Contributions of microRNAs to Peripheral Insulin Sensitivity

Abstract

An extensive literature base combined with advances in sequencing technologies demonstrate microRNA levels correlate with various metabolic diseases. Mechanistic studies also establish microRNAs regulate central metabolic pathways and thus play vital roles in maintaining organismal energy balance and metabolic homeostasis. This review highlights research progress on the roles and regulation of microRNAs in the peripheral tissues that confer insulin sensitivity. We discuss sequencing technologies used to comprehensively define the target spectrum of microRNAs in metabolic disease that complement studies reporting physiologic roles for microRNAs in the regulation of glucose and lipid metabolism in animal models. We also discuss the emerging roles of exosomal microRNAs as endocrine signals to regulate lipid and carbohydrate metabolism.

microRNAs Exert Broad Regulation of Gene and Protein Expression

Coordinated regulation of food intake and energy expenditure maintains metabolic homeostasis. To this end, the liver and adipose tissues sense complex endocrine signals and respond to nutritional demands. Metabolic disorders such as obesity and insulin resistance reflect the combined impacts of genetic and environmental factors that alter the ways the liver and adipose tissues govern energy balance. Therefore, precise strategies must activate or repress metabolic pathways in the liver and adipose tissues to maintain the energetic needs of connected organ systems.

microRNAs are noncoding RNAs of 20 to 25 nucleotides that bind target mRNAs in the 3′ untranslated region (UTR) to induce mRNA degradation and inhibit protein translation (1, 2). microRNA targeting is defined mainly by the microRNA’s seed region consisting of nucleotides 2 through 8 of the microRNA (3-6), which undergoes complementary base-pairing with specific sequences in target mRNAs. The microRNA “seed” guides the effector proteins, including Argonaute-2 (AGO2) and its interacting partner GW182, which comprise the RNA-induced silencing complex to the 3′UTR of target mRNAs. Ultimately, the physical association of the microRNA 5′ seed sequence and target mRNA 3′UTR causes a net reduction in protein output from a given gene.

Plants and animals express microRNAs, but microRNAs emerged independently in each kingdom, likely from a common ancestor 500 million years ago (7). Although microRNAs are found widely in the animal kingdom, they are most abundant in vertebrates. Of the microRNAs confidently annotated in humans (8), about 300 fall into 1 of 177 distinct microRNA families, and only 27 families show conservation among all bilateral animals.

Target nomination algorithms (eg, TargetScan, PicTar, miRanda, miRWalk) predict that microRNAs target hundreds of mRNA and microRNA target spectrums frequently overlap (6, 9). mRNAs can also be targeted by many distinct microRNAs to alter gene expression collaboratively. Conserved microRNAs regulate ~30% to 80% of human genes (10), arguing that microRNAs strongly influence transcriptional mechanisms and protein translation across fundamental cell functions of growth and metabolism.

A significant body of work suggests that microRNAs contribute to metabolic homeostasis. And, countless studies demonstrated microRNAs convey environmental and nutritional signals to regulate gene expression and facilitate cellular responses to metabolic substrates that include lipids, glucose, and insulin (11-15). microRNAs also impact brain function (16), but a large confluence of studies pinpoint roles for microRNAs in the liver and adipose tissues. For example, broad disruption of microRNA function impacts glucose and lipid metabolism in diverse ways. Ablation of Ago2 in the liver allows protection from the metabolic effects of a high-fat diet (17). Metabolic phenotypes of other interacting partners of AGO2, including microRNA processing enzymes and GW182 proteins, have not been reported because of the limited availability of genetic mouse models. However, knockout of the microRNA biogenesis factor Drosha DGCR8 (18, 19) or the processing enzyme Dicer (20-25) disrupts metabolism and causes energy metabolism disorders. These observations strongly support a rate-limiting role for microRNAs in the functions of the liver and adipose tissues.

microRNA Expression Contributes to Whole-body Energy Homeostasis

Expression of the most abundant microRNAs exhibit broad roles in maintaining metabolic homeostasis. Despite the annotation of > 2000 microRNAs, global analysis of microRNA expression patterns in 334 normal tissue samples revealed that the top 10 microRNAs compose ~75% of all microRNA expression (9, 26). The abundant expression of the dominant microRNAs coupled with the large target spectrum of these microRNAs points to a general role in maintaining metabolic homeostasis. Not surprisingly, dysregulation of specific microRNAs in response to genetic or environmental factors (Table 1) can contribute to aberrant gene expression patterns underlying metabolic dysfunction (7).

