Cytotoxic effects of upconversion nanoparticles in primary hippocampal cultures

Abstract

The widespread use of nanomaterials causes public concerns associated with their potential toxicological hazards. New-generation nanomaterials – upconversion nanoparticles (UCNPs) – hold promise for theranostics applications due to their unique optical properties, enabling imaging at the sub-centimetre depth in live biological tissue. In brain tissue, nanoparticle-aided optical imaging and treatment are deemed desirable. To this aim, we carried out cytotoxicity studies of UCNPs in primary hippocampal cultures. The most common core/shell UCNPs (NaYF₄:Yb³⁺:Tm³⁺/NaYF₄) were synthesized using a solvothermal method and hydrophilized with amphiphilic polymaleic anhydride octadecene (PMAO); polyethyleneimine (PEI). Bare UCNPs were produced by using tetramethyl ammonium hydroxide (TMAH). PMAO-, PEI- and TMAH-UCNPs (0.8 mg mL⁻¹) were incubated for 72 hours with primary hippocampal culture and exhibited noticeable cytotoxicity. Our studies showed profound morphological modification of all treated cells with the maximum and minimum uptake observed in PMAO- and TMAH-UCNP-treated cells, respectively. The spontaneous calcium activity in cells treated with TMAH-UCNP, PMAO-UCNP dropped to (17 ± 3)%, (6 ± 3)% of its original level and was completely inhibited in the PEI-UCNP-treated cultures. This study demonstrated that bare and polymer surface-coated upconversion nanoparticles are toxic to dissociated hippocampal cells, evident through aberrant morphological changes, deviant variations of Ca²⁺ activity, and cell death.

---

Introduction

Optical imaging of biological tissues provides highly informative, non-invasive and inexpensive means to assess tissue physiological status and functionality at the molecular level. Deployment of a new class of molecular probes whose excitation/emission falls into the biological tissue transparency window (650–1300 nm) allows deeper imaging due to the minimized absorption and scattering of biotissue in this near-infra-red (NIR) spectral range. The development of NIR-emitting probes is traditionally directed towards organic dyes,¹ but recent studies have demonstrated promise of inorganic nanoparticle-based probes, such as nanorubies² and upconversion nanoparticles (UCNPs).³ The useful properties of photoluminescent (PL) nanoparticles include virtually unlimited photostability,² chemical inertness, physical stability of the solid core, etc.⁴ The unique PL properties of these new-generation synthetic nanomaterials, in particular UCNPs, have enabled high-contrast optical biomedical imaging by suppressing the optical background due to biological tissue autofluorescence and evading high tissue absorption.⁵ Despite the critical supra-linear dependence of the UCNP emission versus the excitation intensity,⁶ leading to the signal diminishment at the centimetre-depth in biological tissue,⁷ such applications of UCNPs as small animal imaging have been successfully demonstrated.⁵

In general, nanoparticle physical and chemical characteristics can be programmed,⁸ while the large effective surface decorated with reactive functional groups allows attachment of task-specific guiding, therapeutic, or/and contrast-rendering modules.⁹ These properties allow versatile design of biohybrid photoluminescent nanocomplexes for theranostics applications.^10,11 Considering the limited accessibility of brain to surgical intervention and imaging, UCNP-aided visualization of specific molecular structures, diagnostic imaging and therapy of brain are of particular interest. At the same time, challenges of targeted delivery, optical imaging and photosensitization of UCNPs and other nanoparticles and macromolecules in the brain must be addressed. Although the delivery of large-size molecules and nanomaterials is hampered by the brain-blood barrier, this barrier is compromised in several pathological conditions, such as brain tumors.^12–14

In view of these reasons, it is important to evaluate safety and biocompatibility of the emerging nanoparticles, including upconversion nanoparticles, for brain cells and tissues. At present, it is generally held that UCNPs comprising rare-earth doped Na⁺ and Ca²⁺ oxide¹⁵ or fluorides matrixes (e.g. NaYF₄,¹⁶ NaLuF₄,⁵ etc.) are nontoxic or mildly toxic to cells and tissues.^17–20 Mild deteriorative effects of the concentration on the cell survival rate have been reported only for the highest UCNP doses of 500–2000 μm mL⁻¹.^21,22 At the same time, UCNPs conjugated with the chemotherapeutic targeting agents (doxorubicin, photoactivatable platinum(IV) prodrug, etc.) demonstrated significant cytotoxicity in the tumor cell lines.¹⁹ It is generally consented that the nanoparticle interaction with cellular systems is primarily governed by the surface moieties, although the cytotoxic properties of the core (bare nanoparticle) are also important in case the integrity of the surface coating is compromised.²²

Despite a wealth of information published on this topic, evidence of the biosafety of UCNPs is still largely circumstantial due to the considerable disparity of the UCNP characteristics used in the reported cytotoxicity tests. For example, the tested concentrations of the nanoparticles ranged from 5 to 2000 μg mL⁻¹, incubation time ranged from 1 h to 9 days,^22,23 the tested cell types varied from normal, tumor, stem and differentiated cells, as well as the composition, size, shape, surface charge, synthesis and surface modification procedures of the UCNPs themselves. It is important to note that all reported studies of the UCNP cytotoxicity were carried out using the established cell lines with the propensity for fast and unlimited proliferation. These cells are less susceptible to adverse conditions and toxic agents owing to the gene activity modification responsible for the cell proliferation and cell cycle progression.²⁴ It is also consented that immortalized cultures of one cellular phenotype reproduce poorly the signaling properties of the parent cells. In particular, neural cell lines exhibit no functional activities, which cells in the central nervous system do.²⁵ The neural network is considered a functionality unit of the central nervous system sustaining the higher nervous system activity, such as transmission and storage of information, learning and memory.²⁶ Hence, the neural system response to nanomaterials needs to be investigated at least at the hierarchical level of the primary hippocampal cultures, which ensures adequate assessment of UCNP toxic effects, including dysfunctional network activity. This activity is conventionally visualized by means of Ca²⁺-imaging, which reports on the changes of the intracellular calcium concentration in each cell of the neural network, and the contribution of each neuron or astrocyte to the network activity.

