| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Westmead Institute for Cancer Research (R.L.A.-M., J.D.G., A.R.H., P.A.M., L.L.S., N.G., A.deF., C.L.C.), University of Sydney at the Westmead Millennium Institute, and Department of Gynaecological Oncology (L.L.S., N.G., A.deF.), Westmead Hospital, Westmead, New South Wales 2145, Australia; and Institut National de la Santé et de la Recherche Médicale Unité 673 (A.G.), Université Paris V, Unité de Gynécologie, Assistance Publique-Hôpitaux de Paris Hôtel-Dieu, Paris, France
Address all correspondence and requests for reprints to: Dinny Graham, Westmead Institute for Cancer Research, Westmead Hospital, Westmead, New South Wales 2145, Australia. E-mail: dinny_graham{at}wmi.usyd.edu.au.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Nuclear receptors have been reported to move into nuclear aggregates or foci in the presence of ligand. The most commonly used approach in these studies has been the visualization of transiently transfected green fluorescent protein (GFP)-tagged receptors, and this has revealed ligand-regulated movement into foci of estrogen receptor-
(ER) (11), androgen (AR) (12, 13), glucocorticoid (GR) (14), mineralocorticoid (MR) (15), vitamin D (16, 17), and retinoid X receptors (16). Less commonly, nuclear distribution of endogenous receptors has been detected using immunofluorescence (18), and there have been very few studies examining nuclear receptor distribution in animal or human tissues (10, 19).
The nucleus is a highly ordered structure, containing numerous specialized subnuclear components localized in discrete domains with biological functions that are still largely unknown. The specific nuclear functions of transcription, RNA processing, and replication are organized into compartments within the nucleus (20), and transcription is localized at the surface of transcriptionally active euchromatin territories where RNAPolII activity is detected (20, 21, 22). Chromatin remodeling by histone deacetylases is associated with distinct nuclear domains (23). Transcription factors and coregulators, such as p300, p53, steroid receptor coactivator- (SRC-), and glucocorticoid receptor interacting protein 1, also aggregate in the nucleus in clusters, discrete domains, or foci (24, 25). The ubiquitin-proteasome pathway of protein degradation is localized in nuclear compartments (26, 27). An increasing number of nuclear bodies with unknown function, such as PML bodies and Cajal bodies, show discrete localization in the nucleus (22, 28, 29, 30). These nuclear bodies are thought to increase efficiency of transcription and RNA processing by bringing together molecules of related function, or alternatively, they could be storage regions for inactive factors that function at sites of lower concentration (28).
The function of PR located in the different nuclear distributions in human tissues in normal physiology remains unclear. Our previous study, showing that PR distribution in vivo is regulated by cyclical progesterone (10), suggests that PR foci may be involved with active transcription, because they are most prominent in a phase of the menstrual cycle where serum progesterone levels are high and at a time when clearly defined morphological and functional effects of progesterone are detectable in the endometrium (31, 32). However, because ligand-mediated degradation of PR is mediated by the 26S proteasome (33) and transcriptional activity of PR is linked to proteasome activity (34), the possibility that PR foci are in the process of proteasomal degradation also requires exploration.
Little is currently known about the subnuclear distribution of endogenous PR and the role of ligand in regulating PR distribution. Focal PR subnuclear distribution has not previously been demonstrated in mouse tissues or in the human breast, and investigation of the regulation and function of PR foci forms the focus of this study. The aim of this study was to explore the hypothesis that the subnuclear distribution of PR isoforms is regulated by progestins and to investigate the function of PR isoforms when distributed into nuclear foci, in particular the possibility that PR isoforms localized to nuclear foci are involved in active transcription.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
), whereas in proestrus evening, when serum progesterone levels are highest, PRA and PRB were focal (Fig. 1B). This supported our previous findings in the menstrual cycle in humans (10) and suggested that serum progesterone may drive this redistribution of PR within the nucleus.
|
). Exposure for 6 h to physiological levels of estradiol resulted in an overall increase in PR levels, but, as had been observed in vehicle-treated controls, PR was evenly distributed in the majority of nuclei in the uterine epithelium of estradiol-treated mice, and few PR foci were seen (Table 1). Exposure to combined progesterone and estradiol resulted in a redistribution of PR into a predominantly focal pattern of detection (Table 1). The subnuclear distribution of PR was independent of PR expression level. These data provide clear evidence that PR is detected both in an even nuclear distribution and in discrete nuclear foci in normal mouse tissue, and that the process of movement into foci is ligand-dependent.
|
). PR foci in endometrial cancers were larger, with a median length of 1.04 µm (0.78–1.94 µm interquartile range) (Fig. 2), and the difference between the size of foci in normal and malignant tissues was statistically significant (P < 0.01, Kruskal Wallis).
|
), supporting the observations made in the endometrium (10) and in the mouse uterus (Table 1
). We also observed PR distributed into foci in human breast cancers. Around 50% of PR-positive breast cancers displayed PR distribution into nuclear foci, and there was no difference between pre- and postmenopausal breast cancers in the likelihood that PR would be focally distributed (Table 2). These data showed that not all PR-positive breast cancers had focal PR and also suggested that, in contrast to normal tissue, movement of PR into foci in breast cancers was independent of the systemic endocrine environment, because premenopausal women with cycling hormones were no more likely to display focal PR than postmenopausal women lacking the levels of serum hormones observed before menopause.
