Emerging Functions of microRNA-146a/b in Development and Breast Cancer

Abstract

MicroRNAs (miRNAs) are a class of small non-coding RNAs that regulate gene expression through translational repression or mRNA degradation. These molecules play critical roles in regulating normal developmental processes, but when deregulated, are causally linked to the pathogenesis of numerous diseases, including cancer. MicroRNA-146a and -146b are encoded by two different genes, but differ by only two bases and appear to function redundantly in many systems. Initial studies branded miR-146a/b as important mediators of inflammatory signaling, documenting the ability of these miRNAs to influence differentiation, proliferation, apoptosis and effector immune mechanisms within the hematopoietic system. Numerous contemporary studies now implicate miR-146a/b as pleiotropic regulators of tumorigenesis, as a polymorphism in miR-146a and altered expression of both miR-146a/b have been linked with cancer risk, tumor histogenesis and invasive and metastatic capacity in diverse cancers. Despite the numerous reports concerning miR-146a/b in human cancers, the mechanistic contributions of these miRNAs in both normal and neoplastic mammary gland development and biology remains poorly characterized.

Introduction

The exons of protein-coding genes make up less than 2% of the human genome [1]. Though these sequences are certainly the best-studied elements of our genetic constitution, it is now apparent that the expansive non-coding regions, originally classified as ‘junk’ DNA [2], have contributed to the current structure and organization of the human genome and play critical roles in maintaining cellular homeostasis. Indeed, the perceived desert of vestigial genetic information is now known to contain functionally important transposable elements, regulatory sequences, and a diverse cadre of non-coding RNAs (ncRNAs) [1]. Genetic changes in protein-coding sequences, and the functional consequences in the proteins derived thereof, have been extensively studied in human cancers. The study of malignancy-associated perturbations in ncRNAs has revealed recurrent abnormalities in these genetic elements that appear to have drastic effects on cellular function and are causally related to cancer pathogenesis. The microRNAs (miRNAs) are the most widely studied class of ncRNAs. MiRNAs are a discrete class of small ncRNAs of 19-24 nucleotides in length that regulate gene expression post-transcriptionally by suppressing mRNA translation or promoting mRNA degradation [3]. To date, more than 1,424 miRNAs have been identified in the human genome, and they are believed to collectively regulate the translation of more than 60% of protein-coding genes [4]. Not surprisingly, miRNAs have been experimentally linked to both normal physiology and pathology, including regulation of developmental processes, proliferation, differentiation and apoptosis.

The biogenesis of miRNAs (Fig. 1) begins by RNA polymerase II or III-dependent transcription. MiRNAs that reside in close proximity to repetitive elements (e.g., Alu retrotransposons) have been noted to use RNA polymerase III, while those at non-repetitive loci and those embedded within introns (‘mirtrons’) of protein-coding genes generally are transcribed by RNA polymerase II [5, 6]. Primary miRNAs (pri-miRNA) are produced from miRNA genes which contain their own promoters, whereas mirtrons are formed from splicing stem-loop structures present in pre-mRNA transcripts. Both of these classes undergo additional processing to form precursor miRNAs (pre-miRNA), with pri-miRNAs requiring the RNase III enzyme Drosha, and the mirtron-derived lariat structures employing the lariat-debranching enzyme DBR1 [7, 8]. Exportin 5 (XPO5) shuttles pre-miRNAs to the cytoplasm, at which point they are processed into mature miRNAs by another RNase III enzyme, Dicer [9–12]. The dsRNA binding protein TRBP binds to Dicer and loads the miRNA into the Argonaute 2 (AGO2) endonuclease [13, 14]. The Dicer-TRBP-AGO2 complex, referred to as RNA-induced silencing complex (RISC), directs complementary binding of miRNAs to target mRNAs (generally in the 3′ untranslated region of the mRNA), and facilitates translational repression by promoting mRNA degradation or preventing translation initiation [3, 14]. MiRNAs may regulate targets cooperatively, and many demonstrate remarkable promiscuity, targeting several or even hundreds of mRNAs [15].

Figure 1.