Table 1.

Putative in vivo roles for microRNAs involved in peripheral insulin sensitivity

MicroRNA	Tissue	Mechanism of action	Target(s)	Ref
Let-7	Skeletal muscle, liver, pancreas	↓Glucose tolerance, ↓insulin secretion	Igf1r, Insr, Irs2, Hmga2	(31, 32)
miR-26 family	WAT	↓Body weight, ↓blood glucose, ↓adipogenesis	Fbxl19	(43)
miR-26a	Liver	↑Insulin sensitivity, ↓gluconeogenesis, ↓fatty acid synthesis	Acsl3, Acsl4, Gsk3b, Pck1	(39)
miR-29	Liver, skeletal muscle	Tissue-dependent on insulin action	Pgc1a, G6pase	(36, 37)
miR-30	BAT, WAT	↑Insulin sensitivity, ↑adipocyte differentiation, ↓WAT inflammation	Nrip1, Ubc9, Stat1	(26, 45-48)
miR-33ab	Brain, liver, WAT	↓Body weight, ↓food intake, ↑insulin sensitivity	Abca1, Hmga2	(13, 14, 95, 96)
miR-34a	WAT, liver, AT Exos	↑WAT inflammation, ↓M2 macrophage	Klf4, Klb, Nampt	(53-57)
miR-99b	AT Exos	Reduced liver serum FGF21	Fgf21	(20)
miR-103/107	Liver, WAT	↑Insulin resistance	Cav1	(15)
miR-122	Liver	↑Cholesterol, ↑fatty acid synthesis, ↓hepatic inflammation	Gys1, Aldoa, Ccng1, P4ha1, Agpat, Dgat1	(76, 99-101)
miR-128-1	Liver	↓Insulin sensitivity, ↓glucose tolerance, ↓energy expenditure	Ldlr, Abca1, Sirt1, Irs1, Insr	(27, 28)
miR-132	Liver	↑Cholesterol, ↑fatty liver, ↑body weight	Foxo3, Ep300, Sirt1, Pten	(33, 34)
miR-133	BAT	↓BAT thermogenesis	Prdm16	(61, 63, 65)
miR-143	Liver, WAT	↓Insulin sensitivity, ↓glucose tolerance	Orp8	(11)
miR-155	BAT	↓BAT thermogenesis	Cebpb	(67)
miR-223	WAT Mϕ	↓WAT inflammation, ↑M2 macrophage, ↑insulin sensitivity	Nfat5, Rasa1	(92)
miR-196a	WAT	↑WAT thermogenesis, ↓body weight	Hoxc8	(64)
miR-455	WAT	↑BAT thermogenesis	Hif1an, Runx1t1, Ndn	(68)
miR-690	WAT Mϕ, AT Exos	↓WAT inflammation, ↑M2 macrophage, ↑insulin sensitivity	Nadk	(127)
miR-802	Liver	↓Liver insulin sensitivity, ↓glucose tolerance	Hnf1b	(12, 38)

Abbreviations: AT, Exos, adipose tissue-derived exosomal microRNA; BAT, brown adipose tissue; HDL, high-density lipoprotein; WAT, white adipose tissue.