In this paper, we address the biocompatibility of upconversion nanomaterial. The most conventional composition of UCNP was chosen as a core/shell beta-phase sodium yttrium fluoride co-doped with 18% (molar ratio) ytterbium; 1.4% erbium and 0.6% thulium (β-NaYF₄:Yb³⁺:Er³⁺:Tm³⁺/NaYF₄), followed by surface modifications.²² Our selection of the tested surface modifications was guided by the reported studies by Guller et al.,²² where the following three surface modifications were graded from the least to most cytotoxic: (1) bare UCNP produced by the ligand-exchange reaction with a phase-transition agent, tetramethylammonium hydroxide (TMAH-UCNPs); (2) amphiphilic polymaleic anhydride octadecene (PMAO), and; (3) positively-charged polymer, polyethyleneimine (PEI). The functionality of the neural network activity was assessed by labeling, imaging and analyzing the Ca²⁺ distribution in single neurons and in ensemble.

Materials and methods

Synthesis of upconversion nanoparticles

In order to synthesis beta-phase sodium yttrium fluoride co-doped with 18% (molar ratio) ytterbium; % erbium and 0.6% thulium (β-NaYF₄:Yb³⁺:Er³⁺:Tm³⁺/NaYF₄) NPs, the following procedures were carried out. A mixture of Y₂O₃ (0.78 mM), Yb₂O₃ (0.2 mM), and Er₂O₃ (14 mM), Tm₂O₃ (6 mM) was suspended in 20 mL 70% trifluoroacetic acid, gently refluxed until a clear solution was obtained (6 h), and then cooled to room temperature. The solution was then evaporated, and the residue was dried in vacuum (0.1 Torr) for 3 h. The resulting slurry was thoroughly ground in an agate mortar to obtain a fine homogenous powder. A rare-earth trifluoroacetate mixture and sodium trifluoroacetate (2 mM) was added to 10 mL oleic acid and 10 mL 1-octadecene in a three-neck flask equipped with a thermometer and glass magnetic stirrer. Next, the solution was heated at 120 °C and stirred in a vacuum for 30 min for degassing and water removal. The mixture was then gradually heated at a rate of 100 C min⁻¹ to 318 °C on a Wood's alloy bath and maintained at this temperature for 26 min in an argon atmosphere. Then, the mixture was cooled quickly by adding 15 mL 1-octadecene to the flask. In order to cool the solution to room temperature, 150 mL isopropanol was added, and the mixture was centrifuged at 6000 rpm for 20 min. The resulting nanocrystals were washed with 100% ethanol three times, dried, dissolved in 10 mL chloroform, precipitated with 50 mL isopropanol, and centrifuged at 4000 rpm for 10 min. The residue was dried at room temperature.

Surface modification of UCNPs

The following materials were purchased from Sigma-Aldrich and used without further purification: potassium persulfate, sodium dihydrogen phosphate, sodium hydrogen phosphate, sodium chloride, boric acid, sodium tetraborate, sodium hydroxide, and poly-N-vinylpyrrolidone. The following materials were of analytical grade and purchased from Aldrich: 2,2-azo-bis-isobutyronitrile, hexane, isopropyl alcohol, and chloroform. Styrene was purified by mixing with 5% sodium hydroxide aqueous solution (to remove the stabilizer), washing in water until the pH was neutralized, drying over calcium chloride, and distilling twice in a vacuum. The buffer stability of the produced colloids and their further fictionalization were achieved using the approaches described below.

Design of biocompatible UCNPs

Four main strategies to design biocompatible samples of UCNPs were employed, including: (1) ligand exchange reaction; (2 and 3) solvent evaporation method with using amphiphilic polymers – poly(maleic anhydride-alt-1-octadecene) (PMAO) and polyethyleneimine (PEI).

(1) The ligand exchange reaction was carried out with tetramethylammonium hydroxide (TMAH) to evaluate the intrinsic toxicity of the as-synthesised UCNPs. TMAH is a low molecular phase-transition catalyst that is dissociated in water by producing OH– ions. TMAH is adsorbed on the surface of UCNPs,²⁷ partially displacing oleic acid moieties from the OA-UCNP surface. UCNPs were transferred into an aqueous fraction by preparation of an UCNP micro-emulsion in aqueous solution of TMAH, followed by solvent evaporation.

(2 and 3) Amphiphilic polymers were adsorbed onto the UCNP surface via hydrophobic interactions between the oleate ligand and the hydrocarbon chain of the polymer. The hydrophilic counterpart of the amphiphilic polymers was directed outwards (i.e., into water), rendering OA-UCNP hydrophilic. We used the following amphiphilic polymers: poly(maleic anhydride-alt-1-octadecene) (UCNP-PMAO), polyethyleneimine (UCNP-PEI). We deployed a solvent evaporation method for UCNP surface modification. The solvent evaporation method is usually used for modification of inorganic nanocrystals surface-coated with water-insoluble polymers.²⁸ In accordance with this method, UCNPs were dispersed in volatile solvent (e.g., chloroform), and polymer solution dissolved in chloroform was then added. After ultrasonic treatment, the mixture was stirred at room temperature for 1.5 h to adsorb the polymer chains on the surface of UCNPs. The resulting mixture was then added drop-wise to water or aqueous solution with stirring and sonication to dilute the mixture by at least ten-fold. The intermolecular forces between the polymer chains enabled the formation of particles in water (or aqueous solution) i.e. insoluble polymer particles containing UCNPs. Evaporation of the solvent resulted in an aqueous dispersion of UCNPs surface-modified with the polymer.

