Molecular dynamics simulation study of Glycine tip-functionalisation of single-walled carbon nanotubes as emerging nanovectors for the delivery of anticancer drugs

ABSTRACT In this work, molecular dynamics (MD) simulations have been performed to explore dynamic properties of three anticancer drugs including Exemestane (EXE), Letrozole (LTZ) and Fulvestrant (FLV) interacting with single-walled carbon nanotubes (SWCNTs) as drug delivery systems in a biological environment. Furthermore, the effect of functionalisation of SWCNTs with Glycine (Gly) group on the drug adsorption process is investigated. The MD simulation results show that among three investigated drugs, FLV with high hydrophobic characteristic exhibits the strongest afﬁnity for hydrophobic SWCNT (16, 8) in terms of the binding free (ΔGbin) amount energy. Moreover, strong binding of FLV drug molecules on the functionalised single-walled carbon nanotube (f-SWCNT) with (16, 0) chirality is facilitated by more active sites available for hydrogen bond (HB) formation between drug molecules and the functional groups of SWCNT. Because of more number of HBs in the simulation system, there are more numbers of hydrophilic interactions between the adsorbed drug molecules and the functional groups of the nanotube.


Introduction
Nowadays, more effective cancer treatment is of increasing importance and demand due to the continuously growing number of cancer cases and deaths in the global human population. In cancer therapy, a major challenge is to deliver anticancer drug molecules precisely to tumour sites for maximum treatment efficacy while minimising side effects to normal organs [1,2]. In recent years, advanced drug delivery systems such as nanoparticles provide a platform to modify the basic properties of drug molecules viz. solubility, half-life, biocompatibility, and their release characteristics [3].
Considerable efforts following the pioneering discovery of carbon nanotubes (CNTs) have yielded that the biological applications could be expected for these novel materials [4,5]. Single-walled carbon nanotubes (SWCNTs) are one of the most appealing and fascinating drug delivery vehicles because of their unique size, shape and ability to functionalise with other molecules [6][7][8]. Drug delivery system based on SWCNTs can pass through cell walls, transfer, and liberate drugs in particular tissues. Although, a series of issues related to the low dispersion of pristine CNTs in aqueous solution is a typical drawback of their applications. To overcome this problem, surface modification or functionalisation plays a crucial role in improving the solubility and the other physicochemical properties of CNTs. It has been reported that after modifications, functionalised carbon nanotubes' (f-CNTs) solubility increases significantly [9][10][11] and alters their cellular interaction pathways, resulting in much-reduced cytotoxic effects.
Several theoretical researches have demonstrated the potential application of SWCNT as an ideal drug delivery system for various therapeutic drugs [12][13][14]. Khatti and Hashemianzadeh [15] used molecular dynamics (MD) simulation to explore carboplatin filling inside pristine SWCNT and wrapping outside the pristine and functionalised SWCNT. In another investigation, the adsorption mechanism of the anticancer drug paclitaxel (PTX) on pristine and functionalised SWCNTs in aqueous solutions is studied using the molecular dynamics approach [16]. Hashemzadeh et al. showed that the solubility of SWCNTs in aqueous solution is increased by functionalisation. MD simulations of vinblastine (VLB)-CNT non-covalent binding at 277 and 300 K were examined to determine the effect of a reduced temperature on the CNT-based delivery of VLB [17]. Additionally, this work provides an in-depth knowledge on drug-carrier interactions through the discussion of loading position, interaction energy, loading capacities, SWCNT functionalisation, and chirality effects. Another group led by Kamel [18] demonstrated the effect of ethanol as a co-solvent on the inclusion complex formation between Flutamide (FLU) drug molecule and the carbon nanotube. The results revealed that the presence of ethanol at the concentration of 0.5 M enhances the stability of the simulation system. The same research group [19] investigated the effect of the functionalised carbon nanotube chirality on the dynamic process of FLU drug molecule adsorption on the nanotube surface.
In the current study, the delivery performance of SWCNTs and functionalised single-walled carbon nanotubes (f-SWCNTs) with Glycine (Gly) groups for Exemestane (EXE), Letrozole (LTZ) and Fulvestrant (FLV) anticancer drugs are explored by MD simulation. Functionalised SWCNTs can be achieved by using Glycine groups; because Glycine has various active sites that may give some insight to further study of interaction between complicated biomolecules and SWCNTs.
