Loading capacity and entrapment efficiency

The loading capacity of Dox loaded HSNs with different modification were

determined by the absorbance of doxorubicin at wavelength 485 nm by UV-Vis spectrometry. The baseline was measured in the 1 mg/mL of PEG-HSNs PBS solution, and the calibration curve was prepared by dissolving different amount of Dox in 1 mg/mL of PEG-HSNs PBS solution, including 20, 40, 60, 80, 100 Dox ug/mL. The Dox loaded particles (~3 mg) were first dispersed in 600 uL of PBS, and 200 uL of the solution (about 1 mg of particles) was diluted to 1 mL by addition of 800 uL of fresh PBS for UV-Vis measurement. After measurement, the Dox loaded particles were collected in pellet and dried in oven at 60°C for weighing. The loading capacity was calculated by the equation

(2.1), and the entrapment efficiency was calculated from the equation (2.2).

Drug loading =mass of Dox loaded in particles

mass of Dox loaded particles (2.1) Drug loading = mass of Dox loaded in particles

mass of Dox initial used for loading (2.2) 2.2.2.4 Doxorubicin release study

For drug release study, 1.5 mg of Dox loaded particles were first dispersed in 2 mL PBS (pH 5.5 or 7.4) and taken in a dialysis device (Thermal Slide-A-Lyzer^TM MINI dialysis Device, 2mL, MWCO: 10 KDa). The dialysis tube with Dox loaded particles was introduced into a release medium containing 40 mL PBS solution (pH 5.5 or 7.4). The whole device was kept and stirred at 37 °C. The amounts of drug released was estimated from the measurement of the residual drug in the particles at each sample point by UV-Vis absorbance. At each point, sink condition was maintained by replacing 2 mL of the

release PBS with fresh PBS. The percent drug release was calculated by the equation (2.3).

%Drug release = (1- Absorbance (t)

Absorbance (to) ) (2.3) 2.2.2.5 Cellular uptake

The uptake efficiency of PEG-RITC HSNs and PEG-TA-FITC HSNs by MDA-MB-231 cells was determined a FACS Calibur flow cytometer and Cell Quest Pro software (Becton Dickenson, Mississauga, CA). The green emitting fluorescein dye FITC incorporated in particles serves as a marker to quantitatively determine their cellular uptake. 2×10⁵ MDA-MB-231 cells were seeded in 6-well plates and allowed to attach overnight. Then, cells were incubated with 25, 50, 100, 250, 500, 750 μg/mL of particles in DMEM F12 medium for 4 h. Treated cells were then washed twice with PBS and then harvested by trypsinization. After centrifugation, the cells were dispersed in trypan blue solution to quench the fluorescence of particles adsorbed on the cell surface and flow cytometry analysis was carried out.

2.2.2.6 Cytotoxicity assay

The biocompatibility of particles and the cytotoxicity of free drugs and drug loaded particles were evaluated by WST-1 assay. 5×10³ MDA-MB-231 cells per well were seeded in 96-well plates for proliferation assays. The PEG-TA HSNs in medium with different particle concentration were incubated with cells for 24 hours (25, 50, 100, 250, 500, 750 μg/mL). To evaluate the anti-cancer efficacy of Dox@PEG-TA HSNs, the cells

were incubated with Dox@PEG-TA HSNs and the corresponding concentration of free Dox solutions in 2, 4, 8, 20, 40, 60 Dox ug/mL for comparison. After incubation for 24 hours, the cells were washed twice with culture medium followed by incubation with WST-1 reagent (Clontech) for 2 hours at 37 °C for proliferation assay. The formazan dye generated by the live cells was proportional to the number of live cells and the absorbance at 440 nm was measured using a microplate reader (Bio-Rad, model 680).

2.2.2.7 Confocal microscopic examination of intracellular drug release

The MDA-MB-231 cells were seeded on the glass overslipes at density of 2×10⁵ cell per well and cultured overnight. The cultured medium was replaced with Dox@PEG-TA-FITC HSNs (100 ug/mL) and incubated for 24 hours. After 24 hours, the MDA-MB-231 cells were fixed with 4% paraformaldehyde and permeabilized with 0.1% Triton X-100

for 5 minutes. After PBS washed, the cells were counterstained for nuclei with 4′,6-diamidino-2-phenylindole (DAPI, DNA marker) for 1 min and washed with PBS. The fluorescence images were obtained with a confocal laser scanning microscope (TCS SP8, Leica).