We obtained four samples of aqueous colloids of surface-modified UCNPs from these surface modification and complexation procedures. Aqueous colloids of the modified UCNPs were stable over at least two weeks, and stability was immune to electrolytes (0.15 M NaCl and buffer solutions). It is worth noting that there were practically no changes in the photoluminescent properties of UCNPs after their modification.

Hydrophilization of UCNPs with tetramethylammonium hydroxide pentahydrate (TMAH-UCNPs)

20 μL UCNP dispersion (10 mg mL⁻¹) in chloroform was added drop-wise to 1 mL 1% aqueous solution of TMAH, and the two immiscible phases were thoroughly shaken. When UCNPs were transferred to the water phase, chloroform was evaporated. To remove excess TMAH, the UCNP aqueous suspension was washed three times with water (consecutive centrifugation at 13 400 rpm for 10 min for the redispersion steps). The pellet was then dispersed in 1 mL water.

Surface modification of UCNPs with polymers

Intercalation of UCNPs using amphiphilic polymers. In order to coat the as-synthesized UCNPs with the amphiphilic polymer poly(maleic anhydride-alt-1-octadecene) (PMAO), the solvent evaporation technique was used as described elsewhere.²⁹ The mixture was stirred, sonicated for 5 min, and incubated for 1 h, while stirring at room temperature. The mixture was then added drop-wise to 1 mL water, PBS buffer (pH 7.2) or Na–borate buffer (pH 8.2) under vigorous stirring and sonication. The solvent was evaporated by heating for 30 min, and the mixture was centrifuged at 13 400 rpm for 10 min with water addition. This procedure was repeated three times to remove free polymers, and the mixture was then dispersed in 1 mL water, PBS buffer (pH 7.2) or Na–borate buffer (pH 8.2).

Characterization of UCNPs

Transmission electron microscopy (TEM) was used to determine the primary composition of the as-synthesized upconversion nanocrystals. An aqueous colloid of UCNPs was dropped on a TEM copper grid, dried and examined under a Philips CM10 TEM (Philips, Eindhoven, The Netherlands). Freeware ImageJ was used to obtain and analyze the size distribution of the UCNP sample. The hydrodynamic diameters of the surface-modified nanoparticles were measured by a sub-micron particle analyzer (Coulter N4-MD, Coultronics, France). The zeta-potential was measured by a method of dynamic light scattering (T90Plus, Brookhaven, USA).

Optical microscopy of primary culture incubated with UCNPs

The sample represented fixed primary cultures treated or untreated with surface-modified UCNPs sandwiched between a microscope glass slide and 170 μm coverslip. The sample was mounted on an inverted confocal laser-scanning microscope (Zeiss LSM 710, Germany) with a C-Apochromat 63×/1.20 W Corr UV-VIS-IR M27 objective lens. The distribution of UCNPs was imaged by capturing their emission using the excitation at 980 nm from a femtosecond pulsed Ti:Sapphire laser (Chameleon vision II, Coherent Inc, CA) in the spectral range of 371–531 nm (corresponding to the blue and green emission of NaYF₄:Yb³⁺:Tm³⁺/NaYF₄ nanoparticles) set by a bandpass interference filter. A confocal pinhole of 1 airy unit was used to obtain axial optical slices of the axial resolution of 1.6 μm.

Ethics statement

All experimental protocols in this research were reviewed and approved by the Bioethics Committee of Nizhny Novgorod State Medical Academy; experiments were conducted in strict accordance with Act 708n (23.08.2010) the National Ministry of Public Health of Russian Federation approving the rules of laboratory practice for the care and use of laboratory animals and Council Directive 2010/63EU of the European Parliament (22^nd of September, 2010) on the protection of animals used for scientific purposes.

Cell cultures

Hippocampal cells were dissociated from embryonic mice (E18) and plated at high initial density of approximately 9000 cells per mm² on coverslips pre-treated with adhesion-promoting molecules (Sigma P3143). We used such high cell density to reproduce the in vivo tissue conditions and enable long-lasting recordings of the network activity. C57BL6J mice were sacrificed by the cervical vertebra dislocation. Embryos were surgically removed and decapitated. The entire hippocampi were exposed to mechanical treatment in phosphate-buffered saline (PBS), followed by enzymatic digestion for 25 min with 0.25% trypsin (Invitrogen 25200-056) at 35.5 °C. Then the cells were carefully stirred (10 passes) using a 1 mL pipette tip to obtain a suspension, which was centrifuged at 1000× rpm for 3 min. The resultant cell pellet was immediately re-suspended in Neurobasal™ medium (Invitrogen 21103-049) with 2% B27 (Invitrogen 17504-044), 0.5 mM L-glutamine (Invitrogen 25030-024), and 5% fetal calf serum (PanEco K055). The dissociated cells were seeded in a 25 μL droplet at the centre of the coverslips, forming a dense monolayer^30–32 and leave the cells for 2 h until adhered to the bottom, then fill the dishes with medium. After 24 h, the cultural medium was replaced by a medium containing Neurobasal™ medium with 2% B27, 1 mM L-glutamine, and 0.4% fetal calf serum without any antibiotics or antimycotics. The glial growth was not suppressed, because glial cells were essential for the long-term culture maintenance. One-half of the medium was changed every 2 days. The cells were maintained at temperature of 35.5 °C, 5% CO₂ and the saturating humidity in a cell culture incubator (MCO-18AIC, Sanyo). We note that the incubation of upconversion nanoparticles under test with fetal calf serum results in the formation of protein corona on these nanoparticles, which mediates the nanomaterial interaction with cells. Therefore, the choice of the culture medium was partly governed by our intent to more fully reproduce the physiological context of the brain, including the mediating effect of the protein corona.

