R ESTORATION OF TUMOR SUPPRESSOR MI RNA S ACTIVITY
A few successful approaches have been employed to restore levels of tumor suppressor miRNAs, including (i) miRNAs mimics, (ii) miRNAs expression vectors and (iii) small-molecule activators
4.1.1 miRNAs mimics miRNAs mimics are non-natural double-stranded miRNA-like RNA fragments Such an RNA fragment is an oligonucleotide designed to have its 5’ end partially complementary to the selected sequence in the 3’-UTR of the target mRNA 56 Once introduced into cells, this RNA fragment, mimicking an endogenous miRNA, can bind specifically to its target mRNA thus regulating protein expression (a, Figure 1.9) In 2010, the Wiggins group has shown successful restoration of tumor suppressors miRNA-34a, leading to reduced cellular proliferation in the xenograft model of colon cancer 57 Following this study, other in vivo studies have been focused on the improvement of delivering systems into cells of miRNA mimics, such as a viral vehicle transporting miRNA-34a to the xenograft model mentioned above, 57 a neutral lipid emulsion in a murine model of non-small cell lung cancer (NSCLC) for the same miRNA, 58 a lentivirus to deliver miRNA-34a in an aggressive mouse model of NSCLC in lung cancer 59 In
In 2013, the first clinical trial of the miRNA-34a mimic (MRX34) in liver cancer patients demonstrated promising results and a safe profile This approach offers several advantages: miRNA mimics can be tailored to either repress translation or induce both repression and mRNA degradation, depending on their sequence complementarity with the target mRNA Additionally, miRNA mimics selectively target specific mRNAs, influencing only the corresponding proteins, unlike native miRNAs, which can affect multiple mRNAs with complementary sequences However, the lack of efficient delivery systems poses a significant challenge for the clinical advancement of many miRNA mimic candidates.
To enhance tumor suppressor miRNA levels, one effective strategy involves the use of miRNA expression vectors, which are designed with specific promoters for targeted expression in particular tissues or tumors For instance, research by Kota et al revealed that miRNA-26a, which is significantly down-regulated in hepatocellular carcinoma (HCC) cells, can be delivered systemically in a mouse model through an adeno-associated virus (AAV) This method effectively inhibited cancer cell proliferation and triggered tumor-specific apoptosis, leading to substantial tumor growth suppression without causing toxicity While clinical applications of this miRNA restoration approach remain unreported, AAV-based gene therapy shows considerable potential for developing targeted miRNA therapeutics.
Pre-miRNA miRNA duplex mature miRNA
Nucleus Cytoplasm miRNA mimics miRNA vectors
3'UTR target mRNA miRNA level increases to normal
Diminution of oncogene expression miRNA as tumor suppressor genes a. b.
To address the underexpression of tumor suppressor miRNAs, two key approaches can be employed: first, the application of oligonucleotides that mimic the activity of endogenous miRNAs; and second, the delivery of genes encoding miRNAs through viral vectors, specifically adenovirus-associated vectors (AAV).
Restoring miRNA levels can be achieved easily with synthetic oligonucleotides like miRNA mimics or expression vectors; however, using small molecules to enhance miRNA expression presents challenges Limited studies have identified small-molecule candidates that increase intracellular miRNA levels through high-throughput screenings, with two notable examples illustrated in Figure 1.10 These small molecules can non-specifically activate the transcription of miRNA genes, resulting in a global upregulation of miRNAs To date, there have been no reports on the rational design of small molecules aimed at specifically up-regulating individual miRNAs.
Figure 1.10 Chemical structure of compounds 1.1 and 1.2 acting as activators of miRNA production, discovered after the screening of chemical libraries
In conclusion, restoring tumor suppressor miRNAs presents a promising therapeutic strategy, although it is still in its early stages with limited successes, particularly highlighted by clinical trials of miRNA-34a in liver cancer patients The primary challenge for translating tumor suppressor miRNAs into effective clinical therapies lies in enhancing in vivo delivery systems.
I NHIBITION OF ONCOGENIC MI RNA S
Recent research has focused significantly on inhibiting overexpressed oncogenic miRNAs, contrasting with the restoration of tumor suppressor miRNAs Current strategies for this inhibition are primarily categorized into two approaches: oligonucleotide-based methods, which include antisense oligonucleotides, miRNA sponges, and miRNA masks, and small molecule-based approaches.
Oligonucleotide-based strategies focus on the direct inhibition of miRNA activity By designing oligonucleotides that are complementary to either the mature miRNA strand, known as antisense oligonucleotides or miRNA sponges, or to the targeted mRNA, referred to as miRNA masks, these approaches effectively block the functions of miRNAs.
