| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Studies |
Westmead Institute for Cancer Research, University of Sydney, Westmead Hospital, Westmead, NSW 2145, Australia
Address all correspondence and requests for reprints to: Dr. C. L. Clarke, Department of Medical Oncology, Westmead Hospital, Westmead, NSW 2145, Australia. E-mail: chrisc{at}westmed.wh.su.edu.au
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The importance of the N-terminal region of PR is highlighted by the fact that although both PR A and PR B are able to mediate progestin-activated transcription in transfection studies, they differ in vitro in their efficacy and specificity for target genes. Although PR B activates the progestin responsive reporter MMTV-CAT more effectively (11), PR A stimulates a tyrosine aminotransferase progestin-responsive reporter to a greater extent than PR B (12) and acts synergistically with estradiol to induce the ovalbumin promoter, whereas PR B is inhibitory under the same conditions (13). Furthermore, there is evidence that although PR B is normally a transcriptional activator, the ability of PR A to function as an activator is more restricted and, in settings where both PR B and PR A activate transcription, PR B is more effective than PR A (14, 15, 16). In addition to the differences in functional activity of PR A and PR B, PR A can act as a dominant repressor of PR B function. Cotransfection of increasing amounts of PR A with PR B resulted in suppression of PR B activation of MMTV-CAT (12, 14). Furthermore, PR A can repress transcriptional activation, not only of PR B but also of cotransfected glucocorticoid, mineralocorticoid, androgen, and estrogen receptors (15, 17, 18), although not all the published reports are consistent (19). Clearly, the effects are complex and depend strongly on the cellular and promoter context. Whether the differences in function of PR A and PR B revealed by the in vitro studies detailed above are borne out in vivo remains to be determined.
The potential functional importance of the N-terminally truncated PR A and, consequently, the relative expression of PR A and PR B in determination of cellular response to progesterone in vitro, prompted a previous study, which measured PR A and PR B expression in breast cancers (20). It was notable that in addition to PR A and PR B, which were expressed at variable relative levels in vivo, a significant proportion of breast cancers contained a smaller PR protein of molecular mass 78 kDa (20). The significance of PR78kDa expression is unknown, and in particular, there are no data on whether PR78kDa is able to bind ligand and therefore potentially exhibit transcriptional activity. If this smaller PR species exhibits similar differences in function as have been evidenced in vitro for PR A relative to PR B, it is possible that this PR species may be an important component in determination of progesterone response in breast cancer.
The source of PR78kDa is not yet clear, although its expression was not noted in previous studies of normal endometrium (8) and may be a feature of malignant tissues. It is not known whether posttranslational mechanisms (such as proteolytic activity within breast tumors) lead to PR78kDa formation or whether posttranscriptional processes (such as alternative translational start site usage or alternative transcript splicing) are involved. There is evidence from breast tumor mRNA studies that several alternatively spliced estrogen receptor transcripts, comprising whole exon deletions, exist (21, 22). A number of the exons of PR are between 100 and 150 bp in length (23), and deletion of any of these could potentially result in the expression of a PR protein of 78 kDa.
The purpose of the present study was to characterize the PR78kDa protein in breast tumors, to determine whether it is able to bind the ligand (and therefore, potentially possess transcriptional activity), and to determine whether posttranscriptional mechanisms contribute to its formation in breast cancers.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Analytical reagent grade chemicals were obtained from the sources previously listed (24). Anti-PR monoclonal antibodies (25) were concentrated by ammonium sulphate precipitation of hybridoma supernatants and dialysed in PBS containing 0.02% sodium azide. The dilutions used for saturating detection of PR on immunoblots were determined for each batch. Abbott PgR-EIA and ER-EIA enzyme immunoassay (EIA) kits were from Abbott Laboratories, Diagnostics Division (Chicago, IL).
