Octadec-1-ene (OD; 90%), NH₄F (99.99%), anhydrous YCl₃ and YbCl₃ (99%), ErCl₃∙6 H₂O (99%), N,N-dimethylacrylamide (DMA; 99%), fluorescein isothiocyanate (FITC), and phosphate buffered saline (PBS) were purchased from Sigma-Aldrich (St Louis, MO, United States). Sodium trihydrate of (4-amino-1-hydroxy-1-phosphonobutyl)phosphonic acid (alendronate; Ale) was obtained from TCI (Tokyo, Japan). Poly (methyl vinyl ether-co-maleic acid) (PMVEMA; M _w = 60,000 g/mol) was obtained from Scientific Polymer Products (Ontario, NY, United States). Oleic acid (OA; 98%), methanol (99.5%), hexane (99.5%), dichloromethane (99.9%), and paraformaldehyde were from Lach-Ner (Neratovice, Czech Republic). Ale-modified PEG (PEG-Ale) and poly (N,N-dimethylacrylamide-co-2-aminoethylacrylamide)-alendronate (PDMA-Ale) were prepared as described in literature (Kostiv et al., 2017; Oleksa et al., 2020; Nahorniak et al., 2023). Dulbecco’s modified Eagle’s medium (DMEM), Alamar Blue assay and 4′,6-diamidino-2-phenylindole (DAPI) were purchased from Thermo Fischer Scientific (Waltham, MA, United States). Fetal bovine serum (FBS) was from Merck (Darmstadt, Germany) and primocin and penicillin streptomycin were from Gibco, Life Technologies (Grand Island, NY, United States). Artificial lysosomal fluid (ALF, pH 4.5) was prepared as previously reported (Colombo et al., 2008). Suprapur^® nitric acid was purchased from Merck (Kenilworth, NJ, United States). Mesenchymal stromal cells (rMSCs) were isolated from rat bone marrow according to the previous report (Nahorniak et al., 2023). C6 cells were kindly provided by Dr. Čestmír Altaner from the Cancer Research Institute SAS (Bratislava, Slovakia). All other reagent grade chemicals were purchased from Sigma-Aldrich and used as received. Polyethersufone filter (MWCO 30 kg/mol) was obtained from VWR (Radnor, PA, United States). Ultrapure water (conductivity ˂0.1 μS/cm) was obtained on a Milli-Q Gradient A10 purification system (Millipore; Molsheim, France) and used throughout the experimental work.

UCNPs with hexagonal-phase were prepared by high-temperature coprecipitation method. To a solution of OA (6 mL) in OD (15 mL), YCl₃ (0.78 mmol), YbCl₃, (0.2 mmol) and ErCl₃∙6 H₂O (0.02 mmol) were added under magnetic stirring. The reaction mixture was heated to 160°C under vacuum (2.1 kPa) and maintained at this temperature for 30 min. The solution was then cooled to room temperature (RT) under Ar atmosphere and a solution of NH₄F (148 mg; 4 mmol) and NaOH (100 mg; 2.5 mmol) in methanol (8 mL) was added. To evaporate the methanol, the reactor pressure was slowly reduced to 2.1 kPa while stirring and the mixture was heated to 65°C. The reaction temperature was then slowly raised to 120°C and maintained at this value for 30 min to evaporate the residual water. Finally, the mixture was heated at 300°C for 1.5 h under Ar atmosphere and cooled to RT. The UCNPs were separated by centrifugation (3,460 rcf) for 1 h, washed twice with a hexane/ethanol (1:1 v/v) mixture, three times with ethanol, twice with ethanol/water (1:1 v/v), and twelve times with water (14 mL each), and redispersed in water.

Modification of UCNPs with polymers was performed according to a previously published procedure (Nahorniak et al., 2023). Briefly, a dispersion of UCNPs (30 mg) in water (1.64 mL) was added dropwise to an aqueous solution (2 mL) of PEG-Ale (M _w = 5,000 g/mol) or PDMA-Ale (15 mg; M _w = 11,000 g/mol) under sonication (Ultrasonic Homogenizer UP200S Hielscher; 20% power) for 1 min. The mixture was magnetically stirred at RT for 24 h. The resulting polymer-coated UCNPs denoted as UCNP@Ale-PEG and UCNP@Ale-PDMA were sedimented by centrifugation and washed twice with water (1.5 mL) followed by centrifugation (14,100 rcf).

