| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ARTICLES |
Department of Endocrinology and Reproduction, Faculty of Medicine and Health Sciences, Erasmus University Rotterdam, Rotterdam; and Hubrecht Laboratory, Netherlands Institute for Developmental Biology (J.P.d.W.), Utrecht, The Netherlands
Address all correspondence and requests for reprints to: Dr. J. W. M. Martens, Department of Endocrinology and Reproduction, Erasmus University Rotterdam, P.O. Box 1738, 3000 DR Rotterdam, The Netherlands. E-mail: martens{at}endov.fgg.eur.nl
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionNote Added in ProofReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionNote Added in ProofReferences |
|---|
Activin and inhibin are both members of the transforming growth
factor-ß (TGFß) superfamily of growth and differentiation factors.
Members of this family consist of two identical or highly homologous
subunits, linked by one disulfide bridge (16, 17). Activin is
consistent with this model; it is a dimer of two inhibin ß-subunits.
Inhibin is the only exception in this family, as it consists of an
inhibin ß-subunit, which is also present in activin, linked to a
distantly related inhibin 
-subunit. Furthermore, no homodimers of
the -subunit of inhibin have been described.
Members of the TGFß superfamily exert their actions through combinations of type I (55 kDa) and type II (68 kDa) receptors. Both receptors are characterized by a small extracellular ligand-binding domain, a single transmembrane domain, and an intracellular Ser/Thr-specific kinase domain. For both TGFß and activin, the type II receptor is a constitutively active kinase that has high affinity for the ligand. This ligand type II receptor complex subsequently interacts with the type I receptor (18, 19). After association, the type I receptor is phosphorylated by the kinase domain of the type II receptor in its juxta-membrane region, also known as the GS box (18, 19, 20). The phosphorylation of the GS box apparently leads to activation of the type I receptor, resulting in stimulation of downstream pathways.
Two activin type II receptors (ActRIIA and ActRIIB) and two activin type I receptors (ActRIA and ActRIB, also known as ALK-2 and ALK-4, respectively) are known (21, 22, 23, 24). However, inhibin receptors have not been identified to date, and the mechanism of action of inhibin has not been clarified. It might be that inhibin signals through its own type I and type II receptors. Alternatively, the special position of inhibin in the TGFß superfamily may indicate that its signaling mechanism is different, as is also suggested by the fact that inhibin appears to interfere specifically with activin signaling. We investigated whether inhibin blocks activin signal transduction downstream of the activin-receptor complex or interferes with activin signaling at the level of the activin receptor as suggested earlier (13, 25). As a model, the Chinese hamster ovary (CHO) cell line K1 was used; its cell growth is sensitive to both activin and inhibin (11). In these cells the interference of inhibin with activin-induced immediate early responses and the effect of inhibin on the activin receptor complex were studied. This showed that inhibin acts directly at the level of the activin-receptor complex in CHO cells. In addition, our data indicate that inhibin can interfere with activin signaling through displacement of activin from the activin type II receptor.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionNote Added in ProofReferences |
|---|
Transfection and luciferase assay
CHO K1 cells were maintained in DMEM-Ham’s F-12 (DF; Life
Technologies, Gaithersburg, MD) supplemented with 10% FCS (SEBAK,
Aidenbach, Germany) and cultured at 37 C in 5% CO2 in air.
Cells were plated before transfection in 12-well plates (Costar,
Cambridge, MA) at a concentration of 4 x 104
cells/well. The next day the cells were transiently transfected with
either p3TP-Lux or p5T-Lux (0.5 µg/well) using the calcium phosphate
transfection method (33). In case of cotransfection with activin
receptor expression plasmid, up to 1.0 µg of this plasmid was used
per well. To correct for transfection efficiencies, the
ß-galactosidase expression plasmid pCH110 (0.5 µg/well) (34) was
added. At 50% confluence, cells were deprived of serum and cultured in
DF containing 0.1% BSA for at least 8 h. Subsequently, the cells
were incubated with vehicle, human recombinant activin A (Innogenetics,
Ghent, Belgium), human recombinant inhibin A (Genentech, South San
Francisco, CA), TGFß (Sanbio, Breda, The Netherlands), ß-phorbol
12-myristate 13-acetate (PMA; Sigma Chemical Co., St. Louis, MO), or
steroid-free bovine ovarian follicular fluid as indicated. After
16 h, the cells were lysed, and luciferase and ß-galactosidase
activities were measured (35, 36).
