohistochemistry (IHC) in the TG mice (Fig. 3h) in contrast to WT mammary glands. Developing tumors were also Esr1 negative (Fig. 3i). Pgr was normally expressed in the TG mammary gland at 2 months of age (Fig. 3j), but thereafter the secreting alveolar structures did not express Pgr (Fig. 3k). However, Pgr expression could be detected in some of the nonsecreting MIN foci, and occasional Pgr positive areas were observed even in tumors, although most of the tumor cells were Pgr negative (Fig. 3l). No androgen receptor expression was found in the TG mammary glands or mammary tumors (data not shown).
Transdifferentiation of the mammary glands of the hCGß-TG mice
Keratin 6 is normally expressed in the epidermis and some TEB structures but not in the mature mammary gland, as demonstrated by absent staining in WT controls (Fig. 4a). In TG mice, keratin 6 was expressed in single cells or small groups of cells at 3–5 months of age in 11 of 12 samples studied, well before the formation of apparent keratin nodules or tumors (Fig. 4, b and c). Keratin 6 was expressed as a uniform layer in all of the tumors studied by IHC in the keratin-forming tumor foci in the cytoplasms of the cells surrounding the acellular keratin core (Fig. 4, d and e). Thus, the tumors showed transdifferentiation into epidermal and pilar structures.
|
, inset) and detected in TG mice in the same manner as SMA staining, scattered and disorganized and as a continuous lining in alveoli of the TG mice (Fig. 4, g and h). In the tumors, keratin 5 was expressed in an organized pattern lining the squamous metaplastic nodules in one or several cell layers (Fig. 4, i and j).
ß-Catenin is stabilized in hyperplastic areas
By immunofluorescence, ß-catenin staining was observed in the cell membranes of the TG and WT mice in which it is bound in catenin-cadherin complexes (Fig. 5, a and b). However, increased staining of ß-catenin could be identified in the cytoplasm of the hyperplastic areas of 3-month-old TG mice (Fig. 5c), whereas the secreting alveoli did not stain strongly for this protein (data not shown). At later stages, strong ß-catenin staining was observed in areas of squamous metaplasia (Fig. 5, d and e).
|
). Histological staining confirmed the presence of alveoli and collapsed structures filled with cellular debris (Fig. 2h). Furthermore, abundant fibrosis or epithelial to mesenchymal transformation was observed in histology. We performed keratin 5 and 6 IHC on these sections and observed reduced but persistent expression of keratin 5 and 6 in the TG mammary epithelium in single cells often lining the cysts or small groups (Fig. 4, f and k). However, pilar-type areas with excessive keratinization were absent. The pituitaries of the TG mice were regressed to normal size, suggesting reduced levels of prolactin in serum. Data on ovariectomized mice indicate that these potentially premalignant changes remained in the mammary gland and are ovarian independent. Thus, the damage previously caused by hormones cannot be fully reversed by withdrawal of ovarian hormones and prolactin at later stages.
hCGß mice show disturbances in mammary gland Wnt expression
We next addressed the question whether any of the Wnts would be up-regulated and potentially responsible for the activation of the canonical wnt pathway leading to the phenotype. We analyzed the mRNA expression of the Wnts naturally expressed in the mouse mammary gland, including Wnt2, -4, -5a, -5b, -6, -7b and -10b. Wnt5b was significantly up-regulated, reaching a 5-fold higher expression in TG animals at the age of 3 months, as compared with age-matched WT littermates. The expression was further increased in the TG tumors that expressed 9-fold higher levels of Wnt5b than the 3-month-old WT mice (Fig. 6A). Wnt7b is normally expressed in TEBs, and the expression is decreased in pregnancy. In our study, Wnt7b was up-regulated in the tumors, showing more than 2-fold higher mRNA expression as compared with WT mammary glands. Three-month-old TG mammary glands expressed Wnt7b mRNA at level similar to WT controls (Fig. 6A). Protein level of Wnt7b in tumors was increased 4-fold (Fig. 6C), compared with WT mice (P = 0.002), whereas its expression was similar in 3-month-old TG and WT mice. Wnt6 was not overexpressed in 3-month-old animals. In four of six tumor specimens studied, Wnt6 expression was nearly absent (Fig. 6A), but in two samples, it was clearly up-regulated.
|
). The other Wnts included in the study, Wnt2, -5a, and -10b, all produced in the stroma, exhibited decreased expression in the tumor samples and similar or decreased expression in 3-month-old TG mammary glands (not shown), a finding that could be attributed to a decreased epithelial/stromal ratio. Wnt1 expression was not found in any mammary gland. Overall, the expression pattern of Wnts in the hCGß mammary glands did not resemble the natural expression pattern seen during virgin state, pregnancy, or lactation.
