- Email: [email protected]
Detection of mRNA levels for the estrogen alpha, estrogen beta and androgen nuclear receptor genes in archival breast cancer tissue
Cancer Letters 237 (2006) 248–255 www.elsevier.com/locate/canlet
Detection of mRNA levels for the estrogen alpha, estrogen beta and androgen nuclear receptor genes in archival breast cancer tissue Robert A. Smitha, Rod A. Leaa, Stephen R. Weinsteinb, Lyn R. Griffithsa,* a
Genomics Research Centre and Wesley Research Institute, School of Health Science, Griffith University Gold Coast, PMB 50 Gold Coast Mail Centre, Qld. 9726, Australia b Department of Pathology, Gold Coast Hospital, Southport, Qld., Australia Received 2 March 2005; received in revised form 3 June 2005; accepted 6 June 2005
Abstract Previous studies in our laboratory have shown association of nuclear receptor expression and histological breast cancer grade. To further investigate these findings, it was the objective of this study to determine if expression levels of the estrogen alpha, estrogen beta and androgen nuclear receptor genes varied in different breast cancer grades. RNA extracted from paraffin embedded archival breast tumour tissue was converted into cDNA and cDNA underwent PCR to enable quantitation of mRNA expression. Expression data was normalised against the 18S ribosomal gene multiplex and analysed using ANOVA. Analysis indicated a significant alteration of expression for the androgen receptor in different cancer grades (PZ0.014), as well as in tissues that no longer possess estrogen receptor alpha proteins (PZ0.025). However, expression of estrogen receptors alpha and beta did not vary significantly with cancer grade (PZ0.057 and 0.622, respectively). Also, the expression of estrogen receptor alpha or beta did not change, regardless of the presence of estrogen receptor alpha protein in the tissue (PZ0.794 and 0.716, respectively). Post-hoc tests indicate that the expression of the androgen receptor is increased in estrogen receptor negative tissue as well as in grade 2 and grade 3 tumours, compared to control tissue. This increased expression in late stage breast tumours may have implications to the treatment of breast tumours, particularly those lacking expression of other nuclear receptor genes. q 2005 Elsevier Ireland Ltd. All rights reserved. Keywords: Estrogen receptor; Androgen receptor; Nuclear receptor; Expression; Breast cancer
1. Introduction The nuclear receptor (NR) genes are a large family of genes, coding for similar proteins, whose functions * Corresponding author. Tel.: C61 7 5552 8664; fax: C61 7 5552 8908. E-mail address: [email protected] (L.R. Griffiths).
are to accept incoming signals from various messenger molecules [1]. NR genes are typically found at the nuclear membrane or, less frequently, free in the cell cytoplasm [1]. There is a large variety of NR genes, each geared to specific ligands, including vitamin D, retinoic acid and steroid hormones. The steroid nuclear receptors are a sub-family of the NR superfamily, mediating the cellular effects of steroid hormones.
0304-3835/$ - see front matter q 2005 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.canlet.2005.06.013
R.A. Smith et al. / Cancer Letters 237 (2006) 248–255
Steroid receptors have three major structural domains, including a central hinge region and highly conserved DNA and ligand binding domains [2]. Once bound to specific ligands, NR genes form a homo- or hetero-dimer with another liganded receptor and are translocated into the nucleus, where they bind to specialised DNA sequences, termed hormone response elements, on target genes [1]. After binding to hormone response elements the receptor complex undergoes chromatin remodelling in order to increase or decrease the expression of hormone responsive genes [3]. Precise effects of NRs on genes are often tissue specific and are affected by the expression of various co-enzymes and stimulation by other NRs and signalling pathways [4,5]. Activated steroid receptors are known to alter the behaviour of genes involved in cellular metabolism and are of particular relevance for breast cancer [4,6]. The estrogen receptor (ESR) plays a large role in cellular metabolism, especially in the breast. The primary effect of ESR in breast tissue is to act as a mitogenic factor, which it also does in several other tissue types [7]. The receptor has a number of other effects, including maintaining bone density in both men and women [8]. Thus ESR is found in many areas of the body, including the brain, though the precise effects of estrogen in brain are not yet well understood. There are two estrogen receptor genes, ESRa and ESRb and their expression seems to be broadly affected by the same factors, though there are some known differences [2,7]. Both ESR types bind to estrogen, though different specificities have been reported [2,7]. Both ESR isoforms are known to form homo- and hetero-dimers with each other after activation, which may induce altered transcriptional events [2]. Not all of the effects of the estrogen receptor are due to direct genomic binding of the activated receptor. There are a number of proteinprotein interactions that activated ESR will undergo, that also result in altered transcriptional events. These include interactions with the Sp1, Ap1 and NFkB proteins, as well as with the well characterised Ras, Raf, MEK pathway [7]. While many breast cancers respond normally to estrogen, some advanced cancers will cease to express ESR, either because the gene becomes disabled and they already produce the needed growth factors, or else the receptor is mutated into a permanently active state [9]. In these breast cancers the cell does not
249
respond to estrogen stimulation and will often respond poorly to other drugs that rely on the manipulation of estrogen mediated pathways, like tamoxifen. Much data indicates that the continued presence of ESR along with other NRs, particularly the progesterone receptor, in a breast tumour indicates a greater chance for responsiveness to certain treatments and this is due to the interactions between the NR pathways [9]. The androgen receptor (AR) gene is located at Xq11-12 and is some 90 kb long with eight exons. The androgen receptor itself plays a role in a number of tissues, most obviously in the development of male traits in animals [10]. Like the estrogen receptors, the progesterone receptor (PgR) and the glucocorticoid receptor (GR), AR has alternative isoforms that assist in mediation of its function. The two isoforms, ARa and ARb are splice variants produced from two separate promoters on the AR gene and are of 87 and 110 kDa, respectively [11]. The precise activities of the two isoforms are not entirely clear and ARa was at first thought to be inert, though some more recent evidence disputes this [11]. In breast tissue, AR is anti-mitogenic, an effect at least partly due to direct interference with the estrogen receptor (ESR) signalling pathways [11–13]. Due to this anti-proliferative role in breast tissue, androgens have been used to treat breast cancer with some success [11,12,14]. However, androgen stimulation has been known to induce tissue growth in a number of breast cancer cell lines [11,13,15]. The precise reason for this is poorly understood, but it may be caused by aromatisation of androgens into estrogens, mutations of the AR gene itself or alterations in the behaviour of NR co-factors [6,11,15]. Thus, by observing patterns of AR expression in breast tumour tissue, as compared to normal tissue, we may begin to understand how cancer progression affects this significant signalling molecule. When assaying the expression of genes in cancer, it is important to remember that there is often a generalised increase in mRNA production due to heightened cellular metabolism and certain genes often used to gauge alterations in expression, such as the commonly used b-actin housekeeping gene, may also be affected in other ways [16–18]. The gene used for the purpose in this study, the 18S ribosomal gene is necessary for protein production, and is expressed at a relatively stale level among different cell lines,
250
R.A. Smith et al. / Cancer Letters 237 (2006) 248–255
Table 1 Population ESRa immunohistochemical staining status Tumour Grade
ESRa Positive
ESRa Negative
Total
Control Grade 1 Grade 2 Grade 3
5 6 5 2
0 0 2 5
5 6 7 7
including several breast cancer cell lines, thus being a suitable measure for the different tumours in the study [18]. It is also used to help verify the reproducibility of the RT-PCR technique.
