Distinct clinical phenotypes for Crohn's disease derived from patient surveys.

BACKGROUND: Defining clinical phenotypes provides opportunities for new diagnostics and may provide insights into early intervention and disease prevention. There is increasing evidence that patient-derived health data may contain information that complements traditional methods of clinical phenotyping. The utility of these data for defining meaningful phenotypic groups is of great interest because social media and online resources make it possible to query large cohorts of patients with health conditions. METHODS: We evaluated the degree to which patient-reported categorical data is useful for discovering subclinical phenotypes and evaluated its utility for discovering new measures of disease severity, treatment response and genetic architecture. Specifically, we examined the responses of 1961 patients with inflammatory bowel disease to questionnaires in search of sub-phenotypes. We applied machine learning methods to identify novel subtypes of Crohn's disease and studied their associations with drug responses. RESULTS: Using the patients' self-reported information, we identified two subpopulations of Crohn's disease; these subpopulations differ in disease severity, associations with smoking, and genetic transmission patterns. We also identified distinct features of drug response for the two Crohn's disease subtypes. These subtypes show a trend towards differential genotype signatures. CONCLUSION: Our findings suggest that patient-defined data can have unplanned utility for defining disease subtypes and may be useful for guiding treatment approaches.

Mucosal Gene Expression in Pediatric and Adult Patients With Ulcerative Colitis Permits Modeling of Ideal Biopsy Collection Strategy for Transcriptomic Analysis.

Background: Transcriptional profiling has been performed on biopsies from ulcerative colitis patients. Limitations in prior studies include the variability introduced by inflammation, anatomic site of biopsy, extent of disease, and medications. We sought to more globally understand the variability of gene expression from patients with ulcerative colitis to advance our understanding of its pathogenesis and to guide clinical study design. Methods: We performed transcriptional profiling on 13 subjects, including pediatric and adult patients from 2 hospital sites. For each patient, we collected 6 biopsies from macroscopically inflamed tissue and 4 biopsies from macroscopically healthy-appearing tissue. Isolated RNA was used for microarray gene expression analysis utilizing Affymetrix Human Primeview microarrays. Ingenuity pathway analysis was used to assess over-representation of gene ontology and biological pathways. RNAseq was also performed, and differential analysis was assessed to compare affected vs unaffected samples. Finally, we modeled the minimum number of biopsies required to reliably detect gene expression across different subject numbers. Results: Transcriptional profiles co-clustered independently of the hospital collection site, patient age, sex, and colonic location, which parallels prior gene expression findings. A small set of genes not previously described was identified. Our modeling analysis reveals the number of biopsies and patients per cohort to yield reliable results in clinical studies. Conclusions: Key findings include concordance, including some expansion, of previously published gene expression studies and similarity among different age groups. We also established a reliable statistical model for biopsy collection for future clinical studies.

Mendelian Disease Associations Reveal Novel Insights into Inflammatory Bowel Disease.

Background: Monogenic diseases have been shown to contribute to complex disease risk and may hold new insights into the underlying biological mechanism of Inflammatory Bowel Disease (IBD). Methods: We analyzed Mendelian disease associations with IBD using over 55 million patients from the Optum's deidentified electronic health records dataset database. Using the significant Mendelian diseases, we performed pathway enrichment analysis and constructed a model using gene expression datasets to differentiate Crohn's disease (CD), ulcerative colitis (UC), and healthy patient samples. Results: We found 50 Mendelian diseases were significantly associated with IBD, with 40 being significantly associated with both CD and UC. Our results for CD replicated those from previous studies. Pathways that were enriched consisted of mainly immune and metabolic processes with a focus on tolerance and oxidative stress. Our 3-way classifier for UC, CD, and healthy samples yielded an accuracy of 72%. Conclusions: Mendelian diseases that are significantly associated with IBD may reveal novel insights into the genetic architecture of IBD.

A probabilistic pathway score (PROPS) for classification with applications to inflammatory bowel disease.

