Singlet oxygen generation under NIR light and visible light excitations of photosensitizers on upconversion nanoparticle surface

Abstract

In this study a one-pot synthesis of hydrophilic, amine-functionalized NaYF₄(Yb,Er) upconversion nanoparticles is demonstrated. A photosensitizing molecule, chlorin e6, is deposited onto the nanoparticle surface to form a hybrid photosensitizer, which can generate singlet oxygen under both near-infrared and visible light excitations. Effects of beef tissue in the light path under both excitations are investigated and compared.

---

Introduction

Lanthanide-based upconversion nanoparticles (UCNPs) have gained a lot of attention in recent years because of their unique properties of near-infrared (NIR) excitation, large anti-stokes shift, high resistance to photo-oxidation, non-blinking emission, sharp emission bands, and low toxicity.^1,2 These properties make UCNPs suitable for bio-imaging, biosensing, bio-labeling and photodynamic therapy (PDT).^1–14 Over the years, UCNPs are usually synthesized by methods involving high boiling-point organic solvents and temperatures up to as high as 330 °C.^4–11,15,16 UCNPs produced from these methods are hydrophobic, and have to be transferred into aqueous media for biological applications by ligand exchange or by the use of surfactants, making the whole process laborious and time-consuming.^4–11 There have been a few reports on in situ synthesis of polymer-coated UCNPs, which aim to develop more facile synthetic methods while maintaining the quality of UCNPs.^17–19 These methods employ only moderate temperatures, and the resulting polymer coating readily provides functionality on the UCNP surface. Still, there is room for further investigations along this line of research, ideally completing the synthesis of UCNPs in one pot without compromising their photoluminescence properties.

Photodynamic therapy (PDT) is emerging as a possible alternative to the chemotherapy for the treatment of cancer. It is an interesting approach because PDT is immunostimulative while chemotherapy is immunosuppressive.¹⁴ However, limited penetration of visible light required for the excitation of photosensitizing molecules is hindering the use of PDT to treat tumors deep inside the tissue. Reports of UCNPs as successful PDT agents are abundant.^3–11 Typically, photosensitizing molecules, which can be excited in spectral regions matching the emissions of UCNPs, are brought close to the UCNP surface, either through physical adsorption or covalent binding. It has been demonstrated that such hybrid photosensitizers (UCNP plus photosensitizing molecules) can be excited by an NIR light source, facilitated through energy transfer from the UCNPs to the photosensitizing molecules. One would expect that the NIR excitation can provide deep tissue penetration, an important issue in advancing photodynamic therapy. However, within this realm of research, there have been few reports comparing the singlet oxygen generation efficiency under the direct excitation (i.e. visible light) of the photosensitizing molecules on the UCNP surface and the indirect excitation (i.e. NIR light) through UCNPs.⁴ It is important to investigate the efficiency of singlet oxygen generation of both processes in order to confirm that the indirect excitation scheme of hybrid photosensitizers is indeed superior to the direct excitation counterpart.

Herein we report a one-pot synthesis of polyethyleneimine (PEI) coated NaYF₄(Yb,Er) UCNPs. The NaYF₄(Yb,Er)@PEI UCNPs are fairly uniform in size and can be readily dispersed in aqueous media. Combined with the photosensitizing molecule chlorin e6 (Ce6), the resulting hybrid photosensitizer, NaYF₄(Yb,Er)@PEI-Ce6, can generate singlet oxygen under both direct excitation (660 nm) and indirect excitation (980 nm). Efficiencies of singlet oxygen generation under different excitation conditions are investigated and compared.

Experimental

Chemicals and material

Ethylene glycol, yttrium chloride hydrate (YCl₃·6H₂O), ytterbium chloride hydrate (YbCl₃·6H₂O), erbium chloride (ErCl₃·6H₂O), 9-anthracene carboxylic acid (9-ACA), and polyethyleneimine (PEI, MW ∼1300, 50% w/w solution in water) were obtained from Sigma-Aldrich. Sodium chloride (NaCl), ammonium fluoride (NH₄F), and chlorin e6 (Ce6) were ordered from Fisher Scientific. All chemicals were used as received without further purification.

Synthesis of NaYF₄(Yb,Er)@PEI UCNPs

In a typical run, 48.5 mg of YCl₃·6H₂O, 13.9 mg of YbCl₃·6H₂O, 1.1 mg of ErCl₃·6H₂O, 23.4 mg NaCl and 50.0 mg of PEI were added in 3 mL ethylene glycol and the mixture was vortexed for 3 h to form a homogeneous solution (solution 1). Separately, 30.0 mg of NH₄F was dissolved in 2 mL of ethylene glycol in a Teflon container (solution 2). Solution 1 was added to solution 2 dropwise under constant stirring. Mixture of the two solutions was stirred for 20 min. The Teflon container was then sealed tightly and placed inside a stainless steel capsule. The whole assembly was heated at 200 °C for 24 h in an oven. The resulting nanoparticles were centrifuged down at a speed of 15k rpm for 30 min after cooling down to room temperature, and washed 3 times by ethanol and twice by DI water with centrifugation. The final UCNPs were well dispersed in aqueous media.

