2.2.1 Preparation of NaGdF₄:20 mol%Yb,2mol%Er,xmol%Fe UCNPs

NaGdF₄:Yb,Er,Fe UCNPs were synthesized via a one-step hydrothermal method using OA as the coordinating ligand. In a typical synthesis, 0.7 g (17.5 mmol) NaOH, 8 mL (25 mmol) OA, and 12 mL (206 mmol) anhydrous ethanol were mixed under agitation to form a white viscous solution. Following this, 4.8 mmol NH₄F (dissolved in 8.5 mL deionized water) was added to the above-mentioned solution and stirred to form a transparent solution. Then, 1.2 mL of an Ln(NO₃)₃ solution (0.8 M, Ln: 58 mol% Gd; 20 mol% Yb; 2 mol% Er) and 0 mol%,10 mol% or 20 mol% FeCl₃ solution was added under vigorous stirring. The mixture was then agitated for another 20 min at room temperature (∼25 °C) to form a translucent colloidal solution. Subsequently, 30 mL of the mixed solution was transferred to a Teflon-lined high-pressure reactor, and the system was kept at 180 °C for 12 h. After naturally cooling to room temperature, the sample was centrifuged at 2000 rpm for 3 min, washed with cyclohexane and ethanol alternately thrice, and then dispersed in cyclohexane for the subsequent experiments.

2.2.2 Preparation of UCNPs@Ce6@mSiO₂

The Stober method was used to coat the silica shell. The synthesized UCNPs were dissolved in cyclohexane (10 mg in 2 mL) and added to 20 mL of an aqueous solution containing 100 mg CTAB. The solution was stirred vigorously overnight at room temperature to evaporate cyclohexane. Then, 10 mL distilled water, 3 mL ethanol, and 150 μL NaOH solution (2 M) were added and heated to 70 °C while stirring. When the temperature was stable, a pretreated Ce6 solution (2 mg dissolved in 0.5 mL DMSO containing 12 μL APTES, 6 mg EDC-HCl, 4 mg NHS) was added and allowed to react at 37 °C for 2 h, following which 120 μL TEOS was added dropwise into the solution, which was kept at 70 °C for 2 h. After cooling to room temperature, the resulting nanocomposite was centrifuged at 3000 rpm for 5 min and washed with ethanol three times. Subsequently, the surfactant CTAB was removed using the ion exchange method. The obtained UCNPs@Ce6@mSiO₂ nanocomposite was dispersed in 50 mL of an ethanol solution containing 0.3 g NH₄NO₃ and stirred at 60 °C for 24 h. Finally, the sample was centrifuged and washed three times with ethanol to remove the remaining CTAB and then dispersed in ethanol. Finally, the sample was centrifuged and washed thrice with ethanol and then dispersed in ethanol. The filtrate was collected to measure the Ce6 content via UV–vis spectroscopy.

2.2.3 Preparation of UCNPs@Ce6@mSiO₂-COOH

First, 100 mg of UCNPs@Ce6@mSiO₂ was added to 100 mL of ethanol and sonicated for 10 min to establish uniform dispersion. Then, 3 mL APTES was added and the solution was transferred to a 50 °C water bath and stirred for 24 h. The reaction product (amino-modified UCNPs@Ce6@mSiO₂) was collected via centrifugation and washed with ethanol. Next, 50 mg amino-modified UCNPs@Ce6@mSiO₂, 40 mg succinic anhydride, and 60 μL triethylamine were dispersed in 5 mL DMF and allowed to react at 40 °C for 24 h. After centrifugation and ethanol wash, the carboxyl-modified UCNPs@Ce6@mSiO₂ was obtained.

2.2.4 Preparation of UCNPs@Ce6@mSiO₂-DOX

DOX (3 mL, 1 mg/mL) was mixed with UCNPs@Ce6@mSiO₂ in phosphate-buffered saline (PBS) to load DOX onto UCNPs@Ce6@mSiO₂. After stirring overnight at 37 °C, UCNPs@Ce6@mSiO₂@DOX was separated via centrifugation and gently washed thrice with PBS. The filtrate was collected to measure DOX content via UV–vis spectroscopy.

