Age and physiologic status, such as menopause, are risk factors for breast cancer. Less clear is what factors influence the diversity of breast cancer. In this study, we investigated the effect of host age on the distribution of tumor subtypes in mouse mammary chimera consisting of wild-type hosts and Trp53 nullizygous epithelium, which undergoes a high rate of neoplastic transformation. Wild-type mammary glands cleared of endogenous epithelium at 3 weeks of age were subsequently transplanted during puberty (5 weeks) or at maturation (10 weeks) with syngeneic Trp53-null mammary tissue fragments and monitored for one year. Tumors arose sooner from adult hosts (AH) compared with juvenile hosts (JH). However, compared with AH tumors, JH tumors grew several times faster, were more perfused, exhibited a two-fold higher mitotic index, and were more highly positive for insulin-like growth factor receptor phosphorylation. Most tumors in each setting were estrogen receptor (ER)-positive (80% JH vs. 70% AH), but JH tumors were significantly more ER-immunoreactive (P = 0.0001) than AH tumors. A differential expression signature (JvA) of juvenile versus adult tumors revealed a luminal transcriptional program. Centroids of the human homologs of JvA genes showed that JH tumors were more like luminal A tumors and AH tumors were more like luminal B tumors. Hierarchical clustering with the JvA human ortholog gene list segregated luminal A and luminal B breast cancers across datasets. These data support the notion that age-associated host physiology greatly influences the intrinsic subtype of breast cancer. Cancer Res; 74(23); 7149–58. ©2014 AACR.

Breast cancer is a heterogeneous disease, or rather, a collection of neoplastic diseases. As tumor characteristics are highly correlated with prognosis, there has been considerable investment in defining classes and standardizing the characterization of breast tumors based on traditional pathology, marker analysis, imaging features, and molecular analysis. Human epidermal growth factor receptor 2 (HER2) amplification distinguishes a specific breast cancer subtype that provides a robust molecular target in therapy (1). Estrogen receptor (ER) status is important not only because ER-positive (ER+) tumors are managed with antiestrogen treatments but also because ER status is associated with different age groups and with different outcomes. ER-negative (ER) tumors are more frequent in young women and have a worse prognosis (2). Among ER breast cancers there are also several diseases. ER and progesterone receptor (PR)–negative cancers tend to be poorly differentiated and aggressive. Triple-negative breast cancers that lack both hormone receptors and HER2 amplification are usually invasive ductal carcinomas that have a high rate of relapse (3).

Molecular profiling discriminates between five and 20 tumor types depending on the algorithms and datasets used (reviewed in ref. 4). As defined by Perou and colleagues (5–7), transcriptomic analysis of human breast cancer comprises at least six different intrinsic molecular subtypes: luminal A, luminal B, HER2, basal-like, claudin-low, and normal-like. Molecular classification is not synonymous with clinical marker status using ER, PR, and HER2. For example, although ER+ tumors predominate in postmenopausal women and have a relatively low relapse rate, molecular characterization revealed that ER+ breast cancers can be divided into the so-called luminal A cancers, which are associated with a longer disease-free survival than that of luminal B cancers. The luminal A and B subtypes are known to maintain the expression of ESR1 transcript and protein, along with the canonical mammary luminal fate regulators GATA3 and FOXA1 (8, 9), but the luminal B subtype has a strong signature of proliferation.

Although tumor characteristics are clearly linked with prognosis, it is less clear how these distinct features arise. One source of biologic distinction between breast tumors is thought to be the cell of origin at transformation (10). A “cells of origin” autonomous view of breast cancer is bolstered by several studies in which disruption of the progenitor fate regulator, BRCA1, via promoters specific to the basal mammary gland compartment, produced tumors that resembled human basal-like breast cancers (11–13). Alternatively, deregulation of key transcriptional regulator, whether by mutation or epigenetic alteration, may divert cells to a particular state. For example, GATA3, a necessary transcriptional factor for differentiation of luminal cells, is associated with the luminal subtype of breast cancers (14–16). External factors are suggested by epidemiologic studies of girls exposed to ionizing radiation, who develop triple-negative breast cancer at an early age (17, 18), and the association of specific tumor types with obesity (19). Several recent studies have turned attention on the stroma to derive prognostic value by using expression profiling of stromal and extratumoral tissues (20). Using microdissected stroma from breast cancer, Finak and colleagues (21) found a stroma-derived prognostic predictor that stratifies disease outcome based on a signature of immune mediators, hypoxia, and angiogenesis. Analysis of the expression profiles from invasive breast cancer and ductal carcinoma in situ provides evidence that stromal biology is a key determinant of progression (22). Consistent with the influence of distinct microenvironment subtypes, an active versus inactive cancer-adjacent microenvironment is associated with aggressiveness and outcome of ER+ human breast cancers (20).

The stromal factors that contribute to breast cancer subtype distribution are not mutually exclusive and may be further influenced by more systemic processes. We used a mouse model to show that mice exposed to ionizing radiation and subsequently transplanted with unirradiated donor tissue preferentially develop ER breast cancers (23). These tumors exhibit a gene expression program characterized by signatures of inflammation and stem cell biology. Somewhat surprisingly, the murine host radiation profile also segregates sporadic human breast cancers according to intrinsic subtype, i.e., basal from luminal, and within basal cancers, the irradiated murine host profile mediated by TGFβ associated with claudin-low cancers (24). This bioinformatic analysis suggests that common processes may underlie the etiology of both radiation-preceded and sporadic basal-type cancers. These data led to the hypothesis that host biology profoundly influences the etiology of breast cancer intrinsic subtypes (25).

As puberty is a period of significant physiologic changes that alter the hormonal milieu, which in turn directs mammary morphogenesis, in this study, we used a mammary chimera model to compare the development of Trp53-null cancer as a function of host age. More tumors arose from Trp53-null outgrowths in adult mice compared with those transplanted during puberty, but the influence of puberty was manifested in development of rapidly growing, strongly ER+ luminal tumors. We found that the transcriptional profile of the features that differ between tumors arising from juvenile versus adult transplantation segregates human luminal intrinsic breast cancer subtypes.

