Autoencoder-Based Detection of Dynamic Allostery Triggered by Ligand Binding Based on Molecular Dynamics.

Dynamic allostery on proteins, triggered by regulator binding or chemical modifications, transmits information from the binding site to distant regions, dramatically altering protein function. It is accompanied by subtle changes in side-chain conformations of the protein, indicating that the changes in dynamics, and not rigid or large conformational changes, are essential to understand regulation of protein function. Although a lot of experimental and theoretical studies have been dedicated to investigate this issue, the regulation mechanism of protein function is still being debated. Here, we propose an autoencoder-based method that can detect dynamic allostery. The method is based on the comparison of time fluctuations of protein structures, in the form of distance matrices, obtained from molecular dynamics simulations in ligand-bound and -unbound forms. Our method detected that the changes in dynamics by ligand binding in the PDZ2 domain led to the reorganization of correlative fluctuation motions among residue pairs, which revealed a different view of the correlated motions from the PCA and DCCM. In addition, other correlative motions were also found as a result of the dynamic perturbation from the ligand binding, which may lead to dynamic allostery. This autoencoder-based method would be usefully applied to the signal transduction and mutagenesis systems involved in protein functions and severe diseases.

Adjuvant chemotherapy improves survival of patients with high-risk upper urinary tract urothelial carcinoma: a propensity score-matched analysis.

BACKGROUND: The purposes of this study were to determine whether adjuvant chemotherapy (AC) improved the prognosis of patients with high-risk upper urinary tract urothelial carcinoma (UTUC)and to identify the patients who benefited from AC. METHODS: Among a multi-center database of 1014 patients who underwent RNU for UTUC, 344 patients with ≥ pT3 or the presence of lymphovascular invasion (LVI) were included. Cancer-specific survival (CSS) estimates were calculated by the Kaplan-Meier method, and groups were compared by the log-rank test. Each patient's probability of receiving AC depending on the covariates in each group was estimated by logistic regression models. Propensity score matching was used to adjust the confounding factors for selecting patients for AC, and log-rank tests were applied to these propensity score-matched cohorts. Cox proportional hazards regression modeling was used to identify the variables with significant interaction with AC. Variables included age, pT category, LVI, tumor grade, ECOG performance status and low sodium or hemoglobin score, which we reported to be a prognostic factor of UTUC. RESULTS: Of the 344 patients, 241 (70%) had received RNU only and 103 (30%) had received RNU+AC. The median follow-up period was 32 (range 1-184) months. Overall, AC did not improve CSS (P = 0.12). After propensity score matching, the 5-year CSS was 69.0% in patients with RNU+AC versus 58.9% in patients with RNU alone (P = 0.030). Subgroup analyses of survival were performed to identify the patients who benefitted from AC. Subgroups of patients with low preoperative serum sodium (≤ 140 mEq/ml) or hemoglobin levels below the normal limit benefitted from AC (HR 0.34, 95% CI 0.15-0.61, P = 0.001). In the subgroup of patients with normal sodium and normal hemoglobin levels, 5-year CSS was 77.7% in patients with RNU+AC versus 80.2% in patients with RNU alone (P = 0.84). In contrast, in the subgroup of patients with low sodium or low hemoglobin levels, 5-year CSS was 71.0% in patients with RNU+AC versus 38.5% in patients with RNU alone (P < 0.001). CONCLUSIONS: High-risk UTUC patients, especially subgroups of patients with lower sodium and hemoglobin levels, could benefit from AC after RNU.

Association between UGT1A1*28*28 genotype and lung cancer in the Japanese population.

