Carbon nanotubes (CNTs) have been widely utilized as a novel drug carrier with promising future applications in biomedical therapies due to their distinct characteristics. In the present work, carboxylic acid-functionalized single-walled carbon nanotubes (f-SWCNTs) were used as the starting material to react with anticancer drug, BA to produce f-SWCNTs-BA conjugate via stacking interaction. The conjugate was extensively characterized for drug loading capacity, physicochemical properties, surface morphology, drug releasing characteristics, and cytotoxicity evaluation. The results indicated that the drug loading capacity was determined to be around 20 wt% and this value has been verified by thermogravimetric analysis. The binding of BA onto the surface of f-SWCNTs was confirmed by FTIR and Raman spectroscopies. Powder XRD analysis showed that the structure of the conjugate was unaffected by the loading of BA. The developed conjugate was found to release the drug in a controlled manner with a prolonged release property. According to the preliminary in vitro cytotoxicity studies, the conjugate was not toxic in a standard fibroblast cell line, and anticancer activity was significantly higher in A549 than HepG2 cell line. This study suggests that f-SWCNTs could be developed as an efficient drug carrier to conjugate drugs for pharmaceutical applications in cancer chemotherapies.

1. Introduction

The revolutionary development of drug delivery technology has become one of the most prominent areas in biomedical science, contributing to a profound beneficial impact on human healthcare. This interdisciplinary technology can be defined as a method or process of delivering drugs and biomolecules to the targeted site of the cell for a specific period of time (sustained release function) with reduced side effect. The current methods of conventional drugs administered via liquids or tablets are generally less efficient and suffered from poor biodistribution, low solubility, long-term toxicity, and limited drug efficacy due to partial biodegradation, swelling, and erosion [1]. This has caused the pharmaceutical industry to develop novel drug delivery systems using a wide range of biocompatible drug carriers with the aim to improve therapeutic efficacy and reduced toxicity. In meeting this demand, various forms of efficient and biocompatible drug delivery systems have been developed extensively and can be generally classified into four major categories: nanomaterials [24], viral carriers [5, 6], organic cationic compounds [7, 8], and recombinant proteins [9, 10].

In the recent years, nanomaterials such as carbon nanotubes (CNTs) have been receiving considerable amount of attention as a new nonviral carrier alternative [1113] compared to viral and cationic carrier. These allotropes of carbons are extensively studied and investigated as novel drug delivery vehicles due to their good biocompatibility, ultrahigh surface area, good mechanical strength yet ultralight weight, low cytotoxicity, and excellent chemical and thermal stability. Furthermore, their outer surface can be chemically functionalized with biocompatible materials [14] whilst inner volume allows the loading of small biomolecules such as proteins and genes for effective drug delivery [15]. Recently, it was reported that chemically functionalized CNTs can be utilized as a novel form of drug carrier by attaching different range of functional groups to their sidewalls [16]. The research team discovered that the functionalized CNTs were able to cross cell barriers in mice through the enhanced permeability and retention effect without causing any harm to the cell under conditions which would normally prevent or delay this biological process. Equipped with all these unique advantages of CNTs and their tremendous breakthroughs in biomedical nanotechnology [1719], it is no doubt that CNTs can be used as a promising novel drug targeted delivery system for advance therapeutic treatment.

Betulinic acid (3β-hydroxy-lup-20(29)-en-28-oic acid, Figure 1) is a naturally occurring pentacyclic lupane-type triterpene extracted from numerous botanical sources found widely distributed in the plant kingdom. Among them, the birch tree, Betula alba, can be considered as one of the highly reported sources of betulinic acid (BA) [20]. It has been used traditionally as a folk remedy by the Native Americans to treat intestinal problems like dysentery and diarrhoea. Recent studies reported that BA possesses many favourable therapeutic activities such as anticancer, hepatoprotective potential, antimalarial, anti-inflammatory, antihuman immunodeficiency virus (HIV), anthelmintic, and antioxidant effects [2125]. BA was also well known for its high selective cytotoxic activity against human melanoma-derived cell lines [26] as well as other types of cancerous tumour [2729]. Healthy normal cells like peripheral blood lymphoblasts, melanocytes, normal human fibroblasts, and astrocytes were also reported to be resistant against BA treatment in vitro [30]. However, its optimum potential is greatly limited by poor solubility in aqueous solvents. As such, an ideal delivery system can further enhance the bioavailability of BA as a potent anticancer agent.

