Computational analysis can accurately detect drug-gene interactions (DGIs) cost-effectively. However, transductive learning models are the hotspot to reveal the promising performance for unknown DGIs (both drugs and genes are present in the training model), without special attention to the unseen DGIs (both drugs and genes are absent in the training model). In view of this, this study, for the first time, proposed an inductive learning-based model for the precise identification of unseen DGIs. In our study, by integrating disease nodes to avoid data sparsity, a multi-relational drug-disease-gene (DDG) graph was constructed to achieve effective fusion of data on DDG intro-relationships and inter-actions. Following the extraction of graph features by utilizing graph embedding algorithms, our next step was the retrieval of the attributes of individual gene and drug nodes. In this way, a hybrid feature characterization was represented by integrating graph features and node attributes. Machine learning (ML) models were built, enabling the fulfillment of transductive predictions of unknown DGIs. To realize inductive learning, this study generated an innovative idea of transforming known node vectors derived from the DDG graph into representations of unseen nodes using node similarities as weights, enabling inductive predictions for the unseen DGIs. Consequently, the final model was superior to existing models, with significant improvement in predicting both external unknown and unseen DGIs. The practical feasibility of our model was further confirmed through case study and molecular docking. In summary, this study establishes an efficient data-driven approach through the proposed modeling, suggesting its value as a promising tool for accelerating drug discovery and repurposing.

Over-expression of glutathione S-transferase(GST)can promote Cisplatin resistance in hepatocellular carcinoma(HCC)treatment.Hence,inhibiting GST is an attractive strategy to improve Cisplatin sensi-tivity in HCC therapy.Although several synthesized GST inhibitors have been developed,the side effects and narrow spectrum for anticancer seriously limit their clinical application.Considering the abundance of natural compounds with anticancer activity,this study developed a rapid fluorescence technique to screen"green"natural GST inhibitors with high specificity.The fluorescence assay demonstrated that schisanlactone B(hereafter abbreviated as C1)isolated from Xue tong significantly down-regulated GST levels in Cisplatin-resistant HCC cells in vitro and in vivo.Importantly,C1 can selectively kill HCC cells from normal liver cells,effectively improving the therapeutic effect of Cisplatin on HCC mice by down-regulating GST expression.Considering the high GST levels in HCC patients,this compound demon-strated the high potential for sensitizing HCC therapy in clinical practice by down-regulating GST levels.

Objective:To investigate the clinical efficacy and prognostic influencing factors of radical surgery for duodenal gastrointestinal stromal tumor (GIST).Methods:The retrospective cohort study was conducted. The clinicopathological data of 741 duodenal GIST patients who under-went radical surgery in 17 medical centers, including 121 cases in Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine, 121 cases in Chinese PLA General Hospital, 116 cases in Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, 77 cases in Fudan University Shanghai Cancer Center, 77 cases in West China Hospital, Sichuan University, 31 cases in Guangdong Provincial People′s Hospital, 24 cases in Fujian Cancer Hospital, 22 cases in Fujian Medical University Union Hospital, 25 cases in Shengjing Hospital of China Medical University, 19 cases in Xiangya Hospital, Central South University, 23 cases in Zhejiang Cancer Hospital, 17 cases in Liaoning Cancer Hospital＆Institute, 17 cases in the First Affiliated Hospital of Xiamen University, 15 cases in Sun Yat-sen University Cancer Center, 14 cases in the First Affiliated Hospital of Nanjing Medical University, 14 cases in Zhongshan Hospital Affiliated to Xiamen University and 8 cases in General Hospital of Chinese People′s Liberation Army Air Force, from January 2010 to April 2020 were collected. There were 346 males and 395 females, aged 55(range, 17?86)years. Observation indicators: (1) neoadjuvant treatment; (2) surgical and postoperative situations; (3) follow-up; (4) stratified analysis. Follow-up was conducted using outpatient examination or telephone interview. Patients were followed up once every 3?6 months during neoadjuvant therapy and once every 6?12 months after radical surgery to detect tumor recurrence and survival of patient up to April 2022. Measurement data with normal distribution were represented as Mean± SD. Measurement data with skewed distribution were represented as M(range), and comparison between groups was conducted using the Mann-Whitney U test. Count data were described as absolute numbers or percentages, and comparison between groups was conducted using chi-square test or Fisher exact probability. The Kaplan-Meier method was used to draw survival curves and calculate survival rates. Log-rank test was used for