Development of Silica-Based Biodegradable Submicrometric Carriers and Investigating Their Characteristics as in Vitro Delivery Vehicles

Nanostructured silica (SiO2)-based materials are attractive carriers for the delivery of bioactive compounds into cells. In this study, we developed hollow submicrometric particles composed of SiO2 capsules that were separately loaded with various bioactive molecules such as dextran, proteins, and nucleic acids. The structural characterization of the reported carriers was conducted using transmission and scanning electron microscopies (TEM/SEM), confocal laser scanning microscopy (CLSM), and dynamic light scattering (DLS). Moreover, the interaction of the developed carriers with cell lines was studied using standard viability, proliferation, and uptake assays. The submicrometric SiO2-based capsules loaded with DNA plasmid encoding green fluorescence proteins (GFP) were used to transfect cell lines. The obtained results were compared with studies made with similar capsules composed of polymers and show that SiO2-based capsules provide better transfection rates on the costs of higher toxicity.

twice with MilliQ water via centrifugation at 1500 rpm for 45 min and discarding of the supernatant, and were finally dispersed in 500 μL of MilliQ water.

Structural characterization of the obtained materials
The whole synthesis process was monitored with Dynamic Light Scattering (DLS) and Laser Doppler Anemometry (LDA) using a commercial setup (Zetasizer Nano ZS, Red badge, ZEN3600, Malvern; 173° backscatter settings, 633 nm laser) for all types of capsules loaded with DEX-blue. As it can be seen in Figure S1, no aggregation was observed after the different synthesis steps, and the SiO2 capsules showed a negative zeta-potential until the final coating with PARG. The evolution of the hydrodynamic diameter dh and the zeta-potential ζ is illustrated for SiO2 capsules with low amount of TEOS and encapsulated DEX-blue. Figure S1. Evolution of A. the hydrodynamic diameter dh and B. the zeta-potential ζ (measured at pH = 7) for SiO2-based capsules (low TEOS) after the different synthesis steps. Empty symbols correspond to the dissolved capsules. The error bars represent the standard deviation values from three independent measurements. For the zeta-potential the error bars are smaller than the size of the symbols. ʺSt. coresʺ are CaCO3 cores stabilized with 5 kDa CH3O-PEG-SH. "Dis. cores" are SiO2 shells after dissolution of CaCO3. SiO2-based submicrometric capsules loaded with other cargos showed a similar behavior.
Polyelectrolyte capsules (DEXS/PARG)4 and (PSS/PAH)4 also did not show significant aggregation during the different layer deposition. The evolution of the hydrodynamic diameter and the zeta-potential after the subsequent polyelectrolyte layer deposition are shown in Figures S2 and S3, illustrated again for DEX-blue-loaded capsules. In general terms, solutions with lower polyelectrolyte (2 mg/mL), and higher NaCl (0.5 M) concentrations, which are normally used for synthesis of micrometric polyelectrolyte capsules, gave rise to aggregated particles [5]. With our reported polyelectrolyte solutions (10 mg/mL, NaCl = 0.05 M for PSS, PAH DEXS, and 5 mg/mL, NaCl = 0.05 M for PARG) only minor increase of hydrodynamic diameter was observed for the non-biodegradable submicrometric capsules after the first B Synthesis steps and second PAH deposition, whereas no major aggregation was observed for the biodegradable capsules.
Capsules loaded with other cargos showed a similar behavior.         capsules loaded with DQ-Ovalbumin (excitation at 488 nm, emission with band pass (BP) 505-580 nm and LP 620). The green fluorescence signal is coming from the not fully quenched DQ-OVA, red fluorescence signal is coming from the dye dimmers [7,8] Both signals are colocalized within one non-degraded capsule. Capsules were dispersed in water. The scale bars correspond to 20 μm.
Scanning electron microscopy (SEM) was carried out on an Inspect SEM (FEI, USA) with an acceleration voltage of 20 kV. The day before measurements, 35 μL of submicrometric capsules dispersed in water were dropped onto a glass cover slip and let dry overnight. The next day, the dry samples were coated with a thin film of gold and imaged with SEM.
Capsules were also analyzed with transmission electron microscopy (TEM Philips 200CM, SEM-FEG Hitachi S4800). Images of individual capsules are shown in Figure S11. Capsule diameters dc as obtained from the microscopy data as well as the final DLS and zeta-potential data of all capsules are given in

Estimation of the concentration of encapsulated cargo
Given the small size of the obtained capsules we were not able to reliably determine their concentration with a Neubauer counting chamber. Thus, capsule doses were quantified by the amount of encapsulated cargo. Therefore, uptake, cytotoxicity, proliferation, and transfection experiments were carried out with capsules containing the same amount of encapsulated molecular cargo.