Pristine CNTs (without functionalization) are generally water insoluble and easily form bundles through strong van der Waals forces and, therefore, they cannot be dispersed effectively in most aqueous solutions. They pose a great barrier for drug delivery applications because the insoluble CNTs may cause aggregation inducing characteristic cell changes and apoptosis when administered into human body. This is further supported by studies showing that raw CNTs were observed to accumulate mainly in the liver, lungs, and spleen of the animals [31] while the chemically functionalized ones were found to be excreted in urine and faeces of the animals [32]. Thus, researchers strongly believed that side wall functionalization using carboxylic acid is one of the most powerful approaches to improve the solubility of CNTs [33, 34].

To the best of our knowledge, this is the first time we report on the conjugation of anticancer drug, BA, by using a more suitable delivery system, namely, carboxylic acid-functionalized single-walled carbon nanotubes (f-SWCNT), to enhance the bioavailability of BA. The aim of this work was to evaluate the drug loading and release properties as well as the release kinetics of the synthesized conjugate. In addition, we also studied the chemical structure and surface morphology of the resulting conjugate. Finally, in order to evaluate the cytotoxicity effects of the conjugate, two human cancer cell lines (HepG2 and A549) were used in comparison with normal fibroblast cell lines (3T3).

2. Materials and Methods

2.1. Materials and Characterizations

BA (≥98% as determined by HPLC, C30H48O3, 456.70 g/mol) was purchased from Sigma-Aldrich (Saint Louis, USA). The short f-SWCNTs with a diameter of 1-2 nm and a purity of >90% used in the current work are commercially available from Chengdu Organic Chemicals Co. Ltd. (Chengdu, China). Methanol of 99.8% purity was supplied by HmbG Chemicals (Germany) and used as solvent for BA. All chemicals and solvents were of analytical grade and used as received.

The samples (BA, f-SWCNTs, and f-SWCNTs-BA) were characterized by ultraviolet-visible spectrophotometer (UV-vis, Lambda 35, Perkin Elmer, Boston, USA), Fourier transform infrared spectroscopy (FT-IR, model Smart Orbit, Thermo Nicolet Nexus FTIR, Vernon Hills, USA), thermogravimetric analysis (TGA, Q500, TA Instruments, New Castle, DE), Powder X-ray diffraction (PXRD-6000, Shimadzu, Kyoto, Japan), Raman (RamanMicro 200, Perkin Elmer, Waltham, USA), and field emission scanning electron microscope (FESEM, JSM-7600F, JEOL, Tokyo, Japan).

2.2. Calibration Curve

The standard solutions of BA were prepared as follows. A stock solution (16 mL) was prepared by dissolving 8 mg of pure BA in methanol to obtain the drug concentration of 0.5 mg/mL. Aliquots of standard solutions of BA were transferred into a 20 mL volumetric flask and the mixture was allowed to stand for 7 min at room temperature before recording the absorbance. The absorbance was spectrophotometrically analyzed at 210 nm against reagent blank using UV-vis spectrophotometer where the drug presents a maximum absorption. A calibration curve was plotted to get the linearity in the concentration range of 0.0–0.5 mg/mL and regression value. The curve is described by the regression equation (mg/mL) + 0.0253, where is the absorbance and is the concentration in mg/mL, with a correlation coefficient () of 0.9937.

2.3. Preparation of f-SWCNTs-BA Conjugate

The purchased f-SWCNT (20 mg) was dispersed in BA solution (5 mg in 99.8% methanol, 40 mL) and sonicated for 30 minutes. Subsequently, the dispersion was magnetically stirred at room temperature for 22 hours using a magnetic bar to facilitate loading of BA. Thereafter, the mixture was subjected to centrifugation at 4000 rpm for 15 minutes and then washed with methanol and followed by deionized water three times and centrifuged to remove free/unbound BA. The supernatant was collected to determine the drug loading capacity, whereas the solid sample was dried at 60°C in an oven for 24 hours to obtain f-SWCNTs-BA conjugate. The resulting f-SWCNTs-BA conjugate was stored at room temperature in a vacuum desiccator for further use of studies.