survival analysis. The COX regression model was used for univariate and multivariate analyses. Propensity score matching was done by the 1∶1 nearest neighbor matching method, and the matching tolerance was 0.02. Results:(1) Neoadjuvant therapy. Of the 741 patients, 34 cases received neoadjuvant therapy for 8(range, 3?44)months. Cases assessed as partial response, stable disease and progressive disease before the radical surgery of the 34 cases were 21, 9, 4, respectively. The tumor diameter of the 34 patients before the neoadjuvant therapy and before the radical surgery were 8.0(range, 3.0?26.0)cm and 5.3(range, 3.0?18.0)cm, with the regression rate as 31.9%(range, ?166.7% to 58.3%). (2) Surgical and postoperative situations. Of the 741 patients, 34 cases underwent radical surgery after receiving neoadjuvant therapy, and 707 cases underwent radical surgery directly. All the 741 patients underwent radical surgery successfully, in which 633, 102 and 6 cases received open surgery, laparoscopic surgery and endoscopic treatment, respectively. Of the 633 cases receiving open surgery and the 102 cases receiving laparoscopic surgery, cases with surgical resection range as pancreatoduodenectomy (PD) was 238, and cases with surgical resection range as duodenal limited resection, including duodenal wedge resection, distal gastrectomy, segmental duodenal resection, local resection of duodenal tumor or segmental duodenum combined with subtotal gastrectomy, was 497, 226, 55, 204, 12. Of the 741 patients, 131 cases had post-operative complications including 113 cases with grade Ⅰ?Ⅱ complications and 18 cases with ≥ grade Ⅲ complications of the Clavien-Dindo classification. The duration of postoperative hospital stay of the 741 patients was 13(range, 4?120)days. Of the 707 patients receiving direct radical surgery, 371 cases were evaluated as extremely low risk, low risk, medium risk of the modified National Institutes of Health (NIH) risk classification after surgery, and 336 cases were evaluated as high risk in which 205 cases receive postoperative adjuvant imatinib therapy with the treatment time as 24(range, 6?110)months. (3) Follow-up. All the 741 patients were followed up for 58(range, 7?150)months. During the follow-up, 110 patients had tumor recurrence and metastasis. The 1-, 3-, 5-year overall survival rates and 1-, 3-, 5-year disease-free survival rates of the 741 patients were 100.0%, 98.6%, 94.5% and 98.4%, 90.9%, 84.9%, respectively. The 1-, 3-, 5-year overall survival rates and 1-, 3-, 5-year disease-free survival rates of the 707 patients receiving direct radical surgery were 100.0%, 98.5%, 94.3% and 98.4%, 91.1%, 85.4%, respectively. (4) Stratified analysis. ① Analysis of prognostic factors in patients undergoing radical surgery directly. Results of univariate analysis showed that primary tumor location, tumor diameter, mitotic count, modified NIH risk classification and tumor gene information were related factors affecting the overall survival of 707 patients with primary duodenal GIST who underwent direct radical surgery ( hazard ratio=0.43, 0.18, 0.22, 0.06, 0.29, 95% confidence intervals as 0.20?0.93, 0.09?0.35, 0.10?0.50, 0.03?0.12, 0.09?0.95, P<0.05). The primary tumor location, tumor diameter, mitotic count, modified NIH risk classification were related factors affecting the disease-free survival of 707 patients with primary duodenal GIST who underwent direct radical surgery ( hazard ratio=0.65, 0.25, 0.25, 0.10, 95% confidence intervals as 0.41?1.03, 0.17?0.37, 0.15?0.42, 0.07?0.15, P<0.05). Results of multivariate analysis showed that primary tumor located at the horizontal segment of duodenum, mitotic count >5/50 high power field, tumor gene KIT exon 9 mutation were independent risk factors affecting the overall survival of 365 patients with primary duodenal GIST after removing 342 patients without tumor gene information who underwent direct radical surgery ( hazard ratio=2.85, 2.73, 3.13, 95% confidence intervals as 1.12?7.20, 1.07?6.94, 1.23?7.93, P<0.05). Tumor diameter >5 cm and mitotic count >5/50 high power field were independent risk factors affecting the disease-free survival of 707 patients with primary duodenal GIST who underwent direct radical surgery ( hazard ratio=3.19, 2.98, 95% confidence intervals as 2.05?4.97, 1.99?4.45, P<0.05). ② Effect of postoperative adjuvant therapy on prognosis of high-risk patients of modified NIH risk classification. Of the 336 patients evaluated as high risk of the modified NIH risk classification, the 5-year overall survival rate and 5-year disease-free survival rate were 94.6% and 77.3% in the 205 cases with postoperative adjuvant therapy, versus 83.2% and 64.4% in the 131 cases without postoperative adjuvant therapy, showing significant differences between them ( χ2=8.39, 4.44, P<0.05). Of the 205 patients evaluated as high risk of the modified NIH risk classification who received postoperative adjuvant therapy, there were 106 cases receiving postoperative adjuvant therapy <36 months, with the 5-year overall survival rate and 5-year disease-free survival rate were 87.1% and 58.7%, and there were 99 cases receiving post-operative adjuvant therapy ≥36 months, with the 5-year