Phase-contrast images of the cultures were taken weekly to record the culture status using a DMIL HC (Leica, Germany) inverted microscope with a 10×/0.2 Ph1 objective. UCNPs of the concentration of 0.8 mg mL⁻¹ were incubated with the culture medium on the day 14 of the culture development in vitro (DIV). The spontaneous calcium activity was recorded on the 3^rd day, following the UCNP incubation.

Cell viability detection

The viability of dissociated hippocampal cells was evaluated according to the percentage ratio between the number of dead cells stained by propidium iodide (Sigma, P4170, Germany) and the total number of cells stained by bisBenzimide (Invitrogen, H3570, USA) 3 days after UCNPs application.³²

Immunocytochemistry

Staining. The cultured cells were fixed for 15 min in 4% formaldehyde containing PBS (pH 7.4). In order to increase the permeability of the plasma membranes and antibody's binding efficiency with intracellular antigens, the cultures were treated with ice-cold methanol for 3 min. Next, the fixed samples were treated with 0.1% Triton X-100 (Sigma 93443-100ML) and 2% bovine serum albumin (BSA) for 30 min. Subsequently, the cells were incubated for 2 h at room temperature in PBS containing 1% BSA and the appropriate mixture of the primary antibodies: mouse monoclonal MAP2a + MAP2b (ab36447, Abcam) to stain neurons and chicken polyclonal GFAP (ab4674, Abcam) to stain astrocytes. After washing in PBS, the samples were incubated for 2 h at room temperature with the following secondary antibodies: rabbit anti-mouse conjugated Alexa Fluor 555 (A21427, Invitrogen) and goat anti-chicken conjugated Alexa Fluor 647 (ab150171, Abcam).

The immunostained cultures were examined under the confocal laser-scanning microscope (Zeiss LSM 710, Germany), as described above. The laser intensity, gain and offset were fixed for each acquisition. Quantitative evaluation was performed using ImageJ (Research Service Branch, NIH).

Ca²⁺ imaging

In order to carry out the functional calcium imaging, Oregon Green 488 BAPTA-1AM (OGB-1) (0.4 μM, Invitrogen, O-6807, USA) was dissolved in dimethylsulfoxide (DMSO) (Sigma, D8418, Germany) with 4% Pluronic F-127 (Invitrogen, P-3000MP, USA) and then added to the culture medium for 40 min. After incubation for 45 min to allow for the cell complete uptake of OGB-1, the cultures were washed for 15 min with dye-free medium. The confocal laser-scanning microscope (Zeiss LSM 710, Germany) with EC Plan-Neofluar 20×/0.50 M27 objective lens was used to visualize spontaneous calcium activity of the primary cultures. The distribution of cytosolic Ca²⁺ was imaged by means of its fluorescence probe OGB-1 using the excitation at 488 nm from an argon laser and emission in the spectral range of 505–663 nm set by a bandpass interference filter. Time series of 512 × 512-pixel images (field of view, 424 μm × 424 μm) were recorded at a rate of 1.3 frames per second. A confocal pinhole of 1 airy unit was used to obtain the axial optical slice resolution of 1.6 μm. Quantitative evaluation of Ca²⁺ transients was performed off-line using a custom-made software in C++ Builder. Cell regions of interest were manually selected using the original fluorescent images. The Ca²⁺ fluorescence of each cell in each frame was calculated as an average fluorescence intensity (relative units from 0 to 255) of the pixels within the selected cell region. Single Ca²⁺ signals were identified using the following algorithm. Firstly, each trace from all of the cells was filtered by averaging two neighboring points in the sample set. Secondly, we calculated the signal derivative by evaluating the difference between each pair of the consequent points. The pulses were identified from the derivative of the trace using a threshold detection algorithm. The left- and right-most points whose trace derivative values exceeded the threshold were taken as the pulse start and end points, respectively.^32,33

Statistical analysis

All data quantification is presented as the mean ± standard error of the mean (SEM). Statistical analysis was performed using a two-way ANOVA implemented in the SigmaPlot 11.0 program (Systat Software Inc.). Student–Newman–Keuls (SNK) was used as a post hoc ANOVA test. The difference between groups was considered significant if the p-value was less than 0.05.

Results and discussion

During the development period, primary hippocampal neural networks undergo substantial reconfigurations in stages characterized by morphological and functional changes, which have little effect on the viability of the constituent cells. The percentage of dead cells in the primary hippocampal culture of DIV 14 was estimated to be 3.2 ± 0.4% cells³¹ and attributed to neurons redundant in the network activity. The percentage of dead cells remains constant for 30 days under stable conditions. When these conditions are changed, for example, because of introduction of extraneous nanomaterial, such as upconversion nanoparticles, the viability of the hippocampal culture needs to be evaluated, and this evaluation is addressed in this study.

The fourteen-days primary hippocampal cultures were incubated for 72 h with the tested colloidal nanomaterial, which represented surface-modified UCNPs in PBS. The choice of the incubation time of 72 h was governed by unusually low metabolic rate of the dissociated cells, especially, neurons in comparison with the established cell lines. Moreover, according to our previous studies, neurons of the hippocampal cultures were most susceptible to unfavorable conditions within the first 72 h after treatment.³²

