|
Title
|
MicroRNA-206 is differentially expressed in Brca1-deficient mice and regulates epithelial and stromal cell compartments of the mouse mammary gland miRNA-206 expression in Brca1-deficient mice
|
|
Abstract
|
Depletion of Brca1 leads to defects in mouse mammary gland development and mammary tumors in humans and mice. To explore the role of microRNAs (miRNAs) in this process, we examined the mammary glands of MMTV-Cre Brca1Co/Co mice for differential miRNA expression using a candidate approach. Several miRNAs were differentially expressed in mammary tissue at day 1 of lactation and in mammary epithelial cell lines in which Brca1 messenger RNA (mRNA) levels have been reduced. Functional studies revealed that several of these miRNAs regulate mammary epithelial cell function in vitro, including miR-206. Creation and analysis of MMTV-miR-206 transgenic mice showed no effect on lactational mammary development and no tumors, but indicates a role in mammary tissue remodeling in mature mice, potentially involving Igf-1 and Sfrp1. These results indicate the potential of miRNAs to mediate the consequences of Brca1 loss and suggest a novel function for miR-206.
|
|
Body
|
Introduction
Mutations in the breast cancer susceptibility gene BRCA1 significantly increase the risk of developing breast cancer. Although the many characteristics of BRCA1-associated tumors have been elucidated, the early molecular changes that arise as a consequence of disruption of BRCA1 expression in the mammary gland and lead to mammary tumorigenesis are not well understood.
Targeted repression and/or deletion of Brca1 transcript leads to aberrant lobular–alveolar development in vivo,1, 2 and defective mammary epithelial differentiation in vitro.3, 4 Notably, tumors arising from conditional Brca1-knockout mice are reminiscent of human BRCA1-associated breast cancers sharing similar morphological and molecular characteristics, and clustering with the basal molecular subtype.5, 6, 7, 8, 9, 10 The molecular basis of these phenotypes however, is yet to be defined. To elucidate the molecular consequences of Brca1 depletion in the mouse mammary gland, we have previously investigated the role of coding genes in Brca1 loss. This analysis identified many genes differentially expressed in conditional Brca1-knockout mice at day 1 of lactation, including the receptor tyrosine kinase, c-Kit, which marks luminal progenitor cells in preneoplastic BRCA1 mutation carrier tissue and may mediate the expansion of luminal progenitor cells in BRCA1 loss.11, 12, 13 This supports the hypothesis that luminal cells are the cells of origin of BRCA1-associated tumors14, 15, 16, 17 and that BRCA1 may have a role in regulating mammary epithelial cell fate (reviewed in reference 18).
We now investigate the role of miRNAs in Brca1-deficient mouse mammary glands. miRNAs are small evolutionary-conserved RNAs that are implicated in many biological processes and diseases (reviewed in reference 19). miRNAs are differentially expressed during mammary gland development20, 21 and regulate the expression of milk transcripts,22, 23 self-renewal of mammary epithelial progenitor cells,24 ductal outgrowth25 and the modulation of key transcriptional networks underpinning the development of the mammary gland.25, 26, 27 The role of these molecules in Brca1-associated mammary epithelial defects however is unknown. In this study, we demonstrate the effects of Brca1 deficiency on the expression of miRNAs in the mouse mammary gland using both in vitro and in vivo models. We show that Brca1 loss in the lactating mammary gland results in the differential expression of miRNAs, and explore their roles in mammary gland morphogenesis.
Results and discussion
Identification of miRNAs that are differentially regulated in the mammary glands of MMTV-Cre Brca1Co/Co mice
To determine the impact of Brca1 deficiency on miRNA expression in the mouse mammary gland, we assessed the levels of nine miRNAs from MMTV-Cre Brca1Co/Co mice mammary glands at day 1 of lactation. These miRNAs were selected using a candidate approach with the following selection criteria: (i) expressed in breast tumors, in particular, tumors associated with BRCA1 mutation/repression (triple negative or basal subtype)28 and (ii) differentially expressed during mouse mammary gland development, in particular, those that inversely mirror the expression of Brca1.21
Out of the nine miRNAs that were screened, four were upregulated (miR-135b, miR-155, miR-205 and miR-206: Figure 1a) and five were downregulated (miR-31, miR-148a, miR-181c, miR-200b and miR-210: Figure 1b). Upregulation of miR-155, a known oncomiR,29 is interesting, given that it has previously been reported to be overexpressed in BRCA1-associated breast cancers and to be transcriptionally repressed by BRCA1.30 Three of the downregulated miRNAs (miR-148a, miR-181c and miR-210) are normally highly expressed during lactation in the mouse mammary gland.21 As Brca1 depletion in these mice results in a defect in lobular–alveolar differentiation,1 we speculate that reduced expression of these miRNAs could contribute to the phenotype observed in Brca1-deficient mammary glands.
To explore the potential functional consequences of the observed differential expression of miRNAs, we conducted in silico analysis of their predicted mRNA targets. We used the m3RNA algorithm31 to identify putative miRNA target genes (Supplementary Table 1) and utilized Ingenuity Pathway Analysis to infer functional consequences (Supplementary Table 2). We compared a previously generated list of coding genes that were differentially expressed in the mammary glands of conditional Brca1-knockout animals vs controls at day 1 of lactation,11 with the predicted mRNA targets of the miRNAs altered in the same tissues (Figures 1a and b). This generated a list of mRNAs whose expression is inversely correlated to the expression of the miRNA and are thus potential targets of these differentially expressed miRNAs. Interestingly, in both instances where miRNAs are downregulated and mRNAs upregulated and vice versa, cancer is a key disease category that is influenced, suggesting that these miRNA–mRNA interactions may contribute to tumorigenesis (Supplementary Table 2).
