Previous reports have described gymnotic delivery at low-micromolar concentrations of ONs, usually with extended incubation periods, to achieve effective gapmer antisense ON activity in cells.5-9
However, only recently has gymnotic delivery started to be considered as a useful approach for screening of ONs as potential therapeutics, since data obtained by gymnotic delivery of ONs correlated much better with in vivo results as compared with ON cell delivery aided by transfection agents, such as Lipofectamine 2000.5,10,11
Practically all the previous gymnotic delivery experiments have focused on antisense targeting of mRNA, whereas unaided delivery of anti-miRs has not been studied, except for Lys-modified PNA ONs as previously reported by us.15,16,40
The data presented in show that in addition to Cys-K-PNA-K3
, OMe PO, OMe PS, LNA/OMe PO and LNA/OMe PS are also all capable of rapidly entering Huh7 liver cells without the use of transfection agents and inhibiting miR-122 at sub-micromolar ON concentrations over a sustained period. Several interesting observations can be made about the rank order of activity of the various ONs. First, the 23-mer LNA/OMe PO was more active than the OMe PO 31-mer with lipofection (). This is consistent with our previous finding that LNA/OMe PO was able to upregulate miR-122 target mRNAs when lipofected into Huh7 cells or primary rat hepatocytes, while the OMe PO anti-miR failed to upregulate such mRNAs in a dose-dependent manner,15
and is validation for the use of the dual-luciferase assay for screening of anti-miR activity. Interestingly, OMe PO and LNA/OMe PO ONs showed similar levels of activity when delivered in the absence of transfection agents (). The OMe PO ON is a 31-mer previously validated against miR-122.20
The 8 additional residues beyond the normal 23-mer anti-miR are complementary to the pre-miR122 sequence, as it was reported that a 23-mer was poorly active.45
Activity of both these all PO ONs, as measured by the RLuc/Fluc ratio, started to decline after 3 d of cell incubation, which can probably be explained by the known instability of phosphodiester linkages to nuclease degradation, whereas the PNA ON and ONs containing PS linkages continued to show an increased RLuc/Fluc ratio on day 4. Although when delivered in the absence of transfection agent, addition of LNA residues did not significantly enhance activity over OMe in the PO context, by contrast significantly increased anti-miR activity was obtained by addition of LNA residues to the OMe ON in the PS context ().
Surprisingly, the commercially obtained LNA/DNA PO anti-miR-122 knockdown probe failed to be internalized and inhibit miR-122 by gymnotic delivery at the tested concentrations (), whereas it was functional when lipofected into cells, albeit with less potency than for the LNA/OMe PO ON (). We have for several years advocated the use of mixmers of LNA and OMe as steric blocking agents,15,46-48
since the binding strength of OMe residues is higher than that of 2’-deoxynucleotides, because both OMe and LNA residues intrinsically adopt a C
, ribose-like conformation. Since the LNA content of the LNA/DNA PO ON is not revealed by the supplier, we cannot rule out the possibility that the increased potency is partly due to a significantly different LNA content between the two ONs, but it seems likely that the commercial LNA/DNA PO probe will have a similar LNA content to the LNA/OMe PO anti-miR (40% LNA). Therefore we believe that the increased potency of LNA/OMe PO over LNA/DNA PO is mostly due to the presence of OMe residues.
