PEI-complexed LNA antiseeds as miRNA inhibitors

Abstract

Antisense inhibition of oncogenic or other disease-related miRNAs and miRNA families in vivo may provide novel therapeutic strategies. However, this approach relies on the development of potent miRNA inhibitors and their efficient delivery into cells. Here, we introduce short seed-directed LNA oligonucleotides (12- or 14-mer antiseeds) with a phosphodiester backbone (PO) for efficient miRNA inhibition. We have analyzed such LNA (PO) antiseeds using a let-7a-controlled luciferase reporter assay and identified them as active miRNA inhibitors in vitro. Moreover, LNA (PO) 14-mer antiseeds against ongogenic miR-17–5p and miR-20a derepress endogenous p21 expression more persistently than corresponding miRNA hairpin inhibitors, which are often used to inhibit miRNA function. Further analysis of the antiseed-mediated derepression of p21 in luciferase reporter constructs - containing the 3′-UTR of p21 and harboring two binding sites for miRNAs of the miR-106b family - provided evidence that the LNA antiseeds inhibit miRNA families while hairpin inhibitors act in a miRNA-specific manner. The derepression caused by LNA antiseeds is specific, as demonstrated via seed mutagenesis of the miR-106b target sites. Importantly, we show functional delivery of LNA (PO) 14-mer antiseeds into cells upon complexation with polyethylenimine (PEI F25-LMW), which leads to the formation of polymeric nanoparticles. In contrast, attempts to deliver a functional seed-directed tiny LNA 8-mer with a phosphorothioate backbone (PS) by formulation with PEI F25-LMW remained unsuccessful. In conclusion, LNA (PO) 14-mer antiseeds are attractive miRNA inhibitors, and their PEI-based delivery may represent a promising new strategy for therapeutic applications.

Introduction

Small non-coding RNAs (sRNA), such as miRNAs, are involved in the regulation of virtually all essential cellular processes, and their deregulation has important implications in pathogenesis. The impact of miRNAs on human diseases has been well established e.g., in the cancer field (for a review see ref. 1), where several miRNAs are either overexpressed in tumors (oncogenic miRNAs) or show reduced levels (miRNAs with tumor suppressor activity). This results in the downregulation of tumor suppressor genes or in the upregulation of cellular proto-oncogenes, respectively.² A well characterized and important cellular tumor suppressor is the cell cycle regulator p21, which is post-transcriptionally regulated by miRNAs of the oncogenic miR-17–92 cluster.³ Oncogenic miRNAs or miRNA clusters are often found in chromosomal regions that are amplified in tumor cells, which promotes their overexpression.^4,5 For example, the miR-17–92 cluster is overexpressed in lymphoma as well as in a wide range of human solid tumors.⁶ MiRNAs of this cluster (mainly miR-17–5p, miR-19 and miR-20a) inhibit apoptosis, promote proliferation and induce tumor angiogenesis.^7,8

Several strategies have been developed to block cellular miRNA function, including the use of miRNA sponges, masking of the mRNA target site or inhibition of miRNAs by antisense oligonucleotides (generally termed antimiRs; for a review see ref. 9). For therapeutic intervention the use of antisense strategies against miRNAs has been established in mice and non-human primates.^10,11 However, efficient delivery of antimiRs is still the main bottleneck for therapeutic applications. Several efforts have been made to deliver antimiRs into cells and tissues, e.g., the conjugation of modified antimiRs with cholesterol, which were termed antagomiRs.¹² Moreover, important improvements with regard to antimiR efficacy, stability and delivery have been achieved by the incorporation of modified nucleotides (for a review see ref. 13). In particular, 2’-fluor (2’-F), 2’-oxymethyl (2’-OMe), 2’-methoxyethyl (2’-MOE) and locked nucleic acid (LNA) modifications have been incorporated into antimiRs for in vivo applications. So far, the most efficient variants are 2’-F/2’-MOE mixmers directed against the complete miRNA guide strand,^11,14 15- or 16-mer LNA/DNA mixmers with 2´-OMe cytosines and a phosphorothioate (PS)^10,15,16 or phosphodiester (PO) backbone,¹⁷ and tiny LNA 8-mers with a PS backbone.^18,19 All these antimiRs were applied by “gymnotic delivery,”²⁰ i.e., without any formulation. In contrast, the formulated delivery of unconjugated modified antimiRs into tissues has not been explored so far.

Here we introduce 14-mer LNA (PO) antiseeds as novel miRNA inhibitors that are particularly potent and can inhibit miRNAs for up to 7 d in cell culture. We have further established their complexation with a cationic polymer, the low molecular weight polyethylenimine PEI F25-LMW, to enable delivery into a broader spectrum of organs and tissues. Since we have shown recently that PEI-complexed siRNA or miRNA can be delivered into tumor xenografts leading to antitumor effects in mice,^21,22 it will be promising to explore the therapeutic potential of PEI-complexed LNA (PO) antiseeds.

Results

LNA antiseeds as miRNA inhibitors

The LNA/DNA mixmers used so far, as well as the tiny LNA 8-mers described recently,¹⁹ are generally designed with a phosphorothioate (PS) backbone that is required for unformulated “gymnotic delivery” in vivo.²⁰ However, gymnotic delivery cannot be applied to all organs and tissues, which is a serious limitation for several therapeutic applications. Here, we explored short LNA antiseeds containing a phosphodiester (PO) backbone. These oligonucleotides are complementary to the 5′-terminal seed (nt 1–8) plus the 4–6 subsequent nucleotides of the mature miRNA strand.

