Covalent conjugation of (bis)phosphonate group-containing molecules, sodium Alendronate (Aln) and 3-AminoropylPhosphoric Acid (ApA), to Cellulose nanocrystals (CNCs) was performed via oxidation/Shiff-base reaction. Further fluorescent labelling with Rhodamine B Iso ThioCyanate (RBITC) was performed to follow CNCs interaction and potential internalization with/in human osteoblasts by confocal microscopy. Complementary analyses were applied to identify the conjugation (Atenuated Total Reflectance–Fourier Transform Infrared and UV–VIS spectroscopies), physico-chemical (Dynamic Light Scattering and Nanoparticle Tracking Analysis) and morphological (Transmission Electron Microscopy) features of native and ApA/Aln-modified CNCs in physiologically relevant environments (Phosphate Buffer Saline, Advanced Dulbecco’s Modified Eagle Medium). While conjugation did not affect the CNCs` size, the RBITC-labelling promotes their aggregation. Faster (1 h vs. 2 h) uptake by osteoblasts of RBITC-CNCoxAln, compared to RBITC-CNCoxApA, and no-internalization (in 24 h) of native RBITTC-CNC, indicate a higher affinity of Aln-modified CNCs to the cells, while all CNCs (in 0.25–0.06 wt%) promote the cell growth. Aln/Apa-modified CNCs shows high potential in drug-delivery for bone therapies, and theranostics.
The online version of this article (doi:10.1007/s10570-017-1432-5) contains supplementary material, which is available to authorized users.
Introduction
Various nanoparticles (NPs) have been considered increasingly in the last decade for their usage as biosensors, imaging agents and drug delivery vehicles (Mahmoudi et al. 2012). Most of these applications require successful internalization of NPs into the targeted cells, therefore, deep understanding of the interactions between NP and biomolecules/cell membrane is an important prerequisite for designing and engineering NPs with intentionally enhanced or suppressed cellular uptake.
Internalization of NPs into cells can initialize the well-tolerated, non-cytotoxic, and even therapeutic (Panariti et al. 2012; Tsai et al. 2013) or conversely, cytotoxic (i.e. cell dead or decreased cellular activity) effect (Huerta-García et al. 2015). A variety of NPs have thus been demonstrated to be internalized by osteoblasts, such as calcium phosphate nano shells (Schmidt et al. 2008), implants´ wear debris (Lohmann et al. 2000), quantum dot/hydroxyapatite composites (Hsieh et al. 2009), polymeric particles (Tautzenberger et al. 2011). Gold NPs have been shown as an effective material with no osteogenesis potential, nor apoptosis to osteoblasts (the MG63 cell line) (Tsai et al. 2013). Conversely, the anatase titanium dioxide (the metallic implant debris-originated) was internalized readily by primary osteoblasts, and affected the adhesion strength, migration and proliferation of the cells negatively (Ribeiro et al. 2016).
Anzeige
Nanocellulose (NC) has also gained significant attention in pharmaceutical and biomedical applications recently, above all in wound healing (Napavichayanun et al. 2016), tissue engineering (Gorgieva et al. 2017; Markstedt et al. 2015; Brown et al. 2013), cell therapy (Lou et al. 2014), gene (Ndong Ntoutoume et al. 2017), drug delivery (Barbosa et al. 2016; Dong et al. 2014; Villanova et al. 2011; Lin et al. 2016), and diagnostics (Edwards et al. 2016). Such a wide application spectrum is related mainly to their nanometer-sized features, large surface area, specific biomechanical characteristics, surface chemistry, ease of conjugation, high biocompatibility, and low (if any) cytotoxicity (Alexandrescu et al. 2013) with tolerogenic potential to the immune system (Tomić et al. 2016). Due to a general acceptance as ˝biosafe˝ nanomaterial (Čolić et al. 2014; Tomić et al. 2016), the cellulose nanocrystals (CNCs), with typical sizes of <300 nm in length and around 10 nm in diameter (Habibi et al. 2010), have also been readily evaluated as catalysis (Zhou et al. 2013) in biomedical engineering (Sinha et al. 2015), as well as targeted drugs (Taheri & Mohammadi 2015) and gene (Hu et al. 2015) delivery. Recent studies also demonstrated the potential of CNCs to target tumours via the Enhanced Permeability and Retention (EPR) effect and delivery of organic compounds or drugs into cancer cells (Drogat et al. 2011). The minimal, nonspecific uptake of FITC-labelled CNCs have been thus reported for mammalian KB cells, mouse brain endothelial cells (bEnd.3), mouse macrophages (RAW 264.7), benign and malignant human breast cancer cells (MCF-10A, MDA-MB-231, MDA-MB-468), human prostate cancer cells (PC-3), and rat brain tumour cells (C6) (Dong et al. 2012). On the other hand, targeting of CNCs through folic acid conjugation leads to specific uptake by (folate-receptor positive) cancer cells (Dong et al. 2014). The Poly(2-(DiMethylAmino) Ethyl Methacrylate (PDMAEMA) brushes-modified CNCs are reported as a pDNA delivery in vitro as well as in vivo (Hu et al. 2015). Another study demonstrated the chlorin-PEI-labelled CNCs as potential photosensitizers with antitumour activity against the HaCat cell line (Drogat et al. 2012). On the other hand, the highly charged (carboxylated) CNCs (~3.8 mmol g−1) induced an exert of stress to different cell lines (i.e. cells from the human colon/Caco-2/, kidney/MDCK/, cervix/HeLa/and macrophage/J774/), causing significant (>75%) reduction in their mitochondrial activity (Hosseinidoust et al. 2015).
