| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
and c-myc Transgenic Mice
From the Department of Pathobiological Sciences,* School of Veterinary Medicine, University of Wisconsin–Madison, Madison, Wisconsin; and the Department of Anatomy, Microbiology, and Pathology, University of Minnesota Medical School–Duluth, Duluth, Minnesota
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionReferences |
|---|
(TGF-) or c-myc. In transgenic glands, before lesion development, epithelial cell turnover was enhanced overall compared with nontransgenic glands, indicating that aberrant cell turnover in normal epithelia may contribute to tumorigenesis. In addition, in tumor-containing glands, proliferation in normal epithelia was higher than in tumor-free transgenic glands, suggesting these cell populations influence one another. Finally, although c-myc glands displayed a uniformly high epithelial cell turnover regardless of age, cell turnover was reduced with aging in nontransgenic and TGF- mice, indicating that some growth and death regulatory mechanisms remain intact in TGF- epithelia. These observations support the evaluation of cell turnover as a biomarker of cancer risk and indicator of prevention/treatment efficacy in preclinical models and warrant validation in human breast cancer.
The net gain of cells, either through excessive proliferation or failure of cell death, is a hallmark feature of cancer. Abnormal patterns of cell turnover observed in breast tissue with hormone replacement therapy may indicate a higher risk of cancer development.1,2 Understanding the dynamics of cell proliferation and death during carcinogenesis and in normal breast epithelia that remain healthy throughout life is critical to our understanding of the cell turnover abnormalities that accompany cancer development.
Genetically engineered mouse models of mammary cancer have been generated to mimic human disease. These mouse models demonstrate similarity in the natural history of disease progression, genomic alterations, gene and protein expression patterns, hormone-responsiveness, and metastasis to distant sites.3 In addition, the histopathology of mammary lesions in genetically engineered mice, recently reviewed by a panel of pathologists who compared tumors in mice to humans, share many features with human breast cancer.4 Mouse models of mammary cancer also have been used in preclinical testing of preventive and therapeutic agents.5 Unlike xenograft models, early-to-late-stage-specific responses to chemo- or nonchemopreventives can be evaluated in genetically engineered mice. In addition, preventives can be tested for efficacy against specific oncogenes representing molecular anomalies found in clinical breast tumors, and in a physiologically relevant context. Encouraging global gene expression data revealed common molecular anomalies in tumors originating from several different transgenic mouse models of mammary cancer in which different oncogenes were overexpressed,6 suggesting that potential targets for prevention and therapy can be directed toward a few general molecular pathways. Regardless of the mode of intervention or the underlying molecular maladies, inhibition of proliferation is likely to be a common feature of successful treatment of breast cancer development. Data from a recent study of 21 patients support this assertion in that a high level of breast epithelial cell proliferation after neoadjuvant therapy was associated with cancer recurrence or death.7 Extrapolation of these observations suggests that evaluation of epithelial cell turnover activities of normal and abnormal epithelium of breast biopsies may provide another measurement of cancer risk. These data could be obtained readily from biopsies collected for the purpose of establishing the underlying character of suspicious regions found on imaging or palpation. Epithelia obtained from reduction mammoplasties could act as controls. Alternatively, a core needle biopsy of unaffected regions would provide information about site and patient-to-patient variation.
Our objective in this study was to test the hypothesis that increased cell turnover in normal mammary epithelium is a common feature preceding and accompanying cancer development. To accomplish this we evaluated proliferative and apoptotic activities in lobules, ducts, hyperplasias, and tumors in transgenic mice overexpressing transforming growth factor (TGF)- or c-myc in mammary epithelia before and during tumorigenesis. We also characterized normal proliferation and apoptosis in wild-type non-TG mice that remained mammary cancer-free. We report that changes in cell turnover preceded the development of mammary hyperplasias and tumors and that cell turnover events were qualitatively similar for both mouse models.
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
WAP-TGF- line 3573-2 transgenic mice, designated TgN(WAPTgfa)215Bri and WAP-c-myc line 3507-1, TgN(WAPMyc)212Bri, have been described.8
These mice were generated in the (C57BL/6xSJL)F2 background and were backcrossed into the FVB/N (Taconic, Germantown, NY) strain at least 15 generations before use in this study. In this report, wild-type non-TG FVB/N mice are designated as non-TG, WAP-TGF- mice are designated as TGF-, and WAP-c-myc mice are designated c-myc mice. All mice were housed separately from males and in Association of and Accreditation of Laboratory Animal Care, International-accredited facilities and handled in accordance with the Guide for the Care and Use of Laboratory Animals. All studies were approved by the institutional animal care and use committee.