2. Materials and methods 2.1. Samples The sample population was comprised of 25 archived breast tissue sections embedded in paraffin and fixed with 10% buffered formalin on slides, with H&E stained slides as a reference. All tumour samples were infiltrating ductal carcinoma. There were 6 samples from tumour grade 1, 7 samples each from grades 2 and 3 and 5 samples of benign breast tissue taken from unaffected patients as the control population. The average age of the individuals from whom the biopsies were obtained were 56.88 years, 59.18 years, 60.45 years and 55.93 years for the control and grades 1, 2 and 3 groups, respectively. The archival breast tissue samples were obtained through collaboration with the Pathology Department of the Gold Coast Hospital, with relevant ethical approvals. For consistency, cancer grade for each sample was determined by a single pathologist from the Gold Coast Hospital’s Pathology Department. Pathological tests also included immunohistochemical staining to detect ESRa proteins and a summary of the population’s ESRa status can be found in Table 1. 2.2. Expression assay RNA was extracted as outlined in our previous publication [19]. Each slide underwent separation of tumour tissue by microdissection, using the H&E stained slides to distinguish tumour from surrounding tissue. Paraffin was then removed with Xylene
washes, the dissected tissue homogenised and then digested with TRIzol reagent. RNA was separated from DNA by adding chloroform and centrifuging. Extracted RNA was then pelleted, Rnasin and Dnase treated to protect the RNA and remove any contaminants. Finally, RNA was further purified using a Qiagen Rneasy Mini Kit. cDNA was then generated using Superscript III from Invitrogen, the reaction primed using random hexamers to maximise cDNA yield from partly degraded RNA. Newly transcribed cDNA underwent PCR to amplify portions of the cDNA that correspond to the mRNA for ESRa, ESRb and AR, with fluorescently tagged primers. Individual samples underwent PCR in triplicate for each gene before results were pooled. Because of the large size of introns in the NR genes, intron spanning primers would be unlikely to amplify in a PCR optimized for small fragments, therefore triple primer sets were designed so that genomic DNA contamination would be indicated by amplification from a second reverse primer, placed a short distance into the intron. However, due to difficulties in obtaining functional triple ESRa and AR primer sets for the detection of DNA contamination, primers from a previous study, corresponding to an intronic/exonic fragment of the glucocorticoid receptor gene (GR) were used for the purpose [19]. No genomic DNA contamination was observed in any sample. Primer details appear in Table 2. The fragment amplified for all receptors was approximately 100 base pairs long, allowing cDNA derived from partially degraded RNA to be fully utilised. Table 2 Primer compositions for PCRs. Primer name a
AREX1-F1 AREX1-R1 GREX2-F1a GRIntron2 ESRaEX1-F1a ESRaEX1-R1 ESRbEX2-F1a ESRbEX2-R1 ESRbIntron2 18S-Ab 18S-B a b
Sequence (5 0 -3 0 ) CCTGATGTGTGGTACCCTGG CCGGAGTAGCTATCCATCCA GAGTACCTCTGGAGGACAGA ATGTCCATTCTTAAGAAACAGGA CCAAAGCATCTGGGATGGCC GGATCTTGAGCTGCGGACGG CCAACACCTGGGCACCTTTC CCAGGGACTCTTTTGAGGTTC TGGCTAGCAACTATAATTCAGAATGAA CTTAGAGGGACAAGTCGCG GGACATCTAAGGGCATCACA
Primer labelled with TET at 5 0 . Primer labelled with HEX at 5 0 .
R.A. Smith et al. / Cancer Letters 237 (2006) 248–255
251
Table 3 Expression data for nuclear receptors by tumour grade Tumour grade
Average ratio (AR/18S)
SD (AR)
Average ratio (ESRa/18S)
SD (ESRa)
Average Ratio (ESRb/18S)
SD (ESRb)
Grade 1 samples Grade 2 samples Grade 3 samples Control samples
0.372213 0.548775 0.530409 0.274586
0.05684 0.1102 0.1012 0.058
0.195848 0.239395 0.366408 0.429228
0.0546 0.0433 0.0621 0.0495
0.223501 0.223477 0.205338 0.223349
0.03424 0.01767 0.01877 0.01865
Fragments produced by any contaminating genomic DNA present was 150 bp in length for samples using the GR primers and 161 for the ESRb intronic primer. To control for PCR efficiency, the nuclear receptors were also multiplexed with the ribosomal 18S gene. Gene expression was quantified using an ABI 310 Genetic Analyser, utilizing peak height as the measure of expression. 2.3. Statistical analysis Expression data obtained for all receptors was normalised by expressing it as a ratio of the expression of the receptor to 18S, which had been multiplexed with each tumour and control sample. As mentioned above, this controls for PCR efficiency by normalising the data. Normalised values for each triplicate sample were then averaged to give the final data used. Normalised final data was analysed using One-way Analysis of Variance (ANOVA) to determine if there was a significant difference of expression between tumour grades. A second ANOVA test was performed to determine whether expression of the nuclear receptors was significantly affected by ESRa status. Homogeneity of variance tests were also performed to discover whether nonparametric tests would be required in addition to the original ANOVA. Appropriate post-hoc tests were subsequently performed to elucidate any differences found. The conventional a-level of 0.05 was specified as the significance threshold. The software package SPSS version 10.1 was employed for all statistical analyses.
3. Results The expression levels for ESRa, ESRb and AR were determined for all archival tissue samples, with
the normalised data summarised in Table 3. Despite initial observations of the data for ESRa indicating a drop in expression for grade 1 tumours (see Fig. 1) ANOVA results determined that this change was not significant (PZ0.057). Data for ESRb would seem to indicate that there is little variance in ESRb expression, regardless of what stage of cancer the tissue is derived from (see Fig. 2). This was borne out by ANOVA testing which returned an insignificant result for ESRb (PZ0.622). A statistical summary for all ANOVA analyses can be found in Table 4. Data obtained for AR (see Fig. 3), showed that the average AR/18S expression ratio was elevated in both grade 2 and grade 3 tumours. One way ANOVA indicated that the differences between tumour grades were significant, with a P-value of 0.014. There was considerable variation among the ratios for some samples and a test for homogeneity of variance was performed to test for a violation of ANOVA assumptions. However, this test proved not to be significant, with the P-value being 0.078.
Fig. 1. ESRa Expression Across Different Cancer Grades: The expression of ratio of ESRa to 18S in the tissue population is graphed here by grade. Despite the drop in ESRa expression in grade 1 tumours, no significant differences were detected (PZ0.057).
252
R.A. Smith et al. / Cancer Letters 237 (2006) 248–255
and AR expression (PZ0.025), with expression increasing in ESRa negative samples. 4. Discussion
Fig. 2. ESRb Expression Across Different Cancer Grades: The expression of ratio of ESRb to 18S in the tissue population is graphed here by grade. There is very little variation in ESRb expression between the grades, the P value being only 0.622.
