Non-Coding RNAs as Regulators of Mammary Development and Breast Cancer

Abstract

Over the past decade, non-coding RNAs (ncRNAs) have become a new paradigm of gene regulation. ncRNAs are classified into two major groups based on their size: long non-coding RNAs (lncRNAs) and small non-coding RNAs (including microRNAs, piRNAs, snoRNAs, and endogenous siRNAs). Here we review the recently emerging role of ncRNAs in mammary development, tumorigenesis, and metastasis, with the focus being on microRNAs (miRNAs) and lncRNAs. These findings shed new light on normal development and malignant progression, and suggest the potential for using ncRNAs as new biomarkers of breast cancer and targets for treatment.

Introduction

The human genome contains approximately 20,000–25,000 protein-coding genes, far fewer than had been anticipated before its sequencing. In fact, only 1.5% of the human genome encodes proteins, whereas the rest consists of introns, regulatory sequences, and non-coding RNA genes [1–3]. To date, more than 1,500 miRNAs (www.mirbase.org) and over 8,000 long non-coding RNAs (lncRNAs) [4] are known to be encoded by the human genome.

The first miRNA, lin-4, was reported in 1993 by Victor Ambros and Gary Ruvkun, who discovered the first example of this type of molecule by genetic screening in C. elegans, and demonstrated that lin-4 turns off the expression of a specific gene, lin-14, whose protein product is crucial for development [5, 6]. The second miRNA, let-7, was found 7 years later, also through forward genetics approaches in worms [7, 8]. By now, miRNAs have been recognized as widely acting regulatory molecules. These endogenously expressed small non-coding RNAs are 18–25 nucleotides long, and they regulate gene expression through base pairing with target mRNAs, leading to degradation of target mRNAs and/or inhibition of their translation [9–11]. MiRNAs have been shown to orchestrate a variety of cellular processes, and the oncogenic or tumor-suppressing functions of a number of miRNAs have been characterized [12–17]. Moreover, we and others have demonstrated the existence of pro-metastatic and anti-metastatic miR-NAs [18, 19]; the molecular mechanisms by which these individual miRNAs function in malignant progression require further investigation.

In addition to the well-annotated protein-coding and miRNA genes, lncRNAs, transcripts that are more than 200 nt in length and lack significant open reading frames (ORFs), have also emerged as important regulatory molecules of cellular processes [20, 21]. A lncRNA could act as a scaffold that keeps proteins together, or as a guide that helps recruit proteins to specific genomic DNA sequences [20]. For instance, several well-characterized lncRNAs, including AIR, HOTAIR, and XIST, interact with chromatin-remodeling complexes and target them to specific genes, thereby affecting the ability of these complexes to regulate gene transcription [22–24].

In this article, we review recent progress in the regulation of mammary development and breast cancer by ncRNAs. These findings confirm and extend the importance of cellular regulatory RNAs in physiological and pathological processes, and illustrate the value of analyzing normal and neoplastic cells for new leads that may improve breast cancer outcomes.

Non-coding RNAs in Mammary Development

The basic components of a mature female mammary gland are the alveoli that are lined with milk-secreting cuboidal epithelial cells and surrounded by myoepithelial cells. The morphogenesis of mammary glands begins during embryonic development and proceeds through puberty, pregnancy, lactation, and involution [25]. Most of the mammary gland development and differentiation steps occur after birth. During these steps, the mammary gland undergoes profound morphological and functional changes [26], which involve cell proliferation, differentiation, and apoptosis in conjunction with extensive changes in gene expression [27–33], and are regarded as a succession of cell fate determinations [34]. Endocrine and paracrine signals, as well as cell-matrix and cell-cell interactions, regulate these processes.

