JPM, Vol. 12, Pages 1988: Correlation between Androgen Receptor Expression in Luminal B (HER–2 Negative) Breast Cancer and Disease Outcomes

1. IntroductionHormone receptor (HR)–positive breast cancers account for about 75% of all breast cancers [1,2,3]. They can be classified as luminal A or B (HER–2 negative) based on their expression of the proliferation index Ki67 and progesterone receptor. Luminal A breast cancers respond well to endocrine treatment efficacy and have good prognosis. Luminal B (HER–2 negative) breast cancers have lower levels of HR expression and higher proliferation indexes than luminal A as well as a higher risk of recurrence and metastasis. Based on these characteristics, the comprehensive treatment plan for luminal B (HER–2 negative) breast cancers includes surgery, radiotherapy, endocrine therapy, and chemotherapy. Most luminal B breast cancers are sensitive to endocrine therapy; however, some patients develop primary or secondary resistance to endocrine therapy. Currently, the prediction and reversal of endocrine resistance and the search for novel therapeutic agents are popular research topics. Some luminal B breast cancers are less sensitive to chemotherapy and have a poor prognosis even with chemotherapy. There is a lack of sensitive indicators to predict sensitivity to chemotherapy and whether patients can benefit from specific chemotherapy regimens. This suggests that large–scale studies are needed to review the factors associated with the treatment and outcomes of luminal B (HER–2 negative) breast cancers and to search for novel predictors.

The systemic treatment of breast cancer can be classified as neoadjuvant or adjuvant therapy, depending on its timing. Neoadjuvant therapy plays a vital role in the treatment of breast cancer, not only to reduce the extent of surgery, and obtain information on treatment responsiveness and prognosis, but also to adjust subsequent treatments based on the treatment response to improve the survival of patients in non–pathological complete response (non–pCR) patients. In HER–2–overexpressing breast cancer and triple–negative breast cancer (TNBC) patients, the rate of pCR can reach 40–50% with neoadjuvant treatment, and the long–term survival of patients achieving pCR is significantly better. Luminal B (HER–2 negative) breast cancer lacks HER–2 targets, and the current standard neoadjuvant treatment regimen is chemotherapy; neoadjuvant endocrine therapy is still being investigated. Luminal B (HER–2 negative) breast cancer has poor sensitivity to neoadjuvant therapy, with a pCR rate of 5–10%, and the predictive value of pCR for long–term prognosis is not as high as that for HER–2 positive breast cancer and TNBC. Therefore, there is a need to find novel indicators to predict the efficacy of neoadjuvant therapy and outcomes for luminal B (HER–2 negative) breast cancers. The literature suggests that Ki–67 can be used to evaluate the efficacy of neoadjuvant endocrine therapy in patients with luminal B (HER–2 negative) breast cancer. However, the cutoff values and predictive sensitivities in different reports vary widely, and its predictive ability lacks validation in neoadjuvant chemotherapy.

The androgen receptor (AR) is widespread across various breast cancer subtypes, and over 70% of HR–positive breast cancers express AR. Most retrospective studies have found that positive AR expression is associated with a better long–term prognosis of breast cancer [4,5]. However, the correlation between its expression and the efficacy of neoadjuvant therapy is inconsistent across breast cancer subtypes. In TNBC, AR positivity is associated with a poor response to neoadjuvant therapy, whereas in HER–2–positive breast cancer, AR expression suggests a good response to neoadjuvant therapy. Studies on the correlation between AR and prognosis and its function in luminal B (HER–2 negative) breast cancer were contradictory. AR may play an oncogenic role, as AR agonists have the potential to reverse endocrine resistance [6]. However, high AR expression may be associated with endocrine resistance, as patients with ER–positive breast cancer with high ratios of AR expression to estrogen receptor (ER) expression (AR/ER) have poorer disease–free survival (DFS) [7]. In these studies, the stage and status of cancer differ, and the ranges of AR, AR/ER, and ratio of AR expression to PR expression (AR/PR) differ as well, which could lead to opposite conclusions. In HER2–positive breast cancer, the expression of AR is elevated, regardless of the status of hormone receptor. It may be related to the positive feedback cycle between AR and HER2. The relationship between AR expression and prognosis in luminal B HER–2 positive breast cancer is not clear, and the conclusions of different studies are inconsistent [8,9]. Therefore, a larger sample size and more in–depth mechanistic studies are needed to investigate the effects of AR on the development and prognosis of breast cancer.

