MiR-146a/b: a family with shared seeds and different roots

Abstract

MicroRNAs are small, noncoding, RNAs known for their powerful modulation of molecular processes, making them a major focus for studying pathological mechanisms. The human miR-146 family of microRNAs consists of two member genes, MIR146A and MIR146B. These two microRNAs are located on different chromosomes and exhibit differential regulation in many cases. However, they are nearly identical in sequence, sharing a seed region, and are thus predicted to target the same set of genes. A large proportion of the microRNA (miR)-146 literature focuses on its role in regulating the innate immune response in the context of various pathologies by modulating two widely studied target genes in the toll-like receptor signaling cascade. A growing subset of the literature reports a role of miR-146 in cardiovascular and renal disease, and data suggest there is exciting potential for miR-146 as a diagnostic and therapeutic target. Nevertheless, the published literature is confounded by unclear and imprecise language concerning the specific effects of the two miR-146 family members. The present review will compare the genomic origin and regulation of miR-146a and miR-146b, discuss some approaches to overcome analytical and experimental challenges, and summarize findings in major areas of miR-146 research. Moving forward, careful evaluation of miR-146a/b specificity in analytical and experimental approaches will aid researchers in elucidating the functional relevance of differential regulation of the miR-146 family members in health and disease.

microRNAs (miRNAs) are a class of small, endogenous, noncoding RNAs known to regulate the expression of protein-coding genes. They are found throughout the genome and are transcribed much like protein-coding sequences (34, 49). miRNA biogenesis (46), target recognition (5), and function (13), reviewed extensively elsewhere, are described briefly here. Posttranscriptional processing of the primary miRNA transcript results in formation of a bioactive mature miRNA ~18–22 nucleotides in length (46). The mature miRNA combines with an RNA-induced silencing complex to exert its biological function via binding to the 3′-untranslated region (3′-UTR) of the mRNA transcript and suppressing translation or promoting the degradation of the message. They bind their targets via seed region complementarity, corresponding to nucleotides 2–8 in the mature miRNA sequence. Many human miRNAs and their target sequences share sequence homology across species, making them good candidates for study in animal models (27). Previous studies have shown that many miRNAs can target a single mRNA transcript, and a single miRNA can target many mRNA transcripts (32, 41, 62, 72). miRNAs are thus intriguing molecules to study for their ability to be master regulators of pathophysiology (13, 58) and their attractive potential as therapeutic targets (69).

The publicly available database of confirmed and suspected miRNAs (29) now contains information on nearly 36,000 mature miRNA products from 223 species; including mature sequences for 2,588 human (hsa-), 1,915 mouse (mmu-), and 765 rat (rno-) miRNA. Many miRNAs have been grouped into families based on shared seed sequences, compounding the difficulty of teasing apart biological effects of a specific miRNA, one such family being miR-146. First described in humans by Taganov et al. in 2006 (85), the miR-146 family consists of two miRNAs with nearly identical sequences, miR-146a-5p and miR-146b-5p. The present review will compare the genomic organization and regulation of the two miR-146 family members, discuss several obstacles and caveats in studying miR-146a and miR-146b, and summarize findings in major areas of miR-146 research.

The miR-146 Family

Genomic location and organization.

Mir146 was first identified in mouse cardiac tissue in a study published by Lagos-Quintana et al. in 2002 (51). Three years later, Cai et al. (12) prepared cDNA libraries from size-selected small RNAs (18–24 nt in length) from the BC-1 human B cell lymphoma cell line and classified their sequences based on analysis from the GenBank database (National Institutes of Health) and miRNA registry (Sanger). They found 34 cellular miRNAs, “several of which [had] been reported only in rodents,” including miR-146. The authors note that per the Sanger Institute RNA family database, Rfam, MIR146A was a predicted human miRNA based on analysis of rodent or zebrafish miRNA clones. Taganov et al. (85) are credited with confirming the presence of a human homolog by providing the first characterization of the genomic location and regulation of the human miR-146 family in 2006.