The mammary morphogenesis has been extensively studied at the physiological, molecular, and genetic levels. In the past few years, ncRNAs, particularly miRNAs, have begun to be implicated in regulating mammary gland development (Table 1). Avril-Sassen et al. reported the first comprehensive miRNA profiling in the postnatal mammary gland throughout juvenile, puberty, mature virgin, gestation, lactation, and involution stages. MiRNAs were found to be represented in seven temporally co-expressed clusters, suggesting co-regulation of miRNA groups during mammary gland development [35]. This study has provided a valuable source for the future functional characterization of individual miRNAs. Consistent with another report that the let-7 miRNA family is underexpressed in mammary progenitor cells and that enforced let-7 expression induced loss of self-renewing cells [36], expression of many let-7 family members showed a peak in puberty and the mature virgin phase, followed by a marked reduction during lactation [35]. On the other hand, miR-22 and miR-205, miRNAs previously reported to promote expansion of mammary progenitor cells [36, 37], displayed enrichment during gestation [35]. Thus, these miRNAs may control the balance between mammary progenitors and differentiated cells. How such ncRNAs themselves are regulated during the course of differentiation remains unknown.

Table 1.

Non-coding RNAs in mammary development

Non-coding RNA	Function	Target gene or effector	Reference
miRNA
let-7	Inhibit self-renewal and promote differentiation; deplete mammary stem and progenitor cell population	RAS, HMGA2	[36]
miR-205	Expand mammary progenitor cell population	PTEN, ZEB1, ZEB2	[37]
miR-101a	Inhibit mammary epithelial cell proliferation and reduce β-casein expression	COX-2	[38]
miR-126-3p	Inhibit mammary epithelial cell proliferation and reduce β-casein expression	PR	[40]
miR-138	Suppress mammary epithelial cell viability and proliferation	PRL-R	[41]
miR-221	Suppress mammary epithelial cell viability and proliferation	GHR	[42]
miR-132/212 family	Regulate epithelial-stromal interaction	MMP-9	[43]
lncRNA
Zfas1	Suppress mammary epithelial cell proliferation and differentiation		[44]

Increased miR-101a levels were observed throughout differentiation and involution of mammary tissues [38]. Ectopic expression of miR-101a inhibited mammary epithelial cell proliferation and reduced expression of β-casein, a milk protein and a marker of cell differentiation; such changes correlated with the ability of miR-101a to directly downregulate cyclooxygenase-2 (COX-2) [38]. These findings suggest that miR-101a may regulate cell proliferation by targeting COX-2 expression, which may be important for the differentiation and involution of mammary glands. Future studies are needed to elucidate the relationship between miR-101a, COX-2, mammary cell proliferation and differentiation.

Mammary gland development is controlled by locally acting breast mitogens such as EGF, as well as endocrine hormones such as estrogen, progesterone, and prolactin [39]. It has been demonstrated that the mRNAs encoding receptors of these endocrine hormones are direct targets of specific miRNAs. For instance, the miR-126-3p miRNA has been shown to directly downregulate expression levels of progesterone receptor (PR), leading to a reduction in both mammary epithelial cell proliferation and in β-casein expression [40]. Interestingly, miR-126-3p is highly expressed in the virgin mammary gland and during involution, but is markedly downregulated during pregnancy and lactation, indicating that this miRNA could play an important role in postnatal mammary gland development [40]. Similarly, another two miRNAs, miR-138 and miR-221, can suppress the viability and proliferation of mammary epithelial cells, probably through their ability to downregulate prolactin receptor (PRL-R) and growth hormone receptor (GHR), respectively [41, 42].