Sampling Design, Age, and Degree of Lobularization
The numbers of mice and epithelial histotypes assessed for proliferation and apoptosis is presented in Table 1
. Eighteen non-TG, 20 TGF-, and 17 c-myc virgin mice were evaluated for proliferation. For apoptotic assessments, the same mammary glands from non-TG mice were evaluated. For TGF-, mice the same mammary glands were evaluated, and mammary glands from four additional animals also were assessed. For c-myc mice, an additional mouse 
12 weeks was evaluated and mammary glands from a single mouse >16 weeks were not available for evaluation. When nulliparous female mice of the FVB/N strain reach 12 weeks of age, arborization of the ductal epithelial tree is complete, and ducts extend to the edges of the mammary fat pad. Therefore, data analysis was carried out on mice segregated into two physiologically relevant age groups: mice 12 weeks with immature mammary glands and mice >16 weeks with fully mature mammary glands. The age distributions for mice 12 weeks and >16 weeks are as follows. For mice 12 weeks, ages ranged from 51 to 81 days with mean ± SE of 65 ± 5, 67 ± 6, and 65 ± 1 for non-TG, TGF-, and c-myc mice, respectively. For mice >16 weeks, ages ranged from 132 to 286 days with mean ± SE of 194 ± 11, 176 ± 9, and 194 ± 11 days for non-TG, TGF-, and c-myc mice, respectively. There were no statistical differences between groups (data not shown). A single cross-section of one to two mammary glands was evaluated for each mouse. The total number of epithelial histotypes evaluated for cellular proliferation and apoptosis also are presented in Table 1
. The morphology of epithelial histotypes (ie, lobules, ducts, hyperplasias, and tumors) is shown in Figure 1
. In addition, animals were categorized by the degree of lobular development. A modified version of the staging system developed by Fata and colleagues11
was used in which lobular development was graded from 0 to 3 with increasing lobularization. In this study lobular development with an average of <1.2 alveolar lumens/lobule was graded as stage 0 whereas versus 
1.2 alveolar lumen/lobules in which animals were graded as stage 1.
|
|
Female virgin mice were injected with 200 mg/kg body weight BrdU (no. B5002; Sigma, St. Louis, MO), a nucleotide analog incorporated into DNA of proliferating cells, and sacrificed 1 to 2 hours later. The fourth mammary gland from each mouse was collected. Tissues were fixed in 10% neutral buffered formalin at room temperature for 18 to 24 hours, dehydrated in ethanols, embedded in paraffin, sectioned (6 µm), and mounted on glass slides. Mice were not staged for estrus, but the extent of lobular development was determined by counting the number of alveolar lumens/lobule. This method was a direct modification of a system developed by Fata and colleagues11 who found that a higher degree of lobular development was positively correlated with circulating progesterone levels.
BrdU Labeling for the Identification of Proliferating Cells
Methods used for immunohistochemistry were described previously.9 Briefly, slides were deparaffinized in Citri-Solv (Fisher Scientific, Pittsburgh, PA), hydrated in graded alcohols, and exposed to 0.5% H2O2 in methanol to block endogenous peroxidase activity. Slides then were boiled in a solution of 0.1 mol/L Tris for antigen retrieval and incubated overnight with rat monoclonal anti-BrdU (MAS-250; Accurate-Scientific, Westbury, NY) at a dilution of 1:40 in 0.5% powdered milk dissolved in phosphate-buffered saline (PBS). The following day slides were rinsed with PBS, treated with rat-specific link antibody (BioGenex, San Ramon, CA) for 30 minutes, rinsed, incubated with the enzyme label peroxidase streptavidin (BioGenex) for 30 minutes, and treated with 3,3'-diaminobenzidine (no. D-4293; Sigma) for up to 10 minutes. Slides were counterstained in Nuclear Fast Red (Polyscientific, Bay Shore, NY) for 1 minute, dehydrated, and mounted under a glass coverslip. Cells labeled with BrdU displayed brown nuclear stain. Positive control tissues included duodenum or lymph nodes from mice used in this study. Negative controls were mammary gland or duodenum incubated with a nonspecific antibody.
