In this paper, as an active ingredient, puerarin chitosan nanoparticles (Pur-CS/TPP-NPs) are prepared by an ionic gelation method. The chitosan (CS) concentration, pH of the CS solution, sodium tripolyphosphate (TPP) concentration, stirring speed, stirring time, ultrasonic power, and dosage are used as single factors for investigation, and the encapsulation efficiency, drug loading capacity, particle size, and polydispersity index (PDI) are used as indicators for investigation. The optimal prescription is determined using the Box–Behnken effect surface design method. The characterization of the best formulation, which is determined via an in vitro release assay and liquid chromatography/tandem mass spectrometry (LC-MS/MS) analysis methods, is used here for pharmacokinetic studies. An in situ single-pass intestinal perfusion model is used to investigate drug absorption in the intestine. After characterization, the morphologies of the nanoparticles are intact. It can be seen from the in vitro release experiments that the equation fitted by the nanoparticles is the Higuchi model, the nanoparticle release process is very stable and without sudden release, indicating that the nanoparticles are well-released in vitro. The pharmacokinetic results and the in situ single-pass intestinal perfusion model study show that the bioavailability and absorption of Pur-CS/TPP-NPs were significantly higher than Pur. Thus, all the results show that the prepared nanoparticles can significantly improve the bioavailability of Pur, and we hope to lay the foundation for the development of new products of Pur.

The protection of unstable anticancer molecules and their delivery to lesions are challenging issues in cancer treatment. Topotecan (TPT), a classic cytotoxic drug, is widely used for treating refractory lung cancer. However, the therapeutic effects of TPT are jeopardized by its active lactone form that is intrinsically hydrolyzed in physiological fluids, resulting in low bioavailability. Herein, the TPT-loaded crosslinked cyclodextrin metal-organic framework (TPT@CL-MOF) was engineered to improve the local bioavailability of TPT for the treatment of lung cancer. CL-MOF exhibited the efficient loading (12.3 wt%) of TPT with sustained release characteristics. In particular the formulation offered excellent protection in vitro against hydrolysis and increased the half-life of TPT from approximately 0.93 h to 22.05 h, which can be attributed to the host–guest interaction between cyclodextrin and TPT, as confirmed by molecular docking. The TPT@CL-MOF could effectively kill the cancer cells and inhibit the migration and invasion of B16F10 cells in vitro. Moreover, TPT@CL-MOF was efficiently distributed in the lungs after intravenous administration. In an in vivo study using a B16F10 pulmonary metastatic tumor model, TPT@CL-MOF significantly reduced the number and size of metastatic lung nodules at a reduced low dose by five times, and no noticeable side effects were observed. Therefore, this study provides a possible alternative therapy for the treatment of lung cancer with the camptothecin family drugs or other unstable therapeutically significant molecules.

Background: As a novel drug delivery vehicle, transfersomes (TFSs) have great deformability, good biocompatibility, and biodegradability, which can significantly improve drug delivery efficiency. However, their application is limited by their poor stability and vulnerability to leaks. Methods: This paper prepared and assessed the stability of transfersomes co-loaded with Pueraria total flavonoids (PTFs) and tanshinone (TAN), known as PTF/TAN-TFSs, in order to overcome this constraint. Results: The findings showed that the co-loaded system (PTF/TAN-TFSs) had better structural stability than the single-loaded systems (PTF-TFSs and TAN-TFSs), with drug leakage rates decreased by roughly 5% and 7%, respectively, and a more uniform size distribution with a particle size of 44.52 ± 0.47 nm and a PDI of 0.22, while the single-loaded systems had PDI values larger than 0.3. The co-loaded system showed a better sustained-release profile than the suspension with a 15.28% decrease in the maximum cumulative release of the poorly soluble component Tan IIA. Additionally, the co-loaded system demonstrated enhanced solubilization capability, making the poorly soluble component TAN more soluble by 89.52 times compared to the free medication and by 1.97 times compared to the single-loaded system. Furthermore, the co-loaded system significantly improved the intestinal absorption of Tan IIA, increasing the Papp and Ka values in the ileum segment by 2.02-fold and 1.80-fold, respectively. Conclusions: The stability of the system was successfully improved by co-loading many components with varying hydrophobicities. This method provides a fresh way to enhance transferosome stability, broadening their application in the field of drug delivery.

Goji berries (Lycium barbarum) are widely consumed as functional foods and medicinal–food products, and their quality and authenticity are closely associated with geographical origin. Rapid and reliable origin identification is therefore essential for quality control, authenticity assessment, and traceability. In this study, a hyperspectral imaging (HSI)-based deep learning framework was developed for the non-destructive geographical origin identification of goji berries by integrating spectral and spatial information. Spectral features were extracted using a one-dimensional convolutional neural network equipped with a spectral–channel attention module (SCAM), which adaptively enhances informative wavelengths and discriminative spectral representations. Spatial features were learned via HSI, using DenseNet to characterize complementary morphological and textural information. A cross-attention fusion module was further introduced to establish bidirectional interactions between spectral and spatial features, enabling adaptive integration of complementary information and reducing feature redundancy. The proposed framework achieved a classification accuracy of 98.24%, outperforming conventional machine learning methods and representative deep learning models. These results demonstrate that cross-attention-based spectral–spatial fusion can effectively exploit the complementary characteristics of HSI data, providing a rapid, non-destructive, and reliable strategy for goji berry authentication and geographical traceability, with potential applications in the quality control of functional foods and other agricultural products.