Development of biodegradable starch nanocrystals/gum Arabic hydrogels for controlled drug delivery and cancer therapy

Maize starch (C₆H₁₀O₅)n containing amylopectin 72%, amylose 28%, was purchased from Alpha Chemika, India; sulfuric acid (H₂SO₄) (98%) (Hi Media Co., India); gum Arabic from Elnasr, Sudan; potassium persulfate (K₂S₂O₈) was obtained from Maknur Co., Canada; polyvinylalcohol (PVA) (M_w ∼ 18000–12000) from Panreac Co., Spain; hydroxyurea purchased from Apotex Inc. Canada; dimethylsulfoxide DMSO (Central drug house Co., India), isopropanol purchased from GCC, UK and buffer saline tablets were purchased from Flow Laboratories, UK. All other reagents of an analytical grade for in vitro studies were used.

Acid hydrolysis process was prepared according to the procedure reported by Angellier et al [9], and Brenda et al [12] with slight modification. A 150 g weight (15 wt%) of starch granules were blended in 1000 ml of 3.16 M sulfuric acid solution at room temperature under constant stirring at 200 rpm for 3 weeks. They were shaken with an orbital action at a speed for 1 min four times a day to ensure a more homogeneous hydrolysis. After repeated centrifugation, the SNCs were washed in deionized water until the pH reached approximately 7. Droplets of chloroform were added to the suspensions and stored at 4 °C until use.

The product was furnace dried (DZ-2BC vacuum dry box, Tianjin, China) at 40 °C to get the SNC powder. New polymer materials were prepared with different ratios, which are 4:0, 3:1, 2:2, 1:3, 0:4 of SNC:GA [13], by using K₂S₂O₈ as an initiator. Briefly, a known weight of starch was dissolved in 100 ml of deionized water followed by the addition of different quantities of GA under continuous mechanical stirring (Labinco L-81, Netherlands) for 15 min. The cross-linking was then initiated by adding K₂S₂O₈ accelerator 12.5% (depending on the dry weight of polymer) to the mixture at 65 °C. After 3 h the reaction was stopped as it was reported [14]. The mixture of a polymer was then isolated by centrifugation (Sigma 2-16PK, Germany) at 4000 rpm for 20 min, then washed with deionized water and dried in a furnace at 55 °C [3].

The prepared SNC/GA polymer (4 g) was dissolved in 40 ml deionized water. HU was used as a drug model. The concentrations of HU were 25% (w/w) with respect to the polymer amount. A selective weight of HU (1 g) was blended with the minimum volume of deionized water. The mixture of HU/Polymer was suitably blended in deionized water by a magnetic stirrer, and an aqueous solution containing polyvinyl alcohol as stabilizer and emulsifier was added dropwise. Polyvinyl alcohol concentration was 3% (w/v) aqueous solution as reported earlier [15]. The prepared emulsion was scattered by applying sonication and homogenization via ultrasonic dispersion machine (sonifier150, Branson, USA) for ten min at room temperature. The last dried blends were stored until use. Table1 represented the composition of HU/SNC/GA blends.

2.5.1. Fourier transforms infrared spectroscopy (FTIR)

2.5.2. X-ray diffraction analysis (XRD)

Wide angle XRD analysis (XRD-6000, Shimadzu, Japan) was performed on powders obtained from both ST and SNC. The XRD showed the crystalline nature of the SNC and their corresponding unmodified starch. The samples were sited in a 2.5 mm deep cell. The operating divergence slit, tube current, and generator voltage were 0.3 mm, 30 mA, and 40 kv, respectively. The performing conditions for the diffractometer were: Cu Kalfa radiation (λ = 1.54060 A), the diffraction angle (2θ) range from 4°–80° at a step length of 0.02° and counting time 0.15 s.

The degree of crystanallity was calculated using equation (1)

$$\begin{eqnarray}&&{\rm{X}} \% =\displaystyle \frac{{\rm{Ac}}}{{\rm{At}}}\,* \,100=\displaystyle \frac{{\rm{Ac}}}{{\rm{Aa}}+{\rm{Ac}}}\,* \,100\end{eqnarray} \tag{ 1 }$$

where: Ac is the crystalline area, Aa is the amorphous area and At is the total area [16].

