| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Neurobiology, Weizmann Institute of Science (T.D.S.-H., N.B.-A., Y.K.), Rehovot 76100, Israel; and Department of Chemistry and Bioscience, Chalmers University of Technology (T.B.), 405 30 Gothenburg, Sweden
Address all correspondence and requests for reprints to: Dr. Yitzhak Koch, Department of Neurobiology, Weizmann Institute of Science, Rehovot 76100, Israel. E-mail: y.koch{at}weizmann.ac.il.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In both humans and rats, GnRH and its receptor have been found to be synthesized in a myriad of tissues outside of their classical sites of production, such as pituitary (1, 2, 3, 4, 5), ovary (6, 7, 8, 9, 10), endometrium and placenta (11, 12), breast tissue and mammary glands (13, 14, 15, 16), liver, heart, skeletal muscles, kidney (17), spleen lymphocytes (18), and T cells (19). GnRH and GnRH receptor (GnRH-R) production in the various extrahypothalamic organs and glands has not yet been quantified, and their roles in these tissues have not been elucidated. We hypothesize, however, that GnRH carries out local functions in extrahypothalamic tissues, which it performs via autocrine/paracrine mechanisms. In our studies of the role of GnRH in extrahypothalamic organs, we chose to focus on the pituitary and ovary.
Evidence has been presented showing that GnRH is produced in and released by gonadotrope cells (20) or gonadotrope and corticotrope cells (3) of rat pituitaries. Moreover, in vitro findings point to a potential role for the endogenous pituitary GnRH in the maintenance of basal LH levels (20, 21, 22). Indeed, our research group has shown that the addition of a selective GnRH antagonist, antide, to primary pituitary cell cultures causes a stark reduction in the basal secretion of LH (21).
In the rat ovary, GnRH and its receptor are produced by granulosa cells (7, 8, 9, 10). A myriad of functions have been suggested for GnRH in this organ (recently reviewed in Ref. 23), including a role in oocyte maturation (24) and follicular atresia or selection (8), an effect on the corpus luteum (25), as well as an effect on the fertilization process (26). Interestingly, it has been shown that GnRH can induce ovulation in hypophysectomized rats (27, 28, 29). Importantly, unlike the pituitary GnRH-R, ovarian receptors are most likely activated by locally produced GnRH, because this peptide is found in only minute, undetectable amounts in the peripheral circulation (30). Hence, it is of relevance to study changes in locally produced ovarian GnRH throughout the estrous cycle.
The chain of hormonal and neuronal events that lead to the preovulatory gonadotropin surge, the pivotal event in the mammalian reproductive cycle, has been studied intensively, although it is still not completely understood (for a concise review, see Ref. 31). We raise the possibility that local GnRH, produced by pituitary and ovarian cells, could be one of the missing links in this complex chain of events leading to a midcycle gonadotropin surge and subsequent ovulation. In the present study we hypothesized that if GnRH produced locally in the ovary and/or pituitary is involved in the processes leading to ovulation, we might be able to discern periovulatory changes in its expression. Thus, our main venue of research consisted of investigating the pattern of GnRH and GnRH-R expression in the pituitary and ovary of the female rat throughout the estrous cycle compared with the pattern observed in the hypothalamus. We focused on the anticipated time of the proestrous LH surge to test the possible involvement of the locally produced pituitary and ovarian GnRH in this event. mRNA levels were measured relative to stably expressed genes using real-time PCR. We found fluctuations in the levels of GnRH and GnRH-R mRNA in all three tissues throughout the estrous cycle. We therefore suggest that locally produced ovarian and pituitary GnRH might fulfill regulatory roles in the processes leading to ovulation.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Animals were killed by decapitation at the following times: at 1000 and 1600 h on diestrus 1 (n = 12 and 4) and diestrus 2 (n = 10 and 4); on proestrus at 0900 h (n = 4), 1200 h (n = 10), 1400 h (n = 4), 1500 h (n = 4), 1600 h (n = 5), 1700 (n = 8), 1900 h (n = 10), and 2130 h (n = 6); and on estrus at 0900 h (n = 7) and at 1400 h (n = 5). Trunk blood was collected, and serum was separated and frozen until subsequent quantitative determination by RIA for LH. Tissues were immediately removed and placed in 10 vol RNA Later (Ambion, Inc., Austin, TX) until subsequent RNA extraction. The hypothalamus was dissected out to a depth of approximately 3 mm with the following borders: the anterior edge of the optic chiasm, the anterior edge of the mammillary bodies, and the two hypothalamic sulci on either lateral side. Both ovaries were removed, and oviducts were examined for the presence of ova only in the groups killed on the morning of estrus. The anterior pituitary was removed together with the attached posterior lobe.