As before mentioned, the selected anticancer drugs (EXE, LTZ and FLV) are used in the most prevalent cancer in women, i.e. Breast cancer. Many breast cancers have estrogen receptors and the growth of these types of tumours can be stimulated by estrogen. Fulvestrant as a new estrogen receptor antagonist has a specific mode of action since it blocks and accelerates degradation of estrogen and progesterone receptor protein, leading to an inhibition of estrogen signaling [20].
Exemestane is an orally active irreversible steroidal aromatase inhibitor. It acts as a false substrate for the aromatase enzyme, and is processed to an intermediate that binds irreversibly to the active site of the enzyme causing its inactivation [21]. Also, Letrozole as a superior therapy for treating breast cancer prevents the aromatase from producing estrogens by competitive, reversible binding to the haeme of its cytochrome P450 unit [22].

Molecular dynamics simulation details
Molecular dynamics simulations are performed to provide molecular insight into the adsorption mechanism of anticancer drugs on the surface carbon nanotubes with chirality parameters of (16,16), (16,0) and (16,8) SWCNTs, i.e. armchair, zigzag and chiral structures with 2.17, 1.25 and 1.66 nm in diameter with same lengths (4 nm), respectively. The defined smaller diameter is suitable for CNT's quality control. It is known that mechanisms such as smaller diameter CNT have been proposed in drug releasing [23]. The nanotubes with small diameters can easily transfuse within the body occupying a low surface volume and for intracellular transport narrow diameter nanotubes are more appropriate. Many theoretical investigations show that curvature of nanotube sidewall plays an important role in determining the reactivity and nanotubes with narrow diameter are thermodynamically more favorable during the course of drug delivery [15,16,[24][25][26][27]. Also, the effect of the functional groups on the interaction of drug molecules with the carbon nanotubes is evaluated. Functionalisation of SWCNT in each simulation box is done using four Glycine functional groups which are grafted onto the tips of SWCNTs through the hydrogen bonding ability and chemical reactivity of terminal carboxyl groups of the oxidised nanotubes, as depicted in Figure SF1, supplementary material. Moreover, close inspection graphical representation of that linkage is provided in Figure SF2, supplementary material. It is known that oxidation of CNTs is commonly carried out by refluxing raw CNTs in strong acidic solution, e.g. HNO 3 / H 2 SO 4 . Under this condition, the end caps of CNTs are opened, and carboxylic groups are formed at these ends and some defect sites on nanotube sidewalls. Also, the detailed descriptions of all force field parameters for the new bond types, new angles and dihedrals of the glycine linkage are provided in Table ST1, supplementary material. Also, the structure of FLV, LTZ, EXE as well as the Glycine functional group are shown in Figure 1. For study the drug-SWCNT interactions and effect of Gly functionalisation on these interactions, different simulation systems are built; each consisting of three molecules of each studied drug and SWCNT. To set up the initial configurations, SWCNTs have been located in the centre of boxes; then, three drug molecules are placed out of SWCNTs and separated sufficiently far away (>3 nm) to minimise the effect of starting orientations. In the next step, all simulation systems are fully hydrated by water molecules using the simple point-charge TIP3P model [28]. The average value of the H 2 O density employed in the simulations over the course of 14,000 ps is 1044.2 kg m −3 , close to the experimental value of 997.1 kg m −3 and the expected density of the TIP3P model of 1002 kg m −3 . All simulations are carried out in a periodic box with the dimensions of 6 × 6 × 6 nm 3 . In order to relax solvent molecules, all systems are equilibrated by 200 ps which the internal structure of the nanotube is assumed to be rigid with exerting position restraints. After equilibration, the molecular dynamic run is performed for 14 ns using the GROMACS 5.4.1 software package [29] with the all-atom CHARMM27 force field [30]. According to previous works [18,19], this time of simulation is sufficient for achieving equilibrium in all investigated systems. The temperature is maintained at 310 K, using the Vrescale thermostat coupling [31], and isotropic pressure coupling is applied, using the Berendsen algorithm [32].
