Introduction

The cellulose is an abundantly available natural polymer. The cellulose nanofibers (CNF) are generally isolated from various agricultural biomass and renewable waste (Cherian et al. 2008; Deepa et al. 2015). Usually, the waste is hardly utilized and mostly burned, which lead to a massive release of pollutants in the air and deteriorate the soil fertility. An alternative approach for efficient utilization is in the paper and pulp, textile industries, or as animal feed (Salah 2013; Bardet and Bras 2014). The lemongrass waste after oil extraction is of limited use, such as animal feed and composting, which is not a cost-effective approach. Every year, almost 30,000,000 tons of residual biomass produced at the farmer’s field (Kaur et al. 2010). The green synthesis of nanofibers using agricultural waste could be an eco-friendly and cost-effective approach to utilize tons of agricultural waste. Therefore, enzyme-mediated nanofiber synthesis has gained immense importance, as it is a milder treatment and does not require prolonged steps and chemicals (Filson et al. 2009; Kumari et al. 2019).

Nanoparticle-based drug delivery has gained importance, especially sustained and targeted release systems are undergoing rapid development by employing various experimental approaches (Costache et al. 2009). The production of nanofibers is one of the unique findings in the field of nanotechnology. Natural polymers have gained more interest in controlled drug delivery compared with synthetic polymers due to biodegradability, biocompatibility, availability, and low toxic properties (Esmaeili and Rafiee 2015; Clift et al. 2011; Pereira et al. 2013). Compared with traditional medicinal formulations, the fiber-based drug carrier system has numerous advantages like effective drug therapy and controlled release of drugs with reduced toxicity (Zeng et al. 2003; Hu et al. 2014). Nanofibers are extensively used in biomedicine as a carrier system for controlled drug delivery due to unique properties like large surface area, high porosity, flexibility, and economic (Kentaro Abe et al. 2007). The high surface to volume ratio of the nanofibers helps in improving the encapsulation and transportation of drugs (Mehrabi et al. 2017).

The application of CNF as a drug carrier is in high demand, therefore needing to be explored. Controlled drug delivery systems have been used to overcome the challenges of conventional drug formulations. The controlled drug delivery systems allow the delivery of drugs at fixed rates for a predefined time interval (Ali and Ahmed 2018). The nanofibers are utilized in various areas like dental applications, wound healing, drug delivery, and tissue engineering (Bhandari et al. 2017). Besides, due to the presence of multiple binding sites, nanofibers serve as a carrier system for the different drugs like antioxidants, antimicrobials, flavors, enzymes, and functional group compounds (Fonseca et al. 2019; Li et al. 2019; Rezaei et al. 2015; Löbmann and Svagan 2017). These nanofibers protect, stabilize, and prevent aggregation of the compounds, which proves to be beneficial in the controlled release (Bhandari et al. 2017; Valo et al. 2011). The poor solubility of drugs is a known challenge during the drug formulation for drug delivery systems. The camptothecin (CPT) is a plant-based alkaloid, which is available naturally in the stem and bark of Camptotheca acuminate. It has shown antitumor activities against several tumor types (Çirpanli et al. 2009). However, the poor solubility (Gao et al. 2008), instability of lactone ring, toxicity (Natesan et al. 2014), non-tumorogenic actions, rapid clearance (Mi and Burke 1994), and stability at different physiological conditions (Fan et al. 2010) hinder its medical usage. Hence, the need for novel excipients to overcome challenges in drug formulations is continuous (Löbmann and Svagan 2017).

The effectiveness of nanostructures as a drug carrier system is generally affected by the shape, size, and other biophysical/chemical characteristics. For example, polymeric nanomaterials with diameters ranging between 10 and 1000 nm, displayed properties, ideal for an effective carrier system (Silva et al. 2013). In various studies, the CNF has been utilized for the formulation of poorly soluble drugs and considered a promising excipient (Löbmann and Svagan 2017). As a natural nanocarrier, the CNF shows promising potentials in enhancing the efficacy of the drug (Lin and Dufresne 2014). The use of an enzyme-hydrolyzed CNF as a carrier system to effectively encapsulate camptothecin and release the drug in controlled and extended patterns has not been explored previously. Considering the properties of CNF, the aim of the present study is to explore the potential of enzyme-mediated synthesized CNF as a drug nanocarrier. To explore the hypothesis, CPT was taken as an example drug, and its encapsulation efficiency was measured at different composition ratios. FT-IR and isotherm models were employed to explore the drug binding and its mechanism. Further, the controlled release characteristics without the appearance of burst effects and drug release mechanisms were explored using drug release kinetics models.