Animals

All animal experiments were performed at Lawrence Berkeley National Laboratory with institutional review and approval. BALB/c mice were purchased from The Jackson Laboratory and housed 4 per cage, fed with Lab Diet 5008 chow and water ad libidum. Trp53-null BALB/c mice were bred in-house under similar conditions. For transplantation experiments, the epithelial rudiments in inguinal glands of 3-week-old mice were surgically removed. At 5 or 10 weeks of age, both cleared mammary glands of host mice were transplanted with an approximately 1 mm3 fragment of previously frozen Trp53-null BALB/c mammary gland harvested and pooled from three or more inguinal glands of 8- to 10-week-old donor mice. Mice were monitored by palpation two to three times per week for 13 months. Once a palpable tumor was detected, tumors were measured with calipers twice weekly until reaching 1 cm3, at which point the first tumor was resected using survival surgery. Tumors were divided and frozen in liquid nitrogen for RNA extraction, or formalin-fixed followed by paraffin embedding. The mouse was further observed until the ipsilateral recurrence or a second tumor developed in the contralateral fatpad, which was monitored as above. If no tumors developed by experiment termination, which was 13 months after transplantation, then a wholemount was prepared to confirm transplantation efficiency. A gross necropsy was performed upon termination. An informative transplant was defined as that which had an epithelial outgrowth evident by tumor development or in the wholemount at sacrifice.

Immunohistochemistry

Sections were deparaffinized and rehydrated before antigen unmasking according to the manufacturer's instructions (Vector Labs; #H-3300), washed once with phosphate-buffered saline (PBS), and blocked with 0.5% casein and 0.1% Tween20/PBS for 1 hour at room temperature. Primary antibodies to ER (C1355; Millipore/Upstate; #06-935), PR (Fisher Scientific; RM-9102-S0), Foxa1 (Abcam; ab23738), Gata3 (Santa Cruz Biotechnology; SC-268), or phospho-IGF1 receptor (IGFR; Abcam; Ab39398) were diluted in Superblock Blocking Buffer (Pierce; #37515) and incubation was done at 4°C, overnight. The slides were then washed, followed by incubation with peroxidase-conjugated secondary antibody, washed, and counterstained with hematoxylin. Histopathologic characteristics of the tumors were reviewed by two observers blinded to the experimental details of the mouse models. ER status was scored using the Allred method (26).

Expression profiling

Total RNA quality and quantity were determined using Agilent 2100 Bioanalyzer and NanoDrop ND-1000. IlluminaMouse WG-6 v2.0 Expression Beadchips were used according to the manufacturer's protocol. The raw values (non-normalized and non-background-corrected) were exported from Beadstudio and read into R/Bioconductor using the Lumi package. Intensity values were then log2 transformed and quantile normalized. Background subtraction was not performed. The main effect of no background correction for Illumina data is that the fold-change estimation for low expression probes tends to be lower, but low expression yet high fold-change probes are more likely to be false positives. Gene expression data are archived on Gene Expression Omnibus (GEO) under accession number GSE56726.

Unsupervised hierarchical clustering (UHC) was done using Gene Cluster v3.0 software and heatmaps were visualized using Java TreeView v1.1.4r3 software. Expression data were mean centered. Gene clustering was done by an uncentered-correlation was done by Spearman rank correlation. Significance of Analysis of Microarray (SAM) analysis was done in a tandem-bootstrapping scheme as previously described (23). This approach derived a profile of 243 genes modulated by at least 1.5-fold in juvenile host (JH) tumors in 100% (13 of 13) of the secondary bootstraps. Enriched pathways within the 243 genes were identified with Ingenuity Pathway Analysis (IPA).

Statistical analysis

Statistical analysis was performed using Prism (GraphPad). Time to 2 × 2 mm3 tumor occurrence per informative fatpad was plotted using the Kaplan–Meier method with significance determined by the log-rank test. Tumor growth curves in a treatment group were fitted to an exponential curve and then averaged into one curve for each treatment group, as previously reported (23). Differences between treatment groups were determined using the χ2 test, Mann–Whitney test, or two-tailed Student t test, which were considered statistically significant at P < 0.05. The K-S test was used to determine the difference in distributions of human breast cancers that were classified by the centroids and then stratified by plotting in two dimensions.

Age at transplantation affects Trp53-null mammary carcinogenesis

The mouse mammary gland develops after birth, starting at 4 weeks of age with the onset of the ovarian hormones of puberty. The inguinal mammary gland can be divested of its endogenous epithelia at 3 weeks of age (27). Afterward, a mammary fragment from a donor mouse can be transplanted into the empty fatpad and it will grow out to form a ductal tree by 6 to 8 weeks. Here, we use the mammary chimera model to test the effect of age by characterizing the occurrence and type of tumors arising from Trp53-null outgrowths initiated in a host undergoing puberty compared with those that arise from transplantation into an adult. Trp53-null mammary fragments from 8- to 10-week-old syngeneic donors were transplanted into 5-week-old (juvenile) or 10 week-old (adult) BALB/c hosts; tumor occurrence was monitored for 13 months after transplantation (Fig. 1A). Significantly more palpable tumors were obtained in mice that were adults at transplantation (Fig. 1B). The extent of morphogenesis and successful transplantation frequency were comparable regardless of age at transplantation. The average size of sub-clinical neoplastic lesions (i.e., those that did not form frank, palpable tumors) evident in carmine-stained outgrowth wholemounts was similar for transplantation at each host age (data not shown). Age at transplantation did not affect the distribution of tumor histopathology (summarized in Supplementary Table S1), which included adenocarcinomas, spindle cell carcinomas, and squamous carcinomas as previously reported for this Trp53-null model (23). Adult host (AH) tumors (n = 89) had a median latency of 347 days, and 55% of informative fatpads were tumor bearing by experiment termination. In contrast, the median latency of JH tumors (n = 55) could not be calculated because only 27% of informative fatpads developed palpable tumors.

Figure 1.

Trp53-null transplantation into JH delays tumorigenesis compared with AH. A, cartoon of experimental scheme in which the epithelial rudiments of host mice were surgically removed from the inguinal mammary gland. A mammary fragment from a syngeneic Trp53-null BALB/c mouse was transplanted into the empty fatpad at either 5 weeks of age (middle of puberty) or 10 weeks of age (young adult). Tumor development was tracked for 13 months. B, Kaplan–Meier analysis tracking the time to tumor. Transplants into AH (red) had a median tumor latency of 347 days, with 55% (n = 89) of successful outgrowths forming tumors. Transplants into JH fatpads (pink; n = 55) formed tumors at a frequency of 27% after 13 months (log-rank test, P = 0.003).

Figure 1.

Trp53-null transplantation into JH delays tumorigenesis compared with AH. A, cartoon of experimental scheme in which the epithelial rudiments of host mice were surgically removed from the inguinal mammary gland. A mammary fragment from a syngeneic Trp53-null BALB/c mouse was transplanted into the empty fatpad at either 5 weeks of age (middle of puberty) or 10 weeks of age (young adult). Tumor development was tracked for 13 months. B, Kaplan–Meier analysis tracking the time to tumor. Transplants into AH (red) had a median tumor latency of 347 days, with 55% (n = 89) of successful outgrowths forming tumors. Transplants into JH fatpads (pink; n = 55) formed tumors at a frequency of 27% after 13 months (log-rank test, P = 0.003).

