| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ARTICLES |
Unité de Génétique et Biochimie du Développement, Unité de Recherche Associée 1960 Centre National de la Recherche Scientifique, Institut Pasteur, 75724 Paris Cédex 15, France
Address all correspondence and requests for reprints to: Dr. Isabelle Rosinski-Chupin, Unité de Génétique et Biochimie du Développement, Unité de Recherche Assocíee 1960 Centre National de la Recherche Scientifique, Institut Pasteur, 25 rue du Docteur Roux, 75724 Paris cedex 15, France. E-mail: ichupin{at}pasteur.fr
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We have been studying the regulation by androgens of genes expressed in the salivary submandibular glands (SMGs) of rodents. In particular, we have previously shown that the VCSA1 gene, a member of a multigene family called the variable coding sequence (VCS) family, is a target for androgens in the SMG of rats (9, 10). VCSA1 encodes an hormonal precursor, the submandibular rat 1 (SMR1) protein. Maturation of SMR1 leads to the production of a pentapeptide Gln/pyroGlu-His-Asn-Pro-Arg (11). This pentapeptide is secreted into the saliva and the bloodstream of male rats and may be involved in the modulation of transport and/or utilization of minerals (12). VCSA1 mRNAs, which accumulate at about 1000-fold higher levels in the SMG of male compared with female rats, have been localized in the acinar cells of the glands (13). These secretory cells are, along with ductal cells, the major cell type of rat SMG, accounting for about 50% of the total cell number. They were previously shown to respond to androgen stimulation by synthesizing at least one other protein, named SMR2 (13, 14). The presence of an AR in these cells has been demonstrated (15).
In contrast to many target cells for androgens, acinar cells do not seem to depend upon androgens for their differentiation. Differentiation of acinar cells occurs during the first 3 wk of postnatal life (16, 17, 18), when only low levels of androgens circulate in the bloodstream. In addition, there is no obvious sex-linked difference in either the morphology or the relative number of acinar cells. Therefore, these cells are excellent models to discriminate between the respective contributions of tissue-specific effects and androgen responses on gene expression. Studying the variations of VCSA1 expression during the time course of puberty can help to explain how small variations in circulating hormonal inducers can lead to responses of large amplitudes. Indeed, androgen concentrations vary by less than 10-fold in rats in males during sexual development.
Dose-dependent responses to inducers, such as steroid hormones, have been well studied but rarely examined with the aim of knowing whether the physiological effect on gene expression is a response of individual gene templates, or of the mass of the templates. A graded model of transcription has been proposed, in which the level of gene expression in individual cells directly correlates with the concentration of inducer. This model has often been opposed to a all-or-none or binary model in which, once the threshold concentration of inducer has been reached, transcription is triggered maximally. Evidence for one or the other model has been obtained in transfection assays (19, 20, 21, 22). In the present work, we have asked whether in all differentiated acinar cells, VCSA1 expression is induced, progressively and proportionally to the androgen concentration, during the period of sexual maturation.
