Relative Expression of Progesterone Receptors A and B in ... · rupting continued estrogenic stimulation of the endometrium, and pregnancy is protective, because progesterone is the

[CANCER RESEARCH 61, 4576–4582, June 1, 2001]

Relative Expression of Progesterone Receptors A and B in EndometrioidCancers of the Endometrium1

Rebecca L. Arnett-Mansfield, Anna deFazio, Gerard V. Wain, Richard C. Jaworski, Karen Byth, Patricia A. Mote,and Christine L. Clarke 2

Westmead Institute for Cancer Research, University of Sydney at the Westmead Millenium Institute, [R. L. A-M., A. deF., P. A. M., C. L. C.], and Departments of GynaecologicalOncology [A. deF., G. V. W.], Tissue Pathology [R. C. J.], and Medicine [K. B.], Westmead Hospital, Westmead, New South Wales 2145, Australia

ABSTRACT

The nuclear receptor for the female hormone progesterone (PR) iswidely expressed in uterine cancer. PR is expressed as two proteins (PRAand PRB) with different functions, and in vitro evidence reveals PRA toinhibit PRB function, so the cellular ratio of PRA:PRB is likely to be animportant determinant of progesterone action. The relative expression ofPRA and B and their involvement in the pathogenesis of endometrialcancer is not known. The aims of this study were to determine PRA andB expression by dual immunofluorescent histochemistry in endometrialadenocarcinomas compared with expression in normal and hyperplasticglands, and to correlate expression in tumors with clinical features includ-ing grade. Significantly lower PR levels were found in tumors comparedwith normal glands and areas of complex atypical hyperplasia within thesame specimen. The normal glands expressed both of the isoforms atsimilar levels, whereas there was increased predominance of one isoformin hyperplastic areas and in tumors, which suggested that the loss ofcoordinated expression of PR isoforms was an early event in tumorprogression. The majority of tumors [27 (58%) of 46] expressed only onePR isoform, and the proportion expressing either PRA or B was the same[14 (30%) of 46, and 13 (28%) of 46, respectively]. One-half of all tumors([23 (50%) of 46] expressed either PRA only or a predominance of PRA,and a few tumors [10 (22%) of 46] expressed comparable levels of PRAand B. Similar levels of PRA and B were noted only in FIGO grade 1tumors, whereas higher grades (2 and 3) were associated with a predom-inance of one isoform. In summary, expression of only one PR isoform wascommon in endometrial cancers, which indicates that the decreased PRlevels observed in these cancers arise from the loss of one PR isoform.Expression of a single PR isoform was associated with higher clinicalgrade, which suggests a relationship between the loss of PR isoformexpression and features of poorer prognosis. Disruption of relative PRisoform expression was observed in complex atypical hyperplasia, whichsuggests that early alterations in the ratio of PRA:PRB may precedeand/or be implicated in the development of endometrial adenocarcinoma.Alterations in the ratio of PR isoform expression are likely to causedisordered regulation of target genes, resulting in altered progestin actionin the uterus, and this may be involved in the pathogenesis of endometrialcancer.

INTRODUCTION

Endometrial carcinomas are the most common malignancy of thefemale genital tract and the third most common cancer in women.They make up 97% of all uterine cancers and arise from the glandswithin the endometrium (1). Estrogen exposure, when in excess and inthe absence of progesterone influence, causes continued stimulation ofthe endometrium and is strongly associated with increased endome-trial carcinoma risk (1–7). Progesterone provides protection by inter-

rupting continued estrogenic stimulation of the endometrium, andpregnancy is protective, because progesterone is the predominanthormone for the duration of pregnancy and confers a break in thecycle of estrogen exposure. Progestins also provide some protectionagainst the stimulatory effects of estrogenic drugs, and hormonereplacement therapy using combinations of estrogens and progestinsyields a lower risk of endometrial carcinoma (8).

Expression of high levels of PR3 is associated with better disease-free, and overall, survival and is an independent predictor of endo-metrial carcinoma outcome (9). PR is encoded by a single geneencoding two proteins (10), termed PRB and PRA, which are struc-turally different in that PRA is a truncated form of PRB, lacking 164amino acids at the NH2 terminus. There is increasing evidence that,although both of the proteins bind progesterone, they have differentcapacities to activate target genes (11–17). In cell culture, PRA andPRB have different abilities to activate transcription, with PRB beingthe more effective activator (14–16), and PRA has a dominant neg-ative effect on PRB and on glucocorticoid, androgen, and mineralo-corticoid receptors in different mammalian cell lines (14, 15, 18). Wenet al. (17) demonstrated, moreover, that PRA repressed estrogenreceptor transcriptional activation. Taken together, evidence to datesuggests that the PR isoforms have different functions on the activa-tion of progestin-regulated promoters, and these differences change,depending on the target cell and gene promoter.

Because PRA and PRB have differing functions and because it hasbeen shown that PRA and PRB exist at different relative levels indifferent tissues (19–24), differing ratios of PRA:PRB are likely toaffect tissue responsiveness to progesterone. Because PR is an impor-tant indicator of response in endometrial cancer patients, but not allPR-positive patients respond, differential expression of PRA and PRBmay be related to patient outcome and response to hormone therapy.The relative expression of PRA and PRB in endometrial cancers andtheir involvement in the pathogenesis of endometrial cancer is notknown. The aims of this study were, firstly, to compare the ratio ofPRA:PRB in tumors with adjacent normal endometrial glands andwith adjacent endometrial hyperplasia, a precursor lesion of endo-metrioid adenocarcinoma; and, secondly, to determine the ratio ofPRA:PRB expression in a cohort of endometrioid adenocarcinoma ofthe endometrium and correlate this with clinical features includingsurgical stage, FIGO grade, and nuclear grade.