The estrogen alpha, estrogen beta and androgen receptor genes are members of the steroid receptor sub-family of the nuclear receptors superfamily. This study investigated the expression of these genes in different grades of breast cancer tissue, as well as stromal tissue derived from the same samples. The results obtained from one way ANOVA within
Table 4 Summary of statistical results for nuclear receptors Nuclear receptor
Results for BC grade
Results for ESRa status
Statistic
P-value
Statistic
P-value
ESR Alpha ESR Beta AR
FZ2.628 FZ0.738 FZ3.130
0.057 0.622 0.014
FZ0.071 FZ0.135 FZ5.743
0.794 0.716 0.025
With the existence of a significant difference between groups indicated in the original analysis, post hoc testing was performed to identify the precise effect of cancer grade on AR expression. The results of the post-hoc analysis for the androgen receptor confirmed the initial impression that expression ratios were elevated in grade 2 and grade 3 tumour tissues as compared to control tissue (P!0.05). Post hoc tests indicated that the grade 1 tumour tissue formed an apparent intermediate state between control and advanced tumours, not being significantly different from these extremes. After testing expression in cancer grades, the expression levels for ESRa ESRb and AR were compared against ESRa protein status, a useful prognostic factor in breast cancer, as determined by immunohistochemical staining (see Fig. 4). ANOVA results for ESRa and ESRb indicated that there was no significant relationship between ESRa protein status and the mRNA levels present for either estrogen receptor (PZ0.794 and 0.716 respectively). The ANOVA results for AR showed that there is a significant relationship between ESRa protein status
Fig. 3. AR Expression Across Different Cancer Grades: The expression of ratio of AR to 18S in the tissue population is graphed here by grade. Expression of AR increases in all grades compared to control, but the increase only becomes significant in grades 2 and 3 (PZ0.014).
Fig. 4. NR Expression by ESRa Status: The expression of ratio of all three nuclear receptors to 18S in the tissue population is graphed here, grouped using detection of ESRa protein in the tissue (designated K and C). Expression of all nuclear receptors increases in tissue negative for ESRa protein, but this increase is only significant for AR (PZ0.025).
R.A. Smith et al. / Cancer Letters 237 (2006) 248–255
the tested population indicated that there is significant statistical evidence of a relationship between AR expression and cancer grade, as well as AR expression and ESRa status, within the tested population. Data for the estrogen receptors indicated that their expression is not significantly influenced by cancer grade or ESRa protein status. Post-hoc tests indicated that the difference in AR expression involves the significant up-regulation of AR in grade 2 and grade 3 tumour tissue, in comparison to control tissue and in ESRa negative tissues in comparison to ESRa positive tissues. Data from this study has found that the expression of both estrogen receptors does not change in the tested population, at least at the level of mRNA, for either cancer grade or ESRa status. The initial drop off in ESRa expression seen in grade 1 tumours, though not significant, would be consistent with the self downregulation of ESR in response to high levels of stimulation of relatively normal cellular growth pathways in the early tumour. Though the difference in the data is not significant, it is close to significance and it is possible that the relatively low number of samples in this study has insufficient power to detect the change in expression. By contrast, ESRb showed no obvious trend of any kind, its expression remaining static for all grades. Since the effects of ESRb are normally mediated through its isoforms, even a relatively unchanged expression of ESRb does not rule out significant effects for this gene in the progression of breast cancer [7]. Immunohistochemical data for each sample indicated that there is no detectable ESRa protein in two out of the seven grade 2 samples and five out of the seven grade 3 samples. Many advanced breast tumours cease expressing ESRa and this is a poor prognostic indicator, since it indicates that the tumour may be unresponsive to treatments that manipulate the estrogen pathway, such as tamoxifen [9]. However, none of the individual samples in this study failed to express ESRa mRNA, nor was there any significant difference in mRNA levels between ESRa positive and negative tissues. Previous studies have shown strong correlations between ESRa mRNA and ESRa protein levels, though other studies have found that the levels can vary considerably, even to the point of mRNA being present without protein expression [20,21]. This would indicate that PCR techniques alone cannot be
253
used to determine whether or not tumours will be responsive to hormonally based treatments. Since all tissues in this study maintain expression of ESRa mRNA, tissues that fail to produce the finished receptor may be undergoing some post-transcriptional event which prevents final assembly of the ESR This event may be rapid degradation of transcribed mRNA, the removal of any translated proteins or alternative splicing of ESRa mRNA into proteins undetectable with standard ESRa antibodies and which favour cellular growth. If the specific mechanism which prevents the detected mRNA from being translated into detectable ESRa can be discovered, it may be able to be blocked, causing tumours to re-express the protein and become responsive to hormonal treatments again. However, due to the heterogeneity of cancer tissue, it seems probable that not all tumours failing to express ESRa, but maintaining the mRNA, will be doing so by exactly the same mechanism. Therefore therapies based on this premise might have to be tailored to individual tumours. With no immunohistochemical data for ESRb, the data obtained from this study does not verify the presence of completed ESRb proteins derived from the mRNA that is expressed in the tested population. Since isoforms of ESRb are known to modulate the effects of ESRa, it seems likely that advanced tumours have skewed expression of any ESRb proteins to isoforms that favour cellular growth. In contrast to the levels of the estrogen receptors, which remained unchanged in the studied population, the androgen receptor showed a significant upregulation in late stage tumours, as well as in ESRa negative tissue. This result may be due to the fact that all ESRa negative tissues in this study are either grade 2 or 3 tumours, or it may be a direct link between the loss of ESRa protein and the expression of AR. An additional ANOVA comparing ESRa positive and negative tissue in grades 2 and 3 only, showed no significant differences however (PZ0.336) so it seems likely that the higher AR expression in ESRa negative tissue is not a direct result of the loss of ESRa expression. It is known that many breast tumours continue to express AR even when they fail to express ESR or PgR, so these tumours remain sensitive to androgen stimulation even after other endocrine therapies become useless [14]. However, the elevated levels of AR in late-stage breast tumours would indicate that
254
R.A. Smith et al. / Cancer Letters 237 (2006) 248–255
these tumours would be more sensitive to androgen stimulated growth arrest than early stage tumours. This would seem to be at odds with the high proliferation rate of advanced tumours, but the increase in expression may be due to the influence of the estrogen receptor pathways, which are known to have effects on the behaviour of AR in the breast [14]. Such an increase in AR expression may be an aspect of normal growth arrest mechanisms, which remain active within the tumour cells. Previous results from our laboratory indicate that a similar pattern of expression also occurs within the same population for the progesterone and glucocorticoid receptors, both anti-mitogenic in the breast [19]. Circulating levels of androgens in female breast cancer patients are likely to be fairly low, considering losses for aromatase activity in adipose and tumour tissue, so the effects of the increase in AR expression may not be particularly marked, compared to the in vivo effects derived from GR and PgR [6]. At the same time, this lack of effect may be precisely the reason that AR expression is known to continue in tumour cells, as the cells are under little selective pressure to cease expression or become AR resistant. This means that AR may remain a viable target for hormone treatment strategies even in late stage tumours. However, it is known that advanced tumours often become insensitive or completely independent of hormone stimulation or manipulation of steroid receptor pathways. It is possible that the increased androgen receptor expression observed in this study may not result in increased sensitivity to androgens due to disruption of normal AR pathway function. Such disruption could come from a number of sources. As for the estrogen receptors, it is possible that the AR mRNA observed in this study is not being translated into a final protein, preventing any androgenic stimulation of the cell. Alternatively, the mRNA may be undergoing post-transcriptional splicing to inactivate the resultant protein or to produce an alternate isoform of AR that is ineffective or antagonises the primary form. Restriction of AR effectiveness might also be achieved through regulation of the various NR co-activator and co-repressor proteins, restricting or increasing availability as needed to prevent AR stimulation. Finally, it is also possible that the increased AR expression observed in this study is a mechanism
for retarding the effectiveness of other antimitogenic nuclear receptors. The androgen receptor’s normal functions do not include antagonism of PgR or GR, but AR has been known to interfere with other NRs through competition for common chaperone proteins and co-factors [13]. This competition, combined with low levels of in vivo androgen stimulation, will reduce the overall effect of other anti-mitogenic hormones in the breast tumour, though likely not abrogate them entirely. Further studies of relevance to these results should concentrate on the expression of co-activator and co-repressor molecules in the same tissue as well as utilising additional immunohistochemistry to gain a more complete picture of what is occurring in these tumour cells.
