Transfer learning for cytochrome P450 isozyme selectivity prediction

In the drug discovery process, the metabolic fate of drugs is crucially important to prevent drug-drug interactions. Therefore, P450 isozyme selectivity prediction is an important task for screening drugs of appropriate metabolism profiles. Recently, large-scale activity data of five P450 isozymes (CYP1A2 CYP2C9, CYP3A4, CYP2D6, and CYP2C19) have been obtained using quantitative high-throughput screening with a bioluminescence assay. Although some isozymes share similar selectivities, conventional supervised learning algorithms independently learn a prediction model from each P450 isozyme. They are unable to exploit the other P450 isozyme activity data to improve the predictive performance of each P450 isozyme's selectivity. To address this issue, we apply transfer learning that uses activity data of the other isozymes to learn a prediction model from multiple P450 isozymes. After using the large-scale P450 isozyme selectivity dataset for five P450 isozymes, we evaluate the model's predictive performance. Experimental results show that, overall, our algorithm outperforms conventional supervised learning algorithms such as support vector machine (SVM), Weighted k-nearest neighbor classifier, Bagging, Adaboost, and latent semantic indexing (LSI). Moreover, our results show that the predictive performance of our algorithm is improved by exploiting the multiple P450 isozyme activity data in the learning process. Our algorithm can be an effective tool for P450 selectivity prediction for new chemical entities using multiple P450 isozyme activity data.     

SOMEViz: A Web Service for Site of Metabolism Estimating and Visualizing

Phase I metabolism is an important consideration in drug discovery because it profoundly affects the toxicity and activity profile of a drug candidate. In these metabolic processes, CYP450 family is responsible for the majority of biotransformation events. However, it is still an important challenge to predict sites of metabolism (SOM) of a new chemical entity due to the complex reaction mechanism and variety in CYP450 enzymes. SOMEViz is an online service designed for predicting and visualizing human cytochromes P450 (CYP450)-mediated sites of metabolism (SOM) of a molecule, on the basis of a previously reported model [1]. The service provides an access for predicting sites of metabolism of molecules with reasonable accuracy, and predicted results are shown in a user-friendly as well as interactive way, which may help chemists explore metabolism properties of chemicals in the early stage of drug discovery. The web-based GUI of SOMEViz offers user a straightforward way to manage and visualize the sites of metabolism (SOM) prediction results. The service and examples are available free of charge at http://www.dddc.ac.cn/some.     

SCYPPred: a web-based predictor of SNPs for human cytochrome P450

Human cytochrome P450(CYP 450) enzymes mediate over 60% of the phase I-dependent metabolism of clinical drugs. They are also known for the polymorphism functions that have significant impacts on the enzyme activities. In this study, a web-server called SCYPPred was developed for predicting human cytochrome P450 SNPs (Single Nucleotide Polymorphisms) based on the SVM flanking sequence method; SCYPPred can rapidly yield the desired results by using the amino acid sequences information alone. The web-server is accessible to the public at http://snppred.sjtu.edu.cn. Hopefully SCYPPred could be a useful bioinformatics tool for elucidating the mutation probability of a specific CYP450 enzyme.     

Metabolic activity and mRNA levels of human cardiac CYP450s involved in drug metabolism

Background

Tissue-specific expression of CYP450s can regulate the intracellular concentration of drugs and explain inter-subject variability in drug action. The overall objective of our study was to determine in a large cohort of samples, mRNA levels and CYP450 activity expressed in the human heart.

Methodology

CYP450 mRNA levels were determined by RTPCR in left ventricular samples (n = 68) of explanted hearts from patients with end-stage heart failure. Samples were obtained from ischemic and non-ischemic hearts. In some instances (n = 7), samples were available from both the left and right ventricles. A technique for the preparation of microsomes from human heart tissue was developed and CYP450-dependent activity was determined using verapamil enantiomers as probe-drug substrates.

Principal Findings

Our results show that CYP2J2 mRNA was the most abundant isoform in all human heart left ventricular samples tested. Other CYP450 mRNAs of importance were CYP4A11, CYP2E1, CYP1A1 and CYP2C8 mRNAs while CYP2B6 and CYP2C9 mRNAs were present at low levels in only some of the hearts analyzed. CYP450 mRNAs did not differ between ischemic and non-ischemic hearts and appeared to be present at similar levels in the left and right ventricles. Incubation of verapamil with heart microsomes led to the formation of nine CYP450-dependent metabolites: a major finding was the observation that stereoselectivity was reversed compared to human liver microsomes, in which the R-enantiomer is metabolized to a greater extent.