In 2020, an unbiased analysis of genome-wide association study data (27) uncovered miR-128-1 regulated cholesterol metabolism in mice (28). miR-128-1 is an intronic microRNA present in the R3HDM1 gene, located at the center of the positively selected locus on human chromosome 2, ~200 kb from the lactase gene. The discovery of the microRNA helped clarify a previously unexplained mystery whereby the 2q21.3 locus was associated with survival in ancient times and obesity and type 2 diabetes in humans. Chromatin accessibility changes near the locus may have contributed to evolutionary adaptation to famine by promoting energy storage, which now confers susceptibility to metabolic diseases. The elegant findings establish a thrifty phenotype connected to miR-128-1-dependent energy storage may link ancient adaptation to famine and modern metabolic maladaptation associated with nutritional overabundance.

The large and ancient microRNA let-7 family was the first human microRNA discovered (29, 30). The important functions of Let-7 family members during development logically extend to glucose metabolism in the pancreas, liver, and muscle (31, 32). Let-7 and the RNA binding protein Lin28 form a balanced counterregulation axis affecting insulin production and sensitivity. Muscle-specific Lin28 overexpression mice fed a high-fat diet showed improved glucose metabolism and let-7 depletion (31). Conversely, let-7 overexpression in the pancreas or skeletal muscle promotes insulin resistance and impairs glucose tolerance in mice (31). Let-7 inhibits the canonical Akt-mediated insulin signaling cascade by suppressing the expression of the insulin receptor, IGF-1 receptor, and insulin receptor substrate-2 (32). Taken together, let-7 targets the insulin-signaling cascade to regulate glucose homeostasis.

Several other abundantly expressed microRNAs play roles in metabolic disorders associated with an aberrant response to insulin signaling. Targeting of the metabolic regulators Foxo3, Ep300, Sirt1, and Pten by transgenic miR-132 in mice causes fatty liver disease and insulin resistance. Likewise, liver samples from both patients with nonalcoholic fatty liver disease and mouse models of hepatic steatosis or nonalcoholic steatohepatitis displayed dramatic increases in miR-132 and varying decreases in miR-132 targets compared with controls. Preclinical models tested the therapeutic feasibility of antisense targeting of miR-132, which reversed fatty liver disease (33, 34). Other antisense strategies to deplete microRNAs expressed at generally higher levels in fatty liver diseases (35) show metabolic benefits in mouse models.

Genetic approaches also defined the functions of microRNAs in the liver. Leptin-deficient diabetic mice express higher miR-29a-c (miR-29) in the liver, and its predicted targets suggested roles in gluconeogenesis. Along these lines, forced expression of miR-29 in the liver of diabetic mice alleviates hyperglycemia (36). However, the tissue-specific effects of miR-29 complicate its function because targeting Pgc1a in the skeletal muscle reduced insulin action (37). Similarly, genetic and diet-induced insulin resistance caused higher miR-103/miR-107 in the liver of mice. But, miR-103/miR-107 inhibition leads to improved insulin action by enabling caveolin-1 expression to increase insulin receptor availability (15). Similar to miR-103/107, miR-143 (11) and miR-802 (12) exhibit greater expression in the liver of diabetic mice. Transgenic overexpression of miR-143 reduced insulin sensitivity, presumably by targeting an Akt adapter protein (11), oxysterol-binding protein-related protein 8. Other work by Stoffel and Bruning demonstrated miR-802 caused impaired glucose tolerance and attenuated insulin sensitivity in mice (12), which is directly regulated by 2 nuclear receptors, farnesoid X receptor and small heterodimer partner (38). Although not as abundant as miR-103/107 or miR-143, miR-802 targets HNF1b, and the loss of HNF1b results in mature onset diabetes of the young, providing some mechanistic insight into the diabetogenic activities of miR-802.