The highest potency anti-miR was found to be the LNA/OMe PS ON. Whereas LNA/MOE PS anti-miRs have been described for cellular and in vivo use targeting miR-122,25,37
there is only one previous example of an LNA/OMe PS ON used as an anti-miR, but only with delivery by transfection agent into cells.37
By contrast, LNA/DNA PS ONs targeting miRNAs are well established and have indeed entered clinical trials (Santaris Pharma News Release 23 September 2010; www.santaris.com
). Further, LNA/DNA PS gapmers were shown to be subject to gymnosis in mRNA targeting studies.5,10,11
In our study we have used an LNA/OMe PS anti-miR containing only ~15% LNA residues that are all located in the region complementary to the miR-122 seed sequence (). When lipofected into cells, the LNA/OMe PS anti-miR somewhat outperformed the LNA/OMe PO containing 40% LNA we previously reported ( and ). However, the dramatic enhancement in anti-miR activity obtained for the LNA/OMe PS ON over the LNA/OMe PO ON when delivered by gymnosis () suggests that the presence of the PS linkages enhances productive cellular delivery leading to miRNA inhibition. Interestingly, as seen in and , the observed activity for the LNA/OMe PS ON reached saturation. It is unlikely that this is due to saturation of cellular uptake, since similar saturation was observed for LNA/OMe PS delivered by lipofection. Although we did not correlate precisely anti-miR activity saturation in the dual-luciferase assay with the level of miRNA inhibition, our results indicate that most of the miR-122 is inhibited at the saturating concentrations tested. A significant amount of further work will be necessary to determine the optimum LNA content and placement within an OMe anti-miR sequence, as well as the ON length optimum, and also to compare in greater detail LNA/DNA PS and LNA/OMe PS mixmers, as well as possible mixmers with all three nucleoside types. However, the potent and long-lasting gymnotic anti-miR activity at sub-micromolar concentrations in cells suggests that LNA/OMe PS ONs should be given greater consideration for in vivo use.
While long incubation times (days) with high amounts of ONs (low-micromolar concentrations) in continuous cell culture were necessary for gapmer antisense inhibition of mRNA by gymnotic delivery,5-12
here we showed rapid (only 4 h) gymnotic delivery at sub-micromolar concentrations of Cys-K-PNA-K3
, OMe and LNA/OMe PO and PS ONs for miR-122 targeting. Also anti-miR122 activity was sustained for at least 4 d after cell incubation for many of the ONs. Zhang et al. showed that gymnotic delivery of LNA/DNA PS gapmers could be achieved in a wide range of tumor cell lines.11
We previously reported inhibition of miR-155 function in B-cells and in mice by PNA ONs in the absence of transfection agents.16
We have also found recently that at least for PNA ONs, gymnotic uptake does not extend only to Huh7 cells and B-cells but also to HEK293ET cells and H2K mdx mice-derived myoblasts and myotubes (Torres et al., manuscript in preparation). The mechanisms of gymnotic cellular uptake and localization of ONs, such as the highly active Cys-K-PNA-K3
and LNA/OMe PS, and how such anti-miRs find and interact with their miRNA targets still remains to be explored.
Stein and colleagues showed that a 5′-FAM-labeled LNA/DNA PS ON delivered by gymnosis was localized in the cytosol together with GW/P-bodies (site for RNA silencing), while the same ON after lipofection showed bright fluorescence in the nucleus instead.5
By contrast, Zhang et al.11
correlated strong gapmer antisense target down-modulation with nuclear localization following gymnosis of their FAM-labeled LNA/DNA PS ON, although at earlier time-points these ONs localized in the perinuclear space in the cytosol. Koller and colleagues failed to detect MOE/DNA PS gapmer ONs in P-Bodies after gymnotic delivery and instead the bulk of the ON was found in lysosomes after 24 h.12
They also showed that uptake of MOE/DNA PS ON into MHT hepatocarcinoma derived liver cells seemed to be through a clathrin- and caveolae-independent but AP2M1-dependent endocytotic mechanism.12
Since miRNAs have been found within the endosomal pathway49
a very likely possibility is that anti-miR ONs delivered by gymnosis are taken up by cells via a rapid endocytosis mechanism, travel through the whole endocytotic pathway, and are able at some point to encounter the miRNA within one or more of these endosomal compartments. This hypothesis may explain why such short ON incubation times (within 4 h) are effective for miR-122 inhibition in the absence of transfection agents. Our recent investigations into understanding the mechanisms of anti-miR uptake and miRNA targeting will be reported shortly which we hope will provide further insight into improving anti-miR design (Torres et al. manuscript in preparation).