3.1 Biofunctionalized CNTs on the tissue culture plates
The effects of biofunctionalized CNTs on neuronal culture may be implemented in two ways: (1) add soluble CNTs directly to the neuronal cell culture media, or (2) immobilize functionalized CNTs on the substrate. Direct addition of CNTs to the cell culture media would lead to cellular uptake and possibly induce cytotoxicity. This issue may be avoided by implementing the second strategy, where CNTs are supported on the surface of materials. In this approach, the intracellular uptake of CNTs and possible toxicity may be less of a concern and their utility for implantable devices or cell guiding scaffolds can be envisaged [14, 21, 31].
We used the commercial BD PureCoat™ amine culture wares, which have positively charged, chemically modified, enhanced cell culture surfaces. The surfaces of the tissue culture polystyrene plates (named ‘Culture Plate’) were coated with the functionalized CNTs using a carbodiimide crosslinker chemistry. The present technology is easily applicable regardless of the size and capacity of these commercial containers, from 175 cm2 flasks to 96-well plates.
To obtain functionalized CNTs, carboxylation of CNT was first performed using the acid oxidative method and carbodiimide chemistry. After carboxylation of CNTs, 0.02% (w/v) CNT-COOH are well dispersed in aqueous solution (Fig. S1a). The zeta potential results of pristine CNT and CNT-COOH are − 7.97 ± 0.45 and − 41.15 ± 1.20 mV, respectively. The TEM images of the two types of samples are compared in Fig. S1b in the Supporting Information. The TEM images of pristine CNT and CNT-COOH clearly show loss of the outer tubes (respectively, about 35.1 ± 3.4, and 27.8 ± 5.7 nm) on the surface of CNTs according to the defect by strong acid treatment. Pristine CNTs could not be dispersed in aqueous solution, not even in methanol, showed as no UV-visible absorption peaks (Fig. S1c). A well-dispersed aqueous solution containing 1% (w/v) CNT-COOH displayed two peaks at 196 and 259 nm. The UV-visible spectra indicated that the clear changes of the water solubility and the compositional surface characteristics of the CNT derivatives. The substrate change due to the chemical modification of CNTs was monitored by TGA (Fig. S1d). Pristine CNTs were thermally degraded to yield a 91% mass loss at 650°C, whereas the weight losses of the carboxylated CNTs were 12% at 400°C and 84.5% at 730°C. The latter two mass losses corresponded to the decomposition of outer shells with the carboxylic groups and the degradation of the inside graphitic carbon phrase.
The surfaces of the tissue culture polystyrene plates (named as ‘Culture Plate’) were coated with the carboxylated CNTs using a carbodiimide crosslinker chemistry. To obtain directly and evenly functionalized CNTs on the surface of the culture plates, a series of carboxyl CNT solutions were prepared by diluting a stock solution from 0.005–0.0003125% (w/v). A carbodiimide crosslinking reaction was used to link the carboxyl-CNTs to the surface of the aminated culture plates, as shown in Fig. S2. The cell viability results and SEM images indicated that the plates coated with 0.0025% (w/v) functionalized CNT provided the best cell viability, based on the observation of homogeneous monolayer-coated surfaces (Fig. S2e).
The functionalized CNT-coated surfaces prepared under optimal conditions were modified through PEGylation and drug loading. FE-SEM and AFM images revealed that the surfaces of the cell culture plates were evenly coated with functionalized CNTs, and the CNT diameters increased upon PEGylation, as expected (Fig. 2a and b). The SEM images shown reveal that the individual CNTs were fully covered by PEGylation. These images of the functionalized CNTs revealed that the defined surfaces of the samples were smoother and thicker (32.2 ± 7.1 nm of diameter) than those of the CNT-COOH (27.6 ± 6.8 nm), indicating that the CNTs were wrapped with PEG molecules (Fig. 2c).
Figure 2d shows the contact angles on each surface of the modified culture plates. The values changed from 37.07 ± 4.26° for the unmodified culture plate to 57.66 ± 1.95° and 57.29 ± 3.48° for the carboxylated and PEGylated CNTs, respectively. These contact angles did not exceed 90° and corresponded to surfaces appropriate for cell culture.