Conclusions

This study determined cardiac mRNA levels of various CYP450 isozymes involved in drug metabolism and demonstrated the prevalent expression of CYP2J2 mRNA. It revealed that cardiomyocytes can efficiently metabolize drugs and that cardiac CYP450s are highly relevant with regard to clearance of drugs in the heart. Our results support the claim that drug metabolism in the vicinity of a drug effector site can modulate drug effects.     

Reduction in hepatic drug metabolizing CYP3A4 activities caused by P450 oxidoreductase mutations identified in patients with disordered steroid metabolism

Cytochrome P450 3A4 (CYP3A4), the major P450 present in human liver metabolizes approximately half the drugs in clinical use and requires electrons supplied from NADPH through NADPH-P450 reductase (POR, CPR). Mutations in human POR cause a rare form of congenital adrenal hyperplasia from diminished activities of steroid metabolizing P450s. In this study we examined the effect of mutations in POR on CYP3A4 activity. We used purified preparations of wild type and mutant human POR and in vitro reconstitution with purified CYP3A4 to perform kinetic studies. We are reporting that mutations in POR identified in patients with disordered steroidogenesis/Antley-Bixler syndrome (ABS) may reduce CYP3A4 activity, potentially affecting drug metabolism in individuals carrying mutant POR alleles. POR mutants Y181D, A457H, Y459H, V492E and R616X had more than 99% loss of CYP3A4 activity, while POR mutations A287P, C569Y and V608F lost 60-85% activity. Loss of CYP3A4 activity may result in increased risk of drug toxicities and adverse drug reactions in patients with POR mutations.     

Machine learning-driven identification of drugs inhibiting cytochrome P450 2C9

Cytochrome P450 2C9 (CYP2C9) is a major drug-metabolizing enzyme that represents 20% of the hepatic CYPs and is responsible for the metabolism of 15% of drugs. A general concern in drug discovery is to avoid the inhibition of CYP leading to toxic drug accumulation and adverse drug–drug interactions. However, the prediction of CYP inhibition remains challenging due to its complexity. We developed an original machine learning approach for the prediction of drug-like molecules inhibiting CYP2C9. We created new predictive models by integrating CYP2C9 protein structure and dynamics knowledge, an original selection of physicochemical properties of CYP2C9 inhibitors, and machine learning modeling. We tested the machine learning models on publicly available data and demonstrated that our models successfully predicted CYP2C9 inhibitors with an accuracy, sensitivity and specificity of approximately 80%. We experimentally validated the developed approach and provided the first identification of the drugs vatalanib, piriqualone, ticagrelor and cloperidone as strong inhibitors of CYP2C9 with IC values <18 μM and sertindole, asapiprant, duvelisib and dasatinib as moderate inhibitors with IC50 values between 40 and 85 μM. Vatalanib was identified as the strongest inhibitor with an IC50 value of 0.067 μM. Metabolism assays allowed the characterization of specific metabolites of abemaciclib, cloperidone, vatalanib and tarafenacin produced by CYP2C9. The obtained results demonstrate that such a strategy could improve the prediction of drug-drug interactions in clinical practice and could be utilized to prioritize drug candidates in drug discovery pipelines.     

An evaluation of the P450 inhibition and induction potential of daptomycin in primary human hepatocytes

Two in vitro studies assessed the potential of daptomycin (Cubicin), a newly marketed antibiotic, to affect the cytochrome P450 (CYP450) isoforms in primary cultured human hepatocytes. Both induction and inhibition of isoforms 1A2, 2A6, 2C9, 2C19, 2D6, 2E1, and 3A4 were evaluated. The highest concentrations of daptomycin used in both the induction and inhibition assays were approximately eight-fold higher than the peak total drug concentration (50-60 microg/mL), or the peak free drug concentration (estimated 5-6 microg/mL), in plasma at the clinical dose regimen of 4 mg/kg qd. Results in primary human hepatocytes indicate that daptomycin, at concentrations up to 400 microg total drug/mL, demonstrated no biologically significant induction of any of the CYP450 isoform activities in comparison with the negative control or known inducers. At daptomycin concentrations up to 40 microg free drug/mL, no biologically significant inhibition of the activities of these CYP450 isoforms was observed as compared with known inhibitors. The human hepatocyte results demonstrate that daptomycin has no effects on hepatic CYP450-mediated drug metabolism and, therefore, suggest that daptomycin is unlikely to show potential for pharmacokinetic interactions with concomitantly administered drugs that are metabolized by CYP450 isoforms.     

