| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Regular Articles |
1 ILEX-CONICET, División Medicina Experimental, Instituto de Investigaciones Hematológicas e 2 Instituto de Estudios Oncológicos, Academia Nacional de Medicina, Buenos Aires, Argentina; and 3 Program of Gene Function and Expression, University of Massachusetts Medical School, Worcester, Massachusetts
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
It has been shown that BALB/c infected with MMTV(LA) strain show a 90% incidence of mammary tumors (3, 4, 5) . Similar to what was reported previously in BALB/cfRIII mice (6) , most of these tumors regress partially or even completely after delivery to reappear at the same site in subsequent pregnancies. Eventually, they become autonomous and grow independently of the female hormonal status (4 , 7) .
Breast carcinoma is one of the most common carcinomas in pregnant women. Recent reports show that these tumors have poor histological and prognostic features (8 , 9) . In addition, it has been reported that women diagnosed with breast cancer in the first year after childbirth have a significantly reduced survival (10) . These authors assume that physiological changes during pregnancy induce tumor growth during the preclinical stage. Pregnancy-associated breast cancers also show significantly reduced relapse-free and cancer-free survival rates (11) . Finally, a recent National Cancer Institute-sponsored workshop on "Early Reproductive Events and Breast Cancer"4 indicates that it has been well established that breast cancer risk is transiently increased after a term pregnancy. They also indicate that one of the "animal model gaps" is the lack of information about the relationship between pregnancy and the risk of preneoplastic lesions and the levels, determinants, and interactions of pregnancy-related mammotrophic factors, ligands, and receptors.
Herein we show that PD tumor transplants remaining dormant in virgin females for up to 300 days (experimental end point) are able to resume growth when hormone stimulated. Interestingly, after transplantation, these tumor cells show a clear progression to a hormone-independent behavior. Therefore, our data indicate that although hormone stimulation is required to rescue PD tumors from dormancy, pregnancy-independent (PI) clones prevail once tumor growth is established. Therefore, we believe that BALB/c MMTV(LA)-induced mammary tumors may provide a suitable animal model to help understanding pregnancy-associated breast cancer.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Cell Proliferation and Death Assays.
Mice received injection i.p. with 50 mg/kg 5-bromodeoxyuridine (BrdUrd) in saline solution (1x PBS) and sacrificed after 2 h. BrdUrd-labeled tissues of dormant and growing tumors (including small intestine as internal control) were fixed in 10% phosphate-buffered formalin and embedded in paraffin. Immunohistochemical staining for in vivo DNA-incorporated BrdUrd in 4-µm sections was performed using Cell Proliferation kit (Amersham International Plc., Little Chalfont, Buckinghamshire, United Kingdom) following manufacturer’s instructions. Tissue sections were lightly counterstained with 10% hematoxylin and 1000 cells were counted by systematic random sampling at 1000x magnification.
Apoptotic cells were identified by observation of dormant and growing tumor samples H&E-stained slides following morphological criteria (condensation and fragmentation of chromatin, cell shrinkage, separation from neighboring cells, and apoptotic bodies). Apoptosis was evaluated by counting at least 1000 cells split among a minimum of 10 randomly chosen fields at 1000x magnification.
The percentage of BrdUrd-labeled and apoptotic cells was calculated for each section, and the results were expressed as means ± SE. When the proliferation/apoptosis ratio was obtained for dormant and growing tumors, the P value was calculated using Student’s t test.
Morphological and Immunohistochemical Studies.
Tumors were fixed in 10% phosphate-buffered formalin and embedded in paraffin for histological studies. Then, paraffin sections were de-waxed in xylene and rehydrated through graded ethanol. Endogenous peroxidase activity was inhibited using 3% H2O2 in methanol. Sections were blocked in 5% normal goat serum and then incubated overnight with either rabbit polyclonal antiestrogen receptor (ER; clone MC-20) or antiprogesterone receptor (PR; clone C-20; Santa Cruz Biotechnology, Santa Cruz, CA) at 1:50 dilution in PBS or rabbit polyclonal anticytokeratin-1 (K1; Babco, Richmond, CA; 1:200 in PBS). All these procedures were followed by indirect immunoperoxidase detection using the Vectastain Elite ABC kit (Vector Laboratories Inc., Burlington, CA) following the manufacturer’s instructions. Negative controls were performed by replacing the primary antibody with normal rabbit serum. ER and PR content were evaluated by random counting of >1000 epithelial cells at 1000x magnification. The results expressed as percentage of positive cells for each section as means ± SE. The analysis of differentiation levels and presence of metasplasias was calculated using the 
2 test.