The primary transcripts (pri-miR) are transcribed in the nucleus from two genes, MIR146A and MIR146B. In the human, MIR146A is found within a larger long noncoding RNA host gene, MIR3142HG (chromosome 5q33.3), while MIR146B is found in an intergenic region of human chromosome 10 (10q24.32) (Fig. 1A). In the mouse, Mir146a is found in an intergenic region of chromosome 11 (band B1.1) and miR-146b is found within the first intron of the protein coding gene called major facilitator superfamily domain containing 13a (Mfsd13a or Tmem180) on chromosome 19 (band C3). In the rat, both Mir146a and Mir146b are found in intergenic regions of chromosomes 10q21 and 1q54, respectively. The precursor miRNA (pre-miR) sequences for these two miRNAs are nearly identical on the 5′-end and largely different on the 3′-end, regardless of species. Sequencing data compiled on the miRBase database provides evidence for the mature -5p strand to be the bioactive “guide strand” and the mature -3p strand to be the “passenger strand” for both miR-146a and miR-146b (Fig. 1B). Thus, for the remainder of this review the mention of miR-146a and miR-146b will refer to the -5p strands.

Fig. 1.

The microRNA (miR)-146 family consists of 2 members, miR-146a and miR-146b, found on human chromosomes 5 and 10, respectively (A). Basic exon structure for the human genes is shown, with black boxes and hairpin structures indicating the location of the precursor miR-146 sequences. B: sequences for the mature -5p and -3p strands are shown for both miR-146a and miR-146b in human, mouse, and rat. There is high sequence homology across species for miR-146a-5p and miR-146b-5p, respectively. Note that miR-146a-5p and miR-146b-5p differ by only a few bases on the 3′-end in each species (nucleotides differing between miR-146a-5p and miR-146b-5p are underlined). While the -3p sequences for both miR-146a and miR-146b, respectively, are largely homologous across species, within each species the miR-146a-3p and miR-146b-3p sequences are nonhomologous. C: a schematic of predicted and known transcription factor (TF) binding sites in the promoter regions of both MIR146A and MIR146B.

The mature sequences for miR-146a and miR-146b are highly conserved across species (Fig. 1B). Moreover, the two miRNAs differ only by two nucleotides on the 3′-end of the mature strand, not within the seed region. Research interest into the miR-146 family is growing, though published literature on miR-146a has outpaced that of miR-146b by nearly fivefold. It is common to find miRNA family members such as these presented as “miR-146a/b,” for example, because little was known about their function and regulation when they were first annotated. With advancements in miRNA research, i.e., the advent of miRNA deep sequencing (miR-seq), numerous miRNA profiling studies report differential expression of miRs in various pathologies. To that end, many publications report differential expression of the two miR-146 isoforms yet discuss the miR-146a/b family en bloc with no discrimination between isoform-specific effects. While they share homology and a miRNA family number, several characteristics highlighted in the reported literature suggest that they may have unique regulatory functions. The available evidence suggests that these two miRNAs are not transcribed in tandem; nor are they transcriptionally regulated by all the same factors and cofactors; nor is their expression temporally synchronized; nor do they have the same tissue-specific expression profiles. This strongly suggests that transcription and/or processing of miR-146a and miR-146b is intentionally specific and that they may even have unique regulatory functions determined by sequence characteristics outside of the seed region. The subsequent sections will summarize some of what is known of the miR-146 family; highlight the need for careful, precise discussion of the independence of these two miRNAs; and discuss limitations that should be considered when selecting approaches to discriminate the roles of miR-146a and miR-146b in pathology.

Transcriptional regulation.

Unlike many other miRNA families, little detail of the transcriptional regulation of MIR146A and MIR146B has been published; however, those that have been identified are summarized in Fig. 1C. In Taganov’s original paper, both MIR146A and MIR146B were reported to be regulated by the transcription factor nuclear factor κ-light-chain-enhancer of activated B cells (NF-κB). They describe putative binding sites for the transcription factors interferon regulatory factor 3/7 (IRF3/7) and CCAAT-enhancer-binding-protein-β (C/EBPβ) upstream of the first exon of MIR146A. They also reported an NF-κB2 binding site ~34 kb upstream of the primary MIR146B transcript. Two years later Chang et al. (15) confirmed a human MIR146A transcription start site (TSS) and identified an additional transcription factor binding site for v-myc avian myelocytomatosis viral oncogene homolog (c-Myc). In 2011, Chien et al. (17) identified a TSS 813 bp upstream of the pre-miR-146b sequence. In 2014, Li et al. (53) reported two direct binding sites for C/EBPβ identified in the miR-146b promoter region ~2 kb upstream of the human precursor miR-146b sequence.

Caveats, Barriers, and Caution against Ambiguity

Target prediction.