Bis-phosphonates are small (molecular size <300 Da) organic pyrophosphate analogue drugs, where two phosphates are connected by a carbon atom (P–C–P) with various side chains (Lezcano et al. 2014), being used in the treatment of several bone diseases, such as osteoporosis, Paget’s disease, fibrous dysplasia, hypercalcemia of malignancy, heterotopic ossification, osteogenesis imperfecta, multiple myeloma (Lewiecki 2010), etc. Recent studies have reported that Alendronate (Aln), the most commonly administered bisphosphonate, enhances osteogenic differentiation of bone marrow stromal cells (Kim et al. 2009), bone marrow mesenchymal stem cells (von Knoch et al. 2005), and adipose-derived stem cells (Wang et al. 2010). However, its fast dissolution at physiological conditions and side effects from oral administration lead to several attempts for its immobilization, as, for example, on a titanium implant surface (Moon et al. 2011) or within calcium phosphate microspheres (Kim et al. 2010), liposomes (Epstein et al. 2008), chitosan scaffolds (Kim et al. 2012), and poly (lactic-co-glycolic acid) nanoparticles (Thamake et al. 2012).
The aim of this study was, thus, to modify the surface of CNCs with selected (bis)phosphonate molecules and fluorescent labelling, and evaluate their interactions with human osteoblasts, as well as potential internalization within the cells, thus pointing to future application of such conjugates in bone therapies. For that purpose, complementary analytical techniques (considering different time-scale tracing capacities, different spatial resolution capacities, as well sensitivity profiles) were chosen in order to identify the conjugation and labelling (ATR–FTIR, UV–VIS), and to follow the physico-chemical (DLS and NTA) and morphological (TEM) features of native and (bis)phosphonate-modified and labelled CNCs in simulated physiological environments (PBS and ADMEM). The cell viability spectroscopic assessment and confocal microscopy (CM) imaging were performed on human bone derived osteoblasts in a 1–24 h incubation period with native and modified CNCs.
Experimental
Materials
The dispersion (pH ~6) of Cellulose nanocrystals (CNC) of around 5 nm in diameter and with length of 150–200 nm, was used as derived from raw wood pulp by sulfuric acid hydrolysis and provided by the University of Maine, USA. Fluorescent Rhodamine B IsothioCyanate dye (RBITC), N-Morpholino-EthaneSulfonic acid (MES), 3-AminoropylPhosphoric Acid (ApA), sodium Alendronate (Aln), sodium periodate, Phosphate Buffer Saline (PBS), 2,3-(4,5-diMethylThiazol-2-yl)-2,5-diphenylTetrazolium bromide) (MTT), Sodium Dodecyl Sulphate (SDS), glycerol and glycine were purchased from Sigma Aldrich, Germany, without further purification. Advance Dulbecco’s Modified Eagle Medium (ADMEM) and Foetal Bovine Serum (FBS) were obtained from ThermoFisher, Germany. Milli-Q water was used throughout all the experiments. All the chemicals were of analytical grade and used as received. Human bone derived osteoblasts (hFOB 1.19, ATCC CRL-11372TM, ATCC, UK) were used for the cell testing.
Anzeige
Cellulose nanocrystals` (CNCs) modification
CNCs oxidation
1.6 g of sodium periodate (NaIO4) was dissolved in 50 mL of distillate water and added to 50 mL of 4 wt% CNC dispersion. 100 mL of Milly-Q water was added to allow uniform mixing that was continued until a homogenous dispersion was obtained. The resulting dispersion was covered with aluminium foil and shaken for 24 h at room temperature. After this period, the dispersion was dialysed against Milly-Q water for 48 h. The small amount of resulting product was freeze-dried to obtain samples for FTIR analysis. The remaining CNC dispersion containing oxidised CNC (denominated as CNCox) was washed additionally with 0.1 M of acetate buffer (pH 4.5) and stored at 4 °C for further usage.
CNCs` functionalization with ApA and Aln molecules
0.45 g of 3-aminopropylphosphonic acid (ApA) and sodium alendronate (Aln) were dissolved in 1.5 mL of 0.1 M acetate buffer of pH 4.5, respectively. 0.5 mL of ApA-containing or Aln-containing solution was added to 100 mL of CNCox. The reference was prepared in parallel by using the same volume (100 mL, 1 wt%) of native CNC, which was mixed with 0.5 mL of ApA-containing or Aln-containing solution. All samples were left for continuous mixing for 24 h at 45 °C, while the reaction was followed by monitoring the pH. Finally, the products (CNCoxApA and CNCoxAln) were washed separately in Milly-Q water until the total removal of unreacted ApA (or Aln), which was followed spectrophotometrically as described below. A small amount of each sample was freeze-dried for the upcoming FTIR analyses.
CNCs` fluorescent labelling
In order to perform the real-time follow-up interaction/internalization of the CNCs with the cells by using Confocal Microscopy (CM) (Leica DMI6000 CSF, Germany), the CNCs were labelled with RBITC fluorescent dye according to the following procedure: 20 mg of RBITC dye was added to 25 mL of 2% w/w of CNC dispersion, followed by vigorous stirring and pH adjustment to 13 with the drop-wise addition of 1 M of NaOH. The resulting dispersion (RBITC-CNC) was incubated in the dark for 4 days, being afterwards dialysed against water by using a 12–14 kDa MWCO membrane (Spectra/Por®) for several days. Dialysis was followed spectroscopically at the absorbance of 552 nm. The degree of labelling (%) was calculated indirectly from the RBITC calibration curve, using the dialysis (washing) solutions.