Pre-miRNA miRNA duplex mature miRNA
3'UTR target mRNA miRNA as oncogene a Antisense oligonucleotides (ASOs)
Sponge constructs sponge construct b miRNA sponges
3'UTR target mRNA c miRNA masks
3'UTR target mRNA miRNA level decreases to normal
Increase of tumor suppressor gene expression miRNA/ASO duplex
Oligonucleotide-based strategies for inhibiting oncogenic microRNAs (miRNAs) include several innovative approaches Synthetic antisense oligonucleotides can effectively target mature miRNAs, resulting in their degradation or the formation of duplexes Additionally, the introduction of miRNA sponges into cells, which contain multiple tandem binding sites for specific miRNAs, leads to a reduction in the expression levels of these oncogenic miRNAs Furthermore, miRNA masks are designed to mimic the seed sequence of target miRNAs, allowing them to bind to the 3’-UTR of corresponding mRNAs, thereby inhibiting their function.
Antisense oligonucleotides (ASOs), also referred to as anti-miRNAs, serve as competitive inhibitors of microRNAs (miRNAs) The initial demonstration of miRNA functional inhibition using ASOs was conducted by Alexandre et al., targeting miRNA-2 and miRNA-13 in Drosophila However, these ASOs proved ineffective in other organisms like Caenorhabditis elegans due to their low stability Additionally, unmodified ASOs may lead to off-target effects by binding to similar-sequence miRNAs within the same family To address these challenges, researchers have modified the chemical structure of oligonucleotides, enhancing their stability, binding affinity, and specificity through various approaches, including 2’-position sugar modifications (such as 2’-OMe and LNA) and internucleotide linkage modifications like phosphorothioates.
Figure 1.12 Most common structural modifications employed in ASOs (B: nucleobase)
Research has shown that 2’-O-methyl-modified oligonucleotides serve as effective inhibitors, enhancing binding affinities for mature miRNAs and providing better nuclease resistance compared to unmodified ASOs Following this, Krutzfeldt et al developed cholesterol-conjugated, single-stranded RNA analogues with phosphorothiorate linkages, known as antagomirs, targeting the liver-abundant miRNA-122, which is linked to hepatitis C virus proliferation This approach demonstrated specific, efficient, and long-lasting silencing of endogenous miR-122 in in vivo liver cancer models in mice.
The most effective modified structure for targeting miRNA, which minimizes off-target effects, is the locked nucleic acid (LNA) LNAs feature a sugar ring locked by a methylene bridge, maintaining a C 3’-endo conformation LNA-anti-miRNAs with phosphorothioate linkages demonstrated enhanced cellular delivery and successfully silenced miRNA-122 at lower doses compared to cholesterol-conjugated analogues in a mouse liver cancer model Notably, Miravisen, a 15-mer LNA-DNA hybrid, has progressed to phase IIa clinical trials as an anti-miRNA-122 treatment for hepatitis C virus (HCV) infection This marks the first miRNA-targeted drug to enter human clinical trials, paving the way for new therapeutic options for chronic hepatitis C patients.
Many miRNAs belong to families that share a common seed sequence but may differ in other nucleotide regions miRNA sponges function similarly to antisense oligonucleotides (ASOs) by base-pairing with specific miRNAs to inhibit their activity Unlike ASOs, which target a single miRNA, miRNA sponges feature multiple binding sites for a miRNA seed sequence, allowing them to effectively target all members of a miRNA family, marking a significant advancement Recent studies have demonstrated the efficacy of miRNA sponges in various cell lines, including non-small cell lung cancer, embryonic neural stem cells, and B cell lymphoma.
The core mechanism of microRNA (miRNA) function involves the base-pairing interaction between miRNAs and their target mRNAs This interaction can be modified by either miRNAs or mRNAs serving as targets The miRNA mask approach utilizes antisense oligonucleotides (ASOs), which differ from antimiRs by being single-stranded 2’-O-methyl-modified oligonucleotides that do not directly interact with the miRNA Instead, these masks are complementary to the binding site of the miRNA within the 3’-UTR of the target mRNA By blocking access to this binding site, miRNA masks facilitate a miRNA mask/mRNA interaction, effectively altering the typical miRNA/mRNA interaction.
Forty gene protectors have been identified that inhibit the effects of miRNA in a sequence-specific manner This method has been effectively utilized in a zebrafish model to counteract the repressive effects of miRNA-430 on transforming growth factor-β signaling pathways.
In conclusion, the oligonucleotide strategy demonstrates that targeting specific miRNAs can achieve significant biological effects Several miRNA candidates have advanced into product and clinical development; however, challenges remain in delivering these non-small molecules due to suboptimal pharmacodynamic and pharmacokinetic properties for therapeutic use To enhance therapeutic efficacy and specificity, improvements in the chemical design of oligonucleotides and the development of efficient delivery systems are essential.
To address the challenges associated with oligonucleotide therapy, researchers have developed indirect strategies that involve using drugs to modulate miRNA expression by targeting their transcription and processing The identification of small-molecule drugs that specifically target miRNAs and alter their functions holds significant promise, as these drugs could be pivotal in creating targeted cancer therapies and treatments for other diseases related to miRNA dysregulation.