Huang and colleagues showed miR-26a governs liver insulin sensitivity (39). Global or liver-specific ectopic miR-26a expression in mice fed a high-fat diet improved insulin sensitivity coupled with decreased hepatic glucose production and lipogenesis. Conversely, miR-26a silencing impaired liver insulin sensitivity, and expression profiling showed reduced miR-26a levels in liver RNA from humans with diabetes and mouse models of insulin resistance. Mechanistically, miR-26a slows the complications of obesity and contributes to insulin sensitivity by regulating the expression of crucial metabolic genes, including Gsk3b, Pepck (Pck1), and 2 members of the protein kinase C family of enzymes. Overall, the biology of miR-103/107, miR-143, miR-26a, miR-29, and miR-802 can be dictated by diet stress that alters the expression of these microRNAs in the liver to repress glucose and insulin sensitivity in mice.

Adipose Tissue microRNAs Remodel Energy Balance

Adipose tissue is an endocrine organ that stores energy in lipids and regulates whole-body energy homeostasis via the secretion of metabolites and adipokines (40). Adipose tissue microRNAs influence metabolism by regulating adipocyte differentiation states or directly modulating specific metabolic and endocrine functions. More than 40 microRNAs correlate with human obesity and type 2 diabetes (41), whereas numerous microRNAs affect adipocyte differentiation, including miR-103/107 (15), miR-193b-365 (42), miR-26 (43), miR-378 (44), and miR-30 (26, 45-48). In particular, miR-143 (11) and miR-103/107 (15) suppress white adipocyte differentiation as part of the way these microRNAs promote insulin resistance.

Several microRNAs identified in obese human white adipose tissue (WAT) also function in the microenvironment to control inflammation by directly or indirectly modulating pro-inflammatory cytokine release from adipocytes and macrophages (49, 50). miR-34a regulates inflammatory pathways, and increased transcripts have been observed in serum (51) and subcutaneous adipose tissue from subjects with obesity and type 2 diabetes (52). Whole-body knockouts of miR-34a cause susceptibility to some inflammatory effects of a high-fat diet, including greater fat pad weights and epididymal WAT macrophage infiltration (53). A more recent study (54) confirmed higher miR-34a levels in obese human and mouse visceral WAT depots. This elegant study also showed conditional knockout of miR-34a in WAT caused obesity-induced glucose intolerance and systemic inflammation that derived from secretory defects that preluded necessary, protective skewing of M2 macrophages upon diet challenge (54). The additional pro-inflammatory activities of miR-34a in the liver (55-57) cement a robust role in obesity-related metabolic diseases.

Brown adipose tissue (BAT) functions to catabolize free fatty acids and dissipate energy in the form of heat (58). Beige adipocytes that emerge in subcutaneous WAT depots (59) also exhibit inducible thermogenic capacity upon stimulation, similar to BAT cells. The presence of brown and beige fat is mainly associated with protection from adverse outcomes of obesity (60). Not surprisingly, individual microRNAs contribute to brown and beige fat functions. microRNAs can alter brown adipogenesis (42, 44, 61) or stimulate a white to brown fat transition (48, 62-64). For example, the miR-193b-365 (42) cluster of microRNAs and miR-133 (61, 65) regulate brown adipogenesis by acting on PRDM16, an obligate factor for brown adipocyte identity (66). Another study performed a screen for inhibitors of brown adipocyte differentiation and identified miR-155, which impairs brown and beige adipocyte development by inhibiting C/EBPβ expression (67).

Genetic or pharmacologic methods to increase the expression of microRNAs specific to BAT (miR-196a) or those induced by cold exposure (miR-30b/c) enhance browning of WAT by disrupting the expression of conventional transcriptional activators of white fat genes (48, 64). Tseng’s group used integrated microRNA and mRNA microarray discovery methods to identify the BAT-specific microRNA miR-455 that regulates brown adipogenesis (68). miR-455 activates AMPKα1 by targeting HIF1an, and AMPK then promotes the brown adipogenic program and mitochondrial biogenesis. Concomitantly, miR-455 also targets the adipogenic suppressors Runx1t1 and Necdin, initiating adipogenic differentiation. This finding and similar studies (44) align with the idea that a single microRNA may target multiple mRNAs to oversee fat cell maintenance.