|
), and foci numbers were increased after exposure to the synthetic progestin ORG2058 (Fig. 1D). Treatment with the partial antagonist RU38486 for 2 h also resulted in the movement of PR into foci (Fig. 1E), although to a lesser extent than for ORG2058 treatment (Fig. 1, compare D and E). Because newly translated PR protein is associated with a complex of chaperone and heat-shock proteins, we asked whether nascent PR is also localized in foci by depleting cells of PR and determining nascent PR distribution once PR levels had recovered. As expected, PRA and PRB proteins detected on immunoblots had mobilities of 83 kDa and 116 kDa, respectively, and PRB was detected in three closely migrating bands, representing different levels of basal phosphorylation of the unliganded receptor (Fig. 3A). Treatment with progesterone for 24 h resulted in down-regulation of both PRA and PRB to very low levels compared with the vehicle-treated control as demonstrated by immunoblot analysis of whole cell extracts and by immunofluorescent detection in fixed cells (Fig. 3A). After a medium change to fresh phenol red-free medium containing charcoal-stripped fetal calf serum (sFCS), PR protein levels began to recover, and after 8 h low levels of PRA and PRB were detectable. After 24 h (48 h after initial exposure to progesterone), substantial reexpression of new PRA and PRB protein was seen, although expression levels were lower than in matched vehicle-treated controls (Fig. 3A). Nascent PRA and PRB distribution was predominantly even throughout the nucleus, and this was seen in over 70% of cells. In the remainder, some PR foci were noted, with a mixed even and focal PR distribution. The predominantly even distribution of nascent PR suggested that chaperone- and heat shock protein-associated PR was not focal, and this was supported by the demonstration that inhibition of hsp90 association, using geldanamycin, had no effect on ligand-dependent PR foci formation (data not shown).
|
), and this persisted to 6 h (Fig. 3C). After 2 h of ORG2058 treatment, 77.5% of cells contained PRA foci (Table 3, even+focal or foci only), compared with 29.4% in controls (P < 0.01, Spearman’s rank correlation). PRB foci were seen in 66.3% of cells after 2 h of ORG2058 compared with 27.7% in controls (P < 0.01; Spearman’s rank correlation). After 6 h of ORG2058 exposure, 76.7% and 82.2% of cells contained PRA or PRB foci, respectively (Fig. 3C), and both were significantly different from time-matched controls (P < 0.01 and P < 0.01, respectively; Spearman’s rank correlation). Many cells that contained PR foci also displayed some even staining for nuclear PRA and PRB, suggesting that PR protein may be in excess in these cells. This was not surprising because T-47D cells constitutively express PR at a high level.
|
), whereas 60.5% of cells had mixed even+foci PRB staining and just 5.9% were exclusively focal for PRB (Table 3). After 6 h of progestin exposure, the proportion of cells with exclusively focal PRB increased from 5.9 to 17% (Table 3), and this increase was statistically significant (P < 0.01; Spearman’s rank correlation). In summary, both PRA and PRB were distributed into significantly more nuclear foci after progestin treatment than in control cells. Furthermore, exclusive distribution of PR into nuclear foci was only seen in ORG2058-treated cells and was never observed in vehicle-treated cells. This was observed for both PRA and PRB foci.
PR Foci Contain Ligand-Regulated Receptor Dimers
To determine whether PR foci reflected the formation of PRA and PRB heterodimers and/or homodimers, we measured fluorescence resonance energy transfer (FRET) of full-length PRA and PRB fusion proteins tagged at the N terminus with cyan or yellow fluorescent protein (CFP and YFP) upon transient expression in PR-negative cells. Transfected cells treated for 2 h with 10 nM ORG2058 or vehicle were fixed, and expression and interaction between CFP and YFP fusion proteins was imaged by fluorescent microscopy. After background subtraction and correction for CFP and YFP cross talk to the FRET image, a color-encoded image was generated that represented the level of FRET signal at each point in the cell (Fig. 4A). The mean background- and cross talk-corrected FRET ± SEM measured for each receptor pair is shown in Fig. 4B. In untreated cells, the fluorescent PR fusion proteins were evenly distributed in the nucleus, and foci were seldom observed. The FRET signal arising from these cells was very low (Fig. 4, A and B). Exposure to ORG2058 resulted in a striking redistribution of fluorescent PR proteins into prominent foci (Fig. 4A). Foci were detected whether one PR isoform or both PRA and PRB were transfected. Treatment with ORG2058 caused a marked increase in the FRET signal. In all combinations tested, the FRET signal at PR foci was significantly higher than vehicle-treated control nuclei (Fig. 4B; Mann Whitney test; P < 0.001) and was also significantly higher than the FRET signal measured in similarly sized regions of ORG2058-treated nuclei that did not contain foci (P < 0.01; data not shown). Although there were differences in the magnitude of FRET signals between different CFP/YFP receptor pairs, the fold increase in FRET signal was consistent when treated nuclei and foci were compared with the matched control combination. In PRA-expressing cells, the FRET signal increased 3-fold in whole nuclei and 4-fold in PRA foci of ORG2058-treated cells compared with vehicle-treated cells. In PRB-expressing cells, FRET was increased 4-fold in whole nuclei and 6-fold in PRB foci in ORG2058-treated compared with vehicle-treated cells. In cells expressing both PRA and PRB, the FRET signal in ORG2058-treated whole nuclei was increased 2- to 3-fold, and 3- to 4-fold in foci compared with control cells. These data demonstrate that PR foci within progestin-treated nuclei contain the highest density of FRET-producing PR dimers.
