Disclosure: The authors disclosed no potential conflicts of interest.
Immunohistochemical analysis of aromatase in metastatic lymph nodes of breast cancer
Article first published online: 16 DEC 2012
© 2012 The Authors. Pathology International © 2012 Japanese Society of Pathology and Wiley Publishing Asia Pty Ltd
Volume 63, Issue 1, pages 20–28, January 2013
How to Cite
Shibahara, Y., Miki, Y., Ishida, T., Nakamura, Y., Suzuki, T., Ohuchi, N. and Sasano, H. (2013), Immunohistochemical analysis of aromatase in metastatic lymph nodes of breast cancer. Pathology International, 63: 20–28. doi: 10.1111/pin.12015
- Issue published online: 29 JAN 2013
- Article first published online: 16 DEC 2012
- Manuscript Accepted: 8 NOV 2012
- Manuscript Received: 6 SEP 2012
- breast cancer;
- nodal metastasis
Aromatase is the key enzyme in intratumoral estrogen production in postmenopausal breast cancer and third generation aromatase inhibitors suppress this enzymatic reaction effectively. Aromatase inhibitor is administered to metastatic breast cancer patients customarily in which estrogen receptor had been demonstrated only in the primary tumor, not the metastatic sites. The status of aromatase in metastatic sites has not been well-characterized to date. We immunolocalized aromatase in 46 estrogen receptor positive primary breast cancers and paired metastatic lymph nodes, using immunohistochemistry. Immunoreactivity was detected in 44/46 primary tumors and 40/46 metastatic lymph nodes. A significant correlation was detected between the status of aromatase in primary and metastatic sites. Aromatase immunoreactivity was correlated with age, size of primary tumor and Ki-67 index. Aromatase immunoreactivity was also detected in adipose tissue surrounding the lymph nodes. In conclusion, aromatase status in primary tumors generally represents its status in metastatic lymph nodes. This indicates that the endocrine environment of estrogen receptor positive tumors remain stable during the metastatic process.
Recent gene expression studies of breast cancer identified molecularly distinct subtypes based on intrinsic gene sets. Routine immunohistochemical analysis of estrogen receptor (ER), progesterone receptor (PgR), HER2 and Ki-67 can be used to define the major breast cancer subtypes; luminal A (ER or PgR+, HER2−), luminal B (ER or PgR+, HER2+ or Ki-67 >14%), HER2-overexpression (ER and PgR−, HER2+), triple-negative (ER and PgR−, HER2−). The most frequent types of breast cancer are ER+ tumors or luminal A/B types. Third generation aromatase inhibitors (AIs) have become the gold standard of endocrine therapy for luminal types, by suppressing aromatase, the key enzyme in estrogen biosynthesis. They are reported to elongate disease free survival and decrease recurrence rates compared to conventional endocrine therapy.[3-6] However, the truth remains that despite potent estrogen inhibition by AI treatment, many ER+ tumors eventually metastasize and develop resistance to AIs.
When tumors metastasize, biological status of primary tumors is considered to be maintained in the metastasized tumor cells; therefore, aromatase expressed in tumor cells and the surrounding stromal cells at primary tumors, should also be preserved at metastatic tumor foci and its surrounding stromal cells. Several investigators, however, have reported that the ER status in metastatic lymph nodes was decreased compared to that in primary tumors,[8, 9] which raises a question as to whether aromatase status in the metastatic foci is indeed the same as in the primary tumor in the same patients.
We hypothesized that the aromatase status in metastasized lymph nodes would be different from the status in primary tumors, which could explain why breast cancer acquires resistance during its process of progression. Therefore, we compared aromatase immunoreactivity in primary tumors and paired metastasized lymph nodes of breast cancer in the same patients. We also correlated the findings with clinicopathological factors and intrinsic subtype of individual patients.
