Fabrication and in vitro characterization of CPT-loaded amphiphilic CD NPs

CPT-loaded amphiphilic CD nanoparticles have been previously optimized in our laboratories [9], as reported, NPs using two different amphiphilic CDs were prepared and 6-O-capro-β-CD nanoparticles coated with chitosan (CS) to obtain a positively charged surface. In vitro characterization and cell culture studies for 6-O-capro-β-CD, CS-(6-O-capro-β-CD), and poly-β-CD-C6 formulations have been comprehensively evaluated previously [9]. According to the pre-formulation studies, an optimal formulation with desired characteristics was determined as CPT/poly-β-CD-C6 NPs with a 135 nm particle size, very low polydispersity index, and a zeta potential of + 40 mV. In vitro release experiments showed that amphiphilic CD NPs have properties suitable for colon targeting, but the most promising were poly-β-CD-C6 NPs with 52% of encapsulated CPT successfully delivered all the way to the simulated colon. When compared to the equivalent CPT dose in solution, CPT-loaded poly-β-CD-C6 nanoparticles exhibited higher cytotoxicity in HT-29 cells. Permeability studies performed with the Caco-2 cell line revealed a 276% increase in drug permeability and significantly higher intestinal penetration with the cationic CD formulation. In our further research [8], it was also reported that the oral CPT-loaded poly-β-CD-C6 NPs showed antitumoral and antimetastatic effects in a colorectal tumor-bearing animal model.

Drug release from amphiphilic CD nanoparticles

In vitro release studies were performed over 48 hours in order to clearly elucidate the release kinetics (Figure 1). An in vitro release study was carried out at 0–2 hours in simulated gastric fluid (SGF), 2–5 hours in simulated intestinal fluid (SIF), then in simulated colonic fluid (SCoF) settings till the completion of the experiment in order to imitate GIT circumstances in terms of pH and transit duration. The purpose of the release study was to elucidate the ability of the formulation to retain the encapsulated drug in the stomach and small intestine and preferably release it when it reaches the colon. The optimum nanoparticle formulation was considered to deliver most of the effective lactone-form CPT to the colon.

Figure 1: In vitro release profile of CPT from nanoparticle formulations (n = 3, ± SD).

It is known that it takes approximately 5 hours for oral drug delivery systems to reach the colon; the first 2 hours in the stomach and the last 3 hours in the small intestine [30]. At the end of the 5th hour, 6-O-capro-β-CD and CS-(6-O-capro-β-CD) formulations revealed faster release profiles (p > 0.05) than poly-β-CD-C6 nanoparticles. Poly-β-CD-C6 nanoparticles showed a slower release of CPT (48%) until the colonic area as compared to the other formulations (p < 0.05). Previous research has looked into detailed assessments of this topic [9].

Release kinetics study

The in vitro release profiles of CPT-loaded amphiphilic cyclodextrin nanoparticles were fitted with a variety of kinetic models, and the release mechanisms, which are illuminating markers for novel drug delivery systems, were mathematically investigated. In this context, 6 models (first order, Hopfenberg, Korsmeyer–Peppas, Higuchi, Peppas–Sahlin, and Weibull models) and 3 criteria (coefficient of determination (R2), Akaike information criterion (AIC) and model selection criterion (MSC)) were evaluated for the in vitro release profiles. Much of the research in this field generally evaluates the kinetic data of the total release profiles of the nanoparticles, although it is useful to look at potential alterations in the release kinetics at different release mediums (SGF, SIF, SCoF) as well, especially in orally administered drug delivery systems. To achieve this, a thorough and in-depth release kinetic study was conducted, and the parameters were compared for the GIT conditions. Table 1 displays the findings of the release kinetic modeling studies and graphical reports are presented in Figure 2, Figure 3, and Figure 4. Figures 2–4 show that the kinetic models' predicted and observed CPT releases appear to be consistent with formulations for the best correlated models. Thus, the mathematical compatibility of the kinetic models' graphics with good correlation was also proven. Furthermore, as seen in Table 2, the release profiles of CPT from different formulations were compared in terms of similarity (f2) and difference (f1) factors, and the results revealed that the release profiles of nanoparticles, which we obtained using the same formulation parameters with structurally similar amphiphilic cyclodextrin derivatives, showed similar release profiles. Formulation parameters affected the release kinetics of the drug-loaded nanoparticles [31,32].

Table 1: Release kinetic modelling and results of NP formulations.