MATERIALS AND METHODS

Human Tissues.Patients diagnosed as having endometrioid endometrialadenocarcinoma between 1994 and 1999 were selected from the Department ofGynaecological Oncology database at Westmead Hospital, Westmead. Table 1is a summary of the clinical information on cases examined. Patients whosetumors contained more that 10% of another histological subtype of endometrialcarcinoma were excluded from the cohort because different histological sub-types have different PR expression patterns (25, 26). All archival, formalin-fixed, and paraffin-embedded tissues were acquired from the Department of

Received 8/28/00; accepted 4/3/01.The costs of publication of this article were defrayed in part by the payment of page

charges. This article must therefore be hereby markedadvertisementin accordance with18 U.S.C. Section 1734 solely to indicate this fact.

1 Supported by the National Health and Medical Research Council of Australia, theLeo and Jenny Leukemia and Cancer Foundation, and the Westmead Millenium Foun-dation.

2 To whom requests for reprints should be addressed, at Westmead Institute for CancerResearch, Westmead Hospital, Westmead, NSW 2145, Australia. Phone: 61-2-9845-9068;Fax: 61-2-9845-9102; E-mail: [email protected].

3 The abbreviations used are: PR, progesterone receptor; TXR, Texas red; OR, oddsratio; CI, confidence interval; BMI, body mass index.

4576

Research. on February 2, 2020. © 2001 American Association for Cancercancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

Tissue Pathology at Westmead Hospital. A representative H&E-stained sectionof each tumor was regraded by a single pathologist (R. C. J.) according toFIGO grade and nuclear grade, to ensure that tumor grading was consistentacross the study. Areas of hyperplasia were also identified and classifiedaccording to WHO classification. Control tissues, normal colon and normalmyometrium, formalin-fixed and paraffin-embedded, were also obtained fromthe Department of Tissue Pathology, Westmead Hospital.

Tissue Sectioning and Antigen Retrieval.Formalin-fixed, paraffin-embedded archival specimens were cut at 2mm and mounted onto SuperFrostPlus slides (Menzel-Glaser; supplied by Lomb Scientific, Taren Point,N. S. W., Australia) coated with Mayer egg albumin adhesive (27). Sectionswere dried at 37°C for 72 h, followed by storage at 4°C until use. Antigenretrieval was by autoclaving as described previously (27). Briefly, slides wereplaced in 0.01M sodium citrate buffer (pH 6.0) and autoclaved using aTuttnauer 2540 EKA autoclave (Tuttnauer Co. Ltd., Jerusalem, Israel) at121°C, 15 psi for 30 min, then allowed to cool in the sodium citrate buffer for30 min.

Dual Immunofluorescent Staining. Sections were stained for PRB andthen for PRA as described previously (24). Briefly, to detect PRB, sectionswere incubated with a mouse antihuman PR monoclonal antibody that detectsPRB only (hPRa6; Ref. 28) and with a biotinylated goat antimouse antibody(Dako, Sydney, NSW, Australia), and TXR-avidin (Vector Laboratories, Bur-lingame, CA). To reveal PRA, sections were incubated with a mouse mono-clonal antibody to detect PRA (hPRa7; Ref. 24) and with a biotinylated goatantimouse antibody (Dako) and FITC-avidin (Calbiochem, Sydney, Australia).Sections were mounted with Vectashield mountant for fluorescence (VectorLaboratories) and stored in the dark at 4°C.

Under dual fluorescent excitation, PRB proteins that were labeled with TXRappeared orange; PRA proteins, labeled with FITC, appeared green, and nucleiexpressing equivalent levels of PRA and PRB were yellow. Control sectionswere treated and stained in the same way as the test sections. Controls includedadjacent sections to each tumor sample stained using antibody diluent (PBS/0.5% Triton X-100): (a) in place of both primary antibodies to control fornonspecific staining; and (b) to replace the second sequence primary antibodyto ensure no cross-reactivity between the two staining sequences. Human colonand myometrial tissues were used for a negative and positive control, respec-tively.

Fluorescent Analysis.PR staining was examined using an Olympus BX 40fluorescent microscope fitted with filters to detect both TXR (BP 545–580) andFITC (BP 450–480) fluorescence simultaneously, and each of the two fluo-rochromes separately. The whole section was examined in detail under bothindividual fluorochrome excitations and also using the dual filter, to identifythe staining intensity of the tumor, recorded as either low, moderate, or highand to describe the color of staining. All of the tumors had coexisting PRpositive areas with different PR staining intensity. The predominant area,defined as comprising at least 60% of the total stained area, formed the focusof the analysis.

The relative expression of PRA and PRB was assessed by three independentobservers (R. L. A. M., A. deF., C. L. C) and was determined by the level ofFITC and TXR fluorescence over the entire section under single and dualexcitation. Tumors that contained only one isoform were scored as PRA orPRB; tumors for which both PR isoforms were present with one in predomi-nance were scored as PRA. PRB or PRB. PRA; tumors for which both ofthe PR isoforms were present in similar levels were scored as PRA5 PRB.