Material and Methods
The patients with breast cancer examined in our study were all operated on at Tohoku University Hospital, Sendai, Japan, and were diagnosed with pure invasive ductal carcinoma with lymph node metastases. 10% formalin fixed and paraffin embedded tissue sections from both the primary tumors and the corresponding lymph node metastases were screened among 123 cases diagnosed with positive lymph node metastases from 2004 to 2008. The patients who had distant metastases or had received neoadjuvant chemotherapy or radiotherapy at the time of operation were subsequently excluded from this study. In addition, in seven cases, the lymph node metastatic sites were diminished while making multiple tissue sections so these cases were also excluded from the present study. Therefore, a total of 46 formalin-fixed, paraffin embedded primary tumor tissues and paired lymph nodes were available for this study. Clinicopathological features of the cases examined are summarized in Table 1. Research protocols for this study were approved by the ethics committee at Tohoku University School of Medicine (2010-570).
|Age (Years): Median 56, Mean 58.1 (Range 41–82)|
|Primary tumor size (cm)|
|Number of lymph nodes with metastasis|
The characteristics of primary antibody for aromatase (aromatase monoclonal antibody 677) were previously reported by Miki et al. Briefly, after deparaffinization, sections were washed with PBS and treated with 0.3% hydrogen peroxidase in methanol for 20 min. Normal rabbit serum (1%) was applied to the sections for 20 min and primary antibody was applied to the tissue sections for 18 h at 4C. Reacted sections were subsequently incubated with anti-mouse immunoglobulin for 20 min, followed by exposure to peroxidase-conjugated streptavidin for 20 min at room temperature. Immunoreactivity was detected by immersing the tissue sections in 3,30-diaminobenzidine (DAB) solution (1 mM DAB, 50 mM tris–HCl buffer (pH 7.6), 10 mM sodium azide, 0.006% hydrogen peroxidase). Sections were counterstained with hematoxylin.
Other antibodies used in this study were as follows: ERalpha (ER1D5; Immunotech S.A., Marseilles, France), PgR (MAB429; Chemicon International Inc., Temecula, CA, USA), HER2 and Ki-67 (MIB1; DakoCytomation Co. Ltd, Kyoto, Japan).
A Histofine Kit (Nichirei, Tokyo, Japan) was used for immunohistochemistry in our study. The antigen–antibody complex was visualized with 3, 3′-diaminobenzidine (DAB) solution (1 mM DAB, 50 mM Tris–HCl buffer (pH 7.6), and 0.006% H2O2) and counterstained with hematoxylin.
Evaluation of immunohistochemistry
Aromatase immunoreactivity was detected in the cytoplasm. The approximate percentage of stained cells (proportion score) were classified into the following four groups: 0, <1%; 1, −25%; 2, −50%; and 3, >50%. Relative intensity of aromatase immunopositive cells was classified as follows: 0, no immunoreactivity; 1, weak; 2, moderate; and 3, intense immunoreactivity, according to the report by Miki et al. Aromatase immunoreactivity was evaluated as a total score of proportion score and intensity score.
Immunoreactivity of ER and PgR was detected in the nuclei of cancer cells. A number of positive cells were counted in 10 random optic fields, using a light microscope equipped with 50x objective lenses. In small lymph node metastatic sites where 10 optic fields were unavailable, all cancer cells were evaluated. Subsequently, the percentage of immunoreactivity, i.e. labeling index (LI), was determined. The ER and PgR LI of more than 10% were considered positive. The ER and PgR were also scored for Allred score (0–8).
HER2 immunoreactivity was detected at the membrane of cancer cells and was evaluated according to the grading system (HercepTest 0, 1+, 2+, 3+)(DAKO).
Ki-67 immunoreactivity was detected in the nuclei and evaluated using labeling index (LI). More than 10% of positive tumor cells were tentatively considered positive staining
All the statistical analysis was performed using StatView (SAS Institute, San Francisco, CA, USA). A P-value of <0.05 was considered statistically significant.