To address the question of VCSA1 induction at the single cell level, we performed in situ hybridization (ISH) experiments on SMG sections at different periods of postnatal development in male and female rats. We show that while all acinar cells express VCSA1 in adult male SMG, the expression in immature rats and in adult females is mosaic, with high levels of VCSA1 mRNA being detected in some acinar cells and other cells being negative. Androgens primarily affect the number of VCSA1-expressing acinar cells, but an increase in transcriptional activity in the positive cells also occurs. Changes in the methylation pattern of the gene, during the time course of induction, suggest that regulation of VCSA1 expression might involve a remodeling of chromatin structure.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Probes
The following plasmids were used to generate probes: 1)
VCSA1 probes: pHV-VA1, a 234-bp fragment covering the
hypervariable region of VCSA1 [nt 174- 407 of the cDNA
(9)] cloned in a pcDNAII vector (Invitrogen Corp., Leek, The Netherlands) at the EcoRV
site, was used to generate VCSA1-specific RNA-probes for
RNase protection and ISH experiments; pInt2VA1, a 1.9-kb-long
BamHI-EcoRI fragment corresponding to the first
part of intron 2 (23) and cloned in pcDNAII, was used for
in situ detection of VCSA1 heterogeneous
nuclear RNA (hnRNA); 2) Actin probes: rpcAct and rpAct150
corresponding to the 1.8 kb nearly entire cDNA for rat
ß-actin (24) used for Northern blot analysis,
and to a 150-bp fragment covering the first 90 bp of the coding
sequence and 60 nucleotides (nt) upstream, in the 5'UTR, used for RNase
protection, cloned in pcDNAII; 3) GRP probes: pcG-1 for ISH
analysis has been described (13); 4) AR probe: a
fragment corresponding to nt 2351–2642 of the cDNA sequence
(25) (coding for part of the androgen binding site) was
amplified after retrotranscription of rat prostate mRNAs using as
primers oligo 17094 (5'-TCCTGTGCATGAAAGCACTGC-3' and oligo 17118
(5'-GAAAGGATCTTGGGCACTTGC-3') and cloned in pMOS Blue-T vector
(Amersham Life Science, Little Chalfont, UK).
RNase protection experiments
The VCSA1 gene belongs to a multigene family and two
other genes (VCSA2 and VCSA3) have very related
sequences except in a hypervariable sequence localized in exon 3
(26). RNase protection experiments were performed, using
this sequence as a probe and in preliminary experiments we verified
that, under our conditions of RNase protection, RNAs obtained by
in vitro transcription of VCSA2 and
VCSA3 were not detected. To quantify VCSA1 mRNA
during development, we used ß-actin mRNAs as an internal
control. A preliminary Northern blot analysis revealed that, relative
to a fixed amount of total RNA, ß-actin mRNA levels
decrease during the first 3 wk after birth by about 6-fold; this
decrease was taken into account when comparing results from rats at
different ages. No significant difference in SMG ß-actin
mRNA levels was observed between male and female rats at a given age.
SMG RNAs were prepared from individual SMGs as previously described
(9). pHV-VA1 and rpAct150 were used to generate
32P-labeled antisense RNAs. These probes were
hybridized simultaneously for 16 h with 5–20 µg of SMG total
RNA at 45 C in the presence of 80% formamide; 40
mM
piperazine-N-N'-bis-[2-ethanesulfonic
acid], pH 6.4; 0.4 M sodium acetate; pH
7; and 1 mM EDTA and digested by RNase one
(Promega Corp., France) according to the manufacturer’s
instructions. Signals were quantified using a PhosphorImager
(Molecular Dynamics, Sunnyvale, CA). For
quantification in high-expressing tissues (6-wk-old and adult male
rats), VCSA1 probes with lower specific activity were used
to remain in the ranges of simultaneous detection of
ß-actin and VCSA1 mRNAs.
ISH. ISH experiments on SMG sections were performed as previously described (13), with digoxigenin- or 33P-labeled complementary RNA probes. Partial hydrolysis to reduce probe size was only performed for the VCSA1 intron 2 probe. For detection of radio-labeled probes a NTB2 (Kodak, Rochester, NY) emulsion was used. Digoxigenin-labeled probes were revealed by Nitro blue tetrazolium/5-bromo-4-chloro-3-indolyl-phosphate alkaline phosphatase detection as previously described (13). When simultaneous detection of radiolabeled and dig-probes was performed, the dig-probe was revealed first, and the revelation time was chosen to not interfere with the further detection of the radiolabeled probe.
The proportion of SMG cells expressing VCSA1 mRNA at a given age was determined on an average of 300–500 acinar cells by microscopic observation. When the proportions of labeled cells were very low, the number of acinar cells in a field was estimated and a minimum of 100 fields was observed.
The detection limit was tested by comparing ISH results on sections of the same gland, after different times of exposure, ranging from 20 h to 21 d. The number of positive cells was found to plateau by 3–6 d of exposure, indicating that the absence of VCSA1 mRNA detection in some cells is really due to an absence of expression in these cells and not to a threshold limit of detection.
