Model-Based Virtual PK/PD Exploration and Machine Learning Approach to Define PK Drivers in Early Drug Discovery.

While poor translatability of preclinical efficacy models can be responsible for clinical phase II failures, misdefinition of the optimal PK properties required to achieve therapeutic efficacy can also be a contributing factor. In the present work, the pharmacological dependency of PK end points in driving efficacy is demonstrated for six common pharmacological processes via model-based analysis. The analysis shows that the response is driven by multiple pharmacology-specific PK end points that change with how the response is defined. Moreover, the results demonstrate that the most important chemical structural features influencing response are specific to both target and downstream pharmacology, meaning the design and screening criteria must be defined uniquely for each target and corresponding pharmacology. The model-based virtual exploration of PK/PD relationships presented in this work offers one approach to identify target pharmacology-specific PK drivers and the associated potency-ADME space early in discovery to increase the probability of success and, ultimately, clinical attrition.

Applications of Model-Based Target Pharmacology Assessment in Defining Drug Design and DMPK Strategies: GSK Experiences.

Many critical decisions faced in early discovery require a thorough understanding of the dynamic behavior of pharmacological pathways following target engagement. From fundamental decisions on the optimal target to pursue and the ultimate drug product profile (combination of modality, potency, and compound properties) expected to elicit the desired clinical outcome to tactical program decisions such as what chemical series to pursue, what chemical properties require optimization, and what compounds to synthesize and progress, all demand detailed consideration of pharmacodynamics. Model-based target pharmacology assessment (mTPA) is a computational approach centered around large-scale virtual exploration of pharmacokinetic and pharmacodynamic models built early in discovery to guide these decisions. The present work summarizes several examples (use cases) from programs at GlaxoSmithKline that demonstrate the utility of mTPA throughout the drug discovery lifecycle.

EFFECTS OF TRICAINE METHANESULFONATE IN A MANAGED COLLECTION OF MOON JELLYFISH (AURELIA AURITA).

The moon jellyfish (Aurelia aurita) is a scyphozoan frequently maintained in public and private aquaria. Little research has been conducted to investigate the effects of various drugs, such as anesthetics, in this species. Tricaine methanesulfonate (MS-222), a common immersion anesthetic for fish and amphibians, was evaluated in a managed population of moon jellyfish. Twenty-four clinically healthy jellyfish were assigned into three groups of eight for trials of 0.3 g/L MS-222 (low concentration [LC]), 0.6 g/L MS-222 (high concentration [HC]), and a saltwater control. The goal was to evaluate the effects of MS-222 administration on moon jellyfish movement and response to stimuli. Movement and response to stimuli were measured via rocking and probe stimulus tests and observations of bell contraction quality and body tone. These tests were performed at baseline and throughout both drug exposure and recovery periods. A threshold drug effect was defined based on systematic scoring criteria. Additionally, elastomer tags were administered to four of eight animals in each MS-222 group to evaluate response to tag placement after drug exposure. Threshold drug effect was achieved in six of eight individuals in the LC group and eight of eight individuals in the HC group. The LC group had median threshold and recovery times of 12.2 and 10.1 min, respectively, while the HC group had median threshold and recovery times of 4.0 and 19.9 min, respectively. The HC group had significantly faster time to threshold drug effect (P < 0.001) and longer recovery times (P= 0.005) than the LC group. In both the LC and HC tagged group, three of four jellyfish had no reaction to tag placement. All animals recovered uneventfully, and there were no mortalities. MS-222 at 0.3 and 0.6 g/L decreased movement and response to stimuli in moon jellyfish.

Whole blood fatty acid concentrations in the San Cristóbal Galápagos tortoise (Chelonoidis chathamensis).

To continue releasing San Cristóbal Galápagos tortoises housed in managed-care facilities at the Giant Tortoise Breeding Center of Galápagos National Park (Galapaguera de Cerro Colorado) to the Otoy Ecological Farm, health assessments and physical examinations were conducted. As a part of these wellness examinations, blood was drawn from 11 tortoises to analyze fatty acid concentrations. Fatty acid levels can provide insight into the nutritional profiles, immune status, and reproductive health of vertebrates. To the co-author's knowledge, there is no current information about fatty acids in this species. It was hypothesized that there would be inherent differences based on the different geographic ranges, diets, sex, and age of turtles. It was noted that the ω-6/ω-3 ratio was higher for the breeding center than for the ecological farm and that overall polyunsaturated fatty acids (PUFAs) did not have any significant differences. The ω-6/ω-3 findings can contribute to a global picture of these fatty acids across taxa, as reptiles are underrepresented in this area of research. Additional results are a resourceful starting point for future investigations into how fatty acids are affected in Galápagos tortoises.