Biogenesis and function of miRNAs. MiRNAs are transcribed as primary miRNAs (pri-miRNA) or are derived from stem-loop structures (“Mirtrons”) within pre-mRNAs. Drosha processes pri-miRNAs whereas the stem-loop-derived structures depend on the lariat-debranching enzyme DBR1 to form precursor miRNAs (pre-miRNAs). Exportin 5 (XPO5) exports pre-miRNAs into the cytoplasm where they are further processed by the Dicer/TRBP complex into mature miRNAs. Argonaute 2 (AGO2) joins the Dicer/TRBP complex to generate the RNA-induced silencing complex (RISC) which mediates binding to complimentary sequences in target mRNAs and translational repression or mRNA degradation

In addition to the role of miRNAs in development and normal cellular homeostasis, many miRNAs can act as oncogenes or as tumor suppressor genes, and when dysregulated, have been shown to affect key aspects of tumor initiation, progression and response to therapy. The remainder of this review will focus on the biologic functions of two pleiotropic miRNAs, miR-146a and miR-146b, with a special emphasis of the roles of these miRNAs in mammary gland development and neoplasia.

Structure and Function of miR-146a and miR-146b

Human miR-146a (hsa-mir-146a; MIR146A) is located at 5q34 and was originally discovered by sequence homology to a novel miRNA found in the murine heart [16, 17]. MiR-146b (hsa-mir-146b; MIR146B) is located at 10q24.32, and was discovered through a combined bioinformatics and sequence-directed cloning method [18]. The stem-loop 5′ proximal strand of both miR-146a and miR-146b (miR-146a-5p and miR-146b-5p) are far more prevalent than the 3′ proximal ‘passenger’ strands (miR-146a-3p and miR-146b-3p), and unless noted otherwise, the remainder of this paper will discuss the biologic activities of the - 5p derived miRNAs (hereafter referred to as miR-146a and miR-146b). Structurally, the mature miR-146a and miR-146b differ by only two nucleotides at the 3′ end, a region that is believed to play a minor role in target recognition, and in many instances miR-146a and miR-146b have demonstrated redundant activities [19] (Fig. 2).

Figure 2.

Structure of miR-146a/b. Predicted stem-loop structures of miR-146a and miR-146b. Structures predicted by mfold (The RNA Institute at SUNY Albany). The mature miR-146a-5p and miR-146b-5p differ by only two bases in the 3′ region

MiR-146a and miR-146b in Hematopoiesis and Immune Function

One of the earliest characterized functions of miR-146a/b came from a study by Taganov et al. aimed at identifying miRNAs induced in human mononuclear cells treated with lipopolysaccharide (LPS). MiR-146 was among three miRNAs whose expression was strongly induced after LPS challenge. Characterization of the signaling pathways involved in LPS-induced miR-146 expression revealed two functional and necessary nuclear factor kappa-B (NFκB) binding sites in the MIR146A promoter, though no discernible NFκB sites were noted in the LPS-inducible MIR146B promoter [20]. This same study noted that the 3′UTRs of the Interleukin (IL)-1 receptor associated kinase (IRAK1) and TNF receptor-associated factor 6 (TRAF6) mRNAs contained functional miR-146 binding sites and that both miR-146a and miR-146b could suppress translation of a luciferase reporter containing the wild-type IRAK1 and TRAF6 UTRs [20]. Both IRAK1 and TRAF6 are important adaptor molecules necessary for induction of the pro-inflammatory NFκB and activating protein-1 (AP-1) signaling pathways in response to activation of toll-like receptors (TLR) and the IL-1 receptor. Subsequent studies have documented an important role of miR-146a in response to additional TLR ligands and cytokines, illuminating the important role of miR-146a/b in modulating the innate immune response [20–24]. As discussed later, the ability of miR-146a/b to modulate NFκB and other pro-inflammatory signaling pathways may bring about many of its tumor suppressor activities.

