• anastrozole;
  • breast neoplasms;
  • tamoxifen;
  • Wnt proteins


  1. Top of page

Background and aim: Identifying potential predictive biomarkers of response to endocrine therapy would benefit most estrogen receptor-positive (ER+) breast cancer patients. Our aim was to compare the expression of the wingless type (Wnt) pathway-related proteins (adenomatous polyposis coli [APC], E-cadherin, beta-catenin, cyclin D1, and c-myc) in postmenopausal women with ER+ invasive ductal carcinomas (IDC), prior to and after tamoxifen (n= 18) or anastrozole (n= 15) treatment, in a double-blind, placebo-controlled (n= 25), prospective study for 26 days prior to surgery. Methods: Tissue microarray blocks were constructed from pre- and post-treatment biopsy samples. Nuclear immunostaining of c-myc, APC, estrogen and progesterone receptor levels were assessed using the Allred scoring system, and cytoplasmic immunostaining of cyclin D1, beta-catenin and E-cadherin was assessed with the Hercep-Test system. An anova statistical analysis estimated general equations and analysis of variance with a significance level of 0.05. Results: Tamoxifen increased c-myc (P= 0.0061) and APC (P= 0.0452) expression. Anastrozole did not significantly affect expression of any Wnt-related proteins. Conclusions: In post-menopausal ER+ IDC, tamoxifen for 26 days prior to IDC surgery influenced statistically the expression of important cell cycle regulators as APC and c-myc, whereas anastrozole therapy did not interfere with this pathway during the same period. These Wnt related proteins may contribute to selective estrogen receptor modulator resistance.


  1. Top of page

Although breast carcinomas comprise a heterogeneous disease,1 one classic characteristic of more than 60% of invasive ductal carcinomas (IDC) is the overexpression of the estrogen receptor (ER), known as the “luminal” molecular subtype in more recently developed molecular approaches.2–4 Therefore the anti-estrogen approach with adjuvant tamoxifen and anastrozole therapies remains an important therapeutic tool in most breast cancer cases.5 Tamoxifen is a selective estrogen receptor modulator that blocks the ER and stops ER-modulated breast cancer growth. Briefly, by occupying the ER ligand-binding region, tamoxifen alters the receptor complex conformation and inhibits signal transduction cascades that stimulate cell replication.6 Conversely, anastrozole, an aromatase inhibitor, blocks estrogen synthesis in peripheral tissues. Anastrozole does not have the estrogen-like side effects of tamoxifen.7,8

Although adjuvant therapy with tamoxifen for 5 years reduces recurrence rates by almost 50%, one-third of these patients will present with disease recurrence within 15 years.9 Tamoxifen-resistant disease may represent up to 25% of all breast carcinomas.10 Aromatase inhibitors became another therapeutic option for ER+ breast cancer and show a clear survival benefit in women with advanced (metastatic) breast cancer compared with endocrine therapy, as shown by a recent meta-analysis,8 albeit not free of the adverse effects related to estrogen modulation.7

An important challenge for successful breast cancer treatment is the identification of more specific biomarkers that could predict the therapeutic response to tamoxifen or anastrozole therapies,11 e.g., molecularly characterizing individual tumors, which could determine their metastatic potential and sensitivity to therapeutic agents.10 The wingless type (Wnt) signaling pathway is important in cell differentiation and proliferation, cell movement, and polarity12,13 Defects in this pathway are implicated in the pathogenesis of several tumor types, including breast cancer.13,14 Unlike colorectal cancers, genetic mutations in Wnt pathway components in breast cancer are seldom described.15–17 However, evidence suggests that this pathway is disregulated in breast cancer,13,18–20 probably through epigenetic mechanisms. For instance, with regard to the adenomatous polyposis coli (APC) gene, which is classically involved in adenomatous polyposis syndrome and colorectal cancer,15 genetic disregulation within the classical mutation regions (exon 15, codons 1,286 to 1,513) are rare in breast cancer,14,21 but over 18% of IDC showed somatic mutations outside this cluster region.21