5. Conclusion The results of this study indicate that the expression of the estrogen receptor alpha and beta genes remain relatively stable in breast tumours, while the expression of the androgen receptor gene is significantly related to cancer grade, showing an increase in later stage (grades 2 and 3) breast tumours. The results of this study also indicate that breast tumours that no longer produce ESRa proteins maintain the production of statistically normal amounts of ESRa mRNA, but that AR mRNA levels increase in ESRa negative tissue. The precise implications of this information for breast cancer treatment are not fully clear, due to the inability of this study to scrutinise post transcriptional activity in the tumours, but it is possible that the enhanced expression of AR in late stage breast tumours and the continuation of ESRa mRNA expression in tissue that lacks the ESRa protein may provide an additional angle of treatment for these advanced tumours, especially those lacking functional expression of other NRs.
Acknowledgements This research was kindly supported by the Wesley Research Institute, Brisbane, Qld., Australia.
R.A. Smith et al. / Cancer Letters 237 (2006) 248–255
References [1] C.J. Fryer, H.K. Kinyamu, I. Rogatsky, M.J. Garabedian, T.K. Archer, Selective activation of the glucocorticoid receptor by steroid antagonists in human breast cancer and osteosarcoma cells, The Journal of Biological Chemistry 275 (2000) 17771–17777. [2] M.-C. Keightley, Steroid receptor isoforms: exception or rule?, Molecular and Cellular Endocrinology 137 (1998) 1–5. [3] I.J. McEwan, Gene regulation through chromatin remodelling by members of the nuclear receptor superfamily, Biochemical Society Transactions 28 (2000) 369–373. [4] R.B. Dickson, M.E. Lippman, Control of human breast cancer by estrogen, growth factors, and oncogenes, Cancer Treatment and Research 40 (1988) 119–165. [5] G. Verrijdt, A. Haelens, F. Claessens, Selective DNA recognition by the androgen receptor as a mechanism for hormone-specific regulation of gene expression, Molecular Genetics and Metabolism 78 (2003) 175–185. [6] W.R. Miller, Biological rationale for endocrine therapy in breast cancer, Best Practice & Research Clinical Endocrinology & Metabolism 18 (2004) 1–32. [7] S. Nilsson, S. Makela, F. Trueter, M. Tujague, J. Thomsen, G. Andersson, et al., Mechanisms of estrogen action, Physiological Reviews 81 (2001) 1535–1565. [8] K.A.J. Grandien, A. Berkenstam, J.-A. Gustafsson, The estrogen receptor gene: promoter organisation and expression, The International Journal of Biochemistry & Cell Biology 29 (1997) 1343–1369. [9] K.-Y. Yoo, K. Tajima, S.-K. Park, D. Kang, S.-U. Kim, K. Hirose, et al., Postmenopausal obesity as a breast cancer risk factor according to estrogen and progesterone receptor status (Japan), Cancer Letters 167 (2001) 57–63. [10] J. Gobinet, N. Poujol, C. Sultan, Molecular action of androgens, Molecular and Cellular Endocrinology 198 (2002) 15–24. [11] P. Ferro, M.G. Catalano, R. Dell’Eva, N. Fortunati, U. Pfeffer, The androgen receptor CAG repeat: a modifier of carcinogenesis?, Molecular and Cellular Endocrinology 193 (2002) 109–120. [12] V. Kuenen-Boumeester, T.H. Van der Kwast, C.C. Claassen, M.P. Look, G.S. Liem, J.G.M. Klijn, S.C. Henzen-Logmans, The clinical significance of androgen receptors in breast
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
255
cancer and their relation to histological and cell biological parameters, The European Journal of Cancer 32A (1996) 1560–1565. S. Ando`, F. De Amicis, V. Rago, A. Carpino, M. Maggiolini, M.L. Panno, M. Lanzino, Breast cancer: from estrogen to androgen receptor, Molecular and Cellular Endocrinology 193 (2002) 121–128. J.M. Bentel, S.N. Birrell, M.A. Pickering, D.J. Holds, D.J. Horsfall, W.D. Tilley, Androgen receptor agonist activity of the synthetic progestin, medroxyprogesterone acetate, in human breast cancer cells, Molecular and Cellular Endocrinology 154 (1999) 11–20. S.N. Birrell, J.M. Bentel, T.E. Hickey, C. Ricciardelli, M.A. Weger, D.J. Horsfall, W.D. Tilley, Androgens induce divergent proliferative responses in human breast cancer cell lines, The Journal of Steroid Biochemistry and Molecular Biology 52 (1995) 459–467. S. Selvey, E.W. Thompson, K. Matthaei, R.A. Lea, M.G. Irving, L.R. Griffiths, b-actin—an unsuitable internal control for RT-PCR, Molecular and Cellular Probes 15 (2001) 307–311. J.L. Aerts, M.I. Gonzales, S.L. Topalian, Selection of appropriate control genes to assess expression of tumor antigens using real-time RT-PCR, Biotechniques 36 (2004) 84–86 see also 88, 90–1. D.L. Morse, D. Carroll, L. Weberg, M.C. Borgstrom, J. Ranger-Moore, R.J. Gillies, Determining suitable internal standards for mRNA quantification of increasing cancer progression in human breast cells by real-time reverse transcriptase polymerase chain reaction, Analytical Biochemistry, in press. R.A. Smith, R.A. Lea, J.E. Curran, S.R. Weinstein, L.R. Griffiths, Expression of glucocorticoid and progesterone nuclear receptor genes in archival breast cancer tissue, Breast Cancer Research 5 (2003) R9–R12. P.A. O’Neill, M.P. Davies, A.M. Shaaban, H. Innes, A. Torevell, D.R. Sibson, C.S. Foster, Wild-type oestrogen receptor beta (ERbeta1) mRNA and protein expression in Tamoxifen-treated post-menopausal breast cancers, British Journal of Cancer 9 (2004) 1694–1702. M.J. Fasco, K. Keyomarsi, K.F. Arcaro, J.F. Gierthy, Expression of an estrogen receptor alpha variant protein in cell lines and tumors, Molecular and Cellular Endocrinology 162 (2000) 167–180.