Summary: Gene-based supervised machine learning classification models have been widely used to differentiate disease states, predict disease progression and determine effective treatment options. However, many of these classifiers are sensitive to noise and frequently do not replicate in external validation sets. For complex, heterogeneous diseases, these classifiers are further limited by being unable to capture varying combinations of genes that lead to the same phenotype. Pathway-based classification can overcome these challenges by using robust, aggregate features to represent biological mechanisms. In this work, we developed a novel pathway-based approach, PRObabilistic Pathway Score, which uses genes to calculate individualized pathway scores for classification. Unlike previous individualized pathway-based classification methods that use gene sets, we incorporate gene interactions using probabilistic graphical models to more accurately represent the underlying biology and achieve better performance. We apply our method to differentiate two similar complex diseases, ulcerative colitis (UC) and Crohn's disease (CD), which are the two main types of inflammatory bowel disease (IBD). Using five IBD datasets, we compare our method against four gene-based and four alternative pathway-based classifiers in distinguishing CD from UC. We demonstrate superior classification performance and provide biological insight into the top pathways separating CD from UC. Availability and Implementation: PROPS is available as a R package, which can be downloaded at http://simtk.org/home/props or on Bioconductor. Contact: rbaltman@stanford.edu. Supplementary information: Supplementary data are available at Bioinformatics online.

Integrated Strategy for Use of Positron Emission Tomography in Nonhuman Primates to Confirm Multitarget Occupancy of Novel Psychotropic Drugs: An Example with AZD3676.

Positron emission tomography (PET) is widely applied in central nervous system (CNS) drug development for assessment of target engagement in vivo. As the majority of PET investigations have addressed drug interaction at a single binding site, findings of multitarget engagement have been less frequently reported and have often been inconsistent with results obtained in vitro. AZD3676 [N,N-dimethyl-7-(4-(2-(pyridin-2-yl)ethyl)piperazin-1-yl) benzofuran-2-carboxamide] is a novel combined serotonin (5-hydroxytryptamine) 5-HT1A and 5-HT1B receptor antagonist that was developed for the treatment of cognitive impairment in Alzheimer's disease. Here, we evaluated the properties of AZD3676 as a CNS drug by combining in vitro and ex vivo radioligand binding techniques, behavioral pharmacology in rodents, and PET imaging in nonhuman primates. Target engagement in the nonhuman primate brain was assessed in PET studies by determination of drug-induced occupancy using receptor-selective radioligands. AZD3676 showed preclinical properties consistent with CNS drug potential, including nanomolar receptor affinity and efficacy in rodent models of learning and memory. In PET studies of the monkey brain, AZD3676 inhibited radioligand binding in a dose-dependent manner with similar affinity at both receptors. The equally high affinity at 5-HT1A and 5-HT1B receptors as determined in vivo was not predicted from corresponding estimates obtained in vitro, suggesting more than 10-fold selectivity for 5-HT1A versus 5-HT1B receptors. These findings support the further integrated use of PET for confirmation of multitarget occupancy of CNS drugs. Importantly, earlier introduction of PET studies in nonhuman primates may reduce future development costs and the requirement for animal experiments in preclinical CNS drug development programs.

Cytochrome p450 inhibitory properties of common efflux transporter inhibitors.

Drug transporter inhibitors are important tools to elucidate the contribution of transporters to drug disposition both in vitro and in vivo. These inhibitors are often unselective and affect several transporters as well as drug metabolizing enzymes, which can make experimental results difficult to interpret with confidence. We therefore tested 14 commonly used P-glycoprotein (P-gp), breast cancer resistance protein (BCRP), and multidrug-resistance associated protein (MRP) inhibitors as inhibitors of cytochrome P450 (P450) enzyme activities using recombinant enzymes. A subset of P-gp and/or CYP3A inhibitors were selected (cyclosporin A, elacridar, ketoconazole, quinidine, reserpine, and tacrolimus) for a comparison of P450 inhibition in human microsomes and hepatocytes. Most P-gp inhibitors showed CYP3A4 inhibition, with potencies often in a similar range as their P-gp inhibition, as well as less potent CYP2C19 inhibition. Other P450 enzymes were not strongly inhibited except a few cases of CYP2D6 inhibition. MRP and BCRP inhibitors showed limited P450 inhibition. Some inhibitors showed less P450 inhibition in human hepatocytes than human liver microsomes, for example, elacridar, probably due to differences in binding, permeability limitations, or active, P-gp mediated efflux of the inhibitor from the hepatocytes. Quinidine was a potent P450 inhibitor in hepatocytes but only showed weak inhibition in microsomes. Quinidine shows an extensive cellular uptake, which may potentiate intracellular P450 inhibition. Elacridar, described as a potent and selective P-gp inhibitor, displayed modest P450 inhibition in this study and is thus a useful model inhibitor to define the role of P-gp in drug disposition without interference with other processes.

Prediction of in vivo rat biliary drug clearance from an in vitro hepatocyte efflux model.