As a first experimental approach, we cloned a let-7a target sequence in the functional forward orientation into the 3′-UTR of a luciferase reporter gene; a luciferase reporter gene with an inverted let-7a target sequence placed at the same position served as negative control (Fig. 1A). To determine the minimum antiseed length for blocking miRNA function, we analyzed the effects of 8-, 10-, 12- and 14-mer all-LNA (PO) antiseeds directed against let-7a (Fig. 1B and C). The luciferase reporter plasmid and LNA antiseeds at various concentrations (5, 20, 50 and 100 nM) were cotransfected into HeLa cells using lipofectamine and luciferase activity was measured as an indicator for derepression. Cells transfected with the luciferase reporter containing an inverted let-7a target site were used the reference for maximum derepression. The 8- and 10-mer antiseeds showed no or only weak derepression activity, whereas the 12- and the 14-mer antiseeds efficiently derepressed the luciferase reporter (Fig. 1C). Next, we compared the derepression activity of the 12- and 14-mer antiseeds with a commercially available miRNA hairpin inhibitor²³ against let-7a (Thermo Scientific Dharmacon). Such hairpin inhibitors carry a central single-stranded sequence complementary to the entire mature miRNA strand and are capped on both ends by hairpins consisting of 8-bp stems and apical tetraloops.²³ As negative control, we employed an unrelated all-LNA 14-mer directed against the RNA subunit of RNase P from E. coli.²⁴ When using concentrations less than 10 nM, the 12- and 14-mer antiseeds showed derepression activities comparable to the hairpin inhibitor, with RLU values reaching approx. 60% of those measured for the inverted control (Fig. 1D). At 10 nM or 20 nM, however, the hairpin inhibitor partially or completely failed to derepress the luciferase mRNA, respectively, while the LNA antiseeds, and in particular the 14-mer, displayed constant activity between 5 and 20 nM (Fig. 1D). In the case of the hairpin inhibitor, concentrations > 20 nM even caused a further drop of luciferase activity to almost zero (0.026 +/− 0.011 or 0.042 +/− 0.018 at 50 or 100 nM, respectively; data not shown). Since cell viability did not seem to be affected by the hairpin inhibitor (data not shown), we speculate that the design of hairpin inhibitors²³ may favor nucleic acid aggregation at higher concentrations when using lipofectamine as the transfection reagent, leading to inefficient cotransfection.

Figure 1. (A) Schematic view of the let-7a forward luciferase reporter and the inverted control construct; the target sequence fully complementary to mature miRNA hsa-let-7a and the inverted control sequence are depicted (in the sense of the RNA transcript). The sequence of mature miRNA hsa-let-7a is shown underneath, with upper case letters indicating the nucleotides targeted by the let-7a antiseeds. (B) Sequences of the used antiseeds directed against the 5′-region of miRNA let-7a and sequence of the LNA control oligonucleotide directed against bacterial RNase P RNA. (C) Derepression activity of let-7a antiseeds (8- to 14-mers) at the indicated concentrations (5–100 nM) in HeLa cells. Luciferase activity of the control vector harboring the inverted let-7a target sequence was set to 1. Values are derived from at least three independent experiments (+/− S.D.). (D) Comparison of LNA 12- and 14-mer antiseeds containing a PO backbone with a hairpin inhibitor (hp inhib.) targeting let-7a at the indicated concentrations (1 to 20 nM) in HeLa cells using lipofectamine as transfection agent. As control, the unrelated LNA 14-mer (LNA con.) was used. Values are derived from at least three independent experiments (+/− S.D.).

We conclude from our results that 12- and 14-mer LNA antiseeds containing a PO backbone are active miRNA inhibitors with potencies comparable to commercial miRNA hairpin inhibitors (Thermo Scientific Dharmacon) which are often used for miRNA inhibition. In addition, transfection with lipofectamine seems to be more robust with LNA antiseeds than hairpin inhibitors.

Derepression of basal p21 levels in K562 cells

The erythroleukemia cell line K562 expresses the tumor suppressor protein p21 at levels that are barely detectable in western blot analyses, although substantial levels of p21 mRNA are present in these cells as inferred from RT-qPCR analyses (data not shown). This indicates that p21 expression is suppressed on the post-transcriptional level. Recently, it was shown that the p21 mRNA, whose 3′-UTR harbors two conserved binding sites for miRNAs of the miR-106b family, is indeed a target for these miRNAs.³ Of note, the oncogenic miR-17–92 cluster encoding three members of the miR-106b family, namely miR-17–5p, miR-18a and miR-20a, is overexpressed in K562 cells.^21,25 We therefore anticipated that the miR-106b family members miR-17–5p and miR-20a repress endogenous p21 mRNA in K562 cells and tested this by transfecting 14-mer LNA (PO) antiseeds directed against miR-17–5p and miR-20a into these cells by electroporation. Indeed, both antiseeds caused a substantial increase of p21 protein levels, whereas the all-LNA control 14-mer remained without effect (Fig. 2).