Model and equation SGF SIF SCoF 0–2 hours kinetics 2–5 hours kinetics 5–48 hours kinetics R2 AIC MSC n/m* R2 AIC MSC n/m* R2 AIC MSC n/m* 6-O-Capro-β-CD                         First-order
F = 100·[1−Exp(−k1·t)] 0.823 16.307 0.328 – 0.765 24.496 0.947 – 0.929 21.434 2.243 – Higuchi
F = kH·t^0.5 0.991 4.366 3.313 – 0.606 26.554 0.432 – −3.300 41.940 −1.859 – Korsmeyer–Peppas
F = kKP·t^n 0.993 5.205 3.103 0.471 0.982 16.301 2.996 1.319 0.861 26.790 1.171 0.163 Peppas–Sahlin
F = k1·t^m + k2·t^(2·m) 0.994 6.564 2.764 0.450 0.976 19.342 2.235 0.450 0.975 20.277 2.474 0.450 Hopfenberg
F = 100·[1−(1−kHB·t)^n] 0.809 18.610 −0.248 3.000 0.915 22.416 1.467 1.000 −2.188 42.443 −1.959 3.000 Weibull
F = 100· 0.997 3.640 3.495 – 0.946 22.605 1.420 – 0.987 17.087 3.112 – CS-(6-O-Capro-β-CD)                         First-order
F = 100·[1−Exp(−k1·t)] 0.830 17.597 0.406 – 0.846 21.276 1.368 – −0.370 36.415 −0.715 – Higuchi
F = kH·t^0.5 0.981 11.809 2.603 – 0.685 24.130 0.654 – −2.498 41.103 −1.652 – Korsmeyer–Peppas
F = kKP·t^n 0.980 10.967 2.064 0.507 0.949 18.874 1.968 1.039 0.886 25.998 1.369 0.176 Peppas–Sahlin
F = k1·t^m + k2·t^(2·m) 0.983 12.359 1.716 0.450 0.950 20.802 1.486 0.450 0.983 18.406 2.887 0.422 Hopfenberg
F = 100·[1−(1−kHB·t)^n] 0.814 19.964 −0.185 3.000 0.944 19.206 1.885 1.000 −2.105 42.508 −1.933 3.000 Poly-β-CD-C6                         First-order
F = 100·[1−Exp(−k1·t)] 0.805 13.614 0.192 – 0.847 20.320 1.377 – 0.992 16.096 4.389 – Higuchi
F = kH·t^0.5 0.995 −1.199 3.895 – 0.627 23.886 0.485 – 0.404 37.459 0.117 – Korsmeyer–Peppas
F = kKP·t^n 0.999 −18.653 8.259 0.429 0.978 14.572 2.814 1.362 0.862 32.156 1.177 0.318 Peppas–Sahlin
F = k1·t^m + k2·t^(2·m) 0.999 −20.534 8.729 0.450 0.975 17.011 2.204 0.450 0.935 30.393 1.530 0.450 Hopfenberg
F = 100·[1−(1−kHB·t)^n] 0.795 15.806 −0.356   0.927 19.383 0.952 1.000 0.985 21.122 3.384 3.188 Weibull
F = 100· 0.996 −7.005 5.346 – 0.967 18.149 1.920 – 0.991 20.463 3.516 –

Figure 2: Results for release kinetics obtained automatically by the DDSolver software for SGF release medium (*represents best fit models).

Figure 3: Results for release kinetics obtained automatically by the DDSolver software for SIF release medium (*represents best fit models).

Figure 4: Results for release kinetics obtained automatically by the DDSolver software for SCoF release medium (*represents best fit models).

Table 2: Difference and similarity factors between formulations.

CPT-loaded amphiphilic CD nanoparticles difference factor (f1) similarity factor (f2) 6-O-capro-β-CD CS-(6-O-capro-β-CD) 6.63 73.82 CS-(6-O-capro-β-CD) poly-β-CD-C6 13.73 58.86 poly-β-CD-C6 6-O-capro-β-CD 11.72 61.57