In 2009, Mastroianni et al.22 showed that ectopic Wnt signaling drives ductal cell growth in the absence of estrogen in transgenic mice. These authors proposed that ectopic Wnt signaling could functionally substitute for estrogen-dependent growth stimuli, producing estrogen-independent tumors.22 Therefore, it is possible that, in ER+ human breast cancers, aberrant Wnt signaling could be an important factor promoting tumor growth.23

Moreover, some of the mitogenic cell cycle regulators strongly influenced by the Wnt pathway, such as c-myc and cyclin D1, seem to interfere with hormone therapy. c-Myc and cyclin D1 overexpression can potentially affect anti-estrogen therapy sensitivity at several levels. Accumulating evidence indicates that overexpression of c-myc or cyclin D1 is associated with tamoxifen resistance in humans,24,25 and molecular signatures predominantly composed of c-myc-responsive or retinoblastoma- and E2F-responsive genes are associated with poor outcomes in women treated with tamoxifen.26,27

To identify a possible correlation between diverse hormone therapy effects and the Wnt-associated protein profile, selected Wnt biomarkers were immunohistochemically assessed before and after short-term primary hormone (anastrozole or tamoxifen) therapy.

The objective of this study was to compare the expression of APC, E-cadherin, beta-catenin, cyclin D1, and c-myc in tumors from postmenopausal women with ER+ IDC prior to and after tamoxifen or anastrozole neoadjuvant treatment in a double-blind, placebo-controlled, prospective study. The hypothesis in this study was that tamoxifen, which blocks tumor ER proliferation stimuli, would result in higher expression of Wnt-related proteins.


  1. Top of page

Study design and patients

We designed a randomized, placebo-controlled, double-blind, diagnostic study including postmenopausal women with IDC. The study was approved by the Human Investigation Committee of Universidade Federal de São Paulo (UNIFESP), Brazil, under the process number CEP 1153/05, and registered with, identification number NCT01016665. Written informed consent was obtained from all patients prior to their inclusion in the study, and anonymity was assured. All procedures were performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki.

Fifty-eight consecutive IDC patients were enrolled from two university public hospitals in São Paulo (Pérola Byington Hospital and São Paulo Hospital) between April 2005 and June 2008. All patients had ER+ and/or progesterone receptor (PR)+ tumors and were submitted to definitive surgical treatment (radical mastectomy or quadrantectomy with axillary dissection) a mean period of 26 days after the incisional biopsy, which confirmed the diagnosis.

The exclusion criteria were patients with endocrine disease, tumors negative for expression of both ER or PR, history of thromboembolism, and those who were currently receiving hormone therapy or who had been previously treated for breast cancer (surgery, radiation, and/or chemotherapy), had been pregnant in the past 12 months before diagnosis, did not comply with the prescribed medication regimen, or who postponed surgery.

The patients were randomized to tamoxifen, anastrozole, or placebo using sealed, opaque envelopes. All included patients received a daily dose (identical oral pills) of their prescribed medication for 26 consecutive days prior to surgery. The patients received a daily dose of one of the following: tamoxifen 22 mg (n= 15); anastrozole 1 mg (n= 18); or placebo (n= 25). Clinical and demographic data, such as weight, height, age, age at first menstrual period, age at menopause, number of children, and age at birth of the first child were recorded, as well as the tumor size. The data were expressed as means with standard deviations (SD) and ranges. Clinical staging and affected lymph nodes were described as frequencies.


The first tumor sample was obtained at the time of diagnosis by incisional biopsy performed at an outpatient facility, using local anesthetic without epinephrine (2% lidocaine). A second tumor specimen was obtained from all patients during definitive surgery under general anesthesia. Both tumor samples were processed using the paraffin-embedding technique described below.

Histology and tissue microarray construction

All samples were fixed in 10% neutral-buffered formalin, processed, and embedded in paraffin. Respective paired tumor blocks containing samples obtained from all patients prior to any of the three interventions and during definitive surgery were retrieved from the pathology files of our institution.