Materials and methods

Materials

Lemongrass biomass after oil extraction was used as a raw material for the production of cellulose, and later, it was used for the production of CNF. The CNF was prepared according to the protocol published by our group previously (Kumari et al. 2019). The lignocellulosic biomass of lemongrass waste was washed, oven-dried (60 °C) until completely dried, and chopped. Further, the material was exposed to the steam explosion along with 2% NaOH in an autoclave. Then, the material was washed with distilled water until neutral pH followed by de-lignification with 1% acidified sodium chlorite solution at 70 °C for 3 h. The obtained material was washed with distilled water until neutral pH and oven-dried. Then, the resultant fibers were subjected to enzymatic hydrolysis to obtain the CNF. The resultant material was dispersed in 50 mM sodium citrate buffer at pH 4.8, followed by stirring. The enzyme was added to the suspension of cellulose and stirred at 50 °C for 24 h. The reaction was stopped by boiling the sample for 10 min followed by centrifugation and rinsed with a buffer solution of pH 7.4. Then, the material was washed with deionized water and sonicated. The final product was freeze-dried and used for further experiments.

Chemicals

The solvent DMSO was purchased from Merck. The camptothecin was purchased from Sigma-Aldrich. All reagents were of analytical grade or higher.

Preparation of CNF-CPT formulations

The enzyme-hydrolyzed CNF, which was synthesized from biomass of lemongrass, was used in the experiment (Kumari et al. 2019). First, the CNF was dispersed in deionized water, followed by continuous stirring until the complete suspension. The CPT was loaded onto CNF according to the method previously described with few modifications (Nguyen et al. 2015). The diagrammatic representation of the preparation of formulations using CNF has been shown in Fig. 1. The different amounts of CPT, i.e., 3 mg, 5 mg, and 7 mg, were dissolved in DMSO and mixed with 10 mg of CNF dropwise and continuously stirred at room temperature for the binding of CPT and CNF. Then, the samples were centrifuged for 15 min at 10,000 rpm and 4 °C, followed by washing with deionized water. The resultant material was freeze-dried and stored at 4 °C in the closed glass vial. The lyophilized CPT-loaded CNF was re-dispersed in distilled water as per the experimental requirements.

Fig. 1
figure 1

Diagrammatic representation of the preparation of formulations. Cellulose nanofibers (CNF), camptothecin (CPT), and cellulose nanofibers and camptothecin composite (CNF-CPT)

Full size image

Characterization

The surface morphology of all the samples was studied by SEM (QUANTA 250, FEI). The interaction between the drug and CNF was estimated by the Fourier transform-infrared spectroscopy (FT-IR) instrument (PerkinElmer Spectrum BX). The particle size and polydispersity index of the prepared drug-loaded formulations was determined by dynamic light scattering (DLS) technique (Mastersizer 2000, Malvern Instruments, UK).

In vitro quantification of camptothecin

The stock solution of 1 mg/ml of CPT in DMSO was prepared. The stock solution of CPT was serially diluted to obtain five different working concentrations, i.e., 100, 10, 1, 0.1, and 0.01 μg/ml of CPT, to establish a calibration curve. The absorbance was measured at 367 nm using a UV spectrophotometer (Epoch 2 Microplate Spectrophotometer, BioTek Instruments). The regression equation was expressed as y = 0.0084x + 0.0168, R2 = 0.999 (Fig. 3a). Each experiment was performed in triplicate.