These findings highlight the roles microRNAs play in adipocyte gene expression. In addition, microRNAs might have pathogenic actions associated with impaired brown and white adipocyte differentiation, WAT inflammation, energy balance, and insulin resistance (18, 22, 69). Because microRNAs act as key regulators of WAT and BAT gene expression, they may be leveraged in preclinical models to alter the metabolic programs of adipose tissue depots, thus improving insulin sensitivity and systemic metabolism.

High-throughput Methods Reveal Targets of microRNAs in Metabolic Tissues

Many studies of microRNA biology rely on computational predictions coupled with low-throughput validation biochemical studies to identify microRNA targets. In this way, the predicted microRNA binding region is fused to luciferase, and microRNA targeting can be inferred by reduced luminescence reporter activity. However, comprehensive knowledge of microRNA functions requires the identification of in situ targets. Several bioinformatics techniques have been developed to predict microRNA targets by integration of biochemical properties of microRNAs with large-scale sequencing technology (1, 70, 71).

More directly, methods that combine UV crosslinking and immunoprecipitation (CLIP) of RNA binding proteins with high-throughput sequencing map interactions between RNA binding proteins and their RNA target sites (72). CLIP methods have been adapted to identify microRNA–mRNA interactions in a genome-wide manner through purification of AGO-associated RNAs, which include microRNAs bound to their respective gene targets (73, 74). In particular, immunoprecipitation of the AGO protein-coupled to high-throughput sequencing (AGO-CLIP-seq) defines points of AGO-mRNA interactions, which allows the definition of underlying microRNA seed complements. Global analyses of microRNA–mRNA interactions by AGO-CLIP-seq methods revealed frequent microRNA targeting of genomic regions other than the 3′UTR, which may represent > 30% of all microRNA binding sites. Because microRNA target prediction algorithms rely on binding site conservation across species, microRNA interactions within coding regions or 5′-UTRs are rarely predicted. Therefore, AGO-CLIP-seq protocols remain the only available method to perform comprehensive, unbiased, and specific mapping of microRNA target spectra.

In the liver, miR-122 accounts for most of the microRNA expression (75), and its inhibition in vivo improves lipid and cholesterol profiles in wild-type and mice with diet-induced obesity (76). Along these lines, antisense inhibition of miR-122 blocks the expression of lipogenic and cholesterol synthesis genes in the liver (76). AGO-CLIP studies of miR-122 targets demonstrated widespread and noncanonical binding in the mouse liver (77). But, only 3′-UTR and coding exon binding of canonical miR-122 targets impacted steady-state RNA levels. These CLIP datasets predicted the downregulation of miR-122 derepressed programs critical for liver cancer progression, including proliferative responses. Although miR-122 appears dispensable for liver development (77), the studies reveal an extensive network of microRNA targeting that slows the progression of chronic liver disease.

More recently, the Cohen laboratory (78) conducted the first comprehensive biochemical mapping of microRNA–mRNA binding in WAT and BAT. This study annotated stark phenotypic differences between BAT and epididymal WAT marked by divergent microRNA expression profiles and AGO-CLIP binding events. BAT-specific peaks showed enrichment for thermogenesis and the citric acid cycle that likely ties into BAT’s metabolic functions to burn carbohydrate and lipid fuels. In contrast, WAT-specific peaks define microRNA binding within the PI3K-Akt cascade reflective of receptivity to insulin action. Additional depot-specific events included modules that regulate leptin secretion and lipolysis. These valuable resources nominate previously undefined regulatory relationships in WAT. Further studies used AGO cross-linking techniques to map the microRNA targetome in the brain, pancreas, heart, and other organs (73, 74, 79, 80). These studies expand the scope and mechanisms of metabolic control by microRNAs outside of the conventional single-target nomination strategies.