Identification of Apoptotic Cells
Slides were deparaffinized in Citri-Solv (Fisher Scientific), hydrated in graded alcohols, and stained with hematoxylin and eosin (H&E). Apoptotic cells were identified at x400 magnification by the presence of pyknotic nuclei and retracted cytoplasm. For some mammary glands, terminal dUTP nick-end labeling (TUNEL) staining was carried out as described previously10 using the in situ cell death detection kit, POD (no. 1 684 817; Roche, Penzberg, Germany). Apoptotic cells were identified by addition of fluorescein-labeled nucleotides to DNA ends by terminal deoxynucleotidyl transferase. Anti-fluorescein-specific antibody Fab fragments conjugated to peroxidase were applied to the tissue, followed by 3,3'-diaminobenzidine. Slides were counterstained with nuclear fast red and mounted under glass coverslips.
In a pilot study, we compared apoptotic activity in normal and hyperplastic epithelia using two methods: a histological method evaluating the number of pyknotic cells on H&E-stained tissue sections (Figure 1F)
with an immunohistochemical method of identifying cells with degraded DNA nucleotides using a TUNEL assay. In mammary glands from three non-TG, three TGF-, and three c-myc mice, the percentage of cells identified as apoptotic were determined by these two methods in adjacent histosections and were not different for lobules or ducts (P 0.05). As a result, the most direct approach of evaluating histosections for pyknotic cells was used.
Proliferative and Apoptotic Activities
Epithelial histotypes including lobules, ducts, hyperplasias, and cystic and solid tumors were evaluated for BrdU labeling (LI) and apoptotic labeling (AI) indices. As previously described,9
lobules were defined as containing one to four adjacent alveolar lumens in cross-section, each surrounded by a single layer of luminal epithelial cells. Ducts could be found in longitudinal and cross-section. Ducts in cross-section contained secretions within a single lumen, which generally had a larger diameter than an alveolar lumen of a lobule. Terminal endbuds in animals 12 weeks of age were grouped with lobules. Structures containing five or more adjacent lumens were classified as lobular hyperplasias. These structures, which resemble lobules at mid-pregnancy, also have been called hyperplastic alveolar nodules.8
Cystic tumors were lined with a single layer of simple and/or complex papillary epithelium and were present in TGF- mice but were not found in non-TG or c-myc mice. Solid tumors contained tightly packed epithelium with few or no cystic spaces. To minimize selection bias, for each slide representing a single individual, at least 10 normal-appearing ducts and 10 normal-appearing lobules were examined, providing a minimum of 500 cells per histotype per slide with the following exceptions: in one TGF- mouse >16 weeks, only nine lobules were evaluated, and in a single c-myc mouse, there were zero lobules to evaluate because the gland was filled with hyperplasias. Fewer than 10 ducts (n = 5, 7, 8, 9, 9) were present in slides representing five TGF- glands. For hyperplasias, a minimum of 10 hyperplasias and at least 1000 cells were evaluated. For cystic and solid tumors, all epithelial cells visible in cross-section were counted. When slides contained multiple tumors, each was counted separately. BrdU staining of normal ducts, lobules, and hyperplasias is shown in Figure 1
. The number of epithelial histotypes identified and evaluated for proliferative and apoptotic activities is depicted in Table 1
.
Proliferation
To describe the focal nature of proliferative activity in mammary glands of non-TG and transgenic mice, the percentage of epithelial histotypes that contained any BrdU-labeled cells for each individual was determined and averaged over all mice for each mouse group (non-TG, TGF-, c-myc). These values are depicted in Table 2
. Next, within each BrdU-containing epithelial unit, we also determined the percentage of BrdU-labeled cells and report these as BrdU labeling indices. A BrdU labeling index (LI) was calculated by dividing the number of BrdU-labeled cells by the total number of cells counted for each epithelial histotype, ie, lobules, ducts, hyperplasias, and tumors. The mean LI for each histotype within an individual then was averaged over all individuals within a mouse group, and these data are represented in Table 3
.
|
|
Finally, we measured two features of mammary epithelial apoptotic activity. First, we calculated for each epithelial histotype (lobules, ducts, hyperplasias, and tumors) the percent that displayed at least one pyknotic cell. This allowed an evaluation of the focal nature of apoptotic activity in the gland of non-TG or transgenic mice and is reported separately for each epithelial histotype in Table 4
. Second, for those epithelial units with at least one pyknotic cell, we calculated apoptotic indices. An apoptotic index (AI) was calculated by dividing the number of pyknotic cells by the total number of cells counted for each epithelial histotype. The mean AI for each histotype within an individual then was averaged over all individuals within a mouse group. This provided a measure of the percentage of pyknotic cells for each epithelial histotype and is reported as AI in Table 5
.