Model-based Target Pharmacology Assessment (mTPA): An Approach Using PBPK/PD Modeling and Machine Learning to Design Medicinal Chemistry and DMPK Strategies in Early Drug Discovery.

The optimal pharmacokinetic (PK) required for a drug candidate to elicit efficacy is highly dependent on the targeted pharmacology, a relationship that is often not well characterized during early phases of drug discovery. Generic assumptions around PK and potency risk misguiding screening and compound design toward nonoptimal absorption, distribution, metabolism, and excretion (ADME) or molecular properties and ultimately may increase attrition as well as hit-to-lead and lead optimization timelines. The present work introduces model-based target pharmacology assessment (mTPA), a computational approach combining physiologically based pharmacokinetic/pharmacodynamic (PBPK/PD) modeling, sensitivity analysis, and machine learning (ML) to elucidate the optimal combination of PK, potency, and ADME specific for the targeted pharmacology. Examples using frequently encountered PK/PD relationships are presented to illustrate its application, and the utility and benefits of deploying such an approach to guide early discovery efforts are discussed.

TEMPORAL PATTERNS IN ADMISSION OF EASTERN BOX TURTLES (TERRAPENE CAROLINA CAROLINA) AT A NORTH CAROLINA WILDLIFE CLINIC AS A REFLECTION OF CLIMATE.

Significant impacts of global climate change on wildlife have been documented and are projected to continue. Reptiles have been suggested as being especially susceptible to these effects along with other anthropogenic impacts on their environment. A retrospective review of medical records for 2,356 wild eastern box turtles (Terrapene carolina carolina) admitted to the Turtle Rescue Team (TRT) at the North Carolina State University College of Veterinary Medicine between 1996 and 2017 was performed in order to: 1) report common presenting complaints, 2) describe the timeline of when box turtles were admitted to the TRT for each year of the study, and 3) investigate temporal shifts in turtle admissions and associations with changes in environmental temperature over a 22-yr period. The most common presenting complaint was vehicular trauma (n = 1,195) with over 70% of the caseload associated with anthropogenic impacts. Average monthly temperatures from 1996 to 2017 for Raleigh, North Carolina, USA, were extracted from the National Oceanic and Atmospheric Administration climate database. By comparing the pre- and post-2006 years using the Mann-Whitney test, we found that both the annual peak temperature and the annual timing of admission to the TRT have shifted significantly or with a strong tendency towards significance (P = 0.0008 and 0.052, respectively). Annual peak temperature has increased by 1.3°C, and timing of admission has shifted 18 days earlier between pre- and post-2006 years. This supports the hypothesis that box turtle activity patterns are shifting and that these changes are potentially related to climate.

Mathematical model of hypoxia and tumor signaling interplay reveals the importance of hypoxia and cell-to-cell variability in tumor growth inhibition.

BACKGROUND: Human tumor is a complex tissue with multiple heterogeneous hypoxic regions and significant cell-to-cell variability. Due to the complexity of the disease, the explanation of why anticancer therapies fail cannot be attributed to intrinsic or acquired drug resistance alone. Furthermore, there are inconsistent reports of hypoxia-induced kinase activities in different cancer cell-lines, where increase, decreases, or no change has been observed. Thus, we asked, why are there widely contrasting results in kinase activity under hypoxia in different cancer cell-lines and how does hypoxia play a role in anti-cancer drug sensitivity? RESULTS: We took a modeling approach to address these questions by analyzing the model simulation to explain why hypoxia driven signals can have dissimilar impact on tumor growth and alter the efficacy of anti-cancer drugs. Repeated simulations with varying concentrations of biomolecules followed by decision tree analysis reveal that the highly differential effects among heterogeneous subpopulation of tumor cells could be governed by varying concentrations of just a few key biomolecules. These biomolecules include activated serine/threonine-specific protein kinases (pRAF), mitogen-activated protein kinase kinase (pMEK), protein kinase B (pAkt), or phosphoinositide-4,5-bisphosphate 3-kinase (pPI3K). Additionally, the ratio of activated extracellular signal-regulated kinases (pERK) or pAkt to its respective total was a key factor in determining the sensitivity of pERK or pAkt to hypoxia. CONCLUSION: This work offers a mechanistic insight into how hypoxia can affect the efficacy of anti-cancer drug that targets tumor signaling and provides a framework to identify the types of tumor cells that are either sensitive or resistant to anti-cancer therapy.