Inverse PCR Cloning.
We used a modified version of the published inverse PCR protocol (12, 13, 14)
. Briefly, 1 µg of DNA from tumor 2312 subline C was digested with EcoRI (Promega, Madison, WI) in a total volume of 30 µl. Digested DNA was then self-circularized by dilution and ligation using T4-DNA ligase (Life Technologies, Inc., Bethesda, MD) in a total volume of 600 µl at 23–26°C for 16 h. Circular DNA was precipitated with glycogen (1 µl), 3 M NaAcO (pH 5.2; 1:10 vol) and ethanol 100% (2.2 vol), and resuspended in 10 µl of nuclease-free water; 2 µl was used in primary PCR in a 50-µl reaction volume containing deoxynucleoside triphosphates (20 nmol each) forward and reverse primers (10 pmol each), 1x buffer B and ELONGASE Enzyme Mix (2 µl; Life Technologies, Inc.). PCR reaction, using outwards MMTV(LA)-long terminal repeat-specific primers, was performed as follows: 94°C 2 min, then 30 cycles of 92°C for 30 s, 60°C for 30 s, and 68°C for 7–10 min with a final extension step at 68°C for 12 min. Primary PCR product (0.01–1 µl) was then used as the template for the secondary PCR reaction that was performed under the same conditions using nested primers. In the primary PCR reactions, primers D' (5'-AGGAAGGTCGAGTTCTCC-3') and E' (5'-ATGATGAGCTTGTGGGGAAAG-3') were used. In the secondary PCR reaction, the following nested primers were used: A' (5'-AGTGTAGGACACTCTCGGG-3') and B' (5'-GAGTAAACTTGCAACAGTCC-3'). PCR product was separated in a 0.7–1.0% agarose gel and purified using the QIAquick Gel Extraction kit (Qiagen, Hilden, Germany) and cloned using the pGEM-T Easy Vector (Promega, Madison, WI). DNA sequencing was performed using the PRISM Big Dye Cycler Sequencing kit (PerkinElmer) on an ABI Model 373 DNA Sequencer (Applied Biosystem).
PCR and Reverse Transcription-PCR Analysis.
To determine the presence of MMTV(LA) insertion in the Int2/Fgf3 locus in different tumor passages belonging to tumor line 2312, PCR reactions that amplify a chimeric fragment corresponding to the 3' end of MMTV(LA) cDNA and the Int2/Fgf3 sequences next to it were carried out. For the MMTV(LA)-long terminal repeat, the primer B' (5'-GAGTAAACTTGCAACAGTCC-3') was used. For the Int2/Fgf3, a specific primer (5'-TGTATTGTCACAGCCTGAAG-3') was used. The amplification was carried out as follows: 95°C for 5 min followed by 30 cycles of 94°C for 1 min, 58–60°C for 1 min, and 72°C for 2 min, with a final extension step at 72°C for 5 min. The PCR product was analyzed by electrophoresis on a 2.0% agarose gel.
For reverse transcription-PCR analysis, RNA was obtained from tumor tissue using the SV Total RNA Isolation System (Promega) following the manufacturer’s instruction. cDNA was generated by reverse transcription (60 min at 40°C followed by 5 min at 90°C) with 2 µg of total RNA using Moloney murine leukemia virus reverse transcriptase, 1x reverse transcription buffer, oligodeoxythymidylic acid primer, 25 mM deoxynucleoside triphosphates mix and RNase inhibitor (Promega) in final 20 µl volume reaction. For Int2/Fgf3 expression (580 bp amplicon), the sense primer sequence was 5'-CCTGGAGATTACTGCGGTGG-3', the antisense primer was 5'-GCCAGGGCTCCAAGTGAATC-3', and the PCR was performed for 30 amplification cycles (94°C for 33 s, 60°C for 1 min, and 72°C for 1 min) with initial 3 min denaturalization at 94°C and final 5-min extension at 72°C.