Two distinct types of heterogeneity in the relative expression of PRA andPRB proteins were assessed: (a) adjacent cell heterogeneity, in which adjacentcells expressed markedly differing levels of PRA and PRB proteins; and (b)area heterogeneity, in which different regions within the same section showedvariable levels of PRA and PRB proteins. Relative PR isoform expression andthe extent of heterogeneity of relative PRA and PRB expression within a tissuewere each evaluated independently.

Statistical Analysis. Spearman’s rank correlation was used to test forassociation between total PR expression, FIGO grade, nuclear grade, andcell-to-cell homogeneity. Logistic regression was used to test whether therewas an association between relative PR isoform expression, total PR expres-sion, FIGO grade, nuclear grade, and cell-to-cell homogeneity; and ORstogether with their 95% CIs were used to qualify these associations. Kruskal-Wallis one-way ANOVA was used to test whether the distribution of total PRexpression, relative PR isoform expression, and homogeneity was similar

between tumor, hyperplasia, and normal endometrial glands. Multivariatelogistic regression analysis was used to test whether PR concentration and cell-to-cell homogeneity were independent predictors of PR isoform expression.

RESULTS

The cohort of endometrial cancer patients selected for this studyconsisted of 49 patients diagnosed with endometrioid endometrialadenocarcinoma (Table 1). The sections of one case were later iden-tified as containing complex atypical hyperplasia only and, therefore,were excluded from the analysis of tumor samples but were includedin the analysis of hyperplasia. The cases were de-identified beforeanalysis. The majority of cases were stage I, and this is consistent withthe relative frequency of the disease at clinical presentation. Therewere only four cases with stage IV tumors, consistent with thepopulation distribution of stage; for the purpose of data analysis, stageIV was combined with stage III and termed late stage, as has beendone in previous studies (9). FIGO grade distribution of this study issimilar to the distribution of FIGO grade in previous reports (1).Twelve cases contained adjacent normal glands, and 11 cases con-tained complex atypical hyperplasia. Most of the patients were.50years of age (87%), had BMI in the overweight or obese range (64%),and had never taken hormone replacement therapy (58%). Mostpatients (73%) had experienced at least one pregnancy.

Concentration of PR in Endometrial Cancers. Most of the en-dometrial cancers in this cohort were PR positive [46 (96%) of 48],and PR expression was low in the majority of cases [31 (68%) of 46;Table 2]. Adjacent areas of normal endometrium and complex atyp-ical hyperplasia were found in a proportion of tumors and expressionof PR was compared in these nonneoplastic regions. Expression of PR

Table 1 Description of the endometrial cancer cohort

No. ofcases %

StageI 25 52II 9 19III 10 21IV 4 8

FIGO grade1 26 542 12 253 10 21

Nuclear grade1 23 482 17 353 8 17

Age (yr)Average, 66 (range, 36–93 yr)Median, 67 yrPremenopausal,,50 6 13Postmenopausal,$50 42 87

50–64 yr 16 3365–80 yr 20 41.80 yr 6 13

BMI (kg/m2)Normal (,25) 6 13Overweight (25–29) 12 25Obese ($30) 19 39Unknown 11 23

Parity (full-term pregnancies), range: 0–110 11 231–2 11 233 8 174 9 19$5 7 14Unknown 2 4

HRTa

Never 28 58Ever 9 19Unknown 11 23

a HRT, hormone replacement therapy.

4577

PRA AND PRB IN ENDOMETRIAL CANCER



was significantly greater in hyperplasia and normal glands(P 5 0.0001; Table 2), compared with the tumors. The intensity ofexpression of PR was not significantly different in hyperplasia andnormal glands, whereas PR expression in tumors was significantlydifferent from both hyperplasia (P5 0.0007) and normal tissue(P 5 0.0004; Table 2).

PR Isoform Expression. Individual PR isoforms (PRA and PRB)were detected using dual immunofluorescence (Fig. 1,A–C). Underdual excitation, nuclei that expressed equivalent levels of PRA andPRB were yellow (Fig. 1A), whereas nuclei that expressed only PRAwere green (Fig. 1D), and those expressing only PRB were orange(Fig. 1E). The majority [27 (58%) of 46] of tumors expressed only onePR isoform (PRA or PRB; Fig. 1,D or E), and the proportion thatexpressed either isoform was similar [PRB: 13 (28%) of 46; PRA: 14(30%) of 46; Table 2]. In addition to the tumors that expressed onlyPRA, there was a proportion of tumors [9 (20%) of 46] that expressedboth isoforms, but with PRA being in marked excess over PRB, whichindicated that one-half of all of the tumors expressed either PRA aloneor a marked predominance of this isoform. Few tumors [10 (22%) of46] expressed similar levels of PRA and PRB (PRA5 PRB; Fig. 1A;Table 2), and there were no tumors that expressed both isoforms witha marked predominance of PRB.

Expression of a single PR isoform was significantly more commonin tumors with low PR expression (P5 0.02; OR, 2.93; 95% CI,1.19–7.23; Fig. 2A). PRB-only tumors were confined to those casesexpressing low levels of PR. The relative expression of PR isoformswas significantly different between tumor, hyperplasia, and normal(P 5 0.05; Table 2). Cells in normal endometrial glands tended toexpress equivalent levels of PRA and PRB (i.e., PRA 5 PRB),whereas tumors and complex atypical hyperplasia expressed a pre-dominance of one or the other isoform (Fig. 1,J andK). Expressionof PRB alone was only seen in tumor cells and not in hyperplasia ornormal glands (Table 2). The expression of PR isoforms in hyperpla-sia and tumor cells was significantly different from that seen in thenormal areas (Table 2), which suggests that alteration in relativeisoform expression is an early event in the development of endome-trial cancer.