2.2.2.8 Cell cycle analysis

MDA-MB-231 cells were seeded at 2 × 10⁵cells per well in 6-well plates and allowed to attach for 24 h at 37 °C. Then, the PEG-TA HSNs (100 ug/mL), Dox@PEG-TA HSNs (8 Dox ug/mL) and fresh medium were incubated with the cells for 24 hours.

The cells were trypsinized, collected, and fixed with 70% precooled ethanol at 4 °C for 24 h. Fixed cells were washed twice with ice-cold PBS, incubated with 1 μg/mL RNase A for 20 min at 37 °C, and then stained with 10 μg/mL PI for 30 min in the dark. Stained MDA-MB-231 cells were analyzed by flow cytometry. Statistical analyses were performed using ModFit LT software (Verity Software House, Topsham, ME).

2.2.2.9 Western blotting analysis

Collected cell lysates were separated by 10 % SDS-PAGE and then transferred electrophoretically to a polyvinylidene difluoride (PVDF) membrane which was blocked in blocking buffer [1X Tris-buffered saline (TBS)- 0.1 % (v/v) Tween20, 5 % (w/v)

non-fat milk] for 1 h. The membrane was incubated with primary antibodies against p-p38 (Cell Signaling Technology; 1:500), α-tubulin (Santa Cruz; 1:10000) overnight at 4℃.

The PVDF membranes were extensively washed and incubated with secondary immunoglobulin G antibody (1:2000 dilution, Santa Cruz) for 1.5 h at room temperature.

Immunoreactive bands were visualized with the enhanced chemiluminescence substrate

kit (Amersham Pharmacia Biotech, GE Healthcare UK Ltd, Bucks, UK) according to the manufacturer’s protocol.

2.2.2.10 In vivo experiments

The Institute of Cancer Research (ICR) nude mice (8 weeks old) and tumor bearing mice (NOD.CB17-Prkdc^scid/JNarl, 8 weeks old) were intravenously injected with

PEG-HSNs and PEG-TA PEG-HSNs (conjugated with RITC dye) at 200 mg/kg dose for each experiment.

2.2.2.10a Circulation

The circulation of particles in ICR nude mice was observed under two-photon microscopy, and the live-time images was acquired at 10, 30, 60, 120 min after injection.

To detect the vessels behind the ears, the hair on the ear needed to be removed for better detection.

2.2.2.10b Bio-distribution

To study the in vivo bio-distribution, the ICR nude mice was kept for 4 weeks, and the tumor-bearing mice (NOD.CB17-Prkdc^scid/JNarl) (4 weeks old) was established by injecting with MDA-MB-231 tumor cells for 4weeks implantation. After the PBS solution of particles were intravenously injected, the major organs (heart, lung, spleen, liver, and kidney), urine and blood were carefully collected, and the fluorescence images of particles were acquired by using an IVIS Imaging System (Lumina) at indicated time (4 h and 24 h).

Chapter 3 Results and Discussion

3.1 Synthesis and engineering of hollow silica nanospheres

3.1.1 Yield improvement and size control of HSNs

As described in the introduction section, the applications of HSNs from microemulsion method were limited due to the problem of low production yield. Hence based on the growth theory of HSNs, the synthetic system with different compositions and silica source were systematically tried and studied for improvement of yield and size-uniformity.

3.1.1.1 Effects of the amount of silica source on HSNs (M1S1, M1S5)

Several microemulsion systems were studied for synthesis of hollow silica nanospheres in our previous reports ^21,22. For improvement of the yield and uniformity, we primarily focused on the decane-CO 520 systems, which were composed of decane as oil, Igepal CO 520 as surfactant, hexyl alcohol as co-surfactant, and D.I water. In a synthetic system of HSNs, two kind of silica source including TEOS and APTMS would be hydrolyzed through NH4OH catalysis and condensed/nucleated through collision of the water droplets for growing the solid silica nanospheres (SSNs), then the SSNs could be transformed into hollow structure by warm-water etching process.