Well-established techniques are available to predict in vivo hepatic uptake and metabolism from in vitro data, but predictive models for biliary clearance remain elusive. Several studies have verified the expression and activity of ATP-binding cassette (ABC) efflux transporters central to biliary clearance in freshly isolated rat hepatocytes, raising the possibility of predicting biliary clearance from in vitro efflux measurements. In the present study, short-term plated rat hepatocytes were evaluated as a model to predict biliary clearance from in vitro efflux measurements before major changes in transporter expression known to take place in long-term hepatocyte cultures. The short-term cultures were carefully characterized for their uptake and metabolic properties using a set of model compounds. In vitro efflux was studied using digoxin, fexofenadine, napsagatran, and rosuvastatin, representing compounds with over 100-fold differences in efflux rates in vitro and 60-fold difference in measured in vivo biliary clearance. The predicted biliary clearances from short-term plated rat hepatocytes were within 2-fold of measured in vivo values. As in vitro efflux includes both basolateral and canalicular effluxes, pronounced basolateral efflux may introduce errors in predictions for some compounds. In addition, in vitro rat hepatocyte uptake rates corrected for simultaneous efflux predicted rat in vivo hepatic clearance of the biliary cleared compounds with less than 2-fold error. Short-term plated hepatocytes could thus be used to quantify hepatocyte uptake, metabolism, and efflux of compounds and considerably improve the prediction of hepatic clearance, especially for compounds with a large biliary clearance component.

Functional ATP-binding cassette drug efflux transporters in isolated human and rat hepatocytes significantly affect assessment of drug disposition.

Freshly isolated hepatocytes are considered the gold standard for in vitro studies of hepatic drug disposition. To ensure a reliable supply of cells, cryopreserved human hepatocytes are often used. ABC-superfamily drug efflux transporters are key elements in hepatic drug disposition. These transporters are often considered lost after isolation of hepatocytes. In the present study, the expression and activity of ABC transporters BCRP, BSEP, P-gp, MRP2, MRP3, and MRP4 in human and rat cryopreserved hepatocytes were investigated. In commercially available human cryopreserved hepatocytes, all drug efflux transporters except human BCRP (hBCRP) exhibited similar expression levels as in fresh liver biopsies. Expression levels of hBCRP were 60% lower in cryopreserved human hepatocytes than in liver tissue, which could lead to, at most, a 2.5-fold reduction in hBCRP-mediated efflux. Fresh rat hepatocytes showed significantly lower levels of rat BCRP compared with liver expression levels; expression levels of other ABC transporters were unchanged. ABC transporters in human cryopreserved cells were localized to the plasma membrane. Functional studies could demonstrate P-gp and BCRP activity in both human cryopreserved and fresh rat hepatocytes. Inhibiting P-gp-mediated efflux by elacridar in in vitro experiments significantly decreased fexofenadine efflux from hepatocytes, resulting in an increase in apparent fexofenadine uptake. The results from the present study clearly indicate that ABC transporter-mediated efflux in freshly isolated as well as cryopreserved rat and human hepatocytes should be taken into account in in vitro experiments used for modeling of drug metabolism and disposition.

The impact of solute carrier (SLC) drug uptake transporter loss in human and rat cryopreserved hepatocytes on clearance predictions.

Cryopreserved hepatocytes are often used as a convenient tool in studies of hepatic drug metabolism and disposition. In this study, the expression and activity of drug transporters in human and rat fresh and cryopreserved hepatocytes was investigated. In human cryopreserved hepatocytes, Western blot analysis indicated that protein expression of the drug uptake transporters [human Na(+)-taurocholate cotransporting polypeptide (NTCP), human organic anion transporting polypeptides (OATPs), human organic anion transporters, and human organic cation transporters (OCTs)] was considerably reduced compared with liver tissue. In rat cryopreserved cells, the same trend was observed but to a lesser extent. Several rat transporters were reduced as a result of both isolation and cryopreservation procedures. Immunofluorescence showed that a large portion of remaining human OATP1B1 and OATP1B3 transporters were internalized in human cryopreserved hepatocytes. Measuring uptake activity using known substrates of OATPs, OCTs, and NTCP showed decreased activity in cryopreserved as compared with fresh hepatocytes in both species. The reduced uptake in cryopreserved hepatocytes limited the in vitro metabolism of several AstraZeneca compounds. A retrospective analysis of clearance predictions of AstraZeneca compounds suggested systematic lower clearance predicted using metabolic stability data from human cryopreserved hepatocytes compared with human liver microsomes. This observation is consistent with a loss of drug uptake transporters in cryopreserved hepatocytes. In contrast, the predicted metabolic clearance from fresh rat hepatocytes was consistently higher than those predicted from liver microsomes, consistent with retention of uptake transporters. The uptake transporters, which are decreased in cryopreserved hepatocytes, may be rate-limiting for the metabolism of the compounds and thus be one explanation for underpredictions of in vivo metabolic clearance from cryopreserved hepatocytes.