Animals: quantitation of GnRH peptide in various tissues
Six- to 8-wk-old female Wistar rats were killed, and hypothalami, ovaries, and anterior pituitaries were removed as described above. Tissues were pooled (for details, see Fig. 1
), boiled, homogenized, and boiled again (32). Homogenates were centrifuged, and the GnRH concentration in the resulting supernatant was determined as described previously (33).
|
RNA purification
Tissues were removed from the RNA Later, weighed, and homogenized in 0.5 ml Tri-Reagent (Molecular Research Center, Inc., Cincinnati, OH). After the addition of 100 µl chloroform and phase separation by centrifugation (at 4 C), the aqueous layer was washed with an equal volume of 70% ethanol and loaded onto an RNeasy minicolumn (Qiagen, Hilden, Germany). The procedures for RNA isolation and purification as well as on-column deoxyribonuclease treatment (Qiagen) were then carried out as detailed in the manufacturer’s instructions. RNA samples were eluted in nuclease-free water (Qiagen). The RNA concentration was quantified using a NanoDrop machine (NanoDrop Technologies, Wilmington, DE), and its RNA purity was assessed on the same machine using 260:280 and 260:230 nM ratios. All samples had 260:280 nM ratios between 1.8 and 2.1, and 260:230 nM ratios above 1.7. RNA integrity was assessed by examining the 28S and 18S bands of representative samples loaded onto a 1.5% agarose gel stained with ethidium bromide.
RT
For each tissue, equal amounts of all RNA samples were reverse transcribed simultaneously. Ovarian and hypothalamic RNA samples (2 µg each) were reverse transcribed using Moloney murine leukemia virus reverse transcriptase ribonuclease H+ (Promega Corp., Madison, WI) according to the manufacturer’s instructions. Each reaction contained 0.5 µg oligo(deoxythymidine) (Amersham Biosciences, Piscataway, NJ), 0.52 mM of each deoxy-NTP (MBI Fermentas, St. Leon-Rot, Germany), 25 U RNAguard ribonuclease inhibitor (Amersham Biosciences), 5 µl of the 5x Moloney murine leukemia virus RT reaction buffer (Promega Corp.), and 200 U of the enzyme in a total volume of 25 µl. Pituitary RNA samples (4 µg each) were reverse transcribed using the SuperScript II ribonuclease H– reverse transcriptase kit (Invitrogen Life Technologies, Inc., Carlsbad, CA). Each 20-µl reaction contained 0.5 µg oligo(deoxythymidine) (Amersham Biosciences), 0.5 mM of each deoxy-NTP (MBI Fermentas), 40 U porcine liver ribonuclease-inhibitor (Takara Bio, Inc., Shiga, Japan), 2 µl 0.1 M dithiothreitol, 4 µl of the 5x First-Strand Buffer (Invitrogen Life Technologies, Inc.), and 40 U of the enzyme.
All RT reactions were performed at 42 C and contained a negative control, which consisted of nuclease-free water instead of RNA. The linearity of the RT reaction was evaluated using triplicate serial dilutions of an RNA pool, reverse transcribed as detailed above, and assayed in the real-time PCR for two genes: GnRH and cyclophillin. The reaction efficiencies (E = 97% and 100%, respectively) and correlation coefficients (r2 = 0.986 and 0.995, respectively), derived from the RNA dilution series, indicated that the RT reaction was linear under the conditions used. Several RT reactions contained duplicate RNA samples (n = 23 duplicates) to assess the RT-PCR variability.