To investigate if the overexpression of miRNAs observed in lactating mammary tissue of MMTV-Cre Brca1Co/Co were also evident in differentiation-competent mammary epithelial cells with reduced levels of Brca1, we assessed miRNA expression in HC11 cells in which Brca1 expression was reduced. This demonstrated that out of the four miRNAs assessed, only miR-206 expression was significantly altered (Figure 1c). Consistent with this, we also found that the expression of miR-206 was increased in HCC1937 cells, which contain a BRCA1 germline mutation resulting in reduced BRCA1 expression and a truncated BRCA1 protein32 as compared with the wtBRCA1 control (Figure 1d).
In contrast to the analysis of Brca1-deficient mammary glands (Figure 1a), upregulation of miR-135b, miR-155 and miR-205 was not observed in HC11 cells in which Brca1 levels have been repressed using siRNA (Figure 1c). There are several possible explanations for this including the higher levels of Brca1 repression using gene deletion (in vivo) compared with siRNA (in vitro) and the heterogeneous cellular environment in mouse mammary tissue compared with a relatively homogenous cell line. For example, it is possible that repression of Brca1 in the epithelial compartment of the mammary gland causes upregulation of miR-135b, miR-155 and miR-205 in nonepithelial cells of the mammary gland. There are many examples of disease-associated miRNAs with altered expression in the stromal compartment of the mammary gland33, 34 and miR-155 has previously been implicated in the transformation of stromal fibroblasts.35
The expression of miR-206 was also evaluated in mouse mammary epithelial tissue at various stages of the mammary gland development (virgin, pregnant, lactating and involution). miR-206 levels were highest in virgin animals as compared with the other stages of mammary gland development (Figure 1e). Interestingly, this is the opposite of what is observed with Brca1 during the mammary gland development.36 Given the role of miR-206 in breast cancer and that it is upregulated in Brca1 loss, it is possible that it may contribute to the consequences of Brca1 loss in the mouse mammary gland.
Loss of Brca1 transcript expression leads to decreased proliferation,37 increased apoptosis1, 38 and defects in the ability to undergo both in vitro and in vivo mammary epithelial differentiation.1, 2, 3, 4, 11 Interestingly, miR-206 overexpression results in defective mammary epithelial differentiation (Figure 2a), in vitro proliferation defects39, 40, 41 and the induction of apoptosis via the repression of Notch3.42 These results together with the differential expression data above suggest miR-206 is a strong candidate acting downstream of Brca1. We therefore decided to further explore the role of miR-206 in mammary gland development.
Overexpression of Brca1-associated miRNAs affects mammary epithelial morphogenesis in vitro
To address the hypothesis that miRNAs overexpressed in Brca1-deficient mammary tissue may affect mammary epithelial morphogenesis, we determined the effect of ectopic overexpression for each miRNA in HC11 cells. Overexpression of miR-155, miR-205 and miR-206 resulted in a complete loss of HC11 dome formation, whereas, overexpression of miR-135b resulted in an increase in HC11 dome formation (Supplementary Figures 1a and b). A role for miR-155, miR-205 and miR-206 in mammary epithelial morphogenesis is consistent with previous studies in other epithelial cell types.43, 44, 45
MMTV-miR-206 mice show a defect in mammary gland structure after 12 months
Given that miR-206 is upregulated in the mammary glands of Brca1 conditional knockout mice and in HC11 cells treated with a Brca1 siRNA, and limits the ability of HC11 cell morphogenesis in vitro, we prioritized miR-206 for in vivo analysis to further explore its role in mammary gland development and function. Transgenic mice expressing miR-206 under the control of the MMTV promoter were created and expression was confirmed in mammary glands at day 1 of lactation (Figure 2b). Analysis of the glands revealed no apparent differences in gross morphology or tissue architecture during pregnancy, lactation or involution (Figures 2c and d). This suggests either that miR-206 has no role in these processes or that any affect is masked by increased branching morphogenesis induced by the various hormonal cascades during pregnancy and lactation.46
In contrast, MMTV miR-206 transgenic mice at 12–15 months of age displayed a striking mammary phenotype. This was characterized by a significant increase in fatty tissue and a significant reduction of branching within the mammary epithelial tree, suggestive of tissue degeneration, in all of the MMTV miR-206 mice (n=5) and one of the six control mice (Figure 3a). No evidence of mammary tumors was observed prior to the end of study. Interestingly, several transgenic mouse models of Brca1 display a similar phenotype.47, 48, 49, 50, 51 Jones et al.50 argue that the Brca1 phenotype is due to the action of increased estrogen/IGF-1 on the mammary stroma microenvironment and that this may create an environment that is permissive to tumor development. The late onset of tumor development in mouse models of BRCA1 disruption is consistent with this.9 In support of this hypothesis, BRCA1 negatively regulates IGF-1 and loss of BRCA1 is associated with an increase in IGF-1.52 A plausible hypothesis is therefore that loss of Brca1 causes an increase in miR-206, which in turn results in tissue remodeling in older mice and that this permits Brca1-associated tumor development.
There are several potential molecular and cellular mechanisms for the observed phenotype, including a miR-206-associated increase in epithelial cell apoptosis, epithelial to mesenchymal transition, induction of IGF and/or adipose production. To explore this, we first used qRT-PCR to confirm that miR-206 was still overexpressed in the mammary glands of transgenic mice at 12 months of age. Interestingly, higher miR-206 expression corresponded to a stronger phenotype, concurrent with this, we found elevated levels of miR-206 in the single control mouse that displayed the phenotype (data not shown). We then examined the expression of epithelial and mesenchymal compartment markers within the aged mice mammary glands. β-catenin staining that detects the epithelial content of the mammary gland, revealed lesser ductal and end bud structures in the MMTV miR-206 glands as compared with the controls (Figure 3b). The expression of miR-200c, another epithelial marker, was found to be significantly decreased in MMTV miR-206 glands as compared with the controls, further corroborating the β-catenin staining (Figure 3c).
