miRNAs regulate protein abundance and control diverse aspects of cellular processes and biological functions in metabolic diseases, such as obesity and type 2 diabetes (T2D). Let (lethal)-7 miRNAs specifically targets genes associated with T2D and have been implicated in the regulation of peripheral glucose metabolism, yet the direct regulators of let-7 miRNA expression are unknown. In the present study, we report on a putative promoter region for the let-7a-1, let-7f-1 and let-7d gene cluster on chromosome 9 and characterize the promoter activity of this novel area. We show that promoter activity and let-7 miRNA expression is dynamically regulated in response to different factors including serum, glucose, tumour necrosis factor (TNF)-α and caffeine. These findings will contribute to understanding the interaction between precise promoter elements to control the transcription and translation of let-7 miRNA genes.
miRNAs inversely regulate protein abundance and thereby control diverse regulatory aspects of numerous cellular processes in most cell types . Changes in miRNA expression are linked to a variety of biological functions including pathological states such as cancer and metabolic diseases. The let-7 miRNAs were originally described in Caenorhabditis elegans , where they control developmental timing. Let-7 miRNA plays an important role as a tumour suppressor to regulate oncogenes [3,4]. The let-7 family of miRNAs also regulate peripheral glucose metabolism in metabolic disease, since its target genes are associated with type 2 diabetes (T2D) [5–7]. Although the gene regulatory effects of let-7 miRNAs have attracted attention, little is known regarding the regulation of let-7 miRNA expression itself.
The miRNA let-7 family is present in multiple copies in the genome and mature let-7 is highly conserved across species . Several of the miRNA let-7 family members are also conserved across species and the genomic organization and clustering are also conserved . In humans, the miRNA let-7 family includes several more members than are present in C. elegans. Moreover, there is also significant post transcriptional modification by the RNA-binding protein, Lin28 (C. elegans lin-28 homologous protein), which selectively blocks and controls let-7 biogenesis  and the H19 lncRNA (long non-coding H19 gene) which also inhibits let-7 bio-availability .
Expression of miRNA let-7a and let-7d is increased in adult skeletal muscle, as well as in cultured myotubes, obtained from type 2 diabetic patients . Whole-body transgenic mice specifically overexpressing let-7a, 7d and 7f are glucose intolerant . Thus, in the present study, we focused on the transcriptional regulation of let-7a and let-7d, since these were up-regulated in human T2D. In order to elucidate the regulation of these let-7 miRNAs, we investigated the human genomic region postulated to include the promoter of let-7a-1, let-7f-1 and let-7d, which form a miRNA cluster on chromosome 9.
Human embryonic kidney (HEK)293 cell line were obtained from A.T.C.C. and cultured in high-glucose (4500 mg/l) Dulbecco's modified Eagle's medium (DMEM; Invitrogen, Life Technologies) supplemented with 10% (v/v) FBS (Sigma–Aldrich). Twenty-four hours before the experiment, cells were seeded either in 96- or six-well plates.
Primary human skeletal muscle culture
Satellite cells were isolated from the vastus lateralis skeletal muscle biopsies as described . Cell cultures were maintained at 37°C under 7.5% CO2 as myoblast cells and myoblast cells were used to transfect promoter plasmids. To measure gene expression of human skeletal muscle cells, myoblast cells were differentiated into myotubes. To initiate differentiation into myotubes, DMEM/F12 (Invitrogen, Life Technologies) supplemented with 20% (v/v) FBS and 1% penicillin–streptomycin (PeSt) was removed from cells and replaced with ‘fusion medium’ for 4 days. Fusion medium was composed of 76% high-glucose (4500 mg/l) DMEM (Invitrogen, Life Technologies), 20% Medium 199 (Invitrogen, Life Technologies), 2% HEPES (Invitrogen, Life Technologies), 1% PeSt, 0.03 μg/ml zinc sulfate (Sigma), 1.4 mg/ml vitamin B12 (Sigma), 100 μg/ml apo-transferin (Sigma) and 10 μg/ml insulin (Novonordisk). Fusion medium was removed from cells after 4 days incubation and replaced with ‘post-fusion medium’. Post-fusion medium was composed 74% high-glucose (4500 mg/l) DMEM, 20% medium 199, 2% FBS, 2% HEPES, 1% PeSt, 0.03 μg/ml zinc sulfate (Sigma) and 1.4 mg/ml Vitamin B12 (Sigma) for 6 more days. Myotube cells where fusion and multinucleation were observed at day 10 after initiation of differentiation were used for total RNA extraction.
