Steroid receptors in hereditary breast carcinomas associated with BRCA1 or BRCA2 mutations or unknown susceptibility genes

Steroid Receptors in Hereditary Breast CarcinomasAssociated with BRCA1 or BRCA2 Mutations orUnknown Susceptibility Genes

Niklas Loman, M.D.

Oskar Johannsson, M.D., Ph.D.

Par-Ola Bendahl, Ph.D.

Åke Borg, Ph.D.

Mårten Ferno, Ph.D.

Håkan Olsson, M.D., Ph.D.

Jubileum Institute, Department of Oncology, Uni-versity Hospital, Lund, Sweden.

The current study was supported by grants fromthe Swedish Cancer Society, the Nordic CancerUnion, the Mrs. Berta Kamprads Foundation, theUniversity of Lund Medical Faculty, the CTRF, theGunnar Arvid and Elisabeth Nilsson Foundation, theJohn and Augusta Persson Foundation, the Hospi-tal of Lund Foundations, and King Gustav V’sJubilee Foundation.

Address for reprints: Niklas Loman, M.D., JubileumInstitute, Department of Oncology, University Hos-pital, S-221 85 Lund, Sweden.

Received September 24, 1997; revision receivedJanuary 16, 1998; accepted January 16, 1998.

BACKGROUND. The expression of steroid receptors is a common feature of both

male and female breast carcinomas and is also one of the most important prog-

nostic factors for patients with this disease. Steroid receptor levels in BRCA1-

related breast carcinoma have reportedly been low. Little data on steroid receptor

levels have been reported with regard to BRCA2.

METHODS. Steroid receptor levels were analyzed in 27 breast carcinomas associated

with BRCA1 mutations, 14 associated with BRCA2 mutations, and 32 from indi-

viduals who had hereditary breast carcinoma but no detectable mutations of either

BRCA1 or BRCA2. Breast carcinomas from 32 consecutive male patients, 6 of whom

had mutations of BRCA2, were also examined for steroid receptors. Estrogen

receptor (ER) and progesterone receptor (PgR) analyses were performed with

radioligand or enzyme immunoassay techniques on tumor cytosol preparations.

Germline mutation screening and detection were performed using the protein

truncation test, single strand conformation polymorphism, and direct sequencing

on DNA from normal tissue.

RESULTS. The BRCA1-related tumors expressed significantly lower levels of ER than

tumors from the other hereditary groups. The PgR levels were significantly lower in

the BRCA1-related cases than in the hereditary cases not related to BRCA1 or

BRCA2, but not significantly lower than in the BRCA2-related cases. Fourteen of 32

(44%) of the hereditary tumors not related to BRCA1 or BRCA2 had PgR levels

exceeding 100 fmol/mg of protein. The tumors from male patients with BRCA2-

related disease did not have receptor levels that differed from those in non-BRCA2-

related tumors.

CONCLUSIONS. BRCA1- and BRCA2-related breast tumors were distinct in their

expression of steroid receptors. Moreover, a subgroup of tumors not related to

BRCA1 or BRCA2 manifested a strongly positive PgR phenotype rarely seen in

BRCA1- and BRCA2-related tumors. These characteristics may be of relevance to

the treatment and follow-up of high risk individuals in these families and may help

identify a homogenous category of hereditary breast carcinomas not related to

BRCA1 or BRCA2 in which new susceptibility genes may be sought. Cancer 1998;

83:310 –9. © 1998 American Cancer Society.

KEYWORDS: hereditary breast carcinoma, male breast carcinoma, estrogen recep-tor, progesterone receptor, BRCA1, BRCA2.

Susceptibility to breast carcinoma is inherited by transmission ofan autosomal dominant allele in approximately 5–10% of all cas-

es.1,2 Two major tumor suppressor genes associated with hereditarybreast carcinoma have been identified, BRCA1 and BRCA2.3–5 Germ-line mutations in BRCA1 are associated with an increased risk ofbreast and ovarian carcinoma in females; the risk of developing an

310

© 1998 American Cancer Society

invasive breast carcinoma is estimated at approxi-mately 70% by age 70 years.6 Germline mutations inBRCA2 confer an increased risk for both female andmale breast carcinoma. Other malignancies, such ascancer of the prostate, pancreas, larynx, ovarium, cer-vix, and ureter, as well as malignant melanoma of theeye, may also be seen more frequently than expectedin BRCA2 mutation carriers.6 –9 The risk of developingbreast carcinoma was 82% by age 80 years in onestudy of a large BRCA2 kindred10 and almost 100%among Icelandic breast carcinoma families with theBRCA2 founder mutation.11 However, risk estimateshave been lower in population-based studies. In astudy of Ashkenazi Jews carrying any of three recur-rent BRCA1 and BRCA2 mutations, the risk for breastcarcinoma was 56% at age 70 years.12