Molecules whose expression is restricted to the mammary stroma can also be critical regulators of mammary development through epithelial-stromal interaction. Mice lacking the miR-212/132 family (which comprises miR-212 and miR-132) had normal rudimental ductal trees in the pre-pubertal phase, but displayed profound defect in pubertal ductal outgrowth [43]. Transplantation experiments in which wild-type mammary epithelial cells were implanted into the cleared fat pads of miR-212/132 knockout mice recapitulated the ductal outgrowth defect, whereas pre-pubertal miR-212/132-null mammary epithelia transplanted into the cleared wild-type fat pads formed normal mammary glands, which suggested that these two miRNAs function in the stroma but not in the epithelium [43]. Indeed, both miR-212 and miR-132 are expressed exclusively in the mammary stroma and directly target the mRNA encoding the matrix metalloproteinase MMP-9; mammary glands that lack these two miRNAs displayed accumulation of MMP-9 around the ducts and hyperactivation of the TGF-β pathway, which might underlie the impaired ductal outgrowth [43]. These findings represent the first example of a miRNA that functions in a developmental process through the tissue microenvironment.

Although lncRNAs are much less annotated and characterized functionally compared with proteins and miRNAs, some of them have also been recognized as regulators of developmental processes. Mattick and colleagues found that a non-coding transcript antisense to the protein-coding transcript Znfx1, named Zfas1, is among the highest and most differentially expressed lncRNAs during mouse mammary development [44]. Zfas1 hosts three small nucleolar RNAs (snoRNAs), but intriguingly, knockdown of Zfas1 in a mammary epithelial cell line resulted in an increase in cell proliferation and differentiation without altering levels of the three snoRNAs. In support of this result, the Zfas1 transcript is remarkably stable, with a half-life of over 16 h [44]. Thus, in contrast with the assumption that the host transcripts of snoRNAs act only as originators of snoRNAs, a lncRNA transcript which generates snoRNAs can actually have a function independent of the snoRNAs it gives rise to; the molecular mechanisms underlying this snoRNA-independent function remain to be determined.

Non-coding RNAs in Mammary Tumorigenesis

Deregulation of oncogenic and tumor-suppressing pathways results in neoplastic transformation and tumorigenesis, and these pathways have been shown to be controlled by small and large ncRNAs at the transcriptional, post-transcriptional, and epigenetic levels [17, 20, 21]. Iorio et al. performed miRNA microarray analysis of clinical breast cancers. They found that 29 miRNAs were differentially expressed in human mammary tumors compared with normal breast tissues, and that miR-21 and miR-155 were among the most consistently upregulated miRNAs [45]. In fact, both of these two miRNAs have been recognized as oncogenic miRNAs and are overexpressed in a variety of human cancers. Transfection of MCF-7 human breast cancer cells with anti-miR-21 oligonucleotides resulted in a reduction of cell proliferation in vitro as well as tumor growth in xenograft experiments, which was accompanied by increased apoptosis and decreased expression of the BCL-2 anti-apoptotic protein [46]. Subsequently, the mRNAs produced by several tumor suppressor genes have been identified as direct miR-21 targets in breast cancer and other cancer cells, including PDCD4 (programmed cell death 4) [47], TPM1 (tropomyosin 1) [48] and PTEN (phosphatase and tensin homolog on chromosome 10) [49]. In a synthetic miRNA library screen to identify modulators of the TRAIL (tumor necrosis factor–related apoptosis-inducing ligand)-induced apoptosis pathway in the MDA-MB-453 human breast cancer cell line, miR-155 stood out as a potent suppressor of apoptosis [50]. Ectopic expression of miR-155 in breast cancer cells promoted proliferation and colony formation in soft agar in vitro and development of tumors in nude mice; these effects were found to depend on the negative regulation of SOCS1 (suppressor of cytokine signaling 1) by miR-155 [51].