Experimental knockout of the Mir146a gene in mice did not result in apparent morphologic abnormalities, and knockout mice appeared to be as viable and fertile as heterozygous and wild-type homozygous littermates [21]. Though unremarkable early in life, Mir146a knockout mice exhibit several phenotypes consistent with the known roles of miR-146a in regulating inflammatory signaling. Knockout mice exhibited profound sensitivity to LPS, both in secretion of sepsis-associated cytokines and in mortality. Moreover, starting at 6-8 months of age, Mir146a knockout mice developed immunoproliferative and autoimmune diseases characterized by splenomegaly, lymphadenopathy, and had a significantly shorter lifespan. Monocytic and lymphocytic infiltrates were observed in numerous organs, as was evidence of aberrant activation of peripheral CD4⁺ and CD8⁺ T cells. Dramatic expansion of the myeloid line-age was observed in knockout animals, as was anemia, thrombocytopenia and lymphopenia. Older knockout animals were also noted to have a significantly increased risk of malignant hematolymphoid neoplasia [21]. The findings in the Mir146a knockout mice bear a striking resemblance to myelodysplastic syndrome (MDS), a heterogeneous group of disorders typified by abnormal or ineffective hematopoiesis and increased risk of myeloid-lineage cancer. Prior to the description of the Mir146a knockout mouse, Starczynowski and colleagues documented reduced expression of miR-146a (and miR-145) in one of the most common subtypes of MDS, the 5q- subtype [25]. This same study also demonstrated that stable knockdown of miR-145 and miR-146a in murine hematopoietic stem/progenitor cells (HSPCs) could elicit the 5q- syndrome, causing thrombocytopenia and variable neutropenia [25]. Overexpression of TRAF6, a previously identified target of miR-146a, could also mimic the 5q- syndrome when its expression was enforced in HSPCs [20, 25]. Complementary studies employing miR-146a overexpression resulted in impaired megakaryopoiesis [26, 27]. As of the writing of this manuscript, a Mir146b knockout mouse has not been reported.

MiR-146a/b in Epithelial Cell Homeostasis

Recent studies suggest that miR-146a/b may play key roles in regulating many cellular functions including differentiation, proliferation and survival. For example, Liu et al. showed that miRNA-146a was significantly up-regulated in response to inflammatory cytokines such as TGF-β1, IL-1β, INF-γ and TNF-α in human bronchial epithelial cells (HBEC). MiR-146a up-regulation in response to the inflammatory cytokines may protect the cells from apoptosis and may induce cellular proliferation through STAT3 phosphorylation and Bcl-XL up-regulation. By the same mechanisms, miR-146a may also protect cells from smoke-induced DNA damage and repair [28]. Furthermore, miR146a promotes vascular smooth muscle cell (VSMC) proliferation in vitro and vascular neointimal hyperplasia in vivo, by targeting Krüppel-like factor 4 (KLF4) 3′-untranslated region [29]. KLF4 has an anti-proliferative effect on VSMCs by up-regulating p21 [29].

MiR-146a/b in Cancer

The role of miR-146a/b in inhibiting invasion and metastasis of different tumor types has been reported. Welch and colleagues recently coined the term “metastamir” to refer to metastasis regulatory miRNAs that have an impact on critical steps in the metastatic cascade, such as epithelial-mesenchymal transition (EMT), apoptosis, and angiogenesis [30]. This group has classified miR-146a/b as a “metastasis-suppressing metastamir”.

Several studies have elucidated the role of miRNAs in thyroid cancers [31–33], specifically in papillary thyroid cancers (PTC), a common type of thyroid malignancy. PTC is characterized by alterations in the RET/PTC-RAS-BRAF signaling pathways [34]. A study by He, H., et al. showed that three miRNAs (miR-146b, miR-221, and miR-222) were dramatically overexpressed, with 11- to 19-fold higher levels in PTC compared to the adjacent unaffected thyroid tissue [32]. Recently, a comprehensive study of microRNAs in aggressive forms of PTC was done by Yip and colleagues [31]. PTCs were classified as aggressive in patients that exhibited local recurrence and/or distant metastasis compared with PTC without those complications on similar follow-up. They concluded that among tumors with BRAF mutations, overexpression of miR-146b was associated with aggressive behavior. There was no mechanistic insight into how miR-146b may be associated with aggressive form of PTC, but it could potentially serve as a biomarker of advanced disease. Moreover, these findings suggest that miR-146b may further refine the prognostic importance of BRAF mutation [31].