Wnt5b, Wnt7b, and classical wnt targets are up-regulated in the mammary gland in response to hCG stimulation without ovarian contribution
To elucidate the cause of up-regulation of wnt5b and wnt7b in the TG mice, we injected ovariectomized and estradiol- and P-treated WT mice with recombinant hCG or saline. Interestingly, we observed 2- and 2.5-fold up-regulation in wnt5b and wnt7b expression in the mammary gland, respectively (Fig. 6B), without apparent histological changes. This suggests that hCG is directly responsible for regulation of wnt5b and wnt7b in the mouse mammary gland. However, it is possible that effects in the mammary gland are secondary changes due to other unknown systemic alterations because LH receptor expression in the mammary gland is low although present. Together with the induced expression of wnt7b and wnt5b, we observed a significant up-regulation of various well-characterized direct wnt target genes, such as autotaxin, axin2, cyclin D1, and Hig2 (Table 3). Injections with prolactin did not affect the expression of these wnts (data not shown). Wnt5b and wnt7b expression in the uterus was unchanged by hCG administration.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Ovariectomy at 6 wk of age abolishes the mammary gland phenotype at a gross morphology level in the hCGß mice (41), thus mimicking endocrine induction of carcinogenesis in human breast tumors. Oncogenic damage cannot, however, be fully reversed by ovariectomy from the age of 3 months onward, showing the true ovarian independence of these lesions. In line with this, the subsequently appearing tumors are Esr1 and Pgr negative, which is also more typical for breast cancer associated with pregnancy (42, 43).
The hCGß mice resemble most closely the TG mice expressing stabilized ß-catenin and CK2 with the uniform keratin expression pattern and the extent of transdifferentiation (39). Keratin 6 expression has been demonstrated in the body cells of the terminal end buds (32, 44) and during lobuloalveolar growth of early pregnancy, whereas mature mammary epithelial cells rarely express it (45). Esr1 expression varies in mouse models with activated wnt/ß-catenin and depends on the additional oncogenic mutations (46). Esr1 is, however, not definitively needed for wnt-induced tumorigenesis (47).
Wnt5b and -6 are up-regulated during pregnancy, secreted from the mammary gland epithelium, and show a similar expression pattern in normal pregnancy (24). The transformation potential of Wnt5b in mammary gland cell culture is controversial (25, 48). Furthermore, the ability to stabilize ß-catenin is context dependent because classical noncanonical ligand Wnt5a has been shown to induce ß-catenin stabilization in mammary epithelium (49). Wnt5b was overexpressed in the hCGß TG mammary glands at 3 months of age and was still increased in tumors despite the lack of ß-casein expression and nonsecreting tumor cell phenotype. This suggests a potential role for Wnt5b in mammary tumorigenesis of the hCGß mice and differential mechanisms in the regulation of Wnt5b and -6 expression. The expression of wnt5b and wnt7b early in the tumorigenesis suggests that additional genetic mutations are needed for malignant conversion. We have shown here a previously unknown mechanism of hCG promoting wnt5b and -7b expression in the mammary gland without changes in ovarian steroidogenesis. Furthermore, selective expansion of Wnt5b and -7b-producing cells during tumorigenesis was observed in TG females. Wnt5b and Wnt7b expression has been detected near the TEBs in virgin epithelium (23, 32), suggesting a role in progenitor cell regulation. In humans, the WNT5B gene could confer susceptibility to diseases because two common haplotypes show significant association with type 2 diabetes (50).
It has been suggested that Wnt7b could be involved in maintaining mammary cells in an uncommitted state (23). Interestingly, canonical ligand WNT7B has been found to be up-regulated in breast malignancies (16, 51). Because of the midpregnancy mammary phenotype observed in 3-month-old TG mice, Wnt7b was expected to be down-regulated (23), and therefore the presence of the Wnt7b-producing cell population was surprising. Wnt7b has been classified into the transforming class of wnt proteins, but it failed to induce hyperplasia in vivo when retrovirally expressed in mammary glands (26). Furthermore, Wnt7b expression is not regulated by ovarian steroids (23) but was found to be up-regulated by hCG in this study. Wnt7b is found up-regulated in 10% of breast carcinomas (16).