Detection of mRNA levels for the estrogen alpha, estrogen beta and androgen nuclear receptor genes in archival breast cancer tissue Robert A. Smitha, Rod A. Leaa, Stephen R. Weinsteinb, Lyn R. Griffithsa,* a
Genomics Research Centre and Wesley Research Institute, School of Health Science, Griffith University Gold Coast, PMB 50 Gold Coast Mail Centre, Qld. 9726, Australia b Department of Pathology, Gold Coast Hospital, Southport, Qld., Australia Received 2 March 2005; received in revised form 3 June 2005; accepted 6 June 2005
Abstract Previous studies in our laboratory have shown association of nuclear receptor expression and histological breast cancer grade. To further investigate these findings, it was the objective of this study to determine if expression levels of the estrogen alpha, estrogen beta and androgen nuclear receptor genes varied in different breast cancer grades. RNA extracted from paraffin embedded archival breast tumour tissue was converted into cDNA and cDNA underwent PCR to enable quantitation of mRNA expression. Expression data was normalised against the 18S ribosomal gene multiplex and analysed using ANOVA. Analysis indicated a significant alteration of expression for the androgen receptor in different cancer grades (PZ0.014), as well as in tissues that no longer possess estrogen receptor alpha proteins (PZ0.025). However, expression of estrogen receptors alpha and beta did not vary significantly with cancer grade (PZ0.057 and 0.622, respectively). Also, the expression of estrogen receptor alpha or beta did not change, regardless of the presence of estrogen receptor alpha protein in the tissue (PZ0.794 and 0.716, respectively). Post-hoc tests indicate that the expression of the androgen receptor is increased in estrogen receptor negative tissue as well as in grade 2 and grade 3 tumours, compared to control tissue. This increased expression in late stage breast tumours may have implications to the treatment of breast tumours, particularly those lacking expression of other nuclear receptor genes. q 2005 Elsevier Ireland Ltd. All rights reserved. Keywords: Estrogen receptor; Androgen receptor; Nuclear receptor; Expression; Breast cancer
1. Introduction The nuclear receptor (NR) genes are a large family of genes, coding for similar proteins, whose functions * Corresponding author. Tel.: C61 7 5552 8664; fax: C61 7 5552 8908. E-mail address: [email protected] (L.R. Griffiths).
are to accept incoming signals from various messenger molecules [1]. NR genes are typically found at the nuclear membrane or, less frequently, free in the cell cytoplasm [1]. There is a large variety of NR genes, each geared to specific ligands, including vitamin D, retinoic acid and steroid hormones. The steroid nuclear receptors are a sub-family of the NR superfamily, mediating the cellular effects of steroid hormones.
0304-3835/$ - see front matter q 2005 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.canlet.2005.06.013
R.A. Smith et al. / Cancer Letters 237 (2006) 248–255
Steroid receptors have three major structural domains, including a central hinge region and highly conserved DNA and ligand binding domains [2]. Once bound to specific ligands, NR genes form a homo- or hetero-dimer with another liganded receptor and are translocated into the nucleus, where they bind to specialised DNA sequences, termed hormone response elements, on target genes [1]. After binding to hormone response elements the receptor complex undergoes chromatin remodelling in order to increase or decrease the expression of hormone responsive genes [3]. Precise effects of NRs on genes are often tissue specific and are affected by the expression of various co-enzymes and stimulation by other NRs and signalling pathways [4,5]. Activated steroid receptors are known to alter the behaviour of genes involved in cellular metabolism and are of particular relevance for breast cancer [4,6]. The estrogen receptor (ESR) plays a large role in cellular metabolism, especially in the breast. The primary effect of ESR in breast tissue is to act as a mitogenic factor, which it also does in several other tissue types [7]. The receptor has a number of other effects, including maintaining bone density in both men and women [8]. Thus ESR is found in many areas of the body, including the brain, though the precise effects of estrogen in brain are not yet well understood. There are two estrogen receptor genes, ESRa and ESRb and their expression seems to be broadly affected by the same factors, though there are some known differences [2,7]. Both ESR types bind to estrogen, though different specificities have been reported [2,7]. Both ESR isoforms are known to form homo- and hetero-dimers with each other after activation, which may induce altered transcriptional events [2]. Not all of the effects of the estrogen receptor are due to direct genomic binding of the activated receptor. There are a number of proteinprotein interactions that activated ESR will undergo, that also result in altered transcriptional events. These include interactions with the Sp1, Ap1 and NFkB proteins, as well as with the well characterised Ras, Raf, MEK pathway [7]. While many breast cancers respond normally to estrogen, some advanced cancers will cease to express ESR, either because the gene becomes disabled and they already produce the needed growth factors, or else the receptor is mutated into a permanently active state [9]. In these breast cancers the cell does not
249
respond to estrogen stimulation and will often respond poorly to other drugs that rely on the manipulation of estrogen mediated pathways, like tamoxifen. Much data indicates that the continued presence of ESR along with other NRs, particularly the progesterone receptor, in a breast tumour indicates a greater chance for responsiveness to certain treatments and this is due to the interactions between the NR pathways [9]. The androgen receptor (AR) gene is located at Xq11-12 and is some 90 kb long with eight exons. The androgen receptor itself plays a role in a number of tissues, most obviously in the development of male traits in animals [10]. Like the estrogen receptors, the progesterone receptor (PgR) and the glucocorticoid receptor (GR), AR has alternative isoforms that assist in mediation of its function. The two isoforms, ARa and ARb are splice variants produced from two separate promoters on the AR gene and are of 87 and 110 kDa, respectively [11]. The precise activities of the two isoforms are not entirely clear and ARa was at first thought to be inert, though some more recent evidence disputes this [11]. In breast tissue, AR is anti-mitogenic, an effect at least partly due to direct interference with the estrogen receptor (ESR) signalling pathways [11–13]. Due to this anti-proliferative role in breast tissue, androgens have been used to treat breast cancer with some success [11,12,14]. However, androgen stimulation has been known to induce tissue growth in a number of breast cancer cell lines [11,13,15]. The precise reason for this is poorly understood, but it may be caused by aromatisation of androgens into estrogens, mutations of the AR gene itself or alterations in the behaviour of NR co-factors [6,11,15]. Thus, by observing patterns of AR expression in breast tumour tissue, as compared to normal tissue, we may begin to understand how cancer progression affects this significant signalling molecule. When assaying the expression of genes in cancer, it is important to remember that there is often a generalised increase in mRNA production due to heightened cellular metabolism and certain genes often used to gauge alterations in expression, such as the commonly used b-actin housekeeping gene, may also be affected in other ways [16–18]. The gene used for the purpose in this study, the 18S ribosomal gene is necessary for protein production, and is expressed at a relatively stale level among different cell lines,
250
R.A. Smith et al. / Cancer Letters 237 (2006) 248–255
Table 1 Population ESRa immunohistochemical staining status Tumour Grade
ESRa Positive
ESRa Negative
Total
Control Grade 1 Grade 2 Grade 3
5 6 5 2
0 0 2 5
5 6 7 7
including several breast cancer cell lines, thus being a suitable measure for the different tumours in the study [18]. It is also used to help verify the reproducibility of the RT-PCR technique.