Figure 2. Duration of p21 derepression by LNA (PO) 14-mer antiseeds vs. hairpin inhibitors. Expression of p21 was analyzed by western blotting after transfection of LNA 14-mer antiseeds and miRNA hairpin inhibitors (hp inhib.) directed against miR-17–5p or miR-20a in K562 cells. As a control, the unrelated LNA 14-mer (LNA con., see Fig. 1B) was used. Beta-Actin served as a loading control.

Since it was recently shown that LNA antisense effects are rather persistent,²⁶ we compared the duration of p21 derepression caused by the 14-mer LNA antiseeds 17–5p and 20a with that of the respective hairpin inhibitors. 48 h after transfection, all antisense molecules gave rise to substantial derepression of p21 in K562 cells (Fig. 2). However, after 72 h the hairpin inhibitors had substantially lost their capacity to derepress p21, while antiseed 17–5p derepressed p21 for up to 96 h and antiseed 20a for at least 120 h. Even after 168 h a weak derepression of p21 with antiseed 20a was still detectable (Fig. 2). Our results on antiseeds are consistent with the long-term effects of LNA antisense molecules described by others.^10,19

Specificity of the antiseed effect

Next, we analyzed the specificity of derepression effects observed with the 14-mer antiseeds against miR-17–5p and miR-20a by using luciferase reporter assays. For this purpose, the 3′-UTR of p21 and a mutated version thereof with both seeds inactivated by two point mutations (Fig. 3A) were fused to the luciferase coding region. Cotransfection (with lipofectamine) of the luciferase reporter plasmid with one of the antiseeds, 17–5p or 20a, caused a more profound derepression of the p21 reporter than cotransfection with hairpin inhibitor 17–5p or 20a at the same concentration (10 nM; Fig. 3B). This stronger derepression activity of the antiseeds as compared with the individual hairpin inhibitors were in line with the results from our western blot experiments (see Fig. 2). Next, we compared the derepression activity of an equimolar (each 5 nM) mixture of the two antiseeds with that of a mixture of both hairpin inhibitors. While the combination of the antiseeds did not further increase the derepression activity, the combined application of the two hairpin inhibitors resulted in an enhanced derepression activity that was now comparable to that of the single LNA antiseeds (Fig. 3B). This suggests that antiseeds are able to block miRNAs from the same family, while hairpin inhibitors are more specific for individual miRNAs.

Figure 3. (A) Schematic presentation of the reporter carrying the luciferase coding sequence (CDS) followed by the p21 3′-UTR that harbours two miR-106b family seed regions (top). The seed regions are mutated in the variant reporter construct underneath, with lower case letters indicating the mutated nucleotides. (B) Derepression activity of the indicated miRNA inhibitors (hp inhib. or 14-meric LNA antiseeds) directed against miR-17–5p and miR-20a after lipofectamine-mediated cotransfection with the luciferase/p21 3′-UTR reporter plasmid into HeLa cells. Luciferase activity values obtained after cotransfection with the LNA con. 14-mer were set to 1. Values are derived from at least three independent experiments (+/− S.E.M.). (C) Effects of the indicated miRNA inhibitors on luciferase expression using the seed-mutated version of the reporter (see panel A). Values are derived from at least three independent experiments (+/− S.E.M.).

As a further control we tested a tiny LNA 8-mer with a PS backbone targeting the seed region of miRNAs from the miR-106b family. Recently, it has been shown in cotransfection experiments with lipofectamine that only LNA (PS) 8-mers can derepress a luciferase reporter efficiently, whereas shorter or longer LNA (PS) oligonucleotides are inefficient.¹⁹ Indeed, we found that the tiny LNA 8-mer (PS) was able to derepress the p21 reporter with high efficiency (Fig. 3B). When using the luciferase reporter with seed-mutagenized p21 3′-UTR, none of the antiseeds showed a substantial derepression effect (Fig. 3C). The miRNA hairpin inhibitors even somewhat reduced luciferase activity of the mutated reporter (Fig. 3C), suggesting unwanted side effects of such molecules in our experimental setup. Also, the LNA 8-mer (PS) slightly derepressed expression from the seed-mutagenized reporter (Fig. 3C), suggesting that the high derepression in Figure 3B includes contributions from this non-specific stimulation.

Taken together, our results establish 14-mer LNA antiseeds (PO) as interesting new miRNA or miRNA family inhibitors with favorable properties in terms of efficacy, duration and specificity of depression effects.

Functional delivery of LNA antiseeds with PEI F25-LMW

A major bottleneck for the therapeutic application of any nucleic acid, including LNA antiseeds, is their efficient and functional delivery in vivo. Previously, we have established nucleic acid complexation with low molecular weight polyethylenimine (PEI F25-LMW) for the systemic administration of siRNA or miRNA in mice.^22,27-29 While PEI F25-LMW has cell type-dependent transfection efficacy in vitro, it is important to note that it can be used for efficient delivery of nucleic acids in vivo, which is not possible with other liposomal transfection reagents. More specifically, PEI F25-LMW is less toxic than other PEIs, and the formation of PEI-based nanoparticles protects nucleic acids from degradation in biological fluids, mediates their cellular uptake and triggers their release from the endosomal/lysosomal system due to the “proton-sponge effect.”^27,29