|
) is representative of the pattern of localization seen for both PR isoforms. As observed previously, PR distribution in vehicle-treated cells was predominantly even with a low basal level of foci, and treatment with ORG2058 resulted in a highly focal pattern of PR detection (Fig. 5). In the presence of actinomycin D or DRB, PR foci were abolished, and a predominantly even nuclear distribution of PR was detected in both control and ORG2058-treated cells (Fig. 5). Inhibition of proteasomal degradation using MG132 had a delayed effect on the formation of PR foci. The nuclear distribution of PR in ORG2058-treated cells remained focal after 1 or 2 h of cotreatment with MG132 (Fig. 5; and data not shown), but longer cotreatment with MG132 and ORG2058 resulted in disruption of PR foci. Cells pretreated for 2 h with MG132 followed by cotreatment for 6 h with ORG2058 had a predominantly even PR distribution, and foci were no longer detectable (Fig. 5), suggesting that PR complex turnover is important for the continued detection of foci.
|
, A–C). The same results were seen for both PRA and PRB. PR and PML foci were located in separate parts of the nucleus and rarely colocalized in cells treated for 1 h with either vehicle (data not shown) or ORG2058 (Fig. 6, A and B). These observations were confirmed in the merged image (Fig. 6C). The few instances where PR and PML foci colocalized or appeared closely associated were likely to be coincidental. The lack of coassociation of endogenous PR and PML, seen in T-47D cells, was confirmed in normal human tissue by dual detection of PR and PML in human secretory phase endometrium. PR and PML nuclear foci were clearly visible under dual excitation in the human endometrium and did not colocate (Fig. 6D). This supports the lack of association between PR and PML bodies observed in T-47D cells.
|
) and PRB foci (Fig. 6G) were dark (Fig. 6E, arrows), representing active euchromatin. The formation of PR foci was related to maintenance of chromatin structure, as treatment of T-47D cells with the histone deacetylase inhibitor trichostatin A (TSA), which remodels chromatin, was associated with alterations in PR foci formation. Firstly, PR foci were augmented in cells treated only with TSA (Fig. 6J), suggesting a diminished requirement for ligand in foci formation upon chromatin disruption. Secondly, PR foci formed on ligand treatment of TSA-exposed cells were larger than foci formed in the absence of TSA (Fig. 6, compare L and M), suggesting that maintenance of the median length of foci was regulated in part by the physical association of PR at specific chromatin locations.
Ligand Promotes Association of PR Foci with Transcriptional Coregulators
PR foci colocalized with the transcriptional coactivator p300, which is known to associate with ligand-activated PR on chromatin (36, 37). Dual detection of PRB and p300 demonstrated that they were rarely colocalized in vehicle-treated cells, and foci containing only p300 were clearly visible (detected as green foci in the merged image; Fig. 6Q). On ORG2058 treatment, p300 and PRB colocalized in a number of foci (Fig. 6T). PRA was also associated with p300, because both proteins colocated in the same foci in ORG2058-treated cells (data not shown). Recruitment of the coactivator SRC-1 and subsequent acetylation of histone H4 are involved in the transcriptional activity of PR (36, 38), and we asked whether SRC-1 recruitment was required for formation of PR foci. The cyclin-dependent kinase inhibitor roscovitine inhibits PR-dependent recruitment of SRC-1 (39): cotreatment of cells with roscovitine and ORG2058 for 1 or 2 h prevented formation of PR foci (Fig. 6N, and data not shown), demonstrating that PR foci represent ligand-dependent PR-containing transcriptional complexes.
Growth Factors Do Not Affect PR Foci Formation
Progestins and growth factors synergistically enhance transcription of key cell cycle components, including cyclin D1, cyclin E, and p21WAF1 (40), demonstrating cross talk between these signaling pathways. To determine whether growth factors may play a role in PR foci formation, PR foci were examined in T-47D cells treated with growth factors in the presence or absence of progestins. Treatment with epidermal growth factor (EGF; 30 ng/ml) or hepatocyte growth factor (HGF; 10 ng/ml) alone for 15 min to 2 h had no effect on PR nuclear distribution, and PR was detected with a predominantly even distribution in these cells (Fig. 7; and data not shown). Growth factors had no effect on ligand-dependent movement of PR into foci, because PR foci were similarly observed in cells cotreated 2 h with ORG2058 and either EGF or HGF (Fig. 7). Moreover, EGF treatment for 15 min had no effect on PR foci preformed by 2 h pretreatment with ORG2058 (data not shown).
|
) and became more prominent with 30-min incorporation (Fig. 8B). In contrast, in vehicle-treated cells, the same distinct puncta were not observed, although at the longer run-on time point some overlap in staining was seen, as demonstrated by the darker regions in the Boolean logic analysis (Fig. 8B).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
PR Forms Foci in Normal Tissues Exposed to Serum Hormones
In the cycling mouse uterus, PR was predominantly focal in late proestrus, when serum progesterone levels were highest, supporting the view that progesterone is required to drive PR into foci. In the human, in a rare cohort of normal breast tissues dated accurately in the menstrual cycle by reference to the uterus from the same woman (41), all cases in the progesterone-dominated luteal phase were focal, whereas most cases in the follicular phase were not. Taken together, the data from cycling mouse and normal human breast in the menstrual cycle support our previous findings in the human endometrium, that PR is more likely to be focal in target tissues at cycle phases associated with high serum progesterone levels. It was also evident that both PRA and PRB move into foci in human and mouse tissues. In cells containing both PRA and PRB at similar levels, the majority of foci contained both PRA and PRB, although foci containing either PRA or PRB but not both were also observed (data not shown).