Gene-specific primers and TaqMan hybridization probes
Primers were designed on two different exons so as to span one intronic sequence. TaqMan hybridization probes were designed to span an exon-exon junction (TIB-Molbiol, Berlin, Germany). All primer and probe sequences, PCR product sizes, and annealing temperatures used are listed in Table 1. To verify the identities of the PCR products obtained using each primer combination, each product was loaded onto an ethidium bromide-stained agarose gel, and the resulting band was purified and sequenced using the ABI Gene Scanner and the ABI BigDye Terminator Cycle Sequencing Kit (PerkinElmer, Applied Biosystems, Foster City, CA).
|
), and 15 sec at 72 C, at the end of which fluorescence was measured with the Rotor-Gene. SYBR Green-I assays also included a melt curve at the end of the cycling protocol, with continuous fluorescence measurement from 65–99 C. All reactions contained the same amount of cDNA, 10 µl Absolute QPCR Master Mix, primers or primers and probe diluted according to an optimized combination (see Table 1), and UltraPure PCR-grade water (Fisher Biotec, Subiaco, Australia) to a final volume of 20 µl. Each real-time PCR included a no-template control as well as five or six serial 4-fold dilutions, in duplicate, of a cDNA pool containing all experimental samples of the respective tissue. The prenormalized DNA quantity of each gene in every sample was estimated relative to this dilution series. This dilution series also served to assess the reaction performance (E and r2). The threshold cycle (Ct) was set so as to obtain the highest reaction efficiency and correlation coefficient. Because not all samples of a particular tissue (n = 93) could be assayed simultaneously for each gene, a common set of at least five samples was included in every reaction for interreaction calibration.
Preliminary validation of reference genes
A common practice to compensate for differences in the steps preceding the PCR is normalization of gene expression to reference genes. For proper normalization, the expression of reference genes should not vary across the experimental conditions. A panel of four candidate reference genes was therefore tested in all experimental samples (n = 93/tissue) to identify the most stably expressed genes at all tested times. Primer sequences, accession numbers, and reaction conditions for all tested reference genes are listed in Table 1 (see gene names followed by an asterisk). The two most suitable reference genes in each tissue were identified by three approaches.
The first approach, developed and described by Vandesompele et al. (35), identifies the pair of genes whose DNA quantities fluctuate the least relative to each other. It is based on the assumption that the expression ratio of two ideal reference genes that are not coregulated, is constant across experimental conditions. This approach is implemented in the freely available Excel macro, GeNorm (35).
The second approach, developed and described by Andersen et al. (36), is based on the decomposition of expression values of each candidate reference gene into technical and biological constituents. The stability of each gene across all samples is determined both within and between groups (e.g. time groups), and the most stable single gene and pair of genes are identified. This approach is implemented in a freely available Excel Add-in, NormFinder (36).
The most obvious advantage of NormFinder is that it examines the expression stability of each single candidate gene independently and not in relation to the other genes, as GeNorm does. This is important in light of our limited knowledge regarding coregulation. Moreover, NormFinder also tests for combinations of genes that may compensate for each other’s fluctuations. This is helpful in situations where none of the candidate reference genes is found to be stably expressed. The disadvantage of NormFinder, which is more critical when a small number of genes are tested, lies in the possibility that several of the candidate genes display similar expression trends or fluctuations in the experimental data. This causes skewing of the average intergroup variation, which should normally be close to zero for a proper identification of the least variable gene. Although the tendency of GeNorm to select coregulated genes increases with the number of tested genes, NormFinder will have more chances of selecting a stably expressed gene or a combination of two genes whose fluctuations compensate one another.
The third approach, which we term least absolute variation, is based on the low probability that biological variability in gene expression will be precisely countermatched by (opposite) technical variability, yielding low variance in prenormalized DNA amounts of a large set of samples. According to this approach, low intersample variability of prenormalized DNA amounts, assessed by the spread, i.e. the SD, of PCR Ct indicates both biological and technical similarities. To determine whether the obtained spread in Ct values of all samples assayed for each candidate reference gene is indeed small, we compared it to the reproducibility of RT-PCRs. Unlike GeNorm and NormFinder, this approach provides information regarding the degree of stability (magnitude of the SD) of each gene.