Cell-to-Cell Expression of PR Isoforms.In some tumors, amarked variation in the PR isoform expression between adjacent cellswas observed. This cell-to-cell variation in PR isoform expression wasdescribed as low (Fig. 1F), moderate, or high homogeneity (Fig. 1E).Cell-to-cell expression of PR isoforms was homogeneous in mosttumors [26 (57%) of 46; Table 2]. However, there was a stronginverse relationship between homogeneity and PR expression levels(P 5 0.0003; rank correlation coefficient5 20.51; Fig. 2B);i.e.,tumors that had homogeneous PR isoform expression had low overallexpression of PR, whereas tumors that had heterogeneous cell-to-cellPR isoform expression tended to express higher levels of PR. Homo-geneous cell-to-cell PR isoform expression was also significantlyassociated with expression of a single PR isoform (P5 0.004; OR,0.22; 95% CI, 0.08–0.62; Fig. 2C). Heterogeneity of PR isoformexpression was also noted in adjacent regions (Fig. 1,G–I). In theexample shown in Fig. 1G, the adjacent regions were of differentgrades, with the upper region (PRA5 PRB; Fig. 1G) being nucleargrade 1 and the lower region (PRA. PRB; Fig. 1G) being nucleargrade 2.

Cell-to-cell homogeneity of PR isoform expression was signifi-cantly different between tumor, hyperplasia, and normal (P 5 0.02;Table 2). Cells in normal endometrial glands displayed homogeneousexpression of PR isoforms; however, the appearance of variation incell-to-cell PR isoform expression was noted early in the endometrialdisease continuum. Heterogeneity of PR isoform expression wassignificantly greater in hyperplasia than in normal glands (P 5 0.008;Table 2), whereas there was no significant difference between tumorand hyperplasia (P5 0.73; Table 2).

Relationship between PR Concentration, PR Isoform Expres-sion, and Cell-to-Cell Homogeneity of PR Isoform Expression.There was a significant inverse relationship between PR concentrationand expression of a single PR isoform, and PR concentration andcell-to-cell homogeneity of PR isoform expression. Multivariate lo-gistic regression analysis was used to test whether PR concentrationand cell-to-cell homogeneity were independent predictors of PR iso-form expression. The adjusted ORs for the model showed that therewas a negative association between PR concentration and expression

Table 2 PR expression in the endometrium

PR expression was determined by dual immunofluorescence. Kruskal-Wallis one-way ANOVA was used to compare PR isoform expression in the three tissue types.

Tumor% (case/total)

Hyperplasia% (case/total)

Normal% (case/total) P

Total PRLow 68 (31/46) 9 (1/11) 17 (2/12)} 0.0001a,b

Moderate 17 (8/46) 45 (5/11) 17 (2/12)High 15 (7/46) 45 (5/11) 67 (8/12)

0.0007b,c

0.0004b,d

0.5e

PR isoform expressionA 30 (14/46) 64 (7/11) 25 (3/12)} 0.05a,b

A.B 20 (9/46) 18 (2/11) 0 (0/12)A5B 22 (10/46) 18 (2/11) 75 (9/12)B.A 0 (0/46) 0 (0/11) 0 (0/12)B 28 (13/46) 0 (0/11) 0 (0/12)

0.02b,c

0.02b,e

Cell-to-cell homogeneityLow 17 (8/46) 0 (0/11) 0 (0/12) 0.02a,b

Moderate 26 (12/46) 45 (5/11) 0 (0/12)High 57 (26/46) 55 (6/11) 100 (12/12)}

0.73c

0.006b,d

0.008b,e

a P across the three histological types.b Statistical significance level of,0.05.c Comparison of tumor and hyperplasia.d Comparison of tumor and normal.e Comparison of hyperplasia and normal.

4578




of a single PR isoform (P5 0.008; adjusted OR, 0.11; 95% CI,0.02–0.56), and a positive association between homogeneous cell-to-cell PR isoform expression and expression of a single PR isoform(P 5 0.002; adjusted OR, 22.6; 95% CI, 3.22–157.76), confirmingthat these two variables were independent predictors of the expressionof a single PR isoform.

Association of PR Isoform Expression with Clinical Features.PR expression was significantly inversely related to clinical grade,with proportionally lower levels of PR being seen in higher FIGO(P 5 0.02; rank correlation coefficient5 20.36) and nuclear grades(P 5 0.02; rank correlation coefficient5 20.33; Fig. 3A). Expressionof similar levels of both PR isoforms was found primarily in lower-

Fig. 1. Expression of PRA and PRB in endome-trial cancers and in the normal and hyperplasticendometrium. Relative PR isoform expression wasdetermined by dual immunofluorescent histochemis-try in endometrial tumors and the adjacent hyper-plastic and normal glands.A–C, endometrial cancerexpressing similar levels of PRA and PRB;A, dual(PRA and PRB);B, TXR (PRB); C, FITC (PRA)excitation. D, endometrial cancer expressing PRAonly, dual excitation.E, endometrial cancer express-ing PRB only, dual excitation.F, endometrial cancerexpressing both PRA and PRB, with low cell-to-cellhomogeneity, dual excitation.G–I, endometrial can-cer showing adjacent areas with different ratios ofPR isoform expression (i.e.,area-to-area heterogene-ity); G, dual;H, TXR; I, FITC excitation.J, normalendometrium, dual excitation.K, hyperplastic endo-metrium, dual excitation.3400.