Depending on the population studied, germlinemutation in BRCA1 can be detected in 9 –79% of fam-ilies with 3 or more cases of breast and/or ovariancarcinoma, whereas BRCA2 accounts for an additional8 – 64% of remaining families.13 Minor fractions of thetotal number of inherited breast carcinomas are asso-ciated with rare cancer syndromes of identified ormapped genetic origin, including the Li–Fraumenisyndrome and TP53,14 the Cowden disease and PTEN/MMAC1,15,16 and the Peutz–Jegher syndrome gene onchromosome 19p.17 However, a considerable percent-age of hereditary breast carcinomas remain unrelatedto a definite gene. In our previous study of 106 Scan-dinavian breast carcinoma families, the frequencies ofBRCA1 and BRCA2 mutations were 23% and 11%, re-spectively, rendering almost two-thirds of families ge-netically uncharacterized.18 The search for new breastcarcinoma susceptibility genes is currently intense; itinvolves various procedures such as linkage analysisand/or mapping of chromosomal deletions in tumors.These approaches are likely to be successful if occur-rences in the remaining families are due to one or afew additional major tumor suppressor genes, butthey are less useful in a more heterogeneous situation.Alternatively, new insights into the causes of heredi-tary breast carcinoma may be obtained from func-tional studies of the known BRCA proteins and theirinteracting cellular components, as well as from phe-notypic studies of genetically modified animals or oftumors in families with genetic predisposition.

Recent evidence suggests that the BRCA1 andBRCA2 proteins are involved in maintaining genomeintegrity, possibly by participating with the Rad51protein in recombination linked repair of double-stranded DNA breaks and by transcriptional coactiva-tion in complex with RNA polymerase II holoen-zyme.19 –21 Both genes demonstrate a virtuallyindistinguishable spatial and temporal pattern of ex-

pression in many fetal and adult tissues, which is mostprominent in proliferating compartments undergoingdifferentiation.22 In the breast, BRCA1 and BRCA2mRNA expression are induced during puberty andpregnancy, suggesting a regulation by sex hormones.22

Tumors from BRCA1 and BRCA2 mutation carriers arecharacterized by a significantly higher number ofchromosomal aberrations than are found in sporadicbreast carcinoma, which indicates that both genes area critical factor in DNA repair. These aberrations alsoshow gene specific alteration patterns, suggesting thatthe accumulation of somatic changes in tumor pro-gression may follow unique pathways that are relatedto genetic origin.23

Clinical and histopathologic analysis of BRCA1-and BRCA2-related tumors further emphasizes thedistinction between sporadic and differing types ofhereditary breast carcinoma. BRCA1-related breastcarcinoma was found to be poorly differentiated duc-tal carcinoma, often with dense lymphocyte infiltra-tion, DNA aneuploidy, and a high S-phase fraction.24

Similar results were obtained in a subsequent study, inwhich BRCA1 tumors were more often found to be amedullary or atypical medullary type, more pleomor-phic, with less tubule formation and a higher mitoticcount than sporadic tumors, whereas BRCA2 tumorshad less tubule formation than sporadic cases butotherwise fewer specific features.25

Steroids, and especially estrogen, are central reg-ulators of the growth and differentiation of the normalmammary gland and are also thought to be involvedin the development and progression of breast carci-noma. Estrogen receptor (ER) and progesterone re-ceptor (PgR) status are among the most important andmost commonly evaluated prognostic factors in breastcarcinoma as well as the most important predictors ofresponse to endocrine therapy.26 Female breast carci-noma is ER and PgR positive in about 65% and 50% ofall cases, respectively, and ER status especially showsan age-dependent variation in which tumors fromyounger patients are less often positive.27 Male breastcarcinoma most often has an ER positive phenotype.28

We have previously reported a low frequency of ERand PgR positive immunostaining in BRCA1-relatedbreast carcinoma.24 Data on BRCA2 and receptor sta-tus is scarce; Karp et al. reported on four patients ofAshkenazi descent who were all ER positive.29

In the current study, we have used biochemicaltechniques for quantitative measurement of ER andPgR in three groups of hereditary breast carcinoma,consisting of tumors from female BRCA1 and BRCA2mutation carriers as well as tumors from female pa-tients with a defined heredity for breast and/or ovar-ian carcinoma but without signs of mutation in BRCA1

Steroid Receptors in Hereditary Breast Carcinoma/Loman et al. 311

or BRCA2. These groups are compared with one an-other and, in analyses stratified for age and calendartime, also compared with 8948 consecutive tumorsfrom breast carcinoma patients treated in the SouthSweden Health Care Region during the period 1978 –1996. In addition, receptor levels in six breast tumorsfrom male BRCA2 mutation carriers are comparedwith a group of 26 sporadic breast tumors from males.We confirm our previous finding of significantly lowerER and PgR values in BRCA1-related tumors and re-port on the identification of non-BRCA1/BRCA2-re-lated hereditary cases with high ER and PgR levels. Theresults may have implications for defining a new sub-class of hormone-sensitive hereditary breast carci-noma as well as for the clinical management of thesepatients.

