Drug delivery with upconversion nanoparticles for multi-functional targeted cancer cell imaging and therapy

Biomaterials 32 (2011) 1110e1120

Contents lists available at ScienceDirect

Biomaterials journal homepage: www.elsevier.com/locate/biomaterials

Drug delivery with upconversion nanoparticles for multi-functional targeted cancer cell imaging and therapy Chao Wang, Liang Cheng, Zhuang Liu* Jiangsu Key Laboratory for Carbon-Based Functional Materials & Devices, Institute of Functional Nano & Soft Materials (FUNSOM), Soochow University, Suzhou, Jiangsu 215123, China

a r t i c l e i n f o

a b s t r a c t

Article history: Received 24 August 2010 Accepted 30 September 2010 Available online 20 October 2010

Upconversion nanoparticles (UCNPs) with unique multi-photon excitation photoluminescence properties have recently been intensively explored as novel contrast agents for low-backgroundbiomedical imaging. In this work, we functionalize UCNPs with a polyethylene glycol (PEG) grafted amphiphilic polymer. The PEGylated UCNPs are loaded with a commonly used chemotherapy molecule, doxorubicin (DOX), by simple physical adsorption via a supramolecular chemistry approach for intracellular drug delivery. The loading and releasing of DOX from UCNPs are controlled by varying pH, with an increased drug dissociation rate in acidic environment, favorable for controlled drug release. Upconversion luminescence (UCL) imaging by a modified laser scanning confocal microscope reveals the time course of intracellular delivery of DOX by UCNPs. It is found that DOX is shuttled into cells by the UCNP nano carrier and released inside cells after endocytosis. By conjugating nanoparticles with folic acid, which targets folate receptors over expressed on various types of cancer cells, we further demonstrate targeted drug delivery and UCL cell imaging with UCNPs. Besides DOX, this non-covalent drug loading strategy can also be used for loading of photosensitizer molecules on UCNPs for potential near-infrared light induced photodynamic therapy. Our results suggest the promise of UCNPs as interesting nano carriers for multi-functional cancer therapy and imaging. Ó 2010 Elsevier Ltd. All rights reserved.

Keywords: Upconversion nanoparticles Doxorubicin Folate targeting Drug delivery Imaging

1. Introduction In recent years, upconversion nanoparticles (UCNPs), typically lanthanide (Ln3þ) -doped nanocrystals, have attracted significant interest in many fields including biomedicine [1e6]. Under nearinfrared (NIR) excitation, UCNPs emit visible to NIR light with shorter wavelengths, affording them unique optical characteristics such as minimized auto-fluorescence background, resistance to photobleaching, as well as increased light penetration depth in biological tissues, all of those are particularly useful for biomedical detection and imaging [7e11]. Recently, the upconversion luminescence (UCL) imaging technique has been widely used for imaging of cellular specimens [3,5,10,12,13]. Yu et al. [10] have demonstrated that the UCL-based visualization technique has negligible signal fading over time, implying the great ability of UCNPs for long-period observation of cells. Hu et al. [12] have developed UCNPs incorporated with FITC dyes and conjugated with folic acid (FA) targeting molecules for targeted imaging of tumor cells. Zako et al. [13] used

* Corresponding author. E-mail address: [email protected] (Z. Liu). 0142-9612/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.biomaterials.2010.09.069

UCNPs modified with cyclic arginineeglycineeasparatic acid (RGD) peptide for specific cell binding and imaging studies. In vivo biomedical imaging using UCNPs have also been reported by a number of groups [14e19]. By linking FA and RGD to UCNPs, Li and co-workers have achieved efficient in vivo tumor targeting and imaging in two separated studies [17,18]. The UCL imaging could be further combined with other imaging functions such as magnetic resonance (MR) imaging for multimodality biomedical imaging [19,20]. The in vivo long-term biodistribution has also been studied in a recent work, showing that UCNPs with small diameters (w10 nm) were excreted from mice after intravenously administration, without causing noticeable toxicity to the treated animals [21]. In a latest study by our group, we compared the in vivo imaging sensitivities of UCNPs with fluorescent quantum dots (QDs) side by side, and found the detection limit of UCNPs to be at least one order of magnitude lower than that of QDs in our current non-optimized in vivo imaging system, owing to the absence of auto-fluorescence background in UCL imaging [11]. Multi-color multiplexed in vivo UCL imaging was also realized in this work, using UCNPs with varied Ln3þ doping. Therapeutic applications of UCNPs have also been studied in the past few years, mainly focused on the photodynamic therapy (PDT)

C. Wang et al. / Biomaterials 32 (2011) 1110e1120

a

b

1111

c UCNPDOX

Centrifugation Sonication

d

0.2

Absorbance

Free DOX

0.15

0.05 500 700 Wavelength (nm)

900

2.0E+07 Free DOX,5µM UCNP-DOX,5µM

1.5E+07 1.0E+07 5.0E+06 0.0E+00 500

550

600 650 Wavelength(nm)

700

f intensity(a.u)

Intensity (a.u.)

Polymer +DOX

0.1

0 300

e

UCNP UCNP-DOX DOX

3.0E+05 2.0E+05

UCNP UCNP-DOX

1.0E+05 0.0E+00 450

650 Wavelength (nm)

850

Fig. 1. Preparation and characterization of the UCNP-DOX complex. (a) TEM imaging of PEGlated UCNPs in water (NaYF4: 78% Y, 20% Yb, 2% Er nanocrystals). (b) Photos of the PEGcoated UCNP (upper) and UCNP-DOX (bottom) in aqueous solutions under the ambient light and under a 980 nm laser excitation. (c) Photos of UCNP-DOX, Free DOX, and polymerDOX solutions after centrifugation and then sonication. DOX was adsorbed by UCNPs, which were precipitated after centrifugation and then re-suspended after a brief sonication. (d) UVeVIS absorbance spectra of free DOX, bare UCNP, and UCNP-DOX solutions. (e) Fluorescence spectra of free DOX and UCNP-DOX solutions with the same DOX concentration (5 mM) under 480 nm excitation. (f) Upconversion emission spectra of UCNP and UCNP-DOX solutions with the same UCNP concentration (0.2 mg/ml) under 980 nm excitation.