Deposition of Ce6 on UCNPs

In brief, 45 mg of NaYF₄(Yb,Er)@PEI UCNPs were dispersed in ethanol. 4 mL of 80 ppm Ce6 solution was added to the nanoparticle solution, and the mixture was vortexed for 3 h. The nanoparticles were then washed with ethanol to remove excess Ce6, until the supernatant did not contain free Ce6 as checked by UV-vis absorption spectrometry.

TEM measurement

A drop of NaYF₄(Yb,Er)@PEI UCNPs dispersion was dried on the Formavar covered carbon coated copper grid at room temperature. Images were taken on a Phillips Biotwin 12 transmission electron microscope.

Photoluminescence measurements

All photoluminescence measurements were done on a QM-40 spectrofluorometer (PTI, NJ), equipped with a Xenon lamp and an external 980 nm laser (Laser Glow Technology, Canada) as the excitation sources. A quartz cuvette was used in all measurements. Emission signals were collected using 10 mm as slit width.

Monitoring of singlet oxygen using 9-ACA as probe

Briefly, 1 μM solution of 9-ACA in ethanol was prepared. NaYF₄(Yb,Er)@PEI-Ce6 nanoparticles were dispersed in the 9-ACA solution. Fluorescence of 9-ACA solution was measured before and after irradiation by either the 980 nm laser or the 660 nm output of Xenon lamp. The excitation wavelength used to measure 9-ACA fluorescence was 363 nm.

In experiments deemed to study the penetration depths of the visible and NIR light and their effects on the singlet oxygen generation capacity, a piece of beef tissue (3–5 mm in thickness) was sandwiched between two glass slides, and placed in between the light source and the cuvette so that light would pass through the tissue before reaching the cuvette.

Results and discussion

NaYF₄(Yb,Er)@PEI UCNPs were synthesized by a hydrothermal method, adapted from procedures reported in the literature.¹⁸ Salts of all lanthanides required for the synthesis, along with low molecular-weight PEI as stabilizing agent, were mixed thoroughly in ethylene glycol. After raising the temperature of the mixture to 200 °C for an extended period of time, NaYF₄(Yb,Er)@PEI UCNPs were formed. Under 980 nm excitation, these NaYF₄(Yb,Er)@PEI UCNPs emit green light at ∼550 nm and red light at ∼660 nm. Nanoparticle size of the UCNPs was estimated to be ∼35 nm based on the TEM image (Fig. 1).

Fig. 1 (Top left) Upconversion emission spectrum of UCNPs under 980 nm excitation. (Top right) TEM image of NaYF₄(Yb,Er)@PEI UCNPs. Scale bar is 100 nm. (Bottom) FTIR spectrum of NaYF₄(Yb,Er)@PEI UCNPs.

The chelating capability of PEI to lanthanide ions is well reported in the literature.²⁰ In this case, the presence of PEI on the UCNPs' surface is confirmed by the FTIR spectrum shown in Fig. 1. Peaks at 1576 cm⁻¹ (N–H bending), 1465 cm⁻¹ (C–H bending), and 2940 and 2830 cm⁻¹ (C–H stretching), all indicate the presence of PEI on the nanoparticle surface,²¹ which not only allows the nanoparticles to be stabilized and dispersed in aqueous medium but also provides possibility of functionalization for further applications.^18,20

Synthesis of NaYF₄(Yb,Er)@PEI UCNPs has been reported with ethanol as solvent.¹⁸ It was shown that, when low molecular-weight PEI was used, irregular-shape and large UCNPs were obtained, whereas high molecular-weight PEI would produce NaYF₄(Yb,Er)@PEI UCNPs with stronger green emission than red emission. The difference was likely due to the different chelating ability of the PEIs of different molecular weights, which affects the rate of nucleation and growth of UCNPs and hence their luminescence properties.¹⁸ In this study, ethylene glycol was used as the solvent and low molecular-weight PEI as stabilizing agent, which resulted in fairly uniform NaYF₄(Yb,Er)@PEI UCNPs, with the red emission stronger than the green emission.

For PDT applications using UCNPs, spectral overlap between the emission of UCNPs and the absorption of photosensitizing molecules is crucial. In this study, Ce6 absorption has perfect overlap with the red emission of NaYF₄(Yb,Er)@PEI UCNPs, as shown in Fig. 2.