The importance of villous physiology and morphology in mechanistic physiologically-based pharmacokinetic models.

PURPOSE: Existing PBPK models incorporating intestinal first-pass metabolism account for effect of drug permeability on accessible absorption surface area by use of "effective" permeability, P eff , without adjusting number of enterocytes involved in absorption or proportion of intestinal CYP3A involved in metabolism. The current model expands on existing models by accounting for these factors. METHODS: The PBPK model was developed using SAAM II. Midazolam clinical data was generated at GlaxoSmithKline. RESULTS: The model simultaneously captures human midazolam blood concentration profile and previously reported intestinal availability, using values for CYP3A CLu int , permeability and accessible surface area comparable to literature data. Simulations show: (1) failure to distinguish absorbing from non-absorbing enterocytes results in overestimation of intestinal metabolism of highly permeable drugs absorbed across the top portion of the villous surface only; (2) first-pass extraction of poorly permeable drugs occurs primarily in enterocytes, drugs with higher permeability are extracted by enterocytes and hepatocytes; (3) CYP3A distribution along crypt-villous axes does not significantly impact intestinal metabolism; (4) differences in permeability of perpetrator and victim drugs results in their spatial separation along the villous axis and intestinal length, diminishing drug-drug interaction magnitude. CONCLUSIONS: The model provides a useful tool to interrogate intestinal absorption/metabolism of candidate drugs.

Bioavailability, metabolism and disposition of oral pazopanib in patients with advanced cancer.

1. Pazopanib (Votrient) is an oral tyrosine kinase inhibitor that was recently approved for the treatment of renal cell carcinoma and soft tissue sarcoma. 2. In this two-part study, we investigated the metabolism, disposition of [(14)C]pazopanib, and the oral bioavailability of pazopanib tablets in patients with advanced cancer. 3. In part A, three men each received a single oral dose of [(14)C]pazopanib in suspension (400 mg, 70 µCi). Pazopanib was the predominant drug-related component in circulation. Two metabolites derived from hydroxylation and one from N-demethylation were also circulating, but were minor, each accounting for <5% of plasma radioactivity. Faecal elimination predominated, accounting for 82.2% of the administered radio-dose, with negligible renal elimination (2.6% of dose). Pazopanib was primarily excreted as the unchanged drug in faeces (67% of dose). 4. In part B, seven additional patients received a single intravenous administration of 5 mg pazopanib (day 1) followed by oral administration of 800 mg pazopanib tablet once daily for 26 days (days 3 or 5-28). In the three evaluable patients from part B, pazopanib had a slow plasma clearance and a small volume of distribution. The absolute oral bioavailability of the 800 mg pazopanib tablet ranged from 14% to 39%.

Gastric pH and gastric residence time in fasted and fed conscious beagle dogs using the Bravo pH system.

To further characterize the time course of gastric pH with respect to meals and gastric residence times (GRTs) in dogs, continuous pH measurements were recorded with Bravo capsules, which were attached to the dogs' stomach mucosa or administered as free capsules, respectively. Experiments took place in home or study cages, and meals were administered at designated times. Up until 2 h prior to mealtime, the fasted gastric pH remained constantly acidic (∼2.0) regardless whether the dogs were in the study or home cages. However, as feeding time became imminent, the pH was typically elevated for dogs in home cages, whereas the pH remained acidic for dogs in study cages. For both monitoring locations, the gastric pH remained acidic during meal consumption and for at least 10 h after meals. The GRT between fasted (25 ± 32 min) and fed (686 ± 352 min) conditions was significantly different with considerable inter- and intrasubject variability. Fasted gastric pH was similar to that of literature monkey and human values but differed after meals as the dog gastric pH remained acidic unlike monkey and human. In dogs, the fasted GRT was remarkably rapid and under fed conditions, longer than that observed in humans.