|
|
Two-tailed Student’s t-test were used to determine significant differences between the percentage of each epithelial histotype (lobules, ducts, hyperplasias, and tumors), which contained BrdU-labeled cells or pyknotic cells, as well as LI or AI for each mouse group (non-TG, TGF-, and c-myc). Analysis of variance was used to compare proliferative and apoptotic activities between epithelial histotypes within a mouse group. Separate analyses were performed for each age group (12 weeks and >16 weeks). Comparisons between age groups but within genotype were carried out using Student’s t-test. Spearmen correlation coefficients were determined for LI and AI in lobules, ducts, hyperplasias, and tumors from non-TG, TGF-, and c-myc animals. Within each mouse group, Student’s t-tests were used to determine whether cell turnover activities were altered with the extent of lobular development (<1.2 versus 1.2 number of lumens/lobule), and linear regression was used to depict the relationship. Linear regression also was used to describe the relationship between age and measurements of cell turnover activities within each mouse group. Data comparisons with P values 0.05 were considered significant. All statistical analyses were carried out using Prism Version 4.0 (GraphPad, Inc., San Diego, CA).
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
In non-TG mice 12 weeks, BrdU-labeled cells were found in one third of lobules and one half of ducts (Tables 2 and 3)
. In mice >16 weeks, the percentage of lobules with BrdU-labeled cells was reduced to 7% (Table 2)
. This reduction likely reflected the loss of terminal endbuds in FVB/N mice at 
12 weeks of age once ductal arborization of the mammary fat pad is complete. LI was similar in glands from both age groups (Table 3)
. According to our definition of hyperplasia, ie, lobules containing 5 ductules, mammary glands from 2 of 13 non-TG mice were found to contain a total of 17 hyperplasias. Half of the hyperplasias contained BrdU-labeled cells and those with BrdU-labeled cells had a LI of 1.6 (Tables 2 and 3
, respectively). We were also interested in the effects of lobularization and aging on cell turnover dynamics. In non-TG animals, proliferative activity was not altered with degree of lobular development.
TGF- Mice >16 Weeks Display Enhanced Proliferation
Differences in proliferation were not found in TGF- mice compared with non-TG mice 12 weeks. However, in mice >16 weeks, the overall percentage of proliferation in lobules of glands from TGF- mice was sevenfold higher than in non-TG glands (Table 2)
. Instead of a reduction in proliferation typically observed in non-TG mice >16 weeks compared with mice 12 weeks, proliferation remained high in glands from TGF- mice >16 weeks: lobular LI was approximately twofold higher than in age-matched non-TG mice and ductal LI was threefold higher (Table 3)
. In addition, in contrast to non-TG mice, which did not display differences in proliferative levels with lobular development, there was a significant increase in ductal proliferation in TGF- glands with greater lobular development (P = 0.024, Figure 2A
), suggesting a growth factor effect in vivo. Furthermore, correlations between proliferative activities and degree of lobular development were also identified in TGF- glands (r2
= 0.31, P = 0.02 and r2
= 0.38, P = 0.009, for lobules and ducts, respectively). Hyperplasias in TGF- glands were different from hyperplasias found in non-TG glands in that the majority of hyperplasias in TGF- mice 1) displayed a twofold greater level of BrdU-labeled cells than in non-TG glands and 2) had significantly more cells (see details in the Lobules and Hyperplasias are Larger in Transgenic Mice and Develop Earlier section). All TGF- tumors displayed BrdU-labeled cells (Table 2)
. There was an elevated LI for ductal epithelium in TGF- versus non-TG glands. In addition, LI was similar in hyperplasias and tumors found in TGF- glands. Finally, normal lobules residing in tumor-bearing mammary glands displayed significantly increased proliferative levels (70% with BrdU-labeled cells) compared with levels of normal lobules in tumor-free TGF- glands (29%, Figure 3
).