Close modal

Transplantation during puberty results in more aggressive tumors

Although fewer tumors arose from transplantation during puberty, once the tumor was detected, JH tumors grew significantly faster compared with AH (Fig. 2A). The tumor growth rate of tumors arising from transplantation to adults was comparable with that observed in previous experiments (23). Moreover, the difference of tumor growth rate was replicated in cohort of mice transplanted as adults versus juveniles in an independent study (28). Mitotic figures that included the spectrum of condensed chromosome states (Fig. 2B) were quantified in high-powered fields of view (n = 10) per hematoxylin and eosin (H&E)–stained tumor. Consistent with increased growth rate, mitotic figures were increased 2-fold in JH tumors compared with AH tumors (Fig. 2C). We next examined the degree of vascular perfusion. Red blood cells are auto-fluorescent under the UV filter in H&E-stained tissue. We quantified the average area of red blood cell cross-sections encapsulated within a membrane lining to exclude areas of hemorrhage to measure the size of perfused vessels (Fig. 2D). The size of perfused vessels in JH tumors was twice that of AH tumors (Fig. 3E). Thus, the age of the host, juvenile, or adult at the time of Trp53-null mammary transplantation affects the features of tumors that arise during the following year.

Figure 2.

Tumors from transplantation during puberty grow faster than those arising from transplantation in AH. A, tumors from transplants into 5-week-old JH grew faster than those from 10-week-old (adult) hosts. B, mitotic index was calculated by counting mitotic figures representing the full range of condensed chromosomal states visualized by H&E staining. C, tumors from JH (n = 12) exhibited a 2-fold higher mitotic index than those from AH (n = 8; **, P = 0.009, Mann–Whitney test). D, the degree of tumor perfusion was quantified by measuring the average area of red blood cell clusters that were encapsulated in a membrane. Red blood cell clusters in tumor sections stained by H&E were auto-fluorescent under UV. E, tumors arising from transplants into JH (n = 12) had 2-fold larger perfused blood vessels than those from AH (n = 8; *, P = 0.014, Mann–Whitney test).

Figure 2.

Tumors from transplantation during puberty grow faster than those arising from transplantation in AH. A, tumors from transplants into 5-week-old JH grew faster than those from 10-week-old (adult) hosts. B, mitotic index was calculated by counting mitotic figures representing the full range of condensed chromosomal states visualized by H&E staining. C, tumors from JH (n = 12) exhibited a 2-fold higher mitotic index than those from AH (n = 8; **, P = 0.009, Mann–Whitney test). D, the degree of tumor perfusion was quantified by measuring the average area of red blood cell clusters that were encapsulated in a membrane. Red blood cell clusters in tumor sections stained by H&E were auto-fluorescent under UV. E, tumors arising from transplants into JH (n = 12) had 2-fold larger perfused blood vessels than those from AH (n = 8; *, P = 0.014, Mann–Whitney test).

Close modal
Figure 3.

Tumors from transplantation during puberty are more highly immunoreactive for ER. A, 80% of tumors (n = 21) from transplants into JH were ER-immunoreactive compared with 70% of tumors (n = 10) from transplants into AH as classified according to the Allred score. B, representative images of ER immunohistochemistry in tumors from juvenile or AH (bar, 60 μm). C, comparison of the frequency component of the Allred score revealed that tumors from JH were more immunoreactive for ER than tumors from AH (JH: mean, 11%–33%, max, >66%; adult: mean, 1%–10%, max, 11%–33%; ***, P = 0.0004, Mann–Whitney test).

Figure 3.

Tumors from transplantation during puberty are more highly immunoreactive for ER. A, 80% of tumors (n = 21) from transplants into JH were ER-immunoreactive compared with 70% of tumors (n = 10) from transplants into AH as classified according to the Allred score. B, representative images of ER immunohistochemistry in tumors from juvenile or AH (bar, 60 μm). C, comparison of the frequency component of the Allred score revealed that tumors from JH were more immunoreactive for ER than tumors from AH (JH: mean, 11%–33%, max, >66%; adult: mean, 1%–10%, max, 11%–33%; ***, P = 0.0004, Mann–Whitney test).

Close modal

Transplantation during puberty results in a luminal breast cancer subtype

The BALB/c Trp53-null mammary transplant model recapitulates the histologic (29, 30) and molecular diversity of human breast cancer (24, 31). Eighty percent of the tumors arising from JH were ER+ by the Allred score (26), compared with 70% ER+ AH tumors (Fig. 3A), similar to that previously observed from transplantation in adult mice (23, 28). However, comparing ER+ tumors between JH and AH tumors, those from JH tumors were much more highly ER+ (Fig. 3C; JH: mean, 11%–33%, max, >66%; AH: mean, 1%–10%, max, 11%–33%). We then examined other canonical luminal markers Gata3, PR, and Foxa1. Fifty percent of JH tumors (8 of 16) were positive for PR, compared with less than 40% of AH tumors (4 of 11). Similarly, PR immunoreactivity in PR+ tumors from JH tended to be higher (JH: mean, 11%–30%; AH: mean, 2%–10%). Eighty percent of JH tumors (8 of 10) were also positive for Foxa1 while around 70% AH tumors (7 of 11) were Foxa1-positive. The intensity and frequency of Foxa1 positivity were comparable between JH and AH tumors. We also examined Gata3 by immunostaining; however, the results were not informative, with only one positive nuclear staining out of 20 tumors (10 JH and 10 AH). Overall, JH tumors exhibit features of a luminal cancer subtype.

We next conducted genome-wide microarray analysis using total RNA from JH tumors (n = 15) and AH (n = 12) tumors using Illumina Mouse WG-6 V2.0 Beadchips. UHC of all tumors did not segregate JH from AH tumors (data not shown). SAM analysis was done in a tandem-bootstrapping scheme, as previously reported (23), to derive a profile of genes modulated by at least 1.5-fold in 100% (13 of 13) of the secondary bootstraps in JH tumors. This yielded 243 genes (Supplementary Table S2) that represented the biology of JH tumors compared with AH tumors (Supplementary Table S1). UHC of the 27 tumors by these 243 genes resulted in segregation of most JH tumors apart from AH tumors (data not shown). IPA revealed that the global processes occurring in these 243 genes included embryonic development, organ development, and organ morphology (Fig. 4A, IPA score 43), cellular growth and proliferation, cancer, and tumor morphology (Fig. 4B, IPA score 38). Consistent with the high ER protein expression detected in these tumors, there was a clear luminal gene expression program induced by the JH: Foxa1, Areg, and Krt18 (Fig. 4C). In addition, Igf2 and Ccnd1 (cyclin D1) transcripts were also increased, consistent with the increased mitotic index detected in JH tumors (32, 33). Beyond growth induction, the Areg gene product has been reported to have proangiogenic properties, which would be consistent with the increased perfusion observed in JH tumors (34).