2.5.3. Scanning electron microscopy (SEM)

2.5.4. Atomic force microscopy (AFM)

The SR of the samples were performed by immersing dried samples (polymer/HU) 10 mg in 1.5 ml of buffer solution at different pH (4 and 7.4) at 37 °C. These solutions correspond to the pH environments of the stomach and colon, respectively. At specific time intervals, buffer solution medium was removed by centrifuge (Centrifuges3k30, Sigma, USA) at 4000 rev/min for ten min. The sample was then weighed on a CPA2P analytical balance (Sartorius, Germany). The SR of samples was calculated using equation (2):

$$\begin{eqnarray}&&SR=\displaystyle \frac{\left({W}_{wet}\,-\,{W}_{dry}\right)}{{W}_{dry}}\times 100 \% \end{eqnarray} \tag{ 2 }$$

where W _wet:- is the weights of swollen powder, W _dry:- is the dried sample as described before [18].

Prior to the release tests, the mixture of HU/polymer was trimmed in order to obtain disc shape (D = 12 mm, t = 1 mm) [19]. In vitro drug release characteristics from polymer/drug blends were carried out in a buffer solution at pH4, pH7.4 at a temperature of 37 ^◦C. The concentration of HU release was measured at a variety of duration interval by UV–vis spectrophotometer (UV-1650 pc, Shimadzu, Japan) after drawing suitable calibration curve. A known weight of polymer/HU mixture was kept in individual vials containing buffer solution of pH4, pH7.4 (1 mg ml⁻¹). The mixture was heated at 37 ^˚C. At a certain time interval, 1 ml of aliquots was taken out to measure the amount of HU released [15]. The concentration of HU release was determined with UV–vis spectroscopy by interaction with urea kit at definite intervals of time.

The MTT (3-[4,5-Dimethylthiazol-2-yl]-2,5-diphenyl tetrazolium bromide) assay is a colorimetric assay used to evaluate the cytotoxic potential of new drug on cancer cells. LS-174 is a human colon cancer cell line gained from the American Type Culture Collection that is commonly used to study cancer biology. Cells were seeded in 96wеll p1аtes (Lab-Tek and Nunc, USA) at a concentration of 1.0 × 10⁵ cells/ml and incubated (MC-20A CO₂ incubator, SNT, Korea) at 37˚C for 24 to 48 h. Afterward, culture media (minimum essential media, Roswell Park Memorial Institute medium (RPMI)) with L-glutamine) was replaced and different concentrations (1, 10, 100, 500 and1000 μg ml⁻¹) of free polymer, free HU, control and polymer/HU were added to cultured wells at a volume of 10² μl in per wells except for control cells. All samples were cultured in triplicate. The plates were incubated for 24 h at 37 °C in a humidified atmosphere containing 5% CO₂. To avoid any contamination, the microtiter-plates were moved to level 2 biohazard safety cabinets with sterilized environment, and the media in all wells were discarded. To avoid any residual amounts of composites or a standard anticancer drug used that may interact with MTT reagents, the LS-174 cell monolayers were washed three times with phosphate buffer saline solution. Then the cell culture was treated with 20 μl/well MTT (Bioworld Co., USA), plates were incubated at 5% CO₂, 37 °C for 2 h, the formazan particles were formed as a mitochondrial enzymatic process of the unaffected viable LS-174 cells, the dead or viral affected cells were not forming formazan particles because its mitochondria organelles were disrupted. Formazan crystals were dissolved in diluted DMSO (1:1) in isopropanol in each well including blank wells and absorbance was taken at optical density of 490 nm using microplate reader (ELx800 absorbance microplate readers, Biotek, USA) and the cell viability percentage was measured using equation (3) [20, 21] :

$$\begin{eqnarray}&&Cell\,availability \% =\displaystyle \frac{\left({A}_{t}-{A}_{b}\right)}{\left({A}_{c}-{A}_{b}\right)}\times 100\end{eqnarray} \tag{ 3 }$$

where A_t, A_b and A_c were the absorbance value of test compound, blank and control, respectively.