For further experimental analysis, the final dispersions were provided in Milly-Q or physiological media (PBS or ADMEM) by simple centrifugation, withdrawal of primary media, addition of the same volume of respective medium and re-dispersion by using the ultrasonic bath.
CNCs characterisation
ATR–FTIR spectroscopy analysis
The spectra of the native and differently modified CNCs were recorded using a Perkin-Elmer Spectrum One FTIR spectrometer with a Golden Gate ATR attachment using a diamond crystal. The absorbance measurements were carried out within the range of 650–4000 cm−1, with 16 scans and a resolution of 4 cm−1. The Spectrum 5.0.2 software programme was applied for data analysis. All the measurements were carried out in duplicate by using two different samples from each of the compositions. In order to get additional information about the chemical composition of the samples in the complex carbonyl region (1500–1800 cm−1), a deconvolution of the peak of interest was performed by using the Peak Fit 4.12 software package for a nonlinear peak separation. A linear baseline was subtracted from this region section. The Savitzky-Golay algorithm method was applied further for smoothing of noisy data. The local minimas were inspected within the second derivative data, or resulting spectra, from where the hidden peaks were obtained. Finally, the new spectra was curve fitted with mixed, spectroscopic Gaussian and Lorentzian peak types up to obtaining r2 close to 0.9999. The position and area of newly obtained peaks were compared further.
Sodium Dodecyl Sulfate PolyAcrylamide Gel Electrophoresis (SDS-PAGE)
SDS-PAGE was applied to monitor the elution of unbound RBITC dye from the RBITC-labelled CNC sample, according to an adapted method described by (Tenuta et al. 2011). In brief, 600 µL of the loading buffer, prepared from 6% w/v of SDS and 30% of glycerol (in Milly-Q water), was added to 200 µL of each sample (RBITC-CNC). The running buffer was prepared from Tris (3.03 g), glycine (14.4 g) and SDS (1 g) in 500 mL of Milly-Q water. The control sample, being the RBITC dye solution, was prepared in parallel, using the same procedure as for the CNC-labelled dispersions. The analysis was performed on an OmniPAGE mini-SDS-PAGE system (Cleaver, Scientific) with a Consort EV 215 power supply and using prepared polyacrylamide gels (Precise, Thermo scientific, Germany). In each of 12 wells, 10 µL of the sample was loaded, and a power supply of 185 V was applied for a 1 h running period. After each run, the obtained gels were scanned under the Singen G: BOX (ThermoFisher, Germany), being equipped with GeneSys programme software.
UV–VIS spectroscopy analysis
The UV–VIS spectroscopic analysis of the selected samples was performed by using the spectrophotometer Tecan plate reader (Infinite 200). The absorbance scans were performed in the region of 450–600 nm, while the measurement was repeated in triplicate.
Dynamic Light Scattering (DLS) analysis
The zeta-potential and zeta-size analysis of native and differently modified CNC dispersions were carried by Zeta-sizer (Nano ZS ZEN360, Malvern Instruments Ltd., UK) at 20 ± 0.1 °C using the cell ZEN1010, applying the following parameters: A material refractive index of 1.47, dispersion refractive index of 1.33, and viscosity of 0.8872 cP (Gorgieva et al. 2015). A field of 40 V was applied across the nominal electrode spacing of 16 mm. The samples were prepared at concentrations of 0.005% w/v in Milly-Q water and zeta-potential was measured over a pH range from 3 to 11, being adjusted by using 0.1 M NaOH and 0.1 M HCl, respectively. The measurement of both the zeta-potential and zeta-size in PBS (pH 7.4) were performed in parallel. The average values, as well as the Standard Deviation of the mean values, were calculated from at least four individual measurements.
Nanoparticle Tracking Analysis (NTA)
The NTA combines laser light scattering microscopy with a highly sensitive Charge-Coupled Device (CCD) camera, which enables the visualization and recording/tracking of individual NPs in a solution, as well as their movement under Brownian motion related to their size according to the Stokes–Einstein formula (Filipe et al. 2010). The NTA measurements of native and differently modified CNC dispersions of 0.01% w/v concentration were performed at 20 ± 0.1 °C using a NanoSight LM10 instrument with finely focused (red) laser beams with a wavelength of 635 nm. The suspended CNCs, being subsequently injected below a quartz prism on the laser beam path, scattered the light in a manner easily visualised using a 20× magnification objective, which was fitted to a conventional optical microscope with a CCD camera operating at 30 frames per second. A video file of CNCs moving under Brownian motion within a field of approximately 100 µm × 80 µm × 10 µm was, subsequently, captured by the camera. For each sample, at least three runs were applied by keeping the instrument setup parameters constant in order to ensure measurement accuracy.
Transmission Electron Microscopy (TEM) analysis
TECNAI G2 20 TEM microscopy was used for size and morphology visualisation of CNCs before and after the modification processes. For that purpose, 10 µL of native and differently modified CNC dispersions of 0.001% w/v concentration were deposited on the carbon coated-copper grids (Formvar, Agar Sci.) and allowed to dry overnight. Analyses were carried out with an accelerating voltage of 120 kV and sample-dependent exposure time with up to 60.000× magnification.