Ligand Requirement for Focal Distribution of PRA and PRB
PRA and PRB are detected in nuclear foci in the uterine epithelium of ovariectomized mice that have been exposed to estradiol and progesterone. Progesterone is necessary for this to occur, because in ovariectomized mice treated with vehicle or estradiol alone, PR distribution was predominantly even. To test the role of ligand in PR foci formation directly, we asked whether newly expressed PR proteins were detected in nuclear foci by exploiting the reversible ability of the natural ligand progesterone to down-regulate the receptor protein (42, 43), allowing subsequent reexpression of nascent PR protein when medium is replaced (44). Progesterone treatment for 24 h completely down-regulated PRA and PRB in T-47D cells, and after an additional 24 h in fresh ligand-free medium a marked recovery of receptor levels was noted, which was concordant with the literature (42, 45). The subnuclear distribution of newly synthesized, steroidally naive PR, in an environment deprived of hormones, was predominantly even with very little or no focally distributed PR. This was observed for both PRA and PRB and was supported by the observation that PR foci were unaffected by the hsp90 inhibitor geldanamycin, suggesting that in the absence of ligand, the receptor is distributed diffusely throughout the nucleus. Because it is well established that nascent PR is associated with a complement of chaperone proteins (46), it can be concluded that this complex is not visualized as a focal distribution.
When T-47D cells expressing nascent PR were treated with the synthetic progestin ORG2058, PRA and PRB redistributed into nuclear foci. PRA and PRB were located in separate foci and also colocated in the same focus, as we had observed in mouse and human tissues. This supports the view that progestin exposure caused movement of PR into foci in T-47D cells. PR distribution after ligand treatment was not confounded by the known ability of ligand to down-regulate PR protein levels (47), because this pattern of staining was observed earlier than any decrease in PR protein level could be detected (data not shown). Exclusively focal PR staining was only ever observed in progestin-treated cells. In most nuclei containing PR foci, focal distribution was accompanied by the coexistence of evenly distributed PR, probably because the high PR levels in T-47D cells resulted in excess PRs that were not redistributed at the concentration of ligand used. The mixed even+foci distribution was also seen in a small percentage of vehicle-treated cells. It is not known whether these reflected the ligand-independent movement of PR into foci, or whether low levels of ligand either in mouse uterus or in the culture medium promoted PR foci formation, but the former possibility is supported because focal distribution of PRA and PRB in the absence of ligand has been demonstrated in T-47D cells (48). Foci in control cells did not result from growth factor-mediated effects on PR (49, 50), because neither EGF nor HGF exposure caused PR distribution into foci.
FRET measurements using transfected fluorescent PR fusion proteins proved that PR foci represent ligand-dependent dimerized PR. The highest levels of FRET in progestin-treated cells were measured in foci, and this was true for PRA or PRB homodimers and for PRA-PRB heterodimers. PRA and PRB also distributed into foci in the presence of the antagonist, RU38486. Although RU38486 blocks the action of progesterone, this type II antagonist promotes the association of the receptor with DNA and is also well known to have agonist activities (51, 52, 53, 54). The movement of PR into foci was less extensive in RU38486-treated cells when compared with ORG2058-treated cells, and, because RU38486 has a lower affinity than ORG2058 for PR, this result is in keeping with the demonstration that the ability of ligands to promote nuclear distribution of GR was related to their affinity for the receptor (55).
The ligand-regulated localization of PR in nuclear foci was not restricted to T-47D cells. A similar redistribution pattern, from even expression to predominantly focal after ORG2058 treatment, was seen in PR-positive MCF-7 breast cancer cells and ECC-1 endometrial cancer cells (data not shown). Taken together, the distribution of PR into nuclear foci appears to be a consequence of ligand binding. The evidence in this study that ligand is required for movement of the majority of PR into foci is consistent with the data for other nuclear receptors (Ref. 55 and references therein). In addition, for GR it has been shown that the ligand binding domain is required for nuclear aggregation and the extent of movement into foci is related to the affinity of the ligand for the receptor (55), further implicating ligand binding in nuclear localization. The role of ligand in directing nuclear localization of PR is also consistent with the demonstration that dynamic movement of PR within the nucleus is ligand-dependent and is required for recruitment and productive interaction with target promoters in living cells (56).
Association of PR Foci with Components of Known Nuclear Bodies
Sumoylation of PRA has been shown to confer dominant negative transcriptional activity (35), and, because sumoylated proteins are sequestered to PML bodies (57, 58), we reasoned that movement of PR into foci may be associated with PR sequestration into PML bodies. However, endogenous PR did not localize in PML bodies in T-47D cells or in human endometrial tissue. This supported the in vitro findings of Guiochon-Mantel et al. (59), who failed to find endogenous PR located in PML bodies. There was also no association of PR foci with other known nuclear bodies, including Cajal bodies (data not shown).
Association of PR Foci with Active Chromatin and Requirement for Proteasomal Activity
Our results demonstrated that PR foci locate in areas with less intense DAPI staining, suggesting that PR foci associate with transcriptionally active euchromatin. Focal clusters of GFP-MR also associated with euchromatin (15), and AR nuclear foci are mainly localized to the peripheral region of the transcriptionally active euchromatin (12, 60). Chromatin remodeling is important to PR transcriptional activation of target genes (36), and we demonstrate in this study that disruption of chromatin by inhibition of histone deacetylase activity disrupted PR foci formation. PR foci were larger in cells treated with the histone deacetylase inhibitor TSA, and there was reduced reliance on ligand for their formation. Conversely, blocking recruitment of coactivator SRC-1 to PR transcriptional complexes and subsequent inhibition of histone H4 acetylation abolished ligand-dependent PR foci, further linking foci to the formation of active PR-containing complexes on euchromatin. Interestingly, inhibition of SRC-1 recruitment with roscovitine does not block PR association with DNA as measured by chromatin immunoprecipitation (39), suggesting a functional dissociation between the ability to bind DNA and the formation of multiprotein complexes incorporating PR dimers and associated transcription factors and coactivators.
