Transcriptional Inhibition of MicroRNA miR-122 by Small Molecules Reduces Hepatitis C Virus Replication in Liver Cells

Abstract

MicroRNAs (miRNAs) are noncoding RNA molecules of 22–24 nucleotides that are estimated to regulate thousands of genes in humans, and their dysregulation has been implicated in many diseases. MicroRNA-122 (miR-122) is the most abundant miRNA in the liver and has been linked to the development of hepatocellular carcinoma and hepatitis C virus (HCV) infection. Its role in these diseases renders miR-122 a potential target for small-molecule therapeutics. Here, we report the discovery of a new sulfonamide class of small-molecule miR-122 inhibitors from a high-throughput screen using a luciferase-based reporter assay. Structure–activity relationship (SAR) studies and secondary assays led to the development of potent and selective miR-122 inhibitors. Preliminary mechanism-of-action studies suggest a role in the promoter-specific transcriptional inhibition of miR-122 expression through direct binding to the liver-enriched transcription factor hepatocyte nuclear factor 4α. Importantly, the developed inhibitors significantly reduce HCV replication in human liver cells.

Graphical Abstract

INTRODUCTION

MicroRNAs (miRNAs) are short, noncoding RNA molecules that bind to the 3′ untranslated region (3′ UTR) of complementary mRNAs and inhibit gene expression.¹ In the canonical miRNA pathway, RNA polymerase II or III transcribes primary miRNAs (pri-miRNAs) that are subsequently cleaved by Drosha, producing a short hairpin known as precursor-miRNA (pre-miRNA).² The pre-miRNA is then exported from the nucleus to the cytoplasm³ where it is cleaved by Dicer and loaded into an RNA-induced silencing complex (RISC). The activated RISC then modulates gene expression via binding to the target mRNA.^4,5 MiRNAs are predicted to regulate 21% of protein-coding genes in humans,⁶ based on PAR-CLIP studies, and their dysregulation has been implicated in the development of various human diseases.^7–9

MicroRNA-122 (miR-122) comprises ~70% of the miRNAs in the human liver¹⁰ and plays a key role in liver development, cell differentiation, homeostasis, and function.¹¹ Furthermore, miR-122 has been shown to regulate lipid- and fatty acid-associated genes and is implicated in iron and glucose homeostasis; however, the mechanisms by which this occurs are still poorly understood.¹² MiR-122 has also been implicated in cholesterol and lipid metabolism,¹³ and inhibition may be suitable for the treatment of metabolic syndrome and type 2 diabetes.¹⁴ Importantly, miR-122 is an essential component for the replication of the hepatitis C virus (HCV) in human liver cells.^15–20 Long-term HCV infection accounts for ~30% of liver transplants annually in the U.S. and can lead to the development of hepatocellular carcinoma.²¹ Furthermore, the CDC found that new cases of HCV infection have tripled in the U.S. over the last 5 years²² and the World Health Organization estimates a global incidence rate of 23.7 per 100,000 people as of 2015.²³ HCV is a single-stranded RNA virus in which the genome encoding the viral polyprotein is flanked by 5′ and 3′ UTRs.²⁴ The 5′ UTR is conserved across viral genotypes and contains essential structures for genome replication and translation.^25,26 Within the conserved 5′ UTR are two complementary binding sites for miR-122, which play a unique and significant role in regulating the life cycle of the virus.¹⁵ The mechanisms of miR-122-mediated HCV regulation include protection of viral RNA from degradation,^27–29 promotion of internal ribosomal entry site (IRES) formation,^30,31 and enhancement of viral replication.^32,33 As a facilitator of HCV propagation, miR-122 is a promising target for the treatment of HCV. Current antiviral therapies target viral proteins required for various steps throughout the HCV life cycle, including (I) viral entry;³⁴ (II) protease-mediated polyprotein processing;³⁵ (III) RNA replication;^36–38 and (IV) virion assembly (Figure 1A).³⁸ However, the emergence of resistance-associated variants (RAVs) of HCV to direct-acting antiviral agents has prompted the development of treatment regimens that combine agents targeting different viral proteins.³⁹ As such, miR-122 represents an attractive alternative antiviral target. Sequence-specific inhibition of miR-122 function has been demonstrated using antisense oligonucleotides, and locked nucleic acids targeting miR-122 are currently in phase II clinical trials for treatment of HCV.⁴⁰ Significant technological advancements have facilitated the recent Food and Drug Administration (FDA) approval of nucleic acid-based therapeutics, including adeno-associated viral vectors, chemically modified antisense oligonucleotides, and N-acetylgalactosamine-conjugated siRNA, increasing the clinical applicability of nucleic acids. Despite this, compared to small molecules, oligonucleotide therapies face several hurdles that must be overcome during development, including stability,⁴¹ pharmacokinetics,⁴² and delivery.⁴¹ Various small-molecule modulators of miRNA function have been developed in recent years to investigate miRNA biogenesis and to address the shortcomings of nucleic acid-based inhibitors.^43,44 Furthermore, while oligonucleotides bind directly to precursor or mature miRNAs, small-molecule modulators have the potential to target various steps of the miRNA pathway, such as miRNA transcription, maturation, or loading of miRNAs into argonaute 2 (Ago2). Herein, we report the development of two potent, selective small molecules that inhibit transcription of miR-122 and demonstrate their potential as therapeutics by reducing HCV replication in liver cells.

Figure 1.

Overview of the HCV life cycle and inhibitor screening. (A) Current direct-acting antiviral agents for the treatment of HCV infection target proteins involved in (I) viral entry; (II) polyprotein processing; (III) RNA replication; and (IV) virion assembly. Here, we are reporting small-molecule-mediated inhibition of miR-122 impairing HCV RNA replication. (B) Assay for small-molecule inhibitors of miR-122 function. In the presence of miR-122, the luciferase signal is suppressed. Treatment with a small-molecule inhibitor of miR-122 alleviates repression of the reporter, resulting in the expression of luciferase and generation of a luminescence signal.

RESULTS AND DISCUSSION

To generate a reporter of miR-122 function, an expression plasmid bearing a miR-122 binding site downstream of a luciferase gene in the 3′ UTR (Figure 1B) was stably introduced into Huh7 liver carcinoma cells (Huh7-miR122).⁴⁵ When miR-122 binds to the mRNA target site, luciferase expression is downregulated. However, when treated with a miR-122 inhibitor, luciferase activity is restored. Using a miR-122 antagomir as a positive control, a Z′-factor of 0.74 was calculated, indicating the stable cell line was amenable to high-throughput screening.⁴⁶ A primary screen of 336,006 compounds from the Broad Institute’s Diversity Oriented Synthesis (DOS) collection (71,424 compounds) and the NIH’s Molecular Libraries Probe Production Center Network (MLPCN) library (264,582 compounds) was conducted (Figure S1). The primary screen yielded 1023 compounds that were identified as potential inhibitors—compounds that elicited a ≥5-fold increase in luciferase units relative to the dimethyl sulfoxide (DMSO) control (0.3% hit rate). Small molecules that were identified as false-positives, such as pan-assay interference compounds (PAINs)⁴⁷ known to be promiscuous in biological assays and compounds with chemically reactive, metabolically labile, pH sensitive, or hydrolytically unstable groups were removed, leaving 825 remaining compounds. A dose–response analysis yielded 406 potential hits with an EC₅₀ value of less than 10 μM. These remaining small molecules were cross-referenced with active compounds from a high-throughput screen for miR-21 inhibitors using MLCPN library (PubChem AID 2289), and molecules found to be active in both assays were disregarded. SIMLES strings are provided as a CSV file in the Supporting Information for compounds with activity in the miR-21 assay that were identified as hits in the primary miR-122 screen. Solid samples of the remaining 65 compounds were retested in the stable cell line, and inhibitor 1 was identified as one of the most promising of 32 small molecules that passed this assay. Interestingly, akin to one of our originally identified miR-122 inhibitors from a small pilot screen, as well as a majority of the initial hit compounds discovered in the high-throughput screen,⁴⁸ 1 has a central sulfonamide structural motif.