The biogenesis of miRNA is a complex multi-step process that begins in the nucleus and culminates in the formation of a mature miRNA duplex in the cytoplasm Each stage of this process can be manipulated, potentially altering the biogenesis and function of miRNAs Throughout their development, miRNA molecules adopt structural forms rich in secondary structures, including stem-loop structures and internal loops, which are essential for recognition by the microprocessor and Dicer enzyme These unique structures facilitate selective binding interactions with small-molecule ligands.
Figure 1.13 Four general classes of RNA secondary structure
An overview of the potential targets for small molecules intervention to activate or inhibit miRNAs inside cells is illustrated in Figure 1.14 (A-E) along with the multi-step miRNA biogenesis pathway
The miRNA biogenesis pathway is illustrated in Figure 1.14, highlighting each step involved in the formation and function of miRNAs The diagram also indicates which steps have been influenced by small molecules, marked with letters A-E, showcasing the potential for targeted modulation in miRNA activity.
A: transcription inhibition B: Drosha-cleavage inhibition
C: Dicer cleavage inhibition D: RISC complex formation inhibition
Recent reviews have highlighted the targeting of miRNAs using small molecules, showcasing numerous examples that affect transcription, Drosha or Dicer cleavage, and the formation of the RISC complex with mRNA However, a significant challenge in developing small molecules for miRNA pathway inhibition is the difficulty in designing compounds that are specific to particular RNA sequences Most current examples have emerged from random screenings of extensive chemical libraries or from known RNA ligands like aminoglycoside antibiotics This section will focus on reported instances of small-molecule drug discovery aimed at selectively interfering with oncogenic miRNAs, utilizing strategies such as high-throughput screening (HTS) of large compound libraries, focused screening of smaller libraries of RNA-interacting molecules, and the design of small molecules based on the secondary structure of targeted RNA or the three-dimensional structures of relevant enzymes.
Initial research on small molecules targeting oncogenic miRNAs primarily utilized high throughput screenings (HTS), which included both cell-free methods employing fluorescence for detecting small-molecule binding to RNA targets and cell-based assays for screening chemical libraries to identify miRNA modulators Many of these assays relied on luciferase reporter systems, which have been comprehensively reviewed While several innovative screening methodologies have been validated recently, they have yet to be applied to actual libraries This section will detail HTS that resulted in the identification of small molecules that can (i) inhibit miRNA transcription, (ii) block Drosha or Dicer cleavage, and (iii) prevent RISC loading.
(i) Discovery of small molecules inhibiting miRNAs transcription
C HOICE OF ANTIBIOTICS
To investigate the impact of rRNA-binding agents from various antibiotic classes on miRNA biogenesis, we focused on selected representatives from each class Aminoglycosides were prioritized due to their significance as established bacterial ribosomal RNA ligands These antibiotics have been extensively researched for their capacity to bind to other biologically relevant RNA targets, including viral RNAs such as HIV-1's TAR and DIS RNAs, showcasing their effectiveness as non-selective RNA binders Aminoglycosides are oligosaccharides characterized by multiple ammonium groups that attach to the prokaryotic ribosomal acceptor site (A-site) on the 30S subunit, which is essential for the translation process.
Figure 2.1 Antibiotic target sites during bacterial protein synthesis (after Ref 136)
Natural products derived from Streptomyces predominantly feature the conserved 2-deoxystreptamine (2-DOS) aminocyclitol, which is glycosylated with aminosugars at various positions The amine groups on the aminoglycosidic scaffold exhibit diverse pKa values, primarily remaining protonated at physiological pH, thereby enhancing their RNA binding through electrostatic and hydrogen bonding interactions Due to their highly cationic nature and conformational flexibility, aminoglycosides are effective universal RNA binders In our study, we focused on screening the activity of several specific 2-deoxystreptamines, including 4,5-linked variants like neomycin B and paromomycin, 4,6-linked types such as kanamycin B and gentamycin, streptamine-based compounds like streptomycin and spectinomycin, as well as the 6-linked 2-deoxystreptamine apramycin and neamine, a component of neomycin B.
(red part) NH 2 OH NH 2 H
Figure 2.2 Chemical structures of selected aminoglycosides
Tetracycline derivatives represent a unique class of antibiotics that, unlike aminoglycosides, have been overlooked for their interactions with RNA structures beyond rRNA These molecules feature a linear fused tetracyclic scaffold with various functional groups and inhibit bacterial protein synthesis by binding near the ribosomal A-site, which prevents the association of aminoacyl-tRNA Structural studies indicate that the tetracycline-rRNA complex primarily forms through electrostatic and hydrogen bonding interactions with the phosphate-oxygens of the rRNA backbone, lacking sequence specificity Doxycycline, oxytetracycline, and minocycline serve as key examples of this antibiotic class.