3.1.1 Analysis of the crystal structure

Fig. 2 shows the X-ray diffraction (XRD) patterns of nanoparticles doped with different amounts of Fe³⁺. The samples exhibited three main diffraction peaks at 29.74°, 30.32°, and 42.92°, which can be attributed to (1 1 0), (1 0 1), and (2 0 1) facet diffraction of β-NaGdF₄ crystals, respectively, (JCPDS: 27-0699). Because of Fe³⁺ doping, hexagonal Fe₂O₃ diffraction peaks appear at 33.14°, 35.64° and 49.48°, which correspond to (1 0 4), (1 1 0) and (0 2 4) facets, respectively, (JCPDS: 33-0664). When Fe³⁺ content reached 10 mol%, the crystallinity of β-NaGdF₄ was enhanced. However, upon increasing the Fe³⁺ content further, the crystallinity of β-NaGdF₄ decreased. It was probably because Fe³⁺ ions were substituted for Gd³⁺ ions in UCNPs (Ramasamy et al., 2013). When Fe³⁺ doping was relatively low (10 mol%), the Fe³⁺ ions seemed beneficial for β-NaGdF₄ crystallization. However, with a further increase in Fe³⁺ content, the crystallization of β-NaGdF₄ was inhibited. In addition, it is noteworthy that the growth rate of the NaGdF₄ (1 1 0) facet remarkably increased with the increase in Fe³⁺ content. This phenomenon is probably due to the lattice distortion caused by Fe³⁺ doping, which was also observed by Ramasamy (Ramasamy et al., 2013).

Close

Fig. 2

XRD patterns of NaGdF₄:Yb,Er nanoparticles doped with 0, 10, or 20 mol% Fe³⁺ (# - hexagonal Fe₂O₃).

Export to PPT

3.1.2 Elemental analysis

XPS was used to determine the elemental composition of the synthesized samples. Fig. 3a shows the survey XPS spectrum of 20 mol% Fe³⁺ ion-doped nanomaterials. The binding energies of Na 1s and F 1s appeared at 1072.44 eV and 685.05 eV; the peaks located at 149.17 eV and 142.97 eV can be attributed to Gd 4d_3/2 and Gd 4d_5/2 binding energy, respectively. These three elements constituted the β-NaGdF₄ host matrix. Figs. 3b and 2c show the binding energies of Yb 4d and Er 4d, wherein the peaks were located at 186.47 eV and 172.87 eV, respectively. This demonstrated that NaGdF₄:Yb,Er nanomaterials were doped with Yb and Er. In addition, the peaks at 711.87 eV, shown in Fig. 3d, were attributed to the Fe 2p_3/2 binding energy. This confirms the successful Fe³⁺ doping in NaGdF₄:Yb,Er nanoparticles (Ramasamy et al., 2013).

Close

Fig. 3

XPS of 20 mol% Fe³⁺-doped NaGdF₄:Yb,Er nanoparticles. (a) Survey, (b) Yb 4d, (c) Er 4d, and (d) Fe 2p.

Export to PPT

3.1.3 UCL analysis

Fig. 4 shows the fluorescence spectra of NaGdF₄:Yb,Er nanoparticles doped with different amounts of Fe³⁺ ions under 980 nm laser excitation. The main emission peaks appeared at 415, 530, 550, and 665 nm, which were attributed to the ²H_9/2-⁴I_15/2, ²H_11/2-⁴I_15/2, ⁴S_3/2-⁴I_15/2, and ⁴F_9/2-⁴I_15/2 energy transitions of Er³⁺ ions, respectively (Bai et al., 2007; Vetrone et al., 2009). The fluorescence intensity of the nanoparticles was enhanced when the content of Fe³⁺ ions increased. The UCL intensity of 10 and 20 mol% Fe³⁺-doped NaGdF₄:Yb,Er nanoparticles at 550 nm increased by 2.7 and 5.8 folds compared with that of the Fe³⁺-free sample. It is well known that the UCL intensity of rare-earth ions is mainly dependent on electron transition probabilities, and a hypersensitive transition can be triggered by changing environment. Thus, this obvious enhancement in UCL intensity might be induced by Fe³⁺ ions with a small radius, which can tailor the surrounding environment around the Er³⁺ ions in the crystal, leading to hypersensitive transition (Ramasamy et al., 2013). Therefore, a certain amount of Fe³⁺ ions doping could effectively enhance the fluorescence efficiency of NaGdF₄:Yb,Er nanoparticles. Based on UCL, 20 mol% Fe³⁺-doped NaGdF₄:Yb,Er nanoparticles were chosen as core structure in the following experiments.