Cell testing
Cell culture preparation
Osteoblast cells (hFOB 1.19, ATCC CRL-11372TM) were seeded on sterile circular coverslips within 24-well plates using a cell density of 30.000 cells/well, followed by incubation in a cell medium (ADMEM, containing 5 wt% of FBS), for 24 h at 34 °C in the presence of 5 wt% CO2. To follow the cell growth during cell culture preparation, the cell morphology was examined regularly using an inverted optical microscope (Zeiss Axiovert 40 CFL, Germany) prior to incubation with CNC dispersions.
Cell interaction study by Confocal Microscopy (CM)
Before the cell interaction study, CNC dispersions were sterilized by autoclaving for 30 min at 120 °C. The particular volumes (0.5–1.5 mL) of native or differently modified and RBITC-labelled CNC dispersions were pre-mixed with 3 mL of the cell culture medium (ADMEM containing 5 wt% of FBS), which results in 0.1 wt%, being followed by pipetting gently onto the cells prepared on coverslips as described above. The prepared samples were imaged every 5 min for up to 2 h using an upright Confocal Microscope (Leica TCS SP5 II, Germany). For this purpose, the as-prepared glass slides were positioned on a transparent glass holder above a 10× (dry) objective of the CM. The fluorophore (RBITC probe) was excited by argon laser (514 nm), while the emitted light was collected using a highly sensitive Leica Hybrid Detector (HyD) in accordance with the RBTIC probe (540–560 nm). Transmitted light images were captured in parallel using a Dodt detector (Leica) with the same excitation laser as for the fluorescence images in order to visualize non-labelled cells. For both the time series and the z stack, the same imaging settings were used, i.e. 1024 × 1024 pixels and line average 8. After completing the time series imaging, z stack was captured before and after 3× washing of the coverslips with cell medium (ADMEM containing 5 wt% FBS). For the native CNC dispersion, the fluorescent and transmitted images were also captured after 24 h of incubation. Analysis was done off-line using ImageJ software (MacBiophotonics). Z stacks were limited, such that single optical sections were permitted above and below cells. To assess the potential internalization of the RBITC-labelled CNCs by the cells, the z stacks were re-sliced based on the transmitted light information that enabled identification of the cells. Maximal projection for the fluorescent signal, transmitted light signal, and merged signal of the two are presented.
Cytotoxicity (viability) study
The cell viability was checked after their exposure for 48 h to various sample dilutions of native and differently modified and RBITC-labelled CNC dispersions, to assess the cell cytotoxicity. The dilutions were prepared using ADMEM containing 5 wt% of FBS based on binary dilution factors (1:2, 1:4, 1:8 etc.). A 48 h of examination period was chosen due to cell medium exchange (according to the manufacturer`s protocol), which could affect the results significantly (mainly through CNC loss), and, hence, lead to misleading conclusions. As a first control, pure Milly-Q water was used instead of CNC dispersions, while the second control was done by osteoblasts incubated only in ADMEM with 5 wt% of FBS, without any media dilution. For each sample, three parallels were performed. All samples were incubated at 34 ± 0.5 °C in the presence of 5 wt% of CO2.
The MTT cell proliferation assay was applied to measure the cell proliferation rate and, conversely, when metabolic events led to apoptosis or necrosis, the reduction in cell viability. The yellow MTT is reduced by metabolically active cells, in part by the action of dehydrogenase enzymes, to generate reducing equivalents, such as Nicotinamide Adenine Dinucleotide (NADH) and Nicotinamide Adenine Dinucleotide Phosphate (NADPH). The resulting intracellular purple formazan was quantified spectrophotometrically at 570 nm. The MTT reagent thus yields low background absorbance values in the absence of cells. For each cell type, the linear relationship between cell number and the signal produced is established, thus allowing an accurate quantification of changes in the rate of cell proliferation (van de Loosdrecht et al. 1994; Ferrari et al. 1990). An MTT cell proliferation assay was performed according to Mosmann 1983. In short, 100 µL of 0.5 wt% CNC dispersions were pre-diluted with ADMEM, containing 5 wt% FBS in 1:2, 1:4 and 1:8 dilutions, and were pipetted onto the prepared osteoblast cell cultures in respective wells in a 24 well plate. The 4 h incubation at 37 °C in the presence of 5°CO2 was performed, and the solution was afterword’s pipetted out and DMSO added for an additional 5 min to dissolve the formed violet crystals. Four repetitions were performed for each sample.
Results and discussion
Characterization of differently modified CNCs
The site-selective periodate oxidation on native CNCs allows cleavage of the C2–C3 bond of the glucopyranose ring, converting the respective vicinal hydroxyl into two aldehyde groups (Hlady and Buijs 1996; Kenawy et al. 2015), while leaving the bulk material crystallinity unchanged (Grate et al. 2015). The CNCs’ modification with Aln and ApA molecules [presented in Scheme 1, according to the literature (Dash et al. 2012)] was carried out using a Shiff-base coupling reaction between (bis)phosphonate amine (in ApA or Aln) and cellulose (CNC) aldehyde groups which are formed during the (periodate) oxidation process. Further RBITC-labelling of both native and ApA or Aln modified CNCs was performed according to the well-known nucleophilic addition mechanism of cellulose hydroxyls with RBITC’ isothiocyanate group (National Institute of Industrial Research (India) 2006) in a highly alkaline (NaOH) medium, thus forming a stable covalent bond (Vieira Ferreira et al. 1998).