Data analysis and statistics
The relative amount of GnRH or GnRH-R mRNA in each sample was calculated by dividing the prenormalized DNA quantity of these genes (calibrated to account for interreaction differences), obtained from the dilution series (see above), by the geometric mean of the DNA quantities of the two most suitable reference genes (35). This normalized DNA quantity is hereafter referred to as the relative expression level of GnRH or GnRH-R.
All statistical analyses were performed using JMP IN Statistical Discovery software (version 5.1, SAS Institute, Inc., Cary, NC). For each tissue and gene, the distribution of the obtained LH levels and relative amounts of GnRH and GnRH-R in all samples were tested for normality using the Kolmogorov-Smirnov-Lilliefor test. In all cases studied, the log-transformed data displayed a normal distribution (P > 0.05). Statistical evaluation of differences between time groups was performed using both parametric and nonparametric multiple comparison tests (one-way ANOVA and the Wilcoxon rank-sum test, respectively), followed by pairwise comparisons of means using the least significant difference test and Student’s t test at a confidence level of 95%.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
. Compared with the hypothalamus, we found about 430-fold less GnRH in the pituitary and about 170-fold less GnRH in the ovary.
Selection of suitable reference genes
Table 2 indicates, for each tissue, the two most suitable reference genes found by each of the three methods employed. The variability of RT-PCRs was estimated from RNA samples reverse transcribed in duplicate and assayed together in the real-time PCR. It was measured as the SD of duplicate reactions’ Ct values and found to be, on the average, SD = 0.22 (range, 0.001–0.42; n = 23). This corresponds well to the RT-PCR reproducibility estimate recently published by Stahlberg et al. (37): SDmRNA = 0.11–0.60. The SD of Ct values obtained for the best candidate reference genes in all tissue was found to range between 0.30 and 0.62 (Table 2), which overlaps the RT-PCR reproducibility values mentioned above. Therefore, we chose the two most suitable reference genes for each tissue to be those with the lowest SD. Thus, in pituitary and hypothalamic samples, the geometric average of cyclophillin and RPL19 expression was used to normalize GnRH and GnRH-R expression; in ovarian samples, the geometric average of cyclophillin and ß-actin expression was used to normalize GnRH and GnRH-R expression.
|
), when serum LH levels were 29.1 ± 1.75 ng/ml (mean ± SEM). Fluctuations in the relative expression levels of GnRH during the estrous cycle were found in all three tissues (by one-way ANOVA, P < 0.05; by Wilcoxon rank-sums test, P < 0.05 for all three tissues). In the hypothalamus (Fig. 2B), the highest GnRH expression was found on proestrus at 1700 h, whereas the lowest expression level was found on this day at 1900 h.
|
), the highest relative GnRH expression was detected on the morning and early afternoon of estrus. Peak proestrous amounts of GnRH mRNA were noted at noon and 1700 h. On diestrus 1 and 2, morning levels were higher than afternoon levels.
In the ovary (Fig. 2D), peak GnRH expression levels occurred at 1000 h on diestrus 1 and 1400 h on proestrus, whereas the lowest expression levels occurred in the late afternoon-evening of proestrus and on estrus.
Relative levels of GnRH-R expression during the rat estrous cycle
Significant variations in the level of GnRH-R mRNA during the estrous cycle were detected in the pituitary (by one-way ANOVA, P < 0.001; by Wilcoxon rank-sums test, P < 0.0001) and ovary (by one-way ANOVA, P < 0.05; by Wilcoxon rank-sums test, P < 0.005). The hypothalamic GnRH-R mRNA (Fig. 3A) displayed only nonsignificant changes during the estrous cycle, possibly due to relatively high interindividual variability.
|
), the highest relative amounts of GnRH-R mRNA were detected in the mornings of diestrus 1 and 2, followed by a sharp reduction in the afternoon of these days. On proestrus, pituitary GnRH-R expression was low in the morning and rose back to diestrous 1 and 2 (morning) levels at noon, followed by significant transient drops at 1400 and 1600 h and a second peak at 1700 h. Relatively low levels of GnRH-R expression were measured on estrus in this tissue.