The composition of the core/shell UCNPs (NaYF₄:Yb³⁺:Tm³⁺/NaYF₄) was selected as the most popular^3,16, whereas the incubation concentration of 0.8 mg mL⁻¹ was chosen as typical for cytotoxicity tests.²² Three types of the surface-modification were chosen. Small-molecular-weight TMAH was demonstrated to partly displace original oleic acid surface groups used for the coordination of the UCNP during its synthesis.²⁷ As a result, the as-synthesized UCNP surface groups were stripped off making nanoparticles (UCNP-TMAH) bare and miscible with water and cultural medium. This surface-modification procedure was particularly valuable, because it allowed testing cytotoxic properties of the upconversion nanomaterial itself decoupled from the coating material properties, as it was tested in a number of the cytotoxicity tests.²⁰ The UCNP surface modification with amphiphilic polymer represented by polymaleic anhydride octadecene (PMAO) is a preferred approach demonstrated by Pokhrel et al.³⁴ and us.²² PMAO molecules bind to the hydrophobic tails of oleic acid functional groups on the UCNP surface, so that the molecule hydrophilic heads are exposed outwards, thus hydrophilizing the nanoparticle. The cross-linking of PMAO molecules consolidates the coating and shields the UCNP core from environmental exposure, so that the cytotoxicity properties of this biohybrid nanocomplex (UCNP-PMAO) are determined by the PMAO alone. Coating of the nanoparticles with positively-charged polymer, polyethyleneimine (PEI) represents another widely accepted strategy used, for example, in polyelectrolyte coating procedures.^22,35 It is well known that positively charged polymers, and specifically PEI, are efficient for cell internalization and transfection, although exhibiting notable cytotoxicity. According to our reported results on the surface-modified cytotoxicity of UCNPs in human skin cells, the least and most cytotoxicity were found in the case of the keratinocytes treatment with UCNP-TMAH and UCNP-PEI, respectively.²²

The cytotoxicity test results are shown in a histogram of Fig. 1A. A significant decrease in the number of viable cells 72 h post-treatment was observed for all tested UCNPs. The lowest cytotoxicity was detected in case of the DIV 17 culture incubation with UCNP-TMAH, where the percentage of viable cells dropped to (57 ± 9)%, whereas (97 ± 1)% of viable cells were detected in the control group. The greatest cell loss (22 ± 8)% was noted in the primary cultures treated by UCNP-PEI. The UCNP-PMAO material also caused a decrease in the number of viable cells to (47 ± 9)%. Therefore, the UNCP-PEI material exhibited the greatest toxic effect compared with the other studied nanomaterials.

Fig. 1 Viability cell determination in primary hippocampal cultures (A) a histogram of the primary hippocampal culture DIV 17 cell viability presented in terms of the percentage of viable cells (* statistically significant, p < 0.05) 72 h after the incubation with upconversion nanoparticles (UCNPs). UCNP-TMAH, UCNP-PMAO and UCNP-PEI stand for UCNP surface-modified with tetramethyl ammonium hydroxide, polymaleic acid octadecene and polyethyleneimine, respectively. (B, C, D and E) Bright-field (left column) and fluorescence (central and left columns) microscopy images of the primary hippocampal cultures incubated with control, UNCP-TMAH, UNCP-PMAO and UNCP-PEI, respectively. Central and right columns show fluorescence microscopy images of the cells stained with bisBenzimide (blue) and propidium iodide (red) to identify viable and dead cells, respectively. Scale bar, 100 μm.

In order to gain insight into the mechanisms of the observed cytotoxicity of UCNPs, we carried out detailed analysis of the culture responses to the UCNP-induced stress, including morphological and electrophysiological evaluations of the culture status. The primary hippocampal culture is heterogeneous in terms of cell types, with the most abundant populations being neurons and astrocytes. The neurons are more sensitive to toxic agents than the astrocytes. The cell shape changes observed in the bright-field microscopy images of the dissociated hippocampal cells, as shown in Fig. 1B–E (left column), represented a notable manifestation of the UCNP-induced morphological changes of the cells, where the neuron reshaping was particularly obvious (shown by circles). The partial dendrites reduction and rounding bodies of neurons was observed.

We performed immunocytochemical staining of neurons and astrocytes with Microtubule Associated Protein 2 (MAP2) and Glial fibrillary acidic protein (GFAP), respectively to demonstrate the morphological changes of primary hippocampal cultures in more details, with the results shown in Fig. 2. The choice of MAP2 as the neuron marker allow to visualize large-molecular-weight proteins (∼70 kDa) of the cytoskeleton. Only neurons having such proteins are able to demonstrate the normal network activity. The displacement of the large-molecular-weight proteins with small-molecular-weight proteins was a hallmark of the neurodegeneration at the cellular level, and visualized via the depletion of the MAP2 staining. In comparison with the control cell culture exhibiting normal functionality (see Fig. 2A), noticeable disruptions of the neural brunches (dendrites) were clearly observable in the primary cultures incubated with UCNP-TMAH and UCNP-PMAO, as shown, respectively, in Fig. 2B and C. Moreover, the neuron interaction with the UCNP-PEI nanomaterial caused severe shortening of the brunches, eventually leading to their complete disintegration (Fig. 2D). This disintegration was visualized as the profound fragmentation of the initially continuous MAP2-stained neural network (cf. Fig. 2A). The important indicators of the neuron functionality, such as the neurite length, number of branches and the spine density on neurons, were undetectable due to the convoluted neuronal growth pattern typical for the high cell density, which was required to form the functional networks.

Fig. 2 Immunocytochemistry-stained and fluorescent microscopy image panels of hippocampal primary cell cultures incubated: (A) without UCNPs; (B, C and D), with UCNP-TMAH, UCNP-PMAO and UCNP-PEI materials, respectively. Each panel displays fluorescent images of: neurons (green) and astrocytes (red), immunocytochemistry-stained with MAP2 and GFAP antibodies, respectively; fluorescence image of UCNP distribution (blue) in the culture and merged images. These images are correspondingly stained MAP2, GFAP, UCNP- and merged. Scale bar, 20 μm.