Encapsulation efficiency and drug loading capacity of camptothecin

The CPT was used as a model drug for drug loading and release studies. The freeze-dried samples of different composition ratio CPT-CNF formulation were dispersed in dimethyl sulfoxide (DMSO) solvent. The samples were continuously stirred for 2 h and sonicated for 1 h for complete drug dissolution. The samples were centrifuged, and drug content in the supernatant was determined spectrophotometrically at 367 nm using a UV-visible spectrophotometer. The pellet was again re-dispersed in the solvent, followed by a similar method, and drug content was determined. The loading efficiency was calculated with respect to the final amount of drug entrapped in the CNF to the initial quantity of drug used during encapsulation experiments (Eq. 1). The loading capacity was calculated with respect to the final amount of drug entrapped in the CNF to the weight of the nanoparticles (Eq. 2).

$$ \mathrm{Loading}\ \mathrm{efficiency}\left(\%\right)=\frac{\mathrm{Total}\ \mathrm{amount}\ \mathrm{of}\ \mathrm{drug}\ \mathrm{entrapped}\ \mathrm{in}\ \mathrm{pellet}\ }{\mathrm{initial}\ \mathrm{amount}\ \mathrm{of}\ \mathrm{drug}\ \mathrm{taken}\ \mathrm{for}\ \mathrm{drug}\ \mathrm{loading}}\times 100 $$
(1)
$$ \mathrm{Loading}\ \mathrm{capacity}\ \left(\%\right)=\frac{\mathrm{Entrapped}\ \mathrm{drug}}{\mathrm{weight}\ \mathrm{of}\ \mathrm{the}\ \mathrm{nanoparticles}}\times 100 $$
(2)

Adsorption isotherm studies

The equilibrium isotherm could be applied to determine the adsorption mechanism between the drug and carrier systems (Huang and Chen 2009). The two-parameter equations like Langmuir, Freundlich, Temkin, and Dubinin-Radushkevich were applied to analyze the equilibrium isotherm data. The assumption of Langmuir isotherm is homogeneous monolayer coverage of the adsorbate on to the adsorbent, similar energy of adsorption for all sites, free from any binding energy, and no interaction among the adsorbate molecules (Vázquez et al. 2007). The linearized equation of Langmuir isotherm was expressed as Eq. 3.

$$ \frac{1}{\mathrm{Qe}}=\frac{1}{\mathrm{Qo}}+\frac{1}{\mathrm{Qo}.\mathrm{b}.\mathrm{Ce}} $$
(3)
$$ {\mathrm{R}}_{\mathrm{L}=\frac{1}{1+\left(1+\mathrm{b}.\mathrm{C}0\right)}} $$
(4)

where Qe is the amount of adsorbate adsorbed per gram of the sorbent at equilibrium, Qo is the maximum monolayer adsorption (mg/g adsorbent), b is the Langmuir equilibrium constant (L/g adsorbent), C0 is the initial concentration of adsorbate, and RL is the separation factor.

The Freundlich isotherm describes the multilayer heterogeneous coverage onto the surface of the adsorbent with different adsorption sites (Vasudevan and Lakshmi 2012; Vasudevan et al. 2013). The isotherm predicts the adsorption intensity based on the 1/n value, i.e., n = 1 indicates heterogeneous surface, 1/n < 1 indicates normal adsorption, and 1/n > 1 indicates cooperative adsorption (Shaker and Yakout 2016). The linearized form of Freundlich model is expressed in Eq. 5.

$$ \log \mathrm{Qe}=\log {K}_{\mathrm{f}}+\frac{1}{n}\log C\mathrm{e} $$
(5)

where Kf is the Freundlich constant (mg g–1)(mg L–1)–1/nF and 1/n is the heterogeneity factor, Ce is the concentration of adsorbate at equilibrium (mg/g), and n is the adsorption intensity between CPT and CNF.