Practical use of the regression offset approach for the prediction of in vivo intrinsic clearance from hepatocytes.

Systematic under-prediction of clearance is frequently associated with in vitro kinetic data when extrapolated using physiological scaling factors, appropriate binding parameters and the well-stirred model. The present study describes a method of removing this systematic bias through application of empirical correction factors derived from regression analyses applied to the in vitro and in vivo data for a defined set of reference compounds. Linear regression lines were established with in vivo intrinsic clearance (CLint), derived from in vivo clearance data and scaled in vitro intrinsic clearance from isolated hepatocyte incubations. The scaled CLint was empirically corrected to a predicted in vivo CLint using the slope and intercept from a uniform weighted linear regression applied to the in vitro to in vivo extrapolation. Cross validation of human data demonstrated that 66% of the reference compounds had a predicted in vivo CLint within two-fold of the observed value. The average absolute fold error (AAFE) for the in vivo CLint predictions was 1.90. For rat, 54% of the compounds had a predicted value within two-fold of the observed and the AAFE was 1.98. Three AstraZeneca projects are used to exemplify how a two-sided prediction interval, applied to the rat regression corrected reference data, can form the basis for assessing the likelihood that, for a given chemical series, the in vitro kinetic data is predictive of in vivo clearance and is therefore appropriate to guide optimisation of compound metabolic stability.

State-of-the-art tools for computational site of metabolism predictions: comparative analysis, mechanistical insights, and future applications.

In drug design, it is crucial to have reliable information on how a chemical entity behaves in the presence of metabolizing enzymes. This requires substantial experimental efforts. Consequently, being able to predict the likely site/s of metabolism in any compound, synthesized or virtual, would be highly beneficial and time efficient. In this work, six different methodologies for predictions of the site of metabolism (SOM) have been compared and validated using structurally diverse data sets of drug-like molecules with well-established metabolic pattern in CYP3A4, CYP2C9, or both. Three of the methods predict the SOM based on the ligand's chemical structure, two additional methods use structural information of the enzymes, and the sixth method combines structure and ligand similarity and reactivity. The SOM is correctly predicted in 50 to 90% of the cases, depending on method and enzyme, which is an encouraging rate. We also discuss the underlying mechanisms of cytochrome P450 metabolism in the light of the results from this comparison.

Comparison of methods for the prediction of the metabolic sites for CYP3A4-mediated metabolic reactions.

Predictions of the metabolic sites for new chemical entities, synthesized or only virtual, are important in the early phase of drug discovery to guide chemistry efforts in the synthesis of new compounds with reduced metabolic liability. This information can now be obtained from in silico predictions, and therefore, a thorough and unbiased evaluation of the computational techniques available is needed. Several computational methods to predict the metabolic hot spots are emerging. In this study, metabolite identification using MetaSite and a docking methodology, GLUE, were compared. Moreover, the published CYP3A4 crystal structure and computed CYP3A4 homology models were compared for their usefulness in predicting metabolic sites. A total of 227 known CYP3A4 substrates reported to have one or more metabolites adding up to 325 metabolic pathways were analyzed. Distance-based fingerprints and four-point pharmacophore derived from GRID molecular interaction fields were used to characterize the substrate and protein in MetaSite and the docking methodology, respectively. The CYP3A4 crystal structure and homology model with the reactivity factor enabled achieved a similar prediction success (78%) using the MetaSite method. The docking method had a relatively lower prediction success (approximately 57% for the homology model), although it still may provide useful insights for interactions between ligand and protein, especially for uncommon reactions. The MetaSite methodology is automated, rapid, and has relatively accurate predictions compared with the docking methodology used in this study.

Novel aspects of drug metabolism and transport.

The 16th International Symposium on Microsomes and Drug Oxidations (MDO2006) in Budapest, Hungary, had almost 400 attendees and was nicely organized by Laszlo Vereczkey and colleagues. The meeting had a very high standard in the field of drug metabolism, drug transport and related areas and in addition, the social events were much appreciated. At the meeting 70 invited lectures were presented in plenary sessions and in three parallel symposia sessions, and about 178 posters were shown, among them 26 posters in the young investigators workshop. The review herein is given of a majority (57) of the lectures presented at the Symposium.

Structural analysis of CYP2C9 and CYP2C5 and an evaluation of commonly used molecular modeling techniques.