The classic view of miRNA activity describes their biological function as inducing transcript degradation or causing translational repression after binding the 3′-UTR of the targeted mRNA transcript. As noted above, these processes are reviewed in depth elsewhere. We also now understand a variety of possible miRNA roles, including that of signaling molecule (6, 33) or biomarker (30) and noncanonical regulatory functions such as posttranscriptional upregulation of target mRNAs (89). Regardless of downstream effect, the miRNAs identify and bind their targets via complementary base pairing of the miRNA seed sequence with the mRNA transcript. Considering miR-146a and miR-146b have identical seed sequences, our ability to identify predicted miRNA target pairs based on sequence alone is limited. Many target prediction sites base prediction algorithms on seed sequence complementarity. Given the identical seed sequences of miR-146a and miR-146b, this poses an obvious problem. How does one differentiate activity of one miRNA over the other? In this section, we will discuss some of the technical limitations of different research approaches and various issues that should be considered when studying these two miRNAs with nearly identical sequences.

Predicted targets for miR-146a and miR-146b largely overlap based on 3′-UTR binding alone due to their identical seed sequences. Many times, these prediction sites have completely overlapping predicted targets. TargetScan v7.1 (1), a widely used target prediction site, identifies 3,887 human mRNA transcripts targeted by “miR-146-5p” totaling over 5,000 specific target sites, regardless of user input (i.e., “146a” or “146b”). Various prediction software suites will report a specific binding site for one or the other miRNA, but upon closer inspection both could potentially bind the target of interest due to seed sequence homology. There is a possibility that additional nucleotide binding sites that differ on the 3′-end of the mature miRNA sequences could preferentially stabilize the miRNA-mRNA complex and increase the probability that one miR-146 isoform is more effective at regulating a given target; however, this has not been described to date.

The gold standard for determining target binding is the 3′-UTR luciferase reporter assay, a cell-based assay that uses a reporter construct, generally transfected into a stable cell line, to investigate whether the mature miRNA sequence binds and regulates a predicted target sequence (82). For a variety of reasons, an investigator may choose to test only one of the miR-146 family members without making a clear distinction between the two miRNAs. The lack of this information limits our ability to understand the impact of differential pathophysiological regulation of shared predicted targets.

Investigators may rely on RNA sequencing data sets (including ncRNA data) to analyze gene expression profiles. The art is in the interpretation of these “big data.” One logical, rigorous, and highly accessible approach for miRNA research begins with comparing miRNA deep sequencing data with RNA sequencing data to identify potential miRNA-mRNA target pairs relevant to the disease model studied. miRNA activity can induce transcript degradation; therefore, target pairs showing reciprocal expression (i.e., increased miRNA abundance and decreased target mRNA abundance) can be passed through pathway analysis software to help identify potential mechanistic pathways on which to devote additional research efforts. miRNAs are also well known to repress translation without the transcript being degraded. For this reason, miRNA activity may correlate better with protein abundance than with message abundance, allowing researchers to detect target suppression induced by either mRNA degradation or translational repression. Recent efforts have attempted to analyze miRNA-sequencing data alongside large proteomics data sets (87). Given the high degree of difficulty associated with producing high-quality proteomics data, this approach will likely gain traction as quantitative techniques evolve (23, 52, 67). Careful and thorough investigation is required to support the selection of a miRNA target-of-interest, compiling all we know from published literature about potential target function in different tissues or organisms and pairing that with rigorous bench-top validation of meaningful miRNA-mRNA target interaction.

The method by which the presence and abundance of miRNA are detected or modulated is a key consideration as we progress in the field of miRNA research. Small RNA sequencing is a robust method that can provide single base resolution to identify differential expression of miRNA from the same family. However, the relatively high cost of such analysis and the burden of “big data” placed on an investigator may deter some from using this approach. Many investigators opt to incorporate smaller scale approaches to miRNA expression profiles, such as quantitative (q)PCR or qPCR-based microarrays; however, one must be careful to confirm the specificity of such techniques when studying miRNA families with nearly identical sequences. Specific information about the design and sequence of commercially available primers is generally unavailable to investigators, and specificity may not be guaranteed. Additionally, clear and accurate descriptions of research methods are important for proper understanding of the exact miRNA reported or studied.

Selective perturbation of miR-146a and/or miR-146b.

Recent published reports highlight various attempts to knock down and overexpress these miRNAs. Selective perturbation of the miR-146 family members helps shed light on the specific mechanisms of each miRNA. Current technology allows for knockdown of a mature miRNA via RNA silencing approaches and/or overexpression via delivery of exogenous precursor or mature miRNA. These provide the investigator tools to test the functional effects of modulating the abundance of a given miRNA in vitro or in vivo.