Figure 1(a) shows FTIR result using pure SNC, pure GA, and SNC/GA blend. For SNC, a typical spectrum of enclosed absorption bands at 3299.06 cm⁻¹ revealed the O–H groups stretching vibration in the glucose units [4], and the peaks at 2924.20 cm⁻¹ and 1340.13 cm⁻¹ ascribed to the C–H stretching and –OH bending, respectively [4, 22]. The peak at 1638.98 cm⁻¹ was presumably attributed to C–O vibrations present in the starch [23]. The characteristic peak occurred at 1146.95 cm⁻¹ was ascribed to O–C stretching of the C–O–C linkages in the glucosidic rings, which made up the spectra of starch backbone [23]. The peak at 854.80 cm⁻¹ was assigned to a methene vibration [24]. The FTIR spectra of pure GA are dominated by OH stretching vibrations at 3272.57 cm⁻¹ [25]. The C–H stretching modes are riding over the broad peak at 2913.77 cm⁻¹ [26]. The carbonyl symmetric stretching modes (C=O) are observed at 1602.08 cm⁻¹ and 1406.48 cm⁻¹. The characteristic from 900–1200 cm⁻¹ was the fingerprint of carbohydrates [27]. O–H bending peaks were detected at 1015.60 cm⁻¹ [28].

Zoom In Zoom Out Reset image size

For SNC/GA, a small shift in the wave number for hydroxyl group stretching vibration peaks from 3299.06 to 3285.41 cm⁻¹ was found due to the increase in the hydrogen bonding between the SNC and GA [29]. Also, the intensity of OH stretching band in the SNC/GA blend was decreased after blending, possibly due to the occurrence of reaction between SNC and GA [3].

Figure 1(b) demonstrated FTIR results for pure HU, and these results indicated that the two strong peaks of N–H stretching was observed at 3412.71 and 3305.19 cm⁻¹ [30]. The peak at 1633.49 cm⁻¹ was corresponding to C=O [30], and the peak of N–H bending appeared at 1584.35 cm⁻¹ while the peak of C–N showed at 1478.07 and 1099.17 cm⁻¹ [31]. The FT-IR of HU-loaded SNC/GA blend was presented in figure 1(b). The position of the peaks band of HU in HU-loaded SNC/GA blends was changed, where the peaks of –OH group are present at 3306.23 cm⁻¹, the N–H stretching appears at 3741.85 cm⁻¹ while the peak at 2917.26 cm⁻¹ is identical to the C–H stretching [32]. Moreover, the peak at 1641.58 cm⁻¹ was referred to C=O group [30], and the peaks at 1585.93 cm⁻¹ and 1411.09 cm⁻¹ are corresponding to C–O vibrations and –OH bending, respectively and the unsymmetrical vibrations of C–O–C bond appeared at 1145.88 cm⁻¹ [31].

SEM photograph of SNC is shown in figure 2(a). The degraded and destructured granules have been observed for SNC [16] and its surface was deformed by acid treatment. The adhesion between some of the granules strongly affected the granule morphology [33], which showed densely agglomerated broken crystals [12]. The hydrolysis acid causes an increase of the surface roughness [33].

Zoom In Zoom Out Reset image size

The sizes of SNC were determined using SEM and AFM. Starch aggregation has been showed for nanocrystal as shown in figure 2(a). This aggregation was due to the formation of a hydrogen bond resulting from the interaction of hydroxyl groups on the surface of SNC [34]. After three weeks, acid hydrolysis starch produced nanocrystals. The nanocrystal starch diameter was ranged between 42.6–244.264 nm as shown in figure 2(a). Figure 2(f) was used to present the surface topography as a 3D image and illustrated the thickness of SNC, which is 14.3 nm.

The homogeneity and microscopic surface structure of the SNC/GA blend were examined using SEM as shown in figures 2 (b)–(d). When GA was added to SNC, the irregular particles with obvious aggregation were observed [35]. GA SNC blends displayed solid shape with a larger average size than neat starch size. The introduction of SNC/GA blends contributed substantially to a reduction in roughness. Exclusion of cracks and porosity was observed on the surface of GA/SNC blends, which lead to reduce water absorption. Moreover, the reduction in roughness was increased with increasing GA proportion [36]. It was observed that the interfacial adhesion and interaction among SNC and GA facilitated the enhancement in the maintenance of integrity and strength of particles [37]. HU-loaded nanocrystal starch/GA blends display a spherical form and have a smooth surface with more regularities structure as illustrated in figure 2(e). The sphere-shaped particles are appropriate for drug delivery utilization and high possibility for cell entry [38].