Figure 2.3 Chemical structures of selected tetracycline derivatives
During the final stages of protein synthesis, the newly formed peptide chain exits the ribosome through a dynamic tunnel, but macrolide antibiotics, such as erythromycin and azithromycin, interfere with this process They bind to the large subunit 23S rRNA near the peptidyl transferase center, preventing the release of the nascent protein and inhibiting ribosomal subunit assembly, which leads to nucleolytic degradation Similarly, lincosamides like lincomycin and clindamycin operate through a comparable mechanism, disrupting the peptidyltransferase reaction on the 50S ribosomal subunit.
We selected three additional antibiotic compounds for screening, starting with linezolid, a synthetic oxazolidinone antibiotic introduced in the late 1990s Oxazolidinones function by inhibiting protein synthesis through binding to the P-site of the ribosomal 50S subunit.
Chloramphenicol, one of the first naturally occurring antibiotics to be synthesized, along with puromycin, which likely acts as a false substrate in the A-site, highlights the diversity of antibiotic mechanisms While the precise action of puromycin remains unclear, these antibiotics differ significantly from earlier classes, potentially enhancing our understanding of how various RNA ligands can inhibit the processing of oncogenic pre-miRNAs.
Figure 2.4 Chemical structure of antibiotics belonging to the classes of macrolides, lincosamides, oxazolidinones and others
This study focuses on 18 chemically diverse compounds that share a common mechanism of action by targeting rRNA The research aims to identify structural characteristics and structure-activity relationships to facilitate the discovery of potential inhibitors for oncogenic miRNAs.
S CREENING OF ANTIBIOTICS FOR THEIR ABILITY TO INHIBIT ONCOGENIC MI RNA S PRODUCTION
The 18 antibiotics described above were screened for their ability to inhibit the processing by Dicer of four different pre-miRNAs In addition to the two pre-miRNA-372 and pre-miRNA-373 that were mentioned above (Figure 1.23), we decided to screen two other oncogenic pre- miRNAs in order to study either selectivity or activity of these compounds on specific targets Thus, we chose the precursor of miRNA-17 that is highly expressed in gastric cancer tissues and was identified as oncogenic in several other types of cancers (Figure 2.5) 141 as well as pre- miRNA-21 that leads to miRNA-21 overexpressed in breast, lung, colon and prostate cancers (Figure 2.5) 142
Figure 2.5 Sequence and secondary structure of pre-miRNA-17 and pre-miRNA-21 as reported in miRBase
Screening of four RNA sequences was conducted using FRET-based experiments, employing a cell-free assay based on established procedures In this assay, targeted pre-miRNAs were labeled with a fluorophore (fluorescein or FAM) and a quencher (dabcyl or DAB) at their 5’ and 3’ ends, respectively The recombinant Dicer enzyme cleaves the RNA, leading to the detection of fluorescence However, if an RNA ligand effectively binds to the structured pre-miRNA and inhibits Dicer's cleavage, fluorescence is not observed.
Figure 2.6 In vitro FRET-based assay developed for the screening of potential inhibitors of Dicer cleavage of pre-miRNAs
In a preliminary screening of 18 compounds at a concentration of 200 μM, only aminoglycosides and tetracyclines demonstrated the ability to inhibit Dicer cleavage of pre-miRNA-372, as shown in Figure 2.7 In contrast, macrolides, lincosamides, linezolid, puromycin, and chloramphenicol were ineffective in this regard.
1.0 pre-miRNA with Dicer Neomycin Apramycin Streptomycin Paromycin Kanamycin Spectipmycin Neamine pre-miRNA without Dicer
1.0 pre-miRNA with Dicer Oxytetracycline Minocycline pre-miRNA without Dicer
1.0 pre-miRNA with Dicer Erythromycin Azythromycin Clindamycin Lincomycin Chloramphenicol Puromycin pre-miRNA without Dicer
Figure 2.7 illustrates the fluorescence increase observed when 0.25U recombinant Dicer is incubated with 50 nM beacon pre-miRNA-372, comparing results in the absence of antibiotics (red line) and with 200 μM of each antibiotic The fluorescence of pre-miRNA without Dicer and antibiotics is shown in purple The effects of different antibiotic classes are categorized as follows: (A) aminoglycosides, (B) tetracyclines, and (C) macrolides, lincosamides, linezolid, puromycin, and chloramphenicol.
The analysis of four RNA sequences revealed that neomycin and apramycin were the most effective aminoglycoside inhibitors, achieving inhibition rates of 60-80% In contrast, kanamycin, streptomycin, and spectinomycin exhibited minimal to no inhibitory activity, while paromomycin, neamine, and gentamycin demonstrated moderate inhibition ranging from 10-60% Among the tetracyclines, minocycline emerged as the top inhibitor with 30-70% effectiveness, followed by doxycycline at 20-40% and oxytetracycline at 10-30%.
In a study assessing the effectiveness of rRNA-targeting antibiotics on Dicer cleavage of various pre-miRNAs, the inhibition percentages for aminoglycosides and tetracyclines were measured The inhibition was quantified based on the fluorescence increase of 5’-FAM, 3’-dabcyl-pre-miRNA beacons when incubated with 0.25U recombinant Dicer and 200 μM of each antibiotic The baseline fluorescence of pre-miRNAs in the presence of Dicer was set at 0% inhibition (100% fluorescence) Results included error bars indicating standard deviation from three independent experiments conducted in duplicates.