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. We have also found that E2 treatment by itself can reinitiate tumor development (data not shown).
|
), and a slightly higher level of apoptosis (7.5% ± 1.0 by morphological observation in H&E-stained tumor samples; n = 6). When these tumors resumed growth after hormone stimulation, they showed a small decrease in the number of dying cells (5.3 ± 0.8%, n = 5) and an important increase in the number of proliferating cells (13.1 ± 1.3%, n = 5, P < 0.001; Fig. 1B
). Similar numbers were obtained in the following transplant generation (Fig. 1C)
. In these cases, in which tumor transplants were carried out in both no. 4 inguinal fat pads, very similar rates of BrdUrd labeling and apoptosis of dormant as well as growing tumors in the same mouse were detected. Likewise, tumor implants resumed growth at similar time in both flanks after hormone stimulation.
|
and PR (ER, 49.5 ± 2.6%; PR, 45.6 ± 3.7%; n = 5). After hormone stimulation, tumors reassumed growth with similarly high ER (Fig. 1F)
and PR levels (ER, 59.3% ± 4.3; PR, 57.4% ± 4.4; n = 7). However, when transplanted, these cells steadily developed into tumors without the requirement of hormone stimulation and showed low levels of ER (Fig. 1G)
and PR (ER, 9.5% ± 2.7; PR, 9.7% ± 2.8; n = 6; P < 0.001). Southern blot analysis of these PI transplants hybridized with an MMTV-long terminal repeat probe showed very stable band patterns through all of the parallel passages (Fig. 2)
, suggesting that PI tumors arise from PD dormant passages as clonal or quasi-clonal populations.
|
Morphology and Patterns of Progression.
It has been shown recently that an accurate morphological evaluation of mammary tumors from genetically engineered mice (GEM) can reveal the activated molecular pathway (15)
. Therefore, we decided to perform a detailed histological study in our model in an attempt to identify signaling cascades that could be involved in BALB/c MMTV(LA)-induced mammary tumor progression.
Supporting our data reported previously (7) , the most common histological pattern in our tumors is the Type-B (Dunn’s classification; Ref. 16 ) or the glandular/papillary morphology (17) . In addition, taking into account the association reported between the molecular-activated pathway in different GEM and their tumor morphology, we found that 58 of 59 analyzed samples corresponded to the activated Wnt pathway, whereas only one showed morphological features similar to those described for the Erb/Ras pathway (15) . However, in our tumor model, a few morphological features clearly differed from the GEM mammary tumors classified in such a category. For instance, (a) no adenosquamous, pilar, type-P, nor myoepithelial tumors were found in the MMTV(LA)-induced tumors; (b) in GEM/Wnt pathway mammary tumors, fibrosis was generally associated with inflammatory infiltrates [although dense stroma is also commonly found in MMTV(LA) tumors, there was no lymphocytic infiltration in the samples studied]; (c) the presence of necrotic cells has not been reported in GEM mammary tumors, although in MMTV(LA)-induced tumors, their presence is common, particularly in those tumors resuming growth after dormancy and in their passages.
Even taking into account the predominance of the Wnt pathway phenotype in the BALB/c MMTV(LA)-induced tumors, we have observed morphological differences between tumor passages that spontaneously progressed to a PI phenotype [PI/not dormant (ND)] and those which went through long dormancy periods before losing hormone regulation [PI/dormant (D)]. Although the first had a higher tendency to display an acinar phenotype with secretory activity, the PI/D passages commonly showed a solid pattern with keratinized foci (Table 2)
. The lower degree of differentiation found by morphological observation was then confirmed by the lower ß-casein expression in solid tumors (data not shown). Interestingly, in solid tumors as well as in solid areas of mostly glandular/papillary tumors, the presence of squamous metaplasia foci was often detected. These structures were significantly more frequent in PI/D than in PI/ND tumors (Table 2
; Fig. 1, I–K
). Moreover, these squamous structures showed high expression of K1 cytokeratin (Fig. 1L)
. This protein was not only found in the morphologically detectable foci but also in epithelial cells close to these structures.