Estimation of the amount of DEX-blue inside capsules:
The amount of encapsulated DEX-blue was estimated by fluorescence measurements. To avoid possible interference of pronase, the background spectra of DEX-blue and pronase were first collected. For this, a DEXS-blue solution with a stock concentration of 6.5 mg/mL was prepared, and it was further diluted up to CDEX-blue = 0.1 mg/mL, with a final volume V = 100 μL, as follows: (i) DEX-blue (1.5 μL, 6.5 mg/mL) was diluted in water (98.5 μL), (ii) DEX-blue (1.5 μL) was diluted in pronase (98.5 μL, 2 mg/mL), and (iii) water (1.5 μL) was mixed with pronase (98.5 μL, 2 mg/mL). Afterwards, the fluorescence spectra of all three samples were recorded with a fluorimeter (Fluorolog-3, Horiba JOBIN YVON) using the same conditions (excitation/emission wavelength λexc/λem = 400/420 nm, respectively) at the same day. The resulting spectra are shown in Figure S12. Figure S12. Fluorescence spectra I() of DEX-blue mixed with pronase. Pronase and DEX-blue are the controls λexc=400 nm.
No fluorescence signal was detected from the pronase sample. The DEX-blue + pronase sample showed slightly lower fluorescence signal than the DEX-blue sample. These data suggest that pronase does not give additional signal and, thus, it is reasonable to use this fluorescence method to quantify the amount of encapsulated DEX-blue.
To determine the encapsulated DEX-blue concentration, first a calibration curve of DEX-Blue was determined. For this, the fluorescence signal of DEX-blue solutions was recorded at different dye concentrations (from CDEX-blue = 0 to 100 μg/mL). The fluorescence intensity I420 of all dilutions at λexc = 400 nm was then plotted versus the dye concentration CDEX-blue ( Figure S13).   the assumption that such figures are similar for all types of capsules was considered a good approximation. In Table S1 the amount of encapsulated DEX-blue in the stock solutions of the 4 different types of capsules is summarized. This assay kit is based on the fluorescence of a dye that links to the DNA. The concentration of DNA can thus be quantified with a fluorimeter. The possible interference of pronase and DNA was also taken into account. For this, DNA was diluted in pronase (2 mg/mL) up to a concentration of 50 μg/mL. This DNApronase solution was kept for 24 hours at 37 °C, 5% CO2 in an incubator. The amount of DNA was then quantified with the Quant-IT RiboGreen RNA Assay Kit, according to the protocol of the product, which will be described later. As controls, only pronase and only DNA at the same concentrations were used ( Figure S14).    (Table S2). For the in vitro transfection, the capsules were added to ensure a DNA amount mDNA/cell of 50 pg of DNA per cell. Since we worked with a number of N = 40,000 cells/well (for transfection experiments), the total mass of DNA needed per well then was: mDNA = mDNA/cellNcell = 50 pg  40000 = 210 6 pg.
The required volume of the capsule solution, illustrated for the (DEXS-PARG)4 capsules, can then be calculated as: Such figures would also correspond to around 15 pg DEX-blue per cell. Given that DNA is expected not to show appreciable toxic effects, with compared with DEX-blue, such estimation can be used to compare toxicity values.

Cytotoxicity studies
Cell viability studies were performed with a fluorescence-based approach using resazurin similar to previously published protocols. [9] Resazurin is a non-toxic non-fluorescent compound, which in living cells is converted into fluorescent resorufin. with the viability V of cells [9,11]. Each measurement was repeated three times to obtain the mean value and the standard deviation. The mean value of the fluorescence intensity was normalized to the fluorescence of cells that had not been exposed to capsules, and was plotted against the concentration of the capsules (Figures S16 -S19).   After incubation for 24 h in complete medium.
Additionally, the cell viability of SiO2 capsules (low and high TEOS) co-loaded with DNA and DEX-blue was examined (Figures S20-S21).  Besides the side-and forward-scattering signals, the blue fluorescence originating from internalized capsules was also recorded for all events. This fluorescence Iblue corresponded to cells associated with capsules. The blue fluorescence intensity distribution N(Iblue) is plotted in Figure S23.  (Figures S24 -S25). The auto-fluorescence from untreated cells was subtracted by counting the fluorescence exceeding the auto-fluorescence signal.