Figure 4.

Gene expression analysis revealed a luminal transcriptional program in tumors arising from JH. Genome-wide microarray analysis was done on tumors from JH (n = 15) or AH (n = 12). SAM-tandem-bootstrapping was done to identify 243 genes that were modulated by at least 1.5-fold in tumors from pubertal hosts compared with AH. IPA revealed two main biologic networks with the 243-gene profile: embryonic development, organ development, and organ morphology (A; IPA score 43); and cellular growth and proliferation, cancer, and tumor morphology (B; IPA score 38). C, the 243 JvA gene profile represented canonical mammary luminal genes, including Foxa1, Krt18, and Areg. Tumors arising from transplants in JH or AH were immunostained with the pY1161-IGFR antibody. D, cell membranes were weakly positive for pY1161-IGFR near the edge of AH tumors (bar, 60 μm). E, cell membranes were strongly positive for pY1161-IGFR near the edge of JH tumors (bar, 60 μm). F, pY1161-IGFR staining was semiquantitatively assessed as the percentage of a tumor that exhibited positive staining. JH tumors exhibited a 25-fold increase in cells positive for phosphorylated IGFR (Mann–Whitney test, P = 0.026).

Figure 4.

Gene expression analysis revealed a luminal transcriptional program in tumors arising from JH. Genome-wide microarray analysis was done on tumors from JH (n = 15) or AH (n = 12). SAM-tandem-bootstrapping was done to identify 243 genes that were modulated by at least 1.5-fold in tumors from pubertal hosts compared with AH. IPA revealed two main biologic networks with the 243-gene profile: embryonic development, organ development, and organ morphology (A; IPA score 43); and cellular growth and proliferation, cancer, and tumor morphology (B; IPA score 38). C, the 243 JvA gene profile represented canonical mammary luminal genes, including Foxa1, Krt18, and Areg. Tumors arising from transplants in JH or AH were immunostained with the pY1161-IGFR antibody. D, cell membranes were weakly positive for pY1161-IGFR near the edge of AH tumors (bar, 60 μm). E, cell membranes were strongly positive for pY1161-IGFR near the edge of JH tumors (bar, 60 μm). F, pY1161-IGFR staining was semiquantitatively assessed as the percentage of a tumor that exhibited positive staining. JH tumors exhibited a 25-fold increase in cells positive for phosphorylated IGFR (Mann–Whitney test, P = 0.026).

Close modal

Enhanced IGF1 signaling associates with the luminal subtype of breast cancer

Studies in human and mouse have shown that enhanced insulin-like growth factor (IGF) signaling is associated with breast cancer phenotype (reviewed in ref. 35), including poor prognosis (36), and IGF signaling regulates proliferation in the normal mammary gland (37, 38). Because JH tumors were highly proliferative and exhibited an increase in Igf2 transcript, we examined phosphorylated IGF receptor (IGFR) levels in JH tumors. Both Igf1 and Igf2 bind to IGFR and ligand binding of the IGFR leads to its phosphorylation. Phosphorylated IGFR was detected with an antibody targeting the Y1161 residue, which represents the activated state of the receptor. Positive cells were located near the tumor border were characterized by membrane staining (Fig. 4D and E). Phosphorylated IGFR staining was semiquantitatively assessed as the percentage of a tumor that exhibited positive staining. Phosphorylated IGFR was increased 25-fold in JH tumors (P = 0.026) compared with AH tumors (Fig. 4F).

Human orthologs of the JvA243 profile segregates luminal breast cancers

The high ER protein expression and luminal transcriptional program in the JH tumors clearly suggested that these tumors were of a luminal subtype. We then determined the utility of 243-gene JH profile in discerning differences between human breast cancers using the Fredlund dataset, which represents 1,608 breast cancers compiled from 10 independent datasets on the Affymetrix U133A platform. Human orthologs of 161 genes, represented by 314 probes, of the 243 mouse genes were present on the Affymetrix Human U133A platform. To further explore this association, we developed a JvA centroid table, which maintains the relative direction of the genes that are differentially expressed in tumors arising in juvenile versus AH (Supplementary Table S3). Genes in the JvA centroid table were the human homologs mouse JvA genes, which are referred to as the JHh and AHh centroids, respectively. Human breast cancers (n = 1,608) from Fredlund were classified and the correlation values of each sample to each of the two centroids were plotted as x and y coordinates. This classification revealed stratification of human breast luminal versus basal tumors. Luminal human breast cancers shifted away from the positive axis representing AH tumors, while basal-like cancers shifted toward this axis (Fig. 5A; P = 1.11E−33). Interestingly, luminal breast cancer is the most heterogeneous (39); when only luminal A (LumA) and B (LumB) tumors were compared, the distribution of human LumA tumors suggested that they were more similar to murine tumors arising from tissue transplanted during puberty, whereas the distribution of LumB tumors was shifted toward the axis representing AH tumors (Fig. 5B and C; P = 3.17E−16).

Figure 5.

Luminal A human breast cancers are more similar to JH tumors. A, centroids (JHh and AHh) representing the human orthologs of the JvA 243-gene profile were used to stratify 1,063 luminal (blue) or basal-like (red) human breast cancers (x-axis, AH; y-axis, JH). JHh vs. AHh centroid stratification distinguished luminal cancers from basal-like cancers, with the distribution of basal-like cancers shifting toward the positive AHh axis (K-S test, P = 1.11E−33). This suggests that basal-like cancers are more similar to tumors arising from transplantation to AH. B and C, the centroid stratification of luminal A and luminal B cancers shows that luminal A cancers shift toward the positive JHh axis and away from the positive AHh axis, while luminal B cancers are shifted toward the positive AHh axis and away from the positive JHh axis (K-S test, P = 3.17E−16).

Figure 5.

Luminal A human breast cancers are more similar to JH tumors. A, centroids (JHh and AHh) representing the human orthologs of the JvA 243-gene profile were used to stratify 1,063 luminal (blue) or basal-like (red) human breast cancers (x-axis, AH; y-axis, JH). JHh vs. AHh centroid stratification distinguished luminal cancers from basal-like cancers, with the distribution of basal-like cancers shifting toward the positive AHh axis (K-S test, P = 1.11E−33). This suggests that basal-like cancers are more similar to tumors arising from transplantation to AH. B and C, the centroid stratification of luminal A and luminal B cancers shows that luminal A cancers shift toward the positive JHh axis and away from the positive AHh axis, while luminal B cancers are shifted toward the positive AHh axis and away from the positive JHh axis (K-S test, P = 3.17E−16).