Fig. 2 (Left) Spectral overlap between NaYF₄(Yb,Er)@PEI UCNPs emission and Ce6's absorption. (Right) Photograph of stable NaYF₄(Yb, Er)@PEI-Ce6 dispersion in water.

The photosensitizing molecules were deposited on the NaYF₄(Yb,Er)@PEI UCNPs surface through electrostatic interaction between the amine functionalized surface and the acid functionality of Ce6. Thereafter, the nanoparticles were washed three times until there was no Ce6 in the supernatant. The resulting nanoparticles have strong green color because of the Ce6 on the surface (Fig. 2). The strong bonding between Ce6 and UCNPs is evident from the absorption spectra shown in Fig. 3.

Fig. 3 (Left) UV-Vis spectra of the supernatant after multiple washes. (Right) Emission spectra with different amounts of Ce6 on the NaYF₄(Yb,Er)@PEI UCNPs surface.

Due to close proximity of the Ce6 molecules to the NaYF₄(Yb,Er)@PEI UCNPs and because of spectral overlap between absorption of Ce6 and emission of UCNPs, 660 nm emission from UCNPs is transferred to Ce6 to great extent, which is evident in Fig. 3, where, as the amount of the Ce6 deposited on the UCNPs increases, the intensity of 660 nm peak from the UCNPs decreases.

Detection of singlet oxygen is assisted by a fluorescent probe molecule, 9-ACA. Singlet oxygen can convert the fluorescent 9-ACA molecule in the solution into its peroxide derivative, which is non-fluorescent. Decrease in the 9-ACA fluorescence intensity would indicate the presence of singlet oxygen.

Singlet oxygen generation from photosensitizers has been widely reported for photodynamic therapy of cancer and photoinactivation of bacteria. For these applications, NIR light is considered to be the preferred source of excitation, as it has deeper penetration into tissues than UV or visible light. In this study, Ce6 on the surface of the NaYF₄(Yb,Er)@PEI UCNPs can be excited either by the 980 nm laser (facilitated by energy transfer) or by visible light (∼660 nm) directly. We carried out experiments to assess how the presence of a beef tissue would affect the two types of excitation. The rate of decrease in the fluorescence intensity of 9-ACA after the mixture containing the NaYF₄(Yb,Er)@PEI-Ce6 and 9-ACA is illuminated by either light source for a certain period of time is a measure of singlet oxygen generation under the respective conditions.²²

The intensity of 9-ACA fluorescence after the illumination by either visible or NIR light is shown in Fig. 4A and B. Under visible light (660 nm output from the xenon lamp) illumination, the presence of a beef tissue in the light path would greatly reduce the rate of 9-ACA fluorescence decrease by ∼13-fold (from −0.26 to −0.02). This is likely due to absorption and scattering of the visible light by the tissue. In contrast, under the NIR light (980 nm laser) illumination, the beef tissue in the light path would only moderately reduce the rate of 9-ACA fluorescence decrease (from −0.08 to −0.06). It can thus be concluded that singlet oxygen generation under NIR illumination is less affected by the presence of beef tissue.

Fig. 4 Time-based measurements of fluorescence of 9-ACA and UCNPs mixture with and without beef tissue in light path under illuminations of (A) visible light (660 nm) and (B) NIR light (980 nm). Controls are described in text. (C) Singlet oxygen generation dependence on the UCNP concentration. (D) Photograph of the beef tissue, sandwiched between two glass slides, used in the experiment.

Note that various control experiments were performed and results are also shown in Fig. 4. Control 1 illustrates the stability of fluorescence intensity of 9-ACA under 363 nm excitation over the same period of time. Effect of visible light (660 nm) and NIR (980 nm) illuminations over the same time span, without nanoparticles, on fluorescence of 9-ACA is shown as control 2 in Fig. 4A and B, respectively, clearly indicating that fluorescence of 9-ACA is little affected by the light illuminations alone. Furthermore, no change in 9-ACA fluorescence is observed under the same conditions when only NaYF₄(Yb,Er)@PEI UCNPs without Ce6 were used (control 3 in Fig. 4A and B). Also shown as Fig. 4C is the singlet oxygen generation dependence on the concentration of the NaYF₄(Yb,Er)@PEI-Ce6 nanoparticles, demonstrating that as the amount of nanoparticles increases the rate of singlet oxygen generated increases. These control experiments confirm that the hybrid photosensitizer NaYF₄(Yb,Er)@PEI-Ce6 can generate singlet oxygen under the excitations of both visible light and NIR light, and more importantly, the NIR excitation is more tolerant to the presence of biological tissue.