To explore the possibility that the IGF-1 pathway may also contribute to the phenotype in miR-206 mice, the expression of Igf1 was evaluated and normalized against the total volume of epithelial cells using miR-200c.53, 54, 55 The results revealed that Igf1 was significantly upregulated in the MMTV miR-206 glands as compared with the controls (Figure 3d). This supports the hypothesis that the Igf1 pathway may contribute to the phenotype in MMTV-miR-206 mouse mammary glands. As Igf1 is a secreted stromal factor that signals in a paracrine manner to the epithelial cell, this also raises the possibility that miR-206 induces an expansion of the stromal compartment, which could explain the apparent higher fat content seen in the MMTV miR-206 mammary glands as compared with the controls.56, 57, 58 Interestingly, mature Brca1-deficient mice also demonstrate an increase in mammary adipose tissue.50
Although the proposed expansion of the stromal compartment may promote tumorigenesis, no tumors were detected in the mammary glands of MMTV-miR-206 glands. It is possible that this reflects a dependence on other components of the Brca1 pathway or on intact p53 activity, which when lost rescues cellular aging but induces tumorigenesis.55
miR-206 targets the 3′UTR of the Wingless-type MMTV integration site (Wnt) antagonist, Sfrp1
To further explore the Brca1-miR-206 pathway, we identified potential gene targets of miR-206 (Figure 4a). Four genes were shortlisted on the basis of the results from four independent miRNA target prediction programs—Sfrp1, Daam1, Slc39a10 and Zbtb4 (Figure 4a). To validate these predictions, we examined the levels of these genes in the mammary glands of MMTV-Cre Brca1Co/Co mice. Only Sfrp1 had reduced expression in the lactating mammary gland of MMTV-Cre Brca1Co/Co mice (Figure 4b). Sfrp1 is a secreted antagonist of the wingless-type MMTV integration site (Wnt)-signaling pathway,59 whose expression is frequently lost in a variety of cancers, including breast cancer via promoter methylation.60 Repression of SFRP1 is associated with poor prognosis with increasing stage of malignancy.52, 61, 62, 63, 64, 65 Although many reports have focused on assessing the methylation status of the SFRP1 promoter, several studies have shown that some tumors exhibit decreased SFRP1 expression, independent of promoter methylation,52, 66 suggesting additional regulatory mechanisms may be involved. In addition, miR-206 has been shown to regulate Sfrp1 expression during myogenesis in a porcine model.67
A miR-206-binding site was identified in the Sfrp1 3′UTR and shown to be evolutionary conserved (Figure 4c). To determine whether miR-206 could repress the Sfrp1 3′UTR in mice, a luciferase reporter assay was conducted with the Notch 3′UTR, a known target of miR-206, as a positive control. Overexpression of miR-206, decreased the activity of both the Sfrp1 and Notch3 3′UTRs (Figure 4d), confirming that miR-206 can repress both Sfrp1 and Notch3. To further demonstrate that the reduced luciferase activity is a direct consequence of miR-206 targeting the predicted Sfrp1 3′UTR-binding site, mutations were introduced into the miRNA-binding site. Increased expression of miR-206 did not affect the activity of the mutated Sfrp1 3′UTR (Figure 4d), suggesting that the decrease in luciferase signal can be attributed to the binding of miR-206 to Sfrp1 at the predicted binding site. We also found that Sfrp1 expression was reduced in mice at day 1 of lactation and 12–15 months of age (Figure 4e). These results are consistent with the mammary epithelial phenotype observed and suggest that this phenotype is potentially a result of deregulated Wnt signaling. Though, there are conflicting reports of Sfrp1 function in the mammary gland, a recent study has shown that Sfrp1 deficiency induces increased adiposity upon forced weight gain in mice, consistent with our observations in aged MMTV miR-206 mice. Another team demonstrated that the knockout of Sfrp1 in mice mammary glands promotes precocious mammary gland development with branching and alveolar development characteristic of the midpregnant mammary gland,68, 69 a feature that has not been observed in the MMTV miR-206 mice (Figure 3a). A probable reason is that this phenotype has been masked by the upregulation of the Igf-1-signaling cascade.55, 70, 71
Conclusions
In conclusion, this work has demonstrated that miR-206 is overexpressed in Brca1-deficient cells and that overexpression of miR-206, in several ways, mirrors the phenotype in MMTV-Cre Brca1Co/Co mice. This raises the possibility that miR-206 contributes to the effects of Brca1 loss in the mammary gland and that this likely involves the Wnt pathway through Sfrp1 and the Igf pathways. Although we focused on miR-206, our data concur with other reports in suggesting other miRNAs, such as miR-155,29 can regulate downstream functions of Brca1, highlighting that this and other miRNAs are an important facet of Brca1 biology and function. Understanding these mechanisms will be valuable for identifying new biomarkers and therapeutic targets and strategies for BRCA1-associated breast cancer.
|
|
Section
|
Introduction
Mutations in the breast cancer susceptibility gene BRCA1 significantly increase the risk of developing breast cancer. Although the many characteristics of BRCA1-associated tumors have been elucidated, the early molecular changes that arise as a consequence of disruption of BRCA1 expression in the mammary gland and lead to mammary tumorigenesis are not well understood.
