*3.6. Gene Silencing*

As a proof of concept, the gene silencing of the eGFP-luciferase reporter gene was performed using T24/eGFPLuc-200cT cells. This cell line stably expresses the eGFP-luciferase fusion protein, while simultaneously featuring a miR200c target site on the expressed mRNA. Therefore, the gene expression of eGFP-luciferase fusion protein can be blocked by the delivery of miR200c. The luciferase signal was thus used to evaluate the gene-silencing efficacy of miR200c. In order to simulate the dynamic conditions of drug delivery and to avoid a prolonged overexposure of the cells with particles suspended in the medium, we have used extremely short incubation times of 45 min. Cells were subsequently washed by exchanging the medium and the effect of this short exposure was then analyzed after time spans referred to as transfection time in the following. Cells were transfected with MSN and MSN-454-GE11 of different sizes loaded with either a synthetic miR200c mimic or a control siRNA without any target gene (Ctrl, Figure 6b). Gene silencing was never observed with the pure MSN samples (missing the 454 polymer construct) loaded with miR200c for silencing (Figure S10a). This might be caused by a premature release of the RNA or by endosomal trapping of the MSN samples missing the 454 polymer.

Similarly, none of the smaller sized MSN-454-GE11 samples showed any significant silencing efficacy. In contrast, with the larger 160 nm targeted polymer-capped MSN160 nm-454-GE11 sample, a luciferase gene knockdown of up to 65% was observed. These findings are in good agreemen<sup>t</sup> with the confocal fluorescence microscopy images, which showed the fastest internalization for particles with a size of 160 nm.

Flow cytometry data showed an improved cellular internalization of the targeted MSN160 nm-454-GE11 vector in comparison with the non-targeted control MSN-454-PEG when 0.1 eq. of the targeting ligand was used. This targeted vector showed a higher silencing e fficacy compared to the non-targeted control, which is in good agreemen<sup>t</sup> with the increased uptake observed via flow cytometry.

As expected, the same sample loaded with just scrambled RNA (Ctrl) did not show a significant unspecific knockdown, thus excluding major toxic e ffects induced by the carrier system itself. Systematic cell viability studies using an MTT assay also show a good tolerance of the cells towards these samples (Figure S11).

These results show that a particle size of about 160 nm in combination with the copolymer 454 enables the fastest cell internalization and also the best transfection e fficacy when short incubation times of only 45 min are used. Inhibition experiments addressing specific endocytic pathways were inconclusive with respect to a preferred mechanism for a certain particle size. However, endocytosis was substantially more blocked for the larger particles as compared to the 60 nm particles (see Figure S12, Supplementary Materials). Our data also confirm the endosomolytic activity of the 454 polymer since cell adherence/uptake was observed for all pure MSN samples (flow cytometry of pure MSN samples in Figure S13), but no transfection occurred without the 454 polymer (see Figure S10a). As described above, we further found that only MSN160 nm-454-GE11 showed gene-silencing activity, while all smaller MSN-454-GE11 samples were not active after the short incubation period applied here. Confocal fluorescence microscopy showed that MSN160 nm-454-GE11 particles were already internalized after 45 min, while all other MSN-454-GE11 samples needed much longer incubation times to achieve internalization. Thus, particle internalization of the smaller particles is likely too slow and prevents a larger e fficacy.

Recently, miR200c was delivered by some of us using GE11 modified 454 polyplexes with a similar size of 120–150 nm [41]. The gene silencing e fficiency achieved in the same T24/eGFPLuc-200cT cells with these polyplexes was around 60%, which is in the same range as the knockdown e fficiency of the MSN-454-GE11 constructs presented here. This comparison suggests that the size of the targeted constructs is a determining feature controlling the knockdown e fficiency. As MSN160 nm-454-GE11 constructs were shown to o ffer the best cellular-uptake behavior and gene-silencing e fficiency, they were used for the following cell migration and cell cycle experiments.