In original M1S1 system, the size of HSNs was about 40nm with a uniform size distribution, as shown in Figure 1-1a based on the TEM characterization. However, the

yield was as low as 10 mg per 20 mL of decane that was not suitable for further applications. In order to elevate the yield of HSNs, it could be proposed that more silica source could be introduced into the original system for growing higher number of particles (Figure 1-1c). Hence in M1S5 system times, 5 times of silica source were used in original M1 system. Although the yield of HSNs in M1S5 system was elevated to more than five times (75 mg per 20 mL of decane), the size distribution of HSNs was broad from the TEM image shown in Figure 1-1b. The results indicated that additional silica source was not evenly distributed to every particle neither contributed to more HSNs.

That is, the elevated yield of M1S5 system was not exactly from higher number of particles but the weight of the larger particles.

Figure 3-1. TEM image and size distribution of HSNs from (a) M1S1, (b) M1S5.

(c) Schematic illustration of more silica source in the certain system.

(a) (b)

More silica source (5 times) (c)

The wide size distribution of the HSNs from M1S5 system might be related to the unstable nucleation. There were some possible factors for the unstable nucleation: (1) much silica source was simultaneously hydrolyzed and moved into water droplets, (2) the ethanol generated from the hydrolysis and condensation of silica source might disturb the interface of water droplets and change the size of droplets. (3) The collision between the unstable droplets with different sizes and silicate amount might cause uneven silicate exchanging. These factors might give rise to different-time nucleation and thus wide size distribution of the particles.

3.1.1.2 Effects of the composition of microemulsion systems on HSNs

From the hypothesis above, for a monodispersed HSNs the water droplets in the system with high silica source should undergo a thermodynamically favored inter-droplet collision and efficiently exchanging silica source during the nucleation state. The properties of a microemulsion could be tuned by adjusting the composition of system which further affected the collision kinetics and final particles²¹.

For accommodation of more silica source, the density of water droplets were increased by tuning up the volume of surfactant, co-surfactant and D.I water in constant oil (decane). The corresponding systems were assigned as M2S5 and M5S5, detailed compositions were described in the Table 2-1. TEM images and size distribution of HSNs from M1S5, M2S5 and M5S5 were shown in Figures 1-2. The HSNs from M1S5 system

showed two mainly sizes distribution at 55 nm and 110 nm; and the size of particles from M2S5 system was mainly at 50 nm and 70 nm. However, the HSNs from M5S5 system were more uniform than other two systems, and the dominate size was about 45 nm with few particles in 60 nm.

Figure 3-2. TEM image and size distribution of HSNs from (a) M1S5, (b) M2S5, and (c) M5S5.

From M1S5 to M5S5 system, the size of HSNs was getting smaller and more monodispersed. With higher density of water droplets in M5S5 system, large amounts of hydrolyzed silica source could be exchanged more evenly in more droplets so that the collision and contents exchanging between the droplets could be more effective, and it would contribute to a stable nucleation, relatively uniform particles, and smaller size.

Compared to M5S5 system, M1S5 and M2S5 systems with lower density of water

(a) (b) (c)

droplets could not effectively exchange the silicate and distribute the silicate evenly so that resulted in a different-time nucleation and particles with wide size distributions.

3.1.1.3 Effects of the time-separated addition of silica source on HSNs

Although the HSNs from M5S5 system were relatively uniform and with elevated yield, the cost of surfactant CO-520 and co-surfactant hexyl alcohol were also elevated 5 times. It was not economical actually. Therefore, a more efficient and practical method for high-yield and uniform HSNs should be developed.

In previous reports ²¹, the mechanism of HSNs from microemulsion system (M1S1) was investigated in detail. Initially, the hydrophobic silica source TEOS was majorly dissolved in the oil phase. The amphiphilic CO-520 and APTMS resided mostly at the oil/water interface of the droplet. After NH⁴OH was introduced into the system, TEOS and APTMS were catalytically hydrolyzed to form silicate and move into the water droplet, then condensed into particles, as shown in Figure 1-3.