A total of 44 PTs (44/46, 95.7%) were immunohistochemically positive for aromatase. Among them, 36 cases (36/46, 78.3%) demonstrated aromatase immunoreactivity in more than 50% of the tumor cells (proportion score (PS)3). Aromatase immunoreactivity was evaluated in matched metastasized lymph nodes (metastatic LNs) of those 46 patients above. Aromatase immunoreactivity in lymph nodes were detected in 40 cases (40/46, 87.0%), and among these 40 cases, 35 cases (35/46, 76.1%) were in PS3 status. The mean percentage of immunopositive cells was 71.2% in PT and 65.7% in metastatic LNs. Aromatase immunoreactivity of tumor cells in PTs and the paired metastatic LNs demonstrated good concordance (Y = 44.494 + 0.298 *X; R∧2 = 0.067, P < 0.05) using regression analysis. The results of aromatase immunohistochemical scores are summarized in Figure 1.
Among 41 PTs associated with a high proportion score (PS3 or 2) in PTs, 35 cases also demonstrated high PS in metastatic LNs while only six cases showed low PS (PS1 or 0). However, four out of five cases with low PS scores in PTs were associated with increased PS scores in metastatic LNs. In addition, 10 out of 42 cases with high scores in PTs were associated with low intensity scores (IS) in metastatic LNs, whilst three out of four cases with low IS in PTs demonstrated high scores in metastatic LNs. Seven (15.2%), 22 (47.8%), 17 (37.0%) of the cases demonstrated increment, no-changes and decrement of PS, respectively (Fig. 1b), and 35 (76.1%) and 11 (23.9%) cases were associated with increased and decreased IS, respectively (Fig. 1a). In total score (PS + IS), 9 (19.6%), 21 (45.7%) and 16 (34.8%) cases were associated with increment, no changes and decrement in metastatic LNs compared to PTs, respectively (Fig. 1c). Representative illustrations of staining patterns for PTs and the corresponding metastatic LNs (which both were scored as 3+) were illustrated in Figure 2.
Aromatase immunoreactivity was detected in stromal components of PTs as well as metastatic LNs, including fibroblasts and endothelial cells. Relatively weak immunoreactivity was also detected in adjacent lymphocytes in metastasized lymph nodes as well as non-metastasized lymph nodes. The adjacent lymphoid follicles of metastatic LNs were negative for aromatase, as well as the non-metastasized lymph nodes (Fig. 3).
Correlation of aromatase in PT and metastatic LNs with histopathological parameters
Aromatase immunoreactivity in PT was significantly correlated with a small PT size (aromatase cut off value 50% and 75%, P < 0.05) and size of largest metastatic LNs (aromatase cut off values 50 and 75%, P < 0.05). Aromatase immunoreactivity in metastatic LNs was negatively correlated with Ki-67 labeling index in PTs (aromatase cut off value 25%, P = 0.0197) and aromatase immunoreactivity in metastatic LNs with younger age (aromatase cut off value 50 and 75%, age ≤60, P = 0.0171 and 0.0009)(Table 2).