MATERIALS AND METHODSPatientsThe three groups of hereditary breast carcinoma diag-nosed during the period 1978 –1996 were collectedfrom a set of Scandinavian kindreds with familialbreast or breast-ovarian carcinoma analyzed for germ-line mutation in BRCA1 and BRCA2.18 The first group(BRCA1-related) included 25 tumors from 20 patientsin 15 families with 8 different proven BRCA1 germlinemutations. An additional two cases from a kindred(Lund1) with a strong linkage to BRCA1 (LOD score1.6), as well as LOH of the wild-type BRCA1 allele infive of six analyzed tumors from haplotype carriers,were included. The second group (BRCA2-related) in-cluded 14 tumors from 13 patients in 8 families with 7different BRCA2 germline mutations. A third group(non-BRCA1/BRCA2) was composed of 30 families inwhich screening for BRCA1 and BRCA2 had been neg-ative. These were families with at least 3 first-degreerelatives affected by breast and/or ovarian carcinoma,1 of whom was diagnosed before age 50 years. Alter-natively, these families had 2 affected first-degree rel-atives, 1 of whom was diagnosed before age 40 years,or 1 single affected woman younger than 30 years.This pattern could be compatible with a dominantmendelian heritage. One breast carcinoma patientfrom each of the 30 non-BRCA1/BRCA2 families wasincluded in this study. In two cases, bilateral breastcarcinomas were included; thus, a total of 32 freshfrozen tumors were analyzed.

Informed consent was obtained before mutationanalysis was performed, and the families were fol-lowed up through the oncogenetic reception jointlyrun by the Departments of Oncology and Clinical Ge-netics in Lund.

Since 1978, fresh tumor material for hormone re-ceptor analysis and flow cytometry has been sent to

our laboratory for routine analysis after breast carci-noma surgery at the different hospitals in the SouthSweden Health Care Region. Thus, a bank of almost9000 breast carcinoma tumors have been collectedand tissues analyzed in these respects. The groups ofhereditary breast carcinoma tissues were comparedwith one another and with this consecutive referencematerial, which mainly consisted of sporadic breastcarcinoma. For the comparisons, this material wasstratified according to patient age and calendar year ofdiagnosis.

Thirty-two breast carcinoma tumors from malepatients from 1985–1996 were sent to our laboratoryfor receptor analysis. These consecutive cases werealso included separately in the study.

MethodsGermline mutations in BRCA1 and BRCA2 were de-tected by screening the DNA extracted from periph-eral lymphocytes collected from affected individuals.In 31 of the 32 men included in the study, mutationswere identified in DNA obtained from fresh frozentumor material. In these cases, mutations were con-firmed in the germline by verification in archived par-affin blocks of normal tissue. The methods used werethe protein truncation test (PTT), single-strand con-formation polymorphism (SSCP), and direct sequenc-ing. We estimated the sensitivity for detection of mu-tation to be approximately 80%.18

Estrogen receptor (ER) levels were analyzed ac-cording to clinical praxis with isoelectric focusing inpolyacrylamide gel (IF) or enzyme immunoassay (EIA;Abbot Laboratories).30 Similarly, progesterone recep-tor (PgR) levels were analyzed using the multiple-point dextran-coated charcoal technique (DDC), orEIA.31,32 The interassay agreements were found to behighly concordant, although higher levels were ob-tained with the EIAs. Thus, all values with IF and DDCwere recalculated to allow for comparison of datafrom different time periods.27,33,34 The cutoff level fora receptor positive tumor was set at 25 fmol/mg pro-tein in our laboratory.

In some cases, when women have had two tumorsthat have clinically been considered separate, bothwere included separately in this study. Clinical data ontumor stage was collected from case records.

StatisticsAge differences between the groups were evaluatedusing a two-tailed Student’s t test, whereas receptorlevels, due to considerable skewness, were comparedusing the nonparametric Kruskal–Wallis test. Due tolow cell frequencies, Fisher’s exact test was used forevaluation of stage differences among groups. Differ-

312 CANCER July 15, 1998 / Volume 83 / Number 2

ences in distribution of receptor positive and negativetumors were analyzed using chi-square tests.

Receptor levels in breast carcinoma are age-de-pendent. To rule out the possibility that observed sig-nificant differences in hormone receptor levels were aconsequence of differences in age distribution, thegroups with hereditary breast carcinoma were com-pared with reference material, which consisted of 8948consecutive, mainly sporadic breast carcinomas. Westratified the reference material into 5-year intervalsaccording to both age and calendar time at diagnosis,transformed the values toward normality (the squareroot), and computed the mean and standard deviationin each stratum. For each of the three hereditarygroups, we then computed a series of differences be-tween observed transformed receptor level and thecorresponding mean. The approximate normal distri-bution of the sum of those differences led to an age-and calendar time– corrected test of the no-differencehypothesis.

All statistical tests performed were two-tailed; andwhen significance is mentioned without reference to alevel, the level of P , 0.05 is assumed. Statisticalcalculations were made using Stata 5.0 software.35

RESULTSBreast Carcinomas from FemalesAge distribution was similar in the three groups ofhereditary breast carcinoma: BRCA1 median 5 44.0years, range 5 26 – 60 (n 5 27); BRCA2 median 5 45.0,range 5 27– 82 (n 5 14); non-BRCA1/BRCA2 median 547.5, range 5 28 – 81 (n 5 32). Tumor stages (I–IV)were also similar in the three groups (P 5 0.19) (Table1). In the cases in which stage was unknown, this wasdue to either lack of information (in one case) orbecause axillary dissection was not performed (inthree cases).