Scheme 1
CNCs modification with Aln and ApA molecules through conjugation mechanisms via oxidation/Shiff base reaction pathway, and subsequent RBITC labelling. Reagents and conditions used: a NaIO4 at room temperature for 24 h; b Aln and c Apa at pH 4.5 and 45 °C for 24 h; d RBITC at pH 13 and room temperature for 4 days
×
The ATR–FTIR analysis (Fig. 1a) was performed to verify the success of CNCs` modifications. A typical spectral line corresponding to the cellulose can be observed for native CNCs (Gorgieva et al. 2015). Namely, the C–O stretching vibration of the cellulose glucose ring at 1031 cm−1, as well as the broad peaks centred at about 3338 and 2901 cm−1, correspond to –OH and C–H stretching vibrations, respectively. The presence of S–O and O–S–O vibration modes, observed at around 607 and 1247 cm−1, respectively, indicate the presence of sulphonate groups as a result of H2SO4 usage in the CNC preparation.
Fig. 1
a ATR–FTIR spectra and b Deconvoluted carbonyl (C=O) region of native (CNCnative), oxidised (CNCox) and ApA/Aln-conjugated (CNCoxApA/CNCoxAln) CNCs
×
Since the banding vibration of water at around 1640 cm−1 has a pronounced shielding effect over the carbonyl region (Qojewska et al. 2005), being important for the verification of the modification, the latter was deconvoluted using the PeakFit 4.12 programme (Fig. 1b). Through peak deconvolution, the oxidation of CNC by NaIO4 (CNCox) can be identified by the appearance of an additional vibration band at about 1717.7 cm−1. The related area under the curve for this peak was reduced from an initial 1.07 to 0.8 cm−1 and 0.5 cm−1 after conjugation with ApA and Aln, respectively. Besides the dominating water-related peak, the two peaks positioned in between 1575–1590 cm−1 and 1670–1675 cm−1 regions appeared in the deconvoluted spectra of all modified CNCs, most likely related to cellulose oxidation products. It has been shown that cellulose periodate oxidation gives rise to two characteristic FTIR bands located at about 1730 and 880 cm−1, that can be assigned to carbonyl groups and hemiacetal bonds between the aldehyde group and neighbouring hydroxyl group, respectively (Calvini et al. 2006). As the hemiacetal bond vibration band is missing in our case, the exhaustion of aldehyde groups may be due to the formation of Schiff-bases with ApA/Aln-related amine groups, rather than to possible reaction with cellulose hydroxyl groups. A possible explanation for the absence of the latter is most likely connected with the relatively labile double bond between the nitrogen of the amine and carbon on the aldehyde, the formation of which is most efficient at pH 9–10. Since the (bis)phosphonate coupling reaction was carried out at pH 4.5, which can lead to a lower reaction efficiency, it is rather straightforward that it is not present (at least not in observable quantities by the used analytical methods). Consequently, very low bands` intensity was observed at about 1575.2/1581.6 cm−1 and 1671.4/1672.8 cm−1 for the corresponding ApA/Aln-modified CNC samples. Due to the relatively low extent of the labelling (7% as determined by spectroscopic analysis), the accumulated FTIR spectra of the labelled sample did not show any changes related to the labelling reaction.
Non-covalent (i.e. physical) adsorption of fluorescent dye onto CNCs may limit their applicability in biological environments due to the potential release of the dye that may cause misleading interpretations of the results or even toxic effects. Therefore, the covalent coupling of RBITC (Vieira Ferreira et al. 1998) to the native and differently modified CNCs was also examined by using SDS-PAGE and UV–VIS spectroscopy. The SDS-PAGE (Supporting Information Fig. S1) confirmed the stability of the label, since the RBITC typical band appears at the gel’ channel bottom only for the reference (the RBITC dye). Around 150–200 nm long, CNCs were not able to penetrate through the gel channels and remained on the top along with the label giving the fluorescence bend. The obtained absorbance spectra for selected sample (CNCoxApA) (Supporting Information Fig. S2) indicated a shiff of the major RBITC related peak position from 552 nm (in RBITC dye as control) to 558 nm for the RBITC-CNCoxApA sample, as well as the appearance of an additional peak (shoulder) at a lower wavelength (528 nm) for the RBITC-CNCoxApA. Both changes can be attributed to the covalent attachment of dye to the CNCs (Mahmoud et al. 2010).
In order to confirm the CNC modification and labelling further, the zeta-potential analysis of CNC samples was performed in Milly-Q water and a pH range between 3 and 11, adjusted with 0.1 M NaOH or 0.1 M HCl. As seen from Fig. 2a, the zeta-potential of CNCs is reduced to −49.7 ± 11 mV after their oxidation at pH higher than 7, which could be related to the presence of a trace amount of surface sulphate and carboxylic groups (39.13 ± 2.1 mmol/kg as determined by potentiometric titration) due to the extraction process. In the case of ApA-conjugated CNCs (CNCoxApA), the inflection point at pH 5 and the slope of the curve by further increasing of pH, imply indirectly on the presence of ApA in which the secondary amine group (resulting from Shiff-base reaction) is getting protonated. Its’ zeta-potential is increased up to −52.6 ± 11 mV at pH 11, which is attributed to different ionization of phosphonate groups present in ApA (pH7 < pKa < pH9) (Gorgieva et al. 2017), as well as sulphate and carboxylic groups present on CNC (pH5 < pKa < pH7). Moreover, the net positive charge of RBITC decreases the negative zeta-potential additionally in all RBITC-labelled CNCs, thus providing an additional evidence for RBITC attachment on the CNC surface.