The Temkin adsorption model reflects indirect interaction between adsorbate-adsorbate throughout the adsorption process and depicts the monolayer coverage. Taking into account the concentrations of intermediate ion, the model concludes that adsorbate binds and covers the adsorbent’s surface; the heat of adsorption would fall linearly instead of logarithmically (Temkin 1940). When the experimental data best fit the Temkin isotherm, the following parameters were predicted, such as B (Temkin isotherm constant related to the heat of sorption). The positive value of B value would demonstrate that CPT binds to CNF in an exothermic process, whereas a negative value would demonstrate an endothermic process. The linearized equation is expressed as Eq. 6.

$$ \mathrm{Qe}=\frac{RT}{B_{\mathrm{t}}}\ln \left({A}_{\mathrm{t}}\right)+\frac{RT}{B_{\mathrm{t}}}\ln \left({C}_{\mathrm{e}}\right) $$
(6)
$$ B=\frac{RT}{B_{\mathrm{t}}} $$
(7)

where T is the absolute temperature in kelvin (298 K), R is the universal gas constant (8.314 J/(mol k)), At is the equilibrium binding constant, B is the constant related to heat of sorption (J/mol), and Bt is the Temkin isotherm constant related to the heat of sorption.

The Dubinin-Radushkevich (D-R) is usually applied to emphasize the process of adsorption taking place on both homogeneous as well as heterogeneous surfaces. It provides information related to interactions between the adsorbate and adsorbent, energy distribution, and type of adsorption (Sarici-Özdemir and Önal 2018). The prediction of the physical adsorption process best fits a multilayer coverage exhibiting Van der Waals interactions, whereas the chemical adsorption prediction best fits monolayer coverage, demonstrating the covalent bond arrangement (Itodo and Itodo 2010). The linear form of D-R is expressed as Eq. 8.

$$ \ln \mathrm{Qe}=\ln \mathrm{Qm}-\upbeta {\varepsilon}^2 $$
(8)
$$ \varepsilon =\mathrm{RTln}\ \left[1+\frac{1}{C\mathrm{e}}\right] $$
(9)

where Qe is the amount of drug adsorbed per unit weight of sorbent (mg/g), Qm is the maximum adsorption capacity (mg/g), β is the D-R sorption mean energy constant (mol2/kJ2), ε is the D-R constant, T (Kelvin) is the temperature, and R = 0.008314 is the gas constant (kJ/mol K).

It was found that when mean free energy value is below 8 kJ/mol, physical adsorption takes place, and when the mean free energy value is more than 8 kJ/mol, then the chemical adsorption occurs (Javadian et al. 2013). The energy value (E) is estimated by using Eq. 10.

$$ \mathrm{E}=1/\sqrt{2\upbeta} $$
(10)

In vitro release studies at different physiological condition

The in vitro drug release profile of drug from CNF was evaluated under sink environment at different simulated physiological conditions. Briefly, freeze-dried CPT-loaded nanofibers were re-dispersed in 10 ml of simulated media of different pH, with constant shaking and incubated at 37 °C, respectively. The pH 1.2 was prepared to mimic the gastric fluid, pH 6.8 to mimic the intestinal fluid, and pH 7.4 corresponds to the physiological conditions (phosphate buffer saline (PBS) solution). The buffer solution of pH 1.2 was prepared using NaCl, HCl, and deionized water, whereas the buffer solution of pH 6.8 was prepared using K2HPO4, NaOH, and deionized water. The release of CPT was determined by taking aliquots of 1 ml of the sample at specified time intervals, and the medium was replaced to retain the sink environment. The drug-loaded particles were separated from the media using centrifugation. The amount of CPT released was quantified using UV-visible spectrophotometer by evaluating the absorbance after precise time intervals at 367 nm. The drug release (%) was estimated with respect to the total amount of drug entrapped into the nanofibers (Eq. 11).

$$ \mathrm{Drug}\ \mathrm{release}\ \left(\%\right)=\frac{\mathrm{Released}\ \mathrm{drug}}{\mathrm{The}\ \mathrm{total}\ \mathrm{amount}\ \mathrm{of}\ \mathrm{drug}\ \mathrm{entrapped}\ \mathrm{in}\ \mathrm{nanofibers}}\times 100 $$
(11)

Mathematical modeling and release kinetics

To understand the mechanism of CPT release from CNF, the CPT release data was evaluated by five release kinetics models (Costa and Sousa Lobo 2001)

  1. (i)

    Zero-order release kinetics (W = Kt), where W is the percentage of drug released at time “t” and K is the release rate constant,

  2. (ii)

    First-order kinetics (log W = log Wo-Kt/2.303), where Wo is the initial concentration of the drug, K is the first-order constant, and t is the release time,

  3. (iii)

    Higuchi kinetics (Wt = KHt1/2), where KH is the Higuchi release rate constant,

  4. (iv)

    Korsmeyer-Peppas kinetics model (Wt/We = Ktn), where Wt/We are the fractional drug release, and K is the constant, and n is the release exponent, which provides evidence about the mechanism of the drug release.