According to the release kinetic parameters in SGF medium, as seen in Table 1, the highest R2, MSC and lowest AIC values were observed in the Weibull model for 6-O-capro-β-CD and CS-(6-O-capro-β-CD) formulations, and in the Korsmeyer–Peppas and Peppas–Sahlin models for the poly-β-CD-C6 formulation. For the poly-β-CD-C6 NPs, two models were found to be compatible with high correlation. There are also studies in the literature indicating that the release kinetic model of nanoparticles can fit to more than one model [33,34]. Since it is the first study to evaluate the release kinetics of amphiphilic cyclodextrin nanoparticles, we have reported that a drug release profile that fits more than one model can be observed for amphiphilic cyclodextrin nanoparticles. In the Weibull model, the “β” (shape parameter of the release curve) value is a criterion used to illuminate the release from a polymeric matrix. “β” ≤ 0.75 indicates Fickian diffusion, while 0.75 < “β” < 1 indicates Fickian diffusion and controlled release combination [35]. The “β” value for the Weibull model was calculated as 0.396 and 0.434 for the 6-O-capro-β-CD and CS-(6-O-capro-β-CD) nanoparticle formulations, respectively. According to the Weibull model, CPT release kinetics from nanoparticles were found to be compatible with Fickian diffusion in SGF medium [36]. In the model-independent principal evaluation of in vitro release profiles, this was considered as the rapid/burst and initial release of the drug adsorbed on the nanoparticle surface or encapsulated in the nanoparticle material matrix. It has been confirmed by mathematical modeling that the release is based on diffusion. This indicates the release of the CPT, which is weakly bound in the nanoparticle matrix and adsorbed on the surface, for the 6-O-capro-β-CD and CS-(6-O-capro-β-CD) formulations. These results we obtained confirm each other with the data we interpreted in our previous studies [9]. The release kinetics, however, also appeared to be consistent with the Korsmeyer–Peppas and Peppas–Sahlin models for poly-β-CD-C6 NPs. While the Korsmeyer–Peppas model expresses diffusion-controlled release from matrix-type nanosystems, the Peppas–Sahlin model is based on the combination of diffusion and erosion of the nanoparticle matrix. In order to further elucidate the kinetics of these models, the diffusional exponent values (n or m) regarding the release kinetics from the nanoparticles were computed [37]. In the Korsmeyer–Peppas model, "n" represents the diffusional exponent illustrating the drug release mechanism, but in the Peppas–Sahlin model, "m" represents the same parameter [38]. In this context, "m" and "n" diffusional exponent values were computed as 0.450 and 0.429, respectively. A diffusional exponent (m/n) ≤ 0.45 indicates that Fickian diffusion is a factor in drug release [39]. For 0.45 < m/n < 0.85, the drug release occurs through a non-Fickian diffusion mechanism, for m/n = 0.85 the release occurs by case II transport and m/n > 0.85 indicates super case II transport [38-42]. It has been determined that there is Fickian diffusion in the release kinetics based on the diffusional exponent values of the Korsmeyer–Peppas and Peppas–Sahlin models [39]. When all the data were analyzed together, it was determined that a single kinetic model was not dominant in the SGF release kinetics, and compliance with different models was observed. However, although the models were different in all formulations, it was observed that the dominant mechanism was diffusion-based release. In our previous studies, it was evaluated that the drug release observed in SGF might be related to the diffusional release of the surface-adsorbed drug and the poorly bound surface drug to the matrix. In the mathematical modeling data, it has been confirmed that the first release seen in the SGF medium is dominantly diffusion-related.

According to the release kinetic parameters in SIF medium, as seen in Table 1, the highest R2, MSC and lowest AIC values were observed in the Korsmeyer–Peppas model for 6-O-capro-β-CD and poly-β-CD-C6 NPs, and in Korsmeyer–Peppas and Peppas–Sahlin models for CS-(6-O-capro-β-CD) formulation. Two models were found to be compatible with high correlation for the CS-(6-O-capro-β-CD) formulation for SIF medium. Similarly, studies showing that nanoparticles can fit more than one model in the literature were also mentioned for SGF data [43]. The Korsmeyer–Peppas model difussional exponent values (n) for 6-O-capro-β-CD and poly-β-CD-C6 NPs were computed as 1.319 and 1.362, respectively. Considering the difussional exponent data over 0.85 indicates that the release mechanism is compatible with super case II transport. Case II transport refers to the release that occurs as a result of relaxation of the polymeric structure [40,42]. These results were interpreted as supporting our idea that the release of the drug adsorbed to the surface is completed by diffusion in the SGF medium, and that the erosion of the nanoparticle material and the relaxation of the polymer chain begins and accelerates the release in SIF. When a further evaluation was made for the CS-(6-O-capro-β-CD), which showed a high correlation between the two models, the diffusional exponent values of the Korsmeyer–Peppas and Peppas–Sahlin models were calculated as 1.039 and 0.450, respectively. Similarly, the n value of Korsmeyer–Peppas above 0.85 indicates that the release mechanism is realized by super case II [44]. This value was interpreted as indicative of the release seen with the erosion of the nanoparticle material and the initiation of polymer relaxation [45]. On the other hand, the “m” value calculated as 0.45 in the Peppas–Sahlin model indicates Fickian diffusion. This situation was evaluated as a very significant and meaningful data when compared with other formulations. Unlike the other two formulations, CS-(6-O-capro-β-CD) is coated on its surface with chitosan, a cationic coating material. The theoretical interpretations so far have been that the drug can be adsorbed in the coating material or weakly bound to the coating polymer structure, and it will be released first. The data obtained from the kinetic modeling provided results that support this interpretation. For the 6-O-capro-β-CD and poly-β-CD-C6 NPs (uncoated formulations), it was confirmed that the release occurred as a result of the relaxation of the nanoparticle material in SIF, while the Fickian diffusion continued for the CS-(6-O-capro-β-CD) formulation, that is, the release of the weakly bound drug adsorbed on the coating material. The diffusional exponent values in SIF for CS-(6-O-capro-β-CD) showed that the release continues as a combination of both the diffusion release of the drug adsorbed to the coating material, chitosan, and the case II release, which occurs as a result of the relaxation of the nanoparticle polymer structure [46].