Close modal

To further explore this association, we then applied the JvA human orthologs gene list to the dataset from (GSE4922; ref. 40), which contains 137 LumA or LumB human breast cancers. JvA human orthologs segregated LumA and LumB tumors. LumB cancers comprised 30% of sub-tree 1 versus 63% of sub-tree 2 (χ2, P < 0.0003; Fig. 6A). Because it is known that LumB cancers are much more proliferative than LumA cancers (reviewed in ref. 8), we determined how the proliferation program in the 243 genes affected segregation. Removal of 31 proliferation-related genes in the 161 human orthologs left 130 human genes represented by 240 probes (Supplementary Table S2). The proliferation-divested profile showed even stronger segregation of LumB cancer in the Ivshina dataset (16% in sub-tree 1, 63% in sub-tree 2; χ2, P < 0.0001; Fig. 6B). The same effect was observed on 110 grade-1 LumA or LumB cancers (41) such that LumB tumors comprised 10% of sub-tree 1, but 39% of sub-tree 2 (χ2, P < 0.001; Fig. 6C). Effective segregation of LumA from LumB cancers without the proliferation-related genes suggests that the JvA 243 profile captures a biologic distinction between LumA and LumB that is more than just increased proliferation. Taken together, the bioinformatic analyses suggest that the JvA gene list identifies a distinct luminal breast cancer subtype.

Figure 6.

Luminal A and B human breast cancers are segregated by the JvA transcriptomic program. A, luminal A or B cancers from Ivshina and colleagues were clustered by the JvA 243 profile and exhibited an enrichment of luminal B cancers in the main bifurcation (30.2% in sub-tree 1 of 2; 63.4% in sub-tree 2 of 2; χ2 test, P < 0.0003). B, removal of proliferation-related genes from the JvA 243 profile did not compromise the segregation of luminal B cancers (16.4% in sub-tree 1 of 2; 62.8% in sub-tree 2 of 2; χ2 test, P < 0.0001). C, grade-1 luminal A or B breast cancers from Fredlund and colleagues were segregated by the JvA profile that was divested of proliferation-related genes. The enrichment from luminal B cancers was also observed (10.3% in sub-tree 1 of 2; 39.4% in sub-tree 2 of 2; χ2 test, P < 0.001).

Figure 6.

Luminal A and B human breast cancers are segregated by the JvA transcriptomic program. A, luminal A or B cancers from Ivshina and colleagues were clustered by the JvA 243 profile and exhibited an enrichment of luminal B cancers in the main bifurcation (30.2% in sub-tree 1 of 2; 63.4% in sub-tree 2 of 2; χ2 test, P < 0.0003). B, removal of proliferation-related genes from the JvA 243 profile did not compromise the segregation of luminal B cancers (16.4% in sub-tree 1 of 2; 62.8% in sub-tree 2 of 2; χ2 test, P < 0.0001). C, grade-1 luminal A or B breast cancers from Fredlund and colleagues were segregated by the JvA profile that was divested of proliferation-related genes. The enrichment from luminal B cancers was also observed (10.3% in sub-tree 1 of 2; 39.4% in sub-tree 2 of 2; χ2 test, P < 0.001).

Close modal

To date, studies using the Trp53-null mammary chimera at different ages, genotypes, or physiologic states suggest that it is a robust model for studying breast cancer diversity (23, 31, 42–44). Our studies using this model suggest that the host environment profoundly influences the type of tumor that eventually develops long after mammary outgrowth is complete. Here, we show that the age at transplantation elicits phenotypically and genomically distinct tumors that have features strongly reminiscent of the human luminal intrinsic subtypes. Although transplantation of AH generated more tumors compared with those transplanted during puberty, tumors arising from JH grew faster, had more mitoses, and were better perfused than those arising from outgrowths initiated in adult mice. Moreover, ER and phosphorylated IGFR staining was significantly enhanced in JH tumors. This phenotype was evident in the distinct genomic profile JvA, defined by comparison of tumors arising in JH versus AH hosts. JH tumors were characterized by elevated Igf2, Fgfr2, and FoxA1. Notably, the JvA was informative in regards to intrinsic tumor types in human breast cancer. Centroids defined by the human orthologs of the JvA profile stratified luminal from basal-like human breast cancers and luminal A from luminal B breast cancers. In support of this, JvA clustered luminal A and B intrinsic subtypes across public datasets.

An intriguing question is how the physiology and microenvironment during puberty promotes the development of highly ER+ mammary cancers. The mammary gland in mice and humans develops into maturity after birth. The onset of puberty initiates the estrus cycle that releases cyclical surges of estrogen and progesterone, two of the main hormones that induce ductal morphogenesis. Concomitantly, growth hormone and IGF stimulate longitudinal growth of the entire organism, while also being essential regulators of mammary gland development (37, 38). Phosphorylated IGFR staining was strikingly elevated in the highly ER+ JH tumors compared with AH tumors. Involvement of IGF signaling is consistent with the highly ER+ luminal phenotype. It has been shown that in mice, serum IGF1 level was significantly higher during the late puberty (5–8 weeks of age) compared with the level during the early puberty (3–5 weeks of age; ref. 45). It is possible that this growth factor surge promotes the carcinogenesis of JH tumor while for AH tumor, such growth-stimulation is lacking. This potential mechanism will be investigated in the further studies where different hormones are evaluated at different ages. Notably, proliferation mediated by IGF1 is important in cancer overall, is strongly associated with poor prognosis in breast tumors (36), and increases susceptibility in familial breast cancer (46).

Given the diversity of Trp53 tumor phenotypes, these specimens provide a meaningful model to further examine specific genomic alterations associated with particular tumor types. One means by which host physiology may influence tumor type is by causing epigenetic changes. The Cancer Genome Atlas project integrated copy number, RNA, methylation status, and miRNA data from approximately 300 to 460 human breast cancer genomes for each breast cancer subtype (9). Breast cancers positive for ER have a distinct methylome profile compared with the more aggressive ER cancers (47). Methylation is a key component of differentiation-associated patterns of transcription in normal development of the mammary gland (48). Both hypo- and hypermethylation changes are associated with disease. DNA methylation can confer long-term memory of early life events (49). Consistent with this, high estrogen exposure during neonatal development induces epigenetic imprinting in the form of gene methylation that is reflected in uterine and prostate cancers arising later on in adulthood (50, 51). Indeed, breast cancer molecular subtypes defined by transcriptomic profiling can also be characterized by methylome profiling (52), suggesting that the two may go hand-in-hand. Identical hypermethylated genes occur in histologically normal ducts that are adjacent to preinvasive and invasive lesions, suggesting a common loco-regional etiology mediated by methylation (9, 53, 54).