Focal localization may be part of the proteasomal degradation process (26, 27). We found that PR foci are not disrupted after short-term exposure to the proteasome inhibitor MG132 but were abolished after a longer exposure. This suggests that formation of foci, at least in the early stages, is not linked to the degradation process. Our results on the role of 26S proteasome in the early stages of focus formation are consistent with a study by Tyagi et al. (60) in which the discrete distribution of GFP-AR did not alter in the presence of MG132. Similarly, in live cells treated with MG132, GFP-ER accumulates in bright nuclear foci (61). Prolonged exposure to MG132 abolished PR foci, demonstrating that proteasome activity is linked to PR action at later times. These findings are consistent with the demonstration that inhibition of the 26S proteasome by MG132 blocked PR-mediated transcription by inhibiting recruitment of RNAPolII (34). Overall, the data from this study suggest that PR foci are associated with euchromatin and are reliant on chromatin structure for assembly, and that the initial stages of movement of PR into foci are not associated with proteasomal activity but that protein turnover is important for the longer-term maintenance of PR foci.
Function of Nuclear Receptors in Foci
Although nuclear foci may have a number of functions (62), it is clear that each nuclear receptor displays specific nuclear distribution and dynamics, and therefore the functional implications of nuclear foci formation may be different for each receptor. MR foci may not be located at transcriptional sites because the use of transcriptional inhibitors failed to alter the distribution of MR, and MR foci only partially overlapped with hyperphosphorylated RNAPolII (63). Although there was a correlation between transcriptional activation of a glucocorticoid responsive reporter and the presence of GFP-tagged GR nuclear foci (14), there was no relationship between GR and RNAPolII or nascent RNA, despite there being a strong positive correlation between the spatial distribution of RNAPolII and nascent RNA representing sites of active transcription (18, 62). ER foci appear to indicate ligand-driven association with chromatin (11, 64), but their link to transcriptional activation remains controversial (64, 65). The transcriptional activity of PR in foci was demonstrated by the colocation of PR foci, in cells treated with ligand, with the transcriptional coactivator p300, and with newly synthesized RNA using in situ transcriptional run-on assays. Although the dual immunofluorescent detection of PR and p300 as yellow foci provides only correlative evidence of their association in active complexes, the ligand-dependence of this codetection, the loss of foci resulting from blocking SRC-1 recruitment, and the evidence that PR foci contain PR dimers associated with active transcription suggests that these foci are likely to represent true protein associations.
PR Foci in Cancer
Nuclear organization plays an important part in the regulation of transcription, and disruption to subnuclear structures such as have been detected in cancer (66) results in changes likely to have profound effects on transcription and subsequent modifications in target gene expression (67). PR foci have previously been detected in endometrial cancers, and the occurrence of PR in foci in those cancers was associated with cancer grade (10), which is a key predictor of outcome. We found that there were a number of aberrations in foci in cancer compared with foci in normal tissues. Firstly, the size of foci in endometrial cancers was greater than in normal endometrium, suggesting that foci in cancers are comprised of a larger number or different complement of proteins than in normal tissues. This difference in size between normal and cancer tissues is likely to be related to the known alterations in chromatin structure in cancers and was confirmed in this study when disruption of chromatin structure resulted in larger PR foci. Secondly, in breast cancers we noted that there was no difference between pre- and postmenopausal cases in foci formation, indicating that the reliance on serum progesterone for foci formation demonstrated in cycling normal tissues was lost in cancers. This supported our previous finding in normal tissues adjacent to endometrial cancers: whereas normal tissue tended to have even PR distribution, adjacent cancers were focal, which confirms a different reliance on circulating hormones for foci formation between normal and cancer tissues sharing the same systemic environment (10). The reduced reliance on ligand for foci formation in cells treated with the chromatin disruptor TSA also provided correlative evidence that altered chromatin structure in cancer may result in the competence of PR to form foci even in the relative absence of ligand. The final difference between normal tissue and cancer in PR foci formation was the relative distribution of PRA and PRB in foci. Whereas both PRA and PRB formed foci in normal tissues, in endometrial cancers PRB was more common than PRA in foci (10). Taken together, the results of this and our previous study suggest that there is aberrant PR focus formation in cancer compared with normal tissues, and this is likely to have an impact on PR-mediated transcription and subsequent target gene expression.
Proposed Model of Subnuclear Distribution of Progesterone Receptor
The proposed model of the subnuclear distribution of PRA and PRB is shown in Fig. 9. In the absence of hormones, PR was evenly distributed in the nucleus. Therefore we propose that unliganded, transcriptionally inactive PR distributes evenly in the nucleus. This is supported by the data in the human normal endometrium where PRA and PRB were distributed evenly in the proliferative phase of the menstrual cycle (10), in the normal breast when serum levels of progesterone are very low, and in the cycling mouse uterus in the diestrus phase of the estrous cycle. In the presence of ligand, both PRA and PRB move rapidly into dimers and are detected as foci. PR foci associated with p300, blocking SRC-1 recruitment blocked their formation, and PR colocated with newly synthesized RNA, indicating its role in transcription in the presence of ligand. This study has established that subnuclear location of PR is regulated by ligand binding in humans, mice, and cell lines. The association between PR foci and transcriptional activity demonstrates that regulation of subnuclear location is likely to be a physiologically important aspect of PR function.
|
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Mouse Tissues
BALB/c mice were obtained from the Animal Resource Centre (Perth, Western Australia, Australia), and were housed in humidity- and temperature-controlled rooms with a 12-h light, 12-h dark cycle, with food and water provided ad libitum. All experiments were approved by the Institutional Animal Care and Ethics Committee.