Specimen pairs were cut into 4-μm sections, mounted on lysine-coated slides, stained with hematoxylin and eosin, and examined to confirm the diagnosis of carcinoma according to the World Health Organization (WHO) criteria (1981). The same slides were used to determine the area of interest (the most representative of the tumor) to be included in the tissue microarray (TMA) marked on the slide. Using a marking pen, the corresponding region was circled on the archival “donor” paraffin block. Tumor TMA blocks were obtained by punching 2-nm tissue cores from each donor paraffin block. The samples were then arrayed onto a recipient blank block using a manual tissue arrayer (Beecher Instruments, Sun Praine, WI, USA). Control tissues were included in each of these paraffin blocks.

Imunohistochemistry assays

After construction, 3-μm tissue sections were cut and transferred to silanized slides then left to dry overnight at 56°C. The next day, the slides were dewaxed in xylene, rehydrated in graded alcohol solutions, and washed with water. Antigen retrieval was performed using a pressure cooker (Eterna; Nigro, São Paulo, Brazil) and 10 mM citrate buffer, pH 6.0. Samples were quenched with 6% hydrogen peroxide and incubated overnight at 4°C with various primary antibodies for immunohistochemical analysis. The following day, the slides were rinsed with phosphate-buffered saline (PBS) and incubated with the secondary antibody (biotinylated goat anti-mouse/rabbit immunoglobulin) diluted 1:200, for 30 min at 37°C. The slides were rinsed again with PBS and incubated with streptavidin-biotinylated-peroxidase complex (1:200, Duet mouse/rabbit horseradish peroxidase, cat No. 0492; DakoCytomation, Carpinteria CA, USA), for 30 min at 37°C. The slides were developed with 0.06% Diaminobenzene (DAB) as chromogen with 0.06% hydrogen peroxide, and counterstained with Harris’ hematoxylin. Positive and negative control slides were included. Colon or breast carcinoma sections known to be positive for each marker were used as positive controls. Sections used as negative controls were incubated with PBS instead of primary antibody. All cells stained brown were considered positive. Table 1 lists the primary antibodies and dilution rates utilized.

Table 1.  Antibodies, clones, and antibody dilutions used in the immunohistochemical assay
AntibodiesClonesDilutionCat no./Source
  1. Cat no., catalog number; APC, adenomatous polyposis coli.

Estrogen receptorSP11:500RM9101, Neomarkers, Fremont, CA, USA
Progesterone receptorPgR6361:400M3569, Dako, Carpinteria, CA, USA
APC (C20)kRabbit polyclonal1:600sc 896, Santa Cruz, Santa Cruz, CA, USA
E-cadherin361:600C20820, BD Transduction, San Jose, CA, USA
Beta-catenin141:80019220, BD Transduction, San Jose, CA, USA
Cyclin D1Rabbit monoclonalReady for useBSB 5366, BioSB, Santa Barbara, CA, USA
c-Myc9E10.31:50MS139, Neomarker, Fremont, CA, USA

Immunohistochemical evaluation of Wnt-related proteins

Biomarker scoring

All slides were examined by two investigators (AFLW and YKJ) and scored semi-quantitatively. ER, PR, APC and c-myc expression levels were evaluated according to Allred's criteria using two parameters: the proportion of positive cells and the intensity of the staining.25 These parameters were used for each immunohistochemical reaction independently. The distribution of the proportional fraction of stained cells on each slide was scored using a scale from zero to five. The intensity of staining was scored from zero to three. The sum of these two partial scores resulted in a final score: zero indicated that no cells were stained, and scores ranging from two to eight indicated different levels of scoring. All cases with a final score equal to or greater than three were considered positive. E-cadherin and beta-catenin were scored with the HercepTest (DakoCytomation) score with score ranging from 0, 1 =+/+++, 2 =++/+++, and 3 =+++/+++. Scores of zero and one were considered negative and scores of 2 and 3 positive.