Close

Fig. 4

Fluorescence spectra of NaGdF₄:Yb,Er nanoparticles doped with 0, 10, or 20 mol% Fe³⁺ under 980 nm excitation.

Export to PPT

3.2.1 Morphology analysis

As shown in Fig. 4a, NaGdF₄:Yb,Er,Fe nanoparticles were ellipsoidal with a size of approximately 25 nm. After the UCNPs were coated with silica, the particles exhibited good dispersion and a quasi-spherical shape, and the particle size was approximately 100–150 nm. The average for UCNP@mSiO₂ was approximately 130 nm. Each core-shell nanoparticle had one or more UCNP core, and the silica layer was uniformly coated on the surface of the nanoparticle at a thickness of approximately 50 nm. It is worth noticing that the pore structure could be clearly observed (Fig. 5b). This indicated that the silica layer was successfully coated on the surface of the UCNPs and possibly formed a mesoporous structure.

Close

Fig. 5

TEM images of (a) UCNPs and (b)UCNPs@mSiO₂.

Export to PPT

3.2.2 Specific surface area and pore size analyses

To further clarify the specific surface area and pore size distribution of the mesoporous silicon layer, as well as obtain nitrogen adsorption-desorption isotherms, nitrogen adsorption and desorption experiments were performed on UCNPs and UCNP@mSiO₂ nanoparticles. It is obvious from Fig. 6 that the physisorption isotherm of UCNP@mSiO₂ could be classified as type V, confirming the existence of the mesoporous structure (Jiang et al., 2014; Tian et al., 2019). This was also in agreement with previous results. The specific surface area of the nanoparticles coated with silica was significantly higher than that of the nanoparticles without the silicon coating (P < 0.01). The specific surface area of the nanomaterials rapidly increased from 5.149 m²/g to 559.257 m²/g. UCNPs@SiO₂ had a pore volume of 2.192 mL/g and a pore diameter of 3.087 nm, as calculated using the BJH model. These findings demonstrated that the nanocomposite material formed a porous structure after silica coating.

Close

Fig. 6

Nitrogen adsorption-desorption curves of OA-UCNPs and UCNPs@mSiO₂.

Export to PPT

3.2.3 Chemical bond analysis

The Stober method was used to coat the surface of NaGdF₄:Yb,Er,Fe nanoparticles with mesoporous silica, which greatly increased the specific surface area and pore volume. Since there are a large number of hydroxyl groups on the surface of pores and the silica layer, we further modified the amino groups using APTES, and with the aid of the ring cleavage of succinic anhydride, carboxyl groups were attached to the surface of the particles based on the amino modification.

To clarify the mSiO₂ and carboxyl group modifications of the UCNP surface and analyze the chemical bonds on the surface of the nanoparticles, FTIR was used; the results are shown in Fig. 7. In the FTIR spectrum of OA-UCNPs, the absorption peak at 3009 cm⁻¹ originated from the C—H vibration of the alkenyl group, while the weak absorption peak at 1652 cm⁻¹ was attributed to the alkenyl C⚌C vibration. The absorption peaks at 2925, 2853, 1466, and 723 cm⁻¹ were attributed to the asymmetric vibration, symmetric vibration, “scissor swing”, and “Swing” vibration of C—H in the alkyl group, respectively. The absorption peaks of 1564 and 1414 cm⁻¹ were the stretching vibration peaks of the carboxyl group —COO— in the alkyl chain. This illustrated that the NaGdF₄:Yb,Er,Fe nanoparticles were coated with OA before mSiO₂ coating. In the FTIR spectrum of UCNPs@mSiO₂-COOH nanoparticles, the stretching vibration peaks of the hydroxyl group appeared from 3000 to 4000 cm⁻¹, the Si—O—Si bending vibration peak appeared at 459 cm⁻¹, and the asymmetric and symmetric stretching vibration peaks of Si—O were observed at 798 and 1079 cm⁻¹. In contrast, the stretching vibration peaks of Si—OH appeared at 963 cm⁻¹, and the vibration peaks of —CH₂ and —CH₃ appeared at 2924 and 2854 cm⁻¹. These findings demonstrated that the nanoparticles were successfully modified by silica (Cao et al., 2015). In addition, the peak at 1702 cm⁻¹ corresponded to the C⚌O stretching vibration, indicating that the carboxyl group was successfully modified on the silica surface (Qian et al., 2014). The above-mentioned results indicated that UCNPs@mSiO₂ core-shell structure with carboxyl group modification was successfully prepared.