MiR-146b has also been reported to suppress glioblastoma multiforme (GBM). GBM (WHO grade IV) is known to be the most malignant and prevalent form of primary intracranial tumor. In one study, the expression profile of miR-NAs in human GBM tumor tissue was evaluated using miRNA microarray and it was concluded that miR-146b inhibits GBM cell migration and invasion by targeting matrix metalloproteinase 16 (MMP16), a protease which is selectively expressed in the brain and has a proteolytic activity against the extracellular matrix (ECM) [35]. Another study demonstrated that miR-146b can inhibit cell migration and invasion by targeting epidermal growth factor receptor (EGFR) [36]. Elevated EGFR expression and invasiveness are hallmarks of glioma and increase with malignancy grade [37]. Furthermore, miR-146a has been also shown to inhibit the development of gliomas and to increase the survival of mice bearing human glioblastoma xenografts. The same study revealed that miR-146a inhibits the formation of glioma stem-like cells from malignant astrocytes by targeting Notch-1 [38].

The anti-metastatic activity of miR-146a was also observed in pancreatic cancers, an aggressive cancer with especially poor prognosis. Dietary compounds, including the soy isoflavone genistein and 3,3′-diindolylmethane (DIM), have been shown to enhance the antitumor activity of chemotherapeutic agents by inhibition of NF-κB signaling. Mechanistically, it was found that treatment with either B-DIM (BioResponse formulated DIM with greater bioavailability) or G2535 (a mixture of genistein and other isoflavones) could inhibit invasion of pancreatic cancer cells through the induction of miR-146a expression and with corresponding downregulation of its targets IRAK-1/NF-κB, EGFR, and metastasis-associated protein 2 (MTA-2) [39].

In addition to PTC, GBM and pancreatic cancers, there is some evidence that lower miR-146a expression is associated with the progression and poor prognosis of gastric cancers. In gastric cancer cells, the reduction of miR-146a expression was associated with up-regulation of both EGFR and IRAK1, two genes that play critical roles in tumor development and cancer progression [40].

Because of the amassing evidence that miR-146a/b regulate the process of invasion and metastasis, they may become useful prognostic markers and/or targets for anti-metastatic therapy in the future.

The rs2910164 miR-146a Polymorphism and Cancer Risk

Much research effort has been directed toward understanding the role of single nucleotide polymorphisms (SNPs) present in precursor and mature miRNAs and their influences on susceptibility and progression of various diseases. A common miR-146a polymorphism, rs2910164, leads to a change from a G:U pair to a C:U mismatch in the stem region, and was found to alter miRNA expression and lead to altered regulation of target mRNAs. The rs2910164 polymorphism correlates with risk of developing several different cancers. The first study to report that this particular SNP was associated with cancer risk was from Jazdzewski et al., who found that the C allele nearly halved the levels of precursor and mature miR-146a and caused impaired negative regulation of miR-146a target genes, including TRAF6 and IRAK1. The G allele was found to be more effectively processed within the nucleus, and thus to produce higher levels of mature miR-146a. Individuals heterozygous for this SNP were found to have a predisposition to PTC, while the GG and CC homozygous genotypes proved to be protective against PTC. This same study noted that approximately 5% of PTCs had sustained somatic mutations at the rs2910164 locus [41]. A later study by the same group elaborated these findings and demonstrated that in addition to the miR-146a produced from the leading strand, heterozygous individuals produced two distinct miRNA species from the passenger strand, termed miR-146a*G and miR-146a*C. These passenger-strand derived miRNAs were overexpressed in tumor tissue when compared to paired normal thyroid tissue and led to an exaggerated DNA damage response and up-regulation of anti-apoptotic NFκB pathway genes [42]. Since its association with PTC, the rs2910164 SNP has been associated with altered risks and prognosis for a number of other cancers, including hepatocellular carcinoma [43, 44], gastric cancer [45, 46], squamous cell esophageal cancer [47], prostate cancer [48], cervical cancer [49, 50], non-small cell lung cancer [51], renal cell carcinoma [52], and glioblastoma multiforme [53].

