SEARCH

SEARCH BY CITATION

Keywords:

  • chick pituitary;
  • hormones;
  • Rathke's pouch;
  • SHH;
  • Fgf10;
  • cLim3;
  • gene expression

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS AND DISCUSSION
  5. EXPERIMENTAL PROCEDURES
  6. Acknowledgements
  7. REFERENCES

The chick embryo is an ideal model to study pituitary cell-type differentiation. Previous studies describing the temporal appearance of differentiated pituitary cell types in the chick embryo are contradictory. To resolve these controversies, we used RT-PCR to define the temporal onset and in situ hybridization and immunohistochemistry to define the spatial localization of hormone expression within the pituitary. RT-PCR detected low levels of Fshβ (gonadotropes) and Pomc (corticotropes, melanotropes) mRNA at E4 and Gh (somatotropes), Prl (lactotropes), and Tshβ (thyrotropes) mRNA at E8. For all hormones, sufficient accumulation of mRNA and/or protein to permit detection by in situ hybridization or immunohistochemistry was observed ∼3 days later and in all cases corresponded to a notable increase in RT-PCR product. We also describe the expression patterns of signaling (Bmp2, Bmp4, Fgf8, Fgf10, Shh) and transcription factors (Pitx1, Pitx2, cLim3) known to be important for pituitary organogenesis in other model organisms. Developmental Dynamics 239:1197–1210, 2010. © 2010 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS AND DISCUSSION
  5. EXPERIMENTAL PROCEDURES
  6. Acknowledgements
  7. REFERENCES

The mature pituitary gland exerts critical roles in vertebrate homeostasis by integrating signals from the hypothalamus to regulate metabolism, reproduction, and growth through the production and appropriate release of peptide hormones. Six major hormone-producing cell-types of the mature pituitary gland reside within the adenohypophysis (anterior and intermediate lobes). Each cell type can be identified based on the hormone that it synthesizes and secretes: corticotropes secrete adrenocorticotropin (ACTH) and melanotropes secrete melanocyte-stimulating hormone (MSH), both of which are derived from cell-type-specific processing of the pro-hormone, pro-opiomelanocortin (POMC); gonadotropes secrete luteininzing hormone (LH) or follicle-stimulating hormone (FSH); thyrotropes secrete thyroid-stimulating hormone (TSH); somatotropes secrete growth hormone (GH); and lactotropes secrete prolactin (PRL).

Development of the pituitary gland has been well described at a morphological level in several species and histological studies have led to a model of pituitary organogenesis that is relatively well conserved between species (Jacobson et al.,1979; Stefanovic et al.,1993; Pikalow et al.,1994; Dubois et al.,1997). Fate-mapping studies in chick and Xenopus (Couly and Le Douarin,1988; el Amraoui and Dubois,1993; Eagleson et al.,1995) demonstrated that the cells that contribute to the adenohypophysis originate at the ventral lip of the anterior neural ridge (Jacobson et al.,1979; Stefanovic et al.,1993; Pikalow et al.,1994; Dubois et al.,1997) and migrate to the region of the oral epithelium that will thicken to form the hypophyseal placode (reviewed in Baker and Bronner-Fraser,2001). Upon head folding, the oral ectoderm cells of the hypophyseal placode invaginate towards the ventral diencephalon giving rise to Rathke's pouch, the anlage of the adenohypophysis (Jacobson et al.,1979; Stefanovic et al.,1993; Pikalow et al.,1994; Dubois et al.,1997). Following elongation and closure of the ventral aspect of Rathke's pouch, cells within it undergo extensive proliferation and, subsequently, differentiate into the six hormone-secreting cell types.

Our understanding of the molecular events that lead to the differentiation of specific cell types is still incomplete. The ventral diencephalon, which is in direct contact with Rathke's pouch throughout its development, has been identified as one of the primary sources of signaling molecules that direct cell-type differentiation within Rathke's pouch. Explant studies using mouse and chick tissues and gain- and loss-of-function experiments in mice, zebrafish, and Xenopus have demonstrated that signals emanating from the ventral diencephalon are essential for the proper differentiation and expansion of cell types within the pituitary gland (reviewed in Dasen and Rosenfeld,2001). In addition, the appropriate expression of transcription factors in the ventral diencephalon and Rathke's pouch is essential for the correct spatial and temporal appearance of terminally differentiated cell types (Dasen and Rosenfeld,2001; Savage et al.,2003).

The chick embryo offers an ideal model system to further refine our understanding of the role of signaling molecules and transcription factors in pituitary cell type differentiation (reviewed in De Groef et al.,2008). During the initial stages of pituitary development, both the hypophyseal placode and Rathke's pouch are accessible and a number of different techniques to manipulate gene expression exist, including the introduction of retroviral vectors, electroporation of expression plasmids, antisense and morpholino treatment, and the possibility of implanting protein-coated beads or cell pellets expressing signaling molecules of interest. However, the analysis of the effects of these manipulations on cell-type differentiation would be limited by the considerable controversy that exists regarding the temporal appearance of the individual differentiated cell types in the chick embryo. For example, somatotropes are reported to appear as early as embryonic day (E) 4.5 of embryogenesis (Thommes et al.,1987) or as late as E16 (Sasaki et al.,2003), lactotropes as early as E6 (Gasc and Sar,1981; Thommes et al.,1987) or as late as E20 (Sasaki et al.,2003), thyrotropes as early as E6.5 (Thommes et al.,1983) to as late as E10 (Muchow et al.,2005), gonadotropes as early as E4 (Kameda et al.,2000) or as late as E10 (Allaerts et al.,1999), and finally POMC expression has been reported to initiate from E7 (Sasaki et al.,2003) to E12 (Allaerts et al.,1999). These studies have focused primarily at the level of protein expression and often the analysis was restricted to only a subset of the six differentiated cell types.

To address these issues, we performed an extensive analysis of the mRNA and protein expression profiles of the hormones that define the six differentiated cell types of the chick adenohypophysis. We initiated our studies by performing RT-PCR analysis of pituitary hormone expression between E4 and E18. We then characterized the spatial distribution of these hormones by in situ hybridization and immunohistochemistry. The expression patterns of multiple hormones were characterized within a single pituitary to permit a more precise characterization of the differences in spatial distribution of hormones. This is the first comprehensive report of the mRNA and protein-expression profiles of these hormones and the first report of the ontogeny of MSH expression in the chick pituitary gland. In addition, we examined the mRNA expression patterns between E3 and E14 of five signaling molecules (Bmp2, Bmp4, Fgf8, Fgf10, and Shh) and three transcription factors (cLim3, Pitx1, and Pitx2) that have been shown to have critical roles in the development of the adenohypophysis.