Close

Fig. 7

FTIR spectra of OA-UCNPs and UCNPs@mSiO₂-COOH nanoparticles.

Export to PPT

3.2.4 UV–vis spectroscopy

After successfully coating the mesoporous silica and modifying the carboxyl group, the absorption spectra of Ce6, DOX, UCNPs, UCNPs@mSiO₂, UCNPs@Ce6@mSiO₂, and UCNPs@Ce6@mSiO₂-DOX were measured using a UV–visible absorption spectrometer; the results are shown in Fig. 8. The photosensitizer Ce6 had obvious absorption peaks in the range of 338–434 nm and 620–660 nm, while the chemotherapeutic drug DOX had obvious absorption peaks in the range of 408–559 nm. There was no obvious absorption peak in the UV–vis absorption spectra of UCNPs and UCNPs@mSiO₂; however, after Ce6 was embedded in the mesoporous silica layer, obvious absorption peaks appeared at 412 and 664 nm, which were consistent with those of Ce6, indicating that Ce6 was successfully loaded onto the silicon dioxide layer. After DOX was also loaded, an absorption peak appeared at 504 nm, which was consistent with the DOX absorption peak range.

Close

Fig. 8

UV–visible absorption spectra of Ce6, DOX, UCNPs, UCNPs@mSiO₂, UCNPs@Ce6@mSiO₂, and UCNPs@Ce6@mSiO₂-DOX.

Export to PPT

3.2.5 UCL analysis

Fig. 9 shows the UCL spectrum upon excitation with NIR at 980 nm; UCNPs showed emission peaks at 409, 530, 550, and 654 nm, corresponding to ²H_9/2-⁴I1_5/2, ²H_11/2-⁴I1_5/2, ⁴S_3/2-⁴I_15/2, and ⁴F_9/2-⁴I_15/2 energy level transitions, respectively (Bai et al., 2007; Song et al., 2011; Yao et al., 2017). After the mesoporous silica layer was coated, the locations and intensities of UCL emission peaks did not change significantly (P > 0.05), indicating that the microporous silica coating had little effect on the luminescence UCNPs. After Ce6 loading, the fluorescence intensities at 530 and 550 nm did not change notably (P > 0.05), but the emission at 409 and 654 nm was significantly weakened (in the inset) (P < 0.01). These two emission peaks overlapped with the absorption peaks of Ce6 (Fig. 8), which indicated that fluorescence resonance energy transfer (FRET) may have occurred (Hu et al., 2019). After further loading DOX, the fluorescence intensities at 530 and 550 nm decreased. This might be due to the excessive loading of DOX on the surface of the nanoparticles, which caused a reduction in the NIR absorbed by the UCNP core. In contrast, as shown in Fig. 8, DOX showed light absorption at 408–559 nm, which overlapped with the green emission of UCNPs. Therefore, FRET may have also occurred, further resulting in a reduction in the emission intensity.

Close

Fig. 9

UCL spectrum of the synthesized nanoplatforms under 980 nm laser excitation.

Export to PPT

3.4.1 Cytotoxicity in the darkness and under 980 nm infrared light irradiation

To clarify the cytotoxic effects of UCNPs before and after mesoporous silica shell coating, MTT tests were performed using different cell lines treated with nanomaterials at concentrations of 0, 10, 30, 90, 270, and 810 μg/mL. As observed from Fig. 11, the viabilities of HeLa, MCF-7, and HepG2 cells treated with 10 μg/mL nanomaterials were about 96.9, 77.9, and 74.8%, respectively. As the nanomaterial concentration increased, the viabilities of the cancer cells decreased remarkably. When nanomaterial concentration reached 810 μg/mL, the average survival rates of the three cancer cell lines declined to 53.4, 37.6, and 26.0%, respectively. The viabilities of HeLa, MCF-7 and HepG2 cells exhibited obvious dose-effect manner. Cell viability decreases when the dose (or concentration) of the nanomaterials is increased. The toxicity was probably caused by OA, which is toxic to many cells in the human body at high concentrations (Fermor et al., 1992).

