Hyaluronic acid-capped compact silica-supported mesoporous titania nanoparticles for ligand-directed delivery of doxorubicin

Accepted Manuscript Full length article Hyaluronic acid-capped compact silica-supported mesoporous titania nanoparticles for ligand-directed delivery of doxorubicin Biki Gupta, Bijay Kumar Poudel, Hima Bindu Ruttala, Shobha Regmi, Shiva Pathak, Milan Gautam, Sung Giu Jin, Jee-Heon Jeong, Han-Gon Choi, Sae Kwang Ku, Chul Soon Yong, Jong Oh Kim PII: DOI: Reference:

S1742-7061(18)30523-3 https://doi.org/10.1016/j.actbio.2018.09.006 ACTBIO 5650

To appear in:

Acta Biomaterialia

Received Date: Revised Date: Accepted Date:

24 April 2018 4 September 2018 6 September 2018

Please cite this article as: Gupta, B., Poudel, B.K., Ruttala, H.B., Regmi, S., Pathak, S., Gautam, M., Jin, S.G., Jeong, J-H., Choi, H-G., Kwang Ku, S., Yong, C.S., Kim, J.O., Hyaluronic acid-capped compact silica-supported mesoporous titania nanoparticles for ligand-directed delivery of doxorubicin, Acta Biomaterialia (2018), doi: https:// doi.org/10.1016/j.actbio.2018.09.006

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Page 1 of 52

Hyaluronic acid-capped compact silica-supported mesoporous titania nanoparticles for ligand-directed delivery of doxorubicin

Biki Guptaa,1, Bijay Kumar Poudela,1, Hima Bindu Ruttalaa, Shobha Regmia, Shiva Pathaka, Milan Gautama, Sung Giu Jinb, Jee-Heon Jeonga, Han-Gon Choic, Sae Kwang Kud, Chul Soon Yonga,**, Jong Oh Kima,*

a

College of Pharmacy, Yeungnam University, 214-1, Dae-Dong, Gyeongsan 712-749, Republic

of Korea b

Department of Pharmaceutical Engineering, Dankook University, 119 Dandae-ro, Dongnam-gu,

Cheonan, 31116, Republic of Korea c

College of Pharmacy, Institute of Pharmaceutical Science and Technology, Hanyang

University, 55, Hanyangdaehak-ro, Sangnok-gu, Ansan 426-791, Republic of Korea d

College of Korean Medicine, Daegu Haany University, Gyeongsan, 712-702, Republic of

Korea *

Corresponding author: Prof. Jong Oh Kim, Ph.D.

Tel: +82-53-810-2813, Fax: +82-53-810-4654, E-mail: [email protected] **

Co-corresponding author: Prof. Chul Soon Yong, Ph.D.

Tel: +82-53-810-2812, Fax: +82-53-810-4654, E-mail: [email protected] **

Co-corresponding author: Prof. Sae Kwang Ku, Ph.D.

Tel: +82-53-819-1549, Fax: +82-53-819-1860, E-mail: [email protected]

1

These authors contributed equally to this work.

Page 2 of 52

ABSTRACT Mesoporous titania nanoparticles (MTN), owing to their high surface area to volume ratio and tunable pore sizes, appear capable of delivering sizable amounts of drug payloads, and hence, show considerable promise as drug delivery candidates in cancer therapy. We designed silica-supported MTN (MTNst) coated with hyaluronic acid (HA) to effectively deliver doxorubicin (DOX) for breast cancer therapy. The HA coating served a dual purpose of stabilizing the payload in the carriers as well as actively targeting the nanodevices to CD44 receptors. The so-formed HA-coated MTNst carrying DOX (HA/DOX-MTNst) had spheroid particles with a considerable drug-loading capacity and showed significantly superior in vitro cytotoxicity against MDA-MB-231 cells as compared to free DOX. HA/DOX-MTNst markedly improved the cellular uptake of DOX in an apparently CD44 receptor-dependent manner, and increased the number of apoptotic cells as compared to free DOX. These nanoplatforms accumulated in large quantities in the tumors of MDA-MB-231 xenograft tumor-bearing mice, where they significantly enhanced the inhibition of tumor growth compared to that observed with free DOX with no signs of acute toxicity. Based on these excellent results, we deduced that HA/DOX-MTNst could be successfully used for targeted breast cancer therapy.

Keywords: mesoporous titania nanoparticles, inorganic nanoplatforms, ligand-directed targeting, doxorubicin, breast cancer, hyaluronic acid

Page 3 of 52

1. Introduction Inorganic-polymer nanocomposites have gained considerable interest in recent years as multifunctional biomaterials that integrate and synergize the versatile properties of inorganic (such as stability, photonics, antimicrobial properties, and molecule loading) and organic (such as

biocompatibility,

amphiphilicity,

cellular

targeting,

and

thermodynamic

stability)

nanocomponents for cancer therapy, such as in drug delivery, imaging, diagnostics, and chemophototherapy. Titanium dioxide (TiO2, titania) has extensive applicability in material chemistry for dye-sensitized solar cells, sensors, displays, adsorbents, energy storage, and pollutant photocatalysis because of its unique semiconducting property, ultraviolet (UV)-photoactivity, chemical inertness, thermal stability, and electrochromism [1-6]. In addition to these properties, mesoporous titania possesses a high surface area and tunable pore topology, similar to mesoporous silica-based materials [7,8]. However, its use in the biomedical field is not very widespread compared with that of mesoporous silica, which could partly be attributed to the difficulty in synthesis of mesoporous titania as stable “nanoparticulate” carriers (< 200 nm) [9]. Indeed, most bioapplication reports on mesoporous titania pertain to the use of continuous mesoporous

films/membranes

rather

than

discrete

particles

or

solid

titania

nanocrystals/semicrystals and mesoporous microspheres [10,11]. The large-sized films with indeterminate structures are not conducive to active or passive cellular internalization and controllable delivery through the cell membrane for cancer diagnosis and treatment. This is because cellular interactions are modulated by the size, shape, and surface chemistry, and a broad mechanistic understanding and screening of shape-cell interactions is required [12]. Additionally, the lack of mesopores in titania nanocrystals greatly limits their drug-loading capacity and might further necessitate extraneous and uneconomical steps such as surface

Page 4 of 52

modifications, drug/molecule conjugation, and purification [10,13]. Mesoporous titania nanoparticles (MTN) can be used to deliver substantial payload by adsorbing the molecules onto the pore surface, which would release the cargo by Fickian diffusion kinetics following administration in the body [14,15]. Furthermore, this process can be controlled by capping the pores with biocompatible polymers. The commonly employed mesoporous carriers, such as MSN, have numerous potential advantages as drug delivery vehicles. However, a number of issues are often associated with the use of these carriers. For instance, toxicity associated with MSN has been viewed as a major impediment to its successful clinical application, demanding comprehensive evaluation during phase 1 clinical trials [16]. Consequently, alternative carriers with potentially low toxicity profile could prove significantly beneficial for drug delivery. Rutile particles of mesoporous titania have exhibited high LD50 values [17, 18]. Moreover, in vitro cytotoxicity and hemolytic potential of TiO2 have been observed to be remarkably lower as compared to non-porous and porous SiO2 [19]. Additionally, MTN could potentially unlatch new avenus in photodynamic therapy since TiO2 is highly capable of generative reactive oxygen species (ROS) under UV light irradiation [20]. Therefore, carefully designed MTN system could prove significant as a nanoscale therapeutic device than contemporary nanocarrier systems. The difficulty in preparing MTN is inherently related to the conventional surfactant-assisted soft template-based methods currently used that normally produce amorphous titania, which subsequently crystallizes following heat treatment, resulting in the collapse of the uniform mesoporous order [21]. Other reported novel methods for conferring mesoporosity have been explored for efficient photocatalysis. These methods produced continuous structures that were preferable for catalysis because the continuity results in higher activity by easing electron transfer within the titania. Some research groups have used inert solid supports such as metal,

Page 5 of 52

metal oxides, and silica nanoparticles (NPs) to construct the so-called core-shell architecture. Wang et al. [22] used silica-coated silver NPs with a mesoporous titania shell for drug-loading and as surface-enhanced Raman scattering tags. Similarly, diverse particles such as gold nanorods, mesoporous silica, graphene oxide, carbon nanospheres, metal-organic frameworks, and iron oxide NPs have been used as solid supports to grow mesostructured titania shell [2327]. Drug release from the uncapped MTN would be drug solubility-dependent following systemic administration, and highly water-soluble drugs might show burst release before reaching the diseased site, thereby escalating acute toxicity. This could be prevented by the prudent application of suitable surface coats. Sub-optimal treatment caused by limited uptake of MTN via passive uptake and non-specific distribution of nanoparticles can be minimized by surface coatings, which confer a new biological identity to the NPs and facilitate avoidance of clearance by the reticuloendothelial system (RES) [28,29]. This strategy can effectively protect the cargo from premature and off-target release as well as metabolism. However, stabilizing the NPs against agglomeration and minimizing the surface-specific toxicity by judicious coatings (such as with polymers with active targeting ligands) can also initiate receptor-mediated endocytosis in cancer cells overexpressing the corresponding receptors for targeted chemotherapy. Considering these points, the present study investigated the application of silica-supported MTN (MTNst) capped with highly biocompatible and non-immunogenic hyaluronic acid (HA) for doxorubicin (DOX) delivery in the treatment of breast adenocarcinoma. Breast cancer is the leading cause of cancer-associated deaths in women, accounting for 26% of newly diagnosed cancer

cases

[30,31].