Proliferation studies
The effect of SiO2 capsules (low and high TEOS) on the cell proliferation was studied by directly measuring the DNA synthesis according to previously published protocols. [9] This can be quantified by the incorporation of the thymidine-analog EdU (5-ethynyl-2'-deoxyuridine). The detection of EdU can be made by a copper-catalyzed click-reaction between its alkyne group and an azide group-containing fluorophore.
HeLa cells were seeded into 96-well plates (5,000 cells/well) and left overnight. Afterwards, the growth medium was exchanged with the medium containing SiO2 (low and high TEOS) capsules loaded with DEX-AF647 at different concentrations (mDEX-AF647 = 0-3 pg/cell). Colchicine (Sigma Aldrich, #C3915), a compound suppressing cell proliferation via inhibition of chromosome segregation during mitosis [15] was used at a concentration of 10 μM as negative control. HeLa cells were incubated with the different  Figure S28) [9]. Finally, the fraction p of proliferated cells from all cells was calculated for each image and averaged for each concentration. Figure S29 shows the mean values ± standard deviations resulting from three individual experiments [9].

Capsule degradation
In order to trigger the mechanism of the protein expression upon delivery of DNA plasmids, the DNA has to be delivered to the nuclei of cells. It is known that capsules are internalized by cells and afterwards they are located inside the endocytic vesicles [16] and, thus, not free in the cytosol. These vesicles have different environment than the cytosol and their membrane is permeable neither for many types of delivered cargo nor the products of the capsule degradation [17]. Thus, it is important to overcome this barrier and to deliver the genes into the cytosol of cells, that they can reach the nucleus.
In order to observe the degradation of SiO2 capsules they were loaded with the commercially available protein ovalbumin, which was saturated with BODIPY dye [8]. In its non-degraded state, DQ-OVA is self-quenched and yields fluorescence signal in the red region of spectrum. This fluorescence is coming from the dye dimmers (640 nm). [7] As soon as DQ-OVA starts to degrade, the distance between the dye molecules increases and the protein thus gives bright green fluorescence (510 nm) [18].  The scale bar corresponds to 20 μm. B. Scheme for the enzymatic cleavage of DQ-OVA, as discussed previously [8]. The sketch is adopted from Rivera Gil et al [8].
In order to observe the degradation of SiO2 capsules, HeLa cells were seeded into 8-well plates (ibidi) at an amount of 25,000 cells/well, with 250 μL of medium added per well, 1 cm 2 surface area per well. The next day, SiO2 capsules as prepared with low amount of TEOS were added to the cells in each well (2 μL of the capsule stock solution) and HeLa cells were incubated with capsules for 24 hours. Afterwards, cells were observed under CLSM ( Figure 4B).

Stability of DNA plasmids
Transfection efficiency of cells can be influenced by different factors. One of them is the degradation of the genetic material, here plasmids encoding GFP (green fluorescent protein, pEGFP-N1, 4700 base pairs) in the intracellular environment. In order to mimic lysosomes, pronase at concentration of 2 mg/mL was used. To study the degradation, plasmids were transferred into (i) PBS at pH 4, (ii) pronase at pH 7, (iii) In order to reach optimal transfection conditions, high and low amounts of Lipofectamine Reagent was used. The differently treated DNA plasmids were mixed with the Lipofectamine reagent (low and high amount). The ratio Lipofectamine:DNA (CDNA =1.8 mg/mL)=2:1 μL/μL was used for the low amount, whereas the ratio Lipofectamine:DNA (CDNA=1.8 mg/mL)=4:1 μL/μL was used for the high amount.
Afterwards, DNA-Lipofectamine mixtures were added to the cells at the amount mDNA = 6.25, 12.5, 25, 50 pg/cell. Cells were placed in the incubator for at least 24 hours. After incubation, cells were prepared for flow cytometry analysis similar as in the uptake study (see §10 of this Supporting Information). Besides forward and side scattering signals, green fluorescence due to expressed GFP was recorded using an LSRFortessa flow cytometer equipped with a HeNe laser (488 nm). Gatings were used for the side and forward scattering signals ( Figure S31) in order to get rid of signal from cell debris and cell duplets.
Green fluorescence signal only from single cells was collected, and green fluorescence intensity distributions for all types of transfected cells were then plotted ( Figure S32).

A B
According to the obtained data, it can be concluded that the surrounding environment affects the ability of plasmid DNA that is used to transfect cells. In detail, untreated DNA showed the maximum of transfection efficiency and transfection was dose-dependent: more added DNA resulted in a higher rate of transfected cells. Acidic environment and the presence of pronase decrease the transfection efficiency of cells, probably due to the degradation of plasmids. Thus, since the pH value of endocytic vesicles is reported to be acidic [19,20], the possible degradation of the plasmid cargo can already start in these compartments.

In vitro transfection studies
HeLa      Figure S23) the mean blue fluorescence intensity per cell <Iblue> and the corresponding standard deviation were calculated and plotted versus the incubation conditions (Figures S38 to S41) for the different types of capsules.             1.5