UCNP@PMVEMA nanoparticles were prepared by adding aqueous dispersion of UCNPs (15 mg; 1 mL) to aqueous PMVEMA solution (50 mg/mL; 15 mL; pH 7.4 adjusted by adding 2 M NaOH) with shaking at RT for 30 min and further stirring at 70°C for 16 h. Nanoparticles were sedimented by centrifugation (14,100 rcf) and washed with water three times.

For the preparation of FITC-alendronate (FITC-Ale), FITC (105 mg) was dissolved in dimethyl sulfoxide (1 mL), a solution of Ale (70 mg) in 0.2 M sodium carbonate-bicarbonate buffer (3 mL; pH 10.4) was added, and the mixture was stirred at RT for 72 h in the dark. Then, the resulting FITC-alendronate (FITC-Ale) was purified by gel filtration on a Sephadex G-25 column with water as eluent and the solution was passed through a column filled with sulfonated polystyrene to replace Na⁺ with H⁺ ions in the phosphonate groups of FITC-Ale. Finally, water was removed at 40°C on a vacuum rotary evaporator. The structure of FITC-Ale was confirmed by ¹H NMR (SI Supporting Information, Supplementary Figure S1); the purity was 52% and the yield was 29 mg.

In the next step, neat UCNPs or UCNPs coated with PEG-Ale and PDMA-Ale were functionalized with FITC-Ale. An aqueous solution of FITC-Ale (0.05 mg; 0.5 mL) was added to an aqueous dispersion (3.64 mL) of particles (30 mg) and the mixture was stirred at RT for 21 h in the dark. In the case of UCNP@PMVEMA particles, an aqueous solution of FITC-Ale (0.5 mg; 0.225 mL) was added 30 min after the addition of the PMVEMA solution to the neat UCNPs under the same conditions as described above. The resulting FITC-modified particles were separated by centrifugation and washed twice with water (1.5 mL) followed by centrifugation (14,100 rcf).

Dispersion of UCNPs in water, 0.01 M PBS (pH 7.4), DMEM or ALF (1 mg/mL) was placed in a plastic vial sealed with a rubber septum and allowed to age at 25 or 37°C for a period of time with agitation (250 rpm). Subsequently, the dispersion was centrifuged (14,129 rcf) for 30 min to separate UCNPs and the supernatant was filtered (MWCO 30 kg/mol) to remove residual particles. A combined fluoride electrode (Thermo Fisher Scientific; Waltham, MA, United States) was used to measure the dissolution of UCNPs expressed as the concentration of released F ̄ ions in the medium according to the manufacturer’s instructions. The molar fraction of F ̄ ions was expressed as the mean ± standard error of the mean (n = 3).

The concentration of Y³⁺, Yb³⁺ and Er³⁺ ions leached from neat and polymer-coated UCNPs was determined spectrophotometrically according to our previous publication (Nahorniak et al., 2023). Briefly, supernatants from particle dispersions were separated via centrifugation (14,010 rcf) for 30 min and mixed with xylenol orange buffer solution (2 mL; pH 5.8). The concentration of free Y³⁺, Yb³⁺ and Er³⁺ released from the UCNPs was measured by UV-Vis spectroscopy at 350–650 nm from the ratio of absorbance at 570 and 443 nm. The calibration curve was obtained as the dependence of the integrated intensity ratios of absorption bands from 18 μM xylenol orange in acetate buffer (pH 5.8) at different concentrations of YCl₃ (0–70 μM Y³⁺).

ICP-MS measurements were performed on a PerkinElmer NexION 350D ICP-MS instrument (Woodbridge, Canada) equipped with Universal Cell Technology™. The sample system consisted of an internal peristaltic pump with Tygon^® tubing (0.38 mm inner diameter), a polytetrafluoroethylene concentric nebulizer and a 100-mL glass cyclone spray chamber.