The IC50 evaluation experiments confirmed these observations, as detailed in Table 2.1, which reports the inhibition of cleavages for pre-miRNA-372, pre-miRNA-373, pre-miRNA-17, and pre-miRNA-21 These evaluations utilized a FRET-based assay, consistent with previous methods, and involved testing a range of 12 different concentrations for each compound.
Table 2.1 IC 50 values calculated on pre-miR-372, pre-miRNA-373, pre-miRNA-17 & pre-miRNA-21
IC 50 (àM) a Pre-miR-372 Pre-miR-373 Pre-miR-17 Pre-miR-21
Kanamycin B nd nd nd nd
Spectinomycin nd nd nd nd
Erythromycine nd nd nd nd
Azythromycine nd nd nd nd
Clindamycin nd nd nd nd
Lincomycin nd nd nd nd
Oxazolidinones Linezolid nd nd nd nd
Others Puromycin nd nd nd nd
In IC50 experiments, chloramphenicol was tested alongside 50 nM of pre-miR-372, -373, -17, and -21 beacons, and 0.5U of recombinant Dicer in buffer A, which consists of 20 mM Tris-HCl (pH 7.4), 12 mM NaCl, 2.5 mM MgCl2, and 1 mM DTT The IC50 values were not determined due to the absence of inhibition within the tested concentration range, with reported values carrying an uncertainty of ± 5%.
Neomycin has been identified as the most effective inhibitor of pre-miRNA-372, with an IC50 value of 88.2 μM, followed by apramycin at 162 μM Other tested compounds exhibited IC50 values greater than 200 μM or showed no activity For pre-miRNA-17, neomycin again proved to be the best inhibitor at 93.7 μM, followed by minocycline at 129 μM, while apramycin was inactive against this RNA A similar pattern was observed for pre-miRNA-21.
Neomycin emerged as the most effective inhibitor of Dicer cleavage, with an IC50 of 143 μM, followed closely by gentamycin at 159 μM The IC50 values for various RNA sequences indicate that different compounds exhibit varied inhibitory effects depending on the RNA type For instance, apramycin demonstrated greater efficiency against pre-miRNA-372 (IC50 = 162 μM) and pre-miRNA-21 (IC50 = 170 μM), while showing no inhibition for pre-miRNA-17 and a higher IC50 of 335 μM for pre-miRNA-373 Minocycline displayed selective inhibition, particularly for pre-miRNA-17 (IC50 = 129 μM), compared to other sequences Streptomycin showed significant inhibition primarily for pre-miRNA-372 and pre-miRNA-21 Overall, neomycin consistently proved to be the most potent inhibitor across different RNA sequences, with IC50 values ranging from 76.7 μM for miRNA-21 to 143 μM for miRNA-373 These findings highlight the potential of aminoglycosides and tetracyclines to inhibit Dicer cleavage, revealing a degree of selectivity for specific RNA structures Further investigation into the binding affinity of these compounds to pre-miRNA sequences is warranted to better understand their mechanism of action.
M EASUREMENT OF BINDING AFFINITY ( DISSOCIATION CONSTANTS , K D )
The inhibition of Dicer cleavage of specific pre-miRNAs is influenced by the binding of compounds to the RNA sequence, suggesting a correlation between IC50 values and binding affinity for each targeted RNA To investigate whether differences in binding affinity contribute to variations in the inhibitory activity of these antibiotics, we assessed the affinity of all compounds for monolabeled pre-miRNA fragments labeled with a fluorescein fluorophore at the 5’-end, without the presence of the Dicer enzyme By evaluating fluorescence changes in relation to compound concentration, we were able to determine the dissociation constant (K D) values.
Figure 2.9 Fluorescence-based assay for the evaluation of the dissociation constant (K D ) as a measure of the binding affinity of each RNA ligands for targeting pre-miRNAs
Table 2.2 reports the K D values for each compound in the case of pre-miRNA-372, pre-miRNA-
373, pre-miRNA-17 and pre-miRNA-21 Concerning aminoglycosides, we observed that the
The study identifies neomycin and apramycin as the most effective inhibitors for pre-miRNA-372, while neomycin and minocycline are the top inhibitors for pre-miRNA-17, highlighting the role of RNA binding in their inhibition mechanisms However, a direct correlation between dissociation constant (K D) and half-maximal inhibitory concentration (IC50) was not established, as some highly effective ligands, like gentamycin and apramycin, did not exhibit the best inhibitory performance despite their favorable K D values.
Tetracyclines exhibit strong RNA binding capabilities, with dissociation constants (K D) ranging from low micromolar to nanomolar, making minocycline the most effective ligand (K D = 597 - 941 nM), followed by doxycycline and oxytetracycline Despite their strong RNA affinity, tetracyclines are less effective than aminoglycosides in inhibiting Dicer cleavage Additionally, compounds like macrolides (erythromycin and azithromycin) and puromycin also demonstrate good RNA binding without inhibiting Dicer cleavage, similar to the behavior observed with pre-miRNA-17, pre-miRNA-373, and pre-miRNA-21.