The interaction of SWCNT with other components of the system is described using the standard Lennard-Jones (LJ) potential for carbon atoms. The LJ parameters for carbon atoms are set up according to CHARMM27 force field, where the atom type CA of benzene included in this force field is chosen to represent the aromatic carbon atoms. Table ST2, supplementary material, shows the values for Lenard Jones parameters in the non-bonded interaction for SWCNTs [16]. The topology and parameters of the studied anticancer drug molecules and Glycine functional groups for the CHARMM27 force field are obtained from the Swiss Institute of Bioinformatics [33]. The atoms in the drug delivery vehicles as well as the anticancer drugs are all modelled as charged Lennard-Jones particles. The structure files of the nanotubes are built using the Nanotube Modeler package [34]. All bonds involving hydrogen atoms are constrained by the LINCS algorithm. The long-range electrostatic interactions are treated with the particle-mesh Ewald method (PME) [35] with a non-bonded cut off 14 Å. The equations of motion are integrated using the velocity form of the Leap-Frog integration algorithm method with a time step of 2 fs. Visualisation and analysis of MD trajectories are implemented by Visual Molecular Dynamics (VMD) software [36].

Results and discussion
The proper equilibration of the simulation systems is monitored by analysis the distance between drugs and SWCNT surface during a trajectory over a time of 14 ns. The obtained results are presented in Figure SF3, supplementary material. Close inspection of Figure SF3, supplementary material, reveals that the distance curves of the equilibrated systems do not fluctuated much during 14,000 ps trajectories. The trends in the distance curves show that the investigated drugs obtain the energetically favorable state on the nanotube. Therefore, 14 ns is the sufficient simulation time for achieving equilibrium in all investigated systems.
The binding free energy (ΔG bind ) of the simulation systems is calculated by MM/PBSA (Molecular Mechanics/Poisson-Boltzmann Surface Area) method through g_mmpbsa module [37] to provide valuable insights into the drug adsorption behaviour on the SWCNTs and f-SWCNTs at the microscopic level. The decomposition of the binding free energy of the simulated systems including the van der Waals (vdW) and the electrostatic contributions over the 14,000 ps simulation time are listed in Table 1. Also, we have calculated the binding free energy value of the individual drug molecule by dividing the total energy value by the number of drug molecules at all simulated systems and the obtained results are tabulated in Table 1.
Results show that the relative binding free energy order for the adsorption of EXE molecules on SWCNTs and f-SWCNTs  Table 1 reveals that most negative binding free energy value of LTZ-SWCNT systems is obtained for the interaction of LTZ molecules with zigzag carbon nanotube.
Inspection of Table 1 shows that the strength of the intermolecular interaction of EXE drug molecules with (16,16) and (16, 0) SWCNTs is increased upon functionalisation of the nanotubes; in fact, the binding free energy for one drug molecule is decreased from −75.640 kJ/mol to −79.094 kJ/mol and from −67.205 kJ/mol to −81.722 kJ/mol in the armchair and zigzag structures, respectively. As well, the minimum binding free energy value is obtained for the interaction of SWCNT (16,8) with LTZ drug molecules. After functionalisation, the binding free energy of LTZ-f-SWCNT (16,8) system is −160.105 kJ/mol which is about 35 kJ/mol bigger than LTZ-SWCNT (16,8) system.
The binding free energies range from −125.517 to −596.057 kJ/mol and from −160.105 to −373.386 kJ/mol for the adsorption of three drug molecules on the pristine and functionalised carbon nanotubes, respectively. Simulation results show that among the three considered drugs, Fulvestrant drug exhibits the strongest affinity for the adsorption on the pristine SWCNT (16,8) and f-SWCNT (16, 0) in terms of the binding free energy values.
The vdW energy values of the simulation systems are normalised in some way, by dividing their total value by the number of atoms and the obtained results are tabulated in Table 1. It is found that FLV drug has more negative vdW energy values at all simulation systems. It can be stated that the most negative E vdW energy for FLV is due to largest number of atoms in this molecule.
It is observed that the binding strength differs for various SWCNT chiralities and the same drug molecule. The density of carbon atoms within the SWCNT walls is the same no matter what chirality we consider. Since all pristine carbon nanotubes with the same length contain conjugated aromatic rings; therefore, it can be stated that the physical properties of the pristine nanotubes is behind these differences. For this purpose, the aromaticity of the different pristine SWCNTs is evaluated using the harmonic oscillator model of aromaticity (HOMA) [38] by applying the density functional theory (DFT) calculations under B3LYP/6-31G** level using Gaussian03 program [39]. We have been calculated HOMA index for the central six-membered rings of the pristine SWCNTs and the obtained results are presented in Table ST3, supporting information. The HOMA philosophy is measuring the aromaticity by comparison with the values of a specific aromatic manifestation in indisputable aromatic systems such as benzene. A higher HOMA aromaticity index corresponds to a more delocalised p-electronic system. Inspection of Table ST3, supporting information, reveals that all pristine carbon nanotubes are aromatic, although chiral nanotubes are less aromatic than the other nanotubes. Therefore, it is clear that FLV with high hydrophobic characteristic prefers to adsorb on the (16,8) SWCNT that have less aromaticity.