  5. (v)

    Hixson-Crowell kinetics model (Wo1/3 − Wt1/3 = Kt), where Wt is the amount of drug released in time “t,” Wo is the initial amount of drug, and K is the constant.

Results and discussion

Morphology, particle size, and polydispersity index of prepared formulations

The understanding of the mechanism of the drug binding can be analyzed through morphological studies as any dimensional change may provide useful information. Scanning electron microscopy (SEM) of the CNF before and after loading of CPT is shown in Fig. 2. SEM images show the CNF interconnected network-like structure and porous matrix (Fig. 2a). After drug loading, the drug crystals were found adsorbed onto the CNF surface with a matrix-type layered structure (Fig. 2 b, c, and d), which may be due to the physical adsorption of the drug. Previous reports also show similar data (Bhandari et al. 2017). The particle size distribution of the enzyme-hydrolyzed CNF is 105.7 nm published previously (Kumari et al. 2019). However, the increase in particle size was observed after the loading of CPT (Table 1). The particle size distribution was estimated by a dynamic light scattering (DLS) technique. While dealing with polydisperse nanoparticles, DLS is unable to differentiate between nanoparticles with the difference in the diameter as the scattered light of larger particles or aggregates overlays the smaller particles. The DLS generally measures the hydrodynamic diameter of the particles, including hydration shell, polymer shells, or stabilizer, which leads to increment in the particle size (Fissan et al. 2014). After CPT loading, the increment in particle size may be due to the aggregation of the drug molecule onto the CNF. The term polydispersity index (PdI) is used to describe the percentage of size distribution within the sample. The PdI generally ranges from 0.0 (uniform sample) to 1.0 (highly polydisperse samples with numerous particle size distribution). The value of 0.2 and less than 0.2 are acceptable for the polymer-based nanoparticles (Clarke 2013). The PdI value of 0.2 in 10:3 composition ratio of CNF:CPT is acceptable for drug delivery applications (Table 1). It was reported that PdI value of 0.3 and below is found to be suitable and indicates a homogeneous population of lipid-based carriers (Chen et al. 2011; Putri et al. 2017)

Fig. 2
figure 2

SEM micrograph of a cellulose nanofibers alone and drug-loaded cellulose nanofibers at different composition ratio. b 10:3 CNF-CPT. c 10:5 CNF-CPT. d 10:7 CNF-CPT

Full size image
Table 1 Encapsulation efficiency, drug loading, particle size, and polydispersity index of the prepared formulations
Full size table

Encapsulation efficiency, drug loading, and isotherm studies of the prepared formulations

A prerequisite for healthcare formulations using the polymeric carrier system is high entrapment efficiency and drug loading capacity (Aditya et al. 2013). To check the suitability of the nanocarrier as a drug carrier, its ability to encapsulate and slow the release of the active therapeutics was quantitatively established. The encapsulation efficiency (EE) and loading capacity (LC) were calculated based on the solid-phase analysis at different time points (Fig. 3b). The EE and LC were evaluated at different time points to determine the time taken for the maximum incorporation of the drug into the CNF matrix. It was found that the highest EE, i.e., 65.28%, was observed in the 10:3 composition ratio of CNF-CPT, in comparison with 10:5 and 10:7. The quantification of the CPT was done using a calibration curve (Fig. 3a). The EE was decreased with an increase in the amount of CPT. The EE corresponds to the percentage of CPT incorporated into the CNF, from a given amount of CPT used, which is comparatively higher in the CNF-CPT (10:3) ratios. This suggests that as the CPT increases in CNF-CPT ratios, incorporation of the CPT from the solvent into the CNF matrix is limited by the solubility of the drug in water and independent of the quantity of the CNF present (Tam et al. 2016). In the case of a 10:3 ratio of CNF-CPT, the low amount of drug was well solvated and easily incorporated into the matrix of CNF after 3 h (Kumari et al. 2010; Chi et al. 2019). However, as the CPT content increases in other CNF-CPT ratios, the drug content limits its solubility and proper incorporation in the CNF. Due to the presence of high CPT content, the time taken by the 10:5 and 10:7 CNF-CPT ratio is only 30 min. Similar results have been observed previously, which displayed the EE in CNF generally governed by the drug solubility in the solvent and antisolvent and not on the polymer to drug ratios (Chow et al. 2015; Tam et al. 2016).