Targeted repression and/or deletion of Brca1 transcript leads to aberrant lobular–alveolar development in vivo,1, 2 and defective mammary epithelial differentiation in vitro.3, 4 Notably, tumors arising from conditional Brca1-knockout mice are reminiscent of human BRCA1-associated breast cancers sharing similar morphological and molecular characteristics, and clustering with the basal molecular subtype.5, 6, 7, 8, 9, 10 The molecular basis of these phenotypes however, is yet to be defined. To elucidate the molecular consequences of Brca1 depletion in the mouse mammary gland, we have previously investigated the role of coding genes in Brca1 loss. This analysis identified many genes differentially expressed in conditional Brca1-knockout mice at day 1 of lactation, including the receptor tyrosine kinase, c-Kit, which marks luminal progenitor cells in preneoplastic BRCA1 mutation carrier tissue and may mediate the expansion of luminal progenitor cells in BRCA1 loss.11, 12, 13 This supports the hypothesis that luminal cells are the cells of origin of BRCA1-associated tumors14, 15, 16, 17 and that BRCA1 may have a role in regulating mammary epithelial cell fate (reviewed in reference 18).
We now investigate the role of miRNAs in Brca1-deficient mouse mammary glands. miRNAs are small evolutionary-conserved RNAs that are implicated in many biological processes and diseases (reviewed in reference 19). miRNAs are differentially expressed during mammary gland development20, 21 and regulate the expression of milk transcripts,22, 23 self-renewal of mammary epithelial progenitor cells,24 ductal outgrowth25 and the modulation of key transcriptional networks underpinning the development of the mammary gland.25, 26, 27 The role of these molecules in Brca1-associated mammary epithelial defects however is unknown. In this study, we demonstrate the effects of Brca1 deficiency on the expression of miRNAs in the mouse mammary gland using both in vitro and in vivo models. We show that Brca1 loss in the lactating mammary gland results in the differential expression of miRNAs, and explore their roles in mammary gland morphogenesis.
|
|
Title
|
Introduction
|
|
Section
|
Results and discussion
Identification of miRNAs that are differentially regulated in the mammary glands of MMTV-Cre Brca1Co/Co mice
To determine the impact of Brca1 deficiency on miRNA expression in the mouse mammary gland, we assessed the levels of nine miRNAs from MMTV-Cre Brca1Co/Co mice mammary glands at day 1 of lactation. These miRNAs were selected using a candidate approach with the following selection criteria: (i) expressed in breast tumors, in particular, tumors associated with BRCA1 mutation/repression (triple negative or basal subtype)28 and (ii) differentially expressed during mouse mammary gland development, in particular, those that inversely mirror the expression of Brca1.21
Out of the nine miRNAs that were screened, four were upregulated (miR-135b, miR-155, miR-205 and miR-206: Figure 1a) and five were downregulated (miR-31, miR-148a, miR-181c, miR-200b and miR-210: Figure 1b). Upregulation of miR-155, a known oncomiR,29 is interesting, given that it has previously been reported to be overexpressed in BRCA1-associated breast cancers and to be transcriptionally repressed by BRCA1.30 Three of the downregulated miRNAs (miR-148a, miR-181c and miR-210) are normally highly expressed during lactation in the mouse mammary gland.21 As Brca1 depletion in these mice results in a defect in lobular–alveolar differentiation,1 we speculate that reduced expression of these miRNAs could contribute to the phenotype observed in Brca1-deficient mammary glands.
To explore the potential functional consequences of the observed differential expression of miRNAs, we conducted in silico analysis of their predicted mRNA targets. We used the m3RNA algorithm31 to identify putative miRNA target genes (Supplementary Table 1) and utilized Ingenuity Pathway Analysis to infer functional consequences (Supplementary Table 2). We compared a previously generated list of coding genes that were differentially expressed in the mammary glands of conditional Brca1-knockout animals vs controls at day 1 of lactation,11 with the predicted mRNA targets of the miRNAs altered in the same tissues (Figures 1a and b). This generated a list of mRNAs whose expression is inversely correlated to the expression of the miRNA and are thus potential targets of these differentially expressed miRNAs. Interestingly, in both instances where miRNAs are downregulated and mRNAs upregulated and vice versa, cancer is a key disease category that is influenced, suggesting that these miRNA–mRNA interactions may contribute to tumorigenesis (Supplementary Table 2).
To investigate if the overexpression of miRNAs observed in lactating mammary tissue of MMTV-Cre Brca1Co/Co were also evident in differentiation-competent mammary epithelial cells with reduced levels of Brca1, we assessed miRNA expression in HC11 cells in which Brca1 expression was reduced. This demonstrated that out of the four miRNAs assessed, only miR-206 expression was significantly altered (Figure 1c). Consistent with this, we also found that the expression of miR-206 was increased in HCC1937 cells, which contain a BRCA1 germline mutation resulting in reduced BRCA1 expression and a truncated BRCA1 protein32 as compared with the wtBRCA1 control (Figure 1d).
In contrast to the analysis of Brca1-deficient mammary glands (Figure 1a), upregulation of miR-135b, miR-155 and miR-205 was not observed in HC11 cells in which Brca1 levels have been repressed using siRNA (Figure 1c). There are several possible explanations for this including the higher levels of Brca1 repression using gene deletion (in vivo) compared with siRNA (in vitro) and the heterogeneous cellular environment in mouse mammary tissue compared with a relatively homogenous cell line. For example, it is possible that repression of Brca1 in the epithelial compartment of the mammary gland causes upregulation of miR-135b, miR-155 and miR-205 in nonepithelial cells of the mammary gland. There are many examples of disease-associated miRNAs with altered expression in the stromal compartment of the mammary gland33, 34 and miR-155 has previously been implicated in the transformation of stromal fibroblasts.35
The expression of miR-206 was also evaluated in mouse mammary epithelial tissue at various stages of the mammary gland development (virgin, pregnant, lactating and involution). miR-206 levels were highest in virgin animals as compared with the other stages of mammary gland development (Figure 1e). Interestingly, this is the opposite of what is observed with Brca1 during the mammary gland development.36 Given the role of miR-206 in breast cancer and that it is upregulated in Brca1 loss, it is possible that it may contribute to the consequences of Brca1 loss in the mouse mammary gland.