The rs2910164 miR-146a Polymorphism and Familial Breast Cancer

With respect to breast cancer, the role of the rs2910164 polymorphism in modulating risk is less clear. Two studies that focused specifically on familial breast cancers have noted that individuals with documented family history of breast cancer who carried at least one variant (C) allele were significantly more likely to develop breast cancer at a younger age, even after accounting for BRCA1/2 mutation status [54, 55]. These results were not replicated in a large study of 844 German and 760 Italian familial breast cancer patients who tested negative for disease-causing or unclassified variants of BRCA1 and BRCA2 [56]. Of interest, both the BRCA1 and BRCA2 mRNAs are predicted targets for miR-146, and studies have documented the ability of miR-146a and miR-146b to silence BRCA1 [54, 57]. A study by Shen and colleagues noted that the expression levels of mature miR-146a were 60% higher from the C allele than from the G allele, which gave the variant C allele an enhanced ability to silence a luciferase mRNA tagged with the BRCA1 3′UTR, offering a potential mechanism of how the rs2910164 G>C polymorphism may promote breast cancer risk [54]. Furthermore, Pastrello, et al. studied the role for genetic variations in three miRNAs, miR-146a, miR-17 and miR-369, as major determinants in cancer predisposition of familial BRCA1/BRCA2-negative breast/ovarian cancer patients. The results showed that no allelic variant was detected for hsa-mir-17 and hsa-mir-369 and allelic and genotype frequencies for the miR-146a rs2910164 SNP were comparable with that of the controls from the same population, ruling out a role for genetic variations in these three miRNAs as major determinants in cancer predisposition of BRCA1/BRCA2-negative patients. In the same study, the age of diagnosis was considered, and it emerged that patients with at least one miR-146a C allele (CC or GC) developed cancer at younger age compared with those of the GG group [55]. To address the possibility that the rs2910164 polymorphism was a modifier of breast cancer risk only in BRCA1 or BRCA2 mutation carriers, Garcia et al. examined 1,166 and 560 BRCA1- and BRCA2-mutation associated breast cancers, respectively, but failed to note any association of rs2910164 allelotype with penetrance or age of onset [58]. A recent meta-analysis of four studies, encompassing 3,007 breast cancer cases and 3,718 controls, also suggests that the rs2910164 polymorphism does not play a significant role in breast cancer risk, at least when breast cancer patients are examined as a homogenous group [59].

MiR-146a/b in Mammary Gland Development

Although many studies report the role of miR-146a/b in cancer and development of other tissues, their role in mammary gland development is still unknown. Bockmeyer et al., studied the molecular characterization of normal luminal and basal mammary epithelial cells, in which the specific subpopulations were isolated by laser-assisted microdissection from immunohistochemically stained tissue sections of healthy human breast specimens. The results revealed cell type-specific miRNA expression profiles [60]. Interestingly, miR-146b was one of the eight up-regulated miRNAs in basal mammary epithelial cells [60]. Another study, conducted in our laboratory, used unique committed and multipotent progenitor clones derived from the mouse mammary epithelial cell line Comma-D, including alveolar progenitor, ductal progenitors, and multipotent progenitor clones [61]. MiRNA screening of the specific mammary progenitor clones revealed significantly higher expression of miR-146b in the alveolar progenitors versus the ductal and multipotent progenitors (unpublished data). Interestingly, hormonal stimulation with prolactin alone or with a combination of estrogen and progesterone significantly up-regulated miR-146b in mammary epithelial cells derived from virgin mouse mammary glands (unpublished data). These findings suggest that miR-146b may regulate normal mammary development and play a role in specification or maintenance of the alveolar lineage. Aside from these reports, very little is known about miR-146a and miR-146b in mammary gland development.

MiR146a/b in Breast Cancer

The Bockmeyer et. al study not only focused on the miRNA profiles of healthy basal and luminal epithelial cells, but also examined the expression of basal- and luminal-specific miRNAs in different breast cancer subtypes. Interestingly, miR-146b, which is considered to be a basal-specific microRNA, was expressed at higher levels in basal-like breast cancers compared to luminal A and B subtypes [60].