Structures of cytochrome P450 3A4

Cytochrome P450 3A4 (CYP3A4) catalyzes the initial step in the clearance of many pharmaceuticals and foreign chemicals. The structurally diverse nature of CYP3A4 substrates complicates rational prediction of their metabolism and identification of potential drug interactions. The first molecular structures of human CYP3A4 were recently determined, revealing an active site of sufficient size and topography to accommodate either large ligands or multiple smaller ligands, as suggested by the heterotropic and homotropic cooperativity of the enzyme.     

CYP3A4 and pregnane X receptor humanized mice

Marked species differences exist in P450 expression and activities. In order to produce mouse models that can be used to more accurately predict human drug and carcinogen metabolism, P450- and xenobiotic receptor humanized mice are being prepared using bacterial artificial chromosomes (BAC) and P1 phage artificial chromosomes (PAC) genomic clones. In some cases, transgenic mice carrying the human genes are bred with null-mice to produce fully humanized mice. Mice expressing human CYP1A1, CYP1A2, CYP2E1, CYP2D6, CYP3A4, and CYP3A7 were generated and characterized. Studies with the CYP3A4-humanized (hCYP3A4) mouse line revealed new information on the physiological function of this P450 and its role in drug metabolism in vivo. With this mouse line, CYP3A4, under certain circumstances, was found to alter the serum levels of estrogen resulting in deficient lactation and low pup survival as a result of underdeveloped mammary glands. This hCYP3A4 mouse established the importance of intestinal CYP3A4 in the pharmacokinetics of orally administered drugs. The hCYP3A4 mice were also used to establish the mechanisms of potential gender differences in CYP3A4 expression (adult female > adult male) that could account for human gender differences in drug metabolism and response. The pregnane X receptor (PXR) is also involved in induction of drug metabolism through its target genes including CYP3A4. Since species differences exist in ligand specificity between human and mice, a PXR-humanized mouse (hPXR) was produced that responds to human PXR activators such as rifampicin but does not respond to the rodent activator pregnenalone 16alpha-carbonitrile.     

Comparative human-horse sequence analysis of the CYP3A subfamily gene cluster

Cytochrome P450 enzymes (CYP450s) represent a superfamily of haem-thiolate proteins. CYP450s are most abundant in the liver, a major site of drug metabolism, and play key roles in the metabolism of a variety of substrates, including drugs and environmental contaminants. Interaction of two or more different drugs with the same enzyme can account for adverse effects and failure of therapy. Human CYP3A4 metabolizes about 50% of all known drugs, but little is known about the orthologous CYP450s in horses. We report here the genomic organization of the equine CYP3A gene cluster as well as a comparative analysis with the human CYP3A gene cluster. The equine CYP450 genes of the 3A family are located on ECA 13 between 6.97-7.53 Mb, in a region syntenic to HSA 7 99.05-99.35 Mb. Seven potential, closely linked equine CYP3A genes were found, in contrast to only four genes in the human genome. RNA was isolated from an equine liver sample, and the approximately 1.5-kb coding sequence of six CYP3A genes could be amplified by RT-PCR. Sequencing of the RT-PCR products revealed numerous hitherto unknown single nucleotide polymorphisms (SNPs) in these six CYP3A genes, and one 6-bp deletion compared to the reference sequence (EquCab2.0). The presence of the variants was confirmed in a sample of genomic DNA from the same horse. In conclusion, orthologous genes for the CYP3A family exist in horses, but their number differs from those of the human CYP3A gene family. CYP450 genes of the same family show high homology within and between mammalian species, but can be highly polymorphic.     

Partition dataset according to amino acid type improves the prediction of deleterious non-synonymous SNPs

Many non-synonymous SNPs (nsSNPs) are associated with diseases, and numerous machine learning methods have been applied to train classifiers for sorting disease-associated nsSNPs from neutral ones. The continuously accumulated nsSNP data allows us to further explore better prediction approaches. In this work, we partitioned the training data into 20 subsets according to either original or substituted amino acid type at the nsSNP site. Using support vector machine (SVM), training classification models on each subset resulted in an overall accuracy of 76.3% or 74.9% depending on the two different partition criteria, while training on the whole dataset obtained an accuracy of only 72.6%. Moreover, the dataset was also randomly divided into 20 subsets, but the corresponding accuracy was only 73.2%. Our results demonstrated that partitioning the whole training dataset into subsets properly, i.e., according to the residue type at the nsSNP site, will improve the performance of the trained classifiers significantly, which should be valuable in developing better tools for predicting the disease-association of nsSNPs.     