2. Materials and methods 2.1. Samples The sample population was comprised of 25 archived breast tissue sections embedded in paraffin and fixed with 10% buffered formalin on slides, with H&E stained slides as a reference. All tumour samples were infiltrating ductal carcinoma. There were 6 samples from tumour grade 1, 7 samples each from grades 2 and 3 and 5 samples of benign breast tissue taken from unaffected patients as the control population. The average age of the individuals from whom the biopsies were obtained were 56.88 years, 59.18 years, 60.45 years and 55.93 years for the control and grades 1, 2 and 3 groups, respectively. The archival breast tissue samples were obtained through collaboration with the Pathology Department of the Gold Coast Hospital, with relevant ethical approvals. For consistency, cancer grade for each sample was determined by a single pathologist from the Gold Coast Hospital’s Pathology Department. Pathological tests also included immunohistochemical staining to detect ESRa proteins and a summary of the population’s ESRa status can be found in Table 1. 2.2. Expression assay RNA was extracted as outlined in our previous publication [19]. Each slide underwent separation of tumour tissue by microdissection, using the H&E stained slides to distinguish tumour from surrounding tissue. Paraffin was then removed with Xylene
washes, the dissected tissue homogenised and then digested with TRIzol reagent. RNA was separated from DNA by adding chloroform and centrifuging. Extracted RNA was then pelleted, Rnasin and Dnase treated to protect the RNA and remove any contaminants. Finally, RNA was further purified using a Qiagen Rneasy Mini Kit. cDNA was then generated using Superscript III from Invitrogen, the reaction primed using random hexamers to maximise cDNA yield from partly degraded RNA. Newly transcribed cDNA underwent PCR to amplify portions of the cDNA that correspond to the mRNA for ESRa, ESRb and AR, with fluorescently tagged primers. Individual samples underwent PCR in triplicate for each gene before results were pooled. Because of the large size of introns in the NR genes, intron spanning primers would be unlikely to amplify in a PCR optimized for small fragments, therefore triple primer sets were designed so that genomic DNA contamination would be indicated by amplification from a second reverse primer, placed a short distance into the intron. However, due to difficulties in obtaining functional triple ESRa and AR primer sets for the detection of DNA contamination, primers from a previous study, corresponding to an intronic/exonic fragment of the glucocorticoid receptor gene (GR) were used for the purpose [19]. No genomic DNA contamination was observed in any sample. Primer details appear in Table 2. The fragment amplified for all receptors was approximately 100 base pairs long, allowing cDNA derived from partially degraded RNA to be fully utilised. Table 2 Primer compositions for PCRs. Primer name a
AREX1-F1 AREX1-R1 GREX2-F1a GRIntron2 ESRaEX1-F1a ESRaEX1-R1 ESRbEX2-F1a ESRbEX2-R1 ESRbIntron2 18S-Ab 18S-B a b
Sequence (5 0 -3 0 ) CCTGATGTGTGGTACCCTGG CCGGAGTAGCTATCCATCCA GAGTACCTCTGGAGGACAGA ATGTCCATTCTTAAGAAACAGGA CCAAAGCATCTGGGATGGCC GGATCTTGAGCTGCGGACGG CCAACACCTGGGCACCTTTC CCAGGGACTCTTTTGAGGTTC TGGCTAGCAACTATAATTCAGAATGAA CTTAGAGGGACAAGTCGCG GGACATCTAAGGGCATCACA
Primer labelled with TET at 5 0 . Primer labelled with HEX at 5 0 .
R.A. Smith et al. / Cancer Letters 237 (2006) 248–255
251
Table 3 Expression data for nuclear receptors by tumour grade Tumour grade
Average ratio (AR/18S)
SD (AR)
Average ratio (ESRa/18S)
SD (ESRa)
Average Ratio (ESRb/18S)
SD (ESRb)
Grade 1 samples Grade 2 samples Grade 3 samples Control samples
0.372213 0.548775 0.530409 0.274586
0.05684 0.1102 0.1012 0.058
0.195848 0.239395 0.366408 0.429228
0.0546 0.0433 0.0621 0.0495
0.223501 0.223477 0.205338 0.223349
0.03424 0.01767 0.01877 0.01865
Fragments produced by any contaminating genomic DNA present was 150 bp in length for samples using the GR primers and 161 for the ESRb intronic primer. To control for PCR efficiency, the nuclear receptors were also multiplexed with the ribosomal 18S gene. Gene expression was quantified using an ABI 310 Genetic Analyser, utilizing peak height as the measure of expression. 2.3. Statistical analysis Expression data obtained for all receptors was normalised by expressing it as a ratio of the expression of the receptor to 18S, which had been multiplexed with each tumour and control sample. As mentioned above, this controls for PCR efficiency by normalising the data. Normalised values for each triplicate sample were then averaged to give the final data used. Normalised final data was analysed using One-way Analysis of Variance (ANOVA) to determine if there was a significant difference of expression between tumour grades. A second ANOVA test was performed to determine whether expression of the nuclear receptors was significantly affected by ESRa status. Homogeneity of variance tests were also performed to discover whether nonparametric tests would be required in addition to the original ANOVA. Appropriate post-hoc tests were subsequently performed to elucidate any differences found. The conventional a-level of 0.05 was specified as the significance threshold. The software package SPSS version 10.1 was employed for all statistical analyses.
3. Results The expression levels for ESRa, ESRb and AR were determined for all archival tissue samples, with
the normalised data summarised in Table 3. Despite initial observations of the data for ESRa indicating a drop in expression for grade 1 tumours (see Fig. 1) ANOVA results determined that this change was not significant (PZ0.057). Data for ESRb would seem to indicate that there is little variance in ESRb expression, regardless of what stage of cancer the tissue is derived from (see Fig. 2). This was borne out by ANOVA testing which returned an insignificant result for ESRb (PZ0.622). A statistical summary for all ANOVA analyses can be found in Table 4. Data obtained for AR (see Fig. 3), showed that the average AR/18S expression ratio was elevated in both grade 2 and grade 3 tumours. One way ANOVA indicated that the differences between tumour grades were significant, with a P-value of 0.014. There was considerable variation among the ratios for some samples and a test for homogeneity of variance was performed to test for a violation of ANOVA assumptions. However, this test proved not to be significant, with the P-value being 0.078.
Fig. 1. ESRa Expression Across Different Cancer Grades: The expression of ratio of ESRa to 18S in the tissue population is graphed here by grade. Despite the drop in ESRa expression in grade 1 tumours, no significant differences were detected (PZ0.057).
252
R.A. Smith et al. / Cancer Letters 237 (2006) 248–255
and AR expression (PZ0.025), with expression increasing in ESRa negative samples. 4. Discussion
Fig. 2. ESRb Expression Across Different Cancer Grades: The expression of ratio of ESRb to 18S in the tissue population is graphed here by grade. There is very little variation in ESRb expression between the grades, the P value being only 0.622.