relying on the resonance energy transfer from UCNPs to PDT photosensitive molecules under the NIR light irradiation [6,22,23]. However, to our best knowledge, delivery of chemotherapy drugs with UCNPs has been rarely explored. Herein, we load PEGylated UCNPs with a commonly used chemotherapy molecule, doxorubicin (DOX), via a hydrophobic interaction based supramolecular chemistry strategy, for targeted intracellular drug delivery and UCL imaging. Our work reveals the great potential of UCNPs for multifunctional drug delivery and biomedical imaging applications. 2. Methods and materials 2.1. Preparation of materials 2.1.1. Synthesis of UCNPs The UCNP synthesis was carried out following a literature protocol with slight modifications [24]. Y2O3, Yb2O3, Er2O3 were purchased from Shanghai Chemical Industrial Co. and used as starting materials without further purification. All the rare-earth trifluoroacetates were prepared by dissolving the respective rare-earth oxides in trifluoroacetic acid (CF3COOH, Shanghai Chemical Industrial Co.). Oleic acid (OA, 90%) and 1-Octadecene (ODE >90%) were purchased from SigmaeAldrich. Typically, 1 mmol of Re (CF3COO)3 (Y: Yb: Er ¼ 78%: 20%: 2%); 20 mmol of NaF and 20 ml solvent (10 ml OA/10 ml ODE) were brought to a 100 ml three-necked flask simultaneously and degassed at 100  C for 1 h under vacuum. In the presence of

nitrogen, the mixture was rapidly heated to 320  C and kept at this temperature for 30 min under vigorous magnetic stirring. After cooling down to room temperature, the products were precipitated by addition of ethanol, separated by centrifugation, washed by cyclohexane, and then washed three times with ethanol. The nanoparticle products could be easily re-dispersed in chloroform. 2.1.2. Synthesis of C18PMH-PEG C18PMH-PEG was synthesized following a literature procedure [25]. 10 mg (1 eq) Poly(maleic anhydride-alt-1-octadecene) (C18PMH, SigmaeAldrich) and 143 mg (1 eq) mPEG-NH2(5 K) (PegBio, Suzhou, China) were dissolved in 5 ml dichloromethane with 6 ml triethylamine (TEA, Sinopharm Chemical Reagent Co.) and 11 mg 1-Ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC, Fluka) added. After 24 h of stirring, the dichloromethane solvent was blowed dry by N2. The leftover solid was dissolved in water, forming a transparent clear solution, which was dialyzed against distilled water for 2 days in a dialysis bag with molecular weight cut-off (MWCO) of 14 kDa to remove unreacted mPEG-NH2. After lyophilization, the final product in a white solid was stored at 20  C for future use. 1H NMR (400 MHz, CDCl3) d: 3.8e3.5 ppm (m, br, CH2 of mPEG), 1.3e1.1 ppm (m, CH2 of C18 chains), 0.88 ppm (m, br, CH3 of PMHC18). The actual PEGylation ratio on the C18PMH polymer was measured by NMR spectrum to be w89%. 2.1.3. Synthesis of C18PMH-PEG-FA 10 mg (1 eq) C18PMH, 143 mg (1 eq) mPEG-NH2(5k), and 50 mg (0.5 eq) NH2mPEG(5k)-BOC (Polymere, Germany) were mixed together in dichloromethane under agitation to form a homogeneous solution, EDC(2 eq) and TEA(8 eq) were then added under magentic stirring. After stirring for 24 h at room tempeture, the

1112

C. Wang et al. / Biomaterials 32 (2011) 1110e1120

a

c

b

d

H+

PEGylation Doxorubicin

* Coating polymer: C18PMH-PEG HOOC O

O

* mHN

OH COCH2OH OH

H OCH3 O OH H 3C O HO H 2N

H 2N

O

Doxorubicin (DOX)

O n

O

N

N

HN N

HN

O

HO HN

O

O HO Folic Acid (FA) O

H

Scheme 1. Schematic illustration of the UCNP-based drug delivery system. (a) As-synthesized oleic acid capped UCNPs; (b) C18PMH-PEG-FA functionalized UCNPs; (c) DOX loading on UCNPs. DOX molecules are physically adsorbed into the oleic acid layer on the nanoparticle surface by hydrophobic interactions; (d) Release of DOX from UCNPs triggered by decreasing pH.

dichloromethane solvent was blowed dry by N2. Subsequently, 2 ml trifluoroacetic acid (TFA, Sinopharm Chemical Reagent Co.) was added under magentic stirring 3 h at room tempeture to de-protect the Boc group. After evaporating the TFA solvent, the leftover solid was dissolved in water, which was dialyzed for 2 days in a dialysis bag (MWCO ¼ 14 kDa) to remove unreacted PEG polymers and other reagents. After lyophilization, the final product (C18PMH-PEG-NH2) in white solid was stored at 20  C for future use.

UCNP UCNP-DOX,pH5 UCNP-DOX,pH7.4 UCNP-DOX,pH8

0.15 0.1 0.05 0 300

500 700 Wavelength (nm)

10 8 6 4 2 0

900

c Loading Capacity(w/w%)

Loading Capacity(w/w%)

b

0.2

UCNPDOX,pH5

UCNPDOX,pH7.4

UCNPDOX,pH8

d 8 6 4 2 0

0

50

100 150 200 250 Doxorubicin (uM)

300

Percentage of Released DOX from UCNP (%)

Absorbance

a

The folic acid conjugated C18PMH-PEG (C18PMH-PEG-FA) was prepared by conjugating the amine-functionalized C18PMH-PEG-NH2 with activated FA. Briefly, 35 mg of FA was mixed with 15 mg EDC and 23 mg NHS in 5 ml anhydrous dimethyl sulfoxide (DMSO, SigmaeAldrich) for 15 min at room temperature. 20 mg of C18PMH-PEG-NH2 in 5 ml DMSO was added afterwards (molar ratio of NH2/FA/ EDC/NHS ¼ 1:2:2:5). After reaction at room temperature under stirring for 8 h, water was added and the product was purified by dialysis. The final product was