Transcriptional Control of microRNAs and its Metabolic Impact

Like mRNAs, the majority of microRNA genes are transcribed by RNA polymerase II, generating primary microRNA transcripts (81). In 2005, 2 pioneering papers first identified mammalian transcription factors that precisely control microRNA transcription (82, 83). The numerous studies that followed revealed the complex network of microRNA transcriptional control at a molecular level. Next, we outline key examples from studies that explored transcriptional regulation of microRNAs.

Peroxisome proliferator-activated receptor gamma (PPARγ) belongs to the nuclear receptor superfamily of ligand-regulated transcription factors (84). PPARγ is obligate for adipocyte formation, and knockout in mice (85) or loss of function variants in humans (86-88) lead to fulminant type 2 diabetes phenotypes. Not surprisingly, PPARγ regulates microRNA expression in adipocytes (89, 90). Many highly expressed microRNAs in WAT (91) can be regulated by PPARγ to enhance beige fat adipogenesis (26, 45, 46) and exert anti-inflammatory effects in obese mice (45). In bone marrow-derived macrophages, PPARγ induces miR-223 transcription and promotes anti-inflammatory M2 macrophage responses (92). PPARγ regulation of microRNAs in the liver emerged from a systematic analysis of microRNA and mRNA profiles in the fibrotic environment, identifying microRNA “hubs” (93) that target fibrosis-associated genes. These studies pinpoint a critical role for PPARγ regulation of microRNAs in mouse models of obesity and type 2 diabetes.

Transcriptional regulation of de novo lipid synthesis is partially mediated by insulin activation and feedback inhibition of sterol regulatory binding proteins (SREBPs) (94). Both Srebpf1 (Srebp1) and Srebf2 (Srebp2) genes are host genes for a highly conserved intronic microRNA. SREBF1 on chromosome 17 harbors miR-33b in intron 17, and SREBF2 on chromosome 22 contains miR-33a in intron 16. Mature miR-33a and miR-33b are highly homologous with largely overlapping target spectrums (13, 14, 95) expressed in the brain, liver, and adipose tissues. miR-33 expression occurs concurrently with Srebp transcripts and allows co-regulation of these microRNAs and the SREBP proteins during insulin and lipid signaling. In agreement with this concept, miR-33a and miR-33b expression are especially abundant in macrophages and the liver (13, 14). Accordingly, inhibition of miR-33 in mouse models of atherosclerosis by antisense methods or genetic ablation generate potent beneficial effects. These findings have suggested miR-33 is a therapeutic target in cardiovascular disease. However, genetic deletion of miR-33 also promotes obesity and insulin resistance (96), suggesting that miR-33 furnishes complex metabolic functions that need reassessment.

As mentioned previously, miR-122 is the most abundant microRNA in the liver. Its expression mostly induces beneficial effects, performs anti-inflammatory actions, and slows tumor progression (97-99). Multiple hepatic nuclear receptors regulate miR-122 transcription. Hepatocyte nuclear factor 4alpha (HNF4α) is the most well-defined upstream regulator. Genomic recruitment of HNF4α to the miR-122 promoter enhances its transcription (100) to allow miR-122 to control many glucose and lipid metabolism that oppose hepatic insulin resistance and carcinogenesis (101, 102). In addition, other nuclear receptors such as RAR-related orphan receptor alpha and farnesoid X receptor positively regulate miR-122 transcription to suppress hepatic fat accumulation and liver cancer, respectively (103, 104). In contrast, the xenobiotic nuclear receptor constitutive androstane receptor inhibits miR-122 transcription by competing for binding sites for HNF4α (105, 106), which presumably mediates the carcinogenic activity of constitutive androstane receptor through derepression of the c-Myc gene (105).