Screening for statistical differences in ER and PgRlevels among the three groups, using the Kruskal–

Wallis test, revealed differences in both receptor types.P 5 0.0006 (ER) and P 5 0.0014 (PgR), respectively.

Pairwise comparisons of ER levels (fmol/mg pro-tein) between the groups revealed significantly lowervalues among BRCA1 tumors (median 5.4, range0 –270) compared with BRCA2 tumors (median 48.5,range 0 –550) (P 5 0.014) and non-BRCA1/BRCA2 tu-mors (median 44.5, range 0 – 410) (P 5 0.0002). Nostatistical difference was observed between BRCA2 tu-mors and non-BRCA1/BRCA2 tumors (Fig. 1).

BRCA1 tumors were found to manifest signifi-cantly lower PgR levels (fmol/mg protein) (median 1.1,range 0 –196) compared with non-BRCA1/BRCA2 tu-mors (median 62.0, range 0 –1960) (P 5 0.0004). Anonsignificant trend toward lower PgR values com-pared with BRCA2 tumors was observed (median 519.0, range 5 0 – 600) (P 5 0.075). The difference inPgR levels between BRCA2 tumors and non-BRCA1/BRCA2 tumors was not statistically significant (P 50.13). However, whereas only 1 of 26 BRCA1 tumors(3.8%) and 1 of 13 BRCA2 tumors (7.7%) manifestedhigh PgR values (.100 fmol/mg protein), this wasevident in 14 of 32 (44%) of the non-BRCA1/BRCA2tumors (Fig. 2).

Similar results were obtained when the frequencyof ER positivity (ER1) and PgR positivity (PgR1) wascompared among the three groups of hereditarybreast carcinoma. Receptor positivity was defined as areceptor value $ 25 fmol/mg protein. BRCA1 tumorswere significantly less ER1 (15%) compared withBRCA2 tumors (57%) (P 5 0.005) and non-BRCA1/BRCA2 tumors (59%) (P , 0.001). BRCA1 tumors werealso significantly less PgR1 (19%) than non-BRCA1/

TABLE 1Clinical Stages of Disease in Three Groups of Hereditary BreastCarcinomas

Stage

No. of cases

BRCA1 BRCA2 non-B1/B2 Total

I 12 2 11 25II 14 9 16 39III 0 2 2 4IV 0 0 1 1Unknown 1 1 2 4

Total 27 14 32 73

FIGURE 1. This dotplot shows estrogen receptor (ER) levels in the different

groups of patients with hereditary breast carcinoma. Median values and the

threshold for receptor positivity (25 fmol/mg protein) are marked in the plot.

Log scale on y-axis.


BRCA2 tumors (59%) but not significantly differentfrom BRCA2 tumors (31%) (P 5 0.42). Again, there wasa trend toward a significant difference between PgRstatus in BRCA2 and non-BRCA1/BRCA2 cases (P 50.082). There was no such trend in ER status (Table 2).

To compare receptor levels between differentgroups of hereditary and sporadic breast carcinomaand to preclude the impact of age on receptor values,an age- and calendar year–stratified analysis was per-formed, comparing the three hereditary groups with areference group of 8948 consecutive breast tumorscollected and analyzed between the years 1978 and1996. BRCA1 tumors had significantly lower ER levels(P 5 0.0008) and PgR levels (P 5 0.0004) than thereference group, which mainly consisted of sporadiccases. Receptor levels in the BRCA2 and non-BRCA1/BRCA2 tumors did not differ significantly from thereference group (Figs. 3 and 4).

Regarding the intragenomic localization of thedifferent mutations within the two genes and its effecton ER and PgR status, no clear trend was observed.The receptor negative and positive cases appeared

evenly distributed with respect to both genes. Fourbilateral cases were included in the material. Of thethree cases with BRCA1 mutation, one had synchro-nous tumors with an ER2/PgR1 and an ER2/PgR2phenotype, another had two synchronous ER2/PgR2tumors, and a third presented with an ER2/PgR2 andan ER2/PgR1 tumor at ages 32 and 37 years, respec-tively. The single case with BRCA2 mutation and bi-lateral disease had an ER2/PgR2 and an ER2/PgR1tumor at ages 72 and 75 years, respectively.

Breast Carcinomas from MalesComparing the ages of breast carcinoma onset be-tween a group of 6 males with BRCA2 mutation and agroup of 26 males without detectable BRCA2 mutationrevealed a weak trend (P 5 0.15) toward younger agein the former group (median 64.1 years, range 46 –78,vs. median 72.6, range 55–93) (Fig. 5). No obviousdifferences were observed between the groups regard-ing the ER and PgR levels. BRCA2 tumors expressed ERat a median level (fmol/mg protein) of 185, range1.9 –250; PgR median 90, range 34 –260. This should becompared with the cases not related to BRCA2: ERmedian 245, range 0 –550; PgR median 145, range0 – 810 (Figs. 6 and 7).