To construct the let-7 promoter plasmid, we obtained genomic co-ordinates of the let-7 family genes and upstream region information from the University of California, Genome Browser by the ENCODE (ENCyclopedia Of DNA Elements) Project (http://genome.ucsc.edu/cgi-bin/hgGateway?redirect=manual&source=genome-euro.ucsc.edu). We synthesized (GeneArt, Life Technology) the fragment upstream of the let-7a-1/let-7d/let-7f-1 miRNA genomic region and cloned it into a plasmid containing a luciferase reporter gene, pGL4.23 (Promega) between KpnI and XhoI site (Figure 1). This plasmid was named as let7p, subsequent promoter truncated clones (construct 1–4 and let7p 631bp) were amplified using this template clone and HotStar HiFidelity Polymerase Kit (Qiagen) and cloned similarly into pGL4.23.
Let-7a-1, f-1 and d genomic position and putative promoter region on chromosome 9
Reporter gene analysis
The let-7p plasmid (100 ng) was transfected into 2×104 HEK293 cells per well of a 96-well plate by using a transfection reagent (Lipofectamine 2000, Life Technologies) with 1 ng of Renilla luciferase vector (Promega) for normalization. After 24 h, cells were incubated for 6 h in the absence or presence TNF-α (tumour necrosis factor α; Sigma), caffeine (Sigma) or insulin (120 nM). The luciferase activity of the reporter gene was measured with the Steady-Glo Luciferase Assay System (Promega) and normalized with the Renilla activity. In the case of myoblast cells, Lipofectamine 3000 (Life Technologies) was used as a transfection reagent, whereas other conditions of the reporter gene analysis experiments were the same as in the case of HEK293 cells.
RNA isolation and determination of miRNA and mRNA expression
For RNA extraction, HEK293 cells or human cultured myotubes were homogenized in TRIzol and total RNA was isolated according to manufacturer's instructions (Life Technologies) using Phase-Lock Heavy Gel tubes (5PRIME) and stored at–80°C until further use. Equal amounts of total RNA (20 ng) were used as input for each sample to synthesize the first strand of cDNAs with reverse transcriptase (RT), and quantitative (q) RT-PCR was performed using TaqMan primers of individual miRNAs according to manufacturer's protocol (Life Technologies). A qPCR reaction was performed using a StepOne Plus Real-Time PCR systems (Life Technologies). For normalization we used RNU44 as reference. All reagents were from Life Technologies.
Bioinformatic analysis to identify putative promoter region
miRBase release 21 (http://www.mirbase.org/) identifies several mature let-7 family sequences in the human genome. Mature miRNA let-7 (let-7a-1, let-7a-2 and let-7a-3) are produced by three separate precursors. Let-7a-1, let-7d and let-7f-1 exist in the same pri-miRNA cluster. Therefore, we focused our efforts on the identification of the promoter region of this cluster, since our primary emphasis was the regulation of let-7a and let-7d. A large highly CpG (-C-phosphate-G-) rich island was seen in typical transcription regulating elements 10-kbp upstream of the let-7a-1/let-7d/let-7f-1 cluster (Figure 1). The levels of enrichment of the H3K27Ac (histone H3 acetylated at Lys27) histone mark from ChIP coupled with high-throughput sequencing (ChIP-seq) data support the hypothesis that this region is an active regulatory element (Figure 1). ChIP-seq data further indicate that several transcription factors bind to this region . Although this region is clearly some distance from the pri-miRNA locus, human EST data supported that long transcripts that included pri-miRNAs are transcribed from this region (Figure 1; Table 1). Furthermore, the mammalian promoter level expression atlas showed enrichment of sequence tags that suggest transcribed products (http://fantom.gsc.riken.jp/5/; Figure 2). In the FANTOM5 (Functional ANnoTation of Mammalian Genome) promoterome atlas, several transcripts were identified and sequenced in skeletal muscle cells (Figure 2). Interestingly, the let-7a-1/let-7f-1/let-7d cluster appears to be encoded as a single transcript driven by this region . Therefore, we concluded that this region 10-kbp upstream of miRNA let-7a-1/let-7d/let-7f-1 cluster was likely to contain the promoter regulating expression of miRNA let-7a-1/let-7f-1/let-7d genes.