|
|
Unlike the relatively low percentage of BrdU-labeled cells found in non-TG mice, nearly 100% of lobules, ducts, hyperplasias, and tumors contained BrdU-labeled cells in c-myc mice (Table 2)
. Differences in proliferative levels were observed already in 12-week-old c-myc mice in that alveolar LI was approximately threefold elevated compared with age-matched non-TG mice (16.2 versus 5.6%, respectively; Table 3
). Increased proliferation in lobules compared with ducts also was found in c-myc mice >16 weeks, suggesting c-myc effects were greatest in lobules. Unlike TGF- glands, epithelial proliferation in c-myc glands displayed progressive increases in LI from ducts to tumors so that LI in ducts < lobules < hyperplasias < tumors (6.6, 12.9, 18.1, and 24.9%, respectively; Table 3
). In addition, the percentage of normal ducts containing BrdU-labeled cells was elevated significantly in c-myc glands with tumors (97%) compared with tumor-free glands (82%, Figure 3
). Finally, not only did c-myc glands display, on average, larger lobules than non-TG and TGF- glands (see the Lobules and Hyperplasias are Larger in Transgenic Mice and Develop Earlier section), but the extent of proliferation was near 90% regardless of lobular size (Table 2
; Figure 2
).
Proliferation of Normal Epithelium in Aging Mice
We also were interested in the effects of aging on proliferation. In non-TG mice, proliferation was decreased with aging in mice between 18 and 40 weeks, as manifested by a reduction in the percentage of BrdU-labeled cells (Figure 4)
. In TGF- or c-myc mice >16 weeks, there were no changes in proliferative levels with aging.
|
In non-TG mice 12 weeks, levels of pyknosis were one third to one half greater than in mice >16 weeks (Tables 4 and 5)
. Although the extent of lobular development was not associated with altered proliferative activities in non-TG mice, higher levels of pyknotic lobular and ductal cells were observed in mice with mammary glands with more extensive lobular development (P = 0.017 for alveoli, Figure 2B
; P = 0.008 for ducts, data not plotted).
Transgenic Mice Displayed Altered Apoptotic Activity in Both Age Groups
TGF- mice 12 weeks displayed a reduced level of apoptosis compared with age-matched non-TG mice. This reduction was manifested in the percentage of alveoli and ducts with pyknotic cells: 14 versus 45% for lobules and 38 versus 82% in ducts for TGF- versus non-TG glands, respectively (Table 4)
. This relationship shifted when non-TG mice reached >16 weeks of age, as the apoptotic percentage dropped to half that of younger non-TG mice 12 weeks. In TGF- mice >16 weeks, apoptotic levels increased to levels 1.5- to 3-fold higher than age-matched non-TG mice (Table 4)
. Nearly all TGF- hyperplasias and tumors contained pyknotic cells although AI levels were the same as those found in normal epithelia of both non-TG and TGF- glands (Table 5)
.
Apoptotic Activity in Aging Mice
The percentage of lobules with pyknotic cells was reduced with age in non-TG mice between 18 to 40 weeks (Figure 5A)
. Glands from TGF- mice of similar age mimicked the pattern displayed in non-TG glands, in that apoptotic levels also decreased with age (Figure 5B)
. However, tumors that developed in older TGF- mice displayed a greater level of pyknosis than those found in younger mice (Figure 5B)
. However, apoptosis in c-myc mammary epithelia, unlike that of non-TG and TGF- mice, was not altered with adult aging.
|
We also wished to determine whether the average cell number for lobules and hyperplasias was different in non-TG and transgenic mice. A lobule consisted of a grouping of alveoli made up of secretory epithelial cells with the potential to produce milk products during lactation. An average lobule in non-TG mice consisted of a total of 23.1 ± 1.1 (mean ± SE) cells together comprising, on average, 1.2 alveolar lumens. A typical lobule in TGF- mice was larger, consisting of 27.8 ± 2.1 cells arranged in 1.4 alveolar lumens, whereas lobules in c-myc mammary glands typically consisted of a total of 28.0 ± 1.9 cells arranged in 2.0 alveolar lumens. The average cell number/lobule was significantly higher in c-myc glands compared with non-TG glands (P = 0.03). Hyperplasia incidence, latency, size (number of lobules/hyperplasia), and cell turnover dynamics also differed significantly in non-TG compared with transgenic mice (Tables 2 through 5)
. First, the incidence of hyperplasias was lower in non-TG mice and latency longer. Only 2 of 18 non-TG mice displayed a total of 17 hyperplasias. Second, hyperplasias were found exclusively in mice >16 weeks and not in mice 12 weeks unlike c-myc mice. Third, hyperplasias were significantly smaller in size in non-TG glands, which contained 5 to 13 ductules (6.7 ± 0.5, mean ± SE). In contrast, 9 of 20 TGF- mice displayed hyperplasias and all were >16 weeks. Hyperplasias were more numerous (n = 39) and larger (range, 5 to 31 alveolar lumens; 11 ± 1, mean ± SE; P = 0.006) than those found in non-TG mice. In c-myc mice, hyperplasias were identified in three of four mice <12 weeks and in all 13 c-myc mice >16 weeks of age and were significantly larger, containing between 5 to 74 ductules (15 ± 1.3, mean ± SE; P = 0.006 compared with non-TG). Finally, the cell turnover characteristics were markedly lower in hyperplasias found in non-TG versus transgenic glands. BrdU-labeled cells and pyknotic cells could be identified in nearly all hyperplasias in transgenic mice but were found in 43 and 75%, respectively, in non-TG mice (Tables 2 and 4)
. LI also was higher for hyperplasias in transgenic glands compared with non-TG glands (Table 3)
, whereas AI was elevated significantly only for hyperplasias in c-myc mice (Table 5)
.