Table 2.2 Dissociation constants (K D ) calculated on pre-miRNA-372, pre-miRNA-373, pre-miRNA-17 and pre-miRNA-21
K D (àM) a Pre-miR-372 Pre-miR-373 Pre-miR-17 Pre-miR-21
Spectinomycin nb nb nb nb
Clindamycin nb nb nb nb
Lincomycin nb nb nb nb
Oxazolidinones Linezolid nb nb nb nb
Chloramphenicol nb nb nb nb a Binding studies were performed on 5’-FAM-pre-miR-372, -373, -17, -21 in buffer A K D values are given with an uncertainty of ±
From a general perspective, there were no significant differences in the K D values of the antibiotics across the four RNA sequences, indicating a lack of selectivity as they bind to various RNA structures Nonetheless, variations were noted in the inhibition activity of Dicer processing.
71 highlighting the fact that other parameters than RNA affinity play a role in the mechanism of action of these compounds
To evaluate the selectivity of antibiotics for specific nucleic acid structures, binding assays were conducted using potential competitors The study focused on antibiotics with pre-miRNA binding affinity, measuring their specificity in the presence of a significant excess of tRNA or dsDNA For instance, when assessing pre-miRNA-372, neomycin, paromomycin, and gentamycin exhibited specificity ratios of 1, indicating high selectivity, while other aminoglycosides and tetracyclines showed ratios ranging from 2 to 3, with minocycline being the least specific at a ratio of 3.2 Similar tests with dsDNA confirmed that neomycin, paromomycin, and gentamycin remained the most selective, again with a ratio of 1, while other antibiotics displayed ratios from 2 to 5, with erythromycin and azithromycin being the least selective at ratios of 4.5 and 5, respectively Overall, the findings indicated that most tested compounds preferentially bind to pre-miRNA stem-loop structures over other nucleic acid forms like tRNA and DNA.
Table 2.3 Competition assays performed with pre-miR-372 in the presence of tRNA and dsDNA a
Class Antibiotic K D (àM) K ’ D with tRNA (àM) b K ’ D /K D K ’ D with dsDNA (àM) c K ’’ D /K D
Fluorescence measurements were conducted on pre-miRNA-372 in buffer A, which consists of 20 mM Tris HCl (pH 7.4), 12 mM NaCl, and 3 mM MgCl2 The KD values obtained have an uncertainty of ± 10% Additionally, measurements were taken in the presence of a 100-fold excess of a natural tRNA mixture and a 100-fold excess of a 15-mer DNA.
In summary, the analysis of binding affinities revealed that while the most effective RNA ligands are also the top inhibitors of pre-miRNA processing by Dicer, a direct correlation between K D and IC50 could not be established Additionally, the same compound exhibited varying selectivity of inhibition based on the RNA type, yet these variations did not correlate with binding affinity This indicates that other factors must be considered to fully understand the observed inhibitory effectiveness of these compounds.
E NZYMATIC FOOTPRINTING AND IDENTIFICATION OF THE INTERACTION SITES
The lack of correlation between IC50 and K D can be attributed to two potential explanations Firstly, the inhibition of Dicer may not solely result from RNA binding but rather from a direct interaction with Dicer itself; however, this hypothesis is unlikely since IC50 values would remain consistent across different RNA sequences if that were the case Secondly, the variation in the interaction sites of each ligand on the RNA target could lead to differing levels of inhibition on Dicer's binding and cleavage activities.
To understand the molecular basis of varying inhibition efficacy, we investigated the interaction sites of ligands on pre-miRNA-372 through enzymatic footprints Our focus was on neomycin, a potent Dicer cleavage inhibitor, and minocycline, which, despite being a strong RNA ligand, shows limited inhibitory efficacy Enzymatic footprinting results revealed that neomycin inhibits RNase V1 cleavage at specific sites, including C27 and G16, while no inhibition was observed with RNase S1, indicating that neomycin primarily binds in the stem region of pre-miRNA-372, particularly near the bulge at G16 and the loop at C27 This binding pattern aligns with previous findings that neomycin preferentially interacts with stem-loop RNAs at distorted double helix regions Additionally, fluorescence intensity analysis demonstrated the differences between uncomplexed pre-miRNA-372 and its complex with neomycin.
The RNase footprinting analysis of pre-miRNA-372 reveals its sequence and secondary structure, as illustrated in Figure 2.10 The interactions between pre-miRNA-372 and neomycin were assessed using RNase V1 and RNase S1 probing techniques The gel results show intact RNA in lane 1, with lanes 2 and 3 displaying the alkaline hydrolysis ladder and T1 digestion ladder, respectively Lane 4 depicts the cleavage pattern of uncomplexed pre-miRNA-372, while lanes 5 to 7 illustrate the cleavage patterns of pre-miRNA-372 when complexed with increasing concentrations of neomycin (1 µM to 100 µM) Additionally, a representative plot highlights the intensity of bands for uncomplexed pre-miRNA-372 compared to its neomycin-complexed forms, with a focus on nucleotides 10-18 in the zoom box.