One of the most distinctive properties of CNTs is their electrical conductivity. The electrical resistances of the various substrates fabricated were measured over an area of 1.0 x 1.0 cm2. The CP-CNT-COOH prepared using the highest CNT concentration yielded the highest electrical conductivity of 0.94 ± 0.01 µS/cm (2.78 ± 0.35 MΩ/sq), that is a remarkable result from mono-layered CNT coating. The CP-CNT-PEG-NH2 yielded a good electrical conductivity of 0.75 ± 0.88 µS/cm (42.69 ± 14.61 MΩ/sq). The culture plate itself was non-conductive, yielding a value of about 5.69 x 10− 15 S/cm2 (Fig. 2e and f). These results suggested that the culture plates coated with functionalized CNTs provided good hydrophilic and electrically conductive surfaces that might be useful for neural culture. CNTs are used to modify substrate surfaces and enhance their functional capacity for use in implantable devices, cell guiding matrices, or tissue engineering scaffolds.
3.2 Drug delivery system by biofunctionalized CNTs
One of advantages of the PEGylated CNT-coated culture plate substrate developed here is its capacity for use as a smart drug delivery system. A drug, bpV, was embedded in the functionalized CNTs and immobilized onto the prepared culture plates (Fig. 1e). The drug release profiles were quantified using UV-Vis spectrometry, which displayed three distinct typical peaks at 195, 221, and 269 nm, corresponding to the absorbances of bpV. The standard curve according to the concentration of bpV is shown in Fig. 3a and the drug release profile for 30 days is shown in Fig. 3b. Standard curve of bpV(phen) was obtained using serial dilutions from 100 to 0.001 µM, and a linear regression yielded the following expression for the curve: y = 0.0282x–0.0053, with R2 = 0.997. The bpV compound is a highly hydrophilic drug and was released relatively quickly. The sustained release profile unloaded half of the initial loading level into the medium within 11.5 hours, and more than 75% was released during the first day. Most of the drug was released within 3 days. After one week, the release curve was dominated by the long-term release of small amounts of bpV, which was probably entrapped in the PEG molecule brushes.
The surface of the culture plates was analyzed for their drug-loading capacities by quantifying all of the chemicals present based on ATR-FTIR spectroscopy and XPS results. The ATR-FTIR spectra were characterized by a broad peak from 3500 to 3000 cm− 1, which could be assigned to the O-H stretch of the carboxyl groups. The peak below 1500 cm− 1 was associated with stretching vibrations characteristic of the C–C bonds in the nanotubes. The carboxyl groups of CP-CNT-COOH produced a strong peak at 1750 cm− 1 that corresponded to a C-O stretch, indicating successful surface oxidation. The peaks at 1550 and 1200 cm− 1 corresponded to the C-O-C ether stretch and the C-O alcohol stretch in the PEG groups. The peaks at 1100 cm− 1 corresponded to the V-O stretching vibrations of the vanadate groups. The ATR-FTIR spectra revealed differences between the surface of the CNT-coated substrates, based on carboxylation, PEGylation, and drug loading (Fig. 3c). These data indicated the degree of drug loading on/inside the PEG groups of the CNTs. The drug might be entrapped between the brushes of the PEG groups on the nanotubes until release during cell culturing.
Figure 3d shows the XPS spectrum, confirmed the attachment of various CNTs to the culture plates. The C1s, N1s, and O1s peaks in the XPS spectra corresponded to bonding energies within the ranges 280–295, 395–410, and 525–540 eV, respectively (Fig. S3). The C1s spectra obtained from the CNT-modified culture plates presented a larger number of carbon atoms due to an increase in the sp2 carbon atoms upon covalent binding between the CNT molecules and the culture plates. The N1s peak positions of the nitrogen atoms appeared at 400 eV, and the N1s peak of the CNTs decreased due to amide bonding at the primary amine molecules of the BD culture plate surface. The O1s XPS peak (at 532 eV) of the bpV-loaded substrate shifted significantly relative to the position obtained from the unloaded substrate due to the presence of vanadate group oxygen molecules. The deconvoluted high-resolution C1s spectrum indicated that bpV was loaded onto the CP-CNT-PEG-NH2 plate, and the C-O peak appeared at 286 eV. Table 1S presents a summary of the XPS results.