100 pH7.4 pH5

80 60 40 20 0

0

100

200 300 Time (min)

700

800

Fig. 2. Drug loading and releasing in the UCNP-DOX system. (a) UVeVIS absorbance spectra of UCNPs loaded with DOX at different loading pH values. (b) Quantification of DOX loading at three pH. Higher drug loaded was achieved at increased pH. (c) Quantification of DOX loading at different DOX concentrations (loading pH ¼ 8). UCNP solutions with the same concentration (0.2 mg/ml) were used in this experiment. A maximal loading of 8% by weight was obtained in our UCNP-DOX system. (d) DOX release from UCNPs over time in buffers at two pH values indicated. Significantly accelerated drug release was observed at a slightly acidic pH. Error bars were based on standard deviations of triplicated samples.

C. Wang et al. / Biomaterials 32 (2011) 1110e1120

1113

Fig. 3. Laser scanning confocal microscopy (LSCM) images of HeLa cells incubated with (a) UCNP-DOX ([DOX] ¼ 2 mM) and (b) free DOX (DOX ¼ 2 mM) for 30 min, 2 h, 6 h and 12 h at 37  C. UCL emissions from UCNPs (green colored) and DOX fluorescence (red colored) were recorded in the wavelength ranges of 500e700 nm and 500e600 nm, under 980 nm and 488 nm laser excitations, respectively. All images were taken under the identical instrumental conditions and presented at the same intensity scale. lyophilized and stored at 20  C until use. 1H NMR (400 MHz, CDCl3), d: 7.73 ppm (s, NH in the aromatic ring of FA), d: 7.53 ppm (d, CH in the aromatic ring of FA), d: 6.98 ppm (d, CH in the aromatic ring of FA), d: 3.8e3.5 ppm (m, br, CH2 of mPEG), 1.3e1.1 ppm (m, CH2 of C18 chains), 0.88 ppm (m, br, CH3 of PMHC18). 2.1.4. PEG functionalization of UCNPs 500 ml stock solution of UCNPs was precipitated by centrifuge. The nanoparticles were washed twice with ethanol and dispersed in chloroform. Another solution of 5 mg C18PMH-PEG or C18PMH-PEG-FA polymer in 2 ml chloroform was then added. The mixture was then stirred for 2 h. After blowing-dry chloroform, the residue was readily dissolved in water. The resultant solution was filtered through a 0.22-mm syringe filter to remove large aggregates. 2.2. Drug loading and releasing 2.2.1. Doxorubicin loading on PEGylated UCNPs Doxorubicin (DOX, BeJing HuaFeng United Technology CO.,Ltd) loading onto PEGylated UCNPs was done by mixing DOX (100 mM) with PEGylated UCNPs (0.2 mg/ ml) in phosphate buffer solution (PBS, 20 mM) with varied pH overnight. For the DOX loading saturation experiment, different concentrations of DOX were mixed with UCNPs (0.2 mg/ml) in PBS (20 mM, pH ¼ 8) overnight. Free DOX was removed by centrifugation at 14800 rpm for 10 min. The supernatant was discarded and the precipitate (UCNP-DOX) was washed 3 times with PBS by centrifugation. The formed complexes were re-suspended by a brief sonication to form a homogeneous clear solution and stored at 4  C. Loading of DOX on FA conjugated UCNPs was carried out by the same protocol with similar drug loading efficiencies achieved. 2.2.2. Characterization of UCNP-DOX UV-vis-NIR absorbance spectra of the UCNPs-DOX were measured by UV765 (produecd by ShangHai Precision&Scientific Instrument Co.,Ltd). The concentrations of DOX loaded on UCNPs were determined by the DOX characteristic peak at 480 nm after subtracting the absorbance contributed by functionalized UCNPs at the same

wavelength, similar to the measurement of DOX loading on functionalized singlewalled carbon nanotubes (SWNTs)[26,27]. Fluorescence spectrum of the UCNPs and UCNP-DOX was measured by using a FluoroMax 4 fluorometer. The UCL spectra of UCNPs were recorded by inducing a 2 W continuous-wave (CW) 980 nm diode laser as the excitation source. 2.2.3. DOX release from UCNPs The UCNP-DOX solutions were incubated in PBS at pH of 5 and 7.4 for different periods of time. DOX released from UCNPs was collected by centrifugation at 14800 rpm for 10 min. The amounts of released DOX in the supernatant solutions were measured by UV-vis-NIR absorbance spectrum. 2.2.4. Loading of photosensitizer molecules Porphyrin derivatives Chlorin e6 (Ce6) and meso-Tetra (4-carboxyphenyl) Porphine (TCPP) were purchased from Frontier Scientific Inc. and dissolved in DMSO at 5 mM as stock solutions for further use. For the molecular loading, PEGylated UCNPs (0.2 mg/ml) in water was mixed with Ce6 or TCPP at final concentrations of 100 mM overnight. Free Ce6 or TCPP was removed by centrifugation at 14800 rpm for 10 min. The supernatant was discarded and the perciptate (UCNP-Ce6 or UCNP-TCPP) was washed 3 times with water by centrifugation. The formed complexes were re-suspended by a brief sonication to form homogeneous clear solutions and stored at 4  C.

2.3. Cellular experiments 2.3.1. Cell culture A human nasopharyngeal epidermal carcinoma cell line (KB cell) and a human cervical cell line (HeLa cell) were obtained from American Type Culture Collection (ATCC). All cell culture related reagents were purchased from Invitrogen. HeLa cells were grown in normal RPMI - 1640 culture medium with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin. KB cells were cultured in folic acid free RPMI e 1640 supplemented with 10% FBS and 1% penicillin/streptomycin.