In this single–center study, cases of luminal B (HER–2 negative) breast cancer at the Peking University First Hospital were retrospectively analyzed to investigate the correlation between AR and survival and neoadjuvant treatment outcomes in luminal B (HER–2 negative) breast cancer, and to determine AR cutoff values with clinical value.

4. DiscussionThe 2011 St. Gallen International Expert Consensus proposed classifying breast cancers based on molecular expression markers. Those with positive ER expression and low PR expression or high Ki67 expression were defined as luminal B breast cancer, which accounted for approximately 50–60% of all breast cancers. The systemic treatment of luminal B (HER–2 negative) breast cancer primarily involves endocrine therapy and chemotherapy. The 8–year DFS of patients with luminal B (HER–2 negative) breast cancer is 78.9% [18]. Some patients have a poor response to neoadjuvant chemotherapy and develop recurrence after surgery. The pCR rate of neoadjuvant therapy for this subtype is low, and the relationship with long–term survival is unclear; therefore, large–scale studies are needed to investigate the factors associated with response and prognosis and find sensitive predictors. In this single–centered retrospective study, we evaluated the treatment efficacy and prognosis of patients with luminal B (HER–2 negative) breast cancer in our hospital, investigated the correlation between AR and its associated indicators with patient survival, and aimed to find a prognostic and predictive cutoff value.AR is a widespread molecular marker across various breast cancer subtypes, and approximately 70–90% of luminal–type breast cancers express AR [19,20]. In ER–negative, AR–positive breast cancers, activated androgen–AR complexes enter the nucleus and bind to androgen response elements in chromatin to induce tumor cell proliferation, whereas in ER–positive, AR–positive breast cancers, activated AR competes with ER to bind estrogen response elements in chromatin and induces apoptosis [21]. A retrospective study suggested that AR expression in luminal–type breast cancer is associated with a better prognosis [22].A total of 985 patients with luminal B (HER–2 negative) breast cancer were enrolled in this study, with a median follow–up of 42 months. There were a total of 83 DFS events and 29 OS events, and the 3–year DFS was 97.3%, 3–year OS was 98.7%, 5–year DFS was 93.5%, and 5–year OS was 97.4%. In the SOFT trial, the 5–year DFS of patients ranged 84.7–86.6% [23], and in the MINDACT study, the 5–year distant metastasis–free survival (DMFS) was 97.3% in low–risk patients and 90.6% in high–risk patients [24]. In this study, patient survival was similar to those in other studies.In this study, over 60% of patients had AR expression levels above 90%, and only 2.2% did not express AR. In the literature, the definition of negative and positive AR expression is not consistent, with 1%, 10%, or more, or the Allred score all being used. It is generally accepted that 70–90% of patients with luminal–type breast cancer are positive for AR expression. The high percentage reported in this study may be associated with the difference between Chinese and Western populations and using an AR cutoff value of 1%. The inconsistency of AR cutoff values in the literature affects the comparability of the study results. It informs the need for finding AR prognosis–related cutoff values. The literature reports that male breast cancer patients represent approximately 1% of the patient population, and almost all are ER+ AR+ [25]. A total of seven male breast cancer patients, none of whom underwent neoadjuvant therapy, were included in this study. All had ER expression ≥70% and AR expression ≥90%, which is consistent with the literature. We observed that the overall AR expression levels of patients undergoing neoadjuvant therapy were lower than those not undergoing neoadjuvant therapy. Other characteristics such as age, menopausal status, tumor stage, histological grading, and molecular markers differed significantly, suggesting a significant difference between the neoadjuvant and non–neoadjuvant therapy populations and that the significance of AR may differ between them.Multifactorial analysis of survival in the overall population suggested that age above 40 years was a protective factor for DFS events. Late T stage, axillary lymph node metastasis, and high histological grade were risk factors for DFS events. In this study, over 60% of patients had at least 90% AR expression levels, and the median AR level of the overall population was 90%. Therefore, a preliminary analysis was performed using an AR expression level of 90% as a cutoff value, and it was observed that the AR expression level correlated with DFS and OS. Further analysis showed that when patients were grouped using an AR of 65% as a cutoff value, there were significant differences in DFS and OS between the two groups, and patients with AR ≤ 65% had significantly worse survival. Recciardelli et al. [26] reviewed 46 studies and included a training cohort (n = 219) and a validation cohort (n = 418) to determine AR expression cutoff values. Their findings revealed that although most studies selected 1% or 10% as cutoff values, these were not the best cutoff values to differentiate between patients with different outcomes, and they suggested that a positivity cutoff value of 78% had better sensitivity (57.1%) and specificity (75.0%). Patients with AR > 78% had better survival, and those with positive ERα and AR > 78% had the best outcomes. Yang et al. [27] studied 957 breast cancer patients and found that AR ≥ 35% was significantly associated with longer progression–free survival. The population in our study was larger and was a single luminal B (HER–2 negative) breast cancer population; however, the sensitivity of the optimal cutoff point was lower than that of the Recciardeli study. We used 10%, 30% and 78% of AR expression levels as positive cutoff values for survival analysis and tested them in this study cohort. It was found that when taking 10% for analysis, there was no statistical difference in DFS and OS between the two groups of patients in the total luminal B HER2 negative population. When using other AR expression levels as the cutoff value, the populations with AR > 78% or AR > 30% had better DFS and OS, and the results were statistically different. The literature and the present study suggest that AR, unlike ER and PR, should not have a cutoff value of 1% or 10%; however, it should be increased to at least 30%. More studies are still needed to determine the optimal cutoff value. In addition to the heterogeneity of the study population, it is difficult to determine the cutoff value of AR expression level in prognosis prediction. One of the potential factors is that AR may have different biological functions and signal pathways in breast cancer.In the population of patients who underwent neoadjuvant therapy, 37 DFS events occurred, with a relapse rate of 25.9%. pCR was achieved in only seven individuals in this study, for a pCR rate of 4.9%. Prognostic analysis grouped by pCR or non–pCR suggested that pCR did not correlate with long–term survival. Therefore, further investigation of prognostic correlates in the neoadjuvant therapy population is needed. Currently, there have been few studies on AR in neoadjuvant therapy for breast cancer, and some studies have suggested that AR positivity may be associated with poor neoadjuvant outcomes in TNBC. Witzel et al. [28] reviewed AR expression in the TECHNO and PREPARE clinical trials. Their findings showed that high AR mRNA levels were generally common among HER–2–overexpressing breast cancers and luminal–type breast cancers but extremely rare in TNBC. High AR mRNA levels were associated with low pCR rates but better DFS and OS. The present study did not find any correlation between AR and prognosis in the neoadjuvant therapy population; hence, further discovery of other relevant factors and predictors is needed.