Preparation of tumor cytosols
Biopsies were collected at the time of breast lump excision and frozen at -70 C. The tissues were pulverized in a precooled vessel, homogenized at 4 C in a buffer containing 10 mmol/L Tris (pH 7.4–7.5), 1.5 mmol/L EDTA, 5 mmol/L sodium molybdate, and 1 mmol/L monothioglycerol and centrifuged at 436,000 x g for 20 min at 4 C. Cytosol protein concentrations were determined by a method based on that of Lowry et al (26). PR and ER concentrations were determined by EIA. The cytosols were stored up to 18 months at -70 C.
T-47D cells and cytosol preparation
T-47D breast cancer cells (27), cultured as described previously (24), were harvested, counted, and collected by centrifugation; and cell pellets were frozen at -70 C. Cell pellets were thawed on ice in PEMTG buffer (50 mmol/L potassium phosphate, 10 mmol/L EGTA, 10 mmol/L sodium molybdate, 12 mmol/L thioglycerol, and 10% glycerol, pH 7.0) containing protease inhibitors (final concentrations: 0.5 mmol/L phenylmethylsulfonyl fluoride, 86 µmol/L leupeptin, 77 µg/mL aprotinin, 1.4 µmol/L pepstatin A, and 100 µg/mL bacitracin) and homogenized. Samples were centrifuged at 100,000 x g for 1 h at 4 C, and cytosol protein concentrations were determined by the Bio-Rad protein assay (AMRAD, Melbourne, Australia).
Protein electrophoresis and immunoblot analysis
Tumor cytosol proteins (300 µg per lane) were separated on 7.5% SDS-PAGE gels, as previously described (24). PR proteins were detected using anti-PR monoclonal antibodies; hPRa 7 recognizes both PR A and PR B proteins, hPRa 6 detects PR B protein but not PR A (25). For routine detection of PR, a combination of hPRa 6 and hPRa 7 was used. Where immunoblot analysis of the same samples with hPRa 6+7 or hPRa 6 alone were compared, duplicate aliquots of four tumor cytosols (300 µg protein each) and T-47D cytosol (10 µg protein) were electrophoresed and blotted as described above. The blot was divided (one half incubated with hPRa 6 alone, the other with both antibodies). Specific PR protein bands were visualized on autoradiographic film using a horseradish peroxidase conjugated goat antimouse secondary antibody (Dako, Carpinteria, CA) at 1:5000 dilution and enhanced chemiluminescent detection substrates (Amersham, Sydney, Australia). T-47D breast cancer cell cytosol was included as a positive control. Band intensities were measured using a densitometer and Imagequant software (Molecular Dynamics, Melbourne, Australia). The linear range of detection of PR on immunoblots was established by a standard curve constructed using increasing concentrations of T-47D cytosol PR. Multiple exposures of each immunoblot were made, and results were taken only from those which fell within this linear range.
Heat treatment of breast tumor cytosols
Tumor cytosol (300 µg protein) was heated at 37 C in the presence or absence of protease inhibitors (final concentrations as stated above). After treatment, samples were mixed with SDS sample buffer and analyzed by immunoblotting. The band intensities of the PR A and B proteins were measured densitometrically.
Photoaffinity labeling with [3H]-R5020
To determine the binding of ligand to PR in tumor cytosols, tumor cytosols (300 µg protein) with known levels of PR78kDa were combined with tritiated synthetic progestin ([3H]-R5020) to a final ligand concentration of 25 nmol/L. T-47D cytosol was used as a control. Samples were incubated at 4 C for 2 h in the presence or absence of 100-fold molar excess unlabeled R5020. Unbound R5020 was removed by incubating the samples with dextran-coated charcoal (0.5% charcoal, 0.05% dextran) for 5 min on ice. The suspensions were centrifuged at 4 C for 3 min, and supernatants were transferred to fresh tubes and immediately exposed to UV light for 15 min to cross-link the ligand. PR proteins were separated by SDS-PAGE, were visualized by immunoblotting, and specific PR immunoreactive bands were excised. The amount of [3H]-R5020 bound was determined by scintillation counting, and specifically bound counts were calculated by subtraction of counts in samples containing excess unlabeled R5020.