Following the identification of parent compound 1, a structure–activity relationship study was initiated to improve the potency of 1 and to better understand the chemical functionalities required for its activity. For preparing the analogues, a modular synthetic approach was adopted and two moieties of 1 were derivatized: the phenyl ring and the imidazole. Substituting the 4-methylimidazole with a dimethylamine (2) or an ethylamine group (3) resulted in a significant loss in activity. Drastic reductions in activity were also observed when replacing the 4-methylimidazole with a structurally similar pyrrole (4), 3-methylpyrrole (5), 4-methylpyrazole (6), or 3-methylpyrazole (7), indicating the importance of the imidazole motif.

We next studied the requirement of the methyl group itself through substitution of 4-methylimidazole with an imidazole (8), 4-phenylimidazole (9), or 4-nitroimidazole (10). Removal of the methyl group or replacement with a phenyl moiety resulted in a large decrease in activity, while the nitro group elicited a more modest reduction. Virtually complete loss of activity was also observed for 2-ethyl- (11) or 4,5-dichloroimidazole (12) modifications. Further substitutions of the imidazole by more sterically demanding benzimidazole (13) and indole (14) motifs also led to a decrease in activity compared to 1, again indicating the importance of a small 4-methylimidazole (Table 1).

Table 1.

Structures and Activities of Derivatives 1–14^d

compound	chemical structure	primary screen (RLU)a	Renilla luciferase assayb	psiCHECK-empty assayc
1		100 ± 3%	80%	127%
2		42 ± 3%	--	--
3		42 ± 3%	--	--
4		29 ± 1%	--	--
5		57 ± 4%	87%	--
6		32 ± 1%	--	--
7		44 ± 1%	--	--
8		33 ±4%	-	--
9		35 ± 1%	--	-
10		61.1 ±0.4%	67%	-
11		40 ± 1%	--	-
12		4 ± 1%	--	-
13		75 ± 3%	45%	-
14		55 ± 1%	94%	-

^aRLU values represent the Renilla luciferase (Rluc) luminescence signal in the Huh7-miR122 stable cell line and are first normalized to firefly luciferase (dual reporter system) and then to the activity of the original hit compound 1.

^bIn vitro Renilla luciferase data are normalized to DMSO.

^cpsiCHECK-empty assay data are normalized to DMSO. Data for all assays represents the average ± standard deviation from at least three independent experiments.

^dGray shading indicates the top compounds, 28 and 33.

Next, a series of analogues were synthesized with varied substitution of the phenyl ring. Replacing the isopropyl group with a hydrogen (15) or a methyl (16) group abrogated activity. The rotationally restricted 1,2,3,4-tetrahydronaphthamidyl analogue (17) also showed a significant loss in activity. Furthermore, O-ethyl (18) and O-propargyl (19) analogues exhibited reduced activity (Table 2).

Table 2.

Structures and Activities of Derivatives 15–19

compound	chemical structure	primary screen (RLU)a	Renilla luciferase assayb	psiCHECK-empty assayc
15		24 ± 3%	80%	-
16		32 ± 5%	--	--
17		33.3 ± 0.5%	--	--
18		74 ± 10%	70%	--
19		50 ± 3%	97%	-

^aRLU values represent the Renilla luciferase luminescence signal in the Huh7-miR122 stable cell line and are first normalized to firefly luciferase (dual reporter system) and then to the activity of the original hit compound 1.

^bIn vitro Renilla luciferase data are normalized to DMSO.

^cpsiCHECK-empty assay data are normalized to DMSO. Data for all assays represents the average ± standard deviation from at least three independent experiments.

To further characterize the activity of compound 1 and to inform additional structural modification, compound activity was assessed using a biochemical luciferase assay. Inhibitory effects observed in biochemical luciferase can reveal compounds that exhibit ligand-based luciferase stabilization, which is a potential source of false-positive hits in high-throughput screens that utilize cell-based luminescence assays.⁴⁹ As such, analogues that elicited ≥50% activity of parent compound 1 in the initial cell-based screen were tested in a biochemical Renilla luciferase (Rluc) assay. Compounds that exhibited luciferase inhibition in this biochemical assay were considered potential false-positives. Here, 10, 13, and 18 induced ≥30% decrease in Renilla luminescence, while analogues 1, 14, and 19 only showed modest enzyme inhibition. To further validate the specificity of these inhibitors for acting through the miRNA pathway, Huh7 cells were transfected with a psiCHECK-empty reporter (where the miR-122 binding site was replaced with a linker not targeted by any known miRNAs) and exposed to each analogue at 10 μM for 48 h. Parent compound 1 induced a 27% increase in the psiCHECK-empty assay, consistent with Rluc inhibition observed in the in vitro assay, and potential enzyme stabilization in cells (Table 1).

A complete scaffold change of the imidazole motif was also investigated. A series of analogues with pyridin-2-amines and anilines were synthesized. Both the unsubstituted pyridine ring (20) and methylation at the 2-position (21) showed 89% activity relative to the parent compound, while 3-methylpyridin-2-amine (22) elicited a significant increase in activity. Aniline (23) and 2-methylaniline (24) analogues showed a reduction in activity. Unfortunately, when assessed in the in vitro Rluc assay, 20–22 induced significant reductions in the luminescence signal, indicating they did not display enhanced miRNA-122 inhibition activity (Table 3).

Table 3.

Structures and Activities of Derivatives 20–24

compound	chemical structure	primary screen (RLU)a	Renilla luciferase assayb	psiCHECK-empty assayc
20		88 ± 1%	8%	--
21		89 ±11%	2%	--
22		135 ±4%	17%	--
23		47 ±3%	--	--
24		32 ±3%	-	-

^aRLU values represent the Renilla luciferase luminescence signal in the Huh7-miR122 stable cell line and are first normalized to firefly luciferase (dual reporter system) and then to the activity of the original hit compound 1.

^bIn vitro Renilla luciferase data are normalized to DMSO.

^cpsiCHECK-empty assay data are normalized to DMSO. Data for all assays represents the average ± standard deviation from at least three independent experiments.

The next modification made was methylation of the secondary amine in the sulfonamide. On inspecting in vitro Renilla luciferase assay data and analogues bearing imidazole and pyrazole motifs, we hypothesized that a tertiary sulfonamide would be important for minimizing Rluc inhibition because secondary sulfonamides form a subset of Rluc inhibitors.⁵⁰ To convert 20–22 to tertiary sulfonamides 25–27, these analogues were N-methylated to minimize further steric perturbation that might have had a negative impact on miR-122 inhibition. Compound 25 showed a modest reduction in activity, while 26 elicited a 7% increase compared to the parent compound. Unfortunately, 27 induced a disappointing 64% activity relative to 1, but all three compounds showed no Rluc inhibition, supporting our hypothesis (Table 4). Gratifyingly, replacement of the pyridine moiety in 26 and 27 with a benzene ring in compounds 28 and 29 yielded 104 and 122% activities, respectively. Compound 28 showed no Rluc inhibition, while 20% inhibition was observed with 29. Methylation of the aniline ring at the para position (30) yielded a modest decrease in activity compared to 1, while 2,3-dimethyl- (31) and 2,5-dimethyl- (32) aniline derivatives showed 80% activity relative to the parent compound. Unfortunately, while 30–32 showed no inhibition in the in vitro assay, all three compounds displayed >200% activity in the psiCHECK-empty assay, indicating that they increase luciferase activity by miRNA-independent mechanisms in cells. Furthermore, compounds 26 and 28 showed little to no inhibition in the psiCHECK-empty assay, consistent with Rluc inhibition observed in the in vitro assay (Table 4).