Close

Fig. 11

Toxicity of UCNPs and UCNPs@mSiO₂ against (a) HeLa, (b) MCF-7, and (c) HepG2 cells. (d) Toxicity of infrared irradiation at 980 nm (2 W/cm², 10 min) against HeLa, MCF-7, and HepG2 cells.

Export to PPT

After coating with mesoporous silica, although the UCNP nanoplatform still exhibited slight cytotoxicity at high concentrations, the overall cell activity was significantly enhanced (P < 0.01). Even at a concentration of 810 μg/mL, cell viability was maintained at approximately 80%. This was possibly because water-soluble silica has excellent photochemical stability and negligible cytotoxicity (Yang and Li, 2020), which substantially reduces the cytotoxicity of nanocarriers. This demonstrates that the silica coating can effectively reduce the cytotoxicity of oleic acid-modified nanoparticles, thereby improving the biocompatibility of nanomaterials.

As the photodynamic analysis required NIR irradiation at 980 nm, the toxicity of NIR irradiation at 980 nm (2 W/cm²) on HeLa, MCF-7, and HepG2 cells was evaluated. The results are shown in Fig. 11(d). We observed that NIR irradiation at 980 nm did not cause obvious cytotoxicity to the three cell lines, and the viabilities of all cells were above 95%.

3.4.2 UCNPs@Ce6@mSiO₂-DOX nanoplatform for PDT-chemotherapy of cancer cells

Fig. 12 shows the toxicity to HeLa, MCF-7 and HepG2 cells in the PDT group (UCNPs@Ce6@mSiO₂ + 980 nm), chemotherapy group (UCNPs@mSiO₂-DOX), and PDT-chemotherapy combination group (UCNPs@Ce6@mSiO₂-DOX + 980 nm), respectively. It can be observed from the figure that all the three experimental groups showed significant dose-effect relationship and high cancer cell toxicity (P < 0.01). At a concentration of 200 μg/mL, the viabilities of HeLa, MCF-7, and HepG2 cells in the PDT group decreased to 38.9, 28.6, and 29.3%, respectively; in the chemotherapy group, the cell viabilities declined to 33.7, 17.1, and 23.7%, respectively; in the PDT-chemotherapy combined group, the cell viabilities dropped to 18.5, 11.4, and 9.3%, respectively. Compared with those in the control group, the differences were statistically significant (P < 0.01). It is worth noting that under the same dose, the cell viability in the PDT-chemotherapy combined group was significantly lower than that in the other groups (P < 0.0001). This demonstrates that the UCNPs@Ce6@mSiO₂-DOX nanoplatform successfully combines the two treatment methods, at the same drug concentration. The nanoplatform has a strong killing capability against cancer cells in PDT-chemotherapy combination therapy.

Close

Fig. 12

PDT, chemotherapy, and PDT-chemotherapy combination applied to (a) HeLa, (b) MCF-7, and (c) HepG2 cells (*p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001).

Export to PPT

Based on the previous analysis, we propose the basic principle of UCNPs@Ce6@mSiO₂-DOX nanoplatform for PDT-chemotherapy combination in the treatment of cancer cells. In a typical PDT treatment, when the UCNPs@Ce6@mSiO₂-DOX nanoplatform is irradiated at 980 nm, UCNPs are excited and can emit green light to excite the photosensitizer Ce6, which can produce ¹O₂ and ROS (Agostinis et al., 2011; Zhang et al., 2015; Zhang et al., 2017a). These intracellular ROS could easily damage biomolecules by causing oxidative stress or via direct reaction with DNA, leading to apoptosis (Tang et al., 2018). In contrast, due to the unique acidic environment in cancer cells, DOX carried by the UCNPs@Ce6@mSiO₂-DOX nanoplatform can be effectively released, and DOX can effectively promote cancer cell apoptosis by inhibiting DNA and RNA synthesis (Zhang et al., 2017b). It also can directly act on the nucleus and lead to apoptosis. This indicates that the UCNPs@Ce6@mSiO₂-DOX nanoplatform can successfully achieve the combined therapy of PDT and chemotherapy to kill cancer cells.