Northern blot analysis of junB messenger RNA (mRNA)
Serum-deprived cells were treated as described and harvested 0,
0.5, 1.5, or 6 h after the addition of activin or activin and
inhibin. Total RNA was extracted by the TRIzol procedure (Life
Technologies), and 20 µg of this were fractionated on a denaturing
agarose gel and then transferred to a nylon membrane (Hybond
N+, Amersham, ’s-Hertogenbosch, The Netherlands) (37). The
junB mRNA was detected using mouse junB
complementary DNA (cDNA; p465.20) as a probe (38). Hybridization with a
glyceraldehyde-3-phosphate dehydrogenase cDNA served as a measure of
the amount of RNA applied to each lane. The labeled probe bound to the
blot was quantified using a PhosphorImager and the ImageQuant software
package (version 3.3, Molecular Dynamics/B&L Systems, Zoetermeer, The
Netherlands).
Reverse transcriptase-PCR (RT-PCR)
RT-PCR was performed on total RNA of CHO cells and K562 cells as
previously described (39) using the oligonucleotide primers described
below. As a control, PCR was performed on cDNA clones of the mouse
ActRIIA, ActRIIB2, and ActRIIB4 described
above. For ActRIIA, primers derived from the rat cDNA sequence (40)
were used for the amplification (forward primer, 5'-CAGGGAACTG
GATATCTAGA GAGAACTTC-3'; reverse primer, 5'-TGGTCCTGGG TCTCGAGTAG
GAACAAGTAC-3'); for ActRIIB, primers derived from the human ActRIIB
(41) were used (forward primer, 5'-CGAATTCCGC TGCTGCCCAT TGGAGGC-3';
reverse primer, 5'-TGTAAGCTTG TGGCCCTCAC CACGACACC-3'). The ActRIIA
primers amplify a fragment of 685 bp; the ActRIIB primers amplify a
790-bp fragment for the ActRIIB2 and a 766-bp fragment for
the ActRIIB4. The PCR reaction consisted of 40 cycles of
denaturation at 94 C for 1 min, primer annealing at 50 C for 2 min, and
subsequent extension reaction at 72 C for 2 min using 0.2 U SuperTaq
(HT/Biotechnology, Cambridge, UK). Reaction products were analyzed by
agarose (2%) gel electrophoresis and visualized by ethidium bromide
staining.
Cross-linking
COS-1 cells were maintained in DF supplemented with 7.5% FCS
and cultured at 37 C in 5% CO2 in air. Wild-type CHO cells
or activin receptor-transfected COS cells (
2–4 x
106 cells) were preincubated for 30 min at 4 C without
additions or with unlabeled activin A or inhibin A followed by an
incubation of 2 h at 4 C with 10 ng [125I]activin A
(180 pM; 750,000 cpm) in 2 ml DF. Activin A was
iodinated using chloramine-T as previously described (29). After the
incubation, the cells were washed twice with HEPES-buffered saline.
Bound activin was cross-linked to its receptor with 1 mM
bis-sulfosuccinimidyl suberate (BS3, Pierce Chemical Co.,
Rockford, IL) as described previously (29). Subsequently, the cells
were lysed, and the cross-linked complexes were purified. For CHO
cells, this was achieved by incubating the extracts with wheat-germ
agglutinin agarose overnight; for activin receptor-transfected COS-1
cells, the complexes were first incubated overnight with specific
antibodies to either the HA or the KT3 tag present on the
C-terminus of the ActRII. The resulting complexes were subsequently
isolated by an incubation with either protein A- or protein
G-Sepharose, as previously described (29), and subsequently separated
on a reducing SDS-PAGE followed by autoradiography. The amount of
iodinated activin cross-linked to either activin receptor was
quantified using a PhosphorImager and the ImageQuant software
package.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionNote Added in ProofReferences |
|---|
, A and C).