RESULTS AND DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS AND DISCUSSION
  5. EXPERIMENTAL PROCEDURES
  6. Acknowledgements
  7. REFERENCES

RT-PCR Analysis of Hormone mRNA Expression During Pituitary Development

We initiated our study by characterizing pituitary hormone mRNA expression using RT-PCR analysis. Expression was analyzed for each day from E4 to E9 and then every second day from E10–E18. Prior to E14, dissected pituitaries were pooled to permit analysis of all hormones from a single sample. At E4 and E5, the block of tissue that was dissected also contained ventral diencephalon as well as some adjacent mesoderm, in addition to Rathke's pouch.

RT-PCR analysis was performed on total RNA extracted from E4 limb bud, E4 Rathke's pouch, and E14 pituitary gland to test the specificity of the primers (Fig. 1). Only the Pomc-specific primers yielded an RT-PCR product in the E4 limb bud sample, which was similar in size and quantity to the RT-PCR product observed in the E4 Rathke's pouch sample. Pomc mRNA has been observed at an equivalent stage of mouse limb development, where ACTH, which is encoded by the Pomc transcript, is proposed to have a role in myoblast proliferation (De Angelis et al.,1993). Pomc and Fshβ were the only RT-PCR products present in the E4 Rathke's pouch sample, while all primer pairs amplified an appropriately sized product from the E14 pituitary RNA. In the RT-PCR analysis performed on the poly(A)+ RNA, low levels of Fshβ, Tshβ, and Pomc RT-PCR products were detected at E4 (Fig. 1). Fshβ and Pomc expression showed a marked increase at E8 and then continued to increase gradually until E18, the last stage examined in our study. Two RT-PCR products were detected using the Tshβ-specific primers: a ∼475-bp fragment (E4–E10) and the expected 404-bp fragment (E8–E18; Fig. 1). Sequencing of the products confirmed that the 404-bp fragment was indeed Tshβ, while the 475-bp product corresponded to a sequence on chromosome 1 that encodes a putative RNA binding protein (NW_001471513.1). This sequence contains a 11/12 nucleotide match to the 3′ half of the forward primer and a 100% match to the last 13 nucleotides of the reverse primer. This product was not observed in the total RNA sample and was not observed after E10. Tshβ expression increased substantially between E9 and E10 and then remained relatively constant throughout the remainder of embryogenesis. Prl and very low levels of Gh mRNA were detected at E8. There was a marked increase in Gh expression between E9 and E10, after which levels of Gh mRNA continued to increase throughout the remainder of embryogenesis. A marked increase in Prl mRNA expression was noted only at E16. Interestingly, the timepoint of the major increase in hormone mRNA expression observed by RT-PCR corresponded to when we were able to detect the hormone in the pituitary gland by in situ hybridization or immunohistochemistry (see below). Gapdh mRNA levels remained constant at all stages examined.

thumbnail image

Figure 1. RT-PCR analysis of hormone gene expression during embryonic development of the chick pituitary gland. Agarose gel electrophoresis of RT-PCR products from total or mRNA isolated from embryonic day (E) 4 limb bud (L) or E4–E6 Rathke's pouch and E8–E18 pituitary gland showed that Pomc and Fshβ mRNA were present from E4 and Gh and Prl were present from E8 onwards. Two products were observed for Tshβ, the expected 404-bp product was present from E8 onwards, and the 475-bp product (indicated by *) was present only from E4–E10. Gapdh was detected at all stages examined.

Download figure to PowerPoint

Temporal Appearance and Spatial Organization of Gonadotropes.

FSH and LH, which are expressed in gonadotropes, and TSH, which is expressed in thyrotropes, are heterodimeric proteins each comprised of a common α glycoprotein subunit (αGSU) and a hormone-specific β-subunit, FSHβ or LHβ in gonadotropes, and TSHβ in thyrotropes. We examined the protein expression patterns of both the α- and β-subunits of FSH and TSH by in situ hybridization and immunohistochemistry. Figure 2a shows a schematic of sagittal sections through Rathke's pouch and the pituitary gland at representative timepoints in development. Using in situ hybridization, we detected αGsu mRNA in the ventral diencephalon at E3 (data not shown) and in Rathke's pouch beginning at E4.5 (Fig. 2b). As previously described by Takagi et al. (2008), αGsu mRNA exhibited an extremely dynamic expression pattern between E4.5 and E7 (Fig. 2). Initially, cells expressing αGsu mRNA were localized to the dorsal half of the posterior side of Rathke's pouch. Beginning at E4.5, we observed αGsu RNA initially in the dorsal posterior side of Rathke's pouch and then at E5 a second domain of expression in the oral ectoderm at the anterior boundary of Rathke's pouch (Fig. 2b, 2e1, and data not shown). At E5.5, αGsu expression was maintained in the ventral-anterior boundary of Rathke's pouch and was now observed in the mid-region of the posterior side of Rathke's pouch (Fig. 2b2). By E6, significant αGsu expression was limited to the cells that were beginning to acquire a differentiated glandular cell morphology (Fig. 2b3). At E7 and E8, clusters of αGsu-expressing cells were concentrated in the morphologically differentiated cells at the lateral edges of the pituitary and rarely observed in the cells immediately adjacent to the remaining cleft of Rathke's pouch (Fig. 2c1, 2c2, 2e2, 2e3). From E11 through E18, αGsu-expressing cells were scattered throughout the entire gland and there was an enrichment of positive cells in the most ventral portion of the cephalic lobe (Fig. 2c3, 2e4, 2e5). αGsu-positive cells were present in the pars tuberalis (data not shown and Fig. 2e4, 2e5). The pattern of αGSU protein expression was identical to its mRNA expression throughout embryogenesis (Fig. 3a) although significant levels of αGSU protein were not detected in the pituitary until E8, more than two days after its mRNA was observed in Rathke's pouch (Fig. 3a). Initially, αGSU-positive cells were detected scattered throughout the cephalic and caudal lobes of the pituitary gland although a higher percentage of cells within the caudal lobe appear to be αGSU-positive. At E18, the αGSU-expressing cells in the cephalic lobe were enriched at the periphery (Fig. 3a6, 3a7).