4579




grade tumors (PRA5 PRB; Fig. 3B). Tumors of higher FIGO gradeexpressed one PR isoform only or a predominance of PRA, and thiswas statistically significant (P 5 0.03; OR, 0.15; 95% CI, 0.03–0.79).There was a trend for this relationship to hold also with nuclear gradeas can be seen by the absence of the PRA5 PRB expression profilein nuclear grade 3 tumors (P 5 0.12; OR, 0.42; 95% CI, 0.14–1.27).Cell-to-cell PR isoform expression was homogeneous in tumors withhigher FIGO grade (P5 0.05; rank correlation coefficient5 0.29;Fig. 3C) and nuclear grade (P5 0.04; rank correlation coeffi-cient 5 0.3; Fig. 3C).

There was no statistical difference between tumors with differentclinical stage in PR expression (P 5 0.61; rank correlation coeffi-cient 5 0.08), relative PR isoform expression (P 5 0.87; OR, 1.07;95% CI, 0.48–2.39), or cell-to-cell homogeneity (P 5 0.75; rankcorrelation coefficient5 20.05; not shown). PR isoform expressionwas compared with patient age, BMI, parity, and hormone replace-ment therapy. No association was noted.

DISCUSSION

PR Expression Levels in Endometrial Carcinoma.We haveshown that although PR is expressed in the majority of tumors, theexpression is low, and this is consistent with earlier studies (reviewedin Ref. 29). In this study, a direct comparison of the expression levelsof PR in normal and hyperplastic tissue, within the same specimen asthe cancer, confirmed that the cancer had reduced PR expressionlevels. However, the reduction in levels of PR was not an early eventin endometrial cancer development, as there was no difference in PR

expression levels between hyperplastic and normal areas, whereas PRlevels in tumor areas were significantly lower than in both hyperpla-sias and normal glands.

PRA and PRB Expression in Endometrial Carcinoma.Mostendometrial cancers expressed only one PR isoform, either PRA orPRB, and expression of PRA only or PRB only was equally common.In addition, an additional proportion of tumors, although expressingboth isoforms, showed a marked predominance of PRA, resulting inexpression of PRA only, or PRA predominantly, in one-half of allendometrial cancers. Expression of one PR isoform only, or a pre-dominance of one isoform, was found in over three-quarters of all ofthe tumors and was strongly associated with low tumor levels of PR,which suggests that loss of PR expression was associated with pref-erential loss of one isoform and consequent predominance of one PRisoform. There were no tumors expressing both isoforms with apredominance of PRB. These results show that although loss of PRAand PRB expression takes place frequently in endometrial carcinomas,loss of PRB is slightly more common overall than loss of PRA.

In the normal endometrium, PR is highly expressed in the glandularepithelium during the proliferative phase of the cycle, and this highPR expression is associated with similar PRA and PRB expression(24); in support of these findings, this study showed that normal areaswithin the tumor specimen expressed similar relative levels of PRAand PRB. Endometrial cancers had low PR expression and predomi-nance of one PR isoform, in contrast with the patterns of PR expres-sion seen in the normal endometrium, although, interestingly, endo-metrial cancers with higher PR expression, mimicking PR levels seenin the normal tissues, showed a greater likelihood of similar PRA and

Fig. 3. The association between PR expression levels, relative isoform expression, andcell-to-cell homogeneity of PR isoform expression and clinical grade. Total PR expressionlevels and relative expression of PRA and PRB were determined by dual immunofluo-rescent histochemistry. Comparison of FIGO grade and nuclear grade with PR expressionlevels (A), scored as low (M), moderate (p), and high (f); relative PR isoform expression(B), scored as PRA only (`), PRA. PRB (p), PRA5 PRB (M), PRB only (f); andcell-to-cell homogeneity of PR isoform expression (C), scored as low (f), moderate (`),and high (M).

Fig. 2. Relative PR isoform expression and cell-to-cell homogeneity of PR isoformexpression in the endometrium. PR isoform expression was determined by dual immu-nofluorescent histochemistry. Relative expression of PRA and PRB was scored as PRAonly, PRA. PRB, PRA5 PRB, or PRB only. Cell-to-cell homogeneity of PR isoformexpression was scored as low, moderate, and high.A, comparison of relative PR isoformexpression and PR expression levels in the tumor cohort.`, PRA only;p, PRA . PRB;M PRA 5 PRB;f, PRB only.B, comparison of homogeneity of PR isoform expressionand of overall PR expression levels;M, low PR expression level;p, moderate PRexpression level;f, high PR expression.C, comparison of cell-to-cell homogeneity of PRisoform expression and relative PRA and PRB expression. Symbols have same meaningas inA.

4580




PRB expression levels. Loss of expression of one PR isoform was alsonoted in hyperplasias, and there was a statistically significant differ-ence in relative PR isoform expression between the normal andhyperplastic areas, which suggests that loss of coordinated expressionof PR isoforms is an early event in the development of endometrialcancer.