The oncogenic roles of miR-21 and miR-155 have been further validated in genetically engineered mouse models (GEMMs). Croce and colleagues generated Eμ-miR-155 transgenic mice in which the expression of miR-155 is under the control of a V_H promoter-Ig heavy chain Eμ enhancer, which becomes active at the late pro-B cell stage of B cell development. These mice initially exhibited a pre-leukemic pre-B cell proliferation in spleen and bone marrow, which was followed by lymphoblastic leukemia and high-grade lymphoma, providing strong evidence that miR-155 is directly implicated in the initiation and progression of these diseases in which this miRNA is overexpressed [52]. Slack and colleagues developed a miR-21 transgenic line in which Cre and Tet-Off technologies were employed to achieve tissue-specific and doxycycline-controlled expression of miR-21; this model exhibited 16-fold overexpression of miR-21 in hematopoietic tissues and development of pre-B-cell lymphoma, both of which could be completely reversed within 1 week of doxycycline treatment, demonstrating that miR-21 is responsible for both initiation and maintenance of lymphoma [53]. Taken together, miR-21 and miR-155 are bona fide oncogenic ncRNAs in lymphoma; whether these two miRNAs are sufficient and/or required for breast cancer initiation and progression remains to be determined in GEMMs.

Interestingly, the second miRNA discovered in C. elegans and the first miRNA discovered in human, the let-7 miRNA family, has been shown to be downregulated in many human cancers as well as in stem cells [54, 55]. In particular, a recent study found that the expression of let-7 is markedly reduced in breast tumor-initiating cells (BT-ICs) and increases with differentiation. Knockdown of let-7 induced self-renewal of non-BT-ICs; conversely, restoring let-7 expression in BT-ICs decreased their proliferative potential, their ability to form mammospheres, as well as tumor formation and metastasis in vivo. Known oncogenic targets of let-7, H-RAS and HMGA2, mediate the effect of let-7 on suppressing tumorigenicity [56]. In breast cancer cells, CCND1 (encoding cyclin D1), an oncogene overexpressed in many human tumor types, has been recently identified as a direct target of the miR-17-5p/miR-20a miRNA cluster. miR-17/20 functions to suppress proliferation and colony formation of breast cancer cells by inhibiting cyclin D1 protein synthesis; on the other hand, cyclin D1 induces miR-17-5p and miR-20a expression [57]. Therefore, this miRNA cluster serves as a negative feedback mechanism to limit cyclin D1 expression, which may impede tumor formation or progression. Additional miRNAs, such as miR-145 [58], miR-34a [59], miR-214 [60], and miR-205 [61, 62], have also been shown to function as putative tumor-suppressing miR-NAs in breast cancer cells (Table 2). In vivo confirmation of the role of these miRNAs in mammary tumorigenesis still awaits new transgenic and knockout mouse models.

Table 2.

Non-coding RNA in mammary tumorigenesis and metastasis

Non-coding RNA	Function	Target gene or effector	Reference
miRNA
miR-21	Increase tumor growth and decrease apoptosis	TPM1, PDCD4, PTEN, MASPIN	[45–49, 53]
miR-155	Promote proliferation and oncogenic transformation	SOCS1	[50–52]
let-7	Inhibit tumorigenicity and stemness	H-RAS, HMGA2	[56]
miR-17/20	Suppress proliferation and oncogenic transformation	CCND1	[57]
miR-145	Promote apoptosis	ESR1	[58]
miR-34a	Induce growth arrest and apoptosis	AXL	[59]
miR-214	Inhibit cell proliferation and invasion	EZH2	[60]
miR-205	Suppress cell growth and invasion; suppress EMT and induce MET	HER3, VEGFA, ZEB1, ZEB2	[61, 62, 91]
miR-10b	Induce tumor invasion and metastasis	HOXD10	[72–76]
miR-9	Promote migration, invasion, and metastasis	CDH1	[72, 77]
miR-373	Promote migration, invasion, and metastasis	CD44	[78]
miR-103/107	Promote migration, invasion, and metastasis	Dicer	[81]
miR-221/222	Promote EMT, migration, and invasion	ESR1, Dicer, TRPS1	[82–84]
miR-126	Suppress proliferation and tumorigenesis		[85]
miR-206	Inhibit cell growth, motility, and invasiveness	ESR1	[85, 114]
miR-335	Inhibit motility and invasiveness	SOX4, TNC	[85]
miR-31	Suppress multiple steps of breast cancer metastasis	ITGA5, RDX, RHOA	[86, 87]
miR-146a/b	Reduce migration, invasion, and metastasis	IRAK1, TRAF6	[89, 90]
miR-200 family	Suppress EMT and induce MET; inhibit migration but promote metastatic colonization	ZEB1, ZEB2, sec23a	[91–96]
lncRNA
GAS5	Promote growth arrest and apoptosis	GREs	[63, 64]
MALAT1	Promote tumor formation		[65–68]
BC200	Is associated with invasive breast cancer		[69]
HOTAIR	Induce metastasis	HOXD genes	[23, 97, 98]