After 24 h incubation with all tested UCNPs at a concentration of 20 μg/mL, rMSCs or C6 cells were treated with concentrated nitric acid (3 mL) and the mixture was placed in a Teflon digestion vessel (Speedwave 4; Berghof, Germany). After dilution, digested sample was spiked with internal standard solution (¹⁰⁰Rh) and placed in a 50-mL volumetric flask. Calibration solution and internal standard solution were prepared from a starting concentration of 1.000 ± 0.002 g/L (Merck, Darmstadt, Germany).

The cell cultures were prepared as previously reported (Nahorniak et al., 2023) and the cytotoxicity of all prepared UCNPs was determined by Alamar Blue assay. Briefly, rMSCs and C6 cell line were incubated with UCNPs for 24 h in complete culture medium, after which the medium was aspirated and a 10% Alamar Blue solution in culture medium was added and incubation continued for 3 h. The solution was pipetted into 96-well plates and fluorescence was measured using a FLUOstar Omega reader (BMG LABTECH; Ortenberg, Germany) at an excitation of 530 nm and emission of 590 nm. In the meantime, fresh culture medium was added to the cells and left for another 48 h; fluorescence was measured after 72 h as a percentage of control (readouts of nanoparticle-free cells).

Cells grown on coverslips were incubated with UCNPs (20 μg/mL) for 24 h, fixed with 4% paraformaldehyde for 20 min, stained with DAPI and mounted. The presence of UCNPs in rMSCs or C6 cells was monitored using an Olympus FV1200 MPE multiphoton laser scanning microscope (Tokyo, Japan). Briefly, cells were fixed with 4% paraformaldehyde in PBS, stained with DAPI and light-field images were taken at 980 nm excitation and 540 nm emission using an infrared pulsed laser with negative chirp for multiphoton excitation. DAPI-stained cells were observed in the blue channel using a 405 nm laser diode and the bright field and the blue channel were superimposed. FITC-modified UCNPs were imaged using a Zeiss fluorescence microscope (Oberkochen, Germany) in a green filter (488 nm).

Micrographs of the synthesized UCNPs were recorded with a Tecnai Spirit G2 transmission electron microscopy (TEM; FEI; Brno, Czech Republic). The TEM was coupled with energy-dispersive X-ray (EDX) detector (Mahwah, NJ, United States). Particles were also characterized using a Thermo Nicolet 870 FTIR spectrometer (Madison, WI, United States) equipped with a liquid nitrogen-cooled mercury cadmium telluride detector using a GoldenGate single reflection diamond ATR system (Specac; Orpington, United Kingdom). Hydrodynamic size, polydispersity (PD) and ξ-potential were measured using dynamic light scattering (DLS; ZSU 5700 Zetasizer Ultra Instrument; Malvern Instruments; Malvern, United Kingdom). ¹H NMR spectra were measured on a Bruker Avance III 600 spectrometer (Billerica, MA, United States). Thermogravimetric analysis (TGA) was conducted on a Perkin Elmer TGA 7 analyzer (Norwalk, CT, United States); the particles were heated in air from RT to 650°C at 10°C/min. UV-Vis spectra were recorded on a Specord 250 Plus UV-Vis spectrophotometer (Analytik; Jena, Germany). Emission and excitation spectra were recorded using a FS5 Edinburgh Instruments spectrofluorometer (Livingston, United Kingdom) coupled with UV-Vis Xe lamp and 980 nm CW laser with 2 W output power (MDL-III-980).