|
). One virgin control mouse was impregnated 258 days after tumor transplantation, whereas another received implantation with pellets of estrogen and progesterone 300 days after tumor implantation. In both cases dormant tumors started to grow (subline 2312 B; Fig. 3B
and subline 2312 C; Fig. 3C
). In the three cases, the second transplant generation tumors were able to develop in virgin females, confirming their PI status. Southern blot analysis of MMTV insertions revealed different band patterns in the three PI sublines. We then used the inverse PCR approach to clone the viral insertion site that corresponded to the strong band at about 5.4 kb obtained in the subline 2312 C. After sequencing, it was determined that the fragment corresponded to a MMTV(LA) insertion that mapped in the 3'end of the Int2/Fgf3 gene (Fig. 3D)
. To confirm the presence of this insertion in the DNA from which the clone was obtained and to determine whether this mutation was also present in tumors from the other sublines, PCR reactions were carried out with primers corresponding to the MMTV(LA) sequence and the Int2/Fgf3 3' region. We then found that both sublines that had arisen independently from dormant tumors (sublines 2312 B and C) showed the same insertion site. However, this mutation was absent in PI tumors from subline 2312 A (Fig. 3E)
. By reverse transcription-PCR we confirmed that this insertion resulted in higher expression of Int2/Fgf3 in sublines B and C when compared with subline A (Fig. 3F)
. The presence of the same insertion in sublines B and C strongly suggests that this mutation was present in a subpopulation of the primary tumor. Therefore, after dormancy, clones containing that insertion were twice selected independently. These results support the hypothesis that dormancy exerts a specific selection pressure on a polyclonal PD tumor cell population and elicits the predominance of clones harboring specific mutations.
|
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The association between mammary tumor progression and loss of ER and PR has been observed in humans (24 , 25) as well as in animal models (18 , 19 , 21 , 26) . Interestingly, the apparent paradox shown by our results (i.e., that a hormone-rich environment such as pregnancy is determinant for PI ER–PR–tumor cells to arise) is commonly found in pregnancy-associated breast carcinomas that show low expression of hormone receptors (8 , 9) . In fact, it has been suggested that an explanation for the poor breast cancer survival observed in women recently after childbirth is the lower proportion of hormone-receptor-positive tumors (27) . Therefore, we believe that this particular trait makes this model particularly suitable for understanding what triggers breast cancer development during or shortly after pregnancy.
The results showed herein confirm our previous data showing that in BALB/c, PD MMTV(LA)-induced mammary lesions can arise from polyclonal cell populations, although PI tumors are mostly monoclonal (7) . Similarly, it has been demonstrated that hormone-dependent GR mouse mammary tumors are oligoclonal with respect to MMTV production (28) , extra MMTV integration (18 , 29) , DNA rearrangements, and gene activation (22) . In addition, in this strain as well as in BR6 mammary tumors (21 , 30) , Int2/Fgf3 involvement in the rise and progression of pregnancy-dependent lesions has been well established.
Unfortunately, our data does not explain how the selection of PI clones takes place during dormancy. However, several clues are provided. First, PD tumors in virgin females do not only consist of noncycling (G0) cells, they also contain a small cycling cell population. Similar results were obtained with another model of endocrine-related tumor dormancy (31) . This is relevant because in a dynamic cell population in which proliferation as well as apoptosis still occur, a cell subpopulation harboring a selective significant mutation can increase in relative size without increasing tumor mass. Therefore, an apparently stable situation such as tumor dormancy becomes a rather dynamic state during which relevant processes of cell selection take place.
It has been shown that squamous differentiation is the most characteristic histological pattern of GEM mammary tumors with activated Wnt/ß-catenin signaling components (15 , 32) . However, spontaneous mouse mammary tumors that are primarily associated with Wnt activation (33) , seldom have squamous metaplasia (16) , and squamous differentiation is less frequent in Wnt 1 and Wnt10b transgenic animals than in GEM with mutations in genes downstream from Wnt (15) . Our morphological data showing a higher tendency to form solid cords together with trans-differentiated foci and a very low incidence of pure acinar structures in tumors arising from long dormancy periods suggest that direct Wnt activation would not be selected during dormancy. Alternatively, mutations that alter expression of genes involved in the Wnt pathway but are either downstream or activate it in a still unclear way are more prone to be selected. This idea is supported by our results in which clones containing an MMTV insertion in the Int2/Fgf3 locus and overexpressing this gene were specifically selected during tumor dormancy.