In figure 2(g), the XRD pattern for native ST and SNC. Native ST the A-type crystallinity pattern is being displayed. The crystallinity pattern does not change after treatment. However, SNC displayed an increase in the reflection peak intensity compared to their native ST. It was expected due to selective hydrolysis of the amorphous regions of starch via acid that lead to a more distinct crystalline region. Similar observations were previously described by Mukurumbira for production of SNC from amadumbe ST [34]. Increasing the intensity of the peaks at 2θ = 19.9° was the most noticeable. This peak is related to individual helices with a crystalline arrangement; these results are in agreement with the previous report by de la Concha [12]. Increasing the intensity of the peaks at 2θ of 15°, 17°, 18°, and 23° was also observed in SNC which was produced by acid hydrolysis, with an unresolved large doublet between 17°−20°. The proportion of crystallinity in ST increases with acid hydrolysis. This meant that the amorphous parts degraded by acid preferentially as reported by Chen et al [39].

The results of the swelling ratio increased with time. The swelling increased rapidly during the first hours of reaction followed by slow increase later, which appeared almost flattened as shown in figure 3 because the porous network of SNC became saturated with water molecules with no more room for further accommodation, which is similar to the results reported by Kaith et al [40]. It is clear that the swelling of C0 blend improved with raising of the pH value of the medium as shown in figure 3. This may be attributed to electrostatic interaction and hydrogen bonding as reported by Nizam El-Din et al [41]. The increase in pH value leads to increase in swelling of anionic SNC because of 2 significant phenomena that happen concurrently. First, the increase in the charge density due to ionized acidic groups lead to increase in cationic concentration differences between the inner and outer surface of SNC that resulted to increase osmotic pressure between SNC and external solution as previously reported [42]. These changes ultimately caused easier penetration of external solution in the SNC network resulted in higher swelling. A second possible phenomenon is electrostatic repulsion amongst anions on SNC surface which also increased the expansion rate of SNC chains leading to higher swelling as reported by Solanki et al [42]. At pH4, there was very little chance of electrostatic repulsion amongst anions on SNC surface which suggest the expansion of SNC chains lead to slower swelling than swelling at pH7.4 as previously reported [42].

Zoom In Zoom Out Reset image size

The swelling ratio of C1, C2 and C3 blends were determined and are shown in figure 3 at pH4 and pH7.4, respectively. The GA with SNC blend may lead to a decrease the water sensitivity of the blends. This phenomenon can be attributed to the formation of a rigid SNC network due to strong hydrogen bonding between SNC, strong interactions between SNC and GA chains as reported earlier [5] and increase crystallinity with SNC as previously reported [43], resulting in decreasing the swelling and water absorbance of the blends as reported by Ghasemlou et al [5].

The swelling behavior of C1, C2, and C3 blends showed a predominant pH-dependence. It was swelled markedly at pH4 but swells rapidly with time at pH7.4 until collapse, which is similar to the results reported by Qi et al [44]. The hydrophilic natures of SNC and GA (mainly because of D-glucopyranosuronic acid moieties) along with the ionic groups (OH–) were the main causes for the absorbency. At pH7.4, the electrostatic repulsion between OH groups caused an improvement of the swelling ratio as reported by Gils et al [45]. The lower swelling ratio can be observed at C2 blend. On the other hand, the swelling behavior was restricted at pH4 because of strong hydrogen-bond interactions and the maintenance of the network structure, which resulted from cross-linking between SNC and GA blends, which is similar to the results reported by Lin et al [37]. It was also observed that the lowest swelling ratio occurred at C3 blends, and the reason may be due to the nature of starch and gum Arabic in different pH.

The release in figure 4 increased with the increase in time. Short release time of pure HU is due to a simple diffusion process. When HU is loaded with polymers (C0, C1, C2, C3), the small burst release of the HU from HU/polymers was recognized due to the HU molecules entrapped at or near the surface of SNC as reported earlier [46].