Anti-sense technologies utilize synthetic oligonucleotide sequences with complementarity to the mature miRNA with special chemical modifications that increases overall stability of the knock-down effect, and this can be successfully administered in vivo to knock down a specific miRNA (22, 25, 26, 47, 48). Because these anti-miRNAs work through sequence complementarity, it is important to test the specificity of the targeting oligonucleotides for a miRNA-of-interest as well as those of other highly similar miRNAs before specificity can be confirmed.

Precursor or mimic technologies function to increase the abundance of a mature sequence, either by providing an exogenous source of precursor miRNA for subsequent posttranscriptional processing to form mature miRNA or by supplying an exogenous mimic of the endogenous mature miRNA-of-interest. These technologies are a more robust way to ensure specific increase in abundance of the miRNA-of-interest. miR-146a and miR-146b have distinct primary and precursor sequences, and thus supplying an exogenous precursor containing a specific mature -5p sequence ensures a greater level of confidence in the specificity of the approach. These methods have been largely successful in upregulating target-miR expression in vitro, but genomic editing/transgenesis remains the dominant approach for studying miRNA upregulation in vivo.

Publication ambiguity.

For over a decade, miR-146a and miR-146b, as with many other miRNA families, have been reported together; sometimes presuming no difference in activity (often reported as “miR-146a/b”) and regularly making no distinction between the two (i.e., “miR-146”). Moreover, a complete description of the methods used is key to determining whether a distinction is possible using the selected approach. If there is any uncertainty, the author should make a statement identifying the factors limiting interpretation of the work. For example, one report highlights the finding that patients with coronary artery disease had increased abundance of both miR-146a and miR-146b. While in vitro treatment of cells with miR-146a and miR-146b mimics and inhibitors showed expected reciprocal expression of target gene expression, the in vivo data did not reveal this expected pattern, rather, showing a positive correlation between miR abundance and target expression (86). The authors of the report rightly address this discrepancy in the discussion section and suggest an explanation. Confusion ensues when a paper reports hsa-miR-146a to be upregulated in lung cancer (97) and is later cited as support that hsa-miR-146b “is known to be downregulated in lung cancer” (3). As will be discussed in the next section, a growing body of literature suggests that these two miRNAs have unique and important roles in pathophysiology. Therefore, it is vital that a clear distinction is made between miR-146a and miR-146b when data are reported. Moreover, a complete description of the methods used is key to determining whether a distinction can even be made with any accuracy using the selected approach. If there is any uncertainty, the author should make a statement identifying the factors limiting interpretation of the work. Publication of ambiguous or unclear results adds to the difficulty of determining miR-146a or miR-146b specific activity and limits understanding of their regulatory actions.

miR-146 Family in Pathophysiology

With the increasing accessibility to miRNA profiling data sets, there are a growing number of studies that report differential expression or correlation of the miR-146 family members with pathological conditions (Table 1). The studies that focus on these miRNAs fall mainly into two broad categories: cancer and inflammation. Many studies identifying miRNA-associated pathologies report expression profiles for a panel of miRNAs. As noted above, it is important to be aware of the method by which miRNA family members are reported, differentiated, and quantified; as these distinctions are vital to the proper interpretation of each study.

Table 1.

Summary of the findings of miR-146a and miR-146b activity in associated pathologies