Aided by the rapid development and widespread availability of new tiling microarray and high-throughput RNA sequencing technologies, investigators are beginning to probe lncRNAs that are differentially expressed in breast cancer and other cancers compared with normal tissues (Table 2). GAS5 (growth arrest-specific 5) has been observed to be downregulated in clinical breast cancer tissues, and overexpression of this lncRNA in the MCF-7 breast cancer cell line promoted growth arrest and apoptosis [63]. Interestingly, GAS5 is induced in cells that are growth-arrested due to lack of nutrients or growth factors, and it competes with glucocorticoid response elements (GREs) for binding to glucocorticoid receptor (GR), thereby blocking transcriptional activation by GR, which, in turn, leads to reduced cell metabolism [64]. Hence, downregulation of GAS5 in cancer may help to keep tumor cells viable and metabolically active even under starvation conditions. The highly conserved 8 kb lncRNA MALAT-1 (metastasis associated lung adenocarcinoma transcript 1) is overexpressed in many different cancer types, including lung, breast, colon, prostate, pancreatic, and hepatocellular carcinomas [65–67]. Overexpression of MALAT-1 in the NIH3T3 cell line and a human melanoma cell line promoted colony formation in soft agar and tumor formation in nude mice, while silencing MALAT-1 in tumor cells by RNAi inhibited tumorigenicity [68], which suggested that MALAT-1 is an oncogenic lncRNA. BC200, a neuron-specific lncRNA, is upregulated in invasive breast carcinomas and correlates with tumor grade [69]. It is tempting to speculate that this lncRNA might contribute to the development of aggressive mammary tumors.

Non-coding RNAs in Breast Cancer Metastasis

Ninety percent of solid tumor-related deaths are caused by metastasis, a multi-step process which begins as primary tumor cells invade adjacent tissues and enter the blood and lymphatic vessels. These cells travel through the vasculature, extravasate into the parenchyma of distant organs, and finally proliferate from minute growths (micrometastases) into macroscopic secondary tumors [70]. A growing list of miRNAs has been implicated in metastasis of various tumor types and has been recently reviewed [19, 71]; here we focus on miRNAs which are functionally important in breast cancer metastasis.

In an initial screen for miRNAs differentially expressed in human breast cancer cells, the three most significantly upregulated miRNAs were identified: miR-10b, miR-9, and miR-155 [72]. Among them, miR-10b is highly expressed in cultured metastatic cancer cells as well as in metastatic breast tumors from patients. Overexpression of miR-10b triggered tumor invasion and distant metastasis from primary mammary tumors in xenograft models. miR-10b is induced by the epithelial-mesenchymal transition (EMT)-inducing transcription factor Twist, and this miRNA inhibits synthesis of the HOXD10 protein, permitting expression of the metastasis-promoting protein RhoC [72]. This was the first report of a miRNA that regulates metastasis. Recently, several independent groups showed that miR-10b induces metastatic behaviors of various types of tumor cells in vitro and in vivo, and that in advanced-stage tumors, miR-10b correlates with high-grade malignancy in clinical cancers, including breast cancer, pancreatic cancer, and glioblastoma [73–76]. A second miRNA that stood out in the initial screen, miR-9, has been shown to directly target CDH1, the E-cadherin-encoding mRNA, leading to increased cell migration/invasion and to a context-dependent EMT-like conversion [77]. miR-9-mediated downregulation of E-cadherin leads to the activation of β-catenin signaling, which in turn contributes to increased expression of VEGF; this then results in enhanced tumor angiogenesis. Overexpression of miR-9 in otherwise non-metastatic breast cancer cells induced metastasis in mice, whereas silencing miR-9 in highly malignant cells suppressed metastasis formation [77]. Whether miR-9 promotes metastasis in an E-cadherin-independent manner (i.e., in E-cadherin-negative tumor cells) remains to be investigated.