HA

is

a

major

high-molecular-weight

(MW),

hydrophilic

Page 6 of 52

glycosaminoglycans found in the extracellular matrix [32]. CD44 is a widely expressed transmembrane glycoprotein found in breast adenocarcinoma and a principle cell surface receptor for HA. Although CD44 is expressed at low levels in normal cell types, these cells are not in direct contact with the systemic blood flow, which makes CD44 an excellent therapeutic target and HA a promising ligand for breast cancer treatment [33]. In addition to its high-affinity binding to breast cancer cells, HA prevents the adsorption of opsonin and other RES recognition proteins owing to its numerous negatively charged carbonyl groups, and thus, promotes the drug circulation half-life [34,35]. In this study, DOX was selected as a model payload because of its potent chemotherapeutic relevance in breast adenocarcinoma. The successful fabrication of spheroidal HA-capped solid silica-supported MTN loaded with DOX (HA/DOX-MTNst) was confirmed via physicochemical analyses of size, morphology, surface properties, chemical compositions, textural and porosity properties, and pH-responsive drug release profiles. Additionally, in vitro and in vivo tests were carried out in a human MDA-MB-231 breast cancer cell line and xenograft tumor-bearing mice, respectively, to evaluate the efficacy of the NPs in targeted chemotherapy of breast cancer.

2. Materials and methods 2.1. Materials and cell lines DOX was obtained from Dong-A Pharmaceutical Company (Yongin, South Korea). Tetraethyl orthosilicate (TEOS, 98%) and titanium(IV) butoxide (TBOT, 97%) were purchased from Sigma-Aldrich Corp., (St. Louis, MO, USA); ammonia solution (28%) was from Junsei Chemical Co., (Tokyo, Japan); KlucelTM hydroxypropylcellulose (HPC) LF Pharm from Ashland Inc. (Covington, KY, USA); poly-(L-lysine) hydrochloride (PLL, MW 1600 Da) from Alamanda

Page 7 of 52

Polymers Co., (AL, USA); and HA (molecular weight [MW] 3000 Da) was from B&K Technology Group (Xiamen, China). The other chemicals were of analytical grade and were used with no further purification. The MDA-MB-231 breast adenocarcinoma cell line was purchased from Korean Cell Line Bank (KCLB), Seoul, South Korea. The cells were cultivated in an incubator maintained at 37 °C in an atmosphere of 5% CO2 in Hyclone high-glucose Dulbecco’s modified Eagle’s medium (DMEM, GE Healthcare Life Sciences, Logan, UT, USA) supplemented with 10% fetal bovine serum (FBS, GE Healthcare Life Sciences), 100 U/mL penicillin G sodium, and 100 μg/mL streptomycin sulfate.

2.2. Preparation of NPs 2.2.1. Preparation of silica templates The compact silica templates were fabricated using the method described by Stӧber [36] with suitable modifications. TEOS (428 µL) was added to a solution containing ethanol (21.7 mL), deionized water (5.6 mL), and aqueous ammonia solution (28%, 460 µL), followed by vigorous stirring at 900 rpm for 3 h. The resulting turbid white silica particles were collected by centrifugation at 10000 × g for 15 min and subsequently washed with ethanol and deionized water.

2.2.2. Preparation of MTNst MTNst were prepared using a modification of the method reported by Joo et al. [37] for the synthesis of mesoporous TiO2 hollow shells. Briefly, the silica particles formulated in section 2.2.1 were dispersed in a solution of HPC (50 mg) in ethanol (25 mL) and deionized water (1 mL) by stirring at 700 rpm for 30 min. Then, TBOT (1 mL) as a 5 mL ethanolic solution was

Page 8 of 52

added dropwise to the silica dispersion under continuous stirring. Then, the temperature was increased to 85 °C with further stirring at 900 rpm under reflux conditions for 100 min. The precipitate was collected by centrifugation at 8000 × g for 20 min, subsequently washed with ethanol, and then redispersed in 10 mL ethanol. Then, 5 mL deionized water was added to the ethanolic solution of the NPs, which was then subjected to solvothermal treatment in a polytetrafluoroethylene (PTFE)-lined stainless steel pressure vessel (Parr) at 150 °C for 6 h. The final MTNst carriers were recovered by centrifugation at 8000 × g for 15 min, washed thoroughly with ethanol, followed by deionized water, and then resuspended in deionized water.

2.2.3. Preparation of HA/DOX-MTNst An aqueous solution of DOX (10 mg/mL, 4 mL) was added dropwise to the MTNst dispersion (10 mg/mL, 4 mL). The volume was made up to 10 mL with deionized water, and the solution was stirred overnight at 800 rpm at 25 °C. The free DOX was removed from the DOXMTNst by washing with deionized water. To 8 mL DOX-MTNst dispersion, 200 µL of an aqueous solution of PLL (10 mg/mL) was added and the mixture was mildly sonicated for 60 min. The resulting PLL/DOX-MTNst were centrifuged and washed with deionized water to remove the unreacted PLL. Finally, an aqueous solution of HA (10 mg/mL, 1 mL) was added to PLL/DOX-MTNst with mild sonication for 60 min to form the HA/DOX-MTNst. The excess HA was removed by centrifugation and washing with deionized water. PLL/MTNst and HA/MTNst were formulated using a similar approach barring the addition of DOX.

2.3. Characterization of NPs 2.3.1. Hydrodynamic size, size distribution, and ζ-potential

Page 9 of 52

Dynamic light scattering (DLS) measurements were carried out using a ZetaSizer Nano S90 (Malvern Instruments, UK) to analyze the hydrodynamic particle size, polydispersity index (PDI), and ζ-potential of the MTNst, DOX-MTNst, PLL/DOX-MTNst, and HA/DOX-MTNst. All measurements were performed in triplicates at a scattering angle of 92 o and temperature of 25 oC after adequate dilution with purified water.

2.3.2. Transmission electron microscopy (TEM) The morphological characteristics of MTNst and HA/DOX-MTNst were evaluated using transmission electron microscopy (TEM). The NPs were mounted on carbon film-coated copper grids, dried under mild infrared (IR) radiation, and then TEM images were captured using an H7600 transmission electron microscope (Hitachi, Tokyo, Japan).

2.3.3. Energy dispersive spectroscopy (EDS) Energy dispersive spectroscopy (EDS) was performed to analyze the elemental components of the MTNst. Freeze-dried MTNst samples were fixed on a brass stub and examined under a scanning electron microscope (S-4100; Hitachi, Japan) to select the target region for the EDS analysis performed using the EX-250 EDX analyzer (Horiba, Japan) and PCI image analyzer (Hitachi, Japan). Point analysis model was used to define the regions for analysis in the sample.

2.3.4. Porosity and surface area measurement The surface area, pore size, and pore volume of the MTNst were determined using nitrogen adsorption-desorption isotherms (ASAP 2010, Micromeritics, USA) at 77.35 K. Prior to the measurement, samples were degassed at 150 °C for 24 h. The conventional Brunauer-Emmet-

Page 10 of 52

Teller (BET) and Barrett-Joyner-Halenda (BJH) methods were used to calculate the specific surface area and determine the porosity, respectively.

2.3.5. Thermal analysis 2.3.5.1. Thermogravimetric analysis (TGA) Differential thermal degradation profiles of MTNst, PLL/MTNst, and HA/MTNst were investigated using the thermogravimetric analysis (TGA). The NPs (~2 mg) were placed in a platinum crucible, heated at a constant rate of 10 °C/min over a temperature range of 30 to 1000 °C under an atmosphere of nitrogen using an SDT Q600 TGA and DSC (TA Instruments, New Castle, DE, USA).

2.3.5.2. Differential scanning calorimetry DSC was performed to observe and record the thermal behavior of free DOX as well as freeze-dried HA/MTNst and HA/DOX-MTNst using the DSC Q200 (TA Instruments, DE, USA), which heated the samples over a 40 to 240 °C temperature range at 10 °C/min in a dynamic nitrogen environment.

2.3.6. X-ray diffraction (XRD) scans The crystalline features of free DOX, freeze-dried drug-free carrier (HA/MTNst), and freezedried HA/DOX-MTNst were analyzed and recorded using XRD scanning with a vertical goniometer and X-ray diffractometer (X’pert PRO MPD diffractometer, Almelo, The Netherlands), operated over a diffraction angle (2θ) ranging from 10 to 60° at the scan speed, voltage, and current set at 5 °/min, 40 kV, and 30 mA, respectively.