Fig. 3
figure 3

a Graphical view of calibration curve of camptothecin and b encapsulation efficiency (%) of the drug-loaded fibers at different CNF-CPT ratios at different time intervals

Full size image

The drug loading capacity corresponds to the amount of CPT-loaded per unit weight of the CNF. Results show that the drug loading increases with the amount of drug increase in the CNF-CPT ratios, i.e., 10:3, 10:5, and 10:7 (Table 1), with the constant amount of CNF being used for encapsulating the different amount of drug. Presumably, it is because the interconnected CNF matrix structures were favorable to load hydrophilic CPT and the occurrence of intermolecular interaction via hydrogen bonding (Mishra et al. 2016). So, with the increase in CPT content in the CNF-CPT ratio, the loading capacity of the CNF increases. Few previous reports demonstrate lower loading capacity in the range of 1–5% (Chen et al. 2012), which is much lower compare with enzyme-mediated CNF shown in the present study (Table 1). These results confirm high encapsulation efficiency with a better loading capacity in the 10:3 CNF-CPT ratio. Therefore, this formulation has been chosen for controlled release characteristics.

To demonstrate the mechanism of adsorption of the CPT with the CNF, different isotherms were applied, such as Langmuir, Freundlich, Dubinin-Radushkevich, and Temkin as the CNF make the network-like structure and consist of the porous matrix. The calculated parameters like isotherm constants (b, Kf, ε) and other parameters are shown in Table 2. The correlation coefficient value (R2) was used to find which model best depicts the adsorption mechanism. Considering the R2 value, the adsorption of the CPT onto the CNF fits well with all the four isotherms, i.e., D-R (0.99), Temkin (0.99), Langmuir (0.94), and Freundlich isotherm (0.92) (Table 2). The good fit of the D-R isotherm model is evident that the adsorption occurred on homogeneous as well as heterogeneous surfaces (Pandian et al. 2017). The data demonstrate that the adsorption follows a pore-filling mechanism (Vijayaraghavan et al. 2006). It assumes multilayer characteristics which involve Van der Waals force, relevant for physical adsorption processes (Ayawei et al. 2017). The mean free energy less than 8 kJ/mol, i.e., 0.0016, demonstrates the involvement of physical adsorption in the uptake of CPT onto the CNF, also seen in the SEM analysis (Fig. 2b–d; Table 2).

Table 2 Langmuir, Freundlich, Dubinin-Radushkevich, and Temkin isotherm parameters for loading of CPT on CNF
Full size table

The Temkin isotherm constant B displayed the maximal binding energy or the heat of adsorption that takes place with the coverage of the adsorbent. Based on the Temkin isotherm data, the positive value of constant B is 240.37 J/mol, which indicates the exothermic adsorption process, representing the physical adsorption. Langmuir isotherm demonstrates monolayer and homogeneous adsorption process with all binding sites possessing an equal affinity for adsorbate, i.e., CPT. The separation factor (RL) for CPT used in this study (RL) ranges between 0 > RL > 1 (Table 2), indicating linear adsorption (Dada et al. 2012). The difference adsorption capacity (Table 2) may be due to the difference in adsorption properties (Bée et al. 2011; Dubey et al. 2016).