Figure 3-3. The nucleation and growth mechanism of silica nanospheres. ²¹

There were four states of particles growth, as shown in Figure 3-4, including: (1) induction period, (2) nucleation burst, (3) growth by diffusion-coalescence, (4) slowdown of growth stage. What we were interested was the first two states: the induction period and the nucleation burst which would dominate the nucleation of the particles. Before achieving the concentration of nucleation, hydrolyzed silica source moved and accumulated into droplets without highly condensation, and the droplets would collide each other and exchange the contents for a thermodynamically favored state. As hydrolyzed silica source accumulating to the concentration of nucleation, the nucleation occurred and resulted in cores/sites for growth of particles.

Figure 3-4. Stages of the size evolution for SSNs: (1) induction period, (2) nucleation burst, (3) growth by diffusion-coalescence, (4) slowdown of growth stage. ²¹

For the system with high amounts of silica source, a thermodynamically-favored droplets collision at pre-nucleation state and a stable nucleation was important for synthesizing the uniform HSNs. Therefore, we tried to separate the high silica source into 2/5 and 3/5 part of all. Only 2/5 of silica source was added to prepare the synthesis

solution at first, and the last 3/5 of silica source was introduced after NH4OH (aq) had been added for 12 minutes. By this time-separated addition, particles from M1S1, M1S5, M2S5, and M5S5 were all with uniform size, and also with elevated yield in M1S5, M2S5 and M5S5 system compared to M1S1 system showed in Table 3-1.

The size, yield, size distribution, and TEM images of HSNs from M1S1, M1S5, M2S5, M5S5 systems by time-separated addition of 5 times of silica source.

There were two points for the time-separated addition of silica source: (1) the less amount of silica source in first addition and (2) the time for addition of the second part of silica source. In the system with less silica source in hydrolysis and condensation, the droplets would be more stable and effectively exchange their contents so that the silica source was evenly and widely distributed in every droplet. After NH4OH (aq) had been added for 12 minutes, it was almost the time for the nucleation. The second part of silica

source was added to promote the nucleation so that higher number of nucleated cores was simultaneous produced for relatviely uniform-growth of more particles compared to one-step addition method, as shown in Figure 1-5. In briefly summary, a stable nucleation would dominate the uniformity of particle size, and higher number of nucleation cores would contribute to higher number of particles.

Figure 3-5. Sschematic illustration of the different nucleation state caused by time-separated addition method and one-step addition of silica source.

3.1.1.4 Size control of HSNs by additional ethanol

Several methods to adjust the microemulsion system for controlling the size of HSNs were systematically investigated in our previous report ²¹, including oil phases of alkanes with different alkyl chains, varying oil molar volumes and co-solvent amounts or surfactant mixture ratios. Instead of changing the system composition, particle size could be controlled and adjusted from 50 nm to 115 nm by simply addition of extra ethanol in constant M2S5 system. In original M2S5 system, ethanol from APTMS ethanolic solution was about 125 uL and the particle size was about 50 nm. As the total volume of ethanol was adjusted to 250 uL, 400 uL, and 500 uL in M2S5 system, the particle size could be

One-step addition of silica source Time-separated addition of silica source

Before nucleation burst

tuned up to 71.5 nm, 90.0 nm, and 114.6 nm as showed in Figure 3-6.

Figure 3-6. The TEM images of HSNs from M2S5 system with (a) 125 uL, (b) 250 uL, (c) 400 uL, (d) 500 uL of ethanol.

Ethanol as an amphiphilic solvent would stay in oil phase (decane), aqueous droplets and interface between water and oil phase, indicating that ethanol would bring partial organic silane (TEOS and APTMS) into water droplets. As more ethanol in the synthesis system, more silica source were dissolved in droplets at first. Under catalysis by NH4OH

(aq), nucleation would occurred more early with sufficient silica source resulting in bigger cores for growth into larger particles. By simply tuning the amount of ethanol, size of HSNs could be effectively controlled with wide size range in the same system.