Page 11 of 52

2.3.7. Drug loading capacity and loading efficiency The loading capacity (LC) and loading efficiency (LE) of DOX inside the HA/DOX-MTNst system were determined by titrating the unloaded DOX during the preparation of DOX-MTNst. DOX-MTNst (1 mL) was filtered using Amicon centrifugal filter units (MW cutoff [MWCO] 10 000 Da; Millipore, MA, USA) at 4000 × g for 10 min. The content of free DOX in the supernatant was determined using high-performance liquid chromatography (HPLC) and, consequently, the LC and LE of DOX in the NP carriers were computed using the following equations: LC = (WT – WU)/(WC + WT – WU) × 100 LE = (WT – WU)/WT × 100 where, WU, WC, and WT are the weights of the unloaded or free drug, carrier, and total drug input, respectively. For the HPLC analysis, an Agilent 1260 Infinity HPLC set-up was used. An Inersil ODS-3 column (5 µm, 4.6 × 150 mm, GL Sciences Inc., Japan) was used, with methanol/water/acetic acid (50:49:1 v/v/v, pH 3.0) as the mobile phase at a flow rate of 1 mL/min, and the column temperature was set at 25.0 ± 1.0 °C. The eluent was analyzed at an ultraviolet (UV) detection wavelength of 480 nm [38].

2.4. In vitro drug release study The in vitro release behavior of DOX from HA/DOX-MTNst under pH 7.4 and 5.0 conditions was evaluated using the diffusion method [39]. Dialysis bags (MWCO 3500 Da) containing 2 mL samples of HA/DOX-MTNst were dipped in 50 mL Falcon tubes containing 25 mL phosphate-buffered saline (PBS, pH 7.4) or acetate-buffered saline (ABS, pH 5.0), with

Page 12 of 52

previously dissolved 1% Tween 20 (v/v). The Falcon tubes were agitated synchronously (100 strokes/min) inside a water bath shaker maintained at a constant temperature of 37 ± 0.5 °C, and aliquots of the release medium (0.5 mL) were withdrawn at predetermined time points and replaced with equivalent amounts of fresh medium. The concentration of DOX in each of the sampled aliquots was quantified using the HPLC method discussed in the previous section. Finally, the statistical analyses of the in vitro drug release data at pH 7.4 and pH 5.0 were performed using the KinetDS software.

2.5. In vitro hemolysis assay In vitro hemolysis induced by free DOX, HA/MTNst, and HA/DOX-MTNst was analyzed by incubating different dilutions of these formulations with a red blood cell (RBC) suspension. Briefly, whole blood samples were collected from Sprague-Dawley (S-D) rats into heparinized tubes and centrifuged at 3000 × g for 10 min to isolate RBCs, which were resuspended in normal saline. Then, 100-µL samples of DOX, HA/MTNst and HA/DOX-MTNst (1, 10, and 100 µg/mL) were placed in separate tubes, each containing 1 mL of the RBC suspension, and a sufficient amount of normal saline was added to each tube to obtain a 10 mL volume. Then, 9 mL each of the normal saline and 1% Triton X-100 solutions were added to additional tubes containing 1 mL of the RBC suspension as the negative and positive controls, respectively. All the sample tubes were incubated at 37 °C for 30 min and then centrifuged at 3000 × g for 10 min. The absorbance of the supernatant was measured at 540 nm using a UV spectrophotometer, and the percentage hemolysis of each test sample was calculated using the following formula: % hemolysis = [(AT – AN)/(AP – AN)] × 100

Page 13 of 52

Where AT, AN, and AP represent the absorbance of the test sample, negative control, and positive control, respectively.

2.6. In vitro cytotoxicity analysis The 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay was used to evaluate the cytotoxic effects of free DOX and HA/DOX-MTNst, as well as the drug-free HA/MTNst on the MDA-MB-231 breast cancer cell line. Briefly, 1 × 104 cells/ well were suspended in high-glucose DMEM supplemented with 10% FBS, seeded in 96-well plates, incubated overnight at 37 °C in a 5% CO2 atmosphere, and then treated with various dilutions of free DOX, HA/MTNst, or HA/DOX-MTNst in serum-free DMEM, followed by 48 h incubation under the specified conditions. Then, the cells were washed twice with PBS and treated with the MTT reagent (100 µL/well, 1.25 mg/mL in serum-free DMEM), followed by 3 h incubation. Finally, dimethyl sulfoxide (DMSO, 100 µL) was added to each well, and the proportion of viable cells against untreated control cells was computed using the following formula: cell viability = (Asample/Acontrol) × 100%, where A represents the absorbance at 570 nm.

2.7. In vitro cellular uptake analysis 2.7.1. Fluorescence-activated cell sorting Fluorescence-activated cell sorting (FACS) was used to illustrate the in vitro cellular uptake characteristics of HA/DOX-MTNst by MDA-MB-231 cells. Briefly, 2 × 105 cells were seeded in each well of the 12-well plates and incubated overnight under the specified conditions. The cells were then treated with HA/DOX-MTNst at concentrations of 0.5, 1, and 2 µg/mL; incubated for 0.5, 1, and 2 h; washed twice with PBS, detached from the well surfaces, and then resuspended

Page 14 of 52

in PBS. The BD FACSCalibur flow cytometer (BD Biosciences, San Jose, CA, USA) was used to analyze the extent of the cellular uptake of HA/DOX-MTNst in each cell sample. Untreated cells were used as the internal control, and 10 000 events were acquired and analyzed per sample. Additionally, the cells were incubated with DOX and HA/DOX-MTNst (equivalent DOX concentration, 1 µg/mL) for 2 h to compare the cellular uptake of HA/DOX-MTNst with that of the free drug.

2.7.2. Confocal laser scanning microscopy (CLSM) Confocal laser scanning microscopy (CLSM) was used to further elucidate the cellular uptake characteristics of HA/DOX-MTNst. Briefly, 3 × 105 cells were seeded on glass coverslips in 12-well plates and incubated overnight under the specified conditions. The cells were then incubated with coumarin 6-loaded HA/MTNst (HA/Cou-MTNst, 1 µg/mL equivalent coumarin 6 concentration) for 2 h with or without pretreatment with 100 µM HA, followed by two washes with PBS and then incubation with 100 nM LysoTracker® Red (ThermoFisher Scientific, MA, USA) for 10 min. The cells were rewashed with PBS and subsequently fixed with 4% paraformaldehyde at room temperature. Finally, the glass coverslips were mounted on glass slides using gel/mount solution (M01, Biomeda, USA) and the CLSM images were captured using the Nikon A1 confocal microscope system (Nikon Instruments Inc., Japan).

2.8. Hoechst imaging assay The Hoechst imaging assay was performed to investigate potential apoptosis-related morphological changes in the MDA-MB-231 cell nuclei. Briefly, 4 × 105 cells were seeded in each well of the six-well plates and incubated overnight under the specified conditions. Then, the

Page 15 of 52

cells were incubated with DOX or HA/DOX-MTNst (equivalent DOX concentration, 1 µg/mL) for 24 h, washed twice with PBS, treated with Hoechst 33342 staining solution (Sigma-Aldrich, St Louis, MO, USA; 2.5 µg/mL in DMEM, 2 mL/well), and then incubated for 10 min in the dark at room temperature. The staining solution was subsequently drained, the cells were washed with PBS, viewed under a Nikon Eclipse Ti fluorescence microscope (Nikon, Japan), and the images were captured using the NIS-Elements BR 4.20.00 microscope imaging software (Nikon, Japan).

2.9. Live/dead assay The effects of treatment with free DOX and HA/DOX-MTNst on the viability of MDA-MB-231 cells were determined using the live/dead assay using Live/Dead viability/cytotoxicity kit (ThermoFisher Scientific, MA, USA), which consists of calcein AM (green) and ethidium homodimer 1 (red) for staining live and dead cells, respectively. Briefly, 3 × 105 cells were seeded in each well of the 12-well plates, incubated overnight under the specified conditions, followed by incubation with free DOX and HA/DOX-MTNst (equivalent DOX concentration, 1 µg/mL), and re-incubated for 24 h. Cells incubated with drug-free media were used as the untreated control. Then, the cells were washed twice with PBS, stained with calcein AM and ethidium homodimer 1 as per the manufacturer’s guidelines, and the images were captured using a Nikon Eclipse Ti fluorescence microscope (Nikon, Japan).

2.10. In vivo animal studies 2.10.1. Development of MDA-MB-231 xenograft tumor model

Page 16 of 52

Six-week-old female BALB/c nude mice were used to establish the MDA-MB-231 xenograft tumor model. The mice were kept under ambient conditions (20 ± 2 °C and 50-60% relative humidity) in the segregated animal facility of Yeungnam University (YU, South Korea) throughout the study period. All the animal handling procedures were in strict accordance with the protocols approved by the Institutional Animal Ethical Committee, YU. The MDA-MB-231 xenograft tumors were developed using our previous method [39]. Briefly, 1 × 107 cells were dispersed in 100 µL serum-free DMEM and injected subcutaneously into the right flank region of the mice with an equal volume of Matrigel® matrix basement membrane HC, phenol red free (Corning Inc., NY, USA).