thumbnail image

Figure 2. mRNA expression patterns of pituitary hormones during embryonic development of the chick pituitary gland. a: Schematic diagram showing sagittal sections through Rathke's pouch (rp) and/or pituitary gland at selected timepoints during chick embryogenesis. Neuroectoderm of the ventral diencephalon is shown in green (vd) and the pituitary gland is shown in blue. The pars tuberalis (pt) is indicated by the blue lines in the top right panel. The black dotted lines in the E11-14 and E15-18 drawings indicate the boundary between the caudal (Ca) and cephalic (Ce) lobes. The white vertical and horizontal dotted lines show the plane of section for a transverse and coronal section, respectively. b–i: Digoxigenin-labeled antisense riboprobes for αGsu (b,c,e), Fshβ (d), Pomc (f), Tshβ (g), Gh (h), and Prl (i) were hybridized to sagittal (S) or transverse (T) 14-μm cryosections as indicated. Positive cells are dark blue in color. All sagittal sections are oriented so that rostral is to the right and dorsal is up. Transverse sections are oriented so that the dorsal aspect of the gland is at the top of the image. pt, pars tuberalis; rp, Rathke's pouch; vd, ventral diencephalon. Scale bars = 100 μm.

Download figure to PowerPoint

thumbnail image

Figure 3. Protein expression patterns of pituitary hormones during embryonic development of the chick pituitary gland. Immunohistochemical analysis was performed on 8-μm paraffin sections using antibodies against αGSU (a), FSHβ/LHβ (b), TSHβ (c), ACTH (d), MSH (e), GH (f), and PRL (g). The ages of the embryos from which the sections were obtained is indicated in the top panel of each column. All sections are sagittal, except for a6-g6, which are transverse (T) and a7-g7, which are coronal (C). Sagittal sections are oriented so that dorsal is up and rostral is to the right. Transverse sections are oriented so that dorsal is up and the coronal sections are oriented so that the cephalic (ce) lobe is at the bottom and the caudal (ca) lobe is at the top. Brown staining indicates positive immunoreactivity. Insets show a higher magnification view of cells positive for hormone immunoreactivity. Scale bars = 100 μm.

Download figure to PowerPoint

To identify differentiated gonadotropes within Rathke's pouch and the developing adenohypophysis, we performed in situ hybridization with an antisense riboprobe for Fshβ (Fig. 2d) and immunohistochemistry with an anti-human FSHβ/LHβ antibody that recognizes both FSHβ and LHβ (Fig. 3b). At E7, Fshβ mRNA-positive cells were rare (Fig. 2d1). A similar result was obtained using the FSHβ/LHβ antibody (Fig. 3b1). At E8, an increased number of cells expressing higher levels of FSH mRNA and protein were observed in the ventral region of the gland (Figs. 2d2 and 3b2). As expected, the FSHβ/LHβ expression domain overlapped with the αGSU-positive cell population expression. Gonadotropes appeared to originate in the cephalic lobe, but from E10 onwards, FSHβ-expressing cells were enriched in the caudal lobe, although there were still a significant number of FSHβ-positive cells in the cephalic lobe, as previously reported (Kameda et al.,2000; Puebla-Osorio et al.,2002). Similarly to αGSU, cells immediately adjacent to the residual cleft did not express FSHβ/LHβ and positive cells appeared to be concentrated at the lateral edges of the gland and the pars tuberalis as observed in the transverse section through the cephalic lobe (Fig. 3b6).

Kameda and colleagues observed αGsu mRNA in Rathke's pouch as early as E3.5 by in situ hybridization and LH-immunoreactivity with an antibody that recognized both the α- and β-subunits has been observed as early as E4, although significant staining in the morphologically differentiated cells was not seen until E6 (Gasc and Sar,1981; Kameda et al.,2000). In contrast, studies using antibodies specific for the chicken β-subunits detected immunoreactivity of LHβ protein between E8 and E10 (Allaerts et al.,1999; Puebla-Osorio et al.,2002; Maseki et al.,2004) and FSHβ protein at E7 (Maseki et al.,2004) or E13 (Puebla-Osorio et al.,2002). Our finding that differentiated gonadotropes expressing the β-subunits were not present until E7 is in agreement with reports from Sasaki et al. (2003) and Maseki et al. (2004) indicating that the anti-human FSHβ/LHβ antibody that we used is as effective at detecting β-subunit immunoreactivity as the anti-chick LHβ and FSHβ antibodies used in the other studies. Collectively, these data suggest that the immunoreactivity observed by Kameda and colleagues at E3.5 using the LH antibody that recognizes both α and β subunits most likely corresponds to αGSU. The discrepancy in the onset of αGsu mRNA expression between our study and that of Kameda et al.'s (2000) may in part be due to the staging of the embryos. In both studies, the age of the embryos is based on the number of days that the fertilized eggs have been incubated. However, the actual rate of development of the embryos can be affected by several factors, including how long after laying the eggs were placed in the incubator and the exact temperature of the incubator. It is clear from the images that Rathke's pouch in their E4 embryos is morphologically more advanced than the Rathke's pouch in our studies.

Temporal Appearance and Spatial Organization of Thyrotropes

In situ hybridization analysis detected a few Tshβ-positive cells at E9 (Fig. 2g1). However, significant numbers of cells strongly expressing Tshβ mRNA were not observed until E10 (Fig. 2g2), which corresponds to the age at which a marked increase in Tshβ mRNA was observed by RT-PCR (Fig. 1). Between E10 and E18, Tshβ-expressing cells were localized throughout the cephalic lobe but essentially restricted from the caudal lobe of the adenohypophysis (Fig. 2g2-4). To characterize the protein expression of TSHβ during chick embryogenesis, we used an antibody raised against rat TSHβ whose specificity was previously validated for use in the chick (Muchow et al.,2005) (Fig. 3c). A few clusters of cells expressing the TSHβ protein were detected at E10 in the cephalic lobe. However, significant levels of protein were not observed until E12 (Fig. 3c4). Between E12 and E16, the relative number of TSHβ-positive cells in the ventral region increased and they were now present throughout the cephalic lobe (Fig. 3c4,5 and data not shown). At E18, the majority of the TSHβ-positive cells were localized at the ventral and lateral edges of the cephalic lobe (Fig. 3c6,7). In contrast to αGSU and FSHβ/LHβ, TSHβ-positive cells were not observed in the pars tuberalis, indicating that all of the αGSU-positive cells located in this region of the gland are likely to be gonadotropes.

