| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Other Original Articles |
Children’s Hospital, University of Leipzig (A.G.), 04317 Leipzig, Germany; Institute of Pathology, University of Leipzig (L.-C.H.), 04103 Leipzig, Germany; Hospital St. Georg (U.K.), 04129 Leipzig, Germany; Faculté de Médicine Necker, INSERM, U-344 (P.A.K.), 75730 Paris, France; Unit on Genetics and Endocrinology, Developmental Endocrinology Branch (S.E.T., C.A.S.), National Institute of Child Health and Human Development, and Department of Pathology (J.G.) and Laboratory of Tumor Immunology and Biology, Molecular and Cellular Endocrinology Section (B.K.V.), National Cancer Institute, National Institutes of Health, Bethesda, Maryland 20892; and Department of Endocrinology (S.R.B.), University of Duesseldorf, D-40225 Duesseldorf, Germany
Address all correspondence and requests for reprints to: Annegret Glasow, Ph.D., Leukaemia Research Fund Center at the Institute of Cancer Research, Chester Beatty Laboratories, 237 Fulham Road, London, United Kingdom SW3 6JB. E-mail: aglasow{at}icr.ac.uk
Abstract
PRL is a major growth and differentiating hormone in the human breast, with activation of the PRL-PRL receptor complex increasingly recognized as an important mechanism in the induction and progression of mammary tumors. Although constitutive activation of various hormone and growth factor receptors is newly recognized as a common cause of tumor development, the PRL receptor gene has not been analyzed for similar aberrations in breast and other tumors. Therefore, using bacterial artificial chromosomes containing the PRL receptor gene and intron-spanning PCR, we determined the exon-surrounding intron sequences providing primers for the first analysis of the entire coding region of the human PRL receptor gene. We examined the presence of PRL receptor in 41 breast tumors by immunohistochemistry and attempted a correlation of its expression to pathological grading of the disease. Then tumor cells were isolated by laser capture microdissection to examine DNA from 30 patients for PRL receptor mutations.
The PRL receptor immunoreactive score did not correlate to the tumor size, histopathological grading, age, or family history of patients. PRL receptor immunoreactivity was predominantly found in steroid hormone receptor-positive tumors, but without overall correlation of immunoreactive score. In both PRL receptor-positive and PRL receptor- negative breast cancer cells, direct sequencing of the coding sequence of the PRL receptor gene did not detect any somatic or hereditary gene aberrations.
In conclusion, PRL receptor mutations do not appear to be common in human breast cancer, suggesting that constitutive activation of the PRL receptor can be excluded as a major cause of mammary tumor genesis. The molecular structure of the PRL receptor seems to remain intact in tumor tissue, and systemic and local production of PRL may participate in tumor cell growth and proliferation through functional receptors.
BREAST CANCER IS the most common cancer in women. According to the American Cancer Society, there are about 180,000 new cases and more than 40,000 deaths from breast cancer each year (1). The intensive search for causes of this cancer revealed no single simplistic etiology, but numerous epidemiological and hormonal risk factors. Inherited and sporadic genetic mutations as well as synchronous interaction of multiple hormones sensitizing the cells to the effects of carcinogens could predispose women to the development of cancer (2, 3). Although mutations of the tumor supressor genes BRCA1 and BRCA2 appear to be directly involved in some familial cases of mammary carcinogenesis, one third of familial breast cancer cases, and most sporadic breast cancer cases are independent of BRCA mutations (4). Mutations of other predisposition genes occurring less frequently have been identified as the ER and AR genes (5, 6). The oncogenic potential of cytokine receptors constitutively activated by mutations has been established (7), although the PRLR, which belongs to the class 1 cytokine receptor superfamily and transfers growth and differentiating effects of PRL in the human breast, has not been investigated to date in view of mutations or polymorphisms.