Associated Pathology or Process		miRNA	Target	Tissue/Cell Type	Reference
Cancer, inflammation, and innate immune response	expression profile of regulatory T cells compared with naïve T cells	146 ↑	N/A	regulatory T cells	Cobb (19)
	inhibition of NF-κB signaling	146a/b ↑	IRAK1/TRAF6 ↓	leukemic monocyte cell line	Taganov (85)
	Treg treatment of graft-vs.-host disease	146b ↓	TRAF6 ↑	regulatory T cells	Lu (64)
	breast cancer malignancy, increased patient survival	146b ↑	IRAK1/TRAF6 ↓	breast cancer cell	Xiang (93)
	5q- syndrome	146a ↓	TRAF6 ↑	hematopoietic stem and progenitor cells	Starczynowski (83)
	FOXP3-induced cellular apoptosis	146a ↑	IRAK1/TRAF6 ↓	breast cancer cell	Liu (60)
	PDGF-stimulated tumor growth	146b ↑	EGFR ↓	glioblastoma cells	Shao (80)
Cardiovascular disease	coronary artery disease	146a/b ↑	IRAK1/TRAF6 ↓ (in vitro)	peripheral blood mononuclear cells	Takahashi (86)
	atherosclerotic plaque formation	146a/b ↑	IGSF1, SORT1, and NOVA1 ↓	patient arteries	Raitoharju (75)
	angiopoietin-1 mediates suppression of LPS-induced inflammatory response	146b ↑	IRAK1/TRAF6 ↓	human umbilical vein endothelial cells	Echavarria (24)
	cardioprotection in microbial sepsis	146a ↑	IRAK1/TRAF6 ↓	cardiac tissue; cardiac myocytes and monocytic cells	Gao (28)
	risk for coronary artery disease positively correlated with presence of SNP	146a ↓	N/A	N/A; SNP typing	Jazdzewski (39), Xiong (95)
	left-ventricular remodeling after myocardial infarction	146a ↑	N/A	circulating plasma	Liu (61)
	cyanotic congenital heart disease and chronic hypoxia	146b ↑	RNase L ↓	myocardial tissue, rat cardiomyocytes	Li (54)
	attenuation of proinflammatory stress in hyperlipidemia	146a ↑	IRAK1/TRAF6 ↓	monocytes and macrophages	Li (55)
Acute and chronic kidney disease	CKD patients on hemodialysis	146b ↑ and 146a ↓	N/A	peripheral blood mononuclear cells	Zawada (99)
	acute kidney injury and kidney fibrosis	146b ↑	N/A	renal cortex tissue	Pellegrini (73)
	unilateral ureteral obstruction model of kidney fibrosis	146b ↑	Smad4 ↓	renal cortex tissue	Morishita (70)
	TGF-β1 induced renal interstitial fibrosis	146b ↓	N/A	renal epithelial cells	Kriegel (42)
	cisplatin-induced acute kidney injury	146b ↑	ErbB4 ↓	renal tissue and renal tubular epithelial cells	Zhu (101)

miR, miRNA: microRNA. N/A, not available.

Cancer, inflammation, and the innate immune response.

The involvement of miR-146a and miR-146b in inflammation is the most advanced field of miR-146 research. Beginning with Taganov’s seminal paper in 2006, which spearheaded interest in miR-146 research, the role of this miRNA family in toll-like receptor (TLR) signaling and the innate immune response has been widely studied (16, 21, 66, 74, 78, 81). The same year, Cobb et al. (19) reported the important role of “miR-146” in the development of mature regulatory T cells (Tregs), including miR-146 in a characteristic subset of miRNAs uniquely expressed in Tregs distinct from naïve T cells. By 2008, Sheedy et al. (81) published a review that made special note of the growing knowledge of the activity of this miRNA family in the innate immune response, a common pathway suggested by early miR-146 publications (i.e., TLR-4 signaling). A large proportion of the reported studies identify interleukin-1 receptor-associated kinase 1 (IRAK1) and tumor necrosis factor receptor-associated factor 6 (TRAF6) as the key targets of the miR-146 family, as summarized in Fig. 2 (9, 36, 93). IRAK1 and TRAF6 have long been known to play a vital role in proper TLR signaling, one mechanism by which the innate immune system senses a pathogenic stimulus. Bacterial lipopolysaccharide (LPS, also: endotoxin) stimulates TLR-4 and sets off a signaling cascade acting through IRAK1/TRAF6 culminating in the activation of NF-κB and AP-1 transcription factors and expression of immune response genes (2). While several classes of negative feedback controllers of TLR-4 signaling had been previously reported, Taganov et al. (85) were the first to suggest that the miR-146 family acted as such. THP-1 cells (derived from an acute monocytic leukemia patient) were stimulated with LPS, and the expression profile of miRNAs was analyzed by microarray. The investigators found that both MIR146A and MIR146B are endotoxin-responsive genes, and promotor analysis identified MIR146A upregulation as NF-κB dependent. After validating IRAK1 and TRAF6 as targets of miR-146a and miR-146b with luciferase reporter constructs, they proposed that the upregulation of these miRNAs serves as negative feedback regulation of TLR-4 signaling (85). At the same time confirming miR-146a/b are endotoxin responsive, Kutty et al. (50) also reported preferential upregulation of miR-146a or miR-146b depends on the inflammatory stimulus. In vitro treatment of human retinal epithelial cells with various cytokines revealed maximal miR-146a upregulation is dependent on interleukin-1β and miR-146b dependent on interferon-γ. These data highlight the unique role the individual miR-146 family members may play in mediating the immune response.

Fig. 2.