This work had two separate aims: to evaluate different modeling techniques and to make a detailed structural characterization of CYP2C9. To achieve these goals, the consensus principal component analysis (CPCA) technique and distance measurements were used to explore available crystal structures, newly built homology models, and repeated molecular dynamics simulations. The CPCA was based on molecular interaction fields focused on the active site regions of the proteins and include detailed amino acid analysis. The comparison of the CYP2C9 and CYP2C5 crystal structures revealed differences in the flexible regions such as the B-C and F-G loop and the N and C termini. Cross homology models of CYP2C9 and CYP2C5, using their respective crystal structures as templates, indicated that such models were more similar to their templates than to their target proteins. Inclusion of multiple templates slightly improved the similarity to the crystal target in some cases and could be recommended even though it requires a careful manual alignment process. The application of molecular dynamics simulations to highly flexible proteins such as cytochromes P450 is also explored and the information is extracted by the CPCA. Advantages and drawbacks are presented for the different modeling techniques. Despite the varying modeling success, the models give insight and understanding by the mutual forming and discarding of hypotheses. This is a dynamic process since the crystal structures are improving with time and, therefore, the answers to the models are also changing accordingly.

Conformer- and alignment-independent model for predicting structurally diverse competitive CYP2C9 inhibitors.

A conformer- and alignment-independent three-dimensional structure-activity relationship (3D-QSAR) model has been derived that is based on flexible molecular interaction fields calculated in GRID and the subsequent description of these fields by use of alignment-independent descriptors derived in ALMOND. The training set consisted of 22 diverse and flexible competitive inhibitors of the drug-metabolizing enzyme CYP2C9 and generated a model with r(2) of 0.81 and q(2) of 0.62. The predicitive capacity of the model was externally evaluated with a test set of 12 competitive inhibitors and 11 out of 12 were predicted within 0.5 log unit. The most relevant points of interaction in the model correlated well to the amino acids involved in CYP2C9-substrate/inhibitor binding in the active site of a CYP2C9 homology model, further validating the mechanistic sense of our model. This approach offers the possibility to derive predicitve 3D-QSAR models without the need for an alignment rule for chemically diverse ligands and in the absence of target protein crystal structure information.

Predicting drug metabolism: a site of metabolism prediction tool applied to the cytochrome P450 2C9.

The aim of the present study is to develop a method for predicting the site at which molecules will be metabolized by CYP 2C9 (cytochrome P450 2C9) using a previously reported protein homology model of the enzyme. Such a method would be of great help in designing new compounds with a better pharmacokinetic profile, or in designing prodrugs where the compound needs to be metabolized in order to become active. The methodology is based on a comparison between alignment-independent descriptors derived from GRID Molecular Interaction Fields for the CYP 2C9 active site, and a distance-based representation of the substrate. The predicted site of metabolism is reported as a ranking list of all the hydrogen atoms of each substrate molecule. Eighty-seven CYP 2C9-catalyzed oxidative reactions reported in the literature have been analyzed. In more than 90% of these cases, the hydrogen atom ranked at the first, second, or third position was the experimentally reported site of oxidation.

Discriminant and quantitative PLS analysis of competitive CYP2C9 inhibitors versus non-inhibitors using alignment independent GRIND descriptors.

This study describes the use of alignment-independent descriptors for obtaining qualitative and quantitative predictions of the competitive inhibition of CYP2C9 on a serie of highly structurally diverse compounds. This was accomplished by calculating alignment independent descriptors in ALMOND. These GRid INdependent Descriptors (GRIND) represent the most important GRID-interactions as a function of the distance instead of the actual position of each grid-point. The experimental data was determined under uniform conditions. The inhibitor data set consists of 35 structurally diverse competitive stereospecific inhibitors of the cytochrome P450 2C9 and the non -inhibitor data set of 46 compounds. In a PLS discriminant analysis 21 inhibitors and 21 non-inhibitors (1 and 0 as activities) were analyzed using the ALMOND program obtaining a model with an r2 of 0.74 and a cross-validation value (q2) of 0.64. The model was externally validated with 39 compounds (14 inhibitors/25 non-inhibitors). 74% of the compounds were correctly predicted and an additional 13% was assigned to a borderline cluster. Thereafter, a model for quantitative predictions was generated by a PLS analysis of the GRIND descriptors using the experimental Ki-value for 21 of the competitive inhibitors (r2 = 0.77, q2 = 0.60). The model was externally validated using 12 compounds and predicted 11 out of 12 of the Ki-values within 0.5 log units. The discriminant model will be useful in screening for CYP2C9 inhibitors from large compound collections. The 3D-QSAR model will be used during lead optimization to avoid chemistry that result in inhibition of CYP2C9.