PRL belongs to the class of breast cancer susceptibility hormones that participate in the regulation of growth in normal and malignant breast cells. It is a major tumor mitogen in rodents, and excess PRL has been shown to induce breast cancer in mice (8, 9, 10). Accordingly, immunization with antibodies against the PRL receptor (PRLR) diminished the incidence of breast cancer in mice (11). In humans, incubation with PRL increases cell proliferation in mammary carcinoma cell lines (12) and breast tumor explants as well as primary cultures (13). Therefore, PRL may favor neoplasia, as carcinogens affect particularly mitotic cells (14). Promotion of mitogenesis could be mediated by PRL’s ability to activate the Raf/MAPK kinase/MAPK cascade (15), to regulate cyclin D1 expression (16), and possibly via the Janus kinase/signal transducer and activator of transcription pathway (17). Besides its direct effects on cell proliferation, PRL has been shown to inhibit apoptosis in rat Nb2 lymphoma cells, resulting in increased cell numbers (18). PRL serum concentrations seem to correlate to the incidence as well as the progression of human breast tumors (19, 20, 21, 22, 23).
The PRLR has been detected by in situ hybridization, immunohistochemical techniques, and RT-PCR in the majority (70–100%) of human mammary carcinomas (8, 24, 25). Increased PRLR expression has been shown by quantitative RT-PCR and immunocytochemistry in human cancerous vs. normal contiguous breast tissue (25), although other researchers found no grossly altered levels of PRLR expression by in situ hybridization or immunohistochemistry (26). In rats, a correlation between growth of dimethyl-benzanthrazene-induced breast cancer and the specific binding of PRL to the tumor tissue was found (27). In humans, the dissociation constant (Kd) of PRLR was reported to differ up to 100-fold between individual mammary carcinomas (28). Furthermore, elevated levels of members of the signal transducer and activator of transcription family involved in PRL signaling were found in human breast carcinoma nuclear extracts (29). Finally, an oncogenic, growth factor-independent activation of the PRLR was shown to be induced by a 178-amino acid deletion of the extracellular receptor domain in the murine myeloid cell line 32Dcl3 (30).
Although elevated PRL levels will exert a growth-promoting effect on breast cells via PRLR, PRLR aberrations may constitute another mechanism of tumor formation. Even in the presence of normal PRL levels, sequence aberrations or expression aberrations could modify the effects of PRL through a change in binding and signaling efficiency or internalization time. Therefore, this study was performed to investigate possible mutations and polymorphisms of the PRLR gene in human mammary carcinomas. Genomic DNA was obtained, specifically from breast cancer cells isolated by laser capture microdissection (LCM). Direct cycle sequencing of the PRLR-coding sequence was applied to screen for mutations. To include PRLR-positive and PRLR-negative carcinomas, PRLR expression was determined by immunostaining. Finally, breast carcinoma cells stained for the PRLR were correlated to the ER/PR score and pathological parameters to evaluate a potential prognostic meaning of the PRLR score.
Subjects and Methods
Patients and study design
Paraffin-embedded (maximum, 5 yr old) and frozen tumor tissue from 30 patients with breast carcinoma was screened for PRLR mutations. Tissues included 27 invasive ductal and 3 invasive lobular carcinomas. Thirteen of 30 patients had a first degree relative with breast cancer. Immunohistochemical analyses were performed on the paraffin-embedded tissue of these patients and on an additional 11 invasive ductal carcinomas (total of 41, 20 with and 21 without family history of breast cancer). The age of patients varied between 38–83 yr (mean ± SD, 57 ± 11.73). All tissues were taken from routine histology preparations of the Division of Gynecologic Histopathology, Institute of Pathology, University of Leipzig (Leipzig, Germany). Pathological classification as tumor size (pT), nodal involvement (pN), and histological grading refers to the TNM staging system (31). As a normal control group, blood samples of 20 healthy volunteers (5 males and 15 females) were used to isolate DNA from leukocytes. The ethical committee of the University of Leipzig approved the study.