Huang, Agami, and colleagues exploited a functional genomics approach to screen for miRNAs that promote cell motility. They infected the non-metastatic breast cancer cell line MCF-7 with a miRNA expression library and then used Transwell migration assays as the functional readout. Both miR-373 and miR-520c increased cell migration and were further found to induce metastasis in vivo. In clinical breast cancers, miR-373 is upregulated in lymph node metastases compared with paired primary tumors [78]. The key enzyme of miRNA biogenesis, Dicer, has been identified as a suppressor of metastasis [79] and epithelial-mesenchymal transition (EMT), an embryonic program that has been proposed to be resurrected by carcinoma cells in order to acquire motility and invasiveness [80]. Interestingly, a miRNA family that attenuates miRNA biosynthesis by targeting Dicer, miR-103/107, induces EMT and metastatic dissemination of otherwise non-metastatic breast cancer cells [81]. miR-221/ 222, a miRNA cluster that targets ESR1 (encoding estrogen receptor) [82], has also been shown to directly downregulate Dicer [83] and TRPS1 (trichorhinophalangeal syndrome type 1) [84], leading to induction of EMT in breast cancer cells [84]. These findings provided an explanation for the observed high expression of miR-221/222 and low expression of Dicer in ER-negative, basal-like breast cancer, and suggested a possible metastasis-promoting role of miR-221/ 222 in this tumor type which is highly aggressive and non-responsive to hormonal therapy.

Several miRNAs have been established as anti-metastatic molecules. miR-126, miR-206, and miR-335 were found to be downregulated in the highly metastatic variants of the MDA-MB-231 breast cancer cell line [85]. Overexpression of these miRNAs in the metastatic variants reduced both lung and bone metastases, but their mechanisms of action are different: miR-126 suppressed proliferation and tumorigenesis but had little effect on migration and invasion; in contrast, miR-335 or miR-206 inhibited motility and invasiveness, but did not affect the proliferation or apoptosis of primary tumor cells [85]. By using gain-of-function and loss-of-function approaches, Valastyan et al. demonstrated that miR-31 is a suppressor of breast cancer metastasis. This suppressing effect was not seen on primary tumor growth, but instead was observed at multiple steps of the invasion-metastasis cascade, including local invasion, anoikis, and metastatic colonization [86, 87]. Nuclear factor κB (NFκB) is a transcription factor associated with enhanced survival and metastasis of cancer cells [88]. Two miRNAs, miR-146a and miR-146b, suppress NFκB activity through their ability to target IRAK1 and TRAF6 in the MDA-MB-231 breast cancer cell line, leading to reduced migration and invasion in vitro as well as suppressed experimental lung metastases in vivo [89, 90]. Collectively, these miRNAs can impede metastasis through inhibition of either a single step or multiple steps of metastasis.

miR-205 and the miR-200 family (miR-200a, miR-200b, miR-200c, miR-141, and miR-429), miRNAs that target the EMT-inducing transcription factors ZEB1 and ZEB2, are downregulated in cells that have undergone EMT [91, 92]. Overexpression of miR-200 inhibited transforming growth factor (TGF)-β-induced EMT; conversely, silencing miR-200 in epithelial cell lines resulted in EMT [91, 92]. Interestingly, ZEB1 directly represses the transcription of miR-200 genes [93]. Consistent with its role in suppressing EMT and promoting mesenchymal-epithelial transition (MET), miR-200 has been found to inhibit cell migration [94]. However, when overexpressed in the 4TO7 mesenchymal-like breast cancer cells, miR-200 promoted metastatic colonization, i.e., formation of macroscopic lung metastases [95, 96]. These results suggest that miR-200 plays opposing roles at early and late stages of metastatic process, and underscore the importance of both EMT and MET in metastasis.

