Silk fibroin biomaterial-functionalized carbon nanotubes for high water dispersibility and promising biomedical applications

Materials and preparation of functionalized MWCNTs

Water-soluble SF (powder, 0.1–2 µm diameter, bulk density = 0.2 g·cm⁻³, Mn = 12,000) was prepared in our laboratory (SF obtained by “0.05 M (w/v) Na₂CO₃ aqueous solution at approximately 120℃ under 0.24 MPa for 2 h” was used). ⁴⁰ N-(3-Dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride (EDC, 98%) and N-hydroxysuccinimide (NHS, 98%) were purchased from Aladdin (Shanghai, China). MWCNTs (98%, 10–20 nm outer diameters and 10–20 µm lengths) were purchased from Chengdu Organic Chemistry Limited Company (Sichuan, Chengdu, China). All the reagents were purchased from commercially available sources and used without further purification unless otherwise stated.

Raw-Ms were oxidized with a mixture of H₂SO₄ (98%)/HNO₃ (68%) (3:1, v/v) at 60℃ for 3 h. The carboxylated MWCNTs were filtered, washed with deionized water until pH 7 was reached and dried in a vacuum oven. The as-prepared AO-Ms (100 mg) were suspended in 10 ml of deionized water by sonicating the mixture for 30 min. Subsequently, 4.5 mmol EDC and 4.5 mmol NHS were added to the above suspension and mixed sufficiently. After stirring the mixture for 24 h, SF (1.98 g, to obtain a saturated solution) was added. The reaction was allowed to stir for another 36 h. After centrifugation, sonication and repeated washing, the SF-functionalized MWCNTs (SF-Ms) were obtained. The control sample, which was prepared following the same steps but without adding catalyst (EDC/NHS) as per the reaction was in Figure 1, was called SF/M (mixture of SF and MWCNTs).

Characterization

Functionalization of MWCNTs was confirmed by FTIR spectra recorded on a Nicolet 5700 spectrophotometer (Thermo Fisher Scientific, USA). X-ray photoelectron spectroscopy (XPS; K-Alpha, Thermo Fisher Scientific) was conducted using an Al Ká X-ray source at 250 W. The binding energy scale was calibrated with respect to the C1s peak of hydrocarbon contamination fixed at 285.0 eV. The peak at around 400 eV corresponded to N 1s, which was derived from the C–N bond. The structure was analyzed via transmission electron microscopy (TEM; JEOL JEM-2100, Japan). Thermal properties were determined with TG209F1 (NETZSCH, Germany).

Dispersibility experiment

After mixing 5 mg of MWCNTs in 10 ml of three kinds of solvents (H₂O, methanoic acid and methanol) and sonicating with a KQ-400DE sonicator for 1 h, we took optical micrographs of the samples in these kinds of solvents obtained 6 h after the solution had been sonicated for 1 h. Furthermore, the dispersions of MWCNTs were determined by a UV–vis spectrophotometer (TU-1901 spectrophotometer, Pgeneral, China) and conducted quantitative analysis for the as-prepared samples in the same three kinds of solvents at a concentration of 1 mg/10 ml taken 6 h after the solution had been sonicated for 1 h.

MTT assay

The cytotoxicity of the as-prepared samples was further evaluated via the MTT assay. Murine preosteoblastic MC3T3-E1 cells were grown in a 96-well plate containing α-MEM cell culture media at 37℃ in a 5% CO₂ incubator. After culturing for 1 day, the culture medium was removed and incubated with supernatant (100 µl), followed by ultrasonication for 30 min. After additional incubation for 1 day, 20 µl of MTT (MTT concentration = 5 mg/ml) was added to each well. The cultures were then incubated for 4 h at 37℃ and the supernatants were removed from the wells. About 150 µl of dimethyl sulfoxide was added to each well and 100 µl of mixed liquid from each well was transferred to a 96-well plate. Finally, the plate was read in an automated microplate spectrophotometer with 490 nm as the reference. Every 48 h, the same test was performed as mentioned above until 168 h had passed.