Investigations of hydrazine cleavage of eltrombopag in humans.

After oral administration to humans, eltrombopag undergoes extensive cleavage of its hydrazine linkage to metabolites, which are exclusively eliminated in urine. In vitro, the cleavage pathway was not detected in systems using cytochrome P450 enzymes, renal or hepatic microsomes, or hepatocytes but was readily evident after anaerobic incubation with rodent cecal contents or human fecal homogenate. Antibiotic treatment in vitro and in vivo inhibited eltrombopag cleavage, further indicating that cleavage is via gut microbes. Antibiotic treatment did not alter the systemic exposure of eltrombopag in mice. Oral and intravenous pharmacokinetic characterization in the mice with one of the cleavage products indicated that it was readily absorbed, conjugated, and eliminated in urine, consistent with its fate after oral administration of eltrombopag. Variation in this microbial pathway, for example by antibiotic cotherapy, is unlikely to be clinically significant.

A mechanism-based mathematical model of aryl hydrocarbon receptor-mediated CYP1A induction in rats using beta-naphthoflavone as a tool compound.

ß-Naphthoflavone (BNF) is a synthetic flavone that selectively and potently induces CYP1A enzymes via aryl hydrocarbon receptor activation. Mechanism-based mathematical models of CYP1A enzyme induction were developed to predict the time course of enzyme induction and quantitatively evaluate the interrelationship between BNF plasma concentrations, hepatic CYP1A1 and CYP1A2 mRNA levels, and CYP1A enzyme activity in rats in vivo. Male Sprague-Dawley rats received a continuous intravenous infusion of vehicle or 1.5 or 6 mg · kg(-1) · h(-1) BNF for 6 h, with blood and liver sampling. Plasma BNF concentrations were determined by liquid chromatography-tandem mass spectrometry. Hepatic mRNA levels of CYP1A1 and CYP1A2 were determined by TaqMan. Ethoxyresorufin O-deethylation was used to measure the increase in CYP1A enzyme activity as a result of induction. The induction of hepatic CYP1A1/CYP1A2 mRNA and CYP1A activity occurred within 2 h after BNF administration. This caused a rapid increase in metabolic clearance of BNF, resulting in plasma concentrations declining during the infusion. Overall, the enzyme induction models developed in this study adequately captured the time course of BNF pharmacokinetics, CYP1A1/CYP1A2 mRNA levels, and increases in CYP1A enzyme activity data for both dose groups simultaneously. The model-predicted degradation half-life of CYP1A enzyme activity is comparable with previously reported values. The present results also confirm a previous in vitro finding that CYP1A1 is the predominant contributor to CYP1A induction. These physiologically based models provide a basis for predicting drug-induced toxicity in humans from in vitro and preclinical data and can be a valuable tool in drug development.

Differences in effects on myocardium and mitochondria by angiogenic inhibitors suggest separate mechanisms of cardiotoxicity.

Several multikinase angiogenesis inhibitors demonstrate mitochondrial and/or cardiovascular toxicity, suggesting an on-target pharmacologic effect. To evaluate whether cardiotoxicity is directly related to vascular endothelial growth factor receptor inhibition, we investigated the effects of sunitinib, sorafenib, and pazopanib on myocardial function and structure. We used a rat model to assess myocardial effects of the inhibitors concurrently exposed to the cardiac stressor dobutamine. Echocardiographic abnormalities including premature ventricular contractions, decreases in heart rate, circumferential strain, and radial and circumferential strain rates were noted with sorafenib, but not with sunitinib or pazopanib. Ultrastructural analysis of ventricular cardiomyocytes by transmission electron microscopy revealed mitochondrial swelling, dense deposits, and matrix cavitation in rats given sunitinib and disrupted mitochondrial cristae in rats given sorafenib, but there were no effects with pazopanib. Effects on neonatal rat cardiomyocyte cultures were assessed, which identified decreases in mitochondrial membrane potential with sunitinib treatment, but not with sorafenib or pazopanib. Intracellular adenosine triphosphate depletion was observed with sunitinib and sorafenib, but not pazopanib. Our results show that cardiotoxicity is not necessarily related to a pharmacologic classwide effect of vascular endothelial growth factor receptor inhibition, and the rat myocardial structural and functional changes identified in this study may be instead a result of inhibition of other kinase pathways, the mechanism of which may be associated with mitochondrial toxicity.