The UCNPs were synthetized by a high temperature coprecipitation method with rigorous removal of volatile solvents, which resulted in the formation of fewer nuclei than in the preparation of small (25 nm) spherical particles (Nahorniak et al., 2023). The TEM image showed that the UCNPs consisted of larger particles transformed into hexagonal plates that had the uniform size with dispersity Ð ∼1 (Figure 1A; Table 1). The average diameter of the plates was ∼120 nm and their thickness was 68 nm. The hydrodynamic size was 174 nm and the ξ-potential was positive (28 mV; Table 1, Supplementary Figure S2). The presence of Y, Yb and F in the particles was confirmed by TEM/EDX spectra, which showed large peaks at ∼0.68, 1.04 and 1.91 keV corresponding to F, Na and Y atoms, respectively (Figure 1E). Peaks at 0.23 and 8.04 keV were attributed to C and Cu atoms from the supporting TEM grid. The smaller peaks at 1.54, 7.4 and 8.42 keV belonged to Yb atoms, while the signal of Er atoms (∼6.9 keV) was not visible due to their low concentration.

To ensure the chemical and colloidal stability of hexagonal UCNPs in aqueous media used in biological experiments, the particles were protected against dissolution by coating with polymers, namely, PEG-Ale, PDMA-Ale and PMVEMA; the hydrodynamic particle size then reached ∼160–230 nm and the polydispersity PD was ∼0.1, characterizing a narrow particle size distribution (Figures 1B-D, Table 1, Supplementary Figure S2A). Compared to UCNP@PMVEMA particles, the correlation functions of UCNP@Ale-PEG and UCNP@Ale-PDMA decayed faster confirming their smaller size (Supplementary Figure S2B). The ξ-potential of UCNP@Ale-PEG, UCNP@Ale-PDMA and UCNP@PMVEMA particles was 4, 22 and −46 mV, respectively, (Table 1, Supplementary Figure S2C). This was due the electroneutrality of PEG and the positive and negative charges of PDMA and PMVEMA, respectively. At the same time, negatively charged particles are known to have a lower frequency of binding to plasma proteins, i.e., they avoid clearance by the reticuloendothelial system. According to TGA, the amount of PEG, PDMA and PMVEMA coating was ∼2, 3 and 14 wt% (Table 1). It should still be noted that PDMA contained a reactive N-(2-aminoethyl)acrylamide comonomer suitable for potential attachment of biomolecules. The ATR FTIR spectra of hexagonal UCNPs, UCNP@Ale-PEG, UCNP@Ale-PDMA and UCNP@PMVEMA particles confirmed the presence of polymers on the nanoparticle surface, where the characteristic ν(OH), ν(C-H) symmetric and asymmetric, ν(CH₂) symmetric and C-O stretching vibrations were observed at ∼3,450–3,300, 2,927–2,884, 2,850–2,860 and 1,105–1,080 cm^-1, respectively (Figure 2A). The spectrum of hexagonal UCNPs@Ale-PDMA exhibited absorption band at 1,633 cm^-1, which was attributed to ν(C=O) stretching vibrations of the amide DMA (Oleksa et al., 2020). Then, after modification of the hexagonal UCNPs by PMVEMA, the asymmetric (1,578 cm^-1) and symmetric (1,403 cm^-1) stretching vibrations of the COOH groups of PMVEMA were found in the spectrum (Ahmad et al., 2020).

The upconversion luminescence of the neat and polymer-coated hexagonal UCNPs in water was determined under excitation in the near-infrared region at a wavelength of 980 nm (Figure 2B). The characteristic upconversion emission of Er³⁺ was observed, which was attributed to the transitions ²H_11/2 → ²I_15/2 (409 nm), ⁴S_3/2 → ²I_15/2 (525 nm), ²H_9/2 → ⁴I_15/2 (542 nm), and ⁴F_9/2 → ²I_15/2 (656 nm). The upconversion emission spectra of the hexagonal nanoparticles (120 nm in diameter) showed similar peaks and emission intensities to those observed for previously described spherical 25-nm particles (Supplementary Figure S3) (Nahorniak et al., 2023). However, the uncoated large hexagonal UCNPs exhibited ∼30 times higher emission intensity at 542 nm than the small spherical UCNPs, making them more attractive and useful for bioassays and biological imaging. Uncoated and PEG-Ale- and PDMA-Ale-coated UCNPs exhibited weaker upconversion emission intensity than UCNP@PMVEMA, independent of the thickness of the PDMA and PEG coating. This demonstrated the quenching effect caused by irregularities in the coating of these particle that allowed direct access of water to the particle surface. The UCNP@PMVEMA particles had the highest upconversion emission intensity due to their largest hydrodynamic diameter of all particles.