FTIR spectroscopy

Figure 2(a) shows the FTIR spectra for the as-received raw-M, AO-M, SF/M and SF-M samples. The peak at 1631 cm⁻¹ in spectrum B was attributed to the C = O stretching frequency of the carboxylate (COO-) group, which was stronger than the one in spectrum A contributing to the acid oxidation, and there was no other evident difference between spectra B and A. The peak at 1631 cm⁻¹ without any red or blue shift (spectrum C) confirmed that the oxidized MWCNT did not graft with SF. The difference between spectra C and B may be caused by the SF attaching to the MWCNT (around 1534 cm⁻¹ and 1250 cm⁻¹, attributed to amide II and amide III bands of SF, see spectra C). Furthermore, the small vibration at 1700 cm⁻¹ may have arisen from the reacted SF. The FTIR spectrum of the SF-M samples in spectrum D indicated that the C = O stretching frequencies shifted from 1631 cm⁻¹ of COO- to 1636 cm⁻¹ (which was obviously stronger) assigned to the amide I, arising from amide functional group. Moreover, the presence of stronger bands at 1535 cm⁻¹, 1256 cm⁻¹ and 3283 cm⁻¹ corresponding to amide II, amide III and N-H bond stretching, respectively, further confirmed the presence of the amide functional group, namely, the SF functional group. A previous study reported similar results.^9,16,40 OPEN IN VIEWER

Thermogravimetric analysis

Thermogravimetric analysis (TGA) measurements of SF-Ms were conducted for the quantitative evaluation of the amino grafted onto MWCNTs (Figure 2(b)). TGA traces of raw-Ms illustrated minimal weight loss, which was 3% below 700℃. The AO-M sample demonstrated a small amount of weight loss: 8%, which can be attributed to the evaporation of the adsorbed water and the hydroxyl and carboxyl on the oxidized MWCNTs through the oxidation, in order to obtain enough carboxyl groups, so the acid treatment was stronger than usual. ⁴¹ In contrast, the curves of SF/Ms and SF-Ms both displayed noticeable weight loss. The weight loss of SF/Ms may be due to the SF attached on the surface of MWCNTs, whereas the weight loss of the SF-Ms was attributed to the SF grafted and attached on the MWCNT surfaces. Therefore, the weight loss in the SF-M sample was about 16% higher than that of the SF/Ms. Thermal analysis indicated that the surfaces of SF-grafted MWCNTs were coated with a large amount of SF and SF/Ms displayed low coating efficiency.^9,33

XPS analysis

XPS analysis patterns for raw-Ms, AO-Ms, SF/Ms and SF-Ms are shown in Figure 3 to determine the chemical species introduced by the modification. Samples were degassed overnight within the XPS chamber (10⁻³ mbar) prior to analysis. As shown in Figure 3(a), unmodified MWCNTs displayed a single peak at 285 eV because of carbon. The peaks around 284.6 and 532.7 eV (Figure 3(b) to (d)) were attributed to carbon and oxygen, respectively. In addition, the peak at 400.2 eV in Figure 3(d) was due to nitrogen, which agreed with the results of FTIR analysis and confirmed that the surfaces of MWCNTs were grafted with a certain amount of SF. SF/M and SF-M samples were washed several times and most of the SF on the surfaces of the SF/Ms was washed away due to the weak combination. Therefore, there was no N1 peak in Figure 3(c). However, the combination of SF on the surfaces of SF-Ms and SF-Ms was much stronger so that there was an N1 peak in Figure 3(d) despite the washing. The atomic ratio of nitrogen was 6.287%, ⁹ as reported in previously published reports, and the steric effect may influence the grafting efficiency. Moreover, the Mn of the SF used in this study was 12,000, which was much higher than common amines, thereby resulting in a strong steric effect. ¹⁶ OPEN IN VIEWER