Loss of Brca1 transcript expression leads to decreased proliferation,37 increased apoptosis1, 38 and defects in the ability to undergo both in vitro and in vivo mammary epithelial differentiation.1, 2, 3, 4, 11 Interestingly, miR-206 overexpression results in defective mammary epithelial differentiation (Figure 2a), in vitro proliferation defects39, 40, 41 and the induction of apoptosis via the repression of Notch3.42 These results together with the differential expression data above suggest miR-206 is a strong candidate acting downstream of Brca1. We therefore decided to further explore the role of miR-206 in mammary gland development.
Overexpression of Brca1-associated miRNAs affects mammary epithelial morphogenesis in vitro
To address the hypothesis that miRNAs overexpressed in Brca1-deficient mammary tissue may affect mammary epithelial morphogenesis, we determined the effect of ectopic overexpression for each miRNA in HC11 cells. Overexpression of miR-155, miR-205 and miR-206 resulted in a complete loss of HC11 dome formation, whereas, overexpression of miR-135b resulted in an increase in HC11 dome formation (Supplementary Figures 1a and b). A role for miR-155, miR-205 and miR-206 in mammary epithelial morphogenesis is consistent with previous studies in other epithelial cell types.43, 44, 45
MMTV-miR-206 mice show a defect in mammary gland structure after 12 months
Given that miR-206 is upregulated in the mammary glands of Brca1 conditional knockout mice and in HC11 cells treated with a Brca1 siRNA, and limits the ability of HC11 cell morphogenesis in vitro, we prioritized miR-206 for in vivo analysis to further explore its role in mammary gland development and function. Transgenic mice expressing miR-206 under the control of the MMTV promoter were created and expression was confirmed in mammary glands at day 1 of lactation (Figure 2b). Analysis of the glands revealed no apparent differences in gross morphology or tissue architecture during pregnancy, lactation or involution (Figures 2c and d). This suggests either that miR-206 has no role in these processes or that any affect is masked by increased branching morphogenesis induced by the various hormonal cascades during pregnancy and lactation.46
In contrast, MMTV miR-206 transgenic mice at 12–15 months of age displayed a striking mammary phenotype. This was characterized by a significant increase in fatty tissue and a significant reduction of branching within the mammary epithelial tree, suggestive of tissue degeneration, in all of the MMTV miR-206 mice (n=5) and one of the six control mice (Figure 3a). No evidence of mammary tumors was observed prior to the end of study. Interestingly, several transgenic mouse models of Brca1 display a similar phenotype.47, 48, 49, 50, 51 Jones et al.50 argue that the Brca1 phenotype is due to the action of increased estrogen/IGF-1 on the mammary stroma microenvironment and that this may create an environment that is permissive to tumor development. The late onset of tumor development in mouse models of BRCA1 disruption is consistent with this.9 In support of this hypothesis, BRCA1 negatively regulates IGF-1 and loss of BRCA1 is associated with an increase in IGF-1.52 A plausible hypothesis is therefore that loss of Brca1 causes an increase in miR-206, which in turn results in tissue remodeling in older mice and that this permits Brca1-associated tumor development.
There are several potential molecular and cellular mechanisms for the observed phenotype, including a miR-206-associated increase in epithelial cell apoptosis, epithelial to mesenchymal transition, induction of IGF and/or adipose production. To explore this, we first used qRT-PCR to confirm that miR-206 was still overexpressed in the mammary glands of transgenic mice at 12 months of age. Interestingly, higher miR-206 expression corresponded to a stronger phenotype, concurrent with this, we found elevated levels of miR-206 in the single control mouse that displayed the phenotype (data not shown). We then examined the expression of epithelial and mesenchymal compartment markers within the aged mice mammary glands. β-catenin staining that detects the epithelial content of the mammary gland, revealed lesser ductal and end bud structures in the MMTV miR-206 glands as compared with the controls (Figure 3b). The expression of miR-200c, another epithelial marker, was found to be significantly decreased in MMTV miR-206 glands as compared with the controls, further corroborating the β-catenin staining (Figure 3c).