Zoom In Zoom Out Reset image size

The C0 blend has a slow release of HU as shown in figure 4 [46]. It may be due to the high intensity of OH group for SNC as shown in FTIR results which the interaction of OH with –NH₂ group of HU formed rigid nanocrystal structure, which leads to hinder the escape route of drug molecules to reduce swelling and exhibited noticeable sustained release profiles as previously reported [47]. The HU release of C0 blend was studied in two different buffered solutions (pH4 and pH7.4) as shown in figure 4. The HU release from loaded SNC was sensitive to the pH. The release of drug from C0 blend was maximum at pH7.4 while the minimum release was observed at pH4. At pH4, the response with respect to pH was due to the reduced swelling ability of SNC as shown in swelling test, leading to reduction in the release of HU as illustrated in figure 4(a) and reported by Wang et al [18]. On the other hand, when the pH increases to 7.4, a great number of anionic groups on the starch are present, as previously reported [42, 48]. This makes the starch chains repel one-another leading to expanding dimension or high swelling of the C0 blend as reported by Solanki et al [42]. Thus the release of HU increases as shown in figure 4 (b).

For C1, C2, C3 blends shown in figure 4, the reduction in release behavior of HU increased in comparison with C0. The good dispersion of SNC in GA leads to increase physical cross-linking through hydrogen bonding between them as reported by Lin et al [49]. This indicated that the addition of GA could significantly enhance the capacity of SNC for controlling HU release. Similar results were shown by Li et al [35]. At pH4, the reduction in HU release of C1, C2, C3 blends was increased compared to C0, as shown in figure 4(a). This may be due to the reduction in swelling for these blends as shown in swelling results. Moreover, the HU release from C0 is greater than A00, thus it was concluded that the reduction in HU release was increased with the increasing GA proportion observed at C3. At pH7.4, the amount of HU release from C1, C2, C3 blends tended to swell and disturbed the structure of particles, which lead to a higher porosity in the particle structure, and enhanced the drug release rate as previously reported [50] in comparison with pH4 for the same blends, as shown in figure 4(b). In addition, the HU release from A00 is greater than C0. It was determined that the slowest HU release from C1, C2, C3 was observed in C2.

The cell viability assays were performed to evaluate the anticancer activity of pure SNC, pure GA, Pure HU, HU/SNC, HU/GA, and HU/SNC/GA at different ratio. Pure SNC, GA and their mixture showed little or no significant effect on the cell viability percentage as shown in figure 5, possibly due to it is biocompatibility as reported by Pandey et al [15, 51], while the treatment with pure HU and HU-loaded biocompatible materials significantly decreased the viability of LS-174 cells with the increase in HU concentration. On the other hand, the formulated A00, C0, C1, C2, C3 blends showed a significant decrease in viability of LS-174 cells.

Zoom In Zoom Out Reset image size

Figure 5 demonstrated the reduction of cell viability of the C1, C2, C3 blends as compared with pure HU. The results showed that for C1, C2, C3 blends, C3 showed higher cell viability. On the other hand, it was observed that the increase in cell viability was as the following order: C3 > C2 > C1 > C0 > A00 > HU. This may be attributed to the low concentrations of HU release from these blends as shown in drug release results. The toxicity of pure HU, which is higher than C1, C2, C3 blends, can be described through the fact that pure HU molecules are imported to cells by passive diffusion mechanism, while C1, C2, C3 blends should first be captured by cells through a possible endocytosis mechanism and thus HU molecules are released gradually into cells Zohreh et al [52].

At 1 μg ml⁻¹ concentration, C3 blends decrease cell viability up to 90% after 24 h, whereas at high concentration (1000 μg ml⁻¹), cell viability decreases up to 37% after 24 h. On the other hand, HU at 1 μg ml⁻¹ decreases cell viability up to 78% after 24 h, and 25% cell viability were observed at 1000 μg ml⁻¹ concentration in a similar time frame, as shown in figure 5. Hence, cytotoxicity of C1, C2, and C3 is found to be less as compared to pure HU due to sustained and slow release of drug as reported by Pandey et al [53]. It was determined that C3 blend are better than pure HU and other blends for controlled drug delivery as a novel vehicle for cancer therapy.

Figure 6 displays the image of cells after treatment with pure starch nanocrystal, pure HU and C3 blend. The colored formazan indicates good cells health which appears in pure starch nanocrystal (biocompatible materials). On the other hand, most dead cells were observed in pure HU. Moreover, C₃ shows the sluggish rate of cells death. The sluggish rate of cells death was indicated to the sustained release of HU from C3 blends produced at a concentration of 100 μg ml⁻¹.

Zoom In Zoom Out Reset image size