ObjectiveTo construct a large language model （LLM）-based intelligent pre-consultation system for traditional Chinese medicine （TCM） to improve efficacy of clinical practice. MethodsA TCM large language model was fine-tuned using DeepSpeed ZeRO-3 distributed training strategy based on YAYI 2-30B. A weighted undirected graph network was designed and an agent-based syndrome differentiation model was established based on relationship data extracted from TCM literature and clinical records. An agent collaboration framework was developed to integrate the TCM LLM with the syndrome differentiation model. Model performance was comprehensively evaluated by Loss function， BLEU-4， and ROUGE-L metrics， through which training convergence， text generation quality， and language understanding capability were assessed. Professional knowledge test sets were developed to evaluate system proficiency in TCM physician licensure content， TCM pharmacist licensure content， TCM symptom terminology recognition， and meridian identification. Clinical tests were conducted to compare the system with attending physicians in terms of diagnostic accuracy， consultation rounds， and consultation duration. ResultsAfter 100 000 iterations， the training loss value was gradually stabilized at about 0.7±0.08， indicating that the TCM-LLM has been trained and has good generalization ability. The TCM-LLM scored 0.38 in BLEU-4 and 0.62 in ROUGE-L， suggesting that its natural language processing ability meets the standard. We obtained 2715 symptom terms， 505 relationships between diseases and syndromes， 1011 relationships between diseases and main symptoms， and 1 303 600 relationships among different symptoms， and constructed the Agent of syndrome differentiation model. The accuracy rates in the simulated tests for TCM practitioners， licensed pharmacists of Chinese materia medica， recognition of TCM symptom terminology， and meridian recognition were 94.09%， 78.00%， 87.50%， and 68.80%， respectively. In clinical tests， the syndrome differentiation accuracy of the system reached 88.33%， with fewer consultation rounds and shorter consultation time compared to the attending physicians （P＜0.01）， suggesting that the system has a certain pre- consultation ability. ConclusionThe LLM-based intelligent TCM pre-diagnosis system could simulate diagnostic thinking of TCM physicians to a certain extent. After understanding the patients' natural language， it collects all the patient's symptom through guided questioning， thereby enhancing the diagnostic and treatment efficiency of physicians as well as the consultation experience of the patients.

ObjectiveTo observe the apoptosis induced by paeoniflorin （PF） in non-small cell lung cancer （NSCLC） cells and explore its mechanism. MethodCell counting kit-8 （CCK-8） was used to detect the inhibition rates of H1299， H292 and A549 cells with different concentrations of PF （2.5， 5， 10， 20， 25 µmol·L^-1）， and to screen suitable concentrations of PF and experimental cells. The inhibitory effect of PF on lung cancer cells was detected by clone formation assay. The effect of PF on cell apoptosis was detected by flow cytometry with annexin V-FITC/propidium iodide （PI） double staining. With the right concentration of drugs， levels of apoptosis-associated protein B-cell lymphoma-2 （Bcl-2）， Bcl-2-associated X protein （Bax）， cleaved Caspase-3 and Caspase-3 were detected by Western blot. At the same time， the molecular expressions of hypoxia inducible factor -1α （HIF-1α） and Hippo signaling pathway were determined. ResultCompared with the blank group， PF significantly inhibited the growth of H1299， H292 and A549 cells of human lung cancer （P<0.01）. PF significantly induced apoptosis in A549 cells （P<0.01）， decreased the Bcl-2/Bax ratio （P<0.01）， and significantly increased the cleaved Caspase-3 expression （P<0.01）. Compared with those in the blank group， the expression levels of HIF-1α， transcriptional coactivator with PDZ-binding motif （TAZ）， large tumor suppressor 1 （LATS1）， Mps one binding 1 （MOB1） and Yes-associated protein （YAP） in A549 cells of the PF treatment group were significantly decreased （P<0.01）， while the expressions of p-LATS1， p-MOB1 and p-YAP were significantly increased （P<0.01）. At the same time， there was no significant effect on the expression levels of phosphorylated mammalian Ste20-like kinase 1 （p-MST1） and MST1， which did not reach a statistical difference. ConclusionAll data demonstrated that PF showed an anti-tumor effect by improving hypoxic conditions and inhibiting the abnormally activated Hippo signaling pathway， thereby inducing and promoting apoptosis in non-small cell lung cancer.