Since both drugs and Glycine linkages have polar groups, the electrostatic interaction between them probably has important influence on the drug-f-SWCNT interactions. It is observed that the electrostatic interaction energies of drug-f-SWCNTs are greater than those in drug-SWCNTs complexes (see Table 1).
As shown in Figure 1, Glycine has one amine group (-NH 2 ) and one hydroxyl group (-OH) as well as one oxygen with double bond (-C = O) that can participate in formation of hydrogen bond with hydrogen atoms. Due to lack of the hydrogen bond donor and acceptor centres in the pristine SWCNTs, no hydrogen bonding exists between SWCNTs and drug molecules in the drug-SWCNTs simulation systems. Therefore, the introduction of the Gly functional groups into the carbon nanotube structures provides an additional interaction site with drug molecules. Therefore, several hydrogen bonding can be established between the studied functionalised nanostructures and the electronegative atoms of the adsorbed drug molecules.
The snapshots corresponding to MD simulations of the most stable systems for SWCNTs and f-SWCNTs with studied drug molecules are presented in Figures 2 and SF4, supporting information. As can be seen from Figure SF4, supporting information, the studied drug molecules tend to adsorb via their six-membered aromatic rings to the hexagonal ring of the pristine SWCNT by π-π stacking interactions. The geometric considerations of Figure 2 reveal that drug molecules can be arranged around the tip of the nanotube which in the functional groups are exist. Such a connection is the result of the electrostatic interaction of charged Gly groups of the nanotubes with the active centres of drug molecules. The polar interactions. i.e. π-π interactions and HBs can be clearly seen in Figure 2.
To assess the structural stability of the studied drug molecules after adsorption, the number of intermolecular HBs between drug molecules and f-SWCNTs is computed for the considered components using the gmx hbond analysis in the GROMACS as shown in Figure 3. In this analysis, we consider a bond to be a hydrogen bond only if it meets the following two criteria: the donor-acceptor distance is smaller than 0.35 nm and the hydrogen donor acceptor angle is smaller than 30° [ 40]. Table 2 reported the values of maximum, minimum and average of hydrogen bonds for the intermolecular interactions of drug molecules with Gly functional groups at the simulation systems.
As shown in Figure 3, the number of intermolecular HBs between FLV drug molecules and f-SWCNTs exhibits different pattern compared to other drugs. As can be seen from Figure 3, the affinity of Fulvestrant drug molecules with zigzag f-SWCNT is enhanced as a result of the formation of many hydrogen bonds between its active sites and the functional groups of f-SWCNT. Interestingly, the strong attraction of FLV to f-SWCNTs in comparison to EXE and LTZ molecules mainly attributes to the existence of more fluorine, sulfur and oxygen atoms as well as hydroxyl functional groups in its structure which can be caused that these electronegative atoms linked the Glycine functional groups able to form HBs with the SWCNT and strengthens the interaction (see Figure 1). Therefore, the existence of more number of the intermolecular HBs interactions between FLV drug molecule and functionalised nanotubes is expectable. Consequently, due to more number of HBs in this simulation system, there are more numbers of hydrophilic interactions between FLV drug molecules and Glycine functional groups of (16, 0) f-SWCNT. These results are well agreement with the binding free energy value of the considered system. Although, the H-bonds profiles for EXE and Table 1. The average calculated vdW (E LJ ) and the electrostatic (E elec ) energy values as well as the binding-free energy values of three drug molecules (all in kJ/mol) for all the simulation systems (values in parenthesis refer to calculations of the normalised vdW energy and the binding free energy values of the individual drug at the simulated systems, respectively). LTZ are remarkably similar as shown in Figure 3, it is expectable that LTZ drug with more heteroatom donor centres in its structure possess more number of hydrogen bonds during the intermolecular interactions with Gly functional groups in comparison to EXE molecule (see Figure 1). This conclusion can be manifested via the obtained results of Table 2. It is found that FLV drug molecules with the long hydrophobic chain in their structures have a stronger interaction with hydrophobic surface of (16,8) SWCNT. Also, Fulvestrant molecules with polar properties because of more polar groups in their chemical structures interact with Glycine functional groups of (16, 0) f-SWCNT via hydrophilic interactions as previously discussed. Exploration of the detailed interactions between FLV and (16, 0) f-SWCNT indicates that FLV drug has the strongest binding strength with (16, 0) f-SWCNT due to the favorable van der Waals, hydrophilic and hydrophobic interactions.