2.4. Assessment of f-SWCNTs Loading with BA

The amount of free/unbound BA weight () in the solution was quantified by measuring the absorbance at 210 nm relative to a calibration curve prepared under the same condition [35, 36] using a UV-vis spectrophotometer (Figure 2). The drug loading capacity was calculated as follows: The drug loading capacity of BA was estimated to be ~20%. A scheme of the drug conjugation is presented in Scheme 1.

2.5. Measurement of BA Controlled Release In Vitro

The release profiles of BA from the f-SWCNTs were performed in human body-simulated phosphate buffered saline (PBS) solutions at pH 7.4 and 4.8 [37, 38]. PBS is a water-based salt solution containing different kinds of anions like , mono phosphate , and dibasic phosphate . Briefly, about 1 mg of f-SWCNTs-BA conjugate was dispersed into 3.5 mL of release media with specific pH value. Maximum absorbance of 210 nm was selected and the accumulated release amount of BA into the solution was determined at preset time intervals using a UV-vis spectrophotometer.

2.6. Cell Lines and Cell Viability Tests

3T3 (normal standard fibroblast cell line), HepG2 (human liver hepatocellular carcinoma cell line), and A549 (human lung adenocarcinoma epithelial cell line) were purchased from American Type Culture Collection (Manassas, VA, USA). The cells were cultured under standard cell culture conditions (37°C in a humidified atmosphere of 95% room air/5% CO2) in RPMI 1640 medium supplemented with 10% heat inactivated fetal bovine serum (FBS) and 1% penicillin (100 units/mL)/streptomycin (100 μg/mL). Cells were subcultured in 75 cm2 culture flasks or in appropriate plates and used for seeding and treatment after reaching approximately 80% confluence.

The MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] assay, as previously described in [39], was used to measure cell viability. In brief, all the normal and cancer cells were seeded at a density of 1 × 104 cells/well in 96-well culture plates in 200 μL of culture medium and incubated at 37°C to assure cell attachment. After 24 h, the cells were treated with BA, f-SWCNTs, and f-SWCNTs-BA conjugate prepared at various concentrations and the treated cells were incubated for 72 h. Subsequently, 20 μL of MTT solution (2.5 mg/mL) in PBS solution was added to each well, and then the plates were further incubated at 37°C for 3 h. The solution in each well was discarded by suction and replaced by 100 μL of dimethyl sulfoxide and vigorously mixed to dissolve the reacted dye. After 30 min of incubation, the absorbance was measured at a wavelength of 575 nm using a microplate reader (EL 800X, Bio-Tel Instruments Inc., Winooski, VT, USA). All assays were done in triplicate and the cytotoxicity results were expressed as the percentage of cell viability with respect to control cells.

3. Results and Discussions

3.1. Characterization of f-SWCNTs-BA Conjugate

The conjugate was extensively characterized by FTIR, TGA, XRD, Raman, and FESEM. FTIR is a powerful and sensitive infrared spectroscopy which is used to verify the conjugation of BA on f-SWCNTs. In this study, FTIR was performed using compressed KBr pellet method and the samples were recorded in the 4000 to 500 cm−1 range. As presented in the spectra of f-SWCNTs-BA (Figure 3), the absorption peaks observed at 3434 and 1627 cm−1 correspond to the absorption peaks of BA at 3449 and 1641 cm−1. These peaks are attributed to the typical stretching vibrations of C=O and O–H, suggesting the presence of –COOH functional group either from BA molecule or from f-SWCNTs. The peaks at 2916 and 1387 cm−1 imply both the asymmetric and symmetric C–H stretching vibrations which arise from the methyl and methylene group of BA [40], respectively, while the absorption peak at 545 cm−1 could be due to the characteristic absorption of aromatic compound from the benzene rings of BA.