The colloidal stability of neat and polymer-modified UCNPs in water, PBS and DMEM was monitored by analyzing their D _h and ζ-potential as a function of time. No significant changes in D _h and ζ-potential were observed for any of the polymer-coated nanoparticles incubated in water at 25°C and 37°C for 168 h, demonstrating their good colloidal stability (Supplementary Figures S4A, B and Supplementary Figures S5A, B). Similar results were observed for neat UCNPs, but after 168 h at 37°C the particles started to aggregate and their hydrodynamic diameter increased from 120 to 220 nm. This was attributed to a decrease in their ζ-potential, as these nanoparticles were only stabilized by electrostatic repulsions (Supplementary Figure S5A). In PBS, the D _h of UCNP@Ale-PEG increased after 24 and 6 h of incubation at 25°C and 37°C, respectively. UCNP@Ale-PDMA particles started to aggregate immediately after contact with PBS, and their D _h was >1,000 nm. In both cases, the aggregation was probably due to the small amount and rapid exchange of polymer coatings with phosphate ions, as previously observed for PEG-phosphate-modified UCNPs (Boyer et al., 2010). The ζ-potential of both UCNP@Ale-PEG and UCNP@Ale-PDMA approached zero due to the formation of the counterion layer (Supplementary Figure S5C, D). UCNP@PMVEMA particles showed very good colloidal stability in PBS at both temperatures for 168 h, as there were no noticeable changes in their D _h and ζ-potential. Also, neat UCNPs, UCNP@Ale-PEG and UCNP@PMVEMA particles were colloidally stable in DMEM at 37°C and their D _h and PD were constant for 168 h (Supplementary Figure S6A, B). In contrast, UCNP@Ale-PDMA particles were unstable in DMEM and their D _h and PD increased due to opsonization with proteins and aggregation.

The dissolution of neat and polymer-coated hexagonal UCNPs in water and PBS at 25°C and 37°C was measured with a fluoride ion selective electrode from time-depended changes in the concentration of F ̄ ions in the supernatants. The F ̄ ion leakage was determined as the molar fraction of dissolved F ̄ ions (X _F) relative to the total amount of fluorine in the nanoparticles with a nominal composition of NaF₄:Y_0.78:Yb_0.20,Er_0.02 (Figures 3A–D). It can be seen that at higher temperatures, X _F increased (Figures 3B, D) due to the better solubility of NaYF₄ (Lisjak et al., 2016). It is also interesting to note that the UCNP@PMVEMA particles dissolved more in water compared to the neat particles (Figures 3 A, B). This phenomenon has been previously described for small 25-nm UCNPs and also for other polymer coatings probably due to their complexation with lanthanides (Plohl et al., 2017; Nahorniak et al., 2023). In contrast to water, dissolution of all particle types was significantly higher in PBS (Figure 3) because the reaction of phosphates with lanthanides accelerated the hydrolysis of surface atoms (Lisjak et al., 2016; Saleh et al., 2020). The F ̄ leakage from neat UCNPs reached 20 mol% after 44 and 7 h of incubation in PBS at 25°C and 37°C, respectively (Figures 3C, D). UCNP@Ale-PEG and UCNP@Ale-PDMA particles were more resistant to dissolution in PBS; at 25°C, X _F reached 20 mol% after 94 and 81 h, respectively (Figure 3C). In the case of UCNP@PMVEMA particles in PBS, X _F reached only 16 mol% even after 168 h. When the temperature was increased to 37°C, 20 mol% F ̄ was released from UCNP@Ale-PEG, UCNP@Ale-PDMA and UCNP@PMVEMA after 15, 14 and 43 h of incubation, respectively (Figure 3D). As a result, the dissolution rate of UCNP@PMVEMA in PBS at 37°C was lower (0.45 mol%/h) than that of UCNP@Ale-PEG (0.8 mol%/h), UCNP@Ale-PDMA (1 mol%/h) and neat UCNPs (1.3 mol%/h; Table 1). Similar results were also obtained at 25°C, but the dissolution rate of all particles was 2.5–6 times slower than that at 37°C due to the poor solubility of NaYF₄. Thus, all protective coatings retarded the dissolution of UCNPs in PBS. Moreover, the determination of the molar fraction of released Y³⁺, Yb³⁺ and Er³⁺ ions (X _Y,Yb,Er) supported the above statement that UCNP@PMVEMA particles were the most resistant to dissolution in PBS at 37°C. X _Y,Yb,Er of UCNP@PMVEMA, UCNP@Ale-PEG, UCNP@Ale-PDMA and UCNPs after 72 h of incubation was 1.46 ± 0.07, 1.57 ± 0.25, 2.46 ± 0.25 and 4.75 ± 0.63 mol%, respectively (Figure 4).

