Identification and construction of a novel biomimetic delivery system of paclitaxel and its targeting therapy for cancer

Abstract

Dear Editor, Paclitaxel (PTX) is among the first-line therapeutic agents against diverse malignancies. However, problems still exist concerning its poor water-solubility, multidrug resistance, lack of specific targeting ability and severe toxic effects, and these issues are far from fully resolved despite the various PTX formulations available on the market, e.g. the gold standard, PTX albumin nanoparticles Abraxane, and PTX liposomes Lipusu. Some studies that try to solve the multiple problems faced by chemotherapy drug delivery, however, fell into the prevailing trap of overcomplicated formulation design which sacrifices the druggability. To better reconcile this paradox, a novel glycosylated liposomal PTX was designed, inspired by the cytomembrane glycosyls with important roles in maintaining both membrane structure and physiological functions, such as enhancing membrane stability, evading immunological clearance, and recognizing corresponding receptors. Ginsenosides, as natural amphiphilic molecules, are structurally similar to cholesterol, contain glycosyl moieties, and are easily obtainable. Thus, we adopted ginsenoside as a substitute for cholesterol in liposomes, with its hydrophobic region inserted into the lipid bilayer, and its glycosyl exposed on the liposomal surface providing biomimetic functions. This ginsenoside-anchored liposome was named ginposome, and the paclitaxel-loaded ginposome (G-PTX) system exhibited enhanced encapsulation and stability, reduced monocyte phagocytosis, and active targeting towards tumor cells by recognizing their overexpressed glucose transporters. In addition, the ginsenoside could also serve as chemotherapy adjuvant due to its inherent pharmacological functions. First, we screened for the most desirable ginsenoside as the cholesterol substitute. Ginsenoside Rg5, with a disaccharide group at the 3-position of the skeleton, showed effective stabilization of the liposome membrane (Fig. 1a, b), and this was consistent with the molecular dynamic calculation results which confirmed that the glycosyl moiety at the 3-position oriented towards the water molecules without disruption of phospholipid arrangement (Supplementary Fig. S1a, b). Meanwhile, the binding affinity between Rg5 and GLUT1 experienced no interference at high concentration of free glucose (Fig. 1c and Supplementary Fig. S2a, b), which was further verified by the cellular internalization assay under high-glucose condition (Supplementary Fig. S2c–e). Therefore, ginsenoside Rg5 was adopted as a cholesterol substitute for the construction of novel PTX liposomes. The formulation of G-PTX was then optimized according to the stability and in vitro targeting ability. Molecular dynamics simulations showed that Rg5 formed intensive hydrogen bonds with POPC and PTX, which further constituted hydrogen bond networks to stabilize the entire system (Fig. 1d and Supplementary Fig. S3a). With the increase of the Rg5/PTX molar ratio, the stability of G-PTX raised and reached a plateau at 3:1 Rg5/PTX ratio, surpassing the stability of conventional paclitaxel liposome (L-PTX) containing an identical proportion of cholesterol (Supplementary Fig. S3b, c). Highglucose level reduced the cellular uptake of G-PTX at 3:1 Rg5/ PTX ratio, possibly related to the decreased solvent-accessible surface area (SASA) of the disaccharide in Rg5 along with the formation of hydrogen bond networks (Fig. 1e, f). When the Rg5/ PTX ratio reached 4:1, its in vitro targeting function peaked, and further increases of the ratio engendered no more improvement (Supplementary Fig. S3d). To facilitate the quantitative preparation of G-PTX, a 4.5:1 molar ratio (equivalent to 4:1 mass ratio) was selected. The final G-PTX had a spherical shape with an average particle size of ~110 nm, encapsulation efficiency of 97.2% and maintained high stability even when the PTX concentration reached over 5 mg/mL (Fig. 1g and Supplementary Fig. S3e–j). Molecular dynamics simulation also confirmed the marked improvement in integral stability of G-PTX (Fig. 1h and Supplementary Fig. S4a–d). We further investigated its expected long circulation and activetargeting function in vivo. The elevated surface hydrophilicity of GPTX provided by glycosyls partly reduced the formation of protein corona thus decreased the macrophage uptake (Supplementary Fig. S5a, b). Pharmacokinetic results also verified the significantly increased blood PTX concentration and prolonged circulation of GPTX compared with L-PTX and Abr (Abraxane) at the same dose (Supplementary Fig. S5c). Enhanced accumulation of G-PTX at the HGC-27 tumor sites (both subcutaneous xenograft and orthotopic tumors) was also imaged compared to L-PTX (Supplementary Fig. S5d, e). Similar trends were observed in three other subcutaneous tumor models (A549, MCF-7, and HGC-27/T), demonstrating a broadspectrum tumor-targeting ability of G-PTX (Supplementary Fig. S5f–h). The tissue distribution experiment revealed significantly elevated drug content in HGC-27 subcutaneous tumor in G-PTX group compared with L-PTX, Abr and Taxol (Fig. 1i). On the other hand, the drug content in the liver, spleen and muscle tissues of the G-PTX was significantly lower than L-PTX, which demonstrated its high efficacy and low toxicity (Fig. 1i and Supplementary Fig. S6a). Meanwhile, the active tumor targeting of G-PTX was mainly achieved through GLUT1-mediated mechanism (verified through competition assay of GLUT1 specific inhibitor STF-31, and siRNA silencing), and GPTX could be further endocytosed via both clathrinand caveolaedependent pathways (Fig. 1j and Supplementary Fig. S6b–f). In vitro anti-tumor experiments showed that the IC50 values of GPTX were more than one order magnitude higher than those of L-PTX and Abr in various tumor cell lines, especially in drug-resistant (PTX) cell lines (Supplementary Fig. S7a). As for in vivo anti-tumor assay with the initial average tumor volume of 200mm, G-PTX (30mg PTX/kg) showed complete suppression of tumor growth on HGC-27/T (drugresistant) tumor-bearing mice model, reducing the tumor volume to 193mm, displaying sharp contrast with Abr group (893mm) at the same dose (Fig. 1k). In HGC-27 tumor-bearing mice, two-thirds of the tumors in the G-PTX group completely disappeared at the end of experiment, while the average tumor volume in the Abr group remained 365mm, at the same dose (30mg PTX/kg) (Fig. 1k). On several other models, G-PTX also performed significantly better than