To determine the actual drug loading, TGA was used and the samples were heated from room temperature up to 1000°C with a heating rate of 10°C/min under a nitrogen purge of 40 mL/min. The TGA curves of f-SWCNTs and f-SWCNTs-BA are presented in Figure 4. In the total heating process, the f-SWCNTs exhibited a weight loss of about 22.5 wt% (Figure 4(a)), while the f-SWCNTs-BA lost approximately 44.4 wt% of weight (Figure 4(b)). The first weight loss interval below 200°C of both samples was mainly due to the evaporation of adsorbed water [41]. At higher temperatures the weight loss was attributed to the thermal decomposition of the loaded drug, BA. The amount of drug loaded onto the f-SWCNTs is about 21.9 wt% (Figure 4(b)) and this value is nicely agreed with the UV-vis analysis, which was calculated to be about 20 wt%.

The PXRD analysis of the samples was carried out by adjusting the radiation CuKα at 30 kV and 30 mA for recording the PXRD diffractograms between 2 and 90°. This nondestructive analytical method is a valuable tool for characterizing the CNTs and surface functionalization. Figure 5 shows the PXRD patterns for f-SWCNTs, f-SWCNTs-BA, and free BA, respectively. The PXRD analysis of f-SWCNTs and f-SWCNTs-BA shows that there was no change in the tubular structure of the nanotubes. This indicates that the drug loading mechanism mainly occurred as noncovalent functionalization at the surface or sidewall of the nanotubes structure and not by intercalation between nanotubes. Similar XRD profiles of the surface functionalized CNTs with biomolecules were also reported by Mehra and Jain [35] and Jain et al. [42].

Raman spectroscopy was used to investigate the structure of SWCNTs before (f-SWCNTs) and after (f-SWCNTs-BA) functionalization with BA. Figure 6 shows the Raman spectra of f-SWCNTs and f-SWCNTs-BA conjugate. The Raman analysis reveals the presence of the radial breathing mode (RBM) and the two characteristic bands of SWCNTs: the so-called G band (tangential mode) and the D band (disorder mode). The RBM of the SWCNTs is generally caused by uniaxial vibrations of the nanotubes and can be observed at 150 and 250 cm−1 (Figure 6(a)). The G band corresponding to the graphitic lattice vibration mode due to stretching of the C–C bonds in graphitic plane was displayed at 1571 cm−1 for f-SWCNTs and 1572 cm−1 for f-SWCNTs-BA, respectively. However, the D band was not so obvious in both samples, especially for f-SWCNTs-BA in which the D band could not be detected. This could be due to the noncovalent binding between carbon nanotubes and BA that prevents the formation of defects on the surface of the SWCNTs. On the contrary, covalent functionalization generally will produce significantly larger structural defects resulting in increased intensity of the D band [34]. Overall, the Raman spectra of f-SWCNTs-BA exhibited enhanced intensities of the peaks in comparison to that of f-SWCNTs.

The surface morphology of free BA, f-SWCNTs, and f-SWCNTs-BA was observed by FESEM and the images are illustrated in Figure 7. As shown in Figure 7(a), the microstructure of free BA was represented by needle-shaped crystalline particle with a smooth surface. The f-SWCNTs (Figure 7(b)) show tubular-shaped nanoparticles with a smooth surface in a highly aggregated state, whereas the surface morphology of f-SWCNTs-BA (Figure 7(c)) appeared to be rougher and uneven due to the deposition of BA molecules. The morphology of the f-SWCNTs-BA is fundamentally different from f-SWCNTs, therefore leading to the conclusion that conjugation phenomenon had taken place.

