br Clinically the long infusion time of up
Clinically, the long infusion time of up to 24 h (slow drug inflow rate) for chemotherapy is recommended to avoid hypersensitivity or allergic reactions caused by the immune system. However, such long hospitalization causes burdens in outpatient administration, increases administration costs, and reduces patient compliance . Hence, ef-forts should be made to increase the drug inflow rate as much as pos-sible, but slow enough to minimize unwanted reactions. Therefore, by utilizing fibroin to carry the cytotoxic drugs, a short infusion time is possible as the fibroin carrier could prevent the direct contact between the drugs and the immune system. Furthermore, within 3-h infusion, less than 10% α-mangostin released into the medium, thus, preserving the α-mangostin loaded FNPs eﬃcacy as a whole system.
3.7. Physicochemical stability
The physicochemical stability of lyophilized powders of α-man-gostin loaded FNPs were determined in terms of particle size, zeta po-tential, and the remaining drug, for a period of 6 months at 25 °C and 4 °C. At 25 °C, all formulas showed aggregation with a mean size of > 1 μm in only 1 month. Thus, further investigations were excluded. Nevertheless, at 4 °C, the FNPs size and zeta potential were maintained for at least 6 months. Temperature plays an important role in α-man-gostin loaded FNPs’ physical stability. At 25 °C, the kinetic Geneticin, G-418 Sulfate of fibroin molecules in the particles, even in solid state, was higher than that at 4 °C. Hence, more inter- and intramolecular interactions were formed due to excessive contact, consequently, aggregation occurred. In term of chemical stability, > 99% of the α-mangostin retained in FNPs after 6 month storage at 4 °C, suggesting the suitability of the systems in protecting the entrapped drugs from degradation.
3.8. Hemolysis study
Being a polyphenolic compound, α-mangostin potentially acts as a non-ionic surfactant, consequently destroys the cell membranes by disrupting the lipid-protein interface [15,34]. In fact, due to this action, α-mangostin shows high hemolysis activity to both human and rabbit red blood cells, at dose of less than 20 μg/mL [22,23]. Therefore, we conducted this study to determine the hemolysis eﬀects of α-mangostin loaded FNPs in comparison to the free drug.
All blank FNPs demonstrated no hemolysis action, Fig. 5A, in-dicating biocompatibility systems. Theoretically, an increase in in-organic material crystallinity such as hydroxyapatite results in an
increase in hemolysis activity . Interestingly, all blank FNPs showed no potential hematotoxicity at the investigated doses despite their extensive diﬀerences in crystallinity (EDChigh-FNP > EDClow-FNP > PEI-FNP). Thus, FNPs are safe to the red blood cells, regardless of their polymorphism. Although PEI is a well-known hemolysis agent due to its positive charge property, blank PEI-FNP showed no hema-totoxicity due to a low amount of PEI (< 1%) coated the FNP.
The free α-mangostin possessed a remarkable hemolysis, logarith-mically to the drug concentration with a 50% hemolytic concentration of 11.23 ± 2.29 μg/mL, in agreement with previous studies [22,23]. Interestingly, the FNPs reduced the hemolysis activity of α-mangostin significantly by more than 10 folds. Because all blank particles were safe, the hemolysis action of the drug loaded FNPs should come from the α-mangostin itself, which should correlate well with the dissolution profiles.
From dissolution profile, the release of α-mangostin followed the order of EDClow-FNP > PEI-FNP > EDChigh-FNP. Thus, it was ideal that the hemolysis activity followed the same order. Unexpectedly, the order of hemolysis was EDChigh-FNP > EDClow-FNP > PEI-FNP (Fig. 5A). Therefore, besides dissolution profiles, other factors also played a role in hemolysis. One such could be the particle surface zeta potential. It was likely that compare to the negative charge EDClow-FNP, the positive charge EDChigh-FNP bound more tightly and eﬀec-tively to the negatively charged red blood cell surfaces. Consequently, α-mangostin had more chance to eﬃciently interact with and lyse the
erythrocytes. Nevertheless, PEI-FNP also had positive charge, yet it showed no observable toxicity. This was partially due to the low release rate and the thick PEI diﬀusion layer of the particles, which hinder the drug contact with the erythrocytes. Hence, the hemolysis activity of drug loaded nanoparticulate systems is a complex aspect, which relates to various factors. Further in-depth studies should be conducted to confirm these hypotheses.
3.9. In vitro cytotoxicity test
In vitro cytotoxicity test of free α-mangostin and α-mangostin loaded FNPs was conducted in Caco-2 colorectal adenocarcinoma and MCF-7 breast adenocarcinoma, using MTT assay. The blank particles showed no toxicity to both cell lines at a dose of up to 200 μg/mL (equivalent to α-mangostin) confirming the biocompatibility of the particles. Moreover, the eﬀects on the cells should come from α-man-gostin itself. Surprisingly, although all formulas showed sustained re-lease profiles of α-mangostin, they possessed better eﬃcacy (i.e., lower IC50) than the free α-mangostin in both cell lines (p < 0.01, Fig. 5B). This was attributed to the dissimilarity in the cellular uptake mechan-isms. The free α-mangostin in solution enters the cells by passive transport pathway, which is limited by cell lipid bilayers and the eﬄux pumps. On the other hand, α-mangostin loaded FNPs were eﬀectively up-taken into the cells by surface adsorption and endocytosis pathways . Moreover, it is possible that the two amino acid sequences