Table 4.

Structures and Activities of Derivatives 25–33

compound	chemical structure	primary screen (RLU)a	Renilla luciferase assayb	psiCHECK-empty assayc
25		89 ± 4%	100%	88%
26		107 ± 1%	100%	117%
27		64 ± 12%	100%	87%
28		104 ±7%	100%	90%
29		122 ± 1%	78%	103%
30		82 ±1%	100%	218%
31		80 ±4%	93%	225%
32		77 ± 1%	100%	245%
33		580 ± 25%	102%	178%

^aRLU values represent the Renilla luciferase luminescence signal in the Huh7-miR122 stable cell line and are first normalized to firefly luciferase (dual reporter system) and then to the activity of the original hit compound 1.

^bIn vitro Renilla luciferase data are normalized to DMSO.

^cpsiCHECK-empty assay data are normalized to DMSO. Data for all assays represents the average ± standard deviation from at least three independent experiments.

While 26 and 28 exhibited an excellent activity profile, poor solubility of the compounds presented a limitation of potential applications in animal studies. Because compound 28 elicited less of a response in the psiCHECK-empty assay, analogue 33 (Table 4), bearing a 5-pyridin-tetrazole moiety, was investigated. Installation of a carboxylic acid or corresponding isostere, such as tetrazole, has been shown to enhance the water solubility of small molecules.⁵¹ While the water solubility of compound 33 was not improved compared to 28, the compound elicited a near 6-fold increase in the luciferase signal relative to compounds 1 and 28 and no inhibition in the in vitro Rluc assay. Surprisingly, 33 induced an 80% increase in luminescence in the psiCHECK-empty assay compared to DMSO, suggesting it can activate luciferase in cells through an unknown mechanism (Table 4). Despite this observation, compound 33 exhibited a near 6-fold increase in activity over compound 1 in the primary luciferase assay, a magnitude of change judged unlikely to be the result of luciferase interference alone. As such, we proceeded to investigate 33 as a new and improved miR-122 inhibitor.

The potencies of the best analogues of 1 were assessed through dose–response of Huh7-miR122 cells (Figure S2). Compound 28 elicited an EC₅₀ of ~11 μM, while 33 showed an improved EC₅₀ value of ~4 μM. Next, the change in the level of mature miRNA upon treatment with compounds 28 and 33 was investigated using quantitative real-time polymerase chain reaction (RT-qPCR). Huh7 cells were exposed to a DMSO control (0.1%) or to the inhibitors at 25 μM for 48 h, total RNA was isolated using the miRNeasy kit (Qiagen), and RT-qPCR was performed in triplicate using TaqMan probes for miR-122 and RNU19 (control). The data was then normalized to the DMSO and RNU19 control using the 2^−ΔΔCt method.⁵² Compound 28 exhibited a 38% decrease in mature miR-122 levels, while analogue 33 elicited an 87% reduction in miR-122 expression, concomitant with their relative activity in the primary screen (Figure 2A).

Figure 2.

Compounds 28 and 33 reduce miR-122 levels. (A) Mature miR-122 levels in Huh7 cells were evaluated via RT-qPCR following 48 h treatment with compounds 28 and 33 (25 μM). Expression of miR-122 was normalized to a DMSO control. RNU19 expression was used as an internal control to account for variation between experiments. (B) HeLa-miR21 reporter cells were treated with 28 and 33 at 10 μM. After 48 h, luminescence was measured and normalized to a DMSO control. (C) E-cadherin expression was assessed in Huh7 cells via western blot following treatment with 28 (25 μM) or 33 (10 μM) for 48 h. GAPDH expression was monitored as an internal control. (D) Primary miR-122 levels were evaluated via RT-qPCR in Huh7 cells following 48 h treatment with compounds 28 and 33 (25 μM). Expression of pri-miR-122 following small-molecule treatment was normalized to a DMSO control. GAPDH expression was used as an internal control to account for variation between experiments. Data represent the averages ± standard deviations from three independent experiments. *p < 0.05 as determined by a Student’s t-test.

To begin elucidating the mode of action of compounds 28 and 33, several preliminary experiments were performed. A HeLa cell line that stably expresses a miR-21 reporter (HeLa-miR21)⁵³ was treated with 10 μM of each compound for 48 h. Following treatment, no increase in firefly luciferase activity relative to the DMSO control was observed, supporting the conclusion that the compounds are not general inhibitors of the miRNA pathway (Figure 2B) and are not firefly luciferase stabilizers. To investigate cellular effects of these inhibitors, expression of E-cadherin was monitored in Huh7 cells following treatment with 28 and 33. The expression of E-cadherin has a positive correlation with miR-122, which targets the Wnt signaling-associated genes Wnt1 and Snail that repress E-cadherin expression.^54,55 In Huh7 cells, E-cadherin expression is stabilized by overexpression of miR-122. As such, treatment with a miR-122 inhibitor would be expected to reduce E-cadherin protein levels. Treatment with DMSO alone elicited no inhibition of E-cadherin expression as expected, whereas reductions in E-cadherin levels were observed upon treatment with 28 and 33, indicating suppression of functional miR-122 (Figure 2C). Pri-miR-122 expression was assessed by RT-qPCR to determine if these compounds inhibited miR-122 transcription. Inhibitor 28 (25 μM) elicited a ~59% decrease in pri-miR-122 expression (Figure 2D), while 33 induced a more potent ~80% reduction, suggesting that both small molecules affect transcriptional or pretranscriptional regulation, rather than downstream steps of the miRNA pathway.

To investigate the possibility that the compounds directly inhibit miR-122 transcription, a reporter plasmid in which a firefly luciferase gene was placed under the control of the miR-122 promoter was developed (Figure 3A).⁵⁶ Briefly, the miR-122 promoter sequence was PCR amplified from Huh7 genomic DNA and ligated into a multicloning site upstream of the firefly luciferase gene in the pGL3-basic plasmid. As expected, almost no luminescence was observed for the control pGL3-basic reporter, while a significant level of luciferase expression was detected in Huh7 cells transfected with the pGL3-miR122 promoter construct (Figure S3A). Treatment with compounds 28 and 33 led to a reduction in miR-122 promoter activity (Figure 3B), consistent with the observed decrease in pri-miR-122 levels and a transcriptional mode of inhibition.

Figure 3.

Inhibitor assessment in miR-122 promoter assays. (A) Individual transcription factor binding sites within the miR-122 promoter sequence were mutated or deleted to attempt to identify the potential transcription factor target 28 and 33. (B) Huh7 cells were transfected with the parent pGL3-miR122 promoter reporter plasmid or reporter plasmids in which transcription factor binding sites were mutated/deleted and then treated with compounds 28 and 33 at 25 μM. After 48 h, a Bright-Glo assay was performed. Luciferase expression was normalized to cell viability and the DMSO control for each reporter. Data represent the averages ± standard deviations from three independent experiments.

