*3.2. Characterization*

The particles were characterized using a number of techniques including transmission electron microscopy (TEM), dynamic light scattering (DLS), thermogravimetric analysis (TGA), zeta potential measurements, Raman and IR spectroscopy. The particle size and the morphology were analyzed with TEM and DLS. All MSN samples show a spherical particle morphology with a disordered, wormlike pore structure, independent of size (Figure 2). Figure 2a,b shows a size distribution around 160 nm (sample MSN160 nm) for particles obtained with the CTAC synthesis using a molar ratio of TEOS:TEA = 1:10. As the molar ratio of TEOS:TEA was decreased to TEOS:TEA = 1:5 and 1:3, the average particle size decreased to 128 nm (sample MSN130 nm, Figure 2c,d) and 100 nm (sample MSN100 nm, Figure 2e,f), respectively. Particles obtained by the F127 synthesis show a mean particle size of 80 nm (sample MSN80 nm, TEOS:TEA = 1:5, Figure 2g,h). The smallest particles with a mean particle size of 60 nm (sample MSN60 nm, Figure 2i,j) were obtained with F127 and a TEOS:TEA ratio of 1:3. DLS measurements of suspended particles show slightly larger hydrodynamic diameters ranging from 90 to 250 nm (Figure S1) but clearly illustrate the same trend of a decreasing particle size with decreasing TEA concentration.

**Figure 2.** TEM micrographs of (**a**) MSN160 nm, (**c**) MSN130 nm, (**e**) MSN100 nm, (**g**) MSN80 nm, (**i**) MSN60 nm and corresponding particle size distribution histograms obtained from TEM images (**b**,**d**,**f**,**h**,**j**). Scale bar represents 150 nm.

Nitrogen sorption measurements resulted in typical type IV isotherms for all samples, as expected for MSNs (Figure 3a). Surface analysis indicates higher surface areas for particles prepared by the CTAC synthesis (samples MSN160 nm, MSN130 nm and MSN100 nm) ranging between 742–960 m<sup>2</sup>/g, and shows a similar pore size of about 4.8 nm. Slightly smaller surface areas but larger pore sizes were obtained for particles prepared by the F127 synthesis (MSN80 nm, MSN60 nm; 660–685 m<sup>2</sup>/g; pore size 6.0 nm, Figure 3b). Table 1 summarizes the properties of the particles obtained by the different synthesis methods.

**Figure 3.** Characterization ofMSN andMSN-454-GE11. (**a**) Nitrogen sorption isotherms and (**b**) corresponding pore size distributions. For clarity, the pore size distribution curves in panel b are shifted along the *y*-axis. (**c**) Raman spectrum of MSN160 nm. The signal at 2580 cm<sup>−</sup><sup>1</sup> (indicated by \*) indicates the presence of thiol groups. (**d**) IR spectra, (**e**)UV VIS spectra of the capping solution before capping (blackline) and the supernatants after capping, and (**f**) zeta potential measurements of MSN and MSN-454-GE11 at pH = 7.3.


**Table 1.** Synthesis and properties of MSN samples.

a The average particle size was obtained from TEM micrographs by measuring the diameter of around 200 particles of respective samples. b The total pore volume was determined at *p*/*p*0 = 0.9 to exclude contributions of textural porosity. c Data were acquired from the adsorption branch of the nitrogen isotherm.

The presence of the amino and thiol groups originating from functionalization was verified with Raman and IR spectroscopy as well as with TGA. Representative results for sample MSN160 nm are shown in Figure 3c,d, respectively. Raman spectroscopy shows the S–H stretching mode of the thiol groups in the particle shell at 2580 cm<sup>−</sup><sup>1</sup> (Figure 3c), while the primary amines from core functionalization are seen at 1630 cm<sup>−</sup><sup>1</sup> with IR spectroscopy (MSN160 nm, black line, Figure 3d). TGA confirms the inclusion of organic functional groups by a weight loss of 20% (Figure S2). The decomposition of aminopropyl and successively, the mercaptopropyl groups starts at 290 ◦C.