The present study investigated the prognostic, predictive value of AR/ER and AR/PR. Most patients had high AR and ER expression, with a median and mode AR/ER of 1.00. Overall, 14.7% of patients had AR/ER > 1.00, and 31 patients had AR/ER ≥ 2.0, accounting for 3.1% of the study population. There were no statistically significant differences in AR/ER and AR/PR between the neoadjuvant and non–neoadjuvant therapy populations. In the neoadjuvant therapy population, the risk of recurrence was 5.381 times higher in patients with high AR/ER than in those with low AR/ER (95% CI: 1.264–22.91; p = 0.023). Further investigation of the cutoff value showed that the optimal AR/ER cutoff value was 1.06, and patients with AR/ER >1.06 accounted for 15.4% of the population who underwent neoadjuvant therapy. Kaplan–Meier analysis suggested that patients with high AR/ER had significantly worse DFS and no significant difference in OS when classified using an AR/ER cutoff value of 1.06. However, the follow–up period of this study was short, and the OS results were incomplete.

The literature shows that AR/ER and AR/PR correlate with endocrine resistance and survival. The study by Cochrane et al. [9] included 192 patients with ER–positive breast cancer who underwent postoperative adjuvant tamoxifen therapy between 1977 and 1993, of whom 48 patients (25%) experienced tamoxifen treatment failure. A high AR/ER ratio (≥2.0) was found in 22 patients, which predicted a fourfold increased risk of tamoxifen treatment failure (HR = 4.43). In addition, this study found that AR/ER was an independent predictor of DFS (HR = 4.04, 95% CI: 1.68–9.69; p = 0.002) and disease–specific survival (HR = 2.75, 95% CI: 1.11–6.86; p = 0.03). In a study by Rangel et al. [29], 402 patients with ER–positive breast cancer underwent AR, ER, PR, HER–2, and Ki67 testing. ROC curve analysis showed that AR/ER ≥ 2.0 was the cutoff value for distinguishing prognosis and risk factors in patients. A total of 19 patients (6%) had AR/ER ≥ 2.0. These patients exhibited more axillary lymph node metastasis, higher histological grade, lower PR expression levels, and significantly worse disease–free survival and disease–specific survival. Prosigna–PAM50 analysis showed that 63% of patients with high AR/ER (12/19) were categorized with a high risk of recurrence, and 47.4% (9/19) were classified as non–luminal–type. Subsequently, Rangel et al. [30] performed IHC and mRNA analysis to determine AR/ER levels in 47 patients with ER–positive breast cancer and validate them with a public database of 979 patients. They found that ER–positive breast cancer patients with AR/ER ≥ 2 had higher cell proliferation gene expression levels. Most of these patients were of luminal B and HER–2 enriched subtypes, which exhibited a higher degree of proliferation and poorer prognosis. Bronte et al. [31] reviewed a total of 159 ER–positive breast cancer patients between 2000 and 2008 with a median follow–up of 63 months, of whom 89 had data on AR, ER, and survival, and 24 patients with metastasis had pathological findings for primary and metastatic foci. They found that patients with AR/PR ≥ 1.54 in the primary focus had significantly worse OS. In addition, they compared AR/ER in primary and metastatic foci in metastatic patients and found that patients with metastatic foci with AR/ER ≥ 0.90 had better OS, and patients had significantly better OS when AR/ER was maintained at high levels in both primary and metastatic foci. Rajarajan et al. [32] reviewed 270 breast cancer patients with a median follow–up of 72 months. They found that patients with high AR/ER (>1.75, third quartile) had higher circulating testosterone levels and significantly worse DFS, which was a result validated in the Cancer Genome Atlas (TCGA) and the Molecular Taxonomy of Breast Cancer International Consortium (METABRIC) databases. We tested the predictive value of prognosis when the cutoff value of AR/ER was 2.0 and 0.9 in the cohort of our study, but the results were not statistically different.In