To investigate whether this activin-induced response was specifically
affected by inhibin, we also studied the effect of inhibin on
TGFß-induced 3TP reporter activity. TGFß was chosen because it is
known to induce similar responses as activin, including 3TP reporter
activity, in a number of cell types (11, 42, 43). TGFß (150
pM; ED50 = 25 pM) stimulated
luciferase activity in the 3TP-Lux construct 8- to 10-fold (Fig. 1A;
dose-response curves not shown). This TGFß-induced luciferase
activity, however, was insensitive to inhibition by inhibin (Fig. 1A).
Even a relatively high doses of inhibin (2 nM) was
completely ineffective (not shown). Activin and TGFß both act through
either one of the two different response elements in the 3TP promoter.
Therefore, we studied the effect of activin, TGFß, and the protein
kinase C activator, PMA, on a reporter construct containing just TREs,
p5T-Lux. PMA (50 nM) stimulated the reporter activity of
this construct in CHO cells 3-fold, whereas activin (1 nM)
and TGFß (250 pM) had no effect (Fig. 1B). In addition,
we found that the effects of TGFß and PMA on the 3TP-Lux construct
were additive (not shown). Thus, the transcription from the 3TP
promoter induced by both activin and TGFß is likely to be exclusively
derived from the PAI-1 promoter element in the 3TP-Lux construct. As
inhibin only interferes with the activin-induced 3TP promoter
activation, inhibin exerts its effect on a part of the activin
signaling cascade that is not in common with that of TGFß.
|
). Inhibin reduced the
activin-dependent junB expression significantly
(P < 0.005, by Mann-Whitney U test). Equimolar amounts
of inhibin reduced activin-induced junB expression by
28 ± 4.1% and reduced activin-induced 3TP luciferase activity by
27 ± 7%. This indicates that both immediately early responses to
activin are equally sensitive to inhibin. On a molar basis, however,
inhibin is less effective than activin.
|
). Experiments with another
splice variant of ActRIIB, ActRIIB4, were also performed.
This splice variant differs from the ActRIIB2 in a small
part of the extracellular domain that affects the affinity for activin
(24). The results obtained with ActRIIB4 were comparable to
those obtained with ActRIIB2 (not shown). In addition,
ActRIA was tested in combination with all three type II receptors. The
results obtained with this receptor were identical to those with ActRIB
(not shown).
|
, control lanes). In addition, some minor larger
complexes (100–160 kDa) were observed, presumably consisting of
multiple activin subunits and/or activin receptors cross-linked to each
other (13).
Preincubation with unlabeled activin resulted in a gradual decrease in
the amount of labeled activin cross-linked to the ActRI and ActRII
(Fig. 3A). The ratio of the amount of labeled activin cross-linked to
the ActRI over that cross-linked to ActRII decreased (to 35 ± 9%
and 40% ± 6% for ActRIIA and ActRIIB2, respectively)
relative to the ratio in the absence of competitor (P
< 0.05, by Student’s t test; n = 3; Fig. 3C). In the
presence of inhibin, the amounts of labeled activin cross-linked to
both ActRII and ActRI also decreased (Fig. 3A). However, the ratio
between the amount of labeled activin cross-linked to the ActRI and
that cross-linked to ActRII did not change significantly after the
addition of inhibin (99 ± 6% and 107% ± 20% for ActRIIA and
ActRIIB2, respectively) compared to that in the absence of
competitor (Fig. 3C). This difference between activin and inhibin was
observed for all ActRI and ActRII combinations tested. In the presence
of unlabeled inhibin, labeled activin was cross-linked more efficiently
to ActRIIA than to ActRIIB2; inhibin was approximately 15
times less potent than activin in competing for labeled activin bound
to ActRIIA, whereas this difference for ActRIIB2 was only
2.5. This difference was independent of the type I receptor (IA or IB)
transfected into COS-1 cells. This set of experiments indicates that
ActRIIs, but not ActRIs, can be targets for inhibin and that inhibin
preferentially interacts with ActRIIBs.
