Fax: (011) (33) 145212751
Luteinizing hormone receptor status and clinical, pathologic, and prognostic features in patients with breast carcinomas
Article first published online: 18 MAR 2003
Copyright © 2003 American Cancer Society
Volume 97, Issue 7, pages 1810–1816, 1 April 2003
How to Cite
Meduri, G., Charnaux, N., Spyratos, F., Hacene, K., Loosfelt, H. and Milgrom, E. (2003), Luteinizing hormone receptor status and clinical, pathologic, and prognostic features in patients with breast carcinomas. Cancer, 97: 1810–1816. doi: 10.1002/cncr.11294
- Issue published online: 18 MAR 2003
- Article first published online: 18 MAR 2003
- Manuscript Accepted: 18 DEC 2002
- Manuscript Revised: 12 DEC 2002
- Manuscript Received: 9 SEP 2002
- breast carcinoma;
- human chorionic gonadotropin (hCG);
- lutenizing hormone;
- luteinizing hormone receptor (LHR)
It has been established that pregnancy protects against breast carcinoma, and animal models have shown that human chorionic gonadotropin (hCG) mimics this effect by inhibiting the initiation and progression of experimental breast carcinoma. Luteinizing hormone (LH)/hCG receptors (LHR) have been characterized in several human breast carcinoma cell lines and in a limited number of breast carcinoma biopsy specimens. These observations led to the suggestion that hCG may be used as a means of prevention and possibly treatment in patients with breast carcinoma.
The authors used immunocytochemistry to analyze tumors from 160 patients who were followed for a median of 2539 days. Using a cut-off value of 18% immunolabeled cells in each tumor, 72% of tumors were identified as LHR positive. The LHR-positive tumors were found more frequently in premenopausal women, who had tumors with greater cell differentiation and positive estrogen receptor α status. Infiltrating lobular carcinomas were positive for LHR more frequently compared with infiltrating ductal carcinomas. There was no correlation between LHR status and lymph node invasion, tumor size, or progesterone receptor status.
Patients with LHR-positive tumors had a longer metastasis free survival, although the statistical significance was slight (P = 0.07), most likely due to the limited number of events in the patients studied. Conversely, there was no difference between patients with LHR-positive or LHR-negative tumors in the local recurrence free interval.
LHR status seems to be related in part to the degree of differentiation in breast tumors, confirming experimental evidence of the effect of hCG on mammary tissue. The presence of LHR is a tumor characteristic that largely is independent of other clinical and pathologic tumor features. It may be of interest in the future to correlate the presence of LHR with a possible therapeutic response in individual patients to hCG. Cancer 2003;97:1810–6. © 2003 American Cancer Society.
It is known that pregnancy—especially early pregnancy—has a protective effect against the occurrence of breast carcinoma.1 It has been proposed that this effect was due to the terminal differentiation of the breast, in that mammary tissue that is not differentiated fully is more susceptible to carcinogenesis.2–4 Russo et al. established animal models to study this issue.3–5 Those authors found that, in rodents, mammary carcinoma could be induced by administration of dimethylbenz [a] anthracene (DMBA) only in young, nulliparous females. Tumor initiation was prevented by the completion of pregnancy prior to carcinogen exposure. In that model, it was observed that human chorionic gonadotropin (hCG) administration to rats mimicked the preventive effect of pregnancy. Furthermore, if DMBA was administered initially to virgin rats, provoking carcinogenesis, then the secondary administration of hCG inhibited the progression of the breast carcinoma.5 This inhibition was accompanied by an increase in transcription products of genes involved in apoptosis.6 Conversely, in the ovary, which is the classic gonadotrophin target organ, hCG did not increase the transcription of apoptosis-related genes. You et al.7 studied the role of connexins in gap junction formation in the mammary gland. They observed that the mammary epithelial cells expressed mainly connexin 26 (Cx26), whereas the surrounding fibroblasts expressed connexin 43 (Cx43). The Cx26 gene, contrary to other connexin genes, when transfected into HeLa cells, provoked a strong negative effect on cell growth. When introduced into the human breast carcinoma MDA-MB-435 cell line, the Cx26 gene decreased growth and restored cellular differentiation. Thus, Cx26 acted like a tumor suppressor gene.8 You et al.7 also studied Cx26 regulation in the breast and observed that Cx26 mRNA and protein were increased by hCG administration both in vivo and in tissue explants. Insulin-like growth factors (IGFs) are known as potent mitogens for breast carcinoma cells.9–11 Huynh12 observed a decreased concentration of IGF1 under the effect of hCG. Several IGF binding proteins that are known to decrease the concentration of available IGFs were increased by hCG administration. These various studies suggest that hCG exerts growth-inhibitory and proapoptotic effects in the breast.