The coefficient correlation obtained for Freundlich isotherm suggests that CPT also binds in multilayer, indicating the non-restrictive and exponential form of binding. In multilayer adsorption, the CPT molecules are not necessarily similar to the particular number of binding sites on to the CNF surface due to exponential adsorption of CPT molecules, which occurs because of the interaction with other CPT molecules (Liu 2015). Based on the Freundlich isotherm, the 1/n specifies the binding strength between the CNF and CPT. The exponent value showed 1/n > 1 (1.11) depicting the cooperative adsorption (Table 2). The cooperative adsorption usually favored by multilayer coverage of CPT into the CNF. Furthermore, the n value should be in the range of 1 to 10, which is observed from the Freundlich isotherm (Table 2) (approx.1.0), indicating the favorable adsorption between the CPT and CNF (Dada et al. 2012). From SEM images, it was confirmed that the CNF are interconnected network-like structure and porous matrix, which may be forming monolayer as well as multilayer surfaces. Therefore, the good fit of all the isotherms indicates adsorption of CPT on to the CNF matrix according to the availability of overlapping surface matrix. Therefore, both the monolayer and multilayer type of coverage is expected. Similarly, as it is a network-like matrix with possible irregular thickness, therefore, homogeneous and heterogeneous mode of drug binding mechanism is also probable.

Interaction studies of the CPT and CNF

The interaction between the CNF and the drug was also assessed by FT-IR spectroscopy. The FT-IR spectra of CNF, CPT, and CNF-CPT are shown in Fig. 4. Each sample showed strong bands around 3400 cm−1, which were attributed to O–H vibrations and depicted the broad intra and intermolecular hydrogen bonding and hydrophilic behavior of the CNF (Chirayil et al. 2014). It was described that rod-like fibers appeared aggregated due to hydrogen bonding and Van der Waals attractions. The symmetrical and asymmetrical stretching of CH vibrations were observed at 2924 cm−1 in CNF and CNF-CPT (Alemdar and Sain 2008). In the case of CNF-CPT, the two adsorption bands at 2919 and 2850 cm−1 were attributed to CH2 and CH vibrations in CNF and benzene ring of CPT, which confirms the chemical interaction between CPT and CNF (Acevedo-Morantes et al. 2013). The carboxylic group and stretching vibrations of ester and lactone carbonyl groups were observed around 1744 cm−1in CPT and CNF-CPT, which was not present in the spectra of CNF. The adsorption band between 1600 and 1700 demonstrate the aromatic combination, present in the CPT and CNF-CPT complex (Acevedo-Morantes et al. 2013). The peak around 1600 cm−1 is associated with water molecules adsorbed into the cellulose fiber structure in CNF and CNF-CPT samples (Li et al. 2016). It has been reported that the adsorption band between 1600 and 1650 cm−1 demonstrates the stretching vibration of the benzene ring, shown in CPT and CNF-CPT (Zhao et al. 2010). The absorption peak of stretching and bending vibrations were observed at 1431, 1369, 1325, 1054, and 899 cm−1, which were attributed to –CH2, –CH, –OH, and C–O bonds in CNF (Xu et al. 2013). The shifting of peaks in CNF-CPT before and after drug binding showed the intermolecular interaction through hydrogen bonding (Mishra et al. 2016). Hence, the FT-IR results confirm the chemical interaction between the CPT and CNF.

Fig. 4
figure 4

FT-IR spectra of the cellulose nanofibers (CNF), camptothecin (CPT), and camptothecin loaded nanofibers (CNF-CPT)

Full size image

In vitro release of CPT and release kinetics at different pH conditions

Drug release is a complex process during which drug mass transport through diffusion is an important mechanism (Siepmann and Siepmann 2012). The CNF exhibited a slow release of CPT (up to 4.4%) from time 0 to 6 h, followed by which 14.38% release of CPT took place from 6 to 48 h at pH 1.2 condition. Drug release at pH 6.8 also showed a slow release pattern for CPT, i.e., 26.2% release takes place from 0 to 6 h, followed by 63.4% release of CPT from 6 to 48 h. In the case of pH 7.4, the CPT release was observed as 24.8% up to 6 h, and 80.2% CPT release took place from 6 to 48 h (Fig. 5). The results confirmed a slow and extended-release pattern of CPT from the CNF at different simulated pH conditions with no initial burst effect. A slow and extended-release of CPT could be due to the matrix structure and swelling behavior of the carrier system with regard to pH, which is allowing the controlled and stable transport of drugs. The hydrophobic behavior plays a significant role in drug release; the interaction between hydrophobic drugs and hydrophobic core is often responsible for the slow release pattern of the drug from the nanoparticles (Sutton et al. 2007; Tran et al. 2014). The controlled and extended-release pattern without any burst release also demonstrates that CPT was fully adsorbed into the CNF matrices, which is also supported by isotherm studies. The most significant limitations in the drug delivery process are the complications of complete encapsulation and retention, followed by prolonged release and targeted delivery of therapeutics. For complex diseases like cancer, the extended-release nanocarriers are considered as attractive therapeutic agents due to their prolonged drug release behavior (Prosperi et al. 2017). Therefore, in addition to CPT, most likely, these CNF could be beneficial for the delivery of other anticancer drugs.