Statistical analysis

Descriptive statistics were used to summarize the sample characteristics at baseline (mean, SD, minimum, median, maximum, frequency, and percentage). The Kruskal–Wallis test was used to evaluate clinical characteristics.

Changes over time and differences between groups were evaluated with repeated-measures anova using rank transformation.

For the comparison of the positive and negative results between the times before and after treatment, the general estimating equations and anova were used.

The Student's t-test was used to compare mean scores for the control group with scores for the tamoxifen and anastrozole groups (negative expression, score = zero; positive expression, score ≠ zero). The same analysis was used to evaluate the mean scores.

Results from the immune reactions from the three groups were compared before and after treatment and also anastrozole and tamoxifen-combined treatment group versus placebo group. Differences in mean values in order to search for behavioral changes among the paired samples at the two distinct time points for each case were verified by bonferroni test and variability analysis test. The level of significance adopted in this study was 5% (P≤ 0.05).


  1. Top of page

Patient characteristics

Between April 2006 and June 2008, 58 final elected patients, 18, 15, 25 were allocated to the tamoxifen, anastrozole, and placebo (control group) groups, respectively.

The mean patient age was 66 years at diagnosis, and 48 years at menopause. The mean tumor size was 4 cm, and 76% of tumors were classified as grade II IDC. Most patients had stage II (stage IIA, 60%; stage IIB, 25%) carcinoma (14), lymph node status was free of disease in 39% (18), and affected lymph nodes were found in 41% ranging from one to more than 11 lymph nodes. Most tumors were ER-positive (98%) and PR-positive (69%) (Table 1). The Kruskal–Wallis test showed that there were no significant differences in clinical characteristics between groups (age, age at first period, age at menopause, number of children, age at birth of first child, or tumor size; all P > 0.10); therefore the sample was considered homogeneous.

Immunohistochemical expression of Wnt proteins

Wnt-related protein expression in the overall sample

The results for all proteins studied in samples obtained before treatment in all three groups were analyzed independently and no association was found (all P > 0.12), indicating that the groups were homogenous before treatment. Fig. 1 depicts the expression pattern of APC, cyclin D1, beta-catenin, E-cadherin and c-myc in our cases.


Figure 1. Expression of wingless type (Wnt)-related proteins in invasive ductal carcinomas. (a) Cyclin D1. (b) Adenomatous polyposis coli (APC). (c) Beta-catenin. (d) E-cadherin. (e) Negative control. (f) c-Myc expression. Central image is tissue microarray construct.

Download figure to PowerPoint

Frequency of E-cadherin expression did not change after treatment (46 and 44 positive cases, respectively). The same was observed with beta-catenin (44 positive patients at both time points). According to anova tests there was no association between E-cadherin or beta-catenin expression at the two time points (P= 0.3754 and P= 0.9550, respectively) in any of the three groups.

The expression of APC and cyclin D1 did change after treatment. 23% of samples APC-positive before treatment and 43% after treatment (P= 0.02); and 88% were cyclin D1-positive before and 64% after treatment (anova, P= 0.003, P= 0.0000). The frequency of expression of all proteins is shown in Table 2.

Table 2.  Immunohistochemical staining for E-caderin, APC, c-myc, beta-catenin, and cyclin D1 expression among tumors from 58 breast cancer patients according to time and group of treatment
Protein Pre-surgeryPost-surgeryTotal
  1. Values are presented as number (%).