1114

C. Wang et al. / Biomaterials 32 (2011) 1110e1120

Fig. 4. LSCM images of HeLa cells at various time points post 2 h incubation with UCNP-DOX. The cells after 2 h of UCNP-DOX incubation ([DOX] ¼ 2 mM) 37  C were washed and transferred into fresh cell medium for further incubation. LSCM images were taken at 0 min, 20 min, 1 h, and 2 h after washing. All images were taken under the identical instrumental conditions and presented at the same intensity scale.

2.3.2. Cytotoxicity assay The in vitro cytotoxicity was measured using a standard methyl thiazolyl tetrazolium (MTT, SigmaeAldrich) assay. HeLa cells were seeded into 96-well cellculture plate at 5  104/well and then incubated for 24 h at 37  C under 5% CO2. After incubating HeLa cells with various concentrations of free DOX, UCNP, and UCNPDOX for 48 h, the standard MTT assay was carried out to determine the cell viabilities relative to the control untreated cells. For the FA targeted DOX delivery experiment, KB cells cultured in FA free RPMI1640 medium with high FR expression were used as positive cells, while HeLa cells cultured in normal RPMI-1640 medium with low FR expression were used as the negative control. Both KB and HeLa were incubated with series concentrations of UCNP-DOX or FA-UCNP-DOX for 1 h first and then changed to fresh medium after washing. After another 24 h incubation, the cell viability test was carried out by MTT assay. 2.3.3. PI/FACS analysis HeLa cells (1  106 cells) were treated with free DOX, UCNPs and UCNPs-DOX for 48 h. After trypsinization, cells were fixed in 5 ml 70% ethanol at 4  C for 4 h. The fixed cells were then washed with PBS twice and then stained with 500 ml propidium iodide (PI) working solution (50 mg/ml) for 30 min before flow cytometry analysis (FACS Calibur from Becton, Dickinson and Company). The collected data were analyzed by using the FlowJo software. 2.3.4. Confocal imaging of cells Confocal UCL imaging of cells was performed using a modified Leica laser scanning confocal microscope, with a CW NIR laser at l ¼ 980 nm as an additional excitation source. Imaging of DOX was carried out under 488 nm laser excitation, while UCNPs was excited by the external laser at 980 nm. The emissions were collected in the ranges of 500e600 nm and 500e700 nm for DOX and UCNPs, respectively. HeLa cells were incubated with UCNPs-DOX [DOX ¼ 10 mM] for 30 min, 2 h, 6 h, and 12 h before Confocal fluorescence and UCL imaging. For FA targeted cell imaging,

both positive and negative cells were incubated with UCNP-DOX and FA-UCNP-DOX for 30 min before imaging. All cells were washed twice with cell culture medium before confocal imaging.

3. Results and discussion 3.1. Drug loading and releasing on UCNPs Yb and Er doped NaYF4 UCNPs (Y : Yb: Er ¼ 78% : 20% : 2%) were synthesized following a literature procedure with slight modifications [24]. The average diameters of UCNPs were measured to be w30 nm by transmission electron microscopy (TEM) (Fig. 1a). X-ray diffraction (XRD) analysis showed that hexagonal NaYF4: Yb, Er nanocrystals with high purity were obtained (Supporting Information, Fig. S1). The as-prepared UCNPs were capped by oleic acid and not water soluble. A PEG drafted amphiphilic polymer (C18PMH-PEG) was used to transfer hydrophobic UCNPs into the aqueous phase, yielding PEGylated UCNPs with excellent water solubility. Successful PEGylation was also evidenced by infrared spectra and thermogravimetric analysis (Supporting Information, Fig. S2&S3). PEGylated UCNPs prepared by our method has a hydrophobic oleic acid layer on top of the inorganic nanoparticle surface and beneath the PEG coating. We hypothesized that lipophilic molecules such as deprotonated doxorubicin could be adsorbed into this

C. Wang et al. / Biomaterials 32 (2011) 1110e1120

a

1115

b 500 600 Counts

Counts

400 400

300 200

200 100

1. 200

c

400 600 800 PI Intensity (a.u.)

0.

1000

200

400 600 800 PI Intensity (a.u.)

1000

200

400 600 800 PI Intensity (a.u.)

1000

d

200

400

Counts

Counts

150

100

50

29. 29.

300

200 100

200

400 600 800 PI Intensity (a.u.)

65. 65.

1000

Viable cell percentage

e Free DOX 100%

UCNP-DOX

50%

0%

1

10 Doxorubicin(uM)

100

Fig. 5. In vitro cell toxicity tests. (a-d) FACS date of PI stained untreated HeLa cells (a), and cells treated with plain UCNP (b), UCNP-DOX ([DOX ¼ 10 mM]) (c), and free DOX ([DOX ¼ 10 mM]) (d). The percentages of apoptotic cells of cell samples treated with UCNP-DOX and free DOX were 30.0  9.2% and 66.3  2.3%, respectively, compared to less than 1% for the untreated and plain UCNP treated cells. (e) Concentration-dependent cell survival data of HeLa cells treated with free DOX and UCNP-DOX. UCNP-DOX showed a higher IC50 value compared with that of free DOX. Error bars were based on standard deviations of triplicated samples.

layer by hydrophobic interactions (Scheme 1), and may be released in acidic pH upon protonation of DOX. The terminals on PEG chains are available for conjugation of targeting ligands such as folate acid for targeted delivery. To test this hypothesis, we mixed PEGylated UCNPs with DOX at pH 8 overnight. Free DOX was completely removed by centrifugation at 14800 rpm for 10 min and 3 times PBS washing until no visible color was noticeable in the supernatant. The DOX loaded UCNPs (UCNP-DOX) was readily dispersed in water after brief sonication, forming a clear transparent solution with

reddish color (Fig. 1b), indicating DOX binding on UCNPs. To confirm DOX was indeed loaded on UCNPs instead being encapsulated by the amphiphilic coating polymer C18PMH-PEG, solutions of free DOX, UCNPs mixed with DOX, and C18PMH-PEG polymer mixed with DOX at the same pH (pH ¼ 8) were prepared and centrifuged for 10 min (Fig. 1c). Reddish precipitate and nearly colorless supernatant were observed after the mixture of UCNP and DOX was centrifuged, suggesting the binding of DOX on nanoparticles which were pulled down by the centrifugation force. No