Snapshot from FANTOM5 human promoterome viewer, ZENBU
Promoter activity of putative promoter region
The miRNA let-7a-1/let-7d/let7-f-1 gene cluster, including the 10-kbp upstream region and the 2.6-kbp putative promoter region is highly conserved between species, with mouse showing 78% identity and rat 79% identity to the human sequence (Figure 3). We selected a 2589 bp (2.6-kb) conserved region as the initial target for promoter analysis. We considered cloning this region from the human genome; however, in addition to a large CpG island, genomic databases indicated a large number of SNP (single nucleotide polymorphism) sites. In light of the architecture of this region and to avoid introducing technical errors, the 2.6-kbp region was chemically synthesized and cloned into pGL4.23 luciferase promoter vector. When this promotor vector was transfected into HEK293 cells, we noted significant luciferase expression driven by this 2.6-kbp region (Figure 4).
Sequence comparison of the human putative promoter region with other species including mouse (A) and rat (B)
Luciferase expression driven by cloned 2.6-kbp region
To explore the mechanism by which activity of this promoter region is regulated, HEK293 cells were serum starved for 4 h and then exposed to either 10% FBS, insulin (120 nM), glucose (20 mM), TNF-α (1 or 2 μg/ml) or of caffeine (0.5 or 5 mM) for 24 h. Re-introduction of 10% FBS increased promoter activity 1.7-fold (Figure 5A; P<0.01). Exposure of HEK293 cells to glucose dose-dependently increased promoter activity, (1.2- or 1.4-fold, P<0.05, in the presence of 20 or 40 mM respectively; Figure 5B). TNF-α dose-dependently increased promoter activity, (1.7- or 2.3-fold, P<0.05, in the presence of 1 or 2 μg/ml respectively; Figure 5C). Similarly, caffeine increased promoter activity (2.1- or 2.9-fold, P<0.01, in the presence of 0.5 or 25 mM respectively; Figure 5D). Insulin exposure did not alter promoter activity (Figure 5E).
Exogenous factors regulating the let-7 promoter region
Glucose-mediated induction of promoter in human skeletal muscle cells
To determine whether the identified promoter region was active in skeletal muscle, the pre-dominant site of glucose storage under insulin-stimulated conditions and therefore the central organ in the regulation of whole-body glycaemic control , we performed luciferase determination in response to glucose using cultured human primary skeletal muscle cells. The promoter reported plasmid was transfected into these cells and promoter luciferase activity was measured in the absence or presence of glucose (Figure 6). Hyperglycaemic conditions resulted in increased promoter activity in human skeletal muscle cells from two separate donors.
Glucose increases activity of the let-7 promoter region in human skeletal muscle cells
Regulation of miRNA content
To determine whether glucose, insulin or TNF-α would lead to increased let-7 miRNA content in HEK293 cells, we assessed the expression of three different let-7 family miRNA members using the same conditions as above (Figure 7). Exposure of HEK293 cells to 20 mM glucose or 2 μg/ml of TNF-α increased miRNA let-7a-1/let-7d. Interestingly, insulin exposure, which did not increase promoter activity, also increased let-7a-1/let-7d miRNA. Caffeine exposure did not significantly increase endogenous let-7a-1/let-7d expression. Cell content of miRNA let-7b was unaltered in response to all stimuli. However, miRNA let-7b is located on a different chromosome and has a different promoter from let-7a-1/d/f-1.