BACKGROUND:C2 ceramide reduces the formation of Alpha-Synuclein(α-Syn)oligomers as the protein phosphatase 2A agonist,which has an important regulatory effect on cell aging in the central nervous system. OBJECTIVE:To investigate the protective mechanism of C2 ceramide on dopaminergic neurons. METHODS:Twenty-five C57BL/6 mice were randomly divided into control group,model group,C2 ceramide low-,medium-and high-dose groups(n=5 per group).Except for the control group,a mouse model of Parkinson's disease was established by injecting mutant A53T α-Syn oligomers into the left striatum in the other groups.On the 30th day after the striatal injection,three C2 ceramide groups were intragastrically administered with C2 ceramide(1,5,10 μg/g)dissolved in saline at one time,while the control and model groups were administered with the same amount of saline within 30-90 days after modeling,for a total of 60 days.Behavioral changes in each group of mice were observed during this period.On the 90th day after striatal injection,mouse brain tissue was extracted by perfusion under anesthesia,and the changes of dopaminergic neurons in the midbrain substantia nigra were analyzed by immunohistochemical staining.The levels of α-Syn oligomerization and phosphorylation in the midbrain of mice were detected by ELISA,and the changes of enzyme activities related to α-Syn phosphorylation were analyzed. RESULTS AND CONCLUSION:C2 ceramide had an ameliorating effect on Parkinson's disease-like dyskinesia in mice caused by the striatal injection of mutant A53T α-Syn oligomers.High-dose C2 ceramide showed better effects on dyskinesia in mice with Parkinson's disease(P<0.01).The mutant A53T α-Syn oligomers significantly reduced the number of dopaminergic neurons in the substantia nigra of mice(P<0.01),while the number of dopaminergic neurons in the substantia nigra increased significantly in the C2 ceramide high-dose group(P<0.01).The levels of α-Syn oligomers and phosphorylated α-Syn in the brain were significantly reduced in the C2 ceramide high-dose group compared with the model group(P<0.01),while the level of ceramide was increased(P<0.05)and the activity of protein phosphatase 2A was significantly upregulated(P<0.01).To conclude,C2 ceramide can attenuate the neurotoxic effects induced by oligomerized α-Syn by the phosphorylation modification environment of α-Syn in mouse midbrain tissue and protect against the reduction in the number of nigrostriatal dopaminergic neurons in mice,thereby reducing the degree of dyskinesia in Parkinson's disease.

This research utilizes a deep learning-based image generation algorithm to generate pseudo-sagittal STIR sequences from sagittal T1WI and T2WI MR images.The evaluations include both subjective assessments by two physicians and objective analyses,measuring image quality through SNR and CNR in ROIs of five different tissues.Further analyses,including MAE,PSNR,SSIM,and COR,establish a strong correlation between the generated STIR sequences and the gold standard,with Bland-Altman analysis indicating pixel consistency.The findings indicate that the deep learning-generated STIR sequences not only align with but potentially surpass the gold standard in terms of image quality and clinical diagnostic capabilities.Moreover,the approach demonstrates promise for clinical implementation,offering reduced scan time and enhanced imaging efficiency.