CpG methylation determination by Southern blot analysis
Genomic DNA was prepared from spleen and kidney of 11-wk-old
male rats and from SMG of rats at different ages. SMG from 5 rats at
birth were pooled as well as SMG from 3-wk-old male (n = 3) and
female (n = 3). SMG DNA from older rats was prepared and analyzed
individually. Genomic DNA was analyzed by double digestion with
HindIII or Sau3A I and one methylation-sensitive
enzyme: Hga I or HpaII (New England Biolabs,
Hitchin, UK) (see Fig. 7A
). Probes A and C (see Fig. 7) were
obtained by PCR amplification using specific primers and labeled by
single strand PCR in the presence of 32P-dCTP.
Signals were quantified with PhosphorImager.
|
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
).
Surprisingly, in the same conditions, a small proportion of acinar
cells were also detected in females (not shown). These cells were
nearly as intensely labeled as cells in males. Varying exposure times
up to 21 d indicated that the maximal proportion of
VCSA1 expressing acinar cells was obtained by 3 d and
led to the conclusion that there exists two populations of acinar
cells: one expressing VCSA1 at a high level, the other for
which expression is undetectable, even with long exposure times (21-d
exposure, Fig. 1C). In addition, in situ detection of
VCSA1 transcripts in the SMG of three random females
revealed varying numbers of expressing cells. Parallel quantification
of VCSA1 mRNAs on the contralateral glands showed a clear
relationship between global VCSA1 mRNA levels and the number
of acinar expressing cells (data not shown). We concluded that, in
female rats, VCSA1 mRNA levels are determined by the number
of expressing cells and therefore, that VCSA1 expression in
SMG acinar cells is mosaic. This prompted us to determine whether this
situation was specific to females or might be encountered in other
conditions of low concentrations of circulating androgens, such as
found before the puberty period in males.
|
, and 4-wk-old rats, Fig. 2, B and C).
|
|
). We therefore concluded that androgens are
necessary to induce VCSA1 expression in all acinar
cells.
Quantification of VCSA1 mRNAs during SMG development
To determine the kinetics of VCSA1 induction, we
quantified VCSA1 mRNAs by RNase protection assays in rat SMG
during the first 11 wk of postnatal life, which includes the period of
sexual maturation in male rats. Indeed, rat plasma T is minimal around
10–15 d after birth and then increases, reaching maximal values by
7–9 wk of age (29, 30).
As shown in Fig. 3A, we found low
VCSA1 mRNA levels at birth; expression thereafter
continuously increased to reach a plateau by 6–9 wk of age in male
rats. At this time, expression was, on average, 125-fold higher than in
18-d-old rats. These results are consistent with the fact that the
increase in VCSA1 expression depends on both acinar
differentiation and plasma T concentrations.
|
). This showed that maximal mRNA
levels were observed in females between 3 and 4 wk of age, which
corresponds to the end of acinar differentiation. The same result was
obtained when GRP mRNAs were quantified (data not shown).
Surprisingly, an apparent decrease in the mean VCSA1
expression was observed in females at adulthood (11-wk-old females),
compared with 6-wk-old females. In addition, we noticed large
interindividual variations in expression levels, ranging over two
orders of magnitude, for 11-wk-old females. These variations were
specific for VCSA1 and were not observed for
ß-actin or for the nonandrogen-controlled acinar-specific
GRP gene (not shown).
The androgen-induced increase in VCSA1 expression during male
puberty involves an increase in the number of expressing cells and an
increase in expression per cell
Because previous ISH experiments revealed a lower proportion
of acinar cells expressing VCSA1, in prepubescent compared
with adult male rats, induction of VCSA1 expression was also
studied at the single cell level, using ISH (Fig. 4A).