Test for presence of proteases in tumor cytosols
To test whether proteases in breast tumor cytosols may give rise to truncated PR proteins including PR78kDa, T-47D cytosol (300 µg protein), photoaffinity labeled with [3H]-R5020 as described above, was incubated either on ice or at 37 C for 30 min with either 150 µg tumor cytosol (containing high levels of PR78kDa) or buffer. Proteins were separated by SDS-PAGE and visualized by immunoblotting. Radioactively labeled proteins were visualized by using a Molecular Dynamics phosphorimager and Imagequant software, excised, and radioactivity measured by liquid scintillation.
Detection of PR transcripts
Total RNA was isolated by the guanidinium isothiocyanate-cesium
chloride method, as described (24), from T-47D cells and four primary
breast tumors. First strand complementary DNA (cDNA) was reverse
transcribed using oligonucleotide primer PR4 (nt 3462–3441,
5'-CTGGAAATTCAACACTCAGTGC-3'), which anneals to sequences within exon 8
(Fig. 1
). Sequences of the primers used
and the nucleotide locations were derived from the published PR
sequence (10). Aliquots of the reverse-transcription reaction,
equivalent to cDNA from 50 ng total RNA, were used in PCR reactions
using primer sets PR1 [forward, nucleotide (nt) 2271–2288,
5'-GCGCTCTACCCTGCACTC-3'] + PR8 (reverse, nt 2999–2977,
5'-CTGAATGAGAGTTATCTGGTCAT-3') or PR3 (forward, nt 2677–2697,
5'-TCAGAGTTGTGAGAGCACTGG-3') + PR4 (reverse, nt 3462–3441,
5'-CTGGAAATTCAACACTCAGTGC-3'), which anneal to sequences within exons
1, 5, 4, and 8, respectively (Fig. 1). Cycling parameters were as
follows: 1 step at 95 C, 3 min; 30 cycles at 95 C, 30 sec; 60 C, 30
sec; 72 C, 1 min; and a final step at 72 C, 5 min. The amplified
products were separated by electrophoresis on 2% agarose gels,
transferred onto HybondN+ membrane by capillary blotting,
and probed with [32P]-labeled PR cDNA or exon-specific
oligonucleotide probes. Blots were probed sequentially, with stripping
between each hybridization, with oligonucleotide probes homologous to
sequences within exons 2 (PR5), 3 (PR6), 4 (PR3), 5 (PR7), and 6 (PR12)
(Fig. 1).
|
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Immunoblot analysis of T-47D breast cancer cell and breast tumor
cytosols (Fig. 2) revealed the PR A, PR
B, and PR78kDa species described previously (20). To determine whether
PR78kDa contained the N-terminal portion of the protein, immunoblots
were probed with hPRa 6, which recognizes PR B but not PR A and
therefore recognizes an epitope in the N-terminus of PR that is unique
to PR B (25). Neither PR A nor PR78kDa were recognized by the hPRa 6
antibody (Fig. 2). This indicates that PR78kDa lacks the N-terminal
epitope to which hPRa 6 is directed. PR78kDa also failed to be
recognized by another antibody, which recognizes both PR A and PR B
(not shown), suggesting that PR78kDa may be a truncated form of PR
A.
|
To determine whether PR78kDa contained the ligand-binding domain,
its ability to bind ligand was determined in vitro by
photoaffinity labeling. PR B and PR A specifically bound the ligand, as
expected (Fig. 3). In addition,
radioactivity was associated with PR78kDa and was specifically bound,
as evidenced by the fact that radioactive ligand binding was not
detected when excess unlabeled ligand was included. The amount of
specifically bound radioactive ligand was related to the levels of
immunoreactive PR78kDa protein detected (Fig. 3, tumors 1–3),
demonstrating that PR78kDa is able to bind ligand.