After confirming that 28 and 33 inhibited the activity of the miR-122 promoter, an effort was made to identify the transcription factor(s) primarily affected by compound treatment. Liver-enriched transcription factors HNF1α, HNF3β, HNF4α, and HNF6 have been reported to regulate miR-122 transcription.^56–58 Additionally, the AP-1 transcription factor was predicted to bind to the miR-122 promoter.^59,60 To study the effects of 28 and 33 on each individual transcription factor, mutants of the parent pGL3-miR122 promoter reporter plasmid were prepared in which a single transcription factor binding site within the miR-122 promoter sequence was mutated or deleted (Figure 3B). Because there is overlap between HNF1α and HNF3β binding sequences,⁵⁷ a single mutant plasmid was generated for both. The mutated miR-122 promoter constructs were individually tested in Huh7 cells. All mutated constructs showed significant reductions in luciferase activity compared to the parent reporter (Figure S3B), suggesting that the interactions between the transcription factors and their corresponding promoters were abrogated by the introduction of mutations. As expected, mutation of the HNF1α/HNF3β and HNF4α binding sites led to a 70 and 60% decrease in promoter activity, respectively, indicating these transcription factors significantly contribute to miR-122 expression.⁵⁷ AP-1 and HNF6 mutants elicited only 30% reductions in activity (Figure S3B). While the modest impact of HNF6 on miR-122 promoter activity reflects previous results,⁵⁸ we have found no reports demonstrating the effect of AP-1 on miR-122. Treatment with 28 and 33 reduced luciferase activity in cells transfected with HNF1α/HNF3β, HNF6, and AP-1 mutant reporters, indicating that the compounds impact transcription independently of the mutated promoters (Figure 3B). However, in combination with the HNF4α mutant, 28 and 33 elicited a diminished reduction to ~64 and 59%, respectively, relative to the parent construct, indicating that the inhibitors may influence the function of HNF4α-driven transcription of the miR-122 gene.

Hepatocyte nuclear factor 4α (HNF4α) is a member of the nuclear receptor superfamily, a class of ligand-dependent transcription factors, and a key regulator of metabolic homeostasis and cell differentiation that is expressed predominantly in the liver, pancreas, kidney, and intestine.⁶¹ HNF4α is involved in liver cell homeostasis⁶² and is regulated by the endogenous fatty acid linoleic acid, which was found to occupy the ligand binding domain (LBD) of HNF4α in mammalian cells and the livers of mice.⁶³ The importance of ligand binding is illustrated by the observation that treatment with long-chain fatty acids, as well as recently discovered synthetic small molecules, has an inhibitory effect on the expression of HNF4α target genes.^64,65 Having identified inhibition of HNF4α transcriptional activity as a potential mechanism of action for our miR-122 inhibitors, this hypothesis was further examined using the previously identified small-molecule HNF4α inhibitor BI6015 (Figure 4A) as a positive control.⁶⁴ Intriguingly, this compound shares a similar sulfonamide motif with the inhibitors described in this study. The capacity for interaction of each inhibitor, 28, 33, and BI6015, with the LBD of HNF4α (PDB: 3fs1) was evaluated using AutoDock Vina.⁶⁶ The predicted bound conformations for each inhibitor within the LBD of HNF4α, superimposed with the fatty acid ligand, lauric acid, are shown in Figure 4B. Notably, in each conformation of 28, 33, and BI6015, polar contacts are predicted between substituents of the phenyl ring and S256. Interestingly, the 5-pryridin-tetrazole moiety of 33 interacts with R226, a residue that stabilizes the fatty acid ligand through H-bonding with the carboxylic acid group. Each of these identified interactions and the favorable predicted binding affinities, which were −7.9, −5.6, and −7.8 kcal/mol, for 28, 33, and BI6015, respectively, support putative binding of the inhibitors to HNF4α.^67,68 To experimentally validate direct binding of the inhibitors to HNF4α, the intrinsic fluorescence of HNF4αLBD (UniProt; P1235) of tyrosine and tryptophan residues was measured upon titration with each inhibitor.⁶⁹ The change in fluorescence intensity upon small-molecule binding was plotted versus inhibitor concentration and further analyzed using the Hill equation, producing K_D values of 832, 439, and 312 nM for inhibitors BI6015, 28, and 33, respectively (Figure 4C). Lastly, the functional effect of the inhibitors on HNF4α transcriptional activity was examined using a luciferase promoter assay. HEK293T cells, which do not endogenously express HNF4α,⁷⁰ were transfected with a reporter plasmid with the HNF4α DNA response element containing a promoter of CYP26A1 upstream of a firefly luciferase coding sequence,⁷¹ an HNF4α expression plasmid, and a Renilla luciferase control plasmid. Treatment of transfected cells with each of the inhibitors BI6015, 28, and 33 resulted in an approximate 50% decrease in HNF4α-driven luciferase expression relative to DMSO (Figure 4D). Taken together, the computational, biophysical, and cellular data presented here support the hypothesis that inhibitors 28 and 33 target miR-122 transcription through the direct inhibition of HNF4α.

Figure 4.

Compound interaction with HNF4α. (A) Structure of HNF4α ligands with docking visualization. BI6015 (left, turquoise), 28 (middle, green), and 38 (right, purple) in the LBD of HNF4α (PDB: 3fs1). (B) Direct binding to the HNF4α LBD was measured based on fluorescence quenching. Data represent the average ± standard deviations from two independent experiments. (C) HEK293T cells were transfected with the CYP26A1 promoter reporter plasmid, HNF4α expression plasmid, and Renilla luciferase reporter plasmid, then treated with compounds 28, 33, or BI6015 at 25 μM. After 48 h, a Dual Luciferase assay was performed. Relative luciferase expression was normalized to DMSO treatment. Data represent the averages ± standard deviations from three independent experiments. *p < 0.05 as determined by Student’s t-test.

Having characterized the miR-122 inhibitory activity and identified a mechanism of action of inhibitors 28 and 33, we next sought to evaluate the potential for these compounds to impact HCV replication. Interactions between miR-122 and the HCV RNA genome have been shown to promote viral RNA accumulation through several mechanisms, including promotion of viral translation through facilitation of IRES formation,^72,73 genome stabilization as a result of protection from exonuclease-mediated degradation,^27–29 and enhancement of viral RNA replication.^18,32,74 Because multiple aspects of the HCV life cycle depend on or are facilitated by interaction with miR-122, we hypothesized that inhibition of miR-122 transcription would have a significant effect on HCV infection (Figure 5A). Huh7.5 cells were pretreated with 28 and 33 (10 μM) or DMSO (negative control) for 1 h and then infected with HCV. After 48 h, total RNA was isolated, and HCV RNA was quantified using RT-qPCR. Inhibitors 28 and 33 elicited 88 and 90% reductions in viral RNA expression, respectively (Figure 5B). Furthermore, Huh7 cells treated with 28 or 33 for 48 h showed no reduction in cell viability at 10 or 25 μM (Figure 5C). The significant reduction in HCV RNA elicited by treatment with 28 and 33, without any apparent induction of cytotoxicity, suggests that small-molecule regulation of HNF4α function may be a promising approach for the development of new HCV therapies.

Figure 5.

Disruption of HCV replication. (A) Compound treatment results in inhibition of HNF4α-mediated miR-122 transcription, leading to reduced HCV mRNA levels. (B) Huh7.5 cells were pretreated with 28 and 33 at 10 μM and then infected with HCV. After 48 h, RT-qPCR was performed to evaluate HCV replication. Relative expression of HCV RNA for small-molecule-treated cells was normalized to a DMSO control and 18S ribosomal RNA expression. (C) Huh7 cells were treated with 28 or 33 for 48 h, then an XTT assay was performed to evaluate cell viability. All data are normalized to DMSO and represent the averages ± standard deviations from three independent experiments.