We found that the proportion of acinar cells expressing
VCSA1 was relatively constant during the first 12 d
(around 6%). Thereafter, from d 18 onwards, it increased reaching
100% between 6 and 11 wk of age. At each age (except in adult males),
both positive cells and negative cells were found, sometimes in the
same acinar structure (Fig. 2B and data not shown).
Because the increase in the percentage of expressing cells in males
during puberty did not totally account for the increase in
VCSA1 expression, the mean VCSA1 mRNA content per
positive cell during SMG development was calculated (Fig. 4B). This
revealed that the mean mRNA content per expressing cell was stable
during the first three wk and began to increase from the 4th wk
onwards. On average, there are about 25-fold more VCSA1
transcripts per expressing cell in adult male rats compared with
prepubertal 3-wk-old males. This suggests that androgens also affect
VCSA1 mRNA levels in the positive cells.
We conclude that the physiological rise in androgen levels leads to a gradual increase in both the number of expressing cells (binary response) and in the level of expression per cell (graded response). The increase in the number of expressing cells occurs first, at lower concentrations of circulating androgens.
VCSA1 transcription occurs in a higher proportion of nuclei and is
more active in the presence of high concentrations of
androgens
Although the increase in VCSA1 expression through cell
recruitment is likely to indicate a de novo induction of
VCSA1 transcription, the increase in VCSA1 mRNA
content could be due to an increase in transcription rate or/and in
VCSA1 cytoplasmic mRNA stability. To date, a way to measure
transcription rate at the level of single cells does not exist.
Therefore, as an alternative way, we used the in situ
detection of VCSA1 primary transcripts (hnRNAs). Indeed,
primary transcripts, which are temporally associated with the
transcription process and are short lived, because of rapid splicing,
are considered as accurate indicators of ongoing or very recent
transcription. The detection of hnRNAs was performed with an antisense
RNA probe corresponding to part of the VCSA1 intron 2
sequence. Under these conditions, labeling was observed specifically in
the acinar cells, in the SMG of males, before puberty (Fig. 5A) or at adulthood (Fig. 5B), as well as
in the SMG of female rats (not shown). As expected for primary
transcripts, labeling was concentrated on the cell nuclei. In contrast,
no labeling was obtained using the sense RNA probe, demonstrating that
signals are not due to DNA (data not shown). Surprisingly, in adult
11-wk-old males, we found that about 30% of acinar nuclei did not
contain detectable amounts of hnRNA (Fig. 5C), whereas 100% of acinar
cells were previously shown to contain VCSA1 mRNA. This
suggests that, even in the continuous presence of androgens,
VCSA1 may not be continuously transcribed. In adult males,
the proportion of labeled nuclei was higher than in females (not shown)
and immature males (Fig. 5C). In addition, a number of nuclei in adult
male SMG were more heavily labeled than in 4-wk-old rats (Fig. 5C) or
in females, suggesting that VCSA1 is more actively
transcribed in the expressing cells in the presence of high androgen
concentrations. However, although changes in hnRNA levels are usually
thought to reflect altered transcription, similar results could also be
obtained by changes in hnRNA stability.
|
, AR mRNAs can be detected both in
acinar and ductal cells but appear a little more abundant in the acinar
cells. No differential accumulation in particular acinar cells either
in males or in females (Fig. 6, A and B) was observed.
|
). We found that, in the SMG, methylation at the promoter sequence as well as in the 3' region was complete before differentiation of acinar cells, reflecting repression of expression. A progressive demethylation was observed in the promoter region, which depends on the age, but not on the sex of the rat. Furthermore, the high percentage of demethylation (around 70%) observed in this region in adult males and females suggested that acinar cells (which represent about 50% of total SMG cells) were probably not the only cells where demethylation occurred. This suggests that demethylation at the promoter is not completely cell specific. Consistent with this observation, a partial demethylation was also observed in the kidney, but not in the spleen, two organs negative for VCSA1 expression.