|
To determine whether protease activity in tumor cytosols
contributed to the formation of PR78kDa, tumor cytosols known to
contain the 78-kDa protein were heated at 37 C in the presence or
absence of protease inhibitors. Immunoblot analysis showed that short
exposure to elevated temperature decreased the overall level of PR
detected (Fig. 4). PR B levels were 66%
of unheated controls after 2 min and 50% of control after 5 min, and
similar decreases in PR A and PR78kDa levels, relative to control, were
observed (PR A: 61%, 2 min; 52%, 5 min. PR78kDa: 59%, 2 min; 55%, 5
min) (Fig. 4C). Inclusion of protease inhibitors partially abrogated
the decrease in PR B levels, particularly after 5 min of heating (77%
control after 2 min with protease inhibitors, 71% after 5 min) but had
little or no effect on the decrease in PR A (68% control after 2 min
with protease inhibitors, 44% control after 5 min) or PR78kDa levels
(63% control after 2 min with protease inhibitors, 39% control after
5 min) (Fig. 4C). The decrease in levels of PR was not accompanied by
an alteration in the relative expression of PR78kDa (Fig. 4). Inclusion
of ligand at saturating concentrations (ORG 2058, 10 nmol/L) and
prolongation of the heating time to 15 min or 30 min did not affect the
relative level of PR78kDa protein detected, although the overall levels
of PR decreased (not shown).
|
To determine whether tumors with high levels of PR78kDa contained proteases that contributed to its formation by degradation of wild-type PR B or PR A, photolabeled T-47D PR proteins were incubated at 37 C for 30 min with a tumor cytosol containing PR78kDa, and the effect on relative levels of radioactive T-47D PR proteins was noted. No relative loss of PR A or PR B was noted, and there was no alteration in the relative expression of PR78kDa (data not shown).
Expression of exon-deleted PR transcripts in relation to PR78kDa levels
To determine whether the presence of exon-deleted PR transcripts
was associated with expression of truncated PR proteins such as
PR78kDa, RNA was amplified from T-47D cells, which contain no PR78kDa,
and from tumors with known concentrations of PR78kDa. The strategy used
was to amplify portions of the PR open reading frame) encoding exons
2–7, because the size of these exons was such that if one were lost, a
deleted PR transcript would result that had the potential to encode a
protein approximately consistent in size with PR78kDa. Smaller products
consistent in size with PR transcripts lacking exon 2 (Fig. 5, panels c and d), exon 3 (Fig. 5, panel
d), exon 4 (Fig. 5, panels b and c), and exon 6 (Fig. 5, panel a) were
detected in all cases at very low levels relative to wild-type PR
transcripts. No transcripts lacking exons 5 or 7 were detected.
Transcripts lacking exon 4 or exon 6 were cloned and sequenced (not
shown), and the other truncated transcripts were confirmed by probing
of blots with exon-specific probes. Transcripts lacking more than one
exon were detected also, with very low amounts of transcripts lacking
exons 2+3 being evidenced upon prolonged exposure of the autoradiograph
in Fig. 5d (not shown). Transcripts lacking exons 5+6 were evidenced
also upon prolonged exposure of Southern blots probed with PR cDNA of
PCR products obtained with primer combination PR3+PR4 (not shown). A PR
transcript lacking half of exon 4 (4/2) (Fig. 5, panels b and c),
suggesting the possible existence of a cryptic splice site in exon 4,
was detected in all the tumors tested and in T-47D cells. This deletion
was confirmed by cloning and sequencing of the PCR-amplified fragment
(not shown).
