WO2012051341A1 - Hydrothermal process for enhanced stability of mesoporous nanoparticles - Google Patents

Abstract

The invention provides biocompatible mesoporous silica nanoparticles having enhanced stability in biological media, and methods of preparing and using those particles.

Description

HYDROTHERMAL PROCESS FOR ENHANCED

STABILITY OF MESOPOROUS NANOPARTICLES

Cross-Reference to Related Applications

This application claims the benefit of the filing date of U.S. application Serial No. 60/392739, filed on October 13, 2010, the disclosure of which is incorporated by reference herein.

Background

Since the first report of ordered mesoporous silica in 1992 (Kresge et al., 1992), extensive work has been done exploring the syntheses, characterization, and application of these materials. Biomedical applications such as delivery of anticancer drugs (Lu et al., 2007), enzymes (Slowing et al., 2007), and DNA (Torney et al., 2007) are among the most promising, exploiting the nanoparticles ordered pore structure, large surface area, and considerable pore volume. The ideal biomedical nanoparticles (NPs) could be used simultaneously in diagnosis, imaging, and therapy. Toward this goal, several types of multifunctional nanomaterials have recently been reported in the literature, including plasmonic magnetic nanocrystals (Lim et al., 2007; Wang et al., 2008; Lim et al., 2008), core-satellite nanocomposites (Lee et al., 2006), and polymeric nanohybrids (Yang et al., 2007; Kim et al., 2008).

Mesoporous silica nanoparticles (MSNs) can also be used as a multifunctional platform, as has been demonstrated in several recent papers focusing on the synthesis of porous silica particles having fluorescent, magnetic, cellular labeling, and/or therapeutic functions (Zhao et al., 2005; Kim et al., 2006; Giri et al., 2005; Lin et al., 2006). However, to date, the biomedical use of the MSN materials reported is hindered by either their tendency to aggregate upon exposure to physiological conditions (Zhao et al., 2005; Kim et al., 2006) and/or overall large nanoparticle size, with particle diameters exceeding 150 nm (Giri et al., 2005; Lin et al., 2006). Studies have shown that silica particles with diameters greater than 100 nm are rapidly taken up by the reticuloendothelial system (RES), accumulating in the liver and spleen before a loaded drug can be delivered to the target cells/tissue (Wu et al., 2008; Taylor et al., 2008), but that smaller solid silica NPs (< 50 nm diameter) with poly(ethylene glycol) (PEG) surface modification have significantly decreased uptake by RES organs and exhibit a longer blood circulation time (He et al., 2008). Although well-ordered MSNs with diameters as low as 20 nm have been synthesized using a double surfactant system (Suzuki et al., 2004), the resulting nanoparticles aggregate easily and, thus, discourage effort to incorporate other functionality such as magnetic contrast, because the base nanoparticle is not appropriate for biomedical use. In cases where attempts have been made to embed magnetic centers into MSNs, the synthesized materials suffer from insufficient magnetic contrast achieved due to the low mass percentage (< 1 %) of incorporated magnetic nanoparticles or destruction of the well- ordered, high surface area character of the silica structure. The resulting low magnetic response is due to the thick mesoporous silica shell (Kim et al., 2006) or the large size of the mesoporous silica (Lin et al., 2006) and suggests that a large dose (e.g., 175 mg or 500 mg NPs/kg) would have to be injected to perform in vivo magnetic resonance imaging (MRI) (Wu et al., 2008; Kim et al., 2008).

To date, these shortcomings largely limit the practical use for in vivo biomedical applications; to realize the full potential of multifunctional MSNs, a synthesis must be designed that produces small, easily dispersed, and well-ordered silica particles with higher mass percentage of incorporated magnetic centers. The great potential of these materials has generated significant efforts to achieve these goals by multiple research groups (Kim et al., 2008; Deng et al., 2008). Recent efforts to achieve this goal include the work of Zhao and co-workers who reported a novel synthesis of a high-magnetization Fe₃0₄@Si0₂ particle with an ordered porous structure (Deng et al., 2008); however, the relatively large size of the particle (> 500 nm diameter) limits its use for intravenous (iv) injected biological applications. Also, Hyeon and co-workers reported a novel synthesis of uniform multifunctional NPs with small size (about 45 nm diameter) and high aqueous stability (Kim et al., 2008). Unfortunately, these nanoparticles still have low magnetic response and disordered mesostructure, limiting their use for bioimaging, efficient drug loading, or bioseparations. None of the reported materials possess: (1 ) small size (< 50 nm); (2) excellent aqueous dispersity; (3) ordered porous structure with high surface are (> 600 m²/g); and (4) high magnetic response (> 10 emu/g) to facilitate drug delivery, bioseparation, and bioimaging applications.

In addition to design of ideal nanomaterials for nanotherapeutics, the unintentional toxicity of these NPs should be monitored as part of the design process. The 3-(4,5-dimethylthiazol-2-yl)-2,5- diphenyl-tetrazolium bromide (MTT) assay is widely used to study the cytotoxicity of various

nanomaterials, and while it permits an NP to NP comparison, it gives insight only into changes in cell mitochondrial function. While there are four major routes of NP entry into the human body (inhalation, ingestion, transdermal penetration, and injection), biomedical NPs are most likely to be delivered via injection, and thus, blood compatibility, e.g., no or minimal hemolytic activity, is quite important.

Practical biomedical application of mesoporous silica nanoparticles is limited by poor particle dispersity and stability due to serious irreversible aggregation in biological media. Hydrothermally treated mesoporous silica nanoparticles of small size with dual-organosilane (hydrophilic and hydrophobic silane) surface modification were synthesized. These highly organo-modified mesoporous silica nanoparticles were characterized by transmission electron microscopy, X-ray diffraction, N₂ adsorption-desorption, dynamic light scattering, zeta potential, and solid-state ²⁹Si NMR and they prove to be very stable in simulated body fluid at physiological temperature. Additionally, they can be dried to a powdered solid and easily redispersed in biological media, maintaining their small size for a period of at least fifteen days. Furthermore, this preparation method can be expanded to synthesize redispersible fluorescent and magnetic mesoporous silica nanoparticles. The highly stable and redispersible mesoporous silica NPs show minimal toxicity during in vitro cellular assays Most importantly, two types of doxorubicin, water- soluble doxorubicin and poorly water-soluble doxorubicin can be loaded into these highly stable mesoporous silica nanoparticles, and these drugloaded nanoparticles can also be well-redispersed in aqueous solution. Enhanced cytotoxicity to cervical cancer (HeLa) cells was found upon treatment with water-soluble doxorubicin-loaded nanoparticles compared to free water-soluble doxorubicin. These results suggest that highly stable, redispersible, and small mesoporous silica nanoparticles are promising agents for in vivo biomedical applications.

Summary of the Invention

The invention provides a method to prepare biocompatible mesoporous silica nanoparticles with enhanced stability in biological media. The method includes heating a surfactant containing solution having mesoporous silica nanoparticles with a diameter from about 20 to about 250 nm, to about 70°C to 150°C for about 12 hours to about 48 hours. In one embodiment, the particles have a diameter from about 20 to about 150 nm, about 20 to about 100 nm, about 20 to about 50 nm or about 25 to about 35 nm. In one embodiment, the particles are heated to about 85 to about 95°C for about 18 hours to about 30 hours. In one embodiment, the silica nanoparticles are prepared using hydrophilic and hydrophobic silanes. The surfactant is then extracted from the heat-treated mesoporous silica nanoparticles, and the extracted, heat-treated mesoporous silica nanoparticles are washed. The washed, heat-treated mesoporous silica nanoparticles are subsequently filtered, yielding mesoporous silica nanoparticles that have enhanced stability in aqueous biological media relative to mesoporous silica nanoparticles that are not heat treated. Biological media includes media suitable for maintaining (culturing) or growing cells in vitro as well physiologically compatible media. In one embodiment, the mesoporous silica nanoparticles that have enhanced stability are stable for one or more weeks, e.g., the nanoparticles exhibit limited if any aggregration or decomposition, such that the dynamic light scattering profile of the nanoparticles is not substantially altered over time. In one embodiment, the biological media comprises phosphate buffered saline. In one embodiment, the biological media comprises serum, for instance, from about 0.1 % to about 20% serum, such as fetal bovine serum (FBS).

To produce size-tunable, sub-100-nm diameter, multifunctional mesoporous silica nanoparticles (MSNs) with sufficient magnetic loading and visible to near-infrared fluorescence, that are stable and well- dispersed in aqueous solutions, ammonium hydroxide was used as a basic catalyst. Size-controlled synthesis of multifunctional MSNs may be combined with incorporation of superparamagnetic

nanoparticles (e.g., Fe₃0₄) and/or fluorescent dye molecules, such as visible fluorophores or near infrared fluorophores, which may be useful in diagnostic or imaging applications or to facilitate multimodal imaging. The dispersed, stable and well-ordered multifunctional nanoparticles with controllable size may also find use in drug delivery. For example, porous multifunctional nanoparticles (Fe₃0₄@Dye-MSNs@PEG), ranging from 33 to 67 nm in diameter, are not only smaller than previously reported MSNs but also possess a well-ordered mesostructure, excellent aqueous dispersity, and a higher volume fraction of magnetic nanoparticles than previously achieved. The silica core with incorporated Fe₃0₄ and fluorophores can be synthesized without disrupting the porous structure, and the nanoparticle exterior may be functionalized to promote stable aqueous dispersity. Targeting ligands may be introduced onto the nanoparticles so that they deliver drug to specific cells.

In one embodiment, sub-50 nm pegylated MSNs were prepared using a hydrothermal treatment and were found to have long-term stability, e.g., for 10 days at 37°C in various biological media at both room and physiological temperatures. Compared to bare MSNs, the highly pegylated MSNs showed significantly improved biocompatibility and decreased macrophage uptake, making these nanoparticles particularly suited for in vivo stealth drug delivery applications.

In one embodiment, to prepare hydrothermal treated pegylated MSNs, a suitable amount of a surfactant, e.g., a cationic quaternary ammonium salt, such as 0.29 g of n-cetyltrimethylammonium bromide (CTAB), is added to an alkaline catalyst, such as 150 ml_ of 0.256 M NH₄OH solution, at 50°C. However, any strong base may be employed so long as an appropriate pH is achieved. In one embodiment, the CTAB employed has from C₁₀ to C₁₈. Then, 2.5 ml_ of 0.88 M ethanolic TEOS was added to solution under continuously stirring. After one hour, 450 μΙ_ of PEG-silane was added to the as- synthesized colloidal solution and the mixture solution was stirred for 30 minutes and then aged at 50°C for 20 hours. The as-synthesized pegylated mesoporous colloidal solution was filtered with a 0.45 μιη GHP filter and diluted to 50 ml_ with D.I. water. Then, the filtered colloidal solution was heat at 90°C for 24 hours. The as-synthesized pegylated mesoporous silica colloids with hydrothermal treatment were transferred to 50 ml_ of 6 g/L ethanolic ammonium nitrate by centrifugation (30,000 rpm = 66226g, 30 minutes) and heated to 60°C for one hour under stirring. The nanoparticles were then transferred to 50 ml_ of acidic ethanol solution (1 ml_ of HCI/1 L of ethanol) via centrifugation and heated to 60°C for two hours under stirring. The extracted pegylated MSNs with hydrothermal treatment were further washed with 95% ethanol and then 99.5% ethanol once. Finally, the surfactant-free pegylated MSNs were suspended in 99.5% ethanol and filtered using a 0.2 μιη PTFE filter. The products were stored at 4°C before use.

The resulting particles include a mesoporous silica core that can have a drug, e.g., hydrophobic drugs could be incorporated therein such as doxorubicin, irinotecan, or oxaliplatin, or that can include imaging moieties, with a surface that is easily modified. If present, superparamagnetic iron oxide nanoparticles provide a MR contrast agent, and/or a fluorophore provides for a fluorescent imaging agent. A coating, such as a PEG coating, can promote aqueous dispersability and minimize fouling. The resulting particles are small enough to avoid uptake by the reticuloendothelial system (so that they can reach the target cell/tissue/organ that needs treatment) but yet are capable of carrying enough drug in the pores to be effective.

In one embodiment, the MSNs of the invention include a coating. In one embodiment, the pores of the MSNs may be coated. In one embodiment, the exterior surface of the MSNs may be coated. The MSNs of the invention may be modified to include a coating in pores and on the exterior surface. In one embodiment, the pores of the MSNs may be coated with a different material than the exterior surface of the MSNs. In one embodiment, the surface coating may include targeting molecules, such as antibodies or other ligands for cell surface receptors, e.g., integrin-specific targeting peptides.

The use of these nanoparticles may enable large loading capacity and targeted delivery of drugs, which in turn may facilitate the use of drug compounds that cannot be used on their own because they are not stable during delivery or are toxic during delivery. The added imaging capability allows for the facile determination whether or not the drug is being delivered to the correct physiological region and if physiological changes take place after drug delivery (e.g., tumor shrinkage). Moreover, the nanoparticles can deliver drugs or other molecules or compounds, e.g., in medical and nonmedical applications, and/or facilitate imaging such as concurrent imaging. To address practical biomedical limitations of MSNs, e.g., poor particle dispersity and stability due to irreversible aggregation in biological media, hydrothermally treated mesoporous silica nanoparticles of small size with dual-organosilane (hydrophilic and hydrophobic silanes) surface modification were synthesized. Exemplary hydrophobic silanes include but are not limited to methyl silanes, e.g., trimethylchlorosilane or dimethydichlorosilane; linear alkyl-silanes, for instance, octyldimethylchlorosilane or octadecyldimethylchlorosilane; branched alkyl-silanes, e.g., t- butyldimethylchlorosilane; cyclic alkyl-silanes, e.g., cyclohexyldimethylchloro silane; phenyl silanes, for instance, phenyldimethylchlorosilane or phenethyldimethylchlorosilane; or fluorinated alkyl-silanes, such as 3,3,3-trifluoropropyl)dimethylchlorosilane. These highly organo-modified mesoporous silica nanoparticles were thoroughly characterized by transmission electron microscopy, X-ray diffraction, N₂ adsorption-desorption, dynamic light scattering, zeta potential, and solid-state ²⁹Si NMR, and their long- term particle stability in biological media, e.g., highly salted solutions or serum-containing media, as well as their redispersity, examined. The highly organo-modified mesoporous silica nanoparticles were very stable in simulated body fluid at physiological temperature. Additionally, they were dried to a powdered solid, which was readily redispersed in biological media. The redispersed mesoporous silica

nanoparticles maintained their small size in biological media for a period of at least fifteen days. The highly stable and redispersible mesoporous silica nanoparticles showed minimal toxicity during in vitro cellular assays, e.g., in human endothelial cells, skin fibroblasts, red blood cells, and platelets.

Furthermore, this preparation method can be expanded to synthesize redispersible fluorescent and magnetic mesoporous silica nanoparticles. Moreover, the versatility of this synthetic method is demonstrated in preparation of redispersible fluorescent and magnetic MSNs, and drug loaded MSNs.

Two types of doxorubicin, water-soluble doxorubicin and poorly water-soluble doxorubicin, were loaded into these highly stable mesoporous silica nanoparticles, and the drug loaded nanoparticles were well- redispersed in aqueous solution. Enhanced cytotoxicity to cervical cancer (HeLa) cells was found upon treatment with water-soluble doxorubicin-loaded nanoparticles compared to free water-soluble doxorubicin, thus demonstrating the synthesis of a small, ultrastable, and redispersible MS

nanotherapeutic.

Thus, the invention further provides a preparation of stable MSNs prepared by the methods described herein. In one embodiment, the stable MSNs have reduced degradation, e.g., as shown by less free silicic acid or reduced pore collapse, reduced aggregation, for instance, as little or no change in hydrodynamic size, reduced cytotoxicity, and/or reduced hemolysis relative to mesoporous silica nanoparticles not subjected to heat treatment. For example, the stable MSNs of the invention in PBS (or serum) have less than about 40 ppm, 30 ppm or 20 ppm free silicic acid, less than about 20% or 10% change in hydrodynamic size, at least 10%, 30% or 50% lower cytotoxicity, and/or at least 10%, 30% or 50% lower hemolytic activity relative to MSNs that are not heat treated. Also provided is a method of using the particles or aqueous preparation thereof to deliver an effective amount of an imaging agent or drug, or a combination thereof, to a subject.

The highly organo-modified MSNs may be employed in catalytic operations, e.g., high throughput catalytic operations, separation, e.g., to remove pollutants, or to purify waste water. In one embodiment, water having hydrophobic organic pollutants may be purified using hydrophobic, coated MSNs of the invention. The use of those MSNs may be better than other water purification methods due to the higher surface area and higher stability of the MSNs, and because the MSNs may be employed in suspension. For example, hydrothermally treated MSNs with trimethylsilane may be introduced into a water supply contaminated with polycyclic aromatic hydrocarbons and polychlorinated biphenyls. Based on hydrophobic interactions, the molecules would be captured by the MSNs. If the MSNs have magnetic cores, the particles and pollutants may be readily collected and removed from the water system. For water purification of heavy metals, the MSNs are modified with a chelating ligand, and so may be employed to the capture heavy metal ions from polluted water. For example, hydrothermally treated MSNs modified with ethylenediaminesilane are introduced into a water supply contaminated with Pb²⁺, Cu²⁺, or Co²⁺ ions. Based on metal chelation, the free metal ions are captured by the MSNs. If the MSNs have magnetic cores, the particles and pollutants could be easily collected and removed from the water system. In another embodiment, organometallic catalysts may be used to modify MSNs which promote the catalytic activity of specific reactions. For instance, hydrothermally treated MSNs modified with palladium bipyridyl catalysts are introduced into a Sonogashira catalytic reaction. Based on the high stability of MSNs, the palladium bipyridyl catalysts would not be deactivated. If the MSNs have magnetic cores, the particles could be easily collected and removed from the reaction.

Accordingly, the invention provides a biocompatible composition comprising stable and redispersible MSNs comprising a hydrophobic organosilane, e.g., a dry powder composition comprising the MSNs. The particles are readily redispersed in biological media, have a reduced number of silanol groups, and/or decreased degradation (increased stability), e.g., after redispersion in media such as physiologically compatible media, relative to particles without the hydrophobic organosilane, and/or which are not hydrothermally treated during particle preparation. In one embodiment, the MSNs have a diameter of about 30 nm to about 60 nm. In one embodiment, the MSNs comprise a drug, a chelating agent, an optically detectable dye, or magnetic particles, or a combination thereof. Further provided is a pharmaceutical composition comprising the particles or the redispersed particles in a pharmaceutically acceptable liquid carrier, e.g., one suitable for injection, for instance, via a needle or catheter. In one embodiment, the particles comprise one or more drugs, e.g., anti-tumor drugs such as nucleotide and nucleoside analogs, alkylating agents, nitrogen mustards, nitrosoureas, antibiotics, or antimetabolites; hormonal agonists/antagonists, androgens, antiandrogens, antiestrogens, gonadotropin releasing hormone analogues, progestrins, or other antineoplastics. See, Physician's Desk Reference (2001 ). In one embodiment, the particles comprise one or more imaging agents, e.g., Gd³⁺-based agents. Brief Description of the Figures

Figure 1 . Schematic diagram of the one-pot synthetic procedure to produce PEG-modified fluorescent mesoporous silica NPs with incorporated Fe304 NPs (Fe304@Dye-MSNs@PEG).

Figure 2. TEM images of SS NPs with varied diameters: (A) SS-24, (B) SS-37, (C) SS-142, and (C) SS-263.

Figure 3. Particle size distributions of SS NPs with four different sizes: SS-24 (black), SS-37

(red), SS-142 (green), and SS-263 (blue). Data are from TEM micrographs.

Figure 4. N₂ adsorption-desorption isotherms of MS-25 (blue line) and SS-24 (black line) NPs. Figure 5. A photograph of as-synthesized colloidal solutions of MSNs with varied sizes. The MSNs were well-dispersed in aqueous solutions.

Figure 6. TEM images of surfactant-free MSNs with varied sizes: (A) MS-25, (B) MS-42, (C) MS-

93, (D) MS-155, and (E) MS-225. (F) Particle size distributions of five sizes of surfactant-free MS NPs. The data were from TEM micrographs.

Figure 7. Characterization of surfactant-free MSNs. (A) Low-angle (1 .5-8°) XRD patterns of MS NPs with varied sizes. (B) N₂ adsorption-desorption isotherms of MS NPs with varied sizes.

Figure 8. (A) Percentage of hemolysis of RBCs incubated with four sizes of SS NPs at different concentrations ranging from 3.125 to 1600 μg/mL for 3 hours. Data represent the mean + SD from at least three independent experiments. (B) Photographs of hemolysis of RBCs in the presence of four sizes of SS NPs. The presence of red hemoglobin in the supernatant indicates damaged RBCs. D.I. water (+) and PBS (-) are used as positive and negative control respectively.