Here, we aimed at extending our PEI F25-LMW-based delivery platform toward LNA antimiRs for blocking the function of oncogenic or other disease-related miRNAs in vivo. We analyzed the derepression activity of PEI F25-LMW/antiseed complexes against miR-17–5p and miR-20a in HeLa cells and compared it with that of miRNA hairpin inhibitors and the tiny LNA 8-mer (PS) complexes in the same manner. Complexation was done at a PEI/nucleic acid mass ratio of 5:1, which had been determined as optimal in previous siRNA and miRNA experiments²² with regard to complex stability and surface charge (zeta potential). In our experimental setup, HeLa cells were first transfected with the p21 reporter plasmid using lipofectamine, followed by transfection of PEI/nucleic acid complexes assembled at 10 nM of the respective nucleic acid. Similar derepression of the luciferase reporter carrying the p21 3′-UTR was observed with most antimiRs, while the hairpin inhibitor against miR-20a caused the highest derepression activity; no derepression at all was seen with the LNA (PS) 8-mer (Fig. 4A). Analysis of the same PEI/nucleic acid complexes in combination with the luciferase reporter encoding the seed-mutagenized p21 3′-UTR ruled out any non-specific effects (Fig. 4B). We further analyzed PEI/nucleic acid complexes for possible toxic side effects by using a WST-1 assay that senses cell viability via mitochondrial energy metabolism (see Supplementary Materials). We did not observe any significant negative effects on cell viability for up to 250 nM of PEI-complexed LNA antiseeds, hairpin inhibitors or the LNA 8-mer (PS) (Fig. S1).

Figure 4. (A) Derepression of the luciferase/p21 3′-UTR reporter by PEI F25-LMW-complexed LNA (PO) 14-mer antiseeds, miRNA hairpin inhibitors (hp inhib.) and an LNA (PS) 8-mer directed against miR-17–5p and miR-20a in HeLa cells. The LNA con. 14-mer was used as a reference and the luminescence measured for cell lysates harbouring LNA con. was set to 1. Values are derived from four independent experiments (+/− S.E.M.). (B) Effects of the indicated PEI-complexed miRNA inhibitors on luciferase expression using the seed-mutated version of the reporter (see Fig. 3A). Values are derived from four independent experiments (+/− S.E.M.). (C) PEI-complexation experiments using 5′-[³²P]-endlabeled double-stranded siRNA or single-stranded 8- or 14-meric RNA (10⁴ Cerenkov cpm per lane) as probes; PEI complexation was indicated by shifting the free RNA probe in agarose gels to slower migrating PEI/RNA complexes. To obtain PEI/RNA mass ratios of 0.1 to 10, 500 ng of either unlabeled siRNA duplex, 14-meric RNA or 8-meric RNA were combined with the corresponding labeled probes and incubated with different amounts of PEI F25-LMW as indicated. Representative agarose gels are shown. A quantitative analysis based on at least three independent experiments is shown at the bottom (+/− S.E.M.).

The absence of LNA 8-mer (PS) activity (Fig. 4A), in contrast to antiseeds or hairpin inhibitors, prompted us to address the question if complexation by PEI F25-LMW might be inefficient for short single-stranded oligomers relative to siRNA duplexes. We thus compared PEI complexation of a double-stranded siRNA³⁰ with that of single-stranded RNA (14- or 8-mer) in an agarose gel assay (Fig. 4C). Different amounts of PEI F25-LMW were incubated with 500 ng of the respective RNA to obtain PEI/RNA mass ratios of 0.1 to 10. To visualize PEI/RNA complexes in agarose gels, trace amounts of the respective 5′-[³²P]-labeled RNA were added (~500- to 1000-fold excess of unlabeled RNA over labeled RNA). Notably, we observed no significant differences in complexation efficacies for the siRNA duplex or the single-stranded RNA oligomers of different length. In all cases, complex formation started at a PEI/RNA mass ratio of 0.25 (which corresponds to 125 ng PEI) and reached 100% complexation at a PEI/RNA mass ratio of 1:1 (Fig. 4C). The quantitative analysis (Fig. 4C, bottom graph) revealed only minor differences in complexation efficacy of siRNA duplex and 8- or 14-meric RNA single strands. More importantly, at the PEI/RNA mass ratios of 2.5:1 or 5:1 used in our HeLa cell transfection experiments (see Fig. 4A and 5) all three types of RNA were fully complexed with PEI (Fig. 4C).

Figure 5. Derepression of the let-7a-responsive luciferase reporter by the PEI F25-LMW-complexed LNA (PO) 14-mer antiseed against let-7a in HeLa cells. For the control LNA 14-mer (LNA con.), see Fig. 1B. Luciferase activity of lysated derived from cells harbouring the control vector with an inverted let-7a target sequence (Fig. 1A) was set to 1. Values are derived from at least three independent experiments (+/− S.E.M.).

Visualization of PEI nanoparticles containing LNA antiseeds

Because 5′-³²P-labeling by T4 polynucleotide kinase using standard protocols was ineffective, we could not apply the same agarose gel-based complexation test to LNA oligomers. Also, staining of PEI/LNA complexes with SYBR^® Gold or silver was unsuccessful (data not shown). Thus, we analyzed PEI/LNA complexes by atomic force microscopy (AFM). Indeed, upon mixing of PEI F25-LMW and LNA 14-mers (PO) at different mass ratios the formation of polymeric nanoparticles was observed (Fig. 6). At PEI/LNA mass ratios of 2.5:1 or 5:1 the resulting nanoparticles were quite homogeneous in size (diameters of ~80 -150 nm). Polymeric PEI nanoparticles were also formed with the LNA 8-mer (PS) (Fig. 6). Thus, the inability to raise derepressional effects with PEI/LNA 8-mer (PS) complexes (see Fig. 4A) has very likely reasons other than inefficient PEI nanoparticle formation.