luminal–type breast cancers, high AR expression is associated with a good prognosis, and low ER expression is associated with a poor prognosis. Therefore, further investigation is required to determine whether AR/ER is an indirect manifestation of low ER expression. The role of AR in HR–positive breast cancer is controversial, particularly with regard to the relationship between AR and resistance to endocrine therapy. Cochrane et al. [9] found that a high AR/ER ratio was associated with an increased risk of tamoxifen resistance, which was independent of the effects of AR and ER expression levels alone. In MCF7 (ER+ AR+) xenograft tumors, enzalutamide blocked E2–driven tumor proliferation as strongly as tamoxifen, suggesting that high AR/ER may not simply result from low ER expression, but that AR itself does play a role in tumor cell proliferation. The mechanism of endocrine resistance in breast cancer may be the conversion of tumor cells from being estrogen–dependent to androgen–dependent. The mechanism of aromatase inhibitors is to block the conversion of precursors to estrogen and progesterone, causing them to be converted to androgens instead. High androgen levels in patients activate the AR pathway in tumor cells and manifest as endocrine resistance. Indeed, increased anastrozole resistance was found in AR–overexpressing breast cancer patients. In contrast, the AR antagonists bicalutamide and enzalutamide could restore sensitivity to endocrine therapy [33]. In addition to the classic AR signaling pathway, AR also mediates endocrine resistance through other pathways. After knocking down AR in tamoxifen resistant breast cancer cells, tamoxifen sensitivity was restored, which might be the result of upregulation of the estrogen related classical signaling pathway. However, the AR blocker enzalutamide could not completely reproduce the effect of AR knockdown, and the drug could not affect the growth and ER expression of tamoxifen–resistant cells and endocrine–resistant xenograft tumor models [34]. In addition, the interaction between ER and AR has been associated with PR. In ER–positive breast cancer, PR, ERα, and various cofactors interact to form a complex. When ER and PR are activated, high ERβ expression is associated with low invasiveness. This may be because ERα and ERβ form a heterodimer, which reduces the recruitment of regulatory cofactors and thus decreases gene transcription. In ER+ PR+ breast cancers, AR dimers translocate to the nucleus, compete with ERα and PR to bind estrogen–responsive elements, and block ER–mediated signaling pathways. In contrast, in ER+ PR− breast cancers, ERβ acts by downregulating ERα target gene transcription. When PR is absent, AR enhances the effect of ERα gene transcription, producing an oncogenic effect. Clinically, a high AR pathway activity does correlate significantly with worse DFS in patients on endocrine therapy [35]. However, a study by Hickey et al. [8] revealed a contrasting conclusion, proposing that AR exerts an inhibitory effect on the proliferation of HR–positive breast cancer and that AR agonists, rather than antagonists, are potentially effective therapeutic agents for ER–positive breast cancer. The study confirmed that AR competes with ER to bind cofactors and responsive elements in chromatin and block ER–mediated tumor proliferative effects. Thereafter, they established ZR–75–1 xenografts in mice and found that the AR antagonist enzalutamide could only inhibit estradiol (E2)–stimulated tumor proliferation in the short term, whereas the AR agonist dihydrotestosterone (DHT) and the selective androgen receptor modulators (SARM) enobosarm resulted in lasting inhibition of tumor growth, even for over 90 days. In both tamoxifen–resistant ER+ MCF7 xenografts and palbociclib–resistant ER+ MCF7 xenografts, significant tumor suppression was achieved with a combination of CDK4/6 inhibitors and DHT. There may be some heterogeneity in endocrine–resistant breast cancers, and the role of AR remains to be elucidated through in–depth studies.