2.10.2. In vivo/ex vivo imaging and biodistribution analysis Fluorescence-labeled organism bioimaging instrument (FOBI, NeoScience, South Korea) was used to record the in vivo images of MDA-MB-231 tumor-bearing mice and ex vivo images of their principal organs and tumors (n = 3) after administration of Cy5.5-labeled HA/MTNst to demonstrate the biodistribution of the NP drug carriers. Cy5.5-labeled HA/MTNst (1 mg/mL, 100 µL) was administered by intravenous (i.v.) injection via the tail vein of the mice. In vivo images of the mice were captured 24 h after administration of Cy5.5-labeled HA/MTNst. Subsequently, the mice were euthanized, their principal organs (the heart, liver, kidneys, spleen, and lungs) and tumors were extracted, and then the ex vivo images were captured and analyzed to determine the biodistribution of HA/MTNst.

2.10.3. In vivo antitumor investigation

Page 17 of 52

MDA-MB-231 xenograft tumor-bearing mice with tumor volumes of approximately 200 mm3 were randomly assigned to three groups of six mice each (n = 6). Free DOX and HA/DOXMTNst were administered i.v. via the tail-vein to mice in two of the above three groups, whereas the third group was used as the saline-treated control. For both DOX and HA/DOX-MTNst treatment groups, 5 mg/kg body weight equivalent of the drug was administered on day 0, 3, 7, and 10. The tumor volumes and mouse body weight were recorded up to treatment day 28. The tumor volumes were computed using the following formula: volume = ½ × [(longest dimension) × (shortest dimension)2].

2.10.4. Histopathological and immunohistochemical analyses After the 28 day period of the in vivo antitumor study, representative tumor masses and principal organs (the heart, liver, spleen, lungs, and kidneys) were removed from euthanized mice, fixed in formalin, and the histopathological and immunohistochemical (IHC) studies were performed. General hematoxylin and eosin (H&E) staining, followed by evaluation under the Eclipse 80i light microscope (Nikon, Japan) were used to observe the histopathological and histomorphometric changes in the tumor masses and principal organs of the DOX and HA/DOXMTNst-treated mice, compared to those of the control mice. Additionally, IHC staining of the tumor samples using primary antisera and avidin-biotin-peroxidase complex (ABC) methods were performed to identify tumor cell apoptosis, angiogenesis, and tumor cell proliferation. Cleaved caspase-3 and cleaved poly-ADP ribose polymerase (PARP) were used as apoptotic markers, while CD31 and Ki-67 were used as angiogenesis and tumor cell proliferation markers, respectively. The tumor cell volumes, intact tumor cell-occupied regions (%/mm2 of tumor mass), and the mean cleaved caspase-3, cleaved PARP, CD31, and Ki-67-immunolabeled cell

Page 18 of 52

percentages (% / mm2 of tumor mass) for DOX and HA/DOX-MTNst-treated mice were compared with those of the control mice.

2.11. Statistical analysis The results are presented as arithmetic means ± standard deviation (SD). The Student’s t-test was used to test the statistical significance of the mean values, where applicable.

3. Results and discussion 3.1. Preparation and characterization of NPs The schematic illustration of the step-wise fabrication of HA/DOX-MTNst starting from MTNst is shown in Figure 1. The core templates of HA/DOX-MTNst comprised dense silica NPs, which were uniformly sized with ~ 60 nm diameter and a narrow PDI of ~0.05, as observed using DLS characterization (data not shown). The colloidal characteristics of the MTNst, DOXMTNst, PLL/DOX-MTNst, and HA/DOX-MTNst were investigated using DLS (Figure 2A and B). All the NPs exhibited relatively uniform size distribution as indicated by their extremely low PDI values (< 0.15). The hydrodynamic diameter of the blank MTNst was 118.7 ± 2.3 nm, and ζpotential at pH 7 was - 37.1 ± 0.2 mV, which could be due to the presence of surface hydroxyl groups (Ti-OH-) also known as titanol groups, and the twisted octahedron of anatase titania [40]. Noticeably, all ζ-potential measurements were performed at pH 7 because titania exhibits a pHdependent surface charge with the point of zero charges (PZC) at ~ pH 6 [41]. The average diameter of the DOX-MTNst was 119.1 ± 2.9 nm, which was not significantly different from that of the blank MTNst, indicating that the DOX was loaded into the interior of the pores. However, the ζ-potential increased significantly to -16.7 ± 0.7 mV. The PLL/DOX-MTNst prepared by

Page 19 of 52

adsorption of positively charged PLL on the outer surface of the MTNst displayed a considerable increase in particle size (130.4 ± 5.6 nm), while its ζ-potential shifted from a negative to a positive value (44.8 ± 1.3 mV). After coating with the negatively charged HA, the average size was further increased to 153.1 ± 4.2 nm, and the net surface charge reverted to negative (-9.3 ± 0.3 mV), which indicated the successful coating with HA. It should be noted that HA was adsorbed over the PLL coating via electrostatic interactions, rather than chemical conjugation. The TEM analysis (Figure 2C) showed that the NPs were well separated and their sizes were consistent with those of the DLS measurements. The morphology of the MTNst was spheroidal and dark-contrasted, where high electron density MTN homogeneously coated the slightly less contrasted solid silica. HA/DOX-MTNst retained a spheroidal morphology, but the particles were larger and darker owing to the presence of PLL, the HA coating, and DOX loading. The successful synthesis of the NPs was further confirmed using EDS, BET/BJH, TGA, DSC, and XRD analyses. The EDS spectrum of the MTNst is shown in Figure 3A. EDS analysis detected ~15, 47, 22, and 16 weight (%) of Ti, O, Si, and C, respectively. The absence of extraneous peaks of other elements indicated the high purity of the NPs. The porosity and surface area of the MTNst were investigated using the nitrogen adsorption and desorption isotherm (Figure 3B). The hysteresis loop of the MTNst was type IV isotherm with P/Po inflection at 0.6, typical of mesoporous materials [42]. The pore size distribution calculated using the BHJ desorption branch is shown in the inset of Figure 3B. The surface area and pore volume calculated using BET, and average pore size calculated based on the BHJ adsorption branch were 202.61 m2/g, 0.73 cm3/g, and 152.36 Å, respectively. The TGA analyses (Figure 3C) of blank MTNst, PLL/MTNst, and HA/MTNst in the range of 25-1000 °C were performed to estimate the coating amount of PLL and HA. Blank MTNst showed a weight loss of 17.7 wt% because of the

Page 20 of 52

incomplete condensation of TBOT and TEOS and the presence of residual solvents. Assuming similar residual solvent and moisture content, comparison of the weight loss of blank MTNst with PLL/MTNst and HA/MTNst showed that PLL/MTNst contained 3.7% w/w PLL and HA/MTNst contained 10.9% w/w HA. The higher weight percentage of HA could be attributed to the higher molecular weight and charge density of HA. Finally, the DSC (Figure 3D) and XRD (Figure 3E) measurements of pure DOX and freeze-dried HA/MTNst and HA/DOXMTNst were performed to evaluate the crystalline/amorphous state of DOX-loaded in the MTNst. The peak of the DOX thermograph at ~200 °C corresponded to its intrinsic melting endotherm of DOX (Tm = 218 °C) [43]. However, no melting endotherm of DOX was revealed in the DSC thermographs of the HA/DOX-MTNst. This absence of phase transition suggests that the DOX was in a non-crystalline/amorphous state in the pores. This was further confirmed by the XRD studies. The XRD diffractogram of pure DOX identified numerous sharp peaks, indicating its highly crystalline nature. However, HA/DOX-MTNst showed a conspicuous loss of distinctive peaks of DOX. The loss of crystallinity might be attributable to the electrostatic interactions of DOX and the inner pore walls, leading to disordering of the encapsulated DOX. XRD was also used to assess the crystalline structure of the MTN (Figure S1). It revealed diffraction peaks at 2θ of 25.3 ° (101), 37.9 ° (004), 48.1° (200), 54.8 ° (105) for the anatase (tetragonal) phase of titania (JCPDS card, No. 21-1272). The anatase crystallinity might have grown during hydrothermal treatment of the amorphous MTN.

3.2. Drug loading DOX was loaded into the NPs by incubating equal weights of DOX and the MTNst carrier in an aqueous solution with continuous stirring overnight. As mentioned in the previous section,

Page 21 of 52

DOX loading was accompanied by an increase in the ζ-potential of the NPs, which signified that DOX loading involved physisorption and electrostatic adsorption. DOX carries a positive charge (pKa = 8.2), and thus, loading in the negatively charged MTNst is favored [44]. As a result, a high loading capacity of 18.3 ± 1.1% w/w was achieved, which was considerably higher as compared to that in previously reported mesoporous titania systems. For instance, Wu et al. reported an MTN system which accomplished remarkably lower LC of a meagre 2.1% w/w [9]. Such dramatically superior LC could be attributed to the larger pore size and volume of the MTNst. LC of mesoporous materials has been directly associated with their pore volume. [45] The LC of MTNst for DOX was also noticeably greater than a variety of previously reported DOX nanocarriers, including lipid-based nanoparticles [38,39], polymeric nanoparticles [46], and mesoporous silica nanoparticles [47], illustrating undeniable advantages of these carriers as DOX delivery vehicles. The variation of DOX loading efficiency of the NPs over a 30-day period at 25 oC storage temperature has been shown in the supplementary information file (Figure S2) along with changes in particle size, as a measure of the physical stability of the NPs, which indicate that the NPs were fairly stable under the specified storage temperature.