In normal endometrium, the relative expression of PRA and PRB inadjacent cells is remarkably homogeneous (24), and this was alsonoted in the normal areas of the specimens examined in this study. Inendometrial cancers, this homogeneity was also noted, but only inendometrial cancer expressing a single PR isoform. In contrast to thenormal endometrium, cancers expressing both isoforms tended todisplay heterogeneous cell-to-cell levels of PR isoforms. Moreover, acomparison of normal, hyperplastic, and tumor areas showed thatthere was a significant increase in cell-to-cell heterogeneity of PRisoform expression in complex atypical hyperplasia and cancers,compared with normal areas. This increased cell-to-cell heterogeneityof PR isoform expression in hyperplasias suggests that it is an earlyevent in endometrial cancer development. If loss of one PR isoform isa feature of endometrial cancer development, as suggested in thisstudy, then heterogeneous cell-to-cell expression of both of the PRisoforms may reflect the asynchronous loss of one isoform fromadjacent cells, leading eventually to homogeneous expression of asingle isoform.

Endometrial hyperplasia is an acknowledged precursor of endome-trial carcinoma, and recent studies have shown that there are geneticalterations in hyperplasias that increase in frequency with dysplasiagrade, such that the majority of complex hyperplasias have aberrantgenetic profiles (30). The hyperplasias in this study were all of acomplex type, and the findings of loss of coordinated expression ofPRA and PRB and the appearance of cell-to-cell heterogeneity of PRisoform expression support the view that these changes in PR expres-sion are early events in the development of endometrial carcinoma.

In addition to cell-to-cell heterogeneity of PR isoform expression,there was also heterogeneous expression of PR isoforms in differentareas of some tumors. This was associated with a number of factors,including heterogeneity of grade within a tumor, which is a knownfeature of endometrial carcinoma (31). Although the origin of biolog-ical heterogeneity in endometrial carcinomas is unknown, tumors aregenerally thought to be monoclonal in origin, and tumor progressionis associated with increasingly altered subpopulations within the tu-mor, resulting in tumor heterogeneity (32). Subsequent selection ofvariants with some growth or survival advantage can lead to differentclones arising within the tumor (32). The area-to-area heterogeneityof PR isoform expression may be a manifestation of this clonalevolution.

PR Expression and Clinicopathological Features in Endome-trial Carcinoma. High expression of PR was significantly associatedwith tumors of low grade, which are associated with good prognosis(33, 34), whereas low PR expression was a feature of high-gradetumors. This supports previous findings (9, 35, 36) and shows thatreduced PR expression is associated with poorer-prognosis tumorphenotypes.

Relative PR isoform expression was also related to tumor grade;equivalent expression of both PR isoforms was seen only in FIGOgrade 1 tumors, with loss of one PR isoform, resulting in a predom-inance of the other isoform, being common in higher FIGO grades.The relationship between PR isoform expression and FIGO gradesuggests that expression of a single PR isoform is associated withpoorer prognostic features in endometrial cancer. Two studies, usingcell lines as described previously (37, 38), indicated PRB was notexpressed in poorly differentiated endometrial cancer cells, but thisfinding is not borne out in tumors, as shown in this study.

In the present study, there was no association between PR expres-sion levels and surgical stage. This contradicts reports of an inversecorrelation between PR expression level and stage in other studies (9,35, 36). However, the distribution of tumor grades within the clinicalstages in this study was similar;e.g.,the proportion of FIGO grade 1tumors was 64% in stage I, 33% in stage II, and 50% in both stage IIIand stage IV tumors (not shown). Because PR expression was stronglyassociated with tumor grade, and the distribution of tumor grade wassimilar between the clinical stages, it is perhaps not surprising that noassociation was found with clinical stage in this study.

Taken together, the results of this study show that overall expres-sion of PR is reduced in endometrioid endometrial cancer comparedwith epithelial cells of the normal endometrium and that a reductionin PR expression is accompanied by reduced expression of PRAand/or PRB. Normal endometrial cells usually coexpress both PRAand PRB (24), and both isoforms are implicated in progesterone actionin this tissue (39). This study has shown that endometrial cancer cellscommonly express only one PR isoform, and because overexpressionof PRA in cultured breast cancer cells can result in altered cellresponse to progestins (40), expression of one isoform in carcinomasmay alter hormone action in these tissues. The mechanisms thatcontrol the relative expression of PRA and PRB are not known, butthis study has revealed a descriptive association between disruption ofthese mechanisms and progression to malignancy. Definitive associ-ations between PR isoform expression and endometrial carcinogenesismust await additional studies. The findings of this study, that dis-rupted expression of PR isoforms is an early event in endometrialcancer development and that it is associated with poorer prognosticclinical features, suggests that relative PR isoform expression isimportant in the maintenance of endometrial homeostasis and that itsdisruption may contribute to the malignant phenotype.

ACKNOWLEDGMENTS

We thank Harry Lai, Data Manager, Department of Gynaecological Oncol-ogy, Westmead Hospital, for assistance in compiling the cohort. We thank BillSinai and the Department of Tissue Pathology, Westmead Hospital, for accessto tumor and control tissue blocks.

REFERENCES

1. Rose, P. G. Endometrial carcinoma. N. Engl. J. Med.,335: 640–649, 1996.2. Gallup, D. G., and Stock, R. J. Adenocarcinoma of the endometrium in women 40

years of age or younger. Obstet. Gynecol.,64: 417–420, 1984.3. Mack, T. M., Pike, M. C., Henderson, B. E., Pfeffer, R. I., Gerkins, V. R., Arthur, M.,

and Brown, S. E. Estrogens and endometrial cancer in a retirement community.N. Engl. J. Med.,294: 1262–1267, 1976.