For studies in cycling mice, vaginal smears were taken daily to monitor progression through the estrous cycle (69). Eight-week-old BALB/c mice exhibiting two consecutive 4- to 5-d cycles were killed during diestrus (1000 h) and on the evening of proestrus (2200 h) (n = 3 per stage). For studies on hormone-treated ovariectomized mice, 9-wk-old BALB/c mice were ovariectomized, and after a minimum of 7 d, mice were treated with vehicle (mineral oil), estradiol (1 µg in mineral oil) alone, or estradiol and progesterone (1 µg and 1 mg, respectively, in mineral oil) for 6 h. Mice were anesthetized with ketamine (100 µg/g body weight) and xylazine (10 µg/g body weight) for blood collection, then killed by cervical dislocation. Their uteri were removed and immediately fixed in neutral buffered formalin and paraffin embedded using routine methods.
Cell Culture
T-47D cells (E. G. and G. Mason Research Institute, Worcester, MA) were grown in antibiotic- and phenol red-free RPMI 1640 medium, supplemented with 10% fetal calf serum (FCS; JRH Biosciences, Melbourne, Australia), and insulin (0.25 U/ml). PR-negative U-2 OS osteosarcoma cells were maintained in DMEM supplemented with 10% FCS. All cell lines were maintained in exponential growth through regular passaging. Cell lines were routinely tested for Mycoplasma contamination and shown to be negative. For all immunohistochemistry experiments, approximately 5 x 104 T-47D cells were seeded in each well of a six-well plate, each well containing an acetone-washed glass coverslip, and allowed to grow for 3 d before treatment. After treatment, cells were fixed with freshly prepared 3.7% formalin in PBS or with 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4) for 30 min at 4 C and permeabilized with cold 100% methanol for 20 min at 4 C. For transfections, U-2 OS cells were plated in glass four-chamber slides with DMEM containing 10% FCS and were allowed to grow for 24 h before transfection. On the day of transfection, the medium was removed and replaced with DMEM containing 5% sFCS.
For progesterone treatment experiments, T-47D cells were treated for 24 h with either progesterone (100 nM; Sigma-Aldrich, Castle Hill, New South Wales, Australia), or vehicle (ethanol, 0.1% final concentration). Cells were washed with 5% sFCS and then placed in fresh medium containing 5% sFCS for 24 h. Cells were fixed or harvested for protein extraction at 24 h of progesterone treatment and 8 and 24 h after medium replacement. For progestin treatment of steroidally naive PR, cells were treated as above; however, after 12 h in 5% sFCS, cells were treated with either ORG2058 (10 nM; Organon Chemicals through Amersham Australia Pty. Ltd., Sydney, New South Wales, Australia) or vehicle (ethanol, 0.1% final concentration) for 2 or 6 h. In experiments examining regulation of PR distribution by the antiprogestin RU38486 (a gift from Dr. J. P. Raynaud of Roussel-Uclaf, Romainville, France), cells were treated with either RU38486 (100 nM) or vehicle (ethanol, 0.1% final concentration) for 2 h. For transcriptional inhibition experiments, cells were treated with ORG2058 (10 nM) or vehicle (ethanol, 0.1% final concentration) in the presence of either the transcriptional inhibitor actinomycin D (1 h, 4 µM; Sigma-Aldrich), DRB (2 h, 50 µM; Sigma-Aldrich), or vehicle (ethanol, 0.1% final concentration). Proteasomal degradation was blocked in the presence or absence of ORG2058 (1 and 6 h, as indicated) by treatment with the inhibitor MG132 (15 µM; Sigma-Aldrich). Cells were pretreated with MG132 or vehicle for 6 h, followed by 1-h cotreatment with 10 nM ORG2058 or vehicle. For longer cotreatment, cells were pretreated 2 h with MG132 or vehicle, followed by 6-h cotreatment with ORG2058 or vehicle. For chromatin disruption experiments, T-47D cells growing on coverslips were treated with 1 µM TSA (Sigma-Aldrich) or vehicle for 16 h, followed by 6-h treatment with 10 nM ORG2058 or vehicle. Recruitment of SRC-1 to PR was inhibited by cotreatment of cells with 30 µM roscovitine and 10 nM ORG2058 or vehicle for 1 h (Sigma- Aldrich). In experiments examining growth factor effects, cells growing on coverslips in sFCS-containing medium were treated with EGF (30 ng/ml; Sigma-Aldrich) or HGF (10 ng/ml; Sigma-Aldrich) for up to 2 h, as indicated, in the presence of 10 nM ORG2058 or vehicle. The cells were rinsed with PBS and fixed as described above.
Protein Extract Preparation and Immunoblotting
Cells were harvested, and protein extracts were prepared using RIPA buffer as previously described (70). Protein concentration in the extracts was estimated using Bradford dye reagent (Bio-Rad, Regents Park, Australia). Equal protein amounts were loaded onto 7.5% polyacrylamide-sodium dodecyl sulfate gels, and PR proteins were fractionated, transferred to nitrocellulose, and visualized by immunoblotting as previously described (71).