Insulin, glucose and TNF-α increases expression of let-7a and let-7d
Next, we tested selective truncation of the 2.6-kbp promoter to determine which part of this region contained the essential promoter activity. The 2.6-kbp region encompasses a large, 1.3-kbp, CpG island and several evolutionarily conserved CpG islands (Evo CpG). ChIP-seq and DNase assay data indicated H3K27AC marks and three DNase clusters . These marks are usually indicative of functional units. Therefore, we constructed promoter truncation fragments taking into account the functional characters of the promoter (Figure 8). A schematic outline of promoter constructs is shown (Figure 8A). Promoter constructs with sequential 5′ deletions (construct 1–4) were transfected into HEK293 cells and luciferase activity was determined (Figure 7B). As shown in Figure 1, the construct containing the entire 2.6-kbp sequence (let7p, 1–2589) significantly increased luciferase activity. Loss of a 631-bp region, (compare construct 3 and 4; Figure 7B) resulted in a loss of luciferase activity. We then specifically focused on this region (1–2589: 1404–2013) and noted that it was sufficient to drive luciferase activity to 69% of that recorded with the whole region. Our results indicated that this middle region of the 2.6-kbp sequence (1–2589: 1404–2013) contains the essential components for promoter activity in HEK293 cells.
Deletion analysis of the let-7 promoter region
Transcription factor candidates for regulation of let-7a/let-7d promoter
To search for transcription factors likely to regulate let 7 promoter activity, we interrogated ChIP-seq data present in the official database. ENCODE TFBS ChIP-seq data derived from a large collection of ChIP-seq experiments allowed us to identify transcription factors MYC (c-myc-derived protein), SMARCB1 (SWI/SNF-related matrix-associated actin-dependent regulator of chromatin subfamily B member 1)/INI1, and EGR1 (early growth response 1) as possible candidates able to bind to the middle region in the 2.6-kbp sequence which we identified as important for promoter activity in our promoter characterization (Figure 8). Next, we tested how mRNA expression of these three transcription factors was changed in response to insulin, caffeine, glucose or TNF-α (Figure 9). mRNA expression of myc was increased in response to TNF, caffeine and serum, (1.2-, 1.5- or 1.2-fold respectively; Figure 9A). mRNA expression of SMARCB1 and EGR1 gene was decreased following exposure to insulin or serum (0.7- or 0.8- fold respectively; Figures 9B and 9C).
TNF-α or caffeine increase expression of MYC
mRNA levels of let-7 target genes
Finally we determined mRNA levels of several let-7 target genes following exposure to insulin, glucose, TNF-α or caffeine in both human skeletal muscle cells and HEK293 cells (Table 2). In line with increased let-7 activity, several insulin pathway genes, including insulin receptor (INSR) and INSR substrate-2 (IRS2) mRNA were decreased in response to insulin and glucose in cultured human muscle cells. TNF-α exposure similarly resulted in reductions in INSR and IRS2 in human cultured muscle cells. Whereas insulin exposure reduced mRNA expression of several genes in HEK293 cells, glucose, TNF-α or caffeine led to increased mRNA expression in some instances.
|Muscle cells||HEK293 cells|
|Muscle cells||HEK293 cells|
miRNA let-7 family members are highly conserved throughout several organisms. The expression of let-7 is dysregulated in disease states including cancer and metabolic disease . Although let-7 expression is increased in tumour tissue, the expression level in some cancer cells, including human breast cancers and lung cancers, is reduced as compared with normal tissues . Oncogenes such as RAS (rat sarcoma-derived protein) and HMGA2 (high-mobility group A2 protein) are miRNA let-7 target genes, suggesting that the lower expression of miRNA let-7 genes might lead to the activation of these oncogenes in cancer cells [4,18]. Although miRNA let-7 family genes have been widely studied in cancer, let-7 miRNAs may also regulate glucose and energy homoeostasis in inflammatory and metabolic diseases. For example, mice with either let-7 overexpression or let-7 silencing have altered glucose homoeostasis and insulin sensitivity [6,19]. Furthermore, a number of suggested miRNA let-7 target genes regulate glucose metabolism in several different tissues . The RNA-binding protein, Lin28, which inhibits the processing of let-7 family genes, influences glucose metabolism in concert with miRNA let-7 . Expression of let-7a and let-7d is up-regulated in cultured skeletal muscle myotubes, as well as in intact skeletal muscle biopsies from type 2 diabetic patients, as compared with normal glucose tolerant people . These observations prompted us to examine the regulation of miRNA let-7.