Figure 6. Detection of PEI F25-LMW/LNA nanoparticles using atomic force microscopy. Shown are four representative images of PEI/LNA complexes at the indicated mass ratios, containing either the LNA (PS) 8-mer (lower panel on the right) or an LNA (PO) 14-mer (all other panels). For details, see Materials and Methods.

After demonstrating the efficient formation of homogeneous PEI/LNA complexes at mass ratios of 2.5:1 and 5:1, we analyzed our PEI-based delivery strategy in the let-7a-specific luciferase reporter system (see Fig. 1A). The efficient repression of the let-7a reporter in HeLa cells (Fig. 1C, cf. bars for “f” and “antiseed 14-mer” at 50–100 nM) qualified this reporter system as optimal for further validating the relatively weak derepression effects of PEI-complexed LNA antiseeds on the p21 3′-UTR reporter system (see Fig. 4A). We now used a PEI/nucleic acid mass ratio of 2.5:1 and varied the 14-meric let-7a LNA antiseed concentrations between 10 and 250 nM. In this setup, the 14-mer LNA antiseed caused a specific approx. 3-fold derepression of the let-7a reporter at 100 nM or 250 nM (Fig. 5), which is a more robust effect than that observed in Figure 4A. We conclude that PEI F25-LMW efficiently complexes LNA antiseeds and allows for their delivery in vitro, thus providing the basis for future in vivo experiments.

Discussion

The inhibition of miRNAs is an important strategy to analyze their cellular functions and to reduce miRNA overexpression levels in several disease states. This may well lead to novel treatment strategies based on reversion of the consequences of aberrantly high miRNA expression levels. One promising approach to block miRNA is the use of LNA-based antisense strategies. LNA has several advantages over other nucleic acids, including the increase in Tm values of usually 2–4°C per LNA modification compared with DNA, its resistance against intracellular nucleases and nucleases in biological fluids, improved mismatch discrimination (for reviews see refs. 13,31 and 32) and low toxicity profiles in preclinical in vivo studies.¹⁰ The rate constant k_off of LNA/RNA hybrids is dramatically reduced^24,33 and therefore steric blockage of target nucleic acids is anticipated to be very efficient and persistent. Recently, the successful use of LNA-modified antisense oligonucleotides against miRNAs has been described, for example tiny LNA 8-mers with a PS backbone or various LNA/DNA mixmers (see Introduction). However, the identification of short-sized (~14 nt) LNA oligonucleotides that have the capacity to target miRNA families with common seeds, in combination with the development of efficient strategies for their delivery, is an important exercise toward in vivo applications.

In this study we have introduced 14-mer LNA antiseeds with a PO backbone, which displayed several context-dependent advantages over other miRNA inhibitors. By using a miRNA (let-7a)-regulated reporter gene assay in HeLa cells and lipofectamine as transfection agent we found 12- or 14-meric LNA antiseeds to be more potent than shorter variants (8- or 10-mer; Fig. 1C). Also, in this assay system, a commonly used miRNA hairpin inhibitor²³ lost its otherwise similar inhibition capacity at higher inhibitor concentrations (≥ 20 nM), possibly due to aggregation effects (Fig. 1D). In our second test system, where we transfected K562 cells by electroporation, derepression of endogenous p21 expression was much more persistent with LNA (PO) 14-mer antiseeds than with corresponding hairpin inhibitors (Fig. 2). In our third analysis, which was comparable to the first assay system except for the reporter carrying the p21 3′-UTR with two miR-106b target sites, we found that LNA (PO) 14-mer antiseeds were again active (Fig. 3B). These experiments provided evidence that the LNA (PO) 14-mer antiseeds tend to target an entire miRNA family, while the hairpin inhibitors are more specific for individual miRNAs. This was inferred from the observation that hairpin inhibitors, but not antiseeds, against miR-17–5p and miR-20a showed synergistic effects (Fig. 3B). A seed-targeting tiny LNA 8-mer (PS), which has the capacity to block miRNAs from a miRNA family,¹⁹ appeared to even perform best in this assay, although control experiments with the reporter gene that contain mutations in the miR-17–5p and miR-20a target sites pointed to some unspecific stimulatory effect of this type of inhibitor on reporter expression (Fig. 3C). From our results we conclude that hairpin inhibitors are suited to dissect the functions of single miRNAs from a miRNA family, whereas the broader effect of our 14-mer LNA (PO) antiseeds on more than one miRNA family member may well enhance possible therapeutic effects.

In our fourth approach (Fig. 4) we employed PEI F25-LMW as transfection agent instead of lipofectamine, keeping in mind that efficient and functional delivery in vivo remains the critical issue for nucleic acid-based therapeutic applications and common transfection reagents cannot be used here (for review, see e.g., ref. 34).

As shown previously, PEI complexation of nucleic acids leads to polymeric nanoscale particles that (1) allow the complete protection of the nucleic acid against degradation, (2) mediate the cellular uptake of the cargo, (3) trigger the release from the endosomal/lysosomal system due to the proton-sponge effect and (4) release the nucleic acid into the cytoplasm (for a review see ref. 35). Notably, at the amounts used for therapeutic intervention, the low molecular weight PEI F25-LMW²⁸ employed in this study has been shown to exert no toxicity and no immunostimulatory or otherwise non-specific effects in vitro and in vivo.^29,36 In line with these previous findings, we have not observed any toxic effects of PEI F25-LMW-complexed LNA antiseeds in the study presented here (Fig. S1).

