2018, Vol.47, No.9

We developed a dendritic molecular glue (GlueSS-BP) bearing multiple guanidinium ion (Gu+) pendants via a disulfide (SS) spacer as a carrier for intracellular siRNA delivery. GlueSS-BP adheres tightly to siRNA and forms a stable conjugate through multivalent salt-bridge formation between the Gu+ pendants of GlueSS-BP and phosphate ions (PO4) of siRNA. The GlueSS-BP/siRNA conjugate enters living cells via direct translocation. In the cytoplasm, the SS spacers in GlueSS-BP are reductively cleaved off by glutathione (GSH) to efficiently release siRNA, thereby inducing RNA interference (RNAi). Products, formed upon reductive cleavage of GlueSS-BP in the cytoplasm, barely cause cytotoxicity.

Intracellular delivery of small interfering RNA (siRNA), which selectively suppresses expression of a certain gene through RNA interference (RNAi),1 is a promising approach to the curing of both inherited and acquired diseases caused by genetic disorders. A variety of carriers for siRNA have been reported, such as those based on lipids, polymers, and nanoparticles,2 most of which, however, are taken up via endocytosis into living cells.2 This is unfavorable, since siRNA trapped in endosomes is subsequently degraded in lysosomes, resulting in a poor gene knockdown efficiency. Because only cytoplasmic siRNA can induce RNAi,1 carriers that can directly penetrate the cell membrane to deliver siRNA to the cytoplasm are advantageous. Nevertheless, only a few peptide-based carriers have been reported to translocate siRNA directly into cells,3 and it still remains a big challenge to design such carriers with other molecular motifs. A promising strategy is to employ guanidinium ion (Gu+)-rich molecules4 that can strongly bind to and destabilize negatively charged cell membranes, thereby enabling direct intracellular translocation of siRNA. However, it should be noted that polycations, which possess a high efficacy for cellular uptake,5 often cause serious cytotoxicity.6

Herein, we report a dendritic molecular glue GlueSS-BP (Figure 1) that can efficiently translocate siRNA directly into the cytoplasm (Figure 2) with minimal cytotoxicity. Previously, we developed water-soluble molecular glues711 bearing multiple Gu+ pendants that adhere tightly to proteins,8 nucleic acids,9 phospholipid membranes,10 and clay nanosheets11 via the formation of multiple salt bridges between their Gu+ pendants and oxyanionic groups on the targets.12 In our previous communication,9a we reported that a siRNA motif can efficiently template oxidative polymerization of oligomeric linear molecular glues with thiol (SH) groups at both termini, affording stable conjugates called “siRNA-nanocaplets”. Although the siRNA-nanocaplets are taken up into living cells via endocytosis, RNAi is successfully induced because the nanocaplets release siRNA by reductive cleavage of their disulfide (SS) bonds in the cytoplasm that contains glutathione (GSH) abundantly. On the basis of this finding, we designed, as a reductively cleavable dendritic analogue, GlueSS-BP (Figure 1) that bears Gu+ pendants via a SS spacer. Unlike siRNA-nanocaplets, the GlueSS-BP/siRNA conjugate can enter living cells via direct translocation (Figure 2). In the cytoplasm, the SS spacers in GlueSS-BP are supposedly cleaved off to release siRNA, thereby suppressing target gene expression by RNAi (Figure 2) with minimal cytotoxicity.13 We also demonstrated the gene knockdown using the GlueSS-BP/siRNA conjugate for different cell lines.

GlueSS-BP and Glue-BP (Figure 1), a reference dendrimer without SS spacers, were synthesized according to the procedures described in the Supporting Information and characterized unambiguously using several analytical methods.14 In agarose gel electrophoresis, siRNA (1.5 µM) is known to migrate according to its net negative charges (Figure 3a, [Gu+]/[PO4] = 0). However, siRNA did not migrate when mixed with GlueSS-BP ([Gu+]/[PO4] > 20), indicating that GlueSS-BP adheres to siRNA and neutralizes its negative charges. Meanwhile, when the GlueSS-BP/siRNA conjugates, prepared at [Gu+]/[PO4] = 0–60, were incubated with GSH (10 mM) at 37 °C for 2 h in HEPES buffer (10 mM, pH 7.4), the resulting electrophoresis profiles were all identical to that of the intact siRNA (Figure 3b), indicating the release of siRNA by reductive cleavage of the SS spacers in GlueSS-BP. Dynamic light scattering (DLS) and zeta potential measurements also supported formation of the GlueSS-BP/siRNA conjugate (Dh = 19.9 ± 4.6 nm, ζ = 23 ± 9 mV; Figures S4 and S5, respectively).14

We found that GlueSS-BP carries siRNA into the cytoplasm of living cells. For visualization, we used siRNA fluorescently labeled with Cy3 (Cy3siRNA). Human epithelial carcinoma HeLa cells were incubated at 37 °C for 1 h in a fetal bovine serum (FBS)-free Dulbecco's modified Eagle's medium (DMEM) containing Cy3siRNA (100 nM) and GlueSS-BP (5 µM). Then, the cell sample was rinsed with Dulbecco's phosphate buffered saline (D-PBS) and subjected to confocal laser scanning microscopy (λext = 435 nm). The cells displayed a fluorescence emission assignable to Cy3siRNA from their interior (Figure 4a). The diffuse fluorescence profile (Figure 4a) indicates the release of siRNA in the cytoplasm. In contrast, an analogous cell sample prepared using Glue-BP (5 µM) instead of GlueSS-BP displayed a punctate fluorescence emission (Figure 4b). HeLa cells, pre-incubated for 24 h at 37 °C in DMEM (10% FBS) containing buthionine sulfoximine (BSO, 100 µM) as an inhibitor for the GSH biosynthesis before the addition of GlueSS-BP and Cy3siRNA,15 also showed a punctate fluorescence profile (Figure 4c). These observations clearly indicate that cleavage of the SS spacers in GlueSS-BP by GSH in the cytoplasm is essential for the liberation of siRNA.