To explore the possibility that the IGF-1 pathway may also contribute to the phenotype in miR-206 mice, the expression of Igf1 was evaluated and normalized against the total volume of epithelial cells using miR-200c.53, 54, 55 The results revealed that Igf1 was significantly upregulated in the MMTV miR-206 glands as compared with the controls (Figure 3d). This supports the hypothesis that the Igf1 pathway may contribute to the phenotype in MMTV-miR-206 mouse mammary glands. As Igf1 is a secreted stromal factor that signals in a paracrine manner to the epithelial cell, this also raises the possibility that miR-206 induces an expansion of the stromal compartment, which could explain the apparent higher fat content seen in the MMTV miR-206 mammary glands as compared with the controls.56, 57, 58 Interestingly, mature Brca1-deficient mice also demonstrate an increase in mammary adipose tissue.50
Although the proposed expansion of the stromal compartment may promote tumorigenesis, no tumors were detected in the mammary glands of MMTV-miR-206 glands. It is possible that this reflects a dependence on other components of the Brca1 pathway or on intact p53 activity, which when lost rescues cellular aging but induces tumorigenesis.55
miR-206 targets the 3′UTR of the Wingless-type MMTV integration site (Wnt) antagonist, Sfrp1
To further explore the Brca1-miR-206 pathway, we identified potential gene targets of miR-206 (Figure 4a). Four genes were shortlisted on the basis of the results from four independent miRNA target prediction programs—Sfrp1, Daam1, Slc39a10 and Zbtb4 (Figure 4a). To validate these predictions, we examined the levels of these genes in the mammary glands of MMTV-Cre Brca1Co/Co mice. Only Sfrp1 had reduced expression in the lactating mammary gland of MMTV-Cre Brca1Co/Co mice (Figure 4b). Sfrp1 is a secreted antagonist of the wingless-type MMTV integration site (Wnt)-signaling pathway,59 whose expression is frequently lost in a variety of cancers, including breast cancer via promoter methylation.60 Repression of SFRP1 is associated with poor prognosis with increasing stage of malignancy.52, 61, 62, 63, 64, 65 Although many reports have focused on assessing the methylation status of the SFRP1 promoter, several studies have shown that some tumors exhibit decreased SFRP1 expression, independent of promoter methylation,52, 66 suggesting additional regulatory mechanisms may be involved. In addition, miR-206 has been shown to regulate Sfrp1 expression during myogenesis in a porcine model.67
A miR-206-binding site was identified in the Sfrp1 3′UTR and shown to be evolutionary conserved (Figure 4c). To determine whether miR-206 could repress the Sfrp1 3′UTR in mice, a luciferase reporter assay was conducted with the Notch 3′UTR, a known target of miR-206, as a positive control. Overexpression of miR-206, decreased the activity of both the Sfrp1 and Notch3 3′UTRs (Figure 4d), confirming that miR-206 can repress both Sfrp1 and Notch3. To further demonstrate that the reduced luciferase activity is a direct consequence of miR-206 targeting the predicted Sfrp1 3′UTR-binding site, mutations were introduced into the miRNA-binding site. Increased expression of miR-206 did not affect the activity of the mutated Sfrp1 3′UTR (Figure 4d), suggesting that the decrease in luciferase signal can be attributed to the binding of miR-206 to Sfrp1 at the predicted binding site. We also found that Sfrp1 expression was reduced in mice at day 1 of lactation and 12–15 months of age (Figure 4e). These results are consistent with the mammary epithelial phenotype observed and suggest that this phenotype is potentially a result of deregulated Wnt signaling. Though, there are conflicting reports of Sfrp1 function in the mammary gland, a recent study has shown that Sfrp1 deficiency induces increased adiposity upon forced weight gain in mice, consistent with our observations in aged MMTV miR-206 mice. Another team demonstrated that the knockout of Sfrp1 in mice mammary glands promotes precocious mammary gland development with branching and alveolar development characteristic of the midpregnant mammary gland,68, 69 a feature that has not been observed in the MMTV miR-206 mice (Figure 3a). A probable reason is that this phenotype has been masked by the upregulation of the Igf-1-signaling cascade.55, 70, 71
|
|
Title
|
Results and discussion
|
|
Section
|
Identification of miRNAs that are differentially regulated in the mammary glands of MMTV-Cre Brca1Co/Co mice
To determine the impact of Brca1 deficiency on miRNA expression in the mouse mammary gland, we assessed the levels of nine miRNAs from MMTV-Cre Brca1Co/Co mice mammary glands at day 1 of lactation. These miRNAs were selected using a candidate approach with the following selection criteria: (i) expressed in breast tumors, in particular, tumors associated with BRCA1 mutation/repression (triple negative or basal subtype)28 and (ii) differentially expressed during mouse mammary gland development, in particular, those that inversely mirror the expression of Brca1.21
Out of the nine miRNAs that were screened, four were upregulated (miR-135b, miR-155, miR-205 and miR-206: Figure 1a) and five were downregulated (miR-31, miR-148a, miR-181c, miR-200b and miR-210: Figure 1b). Upregulation of miR-155, a known oncomiR,29 is interesting, given that it has previously been reported to be overexpressed in BRCA1-associated breast cancers and to be transcriptionally repressed by BRCA1.30 Three of the downregulated miRNAs (miR-148a, miR-181c and miR-210) are normally highly expressed during lactation in the mouse mammary gland.21 As Brca1 depletion in these mice results in a defect in lobular–alveolar differentiation,1 we speculate that reduced expression of these miRNAs could contribute to the phenotype observed in Brca1-deficient mammary glands.
To explore the potential functional consequences of the observed differential expression of miRNAs, we conducted in silico analysis of their predicted mRNA targets. We used the m3RNA algorithm31 to identify putative miRNA target genes (Supplementary Table 1) and utilized Ingenuity Pathway Analysis to infer functional consequences (Supplementary Table 2). We compared a previously generated list of coding genes that were differentially expressed in the mammary glands of conditional Brca1-knockout animals vs controls at day 1 of lactation,11 with the predicted mRNA targets of the miRNAs altered in the same tissues (Figures 1a and b). This generated a list of mRNAs whose expression is inversely correlated to the expression of the miRNA and are thus potential targets of these differentially expressed miRNAs. Interestingly, in both instances where miRNAs are downregulated and mRNAs upregulated and vice versa, cancer is a key disease category that is influenced, suggesting that these miRNA–mRNA interactions may contribute to tumorigenesis (Supplementary Table 2).