Certain lncRNAs exhibit distinct expression patterns between primary tumors and metastases. A 2.2 kb lncRNA, HOTAIR, is significantly overexpressed in clinical breast cancers. Moreover, HOTAIR levels in primary mammary tumors correlate with metastasis and poor survival rate [97]. HOTAIR serves as a scaffold molecule that bridges the Polycomb Repressive Complex 2 (PRC2) and the LSD1 H3K4 demethylase complex together and recruits these two complexes to target genes, such as the HOXD locus, leading to repressed transcription of specific genes and induction of metastasis [23, 97, 98]. These findings illustrate how a specific lncRNA can remodel the epigenetic machinery to repress metastasis suppressor genes (e.g., HOXD10). A future challenge will be to identify the full complement of functional lncRNAs in breast cancer formation, progression, and metastasis, and to determine their roles in normal development as well as their mechanisms of action.

Non-coding RNAs as Breast Cancer Biomarkers and Therapeutic Targets

mRNA profiles have been used to define molecular subclasses of breast cancer, i.e., luminal A, luminal B, basal-like, HER2-amplified, and normal-like, which are predictive of prognosis [99]. A further interrogation by miRNA expression profiling might enable more precise classification. Using a bead-based flow cytometric method, Lu et al. systematically profiled 217 mammalian miRNAs in 334 samples, including multiple human cancers. A general downregulation of miRNAs in tumor tissues was observed compared with normal tissues, and remarkably, the expression pattern of this relatively small set of miRNAs classified tumor types better than the expression data of 16,000 mRNAs [100]. These results underscore the potential of miRNA profiling in cancer diagnosis.

Iorio et al. performed a miRNA microarray analysis of 76 breast tumors and 34 normal tissues, and found that miRNA expression was associated with clinical parameters such as estrogen receptor (ER)/progesterone receptor (PR) status (e.g., miR-30) and tumor stage (e.g., miR-213 and miR-203) [45]. Several let-7 miRNA family members correlated with PR status (let-7c), high proliferative index (let-7c and let-7d), or lymph node metastasis (let-7f-1, let-7a-3, and let-7a-2) in breast tumors [45]. Mattie et al. identified distinct miRNA subsets in human breast tumors that correlated with their ER/PR status or HER2 status [101], and notably, there is a significant overlap between the miRNAs in these profiles and Iorio’s miRNA profiles [45].

In pursuit of non-invasive/minimally invasive biomarkers for cancer diagnosis, investigators have found stable circulating miRNAs in serum from cancer patients [102, 103]. The increased levels of tumor-associated circulating miR-NAs were first identified in serum of diffuse large B-cell lymphoma patients [104]. Subsequently, higher levels of circulating miRNAs, e.g., miR-155 and miR-195, were found in serum from breast cancer patients than in serum from healthy individuals [105–107]. These studies have provided new promise for improved cancer detection.