To our surprise, HeLa cells incubated with GlueSS-BP (5 µM) and Cy3siRNA (100 nM) for 1 h at 4 °C also emitted fluorescence from their interior (Figure 4d). In contrast, no cellular uptake of Cy3siRNA took place at 4 °C using commercial transfection reagent Lipofectamine 2000 (LF2000, Figure S7),14 which is known to carry nucleic acids inside cells via endocytosis.16 Since energy-dependent cellular processes such as endocytosis are suppressed at 4 °C,17 these results indicate that the GlueSS-BP/Cy3siRNA conjugate enters the cells by direct penetration of the cell membrane (Figure 2). The punctate fluorescence emission observed in Figure 4d is most likely due to the aggregates of the GlueSS-BP/Cy3siRNA located in the cytoplasm. Taking into account that the interaction between Gu+ and anionic moieties on the cell surface is essential for direct translocation of Gu+-rich molecules,5 GlueSS-BP, even after the conjugation with Cy3siRNA, possibly carries a certain number of unbound Gu+ pendants. This seems typical of dendritic molecular glues as carriers, considering that siRNA-nanocaplets consisting of oligomeric linear molecular glues do not migrate directly into the cytoplasm.9a

We then examined the possibility of gene knockdown. Mutant HeLa cells stably expressing luciferase (HeLa-luc) were incubated at 37 °C in a FBS-free DMEM containing a mixture of siRNA for luciferase (siRNAluc; 100 nM) and GlueSS-BP (10 µM) for 4 h, followed by DMEM (10% FBS) for 24 h. The cell sample was then subjected to a luciferase activity assay using a Luciferase Assay System (Promega). As a result, the cells exhibited smaller luciferase activity (34%; Figure 5a, ii, green) than that of untreated HeLa-luc cells. Meanwhile, the suppression did not occur with mismatch siRNA (mis-siRNA; 100 nM) incapable of inducing RNAi for the luciferase gene (Figure 5a, ii, blue). These results suggest the occurrence of gene knockdown by RNAi with the GlueSS-BP/siRNAluc conjugate. When Glue-BP (10 µM) was used instead of GlueSS-BP, not only siRNAluc but also mis-siRNA suppressed the luciferase activity (19% and 29%; Figure 5a, i, green and blue, respectively), indicating that Glue-BP (10 µM) causes acute cell death. In fact, Glue-BP exhibited a LC50 (median lethal concentration) value of 3.9 µM (Figure 6, blue), which is smaller than that of GlueSS-BP (LC50 = 11 µM, Figure 6, red), as evaluated by Cell Counting kit-8 (CCK-8).14 As expected, LF2000/siRNAluc efficiently suppressed the luciferase activity (15%; Figure 5a, iii, green) with minimal cytotoxicity (Figure 5a, iii, blue). Even in the presence of FBS (10%), the GlueSS-BP/siRNAluc conjugate knocked down the luciferase gene (73%; Figure 5b, i), whereas no gene knockdown occurred with LF2000/siRNAluc (Figure 5b, ii). Namely, the GlueSS-BP/siRNA conjugate is more stable than LF2000/siRNA under physiological conditions. Similar gene knockdown experiments with GlueSS-BP (10 µM) and siRNAluc (100 nM) were carried out using luciferase-expressing mutants of mouse melanoma B16F10 cells (B16F10-luc), human hepatocellular carcinoma Huh-7 cells (Huh-7-luc), and human lung adenocarcinoma A549 cells (A549-luc). As shown in Figure 5c, all the examined cell lines showed suppressed luciferase activity without serious cytotoxicity.

In conclusion, we developed a reductively cleavable dendritic molecular glue GlueSS-BP (Figure 1) as a carrier for intracellular siRNA delivery. We demonstrated that GlueSS-BP delivers siRNA into the cytoplasm of living cells via direct translocation (Figure 2), which is favorable for efficient induction of RNAi. GlueSS-BP is promising as a carrier for siRNA owing to its high gene knockdown efficiency, low toxicity, and applicability to a wide range of cell lines. Incorporating a ligand moiety to the GlueSS-BP/siRNA conjugate for site-selective delivery is an interesting subject worthy of further investigation.

We acknowledge the Center for NanoBio Integration, the University of Tokyo. This work was supported by Grant-in-Aid for Young Scientist (B) (26810046) to K.O. We appreciate Dr. T. Suzuki and Prof. T. Suzuki (the University of Tokyo) for electrophoresis experiments. We appreciate Prof. K. Miyata and K. Kataoka (the University of Tokyo) for gene knockdown assays. We appreciate Mr. N. Yoshinaga and Prof. H. Cabral (the University of Tokyo) for DLS and zeta potential measurements. R.M. thanks the Research Fellowships of Japan Society for the Promotion of Science (JSPS) for Young Scientists and the Program for Leading Graduate Schools (GPLLI).

Supporting Information is available on http://dx.doi.org/10.1246/cl.180551.

K. Okuro

T. Aida

H. Nemoto

R. Mogaki