The roles of different porcine cytochrome P450 enzymes and cytochrome b5A in skatole metabolism

Boar taint is the unfavourable odour and taste from pork fat, which results in part from the accumulation of skatole (3-methylindole, 3MI). The key enzymes in skatole metabolism are thought to be cytochrome P450 2E1 (CYP2E1) and cytochrome 2A (CYP2A); however, the cytochrome P450 (CYP450) isoform responsible for the production of the metabolite 6-hydroxy-3-methylindole (6-OH-3MI, 6-hydroxyskatole), which is thought to be involved in the clearance of skatole, has not been established conclusively. The aim of this study was to characterize the role of porcine CYP450s in skatole metabolism by expressing them individually in the human embryonic kidney HEK293-FT cell line. This system eliminates the problems of the lack of specificity of antibodies, inhibitors and substrates for CYP450 isoforms in the pig, and contributions of any other CYP450s that would be present. The results show that pig CYP1A1, CYP2A19, CYP2C33v4, CYP2C49, CYP2E1 and CYP3A and human CYP2E1 (hCYP2E1) are all capable of producing the major skatole metabolite 3-methyloxyindole (3MOI), as well as indole-3-carbinol (I3C), 5-hydroxy-3-methylindole (5-OH-3MI), 6-OH-3MI, 2-aminoacetophenone (2AAP) and 3-hydroxy-3-methyloxindole. CYP2A19 produced the highest amount of the physiologically important metabolite 6-OH-3MI, followed by porcine CYP2E1 and CYP2C49; CYP2A19 also produced more 6-OH-3MI than hCYP2E1. Co-transfection with CYB5A increased the production of skatole metabolites by some of the CYP450s, suggesting that CYB5A plays an important role in the metabolism of skatole. We also show the utility of this expression system to check the specificity of selected substrates and antibodies for porcine CYP450s. Further information regarding the abundance of different CYP450 isoforms is required to fully understand their contribution to skatole metabolism in vivo in the pig.     

Modelling human cytochromes P450 involved in drug metabolism from the CYP2C5 crystallographic template

A historical background to homology modelling of human P450s involved in drug metabolism is outlined, showing that the progress in crystallographic studies of bacterial forms of enzyme and, latterly, determination of a mammalian P450 crystal structure, has enabled the production of increasingly satisfactory models of human P450 enzymes. The methodology for the generation of P450 models by homology with crystallographic template structures is summarized, and recent results of CYP2C5-constructed models of P450s are described. These indicate that selective substrates are able to fit within the putative active sites of each enzyme, where key contacts with complementary amino acid residues are largely consistent with the results of site-directed mutagenesis experiments and metabolic studies. Consequently, the CYP2C5 crystal structure can be regarded at the current paradigm for homology modelling of the drug metabolizing P450s, especially those from the CYP2 family.     

Docking studies of piperine - iron conjugate with human CYP450 3A4

Piperine, a major constituent of Piper nigrum (Black pepper), is one of the well known components in many Ayurvedic formulations. Piperine is most studied bioenhancer because it inhibits drug metabolizing enzymes in rodents and increases plasma concentrations of several drugs, including P-glycoprotein substrates. However, there areno evidences on piperine-iron conjugate to inhibit human CYP450 3A4. We therefore investigated the influence of piperine-Fe conjugate to study the metabolism of iron with CYP450 3A4. Our in silico results showed that Piperine when conjugated with iron, inhibited activity of CYP450 3A4. This improved the binding of piperine-Fe conjugate with CYP450 3A4 and increased bioavailability.     

Targeted label-free approach for quantification of epoxide hydrolase and glutathione transferases in microsomes

The aim of this study was to investigate the expression and organ distribution of cytochrome P450 (CYP450) enzymes, microsomal epoxide hydrolase (MEH), and microsomal glutathione-S-transferase (MGST 1, 2, 3) in human liver, lung, intestinal, and kidney microsomes by targeted peptide-based quantification using nano liquid chromatography–tandem multiple reaction monitoring (nano LC-MRM). Applying this method, we analyzed 16 human liver microsomes and pooled lung, kidney, and intestine microsomes. Nine of the CYP450s (CYP1A2, 2B6, 2C8, 2C9, 2C19, 2D6, 2E1, 3A4, 3A5) could be quantified in liver. Except for CYP3A4 and 3A5 existing in intestine, other CYP450s had little content (<0.1 pmol/mg protein) in extrahepatic tissues. MEH and MGSTs could be quantified both in hepatic and in extrahepatic tissues. The highest concentrations of MEH and MGST 1, 2 were found in liver; conversely MGST 3 was abundant in human kidney and intestine compared to liver. The targeted proteomics assay described here can be broadly and efficiently utilized as a tool for investigating the targeted proteins. The method also provides novel CYP450s, MEH, and MGSTs expression data in human hepatic and extrahepatic tissues that will benefit rational approaches to evaluate metabolism in drug development.     