CONCLUSIONS

In summary, a new bis-arylsulfonamide class of small-molecule miR-122 inhibitors was identified from a high-throughput screen of >300,000 compounds. Several analogues were evaluated in comprehensive structure–activity relationship studies, demonstrating that 5-isopropyl, 2-methoxy, and 4-methyl substituents to the phenyl moiety were important for miR-122 inhibitory activity. Replacement of the imidazole moiety led to the identification of several pyridine and aniline derivatives with excellent activity. Unfortunately, many of these compounds were also identified as false-positives due to direct inhibition of the Renilla luciferase enzyme and thus its stabilization in cells through small-molecule binding. By implementing a series of secondary assays to evaluate Renilla luciferase activity in the presence of potential hit compounds, we established that a tertiary sulfonamide moiety was required to prevent luciferase inhibition—a concept that may be generally applicable to other sulfonamides identified in screens using Renilla luciferase reporters. Compound 28 was found to selectively inhibit miR-122 in the Huh7-miR122 reporter cell line, with no effect on Renilla luciferase activity in either biochemical or cell-based control experiments. Furthermore, 28 did not inhibit miR-21 in the HeLa-miR21 reporter cell line, indicating that it is not a general inhibitor of the miRNA pathway. In an attempt to improve aqueous solubility of 28, we synthesized analogue 33. Unfortunately, 33, and several other analogues that we synthesized (not shown), did not improve water solubility; however, the inhibitor displayed further increased activity. It was demonstrated that treatment of Huh7 cells with inhibitors 28 and 33 resulted in decreased E-cadherin expression detectable by western blot, an expected consequence of miR-122 inhibition. Compounds 28 and 33 were shown to reduce cellular miR-122 levels to 62 and 12% at 25 μM. Quantification of pri-miR-122 from inhibitor-treated cells showed significant inhibition of primary miRNA levels, suggesting that the inhibitors target transcription. This mode of action was further supported by the observed inhibitory effects of the two compounds in a miR-122 native promoter assay. Systematic removal of individual transcription factor recognition sites within the miR-122 promoter led to the discovery that the effectiveness of inhibitors 28 and 33 was correlated with the presence of an HNF4α recognition sequence. Molecular modeling and biophysical characterization of the inhibitors with HNF4α suggest that 28 and 33 are likely to inhibit HNF4α through direct binding to the transcription factor’s ligand binding domain. Through these studies, key H-bonding interactions in the ligand binding domain of HNF4α were identified, providing structural insight that may aid the future design of more effective inhibitors of this transcription factor. This hypothesis is supported by an additional HNF4α-dependent promoter assay demonstrating the inhibitory activities of 28 and 33 in comparison to a previously reported inhibitor of HNF4α. To our knowledge, this represents the first example of validated small-molecule inhibition of a transcription factor that controls miRNA expression and might serve as a general paradigm applicable to other miRNAs. Most importantly, both miR-122 inhibitors 28 and 33 blocked viral HCV RNA replication in human liver cells by ~90%, indicating they may have therapeutic potential in the treatment of HCV infection. MiR-122 presents an attractive target for treatment of HCV infections because, as a host component, it is much less susceptible to induction of drug resistance, and a corresponding therapeutic should be effective against all HCV genotypes,⁷⁵ in contrast to agents that directly target viral components.^76,77 Beyond applications in viral infection, small-molecule-mediated inhibition of miR-122 could help further elucidate its role in metabolism and may present novel opportunities for the treatment of metabolic disorders.⁷⁸

EXPERIMENTAL SECTION

Cell Culture.

Experiments were performed using Huh7, HEK293T, and HeLa-miR21⁴⁸ cell lines cultured in Dulbecco’s modified Eagle’s medium (DMEM; Hyclone) supplemented with 10% fetal bovine serum (FBS; Sigma-Aldrich) and maintained at 37 °C in a 5% CO₂ atmosphere. Huh7-miR122 cells⁴⁵ were cultured in DMEM (Hyclone) supplemented with 10% FBS (Sigma-Aldrich) and 500 μg/mL G418 (Sigma-Aldrich) and maintained at 37 °C in a 5% CO₂ atmosphere. Huh7.5 cells⁷⁹ were grown in Dulbecco’s modified high glucose media (Invitrogen) supplemented with 10% FBS (Atlanta Biologicals), 0.1 mM nonessential amino acids (Gibco), and 1% penicillin–streptomycin (Gibco) and maintained at 37 °C in a 5% CO₂ atmosphere.

Assay for Small-Molecule Inhibitors of miR-122.

Small-molecule screens were carried out using our previously described Huh7-miR122 reporter cell line (a stably transfected cell line expressing a miR-122 binding sequence in the 3′ UTR of a Renilla luciferase gene).⁴⁵ Cells were seeded at 15,000 cells per well in white, clear-bottom, 96-well plates (VWR). Following overnight incubation, cells were treated with the compound at the desired concentration or DMSO control (0.1% final DMSO concentration) in triplicate. The cells were incubated for 48 h, then analyzed for Renilla and firefly luciferase expression with a Dual Luciferase Assay Kit (Promega). Luminescence was measured on a microplate reader (Tecan M1000) with an integration time of 1 s and delay time of 5 s.

Assessment of the Selectivity of Small-Molecule Inhibitors of miR-122.

Selectivity of the small molecules for miR-122 was validated using our previously described HeLa-miR21 reporter cell line (a stably transfected cell line expressing a miR-21 binding sequence in the 3′ UTR of a firefly luciferase gene).⁵³ Cells were seeded at 10,000 cells per well in white, clear-bottom, 96-well plates (VWR). Following overnight incubation, cells were treated with the compound at the desired concentration or DMSO control (0.1% final DMSO concentration) in triplicate. The cells were incubated for 48 h, then analyzed for firefly luciferase expression with a Bright-Glo Luciferase Reporter Assay Kit (Promega). Luminescence was measured on a microplate reader (Tecan M1000) with an integration time of 1 s. Cells were also analyzed for cell viability using an XTT assay. For the XTT assay, XTT reagent was activated with 8 μL of menadione per 1 mL of XTT solution. Absorbance was measured immediately after addition of the reagent and again after 4 h on a microplate reader (Tecan M1000) at 450 and 630 nm (control). Cell viability was determined using the equation cell viability = (Absorbance_{450 nm} − Absorbance_{630 nm})_{4 h} − (Absorbance_{450 nm} − Absorbance_{630 nm})_initial. Cell viability was averaged and then used to calculate relative luminescence for each well using the equation relative luminescence = luminescence/cell viability. To assess inhibition of Renilla luciferase in cells, Huh7 cells were seeded at a density of 15,000 per well in white, clear-bottom, 96-well plates (VWR). Following overnight incubation, cells were transfected with 50 ng of psiCHECK-empty using Lipofectamine 2000 (Thermo Fisher). After 2 h incubation at 37 °C, media was replaced with DMEM supplemented with the compounds at 10 μM or DMSO (0.1% final DMSO concentration) in triplicate. The cells were incubated for 48 h followed by analysis with a Dual Luciferase Reporter Assay Kit (Promega). The luminescence was measured on a microplate reader (Tecan M1000) with an integration time of 1 s and a delay time of 5 s.

In Vitro Rluc Inhibition Assay.