1116

C. Wang et al. / Biomaterials 32 (2011) 1110e1120

Fig. 6. LSCM images of (a) KB cells incubated with FA-UCNP-DOX, (b) KB cells incubated with UCNP-DOX, (c) KB cells blocked with free FA first and then incubated with FA-UCNPDOX, (d) HeLa cells incubated with FA-UCNP-DOX, and (e) HeLa cells incubated with UCNP-DOX. Incubation was carried out at 37  C for 30 min. The concentrations of UCNPs and DOX were 0.2 mg/ml and 2 mM, respectively, in all samples. All images were taken under the identical instrumental conditions and presented at the same intensity scale.

Viable Cell Percentage

a

** 100%

**

UCNP-DOX FA-UCNP-DOX

*

50%

0%

3.125

6.25 12.5 Doxorubicin(uM)

25

was observed (Fig. 1e), likely due to the intermolecular interactions between DOX molecules once they were densely packed on the nanoparticle surface. The UCL spectra of UCNPs and UCNP-DOX were also measured using a 980 nm laser as the excitation light (Fig. 1f). The UCL emission band from 500 to 560 nm decreased by w50% while the peak at w660 nm was essentially unchanged in the

b Viable Cell Percentage

precipitate or color change was noted for free DOX and C18PMHPEG þ DOX sample after centrifugation. The concentration of DOX in the UCNP-DOX sample could be determined by the characteristic DOX absorption peak at w480 nm (Fig. 1d). A significant DOX fluorescence quenching effect (nearly 90% of DOX fluorescence was quenched) in the UCNP-DOX sample

100%

FA-UCNP-DOX UCNP-DOX

50%

0%

3.125

6.25 12.5 Doxorubicin(uM)

25

Fig. 7. Concentration-dependent cell survival data of FR positive KB cells (a) and negative HeLa cells (b) treated with FA-UCNP-DOX and UCNP-DOX. Cells were incubated with series concentrations of UCNP-DOX or FA-UCNP-DOX for 1 h, washed two times with fresh cell medium, and then placed into fresh medium for another 24 h of incubation before the cell viability MTT assay. *P < 0.05; **P < 0.01. Error bars were based on standard deviations of four parallel samples at each data point.

C. Wang et al. / Biomaterials 32 (2011) 1110e1120

1117

b

a Centrifugation

UCNPCe6

Sonication

Free Ce6

Free TCPP

Polymer + CE6

Polymer + TCPP

d

UCNP Free Ce6 UCNP-Ce6

3.0 2.5 2.0 1.5 1.0 0.5

7x105

500 600 700 Wavelength(nm)

800

2.0 1.5 1.0

UCNP UCNP-Ce6

6x105 5x105 4x105 3x105 2x105 1x105 500

600 700 800 Wavelength(nm)

400

900

900

f Intensity(a.u.)

Intensity(a.u.)

2.5

0.0 400

0 400

UCNP Free TPCC UCNP-TPCC

0.5

0.0

e

Sonication

3.5 3.0

Absorbance

Absorbance

c

Centrifugation

UCNPTCPP

7x105

500 600 700 Wavelength(nm)

800

900

UCNP UCNP-TCPP

6x105 5x105 4x105 3x105 2x105 1x105 0 400

500

600 700 800 Wavelength(nm)

900

Fig. 8. Loading of photodynamic sensitizer molecules on UCNPs. (a) Photos of UCNP-Ce6, Free Ce6, and polymer-Ce6 solutions after centrifugation and then sonication. (b) Photos of UCNP-TCPP, Free TCPP, and polymer-TCPP solutions after centrifugation and then sonication. Both Ce6 and TCPP were adsorbed by UCNPs, which were precipitated after centrifugation and then re-suspended after a brief sonication. (c) UVeVIS absorbance spectra of free Ce6, bare UCNP, and UCNP-Ce6 solutions. (d) UVeVIS absorbance spectra of free TCPP, bare UCNP, and UCNP-TCPP solutions. (e&f) Comparisons of UCL emission spectra of UCNP versus UCNP-Ce6 (e) or UCNP-TCPP (f) solutions with the same UCNP concentration (0.2 mg/ml) under 980 nm excitation. The UCNP red emission in the UCNP-Ce6 sample significantly decreased due to the resonance energy transfer from UCNPs to Ce6.