Figure 9. Comparison of the TC₅₀ of SS (square) and MS (circle) NPs with varied sizes.

Figure 10. Concentration-dependent hemolytic activity of MSNs with different sizes: MS-42 (circle), MS-93 (upward pointing triangle), MS-155 (downward pointing triangle), and MS-225 (diamond). Data represent the mean ± SD from at least three independent experiments.

Figure 1 1 . (A) Percentage of hemolysis of RBCs in the presence of five sizes of MSNs at different concentrations ranging from 3.125 to 1600 μg/mL for 3 hours. Data represent the mean + SD from at least three independent experiments. (B) Photographs of hemolysis of RBCs incubated with four sizes of MS NPs. The presence of red hemoglobin in the supernatant indicates damaged RBCs. D.I. water (+) and PBS (-) are used as positive and negative control respectively.

Figure 12. Percentage of hemolysis of RBCs incubated with (A) MS-25 (square) and MS-25 after

6-day PBS aging (empty square); (B) MS-42 (circle) and MS-42 after 6-day PBS aging (empty circle). Low-angle XRD patterns of (C) MS-25 and MS-25 after 6-day PBS aging; (D) MS-42 and MS-42@PEG after 6-day PBS aging.

Figure 13. Photographs of RBCs following interaction with (A) MS-25 and MS-25 after 6-day PBS aging; (B) MS-42 and MS-42 after 6-day PBS aging for 3 hours. D.I. water (+) and PBS (-) are used as positive and negative control, respectively.

Figure 14. (A) Percent hemolysis and (B) a photograph of RBCs in the presence of CTAB at different concentrations ranging from 3.125 to 100 μg/mL for 3 hours. D.I. water (+) and PBS (-) are used as positive and negative control, respectively. Data represent the mean + SD from three independent experiments. Photographs of RBCs incubated with (C) supernatant of MS-25 after 6-day PBS aging at 2000 μg/mL concentration and (D) supernatant of MS-42 after 6-day PBS aging at 2000 μg/mL concentration for 3 hours.

Figure 15. TEM images of surfactant-free (A) MS-25 without PBS aging; (B) MS-25 after 6-day PBS aging; (C) MS-42 without PBS aging; (D) MS-42 after 6-day PBS aging.

Figure 16. N₂ adsorption-desorption isotherms of (A) MS-25 and MS-25 after 6-day PBS aging; (b) MS-42 and MS-42 after 6-day PBS aging.

Figure 17. Degraded silicon concentration from 2000 μg/mL of MS-25 and MS-42 after different

PBS aging times. Data represent the mean ± SD from three replicates.

Figure 18. TEM images of surfactant-free (A) MS-25@PEG and (B) MS-42@PEG. The inset is a high-magnification TEM showing 2D hexagonal mesopores inside MS-25@PEG and MS42-PEG.

Figure 19. Low-angle XRD patterns from (A) MS-25@PEG and MS-25@PEG after 6-day PBS aging; (B) MS-42@PEG and MS-42@PEG after 6-day PBS aging.

Figure 20. (A) Percent hemolysis and (B, C) photographs of RBCs interacted with MS-25@PEG, MS-25@PEG after 6-day PBS aging, MS-42@PEG, and MS-42@PEG after 6-day PBS aging. Data represent the mean + SD from at least three independent experiments.

Figure 21 . Percent hemolysis of RBCs in the presence of MS-25@PEG after 6-day PBS aging (black square) and MS-42@PEG after 6-day PBS aging (red triangle) monitored at different incubation times (3, 6, 12, and 24 hours). The concentration of MS-25@PEG after 6-day PBS aging and MS- 42@PEG after 6-day PBS aging is 1600 μg/mL. Data represent to mean + SD.

Figure 22. Preparation flowchart of bare and pegylated MSNs: MS42-C, MS42-d, MS42@PEG, and MS42@PEG-hy-c.

Figure 23. TEM images of surfactant-free (A) MS42-d and (B) MS42@PEG-hy-c NPs. (C) XRD patterns of surfactant-free MS42-d and MS42@PEG-hy-c NPs. (D) N₂ adsorption-desorption isotherms of surfactant-free MS42-d and MS42@PEG-hy-c NPs.

Figure 24. TEM images of (A) surfactant-free MS42-C and (B) MS42@PEG-c NPs.

Figure 25. Hydrodynamic size distribution of 1 mg/mL MS-42-c, MS42-d, and MS42@PEG- hy-c NPs measured by DLS at R.T. in various media: (A) D.I . H₂0, (B) PBS, and (C) DMEM + 10% FBS. Long-term colloidal stability of (D) MS42-d and (E) MS42@PEG-hy-c NPs in various media at R.T. and 37°C. Data represent mean ± SD from three independent experiments. Inset: a photograph of MS42@PEG-hy-c colloidal solutions after 10 days of aging in D.I . H₂0, PBS, and DMEM + 10%

FBS at 37 °C.

Figure 26. A photograph of MS42-C, MS42-d, and MS@PEG-hy-c colloidal solutions after 30 minute aging in PBS at room temperature.

Figure 27. Long-term particle stability of MS42@PEG-c and MS42@PEG-hy-c NPs in D.I. water and PBS at room temperature.

Figure 28. (A) Degraded free silicon concentration from 1 mg/mL MS42-d and MS42@PEG- hy-c suspsension after 10 day D. I. H₂0 and PBS aging at R.T. and 37°C. Data represent mean + SD from three independent experiments. XRD patterns of (B) MS42-d and (C) MS42@PEG-hy-c NPs after 10 day PBS aging at R.T. and 37°C.

Figure 29. XRD patterns from (A) MS42-d and (B) MS42-hy-c NPs after 10-day aging in D.I. H₂0 at R.T. and 37°C; and (C) MS42-hy-c NPs after different aging times in PBS at 37°C.

Figure 30. (A) Cell viability of human endothelial cells after exposure to different

concentrations of MS42@PEG-hy-c NPs. Data represent mean ± SD from four replicates. (B) Percent hemolysis of RBCs incubated with different concentrations of MS42@PEG-hy-c NPs for 3 hours at 37°C; inset is the photograph of RBCs after NP exposure. (C) Uptake amounts of silicon by Raw 264.7 macrophages after 24-hour exposure of 200 μg/mL of MS42-d and MS42@PEG-hy-c NPs. Data represent mean ± SD from three independent experiments with four replicates.

Figure 31 . Highly PEGylated MSN with long-term colloidal stability synthesized with a hydrothermal treatment.

Figure 32. Effect of MS42@PEG/TMS-hy-c on ROS level generation in human endothelial cells.

Figure 33. TEM images of extracted (A) MS42@PEG/TMS-hy-c, (B) MS42@PEG/TFS-hy-c, and (C) MS25@PEG/TMS-hy-c. (D) powder XRD patterns and (F) hydrodynamic size distributions of extracted MS42@PEG/TMS-hy-c, MS42@PEG/TFS-hy-c, and MS25@PEG/TMS-hy-c. (E) N₂ adsorptiondesorption isotherms of MS42@PEG/TMS-hy-c and MS25@PEG/TMS-hy-c.

Figure 34. (A) Long-term colloidal stability of various MSNs in SBF at 37°C. (B) Long-term colloidal stability of MS42@PEG/TMS-hy-c NPs in various media: PBS, DMEM+10%FBS, and SBF at 37°C.

Figure 35. Digital pictures of (A) dry powder and (B) redispersed colloidal solution (30 mg/mL) of

MS42@PEG/TMS-hy-c and MS25@PEG/TMS-hy-c. (C) Hydrodynamic diameter distributions of colloidal solutions prepared from MS42@PEG/TMS-hy-c, MS42@PEG/TFS-hy-c and MS25@PEG/TMS-hy-c powder measured by DLS at R.T. in SBF. (D) Long-term stability of redispersed MS42@PEG/TMS-hy-c, MS42@PEG/TFS-hy-c and MS25@PEG/TMS-hy-c NPs in SBF at 37°C. Data represent mean ± SD from three independent experiments. Inset: a digital picture of MS42@PEG/TMS-hy-c colloidal solution prepared from MS42@PEG/TMS-hy-c powder after 15-day aging in SBF at 37°C.

Figure 36. TEM images of extracted MS42-d and MS42@PEG/TMS-hy-c before (A,C) and after (B,D) 10-day SBF aging at 37°C. (E) XRD patterns of extracted MS42-d and MS42@PEG/TMS-hy-c after 10-day SBF aging at 37°C. (F) Degraded free silicon amount from 1 mg/mL of MS42-d and

MS42@PEG/TMS-hy-c colloidal solutions in SBF at 37°C. Figure 37. ²⁹Si solid-state MAS NMR spectra of (A) MS42-d, (B) MS42@PEG-c, (C)

MS42@PEG-hy-c, and (D) MS42@PEG/TMS-hy-c.

Figure 38. Viability of (A) human endothelial cells and (B) human skin fibroblasts after 24 hours exposure at different concentrations of MS42-d and MS42@PEG/TMS-hy-c. (C) Percentage of hemolysis of RBCs and (D) percent LDH leakage from human platelet after exposure to 200 μg/mL of MS42-d and MS42 MS42@PEG/TMS-hy-c for 0.5, 1 .5, or 3.0 hours at 37°C.

Figure 39. (A,B) TEM images, (C,D) hydrodynamic diameter distribution, (E,G) powder and (G,H) colloidal solutions (1 mg/mL) of surfactant-free FITC-MS42@PEG/TMS-hy-c and

Fe304@MS@PEG/TMS-hy-c. (I) A photograph of 1 mg/mL of redispersed FITC-MS42@PEG/TMS-hy-c in SBF under UV illumination. (J) A photograph of 30 mg/mL of redispersed Fe304@MS@PEG/TMS-hy-c in D.I. water during exposure to a magnet.

Figure 40. Photographs of (A) 1 mM Dox and DoxHCI solutions, (B) powder and (C) colloidal solutions of DoxHCI-MS42@PEG/TMS-hy-c and Dox-MS42@PEG/TMS-hy-c. (D) Hydrodynamic size distributions of redispersed DoxHCI-MS42-d, DoxHCI-MS42@PEG/TMS-hy-c, and Dox- MS42@PEG/TMS-hy-c. (E) Drug release profile of DoxHCI-MS42@PEG/TMS-hy-c, and Dox- MS42@PEG/TMS-hy-c at pH 7.4 and pH 5.0 (n > 3).

Figure 41 . Photographs of (A) 1 mM Dox and DoxHCI solutions, (B) powder and (C) colloidal solutions of DoxHCI-MS42@PEG/TMS-hy-c and Dox-MS42@PEG/TMS-hy-c. (D) Hydrodynamic size distributions of redispersed DoxHCI-MS42-d, DoxHCI-MS42@PEG/TMS-hy-c, and Dox- MS42@PEG/TMS-hy-c. (E) Drug release profile of DoxHCI-MS42@PEG/TMS-hy-c, and Dox- MS42@PEG/TMS-hy-c at pH 7.4 and pH 5.0 (n > 3).

Figure 42. Cell viability of HeLa cells after different incubation times (24, 48, or 72 hours) with different concentrations of (A) DoxHCI-MS42@PEG/TMS-hy-c and (B) Dox-MS42@PEG/TMS-hy-c. Data represent mean ± SD from at least three independent experiments done in triplicate.

Detailed Description of the Invention

Providing an early and accurate diagnosis of disease and subsequent efficient treatment is a major aim in medicine, and nanotechnology has the potential to advance these goals. In the past decade, great strides have been made in nanomedicine because of the rapid development of nanomaterials. One specific area where nanoscale materials are poised to have an impact is in the targeted use of nanoparticles in tumors, to help physicians perform diagnostics, imaging, and drug delivery (Farokhzad et al., 2009; Xie et al., 201 1 ). Chemotherapeutic drugs are well known for their adverse side effects such as fatigue, nausea and hair loss, which result from the way they work to kill quickly dividing cells in the body rather than cancerous cells specifically. Therefore, targeted drug delivery with controlled drug release is a more desirable route in cancer therapy to reduce these negative effects to healthy cells. Mesoporous silica nanoparticles (MSNs) are small, discrete particles with regularly spaced pores. They are particularly promising for medicine because of their large surface area and pore volume, making them an excellent candidate for drug storage and delivery (Rosenholm et al., 2010; Vivero-Escoto et al., 2010; He et al., 201 1 ).

Although there is significant literature precedent for synthesis and biomedical applications of

MSNs, there are several critical issues that result in limited use of these MSNs in vitro or in vivo. For example, the large particle size (> 100-nm-diameter) and poor particle stability (aggregation) result in rapid uptake by the reticuloendothelial system (RES) (Wu et al. , 2008; He et al. , 201 1 ; Huang et al.,

201 1 ). Another major hurdle is the possible unintentional toxicity of such MSNs (Lin et al. , 2010; Yu et al., 201 1 ). If the designed nanoparticles cause unintentional damage to benign cells or healthy tissues and organs, their use will be greatly limited in therapeutic applications. Prior to in vivo animal experiments, unintentional cytotoxicity must be minimized. Based on the critical considerations described above, a MSN should possess the following characteristics: (1 ) small hydrodynamic size

(<100 nm) ; (2) high drug loading capacity; (3) high stability; with (4) minimal unintentional cytotoxicity.

Recently, most MS NP work has focused on the development of controlled release using novel methods such as chemical-(Kim et al. , 2010; Luo et al. , 201 1 ) or enzyme-cleavable bonding

(Schlossbauer et al., 2009; Bernardos et al., 2010) , pH-responsive linkers, 15 or aciddissolvable ZnO NP (Muhammed et al., 201 1 ) and CaP coating (Rim et al., 201 1 ) as capping agents. However, dispersity of the developed nanoparticles was not carefully examined for any of these controlled drug release MSNs. Although these precedents successfully demonstrated that a drug can be released in a controlled manner, the most important requirement of using nanotherapeutics for in vivo cancer therapy is having a small hydrodynamic size (<100 nm) in aqueous conditions. Even particles that are small initially will be taken up by the RES if they aggregate upon exposure to biological media, limiting their therapeutic efficacy. Additionally, to facilitate clinic use and storage of MS

nanotherapetics, it is essential to develop a synthetic method to prepare redispersible MS

nanotherapetics, those that may be dried and resuspended later with no change in hydrodynamic size or therapeutic efficacy.

To date, only a few reports (Liong et al. , 2008; Wang et al., 2010; Meng et al., 201 1 ; Urata et al., 201 1 ; Canda et al., 201 1 ; and Lin et al. , 201 1 ) have demonstrated that the dispersity of MSNs can be improved by coating the MSN surface with a phosphonate (Liong et al. , 2008), phospholipid (Wang et al. , 2010) or polyethyleneimine-polyethylene glycol copolymer (Meng et al., 201 1 ).

However, the hydrodynamic sizes of these phosphate-, phospholipid- or copolymer-functionalized MSNs were not carefully examined or are known to be larger than 100 nm in buffered saline solutions. In very recent work, Urata et al. (201 1 ) synthesized small MSNs (about 20 nm) with an ethenylene-bridged silsesquioxane framework using bis(triethoxysily)ethylene precursors. Although this ethenylene-bridged MSN showed high resistance to biodegradation, no hydrodynamic size data in highly salted solution were demonstrated to prove the colloidal dispersity and long-term stability.

Another recent work reported by Cauda et al. (201 1 ) developed a "liquid-phase calcination" method to remove surfactant in high-boiling solvents at high temperatures, preventing formation of irreversible NP aggregations. The high temperature calcination in liquid-phase further increased the

condensation within the silica network and, accordingly, the stability of the MSNs. However, the hydrodynamic size of these liquid-phase calcined MSNs was only measured in ethanol; again, no particle stability data were demonstrated in biological media at physiological temperature. Lin et al. (201 1 ) showed that sub-50 nm MSNs (hydrodynamic size <100 nm) with a short-chain polyethylene glycol (PEG) surface modification and hydrothermal treatment exhibited enhanced long-term stability in buffer solutions and cell culture media at physiological temperature as compared to unmodified NPs and PEGylated NPs without hydrothermal treatment (see Example 3).

Mesoporous silica nanoparticles (MSNs) were first imagined as a vehicle for drug delivery by Lin et al. (Lai et al., 2003) ; since then, many reseachers have begun to explore the potential of these nanoparticles (NPs) for various biomedical applications such as intracellular labeling (Lin et al.,

2005), biomolecule delivery (Slowing et al. , 2007), targeting (Rosenholm et al. , 2009), and diagnostic imaging (Kim et al., 2008). Although much of the published work has successfully shown the in vitro therapeutic promise of MSNs, very few in vivo experiments can be found in the literature due to the large size (> 150 nm diameter) and low stability of MS NPs in biological media, resulting in rapid unintentional uptake by the reticuloendothelial (RES) system (Wu et al., 2008; Taylor et al., 2008; Lee et al., 2009). Although synthesis of fine bare MSNs (< 50 nm) has been reported by several groups recently (Kobler et al. , 2008; Lu et al., 2009; Urata et al., 2009), particle dispersity has been difficult to maintain, and aggregation occurs either during synthesis or surfactant removal (Lu et al. , 2009). Removal of the surfactant template from MSNs during the synthetic procedure determines retention of the initial dispersity of MSNs, especially for small MS NPs (i.e., diameters less than 50 nm).

Kuroda et al. showed that using a dialysis process to remove surfactant can prevent aggregation of small MSNs and maintain their hydrodynamic size (Urata et al., 2009). Another strategy to maintain NP dispersity is via surface passivation with polyethylene glycol (PEG) ; PEG has been known to prevent protein adsorption (opsonization) on nanoparticle surfaces, enhance circulation time, and reduce nonspecific uptake by the RES system (Fang et al., 2009; Rio-Echevarria et al. , 2010).

Though the pegylation of MSNs has been shown to reduce their hemolytic activity (Lin et al., 2010; Lin et al. , 2009) and serum binding (He et al. , 2010), the long-term stability of pegylated MSNs in various biological media such as phosphate buffered saline (PBS) or cell culture media at 37°C had not, prior to the present disclosure, been shown.

The hydrodynamic size and aggregation state of MSNs, as determined by dynamic light scattering (DLS), has only examined in deionized (D.I.) water (Lu et al., 2009) or organic solvents (Urata et al., 2009). In most cases, biomedical MSNs will be suspended in or experience highly salted solutions or serum-containing media, conditions likely to have a large influence on the hydrodynamic diameter. In addition, the short-term stability (hours) of MSNs may be different from their long-term (days) stability.

"Mesoporous silicate" refers to a mesoporous structure formed by the acid or base catalyzed condensation of a silicon containing material around a surfactant template, forming typically uniform channel structures. As used herein, the terms "mesoporous silicate", "mesoporous silicate body", "mesoporous silicate particle", and "mesoporous silicate nanoparticle" (MSN) can be used

interchangeably.

The mesoporous silicate body can have an average particle size (or average particle diameter for spherical particles) of about 20 to about 300 nm, about 20 to about 200 nm, or about 30 to about 150 nm (prior to surface modification), and can have an average pore diameter of about 1 to about 10 nm, about 2 to about 8 nm, or about 2.5 nm to about 4 nm. In one embodiment of the invention, the pores have a diameter of at least about 2 nm. In other embodiments, the pores have diameters of greater than about 5 nm, or greater than about 10 nm. The particles can have various pre-determined shapes, including, e.g., a spheroid shape, an ellipsoid shape, a rod-like shape, or a curved cylindrical shape. Average particle size of the nanoparticles can be measured using transmission electron microscopy. Particle size refers to the number average particle size and may be determined using any suitable technique, e.g., using an instrument that uses transmission electron microscopy or scanning electron microscopy. Another method to measure particle size is dynamic light scattering that measures weight average particle size as the hydrodynamic diameter. Nanoparticles may be formed of only silica, or they may be composite nanoparticles such as core- shell nanoparticles. A core-shell nanoparticle can include a core of an oxide (e.g., iron oxide) or metal (e.g., gold or silver) of one type and a shell of silica deposited on the core. Silica nanoparticles may be derived from a silicate, such as an alkali metal silicate or ammonium silicate. The unmodified

nanoparticles may be provided as a sol rather than as a powder.

The particles may include a coating comprising a polymer. The particles can be coated, either by forming covalent bonds to a polymer or by encapsulating the particles within a polymer. The polymer coating can act to slow the rate of diffusion of the cations in the ligand template from the pores of the mesoporous silicate body when it is in contact with a liquid. The polymer can be an adhesive, such as a bioadhesive. The adhesive can adhere the particle to the oral tissue of a mammal, such as a human, a human companion, or a farm animal, when the silicate body is contacted with the mouth of a mammal. Alternatively, adhesive can adhere the silicate body to the skin or other mucus membranes of a mammal when the when the silicate body is contacted with cells or membranes.