Cyclin-dependent kinase 2 negatively regulates human pregnane X receptor-mediated CYP3A4 gene expression in HepG2 liver carcinoma cells

The human pregnane X receptor (hPXR) regulates the expression of critical drug metabolism enzymes. One of such enzymes, cytochrome P450 3A4 (CYP3A4), plays critical roles in drug metabolism in hepatocytes that are either quiescent or passing through the cell cycle. It has been well established that the expression of P450, such as CYP3A4, is markedly reduced during liver development or regeneration. Numerous studies have implicated cellular signaling pathways in modulating the functions of nuclear receptors, including hPXR. Here we report that inhibition of cyclin-dependent kinases (Cdks) by kenpaullone and roscovitine (two small molecule inhibitors of Cdks that we identified in a screen for compounds that activate hPXR) leads to activation of hPXR-mediated CYP3A4 gene expression in HepG2 human liver carcinoma cells. Consistent with this finding, activation of Cdk2 attenuates the activation of CYP3A4 gene expression. In vitro kinase assays revealed that Cdk2 directly phosphorylates hPXR. A phosphomimetic mutation of a putative Cdk phosphorylation site, Ser(350), significantly impairs the function of hPXR, whereas a phosphorylation-deficient mutation confers resistance to Cdk2. Using HepG2 that has been stably transfected with hPXR and the CYP3A4-luciferase reporter, enriched in different phases of the cell cycle, we found that hPXR-mediated CYP3A4 expression is greatly reduced in the S phase. Our results indicate for the first time that Cdk2 negatively regulates the activity of hPXR, and suggest an important role for Cdk2 in regulating hPXR activity and CYP3A4 expression in hepatocytes passing through the cell cycle, such as those in fetal or regenerating adult liver.     

Prediction of the phenotypic effects of non-synonymous single nucleotide polymorphisms using structural and evolutionary information

MOTIVATION: There has been great expectation that the knowledge of an individual's genotype will provide a basis for assessing susceptibility to diseases and designing individualized therapy. Non-synonymous single nucleotide polymorphisms (nsSNPs) that lead to an amino acid change in the protein product are of particular interest because they account for nearly half of the known genetic variations related to human inherited diseases. To facilitate the identification of disease-associated nsSNPs from a large number of neutral nsSNPs, it is important to develop computational tools to predict the phenotypic effects of nsSNPs. RESULTS: We prepared a training set based on the variant phenotypic annotation of the Swiss-Prot database and focused our analysis on nsSNPs having homologous 3D structures. Structural environment parameters derived from the 3D homologous structure as well as evolutionary information derived from the multiple sequence alignment were used as predictors. Two machine learning methods, support vector machine and random forest, were trained and evaluated. We compared the performance of our method with that of the SIFT algorithm, which is one of the best predictive methods to date. An unbiased evaluation study shows that for nsSNPs with sufficient evolutionary information (with not <10 homologous sequences), the performance of our method is comparable with the SIFT algorithm, while for nsSNPs with insufficient evolutionary information (<10 homologous sequences), our method outperforms the SIFT algorithm significantly. These findings indicate that incorporating structural information is critical to achieving good prediction accuracy when sufficient evolutionary information is not available. AVAILABILITY: The codes and curated dataset are available at http://compbio.utmem.edu/snp/dataset/     

Prediction of binding modes for ligands in the cytochromes P450 and other heme-containing proteins

The cytochromes P450 (P450s) are a family of heme-containing monooxygenase enzymes involved in a variety of functions, including the metabolism of endogenous and exogenous substances in the human body. During lead optimization, and in drug development, many potential drug candidates are rejected because of the affinity they display for drug-metabolising P450s. Recently, crystal structures of human enzymes involved in drug metabolism have been determined, significantly augmenting the prospect of using structure-based design to modulate the binding and metabolizing properties of compounds against P450 proteins. An important step in the application of structure-based metabolic optimization is the accurate prediction of docking modes in heme binding proteins. In this paper we assess the performance of the docking program GOLD at predicting the binding mode of 45 heme-containing complexes. We achieved success rates of 64% and 57% for Chemscore and Goldscore respectively; these success rates are significantly lower than the value of 79% observed with both scoring functions for the full GOLD validation set. Re-parameterization of metal-acceptor interactions and lipophilicity of planar nitrogen atoms in the scoring functions resulted in a significant increase in the percentage of successful dockings against the heme binding proteins (Chemscore 73%, Goldscore 65%). The modified scoring functions will be useful in docking applications on P450 enzymes and other heme binding proteins.