LI and AI Are Correlated Only in Hyperplasias of c-myc Mice
Proliferative and apoptotic activities in lobules, ducts, and hyperplasias in adjacent sections in a subset of five non-TG, five TGF- and five c-myc mice were evaluated for LI and AI. Only in hyperplasias in c-myc mice was a positive correlation found (r = 0.57).
|
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
and WAP-c-myc transgenic mouse models compared with wild-type mice despite the distinctly different molecular mechanisms underlying tumorigenesis in these models. We also have demonstrated a pattern of cell turnover in mammary glands of wild-type mice that shifts with lobular development and with adult aging.
In this study, there was a significant reduction in cell turnover activity in glands of non-TG mice >16 weeks compared with non-TG mice 12 weeks. This shift coincided with completion of ductal arborization and the loss of terminal endbuds, which in non-TG and transgenic mice of this study was completed by 12 weeks of age. We identified high levels of proliferation and apoptosis in ductal regions distant from terminal endbuds (Tables 2 to 4)
, which may represent regions where lobules develop and recede coincident with the estrous cycle. Similar to a previous report in C57BL/6 mice, we also found that apoptotic activity was enhanced in non-TG FVB/N mice with more extensive lobular development (Figure 5)
.11
Finally, we believe this is the first report of a gradual decrease in both proliferative and apoptotic activity in mammary glands of aging wild-type mice (Figures 3 and 4)
. It is important to emphasize that this finding was observed in a population of healthy mature mice ranging from 18 to 40 weeks of age. Although we did not test fertility for each mouse in this study, in our colony, breeder females often reproduce well past 40 weeks of age. A reduction in cell proliferation may be an intrinsic cellular and tissue mechanism that protects against the expansion of cells that may be cancer prone. Then again, a decrease in apoptosis suggests that the mammary epithelium may be less capable of eliminating aberrant cells with age. Importantly, the fact that both components of cell turnover are altered with age may be important for understanding means in which one can prevent and treat cancer.
Although increases in cell turnover were a common feature of normal epithelia in both TGF- and c-myc mammary glands, there were distinct differences in the pattern of cell turnover between these two transgenic models. First, in mice 12 weeks, lobular proliferation in TGF- and non-TG glands was similar, whereas levels in c-myc glands were approximately twofold to threefold higher. Proliferation subsided in non-TG mice >16 weeks, but remained high in transgenic mice. Second, in glands from TGF- mice >16 weeks, proliferative activity was greater in ducts than lobules, whereas glands of c-myc mice displayed equivalent levels of proliferation even though in both models, transgene expression was directed by the WAP promoter. Elevated lobular proliferation was correlated with development of lobular hyperplasias in glands from both transgenic mouse lineages (Tables 2 and 3
, Figure 6
).
|
mice were observed already in mice 12 weeks, manifested as a twofold to threefold reduction in apoptosis in normal epithelium of TGF- glands. A reduction in epithelial apoptosis was seen previously in other models of TGF- transgenic mice12,13
and in bitransgenic c-myc-TGF-14
and prolactin-TGF- mice15
compared with the single transgenic c-myc- or prolactin-overexpressing mice. Reduced apoptosis coupled with augmented cell proliferation can promote an environment in which defective cells normally destined for programmed cell death remain alive. If these cells also have a growth advantage, as may be the case with cells expressing growth factors, the result can be hyperplasia and cancer development. However, apoptosis in normal mammary epithelia from TGF- mice was suppressed only in mice 12 weeks and actually increased to levels higher than non-TG mice in animals >16 weeks. In addition, unlike non-TG glands, epithelia in TGF- glands displayed an increase in proliferation coinciding with increased stage of lobularization, suggesting a growth effect not seen in non-TG glands (Figure 2)
. The fact that nearly all TGF- mice eventually develop hyperplasias and tumors indicates that enhanced cellular proliferation overwhelms the enhanced apoptotic responses, resulting in a net gain of cells.