Here we demonstrate that single-stranded 14-mer LNA (PO) antiseeds can be functionally delivered by PEI F25-LMW (Fig. 4A and 5) with efficacies comparable to hairpin inhibitors (Fig. 4A). In contrast, the tiny LNA 8-mer (PS) failed to show an inhibitory effect when using PEI F25-LMW as the delivery agent (Fig. 4A). This has other reasons than incomplete PEI complexation of the LNA 8-mer (PS) because neither (1) its short length, (2) its PS backbone, (3) its single-stranded character nor (4) its LNA modifications had any significant negative effect on particle formation with PEI F25-LMW (Fig. 4C and 6). One possibility among several is that very short oligonucleotides such as 8-mers are inefficiently released from the nanoparticle within cells.

Taken together, our results establish the efficiency of LNA (PO) 14-mer antiseeds as new miRNA inhibitors, and demonstrate the capability of such LNA antiseeds to build functional complexes with PEI that can be delivered into cancer cells in order to derepress tumor suppressor genes such as p21. Thus, our results open up an avenue toward therapeutic application of LNA antiseeds for miRNA inhibition.

Materials and Methods

Oligonucleotides and antibodies

Locked nucleic acids (LNA) antisense oligonucleotides (antiseeds) were purchased from RiboTask (Odense, Denmark). Antiseed sequences were as follows:

Antiseed against miR-17–5p (PO): 5′-CTGTAAGCACT[mU]TG;

Antiseed against miR-20a (PO): 5′-CTATAAGCA[mC]T[mU]TA;

Antiseed against miR-17–5-p/20a (PS): 5′-GCACT[mU]TG.

All residues were LNA, except for isolated 2’-oxymethyl pyrimidines marked as [mU] and [mC]. These changes are expected to reduce self-pairing without disturbing target binding.

LNA antiseeds directed against hsa-let-7a were of different lengths (14- to 8-mers):

Antiseed let7a-14-mer: 5′-AACCTACTACCTCA;

Antiseed let7a-12-mer: 5′-CCTACTACCTCA;

Antiseed let7a-10-mer: 5′-TACTACCTCA;

Antiseed let7a-8-mer: 5′-CTACCTCA.

Escherichia coli RNase P RNA 5′-CAAGCAGCCUACCC²⁴ was used as negative control.

The sequence of the RNA 14-mer, which was used for PEI-complexation studies, is: 5′-GUUCGGUCAAAACU

The sequence of the RNA 8-mer, which was used for PEI-complexation studies, is:

5′-GUUCGGUC

The sequence of the VR1 siRNA, which was used for PEI-complexation studies, is:

5′-GCGAUCUUCUAUUCAACdTdT (sense)

and 5′-GUUGAAGUAGAAGAUGCGCdTdT (antisense)³⁰

MiRIDIAN hairpin inhibitors against hsa-let-7a, hsa-miR-17–5-p and hsa-miR-20a were purchased from Dharmacon (Thermo Scientific, Langenselbold, Germany).

Antibodies against p21 (sc-6246) and β-Actin (sc-47778) as well as the secondary antibody (sc-2005) were purchased from Santa Cruz Biotechnology (Heidelberg, Germany).

Cell culture

Cells (K562 and HeLa) were cultured under standard conditions (37°C, 5% CO₂ in a humidified atmosphere) in RPMI 1640 (K562) or IMDM (HeLa), supplemented with 10% FCS (PAA, Cölbe, Germany).

Transfection procedures

Reporter assays using let-7a antimiRs in HeLa cells

HeLa cells were transfected with the LNA antiseeds, controls or the hairpin inhibitors using Lipofectamine^TM 2000 (Life Technologies Invitrogen, Darmstadt, Germany) according to the manufacturer’s protocol. One day before transfection, 8 × 10⁴ cells were seeded into 24-well plates and cultivated under standard conditions. Twenty-four hours after seeding, cells were co-transfected with 0.5 µg of vector DNA (plasmid “pGL3 control” with a let-7a or an inverted let-7a target sequence) and with 0, 10, 20, 50 or 100 nM of the LNA antiseed, LNA control or hairpin inhibitor, respectively, according to the manufacturer’s protocol. Co-complexation was performed in Opti-MEM^® 1 (Life Technologies Invitrogen). After 4 to 6 h, the transfection medium without serum was aspirated and replaced by Iscove's modified Dulbecco's medium (IMDM) containing 10% FCS. Forty-eight hours after transfection, cells were lysed and prepared for reporter assay measurements.

p21 derepression in K562 cells

For transfection experiments, 1 × 10⁶ K562 cells were electroporated in 4 mm cuvettes with a single pulse at 330 V for 10 ms using a BioRad GenePulser XCell (Biorad, München, Germany). Cells were transfected with 1 µg LNA antiseeds, LNA controls or hairpin inhibitors. Forty-eight hours after transfection, cells were lysed and lysates were analyzed in western blot experiments.