To investigate if the overexpression of miRNAs observed in lactating mammary tissue of MMTV-Cre Brca1Co/Co were also evident in differentiation-competent mammary epithelial cells with reduced levels of Brca1, we assessed miRNA expression in HC11 cells in which Brca1 expression was reduced. This demonstrated that out of the four miRNAs assessed, only miR-206 expression was significantly altered (Figure 1c). Consistent with this, we also found that the expression of miR-206 was increased in HCC1937 cells, which contain a BRCA1 germline mutation resulting in reduced BRCA1 expression and a truncated BRCA1 protein32 as compared with the wtBRCA1 control (Figure 1d).
In contrast to the analysis of Brca1-deficient mammary glands (Figure 1a), upregulation of miR-135b, miR-155 and miR-205 was not observed in HC11 cells in which Brca1 levels have been repressed using siRNA (Figure 1c). There are several possible explanations for this including the higher levels of Brca1 repression using gene deletion (in vivo) compared with siRNA (in vitro) and the heterogeneous cellular environment in mouse mammary tissue compared with a relatively homogenous cell line. For example, it is possible that repression of Brca1 in the epithelial compartment of the mammary gland causes upregulation of miR-135b, miR-155 and miR-205 in nonepithelial cells of the mammary gland. There are many examples of disease-associated miRNAs with altered expression in the stromal compartment of the mammary gland33, 34 and miR-155 has previously been implicated in the transformation of stromal fibroblasts.35
The expression of miR-206 was also evaluated in mouse mammary epithelial tissue at various stages of the mammary gland development (virgin, pregnant, lactating and involution). miR-206 levels were highest in virgin animals as compared with the other stages of mammary gland development (Figure 1e). Interestingly, this is the opposite of what is observed with Brca1 during the mammary gland development.36 Given the role of miR-206 in breast cancer and that it is upregulated in Brca1 loss, it is possible that it may contribute to the consequences of Brca1 loss in the mouse mammary gland.
Loss of Brca1 transcript expression leads to decreased proliferation,37 increased apoptosis1, 38 and defects in the ability to undergo both in vitro and in vivo mammary epithelial differentiation.1, 2, 3, 4, 11 Interestingly, miR-206 overexpression results in defective mammary epithelial differentiation (Figure 2a), in vitro proliferation defects39, 40, 41 and the induction of apoptosis via the repression of Notch3.42 These results together with the differential expression data above suggest miR-206 is a strong candidate acting downstream of Brca1. We therefore decided to further explore the role of miR-206 in mammary gland development.
|
|
Title
|
Identification of miRNAs that are differentially regulated in the mammary glands of MMTV-Cre Brca1Co/Co mice
|
|
Section
|
Overexpression of Brca1-associated miRNAs affects mammary epithelial morphogenesis in vitro
To address the hypothesis that miRNAs overexpressed in Brca1-deficient mammary tissue may affect mammary epithelial morphogenesis, we determined the effect of ectopic overexpression for each miRNA in HC11 cells. Overexpression of miR-155, miR-205 and miR-206 resulted in a complete loss of HC11 dome formation, whereas, overexpression of miR-135b resulted in an increase in HC11 dome formation (Supplementary Figures 1a and b). A role for miR-155, miR-205 and miR-206 in mammary epithelial morphogenesis is consistent with previous studies in other epithelial cell types.43, 44, 45
|
|
Title
|
Overexpression of Brca1-associated miRNAs affects mammary epithelial morphogenesis in vitro
|
|
Section
|
MMTV-miR-206 mice show a defect in mammary gland structure after 12 months
Given that miR-206 is upregulated in the mammary glands of Brca1 conditional knockout mice and in HC11 cells treated with a Brca1 siRNA, and limits the ability of HC11 cell morphogenesis in vitro, we prioritized miR-206 for in vivo analysis to further explore its role in mammary gland development and function. Transgenic mice expressing miR-206 under the control of the MMTV promoter were created and expression was confirmed in mammary glands at day 1 of lactation (Figure 2b). Analysis of the glands revealed no apparent differences in gross morphology or tissue architecture during pregnancy, lactation or involution (Figures 2c and d). This suggests either that miR-206 has no role in these processes or that any affect is masked by increased branching morphogenesis induced by the various hormonal cascades during pregnancy and lactation.46
In contrast, MMTV miR-206 transgenic mice at 12–15 months of age displayed a striking mammary phenotype. This was characterized by a significant increase in fatty tissue and a significant reduction of branching within the mammary epithelial tree, suggestive of tissue degeneration, in all of the MMTV miR-206 mice (n=5) and one of the six control mice (Figure 3a). No evidence of mammary tumors was observed prior to the end of study. Interestingly, several transgenic mouse models of Brca1 display a similar phenotype.47, 48, 49, 50, 51 Jones et al.50 argue that the Brca1 phenotype is due to the action of increased estrogen/IGF-1 on the mammary stroma microenvironment and that this may create an environment that is permissive to tumor development. The late onset of tumor development in mouse models of BRCA1 disruption is consistent with this.9 In support of this hypothesis, BRCA1 negatively regulates IGF-1 and loss of BRCA1 is associated with an increase in IGF-1.52 A plausible hypothesis is therefore that loss of Brca1 causes an increase in miR-206, which in turn results in tissue remodeling in older mice and that this permits Brca1-associated tumor development.
There are several potential molecular and cellular mechanisms for the observed phenotype, including a miR-206-associated increase in epithelial cell apoptosis, epithelial to mesenchymal transition, induction of IGF and/or adipose production. To explore this, we first used qRT-PCR to confirm that miR-206 was still overexpressed in the mammary glands of transgenic mice at 12 months of age. Interestingly, higher miR-206 expression corresponded to a stronger phenotype, concurrent with this, we found elevated levels of miR-206 in the single control mouse that displayed the phenotype (data not shown). We then examined the expression of epithelial and mesenchymal compartment markers within the aged mice mammary glands. β-catenin staining that detects the epithelial content of the mammary gland, revealed lesser ductal and end bud structures in the MMTV miR-206 glands as compared with the controls (Figure 3b). The expression of miR-200c, another epithelial marker, was found to be significantly decreased in MMTV miR-206 glands as compared with the controls, further corroborating the β-catenin staining (Figure 3c).