3.3. In vitro drug release The release behavior of DOX from HA/DOX-MTNst was assessed in ABS (0.14 M sodium chloride [NaCl], pH 5.0,) and PBS (pH 7.4, 0.14 M NaCl, Figure 4A). The release profile showed a strong pH-dependence (~40% cumulative release at pH 5.0 and ~60% cumulative release at pH 5.0 after 24 h). The isoelectric point of anatase TiO2 NPs has been reported to be ~ 4 to 6 [48]. Hence, with a decrease in pH from 7.4 to 5.0, MTNst could progressively lose the negative charge, thereby weakening the ionic interaction between the MTNst and PLL.

Page 22 of 52

Additionally, the HA coating could collapse, which could ultimately increase the rate of drug release. A similar phenomenon has been reported with silica nanoparticles coated with PLL [49]. Further, it can be assumed that the higher DOX solubility at pH 5.0 (pKa = 8.2) also contributed to hastening the release. This pH-responsive release of DOX can be strategically exploited to control the drug release following uptake by endo-lysosomal compartments (pH 4.5-5) of breast tumor cells. Statistical analysis of the drug release data (Table S1) revealed a good fit to the Higuchi and Korsmeyer-Peppas models (r2 > 0.9), which indicated that the drug release was primarily by diffusion. The release exponent values of the Korsmeyer-Peppas model (n > 0.45) indicated an anomalous or non-Fickian diffusion mechanism likely mediated the release of DOX from the HA/DOX-MTNst [50].

3.4. In vitro hemolytic toxicity Generally, titania is considered a low-toxicity nanomaterial at very low concentrations (< 100 µg/mL) [51]. However, inhomogeneity in the geometry, porosity, and surface properties of titania NPs have led to several contradictory reports regarding their toxicity. Thus, a hemolysis assay was performed to evaluate the potential toxicity of DOX, HA/MTNst and HA/DOXMTNst using RBCs from mice (Figure 4B). The results showed a concentration-dependent hemolysis and 100 µg/mL DOX caused an immediate onset (within 30 min) of hemolysis to a plateau of ~10%. This might be because of the strong electrostatic binding of positively charged DOX to the negatively charged RBCs (-15 mV) that immediately increased DOX concentration in the RBCs [52]. HA/MTNst and HA/DOX-MTNst with negatively charged and hydrophilic HA coatings caused minimal hemolysis (< 1%) even at the highest concentration tested (100

Page 23 of 52

µg/mL). This highlights the excellent biocompatibility and safety of HA/DOX-MTNst in chemotherapeutic applications.

3.5. In vitro cytotoxicity The cytotoxicity of free DOX, HA/MTNst, and HA/DOX-MTNst was evaluated in the human MDA-MB-231 breast cancer cell line after 48 h incubation (Figure 5). DOX and HA/DOX-MTNst elicited concentration-dependent cytotoxicity, although the blank HA/MTNst showed > 80% cell viability in the range of 0.1–100 µg/mL, suggesting excellent biocompatibility of the drug-free nanocarriers. The inhibitory effect of HA/DOX-MTNst on the in vitro cell proliferation appeared to be remarkably superior to that of the free DOX. This was clearly reflected in the comparison of the IC50 values of the two treatment groups: 0.179 and 0.755 µg/mL for HA/DOX-MTNst and free DOX, respectively. This indicates that HA/DOXMTNst could have increased the transmembrane DOX delivery in the MDA-MB-231 cells by actively targeting the highly expressed CD44 receptors [53-55]. Furthermore, this effect was likely mediated by HA, and could have bypassed the efflux pump-mediated DOX efflux from MDA-MB-231 cells as various tumor-targeted NPs have been reported to be able to effectively circumvent efflux mediated by P-glycoprotein and other efflux pumps and, thereby, often revert multidrug resistance (MDR) [56-58]. This strongly justifies the use of HA/MTNst as an intracellular delivery device for breast cancer treatment. Further, competitive binding cellular uptake experiments were conducted (as discussed in the following section) to justify whether HA mediated the uptake of the NPs via CD44, and thereby likely mediated the enhanced in vitro cytotoxicity of the NPs.

Page 24 of 52

3.6. Cellular uptake The above assumptions of increased DOX delivery and targetability of HA/DOX-MTNst were confirmed using FACS and CLSM-based studies. As shown in Figure 6A & Figure S3, the intracellular uptake was concentration- and incubation time-dependent. Interestingly, when the uptake of free DOX and HA/DOX-MTNst were compared following administration at the same concentration and incubation time (1 µg/mL and 2 h, respectively), HA/DOX-MTNst exhibited higher internalization. This might have been because of several parameters including different uptake mechanism, inhibition or minimization of transmembrane efflux, and subcellular accumulation. HA/DOX-MTNst is internalized by active endocytosis via abundant CD44 receptors, and the internalization is augmented by strong non-specific interaction of negatively charged NPs with cell membrane [59]. We verified the selective interaction of HA/DOX-MTNst with CD44 cell receptors using competitive binding experiments. Specifically, cells were pretreated with free HA (100 µM) to saturate and block CD44 receptors prior to incubation with coumarin-6 labeled HA/MTNst and the fluorescence signals were visualized using CLSM (Figure 6B). The FACS studies quantified the overall fluorescence associated with cells, including adsorbed fluorescent tags, and CLSM studies also indicated that HA/MTNst were indeed internalized into the cells. Results showed that the HA/MTNst localized in lysosomal compartments and the intensity of the intracellular coumarin-6 signal (green fluorescence) decreased significantly after HA pretreatment compared to when the cells were not pretreated with HA. This indicates selective HA receptor-mediated endocytosis, which was partially suppressed following ligand competition with free HA. Thus, HA-coated NPs exhibit improved internalization by CD44 positive cells compared to uncoated NPs.

Page 25 of 52

3.7. Nuclear apoptosis and live/dead viability DOX translocates to the nucleus where it exerts its actions, and the nuclear morphology following incubation of cells with free DOX and HA/DOX-MTNst was investigated using the nuclear Hoechst 33342 dye. As shown in Figure 7A, untreated control cells exhibited weak and dispersed fluorescence in the nuclei. In contrast, cells treated with free DOX and HA/DOXMTNst showed characteristic apoptotic morphological changes. These alterations included condensed or fragmented chromatin assembled as crescents around the nuclear membrane, extended nuclear herniation, budding, and fragmentation (white arrows). HA/DOX-MTNst showed marked nuclear fragmentation and morphological alterations compared to the free DOXtreated cells, indicating the remarkably enhanced apoptosis of MDA-MB-231 breast cancer cells. The live/dead viability assay, which involves co-staining the cell monolayer with calcein AM (green fluorescence, live cells) and ethidium homodimer-1 (red fluorescence, dead cells) was performed to visualize the cells after treatment with free DOX and HA/DOX-MTNst (Figure 7 B). The fraction of cells stained red (dead cells) increased considerably following treatment with HA/DOX-MTNst compared to treatment with free DOX. These results were fairly correlated with those of the in vitro cytotoxicity assay and further corroborated the observation that the HAcoated MTNst acted as an excellent nano device for effective delivery of DOX in breast cancer therapy.

3.8. In vivo biodistribution In vitro studies of tumor targetability of NPs might not always be accurately predictive of their in vivo capabilities. Therefore, a biodistribution study of Cy5.5-labelled HA/MTNst was

Page 26 of 52

performed in subcutaneous MDA-MB-231 xenografted tumor-bearing BALB/c nude mice following intravenous injection (Figure 8). The Cy5.5-labelled NPs were immediately distributed throughout the body of the mice after the injection, followed by progressive accumulation in MDA-MB-231 tumors after 24 h. The NPs showed higher fluorescent signals at the tumor foci with clear boundary from adjacent normal tissues highlighting the tumor-homing capability of HA coating and enhanced permeability and retention (EPR) effect associated with their small size (< 150 nm) (Figure 8A). Tumors and major organs were excised for ex vivo imaging after 24 h, which revealed the expected localization in the liver, spleen, lungs, and kidneys. Interestingly, the tumors showed relatively higher fluorescent intensity than that of the liver (mononuclear phagocytic uptake), suggesting the specific accumulation of NPs in the tumors (Figure 8B(a) & (b)). The preferential tumor accumulation and decreased heart accretion of HA/MTNst is favorable for DOX delivery as DOX is known to be cardiotoxic [60].