Fig. 2
DLS analysis of 0.005 wt% of native (CNCnative), oxidised (CNCox), and ApA/Aln-conjugated CNCs (CNCoxApa/CNCoxAln) before and after RBITC labelling in, a Milly-Q water at different pHs, adjusted with 0.1 M NaOH or 0.1 M HCl, and, b PBS medium of pH 7.4, at 20 °C ±0.1
×
To obtain reliable and reproducible correlation data for particular physicochemical properties of native and differently modified CNCs in regard to a possible cell response, an accurate characterization of the CNC was crucial. In fact, regardless of the specific internalization mechanism (e.g. phagocytosis, endocytosis, etc.), the cell-nanoparticle interactions are modulated mostly by the respective nanoparticle size, shape, surface charge and surface chemistry (Verma and Stellacci 2010), as well as differing according to the type of used cells and the fraction of these in the respective cell cycle phase they are in during the experiment (Mahmoudi et al. 2012). Considering this, a relevant and complementary characterization methodology was applied before the cell testing.
Although the evaluation of the size of elongated CNC structures by DLS analysis may not be an appropriate method, considering the limitations of this technique to determine sizes of non-spherical particles, it was used for comparison of native with differently modified samples, rather than for their size quantification. In addition, the zeta-potential analysis was performed to evaluate the surface charge of the samples as an important aspect within the biological milieu which affects the adsorption of dissolved biomolecules, thus building a so-called ‘protein corona’ through which they communicate with the cells (Lee et al. 2015). The analysis was done in a protein free PBS media of pH 7.4 in order to avoid the complex proteins interactions with CNCs and reveal the modification effect on both parameters, the size and the zeta-potential. The hydrodynamic volume of respective samples (Supporting Information Fig. S3) showed a mono-modal and relatively broad size distribution of CNCs. The average values of peaks (measured by intensity) for all CNC samples are presented in Fig. 2b (secondary y-axis). The results indicate a more than twice reduction in average size of CNCs after their oxidation (from 540 ± 30 to 215 ± 44 nm), whereas neither the size nor the zeta-potential (−15.1 ± 1 mV) changed significantly (compared to CNCox) after conjugation with ApA or Aln. On the other hand, the average size was increased again after RBITC labelling, being most pronounced in the case of RBITC-CNCoxApA (from 220 ± 1.4 to 811.1 ± 104.4 nm) resulting in a surface with zeta-potential of −19.9 ± 1.4 mV compared to RBITC-CNCoxAln of 462.8 ± 15.5 nm and −16 ± 1.9 mV. The reason for this is most likely in the most intensive aggregation of these samples, which could be related to the presence of hydrophobic moiety in ApA, as well as the shielding effect of static charges in ion-rich dispersion media like PBS.
Due to the capacity to resolve and measure accurately different RBITC-labelled CNC particles simultaneously in relevant cell medium (i.e. ADMEM) and in Milly-Q water as a reference, the NanoSight tool and respective NTA method were used in addition to DLS. The rationale behind selection of both media is the fact that ADMEM, as broadly used cell medium, presents a complex aqueous environment being significantly different (due to protein presence, and different ions content) from pure water, and also the step-up in complexity compared to PBS (Moore et al. 2015). In the case of Milly-Q water dispersions, the ionic repulsive forces dominate the attraction ones, allowing the occurrence of relatively stable dispersions, as presented by the scatter plots in Fig. 3a (left).
Fig. 3
a 2D intensity scatter plots from NTA measurement of 0.01 wt% of RBITC-labelled native (RBITC-CNCnative) and ApA/Aln-modified (RBITC-CNCoxApA/CNCoxAln) CNC dispersions in Milly-Q water (left) and ADMEM (right), being measured at 20 °C ±0.1 and constant time*. *Time referred to the period between dispersion mixing and measurement, btop Particle size/intensity scatter plots, middle Particle size/concentration plot, and bottom sample video frame captured by CCD camera, from the NTA analysis of differently concentrated and RBITC-labelled native CNC dispersion in ADMEM at 20 °C ±0.1 and constant time*. *Time referred to the period between dispersion mixing and measurement
×
The highest mean, as well as mode size value (inserted data), in the case of native RBITC-CNCs (308 ± 114 nm), compared with the Aln/Apa-modified RBITC-CNCs (218 ± 116/245 ± 159 nm), results from the thinning effect of periodate oxidation and further modification. On the other hand, a protein- and ions-rich ADMEM medium shields the charges of modified CNCs, initiating the aggregation process (306 ± 142 nm, Fig. 3a, right), being more significant in the case of ApA-modified CNCs (507 ± 260 nm) bearing a single phosphonate group per molecule compared to Aln-conjugated CNCs (368 ± 154 nm), but, however, less pronounced than in PBS (Fig. 2b) although following the same trend. It was demonstrated previously that ApA modification of cellulose nanofibers (CNFs) induces their aggregation (Gorgieva et al. 2017) which is related to the molecular structure of ApA, i.e. the presence of a single phosphonate group anchored on a hydrophobic tail. We can, thus, presume that the presence of two phosphonate groups in Aln may also prevent the aggregation in Aln-conjugated CNCs sterically.