Although differentiated thyrotropes have been reported in the chick embryo as early as E6.5-E7, studies using an antibody raised against chick TSHβ were unable to detect TSH protein at these early stages (Murphy and Harvey,2001). Our observation that TSHβ was first expressed at E10 is in agreement with these data and data published by Sasaki et al. (2003) and Nakamura et al. (2004) who used an antibody raised against the chick hormone. In fact, studies on dispersed pituitary cells indicate that prior to E11, less than 0.7% of the cells in the pituitary gland are TSHβ-positive (Muchow et al.,2005).

Temporal Appearance and Spatial Organization of Corticotropes and Melanotropes

The Pomc gene encodes the pro-hormone POMC that is processed into several distinct and/or overlapping peptides, including the non-overlapping peptide hormones γMSH and ACTH. In situ hybridization analyses using antisense riboprobes for the entire Pomc mRNA or only the Acth region produced identical expression patterns (Fig. 2f1-5 and data not shown). We detected Pomc mRNA in Rathke's pouch beginning at E7, although Pomc-expressing cells were observed in the nervous system at E5 (Fig. 2f1,2 and data not shown). We believe that the RT-PCR products observed between E4 and E5 are likely derived from the contaminating ventral diencephalon in our dissected tissue. At E7, Pomc mRNA expression was observed in cells throughout the gland, with cells in the dorsal region appearing to express higher levels (Fig. 2f2). At E8, Pomc was highly expressed in cells localized to a rostral domain that will contribute to the cephalic lobe (Fig. 2f3). From E10 onwards, Pomc-expressing cells were localized to the cephalic lobe of the pituitary (Fig. 2f4,5). Only rare Pomc-positive cells were ever detected in the caudal lobe (Fig. 2f4,5).

Using an antibody raised against human ACTH, we detected ACTH-immunoreactive cells beginning at E7 (Fig. 3d). At E7 and E8, ACTH protein expression was restricted to a subset of cells in the dorsal-anterior portion of the pituitary gland (Fig. 3d1,2). Between E10 and E18, the number of ACTH-positive cells increased and they were localized primarily to the cephalic lobe (Fig. 3d3-7). At E18, ACTH-immunoreactive cells were evenly distributed throughout the cephalic lobe, in contrast to αGSU-, FSHβ/LHβ-, and TSHβ-positive cells, which were enriched at the lateral edges of the gland (Fig. 3). Our data are in agreement with the onset of ACTH expression described by Puebla-Osorio et al. (1999) and Sasaki et al. (2003) using antibodies that were raised against chick and porcine ACTH, indicating that the anti-human ACTH antibody has similar specificity and sensitivity.

POMC is also processed into the 12–amino acid peptide hormone γMSH in melanotropes. In the mouse, MSH-positive cells are localized to the intermediate lobe, while in chick, which lacks an intermediate lobe, the melanotropes are intermingled with the other cell types in the anterior lobe of the pituitary gland. As expected, the pattern of MSH-expression overlapped with that observed for Pomc (Fig. 2f). In contrast to the majority of differentiated cell types that were initially observed at the lateral, dorsal, or ventral boundaries of the pituitary, MSH-positive cells originated in the middle of the gland (Fig. 3e). At E8, only rare MSH-positive cells were observed in the region of the pituitary that will become the cephalic lobe. From E10 onwards, the number of MSH-positive cells increased and they were scattered throughout the cephalic lobe. The apparent number of MSH-positive cells was less than the number of ACTH-positive cells (Fig. 3e3-7). To our knowledge, this is the first description of γMSH expression in the embryonic chick pituitary gland.

Temporal Appearance and Spatial Organization of Somatotropes and Lactotropes

Somatotropes and lactotropes were the last cell types to differentiate. By in situ hybridization, low levels of Gh mRNA were detected in the caudal lobe at E11 (Fig. 2h1). A slight increase in the level of Gh mRNA was seen at E12 (Fig. 2h2) and the level of Gh expression was substantially increased by E16 (Fig. 2h3). The spatial distribution of Gh-positive cells was unchanged throughout development and only a few scattered cells were ever observed in the cephalic lobe (Fig. 2h). Using an antibody against human GH, we observed low levels of GH protein in a few scattered cells at E13 and E14, one day after its mRNA was detected (Fig. 3f5, inset; and data not shown). At E15, there were more positive cells but the signal in each cell was low; only at E18 was a significant amount of GH protein detected in individual cells (Fig. 3f6,7). As seen with Gh mRNA, GH protein was essentially restricted to the caudal lobe with only a very few positive cells present in the cephalic lobe.

There is considerable variation in reports documenting the onset of growth hormone (GH) expression in the pituitary, beginning as early as E4.5 (Gasc and Sar,1981; Thommes et al.,1987; Kansaku et al.,1994; Porter et al.,1995; Allaerts et al.,1999; Sasaki et al.,2003). However, studies that detected extra-pituitary GH in kidney, heart, neural tube, and lungs between E3 and E7 did not detect GH in the pituitary at these same stages (Harvey et al.,2000; Murphy and Harvey,2001). Even using the highly sensitive technique of RT-PCR, we were not able to detect Gh mRNA expression prior to E8. Our in situ hybridization data is in agreement with the majority of previous reports that describe GH immunoreactivity in rare cells beginning between E10 and E12, followed by a significant increase in the number of positive cells after E14 (Porter et al.,1995; Woods and Porter,1998; Allaerts et al.,1999). We observed only very low levels of Gh expression prior to E12 and noted a significant increase between E12 and E14.

We detected Prl mRNA by RT-PCR at E8 and observed a substantial increase in the amount of Prl RT-PCR product after E14 (Fig. 1). Using in situ hybridization, we detected Prl mRNA in the pituitary beginning at E14 (Fig. 2i1), coinciding with the increase in Prl mRNA at E14 observed in our RT-PCR analysis. A steady increase in the intensity of the Prl signal was observed at E16 (Fig. 2i2) and E18 (Fig. 2i3). By immunohistochemistry, only a few PRL-positive cells per pituitary gland were detected prior to E15 (Fig. 3g5, inset). By E18, Prl-expressing cells were observed throughout the entire cephalic lobe and were essentially absent from the caudal lobe (Fig. 3g6,7). Most studies report that lactotrope differentiation occurs between E15 and E19 (Kansaku et al.,1994; Woods and Porter,1998; Allaerts et al.,1999; Harvey et al.,2000; Sasaki et al.,2003). The fact that we detected Prl mRNA at E14 by in situ hybridization likely reflects a delay in the accumulation of significant amounts of PRL protein.