3.9. In vivo antitumor effect The antitumor effects were subsequently analyzed in the MDA-MB-231 xenograft tumor mouse model following tail-vein i.v. injections of saline (control), free DOX, and HA/DOXMTNst, followed by tumor volume and body weight monitoring. As shown in Figure 9A, both DOX and HA/DOX-MTNst significantly inhibited tumor growth compared to the control (p < 0.05). However, the HA/DOX-MTNst-treated mice showed significantly higher tumor inhibition than those treated with DOX at equivalent doses, which could be attributed to the higher amounts of DOX delivered to tumor site by selective/targeted uptake, prolonged blood circulation halflife, and the EPR effect [61]. It is worth mentioning that three of the six mice in the HA/DOXMTNst group showed a complete shrinkage of discernable tumor nodules by the end of the

Page 27 of 52

treatment. Similar to several independent reports pertaining to DOX-loaded targeted NPs, HAcoated DOX-MTNst significantly enhanced the antitumor effects [62-64]. The toxic effects were assessed based on body weight monitoring (Figure 9B), and that of mice in the HA/DOXMTNst groups was comparable to that of the control groups. However, mice in the DOX group showed lethargic activity and signs of systemic toxicity, leading to almost 10% body weight loss. Following in vivo antitumor studies, histopathological and IHC analyses were performed on the tumors and major organs excised from each treatment group [38,65]. The micrographs of the H&E-stained tumor sections showed that HA/DOX-MTNst significantly decreased the tumor cell volumes and induced apoptotic alterations in tumor microstructure compared to those of DOX and the control groups (Figure 10A and Table S2). The expression levels of apoptotic markers caspase-3 and PARP increased, while CD31 (an angiogenesis marker) and Ki-67 (a tumor proliferation marker) decreased in the HA/DOX-MTNst treated tumors. This observation indicates the enhanced anticancer effects of the formulation compared to that of the free DOX. The H&E histological profiles (Figure 10B and Table S3) of five major organs – the heart, liver, spleen, lungs, and kidneys – from each treatment groups were evaluated to estimate potential toxicity associated with the treatment. No discernable histological changes were observed in the organs from each treatment groups compared to the control.

4. Conclusions In this study, spheroidal HA-capped and solid silica-supported MTN (HA/MTNst) were successfully fabricated for targeted delivery of DOX to solid breast cancer tumors. The size, morphology, surface properties, chemical compositions, textural and porosity properties, and pHresponsive drug release profiles of the NPs were physicochemical characterized. Additionally, in

Page 28 of 52

vitro and in vivo tests were carried out on breast cancer cells to evaluate the efficacy of NPs in chemotherapy of breast cancer. The NPs exhibited excellent colloidal properties that make it suitable for systemic application in chemotherapy, such as < 150 nm size, high surface area for increased drug loading, pH-responsive release, and CD44 receptor targetability. In vitro studies showed superb biocompatibility and active internalization by CD44 over-expressing cells. In vivo studies showed better antitumor activity and lowered toxic potential than the free drug did. Overall, HA/DOX-MTNst could be successfully used for targeted therapy of breast cancer.

Disclosures No conflict of interest exists for any of the authors.

Acknowledgments This research was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIP) (No. 2018R1A2A2A05021143), and by the Medical Research Center Program (2018R1A5A2025272) through the NRF funded by MSIP.

Appendix A. Supplementary data

Page 29 of 52

References [1] S. Ngamsinlapasathian, S. Pavasupree, Y. Suzuki, S. Yoshikawa, Dye-sensitized solar cell made of mesoporous titania by surfactant-assisted templating method, Sol. Energy Mater. Sol. Cells 90 (2006) 3187-3192. [2] D. Chen, H. Zhang, X. Li, J. Li, Biofunctional titania nanotubes for visible-light-activated photoelectrochemical biosensing, Anal. Chem. 82 (2010) 2253-2261. [3] X. Jiang, B.P. Bastakoti, W. Weng, T. Higuchi, H. Oveisi, N. Suzuki, W.J. Chen, Y.T. Huang, Y. Yamauchi, Preparation of ordered mesoporous alumina-doped titania films with high thermal stability and their application to high-speed passive-matrix electrochromic displays, Chemistry 19 (2013) 10958-10964. [4] Q. Wu, D. Wu, Y. Guan, Hybrid titania-zirconia nanoparticles coated adsorbent for highly selective capture of nucleosides from human urine in physiological condition, Anal. Chem. 86 (2014) 10122-10130. [5] M. Janczarek, E. Kowalska, B. Ohtani, Decahedral-shaped anatase titania photocatalyst particles: synthesis in a newly developed coaxial-flow gas-phase reactor, Chem. Eng. J. 289 (2016) 502-512. [6] S. Fleischmann, A. Tolosa, M. Zeiger, B. Kruner, N.J. Peter, I. Grobelsek, A. Quade, A. Kruth, V. Presser, Vanadia-titania multilayer nanodecoration of carbon onions via atomic layer deposition for high performance electrochemical energy storage, J. Mater. Chem. A 5 (2017) 2792-2801. [7] A. Mitra, A. Bhaumik, B.K. Paul, Synthesis and characterization of mesoporous titanium dioxide using self-assembly of sodium dodecyl sulfate and benzyl alcohol systems as templates, Microporous Mesoporous Mater. 109 (2008) 66.

Page 30 of 52

[8] N.F. Jaafar, A.A. Jalil, S. Triwahyono, N. Shamsuddin, New insights into self-modification of mesoporous titania nanoparticles for enhanced photoactivity: effect of microwave power density on formation of oxygen vacancies and Ti3+ defects, RSC Adv. 5 (2015) 90991-91000. [9] K.C. Wu, Y. Yamauchi, C.Y. Hong, Y.H. Yang, Y.H. Liang, T. Fanatsu, M. Tsunoda, Biocompatible, surface functionalized mesoporous titania nanoparticles for intracellular imaging and anticancer drug delivery, Chem. Commun. (Camb.) 47 (2011) 5232-5234. [10] D.G. You, V.G. Deepagan, W. Um, S. Jeon, S. Son, H. Chang, H.I. Yoon, Y.W. Cho, M. Swierczewska, S. Lee, M.G. Pomper, I.C. Kwon, K. Kim, J.H. Park, ROS-generating TiO2 nanoparticles for non-invasive sonodynamic therapy of cancer, Sci. Rep. 21 (2016) 23200. [11] H. Kanetaka, G. Liu, Z. Li, T. Miyazaki, M. Furuya, T.A. Kudo, M. Kawashita, TiO 2 microspheres containing magnetic nanoparticles for intra-arterial hyperthermia, J. Biomed. Mater. Res. B Appl. Biomater. 105 (2017) 2308-2314. [12] L. Shang, K. Nienhaus, G.U. Hienhaus, Engineered nanoparticles interacting with cells: size matters, J. Nanobiotechnology 12 (2014) 5. [13] G.D. Venkatasubbu, S. Ramasamy, V. Ramakrishnan, J. Kumar, Folate-targeted PEGylated titanium dioxide nanoparticles as a nanocarrier for targeted paclitaxel drug delivery, Adv. Powder Technol. 24 (2013) 947-954. [14] T. Lopez, A. Kozina, E. Ortiz-Islas, K.A. Espinoza, R. Gonzalez, In-situ release of antiepileptic drugs from nanostructured reservoirs, in: F.S. Kaneez (Ed.), Underlying Mechanisms of Epilepsy, InTech, Croatia, 2011, pp. 333-346. [15] M. Jarosz, J. Kapusta-Kolodziej, A. Pawlik, K. Syrek, G.D. Sulka, Drug delivery systems based on titania nanostructures, in: E. Andronescu, A.M. Grumezescu (Eds.), Nanostructures for Drug Delivery, Elsevier Inc., Netherlands, 2017, pp. 299-326.

Page 31 of 52

[16] A. Watermann, J. Brieger, Mesoporous silica nanoparticles as drug delivery vehicles in cancer, Nanomaterials 7 (2017) 189. [17] C.M. Sayes, R. Wahi, P.A. Kurian, Y. Liu, J.L. West, K.D. Ausman, D.B. Warheit, V.L. Colvin, Correlating nanoscale titania structure with toxicity: a cytotoxicity and inflammatory response study with human dermal fibroblasts and human lung epithelial cells, Toxicol. Sci. 92 (2006) 174-185. [18] V.D. Matteis, M. Cascione, V. Brunetti, C.C. Toma, R. Rinaldi, Toxicity assessment of anatase and rutile titanium dioxide nanoparticles: the role of degradation in different pH conditions and light exposure, Toxicol. In Vitro 37 (2016) 201-210. [19] M.A. Maurer-Jones, Y.S. Lin, C.L. Haynes, Functional assessment of metal oxide nanoparticle toxicity in immune cells, ACS Nano 4 (2010) 3363-3373. [20] M. Yin, E. Ju, Z. Chen, Z. Li, J. Ren, X. Qu, Upconverting nanoparticles with a mesoporous TiO2 shell for near-infrared-triggered drug delivery and synergistic targeted cancer therapy, Chem. Eur. J. 20 (2014) 14012-14017. [21] A.A. Ismail, D.W. Bahnemann, Mesoporous titania photocatalysts: preparation, characterization and reaction mechanisms, Mater. Chem. 21 (2011) 11686. [22] Y. Wang, L. Chen, P. Liu, Biocompatible triplex Ag@SiO2 @mTiO2 core-shell nanoparticles for simultaneous fluorescence-SERS bimodal imaging and drug delivery, Chemistry 18 (2012) 5935-5943. [23] P.P. Qiu, W. Li, B. Thokchom, B. Park, M.C. Cui, D.Y. Zhao, J. Khim, Uniform core-shell structured magnetic mesoporous TiO2 nanospheres as a highly efficient and stable sonocatalyst for the degradation of bisphenol-A, J. Mater. Chem. A 3 (2015) 6492-6500.