4. Shoff, S. M., and Newcomb, P. A. Diabetes, body size, and risk of endometrialcancer. Am. J. Epidemiol.,148: 234–240, 1998.

5. Kelsey, J. L., LiVolsi, V. A., Holford, T. R., Fischer, D. B., Mostow, E. D., Schwartz,P. E., O’Connor, T., and White, C. A case-control study of cancer of the endome-trium. Am. J. Epidemiol.,116: 333–342, 1982.

6. Elwood, M., Cole, P., Rothman, K. J., and Kaplan, S. D. Epidemiology of endometrialcancer. J. Natl. Cancer Inst. (Bethesda),59: 1055–1060, 1977.

7. Salvesen, H. B., Akslen, L. A., Albrektsen, G., and Iversen, O. E. Poorer survival ofnulliparous women with endometrial carcinoma. Cancer (Phila.),82: 1328–1333,1998.

8. Beresford, S. A. A., Weiss, N. S., Voigt, L. F., and McKnight, B. Risk of endometrialcancer in relation to use of oestrogen combined with cyclic progestagen therapy inpostmenopausal women. Lancet,349: 458–461, 1997.

9. Fukuda, K., Mori, M., Uchiyama, M., Iwai, K., Iwasaka, T., and Sugimori, H.Prognostic significance of progesterone receptor immunohistochemistry in endome-trial carcinoma. Gynecol. Oncol.,69: 220–225, 1998.

10. Kastner, P., Krust, A., Turcotte, B., Stropp, U., Tora, L., Gronemeyer, H., andChambon, P. Two distinct estrogen-related promoters generate transcripts encodingthe two functionally different human progesterone receptor isoforms A and B. EMBOJ., 9: 1603–1614, 1990.

11. Tora, L., Gronemeyer, H., Turcotte, B., Gaub, M., and Chambon, P. The N-terminalregion of the chicken progesterone receptor specifies target gene activation. Nature(Lond.), 333: 185–188, 1988.

12. Meyer, M., Quirin-Stricker, C., Lerouge, T., Bocquel, T., and Gronemeyer, H. Alimiting factor mediates the differential activation of promoters by human progester-one receptor isoforms. J. Biol. Chem.,267: 10882–10887, 1992.

4581




13. Kazmi, S. M. I., Visconti, V., Plante, R. K., Ishaque, A., and Lau, C. Differentialregulation of human progesterone receptor A and B form-mediatedtrans-activationby phosphorylation. Endocrinology,133: 1230–1238, 1993.

14. Tung, L., Mohamed, M. K., Hoeffler, J. P., Takimoto, G. S., and Horwitz, K. B.Antagonist-occupied human progesterone B-receptors activate transcription withoutbinding to progesterone response elements and are dominantly inhibited by A-recep-tors. Mol. Endocrinol.,7: 1256–1265, 1993.

15. Vegeto, E., Shahbaz, M. M., Wen, D. X., Goldman, M. E., O’Malley, B. W., andMcDonnell, D. P. Human progesterone receptor A form is a cell- and promoter-specific repressor of human progesterone receptor B function. Mol. Endocrinol.,7:1244–1255, 1993.

16. Sartorius, C. A., Melville, M. Y., Hovland, A. R., Tung, L., Takimoto, G. S., andHorwitz, K. B. A third transactivation function (AF3) of human progesterone recep-tors located in the unique N-terminal segment of the B-isoform. Mol. Endocrinol.,8:1347–1360, 1994.

17. Wen, D. X., Xu, Y., Mais, D. E., Goldman, M. E., and McDonnell, D. P. The A andB isoforms of the human progesterone receptor operate through distinct signalingpathways within target cells. Mol. Cell Biol.,14: 8356–8364, 1994.

18. McDonnell, D. P., Shahbaz, M. M., Vegeto, E., and Goldman, M. E. The humanprogesterone receptor A-form functions as a transcriptional modulator of mineralo-corticoid receptor transcriptional activity. J. Steroid Biochem. Mol. Biol.,48: 425–432, 1994.

19. Schrader, W. T., and O’Malley, B. W. Progesterone-binding components of chickoviduct. IV. Characterization of purified subunits. J. Biol. Chem.,247: 51–59, 1972.

20. Loosfelt, H., Atger, M., Misrahi, M., Guiochon-Mantel, A., Meriel, C., Logeat, F.,Benarous, R., and Milgrom, E. Cloning and sequence analysis of rabbit progesterone-receptor complementary DNA. Proc. Natl. Acad. Sci. USA,83: 9045–9049, 1986.

21. Ilenchuk, T. T., and Walters, M. R. Rat uterine progesterone receptor analyzed by[3H]R5020 photoaffinity labeling: evidence that the A and B subunits are notequimolar. Endocrinology,120: 1449–1456, 1987.

22. Graham, J. D., Yeates, C., Balleine, R. L., Harvey, S. S., Milliken, J. S., Bilous,A. M., and Clarke, C. L. Characterization of progesterone receptor A and B expres-sion in human breast cancer. Cancer Res.,55: 5063–5068, 1995.

23. Bethea, C. L., and Widmann, A. A. Differential expression of progestin receptorisoforms in the hypothalamus, pituitary, and endometrium of rhesus macaques.Endocrinology,139: 677–687, 1998.