It is uncertain in transgenic mice whether lobular hyperplasias are the precursor lesion of end-stage tumors. In TGF- mice, heterogeneous end-stage lesions include features reminiscent of fibrocystic changes in women,16
such as adenomas, fibroadenomas, and adenocarcinomas, making it difficult to ascribe a simple tumor precursor histotype. However, transplanted TGF- hyperplasias recapitulate many end-stage lesion and tumor histotypes (T.A.R.-H. and E.P.S., unpublished data), suggesting that cells from hyperplastic lobules have the potential to be tumor precursors in this model. In c-myc mice, intermediate lesions other than lobular hyperplasias with various degrees of dysplasia (Figure 1M)
were not found. In non-TG mice, hyperplasias were intrinsically different from transgenic hyperplasias using several criteria, and they likely represent a variant of normal lobular development. First, incidence was low (2 of 13 animals >16 weeks of age). Second, non-TG hyperplasias contained approximately fivefold fewer alveolar lumens than those found in transgenic mice (Figure 1, B, H, and M)
. Third, proliferative and apoptotic levels in hyperplasias from non-TG mice were significantly lower than in transgenic mice (Tables 2 to 5)
. These data indicate that lobular hyperplasias display variable cell turnover characteristics and those that precede or accompany tumor development are distinguishable from large lobules/small hyperplasias that appear to have no potential for cancer development. Hyperplasias that are larger and highly proliferative may be a cellular feature (biomarker) worthy of future evaluation in preclinical testing of preventive regimens.
TGF- and c-myc are molecules that frequently are activated in breast cancer in women. TGF- is overexpressed in 30 to 70% of human breast tumors examined17,18
and, depending on the study, c-myc amplification, expression, or rearrangement occurs in the majority of breast cancers.19
In this study, c-myc mice displayed a uniformly high epithelial cell turnover (Tables 2 to 4)
before and coincident with lesion development. High levels of cell turnover were maintained regardless of epithelial type, size of lobules, degree of lobularization, and age, indicating that in c-myc-expressing cells the signal was dominant, overriding normal cell growth and death signals. Finally, a stepwise increase in cell turnover from normal epithelia to hyperplasias to tumors suggests a progressive increase in this measure during tumor progression. The overall high level of cell turnover in c-myc in normal and abnormal epithelia may be explained by the autocrine nature of the downstream effects of this transcription factor. The impact of TGF- is more diverse, despite the fact that both transgenes were directed by the whey acidic protein (WAP) promoter.20
This diversity may be explained by the autocrine, paracrine, and endocrine nature of TGF- signaling. Numerous factors impact downstream TGF- signaling,21-23
which are accompanied by more highly variable proliferative and apoptotic responses compared with c-myc expressing cells (Tables 2 to 5)
. It also may explain the diverse set of end-stage TGF- lesions.
We have found that enhanced cell turnover in lobules and ducts precede morphological evidence of tumorigenesis in two distinct transgenic mouse models of breast cancer. The identification and validation of early markers that denote an increased risk of breast cancer development and that can be used to predict the success or failure of a preventive therapeutic regimen is a major goal of breast cancer prevention research. The degree of cell turnover in lesions is already a biomarker used in human breast cancer. In addition, alterations in proliferation have been measured from core needle biopsies of breast tissue taken before and after cancer treatment.24 Our data from transgenic mouse models indicate that cell turnover activities in normal-appearing mammary epithelium before the onset of lesions should be evaluated as a potential biomarker of breast cancer risk in women and of the efficacy of cancer prevention regimens in preclinical animal models.
|
Footnotes |
|---|
Supported in part by the National Institutes of Health (grants K01-RR00145 to T.A.R.-H., R01-CA64843 to E.P.S., and ES09090).
Accepted for publication August 1, 2006.
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