Page 32 of 52

[24] B.Y. Guan, L. Yu, J. Li, X.W. Lou, A universal cooperative assembly-directed method for coating of mesoporous TiO(2) nanoshells with enhanced lithium storage properties, Sci. Adv. 2 (2016) e1501554. [25] J. Zhao, P. Xu, Y. Li, J. Wu, J. Xue, Q. Zhu, X. Lu, W. Ni, Direct coating of mesoporous titania on CTAB-capped gold nanorods, Nanoscale 8 (2016) 5417-5421. [26] P. Sikoraa, K. Cendrowski, A. Markowska-Szczupak, E. Horszczaruk, E. Mijowska, The effects of silica/titania nanocomposite on the mechanical and bactericidal properties of cement mortars, Constr. Build. Mater. 150 (2017) 738-746. [27] N. Zhang, B. Li, S. Li, S. Yang, Graphene-supported mesoporous titania nanosheets for efficient photodegeneration, J. Colloid Interface Sci. 505 (2017) 711-718. [28] B. Gupta, B.K. Poudel, T.H. Tran, R. Pradhan, H.J. Cho, J.H. Jeong, B.S. Shin, H.G. Choi, C.S. Yong, J.O. Kim, Modulation of pharmacokinetic and cytotoxicity profile of imatinib base by employing optimized nanostructured lipid carriers, Pharm. Res. 32 (2015) 2912-2927. [29] B.K. Poudel, B. Gupta, T. Ramasamy, R.K. Thapa, S. Pathak, K.T. Oh, J.H. Jeong, H.G. Choi, C.S. Yong, J.O. Kim, PEGylated thermosensitive lipid-coated hollow gold nanoshells for effective combinational chemo-photothermal therapy of pancreatic cancer, Colloids Surf. B Biointerfaces 160 (2017) 73-83. [30] L.A. Torre, F. Bray, R.L. Seigel, J. Ferlay, J. Lortet-Tieulent, A. Jemal, Global cancer statistics, 2012, Ca-Cancer J. Clin. 65 (2015) 87-108. [31] F. Ravar, E. Saadat, M. Gholami, P. Dehghankelishadi, M. Mahdavi, S. Azami, F.A. Dorkoosh, Hyaluronic acid-coated liposomes for targeted delivery of paclitaxel, in-vitro characterization and in-vivo evaluation, J. Control. Release 229 (2016) 10-22.

Page 33 of 52

[32] B. Alberts, A. Johnson, J. Lewis, M. Raff, K. Roberts, P. Walter, Molecular Biology of the Cell, fourth ed., Garland Science, New York, 2002. [33] R.E. Eliaz, F.C. Jr Szoka, Liposome-encapsulated doxorubicin targeted to CD44: a strategy to kill CD44-overexpressing tumor cells, Cancer Res. 61 (2001) 2592-2601. [34] M.Y. Lee, J.A. Yang, H.S. Jung, S. Beack, J.E. Choi, W. Hur, H. Koo, K. Kim, S.K. Yoon, S.K. Hahn, Hyaluronic acid-gold nanoparticle/interferon α complex for targeted treatment of hepatitis C virus infection, ACS Nano 6 (2012) 9522-9531. [35] X. Zhang, M. Yao, M. Chen, L. Li, C. Dong, Y. Hou, H. Zhao, B. Jia, F. Wang, Hyaluronic acid-coated silver nanoparticles as a nanoplatform for in vivo imaging applications, ACS Appl. Mater. Interfaces 8 (2016) 25650-25653. [36] W. Stober, A. Fink, E. Bohn, Controlled growth of monodisperse silica spheres in the micron size range, J. Colloid Interface Sci. 26 (1968) 62-69. [37] J.B. Joo, I. Lee, M. Dahl, G.D. Moon, F. Zaera, Y. Yin, Controllable synthesis of mesoporous TiO2 hollow shells: toward an efficient photocatalyst, Adv. Funct. Mater. 23 (2013) 4246-4254. [38] B. Gupta, S. Pathak, B.K. Poudel, S. Regmi, H.B. Ruttala, M. Gautam, J.S. Lee, J.H. Jeong, H.G. Choi, C.S. Yong, J.O. Kim, Folate receptor-targeted hybrid lipid-core nanocapsules for sequential delivery of doxorubicin and tanespimycin, Colloids Surf. B Biointerfaces 155 (2017) 83-92. [39] B. Gupta, T. Ramasamy, B.K. Poudel, S. Pathak, S. Regmi, J.Y. Choi, Y. Son, R.K. Thapa, J.H. Jeong, J.R. Kim, H.G. Choi, C.S. Yong, J.O. Kim, Development of bioactive PEGylated nanostructured platforms for sequential delivery of doxorubicin and imatinib to overcome drug resistance in metastatic tumors, ACS Appl. Mater. Interfaces 9 (2017) 9280-9290.

Page 34 of 52

[40] K. Bourikas, T. Hiemstra, W.H. Van Riemsdijk, Ion pair formation and primary charging behavior of titanium oxide (anatase and rutile), Langmuir 17 (2001) 749-756. [41] S. Kasar, S. Kumar, A.S. Kar, S.V. Godbole, B.S. Tomar, Sorption of Eu(III) by amorphous titania, anatase and rutle: denticity difference in surface complexes, Colloids Surf. A Physicochem. Eng. Asp. 434 (2013) 72-77. [42] P.Z. Araujo, V. Luca, P.B. Bozzano, H.L. Bianchi, Aerosol-assisted production of mesoporous titania microspheres with enhanced photocatalytic activity: the basis of an improved process, ACS Appl. Mater. Interfaces 2 (2010) 1663-1673. [43] D. Missirliss, R. Kawamura, N. Tirelli, J.A. Hubbell, Doxorubicin encapsulated and diffusional release from stable, polymeric, hydrogel nanoparticles, Eur. J. Pharm. Sci. 29 (2006) 120-129. [44] S. Shahabi, S. Doscher, T. Bollhorst, L. Treccani, M. Maas, R. Dringen, K. Rezwan, Enhancing cellular uptake and doxorubicin delivery of mesoporous silica nanoparticles via surface functionalization: effects of serum, ACS Appl. Mater. Interfaces 7 (2015) 26880-26891. [45] M. Vallet-Regi, F. Balas, D. Arcos, Mesoporous materials for drug delivery, Angew Chem Int Ed. 46 (2007) 7548-7558. [46] H. Xiong, D. Zhou, Y. Qi, Z. Zhang, Z. Xie, X. Chen, X. Jing, F. Meng, Y. Huang, Doxorubicin-loaded carborane-conjugated polymeric nanoparticles as delivery system for combination cancer therapy, Biomacromolecules 16 (2015) 3980-3988. [47] T. Ramasamy, H.B. Ruttala, P. Sundaramoorthy, B.K. Poudel, Y.S. Youn, S.K. Ku, H.G. Choi, C.S. Yong, J.O. Kim, Multimodal selenium nanoshell-capped Au@mSiO2 nanoplatform for NIR-responsive chemo-photothermal therapy against metastatic breast cancer, NPG Asia Mater. 10 (2018) 197-216.

Page 35 of 52

[48] K. Suttiponparnit, J. Jiang, M. Sahu, S. Suvachittanont, T. Charinpanitkul, P. Biswas, Role of surface area, primary particle size, and crystal phase on titanium dioxide nanoparticle dispersion properties, Nanoscale Res. Lett. 6 (2011) 27. [49] D. Patra, F. Sleem, A new method for pH triggered curcumin release by applying poly(Llysine) mediated nanoparticle-congregation, Anal. Chim. Acta 795 (2013) 60-68. [50] B. Gupta, B.K. Poudel, S. Pathak, J.W. Tak, H.H. Lee, J.H. Jeong, H.G. Choi, C.S. Yong, J.O. Kim, Effects of formulation variables on the particle size and drug encapsulation of imatinib-loaded solid lipid nanoparticles, AAPS PharmSciTech 17 (2016) 652-662. [51] M. Fisichella, F. Berenguer, G. Steinmetz, M. Auffan, J. Rose, O. Prat, Intestinal toxicity evaluation of TiO2 degraded surface-treated nanoparticles: a combined physico-chemical and toxicogenomics approach in caco-2 cells, Part. Fibre Toxicol. 9 (2012) 18. [52] K.M. Jan, S. Chien, Role of surface electric charge in red blood cells interactions, J. Gen. Physiol. 61 (1973) 638-654. [53] S.H. Hong, Y. Choi, Mesoporous silica-based nanoplatforms for the delivery of photodynamic therapy agents, J. Pharm. Invest. 48 (2018) 3-17. [54] X. Han, Z. Li, J. Sun, C. Luo, L. Li, Y. Liu, Y. Du, S. Qiu, X. Ai, C. Wu, H. Lian, Z. He, Stealth CD44-targeted hyaluronic acid supramolecular nanoassemblies for doxorubicin delivery: probing the effect of uncovalent pegylation degree on cellular uptake and blood long circulation, J. Control. Release 197 (2015) 29-40. [55] Y.H. Choi, H.K. Han, Nanomedicines: current status and future perspectives in aspect of drug delivery and pharmacokinetics, J. Pharm. Invest. 48 (2018) 43-60.