DISCUSSIONIt has previously been reported that breast tumorsfrom BRCA1 mutation carriers differ from other breasttumors by having higher histologic grade, DNA aneu-ploidy, and S-phase values.24,25,36 We have previouslyalso observed a significantly lower frequency of ER1and PgR1 in BRCA1 tumors as compared with age-and stage-matched controls in immunohistochemicalstaining of paraffin sections.24 In the current study, inwhich largely different and more extensive tumor ma-terial was studied using biochemical techniques forreceptor analysis, we confirmed and extended ourprevious findings, reported by Johannsson et al., bydemonstrating significantly lower ER and PgR levels inBRCA1 tumors as compared with sporadic tumors andother tumors from genetically predisposed individu-als. Hereditary breast carcinoma was categorized hereinto three distinct groups. Two were homogeneousgroups associated with identified germline mutationsin the BRCA1 or BRCA2 gene, respectively, and a third,supposedly heterogeneous, group comprised tumorsfrom families interpreted as carrying a dominant dis-ease allele but in whom no mutations in BRCA1 andBRCA2 had been found. Half of the BRCA1 tumorsshowed a complete loss of PgR expression, as com-pared with 2 of 13 BRCA2 tumors and 6 of 32 non-BRCA1/BRCA2 tumors. In contrast, close to half of thenon-BRCA1/BRCA2 hereditary tumors manifested

FIGURE 2. This dotplot shows progesterone receptor (PgR) levels in the

different groups of patients with hereditary breast carcinoma. Median values

and the threshold for PgR positivity (25 fmol/mg protein) are marked in the plot.

Log scale on y-axis.

TABLE 2Estrogen Receptor (ER) and Progesterone Receptor (PgR) Status inTumors from Three Groups of Hereditary Breast Carcinoma

No. (%) of cases

ER1 PgR1

BRCA1 4/27 (15%) 5/26 (19%)BRCA2 8/14 (57%) 4/13 (31%)non-BRCA1/BRCA2 19/32 (59%) 19/32 (59%)


high PgR levels (.100 fmol/mg protein), a phenotyperarely seen in BRCA1 and BRCA2 tumors.

No significant differences in steroid receptor lev-els were observed between hereditary tumors relatedto BRCA2 or without detectable mutation in the age-matched comparisons with the reference materialconsisting of consecutive, mainly sporadic tumors. Inthis analysis, differences between hereditary and non-hereditary tumors may have been underestimated be-cause the rate of hereditary tumors in the referencematerial could have been expected to be higher for theyounger age groups. Similar results were reported byKarp et al., who analyzed ER levels in breast carcino-mas from patients of Ashkenazi descent, a populationwith a high frequency of specific founder mutations inBRCA1 (185delAG and 5382insC) and BRCA2(6174delT). In their study, breast carcinoma patientswere screened for these mutations and the rates ofER1 were studied. Tumors from patients with germ-line mutations in BRCA1 were significantly less oftenER1 than tumors from patients without one of thesetwo founder mutations. Four cases of breast carci-noma related to germline mutation in BRCA2 werealso reported; they were all ER1.29

Estrogens are potent mitogens and act via binding

to nuclear receptors, ligand-inducible enhancers thatmodulate the expression of certain genes by bindingto short DNA sequences, or estrogen responsive ele-ments (EREs) located in the vicinity of the regulatedgenes.37 The PgR gene is a major estrogen-regulatedgene, and high tumor levels of PgR protein are indic-ative of a functional ER mechanism and hormone-responsive phenotype. This information is clinicallyimportant and useful in predicting response to endo-crine therapy. Patients with advanced disease and anER2/PgR2 primary tumor seldom respond to tamox-ifen therapy (the response rate is ,10%), whereas theresponse rate among ER1/PgR1 cases is 60 –70%.38

The presence of ER and PgR is also associated with amore well-differentiated tumor type and is thus amarker of good prognosis. BRCA1 tumors are, asshown here, mainly receptor negative and have manyother histopathologically aggressive features, such asnuclear pleomorphism, high mitotic count, and a pau-city of tubular formations. However, this is contradic-tory to reported survival analyses claiming thatBRCA1-related breast carcinoma is associated with abetter prognosis than other hereditary cases or spo-radic breast carcinoma.36,39,40 A similar paradox is ev-ident in hereditary nonpolyposis colon carcinoma,

FIGURE 3. Estrogen receptor (ER) levels are compared among different groups

of hereditary breast carcinoma—(A) BRCA1, (B) BRCA2, and (C) non-BRCA1/

BRCA2—and the reference material consisting of 8948 consecutive tumors ana-

lyzed since 1978. Dots represent hereditary breast carcinoma cases; boxes rep-

resent reference material and show median and interquartile range. Log scale on

y-axis.


which is associated with a genomic instability andmutator phenotype but is nonetheless associated witha better prognosis than sporadic colon carcinoma.41

However, our own experience from studies of BRCA1-related breast carcinoma patients suggests a survivalrate among BRCA1 cases that is equal to or even worsethan the rate among age- and stage-matched controlpatients.42