In mammals, lactotropes and somatotropes have been proposed to be derived from a common progenitor cell known as the somatolactotrope, which is capable of expressing both GH and PRL. In the chick, Gh and Prl mRNA are observed in distinct subdomains of the embryonic pituitary gland from the onset of their expression. Based on the distinct localization of GH and PRL in the caudal and cephalic lobes, respectively, our data clearly support previous studies that suggest that these two cell types are unlikely to originate from a common somatolactotrope progenitor capable of producing both hormones (Fu et al.,2004; Zheng et al.,2006). However, these data do not preclude the possibility that somatotropes and lactotropes arise from a common “pre-hormone” progenitor that becomes distributed throughout both lobes of the pituitary gland prior to hormone expression. In our study, we noted a few GH-producing cells in the cephalic lobe and PRL-producing cells in the caudal lobe and it is a possibility that these cells could represent a small population that expresses both hormones. The temporal and spatial appearance of differentiated cell types has been summarized (see Fig. 6).

Expression of Bmps, Fgfs, and Shh During Pituitary Development

Studies in mice have provided evidence for three extra-pituitary signaling centres that participate in the development and differentiation of cell types in the pituitary gland: the ventral diencephalon and its derivative the infundibulum, the oral ectoderm that is initially contiguous with Rathke's pouch, and the mesenchyme surrounding the pouch (Takuma et al.,1998; Gleiberman et al.,1999; Dasen and Rosenfeld,2001). Several signaling molecules, including BMP2, BMP4, the BMP antagonist noggin, FGF8, FGF10, and SHH have been shown to be important for cell type differentiation. Therefore, we examined the mRNA expression patterns of these factors during the embryonic development of the chick pituitary gland beginning at E3, when Rathke's pouch is formed until E14, at which point all differentiated cell types can be identified based on hormone mRNA expression (Fig. 4). These data have been summarized (see Fig. 6).

thumbnail image

Figure 4. Expression of Bmp2, Bmp4, Noggin, Fgf8, Fgf10, and Shh during initial morphogenesis of the chick pituitary gland. Digoxigenin-labeled antisense riboprobes for Bmp4 (a,b), Bmp2 (c–e), Noggin (f), Fgf8 (g–i), Fgf10 (j–l), and Shh (m–o) were hybridized to sagittal sections obtained from embryonic chick pituitary glands. Blue staining indicates positive regions of expression. oe, oral epithelium; rp, Rathke's pouch; vd, ventral diencephalon. Scale bars = 100 μm.

Download figure to PowerPoint

In the mouse embryo, Bmp4 is expressed in the region of the ventral diencephalon, which is in direct contact with Rathke's pouch during its formation and the initial phases of cell type differentiation (Ericson et al.,1998; Treier et al.,1998). In the chick, Bmp4 mRNA was detected in the ectoderm surrounding the branchial arches and in the heart but not in Rathke's pouch nor in the ventral diencephalon at E4 and E5 (Fig. 4a). At E6, Bmp4 expression was observed in the ventral diencephalon and at lower levels in the oral ectoderm that connects Rathke's pouch to the oral cavity (Fig. 4b). However, significant levels of Bmp4 expression were not noted at other stages of development (data not shown). A few cells expressing low levels of Bmp2 were observed in the rostral side of Rathke's pouch and in the caudal mesenchyme at E4 in some embryos (Fig. 4c and data not shown) but this expression was significantly lower than its expression in the heart, which was located on the same section (data not shown). Between E7–E10, Bmp2-positive cells were observed dispersed throughout the gland (Fig. 4d and data not shown), and at E14, the latest stage examined in our studies, there appeared to be a very low level of Bmp2 expression throughout the pituitary (Fig. 4e). This expression pattern is similar to that observed in the mouse, where Bmp2 is initially observed in the surrounding mesenchyme at the rostral side of Rathke's pouch, the ventral region of Rathke's pouch, and subsequently throughout the developing pituitary gland (Ericson et al.,1998; Treier et al.,2001; Davis and Camper,2007).

Studies in mice have shown that a BMP signal is essential for formation of Rathke's pouch and for corticotrope differentiation (Takuma et al.,1998; Treier et al.,1998). The expression pattern in chick suggests that a BMP signal is unlikely to be required for Rathke's pouch formation. However, the onset of Bmp4 expression in the chick ventral diencephalon at E6 immediately precedes Pomc/ACTH expression in the pituitary at E7 suggesting that a BMP4 signal may be important for differentiation of corticotropes in the chick. The role for Bmp2 expression in the posterior mesenchyme may have a role in initiating αGsu expression, which is first detected in the caudal aspect of Rathke's pouch. Interestingly, the expression and expansion of Bmp2 expression in the pituitary at E7 correlates with the onset of αGSU protein expression throughout the gland. In the mouse, expression of the BMP antagonists noggin, in the ventral diencephalon during Rathke's pouch formation and subsequently in the cartilage that underlies the developing pituitary (Davis and Camper,2007), and chordin, in the mesenchyme surrounding Rathke's pouch, are proposed to generate a gradient of BMP signaling that may be important for the positional determination of dorsal and ventral cell types in the developing pituitary gland (Treier et al.,1998). We did not observe noggin expression in the mesenchyme surrounding Rathke's pouch in the chick, although it was expressed at low levels throughout the pituitary gland (Fig. 4f and data not shown). The expression of both Bmp2 and its antagonist in the pituitary suggests that the role of BMP signaling at these stages in the chick may be quite general as opposed to a specific role in directing cell type differentiation.