1118

C. Wang et al. / Biomaterials 32 (2011) 1110e1120

UCNP-DOX sample compared to bare UCNPs without DOX loading. The change in the UCL spectrum of UCNP-DOX is attributed to the resonance energy transfer from UCNPs to DOX (DOX absorption peak overlaps with the green emission of UCNPs), further evidencing the binding of DOX on UCNPs. This represents a novel partition affording a simple and interesting way to attach drugs to PEGylated UCNPs by non-covalent binding. We next systemically studied the DOX loading and releasing behaviors of UCNP-DOX. UCNP solutions at 0.2 mg/ml were mixed with 150 mM DOX at varied pH for drug loading. It was found that the amount of doxorubicin loaded onto PEGylated UCNPs was pHdependent, a phenomena similar to the DOX loading on singlewalled carbon nanotubes [26,27]. The DOX loading efficienies increased from w1%, w5%ew8% (w/w) as the raise of pH values from 5, 7.4 to 8 in the loading buffers, respectively (Fig. 2a&b). This trend was attributed to the decreased hydrophilicity of DOX at higher pH caused by the deprotonation of the amine group in the DOX molecule, thereby enhancing the hydrophobic interaction between DOX and PEGylated UCNPs. Loading of DOX at even higher pH (e.g. pH ¼ 9) resulted in unstable precipitates which could hardly be re-dispersed in water. To determine the saturation level of DOX loading on UCNPs, we mixed 0.2 mg/ml UCNP solutions with different concentrations of DOX at pH 8. A saturated maximal DOX loading efficiency of 8% (w/w) was uncovered when DOX concentration was above 0.2 mM under the current loading condition. Interestingly, the resonance energy transfer induced UCNP green emission quenching was also found to be DOX loading ratio-dependent (Supporting Information, Fig. S4). The drug releasing profile of UCNP-DOX was tested at neutral and acidic pH, by incubating UCNP-DOX in pH 7.4 and pH 5 phosphate buffers. The released DOX concentrations from UCNPs were determined by UVeVIS measurement of supernatants from UCNPDOX samples after centrifugations at different time points. About 20% of DOX was released from UCNPs in the neutral condition at pH 7.4, while w50% of DOX was released in the acidic solution at pH 5 within 2 h. The protonation of the amino group in the DOX molecule offers DOX a positive charge, weakening its binding to the hydrophobic UCNP surface and triggering drug release. 3.2. Intracellular drug delivery and UCL imaging with UCNP-DOX The upconversion photoluminescence of UCNPs provides a unique means of auto-fluorescence background free optical imaging [3,12]. Using a modified laser scanning confocal fluorescence microscope with an external 980 nm CW laser, we were able to simultaneously image upconversion signals from UCNPs (980 nm excitation) and downconversion fluorescence from DOX (488 nm excitation). HeLa cells were incubated with UCNP-DOX (DOX ¼ 2 mM) for 30 min, 2 h, 6 h, and 12 h at 37  C and washed by fresh cell medium before confocal imaging (Fig. 3). From the images, we observed bright UCL signals first emerged inside cells after 2 h incubation, and increased as the prolonging of incubation time. DOX fluorescence signals were evenly distributed inside cells owing to the intracellular release of DOX molecules from UCNPs. Note that high degree of fluorescence quenching of unreleased DOX on UCNPs may prevent the co-localization of UCNP and DOX signals in the overlaid images. Different from the cellular uptake of UCNP-DOX, free DOX was able to enter cells rapidly with strong DOX fluorescence signals observed even after only 30 min of incubation (Fig. 3b). Spread DOX fluorescence inside cells was observed, suggesting the easy diffusion of DOX across various cell organelles (Fig. 3b). To further understand the drug delivery and releasing behaviors in our UCNP-DOX system, we incubated HeLa cells with UCNP-DOX at 4  C and 37  C for 2 h. Strong UCL signals were detected from cells incubated with UCNP-DOX at 37  C (Fig. 4) but barely any signal

was noted for those at 4  C incubation (Supporting Information Fig. S5), indicating that UCNPs enter cells by the energy dependent endocytosis mechanism [28e30]. Cells were washed after UCNPDOX treatment and placed into fresh medium for further incubation at 37  C, and imaged by the confocal microscope at various time points post washing (Fig. 4). While UCL signals from UCNPs inside cells did not change much since there were no free nanoparticles existing in the cell medium and the quenching of UCNPs by DOX was only partially (w50% for the green emission and no effect to the red emission), the DOX fluorescence inside cells increased remarkably over time (Fig. 4). This interesting phenomenon was likely owing to the release of DOX from UCNPs inside cell endosomes or lysosomes with lower pH. The dissociated DOX exhibited much stronger fluorescence compared with DOX bound to UCNPs, the latter showed a significant fluorescence quenching effect (w90% quenching, Fig. 1e). Once released from UCNPs, DOX molecules would light up and freely diffuse inside cells, showing strong and evenly distributed fluorescent signals as observed from our confocal images (Fig. 4, DOX channel). We then tested the cancer cell killing ability of UCNP-DOX. Cell cycle analysis by propidium iodide (PI) staining and flow cytometry measurement was carried out for HeLa cells incubated with free DOX, plain UCNPs and UCNP-DOX for 48 h (Fig. 5). PEGylated UCNPs was nontoxic to cells (Fig. 5aeb, Support Information Figure S6), while free DOX (DOX ¼ 12 mM) and UCNP-DOX (DOX ¼ 12 mM) induced significant cell apoptosis and death (Fig. 5ced). Standard MTT cell viability assay was also performed to compare the cytotoxicity of UCNP-DOX with free DOX at series of DOX concentrations to HeLa cells (Fig. 5e). The half-maximum inhibitory concentration (IC50) values for UCNP-DOX (12 mM) was found to be higher than that of free DOX (3 mM). The reduced toxicity of UCNPDOX is likely due to the less efficient cellular uptake of UCNP-DOX compared with that of free DOX (Fig. 3), and consistent to previous findings where other nano carriers were involved in DOX delivery [26,31,32]. 3.3. Targeted drug delivery and cancer cell imaging with UCNPs One potential advantage of using UCNPs as drug carriers compared to the use of free drugs is the targeted delivery for selective destruction of certain types of cancer cells. Folate receptor is a widely studied cancer cell marker that up-regulated in many types of cancer cell [33,34]. To prove the FR-mediated targeted delivery, we conjugated FA to PEG terminals of the C18PMH-PEG polymer and used the resulted C18PMH-PEG-FA to functionalize UCNPs. Based on the UVeVIS spectrum of FA-UCNP, it was estimated that w670 FA molecules could be conjugated on one UCNP nanoparticle (Supporting Information, Fig. S7). FA modified PEGylated UCNPs (FA-UCNP) were loaded with DOX under the same loading condition, obtaining a similar drug loading efficiency (w8 % w/w) as that for plain UCNPs. FR-positive KB cells cultured in FA free medium and FR-negative HeLa cells cultured in normal medium were incubated with UCNP-DOX or FA-UCNP-DOX for confocal imaging (Fig. 6). Strong UCL signals were observed for KB cells incubated with FA-UCNP-DOX, while UCNP-DOX incubated cells showed negligible non-specific binding. No obvious UCL signal was noted for FR negative HeLa cells after either F-UCNP-DOX or UCNP-DOX incubation. To further demonstrate the specificity of FRmediated targeting, we blocked FR receptors on KB cells by adding excessive free FA prior to the cell incubation with FA-UCNP-DOX, and observed rather week UCL signals from blocked cells. The above confocal images clearly evidenced highly specific FR targeting by FA conjugated UCNPs. The cytotoxicity of FA-UCNP-DOX in comparison to UCNP-DOX without FA conjugation was evaluated using the MTT assay with KB