To explore the possibility that the IGF-1 pathway may also contribute to the phenotype in miR-206 mice, the expression of Igf1 was evaluated and normalized against the total volume of epithelial cells using miR-200c.53, 54, 55 The results revealed that Igf1 was significantly upregulated in the MMTV miR-206 glands as compared with the controls (Figure 3d). This supports the hypothesis that the Igf1 pathway may contribute to the phenotype in MMTV-miR-206 mouse mammary glands. As Igf1 is a secreted stromal factor that signals in a paracrine manner to the epithelial cell, this also raises the possibility that miR-206 induces an expansion of the stromal compartment, which could explain the apparent higher fat content seen in the MMTV miR-206 mammary glands as compared with the controls.56, 57, 58 Interestingly, mature Brca1-deficient mice also demonstrate an increase in mammary adipose tissue.50
Although the proposed expansion of the stromal compartment may promote tumorigenesis, no tumors were detected in the mammary glands of MMTV-miR-206 glands. It is possible that this reflects a dependence on other components of the Brca1 pathway or on intact p53 activity, which when lost rescues cellular aging but induces tumorigenesis.55
|
|
Title
|
MMTV-miR-206 mice show a defect in mammary gland structure after 12 months
|
|
Section
|
miR-206 targets the 3′UTR of the Wingless-type MMTV integration site (Wnt) antagonist, Sfrp1
To further explore the Brca1-miR-206 pathway, we identified potential gene targets of miR-206 (Figure 4a). Four genes were shortlisted on the basis of the results from four independent miRNA target prediction programs—Sfrp1, Daam1, Slc39a10 and Zbtb4 (Figure 4a). To validate these predictions, we examined the levels of these genes in the mammary glands of MMTV-Cre Brca1Co/Co mice. Only Sfrp1 had reduced expression in the lactating mammary gland of MMTV-Cre Brca1Co/Co mice (Figure 4b). Sfrp1 is a secreted antagonist of the wingless-type MMTV integration site (Wnt)-signaling pathway,59 whose expression is frequently lost in a variety of cancers, including breast cancer via promoter methylation.60 Repression of SFRP1 is associated with poor prognosis with increasing stage of malignancy.52, 61, 62, 63, 64, 65 Although many reports have focused on assessing the methylation status of the SFRP1 promoter, several studies have shown that some tumors exhibit decreased SFRP1 expression, independent of promoter methylation,52, 66 suggesting additional regulatory mechanisms may be involved. In addition, miR-206 has been shown to regulate Sfrp1 expression during myogenesis in a porcine model.67
A miR-206-binding site was identified in the Sfrp1 3′UTR and shown to be evolutionary conserved (Figure 4c). To determine whether miR-206 could repress the Sfrp1 3′UTR in mice, a luciferase reporter assay was conducted with the Notch 3′UTR, a known target of miR-206, as a positive control. Overexpression of miR-206, decreased the activity of both the Sfrp1 and Notch3 3′UTRs (Figure 4d), confirming that miR-206 can repress both Sfrp1 and Notch3. To further demonstrate that the reduced luciferase activity is a direct consequence of miR-206 targeting the predicted Sfrp1 3′UTR-binding site, mutations were introduced into the miRNA-binding site. Increased expression of miR-206 did not affect the activity of the mutated Sfrp1 3′UTR (Figure 4d), suggesting that the decrease in luciferase signal can be attributed to the binding of miR-206 to Sfrp1 at the predicted binding site. We also found that Sfrp1 expression was reduced in mice at day 1 of lactation and 12–15 months of age (Figure 4e). These results are consistent with the mammary epithelial phenotype observed and suggest that this phenotype is potentially a result of deregulated Wnt signaling. Though, there are conflicting reports of Sfrp1 function in the mammary gland, a recent study has shown that Sfrp1 deficiency induces increased adiposity upon forced weight gain in mice, consistent with our observations in aged MMTV miR-206 mice. Another team demonstrated that the knockout of Sfrp1 in mice mammary glands promotes precocious mammary gland development with branching and alveolar development characteristic of the midpregnant mammary gland,68, 69 a feature that has not been observed in the MMTV miR-206 mice (Figure 3a). A probable reason is that this phenotype has been masked by the upregulation of the Igf-1-signaling cascade.55, 70, 71
|
|
Title
|
miR-206 targets the 3′UTR of the Wingless-type MMTV integration site (Wnt) antagonist, Sfrp1
|
|
Section
|
Conclusions
In conclusion, this work has demonstrated that miR-206 is overexpressed in Brca1-deficient cells and that overexpression of miR-206, in several ways, mirrors the phenotype in MMTV-Cre Brca1Co/Co mice. This raises the possibility that miR-206 contributes to the effects of Brca1 loss in the mammary gland and that this likely involves the Wnt pathway through Sfrp1 and the Igf pathways. Although we focused on miR-206, our data concur with other reports in suggesting other miRNAs, such as miR-155,29 can regulate downstream functions of Brca1, highlighting that this and other miRNAs are an important facet of Brca1 biology and function. Understanding these mechanisms will be valuable for identifying new biomarkers and therapeutic targets and strategies for BRCA1-associated breast cancer.
|
|
Title
|
Conclusions
|