In addition, the presence of luteinizing hormone (LH)/hCG receptors (LHR) has been described in human breast carcinoma cell lines.13, 14 We previously used immunocytochemistry and reverse transcriptase-polymerase chain reaction analysis to characterize the LH/hCG receptor in a limited number of biopsy specimens from patients with breast carcinoma.14 All these observations led to the proposal that hCG may be useful in the prevention and perhaps in the treatment of breast carcinoma.15, 16 Indeed, a clinical trial is currently being performed in patients with advanced breast carcinoma (unpublished data). Such treatment would be expected to be successful only in patients who harbor the LH/hCG receptor in their tumors. However, the proportion of patients with LHR-positive tumors has not been determined, and it is unknown whether the presence of this receptor is correlated with specific tumor features or with prognosis. Thus, we undertook a study of receptor distribution in 160 patients in whom the clinical and pathologic features have been characterized extensively and who had a median follow-up of 2539 days.
MATERIALS AND METHODS
Breast carcinoma specimens were obtained from 160 patients who underwent mastectomy or lumpectomy. Tissue blocks were snap frozen and preserved in liquid nitrogen until they were used.
The mouse monoclonal antihuman LH/hCG receptor antibody29 has been described previously and characterized extensively.14 It was used at a concentration of 10 μg/mL. The immunochemical characterization of LHR in frozen sections of breast carcinoma tissues and in all controls has been described previously in detail.14
For immunocytochemical scoring, at least 300 epithelial cells were counted at high magnification in 3 different areas of each section and in 3 different sections of the same tumor. The percentage of stained cells was determined. The staining intensity also was assessed and was subdivided into the following four grades: negative, low, medium, and high.
Estrogen receptor (ER) and progesterone receptor (PR) were assayed according to the manufacturer's instructions (ER radioimmunoassay [RIA] and PR RIA; Abbott Laboratories, Abbott Park, IL).
Metastasis free survival (MFS) and disease free survival (DFS) intervals were calculated from the date of first treatment to the first event. The first event was defined as the occurrence of local recurrent disease and/or distant metastasis. Local recurrence was defined as a tumor arising in the treated breast or in regional lymph nodes. Survival curves were derived from Kaplan–Meier estimates.17 The log-rank test was used to test the equality of survival distribution.18 A cut-off value of 18% labeled cells was established as the most efficient for separating categories of tumors.
Scoring of Tumors
After immunolabeling with the anti-hCG/LH receptor antibody, the percentage of stained tumor cells was determined. Examples of breast tumors are illustrated in Figure 1 that show a very positive tumor (Fig. 1A), a moderately positive tumor (Fig. 1B), a weakly positive tumor (Fig. 1C), and a completely negative tumor (Fig. 1D). The proportion of LHR-positive cells varied between 0–80% in the different tumors. The distribution of tumors according to the percentage of immunostained cells is shown in Figure 2. A cut-off value of 18% labeled cells was established as the most efficient for separating categories of tumors.
We initially classified the tumors for LH/hCG receptor status using an immunocytochemical score that took into account both the percentage of stained cells and the intensity of staining. However, statistical analysis showed that addition of the latter characteristic did not improve correlation studies (data not shown).
LHR Status and Clinical and Pathologic Characteristics of the Patients and Tumors
LHR-positive tumors were found in 115 patients (72%), whereas LHR-negative tumors were present in 45 biopsies (28%) (Table 1). LHR-positive tumors were found more frequently in premenopausal women (P = 0.02). This also was reflected in the fact that women age < 50 years more frequently had LHR-positive tumors (P = 0.05). Tumors that contained ERα were also more frequently LHR positive (P = 0.05).
|Characteristic||No. of patients (%)||P value|
|Total population||LHR ≤ 18%||LRH > 18%|
|Total||160 (100)||45 (28)||115 (72)||—|
|≤ 50 yrs||58 (36.3)||11 (24.4)||47 (40.9)||0.05|
|> 50 yrs||102 (63.8)||34 (75.6)||68 (59.1)||—|
|Premenopausal||66 (41.3)||12 (26.7)||54 (47.0)||0.02|
|Postmenopausal||94 (58.8)||33 (73.3)||61 (53.0)||—|
|≤ 20 mm||50 (32.5)||9 (20.9)||41 (36.9)||0.06|
|> 20 mm||104 (67.5)||34 (79.1)||70 (63.1)||—|
|Lymph node statusb|
|Negative||77 (48.4)||23 (51.1)||54 (47.4)||0.91|
|Positive (1–3)||49 (30.8)||13 (28.9)||36 (31.6)||—|
|Positive (> 3)||33 (20.8)||9 (20.0)||24 (21.0)||—|
|Grade 1||71 (47.7)||13 (31.7)||58 (53.7)||0.02|
|Grade 2||78 (52.3)||28 (68.3)||50 (46.3)||—|
|< 10 fmol/mg||50 (31.3)||9 (20.0)||41 (35.6)||0.05|
|≥ 10 fmol/mg||110 (68.7)||36 (80.0)||74 (64.4)||—|
|< 10 fmol/mg||77 (48.1)||24 (53.3)||53 (46.1)||0.40|
|≥ 10 fmol/mg||83 (51.9)||21 (46.7)||62 (53.9)||—|
There was a clear-cut correlation between the presence of LHR and the modified histologic grade (modified Scarff–Bloom–Richardson grade)19 (P = 0.02). Furthermore, infiltrating lobular carcinomas were LHR positive more frequently than infiltrating ductal carcinomas (P = 0.02). There was no correlation with lymph node status (P = 0.91), tumor size (P = 0.06; at the limit of significance), or the presence of PR (P = 0.40).