|
The levels of exon-deleted PR transcripts were estimated
semiquantitatively, as described in Materials and Methods,
to be less than 10% of wild-type transcripts. The most abundant
exon-deleted transcripts detected were those lacking exon 6 or exon 4/2
(Fig. 6). The relative expression of
exon-deleted PR transcripts was compared with the relative expression
of the PR78kDa protein. PR78kDa ranged from 2% (tumor 1) to 64%
(tumor 4) of total PR (Fig. 6). There was no relationship between
presence and levels of exon-deleted PR transcripts and relative
expression of PR78kDa (Fig. 6). In fact, T-47D cells, which express
only wild-type PR protein, generally contained levels of exon-deleted
transcripts either in the same range as, or higher than those of tumors
with significant PR78kDa concentrations.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
PR78kDa was recognized by monoclonal anti-PR antibodies that recognize both PR A and B, but it was not recognized by an antibody specific for PR B (demonstrating that it lacks the epitope recognized by this antibody and implying that it lacks the B-specific upstream region). This suggested that this species may be a truncated form of PR A. A less plausible possibility was that PR78kDa may have the same start site as PR B but contain two truncations (one in the N-terminal region containing the hPRa 6 epitope and, because PR78kDa is smaller that PR A, another in other regions of the PR protein). On balance, the possibility that PR78kDa is a truncated PR A is more likely and is supported by the observation (not shown) that PR78kDa was not recognized by a third antibody that recognized both PR B and PR A.
One potential mechanism for PR78kDa formation is cleavage of the protein by protease activity. Proteolysis, if present, could occur either in vivo or in vitro. The time of storage of samples before immunoblot analysis was not a factor in detection of PR78kDa, because the relative levels of PR78kDa were not markedly elevated in samples which had been stored for longer time periods (not shown). Furthermore, it was not possible to form PR78kDa from either PR B or PR A in vitro. When tumors either containing or lacking this protein were incubated under conditions designed to promote proteolytic digestion, loss of PR protein was detected, as expected, and this could be prevented by inclusion of protease inhibitors. However, there was no relative increase in the levels of PR78kDa, suggesting that its formation was not a consequence of proteolytic degradation of PR B or PR A during cytosol preparation.
It is possible that proteolytic activity could be involved in PR78kDa formation at a stage coincident with tissue homogenization and cytosol preparation. This possibility cannot be eliminated but may be inconsistent with the evidence that although PR was susceptible to proteolytic activity in cytosols, no relative increase in PR78kDa was detected; and therefore, there was no evidence for the existence of a susceptible site giving rise to PR78kDa. Moreover, previous studies have documented that human PR contains sites that are susceptible to proteolytic digestion and that, upon limited digestion, give rise to PR fragments of 49 kDa and smaller (30, 31) but not to fragments consistent in size with the PR78kDa species observed in this study. It is important also to highlight the fact that levels of PR78kDa are not inversely proportional to levels of either PR A or PR B (not shown), suggesting that, at least by this criterion, it seems not to be derived from either of these species. In vivo proteolysis of PR in tumors is another possibility, and this could not be tested because of the difficulty of determining PR78kDa levels in breast tumors before homogenization. However, if in vivo proteolysis of PR takes place and it results in the formation of smaller proteins, such as PR78kDa, this process may have important implications for response to endocrine agents in breast cancer. Present studies are aimed at using immunohistochemistry to determine whether PR78kDa is present in tissue sections.
Because human PR is posttranslationally modified by phosphorylation, and this process leads to alterations in the electrophoretic mobility of the protein (31, 32), PR78kDa may represent an altered phosphorylation form of PR. It is unlikely to be a hypophosphorylated form of PR B, because the difference in molecular mass between PR B and PR78kDa is approximately 38 kDa. It may be a hypophosphorylated form of PR A, but this is also unlikely because phosphatase treatment of PR A does not lead to a PR protein of a size consistent with PR78kDa (4).