Minocycline exhibited a unique behavior in the presence of RNase V1, showing no significant footprint However, it demonstrated a clear inhibition of RNase S1 cleavage at U31, G32, and U33, indicating that minocycline interacts solely within the loop region, unlike neomycin.
RNase footprinting analyses were conducted to examine the interaction between pre-miRNA-372 and minocycline, utilizing RNase V1 and RNase S1 probing techniques The gel results display various lanes: lane 1 shows intact RNA, lanes 2 and 3 depict the alkaline hydrolysis ladder and T1 digestion ladder, respectively, while lane 4 illustrates the cleavage pattern of uncomplexed pre-miRNA-372 Lanes 5 to 7 reveal the cleavage patterns of pre-miRNA-372 when complexed with increasing concentrations of minocycline (1 µM, 10 µM, 50 µM, and 100 µM) Additionally, a representative plot demonstrates the intensity of bands for uncomplexed pre-miRNA-372 (black line) compared to pre-miRNA-372 complexed with minocycline (color-coded from yellow to red) in the presence of RNase S1.
23-37 are represented in the zoom box
Molecular docking studies revealed the interaction site of neomycin, primarily engaging with residues near the G16 bulge through electrostatic and hydrogen bonding interactions, supporting findings from enzymatic footprints Notably, the distinct interaction sites of neomycin and minocycline may account for their differing effects on Dicer cleavage inhibition Cryo-electron microscopy and crystallographic studies depict human Dicer as an “L”-shaped enzyme, emphasizing the significance of these interactions in understanding its function.
The L-shaped structure features the closely linked PAZ and Platform domains, which accommodate the binding pockets for the 3’-overhang and 5’-phosphate regions of a dsRNA substrate At the upper section of the L, the RNase III domain forms an intramolecular dimer, creating the catalytic core A linker separates the RNase III catalytic domain from the dsRNA-end binding domains, stabilizing the dsRNA backbone phosphate residues.
The protein 76 acts as a ruler by aligning the pre-miRNA substrate with the catalytic center, which is crucial for determining where Dicer cleaves Compounds that can bind to the ruler's binding site or the catalytic site, such as neomycin, are likely to be more effective in inhibiting Dicer cleavage compared to those that bind to the loop region, like tetracyclines, where fewer interactions necessary for Dicer cleavage occur.
Neomycin was docked with the pre-miRNA-372 hairpin loop using Autodock 4, which involved fixing the grid boxes on the entire RNA sequence Additionally, a schematic representation illustrates the L-shaped Dicer enzyme (EMD-1646 150) positioned around pre-miR-372, with the catalytic core indicated by a black circle and the linker denoted by a square bracket.
D ICER CLEAVAGE INHIBITION OF PRE - MI RNA-372
To validate the hypothesis, the inhibition of Dicer's processing of pre-miRNA-372 was examined using denaturing gel electrophoresis As shown in Figure 2.13, Dicer cleaves pre-miRNA-372, resulting in a 24-mer (A24) and a 42-mer (A42), which correspond to the cleavage of two adjacent phosphodiester bonds on opposite strands of the stem Although the precise molecular mechanism of Dicer is still being investigated, it is well established that the enzyme cleaves dsRNA to produce fragments with 2nt 3’-protruding ends The cleavage events observed at A24 and A42 align with these findings.
Neomycin effectively inhibits Dicer cleavage at positions A24 and A42, while a new cleavage site at U28 emerges, indicating that although neomycin has an IC50 of 88 µM, it does not completely inhibit Dicer, allowing for cleavage at U28 This partial inhibition may account for neomycin's higher IC50 compared to other pre-miRNA-372 ligands In contrast, minocycline did not inhibit Dicer cleavages at the tested concentration, likely due to its inefficient binding in the loop region, which fails to prevent Dicer processing of this pre-miRNA.
The gel image in Figure 2.13 illustrates the products formed during Dicer-catalyzed pre-miRNA processing, comparing control reactions (lane 4) with those inhibited by increasing concentrations of neomycin and minocycline (lanes 5-8) Lane 1 shows intact RNA, while lanes 2 and 3 depict the alkaline hydrolysis ladder and T1 digestion ladder, respectively Dicer cleavage sites are marked by black arrows, highlighting specific nucleotides at positions A24, U28, and A42 Additionally, the presence of less specific cleavage sites suggests that the fluorescein label at the 5’-end of pre-miRNA may hinder Dicer's recognition, leading to non-specific cleavage.
The findings highlight the critical role of the interaction site of a pre-miRNA ligand in determining its effectiveness as an inhibitor of miRNA biogenesis, especially during the Dicer processing step.