3.2. Loading, Release, and Kinetic Behaviour of Betulinic Acid

The release properties of BA from the f-SWCNTs nanocarrier were evaluated in human body-simulated environment at pH 7.4 and 4.8 (Figure 8). It was observed that the release rate of BA from f-SWCNTs at pH 7.4 is substantially higher than that at pH 4.8. This indicates that the BA release profile from SWCNTs is pH-dependent. The maximum percentage release of BA from f-SWCNTs-BA conjugate reached 89.2% (within 1400 minutes) and 78.7% (within 600 minutes) when exposed to pH 7.4 and 4.8, respectively. BA is generally known as a weakly acidic compound with p value of 5.5 [43]. When the pH of the release medium exceeds the p value of BA, both the –OH group on BA and the –COOH group on f-SWCNTs would be ionized and, thus, resulting in a higher amount of BA being released into pH 7.4 compared to 4.8. On the other hand, the slow and sustained release properties of the conjugate might be attributed to the strong repulsive forces that occurred between the carboxylate anions of the nanotubes and BA molecules as well as the anions in the buffer solutions. Hence, these differences in the release rate of the active drug at pH 7.4 and 4.8 are probably due to the different release mechanism of BA from the conjugate as discussed above.

The release in both pH solutions was very rapid for the first 60 minutes (Figure 8, inset picture) and, thereafter, a more sustained and slower release was observed throughout the experiment. The initial stage of the burst release that occurred in the first 15–20 minutes may be possibly due to the free unloaded drug coated on the surface of the nanotubes. This is generally known as the main cause attributed to surface adhesion and desorption for the burst release of most surface-loaded drugs [44]. Burst release is useful for some medical applications like wound treatment because a fast release in the initial stage would provide immediate pain relief to the patient followed by sustained release to encourage wound healing [45]. However, this phenomenon is also likely to cause unfavourable side effects in drug delivery such as short biological half-life in vivo, frequent dosing due to shortened release profile, and potential systemic/local acute toxicity in human body [46]. In order to circumvent the burst effect of the system, biocompatible polymers such as chitosan could be utilized to form an outer layer through additional coating steps for prolonging the release in the circulation half-life [47]. As such, we are currently conducting further investigation using several biodegradable and biocompatible polymers as well as surfactants for the coating of the conjugate and then study their in vitro drug release profiles.

In order to achieve and maintain the effective therapeutic concentration of the drug in the target tissues for as long as possible, a well-defined behaviour of drug release kinetics is required. For this purpose, mathematical modelling is a very useful method to predict the release kinetics before the drug delivery systems are being realized and, at the same time, it provides important information into the structure-function relationship of material systems. Generally, the release behaviour can be determined using three different kinetic models, that is, first order (2) [3], pseudo-second order (3) [48], and parabolic diffusion (4) [49]: where and are the equilibrium release rate and the release rate at any time , respectively, is a constant corresponding to release amount, and represent the drug content remained in the f-SWCNTs-BA at release time 0 and , respectively, and is a constant.

By fitting the release data of the active BA to these kinetic models, it was observed that the release kinetic processes of BA from f-SWCNTs-BA conjugate conformed well to pseudo-second order kinetic model for both pH solutions (Figure 9). At pH 7.4, the release of BA from nanocarrier followed the pseudo-second order equation, with a correlation coefficient of 0.9992 and value of 3.46 × 10−4 mg per minute. For PBS solution at pH 4.8, the release kinetics of BA also fits better to the pseudo-second order model (correlation coefficient of 0.9996 and value of 8.15 × 10−4 mg per minute). Table 1 demonstrates the time release profile for BA in PBS solutions at pH 7.4 and 4.8, respectively. The second order release kinetics indicates that the release of BA is concentration-dependent and the half-life increases as the initial concentration of BA is increased. The reaction rate observed in this study can be considered as a moderate type [50] with time span over 1000 min for a complete drug release.

3.3. Cytotoxicity Evaluation

Figure 10 shows the effect of free BA, f-SWCNTs nanocarrier, and f-SWCNTs-BA conjugate on cytotoxicity of 3T3 cells (Figure 10(a)), HepG2 cells (Figure 10(b)), and A549 cells (Figure 10(c)) at various concentrations, ranging from 0.78 to 50 μg/mL at 72 hours’ time point using MTT assay. In vitro cytotoxicity analysis is a commonly used approach to evaluate the biocompatibility of the synthesized nanomaterial for a biomedical application. Therefore, we investigated the cytotoxic effect of the free BA, the nanocarrier, and the f-SWCNTs-BA conjugate concentrations in a healthy fibroblast cell line (3T3) and two human cancer cell lines (HepG2 and A549), respectively.