  2. APC, adenomatous polyposis coli; A, anastrozole; P, placebo; T, tamoxifen.

E-caderinA2 (11)16 (89)4 (22)14 (78)18
P7 (28)18 (72)5 (20)20 (80)25
T3 (20)12 (80)5 (33)10 (67)15
Total12 (21)46 (79)14 (24)44 (76)58
APCA15 (83)3 (17)11 (61)79 (39)18
P19 (76)6 (24)18 (72)7 (28)25
T12 (80)3 (20)5 (33)10 (67)15
Total46 (79)12 (21)34 (59)24 (41)58
Cyclin DA2 (11)16 (89)9 (50)9 (50)18
P3 (12)22 (88)6 (24)19 (76)25
T2 (13)13 (87)6 (40)9 (60)15
Total7 (12)51 (88)21 (36)37 (64)58
c-MycA2 (11)16 (89)3 (17)15 (83)18
P8 (32)17 (68)3 (12)22 (88)25
T4 (27)11 (73)5 (33)10 (67)15
Total14 (24)44 (76)11 (19)47 (81)58
Beta-cateninA3 (17)15 (83)4 (22)14 (78)18
P7 (28)18 (72)6 (24)19 (76)25
T4 (27)11 (73)4 (27)11 (73)15
Total14 (24)44 (76)14 (24)44 (76)58

Some Wnt proteins expression correlated to others along time. Taken all cases together, c-myc expression correlated to E-cadherin (0.0213), APC (0.0117) and beta-catenin (0.0117) comparing the pre- and post-surgery samples results.

Wnt expression among treatment groups (Table 3)

Table 3.  Mean, standard deviation and median values of each Wnt-related protein evaluated according to placebo group (n= 25) and treated groups A/T (n= 33) and statistical relevance
ProteinTimeGroupMeanSDMedian P-value
  1. Wnt, wingless type; A, anastrozole; T, tamoxifen; SD, standard deviation; Pre, sample before treatment; P, placebo; Acc., according to; Post, samples after treatment.

E-cadherinPreP2.560.873Acc. group0.6689
 A/T2.730.723Acc. time0.7440
c-MycPreP1.761.162Acc. group0.5061
 A/T2.090.982Acc. time0.3159
PostP2.400.913Group vs time0.0330
APCPreP1.602.290Acc. group0.4825
 A/T1.211.960Acc. time0.0475
PostP1.441.960Group vs time0.0284
Beta-cateninPreP2.041.212Acc. group0.7424
 A/T2.121.143Acc. time0.4527
PostP2.160.992Group vs time0.9497
Cyclin DPreP5.322.256Acc. group0.4169
 A/T5.272.046Acc. time0.0001
PostP4.282.564Group vs time0.2738
Beta-catenin and E-cadherin

Results obtained from the placebo group were compared to the “combined treatment group” (anastrozole plus tamoxifen). The anova test results reflected a lack of association for E-cadherin (P= 0.1603) and beta-catenin (P= 0.9497) regardless of the group evaluated. The same was seen for beta-catenin in the anastrozole (P= 0.5016) and tamoxifen (P= 0.7774) groups when compared with placebo. anova evaluation confirmed the lack of significance for beta-catenin and E-cadherin for either the anastrozole or tamoxifen groups.

Cyclin D1 expression

Cyclin D1 was affected when the before and after treatment in the “combined treatment groups” was compared with placebo (P= 0.001), but there was no difference between groups during time after subsequent statistical analysis (P= 0.2738).

Myc expression

A difference was identified between the “combined treatment group” and the placebo group regarding c-myc expression after treatment (P= 0.0061). Additional anova tests were able to specify significance; positive values were significantly lower after tamoxifen treatment compared with placebo (P= 0.0117), but not for anastrozole compared with placebo (P= 0.2338). The Bonferroni test showed that after treatment, c-myc positivity was lower in the treatment versus the placebo group (P= 0.0317).

APC expression

After treatment with tamoxifen, APC expression was higher than in the pre-treatment samples (P < 0.02). Changes in APC protein expression did not reach significance after treatment with anastrozole (P= 0.2547), but when comparing tumors from tamoxifen-treated patients with those from placebo-treated patients, 20% had positive APC staining before receiving tamoxifen compared with 67% after treatment and this difference was significant (P= 0.0268) confirmed by Bonferroni test (P= 0.0370). However, no difference in APC expression was found after anastrozole treatment (P= 0.1829). anova confirmed a significant difference when time points and groups were compared (P= 0.0452). However, when tamoxifen was compared to the anastrozole group the difference was not significant.