The protocol was adapted from previous reports.⁸⁰ The experiment was performed in black, clear-bottom, 384-well plates (VWR) in triplicate. The compounds were diluted to a final concentration of 10 μM (0.1% final DMSO concentration) in phosphate-buffered saline (PBS), and recombinant Renilla luciferase enzyme (RayBiotech, 0.108 mg/L final Renilla luciferase enzyme concentration) was added to each well. Then, 15 μL of Renilla-Glo buffer (Promega) supplemented with the coelenterazine substrate (10 μL of substrate per 1 mL of buffer) was added to each well. The plate was incubated in the dark for 10 min, and then luminescence was measured on a microplate reader (Tecan M1000) with an integration time of 1 s.

Quantitative Real-Time PCR (RT-qPCR).

Huh7 cells were seeded at 125,000 cells per well in six-well plates (VWR). Following overnight incubation, cells were treated with the compound at 10 or 50 μM or DMSO (0.1% final DMSO concentration). The cells were incubated at 37 °C for 48 h (DMEM, 5% CO₂). Following removal of media, cells were washed with PBS buffer (1 mL, pH 7.4) followed by RNA isolation using the miRNeasy mini kit (Qiagen). Total RNA was quantified using a Nanodrop ND-1000 spectrophotometer. Fifteen nanograms of each RNA sample was reverse-transcribed using the TaqMan microRNA Reverse Transcription Kit (Thermo Fisher) in conjunction with either the miR-122 (Thermo Fisher; Assay ID: 002245) or RNU19 control (Thermo Fisher; Assay ID: 001003) Taqman reverse transcription (RT) primer (16 °C, 30 min; 42 °C, 30 min; 85 °C, 5 min). Quantitative real-time PCR was conducted with a TaqMan 2× Universal PCR Master Mix (Thermo Fisher) and an appropriate TaqMan miRNA assay (Thermo Fisher) on a BioRad CFX96 RT-PCR thermal cycler (1.3 μL of RT product; 95 °C, 10 min; followed by 40 cycles of 95 °C, 15 s; 60 °C, 60 s). The triplicate threshold cycles (Ct) obtained for each small-molecule treatment were used to determine the relative levels of miR-122 in small-molecule-treated cells relative to the DMSO control using the 2^−ΔΔCt method.⁵² The samples were also analyzed by RT-qPCR to measure the expression levels of pri-miR-122. After RNA isolation, 50 ng of each RNA sample was reverse-transcribed using the iScript cDNA synthesis kit (BioRad) (25 °C, 5 min; 42 °C, 30 min; 85 °C, 5 min). Quantitative real-time PCR was conducted with a TaqMan 2× Universal PCR Master Mix and the appropriate TaqMan primers for hsa-pri-miR-122 (Thermo Fisher; Assay ID: Hs03303072_pri) and GAPDH (Thermo Fisher; Assay ID: Hs02758991_g1) on a BioRad CFX96 RT-PCR thermal cycler (2 μL RT product) following the protocol described above. Triplicate threshold cycles (Ct) obtained for each small-molecule treatment were used to determine the relative levels of pri-miR-122 in small-molecule-treated cells relative to the DMSO control and normalized to the GAPDH control using the 2^−ΔΔCt method.⁵²

For HCV RNA quantitation, Huh7.5 cells were pretreated with either DMSO or specified small molecules (10 μM). Cells were then infected with HCV (MOI = 0.5) in DMEM with compounds. Five hours post-infection, the media was replaced with DMEM containing the small-molecule inhibitor. At 48 h post-infection, the supernatant was removed and cells were washed twice in PBS. RNA was extracted using the Rneasy 96 kit (Qiagen) and RNA copy number was assayed via RT-qPCR as previously described.⁸¹ 18S ribosomal and HCV RNA copy numbers were quantified via comparison to standards. HCV RNA was first normalized to 18S RNA and then to the normalized DMSO control.

Western Blot for E-Cadherin.

Huh7 cells were seeded at 125,000 cells per well in six-well plates (VWR). Following overnight incubation, cells were treated with the compound at 10 μM or DMSO (0.1% final DMSO concentration). The cells were incubated at 37 °C for 48 h (DMEM, 5% CO₂). Following removal of media, cells were washed with PBS buffer (1 mL, pH 7.4) followed by lysis. The samples were then boiled at 95 °C and analyzed on a 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gel (60 V for 20 min, followed by 150 V for 70 min). The proteins were transferred to a poly(vinylidene difluoride) (PVDF) membrane (80 V for 1.5 h in the transfer buffer). The membrane was then washed twice with ice-cold tris-buffered saline Tween-20 (TBST) and incubated with blocking buffer (5% bovine serum albumin (BSA) in TBST) for 1 h at RT. Next, the membrane was washed twice with ice-cold TBST. The membrane was incubated with polyclonal anti-E-cadherin antibody (1:1000 dilution; Cell Signaling Technology) or anti-GAPDH polyclonal antibody (1:1000 dilution; Cell Signaling Technology) at 4 °C overnight. The membranes were washed twice with TBST and incubated with a goat-anti-rabbit-IgG-horseradish peroxidase (HRP) (1:2500 dilution; Cell Signaling Technology) secondary antibody at RT for 1 h. The membranes were washed three times with TBST and then developed by HRP colorimetric staining using a SuperSignal West Pico Chemiluminescent Substrate (Thermo Fisher). The membranes were incubated with development reagents for 5 min at RT and imaged on a ChemiDoc XRS+ system (BioRad).

miR-122 Promoter Activity Assay.

The miR-122 promoter region was PCR amplified from Huh7 genomic DNA using primers containing the restriction sites for KpnI (forward primer, 5′ AAGGGGTACCGAATGCATGGTTAACTACGTCAG 3′; IDT DNA) and XhoI (reverse primer, 5′ AACCCCTCGAGCCTCCCGTCATTTCTCGGTC 3′; IDT DNA).⁵⁶ The insert was then cloned into the pGL3-basic plasmid (Promega) in front of the firefly luciferase gene using KpnI and XhoI restriction sites to generate the pGL3-miR122 promoter construct. Huh7 cells were seeded at 15,000 cells per well in white, clear-bottom, 96-well plates (VWR). Following overnight incubation, cells were transfected with the pGL3-mir122 promoter construct using Lipofectamine 2000 (Thermo Fisher) for 3 h. After transfection, media was replaced with DMEM and treated with the compound at the desired concentration or DMSO control (0.1% final DMSO concentration) in triplicate. The cells were incubated for 48 h, then analyzed for firefly luciferase expression with a Bright-Glo Luciferase Reporter Assay Kit. Luminescence was measured on a microplate reader (Tecan M1000) with an integration time of 1 s. Cells were also analyzed for cell viability using an XTT assay. For the XTT assay, XTT reagent was activated with 8 μL of menadione per 1 mL of XTT solution. Absorbance was measured immediately after addition of the reagent and again after 4 h on a microplate reader (Tecan M200) at 450 and 630 nm (control) and data were analyzed as described above. The mutated miR-122 promoter constructs were obtained via PCR amplification of the pGL3-basic-miR-122 promoter using the corresponding primers (Table S1). PCR reactions for each transcription factor contained pGL3-miR122 promoter (100 ng), 1 μM final concentration of forward and reverse primers, 200 nM dNTPs, 1× Phusion polymerase buffer (Thermo Fisher), and 2 U Phusion DNA polymerase (Thermo Fisher). PCR amplifications were performed on a T100 thermal cycler (BioRad) using the following program: 95 °C, 30 s; 30 cycles of 95 °C, 10 s; 45–60 °C, 30 s; 72 °C, 10 min; then 40–55 °C, 5 min; 72 °C, 30 min. Following PCR amplification, the products were digested with the DpnI restriction enzyme (NEB) and further incubated at 37 °C for 1 h. The DpnI digested products (2 μL) were then transformed in Mach 1 competent cells. Following confirmation of the mutant promoter plasmids by DNA sequencing, the effect of miR-122 inhibitor treatment on promoter activity was assessed as described above.