C. Wang et al. / Biomaterials 32 (2011) 1110e1120

and HeLa cells involved as FR positive and negative cells, respectively. Compare to UCNP-DOX, FA-UCNP-DOX exhibited higher toxicity to the FR positive KB cells, owing to the enhanced cellular uptake of DOX by FA targeted delivery (Fig. 7a). In the control experiment for FR negative HeLa cells, since FA conjugation on UCNPs gave no obvious enhancement to the intracellular DOX delivery (Fig. 6), similar cytotixicity was observed for UCNP-DOX and FA-UCNP-DOX (Fig. 7b). These results suggest the potential of UCNP conjugates for targeted drug delivery and selective cancer cells killing. 3.4. Generality of molecular loading on UCNPs by supramolecular chemistry Lastly, to test whether our drug loading strategy on UCNPs could be applied for other therapeutic molecules, we incubated two commonly used photosensitizer molecules, porphyrin derivatives Ce6 and TCPP [35], with PEGylated UCNPs in water for molecular loading. Colorless or light-colored supernatants were observed after centrifugation of the UCNP-Ce6 and UCNP-TCPP mixtures, suggesting that significant amounts of these two molecules were adsorbed on UCNPs, which were then pulled down by the centrifugation force (Fig. 8 a&b). Based on the UVeVIS spectra, the saturating loading capacities of 8.6%(w/w) and 7.3%(w/w) were achieved for UCNP-Ce6 and UCNP-TCPP, respectively (Fig. 8 c&d). Obvious decreases in the UCL emission intensities of UCNPs were noticed upon adsorption of these two molecules on nanoparticles, especially for the UCNP red emission (660 nm) after Ce6 loading, owing to the resonance energy transfer from UCNPs to Ce6 and TCPP (Fig. 8 e&f). The UCNP-photosensitizer systems fabricated here by a simple supramolecular chemistry method may potentially be useful for NIR irradiation induced photodynamic therapy. 4. Conclusion In this work, a novel multi-functional drug delivery system based on UCNPs is designed and developed for targeted drug delivery and cell imaging. A widely used anti-cancer drug, DOX, is loaded on PEGylated UCNPs by physical adsorption upon simple mixing. The loading and releasing of DOX from UCNPs are determined by varying pH values, favorable for drug delivery and controlled release to cancer cells. The intrinsic UCL emission from UCNPs and fluorescence from DOX enable dual up- and downconversion optical imaging to track and study the UCNP-DOX drug delivery system. Utilizing a well established folate targeting model, targeted selective cell imaging and drug delivery is achieved with FA conjugated UCNPs. The UCNP drug loading strategy presented here relying on a supramolecular chemistry approach provides a facile and flexible way to load and deliver not only chemotherapeutic molecules such as DOX, but also photosensitizer agents such as Ce6 and TCPP for potential NIR light mediated photodynamic therapy. Our work shows the promise of UCNPs as a novel multifunctional nano-platform for cancer therapy and imaging. Acknowledgement This work was supported by a research start-up fund of Soochow University and a 973 grant (project no. 2010CB934502) from the Ministry of Science and Technology (MOST) of China. Appendix. Supplementary material The supplementary data associated with this article can be found in the online version at doi:10.1016/j.biomaterials.2010.09. 069.