Due to the higher probability of collisions in their limited volume, the increment of particles` concentration fosters the aggregation also in the case of native CNCs being dispersed in ADMEM (Fig. 3b). Indeed, an eight-time increased concentration (from 0.00125 to 0.01 wt%) of native CNCs causes further aggregation (being partly visible in the presented video frames) and even their precipitation. The precipitated particles are moved further away from the light path, thus reducing their contribution in size values, which explains the >100 nm lower mean/mode of CNC size (from 445 ± 125 to 306 ± 122 nm). This behaviour of native CNCs is important information for further cell testing, since it gives a reliable direction for experiment set-up.
TEM micrographs (Fig. 4) show typical needle-like and well dispersed CNCs, which are thinned and shortened significantly after the oxidation, being in accordinance with the size reduction detected by DLS and NTA. Resent findings suggest that a surface peeling effect due to the periodate oxidation is responsible for the formation of fine CNC crystals with an even more uniform size (Conley et al. 2016). On the other hand, the conjugation of Aln to CNCox induces the conjugate-to-conjugate aggregation (Hermanson 2013) as the results of hydrophobic interactions between the Aln molecule’s alkyl chains from neighbouring CNCs, which may somehow generate even larger clusters during the drying.
Fig. 4
TEM micrographs of 0.001 wt% of RBITC-labelled native (RBITC-CNCnative), oxidised (RBITC-CNCox), Aln and ApA-modified (RBITC-CNCoxAln, RBITC-CNCoxApA) CNCs
×
Evaluation of interactions between differently modified CNCs and human osteoblast cells
A common shortcoming of several techniques dealing with nanoparticles` internalization, such as activated cell sorting/scanning (FACS), Inductively Coupled Plasma (ICP) mass spectrometry and absorption measurements of lysed cells is to distinguish between internalized and adhered nanoparticles (Gottstein et al. 2013). In this study, the RBITC-labelled native and Aln/Apa-modified CNCs are used to track the internalization by non-destructive Confocal Microscopy (CM) imaging, using the fluorescence (red) channel for RBITC-labelled CNCs and transmitted light image channel for osteoblast cells` visualisation. By analysing the orthogonal projections at certain positions within the images, qualitative information was obtained for potential CNC presence within the cell.
The micrographs obtained by CM are presented in Fig. 5. Three orthogonal profiles assigned as D, E and F are presented in each of the channels (A, B, C) as part of one image containing several frames (the z stack). The black arrows on the right side of optical sections indicate cell thickness (trough whole volume) and the presence of a fluorescence signal within this area, indicating a CNC presence within the cell. From the orthogonal profiles of each set of images, presented below, it can be observed that osteoblast cells internalized both the RBITC-CNCoxAln and RBITC CNCoxApA after 1 and 2 h of incubation, respectively. The distribution pattern and localization within the cell in both samples looks rather similar, however, more detailed information needs to be obtained to make the proper conclusions. The RBTIC-CNCoxAln shows a similar zeta-potential (−16 to −20 mV, Fig. 2), but lower average size (368 ± 154 nm) compared to the RBTIC-CNCoxApA (507 ± 260 nm) (Fig. 3), implying the observed differences in kinetic (1 h vs. 2 h) of their uptake by the cells, being thus rather CNC particle-size than chemical-(i.e. conjugated molecule) related.
Fig. 5
Confocal microscopy images of 0.1 wt% of a RBITC-CNCoxAln and b RBITC-CNCoxApA in cell culture medium (ADMEM containing 5 wt% of FBS) after 1 h (z stack thickness = 34.1 µm, 7 optical sections) and 2 h (z stack thickness = 27.4 µm, 14 optical sections) of incubation with human osteoblast cells. A Single frame from transmisson signal from z stack. B Respective frame depicting signal from particles. C Composite signal from transmission signal and signal from particles. D, E, F Orthogonal views of z stacks at line indicated with black (A, C) and white (B)
×
The Aln is a commonly used bisphosphonate drug used in the treatment of osteoporosis due to its specific interaction with osteoclasts` function. As such, the Aln is likely internalized by osteoclasts (Russell et al. 1999) and, hence, interferes with specific biochemical processes, eventually resulting in apoptosis. On the other hand, the limited number of studies demonstrated bisphosphonates` internalization by osteoblasts (Lezcano et al. 2014). One of them demonstrated the internalization of Alexa Fluor–labelled Aln within the osteoblasts’ cytosol (Lezcano et al. 2014), being related to the presence of high affinity binding sites for bisphosphonates (i.e. Olpadronate) in osteoblastic cells (Morelli et al. 2011). The latter gives the clue for possible Aln and ApA action as a driving force for CNCs` attachment to and further delivery through the osteoblasts` membrane. The repulsive ionic forces between the negative cell membrane and the negatively-charged RBITC-CNCoxAln and RBITC-CNCoxAlA are probably overcome by both the dangling Aln/ApA molecules on the surface possessing hydrophobic (aliphatic) chain, as well as the positively charged amino groups in RBITC. The presence of the positively charged RBITC also enables attachment of the sample to the cellular surface, making it more likely that the interaction of Aln and ApA on the surface induces internalization. Regardless of this attachment, the osteoblast morphology did not change during their 2 h exposure to the samples. The aggregation, typical for ApA-modified CNCs in the cell media (ADMEM), can also be observed as merged CNCs within the osteoblasts (Fig. 5).