Unlike nonalcoholic steatohepatitis and liver cancer, hepatic ischemia and reperfusion (IR) injury increases miR-122 expression. Hepatic elevation of miR-122 and its passive release into blood circulation are strongly associated with the degree of hepatic IR injury (107). Hepatic IR activated hypoxia-inducible factor 1alpha (HIF1α) to induce miR-122 expression and perform adaptive liver injury responses (108). Hepatocyte-specific loss of miR-122 accelerates liver injury upon IR stress, suggesting that the HIF1α–miR-122 axis has a protective role against hepatic IR injury. In addition to the hepatic nuclear receptors and HIF1α, other transcription factors like CEBPα and grainyhead like transcription factor 2 directly regulate miR-122 transcription in liver cancer and alcoholic liver disease (109-111). Paradoxically, and for reasons that remain unknown, circulating miR-122 is elevated in both patients with fatty liver disease (112, 113) and individuals with obesity (114). Nonetheless, these studies reveal an extensive network of transcriptional regulation of microRNA signaling that functions to modulate glucose, lipid, and cholesterol metabolism.

Packaged microRNAs Perform Inter-organ Communication

Adipose tissue-derived exosomal microRNAs (AT Exos) likely represent a new form of adipokine. In addition to their endogenous actions, microRNAs can be transported into the extracellular space within nanoparticles (50-150 nm in size) termed exosomes. Exosomes transport most of the microRNA secretome and allow paracrine and endocrine transfer of microRNA silencing into circulation. A significant proportion of plasma microRNAs originate from adipose tissues, and the amount of AT Exos increases in obesity (115) to likely influence metabolic homeostasis (Fig. 1). The amount of AT Exos may be closely tied to absolute amounts of WAT. In patients with lipodystrophy, circulating microRNAs are significantly depleted relative to healthy controls (20).

Figure 1.

AT Exos mediate organ crosstalk that degrades insulin sensitivity. AT Exos contain microRNAs and a unique signature of other lipid and protein cargo to transfer signals between organs in the periphery. In obese WAT, adipocytes and macrophages secrete exosomes that accumulate in the liver, skeletal muscle, and likely the pancreas. The elevated exosome concentrations in serum positively correlate with insulin resistance in obesity. Please see text for details.

Adipocyte-specific knockout of Dicer in mice (Dicer a-KO) depletes circulating microRNAs that communicate with peripheral tissues. Along these lines, transplantation of wild-type WAT into Dicer a-KO restored plasma microRNAs and improved glucose tolerance (20). Dicer a-KO also showed reduced miR-99b in circulating exosomes coupled with higher liver Fgf21 mRNA, which could be partially corrected with the administration of miR-99b loaded into exosomes (20). In other paradigms, TGF-beta accumulation in WAT of obesity induces secretion of AT Exos containing miR-130b that acts on skeletal muscle oxidative metabolism (116) and exacerbates IR injury in the diabetic heart (117). In healthy volunteers, the release of exosomal miR-92a from brown adipocytes correlated inversely with BAT activity measured by metabolic imaging and represents a biomarker for human BAT function. Paracrine functions of AT Exos also exist. Large adipocytes use exosomes to transfer microRNAs to induce hypertrophic obesity responses in small adipocytes (118).

Inflammation in WAT and liver may also be mediated by AT Exos. Early studies found exosomes secreted by the adipose tissue of ob/ob mice increased circulating IL-6 and TNF-α and contributed to insulin resistance (119). Ying and colleagues built upon these observations and showed that WAT macrophage exosomes from obese mice increase insulin resistance in vivo. Moreover, WAT macrophage exosomes from the local macrophages of lean mice attenuate insulin resistance in obese mice (120). In addition to regulating adipocyte differentiation (67), obesity increases levels of miR-155 in WAT macrophage-derived exosomes. The heightened miR-155 in these exosomes increased in vitro insulin resistance by suppressing PPARγ (120) and SOCS1 (121) to enforce obesity-linked inflammatory responses in distal tissues.

