基于特征的深度学习预测化合物-蛋白质相互作用的研究进展

药物研发过程中,化合物-蛋白质相互作用（compound-protein interaction,CPI）预测是发现苗头化合物、药物重定位等研究的关键技术手段。近年来,深度学习被广泛应用于CPI研究,加速了药物发现中CPI预测的发展。本文重点讨论基于特征的 CPI 预测模型,首先,介绍了CPI预测中常见的数据库、化合物和蛋白质的典型特征表示方法。根据建模中的关键问题,从多模态和注意力机制两个方面,对基于特征的CPI预测模型展开论述。在此基础上,选取其中12个模型,在3个经典数据集上评估了模型在分类任务和回归任务中的性能。本文总结当前该领域面临的挑战,对未来的发展方向进行展望,为CPI预测方法进一步研究提供思路。     

Immune adjuvants for chemotherapy or radiotherapy in the 9L rat brain tumor model

Summary The therapeutic efficacy and toxicity of three biological response modifiers,Corynebacterium parvum (Cp), Chinese blister beetle extract (CBBE), recombinant human IL-1 (rhIL-1), used alone or in combination with chemotherapy or radiotherapy, were investigated in the intracerebral (ic) rat 9L brain tumor model. Used alone, Cp (2mg/rat, ip plus 70g/rat, ic), CBBE (5l of an ethanol extract, ic), or IL-1 (lg/rat, ic or 1g/rat × 3, q 3 d, ic), had no effect on animal survival compared to the untreated or saline treated controls. When combined with chemotherapy or radiotherapy, the three immunotherapeutic agents did not show any additive effects on survival compared to that observed with systemic BCNU (12mg/kg), local ic bleomycin (0.25 unit), or local radiotherapy (16 Gy). While ic IL-1 did not produce evident toxicity, there was fatal toxicity caused by ic Cp or CBBE treatment in a few animals. The combination of Cp and bleomycin produced severe neurotoxicity, resulting in the early death of animals. This study demonstrates a lack of efficacy of the nonspecific immune adjuvants IL-1, Cp or CBBE, used either alone or combined with cytotoxic chemotherapy or radiotherapy, in this rat brain tumor model.     

In Vitro Evaluation of Dexamethasone-β-D-Glucuronide for Colon-Specific Drug Delivery

Dexamethasone--D-glucuronide is a potential prodrug for colonic delivery of the antiinflammatory corticosteroid dexamethasone. Previous studies [T. R. Tozer et al., Pharm. Res. 8:445–454 (1991)] indicated that a glucoside prodrug of dexamethasone was susceptible to hydrolysis in the upper gastrointestinal tract. Resistance of dexamethasone--D-glucuronide to hydrolysis in the upper gastrointestinal tract was therefore assessed. Conventional, germfree, and colitic rats were used to examine enzyme levels along the gastrointestinal tract to compare the stability of two model substrates (p-nitrophenyl--D-glucoside and --D-glucuronide) and to evaluate the prodrug dexamethasone--D-glucuronide. Hydrolytic activity was examined in the luminal contents, mucosa, and underlying muscle/connective tissues in all three types of rats. Enzymatic activity (-D-glucosidase and -D-glucuronidase) was greatest in the lumen of cecum and colon of conventional rats. In contrast, germ-free rats exhibited relatively high levels of -D-glucosidase activity (about 80% of total activity in the conventional rats) in the proximal small intestine (PSI) and the distal small intestine (DSI). Rats with induced colitis (acetic acid) showed reduced levels of luminal -D-glucuronidase activity in the large intestine; however, -D-glucosidase activity was relatively unchanged relative to that of the conventional rat. Mucosal -D-glucuronidase activity was significantly lower in the colitic rats compared with that in the conventional animals. Despite reduced luminal levels of -D-glucuronidase activity in the colitic rats, there was still a sharp gradient of activity between the small and the large intestines. Permeability of the glucoside and glucuronide prodrugs of dexamethasone through a monolayer of Caco-2 cells was relatively low compared to that of dexamethasone. The results indicate that dexamethasone--D-glucuronide should be relatively stable and poorly absorbed in the upper gastrointestinal tract. Once the compound reaches the large intestine, it should be hydrolyzed to dexamethasone and glucuronic acid. Specificity of colonic delivery in humans should be even greater due to lower levels of -D-glucuronidase activity in the small intestine compared with that in the laboratory rat.     

Characterization of Oil-in-Water Emulsions Prepared from Solid-State Emulsions: Effect of Matrix and Oil Phase

Emulsions (o/w) were prepared from solid-state emulsions comprised of various matrix materials and oils and the resultant particle size properties determined. Results suggest that for those matrices that can form solid-state emulsions, the droplet size decreased as a function of time, as previously observed. The final droplet size was dependent on the oil utilized but was independent of the matrix material. The use of mineral oil resulted in the smallest droplet diameter (1.5 µm) while isopropyl myristate resulted in the largest droplet diameter (3 µm). With the exception of mineral oil, the oil/water interfacial tension was found to be directly proportional to the droplet diameter. The rate of emulsification appeared to be bi-phasic. The initial emulsification phase appeared to be independent of the matrix material while the terminal phase was a function of the matrix material. Most importantly, it was found that solid state emulsions could be prepared from a diverse, yet specific, list of matrices.     