The polymer can be any suitable and effective polymer that, when covalently bound to the surface of the silicate body, acts to slow the diffusion of the cations from the pores. Non-limiting examples of a polymer coating include polyethylene glycol (PEG) and poly(lactic acid). Additionally, an adhesive can be suitably prepared using a silicone based pressure sensitive adhesive, such as a (polydimethyl-siloxane- silicate resin) copolymer adhesive depicted by the following formula:

Polydimethylsiloxane

wherein R is— Si(CH₃)₃, and x and y represent independent numbers of repeating units sufficient to provide the desired properties in the adhesive polymer or other polymer layers.

The mesoporous silicates may be prepared from surfactant micelles of C₁₀-C₁₈

alkyl(trialkyl)ammonium salts in water, followed by introduction into the solution of an alkyl orthosilicate, such as tetraethylorthosilicate (TEOS), and one or more functionalized silanes, such as one or more mercaptoalkyl-, chloroalkyl-, aminoalkyl-, carboxyalkyl-, sulfonylalkyl-, arylalkyl-, alkynyl-, or alkenyl- silanes, wherein the (C₂-C₁₀)alkyl chain is optionally interrupted by— S— S— , amido (— C(=0)NR— ), — O— , ester (— C(=0)0— ), and the like. The aqueous mixture is stirred at moderate temperatures until the silicate precipitates, and it is collected and dried. The surfactant "template" is then removed from the pores of the ordered silicate matrix, for example, by refluxing the silicate in aqueous-alcoholic HCI or alcoholic NH₄N0₃. The remaining solvent can be removed from the pores of the silicate by placing it under high vacuum. Functional groups incorporated on the surface of the pores may be quantified and further modified by attaching terminally-functionalized organic linker moieties that can be reacted with functional groups on the caps. The polarity of the interior of the pores can also be adjusted by adding other functionalized silanes to the reaction mixture, including ones comprising non-polar inert groups such as aryl, perfluoroalkyl, alkyl, arylakyl and the like. The exterior of the silicate matrix can be functionalized by grafting organic moieties comprising functional groups thereto, which in turn may be linked to biologically active agents, imaging agents targeting or labeling moieties.

Bioactive agents within the scope of the invention include pharmaceutical agents, diagnostic agents, genes, nutrients (vitamins, etc.), and pesticidal agent (e.g., insecticides, herbicides, and rodentacides). For example, the term includes conventional chemotherapeutic agents useful to treat cancer (see PCT US/00/16052), chelated radionuclides, immunosuppressive drugs, antiinflammatory agents, antibacterial agents, antifungal agents, antiviral agents (see U.S. Patent No. 4,950,758) analgesic agents (see U.S. Patent Nos. 5,298,622 and 5,268,490), polypeptides; hormones, hormonal messengers, and cytokines (e.g. insulin, interleukins, interferons, human growth hormone, PTK, TPMT, TGF-.beta., EPO, TNF, ΝΚ-β, and prostaglandins, and the like), imaging agents, contrast agents, enzymes for enzyme replacement therapy (see, U.S. Patent Nos. 6,106,834 and 6,210,666), antibodies (see U.S. Patent Nos. 5,034,222 and 6,106,834); and RNA or DNA molecules of any suitable length (see PCT US/00/16052, PCT WO96/30031 , and U.S. Patent Nos. 5,591 ,625, 6,387,369, 5,190,931 , 5,208,149, and 5,272,065).

The invention will be further described by the following non-limiting examples.

Example 1

Materials and Methods

Materials. All chemicals were used without further purification. n-Cetyltrimethylammonium bromide (CTAB, 99%), fluorescein isothiocyanate isomer I (FITC, 90%), rhodamine B isothiocyanate mixture of isomers (RITC), tetraethyl orthosilicate (TEOS, 98%), dimethyl sulfoxide (DMSO, 99.9%), and polyvinyl pyrrolidone (PVP, average MW 10 000 g/mol) were obtained from Sigma Aldrich. 3- Aminopropyl-trimethoxysilane (APTMS, 97%) was obtained from Fluka. [Hydroxy(polyethyleneoxy)propyl] triethoxysilane, (PEG-Silane, MW575-750 g/mol, 50% in ethanol) was purchased from Gelest. 3-(4,5- Dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide (MTT) was obtained from Invitrogen. Ammonium nitrate (99.9%), ammonium hydroxide (NH₄-OH, 28-30 wt% as NH₃), chloroform (99.8%), hydrochloric acid (HCI, 36.5 - 38%), and iron(lll) chloride hexahydrate (FeCI₃ · 6H₂0, > 99%) were purchased from Mallinckrodt. DyLight 800 NHS ester (98 - 100%), iron(ll) chloride tetrahydrate (FeCI₂ · 4H₂0, 99 - 102%), and oleic acid were obtained from Fisher Scientific. Absolute anhydrous and 95% ethanol was purchased from Pharmco-Aaper. Ultrapure water was generated using a Millipore Milli-Q system with a Milli-pak filter of 0.22 μιη pore size and used for all the preparation of aqueous solutions.

Preparation of Hydrophobic Magnetite (Fe₃Cv) NPs. The hydrophobic magnetite NPs were synthesized based on a slight modification of a published one-pot chemical coprecipitation method (Molday, 1984). First, the deionized water was purged with nitrogen gas for 10 minutes. Then, 4.80 g of FeCI₃ · 6H₂0, 2.00 g, FeCI₂ · 4H₂0, and 0.85 ml_ oleic acid were added to 30 ml_ of deionized water under nitrogen atmosphere with vigorous stirring. The mixture solution was heated to 90°C. Then, 20 ml_ of ammonium hydroxide (14 wt %) was added rapidly to the solution, and it immediately turned black. The reaction was kept at 90°C for 2.5 hours and then allowed to cool to room temperature. The black precipitate was collected by centrifugation at 10,016 g for 10 minutes and resuspended in chloroform with a end concentration of 54.5 mg oleic acid-capped Fe₃0₄/mL.

Synthesis of PEG-Modified Fluorescent Mesoporous Silica NPs with Incorporated Fe₃0₄ NPs

(Fe₃0₄(a>Dve-MSNs(a>PEG) and PEG-Modified Fluorescent Hollow Mesoporous Silica NPs (H-Dye-MSN):

Dye = FITC, RITC, and DyLight 800. The preconjugated N-1 -(3-trimethoxy-silylpropyl)-N-fluoresceyl thiourea (FITC-APTMS) was prepared by combining 5 μΙ_ of APTMS and 2 mL of 0.023 M FITC ethanolic solution under continuous stirring and dark conditions (Lin et al., 2005). The RITC-APTMS was prepared in the same manner. In a typical synthetic procedure for 33-nm-diameter Fe₃0₄@FITC-MSNs@PEG, 1 .2 mL of oleic acid-capped Fe₃0₄ nanoparticle solution was added to 5 mL of 0.16 M aqueous CTAB containing 0.2 g PVP10. The resulting solution was ultrasonicated at 42 kHz for 1 hour to evaporate the chloroform, resulting in a transparent black solution. This solution was added to a 150 mL of 0.255 M ammonium hydroxide solution and heated to 50°C. Then, 450 μί of FITC-APTMS and 3.0 mL of 0.88 M dilute ethanolic TEOS were added sequentially to the reaction solution under continuously stirring (600 rpm). After 30 minutes, 600 μί of PEG-silane was added, and the solution was allowed to stir for another 30 minutes. The solution was then aged at 50°C for 24 hours. The as-synthesized Fe₃0₄@FITC- MSNs@PEG solution was centrifuged at 53 300g for 30 minutes and, then, washed and redispersed in 50 mL of ethanol. To avoid the Fe₃0₄ dissolution that occurs under acidic conditions, the surfactants were removed using a fast and efficient ion exchange method where the assynthesized Fe₃0₄@FITC-

MSNs@PEG was transferred to 50 mL of ethanol containing 0.3 g of NH₄N0₃ and kept at 60°C for 1 hour (Lang et al., 2004). The extraction step was repeated twice to completely remove the surfactants. To generate PEG-coated hollow fluorescent mesoporous silica NPs (H-FITC-MSNs@ PEG), the Fe₃0₄ NPs and CTAB surfactants were removed using acidic ethanol (pH < 1 .0) under reflux for 6 hours. After the extraction, these NPs were washed three times with ethanol and redispersed/stored in 20 mL of ethanol. The Fe₃0₄@RITC-MSNs@PEG was obtained by substituting RITC-APTMS for FITC-APTMS in the previously mentioned reaction. For Fe₃O₄@DyLight800-MSNs@PEG, the Fe₃0₄@MSNs@PEG was prepared without the addition of dye. Then, the Fe₃0₄@MSNs@PEG was modified with APTMS at 60°C for 6 hours. The Fe₃O₄@DyLight800-MSNs@PEG was synthesized by adding 100 μΐ of 10 mg/mL N- hydroxysuccinimide (NHS)-ester-activated DyLight800 (dissolved in DMF) to an amine-functionalized

Fe₃0₄@MSNs@PEG colloidal solution. The reaction was carried out in a Dulbecco's phosphate-buffered saline (DPBS, Gibco) at room temperature for 1 hour. Once the coupling was complete, all purification and surfactant extraction followed the procedure described above. The solid products for XRD and BET characterizations were obtained by centrifugation followed by drying at 60°C for 24 hours.

Cell Culture and MTT Viability Assay. The PC-12 pheochromocytoma cells (American type

Culture Collection, ATCC, CRL-1721 ) were cultured at 37°C under 5% C0₂ in RPMI-1640 media (Lonza) containing 10% horse serum (Invitrogen), 5% fetal bovine serum (FBS, Sigma), and 1 %

penicillin/streptomycin (Gibco). The HeLa and HCT-1 16 cells (a generous gift from Professor Duncan Clarke, University of Minnesota) were cultured at 37°C under 5% C0₂ in DMEM media (HyClone) supplemented with 10% FBS and 1 % penicillin/streptomycin. All three cancer cell lines were used to screen cytotoxicity of 33 nm Fe₃0₄@FITC-MSN@PEG. Typically, 1 X 10⁵ of cells per well were plated in 24-well plates for the MTT viability assay. After 24 hours for cell attachment, the cells were incubated with varied concentrations of 33 nm diameter Fe₃0₄@FITCMSNs@PEG NPs in media for 12 hours. Then, the

NP-treated cells were washed three times with 500 μΙ_ of PBS and allowed to incubate with MTT media

(0.5 mg/mL) for 2 hours at 37°C. The formazan dye crystals generated by live cells were dissolved in DMSO, and the absorbance values at 570 nm were determined by using a microplate reader with absorbance at 655 nm as a reference. The cell viability was calculated by comparing the absorbance of a nanoparticle treated well to that of the control well.

Hemolysis Assay. Human blood was obtained from Memorial Blood Centers (St. Paul, MN). The whole blood was diluted to 1/10 of its original volume using calcium and magnesium-free DPBS solution. Red blood cells (RBCs) were isolated by centrifugation at 10 016g for 10 minutes, washed, and resuspended five times with PBS solution. Then 0.2 mL of diluted RBCs suspension was added to 0.8 mL of nanoparticle solution at different concentration and mixed by vortexing. Incubation of deionized water and DPBS with RBCs were used as positive control and negative control, respectively. All the samples were kept in static conditions at room temperature for 3 hours. Finally, the mixtures were centrifuged at 10 016g for 3 minutes. The absorbance values of supernatants at 570 nm were determined by using a microplate reader with absorbance at 655 nm as a reference. The hemolysis percentage of RBCs was calculated based on the formula shown below.

Hemolysis percentage = ((sample absorbance - negative control absorbance)/(positive control absorbance - negative control absorbance)) X 100.

Characterization. Transmission electron microscopy (TEM) images were taken on a JEOL JEM-

1210 microscope (operating at 120 kV). Samples were prepared on a Formvar-coated copper grid by evaporating one drop of diluted ethanolic suspension of the NPs. Image analysis was performed using SigmaScan Pro 5.0. Powder X-ray diffraction (XRD) was performed on a Bruker-AXS D-5005 (Siemens) X-ray diffractometer using filtered CuKR radiation (λ = 0.154 nm) at 45 kV and 20 mA. Data were collected by step scan with a step size of 0.018° and a step time of 1 .0 s. Surface area, N₂ adsorption- desorption isotherms, and pore-size distributions were obtained using of an Autosorb-1 analyzer

(Quantachrom Instruments) at 77 K. Prior to measurements, samples were degassed at 120°C for 12 hours. The surface area was calculated using Brunauer-Emmett-Teller (BET) equation in the range P/P₀ < 0.3. The zeta potential was measured using a zeta potential analyzer (Brookhaven Instruments Corporation) in 10 mm polystyrene cuvettes at 25°C. The hydrodynamic size distribution was estimated by dynamic light scattering (DLS) with a 90 Plus/BI-MAS particle size analyzer (Brookhaven Instruments Corporation). Photoluminescence spectra were collected on a JASCO FP-6200 (400-700 nm) or Photon Technology International (200-900 nm) fluorescence spectrometer. The room-temperature magnetization curves were measured using a Quantum Designs MPMS-5S cryogenic susceptometer. The cell viability and hemolysis assays were measured using an iMark microplate reader (Bio-Rad).

Results and Discussion

The procedure for Fe₃0₄@Dye-MSNs@PEG synthesis is illustrated in Figure 1 . First, oleic acid- capped-magnetite NPs were synthesized using a facile co-precipitation method. Compared to the thermal decomposition method developed by Sun et al. (2002), this simple synthesis not only utilizes inexpensive reactants but also generates a gram scale of products. The as-synthesized Fe₃0₄ NPs had an average nanoparticle diameter of 1 1 + 3 nm. The hydrophobic Fe₃0₄ NPs dispersed in chloroform were transferred to an aqueous phase by using n-cetyltrimethylammonium bromide (CTAB) and polyvinyl pyrrolidone (PVP) as stabilizing agents. The CTAB added in this step also serves as the porous structure-directing agent during the later silica condensation. PVP was incorporated into the synthesis of the Fe₃0₄ NPs as a cosurfactant for CTAB during the phase transfer to improve the transfer efficiency of the hydrophobic Fe₃0₄ NPs to the aqueous phase. In a previous report using CTAB as a stabilizing agent to transfer hydrophobic NPs to the aqueous phase, the transferred NPs needed to be filtered through a

0.44 μιη syringe filter before silica coating, indicating that large aggregates were present in solution (Liong et al., 2008). Without the addition of PVP, synthesized Fe₃0₄@Dye-MSNs@PEG included large aggregates of Fe₃0₄ NPs. Upon filtering the PVP-free CTAB-Fe₃0₄ nanoparticle solution through a 0.44 μιη syringe filter, a large amount of NPs were stuck in the filter (data not shown). In contrast, when PVP was added as a cosurfactant during synthesis, the aqueous phase CTAB- and PVP-stabilized Fe₃0₄ NPs passed through a 0.20 μιη syringe filter with no observable aggregates. Additionally, no Fe₃0₄ aggregates were detected once incorporated into the silica NPs. This result is not surprising since PVP is well-known as an amphiphilic polymer and, thus, can serve as a protecting agent for Fe₃0₄ NPs and improve their aqueous dispersity (Graf et al., 2003).

The mesoporous silica nanostructure is formed around the aqueous Fe₃0₄ NPs by carefully controlling pH during the condensation reaction. During the mesoporous silica formation, the

incorporation of dye molecules into the silica framework can be easily accomplished using a co- condensation method. Finally, to increase the water dispersity for biomedical applications, Fe₃0₄@Dye- MSNs were simultaneously modified with PEG-silane to prevent the binding of proteins and minimize NP aggregation. A variety of Fe₃0₄@FITC-MSNs@PEG sizes can be synthesized using the methods detailed herein. As the ratio of the Fe₃0₄ NPs to silicate precursor used in the synthetic process is decreased, the resulting size of the Fe₃0₄@FITC-MSNs@PEG increases, making it simple to control the overall nanoparticle size. The multifunctional silica NPs have good aqueous dispersity, an average size smaller than 70 nm and possess a well-ordered mesoporous structure. Furthermore, of the low temperature preparation of Fe₃0₄ NPs, dissolution of incorporated Fe₃0₄ NPs with acidic ethanol produces mesoporous silica NPs containing additional void space with a route that is easier and more economical than previously reported methods (Yang et al., 2008; Darbandi et al., 2007). These hollow FITC-MSNs have an increased drug loading capacity due to the presence of both the hollow voids and the ordered mesopores.

The mesoporous ordering of Fe₃0₄@FITC-MSNs@PEG with varied diameters was examined using low-angle XRD from 2Θ = 1 .5-8°. The ordering of the porous structure increased as the

concentration of Fe₃0₄ NPs decreased. The TEM images of 33 and 42 nm diameter Fe₃0₄@FITC- MSNs@PEG suggest short-range ordering and wormlike porous structures inside the NPs. The visual conclusions are supported by the broad and unresolved (1 10) and (200) XRD peaks detected when probing the 33 and 42 nm diameter Fe₃0₄@FITC-MSNs@PEG. The sharp XRD patterns from 53 and 67 nm diameter Fe₃0₄@FITC-MSNs@PEG indicate a very well-ordered 2D hexagonal mesoporous structure. This result confirms the long-range ordering of porous structure seen in TEM images. The crystalline structure of Fe₃0₄ both before and after incorporation into Fe₃0₄@Dye-MSNs@PEG was characterized using wide-angle XRD. These patterns can be easily indexed to Fe₃0₄ by the International Centre for Diffraction Data powder diffraction file 98-000-0073 for magnetite. This result shows that the embedded Fe₃0₄ NPs retain their magnetite crystalline structure after template extraction. The N₂ adsorption-desorption isotherms for Fe₃0₄@FITC-MSNs@PEG with varied sizes exhibit the characteristic IV behavior for a well-developed mesoporous structure, with a sharp capillary condensation at P/P₀ about

0.35. In addition, all samples have a secondary adsorption step (at P/P₀ about 0.95) that is attributed to the interparticle spaces formed between NPs after drying. All measured structural characteristics are listed in Table 1 .

Table 1 . Structural Properties and Saturation Magnetization Values Of Surfactant-free Fe₃0₄@FITC-MSNs@PEG with Varied Sizes

³¾ΕΤ- BET surface area calculated from data at P/P₀ = 0.05 - 0.25.

bD_BJH: pore diameter assigned from the maximum on the BJH pore size distribution.

°V_t: total pore volume calculated at P/P₀ at 0.99.

All samples have high Brunauer-Emmett-Teller (BET) surface area (694 - 1045 m²/g) and large total pore volume (1 .08 - 1 .95 cm³/g). The decrease in surface area with decreased size is caused by the increasing volume fraction of magnetic NPs. Compared to recent reports (Kim et al., 2008; Deng et al., 2008), the Fe₃0₄@FITC-MSNs@PEG synthesis detailed herein not only allows access to an average diameter smaller than 50 nm but also maintains the porous integrity and high surface area (885 m²/g). The vast improvement in achieved mesostructure, even in the small size regime, is likely due to the use of NH₄OH as a basic catalyst which favors formation of longer cylindrical silicate micelles within the Gouy- Chapman region (Cai et al., 2001 ).

The Barret-Joyer-Halenda (BJH) method was applied to calculate the pore size distribution in all samples, indicating a maximum pore diameter of about 2.4 nm. The magnetic behavior of Fe₃0₄@FITC- MSNs@PEG with varied diameters was measured using a magnetometer tuned from -20 000 to 20 000 Oe. No hysteresis was detected at 300 K for any of the samples, indicating that the superparamagnetic behavior of the incorporated Fe₃0₄ NPs, an essential characteristic for T₂ MRI contrast agents, is intact. The high magnetic response is further demonstrated by applying a magnet to the outside of a vial containing the Fe₃0₄@FITC-MSNs@PEG aqueous solution. This high magnetization will facilitate the use of these multifunctional NPs as T₂ MRI contrast agents, in bioseparations, and in cell tracking. As shown in Table 1 , the saturation magnetization value of Fe₃0₄@FITC-MSNs@PEG increases as the overall nanoparticle size decreases, further confirming the increasing Fe₃0₄ volume fraction. This volume fraction can be estimated by comparing the saturation magnetization values to that measured for the unincorporated 1 1 -nm-diameter Fe₃0₄ NPs (60.0 emu/g, data not shown). The saturation magnetization value for the 33 nm diameter Fe₃0₄@FITC-MSNs-@PEG is 12.0 emu/g and, thus, the magnetic NP content is 20 wt %. Compared to recent reports (Kim et al., 2006; Lin et al., 2006; Kim et al., 2008), the Fe₃0₄@FITC-MSNs@PEG NPs reported herein are not only smaller in size but also possesses much higher magnetization than previously reported nanomaterials.