Despite its high frequency in GEM Wnt pathway tumors, squamous differentiation has not been commonly reported in the classical MMTV-related mammary tumor models. The association of neoplastic glands and stratified squamous epithelium has been reported to occur occasionally in noninfected mouse strains such as FVB and more frequently in carcinogen-treated mice (17) . On the other hand, association of inflammatory infiltrates and squamous metaplasia has been found in GEM mammary tumors (15) , whereas lymphocytic infiltration was not commonly observed in tumors displaying squamous differentiation in our model. Therefore, inflammatory infiltrates might be the outcome of transgene expression rather than a consequence of Wnt pathway activation.
The association between hormone-dependency and Fgf3 function is a complicated issue. GEM models have demonstrated that Fgf3 overexpression induces pregnancy-dependent hyperplasias. Nevertheless, MMTV insertions in the Int2/Fgf3 locus are more frequently found in BALB/cfC3H (pregnancy-independent) than in BALB/cf RIII (pregnancy-dependent) tumors (1) . Therefore, activation of this gene can be exclusively linked to neither hormone-dependent nor independent tumor behavior. Our results show that Fgf3 expressing tumors cells can also show high levels of hormone receptors and milk proteins and, eventually, they may trans-differentiate. This cell plasticity could be attributable to the ability of Fgf3 to activate a variety of signaling pathways, including mitogen-activated protein kinase, protein kinase B, and protein kinase C (34) as argued previously in an attempt to explain the variety of histological features observed in transgenic mice overexpressing Int2/Fgf3 (32) . New experiments are underway to determine activation of these pathways in our model.
In conclusion, we postulate that during dormancy, cells harboring specific mutations that allow tumor cell survival in a hormone-poor environment are being selected. However, these cells still require extra proliferative signals (e.g., pregnancy hormones) to develop into fully growing tumors. Then, as a response to that stimulus, ER+PR+ tumors arise, but their clonal composition is now different from that of the tumor that was implanted 5–10 months earlier. These cell populations seem to require pregnancy hormone stimulus only to start growing, whereas transplantation itself would provide a nonspecific stimulus that helps to abrogate hormone control. In this scenario, we believe that mutations associated with Wnt pathway activation could facilitate the required cell versatility to adapt, survive, and proliferate in very different circumstances.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Requests for reprints: Edith C. Kordon, División Medicina Experimental, Academia Nacional de Medicina, J. A. Pacheco de Melo 3081, (1425) Buenos Aires, Argentina. Fax: 5411-4803-9475; E-mail: ekordon{at}hematologia.anm.edu.ar
4 http://nci.nih.gov/cancerinfo/ere. 
Received 12/29/03. Revised 4/21/04. Accepted 6/ 2/04.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. Gattelli, M. N. Zimberlin, R. P. Meiss, L. H. Castilla, and E. C. Kordon Selection of Early-Occurring Mutations Dictates Hormone-Independent Progression in Mouse Mammary Tumor Lines J. Virol., November 15, 2006; 80(22): 11409 - 11415. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. J. Epstein Maintenance Therapy to Suppress Micrometastasis: The New Challenge for Adjuvant Cancer Treatment Clin. Cancer Res., August 1, 2005; 11(15): 5337 - 5341. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
) and parous (
) hosts. Southern blot analysis indicates different mouse mammary tumor virus (MMTV) insertion sites in each tumor subline (arrowheads). In C (Southern blot), the big arrow shows the fragment cloned by inverse PCR that corresponded to an insertion in the non-coding 3'end region of the Int2/Fgf3 locus. Diagram of MMTV(LA) insertion site: E1, E2, and E3 exons, full black boxes; 5'and 3'non-coding regions, gray checked boxes; MMTV cDNA, dark gray boxes; small arrows, position of primers designed for testing presence of the insertion in different DNAs; insertion site, nt6936. The long terminal repeat (LTR)-Int2/Fgf3 sequence is shown below (MMTV-LTR sequence was underlined; D). PCR assays for detecting MMTV insertion (E) and reverse transcription-PCR assays for analyzing Int2/Fgf3 and actin expression (F) in tumors from 2312 sublines A, B, and C. Normal mammary gland (M), no template (H2O) was used as negative controls. bp, base pair.