Structural and functional studies of the BRCA1and BRCA2 proteins have suggested several sharedfeatures; both have similar cell cycle–specific expres-sion and subcellular localization, manifest an interac-tion with the Rad51 protein, and possess transcrip-tional activities.19,21,43,44,45 Moreover, transcription ofboth BRCA1 and BRCA2 is thought to be directly orindirectly initiated by estrogen stimulation in cell cul-ture models.46 – 48 In spite of these similarities, thereare indications that different pathogenetic pathwaysare involved in BRCA1- and BRCA2-related tumors.However, recent studies suggest that BRCA1 andBRCA2 are differentially regulated by sex hormones.The ratio of BRCA1 to BRCA2 expression is higher inthe female breast than in the male breast, and treat-ment of mice having undergone ovariectomy with

17b-estradiol or progesterone has given rise to a morepronounced induction of BRCA1 than of BRCA2.22

This may reflect the gender-specific difference inBRCA1- and BRCA2-related tumors. Furthermore,comparative genomic hybridization (CGH) analysishas shown that BRCA1 and, less often, BRCA2 tumorshave specific patterns of chromosomal aberrationsthat are distinct from each other and from sporadicbreast carcinoma.23 This suggests that some initialgenetic events in tumorigenesis may determine a pre-ferred course and combination of somatic geneticchanges in tumor progression and that these changesare different in cases related to BRCA1 and BRCA2.

It has been suggested that the hormonal and pro-liferative properties of breast carcinoma are related tothe situation in the premalignant tissue at the time oftumor induction.49 In mice, BRCA1 is expressed inbreast tissue during adolescence and pregnancy,when the levels of ER are also increasing. The induc-tion of BRCA1-related tumors may occur during aperiod of the life cycle when ER levels are low. By thesame token, the expression of BRCA2 may be essentialfor mammary gland development during a stage whenER and PgR levels have increased, leaving a hormone-

FIGURE 4. Progesterone receptor (PgR) levels are compared among different

groups of hereditary breast carcinoma—(A) BRCA1, (B) BRCA2, and (C) non-

BRCA1/BRCA2—and the reference material consisting of 8948 consecutive tumors

analyzed since 1978. Dots represent hereditary breast cancer cases; boxes rep-

resent reference material and show median and interquartile range. Log scale on

y-axis.


dependent imprint on BRCA2-related breast tumors.Tumors arising in female and male individuals withdysfunctional BRCA2 do not differ from sporadicbreast carcinoma in steroid receptor content, and theyhave in many cases preserved a hormone-sensitivephenotype. BRCA1 tumors may constitute a hormone-independent and rapidly proliferating type of breastcarcinoma with an inherent genome instability thatcould make them susceptible to certain DNA damag-ing drugs and other cellular stresses. Obviously,BRCA1-related, BRCA2-related, and other hereditarytypes of breast carcinoma are different pathogenicand clinical entities that probably require distincttreatment and follow-up programs and possibly dif-ferent prophylactic measures. The high receptor levels

and hormone-responsive phenotype seen in some he-reditary non-BRCA1/BRCA2-related tumors may alsoindicate a distinct genetic origin and subgroup of he-reditary breast carcinoma. Data presented here couldsuggest a rationale for trying antiestrogens as chemo-prevention in the treatment of high risk individuals insuch families. This may also be considered for healthywomen carrying BRCA2 mutations.

In most populations throughout the world, a sig-nificant proportion of families who have a high risk fordeveloping breast carcinoma remain genetically un-characterized.13 Thus, it is of considerable interestthat, in tumors of which an unknown genetic factormight be the underlying cause, steroid receptor levelsare higher than in tumors associated with BRCA1 andBRCA2 mutations. The results of the current studysuggest that familial tumors with high levels of PgRmight compose a distinct subgroup of hereditarybreast carcinomas.

REFERENCES1. Claus E, Risch N, Thomson W. Genetic analysis of breast

cancer in the cancer and steroid hormone study. Am J HumGenet 1991;48:232– 42.

2. Newman B, Austin M, Lee M, King M. Inheritance of humanbreast cancer: evidence for autosomal dominant transmis-sion in high risk families. Proc Natl Acad Sci U S A 1988;85:3044 – 8.

3. Miki Y, Swensen J, Shattuck-Eidens D, Futreal PA, Harsh-man K, Tavtigian S, et al. A strong candidate for the breastand ovarian cancer susceptibility gene BRCA1. Science 1994;266:66 –71.

4. Wooster R, Bignell G, Lancaster J, Swift S, Seal S, Mangion J,et al. Identification of the breast cancer susceptibility geneBRCA2. Nature 1995;378:789 –92.

FIGURE 5. This dotplot shows the age distributions for male breast carci-

noma patients related and not related to BRCA2. Median values are marked in

the plot. The difference is not significant (N.S.). Log scale on y-axis.

FIGURE 6. This dotplot shows ER levels in males with breast carcinomas

related and not related to BRCA2. Median values are marked in the plot. The

difference is not significant. Log scale on y-axis.