Expression of the fibroblast growth factor family in the ventral diencephalon and infundibulum has been implicated in promoting proliferation and differentiation of corticotropes and melanotropes during mouse pituitary development (Ericson et al.,1998; Treier et al.,1998; Norlin et al.,2000) and for specification of the pituitary gland in zebrafish (Herzog et al.,2004). In mice, disruption of Fgf10 signaling results in a poorly formed Rathke's pouch that undergoes apoptosis, resulting in the absence of the pituitary gland (De Moerlooze et al.,2000; Norlin et al.,2000). We examined the expression of Fgf8 and Fgf10 during the initial phases of pituitary development. Between E3.5 and E5.5, Fgf8 and Fgf10 were expressed in a mutually exclusive pattern in the ventral diencephalon. Fgf10 was expressed in the portion of the ventral diencephalon immediately adjacent to the dorsal half of Rathke's pouch (Fig. 4j), while Fgf8 was expressed in a more rostral domain of the ventral diencephalon and in the underlying oral ectoderm that is contiguous with the rostral Rathke's pouch ectoderm (Fig. 4g). At E5.5, Fgf8 was not expressed in Rathke's pouch or in cells immediately adjacent to it (Fig. 4h). Similarly to Bmp2, low levels of Fgf8 were observed throughout the pituitary gland at E12 (Fig. 4i), suggesting that it also may be acting as a general growth factor promoting growth/survival of all cells in the gland rather than the differentiation of a particular cell type. Fgf10 expression in the ventral diencephalon adjacent to Rathke's pouch continued until E7.5, at which point its expression was limited to the region that was in direct contact with the caudal-most aspect of the developing pituitary (Fig. 4l). Fgf10 expression was never observed in Rathke's pouch nor in the developing pituitary gland. In mouse, FGFs have been proposed to establish a dorsal-ventral gradient of signaling important for the differentiation of cell types within Rathke's pouch. Our data suggest that FGF signaling may be important for establishing a rostral-caudal signaling gradient in the chick that ultimately could affect the pattern of cell-type differentiation in the cephalic and caudal lobes.

Several lines of evidence indicate that SHH signaling plays a critical role in pituitary development. In chick and zebrafish, disruption of SHH signaling results in the transdifferentiation of the pituitary gland into lens tissue (Ede and Kelly,1964; Lewis et al.,1999; Kondoh et al.,2000). In mice, loss of Shh expression leads to an absence of midline structures, including Rathke's pouch (St-Jacques et al.,1998). Similarly, antagonizing SHH signaling in mice by overexpression of Hedgehog Interacting Protein (HIP) or incubating zebrafish in the presence of cyclopamine disrupts pituitary development (Treier et al.,2001; Herzog et al.,2003; Sbrogna et al.,2003). In contrast, overexpression of SHH in Rathke's pouch in mice results in an increased number of thyrotropes and gonadotropes (Treier et al.,2001), while in zebrafish, SHH overexpression increases the number of prolactin- and growth hormone–secreting cells (Herzog et al.,2003; Sbrogna et al.,2003).

At E3.5, we observed Shh expression in the ventral diencephalon that was adjacent to Rathke's pouch (Fig. 4m) with cells that are adjacent to the ventricle displaying significantly higher levels of Shh than the cells that were physically closer to Rathke's pouch. At E5.5, Shh mRNA was also expressed in the oral ectoderm that was contiguous with Rathke's pouch and in cells in the ventral diencephalon but not in neuroectodermal tissue that was in direct contact with Rathke's pouch (Fig. 4n). A similar pattern between E3 and E4.5 was recently reported by Sjödal and Gunhaga (2008) and Takagi et al. (2008). These studies also showed that at earlier stages Shh is expressed in tissues that are involved in the earliest inductive events in pituitary development including the mesoderm underlying the hypophyseal placode and in the ectoderm anterior and posterior of the pouch but not in the pouch. From E7 onwards, we observed a low level of Shh expression in the pituitary, albeit at a significantly lower level than its expression in the diencephalon (Fig. 4o).

In the mouse, Shh mRNA is reported to be expressed in the oral ectoderm prior to Rathke's pouch invagination and subsequently in the ventral diencephalon and in the oral ectoderm that is contiguous with Rathke's pouch, but not in the pouch itself (Treier et al.,2001), while in zebrafish, Shh is expressed only in the ventral diencephalon (Herzog et al.,2003; Sbrogna et al.,2003). Although Shh was not observed in the chick oral ectoderm prior to Rathke's pouch formation, its expression pattern in the chick following formation of Rathke's pouch is quite similar to that observed in the mouse at comparable developmental stages and, therefore, suggests that SHH signaling may have a role in cell-type differentiation during pituitary development. Given that the highest levels of Shh were observed in the region of the diencephalon closest to the dorsal-anterior portion of the gland, we would predict that it is most likely to affect the cell-type differentiation within the cephalic lobe. A recent study in zebrafish has shown that the dynamic expression of Gli activator and repressors are required to interpret the SHH signal in order to obtain the correct number and localization of terminally differentiated cells within the pituitary gland (Devine et al.,2009). Characterization of the Gli transcription factors in the chick pituitary will undoubtedly provide insight into how the potential gradient of SHH signaling is translated into changes in gene expression required for terminal differentiation of hormone-producing cell types.

Temporal Appearance of cLim3, Pitx1, and Pitx2 During Chick Pituitary Development

Many different transcription factors have been shown to be important for the development and differentiation of the pituitary gland (Watkins-Chow and Camper,1998; Dasen and Rosenfeld,2001; Savage et al.,2003). The first transcription factor to be restricted to Rathke's pouch is Lhx3/Lim3 (Bach et al.,1995; Zhadanov et al.,1995). Deletion of the Lhx3 genomic locus in mice results in the loss of all pituitary cell types except corticotropes (Sheng et al.,1996), while deletion of both Lhx3 and Lhx4 blocks pituitary development at an earlier stage, such that a definitive Rathke's pouch is never formed (Sheng et al.,1997). As in mouse, the onset of cLim3/Lhx3 expression was coincident with invagination of Rathke's pouch at E3 (Fig. 5a). Unlike mouse, where Lhx3 expression is observed throughout Rathke's pouch, in chick Lim3 expression was initially observed only on the rostral side of Rathke's pouch, the portion of the pouch that is in direct contact with the ventral diencephalon. This is the opposite side of the pouch from where αGsu is first detected. In addition, cLim3 expression extended rostrally into the oral ectoderm but not caudally into Seessel's pouch. At E5, low levels of cLim3 expression were observed in both the anterior and posterior walls of Rathke's pouch (Fig. 5b) and at E7, as the pituitary starts to take on a more glandular appearance, cLim3 appeared to be expressed in all cells of the pituitary, including the cells lining the cleft (Fig. 5c). cLim3 expression in both the caudal and cephalic lobes continued until at least E14, the last day examined in our study (Fig. 4d and data not shown). A similar expression pattern for cLim3 was recently described by Sjödal and Gunhaga (2008) and Takagi et al. (2008).

thumbnail image

Figure 5. Expression of cLim3, Pitx2, and Pitx1 in the pituitary between E3 and E10. Digoxigenin-labeled antisense riboprobes for cLim3 (a–d), Pitx2 (e–h), and Pitx1 (i–l) were hybridized to sagittal sections from E3.5 (a,e,i), E5.5 (b,f,j), E7.5 (c,g,k), and E10 (d,h,l) pituitaries. All sections are oriented so that rostral is to the right. Blue staining indicates positive expression. Scale bars = 100 μm.