Specificity of p21 derepression in HeLa cells

The transfection of HeLa cells was performed using Lipofectamine^TM 2000 (Life Technologies Invitrogen). One day prior to transfection, 4 × 10⁴ cells were seeded into 24-well plates; for co-complexation, 1 µl Lipofectamine^TM 2000 was combined with 10 nM LNA antiseed, LNA control or hairpin inhibitor and 0.5 µg of vector DNA (plasmid “pGL3 control” containing the p21 3′-UTR or the mutated p21 3′-UTR) in Opti-MEM^® 1 medium according to the instructions of the manufacturer; the mixture was added to the cells, proceeding as described above (see Reporter assays using let-7a antimiRs in HeLa cells).

PEI complexation and delivery

Using the p21 reporter system

for complexation of miRNA inhibitors with PEI F25-LMW, cells were pre-transfected with vector DNA (“pGL3 control” plasmid containing the p21 3′-UTR or the mutated p21 3′-UTR) using Lipofectamine^TM 2000 as described above (see Specificity of p21 derepression in HeLa cells). After 6 h, the medium was replaced by IMDM / 10% FCS, and PEI complexes were added. For PEI F25-LMW complexation, LNA antiseeds, LNA control oligonucleotide, tiny LNA 8-mer with PS backbone or hairpin inhibitors were diluted in complexation buffer (150 mM NaCl, 10 mM HEPES pH 7.4). To gain a final concentration of 10 nM, 18 ng of LNA 8-mer (PS), 28.5 ng of LNA 14-mers (control, antiseed 17–5p or antiseed 20a) and 80 ng of hairpin inhibitors 17–5p or 20a were diluted in 50 µl complexation buffer and equilibrated 5 min at room temperature. To obtain a PEI / oligonucleotide mass ratio of 5:1, 90 ng PEI was used for complexation of the LNA 8-mer (PS), 142.5 ng PEI for complexation of the LNA antiseeds and LNA 14-mer control, and 390 ng PEI was required for the hairpin inhibitor complexes. After equilibration of PEI to room temperature in complexation buffer, the oligonucleotide and the PEI solution were mixed, briefly vortexed and incubated for 30–60 min at room temperature to allow complex formation. PEI F25-LMW complexes were then added to the cells (see above), and 48 h later cells were lysed and prepared for reporter assay measurements.

Using the let-7a reporter system

Cells were pre-transfected with vector DNA (“pGL3 control” plasmid containing the let-7a forward or inverted target sequence) using Lipofectamine^TM 2000 as described above (see Reporter assays using let-7a antimiRs in HeLa cells). After 6 h, the medium was replaced by IMDM / 2% FCS, and PEI complexes were added. For PEI F25-LMW complexation, the let-7a LNA 14-mer and the LNA control 14-mer were diluted in complexation buffer (150 mM NaCl, 10 mM HEPES pH 7.4). To gain a final oligonucleotide concentration of 10, 50, 100 or 250 nM, 115, 575, 1150 or 2875 ng of the LNA 14-mers were diluted in 50 µl complexation buffer and equilibrated to room temperature. PEI F25-LMW was used at a final PEI / oligonucleotide mass ratio of 2.5:1; therefore 288, 1,440, 2,880 or 7,200 ng of PEI was diluted in 50 µl complexation buffer and also equilibrated 5 min at room temperature. Complexation was initiated by mixing PEI and oligonucleotide solutions, followed by incubation at room temperature for 30 - 60 min. PEI / LNA complexes were finally added to the cells and luciferase reporter measurements were performed 48 h posttransfection.

Determination of PEI complexation efficacy

To determine the complexation efficacy of PEI F25-LMW, a radioactive assay was employed as described previously.³⁷ For this purpose, 500 ng of single-stranded 14- or 8-meric RNA or double-stranded siRNA were used, mixed with trace amounts (10⁴ Cerenkov cpm) of 5′-[³²P]-endlabeled siRNA, RNA 14-mer or RNA 8-mer, respectively. Complexation was performed with the indicated amounts of PEI (see “PEI complexation and delivery” above and Fig. 4C). Samples were mixed with DNA loading buffer (10 mM Tris/HCl (pH 7.6), 80% glycerol, 0.03% bromophenol blue, 0.03% xylene cyanol blue) and analyzed by 1% agarose gel electrophoresis, run at 50 mA for 2 h (PEQLAB Biotechnologie). Complex formation was analyzed by autoradiography of the agarose gel for 3 h using a Fuji FLA-3000 R Phosphorimager (Fujifilm); the ratio of free to partially and fully complexed RNA was calculated using the software AIDA (Version 3.45).

Atomic force microscopy

PEI/LNA complexes were analyzed by atomic force microscopy (AFM). Complexes were prepared as described above with 500 ng of LNA 14-mer 17–5p (PO) or LNA 8-mer (PS) and the indicated amounts of PEI F25-LMW (see Fig. 6). Complexes were directly transferred onto a silicon chip by dipping the chip into the complex solution. AFM was performed on a vibration-damped NanoWizard (JPK instruments, Berlin, Germany) as described in detail elsewhere.^38,39. Commercially available pyramidal Si₃N₄ tips (NSC16 AIBS, Micromasch, Estonia) mounted to a cantilever with a length of 230 µm, a resonance frequency of about 170 kHz and a nominal force constant of about 40 N/m were used. To avoid damage of the sample surface, measurements were performed in intermitted contact. The scan speed was proportional to the scan size, and the scan frequency was between 0.5 and 1.5 Hz. Images were obtained by displaying the amplitude signal of the cantilever in the trace direction, and the height signal in the retrace direction, both signals being simultaneously recorded. The complexes were visualized either in height or in amplitude mode (512 × 512 pixel).