Furthermore, DMEM culture medium and ALF simulating the environment inside lysosomes with which the particles come into contact after endocytosis were used to model the dissolution of UCNPs under different in vitro conditions (pH, ionic strength and the presence of biomolecules). In DMEM, 1.9, 1.8, 1.8 and 3.3 mol% of F ̄ ions were released from UCNPs, UCNPs@Ale-PEG, UCNPs@Ale-PDMA, and UCNPs@PMVEMA particles, respectively, after 168 h of aging at 37°C (Supplementary Figure S7A). The nanoparticle dissolution in ALF was more pronounced than in DMEM, with X _F reaching ∼13, 8, 9 and 16 mol% for neat, Ale-PEG-coated, Ale-PDMA-coated and PMVEMA-coated UCNPs, respectively, after 168 h at 37°C (Supplementary Figure S7B). The slower dissolution in DMEM that in ALF was due to the formation of a protein corona on the particle surface, which inhibited their dissolution (Saleh et al., 2020). Thus, UCNPs can be expected to dissolve predominantly intracellularly and their toxicity will be related to the number of internalized nanoparticles and protective polymer coating.

When comparing the dissolution of particles in PBS, the UCNP@PMVEMA with the highest amount of coating were the most resistant to dissolution. However, in water these particles demonstrated the highest F ̄ release, suggesting that not only the amount of polymer but rather chemistry of coating and surface ζ-potential influenced the rate of particle dissolution. In contrast to PMVEMA with many carboxyl groups that can act as attachment sites of polymer with the particle, Ale-PEG and Ale-PDMA contained only one phosphonate group for binding with UCNPs. Therefore, a polymer with one anchoring group on the particle surface can more easily exchange with phosphate ions, which accelerates the hydrolysis of surface atoms and induces faster particles dissolution. This agreed with earlier report, where ligand containing four anchoring groups provided higher protection of UCNPs against dissolution than those with only two (Andresen et al., 2020). On the other hand, the negative ζ-potential of UCNP@PMVEMA particles can suppress the diffusion of phosphate ions to the nanoparticle surface due to Coulomb repulsions, inhibiting exchange reaction and particle dissolution. Moreover, the particle size seems to influence the rate of dissolution in aqueous media. In comparison to earlier report on the dissolution of 25-nm particles with the same coatings, the dissolution of hexagonal UCNPs in water and PBS at 37°C was lower (Nahorniak et al., 2023). In PBS, the dissolution of 120-nm particles was lower by ∼14 and 20%–24% for neat and all polymer-coated UCNPs, respectively. In water, the decrease in solubility was lower by 31, 37, 65% and 9% for UCNPs, UCNPs@Ale-PEG, UCNPs@Ale-PDMA and UCNPs@PMVEMA, respectively (Table 2). This was attributed to a smaller surface-to-volume ratio of 120-nm UCNPs, which agrees with the literature (Saleh et al., 2020). As a result, the dissolution of UCNPs is a complex phenomenon dependent on many parameters such as temperature, type of aqueous medium, nanoparticle size, surface chemistry and charge.