Download figure to PowerPoint

thumbnail image

Figure 6. Summary of spatial and temporal expression patterns of signaling molecules and differentiated cell types during chick adenohypophysis development. a: Schematic diagram of chick adenohypophysis development showing the expression patterns of Bmp2, Bmp4, Fgf8, Fgf10, noggin, and Shh. Pink stars indicated low level Bmp2 expression in the caudal mesenchyme and in rare cells within the rostral side of Rathke's pouch. Higher levels of Shh in the neuroectoderm of the ventral diencephalon and oral ectoderm are indicated by the dark purple while lower levels observed in Rathke's pouch at E7.5 are indicated by the lighter purple. b: Summary diagram showing the spatial localization and temporal appearance of pituitary hormones during development of the chick adenohypophysis. For simplicity, αGsu mRNA and protein are indicated by the same symbol. Cells expressing FSHβ and TSHβ will also express αGSU as indicated by the blue ring outlining the yellow and light blue circles, respectively.

Download figure to PowerPoint

Explant studies in mice have shown that FGF8 is capable of inducing Lhx3 (the cLim3 homologue) expression and inhibiting Isl1 expression (Ericson et al.1998). As described above, we observed Fgf10 expression in the region of the ventral diencephalon that was immediately adjacent to the domain of rostral/anterior aspect of Rathke's pouch that expresses cLim3. Therefore, we would predict that a FGF10 signal that emanates from the chick ventral diencephalon is responsible for inducing cLim3 in the adjacent domain of Rathke's pouch.

In the mouse, the LIM-homeodomain protein Islet-1 (Isl1) is expressed in the ventral portion of Rathke's pouch and has a role in thyrotrope differentiation (Ericson et al.,1998; Treier et al.,1998). In the chick, Isl1 expression was observed in the ectoderm that is posterior to Rathke's pouch at E3 and in the ventral posterior pouch at E4.5 (Sjödal and Gunhaga,2008). This expression pattern immediately precedes the onset of αGsu expression in the posterior wall of Rathke's pouch (Fig. 2b2, 2e1; Takagi et al.,2008). As development proceeds, increasing numbers of Isl1-positive cells are detected in the differentiating glandular cells of the cephalic lobe and in scattered cells in the caudal lobe (Liu et al.,2005). Isl1 immunoreactivity was limited to the gonadotropes and thyrotropes (Liu et al.,2005).

Other transcription factors, such as Pitx2, are required for the early development of Rathke's pouch and the subsequent differentiation of cell types (Gage et al.,1999; Lin et al.,1999; Suh et al.,2002). The closely related Pitx1 is required for maintaining appropriate numbers of gonadotropes and thyrotropes (Lanctot et al.,1999; Szeto et al.,1999). The absence of both Pitx1 and Pitx2 results in a severely hypoplastic pituitary gland and an absence of Lim3/Lhx3/PLim expression (Charles et al.,2005). Pitx1 and Pitx2 exhibited virtually identical expression patterns throughout development of Rathke's pouch and cell-type differentiation, although the absolute level of Pitx1 expression seemed to be lower (Fig. 5e–j). At E3 through E5, Pitx1- and Pitx2-expressing cells were detectable throughout Rathke's pouch and extended caudally into Seessel's pouch and rostrally into the oral ectoderm (Fig. 5e,f). At E3, the level of Pitx2 expression in Seessel's pouch and the oral ectoderm was somewhat lower than Pitx2 expression in Rathke's pouch (Fig. 5e). However, at E5, expression in these three sites was equivalent (Fig. 5f,j). In addition, Pitx2 expression was observed in a patch of mesenchymal cells immediately dorsal to Rathke's pouch at E3.5 and E5.5 (Fig. 5e,f) and in the ventral diencephalon in all stages examined (Fig. 5e and data not shown). From E7 onwards, essentially all cells of the cephalic and caudal lobes of the pituitary appeared to be positive for Pitx1 and Pitx2 expression (Fig. 5g,h and data not shown), similarly to cLim3 (Fig. 5c,d,g,h).

We observed that at E3.5, cLim3 expression is limited to the rostral side of Rathke's pouch. The fact that αGSU-positive cells (this study and Kameda et al.,2000), LH-positive cells (Kameda et al.,2000), and cLim3 cells also appear in spatially restricted zones of Rathke's pouch suggests that temporary expression gradients may exist in the chick. Given our finding that cLim3 is initially only expressed in the anterior aspect of Rathke's pouch and not in the posterior wall, it is tempting to speculate that we will uncover other anterior-posterior (rostral-caudal) gradients of gene expression, rather than the dorsal-ventral gradients of transcription factor expression observed in the Rathke's pouch of mice. The localization of hormone-producing differentiated cell types within distinct domains of the caudal and cephalic lobes of the chick embryonic pituitary suggests that early rostral-caudal (anterior-posterior) gradients of transcription factors and signaling molecules may provide early patterning information for hormone progenitor cells that will ultimately populate the cephalic and caudal lobes of the chick adenohypophysis. It will be interesting to determine if we will uncover a relationship between putative zones of transcription factor expression in Rathke's pouch and the subdomains of Rathke's pouch described by Sasaki and colleagues (Sasaki et al.,2003).