FIGURE 7. Progesterone (PgR) levels are shown for male breast carcinoma

tumors related to and not related to BRCA2. Median values are marked in the

plot. The difference is not significant. Log scale on y-axis.


5. Tavtigian SV, Simmard J, Rommens J, Couch F, Shattuck-Eidens D, Neuhausen S, et al. The complete BRCA2 geneand mutations in chomosome 13q linked kindreds. NatGenet 1996;12:333–7.

6. Stratton MR. Recent advances in understanding of geneticsusceptibility to breast cancer. Hum Mol Genet 1996;5:1515–9.

7. Wooster R, Neuhausen SL, Mangion J, Quirk Y, Ford D,Collins N, et al. Localization of a breast cancer susceptibilitygene, BRCA2, to chromosome 13q12–13. Science 1994;265:2088 –90.

8. Berman DB, Costalas J, Schultz DC, Grana G, Mary D, God-win AK. A common mutation in BRCA2 that predisposes toa variety of cancers is found in both Jewish Ashkenazi andnon-Jewish individuals. Cancer Res 1996;56:3409 –14.

9. Friedman LS, Gayther SA, Kurosaki T, Gordon D, Noble B,Casey G, et al. Mutation analysis of BRCA1 and BRCA2 in amale breast cancer population. Am J Hum Genet 1997;60:313–9.

10. Goldgar DE, Neuhausen SL, Steele L, Fields P, Ward JH, TranT, et al. A 45-year follow-up of kindred 107 and the searchfor BRCA2. J Natl Cancer Inst Monogr 1995;17:15–9.

11. Johannesdottir G, Gudmundsson J, Bergthorsson JT, ArasonA, Agnarsson BA, Eiriksdottir G, et al. High prevalence of the999del5 mutation in Islandic breast and ovarian cancer pa-tients. Cancer Res 1996;56:3663–5.

12. Struewing JP, Hartge P, Wackolder S, Baker SM, Berlin M,McAdams M, et al. The risk of cancer associated with spe-cific mutations of BRCA1 and BRCA2 among AshkenaziJews. N Engl J Med 1997;336:1401– 8.

13. Szabo CI, King M-C. Population genetics of BRCA1 andBRCA2. Am J Hum Genet 1997;60:1013–20.

14. Malkin D, Li F, Strong L, Fraumeni J, Nelson C, Kim D, et al.Germ line p53 mutations in a familial syndrome of breastcancer, sarcomas, and other neoplasms. Science 1990;250:1233– 8.

15. Li J, Yen C, Liaw D, Podyspanina K, Bose S, Wang SI, et al.PTEN, a putative protein tysosine phosphatase gene mu-tated in human brain, breast and prostate cancer. Science1997;275:1943–7.

16. Steck PA, Pershouse MA, Jasser SA, Yung WKA, Lin H, LogonAH, et al. Identification of a candidate tumour suppressorgene, MMAC1, at chromosome 10q23.3 that is mutated inmultiple advanced cancers. Nat Genet 1997;16:356 – 62.

17. Hemmiki A, Tomlinson I, Markie D, Jarvinen H, Sistonen P,Bjorkqvist A-M, et al. Localization of a susceptibility locusfor Peutz–Jeghers syndrome using comparative genomic hy-bridization and targeted linkage analysis. Nat Genet 1997;15:87–90.

18. Håkansson S, Johannsson O, Johansson U, Sellberg G, Lo-man N, Gerdes A-M, et al. Moderate frequency of BRCA1and BRCA2 germ-line mutation in Scandinavian familialbreast cancer. Am J Hum Genet 1997;60:1068 –78.

19. Scully R, Chen J, Plug A, Xiao Y, Weaver D, Feunteun J, et al.Association of BRCA1 with Rad51 in miotic and meioticcells. Cell 1997;88:265–75.

20. Scully R, Anderson SF, Chao DM, Wei W, Ye L, Young RA, etal. BRCA1 is a component of the RNA polymerase II holoen-zyme. Proc Natl Acad Sci U S A 1997;94:5605–10.

21. Sharan SK, Morimatsu M, Albrechts U, Lim D-S, Regel E,Dinh C, et al. Embryonic lethality and radiation hypersen-sitivity mediated by Rad51 in mice lacking BRCA2. Nature1997;386:304 –9.

22. Rajan J, Marquis S, Gardner H, Chodosh L. Developmental

expression of BRCA2 colocalizes with BRCA1 and is associ-ated with proliferation and differentiation in multiple tis-sues. Dev Biol 1997;184:385– 401.

23. Tirkkonen M, Johannsson O, Agnarsson BA, Olsson H, Ing-varsson S, Karhu R, et al. Distinct somatic genetic changesassociated with tumor progression in carriers of BRCA1 andBRCA2 germ-line mutations. Cancer Res 1997;57:1222–7.

24. Johannsson OT, Idvall I, Anderson C, Borg Å, Egilsson V,Olsson H. Tumour biological features of BRCA1-inducedbreast and ovarian cancer. Eur J Cancer 1997;33:362–71.