3.3 Cell viability of the biofunctionalized CNT-coated culture plates
The cytotoxicity of cells cultured on the drug-loaded biofunctionalized CNT-coated plates were evaluated over 3 days, and the results were compared with the cell viability results obtained using the PEGylated CNT-coated plates. Culture plates coated with 0.0025% (w/v) CNT did not display any cytotoxicity. The PC-12 cells displayed better cell viability on the CNT-coated substrates than under standard culture conditions. High cell viabilities were observed on the two substrates over a NGF concentration range of 0.1–200 ng/mL, especially at 100 ng/mL NGF (Fig. S4a). A 100 ng/mL NGF concentration was, therefore, used as a baseline in the drug-loaded experiments to exclude cell viability differences between the control and the CNT-coated substrates [22].
The drug concentration-dependent cell viability was measured in the presence of bpV, which was added to the culture plates, as the drug loading concentration was varied from 1 to 200 nM. A bpV concentration of less than 50 nM yielded good cell viability (Fig. S4b). A concentration of 10 nM bpV was selected as the drug concentration that maintained cell proliferation and enhanced neuronal outgrowth. The IC50 of bpV was calculated based on a fit to the cell viability curves obtained from the PEGylated CNT-coated and drug-loaded substrates, and was found to be 8.18 nM. This concentration may be the best concentration for inhibiting PTEN and activating Akt signaling, thereby promoting neuron proliferation and axon regeneration [28, 32–34].
3.4 Real-time monitoring of neuronal outgrowth
The time-dependent extent of neuronal outgrowth under various bpV concentrations was measured in PC-12 cells cultured on 24-well plates (Fig. 4 and Fig. S5). Figure 4a and 4b show that neuronal outgrowth (pink lines) was highly enhanced by the bpV-loaded PEGylated CNTs, compared with data on the culture plate in the presence of NGF. Growth on an untreated control plate was compared with growth on a plate coated with PEGylated CNTs by monitoring every 3 hours over 7 days using the IncuCyte ZOOM, yielding 56 time points.
The drug concentration was varied from 0 to 100 nM, and the total neurite length and the total number of branch points are shown in Fig. 4c and 4d. A merged image of one well in a 24-well plate is shown in Fig. 5A. The cells accumulated in the center of the well, possibly due to the convective activity of the air in the incubator immediately after the culture plates had been transferred from the clean bench to the warm incubator after seeding the cells. Spatial accumulation might have been prevented if the plates had been allowed to sit at room temperature for some period of time until the seeded cells attached. These results indicated that a bpV concentration of 10 nM was optimal for PC-12 cell proliferation and neuronal outgrowth. The PC-12 cells proliferated and extended the neurites on day 5; therefore, cell culture studies over 4 days were used to define the end-point experiments.
3.5 Neuronal outgrowth analysis by immunofluorescence staining
At the 5-day end point, the PC-12 cells were fixed and stained with anti-ßIII-tubulin as an inner cytoskeleton marker and an anti-actin antibody as an outer cytoskeleton marker. Neuronal outgrowth on each substrate was analyzed, and the results were compared (Fig. 5). The percentage of cells that displayed neurites and the average total neurite length per cell were calculated in the cultures grown on the carboxylated or PEGylated CNT-coated plates. These values were found to be 38.7% (a 4.9-fold increase over the results obtained from the control plates) and 164 µm (a 2-fold increase), 69.3% (8.8-fold) and 267 (3.2-fold) µm, respectively. The corresponding values obtained from cultures grown on a glass coverslip were 8.0% and 84 µm. The bpV-loaded PEGylated CNT-coated plate displayed the highest percentage of neurites and the longest neurite length, 81.2% (a 10.3-fold increase over the results obtained from the control) and 385 µm (a 4.6-fold increase). These results provided statistically significant evidence that the PEGylated CNT coating and bpV loading strongly influenced neuronal outgrowth.