In this study, 34 deep eutectic solvents (DESs) were successfully prepared for the extraction of proanthocyanidin from Rhodiolae Crenulatae Radix et Rhizomes. The extraction process was optimized using single factor exploration and Box-Behnken design-response surface analysis. The extraction rate was significantly improved when the molar ratio of choline chloride to 1,3-propanediol was 1:3.5 and the water content was 30% (V/V) in DESs. AB-8 macroporous resin and ethyl acetate were used for separation and refining, and the oligomer-rich proanthocyanidin components were eventually obtained. The ultraviolet (UV) and infrared (IR) spectra showed that the proanthocyanidins were mainly composed of catechin and epicatechin. To further clarify the chemical composition of proanthocyanidin, an ion scan list containing 156 proanthocyanidins precursors was obtained by constructing a proanthocyanidins structural library and mass defect filtering (MDF) algorithm, combined with the full mass spectrometry (MS)/dd-MS² scan mode that turns on the "if idle pick others" function. By using ultra-high performance liquid chromatography and high-resolution MS (UHPLC/HRMS), the analysis used both targeted and non-targeted methods to detect proanthocyanidins. Finally, 50 oligomeric proanthocyanidin (OPC) compounds were identified, including 7 monomers, 22 dimers, 20 trimers, and 1 tetramer, most of which were procyanidins of proanthocyanidins (84%), and a small amount of prodelphinidin (14%) and other types of proanthocyanidins (2%), which enabled the systematic characterization of proanthocyanidin components from Rhodiolae Crenulatae Radix et Rhizomes. Meanwhile, the comparison with the grape seeds OPCs standard (United States Pharmacopeia) revealed that the proanthocyanidins in Rhodiolae Crenulatae Radix et Rhizomes were more abundant, suggesting that the proanthocyanidins in Rhodiolae Crenulatae Radix et Rhizomes has promising applications.

Plant-derived nanovesicles have gained attention given their similarity to mammalian exosomes and advantages such as low cost, sustainability, and tissue targeting. Thus, they hold promise for disease treatment and drug delivery. In this study, we proposed a time-efficient method, PEG 8000 combined with sucrose density gradient centrifugation to prepare ginger-derived nanovesicles (GDNVs). Subsequently, curcumin (CUR) was loaded onto GDNV by ultrasonic incubation. The optimum conditions for ginger-derived nanovesicles loaded with curcumin (CG) were ultrasound time of 3 min, a carrier-to-drug ratio (GDNV:CUR) of 1:1. The study achieved a high loading capacity (94.027% ± 0.094%) and encapsulation efficiency (89.300% ± 0.344%). Finally, the drugs' in vivo distribution and anti-colitis activity were investigated in mice. CG was primarily distributed in the colon after oral administration. Compared to CUR and GDNV, CG was superior in improving disease activity, colon length, liver and spleen coefficients, myeloperoxidase activity, and biochemical factor levels in ulcerative colitis (UC) mice. In addition, CG plays a protective role against UC by modulating serum metabolite levels and gut flora. In summary, our study demonstrated that GDNV can be used for CUR delivery with enhanced therapeutic potential.

Objective: To investigate the correlation between skin elasticity and setup error in optical surface image-guided radiotherapy. Methods: The skin elasticity (R7) data of the head, chest and abdomen were extracted and analyzed its correlation with age by systematic literature review. Fifty-four patients diagnosed with nasopharyngeal carcinoma, breast cancer and cervical cancer were recruited in this study. Firstly, the patients were positioned based on the room laser and markers. Subsequently, the patient position was verified by the Varian On-Board Imager, and then C-Rad Catalyst was adopted to obtain surface images in two states (mask or non-mask) as reference images. In the subsequent fraction treatment, after initial positioning, the local calibration was performed by Catalyst, and setup errors in three directions were recorded. Meanwhile, the patient setup was verified by CBCT twice a week. The Pearson correlation analysis was performed to analyze the correlation between setup error and age. Results: The skin elasticity was negatively correlated with aging (P<0.01). The correlation coefficient between random error and age in head-and-neck cancer were 0.645, 0.624 and 0.866 in the AP, SI and LR directions (all P<0.05) for male patients without mask, respectively. The system error was significantly correlated with age in the LR direction (P<0.05) for male patients, and in the AP direction (P<0.05) for female patients with head-and-neck cancer without mask. The setup error had a significant correlation with skin elasticity in male patients with head-and-neck cancer, and the sequence of absolute value of correlation coefficient was LR>SI>AP. Conclusion In optical surface-guided radiotherapy of head and neck cancer, skin elasticity may be a significant index for assessing the setup errors in male patients.