Furthermore, visual inspection of Table 2 reveals that the average number of hydrogen bonds between drug molecules and functionalised nanotube decreases throughout the adsorption of LTZ on the functionalised nanotube surface with (16,8) chirality. This result confirms that this simulation system shows the weakest interaction of LTZ-f-SWCNTs.
Moreover, the number of atomic contacts between the studied drug molecules and the pristine SWCNT and f-SWCNTs are calculated using gmx mindist module within a certain radius from MD production trajectory with the following equation: (1) Here, N SWCNT and N DRG are the total number of carbon nanotubes and drug atoms, respectively, and r j is the distance of jth atom of drug molecules from ith atom of the SWCNT, which are at a distance smaller than 0.6 nm from each other. The delta function in Equation (1) is the distance of the chosen index groups of interest which the threshold distance of 0.6 nm corresponds roughly to two atomic van der Waals radii. In other words, the d function ensures counting of all drug atoms within 0.6 nm from SWCNT atoms. The obtained results of these calculations are presented in Figure 4. The simulation results show that the number of atomic contacts of EXE drug molecules with SWCNTs (16, 0) is increased by functionalisation of carbon nanotube. Based on these results, it can be stated that by approaching drug molecules towards the surface of functionalised carbon nanotubes, the number of atomic contacts between these fragments is increased and the binding free energy becomes more negative (see Table 1). Close inspection of Figure 4 reveals that the number of atomic contacts of EXE molecules with (16,8) f-SWCNT is lower than that with (16,16) and (16, 0) functionalised nanotubes. These results are well agreement with the obtained binding free energy values of the considered simulation systems. It is found that the maximum number of atomic contacts between FLV molecules and the external surface of f-SWCNT (16, 0) is associated with the maximum number of hydrogen bond between these fragments (see Table 2).
The radial distribution function (RDF), g(r), of drug molecules with respect to the nanotubes surface is calculated using gmx rdf analysis from the last 5 ns production trajectory  as shown in Figure 5. RDF is the probability of finding a drug molecule at a spherical shell of a certain thickness at a distance (r) from the nanotubes surface. The position of the first peak obtained in the plot of g(r) versus r, represents the closest distance between the two groups, and the probability that one group appears at this distance would be demonstrated by the first peak height in this plot [41]. In some cases, the g(r) profile for the distribution of drug molecules around the functionalised carbon nanotube surfaces is relatively much higher in comparison to drug-SWCNTs systems. Close inspection of Figure 5 reveals that the strength of the g(r) profiles between the f-SWCNT (16, 0) and f-SWCNT (16,16) and the EXE drug molecules are higher when compared to the g(r) magnitude of the EXE-SWCNTs. Also, inspection of RDF patterns of EXE-f-SWCNTs reveals that the interaction of EXE drug with (16,8) f-SWCNT has the shortest peak intensity in comparison to other functionalised carbon nanotubes. These results indicate that the chiral functionalised carbon nanotube attracts EXE molecules less favorably. Also, LTZ drug molecules presented the stronger peak g(r) with f-SWCNT (16,8) in comparison to chiral pristine model. These results for EXE and LTZ drug molecules are acceptable as shown by significant changes in the binding free energy values of Table 1. In other words, the same trend is obtained for binding free energy values between the studied drug molecules and the mentioned nanotubes.
Close inspection of RDF pattern of FLV drug with (16, 0) f-SWCNT illustrates that the distribution of drug molecules lies at the radial range of 0.2 up to 1.2 nm with the maximum distribution functions at r = 0.38 nm. The shorter distance of Fulvestrant molecules with respect to f-SWCNT (16, 0) causes these two species located in a better situation to create more number of the hydrogen bond interactions and other kinds of interactions such as van der Waals ones. These results are in agreement with the obtained results of the number of HBs and vdW energy values. Generally, the total number of hydrogen bonding for the interaction of FLV drug with (16, 0) f-SWCNT is more than those with armchair and chiral functionalised nanotubes which causes the minimum distance and maximum atomic contacts of this drug with zigzag f-SWCNT.