A large majority of miR-146 literature focuses on the innate immune and inflammatory responses mediated by NF-κB signaling. miR-146 has been shown to act as a negative feedback regulator of NF-κB signaling by suppressing the translation of 2 targets: interleukin-1 receptor-associated kinase 1 (IRAK1) and tumor necrosis factor receptor-associated factor 6 (TRAF6). Stimulation of the toll-like receptor complex (TLR) leads to the association of IRAK1 and TRAF6 with the TAK/TAB complex, resulting in the subsequent activation of the IKK complex and translocation of NF-κB to the nucleus where it induces expression of inflammatory/immune response genes and induces expression of miR-146. miR-146 then may act as a molecular brake on NF-κB signaling by suppressing IRAK1 and TRAF6 expression.

Dysregulation of the innate immune response has been shown to be oncogenic, and miR-146a/b have been implicated in this process, as outlined in Fig. 3 (76). In 2008, Bhaumik et al. (8) reported increased expression of miR-146a/b in metastatic human breast cancer cells negatively regulated NF-κB signaling and reduced the cells’ metastatic phenotype, providing the first clear functional connection of miR-146a/b to cancer pathology. In 2010, Starczynowski et al. (83) published a study in Nature Medicine in which they reported the loss of MIR146A in 5q- syndrome (a hematological disorder and subtype of myelodysplastic syndrome in which there is a loss of the long arm of chromosome 5q33.1) accounted for, in part, patients’ increased susceptibility to cancer. The data showed a clear, inverse correlation between a reduction in miR-146a abundance and increased expression of the target gene TRAF6, which leads to inappropriate activation of innate immune signaling in hematopoietic stem and progenitor cells (83). Derepression of TRAF6 signaling in a mouse culture model altered hematopoiesis and bone marrow failure, suggesting the importance of miR-146a in relation to the pathogenesis of malignancy (88). Clinical data support this, showing 5q- syndrome associates with poorer prognoses in patients with acute myeloid leukemia (10, 68). Xiang et al. (93) highlighted the tumor suppressor role of miR-146b in modulating negative feedback regulation of NF-κB signaling, suggesting that upregulation of miR-146b may be a molecular brake on tumorigenesis.

Fig. 3.

Diagram depicting 3 simple pathways of miR-146-5p modulation of cancer phenotypes. The previously discussed NF-κB signaling pathway has been shown to be induce the metastatic phenotype of various breast cancer cells. Increased expression of miR-146a/b acts to block NF-κB signaling and thus reduces this metastatic potential. miR-146b also fine-tunes the cell’s response to growth factor stimulus in a cancer setting. Platelet-derived growth factor increases expression of miR-146b, which subsequently blocks epidermal growth factor receptor (EGFR) signaling thereby halting metastatic growth. When miR-146a is lost due to a chromosomal abnormality (5q syndrome), aberrant expression of tumor necrosis factor receptor-associated factor 6 (TRAF6) induces NF-κB signaling and increases cancer susceptibility.

Aberrant signaling by platelet-derived growth factor (PDGF) is another important oncogenic mechanism through which the miR-146 family is involved. In 2011, Shao et al. (80) identified a negative feedback regulatory role for miR-146b in human glioblastoma cells. PDGF-BB (a known ligand of PDGF) induced the expression of miR-146b via a c-fos-dependent mechanism. Increased miR-146b inversely correlated with abundance of the epidermal growth factor receptor, fine-tuning the cells’ response to growth factor stimulus in a cancer setting (80), supporting a previous study that suggested therapeutic potential for miR-146a or miR-146b in controlling breast cancer metastasis (38). Liu et al. (60) reported interesting data suggesting a specific role for miR-146a and not miR-146b in FOXP3-induced apoptosis of breast cancer cells. The authors showed miR-146a upregulated by FOXP3+ T-regulatory cells inhibited NF-κB signaling through suppression of IRAK1 and TRAF6, leading to increased cellular apoptosis, suggesting a new mechanism for FOXP3-mediated tumor suppression critically supported by the miR-146 family (60).

The role of the miR-146 family in the immune response has focused on functions outside of cancer biology as well (35, 64, 90). In a recent study investigating the role of Tregs in the maintenance of proper immune response balance, Lu et al. (64) proposes knockdown of miR-146b (and not miR-146a) helps to prolong the immune-suppressive effect of Tregs in graft vs. host disease, a disease in which adoptive Treg transfer is effective treatment. The data showed knockdown of miR-146b in thymic-derived Tregs (tTregs) led to increased nuclear localization and activation of NF-κB via increased expression of TRAF6. The anti-miR-146b-treated tTregs showed enhanced prosurvival pathway activation and greater efficacy to suppress inflammatory responses (64). In another example, miR-146a has been implicated in the pathogenesis of Alzheimer’s disease (56, 65). miR-146a was shown to be upregulated in brain regions exhibiting neuroinflammation, supporting the role of miR-146a in modulating the immune response in neuropathology. These studies, and others, highlight the complexity and great potential of miRNA research and point to the multitude of cell-specific or condition-specific effects miRNAs may have in pathology.