On the other hand, no internalization was observed for a dispersion of native and RBITC-labelled CNC (profiles D and E) after 2 h (Fig. 6a) and even 24 h (Fig. 6b) of incubation, when a change in cells` morphology (from elongated to rounded) was also observed. The observed RBITC-CNCnative particles within the cell (profile F) are continuing even outside the cell, which is a rather non representative (isolated) example than indication of internalization outset. It can be speculated that lack of internalization in native CNCs as opposed to Aln/ApA-modified CNCs, indicates specific, rather than unspecific internalization, being highly desired in, e.g., targeted drug delivery applications. Besides, the aggregation of RBITC-CNCnative was also identified, especially in images collected after 24 h of incubation.
Fig. 6
Confocal Microscopy images of 0.1 wt% of RBITC-CNCnative after a 2 h* (z stack thickness = 78.3 µm, 4 optical sections) and b 24 h of incubation (z stack thickness = 75 µm, 4 optical sections) in cell culture medium (ADMEM containing 5 wt% of FBS). A Single frame from transmisson signal from z stack. B Respective frame depicting signal from particles. C Composite signal from transmission signal and signal from particles. D, E, F Orthogonal views of z stacks at line indicated with black (A, C) and white (B)
×
The tendency towards aggregation at applied concentrations corresponds well with NTA measurements in ADMEM, since the concentration of CNC incubated with osteoblasts is ≥0.01 wt%, the latter being the aggregation concentration visualised by NanoSight´ camera (as presented in the NTA analysis section). Moreover, the sedimentation of CNC in the cells` surroundings (the non-attached) was evidenced by reduction in fluorescence intensity in the same focal plane even after 1 h of incubation (Supporting Information Fig. S4). This was also compared qualitatively by CTF parameter the value of which was calculated from plot profile data obtained with ImageJ software).
Osteoblasts` viability was determined after 24 h of incubation for each sample using the MTT test assay (Fig. 7). The CNCs were found to promote the cell growth even in its native (non-modified) state. Dilution of the main sample suspension did not affect the viability test significantly, which implies that the materials are safe in all used concentrations, at least for the tested exposure time of 24 h. An important finding in relation to the viability testing is certainly that, although the cell growth is lower for the Aln/Apa-modified CNCs in comparison with the native CNC, all samples outperform all the used control samples. This not only shows that neither the native CNCs, nor their possible degradation products that could appear during the incubation period (either released RBITC, parts of the CNCs etc.), have any toxic effect on the used osteoblasts, which confirms the notation from previous studies (Hosseinidoust et al. 2015) for CNC non cytotoxicity.
Fig. 7
MTT testing results (the absorbance intensity of formazan product measured at 570 nm) of osteoblasts incubated with differently diluted (1:2, 1:4 and 1:8), 0.5 wt% of native (CNCnative), RBITC-labelled (RBITC-CNC), and ApA/Aln-modified (RBITC-CNCoxApa/RBITC-CNCoxAln) CNC dispersions in cell culture medium (ADMEM containing 5 wt% of FBS). ADMEM, PBS and Milly-Q water were used as controls
×
Considering the available related literature, the use of CNCs in different formulations, even as a hybrid material (e.g. in combination with bioglass), have been shown to be suitable as orthopaedic materials. Such CNC-based hybrid materials can accelerate osteoblast attachment, increase their spreading, proliferation, even promote differentiation of mesenchymal cells towards the osteoblast like phenotype (Gorgieva et al. 2017), as well as influence mineralization of the native osteoblast extracellular matrix positively (Chen et al. 2015). Nevertheless, neither of these studies reports or even considers the internalization of materials (or their parts) by osteoblasts. Since in native state (not protected or encapsulated form) the (bis)phosphonates are not retained in the skeleton and are cleared rapidly from the circulation by renal excretion, the proposed modified CNCs with covalently-attached Aln/ApA may prolong the half-life of the drug and, as such, increase its therapeutic potential, as well as ease the patient intake (by reducing the frequency of administration) (Drake et al. 2008).
Conclusion
The complementary analytical tools with different time-scale tracing capacities, as well as resolution and sensitivity profiles (i.e. DLS, NTA and TEM), were used to follow the native and (bis)phoshonate (Ala/Apa)-modified CNCs, as well as their further fluorescent (RBITC)-labelling. The aggregation patterns of labelled native and ApA/Aln-modified CNCs in different media (Milly-Q water, PBS and ADMEM cell medium) were determined, indicating their aggregation phenomena in the ADMEM medium, being less pronounced in the case of Aln-modified CNC, but highly concentration-dependent for native CNCs. The uptake of both (ApA/Aln)-modified CNCs by the human osteoblast cells in the short-term (1/2 h) performed exposure experiments was confirmed by Confocal Microscopy imaging, being the most likely due to the presence of phosphonate groups and hydrophobic alkyl chains of ApA/Aln. In contrast, no internalization of native CNCs was observed even in 24 h of incubation period. The cell viability in all examined dispersions was, however, higher in comparison with the controls, proving the potential of both native and Apa/Aln-modified CNCs for further evaluation towards their usage in biomedicine, above all in theranostic applications, as well as in treatment of various bone cell related diseases, and development of novel intracellular drug delivery systems.
Acknowledgments
This research was supported financially by the Slovenian Ministry of Education, Science, Culture and Sport under the MNT Era-Net programme (Project No. POSSCOG), and QualityNano Project (Grant No. INFRA-2010-262163). We are grateful to dr. Andraž Stožer for the scientific comments and to Rudi Mlakar for technical help in the CM unit at the Institute of Physiology, Faculty of Medicine, Maribor.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