|25% Aromatase PT||25% Aromatase LN||50% Aromatase PT||50% Aromatase LN||75% Aromatase PT||75% Aromatase LN|
|(+), (−)||P||(+), (−)||P||(+), (−)||P||(+), (−)||P||(+), (−)||P||(+), (−)||P|
|Age||≤55||19, 1||0.3693||18, 2||0.435||19, 1||0.0278||15, 5||0.7495||14, 6||0.364||12, 8||nsa|
|>55||22, 4||20, 6||17, 9||18, 8||14, 12||15, 11|
|≤60||16, 2||nsa||16, 2||0.4525||14, 4||nsa||9, 9||0.0171||7, 11||0.0288||5, 13||0.0009|
|>60||25, 3||22, 6||22, 6||24, 4||21, 7||22, 6|
|Histological Grade||1||12, 0||0.3059||11, 1||0.6598||10, 2||nsa||10, 2||0.4614||9, 3||0.3151||9, 3||0.3071|
|2&3||29, 5||27, 7||26, 8||23, 11||19, 15||18, 16|
|Primary tumour size||<2||16, 2||nsa||15, 3||nsa||15, 3||0.7172||14, 4||0.5223||13, 5||0.2344||14, 4||0.0644|
|≥2||25, 3||23, 5||21, 7||19, 9||15, 13||13, 15|
|<2.5||27, 3||nsa||24, 6||0.6942||27, 3||0.0202||22, 8||0.7441||23, 7||0.0043||19, 11||0.531|
|≥2.5||14, 2||14, 2||9, 7||11, 5||5, 11||8, 8|
|<3||32, 2||0.1029||27, 7||0.6598||28, 6||0.4157||23, 11||0.4614||24, 10||0.0383||19, 15||0.7346|
|≥3||9, 3||11, 1||8, 4||10, 2||4, 8||8, 4|
|Number of metastatic LN||1 to 3||26, 3||nsa||22, 7||0.2263||24, 5||0.4623||19, 10||0.315||18, 11||nsa||15, 14||0.2352|
|4–||15, 2||16, 1||12, 5||14, 3||10, 7||12, 5|
|Size of largest metastatic LN||<1 cm||23, 2||0.6476||19, 6||0.2597||20, 5||ns||15, 10||0.0987||18, 7||0.1317||13, 12||0.3769|
|≥1 cm||18, 3||19, 2||16, 5||18, 3||10, 11||14, 7|
|<2 cm||38, 4||0.3794||36, 6||0.1341||35, 7||0.0278||32, 10||0.0622||27, 15||0.0238||26, 16||0.2916|
|≥2 cm||3, 1||2, 2||1, 3||1, 3||0, 4||1, 3|
|LN ER status||(+)||40, 5||nsa||38, 7||0.1739||36, 9||0.2174||33, 12||0.2826||28, 17||0.3913||27, 18||0.413|
|(−)||1, 0||0, 1||0, 1||0, 1||0, 1||0, 1|
|PT PgR status||(+)||28, 4||nsa||27, 5||0.684||25, 7||nsa||23, 9||nsa||22, 10||0.1152||21, 11||0.1988|
|(−)||13, 1||11, 3||11, 3||10, 4||6, 8||6, 8|
|LN PgR status||(+)||33, 5||0.5692||33, 5||0.1294||30, 8||nsa||28, 10||0.6689||25, 13||0.2316||25, 13||0.0505|
|(−)||8, 0||5, 3||6, 2||5, 3||3, 5||2, 6|
|PT ER (LI)||<50||5, 0||nsa||5, 0||0.5692||4, 1||nsa||3, 2||0.6119||3, 2||nsa||2, 3||0.6351|
|≥50||36, 5||33, 8||32, 9||30, 11||25, 16||25, 16|
|PT PgR (LI)||<50||22, 0||0.0502||18, 4||nsa||19, 3||0.2894||17, 5||0.5207||13, 9||nsa||13, 9||nsa|
|≥50||19, 5||20, 4||17, 7||16, 8||15, 9||14, 10|
|PT HER2||0,1,2||34, 4||nsa||31, 7||nsa||29, 9||0.6641||27, 11||nsa||22, 16||0.4525||21, 17||0.4395|
|3||7, 1||7, 1||7, 1||6, 2||6, 2||6, 2|
|PT Ki-67 (LI)||≤10||21, 3||nsa||23, 1||0.0197||20, 4||0.4839||19, 5||0.3304||16, 8||0.5468||16, 18||0.3695|
|>10||20, 2||15, 7||16, 6||14, 8||12, 10||11, 11|
|LN Ki-67 (LI)||≤10||16, 1||0.6375||16, 1||0.2263||14, 3||0.7227||14, 3||0.315||9, 8||0.5335||12, 5||0.2352|
|>10||25, 4||22, 7||22, 7||19, 10||19, 10||15, 14|
Heterogeneous patterns of aromatase status between primary tumor and metastasized lymph nodes
We divided the aromatase staining pattern between PT and metastatic LNs in to two, heterogeneous and homogeneous. Cases which showed a change of more than 2-fold between PT and metastatic LNs in terms of the approximate percentage of aromatase positive cells was tentatively named the ‘heterogeneous’ group; cases which showed a change of less than 2-fold between PT and metastatic LNs was named the ‘homogeneous’ group. Eleven cases (11/46, 23.9%) corresponded to ‘heterogeneous’ group; four increased and seven decreased patterns in metastatic LNs compared to PT. The remaining 35 cases corresponded to ‘homogeneous’ group. Figures 4 and 5 shows heterogeneous aromatase staining between PT (Fig. 4) and metastatic LNs (Fig. 5) in the same case. The ‘heterogeneous’ group of the patients was associated with significantly younger age, smaller size of largest metastatic LNs and lower ER status in PTs compared to the ‘homogeneous’ group of the patients (Table 3).