Ki67 expression levels are predictive of breast cancer prognosis, with higher Ki67 expression levels indicating worse survival. In hormone receptor–positive breast cancer, Ki67 is an important molecular marker to distinguish between luminal A and B breast cancers.

A multifactorial survival analysis of non–pCR patients who underwent neoadjuvant therapy revealed that high residual tumor Ki67 expression was a risk factor for recurrence and metastasis, whereas changes in residual tumor ER, PR, and Ki67 before and after neoadjuvant therapy were not correlated with prognosis. Using the ROC curve, we found that the optimal residual tumor Ki67 cutoff value for distinguishing prognosis was 23%. Non–pCR patients with residual tumor Ki67 36]. Among non–pCR patients in the GeparTrio trial, patients with high residual tumor Ki67 had a higher risk of recurrence and death, whereas the survival of patients with low Ki67 did not differ significantly from that of pCR patients [37]. Due to the low pCR rate of neoadjuvant therapy for HR–positive breast cancer, some researchers have proposed using residual tumor Ki67 level instead of pCR as the index for evaluating response to neoadjuvant endocrine therapy. The preoperative endocrine prognostic index (PEPI) score was derived from the P024 neoadjuvant endocrine therapy clinical trial.. It includes residual tumor size, axillary lymph node status, and Ki67 and ER expression levels. Among them, the cutoff values for Ki67 were 2.7%, 7.3%, 19.7%, and 63.1% [38]. In the ACOSOG Z1031A trial [39], patients with stage II–III ER–positive breast cancer underwent core needle biopsy after 2–4 weeks of neoadjuvant endocrine therapy. Those with Ki67 > 10% were converted to neoadjuvant chemotherapy or immediate surgery, and those with Ki67 0 had a relapse event. In the FELINE neoadjuvant endocrine therapy trial [40], PEPI = 0 was the primary study endpoint, with complete cell cycle arrest (Ki67 41].

The advantage of the present study is that it is an early and large population–based, single–center study of the relationship between AR and related indicators and survival among the luminal B breast cancer population in China, and the number of included cases is larger than most foreign studies. This is the first study of the correlation between AR/ER and AR/PR with survival in China. The number of cases far exceeds that of foreign studies related to AR/ER. Moreover, this is the first study of neoadjuvant therapy and AR, AR/ER, and AR/PR in China, and it is also a less studied and novel research topic internationally. The limitations of the study are that it is a single–center, retrospective study that lacks further testing of gene transcription levels; therefore, its results should be validated in a large–sample, prospective study, and gene–level and translational studies are needed to investigate the mechanism.

留言 (0)

沒有登入
gif