To assess the suitability of UCNPs for biological applications, it is necessary to determine whether the individual particle types, specifically UCNPs, UCNP@Ale-PEG, UCNP@Ale-PDMA and UCNP@PMVEMA both without and with bound FITC are toxic to cells. Two cell types were used to assess cytotoxicity, rMSCs serving as a model of healthy non-tumorigenic cells and a C6 rat tumor cell line representing diseased tumorigenic cells. Cells in cell culture medium were incubated with particles at a concentration of 1–1,000 μg/mL for 24 and 72 h, and cell viability was examined using the Alamar Blue assay, one of the best in vitro assays for the determination of cell viability, cytotoxicity and cell proliferation. It is an alternative to tetrazolium dyes such as 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT), and because it is non-toxic, it can also be used for long-term cell proliferation studies. Alamar Blue is based on the fluorometric redox indicator resazurin (7-hydroxy-3H-phenoxazin-3-one 10-oxide), a blue-colored non-fluorescent compound. After intracellular uptake, the oxidized resazurin is reduced to the fluorescent resorufin (7-hydroxy-3H-phenoxazin-3-one) due to the reducing environment of the cell cytosol. Resorufin produces bright red fluorescence with an excitation of 530–570 nm and emission of 580–610 nm; the fluorescence intensity is then used to quantify cell viability (Longhin et al., 2022).

The presence of nanoparticles in the medium for 24 h did not affect the proliferation of C6 cells regardless of the type of particle coating (Figure 5). Similarly, the viability rMSCs cultured for 24 h in the presence of different UCNPs at two highest concentrations (500 and 1,000 μg/mL) decreased only slightly (to 80%). Culturing C6 cells for 72 h in the presence of uncoated UCNPs and UCNP@PMVEMA particles at a concentration of 1,000 μg/mL resulted in a decrease in cell viability to 40%; with decreasing particle concentration, viability slowly increased to values reaching 80%–90%. Toxicity of UCNP@PMVEMA particles can be ascribed to their high solubility in water (Figures 3A, B) as well as in AFL. In contrast, cells cultured in the presence of UCNP@Ale-PDMA or UCNP@Ale-PEG nanoparticles at the two highest concentrations of 500 and 1,000 μg/mL maintained viability between 70% and 58%, while at lower concentrations cell viability was between 70% and 90%. Likewise, the viability of rMSCs cultured for 72 h in the presence of UCNPs and UCNP@PMVEMA particles was reduced to 30% and maintained <60% up to a concentration of 63 μg/mL. At lower concentrations of neat UCNPs and UCNP@PMVEMA, the viability ranged between 80%–90%. The rMSCs incubated with UCNP@Ale-PDMA maintained 80%–90% viability regardless of particle concentration. For the same cells incubated with UCNP@Ale-PEG at a concentration of 63–1,000 μg/mL, viability gradually decreased from 80% to 60%; at lower concentrations, rMSC viability was ∼90%. In addition, there was no difference in the viability of cells incubated with UCNPs either with or without FITC. These findings corresponded to the results of dissolution of UCNPs in ALF, as the best viability was achieved for particles coated with PEG and PDMA, which protected them well. In our previous study, a significant decrease in rMSCs and C6 cell viability (<20%; MTT assay) was observed already after 24 h of incubation with uncoated small spherical UCNPs at a concentration 1,000 μg/mL (Nahorniak et al., 2023). It should be mentioned that the toxicity of particles is generally influenced by a number of factors, including the ζ-potential, the particle size and the amount of lanthanide and fluoride ions released into the medium. A limitation of our study was that we did not have non-toxic polymer-coated particles with the same ζ-potential as the control. However, the ζ-potential of UCNP@Ale-PDMA particles was comparable to that of uncoated UCNPs, but the cytotoxicity of the former was significantly lower. Thus, it can be concluded that the cytotoxicity of UCNPs is not only determined by the ζ-potential but also by the dissolution of the particles.