In contrast, we observed a differential demethylation in the 3' region depending both on age and sex of the rats, in correlation with VCSA1 differential expression. In addition, we found that demethylation in exon 3 was more than 40% in adult males but less than 10% in adult females, consistent with the fact that acinar VCSA1-expressing cells represent about 50% of total cells in males but less than 5% in adult females. In contrast, in the two organs negative for VCSA1 expression, no demethylation of exon 3 could be observed. Thus, the region around exon 3 is differentially methylated depending on organ and cell type and depending on androgen concentration.
Therefore, VCSA1 is globally more heavily methylated in females or prepubescent males than in adult males. This suggests that regulation of VCSA1 expression could involve epigenetic mechanisms and that changes in VCSA1 transcriptional activity during androgen response are correlated with certain changes in chromatin structure.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We have shown that VCSA1 expression is initially detected around birth when acinar cells begin to differentiate. mRNA levels then increase proportionally with acinar differentiation up to 3 wk of age in both sexes. Although expression reaches a maximum level in females, a second phase of increase is observed in males from 3–4 wk, in parallel with the increase in circulating androgens and sexual maturation. Thus, our data are consistent with the idea that VCSA1 expression depends both on acinar cell differentiation and androgen levels. VCSA1 response to androgens results in a 125-fold higher mRNA level on average in adult male rats compared with 18-d-old prepubescent males.
Surprisingly, a decrease in VCSA1 expression was observed in the adult females, compared with younger females. The decrease in VCSA1 expression is not specific for females because similar low levels of expression were also observed in adult males castrated at 4 wk of age. This indicates that, in the absence of high androgen concentrations, the decrease might occur as a consequence of aging. Therefore, the magnitude of the previously observed sex-linked difference in VCSA1 expression (1000-fold on average) results from two different events: 1) a strong up-regulation of VCSA1 expression in response to androgens in males during the period of sexual maturation; and 2) an apparent down-regulation as a consequence of aging in the absence of high androgen concentrations.
We previously reported the detection of VCSA1 mRNAs in the whole population of SMG acinar cells in adult male rats (13). Here we show that, surprisingly, before puberty but at the end of the acinar cell differentiation period, VCSA1 is expressed only in a subpopulation of acinar cells. This is in contrast with expression of constitutively expressed acinar markers (18) and suggests that differentiation of acinar cells is not sufficient to lead to homogeneous VCSA1 basal expression. Therefore, what we called basal level of expression, determined by mRNA quantification on the bulk population of cells, is due to a small number of cells that express VCSA1 at high levels and not to a generalized expression at a low level. The reason why VCSA1 is expressed in these few cells, whereas the majority of acinar cells are transcriptionally silent, is not yet clear. This might be due to gene activation by a yet unknown stimulus depending on cell fate or position; also we cannot exclude a pathway involving the AR itself. Indeed, in conditions of low androgen concentrations, it might be possible that the different androgen-regulated genes found in acinar cells compete for the binding of a limiting amount of activated AR. Small variations of coactivators or other transacting factors, acting in concert with the AR to promote gene expression might also be involved.
Methylation of DNA has been proposed as a means of regulating gene expression and hypermethylated states of chromatin are often associated with gene silencing (31, 33, 34, 35). Our results show that VCSA1 is globally hypermethylated before acinar differentiation. This hypermethylated state persists after acinar differentiation, in prepubescent and in female rats. One notable exception concerns demethylation at a CpG in the promoter region. This demethylation clearly does not correlate with a mechanism of gene activation because it also occurs in the kidney, an organ negative for VCSA1 expression. Therefore, our data are in favor of a chromatin-mediated repression of VCSA1 expression, involving gene methylation, in most acinar cells in conditions of low androgen concentrations. This repressive context does not seem to preclude the induction of VCSA1 transcription in response to androgens during puberty. Furthermore, after exposure to androgens, a change in VCSA1 methylation status occurs, suggesting that a consequence of androgen exposure is the establishment of a less repressive chromatin structure. Similarly, estrogens were previously demonstrated to induce gene expression even in a hypermethylated context (36) and to induce changes in gene methylation patterns (37).