Obesity changes the profile of circulating microRNAs, and countless studies now present microRNAs as biomarkers of metabolic diseases (122). Many microRNAs correlate strongly with obesity-related measures (123, 124) and reductions after weight-loss surgeries (125). Additional evidence established the notion that adipose-tissue derived microRNAs miR-221, miR-201, miR-222, miR-16, let-7b, miR-34a, miR-103, miR-146b, and miR-148a incorporate into AT Exos (20, 54, 126). In contrast, under healthy conditions, M2 macrophages secrete miR-690-containing exosomes to reduce inflammation and improve insulin sensitivity (127). Although WAT and BAT secrete AT Exos that mediates responses in recipient cells and tissues, more studies will be needed to understand the endocrine functions of microRNAs.

Future Perspectives

Although the understanding of roles for microRNAs in metabolism and insulin sensitivity progressed rapidly, we know only a handful of microRNAs capable of significantly modulating whole-body insulin sensitivity in animal models. Additionally, for these microRNAs, only a few important targets are clearly defined. Although changes in microRNA expression in human insulin resistance and type 2 diabetes exist, the limited available data do not draw a clear picture, and interpretations draw from purely correlative observations. microRNAs that go up or down in the serum of obese rodents and humans also abound, but reproducibility and underpowered approaches slow the development of serum microRNAs as reliable biomarkers (122). Significant pitfalls must be considered, including uniformity in microRNA measurement techniques and confounding factors related to age, body mass index, and sex differences. Experimentally, the development of conditional and tissue-specific microRNA knockout mice will be especially valuable in studying microRNAs that are cell and tissue-biased.

A few compelling demonstrations that AT Exos are endocrine factors for whole-body metabolic control provide a new perspective on organ communication. The liver also produces exosomes in obesity that promote and diminish insulin sensitivity (128). Yet, it remains unclear whether circulating microRNAs secreted by the liver and other tissues, such as the muscle or pancreas, generate metabolic effects. Also, some obesity-associated microRNAs target overlapping pathways, which presents significant challenges for the field to reproduce previously identified circulating obesity-associated microRNAs in rigorous ways. Standardized research methods might help in this regard to identify the full set of circulating microRNAs that affect metabolism and to accurately determine the tissue source.

Reports of microRNA–mRNA relationships based on in vitro data may have limited physiologic relevance and must be appropriately vetted in vivo. The field must consider larger impacts of microRNA expression and explore networks of targeting events in vivo. Future research will require additional effort to define novel pathways altered by currently undescribed microRNAs or noncanonical microRNAs, in addition to more comprehensively defining the multiple pathways and targets through which known microRNAs operate. New approaches such as AGO-CLIP-seq and mass spectrometry proteomics will facilitate the determination of the composition and regulation of microRNA networks to explore how these microRNA networks are coordinated with mRNA and protein turnover. Finally, considerations for regulatory selection events and other environmental influences (eg, circadian biology) that govern microRNA expression might help us understand microRNA targeting in physiological ways. In sum, although the regulatory potential of microRNAs in human metabolic disease is currently apparent, additional efforts to define the scope and mechanisms of this control are necessary.

Glossary

Abbreviations

AGO2

Argonaute 2

AGO

Argonaute proteins

AGO-CLIP-seq

immunoprecipitation of the Argonaute protein-coupled to high throughput sequencing

AT Exos

adipose tissue-derived exosomal microRNAs

BAT

brown adipose tissue

CLIP

crosslinking and immunoprecipitation

Dicer a-KO

adipocyte-specific knockout of Dicer in male mice

HIF1α

hypoxia-inducible factor 1alpha

HNF4α

hepatocyte nuclear factor 4alpha

IR

ischemia and reperfusion

miR

microRNA

PPARγ

peroxisome proliferator-activated receptor-gamma

SREBP

sterol regulatory binding protein

UTR

untranslated region

WAT

white adipose tissue

Funding

This work was funded by the American Heart Association Career Development Award (19CDA34660196), the Nancy Chang, Ph.D. Award for Research Excellence at Baylor College of Medicine, American Diabetes Association #1-18-IBS-105, and National Institutes of Health (NIH) R01DK126656, NIH R01DK126042, and NIH R01DK114356.

Data Availability Statement

Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.