|Primary tumour size||<2||15||3||0.4865|
|Number of metaLN||1 to 3||21||8||0.5012|
|Size of largest metaLN||<2 cm||32||10||nsa|
|LN ER status||(+)||34||11||nsa|
|PT PgR status||(+)||26||6||0.2691|
|LN PgR status||(+)||31||7||0.0789|
|PT ER (LI)||<50||6||5||0.0407|
|PT PgR (LI)||<50||16||6||0.7343|
|PT Ki-67 (LI)||≤10||21||3||0.0861|
|LN Ki-67 (LI)||≤10||15||2||0.1723|
In breast cancer, an interaction of tumor cells with surrounding stromal cells plays a critical role in estrogen production via the aromatase enzyme. Intratumoral aromatase has been detected in stromal cells such as adipocytes and fibroblasts as well as in parenchymal or cancer cells.[10, 14] Reports show that these aromatase positive stromal cells promote the process of invasion and metastasis. Such a model raises the following questions: (i) what is the location of aromatase enzyme expression in the metastasized lymph nodes? and (ii) do tumor cells synthesize aromatase via tumor-stromal interaction at the metastatic sites as in primary tumors?
Aromatase immunoreactivity was recently reported to be decreased in lymph node metastatic sites when compared to the primary sites, using tissue microarray, which they concluded may account for potential resistance to endocrine therapy at metastatic sites. However, it is also known that results of tissue microarray may not be compared to those evaluated in the whole sections, due to the heterogeneous nature of aromatase staining, which calls for investigation using whole sections. Therefore, in our present study, aromatase immunoreactivity was evaluated in paired whole tissue sections of primary breast cancer and the corresponding metastasized lymph nodes in ER+ breast cancer tissues.
Aromatase immunoreactivity was detected in metastatic cancer cells of lymph nodes as well as the surrounding stromal cells or fibroblasts in ER+ breast cancer using whole tissue sections. There was a statistically significant positive correlation between aromatase status in PTs and metastatic LNs; with metastatic LNs demonstrating slightly higher aromatase immunoreactivity compared to PTs. Such consistent patterns of aromatase expression in metastatic LNs indicates that aromatase immunoreactivity in PTs, which can be easily evaluated following surgery or biopsy using immunohistochemistry, can predict aromatase status in metastatic LNs with high probability.
There were, however, cases in which aromatase status between PTs and metastatic LNs were clearly discordant. Interestingly, these cases were associated with significantly lower ER positivity in PTs. Aromatase immunoreactivity of PTs is not a predictive factor of response to endocrine therapy but ER positivity in cancer cells is well-known to be a strong predictor of response to endocrine therapy. Therefore, we infer that heterogeneous aromatase expression between PTs and metastatic LNs may be one explanation for variable response to endocrine therapy among ER+ tumors. Also, we noticed a particular result of interest in a few cases which presented with negative aromatase expression in PTs but positive aromatase expression in metastatic LNs. This data suggests a possible indication of aromatase inhibitors in LN metastatic cases with negative aromatase status in PTs. This possibility could only have been discovered by this immunohistochemical analysis of aromatase status in lymph nodes. To clarify this, further study utilizing ER negative breast carcinoma is warranted.