### *3.3. Polymer Capping with 454-PEG and 454-GE11 and Targeting*

The surface of the MSNs was capped with a modularly designed block copolymer 454 to prevent a premature release of the cargo and to enhance endosomal escape for intracellular delivery of miRNA. The structure of the copolymer 454 is shown in Figure S3. This T-shaped polymer 454 consists of a hydrophobic domain in the center made of two oleic acids attached to lysine units. Branched off at each side from the center are two succinyl-tetraethyl-pentamine (Stp) units [47], providing the polymer with a cationic charge. Each end contains a tyrosine trimer coupled to a terminal cysteine unit, potentially allowing additional functionalization. The combination of the central Stp and oleic acid units in 454 is assumed to facilitate interactions with the endosomal membrane, resulting in endosomal destabilization and thus enabling the nucleic acid cytosolic delivery. This block copolymer was successfully used for the formulation of INF-7-polyplex/siRNA vehicles [48] and was further essential for the highly efficient delivery of siRNA using MSNs, reported previously [21].

An anticancer therapy can potentially be improved by implementing cancer-cell-specific targeting molecules to the external surface of a carrier system. It is well known that EGF receptors (EGFR) are concentrated on many cancer cells. The small peptide GE11 has been used successfully before to specifically address EGFR-expressing cells [41,49,50]. Here, we exploit the mercapto residues of the cysteine groups in the capping polymer 454 for binding GE11 via a PEG linker to study the targeted, particle-size-dependent delivery of the miR200c.

For anchoring GE11 to the cysteine units in the block copolymer, we used a mercapto-reactive maleimide-PEG moiety consisting of 28 ethylene glycol monomer units that was previously linked to the GE11 ligand (Mal-PEG-GE11 reagent). The Mal-PEG-GE11 reagen<sup>t</sup> was then mixed with the 454 block-copolymer (454-GE11). Only 0.1 equivalents (eq.; relative to 454-polymer) of the targeting ligand were used for 454-functionalization to maintain free thiol groups at the end of 454. The latter can potentially undergo a disulfide bridging with the terminal mercapto-groups on the outer surface of the MSNs, thus creating a stable copolymer capping. The attachment of the cationic polymer 454

to the surface is additionally electrostatically favored since zeta potential measurements of our pure MSNs reveal a negative surface charge for all samples at a pH higher than 5.5 (Figure S4). To perform the attachment of the premixed 454-GE11 to the surface of MSN, a suspension of MSN at pH 7.3 was mixed with 454-GE11 and the successful capping was confirmed using UV–VIS, IR, and zeta potential measurements. IR spectroscopy shows a significant increase of the CH stretching vibrations between 2850 and 2930 cm<sup>−</sup><sup>1</sup> and of the C-H bending vibrations of 1460 cm<sup>−</sup><sup>1</sup> as compared to the bare MSN particles (red graph in Figure 3d). Furthermore, the MSN-454-GE11 spectrum shows an increase of the N–H bending vibrations at 1650 cm<sup>−</sup><sup>1</sup> and a new signal at 1720 cm<sup>−</sup><sup>1</sup> related to a C=O stretching mode, indicating the multiple amide bonds of the copolymer. UV–VIS measurements were further used to quantify the amount of the 454-GE11 capping agent, which was similar in all samples and ranged from 17 to 20 wt% (Figure 3e, calibration curve Figure S5 and capping amount of different samples, Table S1).

Figure 3f shows the zeta potential measurements of all MSN samples before and after attachment of the 454-GE11 linker (MSN-454-GE11) at pH = 7.3. Besides the size of the nanoparticles, their surface charge is an important parameter for cellular internalization. Our pure MSN samples display a negative surface charge between −37 and −20 mV, as expected for mercapto-covered MSNs. The zeta potential of all capped MSN-454-GE11 samples is very comparable and has increased to about −24 mV when measured at pH 7.3. The observed difference before and after capping can be seen as an additional indication for a successful attachment of the targeted copolymer. The polymer capping did not affect the particle size distribution in water for most of the MSN-454-GE11 samples when measured by DLS. Only the sample with a mean particle size of 60 nm showed some moderate agglomeration (Figure S6).