The photoluminescence spectra of surfactant-free Fe₃0₄@FITC-MSNs@PEG and Fe₃0₄@RITC- MSNs@PEG exhibit the typical fluorescence emission of FITC and RITC; this result confirms that the dye molecules can be anchored stably to the mesoporous silica framework after the surfactant extraction. Initial multifunctional NP syntheses focused on visible emission dyes, and in fact, different emission wavelengths were observed from Fe₃0₄@FITC-MSNs@PEG and Fe₃0₄@RITC-MSNs@PEG under UV illumination, as expected. However, some applications of multifunctional biomedical NPs require fluorescent emission in the near-infrared (NIR) where the low light absorptivity of biological chromophores allows light to penetrate through several centimeters of tissue. Molecular dyes such as Cy5.5 and indocyanine green (ICG) are widely used in in vitro cell labeling and as NIR imaging contrast agents for in vivo deep-tissue imaging (Ke et al., 2003). Hence, in addition to the previously mentioned visible region fluorophores, the NIR dye DyLight800 was incorporated into the multifunctional NPs. This was accomplished by reacting the primary amine groups from silica-bound aminopropyltrimethoxysilane with NHS-ester-activated Dyl_ight800 fluorophores. The photoluminescence spectrum of surfactant-free Fe₃O₄@DyLight800-MSNs-@PEG showed the NIR dye was successfully incorporated into the silica wall by covalent bonding. The PEG-silane was introduced into the synthetic procedure to enhance the colloidal stability. As is shown in Figure 5b, these surfactant-free NPs remained well-dispersed in deionized water without any visible aggregation after room temperature storage for 15 days. Dynamic light scattering (DLS) revealed that 33 nm diameter Fe₃0₄@FITC-MSNs@PEG NPs have a hydrodynamic diameter of 57 + 1 nm (Figure S3a in the Supporting Information) in deionized water and 73 + 1 nm in PBS solution. The size difference between TEM and DLS measurements is attributed to the hydration sphere and outer PEG layer. Zeta potential measurement showed that 33 nm diameter Fe₃0₄@FITC- MSNs@PEG NPs have a surface charge of -24 + 2 mV in deionized water, further supporting the conclusion that a stable suspension is formed. The hydrodynamic diameter stability of 33 nm

Fe₃0₄@FITC-MSNs@PEG in deionized water and PBS solution were measured at different times. There is no significant size change detected either in deionized water or PBS, further supporting the absence of nanoparticle aggregation. The photographs of 33 nm Fe₃0₄@FITC-MSNs@PEG NPs in deionized water and PBS solution visually confirms the excellent dispersity and high colloidal stability, because no particle precipitation is observed.

Initial evaluation of the in vitro biocompatibility of 33 nm Fe₃0₄@FITC-MSNs@PEG NPs was performed using the MTT viability and hemolysis assays. MTT results showed that HeLa, PC-12, and HCT-1 16 cell viability were not affected even with 12 hours exposures of up to 200 μg/mL NPs. In addition to the MTT assay, because the Fe₃0₄@Dye-MSNs@PEG NPs are designed for intravenous administration, it is important to test the compatibility of these multifunctional NPs with blood cells. In addition to the multifunctional nanoparticles described above, two other types of commonly used silica NPs, nonporous Stober and Fe₃0₄@Si02 core-shell NPs, were synthesized to facilitate comparison of hemolytic activity with that of 33 nm Fe₃0₄@FITC-MSNs at different concentrations from 12.5 to 1000 g/mL for 3 hours. Typically, a large dose of NPs is used to achieve enough imaging contrast and therapeutic efficacy. Hence, the RBCs were incubated with NPs at a high concentration, up to 1000 g mL.

No hemolysis of RBCs was caused by 33 nm Fe₃0₄@FITC-MSNs at a concentration up to 100 μg/mL, but over 90% of RBCs were lysed by nonporous Stober and Fe₃0₄@Si0₂ core-shell NPs at a concentration of 50 μg/mL. In the case of 33 nm Fe₃0₄@FITC-MSNs, detectable release of hemoglobin was observed at concentrations over 200 μg/mL. However, even after 3 hours incubation, only 20% hemolysis was observed from 33 nm Fe₃0₄@FITC-MSNs. This is likely due to the lower number density of silanol groups on the surface of mesoporous silica shell compared to that of the solid silica

nanoparticles. As a separate control, pure Fe₃0₄ NPs showed no hemolytic effect on RBCs (data not shown), indicating that the unlikely liberation of the magnetic center from the multifunctional nanoparticles would not be problematic.

While these measurements clearly reveal that the mesoporous silica coating greatly reduces the hemolytic activity of silica NPs compared to the nonporous silica coating, further blood compatibility can be promoted via surface modification of either porous or nonporous silica NPs (Dorovolskaia et al., 2008). Herein, PEG-silane was used to eliminate the silanol groups on the surface of nonporous Stober Si0₂ NPs and 33 nm Fe₃0₄@FITC-MSNs. Nonporous Stober Si0₂ NPs were reacted with PEG-silane in two different periods, 12 and 24 hours. The resulting PEG-coated silica NPs are referred to as Si0₂@PEG-12 and Si0₂@PEG-24, with expected variations in PEG coverage and ordering. Compared to the 12 hours PEG-silane coating, the hemolytic activity of nonporous Stober Si0₂ NPs was greatly reduced with 24 hours PEG-silane coating; however, detectable hemolysis still can be seen at concentrations over 600 μg/mL. Contrary to Si0₂@PEG NPs, no hemolysis was observed after 3 hours incubation with 33 nm Fe₃0₄@FITC-MSNs@PEG at concentrations ranging from 12.5 to 1000 μg/mL. This result further confirms that the surface silanol groups of Fe₃0₄@FITC-MSNs are much fewer than nonoporous Stober Si0₂ NPs and can be eliminated by PEG modification. Finally, long-term hemolysis of RBCs was tested in the presence of 33 nm Fe₃0₄@FITC-MSNs@PEG at the concentration of 1000 μg/mL, and no significant hemolysis was detected even after 36 hours incubation (Figure S8 of the Supporting Information), confirming the biocompatibility of 33 nm Fe₃0₄@FITC-MSNs@PEG.

Conclusions

In summary, this work demonstrates a one-pot size controllable synthesis of multifunctional mesoporous silica NPs having well-ordered structure and large surface areas (700-1000 m²/g), high magnetization (2-12 emu/g), PEG coating, and diameters less than 70 nm. Dissolution of embedded Fe₃0₄ NPs produces hollow mesoporous silica NPs as well. Visible fluorophores, NIR fluorophores, and magnetic NPs were all incorporated into the porous silica NPs without compromising the mesoporous structure to facilitate multimodal imaging. Importantly, these PEG-modified multifunctional mesoporous silica NPs exhibit excellent colloidal stability in both water and PBS solutions. MTT and hemolysis results further confirm their high biocompatibility. Compared to nonporous silica nanoparticles, mesoporous silica materials can not only reduce red blood cell membrane damage but also provide a large surface area for drug loading. We believe these biocompatiable multifunctional NPs have great potential for bioimaging and drug delivery applications.

Example 2

Materials and Methods

Chemicals. All chemicals were used without additional purification. n-Cetyltrimethylammonium bromide (CTAB, 99%) and tetraethyl orthosilicate (TEOS, 98%) were purchased from Sigma Aldrich (Milwaukee, Wl). [Hydroxy(polyethyleneoxy)propyl] triethoxysilane, (PEG-silane, MW 575-750 g/mol, 8- 12 EO, 50% in ethanol) was obtained from Gelest (Morrisville, PA). Ammonium nitrate (99.9%) and ammonium hydroxide (NH₄OH, 28-30 wt% as NH₃) were obtained from Mallinckrodt (Phillipsburg, NJ). Absolute anhydrous ethanol and 95% ethanol were purchased from Pharmco-Aaper (Brookfield, CT). Calcium- and magnesium-free Dulbecco's phosphate buffered saline (PBS) was obtained from Invitrogen (Grand Island, NY). The de-ionized (D.I.) water was generated using a Millipore Milli-Q system (Billerica,

MA).

Characterization. Transmission electron microscopy (TEM) micrographs were taken on a JEOL 1200 EXM (Tokyo, Japan) with a 100 kV voltage. TEM specimens were prepared by evaporating one drop of ethanolic nanoparticle solution on Ted Pella Formvar-coated copper grids (Redding, CA). Nanoparticle size was measured in the micrographs using Sigma Scan Pro 5.0 software (Ashburn, VA). Powder XRD data were recorded on a Siemens Bruker-AXS D-5005 X-ray diffractometer using filtered Cu Ka radiation (λ = 1 .5406 A) at 45 kV and 20 mA. Data were recorded by step scan with a step size of 0.040° and a step time of 1 .0 second. Nitrogen adsorption-desorption measurements were carried out on a

Quantachrom Autosorb-1 analyzer (Boynton Beach, FL) at 77 K. Samples were outgassed at 120 °C for at least 12 hours before measurements. The specific surface area was calculated using Brunauer- Emmett-Teller (BET) equation at P/P₀ < 0.3. The pore size distribution was calculated from the branch of the adsorption isotherm using a Barrett-Joyner-Halenda (BJH) method. The primary and total pore volume was determined at P/P₀ = 0.50 and 0.99, respectively. The absorbance of hemoglobin was measured using a Bio-Rad iMark microplate reader (Hercules, CA).

Preparation of Nonporous and Porous Silica Nanoparticles (NPs). a. Svthesis of Stober Silica (SS) NPs with varied sizes. Uniform nonporous silica nanospheres were synthesized using a well-known method developed by Stober et al. (1968). The nonporous silica particle diameter was tuned by varying the amounts of ammonium hydroxide and TEOS. The synthesis conditions and composition of the synthesis mixture are described below in Table 2.

Table 2. Synthesis conditions of SS NPs with varied sizes

Samples 95% Ethanol 28-30% NH₄OH (mL) ^T^ Temperature (K)

SS-24 40 0.75 0.50 313

SS-37 40 1 .0 0.50 313

SS-142 40 2.0 0.75 313

SS-263 40 3.0 1 .5 313

The as-synthesized SS NPs were collected by centrifugation and washed with absolute ethanol twice to remove unreacted precursors. All SS NPs for hemolysis assays were suspended and diluted in DPBS before use. The SS NPs with mean diameters of 24, 37, 142, and 263 nm are referred to as SS-24, SS- 37, SS-142, and SS-263, respectively.

b. Synthesis of Mesoporous Silica (MS) and Polyethylene Glycol (PEG)-coated MSNs with varied sizes. The porous silica NPs were prepared using an ammonia base-catalyzed method under highly dilute and low surfactant conditions (Lin et al., 2005). The particle size was controlled by adjusting ammonia concentration, TEOS volume added, and reaction temperature. The detailed synthesis conditions are shown in Table 3. Table 3. Synthesis conditions of MSNs with varied sizes

MS-25 0.128 150 2.0 333

MS-42 0.256 150 2.5 323

MS-93 0.512 150 3.0 313

MS-155 0.768 150 3.5 303

MS-225 1 .024 150 4.0 303

Typically, 0.29 g of CTAB was dissolved in 150 mL of ammonium hydroxide solution (0.128, 0.256, 0.512, 0.768, and 1 .024 M) at the desired temperature (30, 40, 50, 60^<€). Then, 0.88 M ethanolic TEOS was added to the solution under vigorous stirring (600 rpm). After one hour, the mixture solution was aged for at least 12 hours in static condition. The as-synthesized colloid was transferred to 50 mL of ethanolic ammonium nitrate solution (6 g/L) with continual stirring at 60 °C for one hour to remove surfactant. The surfactant extraction step was repeated two times to ensure removal of CTAB. The extracted NPs were washed with ethanol twice and resuspended in absolute ethanol. All MSNs were suspended and diluted in PBS before hemolysis assay use. The MSNs with mean diameters of 25, 42, 93, 155, and 225 nm are designated as MS-25, MS-42, MS-93, MS-155, and MS-225, respectively. To modify the outer surface of MSNs, 600 μί of PEG-silane (50% in ethanol) was added after the formation of MSNs. Herein, two sizes of MSNs, MS-25 and MS-42 were chosen as PEG-modified examples. The code names for PEG-modified MSNs were MS-25@PEG and MS-42@PEG, respectively.

Assessing the Stability of MS-25, MS-42, MS-25(5>PEG, and MS-42(5>PEG. All stability test samples were prepared in DPBS solution at 2 mg/mL concentration and aged for 6 days at room temperature. Then, the aged NPs were centrifuged and the supernatant was saved for hemolysis assays. The obtained NPs were washed with absolute ethanol two times, and dried at 60°C. The solid products were used for XRD and N₂ adsorption-desorption characterization.

Hemolysis Assay. Ethylenediamine tetraacetic acid (EDTA)-stabilized human blood samples were freshly obtained from Memorial Blood Center (St. Paul, MN). First, 5 mL of blood sample was added to 10 mL of PBS, and then red blood cells (RBCs) were isolated from serum by centrifugation at 10016 g for 10 minutes. The RBCs were further washed five times with 10 mL of PBS solution. The purified blood was diluted to 50 mL of DPBS. Prior to nanoparticle exposure, the absorbance spectrum of the positive control supernatant was checked and used only if it was in the range of 0.50-0.55 optical density units to reduce sample difference from different donors. Herein, RBC incubation with D.I. water and PBS were used as the positive and negative controls, respectively. Then 0.2 mL of diluted RBC suspension was added to 0.8 mL of Stober and mesoporous silica nanoparticle solutions at systematically varied concentrations and mixed by vortexing. The silica NPs suspended in DPBS solutions with different concentrations were prepared immediately before red blood cell incubation by serial dilution. All the sample tubes were kept in static condition at room temperature for 3 hours. Finally, the mixtures were centrifuged at 10016 g for 3 minutes, and 100 μί of supernatant of all samples was transferred to a 96- well plate. The absorbance values of the supernatants at 570 nm were determined by using a microplate reader with absorbance at 655 nm as a reference. The percent hemolysis of RBCs was calculated using the formula shown below.

Percent hemolysis = ((sample absorbance - negative control absorbance)/(positive control absorbance - negative control absorbance)) x 100.

Calculations and Data Analysis. Every experimental condition was repeated at least three times in triplicate. All hemolysis data are presented as mean ± standard deviation (SD). The concentration leading to 50% lysis of RBCs (TC₅₀) was determined using ED50plus v1 .0 software. The statistical significance of the data was analyzed using the unpaired and two-tailed Student's t-test (Prism, GraphPad software, San Diego, CA). When p < 0.05, the differences between data sets was considered to be statistically significant. The detailed calculations of NPs number density per gram of SS and MSNs are described below.

(i) Nonporous silica NPs, SS-24:

Density of amorphous silica = 2.2 (g/cm³)

a. The volume of each NP (V/NP) = 4/3 x π x (12 x 10^~7)³ = 7.238 x 10^~18 (cm³/NP)

b. The number of SS-24 per gram, Z₂₄ (No. NPs/g)

Assumption: 1 g of SS-24

Z₂₄ NPs

(7.238 x 10^~18 x Z₂₄) x 2.2 = 1

Z₂₄= 6.3 x 10¹⁶ NPs/g

The process above is also applied to calculate the number NPs per gram of SS-37, SS-142, and SS-263.

(ii) Mesoporous silica NPs, MS-25:

Density of amorphous silica = 2.2 (g/cm³)

Primary pore volume = 0.92 (cm³/g)

a. In this case, the shape of MS-25 is assumed to be spherical, because of the short-ranged ordered porous structure.

The total volume of each NP (V/NP) = 4/3 x π x (12.5 x 10^~7)³ = 8.181 x 10^~18 (cm³/NP)

b. The number of MS-25 per gram, Y₂₅ (No. NPs/g)

Assumption: 1 g of MS-25

Y₂₅ of NPs

{( 8.181 x 10^"18 x Y₂₅) - [ 1 x 0.92 ( cm³/g) ]} x 2.2 (g/cm³) = 1

Y₂₅ = 1 .7 x 10¹⁷ (No. NPs/g)

(iii) Mesoporous silica NPs, MS-42:

Density of amorphous silica = 2.2 (g/cm³)

Primary pore volume = 0.81 (cm³/g)

a. In this case, the shape of MS-42 is assumed to be hexagonal, because of the long-ranged ordered porous structure.

The total volume of each NP (V/NP) = x (21 x 10^"7)² x 6 x (42 x 10^~7) = 4.812 x 10^~17 (cm³/NP)

4

b. The number of MS-42 per gram, Y₄₂ (No. NPs/g)

Assumption: 1 g of MS-42

- Y₄₂ of NPs {(4.812x 10^~18 x Y₄₂) - [ 1 x 0.81 ( cm³/g) ]} x 2.2 (g/cm³) = 1

Y₂₅ = 2.6 x 10¹⁶ (No. NPs/g)

Results and Discussion

Preparation and Characterization of SS and MSNs. In this study, two types of amorphous silica particles, nonporous and porous NPs, were used to investigate the influence of nanoparticle size on hemolytic activity. The monodisperse nonporous silica NPs were prepared using the common base- catalyzed Stober preparation. Typical TEM images in Figure 2 show the SS NPs having four distinct diameters. The size of SS NPs increases as the amount of TEOS and ammonium hydroxide used in the synthetic procedure are increased. On the basis of TEM images, the average diameter of these SS NPs is 24 ± 2.9, 37 ± 4.5, 142 ± 12, and 263 ± 14 nm, respectively (Figure 3). The surface area of the SS NPs with varied sizes was determined by measuring N₂ adsorption-desorption isotherms and application of BET modeling. All the isotherms from SS NPs showed a typical type II isotherm that is characteristic of nonporous silica (one example shown in Figure 4) (Sing et al., 1985). The BET surface area data shows that the external surface area per gram of SS NPs decreases as the size increases, as listed in Table 4.

Table 4. The Surface Area of SS NPs and the Concentration of SS NPs Leading to a 50 % Lysis of RBCs (TC₅₀, expressed as μg and number of NPs per mL of solution)

SS-24 SS-37 SS-142 SS-263

TC₅₀ (Mg mL) 8.8 18 94 307

TC50 (No. NPs/mL) 5.5x10 3.1 X10^{1 1} 2.8x10¹⁰ 1 .5X10¹⁰

[a] 0Ί 00. (100) inter-planar spacing; [b] a, unit cell length of hexagonal packing structure; [c] ; S_BET- Specific surface area calculated from data in the range P/P₀ < 0.3 using BET equation ; [d] D_BJH: pore diameter assigned from the maximum on the BJH pore size distribution; [e] V_T: total pore volume calculated at P/P₀ = 0.99; [f] V_P: primary pore volume obtained from P/P₀ = 0.50; [g] V_TEXT: textural pore volume was calculated using V_T - V_P

The porous silica NPs were synthesized by using a positively charged CTAB-template and NH₄OH catalyst under dilute aqueous conditions. A photograph of the as-synthesized MS colloidal aqueous solutions shows that solution appearance changes from clear to turbid as the size of NPs increases (Figure 5). The Tyndall effect can be seen clearly when light passes through the transparent colloidal solution formed from 25 nm diameter mesoporous silica nanoparticles. The TEM images and size distribution histograms in Figure 6 show the surfactant-free MSNs with average sizes of 25 ± 3.7, 42 ± 6.5, 93 ± 13, 155 ± 19, and 225 ± 18 nm, respectively. The size of MSNs increases as the NP synthesis incorporates higher ammonia concentration, more silica precursor, and lower temperature conditions. The pore structure of all five MSP sizes was examined using low-angle powder XRD (Figure 7A). For the four largest diameter NPs, MS-42, MS-93, MS-155, and MS-225, four characteristic XRD peaks, (100), (1 10), (200), and (210) are present which indicate a two dimensional (2D) long-range order hexagonal structure. The XRD pattern of the smallest porous nanoparticle, MS-25, shows two broad peaks which suggest short-range ordering and a wormlike pore structure inside the NPs. This result can be confirmed visually in TEM images of MS-25 (Figure 6A). All of the N₂ adsorption-desorption isotherms in Figure 7B exhibit a steep adsorption behavior at P/P₀ around 0.35 without a hysteresis loop, known as a type IV isotherm according to lUPAC classification (Sing et al., 1985). To further compare the structural properties of the MSNs with varied sizes, the inter-planar spacing, unit cell, surface area, pore size, and pore volume data from XRD and N₂ sorption measurements are summarized in Table 5.