Objective To evaluate the clinical application of Catalyst system in positioning patients during cervical cancer radiotherapy,and to analyze its correlation with the bladder volume and body mass index (BMI) of patients.Methods A total of 33 patients diagnosed with cervical cancer from July to December 2017 in our hospital were included in the study.All patients were auxiliary positioned by an optical surface imaging system (C-Pad Catalyst) before each treatment.The CBCT imaging was executed twice a week.The setup errors from Catalyst and CBCT in the anterior-posterior (AP),superior-inferior (SI) and leg-fight (LR) directions were recorded.Each patient's bladder volume and BMI were also recorded.Results The setup errors between Catalyst with masks and CBCT had the significant difference in SI (P＜0.05) and LR (P＜0.05).For Catalyst without masks,the setup errors with the bladder volume of 200-300 ml had the significant association in SI (R=-0.316,P＜0.05).For the bladder volume of＞300 ml,the setup errors for Catalyst with masks had the significant association in AP (R=-0.493,P＜0.05),and that without masks had the significant association in SI and LR (R=0.335,P＜0.05,R=-0.348,P＜0.05).For patients of＜25 kg/m2,setup errors for Catalyst with masks had the significant association with the BMI in LR (R=0.197,P＜0.05);for ≥ 25 kg/m2,that with masks had the significant association in AP and SI (R =0.818,P＜0.05;R=-0.498,P＜0.05),that without masks had the significant association in AP and LR (R=0.652,P＜0.05;R=-0.558,P＜0.05).Conclusion Unlike CBCT system,the patient positioning by Catalyst system was easily affected by the bladder volume and BMI of patients.

Objective To explore the effect of divided injection of fat grafting and to provide a new sight for the strategy of clinical particle fat transplantation.Methods Adipose tissue was aspirated from the healthy female abdomen by liposuction.In the control group,0.5 ml of adipose tissue was subcutaneously injected into the nude mice.The experimental group was injected with 0.25 ml first,followed by 0.25 ml injection of adipose tissue on the 7th,14th and 30th days.To assess graft retention rate and effectiveness we measured the wet weight and observed the pathological sections with HE or perilipin immunofluorescence.Results The wet weight of the tissue between the experimental group and the control group had no statistical difference,the experimental group had less necrosis and empty tissue than the control group.The proportion of perilipin positive staining tissue in 7 day group had statistical difference from that of the control group.Conclusions The strategy of preinjection of part of the adipose tissue and then supplement of the residual tissue after 7 days may increase the proportion of active adipose tissue in the graft.

Zinc-α2-glycoprotein (ZAG), encoded by the AZGP1 gene, is a major histocompatibility complex I molecule and a lipid-mobilizing factor. ZAG has been demonstrated to promote lipid metabolism and glucose utilization, and to regulate insulin sensitivity. Apart from adipose tissue, skeletal muscle, liver, and kidney, ZAG also occurs in brain tissue, but its distribution in brain is debatable. Only a few studies have investigated ZAG in the brain. It has been found in the brains of patients with Krabbe disease and epilepsy, and in the cerebrospinal fluid of patients with Alzheimer disease, frontotemporal lobe dementia, and amyotrophic lateral sclerosis. Both ZAG protein and AZGP1 mRNA are decreased in epilepsy patients and animal models, while overexpression of ZAG suppresses seizure and epileptic discharges in animal models of epilepsy, but knowledge of the specific mechanism of ZAG in epilepsy is limited. In this review, we summarize the known roles and molecular mechanisms of ZAG in lipid metabolism and glucose metabolism, and in the regulation of insulin sensitivity, and discuss the possible mechanisms by which it suppresses epilepsy.
Adipocytes ; metabolism ; Animals ; Brain ; metabolism ; Carrier Proteins ; metabolism ; Epilepsy ; metabolism ; Glucose ; metabolism ; Glycoproteins ; metabolism ; Humans ; Insulin Resistance ; Lipid Metabolism ; Neurons ; metabolism ; Signal Transduction