Fig. 5
figure 5

Drug release graph showing a cumulative release of camptothecin at different pH conditions

Full size image

The drug release mechanism of CPT from CNF at different pH levels was evaluated by employing several mathematical models such as zero order, first order, Higuchi, Korsmeyer-Peppas, and Hixson-Crowell (Costa and Sousa Lobo 2001). Table 3 demonstrates the coefficient correlation values (R2) obtained after the model fitting of release data at different pH. The graphical representation of the fitted models has been shown in Fig. 6. The results showed that release data were best fitted to Higuchi model at pH 1.2 (R2 value 0.98), at pH 6.8 (R2 value 0.97), and at pH 7.4 (R2 value 0.99). The best fitted Higuchi model showed that the release of the drug was governed by the diffusion mechanism at all pH conditions. Moreover, the release data was also good fitted to Korsmeyer-Peppas models at pH 1.2 (R2 value 0.94) and 7.4 (R2 value 0.98) and slightly fitted at pH 6.8 (R2 value 0.90).

Table 3 Mathematical models and correlation values (R2) based on release data
Full size table
Fig. 6
figure 6

Drug release kinetics plots. a Zero order. b Hixson-Crowell. c Korsmeyer-Peppas. d Higuchi models

Full size image

In Korsmeyer-Peppas model, n is the diffusional exponent, which is the characteristic parameter of the transport mechanism. The n value of 0.45 is correlated to Fickian diffusion mechanism and drug release governed by diffusion, 0.45 < n < 0.89 correlated to non-Fickian or anomalous transport and governed by diffusion and swelling, 1.0 < n corresponds to non-Fickian (case II) or zero-order release kinetics, and release of drug governed by swelling or relaxation of polymeric chain (Costa and Sousa Lobo 2003; Mhlanga and Ray 2015). From the results, exponent (n) is 0.23 at pH 1.2, 0.51 at pH 6.8, and 0.91 at pH 7.4 demonstrating that release of drug was controlled by non-Fickian diffusion mechanism at pH 1.2, 6.8, and case II transport mechanism, indicating that the drug release was followed by diffusion, swelling, and polymer relaxation due to erosion of the polymer (Gao et al. 2013). The Korsmeyer-Peppas model helps to understand the contribution of diffusion and polymer relaxation in the drug release pattern. The release data were found not best fitted to the zero-order and Hixson-Crowell models at any of the pH conditions and did not show a higher correlation coefficient at pH 1.2 and 6.8. The data was not fitted to the first-order equation, so the graphical representation was not shown in the manuscript.

Conclusion

The present study showed an application of a green synthesized biodegradable CNF as a promising nanocarrier system. A CNF-CPT composite has been prepared. The morphological and isotherm studies showed the physical absorption of CPT into the CNF matrices. The CNF-CPT formulation of a 10:3 ratio displayed a controlled and extended-release of CPT at three different physiological pH conditions, which is generally an essential requirement in the pharmaceutical applications. The controlled and extended drug release profile is governed by diffusion, swelling, polymer relaxation/erosion, and drug carrier interaction mechanisms as predicted by the release kinetics studies. Henceforth, it is concluded that the green synthesized CNF could be a promising nanocarrier for the problematic drugs. However, in the future, cell-based and animal studies are needed to demonstrate the complete efficacy of the nanocarriers.