*3.4. Cell Internalization*

As documented above, we have now established that our MSN carrier system shows physical properties such as surface area, pore size, zeta potential and capping concentration that are all very comparable, thus, leaving solely the particle size as a distinct variable. First, flow cytometry and confocal fluorescence microscopy data were used to investigate how the targeting ligand concentration as well as particle size might influence the cell internalization of the fully assembled MSN-454-GE11 samples.

Figure 4 shows the cellular internalization of the MSNs by T24 cells as examined via flow cytometry after a 45 min incubation time. MSN-454-GE11 samples were labeled with the Atto-633-carboxy fluorescent dye, which was covalently bound to the amino groups in the core of the particles. In order to investigate the best targeting concentration, we mixed different weight equivalents of targeting ligand GE11 with the 454 polymer in the range from 0.1 to 1 eq. before attaching the polymer to the MSN samples. Maleimide-PEG without ligand was used as negative control for polymer capping without targeting ligand (sample MSN-454-PEG). A significant targeting effect can be seen in Figure 4a when 0.1 eq. of the targeting ligand was used, while only a minor improvement is observed at higher concentrations.

Figure 4b documents the particle-size-dependent association of MSN-454-GE11 to cells as investigated via flow cytometry. Flow cytometry showed a similar degree of cell association for all samples, while a slight trend towards stronger association can be seen for smaller particles (quantification and statistical analysis can be found in SI, Figure S7). Since flow cytometry is not able to differentiate internalization from externally adhering particles, we followed this process also with confocal fluorescence microscopy as shown in Figure 5 (for enlarged images with additional orthogonal views of each image showing the particle internalization, see Figure S8, Supplementary Materials). Images were subsequently analyzed via the digital method 'Particle\_in\_Cell-3D' to quantify the cellular uptake of the differently sized MSN vectors. After imaging cells directly after an incubation time of 45 min, we found that only MSN160 nm-454-GE11 particles had penetrated through the cell membrane and were truly internalized. Other smaller MSN-454-GE11 particles were only attached to the outer cell membrane at this time. In contrast, all particle sizes were internalized after a 6 h incubation time. However, also after this time, MSN160 nm-454-GE11 particles showed the highest number of internalized particles of all samples (Figure S9).

**Figure 4.** Cellular internalization determined by flow cytometry. Cells were incubated for 45 min, washed and analyzed. (**a**) MSN160 nm-454-PEG was used for passive and MSN160 nm-454-GE11 for receptor-mediated uptake. 454 was functionalized with different equivalents of Mal-PEG (454-PEG) and Mal-PEG-GE11 reagen<sup>t</sup> (454-GE11), respectively, in the range from 0.1 to 1 eq. The mean fluorescence intensity of the Atto-633 signal (MFI APC) represents the amount of internalized nanoparticles. A high MFI corresponds to a large cell uptake of MSN. For statistical analysis, a two-tailed t-test was performed (*n* = 3, mean ± SD, \* *p* < 0.05, \*\* *p* < 0.01) (**b**) Histograms of cellular internalization of Atto-633 labeled MSN-454-GE11 after 45 min incubation with particle sizes in the range of 160 nm to 60 nm and negative control (NC). The Atto-633 intensity (APC-A channel) is plotted against the number of events detected ('count'). 'Count' represents cumulative counts of cells with indicated Atto-633 fluorescence after appropriate gating by forward/sideward scatter and pulse width. For statistical analysis, see Figure S7.

**Figure 5.** Representative confocal fluorescence microscopy images of Atto-633 labeled MSN-454-GE11 (red) with particle sizes in the range of 160 nm to 60 nm after 45 min (upper panel) and 6 h (lower panel) incubation on WGA488-stained T24 cells (green). For a statistical evaluation of particle-size-dependent uptake, see Figure S9, SI.

## *3.5. Loading of RNA*

Loading of the genetic material into our MSN samples was performed with all samples prior to the attachment of the block copolymer. The adsorption of the miRNA is mainly driven by electrostatic interactions with the cationic, amino-functionalized core of the MSNs. The negatively charged thiol groups at the particle periphery are expected to minimize any external adsorption. For the best results, we performed the RNA adsorption in MES buffer at pH = 5. The MSN particles show a positive zeta potential at this pH due to protonation of the amino groups (zeta potential titration, see Figure S4).

Here, aliquots of 50 μg mg<sup>−</sup><sup>1</sup> RNA/MSN were used for loading and the actual uptake was calculated by difference measurements by determining the remaining RNA concentration in the supernatant. All samples were able to adsorb the amount offered, as no residual RNA could be detected in the supernatant after loading times as short as 30 min. The samples were subsequently capped with 454-GE11 to obtain the final MSN-454-GE11 samples used for subsequent experiments. The stable binding of RNA in these samples is reflected in gel shift results, as shown in Figure 6a). Only little RNA elution is visible upon applying a voltage of 100 V for 1 h to the gel.

**Figure 6.** (**a**) Gel electrophoresis of samples MSN160 nm-454-GE11 (1), MSN130 nm-454-GE11 (2), MSN100 nm-454-GE11 (3), MSN80 nm-454-GE11 (4), and MSN60 nm-454-GE11 (5). (**b**) Gene silencing of T24/eGFPLuc-200cT cells transfected with MSN-454-GE11 containing either Ctrl (white) or miR200c (grey). After an incubation time of 45 min, cells were washed. At 48 h after transfection, gene-silencing effects were analyzed.
