Transforming growth factor β (TGFβ) signaling has been recently shown to reduce antitumor response to PD-L1 blockade, leading to a renewed enthusiasm in developing anti-TGFβ therapies for potential combination with cancer immunotherapy agents. Inhibition of TGFβ signaling in nonclinical toxicology species is associated with serious adverse toxicities including cardiac valvulopathies and anemia. Previously, cardiovascular toxicities have been thought to be limited to small molecule inhibitors of TGFβ receptor and not considered to be a liability associated with pan-TGFβ neutralizing monoclonal antibodies (mAbs). Here, we report the toxicity findings associated with a potent pan-TGFβ neutralizing mAb (pan-TGFβ mAb; neutralizes TGFβ1, 2, and 3) after 5 weekly intravenous doses of 10, 30, and 100 mg/kg, followed by a 4-week recovery period, in mice and cynomolgus monkeys. Mortality was observed due to acute bleeding and cardiovascular toxicity in mice at ≥ 30 mg/kg and prolonged menstruation in female monkeys at 100 mg/kg. Additional findings considered to be on-target exaggerated pharmacology included generalized bleeding and cardiovascular toxicity in mice and monkeys; histopathologic changes in the teeth, tongue, and skin in mice; and abnormal wound healing and microscopic pathology in the bone in monkeys. Importantly, our data indicate that the cardiovascular toxicities associated with the inhibition of TGFβ signaling are not limited to small molecule inhibitors but are also observed following administration of a potent pan-TGFβ inhibiting mAb.

This study developed a semi-mechanistic kidney model incorporating physiologically-relevant fluid reabsorption and transporter-mediated active reabsorption. The model was applied to data for the drug of abuse γ-hydroxybutyric acid (GHB), which exhibits monocarboxylate transporter (MCT1/SMCT1)-mediated renal reabsorption. The kidney model consists of various nephron segments – proximal tubules, Loop-of-Henle, distal tubules, and collecting ducts – where the segmental fluid flow rates, volumes, and sequential reabsorption were incorporated as functions of the glomerular filtration rate. The active renal reabsorption was modeled as vectorial transport across proximal tubule cells. In addition, the model included physiological blood, liver, and remainder compartments. The population pharmacokinetic modeling was performed using ADAPT5 for GHB blood concentration-time data and cumulative amount excreted unchanged into urine data (200–1000 mg/kg IV bolus doses) from rats (Felmlee et al (PMID: 20461486)). Simulations assessed the effects of inhibition (R=[I]/KI=0–100) of renal reabsorption on systemic exposure (AUC) and renal clearance of GHB. Visual predictive checks and other model diagnostic plots indicated that the model reasonably captured GHB concentrations. Simulations demonstrated that the inhibition of renal reabsorption significantly increased GHB renal clearance and decreased AUC. Model validation was performed using a separate dataset. Furthermore, our model successfully evaluated the pharmacokinetics of L-lactate using data obtained from Morse et al (PMID: 24854892). In conclusion, we developed a semi-mechanistic kidney model that can be used to evaluate transporter-mediated active renal reabsorption of drugs by the kidney.

This research describes a rapid solubility classification approach that could be used in the discovery and development of new molecular entities. Compounds (N = 635) were divided into two groups based on information available in the literature: high solubility (BDDCS/BCS 1/3) and low solubility (BDDCS/BCS 2/4). We established decision rules for determining solubility classes using measured log solubility in molar units (MLogSM) or measured solubility (MSol) in mg/ml units. ROC curve analysis was applied to determine statistically significant threshold values of MSol and MLogSM. Results indicated that NMEs with MLogSM >−3.05 or MSol >0.30 mg/mL will have ≥85% probability of being highly soluble and new molecular entities with MLogSM ≤−3.05 or MSol ≤0.30 mg/mL will have ≥85% probability of being poorly soluble. When comparing solubility classification using the threshold values of MLogSM or MSol with BDDCS, we were able to correctly classify 85% of compounds. We also evaluated solubility classification of an independent set of 108 orally administered drugs using MSol (0.3 mg/mL) and our method correctly classified 81% and 95% of compounds into high and low solubility classes, respectively. The high/low solubility classification using MLogSM or MSol is novel and independent of traditionally used dose number criteria.

Renal clearance (CL R ), a major route of elimination for many drugs and drug metabolites, represents the net result of glomerular filtration, active secretion and reabsorption, and passive reabsorption. The aim of this study was to develop quantitative structurepharmacokinetic relationships (QSPKR) to predict CL R of drugs or drug-like compounds in humans. Human CL R data for 382 compounds were obtained from the literature.Step-wise multiple linear regression was used to construct QSPKR models for training sets and their predictive performance was evaluated using internal validation (leave-one-out method). All qualified models were validated externally using test sets. QSPKR models were also constructed for compounds in accordance with their 1) net elimination pathways (net secretion, extensive net secretion, net reabsorption, and extensive net reabsorption), 2) net elimination clearances (net secretion clearance, CL SEC ; or net reabsorption clearance, CL REAB ), 3) ion status, and 4) substrate/inhibitor specificity for renal transporters. We were able to predict 1) CL REAB Moreover, compounds undergoing net reabsorption/extensive net reabsorption predominantly belonged to Biopharmaceutics Drug Disposition Classification System classes 1 and 2. In conclusion, constructed parsimonious QSPKR models can be used to predict CL R of compounds that 1) undergo net reabsorption/extensive net reabsorption and 2) are substrates and/or inhibitors of human renal transporters.

Abstract. The objective of the present study was to evaluate mechanistic pharmacokinetic models describing active renal secretion and reabsorption over a range of Michaelis-Menten parameter estimates and doses. Plasma concentration and urinary excretion profiles were simulated and renal clearance (CL r ) was calculated for two pharmacokinetic models describing active renal reabsorption (R1/ R2), two models describing active secretion (S1/S2), and a model containing both processes. A range of doses (1-1,000 mg/kg) was evaluated, and V max and K m parameter estimates were varied over a 100-fold range. Similar CL r values were predicted for reabsorption models (R1/R2) with variations in V max and K m . Tubular secretion models (S1/S2) yielded similar relationships between Michaelis-Menten parameter perturbations and CL r , but the predicted CL r values were threefold higher for model S1. For both reabsorption and secretion models, the greatest changes in CL r were observed with perturbations in V max , suggesting the need for an accurate estimate of this parameter. When intrinsic clearance was substituted for Michaelis-Menten parameters, it failed to predict similar CL r values even within the linear range. For models S1 and S2, renal secretion was predominant at low doses, whereas renal clearance was driven by fraction unbound in plasma at high doses. Simulations demonstrated the importance of Michaelis-Menten parameter estimates (especially V max ) for determining CL r . K m estimates can easily be obtained directly from in vitro studies. However, additional scaling of in vitro V max estimates using in vitro/in vivo extrapolation methods are required to incorporate these parameters into pharmacokinetic models.