Table 5. The Structural Properties of Surfactant-Free

MSNs with Varied Sizes

MS-25 4.46 5.15 1 164 2.74 2.00 0.92 1 .08

MS-42 4.13 4.77 1038 2.59 1 .41 0.81 0.60

MS-93 4.05 4.67 1089 2.51 1 .28 0.86 0.42

MS-155 4.05 4.67 1 105 2.40 1 .09 0.88 0.21

MS-225 4.05 4.67 1 123 2.41 1 .04 0.89 0.15

MS-25 after

6-day PBS 4.46 5.15 738 - 2.23 0.47 1 .76

aging

MS-42 after

6-day PBS 3.98 4.60 791 - 1 .89 0.51 1 .38

aging

[a] 0Ί 00. (100) inter-planar spacing; [b] a, unit cell length of hexagonal packing structure; [c] ; S_BET- Specific surface area calculated from data in the range P/P₀ < 0.3 using BET equation ; [d] D_BJH: pore diameter assigned from the maximum on the BJH pore size distribution; [e] V_T: total pore volume calculated at P/P₀ = 0.99; [f] V_P: primary pore volume obtained from P/P₀ = 0.50; [g] V_TEXT: textural pore volume was calculated using V_T - V_P The d_Wo spacing, BJH pore size and primary pore volume of all MS samples with long-range ordered structure is around 4.05 nm, 2.50 nm, and 0.85 cm³ g^"1 , respectively. Compared to well-ordered MSNs, the MS-25 NPs have greater d spacing and pore size. In addition, the total pore volume of MSPs per gram increases as the NP diameter decreases. The second adsorption of the isotherms at high relative pressure (P/P₀ > 0.8) represents the formation of interstitial pores among the dried NP agglomerates called textural porosity. The textural pore volume of MSNs increases as the NP size decreases, because more interstitial pores are formed between the smaller NPs. All the MSNs have high total surface area ranging from 1038 to 1 164 m²/g. One important thing to keep in mind is that the obtained total surface area includes both internal and external surface area. Quantitative external surface areas of the MSNs with varied diameter are difficult to determine from the total surface area data.

Dose- and Size-Dependent Hemolytic Activity of SS and MSNs. The hemolysis assay was used to evaluate the cytotoxic effect of nonporous and porous silica NPs on human RBCs because silica materials have been known to cause membrane damage to RBCs (Gerashchenko et al., 2002; Murashov et al., 2006). To determine the concentration leading to 50% lysis of RBCs (TC₅₀) of each nanoparticle, the RBCs were exposed to each NP sample at a range of concentrations from 3.125 to 1600 μg/mL for 3 hours. The highest doses used in this work were chosen to model doses often used for in vivo biodistribution, imaging, and therapeutic experiments. As shown in Figure 8A, the hemolysis percentage of RBCs increases in a dose-dependent manner. The photographs of RBCs after exposure to four diameters of SS NPs for 3 hours are shown in Figure 8B. It is apparent that smaller SS NPs cause observable release of hemoglobin from damaged RBCs at lower nanoparticle exposure concentrations. This result demonstrates that the smaller particles have higher hemolytic activity than the larger particles.

The concentration and number of SS NPs leading to 50% of lysis of RBCs for all SS NPs are listed in

Table 4. Student's t-test analyses reveal significant differences among these TC₅₀ values for all samples

(p < 0.0001 ). These data reveal a near-linear correlation between TC₅₀ and NP diameter (Figure 9). The higher hemolytic activity of smaller SS NPs may be due to the larger surface area per gram, indicating higher number of silanol group present on the cell-contactable surface of the smaller SS NPs. The dose- and size-dependent cytotoxicity of nonporous amorphous silica NPs has been previously investigated using in vitro tetrazolium (meththylthiazolyldiphenyl-tetrazolium bromide, MTT) and lactate dehydrogenase (LDH) assays (Yu et al., 2009; Napierska et al., 2009). Napierska et al. (2009) and Yu et al. (2009) concluded that the cytotoxicity of the tested nanoparticles is strongly correlated to their size, because smaller particles have larger surface area per mass and show higher toxicity than larger particles. While it is true that toxicity mechanisms can be cell type-dependent, the simple hemolysis assay results are similar to those achieved with more complicated and expensive assays, and thus will be a good first line assessment for candidate injectable nanoparticles.

Having demonstrated that the hemolytic activity accurately predicted the toxicity of nonporous silica, a systematic study of the hemolytic activity of MSPs with varied sizes was conducted (Figure 10). As with the nonporous NPs, dose-dependent hemolysis behavior was also observed in all cases with MSNs (Figure 10A). The size effect of MSNs on hemolysis can be clearly seen when considering MS-42, MS-93, MS-155, and MS-225 NPs (Figure 10B). The TC₅₀ for all MSNs are listed in Table 6. Table 6. The Concentration of MSNs Leading to a 50% Lysis of RBCs

(TC₅₀, expressed as μg and number of NPs per mL of solution)

With the exception of the MS-25 NP, the TC₅₀ values decrease as the diameter of MSNs decreases (Figure 9). The outlier characteristic of the MS-25 NP indicates that the size-dependent hemolytic activity holds only for MSNs having a well-ordered mesoporous structure. Statistical comparisons showed significant differences between all samples (p < 0.05) except for the comparison between MS-25 and MS-155 (p=0.1756), suggesting that the MS-25 and MS-155 have indistinguishable hemolytic activity. The MS-25 NP likely has lower than expected hemolytic activity due to the larger than expected pore size and greater primary pore volume (as shown in Table 4) compared to the larger diameter MSNs, resulting in a smaller number of cell-contactable silanol groups on MS-25 NPs. In short, compared to SS NPs with similar size, MSNs show a reduction in hemolytic activity due to the voids on the surface of MSNs. In addition, RBCs can tolerate more MSNs than SS NPs with a similar size

(as shown in Tables 4 and 5). Based on the hemolytic activity difference between the well-ordered and moderately ordered mesoporous NPs (well-ordered and moderately ordered particles may be

distinguished based on the presence or absence of distinct X-ray diffraction peaks (100) (200) (1 10), where well-ordered particles have distinct peaks while the peaks for moderately ordered particles are blurred), it is clear that the pore structure of MSNs, and thus the cell-contactable surface area, also influences the hemolytic activity of MSNs.

Effect of Mesopore Stability of MSNs on Hemolysis. To further explore the role of mesopore structure on silica nanoparticle toxicity, pore stability was examined among the various MSNs. Pore stability is a concern because the surfactant template that originally supports the pore structure is removed before intravenous introduction, in part to allow introduction of a drug cargo and in part because the CTAB surfactant itself is highly cytotoxic. In the course of performing the aforementioned

experiments, experimental observation revealed that there was an increase in the hemolytic activity of MS-25 and MS-42 NPs when the NPs were aged in PBS for 6 days (Figures 1 1 A and B, and Figure 13) before RBC incubation. The significant differences among TC₅₀ values (Table 6) between MS-25 and MS-25 after 6-day PBS aging, MS-42 and MS-42 after 6-day PBS aging were confirmed by student's t- test (p < 0.0001 ). Three possible reasons were considered to explain this unintentional increase in apparent toxicity: residual surfactant, degradation product silicic acids, and pore collapse. The degradation of MS films and particles to monomeric and oligomeric silicic acids during aging in phosphate buffered solutions or simulated body fluids has been reported (Bass et al., 2007; He et al., 2009). Bass et al. (2007) revealed a dynamic change in the porous character of MS films under biological conditions and emphasized that this structural change would have significant implications on the use of these porous materials for drug delivery applications. He et al. (2009) found that compromised structures produced degraded silicic acid products but that these species showed no toxicity to human breast cancer or African green monkey kidney cells. While these two papers showed structural changes and examined the effects of any dissolved species, neither paper directly evaluates the impact of the altered mesoporous structure itself on cytotoxicity.

To assess the three aforementioned possibilities, MS-25 and MS-42 NPs were suspended in PBS at the particle concentration of 2000 μg/mL, and after 6 days, the aged NPs were separated from the solution by centrifugation. Then, the hemolytic activity of supernatant was examined. Compared to a control CTAB-hemolysis experiment, no hemolysis was observed from the supernatant of MS-25 after 6- day PBS aging and MS-42 after 6-day PBS aging (Figure 14). This implies that the amount of residual CTAB surfactant is less than 3.125 μg and not the source of increased hemolysis.

Because the hemolytic activity of MSNs is strongly correlated to their porous structure and their cell-contactable surface area, XRD and N₂ adsorption-desorption measurements were applied to examine the pore integrity of aged MS NPs. The XRD data showed an intensity decrease in the (100) peak and disappearance of peaks at higher angles in both aged MSN populations (Figures 1 1 C and D). In addition, comparisons of TEM images between MS-25 and MS-42 without PBS aging and after 6-day PBS aging clearly showed that parts of the mesopores inside NPs collapsed after 6-day PBS aging (Figure 15). The disappearance of capillary adsorption in N₂ adsorption-desorption isotherms (Figure 16) and the decrease in surface area and primary pore volume (as listed in Table 3) confirmed collapse of the mesopores inside

NPs occurring during the PBS incubation.

To further confirm the dissolution of MSNs upon incubation in PBS and the subsequent pore collapse within the MSNs, the quantity of silicic acid in solution after different PBS aging times was measured using the blue silicomolybdic assay (Coradin et al., 2009). The concentrations of degraded free silicon from MS-25 after 30-minute aging, MS-42 NPs after 30-minute aging, MS-25 after 6-day aging, and MS-42 after 6-day aging were 33 μg/mL, 37 μg/mL, 80 μg/mL, and 83 μg/mL respectively

(Figure 17). These results show that both MS-25 and MS-42 have similar dissolution rates during PBS incubation, and that silica dissolution to silicic acid is time fast and time-dependent. As previously discussed, the supernatant of the aged MS NPs was exposed to RBCs, and no detectable hemolysis was observed. These combined results demonstrate that the dissolved silicic acid species are not the source of increased hemolysis but that the liable feature must be a change in the nanoparticle surface itself. In fact, upon silica dissolution and pore collapse, the cell-contactable surface area is likely to increase, despite the total surface decrease, either by (1 ) elimination of void spaces or (2) lengthwise cracking of the pores to reveal the pore interior. None of the literature using mesoporous nanoparticles for drug delivery, imaging, and therapy has apparently reported this pore collapse and its resultant effect on hemolytic activity. The present work clearly demonstrates that the integrity of mesopores should be considered in toxicological studies of MSNs.

Influence of Surface Coating of PEG on Hemolytic Activity of MSNs. To counter the increase in hemolytic activity of MS NPs collapsed mesopores, a surface modification strategy was employed. Based on the improved biocompatibility of MSNs functionalized with a PEG coating reported in Lin et al. (2009), PEG-silane modified MSNs were synthesized to compare the hemolytic activity with unfunctionalized MSNs. As shown in Figure 18, the TEM images of surfactant-free MS-25@PEG and MS-42@PEG showed obvious porous structure within both NP populations. This result demonstrates that the PEG modification on the outer surface of MSNs does not interrupt the ordering of mesopores. The mesopore stability in these modified NPs was examined after suspending PEG-coated MSNs in PBS solutions for 6 days. The XRD patterns show evidence for some pore collapse after aging (Figure 19). The hemolytic activity of the PEG-coated MSNs and PEG-coated MSNs after aging is shown in Figure 20A. Contrary to bare MSNs, no hemolysis is apparent after a 3 hour blood incubation (Figures 20B and C), even at high nanoparticle doses (i.e., 1600 μg/mL) with collapsed mesopores. In addition, longer term biocompatibility of MSNs was tested in the presence of MS-25@PEG after 6-day aging and MS-42@PEG after 6-day aging at the nanoparticle concentration of 1600 μg/mL. Neither of these aged NPs causes significant lysis of RBCs, even after 24 hour incubation (Figure 21 ), revealing that PEG not only masks the surface silanol groups but also serves as a protecting layer, preventing the access of additional silanol groups from collapsed pores to RBCs. This simple surface modification strategy may ensure the safety of MSNs in biomedical applications.

Conclusions

In summary, systematic comparisons of cytotoxicity of nonporous and porous silica NPs with varied sizes has been demonstrated using a simple hemolysis assay. All the nonporous and porous silica NPs show dose- and size-dependent hemolytic activity on RBCs, except for the case of the smallest MSNs. Generally, smaller particles exhibited higher toxicity than larger particles as the TC₅₀ was expressed as mass concentration. MSNs showed a reduction of hemolytic activity compared to similar sized nonporous counterparts.

In addition to this size effect, the porous ordering structure and stability also influence hemolytic activity of MSNs. The collapse of pores inside the porous silica NPs after aging in PBS was confirmed by XRD, N₂ sorption, and TEM. The pore collapse leads to greatly increased hemolytic activity of mesoporous silica nanoparticles. However, it is possible to ameliorate the enhanced hemolytic activity of

MSNs via surface modification with a PEG-silane.

Example 3

Materials and Methods

Materials. All chemicals were used as received. n-Cetyltrimethylammonium bromide (CTAB), tetraethyl orthosilicate (TEOS), dimethyl sulfoxide (DMSO), ammonium molybdate tetrahydrate, 4- methylaminophenol sulfate, sodium sulfite, oxalic acid, silicon standard (1000 mg/L) and 10 X phosphate buffered saline (PBS) were purchased from Sigma Aldrich. 2-[Methoxy(polyethyleneoxy)propyl] trimethoxysilane, (PEG-silane, MW 596-725 g/mol, 9-12 EO) was obtained from Gelest. Ammonium nitrate (NH₄N0₃), hydrofluoric acid (HF), nitric acid (HN0₃) and ammonium hydroxide (NH₄OH, 28-30 wt% as NH₃) were obtained from Mallinckrodt. Hydrochloric acid (HCI) and acetic acid were obtained from BDH. 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide (MTT) was purchased from Invitrogen. Absolute anhydrous 99.5% ethanol and 95% ethanol were purchased from Pharmco-Aaper. The de- ionized (D.I.) water was generated using a Millipore Milli-Q system. Heat inactivated fetal bovine serum (FBS) and Dulbecco's Modified Eagle's Medium (DMEM) with 4.5 g/L glucose, 1 10 μg/mL sodium pyruvate, 4.00 mM L-glutamine, and phenol red were purchased from Hyclone. Trypsin-EDTA and penicillin streptomycin (PS) were obtained from Gibco. Powder DMEM without phenol red was purchased from SAFC Biosciences.

Characterization. Transmission electron microscopy (TEM) micrographs were taken on a JEOL

1200 EXII with a 100 kV voltage. Powder X-ray diffraction (XRD) patterns were measured on a Siemens Bruker-AXS D-5005 X-ray diffractometer using filtered Cu Ka radiation (λ = 1 .5406 A) at 45 kV and 20 mA. Data were recorded by step scan with a step size of 0.040 ° and a step time of 1 .0 second. N₂ adsorption-desorption isotherms were taken on a Quantachrom Autosorb-1 analyzer at 77 K. Samples were degassed at 120°C for at least 12 hours prior to measurements. The total surface area was calculated using the Brunauer-Emmett-Teller (BET) equation at P/P₀ < 0.3. The pore size distribution was calculated from the branch of the adsorption isotherm using a Barrett-Joyner-Halenda (BJH) method. UV- vis measurements were performed on a Perkin Elmer Lambda 12 spectrometer. Hydrodynamic diameter data were measured at particle concentration of 1 mg/mL using dynamic light scattering (DLS) with a Brookhaven 90Plus/BIMAS particle analyzer equipped a 655 nm laser. Three runs and one minute run duration were set for each measurement. The DLS size distribution was plotted using a lognormal analysis method. Cell viability and hemolysis percentage were measured at 570 nm using a Bio-Rad iMark microplate reader.

Preparation of mesoporous silica (MS) nanoparticles with 42 nm diameter (MS42)

(a) Bare MS42 nanoparticles. The bare MS nanoparticles were prepared as described in

Example 2. First, 0.29 g of CTAB was dissolved in 150 mL of 0.256 M NH₄OH solution at 50°C. After one hour, 2.5 mL of 0.88 M ethanolic TEOS was added under vigorous stirring. After one hour, the stirring was stopped and the colloidal solution was aged for 20 hours at 50°C. After aging, the as-synthesized colloidal solution was passed through a 0.45 μιη GH propylene (GHP) filter and diluted to 40 ml_ with D.I. water. Two methods were used to remove surfactant. One was a centrifugation method, and the other was a dialysis method. With centrifugation, the filtered as-synthesized MS42 colloids were transferred to 50 ml_ of 6 g/L ethanolic ammonium nitrate by centrifugation (30,000 rpm = 66226 g, 30 minutes) and heated to 60°C for one hour with stirring. The MS42 nanoparticles were then transferred to 50 ml_ of acidic ethanol solution (1 ml_ of HCI/1 L of ethanol) via centrifugation and heated to 60°C for two hours under stirring. The extracted MS-42 nanoparticles were further washed with 95% ethanol and then 99.5% ethanol once. Finally, the surfactant-free MS42 nanoparticles were suspended in 99.5% ethanol and stored at 4°C. The MS42 nanoparticle extracted by the centrifugation method is designated as MS42-C. For dialysis purification, the surfactant was removed from the as-synthesized MS42 nanoparticles using a dialysis process described by Urata et al. (2009). The as-synthesized sample was transferred to regenerated cellulose dialysis tubing (with a molecular weight cut off, MWCO, of 12,000 - 14,000, Fisherbrand) and placed into a 250 ml_ acid solution composed of 95% ethanol and 2 M acetic acid. The acid solution was replaced every 24 hours and repeated two times. The particles were then dialyzed against 500 ml_ of D.I. water three more times. Finally, the dialyzed MS42 nanoparticles were filtered through a 0.45 μιη GHP filter and stored at 4 °C until use. The MS42 nanoparticles purified by the dialysis method are designated as MS42-d.

(b) Pegylated MS42 nanoparticles with hydrothermal treatment. Typically, 0.29 g of CTAB was added to 150 mL of 0.256 M NH₄OH solution at 50°C. Then, 2.5 ml_ of 0.88M ethanolic TEOS was added to solution under continuously stirring. After one hour, 450 μΙ_ of PEG-silane was added to the as- synthesized colloidal solution. The mixture solution was stirred for 30 minutes and then aged at 50°C for 20 hours. The as-synthesized pegylated MS42 colloidal solution was filtered with a 0.45 μιη GHP filter and diluted to 50 mL with D.I. water. The filtered colloidal solution was then heated at 90°C for 24 hours in a sealed vessel. The surfactant removal steps followed the centrifugation method described above. Finally, the surfactant-free pegylated MS42 nanoparticles were filtered using a 0.2 μιη

polytetrafluoroethylene (PTFE) filter. The products were stored at 4 °C until use. The pegylated MS42 nanoparticles without and with hydrothermal treatment were purified via centrifugation and are designated as MS42@PEG-c and MS42@PEG-hy-c, respectively.

Quantification of degraded free silicon from MS42-d and MS42(5>PEG-hv-c after 10 day aging in

D.I, water and PBS at R.T. and 37°C using a blue silicomolybdic assay. Surfactant-free MS42-d and MS42@PEG-hy-c nanoparticles were suspended in D.I. water and PBS at 1000 μg/mL concentration. These nanoparticle solutions were aged in PBS at R.T. and 37°C for 10 days; then, the aged NPs were separated from the solution by centrifugation (66226 g, 30 minutes). The degraded silicon concentration in the supernatant was determined using a blue silicomolybdic assay (SMA) (Lin et al.). The silicon quantification was based on a calibration curve (0 ppm, 1 ppm, 5 ppm, 10 ppm, 20 ppm, 40 ppm, 50 ppm, and 60 ppm) made before sample measurements. The silicon quantification was performed in three independent experiments.

In-vitro cell viability (MTT assay) and hemolysis assay

(a) MTT assay. Human endothelial cells were purchased from American Type Culture Center

(ATCC). Typically, 6x10⁴ cells were seeded in 96-well plates and cultured in DMEM supplemented with 10% FBS and 1 % PBS at 37°C under 5% C0₂. After 24 hours, the cells were incubated with 100 μί of different concentrations of MS42@PEG-hy-c nanoparticle suspensions in DMEM + 10% FBS media for 24 hours. After nanoparticle incubation, the cells were washed with 100 μΙ_ of serum-free DMEM two times and incubated with 100 μΙ_ of 0.5 mg/mL MTT media for 2 hours at 37°C under 5% C0₂. Finally, the MTT media was removed and purple formazan crystals produced by live cells were dissolved in 200 μΙ_ of DMSO. Optical density of the produced stain was monitored at 570 nm with 655 nm as a reference using a microplate reader.

(b) Hemolysis assay. Fresh ethylenediamine tetraacetic acid (EDTA)-stabilized human whole blood samples were purchased from Memorial Blood Center. Typically, 5 ml_ of whole blood was added to 10 mL of PBS and centrifuged at 10016 g for 10 minutes to isolate red blood cells (RBCs) from serum. The RBCs were then washed five times with 10 mL of PBS and diluted to 50 mL with PBS. To test the hemolytic activity of MS42@PEG-hy-c NPs, 0.2 mL of diluted RBC suspension was added to 0.8 mL of nanoparticle solution at different concentrations ranging from 15.625 to 1000 μg/mL. D.I. water and PBS were used as positive control and negative control, respectively. All the samples were placed on a rocking shaker at 37°C for 3 hours. After incubation, the samples were centrifuged at 10016 g for 3 minutes. The hemoglobin absorbance in the supernatant was measured at 570 nm with 655 nm as a reference.