Plasmid construction and seed mutagenesis

To introduce a hsa-let-7a site, two cDNA oligonucleotides encoding the let-7a target sequence (MIMAT0000062) were annealed, cleaved with XbaI, and cloned into the unique XbaI site of the reporter plasmid pGL3-Control (Promega, Mannheim, Germany); the XbaI cleavage site is located immediately downstream of the luciferase coding sequence. Insertion of the fragment was identified through linearizing of recombinant plasmids via the unique BlpI site introduced with the insert (see below). Cloning of the insert resulted in two opposite orientations, which were identified by DNA sequencing. Clones giving rise to RNA transcripts with a let-7a target site were classified as “forward orientation”; clones with the opposite insert orientation were classified as “inverted orientation” and were used as controls for the absence of a let-7a target site in the RNA transcript. The two annealed DNA oligonucleotides had the following sequences:

let-7a sense XbaI:

5′-GCTCTAGAGCTGAGGTAGTAGGTTGTATAGTTGCTCAGCGCTCTAGAGC-3′

let-7a antisense XbaI:

5′-GCTCTAGAGCGCTGAGCAACTATACAACCTACTACCTCAGCTCTAGAGC-3′

(XbaI and BlpI sites underlined and the hsa-let-7a sense sequence in italics).

To determine the effects of LNA antiseeds that are directed against mature miR-106 family members on the expression of the p21 protein, the 3′-untranslated region (3′-UTR) of p21 was cloned into the pGL3 control luciferase reporter vector (Promega, Mannheim, Germany) via its XbaI site. The 3′-UTR was initially amplified from K562 genomic DNA with the forward primer 5′-TCTAGACCTCAAAGGCCCGCTCTA-3′ and the reverse primer 5′-TCTAGAGGAGGAGCTGTGAAAGACACA-3′; the amplicon was subcloned into a pCR 2.1-TOPO vector (TOPO TA Cloning^® kit, Invitrogen, Karlsruhe, Germany) prior to insertion into the pGL3 vector. PCR mutagenesis of the two miRNA seed target sites was done according to Brennecke.⁴⁰ The following primers were used for mutagenesis (sites of mutation underlined):

5′-target site forward: 5′-GAAGTAAACAGATGGGACTGTGAAGGGGCCTCACC-3′,

5′- target site reverse: 5′-GGTGAGGCCCCTTCACAGTCCCATCTGTTTACTTC-3′,

3′-target site forward: 5′-CTCCCCAGTTCATTGGACTGTGATTAGCAGCGGAA-3′,

3′-target site reverse: 5′-TTCCGCTGCTAATCACAGTCCAATGAACTGGGGAG-3′. Mutations were verified by DNA sequencing.

Luciferase reporter assays

Luciferase reporter assays were performed using the Promega Luciferase Assay System (Promega, Mannheim, Germany). After aspirating the media, cells were washed with PBS and lysed in 100 µl reporter lysis buffer. In a 96-well plate, 10 µl lysate were mixed with 10 µl substrate, and the luminescence was measured immediately in a Safire^{2 TM} microplate reader (Tecan, Crailsheim, Germany). In the case of the let-7a-responsive reporter assay, the measured luminescence was normalized to that of lysates from cells harbouring the inverted target vector (which was set to 1), yielding the normalized “relative light units” (RLU). In the case of the reporter with p21 3′-UTR, RLU were normalized to the luminescence of the reporter in the presence of the control LNA 14-mer (set to 1) to illustrate the relative weak derepression effects of PEI-complexed oligonucleotides.

Western blotting

Cells were lysed in lysis buffer (125 mM TRIS-HCl pH 6.8, 4% SDS, 1.4 M 2-mercaptoethanol, 0.05% bromophenol blue) and heated at 95°C for 5 min. Samples were loaded onto 15% SDS-polyacrylamide gels (Mini-PROTEAN^® 3 cell mini gel system, BioRad) and run for 1 h at 180 V. Proteins were transferred onto an Immobilon^TM-P PVDF membrane for 30 min at 10 V using a Trans-Blot^® SD Semi-Dry Transfer Cell (BioRad), prior to blocking with 5% milk powder solved in TBST. Primary and secondary antibodies were diluted in TBST 1:200 (p21), 1:5000 (β-Actin) and 1:5000 (goat anti-mouse IgG-HRP, secondary antibody). After a final washing step, blots were incubated with Amersham ECL^TM or ECLplus^TM Western Blotting Detection Reagents according to the manufacturer’s protocol. For detection of chemiluminescence, Kodak^® BioMax^TM light films, Kodak GBX Developer and Replenisher and GBX Fixer and Replenisher were used.

Glossary

Abbreviations:

PO

phosphodiester

3′-UTR

3′-untranslated region

LNA

locked nucleic acid

PEI

polyethylenimine

PS

phosphorothioate

PEI F25-LMW

polyethylenimine F25 low molecular weight

RLU

relative light units

FCS

fetal calf serum

WST-1

water soluble tetrazolium salt 1

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Supplemental Materials

Supplemental materials may be found here: www.landesbioscience.com/journals/rnabiology/article/21165