Based on available information regarding miRNA let-7 gene structure obtained from genomic databases, we cloned the putative promoter region of the let-7a-1/let-7d-1/let-7f gene cluster, located approximately 10-kbp upstream of the miRNA encoding region. Although this region is rather distant, it contains typical promoter characters such as GC islands and transcription factor-binding sites predicted by a motif-prediction program. Furthermore, ChIP-seq analysis using several different transcription factors and histone-binding proteins further indicated that this region was a likely to have promoter function (Figure 1). Furthermore, the structure of the let-7a-1/let-7d-1/let-7f gene cluster and of the putative promoter sequence we identified is conserved in the mouse and rat genome, lending further support to an important role for this genomic region. The chromosomal region identified in this paper also displayed promoter activity in HEK293 cells and in primary human skeletal muscle cells. This activity was specifically up-regulated following exposure to serum, glucose, TNF-α or caffeine, factors that also increased the level of let-7 miRNA species. Insulin did not increase activity of the promoter; however, since the luciferase activity is a function of a stable protein accumulation and endogenous gene expression is subject to a number of different dynamic regulators a direct comparison is challenging. Type 2 diabetic patients have elevated blood glucose and an increased inflammatory tone  and thus results are consistent with our earlier report of increased miRNA let-7a and -d in skeletal muscle from type 2 diabetic patients . Whether these factors directly increase the expression of let-7 miRNA in skeletal muscle in people with T2D or obesity remains to be determined.
Understanding the key transcription factors that regulate let-7 miRNA may give insights into the pathophysiology of several diseases including cancer, immunology and T2D. To gain insight into the critical region involved, we analysed several deletion mutants of the promoter construct, identifying a key 0.6-kbp region. Next, we probed which transcription factors might regulate miRNA let-7 gene expression through this 0.6-kbp region by analysing genomic databases of the ENCODE project with the UCSC genome browser (https://genome.ucsc.edu/). ChIP-seq experiments indicate that MYC, SMARCB1 or EGR1 might bind to this region. The 0.6-kbp region contains several E-box (motif E-Box sequence) sites which are known to be MYC-binding motifs. A previous report highlighting a similar region upstream of miRNA let-7a-1/let-7d/let-7f-1 indicated that MYC inhibited transcription of these miRNAs . Although this region shows some overlap with the region studied in our present report, a large difference in the critical region for promoter activity was identified. The critical region for promoter activity was found to be upstream of the large CpG island, whereas in our report it was found to be downstream. This could partly be due to differences in the cell models studied, i.e. HEK293 cells (present study) compared with immortalized human liver cell line L02 . Furthermore, of the three identified potential transcription factors, MYC mRNA was most consistently up-regulated in response to factors increasing promoter activity. Expression level is only one potential level of regulation for transcription factors and promoter utilization by transcription factors is dependent on specific cell types complicating studies of transcriptional control in mammalian cells. Despite this, a potential role for MYC protein binding was highlighted in the present study and previously , since the promoter area that we identify has several putative MYC-binding sites (E-box or E-box-like site).