Relevance to cardiovascular pathophysiology.

While the innate immune and inflammatory responses account for a clear majority of the miR-146 literature, a substantial subsegment of the reported literature regarding the pathophysiological role of the miR-146 family has focused on cardiovascular inflammatory diseases. An early report from Takahashi et al. (86) examined the abundance of miR-146a/b by real-time PCR in peripheral blood mononuclear cells in patients with coronary artery disease (CAD) compared with non-CAD control patients, finding increased abundance of both miR-146a and miR-146b in CAD patients compared with controls. They suggested the importance of the miR-146 family in modulating TLR4 signaling through IRAK1 and TRAF6 inhibition, thereby mediating the development and progression of atherosclerotic disease. Additional evidence suggesting the miR-146 family is modulated in vascular inflammatory disease was reported in a study examining miRNA profiles in atherosclerotic plaques in arteries from patients enrolled in the Tampere Vascular Study (75). The study analyzed miRNA expression data by microarray (Agilent) after which differentially expressed target miRNAs were verified by qRT-PCR. The report showed miR-146a and miR-146b-5p, among others, to be upregulated in the atherosclerotic arteries compared with control arteries. No further mechanism was examined, though these findings suggest this to be an intriguing area of future study. Apolipoprotein E (ApoE) is known to display anti-inflammatory properties and protect from atherosclerosis and inflammatory disease. Loss of ApoE can enhance NF-κB signaling and lead to proinflammatory stress responses. Li et al. (55) reported delivery of miR-146a mimetic to ApoE knockout mice (ApoE^−/−) attenuated macrophage activation, atherosclerosis, and proinflammatory responses in hyperlipidemia. This study is one, among many, to suggest the great therapeutic potential of new miRNA drug delivery techniques (57, 91, 98).

There is also evidence of a direct link between miR-146a and myocardial infarction (MI). Zidar et al. (102) correlated miR-146a levels in autopsied heart tissue with severity of infarction in patients who died from complications following MI, including ventricular rupture. The authors suggest that miR-146a is upregulated in response to NF-κB activation in the early innate immune response to severe MI. However, the mechanisms leading to, or resulting from, miR-146a modulation were not identified. It was not determined whether its role is protective or damaging to the heart. In 2012, a published report showed increased expression of circulating miR-146a, among other miRNAs, in the serum of patients diagnosed with acute coronary syndrome (ACS) in an emergency setting (71). The authors suggest their data point to a diagnostic role of circulating miR-146a in early assessment of suspected ACS patients. One study by Liu et al. (61) examined the potential to use circulating miR-146a levels as a biomarker to predict left-ventricular remodeling (LVR) after MI. The authors reported abundance of circulating miR-146a in patients with LVR to be significantly higher than those patients without LVR, subsequently showing miR-146a abundance was an independent predictor of LVR development (odds ratio 2.127). Other studies directly link miR-146b with cardiac pathology. Li et al. (54) showed chronic upregulation of miR-146b-5p in the heart tissue of patients with cyanotic congenital heart disease and chronic hypoxia and suggest a protective role for miR-146b in chronic cardiac hypoxia.

A common minor C allele single nucleotide variation (rs2910164, MAF C = 0.2792) found within the passenger strand of pre-miR-146a was reported to affect the processing of the mature miR-146a sequence by altering transcription factor binding efficiency, reducing the abundance of both pre-miR-146a and miR-146a-5p by nearly twofold and affecting the efficiency of translation repression of target genes (39). This common genetic variation at the miR-146a locus within the Chinese Han population correlates with an increased risk for CAD (95). After genotyping 295 CAD patients and 283 controls by restriction fragment length polymorphism PCR, logistic regression analysis revealed heterozygous patients (CG) or patients homozygous for the minor allele (CC) to have significantly increased risk for CAD compared with patients that did not carry the minor allele. A recent meta-analysis of 10 studies examined the correlation of this common SNP with coronary heart disease, adding support for the pathological relevance of miR-146a regulation in cardiovascular disease (94).