Immunohistochemical analysis
Immunostaining of the PRLR. Formalin-fixed paraffin-embedded surgical specimens were provided by the Institute of Pathology, University of Leipzig. After sectioning (5–7 µm) and deparaffinization in degrading alcohols, immunostaining was performed using the labeled streptavidin-biotin technique (DAKO Corp., Hamburg, Germany), but without microwave demasking, as previously described (32). For specific staining, 3-amino-9-ethyl-carbazole was applied as color reagent, resulting in a red deposit. Nuclei were counterstained with Mayer’s hemalum solution (Merck & Co., Darmstadt, Germany). For specific staining of the PRLR, 5–10 µg/ml of the monoclonal mouse-antihuman PRLR antibody B 6.2 (IgG1), recognizing both (90 and 40 kDa) of the PRLR isoforms, were applied (33). Two slices were stained from each tumor. If the result differed, staining was repeated. For negative controls, antibody was replaced by mouse IgG in the same protein concentration. No nonspecific staining was found in these controls. The immunoreactive score (IRS) of the PRLR was estimated as the product of the nuclear staining intensity (0 = no staining reaction, 1 = weak staining, 2 = moderate staining, 3 = strong staining) and the percentage of positively stained cells (1 = <10% positive cells, 2 = 10–50% positive cells, 3 = 51–80% positive cells, 4 = >80% positive cells). It allows a maximum value of 12. To calibrate the staining intensity, reference slides with a score of 12 were applied (staining not shown).
Immunostaining of the ER and PR. The ER and the PR
were stained with the supersensitive ER/PR complete kit (containing the
monoclonal ER
antibody 1D5, IgG1, and the monoclonal PR-A/B antibody
PR 88, IgG1) following the manufacturer’s instructions (BioGenex Laboratories, Inc., San Ramon, CA). Diaminobenzidine was used as
the chromogene. The IRS values for ER and PR were estimated according
to IRS classification A* (34) as previously detailed for
the PRLR.
LCM and DNA isolation
Deparaffinized or frozen sections (7–10 µm) of mammary carcinomas were hematoxylin-eosin stained and dehydrated in rising alcohols. Cancer cells were separated from surrounding tissue by LCM as previously described (35, 36, 37). After fusion of cancer cells from three sections per carcinoma to the transfer film, it was inserted into a microcentrifuge tube. DNA was extracted by adding 50 µl Tris-HCl buffer (10 mM, pH 8.0, containing 1 mM EDTA, 1% Tween-20, and 0.04% proteinase K) at 37 C overnight. After heat inactivation (95 C, 8 min) 3–6 µl of the solution were directly used as template for PRLR-PCR amplification (high fidelity system, Roche, Mannheim, Germany). DNA of the control group was isolated from blood leukocytes after Ficoll gradient centrifugation. EDTA-blood (10 ml) was diluted with 10 ml PBS, pH 7.4, and overlaid on 8 ml Ficoll (Biochrom KG Seromed, Berlin, Germany). After centrifugation (1500 rpm, 20 min), leukocytes were collected by pipetting from the Ficoll layer and were washed twice with PBS. DNA was extracted using the Qiaamp blood kit (QIAGEN, Hilden, Germany).
Identification of exon-surrounding intronic regions
To amplify the introns, primers were chosen within the exons,
based on the recently published PRLR gene map
(38). Introns between exons 4–5, 6–7, 7–8, 8–9, and
9–10 were amplified by PCR using intron-spanning primers within the
exons (primer sequences available on request). PCR products were
purified using Qiaquick columns (QIAGEN). To obtain the
intron sequence between exons 5 and 6, a human bacterial artificial
chromosome (BAC) library (Research Genetics, Inc.,
Huntsville, AL) was screened by PCR. Two pairs of primers amplifying a
part of exons 5 and 6, respectively, were used. BAC 413-H-12 contained
both exons 5 and 6. DNA was isolated from a 250-ml overnight culture
using a plasmid extraction maxi kit (QIAGEN) and a
modified protocol (39). The exon-neighboring intron
regions were sequenced on an A.L.F. sequencer (Pharmacia Biotech, Freiburg, Germany) using Cy5-labeled primers. Based on
these sequences, primer pairs for the amplification of PRLR-coding
sequence were chosen (Table 1
).