BACKGROUND: Lung cancer is the leading cause of cancer death and is closely linked to tobacco smoking. Genetic polymorphisms in genes that encode enzymes involved in metabolizing tobacco carcinogens could affect an individual's risk for lung cancer. While polymorphism of UDP-glucuronosyltransferase1A1 (UGT1A1) is involved in detoxification of benzo(a)pyrene-7,8-dihydrodiol(-), a major tobacco carcinogen, the association between UGT1A1 genotype and lung cancer has not been examined. METHODS: We retrieved the clinical data of 5,285 patients who underwent systemic chemotherapy at Kyoto University Hospital. A total of 765 patients (194 lung cancer patients and 671 patients with other malignancies) with UGT1A1 genotyping data were included in this analysis. We used logistic regression with recessive, dominant, and additive models to identify differences in genotype frequencies between lung cancer and other malignancies. RESULTS: In the recessive model, UGT1A1*28*28 genotype was significantly associated with lung cancer compared to other malignancies (odds ratio 5.3, P = 0.0083). Among lung cancer patients with a smoking history, squamous cell carcinoma was significantly predominant in patients with UGT1A1*28*28 compared to those with other UGT1A1 genotypes (P = 0.024). CONCLUSION: This is the first study to demonstrate a significant association between the homozygous UGT1A1*28 genotype and lung cancer.

Structure-based analyses of gut microbiome-related proteins by neural networks and molecular dynamics simulations.

An imbalance in the gut microbiome is linked to immune disorders, such as autoimmune, allergic, and chronic inflammatory disorders. Elucidation of disease mechanisms is a matter of urgency. It requires precise elucidation of the structure-based mechanisms of protein interactions involved in disease onset. In addition, an understanding of the protein dynamics is vital because these fluctuations affect the function and interaction of disease-associated proteins. Experimental evaluation of not only protein interactions, functions, and structures but also the dynamics are time-consuming; therefore, computational predictions are necessary to elucidate disease mechanisms. Here, we introduce recent studies on structure-based analyses of proteins using computational approaches, particularly artificial intelligence (AI) and molecular dynamics (MD) simulations.

Autoencoder-based detection of the residues involved in G protein-coupled receptor signaling.

Regulator binding and mutations alter protein dynamics. The transmission of the signal of these alterations to distant sites through protein motion results in changes in protein expression and cell function. The detection of residues involved in signal transmission contributes to an elucidation of the mechanisms underlying processes as vast as cellular function and disease pathogenesis. We developed an autoencoder (AE) based method that detects residues essential for signaling by comparing the fluctuation data, particularly the time fluctuation of the side-chain distances between residues, during molecular dynamics simulations between the ligand-bound and -unbound forms or wild-type and mutant forms of proteins. Here, the AE-based method was applied to the G protein-coupled receptor (GPCR) system, particularly a class A-type GPCR, CXCR4, to detect the essential residues involved in signaling. Among the residues involved in the signaling of the homolog CXCR2, which were extracted from the literature based on the complex structures of the ligand and G protein, our method could detect more than half of the essential residues involved in G protein signaling, including those spanning the fifth and sixth transmembrane helices in the intracellular region, despite the lack of information regarding the interaction with G protein in our CXCR4 models.

Combination of host immune metabolic biomarkers for the PD-1 blockade cancer immunotherapy.

BACKGROUNDCurrent clinical biomarkers for the programmed cell death 1 (PD-1) blockade therapy are insufficient because they rely only on the tumor properties, such as programmed cell death ligand 1 expression frequency and tumor mutation burden. Identifying reliable, responsive biomarkers based on the host immunity is necessary to improve the predictive values.METHODSWe investigated levels of plasma metabolites and T cell properties, including energy metabolism markers, in the blood of patients with non-small cell lung cancer before and after treatment with nivolumab (n = 55). Predictive values of combination markers statistically selected were evaluated by cross-validation and linear discriminant analysis on discovery and validation cohorts, respectively. Correlation between plasma metabolites and T cell markers was investigated.RESULTSThe 4 metabolites derived from the microbiome (hippuric acid), fatty acid oxidation (butyrylcarnitine), and redox (cystine and glutathione disulfide) provided high response probability (AUC = 0.91). Similarly, a combination of 4 T cell markers, those related to mitochondrial activation (PPARγ coactivator 1 expression and ROS), and the frequencies of CD8+PD-1hi and CD4+ T cells demonstrated even higher prediction value (AUC = 0.96). Among the pool of selected markers, the 4 T cell markers were exclusively selected as the highest predictive combination, probably because of their linkage to the abovementioned metabolite markers. In a prospective validation set (n = 24), these 4 cellular markers showed a high accuracy rate for clinical responses of patients (AUC = 0.92).CONCLUSIONCombination of biomarkers reflecting host immune activity is quite valuable for responder prediction.FUNDINGAMED under grant numbers 18cm0106302h0003, 18gm0710012h0105, and 18lk1403006h0002; the Tang Prize Foundation; and JSPS KAKENHI grant numbers JP16H06149, 17K19593, and 19K17673.