3.5 Neuronal outgrowth by morphological analysis
A detailed morphological analysis of PC-12 cells grown on each substrate was obtained by fixing the cells and collecting FE-SEM images under high magnification and the images were then analyzed using ImageJ (Fig. 6). Each neuronal cell extended its neurites along the surface of the glass coverslip or CNT-coated culture plate in the presence of NGF. The total neurite length and the maximum neurite lengths obtained from the cultures grown in the presence or absence of CNTs or drugs differed substantially (Fig. 6k and 6l); however, the percentage of cells that displayed neurites, the number of neurites, and the number of branches found on each substrate were not statistically different (Fig. 6i and 6j).
The CNTs may provide beneficial topographically and electrically conductive environments to the neuronal cells, whereas the glass or normal culture plates did not. High-magnification FE-SEM images revealed that the tips of each neurite assumed different morphologies on each substrate (Fig. 4a-4h). On the glass coverslips, the PC-12 cells extended their neurites across the smooth and slippery surfaces, but these neurites could not stretch very far or remain stable for more than 3 days. The PC-12 cells grown on the glass coverslip began to detach on day 3 (Fig. 4e). By contrast, neuronal cells grown on the CNT-coated surfaces attached tightly for longer periods of time, and the CNTs might provide topographical features and electrical effects. The CNT molecules and the surface functional groups on the CNTs affected the PC-12 cell neurite morphology. Well-attached neurites from the neuronal cells grown on the negatively charged carboxyl CNT substrates included many tiny neurites and branches, unlike the neurites observed on the glass coverslip (Fig. 4f). The neurites that grew on the positively charged PEGylated CNT substrates were longer than those that grew on the glass or carboxyl-CNT surfaces (Fig. 4g). Numerous distinct long slender filopodia were observed to form at the tips of the neurites. Interestingly, the ends of the neurites grown on the bpV-loaded PEGylated CNT substrates formed wide and broad lamellipodia, which are found primarily in very mobile cells and are involved in rapid repair processes (Fig. 4h). Indeed, the maximum neurite length was measured on the bpV-loaded PEGylated CNT substrates. These results suggested that the bpV-loaded PEGylated CNT substrates stimulated neuronal outgrowth in the neuronal cells, and this system may be applicable to enhancing neural regeneration after injury.
3.6 Drug release effects on neuronal outgrowth
The effects of indirect drug release on neurite growth were examined using a new method (Fig. S6). Neuronal progenitor PC-12 cells were seeded onto PTFE inserts with 3 µm pores and were incubated in 24-well tissue culture plates that had been coated with functionalized CNTs over 7 days (Fig. 7). The culture plate as control showed little neurite growth. On the other hand, the culture plate prepared with bpV displayed good levels of total neurites (a 2.6-fold increase over the control). The total neuronal outgrowth measured on the PEGylated CNT substrates or the bpV-loaded PEGylated CNT substrates was 3.3-fold or 18.1-fold higher, respectively, over the neuronal outgrowth value measured on the control plate (Fig. 7b). Neuronal outgrowth in the PC-12 cells was stimulated in response to the PTEN inhibitor, bpV(phen), showing the promise as a neuronal regeneration agent. Neuronal outgrowth was enhanced on the CNT-coated substrates, which provided long-term controlled drug release to the neuronal cells.
3.7 Possible mechanisms of neuronal outgrowth on bpV-loaded CNT substrates
Zhang et al. reported that bpV activated Akt protein kinase and induced high PTEN phosphorylation levels, suggesting that bpV may act as a neuroprotective agent during ischemic brain damage [14, 35]. Christie et al. also reported the inhibition of phosphoinositide 3-kinase (PI3K)/Akt signaling by PTEN in adult peripheral neurons during regeneration. On the other hand, the inhibition of PTEN by bpV at the injury site accelerated axon outgrowth in vivo, which improved nerve repair [36].