Page 36 of 52

[56] F. Wang, Y.C. Wang, S. Dou, M.H. Xiong, T.M. Sun, J. Wang, Doxorubicin-tethered responsive gold nanoparticles facilitate intracellular drug delivery for overcoming multidrug resistance in cancer cells, ACS Nano 5 (2011) 3679-3692. [57] T. Sun, Y.S. Zhang, B. Pang, D.C. Hyun, M. Yang, Y. Xia, Engineered nanoparticles for drug delivery in cancer therapy, Angew. Chem. Int. Ed. 53 (2014) 12320-12364. [58] B. Gupta, C.S. Yong, J.O. Kim, Solid matrix-based lipid nanoplatforms as carriers for combinational therapeutics in cancer, J. Pharm. Invest. 47 (2017) 461-473. [59] S. Ahn, E. Seo, K. Kim, S.J. Lee, Controlled cellular uptake and drug efficacy of nanotherapeutics, Sci. Rep. 3 (2013) 1997. [60] M.M. Joseph, S.R. Aravind, S.K. George, K.R. Pillai, S. Mini, T.R. Sreelekha, Galactoxyloglucan-modified nanocarriers of doxorubicin for improved tumor-targeted drug delivery with minimal toxicity, J. Biomed. Nanotechnol. 10 (2014) 3253-3268. [61] Z. Amoozgar, Y. Yeo, Recent advances in stealth coating of nanoparticle drug delivery systems, Wiley Interdiscip. Rev. Nanomed. Nanobiotechnol. 4 (2012) 219-233. [62] A.K. Jain, N.K. Swarnakar, M. Das, C. Godugu, R.P. Singh, P.R. Rao, S. Jain, Augmented anticancer efficacy of doxorubicin-loaded polymeric nanoparticles after oral administration in a breast cancer induced animal model, Mol. Pharm. 8 (2011) 1140-1151. [63] Q. Zhu, L. Jia, Z. Gao, C. Wang, H. Jiang, J. Zhang, L. Dong, A tumor environment responsive doxorubicin-loaded nanoparticle for targeted cancer therapy, Mol. Pharm. 11 (2014) 3269-3278. [64] C. Wang, X. Sun, K. Wang, Y. Wang, F. Yang, H. Wang, Breast cancer targeted chemotherapy based on doxorubicin-loaded bombesin peptide modified nanocarriers, Drug Deliv. 23 (2016) 2697-2702.

Page 37 of 52

[65] H.B. Ruttala, T. Ramasamy, B.K. Poudel, Y. Choi, J.Y. Choi, J. Kim, S.K. Ku, H.G. Choi, C.S. Yong, J.O. Kim, Molecularly targeted co-delivery of a histone deacetylase inhibitor and paclitaxel by lipid-protein hybrid nanocapsules for synergistic combinational chemotherapy, Oncotarget 8 (2017) 14925-14940.

Page 38 of 52

Figure legends Figure 1. Schematic illustration of step-wise fabrication of doxorubicin-loaded hyaluronic acidcapped silica-supported mesoporous titania nanoparticles (HA/DOX-MTNst). MTNst, silica-supported mesoporous titania nanoparticles; DOX, doxorubicin; PLL, poly(Llysine); HA, hyaluronic acid. Figure 2. Dynamic light scattering (DLS) characterization. (A) Hydrodynamic particle size and (B) ζ-potential of MTNst, DOX-MTNst, PLL/DOX-MTNst, and HA/DOX-MTNst. (C) Transmission electron microscopy (TEM) images of (a) MTNst and (b) HA/DOXMTNst. Scale bars in (C) represent 100 nm. Figure 3. (A) Energy dispersive spectroscopy (EDS) spectrum and (B) N 2 adsorption-desorption isotherms of MTNst, (C) Thermogravimetric analysis (TGA) thermograms of MTNst, PLL/MTNst, and HA/MTNst, and (D) DSC and (E) XRD scans of DOX, HA/MTNst and HA/DOX-MTNst. Figure 4. (A) In vitro release profiles of doxorubicin (DOX) from doxorubicin-loaded hyaluronic acid-capped silica-supported mesoporous titania nanoparticles (HA/DOXMTNst) in pH 7.4 phosphate-buffered saline (PBS) and pH 5.0 acetate-buffered saline (ABS). (B) Hemolysis assay of DOX, HA/MTNst, and HA/DOX-MTNst at 1, 10, and 100 μg/mL concentrations. Figure 5. In vitro inhibition of MDA-MB-231 cell proliferation following doxorubicin (DOX) and DOX-loaded hyaluronic acid-capped silica-supported mesoporous titania nanoparticles (HA/DOX)-MTNst treatment. The figure inset shows in vitro cytotoxicity of HA/MTNst against MDA-MB-231 cells.

Page 39 of 52

Figure 6. In vitro cellular uptake characteristics. (A) Fluorescence-activated cell sorting (FACS) analysis showing cellular uptake characteristics of HA/DOX-MTNst: (a) concentrationdependent, (b) incubation time-dependent cellular uptake, and (c) relative uptake of DOX and HA/DOX-MTNst at fixed concentration and incubation time (1 µg/mL and 2 h respectively); (B) Confocal images showing cellular uptake characteristics of coumarin6-labeled HA/MTNst with or without pretreatment with hyaluronic acid (HA, 100 μM). Scale bars in (B) represent 5 μm. Figure 7. (A) Hoechst assay and (B) live/dead assay fluorescence images of untreated (control) and DOX- and HA/DOX-MTNst-treated MDA-MB-231 cells. Equivalent drug concentrations in each treatment group was 1 μg/mL with 24 h incubation time. The scale bars in (A) represent 50 μm, while those in (B) represent 100 μm. Figure 8. (A) In vivo biodistribution and (B) ex vivo tissue distribution of Cy5.5-labeled hyaluronic acid-capped silica-supported mesoporous titania nanoparticles (HA/MTNst) in MDA-MB-231 xenograft tumor-bearing mice. Figure 9. In vivo antitumor effect in MDA-MB-231 xenograft tumor-bearing mice. Changes in (A) tumor volume and (B) weight of mice following saline (control), DOX, and HA/DOX-MTNst treatment. The mice received four doses (each equivalent to 5 mg/kg body weight) at 0, 3, 7, and 10 days (pointed out by the arrows). Scale calibrations in the lower panel of (A) are in centimeters (cm). *p < 0.05 and **p < 0.01. Figure 10. (A) Tumor histopathological and immunohistochemical (IHC) analysis in MDA-MB231 xenograft tumor-bearing mice. (a) histopathological images of H&E staining, and IHC images following immunolabeling with (b) apoptosis markers (cleaved-caspase-3, cleaved-poly(ADP-ribose) polymerase (PARP) and (c) angiogenesis (CD31) and

Page 40 of 52

proliferation (Ki-67) markers of control, DOX-treated and DOX-loaded hyaluronic acid-capped silica-supported mesoporous titania nanoparticles (HA/DOX-MTNst) tumors. (B) Histopathological images of principal organs from mice belonging to control, DOX, and HA/DOX-MTNst groups.

Page 41 of 52

Statement of significance This is the first study to use silica-supported mesoporous titania nanoparticles (MTNst) for doxorubicin (DOX) delivery to treat breast cancer, which exhibited effective and enhanced in vitro and in vivo apoptosis and tumor growth inhibition. Solid silica was used to support the mesoporous TiO2 resulting in MTNst, which efficiently incorporated a high DOX payload. The hyaluronic acid (HA) coating over the MTNst surface served a dual purpose of first, stabilizing DOX inside the MTNst (capping agent), and second, directing the nanoplatform device to CD44 receptors that are highly expressed in MDA-MB-231 cells (targeting ligand). The NPs exhibited highly efficacious in vitro tumor-cell killing and excellent in vivo tumor regression, highlighting the enormous promise of this system for breast cancer therapy.

Page 42 of 52

Page 43 of 52

Page 44 of 52

Page 45 of 52

Page 46 of 52

Page 47 of 52

Page 48 of 52

Page 49 of 52

Page 50 of 52

Page 51 of 52

Page 52 of 52