To demonstrate the concept of dual imaging, i.e., to visualize particle internalization in cells and particle distribution in the cytoplasm by confocal microscopy, a commonly used fluorescent dye, fluorescein isothiocyanate (FITC), was attached to UCNPs, which is useful for prospective in vivo animal control and histology. In addition, optical imaging allows for independent control, where nanoparticles can be excited throughout the organism to easily verify their biodistribution. The FITC was first modified with Ale to complex it with lanthanide ions on the particle surface. Its normalized photoluminescence spectrum and spectra of FITC-modified polymer-coated UCNPs under excitation and emission at 492 and 517 nm, respectively, were shown in Supplementary Figure S8. The main advantage of fluorescent particles is that they are visible in an epifluorescence microscope, which is part of the equipment of every culture laboratory, and therefore their uptake by cells can be monitored during cell culturing; it is also easier to detect particles in histological tissue sections. In addition, with excitation at 980 nm and emission at 542 and 656 nm, UCNPs enable deep tissue in vivo imaging.

To examine the uptake of all differently coated UCNPs, rMSC and C6 cells were cultured in the presence of particles at a concentration of 20 μg/mL for 24 h. Cells were then fixed with paraformaldehyde and imaged under confocal and epifluorescent microscope (Figure 6). Our results confirmed that particle uptake can be visualized by both confocal and fluorescence microscopy. The strongest signal in both confocal and fluorescence microscopy was observed in cells incubated with uncoated UCNPs. Of the coated particles, UCNP@Ale-PDMA were the most abundant in the cells, followed by UCNP@PMVEMA. UCNP@Ale-PEG particles were the least abundant in the cells because PEG is bioinert to a number of biological components found in the human body including proteins and does not exhibit cell-adhesive properties. There were generally more particles in the cytoplasm of rMCSs because they are usually larger than C6 cells. The signal was brightest in cells incubated with uncoated UCNPs, probably due to their high internalization (Figure 6). These findings are consistent with the results obtained by ICP-MS, which determined the concentration of Y and Yb in rMSCs and C6 cells after 24 h of culture with nanoparticles (Figure 7). The highest concentrations of Y and Yb were found in rMSCs containing uncoated UCNPs; Y and Yb concentrations ranged from 11.3 to 15.4 and from 5.2 to 7.0 pg per cell, respectively. The amount of Y per cell ranged from 4.9 to 15.5 pg and from 5.6 to 12.3 pg for UCNP@Ale-PDMA and UCNP@PMVEMA nanoparticles, respectively; similarly, the amount of Yb per cell was 2.6–7.0 pg and 2.8–5.9 pg for UCNP@Ale-PDMA and UCNP@PMVEMA nanoparticles, respectively. The only difference was found for UCNP@Ale-PEG particles, as PEG prevented their internalizing into cells and the concentrations of Y and Yb were in the range of 1.7–2.0 and 0.9–1.1 pg per cell, respectively.

In general, slightly lower concentrations of Y and Yb were found in C6 cells than in rMSCs (Figure 7). The highest amounts of Y and Yb were in C6 cells incubated with UCNP@Ale-PDMA nanoparticles for 24 h, 3.5–10.1 pg and 1.6–4.6 pg per cell, respectively. The average levels of Y and Yb in C6 cells labeled with uncoated particles were 4.4 and 0.7 pg/cell, respectively, while C6 cells incubated with UCNP@PMVEMA particles contained 3.6 and 1.9 pg of Y and Yb per cell, respectively. As in rMSCs, the lowest concentrations of Y (1.7 pg) and Yb (0.9 pg) were found in C6 cells containing UCNP@Ale-PEG nanoparticles. Comparison of the results of hexagonal UCNPs with previously published data on spherical nanoparticles coated with the same polymers (Nahorniak et al., 2023) shows that C6 cells cultured with hexagonal UCNPs contain orders of magnitude more Y and Yb than after culture with small spherical UCNPs. This difference was not observed for rMSCs, where Y and Yb concentrations were in a similar range, probably due to the larger size of rMSCs and their higher capacity for particle endocytosis. The size of the nanoparticles may also play a role; large particles contain higher amounts of Y and Yb, although their total number in the cell may not be greater than that of small spherical nanoparticles.