Results of immunoblotting analysis, as shown in Figs. 8 and S7, revealed the putative roles of CNTs and bpV in neuronal outgrowth. The phosphorylation levels of PTEN, Akt, extracellular signal regulated kinase (ERK), focal adhesion kinase (FAK), and glycogen synthase kinase 3β (GSK-3β), and the expression levels of mammalian target of rapamycin (mTOR), βIII-tubulin, microtubule-associated protein 2 (MAP-2), growth associated protein 43 (GAP-43), hypoxia-inducible factor 1 α (HIF-1α), were measured and compared with the glyceraldehyde 3-phosphate dehydrogenase (GAPDH) control. Figure 8c illustrates a possible mechanism underlying the promotion of neuronal outgrowth in PC-12 cells on the bpV-loaded biofunctionalized CNT substrates. Several signaling pathways in the PC-12 cells were stimulated by NGF and played important roles in stimulating neuronal growth via the classical Akt pathway and ERK cascade [32, 33, 37, 38].
As expected, bpV itself inhibited PTEN activation and increased Akt signaling for neuronal regeneration. CNTs enhanced the expression of FAK and HIF-1 and stimulated neuronal cells grown on the CNT-functionalized surfaces by facilitating electrical conduction. We previously reported that CNT-coated substrates improved FAK expression of PC-12 cells for neural regeneration [12]. An increase in HIF-1α expression due to stimulation of the CNTs may also regulate neuronal regeneration. Le et al. described the protective role of HIF-1 α in regulating cellular apoptosis after neonatal hypoxia-ischemia brain damage [39].
The bpV-loaded biofunctionalized CNTs were highly effective in reducing the ratio of phosphorylated PTEN to the total PTEN due to inhibition by bpV [40]. The fraction of phosphorylated Akt and ERK present among the respective total proteins was also higher on the bpV-loaded biofunctionalized CNTs relative to the other substrates. The mTOR level was down-regulated by bpV and the bpV-loaded substrates, whereas p-GSK3β was up-regulated. Inactivated phosphorylated GSK3 and mTOR, formed through the PI3K–Akt pathway, coordinated the proliferation and/or differentiation of neural progenitor cells to form axons [36, 41]. GSK3 signaling inhibition was essential for promoting microtubule assembly, organization, and dynamics in axons and dendrites. These functions involved the Tau and MAP proteins [42].
Neuronal outgrowth is regulated by the PI3K–Akt signaling pathway. Treatment with CNTs and bpV(phen) significantly increased the levels of phosphorylated Akt, ERK, and the related neuronal outgrowth proteins, MAP-2 and βIII-tubulin. Taken together, these results revealed that the bpV(phen)-loaded biofunctionalized CNT substrates induced neuronal outgrowth in PC-12 cells to promote neural regeneration.
The utility of the biofunctionalized CNTs was demonstrated for use in drug delivery applications to address the current need for nanobiotechnology-based target-specific controlled drug release systems. The substrates were used to investigate the biofunctional activities of neurons in vitro. The present study demonstrated that a smart biomaterial based on drug-loaded biofunctionalized CNTs enhanced neural adhesion and promoted neurite extension through electrical stimulation, nano-scale topographical environments, and smart drug delivery system.
This study demonstrated the utility of CNTs in promoting neural regeneration via the conductive properties of CNTs and the surface chemical and morphological properties of CNT surface coatings [43, 44]. The conductivity and nanotopographic morphology of the CNT-coated substrates modulated the signaling pathways within and among the neuronal cells. Neural extension and cell proliferation in the PC-12 neuronal cells were significantly enhanced on the drug-loaded biofunctionalized CNTs compared to the un-loaded functionalized CNTs, suggesting synergic effects in the presence of both the drug and the CNTs for neuron cell regulation. Adherent neuron cells may sense the nanotopography and surface functional groups of the CNTs, which can stimulate ECM molecules on the cell surfaces, thereby triggering cell signaling pathways. Neuron cells may be stimulated chemically and electrically by the drug-loaded biofunctionalized CNTs to spur axon or dendrite growth and elongation.
Neuroscientists studying neural tissue regeneration after injury or damage have increasingly examined biomedical applications of CNTs in an attempt to improve neural implant graft coatings and tissue engineering guiding matrices that support neural cells. CNTs provide useful nanocarriers for bioactive molecule delivery. The significant findings reported here suggest that CNTs have clinical utility, and further research is warranted to examine CNT biofunctionalization with biomolecules relevant to the functional recovery of injured neural tissue.