25. Breast Cancer Linkage Consortium. Pathology of familialbreast cancer: differences between breast cancers in carriersof BRCA1 or BRCA2 mutations and sporadic cases. Lancet1997;349:1505–10.

26. Habel LA, Stanford JL. Hormone receptors and breast can-cer. Epidemiol Rev 1993;15:209 –19.

27. Ferno M, Borg Å, Johansson U, Norgren A, Olsson H, RydenS, et al. Estrogen and progesterone receptor analyses inmore than 4000 human breast cancer samples. Acta Oncol1990;29:129 –35.

28. Hecht J, Winchester D. Male breast cancer. Am J Clin Pathol1994;102:25–30.

29. Karp SE, Tonin PN, Begin LR, Martinez JJ, Zhang JC, PollakMN, et al. Influence of BRCA1 mutations on nuclear gradeand estrogen receptor status of breast carcinoma in Ash-kenazi women. Cancer 1997;80:435– 41.

30. Ferno M, Borg Å, Sellberg G. Enzyme immunoassay of theestrogen receptor in breast cancer biopsy samples: a com-parison with isoelectric focusing. Acta Radiol Oncol 1986;25:171–5.

31. Norgren A, Borg Å, Ferno M, Johansson U, Lindahl B, Tsio-banelis K. Improved method for assay of estradiol and pro-gesterone receptors with special reference to breast cancer.Anticancer Res 1982;2:315–20.

32. Ferno M, Borg Å, Johansson U. Enzyme immunoassay ofprogesterone receptor in breast cancer biopsy samples. ActaOncol 1989;28:19 –22.

33. Ferno M, Borg Å, Norgren A. A comparison of two steroidreceptor assays in breast cancer: dextran coated charcoaland isoelectric focusing. Anticancer Res 1983;3:243– 6.

34. Ferno M, Andersson C, Fallenius G, Idvall I. Oestrogen re-ceptor analysis of paraffin sections and cytosol samples ofprimary breast cancer in relation to outcome after adjuvanttamoxifen treatment. Acta Oncol 1996;35:17–22.

35. StataCorp. Stata Statistical Software. Version 5.0. CollegeStation, TX: Stata Corporation, 1997.

36. Marcus JN, Watson P, Page DL, Narod SA, Lenoir GM, ToninP, et al. Hereditary breast cancer pathobiology, prognosis,and BRCA1 and BRCA2 gene linkage. Cancer 1996;77:697–709.

37. Tsai M-J, O’Malley BO. Molecular mechanisms of action ofsteroid/thyroid receptor superfamily members. Annu RevBiochem 1994;63:451– 86.

38. Clark GM, McGuire WL. Steroid receptors and other prog-nostic factors in primary breast cancer. Semin Oncol 1988;15:20 –5.

39. Eisinger F, Stoppa-Lyonnet D, Longy M, Kerangueven F,Noguchi T, Bailly C, et al. Germ line mutation at BRCA1affects the histoprognostic grade in hereditary breast can-cer. Cancer Res 1996;56:471– 4.

40. Rubin SC, Benjamin I, Behbakht K, Takahashi H, MorganMA, LiVolsi VA, et al. Clinical and pathological features ofovarian cancer in women with germ-line mutations ofBRCA1. N Engl J Med 1996;335:1413– 6.


41. Kinzler KW, Vogelstein B. Lessons from hereditary colorectalcancer. Cell 1996;87:159 –70.

42. Johannsson OT, Ranstam J, Borg Å, Olsson H. Survival ofBRCA1 breast and ovarian cancer patients: a population-based study from southern Sweden. J Clin Oncol 1998;16:397– 404.

43. Brugarolas J, Jacks T. Double indemnity: p53, BRCA andcancer. Nat Med 1997;3:721–2.

44. Chapman MS, Verma IM. Transcriptional activation byBRCA1. Nature 1996;382:678 –9.

45. Milner J, Ponder B, Hughes-Davies L, Seltmann M, Kouzar-ides T. Transcriptional activation functions in BRCA2. Na-ture 1997;386:772–3.

46. Spillman MA, Bowcock AM. BRCA1 and BRCA2 mRNA

levels are coordinately elevated in human breast cancercells in response to estrogen. Oncogene 1996;13:1639 –45.

47. Vaughn JP, Cirisano FD, Huper G, Berchuck A, Futreal PA,Marks JR, et al. Cell cycle control of BRCA2. Cancer Res1996;56:4590 – 4.

48. Marks JR, Hupers G, Vaughn JP, Davies PL, Norris J, Mc-Donnel DP, et al. BRCA1 expression is not directly respon-sive to estrogen. Oncogene 1997;14:115–21.

49. Olsson H, Sigurdsson H, Borg Å, Ferno M. Relationship ofprogesterone-receptor positivity in malignant breast tumorsto reproductive status of women at tumor initiation: resultsfrom patients with possible radiation-induced tumors.J Natl Cancer Inst 1990;82:529 –31.


Steroid receptors in hereditary breast carcinomas associated with BRCA1 or BRCA2 mutations or unknown susceptibility genes

Documents