T HERMODYNAMIC SIGNATURE OF ANTIBIOTICS
The characterization of antibiotics binding to pre-miRNAs involved measuring thermodynamic parameters, specifically K D values, which enable the calculation of Gibbs free energy (ΔG o ) using the formula ΔG o = RTlnK D Here, R represents the ideal gas constant and T denotes absolute temperature The enthalpy (ΔH o ) and entropy (ΔS o ) contributions were estimated using the Van’t Hoff relation, which describes the temperature dependence of ΔG o Results for the interaction with pre-miRNA-372, along with similar findings for other RNAs, are detailed in table 2.4 All ΔG o values were significantly negative, indicating a spontaneous and thermodynamically favorable interaction The ΔG o values can be partitioned into enthalpic and entropic contributions, which together influence the overall binding affinity while remaining interdependent.
Table 2.4 Thermodynamic parameters for antibiotics/pre-miRNA-372 interactions ΔG°, ΔH°, TΔS° and ΔG°nel are expressed in kJ/mol
Neomycin B -27.6 -6.47 ± 0.8 21.1 ± 0.9 -19.3 70 Paromomycin -22.3 -14.7 ± 1.0 7.66 ± 1.2 -13.6 61 Neamine -21.6 -2.5 ± 0.7 19.1 ± 1.6 -15.6 72 Gentamycin -26.2 2.23 ± 0.6 28.4 ± 2.2 -16.5 63 Streptomycin -25.6 -20.4 ± 0.9 5.14 ± 1.0 -12.0 47 Apramycin -33.8 -3.88 ± 1.1 29.8 ± 1.5 -24.3 72
Aminoglycosides exhibit optimal binding to specific RNAs primarily due to significant entropic contributions Compounds such as neomycin, neamine, gentamycin, and apramycin demonstrate a strong entropic factor, enhancing their binding affinity In contrast, paromomycin and streptomycin show weaker entropic contributions, resulting in lower binding affinity Tetracyclines present a distinct profile in this context.
Aminoglycosides and tetracyclines exhibit distinct behaviors due to their different chemical structures and interaction sites, leading to varying interaction mechanisms To understand these interactions better, we analyzed the free energy associated with electrostatic interactions by measuring the dissociation constant (K D) at varying salt concentrations The results indicate that non-electrostatic interactions play a predominant role in the binding of both classes of compounds, with tetracyclines showing the highest percentages of non-electrostatic free energy contribution (ΔG o nel) Overall, this study highlights the significance of non-electrostatic interactions in the affinity of aminoglycosides and tetracyclines.
Aminoglycosides exhibit binding efficiencies of 47% to 72%, while tetracyclines show a higher range of 78% to 91%, indicating that tetracyclines engage in more specific interactions Thermodynamic studies reveal that both drug classes can be chemically modified to enhance their binding and inhibition activities Tetracyclines, in particular, demonstrate an impressive binding profile with a K D in the nanomolar range and a significant number of specific non-electrostatic interactions Although their binding site is not ideal for inhibiting pre-miRNA processing, some level of inhibition is still observed Their thermodynamic profile, characterized by enthalpic binding, suggests that further chemical modifications could optimize their efficacy.
A comprehensive study on the inhibition activity and RNA binding affinity of various clinically relevant rRNA-binding antibiotics has highlighted key parameters essential for the effective binding of pre-miRNAs and their processing inhibition Among the evaluated antibiotics, neomycin emerged as the most effective inhibitor and a superior RNA ligand, with potential for further optimization through chemical modification Notably, the comparison of aminoglycosides and tetracyclines underscored the critical role of the interaction site on the pre-miRNA sequence/structure in efficiently inhibiting processing by the enzyme Dicer While macrolides and puromycin exhibited strong affinity for targeted RNAs, they failed to inhibit pre-miRNA processing.
Recent studies have demonstrated that certain antibiotics, specifically aminoglycosides and tetracyclines, can selectively inhibit the processing of oncogenic miRNAs by Dicer, while others like lincosamides, linezolid, and chloramphenicol do not bind to pre-miRNAs This selectivity highlights the potential for designing small molecules that target RNA effectively Notably, some of these compounds show discrimination among various RNA sequences and other nucleic acid structures, such as tRNA and dsDNA Tetracyclines, previously unexamined for their RNA binding capabilities, exhibit strong binding affinity and promising thermodynamic profiles in inhibiting pre-miRNA processing, suggesting that chemical modifications could yield more potent and selective inhibitors for oncogenic miRNA production.
The rational design of RNA binders that can effectively disrupt the biogenesis and function of target RNAs presents a significant challenge for medicinal chemists Discovering small molecules that specifically target RNA could lead to innovative drugs for various diseases This study offers valuable insights into the key parameters necessary for strong and selective binding to pre-miRNAs, as well as those crucial for inhibiting their processing It suggests that aminoglycosides and tetracyclines may serve as promising scaffolds for future drug development aimed at inhibiting oncogenic miRNAs.