Conclusion

In summary, we report the synthesis of a hybrid photosensitizer, NaYF₄(Yb,Er)@PEI-Ce6, based on upconversion nanoparticles. The NaYF₄(Yb,Er)@PEI-Ce6 nanoparticles can generate singlet oxygen under both visible excitation at ∼660 nm (directly to Ce6) and NIR excitation at ∼980 nm (indirectly through energy transfer from UCNPs to Ce6). It is observed that the presence of a beef tissue in the light path would significantly reduce singlet oxygen generation of the hybrid photosensitizer under the visible excitation, while only moderately affect it under NIR excitation.

Acknowledgements

Supports from the US NSF (CBET-0931677, CBET-1065633) are gratefully acknowledged.

Notes and references

1. (a) Y. Liu, D. Tu, H. Zhu and X. Chen, Chem. Soc. Rev., 2013, 42, 6924; (b) G. Liu, Chem. Soc. Rev., 2015, 44, 1635–1651; (c) E. Chan, Chem. Soc. Rev., 2015, 44, 1653–1679; (d) E. Chan, E. Levy and B. Cohen, Adv. Mater., 2015, 1–9,  DOI:10.1002/adma.201500248.
2. (a) J. Shen, L. Zhao and G. Han, Adv. Drug Delivery Rev., 2013, 65, 744–755; (b) A. Punjabi, X. Wu, A. Apollon, M. Rifai, H. Lee, Y. Zhang, C. Wang, Z. Liu, E. Chan, C. Duan and G. Han, ACS Nano, 2014, 8, 10621–10630; (c) D. Gargas, E. Chan, A. Ostrowski, S. Aloni, V. Altoe, E. Barnard, B. Sanii, J. Urban, D. Milliron, B. Cohen and J. Schuck, Nat. Nanotechnol., 2014, 9, 300–305.
3. P. Zhang, W. Steelent, M. Kumar and M. Scholfield, J. Am. Chem. Soc., 2007, 129, 4526–4527.
4. K. Lu, X. Liu, Q. Zeng, Y. Zhang, L. Tu, T. Liu, X. Kong, Y. Wang, F. Cao, S. Lambrechts, M. Alders and H. Zhang, ACS Nano, 2012, 6, 4054–4062.
5. C. Wang, H. Tao, L. Cheng and Z. Liu, Biomaterials, 2011, 32, 6145–6154.
6. S. Cui, D. Yin, Y. Di, H. Chen, Y. Ma, S. Achilefu and Y. Gu, ACS Nano, 2013, 7, 676–688.
7. C. Wang, L. Cheng and Z. Liu, Biomaterials, 2011, 32, 1110–1120.
8. Y. Park, H. Kim, J. Kim, K. Moon, B. Yoo, K. Lee, N. Lee, Y. Choi, W. Park, D. Ling, K. Na, W. Moon, S. Choi, H. Park, S. Yoon, Y. Suh, S. Lee and T. Hyeon, Adv. Mater., 2012, 24, 5755–5761.
9. Y. Wang, H. Wang, D. Liu, S. Song, X. Wang and H. Zhang, Biomaterials, 2013, 34, 7715–7724.
10. G. Tian, W. Ren, L. Yan, S. Jian, Z. Gu, L. Zhou, S. Jin, W. Yin, S. Li and Y. Zhao, Small, 2013, 11, 1929–1938.
11. F. Chen, S. Zhang, W. Bu, Y. Chen, Q. Xiao, J. Liu, H. Xing, L. Zhao, W. Peng and J. Shi, Chem.–Eur. J., 2012, 18, 7082–7090.
12. P. Kannan, F. Rahim, R. Chen, X. Teng, L. Huang, H. Sun and D. Kim, ACS Appl. Mater. Interfaces, 2013, 5, 3508–3513.
13. P. Castano, P. Mroz and M. Hamblin, Nat. Rev. Cancer, 2006, 6, 535–545.
14. E. Dennis, J. Dolmans, D. Fukumura and R. Jain, Nat. Rev. Cancer, 2003, 3, 380–387.
15. G. Yi and G. Chow, Adv. Funct. Mater., 2006, 16, 2324–2329.
16. C. Liu, H. Wang, X. Zhang and D. Chen, J. Mater. Chem., 2009, 19, 489–496.
17. F. Wang and X. Liu, J. Am. Chem. Soc., 2008, 130, 5642–5643.
18. F. Wang, D. Chatterjee, Z. Li, Y. Zhang, X. Fan and M. Wang, Nanotechnology, 2006, 17, 5786–5791.
19. Y. Zhang, A. Rufaihah, L. Sim and N. G. Ye, Biomed, 2008, 21, 82–85.
20. S. Kobayashi, K. Hiroishi, M. Tokunoh and T. Saegusai, Macromolecules, 1987, 20, 1496–1500.
21. F. Wang, P. Liu, T. Nie, H. Wei and Z. Cui, Int. J. Mol. Sci., 2013, 14, 17–29.
22. Y. Guo, M. Kumar and P. Zhang, Chem. Mater., 2007, 19, 6071–6072.

---