Protein Binding Assay.

An expression plasmid pE21 backbone containing the sequence of the ligand binding domain (aa 147–377) of HNF4α (UniProt; P41235) with a C-terminal His tag was purchased from Twist Biosciences. The protein fragment was recombinantly expressed and purified from E. coli as described previously.⁶⁹ Purified HNF4αLBD (200 nM) in PBS buffer (1 mL, pH 7.4) was titrated with increasing concentration of the compound. Fluorescence emission spectra were obtained at 25 °C using a Cary 5000 spectrofluorometer. Maximal emission was measured at 330 nm upon excitation at 280 nm and corrected for background (buffer and ligand alone).

CYP26A1 Promoter Assay.

HEK293T cells were seeded at 15,000 cells per well in a white flat-bottom 96-well plate. Following overnight incubation at 37 °C (DMEM, 5% CO₂), each well of cells was transfected with 50 ng of CYP26A1 promoter (Addgene; 135566), 50 ng of HNF4α expression (Addgene; 31100), and 5 ng of Renilla luciferase control (Promega) plasmids using FuGENE HD (Promega) for 4 h. Following transfection, media was replaced with 25 μM compound- or DMSO-containing media (0.25% final DMSO concentration). After 48 h treatment, cells were analyzed using Dual Luciferase Assay Kit (Promega). The luminescence was measured on a microplate reader (Tecan M1000) with an integration time of 1 s and a delay time of 5 s.

HNF4α-Inhibitor Docking.

The crystal structure of the ligand binding domain of HNF4α was obtained from the Protein Data Bank (PDB: 3fs1). Ligand (inhibitor SMILES string encoding) and receptor (crystal structure) files were prepared using the software package MGL Tools (Scripps Research Institute) to generate input files to AutoDock Vina.⁶⁶ The open-source molecular docking program AutoDock Vina was used to assess the binding of inhibitors to the ligand binding domain of HNF4α. A 20 Å search space around the center of mass of the fatty acid ligand was used, with an exhaustiveness variable of 20 and maximum number of binding modes of 10. The predicted binding energy of the lowest energy conformations was given in kcal/mol. The resulting conformations were visualized using the molecular visualization software PyMOL (Schrodinger).

Chemistry.

All reactions were performed in flame-dried glassware under a nitrogen atmosphere and stirred magnetically. Reactions were followed by thin layer chromatography (TLC) using glass-backed silica gel plates (Sorbent technologies, 250 μm thickness). Anhydrous tetrahydrofuran (THF) and acetonitrile were purchased from Acros, anhydrous dimethylformamide (DMF) was purchased from Alfa Aesar. Yields refer to pure compounds unless otherwise stated. Silica gel column chromatography was performed on silica gel (60 Å, 40–63 μm, 230 × 400 mesh, Sorbtech) as a stationary phase. High-resolution mass spectra (HRMS) analysis was performed on a Q-Exactive (Thermo Scientific) mass spectrometer. ¹H NMR and ¹³C NMR spectra were recorded on a 300 MHz or a 400 MHz Varian NMR spectrometer. Chemical shifts are given in δ units (ppm) for ¹H and ¹³C NMR spectra. All compounds are ≥95% pure according to HPLC analyses.

5-Isopropyl-2-methoxy-4-methylbenzenesulfonyl Chloride.

Potassium carbonate (1.4 g, 10.13 mmol, 0.8 equiv) and tetrabutylammonium bromide (2.2 g, 6.82 mmol, 0.5 equiv) were added to a solution of 4-isopropyl-3-methylphenol (2.0 g, 13.31 mmol, 1.0 equiv) in dimethyl carbonate (11 mL, 0.133 mol, 10 equiv) at room temperature. This suspension was refluxed for 36 h and monitored by TLC. After consumption of the reactant was confirmed, the reaction was quenched by adding water (20 mL) and then extracted with ethyl acetate (3 × 20 mL). The organic extracts were combined, washed with brine (20 mL), dried over sodium sulfate (20 g), and concentrated to obtain 1-isopropyl-4-methoxy-2-methylbenzene as a pale-yellow clear oil (2.2 g, 13.31 mmol, quantitative yield). ¹H NMR (300 MHz, CDCl₃) δ ppm 7.19 (d, J = 8.31 Hz, 1H), 6.71–6.81 (m, 2H), 3.81 (s, 3H), 3.12 (sep, J = 6.82 Hz, 1H), 2.36 (s, 3H), 1.24 (d, J = 6.80 Hz, 6H); ¹³C NMR (75 MHz, CDCl₃) δ ppm 157.37, 139.22, 136.42, 125.74, 115.92, 111.37, 55.26, 28.76, 23.57, 19.61; HRMS (electrospray ionization (ESI)) calcd for C₁₁H₁₇O (M + H)⁺ 165.1274, found 165.1275.

Chlorosulfonic acid (1.2 mL, 18.02 mmol, 3.0 equiv) was added dropwise to a solution of 1-isopropyl-4-methoxy-2-methylbenzene (1.0 g, 6.09 mmol, 1.0 equiv) in CHCl₃ (6 mL) at 0 °C. The reaction was stirred at 0 °C for 2.5 h. The reaction was slowly quenched into ice-water mixture (50 mL), extracted by multiple extracts of chloroform (3 × 20 mL), and combined. The combined extracts were washed with brine (20 mL), dried over sodium sulfate (20 g), and concentrated to yield 5-isopropyl-2-methoxy-4-methylbenzenesulfonyl chloride as a white solid (1.2 g, 4.57 mmol, 74% yield). ¹H NMR (400 MHz, CDCl₃) δ ppm 7.76 (s, 1H), 6.87 (s, 1H), 4.01 (s, 3H), 3.10 (sep, J = 6.85 Hz, 1H), 2.43 (s, 3H), 1.23 (d, J = 6.85 Hz, 6H); ¹³C NMR (101 MHz, CDCl₃) δ ppm 154.94, 146.31, 139.54, 129.71, 126.22, 115.02, 56.66, 29.03, 23.21, 20.31.

5-Isopropyl-2-methoxy-4-methyl-N-(o-tolyl)benzenesulfonamide (24).

Triethylamine (160 μL, 1.15 mmol, 1.5 equiv) was added to a stirred solution of o-toluidine (100 μL, 0.91 mmol, 1.2 equiv) in anhydrous THF (5 mL) at room temperature under N₂ atmosphere. 5-Isopropyl-2-methoxy-4-methylbenzenesulfonyl chloride (200 mg, 0.76 mmol, 1.0 equiv) was added to this solution and stirred under reflux overnight. Upon completion of reaction as confirmed by TLC, the reaction mixture was cooled to room temperature and concentrated under reduced pressure. The residue was purified by silica gel column chromatography and eluted with 20% ethyl acetate/hexanes to obtain 24 as a white solid (207 mg, 0.62 mmol, 82% yield). ¹H NMR (300 MHz, CDCl₃) δ ppm 7.66 (s, 1H), 7.17–7.24 (m, 1H), 6.92–7.13 (m, 3H), 6.74 (s, 2H), 3.94 (s, 3H), 3.02 (sep, J = 6.87 Hz, 1H), 2.34 (s, 3H), 2.25 (s, 3H), 1.15 (d, J = 6.80 Hz, 6H); ¹³C NMR (75 MHz, CDCl₃) δ ppm 153.71, 142.95, 139.64, 135.48, 130.85, 130.12, 127.43, 126.87, 125.22, 124.93, 121.76, 114.13, 56.37, 28.91, 23.20, 20.00, 17.69; HRMS (ESI) calcd for C₁₈H₂₄O₃NS (M + H)⁺ 334.1471, found 334.1491.