Gastric pH and gastric residence time in fasted and fed conscious cynomolgus monkeys using the Bravo pH system.

PURPOSE: To measure fasted and fed gastric pH and gastric residence time (GRT) in Cynomolgus monkeys using Bravo radiotelemetry capsules. METHODS: Continuous pH measurements were recorded with Bravo capsules, which were either attached to the monkeys' stomach or administered as free capsules. Meals (either slurry or standard), were administered at designated times with monkeys chair-restrained during slurry meal ingestion. RESULTS: From the attached capsule studies, the fasted gastric pH (~1.9-2.2) was consistent among monkeys. Under fasted conditions, pH spikes were infrequently observed (once every 7.9 min to 3.6 h) with peaks reaching pH 9.4 and having short durations (<1 min). After feeding, the gastric pH rose quickly and remained alkaline for approximately 4.5-7.5 h before returning to baseline. Although significantly different (p < 0.05), there was overlap between the fasted (153 +/- 87 min) and fed (436 +/- 265 (slurry) and 697 +/- 193 (standard) min) GRT due to considerable inter- and intra-subject variability. CONCLUSIONS: Fasted gastric pH was similar between monkeys and literature human values. After a meal, the monkey gastric pH was elevated for a longer duration than that in human. The monkey GRT appears longer than that observed in human under both fasted and fed conditions, although this is likely dependent on the Bravo capsule size.

Enhancement of in vitro and in vivo microdialysis recovery of SB-265123 using Intralipid and Encapsin as perfusates.

This study was conducted to compare the ability of two potential microdialysis perfusates to enhance the recovery of SB-265123, a lipophilic, highly protein-bound compound, both in vitro and in vivo. Initial in vitro experiments established that the recovery of SB-265123 by microdialysis using normal saline as a perfusate was poor (1.7%). Different concentrations of Intralipid and Encapsin also were evaluated in an identical in vitro setting, to determine enhancement of recovery. In vitro recovery was enhanced to approximately 24 and 65% with 5 and 20% Intralipid, and to approximately 59 and 62% with 5 and 20% Encapsin, respectively. A rat in vivo study was conducted with 20% Encapsin to confirm the in vitro observations. In the in vivo study, 75-80% recovery of free SB-265123 was achieved using 20% Encapsin as a perfusate. The results from this study indicate that for SB-265123, a lipophilic, highly protein-bound molecule, Encapsin is an efficient recovery enhancer in vitro. The results from this investigation further demonstrate that a recovery enhancer may be useful for in vivo applications, even with a compound that is highly bound to plasma protein.

Discovery of a novel and potent class of FabI-directed antibacterial agents.

Bacterial enoyl-acyl carrier protein (ACP) reductase (FabI) catalyzes the final step in each elongation cycle of bacterial fatty acid biosynthesis and is an attractive target for the development of new antibacterial agents. High-throughput screening of the Staphylococcus aureus FabI enzyme identified a novel, weak inhibitor with no detectable antibacterial activity against S. aureus. Iterative medicinal chemistry and X-ray crystal structure-based design led to the identification of compound 4 [(E)-N-methyl-N-(2-methyl-1H-indol-3-ylmethyl)-3-(7-oxo-5,6,7,8-tetrahydro-1,8-naphthyridin-3-yl)acrylamide], which is 350-fold more potent than the original lead compound obtained by high-throughput screening in the FabI inhibition assay. Compound 4 has exquisite antistaphylococci activity, achieving MICs at which 90% of isolates are inhibited more than 500 times lower than those of nine currently available antibiotics against a panel of multidrug-resistant strains of S. aureus and Staphylococcus epidermidis. Furthermore, compound 4 exhibits excellent in vivo efficacy in an S. aureus infection model in rats. Biochemical and genetic approaches have confirmed that the mode of antibacterial action of compound 4 and related compounds is via inhibition of FabI. Compound 4 also exhibits weak FabK inhibitory activity, which may explain its antibacterial activity against Streptococcus pneumoniae and Enterococcus faecalis, which depend on FabK and both FabK and FabI, respectively, for their enoyl-ACP reductase function. These results show that compound 4 is representative of a new, totally synthetic series of antibacterial agents that has the potential to provide novel alternatives for the treatment of S. aureus infections that are resistant to our present armory of antibiotics.