overall survival rate and 5-year disease-free survival rate were 100.0% and 91.5%. There were significant differences in the 5-year overall survival rate and 5-year disease-free survival rate between the 106 patients and the 99 patients ( χ2=13.92, 29.61, P<0.05). ③ Comparison of clinical efficacy of patients with different surgical methods. Before propensity score matching, cases with primary tumor located at bulb, descending, horizontal, ascending segment of duodenum, cases with tumor diameter ≤5 cm and >5 cm were 95, 307, 147, 34, 331, 252, in the 583 patients receiving open surgery with complete clinical data, versus 15, 46, 17, 5, 67, 16 in the 83 patients receiving laparoscopic surgery with complete clinical data, showing no significant difference in the primary tumor location ( χ2=0.94, P>0.05), and a significant difference in the tumor diameter ( χ2=17.33, P<0.05) between them. After propensity score matching, the above indicator were 16, 39, 20, 8, 67, 16 in the 83 patients receiving open surgery, versus 15, 46, 17, 5, 67, 16 in the 83 patients receiving laparoscopic surgery, showing no significant difference between them ( χ2=1.54, 0.00, P>0.05). Cases with postoperative complications, cases with grade Ⅰ?Ⅱ complica-tions and ≥grade Ⅲ complications of the Clavien-Dindo classification, duration of postoperative hospital stay, the 5-year overall survival rate and 5-year disease-free survival rate were 17, 12, 5, 11(range, 5?120)days, 92.0%, 100.0% in the 83 patients receiving open surgery, versus 9, 7, 2, 11(range, 5?41)days, 91.6%, 97.3% in the 83 patients receiving laparoscopic surgery, showing no signi-ficant difference in postoperative complications, duration of postoperative hospital stay, the 5-year overall survival rate and 5-year disease-free survival rate ( χ2=2.91, Z=3 365.50, χ2=3.02, 1.49, P>0.05) between them. There was no significant difference in complications of the Clavien-Dindo classification between them ( P>0.05). ④ Comparison of clinical efficacy of patients with primary tumor located at the descending segment of duodenum who underwent surgery with different surgical resection scopes. Before propensity score matching, cases with tumor diameter ≤5 cm and >5 cm, cases with tumor located at opposite side of mesangium and mesangium were 71, 85, 28, 128 in the 156 patients with primary tumor located at the descending segment of duodenum who underwent PD with complete clinical data, versus 92, 41, 120, 13 in the 133 patients with primary tumor located at the descending segment of duodenum who underwent duodenal limited resection with complete clinical data, showing significant differences between them ( χ2=16.34, 150.10, P<0.05). After propensity score matching, the above indicator were 28, 13, 16, 25 in the 41 patients with primary tumor located at the descending segment of duodenum who underwent PD with complete clinical data, versus 28, 13, 16, 25 in the 41 patients with primary tumor located at the descending segment of duodenum who underwent duodenal limited resection with complete clinical data, showing no significant difference between them ( χ2=0.00, 0.00, P>0.05). Cases with postopera-tive complications, cases with grade Ⅰ?Ⅱ complications and ≥grade Ⅲ compli-cations of the Clavien-Dindo classification, duration of postoperative hospital stay, the 5-year overall survival rate and 5-year disease-free survival rate were 13, 11, 2, 15(range, 9?62)days, 94.2%, 64.3% in the 41 patients with primary tumor located at the descending segment of duodenum who underwent PD with complete clinical data, versus 9, 8, 0, 15(range, 7?40)days, 100.0%, 78.8% in the 41 patients with primary tumor located at the descending segment of duodenum who underwent duodenal limited resection with complete clinical data, showing no significant difference in post-operative complica-tions, the 5-year overall survival rate and 5-year disease-free survival rate ( χ2=0.99, 0.34, 1.86, P>0.05) between them. There was no significant difference in complications of the Clavien-Dindo classification ( P>0.05) and there was a significant difference in duration of postopera-tive hospital stay ( Z=614.50, P<0.05) between them. Conclusions:The clinical efficacy of radical surgery for duodenal GIST are ideal. Primary tumor located at the horizontal segment of duodenum, mitotic count >5/50 high power field, tumor gene KIT exon 9 mutation are independent risk factors affec-ting the overall survival of patients undergoing direct radical surgery and tumor diameter >5 cm and mitotic count >5/50 high power field are independent risk factors affecting the disease-free survival of patients. There is no significant difference in the short-term efficacy and long-term prognosis between patients undergoing open surgery and laparoscopic surgery. For patients with primary tumor located at the descending segment of duodenum, the duration of postoperative hospital stay is longer in patients undergoing PD compared with patients undergoing duodenal limited resection. For patients evaluated as high risk of the modified NIH risk classification, posto-perative adjuvant therapy and treatment time ≥36 months are conducive to improving the prognosis of patients.