In addition, we performed similar displacement studies with the
endogenous activin receptors of CHO cells that were incubated with
labeled activin (0.18 nM) alone or in the presence of
unlabeled activin (3.6 nM) or inhibin (3.0 nM).
After cross-linking, the complexes were purified by binding to
wheat-germ agglutinin agarose beads and analyzed on SDS-PAGE (Fig. 3B).
Similar to the results of the COS-1 cell experiments, labeled activin
was predominantly cross-linked to a type I and a type II receptor and
unlabeled activin and inhibin reduced the amount of labeled activin
cross-linked to both type I and type II receptors (Fig. 3B). However,
unlike the findings in COS-1 cells, unlabeled activin did not affect
the ratio of labeled activin cross-linked to ActRI and ActRII (Fig. 3, B and D). Inhibin was 2–3 times less potent than unlabeled activin in
displacing labeled activin from the endogenous ActRII in CHO cells
(Fig. 3D).
To identify the ActRIIs involved in this binding, we analyzed
which of the known ActRIIs are expressed in CHO cells. Expression
levels of ActRIIs were too low to be detected on a Northern blot.
Therefore, we identified the receptors by PCR using primers specific
for either ActRIIA (Fig. 4A) or ActRIIB (Fig. 4B). Both
receptors were detected in CHO cells; the only ActRIIB splice variant
present was ActRIIB2. Another inhibin-responsive cell line,
K562, which predominantly expresses ActRIIB2, as determined
by ribonuclease protection (van Schaik, R. H. N., unpublished results)
was included in this analysis as a control. In these cells,
ActRIIB2 could be amplified to a prominent band (Fig. 4B, lane 7), whereas no ActRIIA could be detected in these cells.
|
). In cells transfected with the wild-type receptor
construct, luciferase activity was low and did not depend on the amount
of expression plasmid. In contrast, the luciferase activity
considerably increased with increasing amounts of expression plasmid
when the mutant ActRIB was introduced. This indicates that the
replacement of threonine 206 by aspartate leads to the activation of
the ActRIB in the absence of activin, as was also recently shown by
Attisano et al. (18). The constitutively active receptor did
not activate the 5T-Lux construct containing only TREs (Fig. 5B),
showing that 3TP promoter activation by this mutant receptor is
promoter specific and is not due to aspecific changes due to, for
example, overexpression of this receptor. Subsequently, CHO cells were
transfected with either wild-type ActRIB or ActRIB(T206D), and 3TP
promoter activation was studied in response to activin and/or inhibin
(Fig. 5C). CHO cells transfected with the wild-type ActRIB responded to
activin with a 4- to 5-fold increase in luciferase activity, and this
activin-dependent luciferase activity was sensitive to inhibin, similar
to that in wild-type CHO cells. In CHO cells transfected with the
mutant ActRIB, basal luciferase activity was increased about 10-fold
compared to that in cells transfected with the wild-type ActRIB.