The estrogen alpha, estrogen beta and androgen receptor genes are members of the steroid receptor sub-family of the nuclear receptors superfamily. This study investigated the expression of these genes in different grades of breast cancer tissue, as well as stromal tissue derived from the same samples. The results obtained from one way ANOVA within
Table 4 Summary of statistical results for nuclear receptors Nuclear receptor
Results for BC grade
Results for ESRa status
Statistic
P-value
Statistic
P-value
ESR Alpha ESR Beta AR
FZ2.628 FZ0.738 FZ3.130
0.057 0.622 0.014
FZ0.071 FZ0.135 FZ5.743
0.794 0.716 0.025
With the existence of a significant difference between groups indicated in the original analysis, post hoc testing was performed to identify the precise effect of cancer grade on AR expression. The results of the post-hoc analysis for the androgen receptor confirmed the initial impression that expression ratios were elevated in grade 2 and grade 3 tumour tissues as compared to control tissue (P!0.05). Post hoc tests indicated that the grade 1 tumour tissue formed an apparent intermediate state between control and advanced tumours, not being significantly different from these extremes. After testing expression in cancer grades, the expression levels for ESRa ESRb and AR were compared against ESRa protein status, a useful prognostic factor in breast cancer, as determined by immunohistochemical staining (see Fig. 4). ANOVA results for ESRa and ESRb indicated that there was no significant relationship between ESRa protein status and the mRNA levels present for either estrogen receptor (PZ0.794 and 0.716 respectively). The ANOVA results for AR showed that there is a significant relationship between ESRa protein status
Fig. 3. AR Expression Across Different Cancer Grades: The expression of ratio of AR to 18S in the tissue population is graphed here by grade. Expression of AR increases in all grades compared to control, but the increase only becomes significant in grades 2 and 3 (PZ0.014).
Fig. 4. NR Expression by ESRa Status: The expression of ratio of all three nuclear receptors to 18S in the tissue population is graphed here, grouped using detection of ESRa protein in the tissue (designated K and C). Expression of all nuclear receptors increases in tissue negative for ESRa protein, but this increase is only significant for AR (PZ0.025).
R.A. Smith et al. / Cancer Letters 237 (2006) 248–255
the tested population indicated that there is significant statistical evidence of a relationship between AR expression and cancer grade, as well as AR expression and ESRa status, within the tested population. Data for the estrogen receptors indicated that their expression is not significantly influenced by cancer grade or ESRa protein status. Post-hoc tests indicated that the difference in AR expression involves the significant up-regulation of AR in grade 2 and grade 3 tumour tissue, in comparison to control tissue and in ESRa negative tissues in comparison to ESRa positive tissues. Data from this study has found that the expression of both estrogen receptors does not change in the tested population, at least at the level of mRNA, for either cancer grade or ESRa status. The initial drop off in ESRa expression seen in grade 1 tumours, though not significant, would be consistent with the self downregulation of ESR in response to high levels of stimulation of relatively normal cellular growth pathways in the early tumour. Though the difference in the data is not significant, it is close to significance and it is possible that the relatively low number of samples in this study has insufficient power to detect the change in expression. By contrast, ESRb showed no obvious trend of any kind, its expression remaining static for all grades. Since the effects of ESRb are normally mediated through its isoforms, even a relatively unchanged expression of ESRb does not rule out significant effects for this gene in the progression of breast cancer [7]. Immunohistochemical data for each sample indicated that there is no detectable ESRa protein in two out of the seven grade 2 samples and five out of the seven grade 3 samples. Many advanced breast tumours cease expressing ESRa and this is a poor prognostic indicator, since it indicates that the tumour may be unresponsive to treatments that manipulate the estrogen pathway, such as tamoxifen [9]. However, none of the individual samples in this study failed to express ESRa mRNA, nor was there any significant difference in mRNA levels between ESRa positive and negative tissues. Previous studies have shown strong correlations between ESRa mRNA and ESRa protein levels, though other studies have found that the levels can vary considerably, even to the point of mRNA being present without protein expression [20,21]. This would indicate that PCR techniques alone cannot be
253
used to determine whether or not tumours will be responsive to hormonally based treatments. Since all tissues in this study maintain expression of ESRa mRNA, tissues that fail to produce the finished receptor may be undergoing some post-transcriptional event which prevents final assembly of the ESR This event may be rapid degradation of transcribed mRNA, the removal of any translated proteins or alternative splicing of ESRa mRNA into proteins undetectable with standard ESRa antibodies and which favour cellular growth. If the specific mechanism which prevents the detected mRNA from being translated into detectable ESRa can be discovered, it may be able to be blocked, causing tumours to re-express the protein and become responsive to hormonal treatments again. However, due to the heterogeneity of cancer tissue, it seems probable that not all tumours failing to express ESRa, but maintaining the mRNA, will be doing so by exactly the same mechanism. Therefore therapies based on this premise might have to be tailored to individual tumours. With no immunohistochemical data for ESRb, the data obtained from this study does not verify the presence of completed ESRb proteins derived from the mRNA that is expressed in the tested population. Since isoforms of ESRb are known to modulate the effects of ESRa, it seems likely that advanced tumours have skewed expression of any ESRb proteins to isoforms that favour cellular growth. In contrast to the levels of the estrogen receptors, which remained unchanged in the studied population, the androgen receptor showed a significant upregulation in late stage tumours, as well as in ESRa negative tissue. This result may be due to the fact that all ESRa negative tissues in this study are either grade 2 or 3 tumours, or it may be a direct link between the loss of ESRa protein and the expression of AR. An additional ANOVA comparing ESRa positive and negative tissue in grades 2 and 3 only, showed no significant differences however (PZ0.336) so it seems likely that the higher AR expression in ESRa negative tissue is not a direct result of the loss of ESRa expression. It is known that many breast tumours continue to express AR even when they fail to express ESR or PgR, so these tumours remain sensitive to androgen stimulation even after other endocrine therapies become useless [14]. However, the elevated levels of AR in late-stage breast tumours would indicate that
254
R.A. Smith et al. / Cancer Letters 237 (2006) 248–255
these tumours would be more sensitive to androgen stimulated growth arrest than early stage tumours. This would seem to be at odds with the high proliferation rate of advanced tumours, but the increase in expression may be due to the influence of the estrogen receptor pathways, which are known to have effects on the behaviour of AR in the breast [14]. Such an increase in AR expression may be an aspect of normal growth arrest mechanisms, which remain active within the tumour cells. Previous results from our laboratory indicate that a similar pattern of expression also occurs within the same population for the progesterone and glucocorticoid receptors, both anti-mitogenic in the breast [19]. Circulating levels of androgens in female breast cancer patients are likely to be fairly low, considering losses for aromatase activity in adipose and tumour tissue, so the effects of the increase in AR expression may not be particularly marked, compared to the in vivo effects derived from GR and PgR [6]. At the same time, this lack of effect may be precisely the reason that AR expression is known to continue in tumour cells, as the cells are under little selective pressure to cease expression or become AR resistant. This means that AR may remain a viable target for hormone treatment strategies even in late stage tumours. However, it is known that advanced tumours often become insensitive or completely independent of hormone stimulation or manipulation of steroid receptor pathways. It is possible that the increased androgen receptor expression observed in this study may not result in increased sensitivity to androgens due to disruption of normal AR pathway function. Such disruption could come from a number of sources. As for the estrogen receptors, it is possible that the AR mRNA observed in this study is not being translated into a final protein, preventing any androgenic stimulation of the cell. Alternatively, the mRNA may be undergoing post-transcriptional splicing to inactivate the resultant protein or to produce an alternate isoform of AR that is ineffective or antagonises the primary form. Restriction of AR effectiveness might also be achieved through regulation of the various NR co-activator and co-repressor proteins, restricting or increasing availability as needed to prevent AR stimulation. Finally, it is also possible that the increased AR expression observed in this study is a mechanism
for retarding the effectiveness of other antimitogenic nuclear receptors. The androgen receptor’s normal functions do not include antagonism of PgR or GR, but AR has been known to interfere with other NRs through competition for common chaperone proteins and co-factors [13]. This competition, combined with low levels of in vivo androgen stimulation, will reduce the overall effect of other anti-mitogenic hormones in the breast tumour, though likely not abrogate them entirely. Further studies of relevance to these results should concentrate on the expression of co-activator and co-repressor molecules in the same tissue as well as utilising additional immunohistochemistry to gain a more complete picture of what is occurring in these tumour cells.