Epigenetic modification may also control gene promoter activity. Hypermethylation of the let-7a-3 locus, located on chromosome 22q12.31, occurs in a variety of different tumour tissues [23,24]. In the case of the let-7a-3 gene locus, hypermethylation of this gene locus functions as an oncogene in epithelial ovarian cancer [25,26]. The promoter region of the chromosome 9 let-7 cluster genes that we identified contains a typical CpG island, prompting us to determine the methylation status of this CpG island. However, the large size of the CpG island in the let-7 promoter made this technically challenging and we were unable to obtain a reliable result. We noted not only increased miRNA levels in skeletal muscle from type 2 diabetic patients, but also increased let-7a and d expression in cultured muscle cells derived from type 2 diabetic patients . These cultures are derived from skeletal muscle satellite cells and the increased expression was noted in cells that had been cultured for several weeks, thus reducing the impact of systemic circulatory factors. The increased expression of let-7a and d could be due to a genetic pre-disposition for increased let-7 miRNA, but could also be controlled by epigenetic mechanisms
To determine potential physiological regulators or let-7 promoter activity, we exposed cultured cells to systemic factors or cellular stressors that modify glucose metabolism. Exposure of HEK293 cells to high glucose or TNF-α up-regulated both promoter activity and let-7 miRNA expression, whereas exposure to insulin increased let-7 miRNA without altering promoter activity. Similarly, hyperglycaemia up-regulated promoter activity in primary human muscle cells. Let-7 miRNA content is dependent on transcription, but also on interaction with Lin28 , as well as with the long non-coding RNA lncH19 . Lin28 stabilizes let-7, whereas H19 reduces let-7 bioavailability without altering miRNA expression level, thus making the regulation of let-7 in cells highly complex. Furthermore, the effect of insulin on let-7 promoter activity may be further complicated, since let-7 miRNAs are reported to interact with the 3′-UTR mRNA of the INSR, insulin growth factor-1 receptor (IGF-1R) and IRS2 and thus down-regulate canonical insulin/IGF signalling pathways . In line with this, stimuli which up-regulated promoter activity and let-7 content also resulted in a reduction in mRNA of several of the predicted let-7 target genes in the insulin/IGF signalling pathways.
Glucose metabolism is influenced by several systemic factors in both healthy subjects, as well as in type 2 diabetic or cancer patients. Previous reports in rodent models of Lin28 or let-7 transgenic [5,19] suggest that the Lin28/let-7 axis is one key regulatory pathway. In the present study, we report let-7 miRNA expression is dynamically regulated through a novel promoter region fairly distant from the mature let-7 miRNA coding region and the activity is altered in response to different factors including glucose, TNF-α and caffeine. We also characterize the promoter activity of this novel area. These findings will contribute to understanding the interaction between precise promoter elements to control the transcription and translation of let-7 miRNA genes.
The authors thank Professor Juleen R Zierath for helpful discussions. Mutsumi Katayama and Anna Krook conceived and designed the research. Mutsumi Katayama performed the experiments. Mutsumi Katayama, Rasmus Sjögren and Brendan Egan analysed data and interpreted results of experiments. Mutsumi Katayama prepared figures and tables. Mutsumi Katayama and Anna Krook drafted the manuscript. All authors edited the paper.
This work was supported by the Swedish Diabetes Association (Krook) the Novo Nordisk Foundation (Krook) the Diabetes Wellness Foundation [grant number 2949/2014SW]; the Swedish Research Council [grant number K2013-55X-12669-16-5]; the Swedish Foundation for Strategic Research [grant number SRL10-0027]; and the European Research Council Ideas program [grant number ERC-2008-575 AdG23285].
ChIP coupled with high-throughput sequencing
Dulbecco's modified Eagle's medium
motif E-Box sequence
early growth response 1
Functional ANnoTation of Mammalian Genome
histone H3 acetylated at Lys27
human embryonic kidney
insulin growth factor-1 receptor
insulin growth factor
insulin receptor substrate-2
Caenorhabditis elegans lin-28 homologous protein
long non-coding H19 gene
SWI/SNF-related matrix-associated actin-dependent regulator of chromatin subfamily B member 1
single nucleotide polymorphism
type 2 diabetes
tumour necrosis factor-α
tumour protein p53