Dual Immunofluorescent Staining
Sections from formalin-fixed, paraffin-embedded tissues were antigen retrieved and stained for PRB, then PRA as described previously (8). Briefly, deparaffinized and rehydrated sections were placed in 0.01 M sodium citrate buffer (pH 6.0) and subjected to antigen retrieval using a Tuttnauer 2540 EKA autoclave (Tuttenauer Co. Ltd., Jerusalem, Israel) at 121 C, 15 psi for 30 min. After blocking with normal goat serum, sections were incubated with a mouse antihuman PR monoclonal antibody that detects PRB only (hPRa6, 1/5, overnight) (72) and with a biotinylated goat antimouse antibody (Dako Pty. Ltd., Sydney, New South Wales, Australia), and Texas red (TXR)-avidin (Molecular Probes, Eugene, OR). To block sites of potential cross-reactivity between the two staining sequences, sections were incubated overnight with goat antimouse Ig Fab (Cappel antibodies; ICN Biomedical, Aurora, CA). Sections were incubated with a mouse monoclonal antibody (hPRa7, 1/10, 1 h), which only detects PRA in formalin-fixed tissues (8, 73) and with a biotinylated goat antimouse antibody (Dako) and fluorescein isothiocyanate (FITC)-avidin (Calbiochem-Novobiochem Corp., Sydney, New South Wales, Australia). Sections were mounted with Vectashield mountant for fluorescence (Vector Laboratories, Inc., Burlingame, CA) and stored in the dark at 4 C.
Dual immunofluorescent staining of PR isoforms in mouse tissues was the same as described above except that a Fab blocking step was applied for 1 h before normal goat serum blocking step to prevent any cross-reactivity of mouse antibody with mouse tissue. Primary antibodies were titrated to find the optimal dilution for staining in mouse tissues: hPRa6 was used at 1/20 and hPRa7 at 1/40.
For experiments examining colocalization of PRA or PRB with other nuclear components, cells fixed on coverslips were antigen retrieved and stained as described above with the following exceptions: PR was detected as the first sequence (visualized with TXR), and the second primary antibody was either the mouse antihuman PML monoclonal antibody (1/100, 1 h; Santa Cruz Biotechnology, Inc., Santa Cruz, CA) or the mouse antihuman p300 monoclonal antibody (1/50, 1 h, Oncogene Research Products, San Diego, CA). Where the primary antibody was a rabbit polyclonal antibody, the secondary antibody used was biotin conjugated goat antirabbit Ig (Dako). For the experiments examining the association between chromatin and PR, distribution cells were stained for PRA and PRB by dual immunofluorescence and counterstained with 1.5 µM DAPI (Sigma-Aldrich) after FITC incubation.
Controls were treated and stained in the same way as the test sections and coverslips. Controls included adjacent sections or coverslips stained using antibody diluent (PBS/0.5% Triton X-100): 1) in place of both primary antibodies to control for nonspecific staining; and 2) to replace the second sequence primary antibody to ensure no cross-reactivity between the two staining sequences.
Single Immunofluorescent Staining
Cells fixed on coverslips were antigen retrieved as described above and stained by incubating with either hPRa6 (PRB) or hPRa7 (PRA), at the concentrations and times described above, followed by incubation with biotinylated goat antimouse antibody (Dako) and FITC-avidin (Calbiochem-Novobiochem). The sections were mounted in Vectashield fluorescent mountant (Vector Laboratories) and stored in the dark at 4 C.
Analysis of PR Distribution by Fluorescence Microscopy
PR staining was examined using an Olympus BX 40 microscope (Olympus, North Ryde, Australia) fitted with filters to detect both TXR (BP 545–580) and FITC (BP 450–480) fluorescence simultaneously and each of the two fluorochromes separately. The whole section or coverslip was examined in detail, both under individual fluorochrome excitations and using the dual filter, to identify the distribution of PR. Total PR distribution was determined by examining the localization pattern of TXR and FITC fluorescence under the dual filter, and individual PR isoform distribution was determined by examining the localization pattern of fluorescence under the individual fluorochrome excitations. Nuclear morphology was visualized with DAPI using a wide UV (BP 330–385) filter.
Analysis of PR distribution into nuclear foci in normal breast and breast cancer tissues was carried out on sections stained for PRA and PRB by dual immunofluorescence, as described above. PR distribution was determined by examining the nuclear localization pattern in an average of three representative high-power fields (x400 magnification) under the dual filter. Cases were considered to be focal if 5% or more of the cells examined displayed discrete foci. Each case was scored by three independent observers.
Analysis of Subnuclear PR Distribution by Confocal Microscopy
Distribution of PR within nuclei was studied using confocal microscopy [OptiScan F900e personal confocal system fitted with a two-line Kr-Ar laser (Optiscan Pty. Ltd., Notting Hill, Victoria, Australia), which provides excitation at 488 nm and/or 568 nm, attached to an Olympus BX 40 fluorescent microscope]. Four areas within each section identified as being representative of PR distribution within the entire section were analyzed further by three-dimensional sequence acquisition with a step size of 0.2 µm. The Optiscan software (F900e version 1.6.0) image profiling tool was used to obtain X-Y and Z profiles of staining patterns. Confocal imaging of cells stained by dual immunofluorescence for BrUTP-containing nascent RNA and PR protein was performed using a Leica TCS SP2 confocal system (Leica, North Ryde, Australia) fitted with Ar (458-nm, 476-nm, 488-nm, and 514-nm excitation lines) and HeNe (633-nm) lasers and attached to a Leica DMRE upright fluorescent microscope. Overlapping detection of PR protein and nascent RNA was determined by mathematical modeling using a Boolean logic protocol that produces a grayscale image where intensity is directly proportional to level of pixel overlap in matched fluorescent scans.