Development of miRNA-based therapeutic strategies is under intensive investigation. The feasibility of restoring cancer-suppressing miRNAs and silencing cancer-promoting miRNAs has been addressed in preclinical models of various tumor types. Several tumor-suppressing miR-NAs have been actively studied as targets for restoration. Intratumoral injection of lipid-based let-7 or miR-34 miRNA mimics reduced tumor growth in a subcutaneous H460 NSCLC model [108, 109]. In a K-RAS model of NSCLC, systemically delivered let-7 or miR-34 was well tolerated and effectively inhibited lung tumors [110]. In an orthotopic model of pancreatic cancer, systemic delivery of miR-34 or miR-143/145 decreased tumor growth with no signs of toxicity [111]. In addition, systemic miR-34 delivery suppressed prostate cancer metastasis and improved survival in mice [112]. On the other hand, miRNAs that promote tumor formation and/or metastasis can be silenced therapeutically using engineered antisense RNA oligonucleotides. In collaboration with Regulus Therapeutics (Carlsbad, CA), we developed a miR-10b antagomir—a new class of chemically modified, cholesterol-conjugated anti-miRNA oligonucleotides, and tested it in a 4T1 mouse mammary tumor model (4T1 cells are highly metastatic and express high miR-10b levels). Systemically delivered miR-10b antagomirs had a potent and sequence-specific metastasis-suppressing effect on 4T1 cells without affecting their ability to grow as primary tumors. This work is the first report of systemic delivery of antagomirs for cancer therapy in mice [113]. Taken together, these studies suggest the potential for manipulating miRNAs as therapeutic strategies. Whether the benefits observed in preclinical models can be recapitulated in human patients remains to be seen. Doses and schedules of treatment with such agents need to be carefully assessed.

Concluding Remarks

Here we have reviewed ncRNAs which regulate mammary development and breast cancer. Every single year, more than 1.3 million women will be diagnosed with breast cancer in the world, and around 465,000 will die from this disease (Susan G. Komen For The Cure; ww5. komen.org). There is an urgent need to diagnose and treat breast cancer appropriately at its early stage. The emerging roles of regulatory RNAs in breast cancer formation and progression have stimulated extensive efforts toward using ncRNAs as cancer biomarkers and therapeutic targets. Circulating miRNAs could be valuable as diagnostic and/or prognostic factors. In the next decade, we anticipate big steps forward in developing ncRNA-based breast cancer therapy.

Abbreviations

AIR

Antisense to IGF2R

BC200

Brain cytoplasmic RNA 200 nt

COX-2

Cyclooxygenase-2

EMT

Epithelial-mesenchymal transition

ER

Estrogen receptor

GAS5

Growth arrest-specific 5

GEMM

Genetically engineered mouse model

GHR

Growth hormone receptor

GR

Glucocorticoid receptor

GRE

Glucocorticoid response element

HER2

Human epidermal growth factor receptor 2

HMGA2

High-mobility group AT-hook 2

HOTAIR

Hox antisense intergenic RNA

IRAK1

Interleukin-1 receptor-associated kinase 1

lncRNA

Long non-coding RNA

LSD1

Lysergic acid diethylamide 1

MALAT1

Metastasis associated lung adenocarcinoma transcript 1

MET

Mesenchymal-epithelial transition

miRNA

microRNA

MMP-9

Matrix metalloproteinase-9

ncRNA

Non-coding RNA

NFκB

Nuclear factor κB

NSCLC

Non-small cell lung cancer

PDCD4

Programmed cell death 4

piRNA

Piwi-interacting RNA

PLR-R

Prolactin receptor

PR

Progesterone receptor

PRC2

Polycomb repressive complex 2

PTEN

Phosphatase and tensin homolog on chromosome 10

siRNA

Small-interfering RNA

snoRNA

Small nucleolar RNA

SOCS1

Suppressor of cytokine signaling 1

TGF-β

Transforming growth factor β

TPM1

Tropomyosin 1

TRAF6

TNF receptor associated factor 6

TRAIL

Tumor necrosis factor–related apoptosis-inducing ligand

TRPS1

Trichorhinophalangeal syndrome type 1

VEGF

Vascular endothelial growth factor

XIST

X-inactive-specific transcript

ZEB1

Zinc finger E-box binding homeobox 1

ZEB2

Zinc finger E-box binding homeobox 2

Zfas1

Antisense to the 5′ end of the protein-coding gene Znfx1