In the ovary, GnRH-R mRNA expression reached peak levels at 1400 h on proestrus, followed by a second lower peak at 1700 h (Fig. 3C). The lowest expression levels occurred at 1400 h on estrus and noon on proestrus.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We obtained extremely low intersample Ct variability for our candidate reference genes (see SD values in Table 2), which is within the reproducibility range of RT-PCRs (37). This might be a somewhat rare situation, reflecting extremely small technical and biological errors. Nonetheless, we believe that under such circumstances, measurements of variation in prenormalized DNA quantities provides the most direct means of assessing the stability of candidate reference genes. Therefore, we chose the two most suitable reference genes to be those with the lowest SD. We believe that these reference genes could be used in future studies employing similar experimental systems.
When comparing the amounts of GnRH produced by the pituitary or ovary to those found in the hypothalamus, one should take into consideration the different modes of action that this peptide undertakes in the different tissues. Thus, although GnRH functions as an endocrine hormone once it is released from the hypothalamus into the portal system, in extrahypothalamic regions it probably plays an autocrine/paracrine role. One might therefore expect lower levels of GnRH production in extrahypothalamic tissues compared with the hypothalamic levels.
The present study is the first demonstration of a GnRH mRNA expression pattern in the pituitary and ovary throughout the estrous cycle of any species, whereas the pattern of hypothalamic GnRH expression has been studied previously (40, 41, 42, 43, 44). Our findings corroborate with previous demonstrations of a peak in hypothalamic GnRH mRNA levels in the later afternoon-early evening hours of proestrus, slightly before the gonadotropin surge (42, 43, 44). Similarly, peak levels of GnRH peptide have been found in the mediobasal hypothalamus (45, 46) and in the portal blood (47) on the evening of proestrus. This increase probably serves either to increase the availability of hypothalamic GnRH in preparation for the imminent GnRH surge or to replenish this peptide’ stores that were depleted by the surge.
Expression of the receptor for GnRH in the mediobasal hypothalamus was previously shown to be modulated during the rat estrous cycle (48), such that peak levels were observed on the morning of proestrus and the evening of estrus. It is noteworthy that very few times points of the estrous cycle were sampled in that study. In our investigation, the entire hypothalamus was analyzed rather than specific nuclei, and only nonsignificant fluctuations in the expression of GnRH-R during the estrous cycle were observed. It is possible that the GnRH-R are differentially modulated at various hypothalamic sublocations. The hypothalamic GnRH-R are probably involved in autoregulatory feedback mechanisms in this tissue.
In the pituitary, the rise in GnRH mRNA we observed at noon on proestrus is interesting in light of the increase in LH mRNA that begins more or less at the same time (49). It is possible that the local pituitary GnRH plays a role in the regulation of LH production, because at this time, GnRH levels in the median eminence (50) and portal system (47) are still low. As mentioned previously, a potential role for pituitary GnRH in the release of LH has been demonstrated in vitro (21). The physiological significance of the high estrous levels of GnRH mRNA in the pituitary is unclear at this point. It might be noteworthy that FSH mRNA has also been shown to increase on the day of estrus (51).
Interestingly, there appears to be a temporal correlation between the expression pattern of GnRH and that of its receptor in the pituitary throughout the estrous cycle, except on estrus. It is possible that endogenous pituitary GnRH participates in the regulation of its receptor in this gland or that the two are coregulated. Pituitary GnRH-R expression and content as well as GnRH binding to pituitary receptors during the rat estrous cycle have been investigated in the past, although results are somewhat contradictory (52, 53, 54, 55, 56). Nevertheless, a heightened GnRH-binding capacity of the pituitary during the day or so before the preovulatory gonadotropin surge appears to be a result of increased receptor synthesis. This increase, which might be generated by locally produced pituitary GnRH, is postulated to induce heightened pituitary responsiveness to hypothalamic GnRH stimulation (57).