Activin did not further increase luciferase activity in these cells,
suggesting that this mutant receptor was already fully active in the
absence of activin. More importantly, the addition of inhibin (0.8
nM) was also without effect (Fig. 5C). Even a high
concentration of inhibin (2 nM) or bovine follicular fluid
containing 10 nM bioactive inhibin did not decrease 3TP
promoter activity (not shown). This experiment shows that inhibin is
unable to block 3TP promoter activation induced by activin downstream
of the ActRIB in CHO cells.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionNote Added in ProofReferences |
|---|
Our results indicate that inhibin cannot interact with ActRI directly or after it has formed a complex with the ActRII. This in contrast to activin, which can interact with the ActRI after it is complexed to a ActRII. Two observations support this conclusion. Firstly, if inhibin interacts with the ActRI directly, it would displace labeled activin from the ActRI and not from ActRII. This is clearly not the case, because the ratio of labeled activin cross-linked to ActRI compared to that cross-linked to ActRII does not change (in either CHO or COS-1 cells) in the presence of unlabeled inhibin compared to the ratio in the absence of competitor. Thus, displacement of labeled activin from ActRI by inhibin is indirect and is due to displacement of labeled activin from ActRII, which is in line with previous observations (13, 21, 24, 25). Secondly, we observed that unlabeled inhibin behaves differently from unlabeled activin in COS-1 cells overexpressing ActRs. This phenomenon can only be explained if inhibin that is bound to ActRIIs is unable to interact with ActRI and if an excess of ActRIIs over ActRIs is present in transfected COS-1 cells. In that case, unlabeled activin binds to both receptors in a ternary complex, leaving predominantly unoccupied type II receptors for the binding of labeled activin. In this way activin affects the ratio of labeled activin cross-linked to ActRI and ActRII. In the presence of unlabeled inhibin, however, no ternary complex is formed between inhibin complexed to ActRII and the type I receptors. Thus, the type I receptors are not occupied, making it possible that labeled activin can bind equally well to type I and type II receptors, as is the case in the absence of inhibin. CHO cells may have no excess of type II receptors, causing a similar suppression of labeled activin bound to both types of activin receptors by activin and inhibin.
The fact that inhibin cannot interact with the ActRI is in line with
studies performed by Xu and co-workers (13). They showed that inhibin
only interacts via its ß-subunit with ActRII and that the inhibin
-subunit cannot interact with any ActR. This suggests that the
-subunit of inhibin has lost domains that are important for activin
receptor interaction and that the -subunit only prevents receptor
dimerization. This is further supported by the fact that cleavage of
the inhibin -subunit to its mature form is not required for inhibin
action, whereas cleavage of the inhibin ß-subunit is a prerequisite
for activin action (47). In addition, compared to the ß-subunit gene,
the -subunit displays much greater genetic variability between
species (Table 1). All of these features indicate the
separate position of inhibin in the TGFß superfamily and support its
unique mechanism of action.
|
This competition model allows for a complex and intricate regulation of
target cell activities, depending on the relative production of -
and ß-subunits and the relative expression of ActRIIA and
ActRIIB2/4. Unfortunately, we were unable to determine the
ActRIIA/ActRIIB mRNA ratio for CHO cells due to the low level of
expression of both ActRIIs. However, the human erythroid cell line
K562, which is highly sensitive to inhibin (5), shows a high
ActRIIB2/ActRIIA ratio (van Schaik, R. H. N., unpublished
results), which is in line with the idea that the ActRIIB is the ActRII
that is most sensitive to inhibin.
The current competition model can explain most data on inhibin action. However, it is difficult to envisage how both inhibin and activin can inhibit a response in the same cell (48, 49) and how inhibin can stimulate a response in the presumed absence of endogenous activin (50). Further, complete insensitivity to inhibin in cells that are responsive to activin (11, 14, 15) may be explained if these cells express the inhibin-insensitive ActRIIA or, alternatively, if they express high levels of ActRIIs so that physiological concentrations of inhibin cannot displace activin. However, it may be necessary to postulate a separate inhibin receptor pathway to explain all of these observations.
The similarity of the intracellular domains of ActRIB and TGFßRI suggests that the activation of these receptors by type II receptors and the signaling pathway downstream of these receptors are identical (42). Indeed, analogous to the TGFßRI (44), the introduction of an aspartic acid for threonine 206 in the GS box of the ActRIB leads to constitutive activation of the kinase domain of this receptor (18)). This supports the view that modifications in or near the GS box are a general mechanism for activation of type I receptors of the TGFß superfamily. Under normal conditions this occurs through phosphorylation of the GS box by the type II receptor (17, 19). It is noteworthy that the introduction of aspartate in the ActRIB results in complete activation of the downstream pathway, because addition of activin to CHO cells transfected with this mutant receptor does not lead to a further increase in the 3TP luciferase response. A similar mutation in TGFßRI resulted in only partial activation of the downstream pathway (44). This suggests that the mechanism or the threshold for activation of the type I receptor by the type II receptor for activin is slightly different from that for TGFß.