Changes in the shape of astrocytes in the cultures treated by UCNP-PEI were also revealed. In the control sample (Fig. 2A), astrocytes were cottony-shaped, distinctly isolated, connecting to each other via brunch terminal tips. They formed a relatively ordered evenly distributed array, as it is schematically presented in Fig. 3A1 and shown in Fig. 2A. The mean separation between the astrocytes was measured to be (44 ± 3) μm (see Fig. 3B). In the UCNP-PEI sample, astrocyte shapes were changed to take aberrant elongated forms. The uniform cell array pattern was also changed to a skewed pattern, with the mean astrocyte separation reduced to (22 ± 7) μm in the plane transverse to the elongation axis, as schematically presented in Fig. 3A2 and discernable in Fig. 2D. The astrocytes displayed a tendency for clumping, forming conglomerates with almost indistinguishable individual cellular elements. These changes of the astrocyte architecture are schematically illustrated in Fig. 3.

Fig. 3 The effect of the surface-modified UCNPs on the astrocyte morphology in the primary hippocampal cultures. A simplified schematic diagrams of the astrocyte networks under normal (A1) and (A2) UCNP-affected conditions. Note the mean distance between the nucleus centers of adjacent astrocytes (denoted l) was reduced in (A2) in comparison with (A1) in the plane transverse to the elongation axis. (B) A histogram of the mean l–s in the DIV 17 cell cultures evaluated in control, UCNP-TMAH, UCNP-PMAO and UCNP-PEI samples; * statistically significant, p < 0.05.

The UCNP-TMAH and UCNP-PMAO colloidal nanomaterials induced morphological modifications in the primary cultures. The astrocyte mean separations were markedly reduced in all directions, with l–s measured to be (17 ± 1) μm and (24 ± 3) μm for UCNP-TMAH and UCNP-PMAO, respectively, incomparison with l = (46 ± 2) μm in the control group (see Fig. 3). Note that these changes are typical for the primary cultures undergoing stress exposure, which corroborates the toxic effects of the tested nanoparticles.

Since the cultural medium contained fetal calf serum, protein corona was formed on UCNPs, which, mediated the nanomaterial interaction with cells and hence affected the cytotoxic effects exhibited by the tested surface-functionalized UCNPs.

Our next test related to the important functional activity of neural networks in the primary hippocampal cultures. Investigation of the functional metabolic calcium activity of the neuron-glial networks was conducted with the calcium-sensitive dye Oregon Green 488 (OG) application, which is widely used in studies concerning the dynamics of calcium homeostasis in nervous cells.^36,37

Beginning from the days 5–7 of the brain development in vitro, single calcium events are detected in cell networks. Calcium oscillations from different cells are random, asynchronous and rare with the percentage of cells exhibiting the calcium activity less than 12%. Functionally mature neural networks are formed in DIV 14 cell cultures, as reported by us.^30,31 There are calcium events characterized by frequency of 2.2 ± 0.3 oscillations per min of the duration of 6.9 ± 0.5 s in cultures. 80% of the cells demonstrate synchronized network activity. The process of the functional network activity formation was demonstrated to be associated with the appearance of complex chemical synapses in the primary hippocampal cultures (detailed in ESI†).

We investigated the spontaneous calcium activity in the primary hippocampal cultures incubated with the surface-modified UCNPs for 72 h. The results are presented in Fig. 4. The percentage number of cells exhibiting spontaneous calcium activity in the control cultures was approximately 90%, the duration of Ca²⁺ oscillations 7.9 ± 0.2 s, Ca²⁺ oscillation frequency 1.82 ± 0.07 min⁻¹. The introduction of the tested nanomaterial to the culture caused significant changes in the calcium activity. We observed a decrease in the number of cells exhibiting spontaneous calcium activity in all experimental groups (UCNP-TMAH, 17 ± 3%; UCNP-PMAO, 6 ± 3%). Moreover, the cells in the primary culture incubated with UCNP-PEI exhibited no calcium activity (Fig. 4). The detailed information on the presented calcium imaging, in addition to video clips of the spontaneous activity are provided in ESI.†

Fig. 4 UNCP-induced changes in the spontaneous calcium activity in the primary hippocampal cultures. (A) One frame from the time-lapse series of the cell culture stained with calcium-dependent fluorescent dye (oregon green). (B) Spontaneous Ca²⁺ oscillation recordings in neurons marked by circles in (A) (control). Temporal raster of the spontaneous calcium activity of (C) – control, (D, E and F) – UNCP-TMAH, UNCP-PMAO and UNCP-PEI treated cultures, respectively. (G, H and I) – Histograms of the culture parameters versus tested cell groups: proportion of cells exhibiting calcium activity; (H) – number of Ca²⁺ oscillations per min, (I) – duration of Ca²⁺ oscillations, respectively (* statistically significant, p < 0.05).

The UCNP-TMAH and UCNP-PMAO nanomaterials affected on the functionality of the primary hippocampal cultures. In particular, the duration and frequency of Ca²⁺ oscillations were changed markedly. The Ca²⁺ oscillation frequency was reduced to (1.3 ± 0.3) min⁻¹ and (1.2 ± 0.1) min⁻¹ in UCNP-TMAH and UCNPs-PMAO experimental groups. The culture treatment with UCNP-TMAH led to a negligible increase in the duration of Ca²⁺ oscillation (7.7 ± 0.7) s, whereas the UCNP-PMAO treatment caused a decrease this parameter to 5.0 ± 0.5 s. Interestingly, the TMAH-UCNP-treated neural network temporal signal exhibited a pulse-burst pattern akin to aged primary cell cultures, in contrast with a coalescent single-pulse signal of the normal neural network (see details in ESI†).

The functional calcium homeostasis alterations is associated with the neural networks destruction in primary hippocampal cultures. The neural structure formation and a short-term opening of calcium channels on the postsynaptic membrane are formed a periodic synchronized activity during the culture development in vitro.^38–40 A spontaneous short-term increase of the intracellular calcium leads to the cellular buffer systems activation and its concomitant reduction in cells. UCNP-induced destruction of the neural networks results in the irreversible degradation of their functional activity.