Detection of Nascent RNA by Run-On Incorporation of BrUTP
Cells growing on coverslips were rinsed with cold TBS (10 mM Tris, pH 7.4; 150 mM NaCl; 5 mM MgCl2), then washed at 4 C, 10 min with glycerol buffer (20 mM Tris, pH 7.4; 5 mM MgCl2; 0.5 mM EGTA; 25% glycerol; 0.5 mM phenylmethylsulfonylfluoride). The cells were permeabilized for 3 min at 4 C using glycerol buffer containing 0.05% Triton X-100 and 25 U/ml RNase inhibitor, then immediately incubated with transcription buffer (50 mM Tris, pH 7.4; 150 mM NaCl; 10 mM MgCl; 25% glycerol; 0.5 mM ATP; 0.5 mM CTP; 0.5 mM GTP; 0.5 mM BrUTP; 0.5 mM phenylmethylsulfonylfluoride, 25 U/ml RNase inhibitor) containing 10 nM ORG2058 or vehicle, as indicated, for 5 or 30 min at room temperature. After run-on incorporation of BrUTP, the cells were rinsed with the permeabilizing buffer and fixed with 3.7% formaldehyde, 0.1% glutaraldehyde in PBS, for 30 min at 4 C or with 4% paraformaldehyde in 0.1 M phosphate buffer for 30 min at 4 C. The fixed cells were subsequently washed with cold PBS, and BrUTP and PR were detected by dual immunofluorescence.
FRET Analysis of PR
Fluorescently tagged PRA and PRB constructs were prepared by insertion of PRA and PRB cDNA sequences derived from pSG5-hPR1 (a gift from Prof. P. Chambon, Institut de Genetique et de Biologie Moleculaire et Cellulaire, Strasbourg, France), in-frame at the 3'-end of the CFP and YFP sequences in pECFP-C1 and pEYFP-C1 (Clontech, Mountain View, CA). The CFP- and YFP-tagged PRA and PRB constructs were transfected separately and in CFP/YFP paired combinations into U-2 OS cells growing in glass four-chamber slides, using FuGene (Roche Diagnostics, Castle Hill, Australia) and following the manufacturer’s recommended protocol. Cells were incubated at 37 C for 18 to 20 h after transfection, then treated for 2 h with either 10 nM ORG2058 or vehicle. The cells were rinsed briefly with PBS, then fixed for 1 h at 4 C with 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4). After fixation, the cells were washed with PBS and mounted using Vectashield mounting medium. Results were analyzed using a Zeiss AxioVert 200M inverted fluorescent microscope (Zeiss, North Ryde, Australia) fitted with filters for detection of CFP (excitation 436/20, emission 480/40), YFP (excitation 500/20, emission 535/30), and CFP to YFP FRET (excitation 436/20, emission 535/30). Emission intensities and extent of FRET were determined using AxioVision FRET software. To control for cross talk from donor (CFP) and acceptor (YFP) to FRET detection, intensity readings of single transfected samples expressing CFP-PRA, CFP-PRB, YFP-PRA, or YFP-PRB were measured, and a median donor and acceptor correction factor was determined from greater than 20 nuclei in each instance. The FRET intensity in dual transfected whole nuclei and in PR foci was measured using the same exposure time in each channel, and background and cross talk corrected FRET values were calculated using the method of Youvan et al. (74). By this method, corrected FRET intensity values were determined using the calculation: Fc = (fretgv – bgfret) cfdon * (dongv – bgdon) – cfacc * (accgv – bgacc), where gv is intensity as gray value, bg is background, cf is correction factor, fret is signal in FRET channel, don is signal in donor channel, and acc is signal in acceptor channel. FRET values were determined in multiple vehicle and ORG2058-treated cell nuclei and in individual PR foci and nonfocal regions of equivalent size within ORG2058-treated nuclei. Corrected FRET pixel intensities were visually represented as color encoded images. Low FRET intensities are represented by "cold" black-blue colors, intermediate intensities are "warm" green-yellow colors, and high FRET intensities are represented by "hot" red-white coloring. The color to intensity range is indicated in Fig. 4.
Statistical Analysis
The SPSS statistical program version 10.0.5 (SPSS Inc., Chicago, IL) was used to perform all statistical tests. The Kruskal-Wallis one-way ANOVA was used to test whether the size of PR foci was different between the normal and malignant endometrium. Spearman’s rank correlation was used to test for an association between PR isoform distribution and treatment with either vehicle or ORG2058 at 2 or 6 h, after PR depletion and recovery. A Mann Whitney U test was used to compare the FRET values from vehicle- and ORG2058-treated cell nuclei and subnuclear foci and nonfocal regions of ORG2058-treated cells.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
This work was supported by the National Health and Medical Research Council of Australia and the U.S. Army Medical Research and Materiel Command under DAMD17-00-0696. R.L.A.-M. was supported by a Department of Education, Training and Youth Affairs Australian Postgraduate Award and by the Westmead Millennium Foundation. N.G. was supported by the Westmead Millennium Foundation and the Westmead Hospital Gynaecological Oncology Research Fund.
Disclosure Statement: The authors have nothing to disclose.
First Published Online October 4, 2006
1 R.L.A.-M. and J.D.G. contributed equally to the manuscript. 
Abbreviations: AR, Androgen receptor; BrUTP, 5-bromouridine triphosphate; CFP, cyan fluorescent protein; DAPI, 4, 6-diamino-2-phenylindole; DRB, 5,6-dichloro-1-ß-D-ribofuranosyl benzimidazole; EGF, epidermal growth factor; ER-, estrogen receptor-; FCS, fetal calf serum; FITC, fluorescein isothiocyanate; FRET, fluorescence resonance energy transfer; GFP, green fluorescent protein; GR, glucocorticoid receptor; HGF, hepatocyte growth factor; MR, mineralocorticoid; PR, progesterone receptor; sFCS, charcoal-stripped FCS; SRC, steroid receptor coactivator; TSA, trichostatin A; TXR, Texas red; YFP, yellow fluorescent protein.
Received for publication January 23, 2006. Accepted for publication September 25, 2006.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