In-vitro macrophage uptake. Mouse macrophage, Raw 264.7, cells obtained from ATCC were cultured in DMEM with 10% FBS and 1 % PS under 5% C0₂ atmosphere at 37°C. Before macrophage uptake experiments, cells were trypsinized and seeded into 24-well plates at 3x10⁵ cells per well. After 24 hour incubation, the cells were exposed to 1 mL of MS42-d and MS42@PEG-hy-c at a concentration of 200 μg/mL in DMEM + 10% FBS media for 24 hours. Cells incubated without mesoporous silica nanoparticles were used as control. The cells were then washed with PBS two times and lysed with 0.5 mL of acid solution containing 2% HN0₃ and 0.1 % HF for 20 hours at 37°C. After digestion, the solutions were centrifugation at 10016 g for 10 minutes. The supernatant was separated for silicon quantification. The silicon uptake amounts by the macrophage cells were determined by the blue silicomolybdic assay described previously.

Results

Small MSNs (< 50 nm) were prepared with well-ordered pore structure as described in Example 2. The template of the as-synthesized bare MSNs with 42 nm diameter (MS42) was removed by two different methods, centrifugation (Lin et al. , 2009) or dialysis (Kobler et al. , 2008), and accordingly, are designated as MS42-C and MS42-d, respectively (Figure 22). In the preparation of pegylated MSNs, a short polyethylene glycol silane (PEG-silane) was added to functionalize the exterior of the MS42 NPs. In some cases, a hydrothermal treatment was introduced (to promote NP stability) before pegylated MSN template removal was accomplished by centrifugation ; this product is designated as MS42@PEG-hy-c (Figure 22). Pegylated MSNs without hydrothermal treatment, for comparison, are named as MS42@PEG-c.

Transmission electron microscopy (TEM) images of MS42-C, MS-42-d, MS42@PEG-c, and MS42@PEG-hy-c showed that all particles had similiar diameters of approximately 40 nm (Figures 23A and B and Figure 24). The TEM images clearly reveal that the pore structure inside

hydrothermally treated NP was not as ordered as the other three MSNs. Compared to MS42-d NPs, MS42@PEG-hy-c NPs also had a lower number of X-ray diffraction (XRD) peaks (Figure 23C) and a shift in the N₂ adsorption-desorption isotherm capillary condensation to lower pressure (Figure 23D). Additionally, the total surface area of MS42@PEG-hy-c decreased from 1 131 (MS42-d) to 731 m²/g.

This data support that pore order was altered by PEG-silane surface modification and/or

incorporation into MSN framework during the hydrothermal treatment.

The hydrodynamic diameter of the as-synthesized and extracted MSNs in D.I. water at room temperature (R.T.) is summarized in Table 8.

Table 8. Hydrodynamic diameter of as-synthesized MS42, MS42@PEG, MS42-d, and MS42@PEG-hy-c measured in D.I. water by DLS at R.T.

aAII the data were obtained from three independent experiments. Each measurement was taken three times with one-minute run duration.

Except for the bare MSNs purified using centrifugation, MS42-C, the hydrodynamic size of surfactant-free bare and pegylated MSNs was smiliar to their as-synthesized counterparts (Figure 25A and Table 8), showing that aggregation occurs during the high-speed centrifugation and redispersion of NPs via ultrasonication. Compared to MS42-d and MS42@PEG-hy-c NPs, the MS42-C NPs aggregated easily and precipitated after a 30 minute aging in PBS (see Figure 26). Figure 25B shows a slight increase in hydrodynamic diameter for MS42-d NPs when they were dispersed in PBS and a significant size change upon dispersion into Dulbecco's modified Eagle's medium (DMEM) with 10% fetal bovine serum (FBS) (Figure 25C). The size change of the MS42-d NPs in both PBS and DMEM + 10% FBS is likely due to charge neutralization of surface silanol groups by ionic species and rapid adsorption of proteins on the nanoparticle surface.

Additionally, further examination of the long-term stability of MS42-d and MS42@PEG-hy-c in D.I. water, PBS, and DMEM + 10% FBS at R.T. and 37°C was performed (Figures 25D and E). Both nanoparticles were very stable in D.I. water, but MS42-d aggregated in highly salted media, and the temperature effect is profound for the bare MSNs, with higher incubation temperature causing more extensive aggregation. For MS42@PEG-hy-c NPs, no significant size change was observed after even 10 days of aging in biological media at either R.T. or 37°C, except for the PBS aging at 37°C. Even though the hydrodynamic size of MS42@PEG-hy-c NPs gradually increased after 6 day PBS aging at 37°C, the aggregates are still small (less than 100 nm). No visible particle precipitation was observed for these NPs after a 10 day incubation in D.I. water, PBS, or DMEM + 10% FBS at 37°C (see inset in Figure 25E). The stability comparison between MS42@PEG-c and MS42@PEG-hy-c NPs in PBS further showed that the the hydrothermal treatment significantly improved the long-term stability of MS42@PEG- hy-c NPs (see Figure 27). This is likely due to greater number of PEG groups either on the outer or inner surface of the nanoparticles as well as more complete silica condensation between silanol groups following hydrothermal treatment.

Figure 28A shows the amount of free silicic acid, a clear indicator of NP degradation, from MS42-d and MS42@PEG-hy-c after 10 day D.I . water and PBS incubation at R.T. and 37°C, as quantified by a blue silicomolybdic assay (Coradin et al., 2004). The degraded amounts from MS42-d were greater than MS42@PEG-hy-c after both 10 days in D. I. water or PBS at R.T. and 37°C.

Additionally, the aging temperature influenced the amount of dissolved silica. Higher temperature led to more silica dissolution in both D. I. water and PBS aging conditions. The XRD patterns of MS42-d and MS42@PEG-hy-c were also examined after 10 day aging in D.I . water and PBS at R.T. and 37°C. In the case of D. I. water aging, the XRD patterns of aged MS42-d and MS42@PEG-hy-c NPs were almost the same as the unaged particles (see Figure 29). In contrast, the pore structure of MS42-d and MS42@PEG-hy-c collapsed after 10 day PBS aging (Figures 28B and C). In both particles, the extent of pore collapse was more extensive at 37°C than at R.T. While the (100) peak of MS42-d NPs after a 10 day PBS aging at 37°C is almost undetectable, it is still apparent in the aged MS42@PEG-hy-c NPs, indicating enhanced pore retention achieved by hydrothermal treatment. The comparison of dissolved silica amounts and the pore structure after 10 day PBS aging of MS42-d and MS42@PEG-hy-c NPs further demonstrates that the PEG modification and hydrothermal treatment of the MSNs reduces the extent of MSN degradation within biological media, yielding a stable nanoparticle with significant promise for in vivo applications.

To evaluate the biocompatibility of these stable MS42@PEG-hy-c NPs, 3-(4,5- dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide (MTT) and hemolysis assays were used. The results in Figures 31 A and B show that human endothelial cells and red blood cells are not compromised even after 24 hour exposure at a relatively high concentration (1000 μg/mL).

Additionally, the stealth property of MS42@PEG-hy-c NPs was studied by monitoring macrophage uptake. Figure 30C showed that MS42@PEG-hy-c NPs exhibited 70% reduction in the amount of macrophage uptake compared to bare MSNs, MS42-d. The uptake percentage of MS42@PEG-hy-c NPs (200 μg/mL) by macrophages after 24 hour exposure was only 0.5%, further confirming that pegylated MSNs with hydrothermal treatment yields greatly reduced protein adsorption, resulting in resistance to nonspecific uptake by macrophage cells.

In summary, the dispersity and the stability of small mesoporous silica nanoparticles in biological media was improved using PEG modification and hydrothermal treatment. Those porous silica nanoparticles with a hydrodynamic diameter less than 100 nm exhibited long-term stability in biological media at 37°C. Additionally, compared to bare mesoporous silica nanoparticles, the hydrothermally treated pegylated mesoporous silica nanoparticles were highly biocompatiable, resistant to protein adsorption, and curb macrophage uptake. This preparation can be easily extended to fabricate multifunctional mesoporous silica nanoparticles having long-term stability, enabling use as stealth theranostic nanoparticles.

Example 4

Brief Summary

An exemplary synthesis to provide a MSN having different coatings on the exterior surface and in pores is described below. A solution of 150 ml_ of 0.256 M ammonium hydroxide and 0.29 g of cetyltrimethylammonium bromide was stirred at 50 °C. After 1 hour, 2.5 ml_ of 0.88 M

tetraethylorthosilicate in 99.5% ethanol (EtOH) was added under vigorous stirring. One hour later, 450 μΙ_ of PEG-silane was added, and then 30 minutes later, 65 μΙ_ of chlorotrimethylsilane or

trifluoropropyltrimethylsilane was added. The hydrophobic silane increases MSN stability, drug loading capability and regulates the drug release profile. Stirring was stopped after another 30 minutes and the solution was aged at 50°C for 20 hours. The solution was filtered with a 0.45 μιη GHP filter, then diluted to 50 ml_ with deionized water (D.I. H₂0) and kept in a sealed container at 90 °C for 24-hour hydrothermal treatment.

Purification was accomplished via the centrifugation method. The solution was centrifuged down then suspended in ammonium nitrate solution to 50 ml_, refluxed for 1 hour at 60°C and washed again with 95% EtOH then HCI in 95% EtOH and diluted to 50 ml_. The solution was again refluxed for 2 more hours at 60 and washed with 95% EtOH then twice more with 99.5% EtOH. The final solution in 99.5% EtOH was diluted to 20 ml_ and filtered with a 0.20 μιη PTFE filter and stored at 4°C.

Experimental Section

Chemicals and Reagents. All chemicals were used without further purification. Tetraethyl orthosilicate (TEOS), n-Cetyltrimethylammonium bromide (CTAB), dimethyl sulfoxide (DMSO), trimethylchlorosilane (TMS), fluorescein isothiocyanate (FITC), and polyvinyl pyrrolidone (PVP, average MW 10,000) were purchased from Sigma-Aldrich (Milwaukee, Wl). 2-[Methoxy(polyethyleneoxy)propyl]- trimethoxysilane, (PEG-silane, MW 596-725 g/mol, 9-12 EO) and 3,3,3-trifluoropropyldimethylchlorosilane (TFS) were obtained from Gelest (Morrisville, PA). Absolute anhydrous 99.5% ethanol and 95% ethanol were purchased from Pharmco-Aaper (Brookfield, CT). Ultrapure de-ionized (D.I.) water was generated using a Millipore Milli-Q system (Billerica, MA). Ammonium hydroxide (NH40H, 28-30 wt% as NH3) and sodium hydroxide (NaOH) were obtained from Mallinckrodt (Phillipsburg, NJ). Acetic acid was obtained from BDH (West Chester, PA). 10x Calcium- and magnesium-free Dulbecco's phosphate buffered saline (PBS), heat-inactivated fetal bovine serum (FBS), trypsin-ethylenediamine tetraacetic acid (EDTA), penicillin streptomycin (PS), and 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide (MTT) were obtained from Gibco/lnvitrogen (Grand Island, NY). High glucose Dulbecco's Modified Eagle's Medium (DMEM) and Minimum Essential Medium Eagle (MEM) were purchased from Hyclone (Logan, UT). Simulated body fluid (SBF) was prepared according to a reported protocol developed by Kokubo et al. (1990). Doxorubicin hydrochloride salt was purchased from LC Laboratories (Woburn, MA).

Mesoporous Silica (MS) Nanoparticle (NP) Fabrication. Unmodified MSNs (MS42-d) and PEGylated MSNs without hydrothermal treatment (MS42@PEG-c) and with hydrothermal treatment (MS42@PEG-hy-c): The unmodified and PEGylated MSNs having a 42 nm diameter were prepared and purified as described in Example 3.

Hydrothermally treated highly organosilane-modified MSNs (MS42@PEG/TMS-hy-c,

MS42@PEG/TFS-hy-c, and MS25@PEG/TMS-hy-c): First, 0.29 g of CTAB were added to 150 mL of 0.256 M NH40H solution, sealed, and continuously stirred for one hour at 50°C. Then, 2.5 mL of 0.88 M ethanolic TEOS solution were added to the solution under continuous stirring. After one hour, 450 μί of PEG-silane were added to the as-synthesized colloidal solution. The mixture solution was stirred for 30 minutes and then 68 μί of TMS (for preparation of MS42@PEG/TMS-hy-c) or 86 μί of TFS (for preparation of MS42@PEG/TFS-hy-c) were added. After another 30 minutes, stirring was stopped and the obtained colloidal solution was aged at 50°C for 20 hours. The as-synthesized modified MSN solution was filtered with a 0.45 μιη GHP filter and diluted to 50 mL with D.I. water. The filtered MSN solution was then heated at 90°C for 24 hours in a sealed vessel. The surfactant removal steps followed the centrifugation method described Example 3. The filtered as-synthesized organo-modified MS colloids were transferred to 50 mL of 6 g/L ethanolic ammonium nitrate by centrifugation (66226g, 30 minutes) and heated to 60°C for one hour with stirring. The NPs were washed once using 95% EtOH and then transferred to 50 mL of acidic ethanol solution (1 mL of concentrated HCI/1 L of ethanol) via centrifugation and heated to 60°C for two hours under stirring. The extracted NPs were further washed with 95% ethanol and then 99.5% ethanol once. Finally, the surfactant-free MS42@PEG/TMS-hy-c or MS42@PEG/TFS-hy- c NPs were suspended in 99.5% ethanol and filtered using a 0.2 μιη polytetrafluoroethylene (PTFE) filter. The final products were stored in one of two ways, either in 99.5% EtOH or as a dry powder at either room temperature (RT) or 4°C until use. The dry powder of products was obtained by evaporation from an ethanolic NP solution under vacuum. For MS25@PEG/TMS-hy-c, the synthesis procedure is similar to MS42@PEG/TMS-hy-c; the differences were using 0.128 M NH40H solution instead of 0.256 M NH40H solution and changing PEG-silane and TMS amounts to 360 and 52 μί, respectively.

Redispersible fluorescent (FITC-MS42@PEG/TMS-hy-c) and magnetic MSNs.

(Fe₃0₄@MS@PEG/TMS-hy-c): For synthesis of FITC-MS42@PEG/TMS-hy-c, 1 .9 mg of FITC were first dissolved in 1 mL of 99.5% ethanol and 2 μί of APTES were then added to the FITC ethanolic solution to prepare the ethanolic FITC-APTES solution. The solution was stirred under dark conditions for 18 hours at room temperature. Next, 0.29 g of CTAB were added to 150 mL of 0.256 M NH40H solution at 50°C. Then, 1 mL of ethanolic FITC-APTES solution and 2.5 mL of 0.88 M ethanolic TEOS solution were added simultaneously to the solution under continuous stirring. After one hour, 450 μί of PEG-silane were added to the as-synthesized colloidal solution. The mixture solution was stirred for 30 minutes and then 68 μί of TMS were added. After another 30 minutes, stirring was stopped and the obtained solution was aged at 50°C for 20 hours. The subsequent hydrothermal treatment and purification steps followed the procedure described in Example 3. For synthesis of Fe₃0₄@MS@PEG/TMS-hy-c, the oleic acid coated Fe₃0₄ was prepared based on a chemical coprecipitation method as described in Lin et al. (2009). First, a 5 mL of aqueous solution containing 0.29 g of CTAB and 0.2 g PVP was prepared. Then, 0.63 mL of 52 mg/mL hydrophobic Fe₃0₄ NPs (in chloroform) were added to the CTAB solution and ultrasonicated for one hour to evaporate the chloroform and allow aqueous suspension of the Fe₃0₄ NPs. The resulting Fe₃0₄ NP aqueous suspension was added to 150 mL of 0.256 M NH40H solution and heated to 50°C for one hour. The subsequent synthetic conditions were based on the procedure described in the synthesis of

MS42@PEG/TMS-hy-c. Then, 3 mL of 0.88 M ethanolic TEOS solution were added. After one hour, 540 μί of PEG-silane were added to the mixture solution, and 30-minute later, 78 μί of TMS were added. After another 30-minute stirring period, the obtained solution was aged at 50°C for 20 hours. The subsequent hydrothermal treatment and purification steps followed the procedure described previously.

Redispersitv and Long-Term Particle Stability of MSNs in Various Media. In this work, there were two ways to disperse purified NPs in D.I. water. For purified MSNs suspended in 99.5% EtOH, the NPs were centrifuged, transferred to D.I. water, washed with D.I. water one time and suspended at 2 mg/mL. For dry powder, 1 mL of D.I. water was added to 30 mg or 10 mg of organo-modified or drug-loaded MSN powder, respectively. The solution was sonicated for five minutes to homogenously redisperse the NPs. All the MSN suspension solutions were diluted to 1 mg/mL in various media (PBS, DMEM+10%FBS, or SBF) by adding 2 ml_ of 2 mg/mL of MSN stock solution in D.I. water to 2 mL of 2x PBS, DMEM+10%

FBS or SBF solutions. For long-term particle stability studies, the MSN solutions were then aged for 15 days at 37°C.

Quantification of Degraded Free Silicon from MS42-d and MS42(5>PEG/TMS-hv-c. Surfactant-free MS42-d and MS42@PEG/TMS-hy-c NPs were suspended in SBF at 1 mg/mL concentration. The MSN solutions were aged in SBF at 37°C. Then, 1 mL of the MSN solution was taken from the solution at different time points. The aged MSNs were separated by passing the aged solution through a Millipore Amicon Ultra centrifugal filter (MWCO 10,000) at 5250g for five minutes. The free degraded silicon concentration in the filtered solution was determined using a blue silicomolybdic assay (SMA) on a Perkin Elmer Lambda 12 UV-vis spectrometer (Waltham, MA) at 810 nm. The details of the SMA have been described in Coradin et al. (2004). The silicon quantification was based on a calibration curve (0 ppm, 1 ppm, 5 ppm, 10 ppm, 20 ppm, 40 ppm, and 50 ppm Si in SBF) made before sample measurements. The silicon quantification was performed in three independent experiments.

Biocompatibilitv of MS42-d and MS42(5>PEG/TMS-hv-c. MTT viability assay: Human endothelial cells (CRL-2922) and human skin fibroblast cells (CRL-2522) were purchased from American Type

Culture Center (ATCC). Typically, 6x10⁴ cells were seeded in 96-well plates and cultured in DMEM (for endothelial cells) or MEM (for fibroblasts) supplemented with 10% FBS and 1 % PS at 37°C under 5% C02. After 24 hours, the cells were incubated with 100 μΙ_ of different concentrations of MS42-d and MS42@PEG-hy-c NP suspensions in serum-free media for 24 hours. After NP incubation, the cells were washed with 100 μΙ_ of serum-free DMEM two times and incubated with 100 μΙ_ of 0.5 mg/mL MTT media for 2 hours at 37°C under 5% C02. Finally, the medium was removed and water-insoluble purple formazan crystals were dissolved in 200 μΙ_ of DMSO. The plate was placed on a rocking shaker for at least 20 minutes and then 100 μΙ_ of the DMSO solution in each well were transferred to a new 96-well plate. Optical density of the produced stain was monitored at 570 nm, with 655 nm as a reference, using a Bio-Rad microplate reader (Hercules, CA). The cell viability was calculated using equation 1 . Cells without NP exposure were used as control.

sample abs_570nm__655i

Viability from MTT assay {%) + Z 100 (1 )

control abs_570nm__655nm

Reactive oxygen species (ROS) measurements. The human endothelial cells were seeded in a 96-well plate (6x10⁴ per well) and cultured in DMEM supplemented with 10% FBS and 1 % PS at 37°C under 5% C0₂. After 24 hours, the cells were washed with serum-free DMEM (1 % PS, no phenol red) twice and incubated with 100 μΙ_ of 20 μΜ of H₂DCFDA (Invitrogen, Eugene, OR) in serum-free DMEM (1 % PS, no phenol red) for 1 hour. The cells were washed with serum-free DMEM (1 % PS) one time and then incubated with 100 μΙ_ of different concentrations of MS42@PEG/TMS-hy-c nanoparticles (0, 50, 100, 200, 400, 600, 800, and 1000 μg/mL) in serum-free MEM (1 % PS, no phenol red) for 24 hours. Then, the treated cells were washed with serum-free MEM (1 %PS, no phenol red) two times. After adding 100 μΙ_ of PBS to each well, the fluorescence intensity of the treated cells were measured by a fluorescence microplate reader (BioTek, Winooski, VT) with excitation/emission at 485/528 nm. The ROS level is expressed as ratio of fluorescence intensity of the sample well (F_sampi_e, cells exposed to

MS42@PEG/TMS-hy-c) to control well (F_contr0|, cells without nanoparticle exposure). Hemolysis assay. Fresh EDTA-stabilized human whole blood samples were obtained from Memorial Blood Center (St. Paul, MN). The washed RBCs were prepared following the procedure in Liao et al. (201 1 ). To examine the hemolytic activity of MS42-d and MS42@PEG/TMS-hy-c NPs, 0.2 mL of diluted RBC suspension (around 4.5 x 10⁸ cells/mL) were added to 0.8 mL of 250 μg/mL of MSN suspension solutions in PBS. The final concentration of MSNs was 200 μg/mL. D.I. water (+RBCs) and PBS (+RBCs) were used as the positive control and negative control, respectively. All the samples were placed on a rocking shaker in an incubator at 37°C for 0.5, 1 .5, or 3.0 hours. After incubation, the samples were centrifuged at 10,016 g for three minutes. The hemoglobin absorbance in the supernatant was measured at 540 nm, with 655 nm as a reference, using a Bio-Rad iMark microplate reader (Hercules, CA). Percentage of hemolysis was determined using equation 2.