114.6 ± 8.56 nm 71.5 ± 6.3 nm

90.0 ± 6.8 nm 49.7 ± 2.10 nm

(a) (b)

(c) (d)

3.1.2 Investigation on the colloidal stability of HSNs

3.1.2.1 Bare HSNs

With elevated yield and well-controlled size, HSNs would be advanced in practical applications. However for biological application such as drug delivery, a good colloidal stability was a crucial property influencing the efficiency and efficacy. The aggregation of bare HSNs from the microemulsion method was also a major problem limiting their applications. During the synthesis of bare HSNs, after breaking the microemulsion by adding ethanol, it could be found that the particles would deposit in lower ethanol/water phase indicating that the slightly particle-aggregation have occurred in form of SSNs. The following processes including high speed centrifugation and water etching process could cause the irreversible binding between particles and decreasing the structural stability inducing severe particles-aggregation. The results from DLS measurement also revealed corresponding phenomenon, as showed in Figure 3-7. It implied that the dispersity improvement of HSNs should be dealt from SSNs.

Figure 3-7. The DLS measurement of bare SSNs and HSNs in D.I water.

3.1.2.2 Bare HSNs synthesized under open system

The aggregation of SSNs could be related to the anchored APTMS in the structure, which reduced the electrostatic repulsive force between negatively charged silica-based nanoparticles, and decreased the structural stability due to the low condensation level of APTMS with only three siloxane bridges. To improve the dispersity, one way is enhancing the inherent stability dependent on the condensation level of silica structure. The silica condensation was also related to the pH value and reaction time during the synthesis, as shown in Figure 3-8 ⁴¹.

Figure 3-8. Effects of pH value on the silica condensation rate, charge properties and charge density on the surface of the silica species. ⁴¹

Therefore, an open synthetic system was employed to synthesize SSNs further to HSNs. In an open synthetic system, NH4OH (aq) would gradually evaporate out of system resulting in a falling pH value. As the pH value gradually close to the neutral, it induced the increasing condensation rate of silica source, as shown in Figures 3-8, and further

enhancing the structural condensation which could improve the structural stability. In addition, Si-O^- on the particles would be transformed into Si-OH reducing the inter-particles electrostatic interaction of negative Si-O- and positive APTMS. The bare HSNs from open system were also about 50 nm and with hollow structure from TEM characterization as showed in Figure 3-9. Bare HSNs (also SSNs) with enhanced structural stability showed improved colloidal stability in EtOH, water, and DMEM medium with about 130 nm hydrodynamic diameter from DLS measurement, as showed in the Figure 3-10. However, in a solution with higher ionic strength such as PBS, aggregation of bare HSNs also occurred, indicating the ions in solution might be adsorbed on the particles surface and alter the colloidal stability of the bare HSNs. Hence, further improvement was inevitable.

Figure 3-9. The TEM images of bare HSNs from open synthetic system.

Figure 3-10. The DLS measurement of (a) SSNs and (b, c) HSNs in different solvent

from open synthetic system.

3.1.2.3 Bare HSNs with one-step surface modification

Another effective way to improve the colloidal stability of nanoparticles was surface-modifying PEG as a stealth polymer to cover the active silica shell and reduce the ionic effect in different environments ⁴². However, the post-modification of PEG-silane on pristine HSNs could not effectively improve the dispersity due to the already aggregation of the SSNs, as shown in Figure 3-11.

Figure 3-11. The DLS measurement of HSNs with post-modification of PEG.

As previous description, the dispersity improvement should be started from the SSNs.

(a) (b) (c)

Hence, a one-step surface modification method was developed. After 12 hours growth and fully formation of SSNs, PEG-silane accompanied with importantly few TEOS was introduced to modify the SSNs to form PEG-SSNs in M2S5 microemulsion system. After warm water etching, the SSNs could also be successfully transformed into

Hence, a one-step surface modification method was developed. After 12 hours growth and fully formation of SSNs, PEG-silane accompanied with importantly few TEOS was introduced to modify the SSNs to form PEG-SSNs in M2S5 microemulsion system. After warm water etching, the SSNs could also be successfully transformed into