Garcia, et al. [57] studied the regulation of BRCA1 by the miR-146 family, and found that miR146a and miR-146b expression was higher in basal-like breast cancer cell lines, a subtype which commonly exhibits low or absent BRCA1 expression [62, 63]. They also showed that miR-146a and miR-146b precursor transfection increased cell proliferation in HeLa and in MDA-MB-468 cells by targeting the 3′UTR of BRCA1 and down-regulating its expression. BRCA1 is already known to inhibit cellular proliferation when overexpressed in different cell types [57]. Furthermore, both miR-146a and miR-146b were upregulated in triple negative versus ER+/PR+, and in the Scarff-Bloom-Richardson (SBR) grade III versus grade II breast tumors [57]. The SBR grading system is known to be an important prognostic factor in breast cancer. Based on this work, miR-146a/b may have a direct role in breast cancer progression.

miR-146 in Breast Cancer Metastasis

The role of NF-κB in the proliferation, survival and metastatic progression of cancer cells has been appreciated for some time [64, 65]. Bhaumik et al. demonstrated that overexpression of miR-146a/b in the highly metastatic human breast cancer cell line MDA-MB-231 caused suppression of constitutive NF-κB activity [66]. As seen previously, miR-146a/b significantly down-regulated IRAK1 and TRAF6, two key adaptor/scaffold proteins in the IL-1 and TLR signaling pathway known to positively regulate NF-κB activity [66]. Furthermore, miR-146a/b-expressing cells displayed only 45% and 38% of the control migration capacity, respectively. Interestingly, although compromised in their motility and invasiveness, miR-146a/b-overexpressing cells exhibited no proliferation impairment or apoptosis sensitivity relative to control cells [66]. Another study by Hurst et al. [67] revealed that breast cancer metastasis suppressor 1 (BRMS1), a nuclear protein that causes suppression of cancer metastasis, could exert its anti-metastatic action by differentially regulating expression of miR-146a and miR-146b. In addition, both miR-146a/b reduced EGFR expression and suppressed metastasis [67]. These studies add miR-146a and miR-146b to the growing list of miRNA genes that regulate metastasis and provide a potential mechanism by which the expression profile of miRNA is altered in metastatic cells.

Concluding Remarks

As pleiotropic regulators of gene expression, dysregulation of miRNAs can lead to dramatic changes in cellular physiology. Not surprisingly, malignancy-associated perturbations in many miRNAs appear to be as common and functionally important as changes in protein-coding genes. An amassing body of data suggests that alterations in the structure or activity of miR-146a/b may be contributory to the pathogenesis of many human cancers, including breast cancer. Though miR-146a/b are known to regulate important aspects of tumorigenesis, including inflammatory signaling and acquisition of invasive potential, the role(s) of these miRNAs in both normal mammary gland development and in breast cancer pathogenesis remain uncharacterized. Mechanistic studies of these miRNAs in specification and maintenance of the epithelial lineages of the mammary gland are certain to illuminate novel epigenetic programs that underlie both normal breast development and tumorigenesis.

Abbreviations

MiRNAs

MicroRNAs

ncRNAs

Non-coding RNAs

AGO2

Argonaute 2

XPO5

Exportin 5

DBR1

Lariat debranching enzyme

TRBP

TAR RNA binding protein

RISC

RNA-induced silencing complex

LPS

Lipopolysaccharide

NFκB

Nuclear factor kappa-light-chain-enhancer of activated B cells

IRAK1

Interleukin (IL)-1 receptor associated kinase

TRAF6

TNF receptor-associated factor 6

AP-1

Activating protein-1

TLR

Toll-like receptors

MDS

Myelodysplastic syndrome

HSPCs

Hematopoietic stem/progenitor cells

HBEC

Human bronchial epithelial cells

SNP

Single-nucleotide polymorphism

ER

Estrogen receptor α

PR

Progesterone receptor

BRCA1/2

Breast cancer type ½ susceptibility protein

TGFβ

Transforming growth factor beta

STAT3

Signal transducer and activator of transcription 3

VSMC

Vascular smooth muscle cells

KLF4

Krüppel-like factor 4

EMT

Epithelial-mesenchymal transition

PTC

Papillary thyroid cancers

BRMS1

Breast cancer metastasis suppressor 1

BRAF

Serine/threonine-protein kinase B-Raf

GBM

Glioblastoma multiforme

ECM

Extracellular matrix

EGFR

Epidermal growth factor receptor

MMP16

Matrix metalloproteinase 16

MTA-2

Metastasis-associated protein 2