EXPERIMENTAL PROCEDURES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS AND DISCUSSION
  5. EXPERIMENTAL PROCEDURES
  6. Acknowledgements
  7. REFERENCES

Embryo Collection

Fertilized White Leghorn eggs (Réal Côté Inc., Quebec, Canada) were incubated in a humidified rotating incubator at 38.5°C. Embryos were handled according to the Canadian Council on Animal Care guidelines. For RT-PCR analysis, dissected tissue was immediately frozen in liquid nitrogen. For E3–E5 embryos, a block of tissue was isolated that contained the complete Rathke's pouch, the portion of ventral diencephalon that is directly attached to the rostral and dorsal sides of Rathke's pouch as well as some surrounding mesenchyme. From E6 onwards, pituitary tissue was collected separately from the ventral diencephalon and mesenchyme. At E3–E6, tissue was pooled from 5-6 embryos, at E7–E15 tissue was pooled from 2-3 embryos and individual pituitaries were collected from E16–E18 embryos. For in situ hybridization, embryos were fixed in formalin at 4°C for a minimum of 18 hr and then washed in 5 and 30% sucrose/PBS. Cryosections (12-16 μm) were collected on Fisherbrand Superfrost Plus microscope slides and stored at −80°C. For immunohistochemistry, embryos were fixed in a solution of ethanol:water:formaldehyde (6:3:1) for 1.5 hr, washed in 70% ethanol, and stored at −20°C. Tissues were dehydrated through a graded series of alcohol washes, transferred into xylene, and embedded in paraffin. Sections (5-10 μm) were collected on Fisherbrand Superfrost Plus microscope slides.

RT-PCR Analysis and cDNA Cloning

Total RNA was isolated from E4 and E14 chick tissue using the Absolutely RNA Isolation Kit (Stratagene, La Jolla, CA). Poly A+ RNA was isolated from E3–E18 chick tissue using the Micro-FastTrack 2.0 Kit (Invitrogen, Carlsbad, CA) and RT-PCR was performed using the QIAGEN (Chatsworth, CA) One-Step RT-PCR kit. Sufficient material was obtained such that all RT-PCR reactions could be performed using the same aliquot of mRNA. The sequence of the oligonucleotide primers used for PCR amplification and the size of the amplified product are shown in Table 1. Reverse transcription was performed at 50°C for 30 min followed by an initial PCR activation step at 95°C for 15 min. Forty cycles of a 3-step PCR cycle consisting of a 1-min denaturing step at 94°C, a 1-min annealing step at 50°C (αGsu) or 55°C (Pomc, Fshβ, Tshβ, Gh, and Prl) and a 1-min extension step at 72°C were performed. RT-PCR products were inserted into the TOPO TA Cloning© vector (Invitrogen). Constructs were sequenced to confirm their identity and orientation.

Table 1. Oligonucleotide Primers Used for RT-PCR
Hormone/geneNCBI gene ID no.Primer sequenceSize of RT-PCR product (bp)
αGSUGI: 15088356Forward 5′-ATGGATTGCAGGAAGTATGC-3′363
  Reverse 5′-TTAAGATTTATGATAGTACGAGG-3′ 
FSHβGI: 374108Forward 5′-ATGAAGACACTTAACTGTTAT-3′396
  Reverse 5′-TCATTGATTGCTTCCATTGTGACTG-3′ 
GapdhGI:46048960Forward 5′-GGTGGTGCTAAGCGTGTTA-3′524
  Reverse 5′-ACCATTGAAGTCACAGGAGACAA-3′ 
GHGI: 211808Forward 5′-ATGGCTCCAGGCTCGTGG-3′650
  Reverse 5′-TCAGATGGTGCAGTTGCTCTCTCC-3′ 
POMCGI: 3869132Forward 5′-ATGCGGGGGGCGCTGTGC-3′754
  Reverse 5′-TCACTGACCCTTCTTGTAGG-3′ 
ProlactinGI: 15294318Forward 5′-ATGAGCAACAGAGGGGCTTC-3′685
  Reverse 5′-TTAGCAATTGCTATCATGG-3′ 
TSHβGI: 2660744Forward 5′-ATGAGTCCCTTCTTCATGATGTCTCTCC-3′404
  Reverse 5′-TCACATGTTACAGAGCTTCTGTGG-3′ 

In Situ Hybridization Analysis of mRNA Expression Patterns

cDNA templates were linearized and antisense digoxigenin-labeled riboprobes were synthesized using the appropriate RNA polymerase according to standard protocols. Probes were resuspended in hybridization solution at a concentration of 0.5-1 μg/ml. Sections were hybridized with antisense riboprobes overnight at 70°C in a humid box as previously described (Simard et al.,2005). Detection of AP reactivity conjugated to anti-DIG antibody was done with a 5-bromo-4-chloro-3-indolyl-phosphate/4-nitroblue tetrazolium (BCIP/NBT) substrate system (Roche, Nutley, NJ). Slides were placed in water to stop the reaction, dehydrated through a graded series of ethanol and xylene washes, and coverslipped.

Immunohistochemical Detection of Pituitary Hormones

Immunohistochemistry was performed as previously described (Simard et al.,2006) using 1:200 dilution of polyclonal antisera to rat αGSU (National Institute of Diabetes and Digestive and Kidney Disease), rat TSHβ (National Institute of Diabetes and Digestive and Kidney Disease), human GH (Dako Corporation, Carpinteria, CA), and human PRL (Dako), and 1:1,000 dilutions of polyclonal antisera to human LH/FSH (Chemicon International Inc., Temecula, CA), synthetic MSH-γ1 (Chemicon), and human ACTH (Chemicon). Immunoreactivity was detected using the streptavidin-biotin-peroxidase method (ABC Elite Kit, Vector Laboratories, Burlingame, CA) and color development was performed with a 1× working solution of 3,3′-diaminobenzidine tetrahydrochloride (DAB; Pierce Biotechnology, Rockford, IL). Sections were counterstained with methyl green and coverslipped.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS AND DISCUSSION
  5. EXPERIMENTAL PROCEDURES
  6. Acknowledgements
  7. REFERENCES

We thank members of the Ryan lab, C. Goodyer and L. Jerome-Majewska, for helpful discussions and comments. M.M.C. is a recipient of a MCH-RI Studentship. A.K.R. was supported by a Career Award in Biological Sciences from the Burroughs Wellcome Fund and a New Investigator Award from CIHR. A.K.R. is a member of the Research Institute of the McGill University Health Centre, which is supported in part by the FRSQ.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS AND DISCUSSION
  5. EXPERIMENTAL PROCEDURES
  6. Acknowledgements
  7. REFERENCES