In conclusion, we showed that inhibin blocks activin signaling in CHO cells at the level of the activin-receptor complex, most likely through competition with activin at the ActRIIs; the ActRIIBs are more sensitive. Whether inhibin acts via a similar mechanism in other cell types or via a separate inhibin receptor pathway remains to be determined. In this regard it will be of particular interest to investigate gonadotropic cells, because this cell type is very sensitive to inhibin.
|
Note Added in Proof |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionNote Added in ProofReferences |
|---|
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received December 10, 1996.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionNote Added in ProofReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  R. S. Pearsall, E. Canalis, M. Cornwall-Brady, K. W. Underwood, B. Haigis, J. Ucran, R. Kumar, E. Pobre, A. Grinberg, E. D. Werner, et al. A soluble activin Type IIA receptor induces bone formation and improves skeletal integrity PNAS, May 13, 2008; 105(19): 7082 - 7087. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Eijken, S. Swagemakers, M. Koedam, C. Steenbergen, P. Derkx, A. G. Uitterlinden, P. J. van der Spek, J. A. Visser, F. H. de Jong, H. A. P. Pols, et al. The activin A-follistatin system: potent regulator of human extracellular matrix mineralization FASEB J, September 1, 2007; 21(11): 2949 - 2960. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Dixit, K.L. Rao, V.V. Padmalatha, M. Kanakavalli, M. Deenadayal, N. Gupta, B.N. Chakrabarty, and L. Singh Mutational analysis of the betaglycan gene-coding region in susceptibility for ovarian failure Hum. Reprod., August 1, 2006; 21(8): 2041 - 2046. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. G. Farnworth, P. G. Stanton, Y. Wang, R. Escalona, J. K. Findlay, and G. T. Ooi Inhibins Differentially Antagonize Activin and Bone Morphogenetic Protein Action in a Mouse Adrenocortical Cell Line Endocrinology, July 1, 2006; 147(7): 3462 - 3471. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. G Farnworth, Y. Wang, P. Leembruggen, G. T Ooi, C. Harrison, D. M Robertson, and J. K Findlay Rodent adrenocortical cells display high affinity binding sites and proteins for inhibin A, and express components required for autocrine signalling by activins and bone morphogenetic proteins. J. Endocrinol., March 1, 2006; 188(3): 451 - 465. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. M. Foster, P. R. Olton, and V. Padmanabhan Diurnal changes in FSH-regulatory peptides and their relationship to gonadotrophins in pubertal girls Hum. Reprod., February 1, 2005; 20(2): 543 - 548. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. A. Sweeney and P. A. Johnson Messenger RNA and Protein Expression Analysis of Betaglycan in the Pituitary and Ovary of the Domestic Hen Biol Reprod, January 1, 2005; 72(1): 172 - 178. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C A Longui, S H V Lemos-Marini, B Figueiredo, B B Mendonca, M Castro, R Liberatore Jr, C Watanabe, C L P Lancellotti, M N Rocha, M B Melo, et al. Inhibin {alpha}-subunit (INHA) gene and locus changes in paediatric adrenocortical tumours from TP53 R337H mutation heterozygote carriers J. Med. Genet., May 1, 2004; 41(5): 354 - 359. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Shimasaki, R. K. Moore, F. Otsuka, and G. F. Erickson The Bone Morphogenetic Protein System In Mammalian Reproduction Endocr. Rev., February 1, 2004; 25(1): 72 - 101. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. A. Prendergast, L. L. Burger, K. W. Aylor, D. J. Haisenleder, A. C. Dalkin, and J. C. Marshall Pituitary Follistatin Gene Expression in Female Rats: Evidence That Inhibin Regulates Transcription Biol Reprod, February 1, 2004; 70(2): 364 - 370. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. C. Chapman and T. K. Woodruff Betaglycan Localization in the Female Rat Pituitary: Implications for the Regulation of Follicle-Stimulating Hormone by Inhibin Endocrinology, December 1, 2003; 144(12): 5640 - 5649. [Abstract] [Full Text] [PDF]
|
|
,
significantly different from control, but not different from
TGFß-treated cells (P < 0.05). The results of
one experiment of three with similar results are shown.