  1. Top of page

Primary endocrine therapy is indicated for women with ER positive breast cancers who are unfit for or refuse surgery.28,29 A recent meta-analysis showed that, in women with advanced breast cancer, aromatase inhibitors are associated with survival benefits compared with other endocrine therapies (mainly tamoxifen).8 Patients treated with neoadjuvant endocrine therapy, whose tumors present with an Allred score ≥ 6 are more likely to respond and achieve a clinical benefit from neoadjuvant endocrine therapy.29–31

The possibility of a response to therapy is essential for treatment planning. Researchers are currently seeking markers of responsiveness to assist with treatment decision-making.11,32 The results of one treatment in terms of clinical outcome are certainly related to the patient's personal characteristics; i.e., her estrogen status and unique genetic profile. Since ER status is the currently established marker,32 we believe that subsequent prognosis prediction in ER+ breast cancer patients resides, in great part, in the expression of other proteins indirectly involved in a molecular relationship with the estrogen mechanism. In this study, we prospectively evaluated the influence of tamoxifen and anastrozole on the expression of Wnt pathway proteins theoretically implicated as being affected by estrogen. As expected, the placebo group had equivalent clinical features compared with the study groups and did not exhibit any statistically significant differences between the pre- and post-surgical samples. The molecular mechanism of antiestrogen drugs is not yet fully known, and another intention of this study was to see whether there was a relationship between in vivo exposure of tumor cells to tamoxifen and anastrozole and selected APC regulators.

There was no significant difference in cadherin expression in the biopsy samples from the placebo group compared with samples obtained during definitive surgery. Likewise, the comparison of the immunoreaction scores for beta-catenin between biopsy and surgical samples was not statistically significantly different. The absence of this variation was expected, since both proteins are known as adhesion molecules highly expressed in ductal carcinomas.33 This shows that any change in beta-catenin and in the Wnt pathway was not caused by the loss of cadherin (which is linked to beta-catenin).33–35 The classical ER signaling pathway does not appear to interfere with cadherin metabolism. Apparently, beta-catenin is also not influenced by blocking the ER or estradiol concentrations. These results are in accordance with data showing that the vast majority of ER-negative ductal carcinomas still express E-cadherin and many lobular E-cadherin-negative carcinomas are positive for the ER.17,36 We included the analysis of cadherin and beta-catenin because beta-catenin is highly related to the APC protein and, when it inadvertently migrates into the nucleus, may trigger proliferation-related genes such as cyclin D1 and c-myc.37 The stability of these adhesion molecules demonstrates that APC alteration does not occur via beta-catenin translocation, and may be an independent event.

Adenomatous polyposis coli expression was significantly reduced after treatment, especially treatment with tamoxifen. The APC protein is a member of the Wnt/beta-catenin signaling pathway, which is involved in the maintenance of the progenitor cell population in tissues like skin and intestine.37 Mutations and altered expression of the tumor suppressor gene APC are frequently found in breast carcinomas,12,14,15,35 with an emphasis on its potential role as a tumor suppressor in mammary cells. In mice, activation of Wnt/beta-catenin signaling contributes to tumorigenesis in the mammary epithelium, either through an Apc mutation or stabilization of beta-catenin. Mice heterozygous for germline mutations in Apc (ApcMin) spontaneously develop mammary tumors (although at a significantly lower incidence than intestinal tumors).38 Our results point out that the role of APC in early breast carcinogenesis may be underestimated, and APC is sensitive to alterations in ER metabolism.16 APC (+/1572T) mice (notwithstanding the constitutive nature of the mutation) have no predisposition toward intestinal cancer, but develop multifocal mammary adenocarcinomas and subsequent pulmonary metastases in both genders.38 Sequencing of the APC gene in cases of APC overexpression would be interesting, and an evaluation of APC protein in normal and pre-neoplastic lesions may indicate the role of this protein in ductal cell balance.