|
All PCR reactions were carried out in a thermocycler 9600
(Perkin-Elmer Corp., Weiterstadt, Germany) using the high
fidelity system (Roche). PRLR gene amplification
conditions are described in Table 1
. The coding exons 3 (A285) to 10
were amplified using primers within the bordering intron regions. After
purification (Qiaquick columns, QIAGEN) products were
sequenced in both directions on an A.L.F. sequencer using Cy5-labeled
PCR primers and AmpliTaq DNA polymerase (Perkin-Elmer Corp.) corresponding to the manufacturer’s protocol.
Statistical analysis
Correlation analyses were performed using SPSS 9.0.1 for Windows (SPSS, Inc., Chicago, IL). We performed Pearson’s correlation analysis and considered P < 0.05 significant.
Results
Immunostaining for PRL, ER, and PR
PRLR-positive immunoreactivity could be detected in 30 of 41
(73%) specimens and was localized on the plasma membrane and in the
cytoplasm, but not within nuclei. PRLR staining was highly
heterogeneous in different parts of the same tumors and among different
patients (Fig. 1, A–D). Infiltrating
carcinoma cells within the stroma were usually strongly stained in
invasive ductal carcinomas. Intraductal tumor cells exhibited a
variable staining pattern. In some tissues, cells close to the basal
membrane were strongly PRLR immunoreactive, whereas luminal epithelial
cells appeared less intensely stained. In addition, the three included
lobular carcinomas were PRLR positive (Fig. 1, C and D). Contiguous
stroma cells showed scattered weak staining. Controls performed with
nonimmune IgG showed no staining (Fig. 1E). No significant correlation
was found between the IRS of the PRLR and histopathological grading,
tumor stage, lymph node involvement, age, or family history of patients
(Table 2). Also, the IRS values of the
PRLR and the ER or PR score showed no significant overall correlation
(PRLR: PR, r = 0.213; PRLR: ER, r = 0.026). Despite this, we
found that all tumors lacking both PR and ER were also PRLR negative
(refer to no. 4, 5, 7, and 14, Table 2). The known correlation between
the IRS values of the PR and the ER was confirmed (r = 0.609;
P = 0.0001; r is Pearson’s correlation
coefficient).
|
|
LCM allowed us to separate mammary carcinoma cells exclusively
from surrounding tissue and to use these cells for DNA isolation (Fig. 2). Sequencing of the intron-exon regions
enabled us to use intronic primers (except exon 3, as shown in Table 1)
for mutational analysis of the PRLR-coding sequence (location 285-2153
bp) (40) and to screen for aberrations of the intron-exon
junctions. Complete sequence analysis of the coding exons 3–10 of the
PRLR gene did not reveal any mutations or polymorphisms in the mammary
carcinoma (n = 30) or in the control group (n = 20).
|
PRL is an important hormonal and growth-promoting factor in mammary cells, and activation of the PRL-PRLR system may constitute a risk factor for human breast cancer development. In the present study we performed a complete mutational analysis of the PRLR-coding sequence and investigated a possible correlation of PRLR status and major tumor classification parameters. Immunostaining confirmed the presence of PRLRs in the majority of human mammary carcinomas (73%), corresponding to other studies that described PRLRs in 70–95% of breast carcinomas using immunohistological methods (24, 41, 42). Binding studies (28) and RIA analysis gave generally lower results (20–60%) (43, 44), whereas mRNA expression was found in 100% of the carcinomas in which leukocyte detection may not have been excluded (25, 45).