24. Mote, P. A., Balleine, R. L., McGowan, E. M., and Clarke, C. L. Colocalization ofprogesterone receptors A and B by dual immunofluorescent histochemistry in humanendometrium during the menstrual cycle. J. Clin. Endocrinol. Metab.,84: 2963–2971,1999.

25. Lax, S. F., Pizer, E. S., Ronnett, B. M., and Kurman, R. J. Clear cell carcinoma of theendometrium is characterized by a distinctive profile of p53, Ki-67, estrogen, andprogesterone receptor expression. Hum. Pathol.,29: 551–558, 1998.

26. Lax, S. F., Pizer, E. S., Ronnett, B. M., and Kurman, R. J. Comparison of estrogen andprogesterone receptor, Ki-67, and p53 immunoreactivity in uterine endometrioid

carcinoma and endometrioid carcinoma with squamous, mucinous, secretory, andciliated cell differentiation. Hum. Pathol.,29: 924–931, 1998.

27. Mote, P. A., Leary, J. A., and Clarke, C. L. Immunohistochemical detection ofprogesterone receptors in archival breast cancer. Biotech. Histochem.,73: 117–127,1998.

28. Clarke, C. L., Zaino, R. J., Feil, P. D., Miller, J. V., Steck, M. E., Ohlsson-Wilhelm,B. M., and Satyaswaroop, P. G. Monoclonal antibodies to human progesteronereceptor: characterization by biochemical and immunohistochemical techniques. En-docrinology,121: 1123–1132, 1987.

29. Nyholm, H. C. J. Estrogen and progesterone receptors in endometrial cancer. Clini-copathological correlations and prognostic significance. APMIS Suppl.,65: 104:5–33, 1996.

30. Kiechle, M., Hinrichs, M., Jacobsen, A., Luttges, J., Pfisterer, J., Kommoss, F., andArnold, N. Genetic imbalances in precursor lesions of endometrial cancer detected bycomparative genomic hybridization. Am. J. Pathol.,156: 1827–1833, 2000.

31. Kurman, R. J. Grading of endometrial carcinoma. Verh. Dtsch. Ges. Pathol.,75:376–377, 1991.

32. Nowell, P. C. Mechanisms of tumor progression. Cancer Res.,46: 2203–2207, 1986.33. Nordstrom, B., Strang, P., Lindgren, A., Bergstrom, R., and Tribukait, B. Carcinoma

of the endometrium: do the nuclear grade and DNA ploidy provide more prognosticinformation than do the FIGO and WHO classifications. Int. J. Gynecol. Pathol.,15:191–201, 1996.

34. Salvesen, H. B., Iversen, O. E., and Akslen, L. A. Prognostic impact of morphometricnuclear grade of endometrial carcinoma. Cancer (Phila.),83: 956–964, 1998.

35. Carcangiu, M. L., Chambers, J. T., Voynick, I. M., Pirro, M., and Schwartz, P. E.Immunohistochemical evaluation of estrogen and progesterone receptor content in183 patients with endometrial carcinoma. Part 1. Clinical and histological correla-tions. Am. J. Clin. Pathol.,94: 247–254, 1990.

36. Nyholm, H. C. J., Nielsen, A. L., Lyndrup, J., Norup, P., and Thorpe, S. M.Biochemical and immunohistochemical estrogen and progesterone receptors in ade-nomatous hyperplasia and endometrial carcinoma: correlations with stage and otherclinicopathologic features. Am. J. Obstet. Gynecol.,167: 1334–1342, 1992.

37. Leslie, K. K., Kumar, N. S., Richer, J., Owen, G., Takimoto, G., Horwitz, K. B., andLange, C. Differential expression of the A and B isoforms of progesterone receptorin human endometrial cancer cells. Only progesterone receptor B is induced byestrogen and associated with strong transcriptional activation. Ann. NY Acad. Sci.,828: 17–26, 1997.

38. Kumar, N. S., Richer, J., Owen, G., Litman, E., Horwitz, K. B., and Leslie, K. K.Selective down-regulation of progesterone receptor isoform B in poorly differentiatedhuman endometrial cancer cells: implications for unopposed estrogen action. CancerRes.,58: 1860–1865, 1998.

39. Graham, J. D., and Clarke, C. L. Physiological action of progesterone in target tissues.Endocr. Rev.,18: 502–519, 1997.

40. McGowan, E. M., and Clarke, C. L. Effect of overexpression of progesterone receptorA on endogenous progestin-sensitive endpoints in breast cancer cells. Mol. Endocri-nol., 13: 1657–1671, 1999.

4582




2001;61:4576-4582. Cancer Res Rebecca L. Arnett-Mansfield, Anna deFazio, Gerard V. Wain, et al. Endometrioid Cancers of the EndometriumRelative Expression of Progesterone Receptors A and B in

Updated version

http://cancerres.aacrjournals.org/content/61/11/4576

Access the most recent version of this article at:

Cited articles

http://cancerres.aacrjournals.org/content/61/11/4576.full#ref-list-1

This article cites 40 articles, 7 of which you can access for free at:

Citing articles

http://cancerres.aacrjournals.org/content/61/11/4576.full#related-urls

This article has been cited by 19 HighWire-hosted articles. Access the articles at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/61/11/4576To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/content/61/11/4576.full#ref-list-1

http://cancerres.aacrjournals.org/content/61/11/4576.full#related-urls

http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]



Documents

Relative Expression of Progesterone Receptors A and B in ... · rupting continued estrogenic stimulation of the endometrium, and pregnancy is protective, because progesterone is the