Wnt4 has been recognized as an essential mediator of Pgr function (29). Wnt4 expression is induced by P and estrogen treatment (29), but because of the lacking Esr1 and Pgr expression in the TG mammary glands and in tumors, wnt4 was not found to be up-regulated.
The data suggest a possible mechanism in which initial hCG stimulation drives the expression of Wnt5b, Wnt7b, and wnt target genes, thus contributing to ß-catenin stabilization, transdifferentiation, and eventual tumorigenesis in the hCGß females. Although the data show dysregulation of wnt 5b and -7b mRNAs, an alternative mechanism for ß-catenin stabilization cannot be excluded in the presence of several hormonal disturbances because the canonical wnt pathway can also be activated through several G protein-coupled receptors without wnt contribution (52).
Activation of the wnt pathway has been shown to induce mammary gland tumors from progenitor cells (53). This is consistent with the tumor histology, in which differentiation along the luminal and myoepithelial cell lineages is observed. Consistent myoepithelial differentiation in the wnt pathway tumors has a counterpart in human breast cancers because 17–37% of them are classified as the basal subtype, characterized by positive keratin 5/6 and vimentin staining but negative ESR1 and HER2 status. ß-Catenin is stabilized in over 50% of human breast carcinomas and belongs therefore to an important molecular pathway involved in breast malignancies (13, 14). Mutations in the ß-catenin signaling pathway are well documented in other cancers (12), but despite abundant ß-catenin stabilization in human breast cancer, such mutations have not been found, indicating a functionally hyperactive pathway. Recent findings on epigenetic silencing of Wnt pathway inhibitors in breast cancer (15, 19) support this view, together with the studies showing a constitutive autocrine Wnt activation in breast cancer cell lines (20). Three-month-old TG mammary glands and the TG tumors in aged mice expressed a pattern of Wnt genes that did not resemble any normal developmental state of this organ.
In summary, hCG induces wnt signaling in the mammary gland. Accordingly, the hCGß-expressing TG mice present with hormonally induced mammary tumors exhibiting all the typical characteristics of wnt pathway tumors and thus serve as a model of hormonally induced ß-catenin-stabilizing breast cancers. The phenotype is initially hormone induced and is associated with precocious puberty and transiently increased circulating levels of estradiol, followed by high levels of serum P and prolactin. Hence, the mice are subjected to an induction mechanism in which altered endocrine factors and breast tumorigenesis are connected with abnormal wnt signaling and ß-catenin stabilization, a common feature in human breast cancer. Disturbed wnt signaling is partially a direct consequence of elevated hCG, and this novel mechanism could mimic the poorly known mechanism of pregnancy-associated breast cancer and its induction. We can thus conclude that the current mouse model provides novel information about hCG action and hormonally induced basal-type tumorigenesis of the breast.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Disclosure Summary: The authors have nothing to disclose.
First Published Online May 17, 2007
Abbreviations: Esr1, Estrogen receptor-; hCG, human chorionic gonadotropin; hCGß, hCG ß-subunit; IHC, immunohistochemistry; MIN, mammary intraepithelial neoplasia; P, progesterone; Pgr, progesterone receptor; q, quantitative; SMA, smooth muscle actin; TEB, terminal end bud; TG, transgenic; WT, wild type.
Received February 23, 2007.
Accepted for publication May 4, 2007.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
N89ß-catenin induces precocious development, differentiation, and neoplasia in mammary gland. J Cell Biol 153:555–568. Cancer Res 59:1869–1876This article has been cited by other articles:
 |
|
 |
|
  F. Sotgia, M. C. Casimiro, G. Bonuccelli, M. Liu, D. Whitaker-Menezes, O. Er, K. M. Daumer, I. Mercier, A. K. Witkiewicz, C. Minetti, et al. Loss of Caveolin-3 Induces a Lactogenic Microenvironment that Is Protective Against Mammary Tumor Formation Am. J. Pathol., February 1, 2009; 174(2): 613 - 629. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Fernandez-Valdivia, A. Mukherjee, C. J. Creighton, A. C. Buser, F. J. DeMayo, D. P. Edwards, and J. P. Lydon Transcriptional Response of the Murine Mammary Gland to Acute Progesterone Exposure Endocrinology, December 1, 2008; 149(12): 6236 - 6250. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