We speculate that the distinct tumor subtypes that arise as a function of host biology in the mammary chimera model could be rooted in factors that influence cell intrinsic processes of either initiation or early progression during specific physiologic states, which in this case is puberty. The relative short age window of 5 weeks between transplantation resulting in JH and AH tumors suggest that events occurring in roughly the first month of transplant outgrowth affects the specific tumor type. The transplanted tissue fragments contain about 2,000 epithelial cells, of which 1 to 4 are stem cells based on functional repopulation assays (28). Trp53-null tissue gives rise to morphologically normal ductal outgrowths by 3 months posttransplantation; over the course of the next 6 to 9 months, morphologically aberrant lesions become evident. This effect of transplantation at different ages is consistent with our studies in irradiated hosts in which tumor type was affected even though only the host was irradiated months before tumor development. A strong radiation signature in the mammary gland expression profiles is present at 1 week after exposure, is much less robust at 4 weeks, and is no longer discernable 12 weeks after irradiation, yet is evident in tumors arising months later (23, 28).

In multistage carcinogenesis models, cancer is initiated by genomically aberrant cells whose proliferation results in acquisition of additional mutations to establish the malignant phenotype. Concomitant subversion of tissue suppressive function, evasion of the immune system, and recruitment of a permissive stroma involves a dynamic interplay between malignant and nonmalignant cells that is equally important (25). Our data from the Trp53 mammary chimera model show that tumor latency and type is highly influenced by differences in host biology at the time of transplantation due to exogenous conditions (e.g., ionizing radiation exposure, diet) or endogenous factors (e.g., host genotype, age). The idea that tumor type is determined by biology outside the epithelial cell genome is provocative in that it suggests that tumor type is not the sum of random genomic alterations and that, once understood, such host biology might be a target for chemoprevention. Extrinsic carcinogenesis processes can be viewed in terms of environmental imprinting or ecologic selection, although it is unclear which is more likely without more detailed studies or modeling. The mammary chimera model provides the means to somewhat uncouple cellular (i.e., initiation) and systemic responses to physiologic or environmental (e.g., life style) factors during carcinogenesis.

Age is clearly a primary factor associated with the risk of detecting clinically evident cancer, but it has been less clear how age intersects with the carcinogenic process. Here, mammary Trp53-null carcinogenesis advanced more rapidly from transplantation to adult compared with pubertal hosts, but tumors arising from transplantation during puberty had distinct hallmarks (e.g., faster growth rate, more mitoses, and greater perfusion) than those tumors arising from transplantation to an adult mouse. Preliminary data from middle-aged mice (i.e., transplanted at 10 months of age) indicate that tumorigenesis is reduced but the frequency of ER tumors substantially increased (unpublished data). The mammary chimera model thus provides novel evidence that both the latency and subtype of cancer is strongly influenced by factors beyond genomic alterations within the tumor cells.

The distinct biology of murine tumors arising from tissue undergoing morphogenesis in the context of puberty suggests that neoplastic transformation at a young age promotes a luminal tumor phenotype. These data support the notion that tumor intrinsic subtype is influenced by physiologic status at the time of putative initiation. Future studies will determine how systemic factors particular to puberty act to promote luminal-type tumors. Potential mechanisms include those affecting initiation, supporting survival of cells with particular mutations, or generating a stroma permissive for particular initiation events.

D.H. Nguyen is a consultant at Creative Bioinformatics Consulting and is Editor-in-Chief and President of Cancer InCytes Magazine. No potential conflicts of interest were disclosed by the other authors.

Conception and design: D.H. Nguyen, M.H. Barcellos-Hoff

Development of methodology: D.H. Nguyen

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): H. Ouyang, J.-H. Mao, L. Hlatky

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): D.H. Nguyen, H. Ouyang, J.-H. Mao, M.H. Barcellos-Hoff

Writing, review, and/or revision of the manuscript: D.H. Nguyen, H. Ouyang, J.-H. Mao, L. Hlatky, M.H. Barcellos-Hoff

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): D.H. Nguyen, M.H. Barcellos-Hoff

Study supervision: J.-H. Mao, L. Hlatky, M.H. Barcellos-Hoff

The authors thank Dr. Sylvain Costes (Lawrence Berkeley National Laboratory) for assistance with statistics and the GeneSys Research Institute (Tufts University School of Medicine) for expression profiling.

This research was supported by a predoctoral fellowship to D.H. Nguyen from the Department of Defense-Breast Cancer Research Program, funding from the Department of Energy, Office of Biological and Environmental Research program on Low Dose Radiation (J.-H. Mao and M.H. Barcellos-Hoff) and the National Cancer Institute of the NIH under Award Number U54CA149233 (L. Hlatky and M.H. Barcellos-Hoff).

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.