During puberty, VCSA1 expression is strongly increased in response to androgens. Previous results on cell systems revealed that dose-dependent responses to inducers can be interpreted according to either a binary or a graded model of transcription (19, 20, 21, 22). According to the binary (or all or none) model of transcription, an increase in the number of VCSA1 expressing cells was expected to be the major determinant of the VCSA1 response to the rise of circulating androgens at puberty. Results consistent with this model were previously observed for gene activation by glucocorticoids, in a transfection model, when expression was studied at the single cell level. In that system, the examination of dose-dependent gene expression induced by glucocorticoid receptors in response to dexamethasone revealed a heterogeneous distribution of cells that either did or did not express the reporter gene. When the concentration of dexamethasone increased, the percentage of expressing cells was shown to increase (20). In contrast, the graded model of expression predicted that VCSA1 expression would be gradually and simultaneously induced in all previously negative acinar cells, in proportion to the increase in androgen concentration. The results we have obtained in this study give evidence for a combination of both types of response. We observed that the first effect of androgens is to modify the relative size of the expressing vs. nonexpressing cell populations. Acinar cells are shifted from a silent state to an expressing state, characterized by high VCSA1 hnRNA and mRNA levels. In a second phase, androgens induced a graded and progressive increase in VCSA1 mRNA content in the expressing cells. This increase in VCSA1 mRNA per cell is partly correlated with a slight increase in hnRNA levels in the expressing cells, which is likely to reflect a higher transcriptional activity under conditions of higher androgen concentrations.
Interestingly, we observed that, in adult male rats, while mRNAs were detected in 100% of acinar cells, hnRNAs were detected in only 70% of cells. A lower proportion of acinar cells positive for hnRNA compared with mRNA was also found in prepubescent male and in female rats. Because mature RNAs are expected to decay at lower rates than premessengers, this could indicate that transcription is shut off in some of the acinar cells. It was recently reported that during gene response to saturating doses of estrogens, the assembly of a transcription complex involving ER and coactivators occurs in a cyclic fashion, resulting in waves of transcription (38). In addition, covalent modification of coregulators was observed that could be involved in this cycling (39). Transcriptional shut off of VCSA1 transcription might be related to a similar cycling of the transcription complex.
A major conclusion of this work is that VCSA1 response to androgens is strongly amplified by the combination of two different mechanisms, i.e. the control of the number of expressing cells and the control of the expression level per cell. This provides experimental evidence that large amplitudes of gene regulation in response to an inducer might generally be linked to a mechanism of transcriptional repression in most cells under non induced conditions. In that context, chromatin-mediated repression of transcription might play a central role by controlling basal gene expression and therefore the amplitude of gene response.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Abbreviations: hnRNA, Heterogeneous nuclear RNA; ISH, in situ hybridization; nt, nucleotides; SMG, submandibular gland; SMR1, submandibular rat 1; VCS, variable coding sequence family.
Received April 16, 2001.
Accepted for publication June 19, 2001.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
1 activation domain. Mol Cell Biol 20:2004–2013This article has been cited by other articles:
 |
|
 |
|
  A. Wisner, E. Dufour, M. Messaoudi, A. Nejdi, A. Marcel, M.-N. Ungeheuer, and C. Rougeot Human Opiorphin, a natural antinociceptive modulator of opioid-dependent pathways PNAS, November 21, 2006; 103(47): 17979 - 17984. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. D. Anway, M. A. Memon, M. Uzumcu, and M. K. Skinner Transgenerational Effect of the Endocrine Disruptor Vinclozolin on Male Spermatogenesis J Androl, November 1, 2006; 27(6): 868 - 879. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J.-F. Huaulme, Y. Courty, F. Rougeon, and I. Rosinski-Chupin Androgen Regulation of SMR2 Gene Expression in Rat Submandibular Gland: Evidence for a Graded but not a Binary Response J. Histochem. Cytochem., October 1, 2003; 51(10): 1317 - 1329. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