According to the release kinetic parameters in the targeted main release medium, SCof, the highest R2, MSC, and lowest AIC values were observed in the Weibull model for 6-O-capro-β-CD and CS-(6-O-capro-β-CD) formulations, and in the first order release and Weibull models for the poly-β-CD-C6 NPs, as seen in Table 1. In the Weibull model, the “β” (shape parameter of the release curve) exponent is a parameter used to elucidate the release from a nanoparticle matrix. “β” ≤ 0.75 indicates Fickian diffusion, while 0.75 < “β” < 1 indicates a complex mechanism (Fickian diffusion and controlled release). For values of “β” higher than 1, it was demonstrated that the drug transport follows a complex release mechanism [35,47,48]. The “β” value for the Weibull model was calculated as 0.493 and 0.401 for the 6-O-capro-β-CD and CS-(6-O-capro-β-CD) NPs, respectively. When evaluated within the framework of the literature, it was determined that the release mechanism of encapsulated CPT in SCof medium is by Fickian diffusion [36]. In addition, this situation has also been interpreted as further relaxation of the nanoparticle matrix structure in the SCoF medium, making diffusion easier and coming to the fore as a primary release mechanism [49]. On the other hand, the other formulation, the poly-β-CD-C6 NPs, was found in accordance with both Weibull and first order kinetics. According to the Weibull model, the “β” value was calculated as 0.762. Within the framework of the information explained above, it was evaluated as a complex (Fickian diffusion and controlled release) release mechanism according to the Weibull model for the poly-β-CD-C6 NPs. As stated in the literature, values of “β” in the range of 0.75–1.0 indicate a combined mechanism which is frequently encountered in release studies. When the power law can adequately represent the whole collection of data in these situations, further confirmation can be gained. The special case of “β” = 1 is compatible with first order release, whereas the concentration gradient in the dissolution medium drives the rate of release [35]. In our calculations, results compatible with first order kinetics were found for the poly-β-CD-C6 and it was considered to fit both models. The partially high “β” value for the Weibull model also confirmed the tendency towards first order kinetics, which is also evaluated in the previous sentence in line with the literature. In this context, it has been evaluated that the first order kinetics associated with diffusion in the SCoF medium for the poly-β-CD-C6 formulation also occurs as a release mechanism. It was observed that the Weibull and first order models were compatible, supported and confirmed each other, providing an explanatory idea about CPT release from the formulation.

Cell culture studies Determination of IC50 values of camptothecin

CT26 and HT29 cells were incubated with increasing concentrations of CPT and different CD nanoparticle formulations containing equal amounts of camptothecin for 48 or 72 hours. When the incubation period was over (48 or 72 h), cell viability was determined with the WST-1 assay. IC50 values are shown in Table 3.

Table 3: IC50 (µM) values of CPT solution and CPT-loaded CD nanoparticle formulations for CT26 murine and HT29 human colon cancer cell lines at 48 h and 72 h (n = 6, mean ± SD).

Formulation CT26 HT29 48 h 72 h 48 h 72 h CPT/6-O-capro-β-CD 1.23 ± 0.02 1.19 ± 0.06 0.76 ± 0.09 0.58 ± 0.14 CPT/CS-(6-O-capro-β-CD) 0.72 ± 0.26 0.59 ± 0.12 0.89 ± 0.07 0.59 ± 0.16 CPT/poly-β-CD-C6 1.35 ± 0.46 0.61 ± 0.14 0.30 ± 0.03 0.25 ± 0.04 CPT solution in DMSO 1.86 ± 0.28 1.27 ± 0.42 1.47 ± 0.06 1.31 ± 0.06

It was observed that the IC50 values of the drug solution and nanoparticle formulations in each cell line were different and the IC50 values of CPT-loaded nanoparticles was lower than the CPT solution in both cell lines. For CT26 cells, the IC50 values of the drug solution was calculated as 1.86 ± 0.28 µM and 1.27 ± 0.42 µM for 48 h and 72 h, respectively. Among different formulations, the CS-(6-O-capro-β