Our studies point to the toxicity of upconversion nanoparticles associated with the coating, rather than the UCNP core, although the core-induced cannot be excluded.

Several comments are now in order. Biohybrid photoluminescent nanocomplexes offer new opportunities in neuroscience and neurotechnology. The exceptional optical properties of these nanocomplexes allow ultrahigh-sensitivity imaging down to the single biomolecule level, paving a way for investigation of the ligand–receptor interactions in neurons towards new drug design. For example, labelling GABA receptors with quantum dots has enabled investigation of the dynamics of neurotransmitter receptors.⁴¹ The high loading capacity of the nanocomplexes is instrumental for targeted drug delivery and concomitant fluorescence-guided surgery of brain tumors.¹² These applications demand investigation of the toxicity of the emerging biohybrid photoluminescent nanocomplexes and development of strategies to ameliorate their potential adverse effects to brain cells and tissues.

As demonstrated in this study, bare and surface-coated UCNPs exerted stress on dissociated hippocampal cells manifested by the disruption of the signaling pathways, aberrant morphology and considerable cell loss. It appeared that bare UCNPs displayed the least cytotoxicity, which lent itself several mitigation strategies. Firstly, it seems reasonable to design of UCNP surface coating, which predominantly determines the entire complex cytotoxicity property, rationally using materials of proven non-toxicity, such as gold.⁴² Surface coating UCNP with silica represents an interesting approach to ameliorate the toxicity effects. Secondly, it seemed that the cytotoxicity of UCNPs depended on the level of the cell uptake, as can be inferred from the UCNP integral distribution in Fig. 2. Fig. 2B and D display the least and most UCNP uptake levels for the respective bare UCNP and PEI-UCNP treatments of the hippocampal culture, which also correlated with the assayed least and most cytotoxicity of these surface-modified UCNPs. Besides, the uptake was largely diminished in case of the primary culture incubation with large size sub-microns (>200 nm) UCNPs of the same surface modifications (results not shown). This was accompanied by the negligible cytotoxicity (see ESI†). This observation suggests the decoration of UCNP with surface moieties, e.g. polyethylene glycol, which hamper the particle internalization.⁴³

Conclusion

In summary, we first performed systematic cytotoxicity studies of upconversion nanoparticles in primary brain cultures. In comparison to cell lines, primary hippocampal cells are fully differentiated and not genetically transformed, thus the effects, demonstrated on such model, are most adequately reflected to real conditions. UCNPs were surface-modified with amphiphilic polymaleic anhydride octadecene (PMAO); positively-charged polymer, polyethyleneimine (PEI); and tetramethyl ammonium hydroxide (TMAH), rendering UCNPs bare. All three types of UCNP exhibited noticeable cytotoxicity measured as (57 ± 9)%, (39 ± 9)% and (36 ± 8)% of the cell viability percentage, respectively, benchmarked against the UCNP-untreated control. The differential immunocytochemical staining of neurons and astrocytes showed profound morphological modification of all TMAH-, PMAO- and PEI-UCNP-treated cells. The spontaneous calcium activity in the cells was also dropped considerably, and completely inhibited in the PEI-UCNP-treated cultures. Bare and polymer surface-coated UCNPs appeared to be cytotoxic to dissociated hippocampal cells evident through aberrant morphological changes, reduction of the Ca²⁺ activity, and cell loss. Since the surface coating of the tested nanoparticles largely determined their cytotoxic properties, rational design of non-toxic surface coating becomes an important research goal to reinstate feasibility of UCNP-based theranostics applications in brain. Therefore, the surface coating of upconversion nanoparticles depending on their application area should be noticed. Development of a new non-toxic coatings will open new possibilities in molecular neuro diagnostics.

Conflict of interest

The authors declare no conflict of interest.

Acknowledgements

This work was supported by Grant to the Government of the Russian Federation (Megagrant), grant number 14.Z50.31.0022 (in part of the chemical syntheses and surface modification of the UCNPs); state project “Provision Scientific Research”(in part of cell culture studies). We wish to acknowledge partial support from the Australian Research Council Centre of Excellence for Nanoscale BioPhotonics (CNBP) CE140100003.