5-Isopropyl-2-methoxy-N,4-dimethyl-N-(o-tolyl)-benzenesulfonamide (28).

Sodium hydride (20 mg, 60% suspension in mineral oil, 0.50 mmol, 1.1 equiv) was added to a 0 °C precooled solution of 24 (150 mg, 0.45 mmol, 1.0 equiv) in anhydrous DMF (5 mL) under N₂ atmosphere and stirred for 30 min at 0 °C. Methyl iodide (40 μL, 0.64 mmol, 1.4 equiv) was added to the resultant suspension at 0 °C and the reaction mixture was warmed to room temperature. This solution was stirred overnight and quenched by ice-cold water (5 mL). The reaction mixture was diluted by additional 50 mL of deionized water and subsequently extracted by multiple washes of ethyl acetate (3 × 15 mL). The organic extracts were combined, washed with brine (15 mL), dried over sodium sulfate (20 g), and concentrated. The residual solid was loaded onto a column packed with silica gel and eluted with 10% ethyl acetate/hexanes to obtain 28 as a white solid (133 mg, 0.38 mmol, 85% yield). ¹H NMR (400 MHz, CDCl₃) δ ppm 7.49 (s, 1H), 7.22–7.25 (m, 1H), 7.15 (td, J = 7.46, 1.22 Hz, 1H), 6.94–7.02 (m, 1H), 6.80 (s, 1H), 6.68–6.72 (m, 1H), 3.91 (s, 3H), 3.28 (s, 3H), 2.96–3.09 (m, 1H), 2.38 (m, 6H), 1.09 (d, J = 6.85 Hz, 6H); ¹³C NMR (101 MHz, CDCl₃) δ ppm 154.41, 142.31, 140.48, 139.17, 138.94, 131.42, 128.57, 128.24, 128.20, 126.42, 125.58, 114.14, 56.04, 39.72, 28.85, 23.20, 19.98, 18.20; HRMS (ESI) calcd for C₁₉H₂₆O₃NS (M + H)⁺ 348.1628, found 348.1649.

N-(2-Hydroxy-3-(5-(pyridin-2-yl)-1H-tetrazol-1-yl)propyl)-5-isopropyl-2-methoxy-4-methyl-N-(o-tolyl)benzenesulfonamide (33).

(±)-Epichlorohydrin (90 μL, 1.15 mmol, 1.5 equiv) was added to a stirred suspension of 24 (250 mg, 0.75 mmol, 1.0 equiv) and potassium carbonate (210 mg, 1.52 mmol, 2.0 equiv) in anhydrous acetonitrile (10 mL) at room temperature under N₂ atmosphere. The resulting suspension was heated to 60 °C and stirred overnight. Upon disappearance of 24 as monitored by TLC, the reaction was quenched by the addition of deionized water (30 mL) and subsequently extracted by multiple washes of ethyl acetate (3 × 20 mL). The ethyl acetate extracts were combined, washed with brine (20 mL), dried over sodium sulfate (20 g), and concentrated. The residual solid was purified by silica gel column chromatography (eluted with 30% ethyl acetate/hexanes) to obtain 5-isopropyl-2-methoxy-4-methyl-N-(oxiran-2-ylmethyl)-N-(o-tolyl)benzenesulfonamide as a white solid (181 mg, 0.45 mmol, 62% yield). ¹H NMR (500 MHz, CDCl₃) δ ppm 7.43 (s, 1H), 7.23–7.24 (m, 1H), 7.15–7.18 (m, 1H), 6.99 (br s, 1H), 6.71–6.81 (m, 2H), 4.01–4.19 (m, 1H), 3.93 (s, 3H), 3.52–3.68 (m, 1H), 3.18–3.19 (m, 1H), 3.01 (sep, J = 6.84 Hz, 1H), 2.65–2.70 (m, 1H), 2.23–2.38 (m, 7H), 1.05–1.09 (m, 6H); ¹³C NMR (126 MHz, CDCl₃) δ ppm 154.14, 142.47, 139.52, 139.21, 131.27, 129.47, 128.60, 128.39, 126.31, 125.30, 114.14, 56.00, 55.23, 50.71, 49.97, 45.94, 28.70, 23.03, 19.81, 18.19; HRMS (ESI) calcd for C₂₁H₂₈O₄NS (M + H)⁺ 390.1734, found 390.1744.

5-Isopropyl-2-methoxy-4-methyl-N-(oxiran-2-ylmethyl)-N-(o-tolyl)benzenesulfonamide (124 mg, 0.32 mmol, 1.0 equiv) was added to a stirred suspension of 2-(1H-tetrazol-5-yl)pyridine, synthesized following a previously reported procedure⁸² (56 mg, 0.38 mmol, 1.2 equiv), and potassium carbonate (66 mg, 0.48 mmol, 1.5 equiv), in anhydrous DMF (5 mL) at room temperature under N₂ atmosphere. The resulting suspension was heated to 100 °C and stirred overnight. Upon completion of the reaction as monitored by TLC, deionized water (50 mL) was added to the reaction mixture and extracted with multiple washes of ethyl acetate (3 × 15 mL). The ethyl acetate fractions were combined, washed with brine (15 mL), dried over sodium sulfate (10 g), and concentrated. The residual solid was loaded onto a column packed with silica gel and eluted with 50% ethyl acetate/hexanes to yield 33 as a white solid (99 mg, 0.19 mmol, 58% yield). ¹H NMR (500 MHz, CDCl₃) δ ppm 8.68 (br s, 1H), 8.29–8.31 (m, 1H), 7.96 (q, J = 7.76 Hz, 1H), 7.48–7.54 (m, 1H), 7.34 (br d, J = 7.63 Hz, 1H), 7.12–7.22 (m, 2H), 6.92–7.04 (m, 1.5H), 6.76–6.80 (m, 1.5H), 5.23–5.41 (m, 1H), 4.91–5.00 (m, 1H), 4.25–4.29 (m, 0.5H), 4.14–4.23 (m, 1H), 4.01–4.07 (m, 1H), 3.86–3.94 (m, 4H), 3.71–3.75 (m, 0.5H), 2.94–3.01 (m, 1H), 2.44 (s, 1.5H), 2.34–2.38 (m, 3H), 2.23 (s, 1.5H), 0.99–1.06 (m, 6H); ¹³C NMR (126 MHz, CDCl₃) δ ppm 154.19, 142.81, 139.27, 138.45, 131.47, 130.01, 129.31, 128.91, 128.70, 126.52, 126.40, 125.94, 125.81, 128.36, 125.14, 124.35, 114.22, 69.71, 69.25, 56.77, 56.33, 56.13, 52.76, 52.52, 28.67, 23.03, 19.83, 18.59, 18.05; HRMS (ESI) calcd for C₂₇H₃₃O₄N₆S (M + H)⁺ 537.2278, found 537.2265.

ABBREVIATIONS USED

HNF4α

hepatocyte nuclear factor 4α

LBD

ligand binding domain

miRNA

microRNA

pri-miRNA

primary microRNA

Rluc

Renilla luciferase