The role of exon-deleted PR transcripts in formation of PR78kDa was investigated. Although PR transcripts lacking one or more exons were detected in breast tumors, as described previously (33), their expression was at very low levels, relative to wild-type, even if the limitations in relative quantitation of exon-deleted and wild-type PR transcripts are considered. When the presence of exon-deleted PR transcripts was examined in tumors with known levels of PR78kDa protein, no relationship was seen, despite the fact that the size of proteins predicted to be encoded by transcripts lacking exons 3 or exon 4/2 was close to that of PR78kDa. It was also noteworthy that T-47D cells, in which there is no evidence for expression of PR78kDa or any other truncated PR protein apart from the wild-type PR B and PR A, contained among the highest levels of the exon-deleted PR transcripts. These data, and the very low levels of exon-deleted PR transcripts observed, argue that exon deletions of PR are unlikely to be implicated in the formation of truncated PR proteins, in general, and PR78kDa in particular.
The significance of exon-deleted PR transcripts is unknown, and it has yet to be shown that these transcripts are translated in vivo, although low levels of truncated PR proteins, other than PR78kDa, are present in breast tumors (20). The finding in this study that alternatively spliced PR transcripts are present in very low abundance, relative to wild-type PR, indicates that in these tumors, at least, they may have little impact on progestin responsiveness if translated into functional proteins. However, the stability of alternatively spliced PR transcripts has yet to be determined: if such transcripts have lower stability than wild-type transcripts, the measured levels of wild-type and alternatively spliced PR transcripts may not reflect the actual synthesis of such transcripts in vivo.
The potential functional capacity of PR78kDa was investigated by determining whether this protein was able to bind the ligand. Photoaffinity labeling studies demonstrated specific binding of ligand to PR78kDa, indicating presence of the ligand binding domain. This suggests that PR78kDa, if expressed in breast tumors in vivo, may have transactivating capacity and, if it is a truncated form of PR A, may exhibit a similar impact on PR function (11, 12, 13, 14) and cross-talk with other members of the nuclear receptor family (15, 17, 18, 19) as have been shown for PR A. The extent of PR78kDa expression in progesterone target tissues has yet to be determined, but its presence was not noted in previous studies on normal endometrium (8).
This study has investigated the origin of a truncated 78-kDa PR protein, detected in a significant proportion of human breast tumors, to determine likely mechanisms for its formation and consequently to understand its role, if any, in breast cancer. PR78kDa is likely to be derived from PR A and is able to bind ligand, indicating that if it is expressed in vivo, it may form part of endocrine signaling pathways in breast cancers. There was no evidence that PR78kDa was derived from proteolytic activity of either PR B or PR A. Similarly, although exon-deleted PR transcripts were detected, which could, if translated, give rise to a PR protein similar in size to PR78kDa, neither the abundance of such transcripts nor their relationship to levels of expressed PR78kDa protein supported a role for exon-deletion in formation of this truncated PR protein. Mechanisms that may be implicated in PR78kDa formation could include mutations in the PR gene in a subset of breast tumors, or in vivo processing of PR, and these possibilities require investigation and further studies.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Was a National Health and Medical Research Council Medical
Postgraduate Scholar.
Received May 29, 1997.
Revised October 3, 1997.
Accepted October 14, 1997.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  A. H. Taylor, P. C. McParland, D. J. Taylor, and S. C. Bell The Progesterone Receptor in Human Term Amniochorion and Placenta Is Isoform C Endocrinology, February 1, 2006; 147(2): 687 - 693. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. M. Katso, J. H. Parham, M. Caivano, W. C. Clay, J. P. Condreay, D. W. Gray, K. M. Lindley, S. J. Mason, J. Rieger, N. C. Wakes, et al. Evaluation of Cell-Based Assays for Steroid Nuclear Receptors Delivered by Recombinant Baculoviruses J Biomol Screen, October 1, 2005; 10(7): 715 - 724. [Abstract] [PDF]
|
|
|
|
 |
|
 R. L. Balleine, S. M. N. Hunt, and C. L. Clarke Coexpression of Alternatively Spliced Estrogen and Progesterone Receptor Transcripts in Human Breast Cancer J. Clin. Endocrinol. Metab., April 1, 1999; 84(4): 1370 - 1377. [Abstract] [Full Text]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 |