5. Conclusion The results of this study indicate that the expression of the estrogen receptor alpha and beta genes remain relatively stable in breast tumours, while the expression of the androgen receptor gene is significantly related to cancer grade, showing an increase in later stage (grades 2 and 3) breast tumours. The results of this study also indicate that breast tumours that no longer produce ESRa proteins maintain the production of statistically normal amounts of ESRa mRNA, but that AR mRNA levels increase in ESRa negative tissue. The precise implications of this information for breast cancer treatment are not fully clear, due to the inability of this study to scrutinise post transcriptional activity in the tumours, but it is possible that the enhanced expression of AR in late stage breast tumours and the continuation of ESRa mRNA expression in tissue that lacks the ESRa protein may provide an additional angle of treatment for these advanced tumours, especially those lacking functional expression of other NRs.
Acknowledgements This research was kindly supported by the Wesley Research Institute, Brisbane, Qld., Australia.
R.A. Smith et al. / Cancer Letters 237 (2006) 248–255
References [1] C.J. Fryer, H.K. Kinyamu, I. Rogatsky, M.J. Garabedian, T.K. Archer, Selective activation of the glucocorticoid receptor by steroid antagonists in human breast cancer and osteosarcoma cells, The Journal of Biological Chemistry 275 (2000) 17771–17777. [2] M.-C. Keightley, Steroid receptor isoforms: exception or rule?, Molecular and Cellular Endocrinology 137 (1998) 1–5. [3] I.J. McEwan, Gene regulation through chromatin remodelling by members of the nuclear receptor superfamily, Biochemical Society Transactions 28 (2000) 369–373. [4] R.B. Dickson, M.E. Lippman, Control of human breast cancer by estrogen, growth factors, and oncogenes, Cancer Treatment and Research 40 (1988) 119–165. [5] G. Verrijdt, A. Haelens, F. Claessens, Selective DNA recognition by the androgen receptor as a mechanism for hormone-specific regulation of gene expression, Molecular Genetics and Metabolism 78 (2003) 175–185. [6] W.R. Miller, Biological rationale for endocrine therapy in breast cancer, Best Practice & Research Clinical Endocrinology & Metabolism 18 (2004) 1–32. [7] S. Nilsson, S. Makela, F. Trueter, M. Tujague, J. Thomsen, G. Andersson, et al., Mechanisms of estrogen action, Physiological Reviews 81 (2001) 1535–1565. [8] K.A.J. Grandien, A. Berkenstam, J.-A. Gustafsson, The estrogen receptor gene: promoter organisation and expression, The International Journal of Biochemistry & Cell Biology 29 (1997) 1343–1369. [9] K.-Y. Yoo, K. Tajima, S.-K. Park, D. Kang, S.-U. Kim, K. Hirose, et al., Postmenopausal obesity as a breast cancer risk factor according to estrogen and progesterone receptor status (Japan), Cancer Letters 167 (2001) 57–63. [10] J. Gobinet, N. Poujol, C. Sultan, Molecular action of androgens, Molecular and Cellular Endocrinology 198 (2002) 15–24. [11] P. Ferro, M.G. Catalano, R. Dell’Eva, N. Fortunati, U. Pfeffer, The androgen receptor CAG repeat: a modifier of carcinogenesis?, Molecular and Cellular Endocrinology 193 (2002) 109–120. [12] V. Kuenen-Boumeester, T.H. Van der Kwast, C.C. Claassen, M.P. Look, G.S. Liem, J.G.M. Klijn, S.C. Henzen-Logmans, The clinical significance of androgen receptors in breast
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
255
cancer and their relation to histological and cell biological parameters, The European Journal of Cancer 32A (1996) 1560–1565. S. Ando`, F. De Amicis, V. Rago, A. Carpino, M. Maggiolini, M.L. Panno, M. Lanzino, Breast cancer: from estrogen to androgen receptor, Molecular and Cellular Endocrinology 193 (2002) 121–128. J.M. Bentel, S.N. Birrell, M.A. Pickering, D.J. Holds, D.J. Horsfall, W.D. Tilley, Androgen receptor agonist activity of the synthetic progestin, medroxyprogesterone acetate, in human breast cancer cells, Molecular and Cellular Endocrinology 154 (1999) 11–20. S.N. Birrell, J.M. Bentel, T.E. Hickey, C. Ricciardelli, M.A. Weger, D.J. Horsfall, W.D. Tilley, Androgens induce divergent proliferative responses in human breast cancer cell lines, The Journal of Steroid Biochemistry and Molecular Biology 52 (1995) 459–467. S. Selvey, E.W. Thompson, K. Matthaei, R.A. Lea, M.G. Irving, L.R. Griffiths, b-actin—an unsuitable internal control for RT-PCR, Molecular and Cellular Probes 15 (2001) 307–311. J.L. Aerts, M.I. Gonzales, S.L. Topalian, Selection of appropriate control genes to assess expression of tumor antigens using real-time RT-PCR, Biotechniques 36 (2004) 84–86 see also 88, 90–1. D.L. Morse, D. Carroll, L. Weberg, M.C. Borgstrom, J. Ranger-Moore, R.J. Gillies, Determining suitable internal standards for mRNA quantification of increasing cancer progression in human breast cells by real-time reverse transcriptase polymerase chain reaction, Analytical Biochemistry, in press. R.A. Smith, R.A. Lea, J.E. Curran, S.R. Weinstein, L.R. Griffiths, Expression of glucocorticoid and progesterone nuclear receptor genes in archival breast cancer tissue, Breast Cancer Research 5 (2003) R9–R12. P.A. O’Neill, M.P. Davies, A.M. Shaaban, H. Innes, A. Torevell, D.R. Sibson, C.S. Foster, Wild-type oestrogen receptor beta (ERbeta1) mRNA and protein expression in Tamoxifen-treated post-menopausal breast cancers, British Journal of Cancer 9 (2004) 1694–1702. M.J. Fasco, K. Keyomarsi, K.F. Arcaro, J.F. Gierthy, Expression of an estrogen receptor alpha variant protein in cell lines and tumors, Molecular and Cellular Endocrinology 162 (2000) 167–180.