In the ovary, we observed relatively high GnRH and GnRH-R expression in the early afternoon of proestrus and high GnRH-R mRNA levels also around the time of the preovulatory surge. In an earlier investigation (58), it was reported that cyclical changes during the estrous cycle in the expression of the ovarian GnRH-R are specific to the stage of follicular development, such that they were observed only in corpora lutea and atretic follicles. In these two types of follicles, peak GnRH-R levels were observed in the evening of proestrus, with a second increase in the morning of estrus observed only in atretic follicles. The researchers suggested that ovarian GnRH might be involved in follicular atresia and possibly also in the induction of ovulation. The fact that we detected peak GnRH expression concomitantly with peak GnRH receptor expression in the ovary is intriguing and raises the possibility that here too, GnRH regulates the expression of its own receptor or that the two are coregulated.
An interesting observation was recently published (59), suggesting that oocytes of the gilthead sea bream produce and release gonadotropins, and that this release can be enhanced by a GnRH analog or reduced by a GnRH antagonist. As the researchers point out, this discovery raises the interesting possibility of a local GnRH-gonadotropin axis within the fish ovary. In preliminary RT-PCR experiments (performed in collaboration with the laboratory of N. Dekel at the Weizmann Institute), we also identified LHß expression in rat and mouse oocytes. One could thus envisage local GnRH-gonadotropin axes within the pituitary and ovaries of mammalian females. Such local regulatory axes could contribute to or finely tune the hypothalamic-pituitary-gonadal axis, for instance by priming the relevant organs in preparation for the preovulatory peak (in the case of the pituitary) and ovulation (in the case of the ovaries). Nonetheless, research in this direction has yet to be conducted.
The present report presents a detailed and precise pattern of GnRH and GnRH-R expression during the rat estrous cycle. We propose that in the adult female rat, the production of GnRH and GnRH-R is locally regulated in the pituitary and ovary, in accordance with the animal’s reproductive state or phase of the estrous cycle. The earlier increase in GnRH production in the pituitary and ovary compared with the hypothalamus during the proestrous stage of the sexual cycle might indicate that this peptide participates in the preparation of these organs for the imminent preovulatory surge, possibly via local GnRH-gonadotropin axes. One should not preclude, however, nonreproductive autocrine/paracrine roles of GnRH in extrahypothalamic tissues.
|
Footnotes |
|---|
First Published Online May 19, 2005
Abbreviations: Ct, Threshold cycle; GnRH-R, GnRH receptor.
Received February 28, 2005.
Accepted for publication May 10, 2005.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
messenger ribonucleic acid in the ovary. Endocrinology 127:2350–2356[Abstract]
and luteinizing hormone ß subunit messenger ribonucleic acids during the rat estrous cycle. Endocrinology 119:1867–1869[Abstract]
This article has been cited by other articles:
 |
|
 |
|
  A. Sengupta and R. Sridaran Expression and Localization of Gonadotropin-releasing Hormone Receptor in the Rat Oviduct During Pregnancy J. Histochem. Cytochem., January 1, 2008; 56(1): 25 - 31. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Gershon, A. Hourvitz, S. Reikhav, E. Maman, and N. Dekel Low expression of COX-2, reduced cumulus expansion, and impaired ovulation in SULT1E1-deficient mice FASEB J, June 1, 2007; 21(8): 1893 - 1901. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W. Zheng, M. Jimenez-Linan, B. S. Rubin, and L. M. Halvorson Anterior Pituitary Gene Expression with Reproductive Aging in the Female Rat Biol Reprod, June 1, 2007; 76(6): 1091 - 1102. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Singh, A. Krishna, and R. Sridaran Localization of gonadotrophin-releasing hormone I, bradykinin and their receptors in the ovaries of non-mammalian vertebrates Reproduction, May 1, 2007; 133(5): 969 - 981. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. Yin and A. C Gore Neuroendocrine control of reproductive aging: roles of GnRH neurons. Reproduction, March 1, 2006; 131(3): 403 - 414. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Kitajima, T. Endo, K. Nagasawa, K. Manase, H. Honnma, T. Baba, T. Hayashi, H. Chiba, N. Sawada, and T. Saito Hyperstimulation and a Gonadotropin-Releasing Hormone Agonist Modulate Ovarian Vascular Permeability by Altering Expression of the Tight Junction Protein Claudin-5 Endocrinology, February 1, 2006; 147(2): 694 - 699. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. H |