References

1. S. A. Hilderbrand and R. Weissleder, Curr. Opin. Chem. Biol., 2010, 14, 71–79.
2. A. M. Edmonds, M. A. Sobhan, V. K. A. Sreenivasan, E. A. Grebenik, J. R. Rabeau and E. M. Goldys, et al., Part. Part. Syst. Charact., 2013, 30, 506–513.
3. S. W. Wu, G. Han, D. J. Milliron, S. Aloni, V. Altoe and D. V. Talapin, et al., Proc. Natl. Acad. Sci. U. S. A., 2009, 106, 10917–10921.
4. A. M. Schrand, S. A. C. Hens and O. A. Shenderova, Crit. Rev. Solid State Mater. Sci., 2009, 34, 18–74.
5. Q. Liu, Y. Sun, T. Yang, W. Feng and C. Li, et al., J. Am. Chem. Soc., 2011, 133, 17122–17125.
6. C. Vinegoni, D. Razansky, S. A. Hilderbrand, F. W. Shao, V. Ntziachristos and R. Weissleder, Opt. Lett., 2009, 34, 2566–2568.
7. A. Nadort, V. K. A. Sreenivasan, Z. Song, E. A. Grebenik, A. V. Nechaev and V. A. Semchishen, et al., PLoS One, 2013, 8, 1–13,  DOI:10.1371/journal.pone.0063292.
8. Y. Q. Lu, J. B. Zhao, R. Zhang, Y. J. Liu, D. M. Liu and E. M. Goldys, et al., Nat. Photonics, 2014, 8, 33–37.
9. V. K. A. Sreenivasan, E. A. Ivukina, W. Deng, T. A. Kelf, T. A. Zdobnova and S. V. Lukash, et al., J. Mater. Chem., 2011, 21, 65–68.
10. M. A. C. Stuart, W. T. S. Huck, J. Genzer, M. Muller, C. Ober and M. Stamm, et al., Nat. Mater., 2010, 9, 101–113.
11. M. K. Yu, J. Park and S. Jon, Theranostics, 2012, 2, 3–44.
12. B. Chertok, B. A. Moffat, A. E. David, F. Yu, C. Bergemann and B. D. Ross, et al., Biomaterials, 2008, 29, 487–496.
13. L. Lilge, M. Portnoy and B. C. Wilson, Br. J. Cancer, 2000, 83, 1110–1117.
14. S. Gioux, H. S. Choi and J. V. Frangioni, Mol. Imaging, 2010, 9, 237–255.
15. K. Kuningas, T. Rantanen, T. Ukonaho, T. Lovgren and T. Soukka, Anal. Chem., 2005, 77, 7348–7355.
16. H. X. Mai, Y. W. Zhang, R. Si, Z. G. Yan, L. D. Sun and L. P. You, et al., J. Am. Chem. Soc., 2006, 128, 6426–6436.
17. C. Wang, L. Cheng and Z. Liu, Biomaterials, 2011, 32, 1110–1120.
18. Y. Chen, K. Ai, Y. Liu and L. Lu, ACS Appl. Mater. Interfaces, 2014, 6, 655–663.
19. Y. Min, J. Li, F. Liu, E. K. L. Yeow and B. Xing, Angew. Chem., Int. Ed., 2014, 53, 1012–1016.
20. J. Zhou, Z. Liu and F. Li, Chem. Soc. Rev., 2012, 41, 1323–1349.
21. G. K. Das, D. T. Stark and I. M. Kennedy, Langmuir, 2014, 30, 8167–8176.
22. A. E. Guller, A. N. Generalova, E. V. Petersen, A. V. Nechaev, I. A. Trusova and N. N. Landyshev, et al., Nano Res., 2015, 8, 1546–1562.
23. T. Y. Cao, Y. Yang, Y. A. Gao, J. Zhou, Z. Q. Li and F. Y. Li, Biomaterials, 2011, 32, 2959–2968.
24. J. Wang, R. Wei, G. Bou and Z. Liu, Biochem. Biophys. Res. Commun., 2015, 456, 743–749.
25. M. Jelitai, M. Anderova, A. Chvatal and E. Madaras, J. Neurosci. Res., 2007, 85, 1606–1617.
26. R. Yuste, Nat. Rev. Neurosci., 2015, 16, 487–497.
27. A. N. Generalova, I. K. Kochneva, E. V. Khaydukov, V. A. Semchishen, A. E. Guller and A. V. Nechaev, et al., Nanoscale, 2015, 7, 1709–1717.
28. M. Li, O. Rouaud and D. Poncelet, J. Pharm., 2008, 363, 26–39.
29. E. A. Grebenik, A. Nadort, A. N. Generalova, A. V. Nechaev, V. K. A. Sreenivasan and E. V. Khaydukov, et al., J. Biomed. Opt., 2013, 18, 1–10,  DOI:10.1117/1.jbo.18.7.076004.
30. S. M. Potter and T. B. DeMarse, J. Neurosci. Methods, 2001, 110, 17–24.
31. V. Pasquale, P. Massobrio, L. L. Bologna, M. Chiappalone and S. Martinoia, Neuroscience, 2008, 153, 1354–1369.
32. M. V. Vedunova, T. A. Mishchenko, E. V. Mitroshina and I. V. Mukhina, Oxid. Med. Cell. Longevity, 2015, 1–9,  DOI:10.1155/2015/453901.
33. M. Vedunova, T. Sakharnova, E. Mitroshina, M. Perminova, A. Pimashkin and Y. Zakharov, et al., Front. Cell. Neurosci., 2013, 7, 1–10,  DOI:​10.3389/fncel.2013.00149.
34. M. Pokhrel, L. C. Mimun, B. Yust, G. A. Kumar, A. Dhanale and L. Tang, et al., Nanoscale, 2014, 6, 1667–1674.
35. G. B. Sukhorukov, E. Donath, H. Lichtenfeld, E. Knippel, M. Knippel and A. Budde, et al., Physicochem. Eng. Aspect, 1998, 137, 253–266.
36. N. Jaafari and M. Canepari, J. Physiol., 2016, 594, 967–983,  DOI:10.1113/jp271830.
37. U. Olcese, G. Iurilli and P. Medini, Neuron, 2013, 79, 579–593,  DOI:10.1016/j.neuron.2013.06.010.
38. O. M. Shirokova, L. E. Frumkina, M. V. Vedunova, E. V. Mitroshina and Y. N. Zakharov, et al., Modern Technologies in Medicine, 2013, 5(2), 6–13.
39. D. A. Wagenaar, J. Pine and S. M. Potter, BMC Neurosci., 2006, 7(11), 1–18,  DOI:10.1186/1471-2202-7-11.
40. D. Ito, H. Tamate, M. Nagayama, T. Uchida and S. N. Kudoh, et al., Neuroscience, 2010, 171, 50–61.
41. H. Bannai, S. Levi, C. Schweizer, T. Inoue, T. Launey and V. Racine, et al., Neuron, 2009, 62, 670–682.
42. R. Gromnicova, H. A. Davies, P. Sreekanthreddy, I. A. Romero, T. Lund and I. M. Roitt, et al., PLoS One, 2013, 1–10,  DOI:10.1371/journal.pone.0081043.
43. T. A. Kelf, V. K. A. Sreenivasan, J. Sun, E. J. Kim, E. M. Goldys and A. V. Zvyagin, Nanotechnology, 2010, 21, 1–14,  DOI:10.1088/0957-4484/21/28/285105.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra01272h

---