1119

Appendix Figures with essential color discrimination. Figs. 1e8 in this article have parts that are difficult to interpret in black and white. The full color images can be found in the online version, at doi:10. 1016/j.biomaterials.2010.09.069. References [1] Wang F, Han Y, Lim CS, Lu YH, Wang J, Xu J, et al. Simultaneous phase and size control of upconversion nanocrystals through lanthanide doping. Nature 2010;463:1061e5. [2] Song K, Tian LJ, Kong XG, Liu K, Zhang QB, Du C, et al. Preparation, characterization and specific biological labeling of silica coated upconversion fluorescent nanocrystals. Spectrosc Spect Anal 2010;30:133e6. [3] Jiang S, Zhang Y. Upconversion nanoparticle-based FRET system for study of siRNA in live cells. Langmuir 2010;26:6689e94. [4] Zhou HP, Xu CH, Sun W, Yan CH. Clean and flexible modification strategy for carboxyl/aldehyde-functionalized upconversion nanoparticles and their optical applications. Adv Funct Mater 2009;19:3892e900. [5] Wang M, Mi CC, Zhang YX, Liu JL, Li F, Mao CB, et al. NIR-responsive silicacoated NaYbF4:Er/Tm/Ho upconversion fluorescent nanoparticles with tunable emission colors and their applications in immunolabeling and fluorescent imaging of cancer cells. J Phys Chem C 2009;113:19021e7. [6] Qian HS, Guo HC, Ho PCL, Mahendran R, Zhang Y. Mesoporous-silica-coated up-conversion fluorescent nanoparticles for photodynamic therapy. Small 2009;5:2285e90. [7] Mai HX, Zhang YW, Sun LD, Yan CH. Highly efficient multicolor up-conversion emissions and their mechanisms of monodisperse NaYF4: Yb, Er core and core/shell-structured nanocrystals. J Phys Chem C 2007;111:13721e9. [8] Wang X, Li YD. Monodisperse nanocrystals: general synthesis, assembly, and their applications. Chem Commun 2007;28:2901e10. [9] Yi GS, Chow GM. Synthesis of hexagonal-phase NaYF4: Yb, Er and NaYF4: Yb, Tm nanocrystals with efficient up-conversion fluorescence. Adv Funct Mater 2006;16:2324e9. [10] Yu MX, Li FY, Chen ZG, Hu H, Zhan C, Yang H, et al. Laser scanning upconversion luminescence microscopy for imaging cells labeled with rareearth nanophosphors. Anal Chem 2009;81:930e5. [11] Cheng L, Yang K, Zhang S, Shao M, Lee S-T, Liu Z. Highly-sensitive multiplexed in vivo imaging using PEGylated upconversion nanoparticles. Nano Res 2010;3:722e32. [12] Hu H, Xiong LQ, Zhou J, Li FY, Cao TY, Huang CH. Multimodal-luminescence core-shell nanocomposites for targeted imaging of tumor cells. Chem-Eur J 2009;15:3577e84. [13] Zako T, Nagata H, Terada N, Utsumi A, Sakono M, Yohda M, et al. Cyclic RGD peptide-labeled upconversion nanophosphors for tumor cell-targeted imaging. Biochem Bioph Res Co 2009;381:54e8. [14] Nyk M, Kumar R, Ohulchanskyy TY, Bergey EJ, Prasad PN. High contrast in vitro and in vivo photoluminescence bioimaging using near infrared to near infrared up-conversion in Tm3þ and Yb3þ doped fluoride nanophosphors. Nano Lett 2008;8:3834e8. [15] Idris NM, Li ZQ, Ye L, Sim EKW, Mahendran R, Ho PCL, et al. Tracking transplanted cells in live animal using upconversion fluorescent nanoparticles. Biomaterials 2009;30:5104e13. [16] Kobayashi H, Kosaka N, Ogawa M, Morgan NY, Smith PD, Murray CB, et al. In vivo multiple color lymphatic imaging using upconverting nanocrystals. J Mater Chem 2009;19:6481e4. [17] Xiong LQ, Chen ZG, Yu MX, Li FY, Liu C, Huang CH. Synthesis, characterization, and in vivo targeted imaging of amine-functionalized rare-earth up-converting nanophosphors. Biomaterials 2009;30:5592e600. [18] Xiong LQ, Chen ZG, Tian QW, Cao TY, Xu CJ, Li FY. High contrast upconversion luminescence targeted imaging in vivo using peptide-labeled nanophosphors. Anal Chem 2009;81:8687e94. [19] Il Park Y, Kim JH, Lee KT, Jeon KS, Bin Na H, Yu JH, et al. Nonblinking and nonbleaching upconverting nanoparticles as an optical imaging nanoprobe and T1 magnetic resonance imaging contrast agent. Adv Mater 2009;21:4467e71. [20] Zhou J, Sun Y, Du XX, Xiong LQ, Hu H, Li FY. Dual-modality in vivo imaging using rare-earth nanocrystals with near-infrared to near-infrared (NIR-toNIR) upconversion luminescence and magnetic resonance properties. Biomaterials 2010;31:3287e95. [21] Xiong LQ, Yang TS, Yang Y, Xu CJ, Li FY. Long-term in vivo biodistribution imaging and toxicity of polyacrylic acid-coated upconversion nanophosphors. Biomaterials 2010;31:7078e85. [22] Chatterjee DK, Yong Z. Upconverting nanoparticles as nanotransducers for photodynamic therapy in cancer cells. Nanomedicine-UK 2008;3:73e82. [23] Zhang P, Steelant W, Kumar M, Scholfield M. Versatile photosensitizers for photodynamic therapy at infrared excitation. J Am Chem Soc 2007;129:4526e7. [24] Liu C, Wang H, Li X, Chen D. Monodisperse, size-tunable and highly efficient beta-NaYF4:Yb, Er(Tm) up-conversion luminescent nanospheres: controllable synthesis and their surface modifications. J Mater Chem 2009;19:3546e53.

1120

C. Wang et al. / Biomaterials 32 (2011) 1110e1120

[25] Prencipe G, Tabakman SM, Welsher K, Liu Z, Goodwin AP, Zhang L, et al. PEG branched polymer for functionalization of nanomaterials with ultralong blood circulation. J Am Chem Soc 2009;131:4783e7. [26] Liu Z, Sun XM, Nakayama-Ratchford N, Dai HJ. Supramolecular chemistry on watersoluble carbon nanotubes for drug loading and delivery. ACS Nano 2007;1:50e6. [27] Liu Z, Fan AC, Rakhra K, Sherlock S, Goodwin A, Chen XY, et al. Supramolecular stacking of doxorubicin on carbon nanotubes for in vivo cancer therapy. Angew Chem Int Ed 2009;48:7668e72. [28] Kam NWS, Liu ZA, Dai HJ. Carbon nanotubes as intracellular transporters for proteins and DNA: an investigation of the uptake mechanism and pathway. Angew Chem Int Ed 2006;45:577e81. [29] Silverstein SC, Steinman RM, Cohn ZA. Endocytosis. Annu Rev Biochem 1977;46:669e722.

[30] Schmid SL, Carter LL. ATP is required for receptor-mediated endocytosis in intact cells. J Cell Biol 1990;111:2307e18. [31] Sun XM, Liu Z, Welsher K, Robinson JT, Goodwin A, Zaric S, et al. Nano-graphene oxide for cellular imaging and drug delivery. Nano Res 2008;1:203e12. [32] Tang S, Huang X, Chen X, Zheng N. Hollow mesoporous zirconia nanocapsules for drug delivery. Adv Funct Mater 2010;20:2442e7. [33] Hilgenbrink AR, Low PS. Folate receptor-mediated drug targeting: from therapeutics to diagnostics. J Pharm Sci-US 2005;94:2135e46. [34] Wang S, Low PS. Folate-mediated targeting of antineoplastic drugs, imaging agents, and nucleic acids to cancer cells. J Cont Rel 1998;53:39e48. [35] Bechet D, Couleaud P, Frochot C, Viriot ML, Guillemin F, Barberi-Heyob M. Nanoparticles as vehicles for delivery of photodynamic therapy agents. Trends Biotechnol 2008;26:612e21.