Some investigators have focused on the role of miR-146b in regulating cardiovascular responses to experimentally induced inflammatory processes. Echavarria et al. (24) suggest a role for miR-146b-5p in mediating the LPS-induced inflammatory response in culture human endothelial cells. The study reported angiopoietin-1 mediated the upregulation of miR-146b-5p and subsequently inhibited LPS-induced TLR-4 signaling through suppression of IRAK1 and TRAF6. In 2015, Gao et al. (28) suggested miR-146a is cardioprotective, attenuating cardiac dysfunction in microbial sepsis. Mice subjected to a sepsis-inducing cecal puncture procedure were treated with lenti-viral miR-146a, and cardiac function was monitored via echocardiography. Fractional shortening and ejection fraction were significantly greater in miR-146a-treated animals than untransfected or scramble transfected controls. The authors propose miR-146a upregulation inhibited NF-κB-mediated expression of inflammatory cytokines and infiltration of proinflammatory cells into the myocardium. NF-κB has been reported to play a role in pathological cardiac hypertrophy and heart failure, suggesting it may be an important link between miR-146 function and heart failure (31, 37, 92).

Relevance to renal pathophysiology.

There is recent movement toward understanding the genetic and epigenetic dysregulation of genes involved in miRNA regulation of renal physiology in both chronic and acute pathophysiology (7, 40, 43–45, 59, 77, 96). Zawada et al. (99) published a paper in which they examined miRNA-sequencing data from peripheral blood mononuclear cells in chronic kidney disease (CKD) patients on hemodialysis and compared expression profiles to healthy controls. They identified 182 differentially expressed miRNAs, finding that miR-146b was upregulated and miR-146a was downregulated in this patient population. These data suggest and support the differential regulation of miR-146 family members in CKD. The miR-146 family has also been associated with acute kidney injury (AKI) in humans. Pellegrini et al. (73) published data in 2016 showing miR-146b was highly upregulated in the renal cortex in patients with AKI, noting that miR-146b was maximally upregulated at the peak of fibrosis. The upregulation of miR-146b was recapitulated in a unilateral ureteral obstruction model of kidney fibrosis in the rat. The authors highlight the use of small-RNA sequencing technology to survey expression profiles in a temporal manner throughout distinct phases of kidney injury. Additionally, in an in vitro model of renal interstitial fibrosis, Kriegel et al. (42) examined the change in miRNA expression following TGF-β1 (transforming growth factor-β 1) treatment of human renal epithelial cells. This study reported the downregulation of miR-146b-5p by nearly 1.5-fold in the treated cells compared with vehicle-treated controls, confirmed by qRT-PCR analysis. In a mouse model of unilateral ureteral obstruction, treatment with exogenous miR-146a attenuated renal fibrosis by inhibiting the proinflammatory and profibrotic NF-κB and TGF-β1 signaling cascades (70), suggesting a protective role for miR-146a. In contrast, a recent report published by Zhu et al. (101) suggests a deleterious effect of miR-146b-5p upregulation in cisplatin-induced AKI. Knocking down miR-146b-5p in rat renal tubule epithelial cells using miR-146b inhibitors resulted in protection from cisplatin-induced apoptosis in vitro. This study is one case, among many, where there was clear evidence that the knockdown of miR-146b-5p was effective, but there was no mention of off-target effects on miR-146a expression or compensatory upregulation of miR-146a.

Ongoing Considerations and Future Directions

There are many questions yet unanswered in miRNA research, especially with regard to the miR-146 family. The advent of small noncoding RNA sequencing technologies allows unprecedented precision for researchers to investigate expression profiles of many noncoding RNAs, including miRNAs. Moreover, anti-sense and precursor technologies are new tools at our disposal to study the effects of differential miRNA expression in vivo and in vitro. New evidence shows increased support for “noncanonical” miRNA activity, for example, non-3′-UTR target binding or translational activation of an mRNA transcript (11, 18, 89, 100). Genomic editing technologies that allow researchers to pinpoint knockdown or knock-in of a sequence of interest are one potential avenue on which to find these answers. CRISPR-mediated genomic editing is becoming the new standard for stable knockdown of miRNAs in vitro and in vivo (14). Many studies are identifying miRNAs as exciting biomarkers in pathology (4, 30, 84), and new strategies are in development for the effective delivery of miRNA mimics, pre-miRNAs, or anti-sense oligos, with dozens tested in cancer therapies and some new technologies even reaching clinical trials. The potential therapeutic benefit of in vivo miRNA modulation is only beginning to be uncovered (6, 20, 63, 79). Rigorous experimental approaches will aid us in understanding how and why these miRNAs are differentially regulated in health and disease as well as ways in which we might exploit this knowledge to identify potential therapeutic pathways.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

M.R.P. prepared figures; M.R.P. drafted manuscript; M.R.P. and A.J.K. edited and revised manuscript; A.J.K. approved final version of manuscript.