Lactate dehydrogenase (LDH) assay. Membrane integrity of human platelets after MS42-d and MS42@PEG/TMS-hy-c exposure was examined using the LDH assay. The LDH activity was measured using a Bio Vision LDH cytotoxicity assay kit (Milpitas, CA). Washed human platelets were obtained based on a protocol described in He et al. (201 1 ). Typically, 0.1 mL of the washed platelets (around 2 χ 10⁸) were added to 0.4 mL of 250 μg/mL of MS NP suspension solutions in PBS. The final NP concentration was 200 μg/mL. All the samples were placed on a rocking shaker in an incubator at 37°C for 0.5, 1 .5, or 3.0 hours. After NP exposure, the solution was centrifuged at 1900g for 8 minutes. Then, 10 μί of supernatant were transferred to a 96-well plate and 100 μί of tetrazolium (WST-8) substrate mix were added. After 30 minutes, absorbance of the mixture solutions was measured at 450 nm with 655 nm as a reference, using a BioRad iMark microplate reader (Hercules, CA). The percent LDH release was calculated using equation 3. Cells without particle exposure and lysed with surfactant were used as the positive control.

Percent LDH release {%) = { ^M^^{fe abs}^-^ ^{~ ne}Z^{ative contml abs}^-^ X ₁₀₀ (³)

^ positive control abs_Uijrlm__6iirlm— negative control abs_i40nm__6iinm J

Drug Loading and Delivery of MS42(5>PEG/TMS-hv-c. Two types of doxorubicin, water-soluble doxorubicin hydrochloride (DoxHCI), and poorly water-soluble doxorubicin (Dox), were loaded into M42@PEG/TMS-hy-c in this study. The poorly water-soluble Dox was prepared by adding equal volume of 0.5 mg/mL of DoxHCI to 0.001 M of NaOH. After 10 hours, the solution was centrifuged at 10,016 g for five minutes. The obtained precipitated Dox was dried under vacuum. For drug loading conditions, about 10 mg of MS42@PEG/TMS-hy-c were added to 1 mL of 0.5 mg/mL of DoxHCI (in D.I water) or 2 mL of 0.25 mg/mL of Dox (in DMSO/H₂0=1 :1 solution) solution and stirred for 24 hours at RT. The drug-loaded NPs were collected by centrifugation (66226 g, 30 minutes) and washed with D.I. water one time. The obtained drug-loaded MSNs were dried under vacuum. To determine the loaded amount of Dox in MSNs, the drug-loaded NP powder (1 .0-2.0 mg) was resuspended and ultrasonicated in 1 mL of DMSO and then sat overnight. The optical density of Dox (in DMSO) at 480 nm was measured using the microplate reader. The loaded Dox amount was calculated based on a calibration curve. Drug delivery from 1 mg/mL of DoxHCI-MS42@PEG/TMS-hy-c and Dox-MS42@PEG/TMS-hy-c was assessed in SBF at 37 using UV-vis absorbance of Dox at 480 nm. Dried drug-loaded MSNs were redispersed in SBF and the high salt concentration induced drug delivery. At each measurement time, a 1 mL of aliquot of the solution was collected and centrifuged down. The obtained NPs were redispersed in SBF and added back to stock solutions of DoxHCI-MS42@PEG/TMS-hy-c and Dox-MS42@PEG/TMS-hy-c. The optical density of the supernatant was measured at 480 nm (655 nm as reference) using the microplate reader. The released free Dox was determined compared to a calibration curve.

In Vitro Therapeutic efficacy of DoxHCI-MS42(5>PEG/TMS-hv-c and Dox-MS42(5>PEG/TMS-hv-c. HeLa cancer cells (6 x 10³) were seeded in 96-well plates and cultured in DMEM supplemented with 10%

FBS and 1 % PS at 37°C under 5% C0₂ for 24 hours. The cells were incubated with 100 μΙ_ of the indicated concentrations of free DoxHCI, DoxHCI-MS42@PEG/TMS-hy-c, or Dox-MS42@PEG/TMShy-c.

After different incubation periods (24, 48, or 72 hours), the medium was removed and the cells were washed twice with serum-free DMEM. The procedure for measuring HeLa cell viability was the same as that described in the MTT viability assay.

Material Characterization. Transmission electron microscopy (TEM): TEM micrographs were taken on a JEOL 1200 EXII (Tokyo, Japan) with a 100 kV voltage. TEM samples were prepared by dipping a Formvar-coated copper grid (Ted Pella, Redding, CA) into an ethanolic MS NP solution and the grid was dried under air. XRD: Powder X-ray diffraction (XRD) patterns were measured on a Siemens Bruker-AXS D-5005 X-ray diffractometer (Karlsruhe, Germany) using filtered Cu Ka radiation (λ = 1 .5406 A) at 45 kV and 20 mA. Data were recorded by step scan with a step size of 0.040° and a dwell time of 1 .0 second.

A/_£-sorption measurements. The N₂ adsorption-desorption isotherms were measured on a Quantachrom Autosorb-1 (Boynton Beach, FL) at 77K. The surface area and pore size of samples were determined by the Brunauer-Emmett-Teller (BET) and Barret-Joner-Halenda (BJH) methods, respectively.

Dynamic light scattering (DLS). The hydrodynamic diameter measurements were carried out at either RT or 37°C using DLS with a Brookhaven 90Plus/BIMAS particle analyzer (Holtsville, NY) equipped with a 35 mW red diode laser (660 nm). All the particles suspended in various media were at a NP concentration of 1 mg/mL and were filtered through a 0.2 μιη GHP filter to remove any possible dust. Three one-minute runs were performed on each measurement. The average DLS diameter was calculated from three independent samples. The DLS size distribution was plotted using a lognormal analysis method.

ζ-Potential Measurements. All MSN solutions were prepared in D.I. water and SBF at a concentration of 1 mg/mL. ζ-potential was measured using a Brookhaven ZetaPALS Zeta-Potential Analyzer (Holtsville, NY). Five runs and ten cycles were set for each measurement. Each sample was measured three times.

²⁹ Si solid state NMR. Solid state 29Si NMR spectra were recorded using a Varian VNMRS spectrometer operating at ¹ H Larmor frequency of 700 MHz. Samples were spun at the magnetic angle (10 kHz) in a BioMAS Varian triple resonance probe. A single 90° pulse was applied to ²⁹Si channel with ¹ H decoupling during data acquisition. A recycle delay of 30 seconds was used between scans. The area of the Q⁴, Q³, and Q² peaks were calculated by Gaussian function fitting in OriginPro 8.5 software (Northampton, MA).

Optical Microscopy. Following 48-hour exposure to DoxHCI-MS42@PEG/TMS-hy-c, the treated HeLa cells were washed with serum-free DMEM two times and observed under a Nikon Eclipse TE 2000- U inverted microscope (Melville, NY). The images were recorded using a Photometries QuantEM 512SC camera (Tucson, AZ) with Meta-Morph imaging software (Molecular Devices, Downingtown, PA).

Results and Discussion MSN Synthesis and Characterization. The synthesis procedure for hydrothermally treated, highly organosilane-modified MS NPs followed in this work is shown in Figure 32. To increase the dispersity and stability of the MSNs, the as-synthesized bare MSNs were simultaneously functionalized with two types of organosilane, a hydrophilic silane (polyethyleneglycol-silane, PEG-silane) and a hydrophobic silane (trimethylchloro silane, TMS or 3,3,3-trifluoropropyldimethylchlorosilane, TFS) before the surfactant removal step because irreversible aggregation often occurs during the surfactant removal process. In addition, hydrothermal treatment is a key step to largely increase the amounts of organosilane modified on the interior or exterior surface of MSNs as demonstrated in Lin et al. (201 1 ) (Example 3). Herein, small

MSNs having 42 nm-diameter were modified with PEG/TMS or PEG/TFS using hydrothermal treatment and were purified by centrifugation (MS42@PEG/TMS-hy-c and MS42@PEG/TFS-hy-c, respectively). In addition, smaller MSNs (25nm-diameter) modified with PEG and TMS (denoted as MS25@PEG/TMS-hy- c) were also synthesized using this hydrothermal comodified method. Transmission electron microscopy (TEM) images of MS42@PEG/TMS-hy-c, MS42@PEG/TFS-hy-c, and MS25@PEG/TMS-hy-c are shown in Figure 33. The TEM images show that both types of NPs have a hexagonal pore structure, regardless of different hydrophobic silane modification or size. However, the pore structure inside these NPs is somewhat disordered due to organosilane incorporation into the silica framework during the hydrothermal silica restructuring treatment. Only one strong peak (100) was detected in the low angle X-ray diffraction (XRD) patterns of MS42@PEG/TMS-hy-c, MS42@PEG/TFS-hy-c, and MS25@PEG/TMS-hy-c (Figure 33D), further confirming the two-dimensional hexagonal pore ordering inside particles observed in TEM images. Compared to MS42@PEG/TMS-hy-c, the (100) peak of MS25@PEG/TMS-hy-c is broader due to short-range ordering of pore structure inside the particle. The N₂ adsorption-desorption measurements of MS42@PEG/TMS-hy-c and MS25@PEG/TMS-hy-c show that both particles exhibit a typical type IV isotherm (Figure 33E). The surface areas of MS42@PEG/TMS-hy-c and MS25@PEG/TMS-hy-c were 629 m²/g and 498 m²/g, respectively. Compared to the high surface area (1 163 m²/g) of unmodified MSNs (MS42-d) reported by Lin et al. (201 1 ), a large surface area decrease occurred for these organo-modified MSNs. Also, there was a large decrease in pore diameter from 2.4 nm for MS42-d to 1 .6 nm for

MS42@PEG/TMS-hy-c. These results confirm that large amounts of short-length PEG and TMS were indeed incorporated and modified either outside or inside the pores during the hydrothermal treatment causing the decrease in pore size and surface area.

Dispersity and Stability of Hydrothermally Treated Dual-Organosilane Modified MSNs. To examine the dispersity of these organo-modified MSNs, they were directly transferred to a highly salted medium, simulated body fluid (SBF) by centrifugation and their hydrodynamic size was measured using dynamic light scattering (DLS). As shown in Figure 33F, the hydrodynamic size distributions of

MS42@PEG/TMS-hy-c and MS42@PEG/TFS-hy-c were almost identical, and each had a hydrodynamic size of approximately 60 nm. Compared to the hydrodynamic size of as-synthesized MS42@PEG/TMS- hy-c and MS42@PEG/TFS-hy-c listed in Table 9, the hydrodynamic size of these modified NPs was still around 60 nm in SBF after nine repeated high-speed centrifugation and ultrasonication steps during the surfactant removal and redispersion process. Table 9. Physicochemical characteristics of highly organo-modified MSNs

The NP concentration is 1 mg/mL. The measurements were taken at R.T.; values presented mean ± SD from triplicate measurements.

In addition, the smaller MS25@PEG/TMS-hy-c NPs were also well-dispersed in SBF and had a hydrodynamic size of 40 nm, which was almost the same as the as-synthesized MS25@PEG/TMS-hy-c. These results showed that modifying the MSNs with hydrophilic and hydrophobic silane can greatly improve their dispersity. However, the single time measurement of the hydrodynamic size of NPs only shows short-term stability (minutes). Long-term (days) particle stability of three types of MSNs: bare

MSNs (MS42-d), highly PEGylated MSNs (MS42@PEG-hy-c), and highly PEG/TMS dual-modified MSNs (MS42@PEG/TMS-hy-c) was measured in biological media, including phosphate buffer saline (PBS), cell culture media (Dulbecco's Modified Eagle's Medium+10% fetal bovine serum, DMEM+10%FBS), and SBF at 37°C. As shown in Figure 34A, the hydrodynamic size of MS42-d and MS42@PEG-hy-c in SBF increases over time. This result shows that these particles form irreversible aggregates in a biologically relevant environment. Interestingly, MS42@PEG-hy-c increased in size at a much slower rate, but MS42@PEG/TMS-hy-c and MS42@PEG/TFS-hy-c did not increase in size at all over a 15 day incubation period (Figure 34A), showing that the additional hydrophobic silane modification/incorporation on/into MSNs improved the stability of hydrothermally treated PEGylated MSNs. For the smaller

MS25@PEG/TMS-hy-c NPs, no size change was also observed even after 15-day aging in SBF at 37°C. In addition, no significant size change of MS42@PEG/TMS-hy-c in other biological media after 15-day aging at 37°C was observed. All these results confirm that dual-silane surface modification via hydrothermal treatment significantly improves the MSN stability in various biologically relevant media, even with a particle size as small as 25 nm.

Another common problem in various nanotherapeutic candidates is irreversible aggregation once the NPs are dried. The redispersity of hydrothermally treated dual-organosilane-modified NP powder was evaluated by drying the NPs from ethanolic suspensions using rotary evaporation (Figure 35A) and simply redispersing the NP powder in D.I. water by ultrasonication for five minutes. One can clearly see well suspended, optically transparent colloidal solutions (Figure 35B) prepared from either MS42@PEG/TMS- hy-c or MS25@PEG/TMS-hy-c powders. The hydrodynamic size distributions of MS42@PEG/TMS-hy-c, MS42@PEG/TFS-hy-c, and MS25@PEG/TMS-hy-c NP solutions prepared from the powder samples are shown in Figure 35C. The hydrodynamic size of dried MS42@PEG/TMShy-c and MS42@PEG/TFS-hy-c was about 60 nm in D.I. water and almost identical to as-synthesized MS42@PEG/TMS-hy-c and MS42@PEG/TFS-hy-c (Table 7). Even for MS25@PEG/TMS-hy-c with smaller diameter, the dried NPs can be reversibly dispersed to aqueous solution, and their hydrodynamic size was maintained at about 40 nm (Figure 35C). In addition, no significant hydrodynamic size change (Figure 35D) or visible particle precipitate (inset photograph of Figure 35D) was observed for any of the NP solutions prepared from MS42@PEG/TMS-hy-c, MS42@PEG/TFS-hy-c, and MS25@PEG/TMShy-c powders after 15-day aging in SBF at 37°C. This result is a vast improvement over previous published work on improving MSN dispersity (Liong et al., 2008; Wang et al., 201 1 ; Meng et al., 201 1 ; Urata et al., 201 1 ; Canda et al., 201 1 ). This is the first work synthesizing highly stable and redispersible MSNs.

Degradation of Hvdrothermally Treated Dual-Organosilane Modified MSNs. A variety of methods were used to study the degradation of MSNs in biological media. TEM images taken before and after 10- day SBF aging at 37°C (Figures 36A, B, C, and D) showed that MS42-d experienced a decrease in diameter and had almost completely lost its porous structure following aging while, MS42@PEG/TMS-hy- c showed some degradation but still retained its porous structure. XRD patterns of the MS42-d and MS42@PEG/TMS-hy-c after 10-day SBF aging further confirmed that pore collapse was much more pronounced for MS42-d in SBF (Figure 36E). Colorimetric silicate assays showed that MS42-d experienced fast bulk silica degradation, followed by a decrease in silicon amount because of the formation and deposition of magnesium/calcium silicate layers on/in MSNs and a subsequent slow dissolution process. This result is similar to a published work reported by He et al. (201 1 ). In contrast, MS42@PEG/TMS-hy-c had a much slower dissolution behavior and degradation rate (Figure 36F) than MS42-d. In addition, ²⁹Si solid-state MAS NMR was used to investigate the degree of silica condensation and connectivity of the organic groups to the silica framework in four types of MSNs: MS42-d,

MS42@PEG-c, MS42@PEG-hy-c and MS42@PEG/TMS-hy-c. Spectra for all four classes of MSNs showed three peaks at -1 10 ppm (Q⁴, Si(OSi)₄), -100 ppm (Q³, Si(OSi)₃(OH), and -90 ppm (Q²,

Si(OSi)₂(OH)₂). Q⁴, Q³, and Q² represented fully condensed silica, silica with one terminal hydroxyl group, and silica with germinal hydroxyls, respectively. For MS42@PEG-c and MS42@PEG-hy-c, two additional peaks were observed at -68 ppm (T³, R-Si(OSi)₃) and -58 ppm (T², R-Si(OSi)₂(OH)), corresponding to the PEG-silane modification and incorporation into the MS NPs (Figure 37B and C). For MS42@PEG/TMS- hy-c, one more additional peak was detected at 15 ppm (M¹ , R3-Si(OSi), indicating the trimethyl group functionalization on/in the MSNs. The relative ratio of partially and fully condensed silicon sites

[(Q³+Q²)/Q⁴] calculated by integrating the peak areas yields an estimate of degree of condensation in the silica framework. The (Q³+Q²)/Q⁴ ratios for MS42-d, MS42@PEG-c, MS42@PEG-hy-c and

MS42@PEG/TMS-hy-c are 1 .6, 1 .3, 0.90, and 0.77, respectively. This result shows that organosilane modification and hydrothermal treatment greatly increased the amount of fully condensed silica. In addition, compared to the highly negative charge on the MS42-d (-34.5 mV), a large decrease in surface charge on MS42@PEG/TMS-hy-c (-14.5 mV) and MS42@PEG/TFS-hy-c (-16.4 mV), as measured using ζ-potential analysis, further confirmed that most of the surface silanol groups were eliminated by PEG and TMS functionalization through the hydrothermal process. In summary, the high stability of these organo- modified MSNs was due to more fully condensed silica and more organosilane anchored during hydrothermal treatment. These changes resulted in higher resistance to silica dissolution (biodegradation) due to increased hydrophobicity (Koyano et al., 1997) and prevention of pore collapse (Lin et al., 201 1 ) and irreversible aggregation from silica deposition (He et al., 2010; Canda et al., 2010) on the particles. Additionally, the high redispersity is very likely attributable to large amounts of organosilane modification completely eliminating either the outer surface or interior silanol groups.

In Vitro Biocompatibilitv of Hvdrothermally Treated Organo-Modified MSNs. The biocompatibility of MS42-d and MS42@PEG/TMS-hy-c was first assessed by incubating the nanoparticles with human endothelial and skin fibroblast cells. Compared to a significant decrease in viability caused by MS42-d as the NP concentration exceeded 200 μς/ηιί, the MS42@PEG/TMS-hy-c NPs did not influence either the human endothelial or skin fibroblast cell viability even after 24-hour exposure at 1000 μς/mL (Figures 38A and B). In addition, no significant reactive oxygen species were generated by MS42@PEG/TMS-hy-c in human endothelial cells after 24-hour exposure (see Figure 38C), showing that the PEG and TMS modification did not produce species that were cytotoxic to mammalian cells.

Since the MS42@PEG/TMS-hy-c is designed to be intravenously injected, its compatibility to red blood cells (RBCs) and platelets was also examined by hemolysis and a lactate dehydrogenase (LDH) assay. Following a similar trend, 200 μg/mL of MS42-d NPs caused significant membrane damage in

RBCs (>90% cell lysis) and platelets (>60% cell lysis) after 3-hour incubation, but almost zero percent hemoglobin and LDH release occurred after 200 μg/mL of MS42@PEG/TMS-hy-c NP exposure. These results show that MS42@PEG/TMS-hy-c NPs are much more biocompatible than unmodified MSNs and that the additional TMS modification does not cause any damage or produce toxic species.