As shown in Figure 10(a), the 3T3 cell viability of free BA was reduced to approximately 25% when the concentration exceeded 50 μg/mL. Although BA is a very promising therapeutic compound as discussed before, it has an implicit low toxicity when compared to other chemotherapy drugs like camptothecin, taxol, vincristine, and vinblastine [51]. Furthermore, 3T3 cell line is a very sensitive in vitro model commonly used for the detection of carcinogenic potential of chemicals and, hence, low cytotoxic effect of BA may cause the significant reduction in 3T3 viable cells when applied at high concentration (50 μg/mL). For the nanocarrier, the cell viability was dramatically reduced to about 40% when the concentration was more than 50 μg/mL. This is mainly attributed to the physicochemical property of CNTs in which when further diluted into the cell culture media they tend to rapidly reaggregate into bundles due to strong van der Waals interaction. On the other hand, the resulting conjugate at concentrations from 0.78 to 25 μg/mL did not express any cytotoxicity after 72 hours of treatment, but the cells exposed to 50 μg/mL showed more than 50% reduction in cell viability. This could be due to the effect of the BA in which the BA compound showed a significant reduction of 80% in cell viability when the concentration exceeded 50 μg/mL. In general, we observed higher cell viability for f-SWCNTs-BA conjugate in comparison with the nanocarrier or with the free BA after 72 hours of treatment.

Figures 10(b) and 10(c) show that exposure to free BA, f-SWCNTs, and f-SWCNTs-BA at doses ranging from 0.78 to 50 μg/mL caused a dose-dependent cytotoxicity manner in both liver and lung cancer cells. The conjugate exhibited greater anticancer activity at lower dosages in A549 cells compared to HepG2 cells. It is also important to be noted that, even though the same concentrations of BA and f-SWCNTs-BA were used in this study, only about 20–22% of the drug was actually loaded in the conjugate. This indicates that the conjugate can achieve better therapeutic effect of killing the cancer cells with only ~20–22% of BA in comparison with pure BA (100%). With the reduction of the drug dosage, many unwanted side effects of the drug can be further minimized, especially in cancer-related treatments.

4. Conclusions

In summary, we have successfully synthesized a controlled and sustained release formulation of BA noncovalently loaded to f-SWCNTs as drug delivery system conjugate. FTIR analysis of the conjugate demonstrated the presence of functional groups of both BA and the nanocarrier, indicating that the drug was loaded onto the f-SWCNTs. This result is further confirmed by Raman spectroscopy studies. The loading of BA onto the nanocarrier was estimated to be about 20% (w/w) and this value is in agreement with the TGA study. Powder XRD analysis showed that the structure of the functionalized nanotubes was unaffected by the loading of drug BA. The release of BA from nanocarrier was observed to occur in a controlled manner governed by the different levels of pH in the release medium and, therefore, this drug carrier could be further developed as a controlled release formulation with a prolonged release property. Pseudo-second order model was found to be the best to describe the release kinetic process of BA from the nanocarrier at both pH levels 7.4 and 4.8 in comparison to the other models. Although the burst effect of BA would most likely be the limiting factor for therapeutic use in controlled delivery applications, additional coating steps for prolonging the release in the circulation half-life using biocompatible polymers are currently underway. The cytotoxicity study demonstrated a decrease in the toxicity potential of the conjugate in a normal cell line but enhanced anticancer activity in both human cancer cell lines. Further investigations in terms of specific cellular mechanism of the resulting conjugate are required to elucidate the conjugate interaction with cells prior to in vivo biological studies involving animal models. The results from this preliminary finding indicate that the synthesized conjugate (~20–22% of loaded BA) has the potential to be developed as the drug delivery system for BA since it demonstrated less cytotoxicity but enhanced anticancer activity than free BA (100% of BA). By reducing the drug dosage, this means causing less harm to the normal cells and, thus, provides a better healthcare and well-being in patient.

Conflict of Interests

The authors confirm that they have no financial and personal interests in the work described.


This research was substantially supported by a grant from the Ministry of Education of Malaysia (MOE) under Grant no. GP-IPB/2013/9425800 and MyPhD scholarship under the MyBrain15 program for Julia M. Tan.