In our present study, a significant negative correlation was detected between the size of PTs and aromatase status of both PTs and metastatic LNs. This result is in good agreement with a prior study by Ellis et al. From this, we may hypothesize that aromatase expression decreases as the tumor grows and progresses, not only at primary tumor sites but also at metastasized lymph nodes which could explain this interesting phenomenon. This evidence strengthens the known fact that early detection and treatment of breast cancer is, above all, the most important treatment strategy for breast cancer. In addition, the lower Ki-67 labeling index in PTs was significantly associated with higher aromatase status in metastatic LNs, which suggests that aromatase positive tumor cells may intrinsically have a less aggressive behavior. This finding also implies that for predicting outcomes of ER+ tumors, small tumor size and Ki-67 labeling index in combination with ER, PgR and HER2 status may give the same amount of information that immunohistochemical analysis of aromatase would give. It is therefore suggested that routine immunohistochemical analysis of aromatase in PTs and metastasized LNs would not aid in the prediction of a prognosis in ER+ lymph node positive patients; instead they help clarify the role of aromatase in the process of metastasis.
Our data shows that almost all of the tumors demonstrated ER positivity in metastatic LNs (45/46, 97.8%). This concordance rate is much higher compared to other reported studies, due to the fact that we only examined ER+ tumors in this study. PgR was associated with a more variable result; 10 out of 46 tumors (21.7%) demonstrated disparities between PTs and metastatic LNs. Many previously reported studies have focused on ER expression in PTs and LNs but few have examined thoroughly the discordance of PgR expression between PTs and metastatic LNs. In this study, two out of 32 PgR+ tumors showed conversion to PgR– in metastatic LNs. Of interest, both of these cases were associated with low ER total score (TS5 and 6) and low aromatase expression (0% and 15%). ER+/PgR– breast cancer has been reported to be a distinct subset of breast cancer showing aggressive behavior and endocrine therapy resistance. Therefore, further examination of PgR in metastatic LNs with its relation to prognosis and acquired resistance is warranted.
Heterogeneous aromatase expression patterns among multiple metastatic LNs were observed in this study where 18 out of 34 cases (18/34, 52.9%) with multiple metastatic LNs demonstrated more than a two-fold difference in aromatase immunoreactivity. The cases showing aromatase heterogeneity were significantly correlated with a larger number of metastatic LNs only but also tended to be correlated with weak ER expression in PTs (data not shown). Much of this phenomenon may be explained by the heterogeneous nature of aromatase staining but awaits further examination.
In summary, both cancer cells and surrounding stromal cells at the sites of metastasized lymph nodes generally expressed aromatase in the same manner and the same intensity as in the primary tumor of the same patients with few exceptions. We may conclude that tumor-stromal interaction is as important at the metastasized lymph nodes as it is at the primary tumor in inducing aromatase. These findings demonstrated that once tumor cells get into the lymphatic system, tumors will seed on to lymph nodes to receive interaction from the stromal components of metastatic sites, where aromatase once again synthesizes estrogen for tumor growth. Somehow, some of the tumor cells stop generating aromatase; therefore, source of nutrition lies elsewhere, which warrants further research.
We thank Katsuhiko Ono and Hikako Kumasaka for (Department of Pathology, Tohoku University School of Medicine, Sendai, Japan) for skillful technical assistance. Yukiko Shibahara has been announced as the winner of the Japanese Society of Pathology’s Centennial Anniversary Award for Young Scientists.
- 3Update of the BIG 1-98 Trial: Where do we stand? Breast 2009; 18 Suppl 3: S78–82., .