GLIDA: GPCR-ligand database for chemical genomic drug discovery.

G-protein coupled receptors (GPCRs) represent one of the most important families of drug targets in pharmaceutical development. GPCR-LIgand DAtabase (GLIDA) is a novel public GPCR-related chemical genomic database that is primarily focused on the correlation of information between GPCRs and their ligands. It provides correlation data between GPCRs and their ligands, along with chemical information on the ligands, as well as access information to the various web databases regarding GPCRs. These data are connected with each other in a relational database, allowing users in the field of GPCR-related drug discovery to easily retrieve such information from either biological or chemical starting points. GLIDA includes structure similarity search functions for the GPCRs and for their ligands. Thus, GLIDA can provide correlation maps linking the searched homologous GPCRs (or ligands) with their ligands (or GPCRs). By analyzing the correlation patterns between GPCRs and ligands, we can gain more detailed knowledge about their interactions and improve drug design efforts by focusing on inferred candidates for GPCR-specific drugs. GLIDA is publicly available at http://gdds.pharm.kyoto-u.ac.jp:8081/glida. We hope that it will prove very useful for chemical genomic research and GPCR-related drug discovery.

CGBVS-DNN: Prediction of Compound-protein Interactions Based on Deep Learning.

Computational prediction of compound-protein interactions (CPIs) is of great importance for drug design as the first step in in-silico screening. We previously proposed chemical genomics-based virtual screening (CGBVS), which predicts CPIs by using a support vector machine (SVM). However, the CGBVS has problems when training using more than a million datasets of CPIs since SVMs require an exponential increase in the calculation time and computer memory. To solve this problem, we propose the CGBVS-DNN, in which we use deep neural networks, a kind of deep learning technique, instead of the SVM. Deep learning does not require learning all input data at once because the network can be trained with small mini-batches. Experimental results show that the CGBVS-DNN outperformed the original CGBVS with a quarter million CPIs. Results of cross-validation show that the accuracy of the CGBVS-DNN reaches up to 98.2 % (σ<0.01) with 4 million CPIs.

Development and validation of a set of six adaptable prognosis prediction (SAP) models based on time-series real-world big data analysis for patients with cancer receiving chemotherapy: A multicenter case crossover study.

BACKGROUND: We aimed to develop an adaptable prognosis prediction model that could be applied at any time point during the treatment course for patients with cancer receiving chemotherapy, by applying time-series real-world big data. METHODS: Between April 2004 and September 2014, 4,997 patients with cancer who had received systemic chemotherapy were registered in a prospective cohort database at the Kyoto University Hospital. Of these, 2,693 patients with a death record were eligible for inclusion and divided into training (n = 1,341) and test (n = 1,352) cohorts. In total, 3,471,521 laboratory data at 115,738 time points, representing 40 laboratory items [e.g., white blood cell counts and albumin (Alb) levels] that were monitored for 1 year before the death event were applied for constructing prognosis prediction models. All possible prediction models comprising three different items from 40 laboratory items (40C3 = 9,880) were generated in the training cohort, and the model selection was performed in the test cohort. The fitness of the selected models was externally validated in the validation cohort from three independent settings. RESULTS: A prognosis prediction model utilizing Alb, lactate dehydrogenase, and neutrophils was selected based on a strong ability to predict death events within 1-6 months and a set of six prediction models corresponding to 1,2, 3, 4, 5, and 6 months was developed. The area under the curve (AUC) ranged from 0.852 for the 1 month model to 0.713 for the 6 month model. External validation supported the performance of these models. CONCLUSION: By applying time-series real-world big data, we successfully developed a set of six adaptable prognosis prediction models for patients with cancer receiving chemotherapy.