A good ligand is hard to find: Automated docking methods

Summary Many approaches have been developed for solving the docking problem: Predict the structure and binding free energy of a ligand-receptor complex given only the structures of the free ligand and receptor. We review major approaches for docking small-molecule ligands to receptors and focus on the successes and limitations of their application to drug design.     

Beyond Zinberg's 'social setting': a processural view of illicit drug use

This article seeks to go beyond existing understandings of 'social setting', or social context, in the addictions field. Using drug-career data, collected during recent ethnographic research, the fluidity and dynamism of recreational drug-using social scenes are described before introducing a number of anthropological and sociological concepts. These concepts are processural, that is, they seek to describe and analyze social and cultural change. Finally, the theoretical and practical implications of recognizing the diversity and ever-changing nature of recreational drug-using social networks are discussed.     

HIV prevalence among IDUs in Australia: a methodological review

A review was carried out of Australian studies which have measured the prevalence of HIV infection among injecting drug users (IDUs). The review considered published studies which had reported on serologically-determined HIV prevalence. There were five studies reported from specialized sexually-transmissible disease of HIV clinics, five studies reported from health services aimed at IDUs, three studies reported from other health services and one multi-centre behavioural study. The main findings from the studies were that HIV prevalence in IDUs has been low in Australia, apart from in male IDUs who also had homosexual contact. HIV prevalence ranged from 20 to 24% in male IDUs reporting homosexual contact and from 0 to 5% in other IDUs.

The studies, while reflecting a range of research methodologies, are subject to a number of limitations. Most of the studies did not provide detailed analyses of HIV prevalence by age and sex or behavioural factors, and several studies used sampling frames which were not clearly defined. There is little available information on temporal trends in seroprevalence and geographical comparisons are rendered difficult by differences in the study methodology. Adoption of standardized, continuing seroprevalence surveys on IDUs would provide a better means of monitoring the occurrence of HIV infection in this group, which has been a key determinant of the course of the HIV epidemic in a number of Western countries.     

Investigation of nimodipine pharmacokinetics in Chinese patients with acute subarachnoid haemorrhage

Summary Nimodipine pharmacokinetics was investigated in 12 Chinese patients with acute subarachnoid haemorrhage receiving an IV infusion of 1.6 or 2 mg/h (based on estimated body weight) for 10 days. Peripheral venous blood samples were collected for up to 4 days and plasma nimodipine was assayed by GC/ECD. The mean value was taken as the steady state concentration (C_ss) and Clearance (CL) (hourly dose/C_ss) was calculated. Eight survivors were given oral nimodipine (60 or 90 mg) every 6h (based on body weight), blood was sampled over 6 h and the plasma nimodipine level determined. The values for C_ss, CL and CL·kg^–1 were 33.5 g·l^–1, 58 l·h^–1 and 1.0 l·h^–1·kg^–1 respectively; in survivors receiving the drug orally, bioavailability of the 30 mg tablet was 9%. In one very sick patient given crushed tablets by naso-gastric tube, the AUC was very low; in vitro studies indicated that adsorption of nimodipine by the tubing was unlikely to have been the cause.The pharmacokinetic findings in Chinese patients are comparable to previously reported values in Caucasians.     

A defasciculating dose of d-tubocurarine causes resistance to succinylcholine

Forty-four patients, ASA physical status I or II, undergoing thiamylal, fentanyl, N2O/O2 anaesthesia were studied to determine the dose-response to succinylcholine (Sch) without prior defasciculation (24 pt - Group 1), or three minutes following d-tubocurarine (dTC), 0.043 mg.kg-1 (20 pt - Group 2). The individual log dose-logit response curve for each patient was determined using a cumulative dose plus infusion technique and integrated EMG monitoring of the first dorsal interosseous muscle. The mean (+/- SEM) ED50, ED90 and ED95 values for Sch in Group 1 were 0.13 +/- 0.01, 0.19 +/- 0.01 and 0.22 +- 0.01 mg.kg-1, and in Group 2 were 0.16 +/- 0.01, 0.25 +/- 0.01 and 0.29 +/- 0.02 mg.kg-1, respectively. The mean ED values in Group 2 were significantly greater than the equivalent values in Group 1 (P less than 0.05). Compared with values in Group 1, ED values in Group 2 represented mean increases of 23, 32, and 32 per cent, respectively. These pharmacodynamic data indicate that the dose of Sch needs to be increased by 32 per cent following a defasciculating dose of dTC 3 mg.70 kg-1 (0.043 mg.kg-1).