TEM analysis

The morphology and nanostructures of raw-Ms, AO-Ms, SF/Ms and SF-Ms were observed by TEM, and the results are shown in Figure 4. The detailed morphology is presented in Figure 4(a) to (d). The unevenness of the SF-M surface differed from the other three surfaces. In EDS analysis, the appearance of a nitrogen peak in Figure 4(c) could be ascribed to the SF attaching on MWCNTs, and the stronger nitrogen peak could be attributed to the modification. To further prove the effect of SF modification, SF/Ms and SF-Ms were compared (Figure 4(e) to (h)). The representative fields were marked as clearly displayed in Figure 4(f). The MWCNT surface was notably coated with SF, which may be attributed to chemical processing. The grafted SF may result in more quantities of coated SF because of the interaction between SF molecules. Meanwhile, SF/Ms showed minimal changes, namely, in the difference in quantity between these two samples, which agreed with the previous images. ³³ The difference between SF/Ms and SF-Ms in Figure 4 reflect the different coating efficiencies of these two kinds of materials, and those between the SF-Ms and SF-Ms in Figure 4 result from different magnifications and different areas. OPEN IN VIEWER

Evaluation of dispersibility

Table 1 and Figure 5(a) show the dispersibility of MWCNTs in three solvents, with mixing of 5 mg of MWCNTs in 10 ml of these solvents and sonicating with a KQ-400DE sonicator for 1 h. Furthermore, the dispersions of MWCNTs were determined by a UV–vis spectrophotometer (TU-1901 spectrophotometer, Pgeneral, China) and the obtained quantitative analysis for the as-prepared samples is illustrated in Figure 5(b). The MWCNT dispersibilities were changed significantly by the modification. SF-Ms displayed high dispersibility, whereas SF/Ms showed similar dispersibility to AO-Ms. However, according to the attaching SF on the MWCNT surfaces, the absorbance of SF/Ms was slightly higher than that of AO-Ms. The dispersion behavior of SF-Ms for methanoic acid was better than in H₂O, which could be attributed to the superior dissolubility of SF in methanoic acid. The remarkably enhanced dispersibility in H₂O and methanoic acid resulted from the grafted SFs on the surface of the MWCNTs. The dispersion activity for methanol was also studied. AO-Ms and SF/Ms were relatively easily dispersed in methanol, thereby suggesting that the attaching SF hardly affected the dispersibility of SF/Ms, which agreed with the results for the other two solvents. In addition, the poor dispersion of SF-Ms in methanol may be ascribed to the poor dissolubility of SF in methanol. The dispersibility for methanol further confirmed the change resulting from the modification. Given the novel SF-M dispersibility and the excellent biocompatibility for SF, studies on reactions with this kind of amino functionalization are of great importance because such functionalized MWCNTs are believed to be highly promising for the preparation of composite materials and biological technologies. ⁴²

OPEN IN VIEWER

Table 1.

Dispersion of multi-walled carbon nanotubes in different solvents after various times (23 ± 2℃)

Solvent	Time	raw-M	AO-M	SF/M	SF-M
H2O	8 h	D	B	A	A
Methanoic acid	8 h	D	B	A	A
Methanol	8 h	C	A	A	C
H2O	20 h	D	C	B	A
Methanoic acid	20 h	D	C	B	A
Methanol	20 h	D	A	A	D
H2O	3 weeks	D	D	D	A
Methanoic acid	3 weeks	D	D	D	A
Methanol	3 weeks	D	A	C	D

OPEN IN VIEWER

Figure 5.

(a) Optical micrographs of the samples in (A1) H₂O, (A2) methanoic acid and (A3) methanol obtained 6 h after the solution had been sonicated for 1 h. The concentration was 5 mg/10 ml; (b) UV–vis spectra of the samples in (B1) H₂O, (B2) methanoic acid and (B3) methanol at a concentration of 1 mg/10 ml taken 6 h after the solution had been sonicated for 1 h. (c) MTT results of raw-M, AO-M, SF/M, SF-M and control samples.

Cytotoxicity analysis

The cytotoxicities of the raw-M, AO-M, SF/M, and SF-M samples were evaluated by MTT assay. The control group, in which cells underwent the same process but without the addition of any MWCNTs, provided a comparison. As shown in Figure 5(c), all samples displayed a certain amount of cytotoxicity and restricted cell viability on the first day of incubation. Further comparison revealed that SF-Ms demonstrated high cell viability, suggesting enhanced biocompatibility. The following results also showed a similar trend, which may be due to two aspects. On the one hand, the SF-M sample possessing better dispersion (as shown in Figure 5(a) and (b)) may prevent the occurrence of sectional toxicity. On the other hand, the favorable biocompatibility of SF could play a critical role as a positive factor. ⁴³