1.
Gusterson
BA
,
Gelber
RD
,
Goldhirsch
A
,
Price
KN
,
Save-Soderborgh
J
,
Anbazhagan
R
, et al
Prognostic importance of c-erbB-2 expression in breast cancer. International (Ludwig) Breast Cancer Study Group [see comments]
.
J Clin Oncol
1992
;
10
:
1049
56
.
2.
Rakha
EA
,
El-Sayed
ME
,
Green
AR
,
Lee
AH
,
Robertson
JF
,
Ellis
IO
. 
Prognostic markers in triple-negative breast cancer
.
Cancer
2007
;
109
:
25
32
.
3.
Rakha
EA
,
Elsheikh
SE
,
Aleskandarany
MA
,
Habashi
HO
,
Green
AR
,
Powe
DG
, et al
Triple-negative breast cancer: distinguishing between basal and nonbasal subtypes
.
Clin Cancer Res
2009
;
15
:
2302
10
.
4.
Perou
CM
,
Børresen-Dale
A-L
. 
Systems biology and genomics of breast cancer
.
Cold Spring Harb Perspect Biol
2011
;
3
pii: a003293
.
5.
Perou
CM
,
Sorlie
T
,
Eisen
MB
,
van de Rijn
M
,
Jeffrey
SS
,
Rees
CA
, et al
Molecular portraits of human breast tumours
.
Nature
2000
;
406
:
747
52
.
6.
Sorlie
T
,
Perou
CM
,
Tibshirani
R
,
Aas
T
,
Geisler
S
,
Johnsen
H
, et al
Gene expression patterns of breast carcinomas distinguish tumor subclasses with clinical implications
.
Proc Natl Acad Sci U S A
2001
;
98
:
10869
74
.
7.
Prat
A
,
Parker
JS
,
Karginova
O
,
Fan
C
,
Livasy
C
,
Herschkowitz
JI
, et al
Phenotypic and molecular characterization of the claudin-low intrinsic subtype of breast cancer
.
Breast Cancer Res
2010
;
12
:
R68
.
8.
Creighton
CJ
. 
The molecular profile of luminal B breast cancer
.
Biologics
2012
;
6
:
289
97
.
9.
Network CGA
. 
Comprehensive molecular portraits of human breast tumours
.
Nature
2012
;
490
:
61
70
.
10.
Visvader
JE
. 
Cells of origin in cancer
.
Nature
2011
;
469
:
314
22
.
11.
Liu
X
,
Holstege
H
,
van der Gulden
H
,
Treur-Mulder
M
,
Zevenhoven
J
,
Velds
A
, et al
Somatic loss of BRCA1 and p53 in mice induces mammary tumors with features of human BRCA1-mutated basal-like breast cancer
.
Proc Natl Acad Sci U S A
2007
;
104
:
12111
6
.
12.
Molyneux
G
,
Geyer
FC
,
Magnay
FA
,
McCarthy
A
,
Kendrick
H
,
Natrajan
R
, et al
BRCA1 basal-like breast cancers originate from luminal epithelial progenitors and not from basal stem cells
.
Cell Stem Cell
2010
;
7
:
403
17
.
13.
Proia
TA
,
Keller
PJ
,
Gupta
PB
,
Klebba
I
,
Jones
AD
,
Sedic
M
, et al
Genetic predisposition directs breast cancer phenotype by dictating progenitor cell fate
.
Cell Stem Cell
2011
;
8
:
149
63
.
14.
Jacquemier
J
,
Charafe-Jauffret
E
,
Monville
F
,
Esterni
B
,
Extra
JM
,
Houvenaeghel
G
, et al
Association of GATA3, P53, Ki67 status and vascular peritumoral invasion are strongly prognostic in luminal breast cancer
.
Breast Cancer Res
2009
;
11
:
R23
.
15.
Kouros-Mehr
H
,
Bechis
SK
,
Slorach
EM
,
Littlepage
LE
,
Egeblad
M
,
Ewald
AJ
, et al
GATA-3 links tumor differentiation and dissemination in a luminal breast cancer model
.
Cancer Cell
2008
;
13
:
141
52
.
16.
Voduc
D
,
Cheang
M
,
Nielsen
T
. 
GATA-3 expression in breast cancer has a strong association with estrogen receptor but lacks independent prognostic value
.
Cancer Epidemiol Biomarkers Prev
2008
;
17
:
365
73
.
17.
Castiglioni
F
,
Terenziani
M
,
Carcangiu
ML
,
Miliano
R
,
Aiello
P
,
Bertola
L
, et al
Radiation effects on development of HER2-positive breast carcinomas
.
Clin Cancer Res
2007
;
13
:
46
51
.
18.
Broeks
A
,
Braaf
LM
,
Wessels
LF
,
van de Vijver
M
,
De Bruin
ML
,
Stovall
M
, et al
Radiation-associated breast tumors display a distinct gene expression profile
.
Int J Radiat Oncol Biol Phys
2010
;
76
:
540
7
.
19.
Yang
XR
,
Chang-Claude
J
,
Goode
EL
,
Couch
FJ
,
Nevanlinna
H
,
Milne
RL
, et al
Associations of breast cancer risk factors with tumor subtypes: a pooled analysis from the breast cancer association consortium studies
.
J Natl Cancer Inst
2011
;
103
:
250
63
.
20.
Roman-Perez
E
,
Casbas-Hernandez
P
,
Pirone
J
,
Rein
J
,
Carey
L
,
Lubet
R
, et al
Gene expression in extratumoral microenvironment predicts clinical outcome in breast cancer patients
.
Breast Cancer Res
2012
;
14
:
R51
.
21.
Finak
G
,
Bertos
N
,
Pepin
F
,
Sadekova
S
,
Souleimanova
M
,
Zhao
H
, et al
Stromal gene expression predicts clinical outcome in breast cancer
.
Nat Med
2008
;
14
:
518
27
.
22.
Lee
S
,
Stewart
S
,
Nagtegaal
I
,
Luo
J
,
Wu
Y
,
Colditz
G
, et al
Differentially expressed genes regulating the progression of ductal carcinoma in situ to invasive breast cancer
.
Cancer Res.
2012
;
72
:
4579
86
.
23.
Nguyen
DH
,
Oketch-Rabah
HA
,
Illa-Bochaca
I
,
Geyer
FC
,
Reis-Filho
JS
,
Mao
JH
, et al
Radiation acts on the microenvironment to affect breast carcinogenesis by distinct mechanisms that decrease cancer latency and affect tumor type
.
Cancer Cell
2011
;
19
:
640
51
.
24.
Nguyen
DH
,
Fredlund
E
,
Zhao
W
,
Perou
CM
,
Balmain
A
,
Mao
J-H
, et al
Murine microenvironment metaprofiles associate with human cancer etiology and intrinsic subtypes
.
Clin Cancer Res
2013
;
19
:
1353
62
.
25.
Barcellos-Hoff
MH
. 
Does microenvironment contribute to the etiology of estrogen receptor negative breast cancer?
Clin Cancer Res
2013
;
19
:
541
8
.
26.
Harvey
JM
,
Clark
GM
,
Osborne
CK
,
Allred
DC
. 
Estrogen receptor status by immunohistochemistry is superior to the ligand-binding assay for predicting response to adjuvant endocrine therapy in breast cancer
.
J Clin Oncol
1999
;
17
:
1474
81
.
27.
DeOme
KB
,
Faulkin
LJ
 Jr
,
Bern
HA
,
Blair
PB
. 
Development of mammary tumors from hyperplastic alveolar nodules transplanted into gland-free mammary fat pads of female C3H mice
.
Cancer Res
1959
;
19
:
515
20
.
28.
Tang
J
,
Fernandez-Garcia
I
,
Vijayakumar
S
,
Martinez-Ruiz
H
,
Illa-Bochaca
I
,
Nguyen
DH
, et al
Irradiation of juvenile, but not adult, mammary gland increases stem cell self-renewal and estrogen receptor negative tumors