The staining intensity showed considerable variations among the same and different patients. Areas of strongly stained ductal carcinoma cells were found beside moderate or nonstained tumorous ducts and confirm the significant heterogeneity of PRLR appearance described previously (46). In accordance with former studies, (47), PRLR was detected on the plasma membrane and within the cytoplasm. Cytoplasmic PRLRs are probably the result of PRL-induced PRLR internalization as shown in PRLR-transfected cells (48).
In accordance with previous findings (26, 49), histopathological grading as well as tumor stage and family history in our study did not reveal a significant overall correlation with the IRS of the PRLR. Some studies, however, reported a higher concentration of PRLR in well differentiated vs. poorly differentiated breast carcinomas (28, 46, 47).
The lack of a significant correlation between the histopathologically proven extent of node involvement and the PRLR IRS suggests that PRLR expression, detected by immunohistochemistry, is not a useful marker predisposing tumor differentiation and lymph node involvement in human breast cancer.
The number of PRLR-positive cells was not related to the age of patients (38–83 yr) in our group, as has been previously reported (49). Likewise, increased mean PRLR levels in women aged 60–70 yr compared with those in either younger or older women were described, but without overall correlation of patient age and PRLR binding (28).
Relation of PRLR and PR/ER IRS
Cross-regulation of PR and ER with PRLR transcription (50) together with the recently demonstrated convergence of the progesterone and PRLR signaling pathway via signal transducer and activator of transcription-5 in breast cancer cells supports a correlated action of these hormones (51).
Conflicting data regarding the correlation between the mRNA and protein expression of ER and/or PR with PRLR are available. Although a significant positive correlation between ER/PR and PRLR numbers has been described by some groups in primary breast cancer and breast cancer cell lines (13, 24, 25, 47, 50, 52), this finding could not be confirmed by others (8, 28, 53, 54, 55).
In our study PRLR staining was negative in steroid hormone receptor-negative mammary carcinomas, whereas tumors positively stained for both ER and PR also expressed the PRLR. Despite this, the IRS of the PRLR showed no significant overall correlation to the IRS of the PR or ER. This may be due to the fact that PRLRs are regulated by various factors (e.g. androgens, glucocorticoids, and PRL) besides estrogen and progesterone levels (56). The well known correlation between ER and PR was clearly confirmed in our study group (57).
Screening for PRLR mutations
As mutations of the PRLR might change binding properties and signaling capacities, thereby varying the potential neoplastic effect of PRL, we investigated possible PRLR mutations in 30 patients with mammary carcinomas. The PRLR is closely related to the GH receptor, for which various mutations in different parts of the gene have been described (58, 59). Therefore, we could not restrict on predisposed exons and screened the whole coding region of the PRLR for mutations. Patients with a positive family history of breast cancer were included because genetic reasons are suspected in breast cancer cases of first degree relatives (2). With the exception of three invasive lobular carcinomas, all investigated tissues belong to the invasive ductal type. This represents the distribution of the disease in the population, where 85% of the tumors are invasive ductal carcinomas. The age of the patient group (median age, 57 yr) is of high clinical relevance due to the fact that more than 77% of the breast cancer patients are more than 50 yr old (60). Furthermore, the incidence of mutations in the early-onset breast cancer susceptibility genes, BRCA1 and/or BRCA2, should be relatively low (61, 62, 63). Bidirectional cycle sequencing was applied as a very sensitive technique for the detection of unknown mutations and heterozygous PRLR aberrations. Failure to detect mutated, and therefore often weakly expressed, alleles was minimized by using genomic DNA instead of cDNA. Furthermore, using genomic DNA, all PRLR splicing variants, characterized to date as a long and an intermediate isoform (40, 47), were included in the sequence screening. Because the development of mammary carcinomas may involve a stepwise sequence of somatic mutations similar to that known for colon cancer, mammary tissue, instead of blood, was employed for mutational screening. Carcinoma cells were separated by LCM from the contiguous stroma and adipose tissue to avoid masking of somatic mutations. In addition, blood samples from a group of 20 healthy volunteers were screened for PRLR polymorphisms.