Although hormone therapy did not interfere with beta-catenin expression and its cellular localization remained in the cytoplasm, APC was importantly affected by hormone therapy (P < 0.003) and, remarkably, was unaffected and stable in the placebo group. It is interesting that augmentation was more evident in the tamoxifen group than in the anastrozole group (P= 0.03). The increase in APC staining may be explained by stable protein accumulating owing to overexpression, decreased degradation, or, more rarely, point mutations that could produce truncated or inactive proteins. Recently, some reports demonstrated the occurrence of APC gene hypermethylation which causes inactivation of the protein.35,39 Alterations in methylation of the APC gene were also accompanied by cyclin D1 overexpression.35 In our study, we demonstrated the presence of concomitant APC and cyclin D1 overexpression after hormone therapy, and one could speculate whether these agents might exert some effect on the methylation status of selected genes.

Nevertheless, APC is a crucial member of the Wnt pathway, and it is reasonable that c-myc and cyclin D1, which are Wnt target genes and also ER sensitive, would be affected by alterations in APC and hormone therapy. The apparently stronger influence of tamoxifen than anastrozole in APC and c-myc overexpression may mean that estradiol exerts some effect on Wnt in a predominantly receptor-dependent way.

Similarly, we expected to find some effect of Wnt imbalance on c-myc immunostaining. c-Myc is also modulated by the Wnt pathway and indirectly regulates the cell cycle machinery. Although c-myc expression was moderate in pretreatment samples, a significant upregulation became evident after detailed anova, and a much larger fraction of tumors exhibited overexpression of the c-myc protein.40 c-Myc expression in neoplastic breast lesions was addressed by our group previously, and we found a lower incidence of 66% c-myc positivity in IDC and 45% in ductal carcinoma in situ.41 The higher scores may be due to advanced IDC in contrast with diagnosed precursor lesions. However, it seems that c-myc alterations are an early and transient event in breast carcinogenesis because c-myc overexpression in the placebo group increased from 68% to 88%. We may infer from these results that tamoxifen treatment somehow attenuates c-myc expression. On the contrary, in a recent study by our group,42 estradiol, administered as contraceptive pills to patients with mammary fibroadenoma, did not affect c-myc expression. However, c-myc is considered a downstream effector of the Wnt pathway and can also activate the Wnt pathway in a positive feedback loop that is still being studied.37 As stated by Cowling and Cole,40 c-myc's role in breast cancer is an attractive model.

Cyclin D1 is particularly interesting in breast cancer, since Zwijsen et al.43 demonstrated that cyclin D1 may activate the ER and induce proliferation in hormone responsive tissues even in the absence of estrogen. Although cyclin D1 could have an agonist activity in ER+ cases, our results did not indicate a mandatory effect of tamoxifen or anastrozole in cyclin D1 expression. Further analysis of cyclin D1's contribution to tailor-made breast cancer therapy must be addressed.

A potential problem with using tamoxifen in breast cancer treatment is the long time period required to reach steady-state plasma levels (up to 5 weeks). In contrast, aromatase inhibitors build up rapidly, reaching therapeutic concentrations within days.29,44 We can speculate whether or not 26 days of tamoxifen would be enough to trigger any reduction in ER-related proteins. In this scope, longer periods of tamoxifen treatment would be required, but ethical issues impede such procedures since it would delay surgery. The hypothesis that the estrogen receptor, not simply estradiol, could be indirectly involved in Wnt protein expression was partially confirmed, suggesting that tamoxifen, by blocking the effects of estrogen, would increase Wnt protein expression and, moreover, that in a “blocked ER environment” a switch in the Wnt protein balance may occur.

In post-menopausal women with ER-positive IDC, tamoxifen therapy for 26 days prior to surgical treatment resulted in significant alteration in Wnt proteins, whereas therapy with anastrozole before did not. Wnt pathway alterations indicate indirect involvement of some of these proteins, notably c-myc and APC, in estrogen effect and may contribute, at least partially, to resistance. Further analysis of APC and other Wnt pathway components must be addressed in a larger clinical series.


  1. Top of page