.
Stem Cells
2013
;
32
:
649
61
.
29.
Jerry
DJ
,
Kittrell
FS
,
Kuperwasser
C
,
Laucirica
R
,
Dickinson
ES
,
Bonilla
PJ
, et al
A mammary-specific model demonstrates the role of the p53 tumor suppressor gene in tumor development
.
Oncogene
2000
;
19
:
1052
8
.
30.
Rajkumar
L
,
Kittrell
F
,
Guzman
R
,
Brown
P
,
Nandi
S
,
Medina
D
. 
Hormone-induced protection of mammary tumorigenesis in genetically engineered mouse models
.
Breast Cancer Res
2007
;
9
:
R12
.
31.
Herschkowitz
JI
,
Zhao
W
,
Zhang
M
,
Usary
J
,
Murrow
G
,
Edwards
D
, et al
Comparative oncogenomics identifies breast tumors enriched in functional tumor-initiating cells
.
Proc Natl Acad Sci U S A
2011
;
109
:
2778
83
.
32.
Wang
TC
,
Cardiff
RD
,
Zukerberg
L
,
Lees
E
,
Arnold
A
,
Schmidt
EV
. 
Mammary hyperplasia and carcinoma in MMTV-cyclin D1 transgenic mice
.
Nature
1994
;
369
:
669
71
.
33.
Klos
KS
,
Kim
S
,
Alexander
CM
. 
Genotoxic exposure during juvenile growth of mammary gland depletes stem cell activity and inhibits Wnt signaling
.
PLoS ONE
2012
;
7
:
e49902
.
34.
Ma
L
,
Gauvillé
C
,
Berthois
Y
,
Millot
G
,
Johnson
GR
,
Calvo
F
. 
Antisense expression for amphiregulin suppresses tumorigenicity of a transformed human breast epithelial cell line
.
Oncogene
1999
18
:
6513
20
.
35.
Kleinberg
DL
,
Wood
TL
,
Furth
PA
,
Lee
AV
. 
Growth hormone and insulin-like growth factor-I in the transition from normal mammary development to preneoplastic mammary lesions
.
Endocr Rev
2009
;
30
:
51
74
.
36.
Creighton
CJ
,
Casa
A
,
Lazard
Z
,
Huang
S
,
Tsimelzon
A
,
Hilsenbeck
SG
, et al
Insulin-like growth factor-I activates gene transcription programs strongly associated with poor breast cancer prognosis
.
J Clin Oncol
2008
;
26
:
4078
85
.
37.
Ruan
W
,
Kleinberg
DL
. 
Insulin-like growth factor I is essential for terminal end bud formation and ductal morphogenesis during mammary development
.
Endocrinology
1999
;
140
:
5075
81
.
38.
Ruan
W
,
Fahlbusch
F
,
Clemmons
DR
,
Monaco
ME
,
Walden
PD
,
Silva
AP
, et al
SOM230 inhibits insulin-like growth factor-I action in mammary gland development by pituitary independent mechanism: mediated through somatostatin subtype receptor 3?
Mol Endocrinol
2006
;
20
:
426
36
.
39.
Ciriello
G
,
Sinha
R
,
Hoadley
KA
,
Jacobsen
AS
,
Reva
B
,
Perou
CM
, et al
The molecular diversity of Luminal A breast tumors.
Breast Cancer Res Treat
2013
;
141
:
409
20
.
40.
Hilton
HN
,
Graham
JD
,
Kantimm
S
,
Santucci
N
,
Cloosterman
D
,
Huschtscha
LI
, et al
Progesterone and estrogen receptors segregate into different cell subpopulations in the normal human breast
.
Mol Cell Endocrinol
2012
;
361
:
191
201
.
41.
Fredlund
E
,
Staaf
J
,
Rantala
J
,
Kallioniemi
O
,
Borg
A
,
Ringner
M
. 
The gene expression landscape of breast cancer is shaped by tumor protein p53 status and epithelial-mesenchymal transition
.
Breast Cancer Res
2012
;
14
:
R113
.
42.
Yan
H
,
Blackburn
AC
,
McLary
SC
,
Tao
L
,
roberts
AL
,
Xavier
EA
, et al
Pathways contributing to development of spontaneous mammary tumors in BALB/c-Trp53+/− mice
.
Am J Pathol
2010
;
176
:
1421
32
.
43.
Abba
MC
,
Hu
Y
,
Levy
CC
,
Gaddis
S
,
Kittrell
FS
,
Hill
J
, et al
Identification of modulated genes by three classes of chemopreventive agents at preneoplastic stages in a p53-null mouse mammary tumor model
.
Cancer Prev Res
2009
;
2
:
175
84
.
44.
Zhang
M
,
Behbod
F
,
Atkinson
RL
,
Landis
MD
,
Kittrell
F
,
Edwards
D
, et al
Identification of tumor-initiating cells in a p53-null mouse model of breast cancer
.
Cancer Res
2008
;
68
:
4674
82
.
45.
Callewaert
F
,
Venken
K
,
Kopchick
JJ
,
Torcasio
A
,
van Lenthe
GH
,
Boonen
S
, et al
Sexual dimorphism in cortical bone size and strength but not density is determined by independent and time-specific actions of sex steroids and IGF-1: evidence from pubertal mouse models
.
J Bone Miner Res
2010
;
25
:
617
26
.
46.
Werner
H
,
Bruchim
I
. 
IGF-1 and BRCA1 signalling pathways in familial cancer
.
Lancet Oncol
2012
;
13
:
e537
e44
.
47.
Brent
RL
. 
Carcinogenic risks of prenatal ionizing radiation
.
Semin Fetal Neonatal Med
2013
;
19
:
203
13
.
48.
Liu
S-C
,
Alomran
R
,
Chernikova
SB
,
Lartey
F
,
Stafford
J
,
Jang
T
, et al
Blockade of SDF-1 after irradiation inhibits tumor recurrences of autochthonous brain tumors in rats
.
Neuro-Oncology
2014
;
16
:
21
8
.
49.
Arnold
JN
,
Magiera
L
,
Kraman
M
,
Fearon
DT
. 
Tumoral immune suppression by macrophages expressing fibroblast activation protein-alpha and heme oxygenase-1
.
Cancer Immunol Res
2013
;
2
:
121
6
.
50.
Li
Y
,
Welm
B
,
Podsypanina
K
,
Huang
S
,
Chamorro
M
,
Zhang
X
, et al
Evidence that transgenes encoding components of the Wnt signaling pathway preferentially induce mammary cancers from progenitor cells
.
Proc Natl Acad Sci U S A
2003
;
100
:
15853
8
.
51.
Gao
C
,
Kozlowska
A
,
Nechaev
S
,
Li
H
,
Zhang
Q
,
Hossain
DMS
, et al
TLR9 signaling in the tumor microenvironment initiates cancer recurrence after radiotherapy
.
Cancer Res
2013
;
73
:
7211
21
.
52.
Klug
F
,
Prakash
H
,
Huber Peter
E
,
Seibel
T
,
Bender
N
,
Halama
N
, et al
Low-dose irradiation programs macrophage differentiation to an iNOS+/M1 phenotype that orchestrates effective T cell immunotherapy
.
Cancer Cell
2013
;
24
:
589
602
.
53.
Rajaram
M
,
Li
J
,
Egeblad
M
,
Powers
RS
. 
System-wide analysis reveals a complex network of tumor-fibroblast interactions involved in tumorigenicity
.
PLoS Genet
2013
;
9
:
e1003789
.
54.
Sofia Vala
I
,
Martins
LR
,
Imaizumi
N
,
Nunes
RJ
,
Rino
J
,
Kuonen
F
, et al
Low doses of ionizing radiation promote tumor growth and metastasis by enhancing angiogenesis
.
PLoS ONE
2010
;
5
:
e11222
.