It should be noted that subsequent to adsorption, the carrier needs to deliver its payload to the cells. Thus, we have simulated the release profile of FLV drug from the surface of f-SWCNT (16, 0) as a best drug carrier using Potential of Mean Force (PMF) calculation. To calculate PMF, we have employed the Umbrella Sampling (US) technique [42]. This method adopts a biased potential function imposing on the structure generated by the steered molecular dynamic (SMD) along the reaction coordinate. First, we have conducted SMD simulation to pull FLV drug far away from f-SWCNT at a constant speed of 0.005 nm/ps with a spring constant of 1000 kJ mol −1 nm 2 . A series of initial configurations is generated along the reaction coordinate which 11 of these configurations have served as the starting configurations for the umbrella sampling windows. The distance between the centres-of-masses of FLV drug and f-SWCNT (16, 0) is harmonically restrained. In each window, the complex structure is subjected to a 100 ps equilibration, followed by a 10 ns production simulation under a harmonic force constant of 1000 kJ mol −1 nm 2 . The centre of mass coordinates of FLV collected during the US simulations are used in constructing its PMF via the weighted histogram analysis method [43]. Figure SF5, supporting information, clearly shows that the release of the drug from the nanotube is spontaneous, since low energy barrier exists.
4. Sensitivity and selectivity mechanism of drug delivery vehicles towards the anti-cancer drug molecules In this section, in order to predict the competitive adsorption behaviour of FLV, EXE and LTZ drug molecules on the surface of the functionalised carbon nanotube with (16, 0) chirality as the best nanocarreir, one simulation box has been defined. The initial configuration in the case of the f-SWCNT (16, 0) is built by placing the mentioned drug delivery vehicle in the centre of box and three FLV, three EXE and three LTZ molecules at certain distances from the nanotube axis. Next, the TIP3P water model has been used for filling the simulation box with the initial size of 6 × 6 × 6 nm 3 . The van der Waals and electrostatic energy calculations between FLV, EXE, LTZ and f-SWCNT in water solution are tabulated in Table ST4, supporting information. From the energetic prospective, all three drug molecules can adsorb on the drug delivery system to some extent, which FLV drug shows more negative van der Waals and electrostatic energies. These results indicate high chance of FLV adsorption on the f-SWCNT with respect to LTZ and EXE drug molecules during the competitive adsorption process.
To study the structural information of FLV, EXE and LTZ adsorbed on the functionalised nanotube, the radial distribution function pattern of the simulation system is shown in Figure SF6, supplementary material. RDF patterns between the studied drug molecules and Gly functional groups of the nanotube show that the first peak is sharp and strong with a maximum value of 246.401 at separation of 1.842 nm for FLV and Gly functional groups of f-SWCNT (16, 0), showing a strong interaction between FLV and f-SWCNT. This result also confirms that the FLV has a strong tendency for adsorption on the functionalised nanotube surface in comparison to other kinds of drug molecules. Figure SF7, supplementary material, shows a quantitative examination of the number of atomic contacts of drug molecules with the nanotube surface. Furthermore, the average number of hydrogen bonds between drug molecules and f-SWCNT (16, 0) have been calculated and shown in Figure  SF8, supplementary material. Figure SF7, supplementary material, reveals that the impact among FLV drug molecules and (16, 0) f-SWCNT increases throughout the simulation trajectory which is accompanied by the formation of more hydrogen bonds between FLV drugs and functional groups of (16, 0) SWCNT (see Figure  SF8, supplementary material). As expected, FLV drug is preferentially adsorbed to functionalised carbon nanotube compared to EXE and LTZ molecules.

Conclusion
In summary, we have examined the molecular dynamics simulations to investigate the effects of Glycine functional groups on the dynamics and binding strength of the anticancer drugs during the adsorption process on the SWCNT. By comparison of time evolution of the binding energy values for drug adsorption in the absence and the present of the functional groups, it is found that FLV drug presents more tendency for the adsorption on the (16,8) SWCNT and (16, 0) f-SWCNT surfaces contrary to the other considered drug molecules in the simulation study. Studies reported in this work are devoted to exploring the possible applications of the carbon nanotubes for the drug transportation and delivery of a variety of therapeutic and diagnostic agents.

Disclosure statement
No potential conflict of interest was reported by the authors.