We did not detect any polymorphisms (point mutations, insertions, or deletions) of the PRLR gene in the control DNA samples or in the DNA from 30 mammary carcinomas. No evidence of an association between somatic and/or germline mutations of the PRLR and breast cancer was found. The normal PRLR sequence of the included 6 PRLR-immunonegative tumors did not support a causal role of PRLR mutations for the lack of PRLR expression in PRLR-negative tumors.
By providing exon-external primer information and the BAC address, this study should contribute to delineation of the PRLR analysis of proximal gene sequences of larger patient groups. Other PRLR-linked aberrations, such as alterations of promoter-activating factors and promoter usage, especially in mammary tumors with varied PRLR expression, should be considered. The importance of the quantitative distribution of PRLR isoforms (40, 47) regarding their role in mammary tumorigenesis needs to be investigated further. Additionally, aberrations of intracellular cascades triggering the PRL signal and the paracrine/autocrine stimulation by PRL synthesized by mammary carcinoma cells itself as well as synergistic effects due to cross-talk with other hormones (e.g. steroids and GH) might be involved in the progression of human breast cancer. Besides the PRL tissue concentration, the balance of locally produced PRL variants, with regard to size, glycosylation, and phosphorylation, also could be an important factor determining the differential effects of PRL on mammary cells. As known, the heterogeneity of PRL affects its bioactivity (64, 65, 66, 67) and might result in distinct functions, as recently shown for a cleaved 16-kDa fragment of PRL with antiangiogenic activity (68).
In summary, we could confirm the presence of PRLR protein in 73% of 41 mammary carcinomas. No overall correlation was found for the PRLR IRS with pathological tumor parameters such as PR/ER IRS, pathological grading, tumor stage, nodal involvement, age, or family history. Therefore, PRLR protein expression of mammary carcinoma cells is unlikely as a useful marker for tumor progression and lymph node metastasis in human breast cancer. We could not identify any somatic or hereditary mutation or polymorphism in the coding region of the PRLR gene. We conclude that mutations of the PRLR are not a common event in human mammary carcinomas and do not seem to play a major role in the pathogenesis of such tumors. This suggests that systemic and locally produced PRL can act on tumor cell growth through functional receptors.
Acknowledgments
We are grateful to Silke Brauer and Sandy Laue for their excellent technical assistance.
Footnotes
This study was supported by a grant from the Deutsche Krebshilfe (10-1662; to S.R.B.).
Abbreviations: BAC, Bacterial artificial chromosome; IRS, immunoreactive score; LCM, laser capture microdissection; PRLR, PRL receptor.
Received October 18, 2000.
Accepted April 25, 2001.
References
This article has been cited by other articles:
 |
|
 |
|
  A. Vaclavicek, K. Hemminki, C. R. Bartram, K. Wagner, B. Wappenschmidt, A. Meindl, R. K. Schmutzler, R. Klaes, M. Untch, B. Burwinkel, et al. Association of Prolactin and Its Receptor Gene Regions with Familial Breast Cancer J. Clin. Endocrinol. Metab., April 1, 2006; 91(4): 1513 - 1519. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. S. Tworoger, P. Sluss, and S. E. Hankinson Association between Plasma Prolactin Concentrations and Risk of Breast Cancer among Predominately Premenopausal Women Cancer Res., February 15, 2006; 66(4): 2476 - 2482. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Mihaila, J. A. Gutierrez, M. L. Rosenblum, I. F. Newsham, O. Bogler, and S. A. Rempel Meningiomas: Analysis of Loss of Heterozygosity on Chromosome 10 in Tumor Progression and the Delineation of Four Regions of Chromosomal Deletion in Common with Other Cancers Clin. Cancer Res., October 1, 2003; 9(12): 4435 - 4442. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