Data regarding parity were available for only 53 patients. Due to the limited number of patients, no statistical analysis could be performed; however, it is noteworthy that 2 of 5 nulliparous women had LHR-negative tumors, whereas only 4 of 48 parous women had LHR-negative tumors.
LHR Status and Clinical Outcome
Patients with LHR- positive tumors and LHR- negative tumors did not undergo different treatments (Table 2). There was no statistically significant difference (P = 0.35) in the DFS (Fig. 3) of patients according to their LHR status. However, analysis of MFS showed a tendency (P = 0.07) toward better prognosis in patients with LHR- positive status (Fig. 4). In fact, the difference was visible between 5 years and 12 years of follow-up and disappeared after longer periods of disease evolution. Conversely, there was no correlation between LHR status and the development of local recurrence (Fig. 5).
|Characteristics||No. of patients (%)||P value|
|Total population||LHR ≤ 18%||LRH > 18%|
|Total||160 (100)||45 (28)||115 (72)||—|
|Tumorectomy||71 (44.9)||15 (34.1)||56 (49.2)||0.09|
|Mastectomy||87 (55.1)||29 (65.9)||58 (50.8)||—|
|Negative||101 (63.1)||32 (71.1)||69 (60)||0.19|
|Positive||59 (36.9)||13 (28.9)||46 (40)||—|
|Negative||101 (63.1)||27 (60)||74 (64.3)||0.61|
|Positive||59 (36.9)||18 (40)||41 (35.7)||—|
|Metastases and local recurrenceb|
Using a cut-off value of 18% cells immunolabeled with anti-LHR antibody, we observed that 72% of breast tumors were LHR positive. LHR-positive tumors were observed more frequently in premenopausal women. This may be due either to receptor down-regulation by high concentrations of LH in postmenopausal women or to receptor induction by progesterone, as observed previously in pig mammary gland.14 The presence of LHR also may be related to cell differentiation, because it is found more frequently in lower grade tumors and is associated with the presence of ERα. Furthermore lobular invasive carcinomas, which are derived from more differentiated mammary cells,20, 21 were LHR positive more frequently than ductal invasive carcinomas. LHR contents appeared to be independent of other clinical and pathologic tumor features.
The prognostic importance of LHR remains unclear to date. Although patients with LHR-positive status had a longer MFS interval, this was at the limit of statistical significance. It is possible that LHR, because it is related to cell differentiation, is associated to some extent with a better prognosis. However, because it is not a major prognostic factor, it would be necessary to study a larger group of patients to show this unequivocally. It is not surprising that other factors (the number of invaded lymph nodes, tumor size, steroid hormone receptor status, c-erb B2 content, etc.) play a more prominent prognostic role.
When examining the clinical outcome of our 160 patients, it appeared that the difference in MFS interval between patients with LHR-positive status and LHR-negative status was more apparent for a period up to 10 years after diagnosis and disappeared thereafter. A similar observation has been made for some other prognostic factors.22, 23 LHR status was not found to be associated with any difference in the local recurrence free interval.
It is not clear whether mammary LH/hCG receptors play a physiologic role only during pregnancy, when hCG concentrations are very high. Activation of LH/hCG receptors in the mammary gland may intervene in the process of differentiation of terminal ductuloalveolar units, most likely in synergy with other hormones of pregnancy.2 The diminished susceptibility to cancerization of the mammary gland after multiple pregnancy may be explained by the reduced sensitivity of differentiated terminal acinoductular structures to mitogens.5
Breast carcinoma cells also produce hCG and LH.24 The α subunit of glycoprotein hormones is synthesized in high concentrations, particularly in ERα-positive tumors. However, βhCG also is present, although in markedly lower concentrations.25 The active dimer, thus, may be synthesized locally.
The effect of these hormones may involve the classic seven-transmembrane receptor. However, for the follicle-stimulating hormone receptor, a variant has been described with a single transmembrane domain that acts as a growth factor receptor.26 Finally, crystallization and X-ray diffraction studies have determined the three-dimensional structure of hCG.27 This structure indicates that hCG is a member of the family of cystine-knot growth factors that also includes nerve growth factor, transforming growth factor β, and platelet-derived growth factor β. Further studies will be necessary to determine whether the effects of hCG on the initiation and progression of experimental carcinoma in rats are mediated by the LH/hCG receptor or whether another signaling pathway is involved. The same issue should be investigated to determine the putative effects of hCG in the current clinical trial performed in patients with advanced breast carcinoma. It will be of interest to determine whether clinical responses vary between patients with LHR-positive tumors and patients with LHR-negative tumors.
- 19Prognostic value of histologic grade nuclear components of Scarff–Bloom–Richardson (SBR). An improved score modification based on a multivariable analysis of 1262 invasive ductal breast carcinomas. Cancer. 1989; 64: 1914–1921., , , , , .