Redispersible Fluorescent, Magnetic, and Anticancer Drug Loaded MSNs. To further demonstrate the versatility of this hydrothermal-assisted dual-organosilane modification method, fluorescent and magnetic functionality were incorporated into these ultrastable, redispersible, small and highly organo- modified MSNs. First, a green fluorescent MSN, called FITC-MS42@PEG/TMS-hyc, was prepared by incorporating FITC was into the MS42@PEG/TMS-hy-c with a commonly used cocondensation method (Lin et al., 2005). The TEM image of FITC-MS42@PEG/TMS-hy-c showed no size or morphology change after FITC incorporation compared to MS42@PEG/TMS-hy-c (Figure 39A). To study the redispersity of FITC-MS42@PEG/TMS-hy-c, the dried powder of FITC-MS42@PEG/TMShy-c (Figure 39D) was redispersed into D.I. water by ultrasonication and its hydrodynamic size was measured using DLS. The hydrodynamic size distribution of as-synthesized FITC-MS42@PEG/TMShy-c and powdered FITC- MS42@PEG/TMS-hy-c confirmed the excellent redispersity of FITCMS42@PEG/TMS-hy-c. Colorless and transparent NP solutions (in D.I. water or SBF) showing Tyndall light scattering behavior were prepared from powdered FITC-MS42@PEG/TMS-hy-c samples (Figure 39E). The fluorescence of FITC-

MS42@PEG/TMS-hy-c under UV illumination was homogeneously distributed in SBF (Figure 39F). Similarly, the magnetic MSNs, denoted as Fe₃0₄@MS@PEG/TMShy-c, were synthesized using Fe304 NPs as cores upon which the MS shell was deposited. Figure 39G showed that Fe₃0₄ NPs were successfully coated with a MS shell and functionalized with PEG/TMS. The Fe₃0₄@MS@PEG/TMS-hy-c NPs can be simply synthesized by adding Fe₃0₄ NPs prior to the silica condensation step in the synthesis procedure of MS42@PEG/TMS-hy-c. Again, the comparison of hydrodynamic size distribution (Figure 39H) between as-synthesized Fe₃0₄@MS@PEG/TMS-hy-c NP solution and Fe₃0₄@MS@PEG/TMS-hy-c solution prepared from powder samples (Figure 391) showed the great resdispersity of

Fe₃0₄@MS@PEG/TMS-hy-c powder. The homogenous (Figure 39J) and clear NP solutions in D.I. water and PBS were prepared from Fe₃0₄@MS@PEG/TMS-hy-c powder. In addition, high magnetic response of the NP solution was observed by placing a strong neodymium magnet close to the vial (Figure 39K). It is worth mentioning that no NP separation from the solution over a long period of time (> 2 hours) further confirmed the truly redispersible nature of Fe304@MS@PEG/TMS-hy-c and extremely high colloidal stability of resuspended Fe₃0₄@MS@PEG/TMS-hy-c NPs.

To explore the capability of the ultrastable MS42@PEG/TMS-hy-c NPs as anticancer drug delivery carriers, two types of doxorubicin, water-soluble doxorubicin hydrochloride (DoxHCI) and poorly water-soluble doxorubicin (Dox) were loaded to MS42@PEG/TMS-hy-c NPs, denoted as DoxHCI- MS42@PEG/TMS-hy-c and Dox-MS42@PEG/TMS-hy-c, respectively. As shown in Figure 40A, DoxHCI can be completely dissolved in water but Dox formed significant aggregates in water. The DoxHCI and

Dox loading into MS42@PEG/TMS-hy-c NPs were simply confirmed by the red color from drug-loaded

MS42@PEG/TMS-hy-c NPs after separation from the NP-incubated doxorubicin solutions by

ultracentrifugation. The loading weight percentages (compared to overall weight of the particle) of DoxHCI and Dox in MS42@PEG/TMS-hy-c NPs were 3.0% and 2.8%, respectively. The DoxHCI-

MS42@PEG/TMS-hy-c and Dox-MS42@PEG/TMS-hy-c were dried under vacuum (Figure 40B) and redispersed into D.I. water by ultrasonication. Both DoxHCI-MS42@PEG/TMS-hy-c and Dox-

MS42@PEG/TMS-hy-c powders can be resuspended, forming transparent aqueous solutions without observable precipitate (Figure 40C). Although doxorubicin-loaded MSNs have precedent in the literature (Meng et al., 201 1 ; Chen et al., 201 1 ; Zhu et al., 2010), none of the previous work has shown the hydrodynamic size or examined the particle dispersity after drug loading. Compared to a large hydrodynamic size (>2000 nm) of resuspended DoxHCI-loaded MS42-d (denoted as DoxHCI-MS42-d) NP solution, the hydrodynamic sizes of redispersed DoxHCI-MS42@PEG/TMS-hy-c and Dox- MS42@PEG/TMS-hy-c NP solutions were both retained around 60 nm (Figure 40D), further showing that MS42@PEG/TMS-hy-c NPs can be used as excellent drug carriers, especially to improve the

dispersity/solubility of poorly soluble drugs or hydrophobic drugs in aqueous solutions.

Drug Delivery and Cytotoxic Efficacy of Hvdrothermallv Treated Organo-Modified MSNs.

Cumulative drug release profiles for DoxHCI-MS42@PEG/TMS-hy-c and Dox-MS42@PEG/TMS-hy-c in SBF (pH=7.4 and 5.0) at 37°C are shown in Figure 40E. DoxHCI-MS42@PEG/TMS-hy-c exhibited very quick burst release (about 50% of drug release within 1 hour at pH=7.4), followed by a sustained, relatively slow release (about 60% of drug release within 48 hours at pH=7.4). The initial rapid release rate from DoxHCI-MS42@PEG/TMS-hy-c NPs is likely attributable to the high solubility of DoxHCI and the DoxHCI drugs held weakly or without interaction to the interior surface of MS42@PEG/TMS-hy-c.

Compared to DoxHCI-MS42@PEG/TMS-hy-c, Dox-MS42@PEG/TMS-hy-c showed a much slower initial release rate and lower released amounts (less than 20% of drug release within 48 hours at pH=7.4) because of the low solubility of Dox and strong hydrophobic interactions between Dox and TMS. This is the first example showing that the release behavior of doxorubicin from MSNs depends on the solubility of doxorubicin. Furthermore, both DoxHCI-MS42@PEG/TMS-hy-c and Dox-MS42@PEG/TMS-hy-c exhibited faster release rate and greater release amounts in mildly acidiccondition (pH=5.0) due to higher solubility of both DoxHCI and Dox at lower pH (more protonated -NH₂ groups on DoxHCI and Dox). This acidic environment will be relevant if the drug delivery nanoparticles are ever taken up into intracellular acidic organelles (Lee et al., 2010; Muhammed et al., 201 1 ; Rim et al., 201 1 ).

To examine whether the released doxorubicin was still able to kill cancer cells, the cytotoxic efficacy of DoxHCI-MS42@PEG/TMS-hy-c and Dox-MS42@PEG/TMS-hy-c on HeLa cells was investigated. The HeLa cells were incubated with either free DoxHCI, DoxHCI-MS42@PEG/TMS-hy-c, or Dox-MS42@PEG/TMS-hy-c at equivalent doxorubicin doses for 24, 48, or 72 hours. The half maximal inhibitory concentration of doxorubicin (IC₅₀) was determined by MTT viability data. The comparison of cytotoxcity and IC₅₀ of free DoxHCI and DoxHCI-MS42@PEG/TMS-hy-c on HeLa cells is shown in Figures 41 A-C and Table 10. Table 10. IC50 of free DoxHCI, DoxHCI-MS42@PEG/TMS-hy-c, and Dox-MS42@PEG/TMS-hy-c after 24, 48, or 72-hour incubation with HeLa cells

These data reveal that free DoxHCI and DoxHCI-MS42@PEG/TMS-hy-c have similar cytotoxicity to HeLa cells after 24-hour incubation, and then MS42@PEG/TMS-hy-c shows enhanced cytotoxicity after 48-hour or 72-hour incubations. This result was further confirmed by optical microscopy images of HeLa cells after incubation with free DoxHCI and DoxHCI-MS42@PEG/TMS-hy-c at equivalent doxorubicin concentration (0.1 μΜ). Compared to the control image (only medium and serum incubation, Figure 41 A), some viable cells were still observed after 48-hour free DoxHCI exposure (Figure 41 B) but all the HeLa cells changed to spherical or irregular morphology (damaged or dead cells) after 48-hour DoxHCI- MS42@PEG/TMS-hy-c exposure (Figure 41 C). The enhanced cytotoxicity of DoxHCI-MS42@PEG/TMS- hy-c is likely due to the slow drug release from DoxHCI-MS42@PEG/TMS-hy-c NPs associated or taken up by HeLa after 24-hour incubation. Therefore, more localized doxorubicin was delivered to HeLa cells. In addition, the comparison of cytotoxicity of DoxHCI-MS42@PEG/TMS-hy-c and Dox-MS42@PEG/TMS- hy-c is shown in Figure 42. Both exhibited dose and time-dependent cytotoxicity to HeLa cells. Based on IC50 data, DoxHCI-MS42@PEG/TMS-hy-c has higher cytotoxic efficacy on HeLa cells than Dox- MS42@PEG/TMS-hy-c; this is likely due to faster and higher percentage of drug release from DoxHCI- MS42@PEG/TMS-hy-c NPs. Although the Dox-MS42@PEG/TMS-hy-c has lower cytotoxic efficacy, the poorly water-soluble doxorubicin was able to disperse well in aqueous solutions and kill cancer cells when delivered by MS42@PEG/TMS-hy-c. The slow release property of Dox-MS42@PEG/TMS-hy-c may be useful for cases that require long-term cancer therapy.

Conclusions

In summary, the external and internal surfaces of MSNs were co-modified with two types of organosilanes, hydrophilic silane (PEG-silane) and hydrophobic silane (TMS or TFS) accompanied by a hydrothermal treatment to increase their dispersity and long-term colloidal stability in biologically relevant media. Importantly, these highly organo-modified NPs can be dried and redispersed into a buffer solution with no significant change in size or stability. To demonstrate the versatility of this hydrothermal-assisted dual-organosilane modification method, the redispersity of ultrasmall (25 nm), fluorescent, and magnetic co-modified MSNs was demonstrated. These highly organo-modified MSNs also exhibit high

biocompatibility as measured by a red blood cell lysis assay and platelet membrane integrity assay as well as unperturbed cell viability to human endothelial and skin fibroblast cells. In addition, a common anticancer drug, doxorubicin with two different forms, water-soluble DoxHCI and poorly water-soluble Dox, were loaded into these highly organo-modified MSNs. Most importantly, these drug-loaded MSNs can also be dried and resdispered in aqueous solution without significant hydrodynamic size change compared unloaded MSNs; this is especially important for the poorly water-soluble Dox-loaded MSNs. This is the first example showing the redispersity of MSNs after drug loading. The Dox-loaded MSNs exhibited slower drug release kinetics and lower percent Dox release compared to DoxHCI-loaded MSNs. Finally, these redispersible drug-loaded MSNs show dose-and time-dependent cytotoxic effects on cancerous (HeLa) cells. The DoxHCI-loaded MSNs further exhibited higher cytotoxicity than free DoxHCI.

These ultrastable, redispersible, and small MS nanotherapeutics have great potential in passive tumor targeting and therapy applications, as well as in non-medical applications for delivery of one or more compounds, e.g., in cosmetics.

References

Barik et al., Pediatr. Res.. 103: 253 (2008).

Bass et al., Chem. Mater.. 19:4349 (2007).

Bernardos et al., ACS Nano. 4:6353 (2010).

Cai et al.. Chem. Mater.. 13:258 (2001 ).

Cauda et al., J. Am. Chem. Soc. 133:6484 (201 1 ).

Cauda et al., Microporous Meso. Mater., 132:60 (2010).

Chang et al., Environ. Sci. Technol., 41:2064 (2007).

Chen et al.. ACS Nano. 4:6001 (201 1 ).

Cho et al.. Toxicol. Lett.. 175:24 (2007).

Coradin et al., Spectroscopy: Int. J., 18:56 (2004).

Coradin et al., Spectroscopy: Int. J., 18:567 (2004).

Darbandi et al., Chem. Mater., 19:1700.

Deng et al., J. Am. Chem. Soc, 130:28 (2008).

Pi Pagua et al., J. Inorg. Biochem., 102:1416 (2008).

Dorovolskaia et al., Nano Lett., 8:2180 (2008).

Fang et al., Small. 5:1637 (2009).

Farokhzad et al., ACS Nano. 3:16 (2009).

Ge et al.. Anal. Chem.. 83:2598 (201 1 ).

Gerashchenko et al., Cytometry, 49:56 (2002).

Giri et al., Angew. Chem., Int. Ed., 44:5038 (2005).

Graf et al., Langmuir, 19:6693 (2003).

He et al., Anal. Chem.. 80:9597 (2008).

He et al., Biomaterials, 31 :1085 (2010).

He et al.. J. Mater. Chem.. 21 :5845 (201 1 ).

He et al.. J. Small. 7:271 (201 1 ).

He et al., Microporous Meso. Mater., 131:314 (2010).

He et al., Small. 5:2722 (2009).

Huang et al., ACS Nano. 5:5390 (201 1 ).

Hudson et al., Biomaterials, 29:4045 (2008).

Jovanovic et al., Biomacromolecules, 7:945 (2006).

Kaewamatawong et al., Toxicol. Pathol. , 33:745 (2005).

Ke et al., Cancer Res.. 63:7870 (2003).

Kim et al., Adv. Mater.. 22:4280 (2010).

Kim et al.. Adv. Mater.. 20:478 (2008).

Kim et al., Angew. Chem. Int. Ed., 45:4789 (2006).

Kim et al., Angew. Chem. Int. Ed., 47:8438 (2008). Kim et al., J. Am. Chem. Soc, 128:688 (2006).

Kobler et al., ACS Nano, 2:791 (2008).

Kokubo et al., J. Biomed. Mater. Res., 24:721 (1990).

Koyano et al., J. Phvs. Chem. B, 101:9436 (1997).

Koziara et al., J. Pharm. Res. 22:1821 (2005).

Kresge et al., Nature, 359:710 (1992).

Lai et al., J. Am. Chem. Soc, 125:4451 (2003).

Lang et al., Chem. Mater., 16:1961 (2004).

Lee et al., Adv. Funct. Mater., 19:215 (2009).

Lee et al.. Anaew. Chem. Int. Ed., 49:8214 (2010).

Lee et al., Angew. Chem., Int. Ed., 45:8160 (2006).

Lee et al., J. Am. Chem. Soc, 132:552 (2010).

Liao et al., ACS Appl. Mater. Interfaces, 3:2607 (201 1 ).

Lim et al., Adv. Mater., 20:1721 (2008).

Lim et al., ChemBioChem. 8:2204 (2007).

Lin et al., Adv. Mater., 19:577 (2007).

Lin et al., Chem. Comm., 47:532 (201 1 ).

Lin et al., Chem. Mater.. 17:4570 (2005).

Lin et al., Chem. Mater.. 18:5170 (2006).

Lin et al.. Chem. Mater., 21 :3979 (2009).

Lin et al., J. Am. Chem. Soc, 132:4834 (2010).

Lin et al., Toxicol. Appl. Pharmacol., 217:252 (2006).

Liong et al., ACS Nano, 2:889 (2008).

Lu et al., Nano Lett. , 7Λ 49 (2007).

Lu et al.. Small, 5:1408 (2009).

Lu et al., Small, 8:1341 (2007).

Luo et al., Anoew. Chem. Int. Ed., 50:640 (201 1 ).

Meng et al.. ACS Nano, 5:4131 (201 1 ).

Molday, U.S. Patent 4,452,773.

Muhammad et al.. J. Am. Chem. Soc, 133:8778 (201 1 ).

Murashov et al., J. Qccup. Environ. Hygiene, 3:718 (2006).

Napierska et al., Small, 5:846 (2009).

Ohulchanskyy et al., Nano Lett., 7:2835 (2007).

Rim et al., Anoew. Chem. Int. Ed., 50: (201 1 ).

Rio-Echevarria et al., J. Mater. Chem., 20:2780 (2010).

Rosenholm et al., ACS Nano, 3:197 (2009).

Rosenholm et al., Nanoscale, 2:1870 (2010).

Roy et al., Proc. Natl. Acad. Sci. U.S.A., 102:279 (2005).

Schlossbauer et al., Angew. Chem. Int. Ed., 48:3092 (2009). Sing et al.. Pure Appl. Chem., 57:603 (1985).

Slowing et al., J. Am. Chem. Soc, 129:8845 (2007).

Slowing et al., Small, 5:57 (2009). Stober et al., J. Colloid Interface Sci.. 26:62 (1968).

Sun et al., J. Am. Chem. Soc, 124:8204 (2002).

Suzuki et al., J. Am. Chem. Soc, 126:462 (2004).

Taylor et al., J. Am. Chem. Soc. 130:2154 (2008).

Tornev et al., Nat. Nanotechnol., 2:295 (2007).

Tu et al., Adv. Mater.. 21:172 (2009).

Urata et al., Chem. Commun., :5094 (2009).

Urata et al., J. Am. Chem. Soc. 133:8102 (201 1 ).

Vallhov et al., Nano Lett.. Z:3576 (2007).

Vivero-Escoto et al., Small. 6:1952 (2010).

Wang et al., ACS Nano. 4:4371 (2010).

Wang et al., Angew. Chem., Int. Ed., 47:2439 (2008).

Wu et al., ChemBioChem, 9:53 (2008).

Xie et al., Adv. Drug Deliv. Rev. , 62:1064 (2010).

Yang et al.. Angew. Chem.. Int. Ed.. 46:8836 (2007).

Yang et al., Langmuir, 24:3417 (2008).

Yu et al., ACS Nano. Z:5717 (201 1 ).

Yu et al., Nanopart. Res., 1 1_:15 (2009).

Zhao et al., J. Am. Chem. Soc. 127:8916 (2005).

Zhao et al., J. Am. Chem. Soc. 131:8398 (2009).

Zhu et al., Small. 3:471 (2010).

All publications, patents and patent applications are incorporated herein by reference. While in the foregoing specification, this invention has been described in relation to certain preferred embodiments thereof, and many details have been set forth for purposes of illustration, it will be apparent to those skilled in the art that the invention is susceptible to additional embodiments and that certain of the details herein may be varied considerably without departing from the basic principles of the invention.

Claims

WHAT IS CLAIMED IS:

1 . A method to prepare biocompatible mesoporous silica nanoparticles with enhanced stability in biological media, comprising:

a) heating a surfactant containing solution having mesoporous silica nanoparticles with a diameter of from about 20 to about 250 nm, to about 70°C to 150°C for about 12 hours to about 48 hours;

b) extracting the surfactant from the heat-treated mesoporous silica nanoparticles;

c) washing the extracted, heat-treated mesoporous silica nanoparticles; and

d) filtering the washed, heat-treated mesoporous silica nanoparticles so as to yield mesoporous silica nanoparticles that have enhanced stability in aqueous biological media relative to mesoporous silica nanoparticles that are not heat treated.

2. The method of claim 1 wherein the biological media comprises phosphate buffered saline.

3. The method of claim 1 wherein the biological media comprises serum.

4. The method of any of one of claims 1 to 3 wherein the mesoporous silica nanoparticles have a diameter of about 30 nm to about 60 nm.

5. The method of any of one of claims 1 to 4 wherein the mesoporous silica nanoparticle comprises an imaging agent.

6. The method of claim 5 wherein the imaging agent is a fluorophore.

7. The method of claim 5 wherein the imaging agent is a magnetic resonance imaging agent.

8. The method of any of one of claims 1 to 7 wherein the mesoporous silica nanoparticle comprises a drug.

9. The method of any of one of claims 1 to 8 wherein the mesoporous silica nanoparticle comprises a coating.

10. The method of claim 9 wherein the coating comprises polyethylene glycol.

1 1 . The method of claim 9 wherein the coating comprises polylactic co-glycolic acid.

12. The method of any one of claims 9 to 1 1 wherein the coating comprises an antibody, targeting peptide, or folic acid.

13. The method of any one of claims 1 to 13 wherein the nanoparticles comprise hydrophobic and hydrophilic silanes.

14. A preparation of stable mesoporous silica nanoparticles prepared by the method of any one of claims 1 to 13.

15. The preparation of claim 14 wherein the stable mesoporous silica nanoparticles have reduced degradation, reduced aggregation, reduced cytotoxicity and/or reduced hemolysis, or any combination thereof, relative to mesoporous silica nanoparticles not subjected to heat treatment.

16. The preparation of claim 14 or 15 wherein the mesoporous silica nanoparticles comprise a drug.

17. The preparation of any one of claims 14 to 16 wherein the mesoporous silica nanoparticles comprise an imaging agent.

18. The preparation of any one of claims 13 to 16 wherein the mesoporous silica nanoparticles comprise a coating.

19. A method of delivering an imaging agent or drug to a subject comprising administering an effective amount of a composition comprising the preparation of any one of claims 14 to 18 to a subject.

20. The method of claim 19 wherein the composition is intravenously administered.

21 . A biocompatible composition comprising stable and redispersible mesoporous silica nanoparticles comprising a hydrophobic organosilane.

22. The composition of claim 21 which is a powder.

23. The composition of claim 21 wherein the mesoporous silica nanoparticles have a diameter of about 30 nm to about 60 nm.

24. The composition of claim 23 wherein the mesoporous silica particles comprise a drug, a chelating agent, an optically detectable dye, or magnetic particles, or a combination thereof.

25. A biocompatible composition comprising stable mesoporous silica nanoparticles comprising a hydrophobic organosilane which particles are dispersed in a biologically compatible medium.

26. The composition of claim 25 wherein the mesoporous silica nanoparticles have a diameter of about 30 nm to about 60 nm.

27. The composition of claim 25 wherein the mesoporous silica particles comprise a drug, a chelating agent, an optically detectable dye, or magnetic particles, or a combination thereof.