Perspectives of data science in preclinical safety assessment.

The data landscape in preclinical safety assessment is fundamentally changing because of not only emerging new data types, such as human systems biology, or real-world data (RWD) from clinical trials, but also technological advancements in data-processing software and analytical tools based on deep learning approaches. The recent developments of data science are illustrated with use cases for the three factors: predictive safety (new in silico tools), insight generation (new data for outstanding questions); and reverse translation (extrapolating from clinical experience to resolve preclinical questions). Further advances in this field can be expected if companies focus on overcoming identified challenges related to a lack of platforms and data silos and assuring appropriate training of data scientists within the preclinical safety teams.

Entrectinib, a TRK/ROS1 inhibitor with anti-CNS tumor activity: differentiation from other inhibitors in its class due to weak interaction with P-glycoprotein.

BACKGROUND: Studies evaluating the CNS penetration of a novel tyrosine kinase inhibitor, entrectinib, proved challenging, particularly due to discrepancies across earlier experiments regarding P-glycoprotein (P-gp) interaction and brain distribution. To address this question, we used a novel "apical efflux ratio" (AP-ER) model to assess P-gp interaction with entrectinib, crizotinib, and larotrectinib, and compared their brain-penetration properties. METHODS: AP-ER was designed to calculate P-gp interaction with the 3 drugs in vitro using P-gp-overexpressing cells. Brain penetration was studied in rat plasma, brain, and cerebrospinal fluid (CSF) samples after intravenous drug infusion. Unbound brain concentrations were estimated through kinetic lipid membrane binding assays and ex vivo experiments, while the antitumor activity of entrectinib was evaluated in a clinically relevant setting using an intracranial tumor mouse model. RESULTS: Entrectinib showed lower AP-ER (1.1-1.15) than crizotinib and larotrectinib (≥2.8). Despite not reaching steady-state brain exposures in rats after 6 hours, entrectinib presented a more favorable CSF-to-unbound concentration in plasma (CSF/Cu,p) ratio (>0.2) than crizotinib and larotrectinib at steady state (both: CSF/Cu,p ~0.03). In vivo experiments validated the AP-ER approach. Entrectinib treatment resulted in strong tumor inhibition and full survival benefit in the intracranial tumor model at clinically relevant systemic exposures. CONCLUSIONS: Entrectinib, unlike crizotinib and larotrectinib, is a weak P-gp substrate that can sustain CNS exposure based on our novel in vitro and in vivo experiments. This is consistent with the observed preclinical and clinical efficacy of entrectinib in neurotrophic tropomyosin receptor kinase (NTRK) and ROS1 fusion-positive CNS tumors and secondary CNS metastases.

Subcutaneous Site-of-Absorption Study with the Monoclonal Antibody Tocilizumab in Minipigs: Administration Behind Ear Translates Best to Humans.

Minipigs have been proposed as animal model to study the subcutaneous (SC) absorption of monoclonal antibodies (mAb), because they are more translatable to humans than other species. However, the minipig SC tissue structure differs markedly depending on its location. This study explored different SC administration sites for mAb SC administration, to explore which site translates best to humans. The study assessed the SC absorption of tocilizumab (Actemra®) following administration at several injection sites: Inguinal area, flank, caudal to the ear, and interscapular area, in comparison with an IV group. After SC administration, tocilizumab absorption was most rapid from the inguinal administration site, and slowest after administration behind the ear, with absorption from the other sites in between. Tocilizumab bioavailability was 98.6, 88.3, 74.1, and 86.3% after administration in inguinal area, flank, behind the ear, and interscapular area, as determined by non-compartmental analysis. Fitting of a single first-order absorption rate constant by compartmental analysis was dissatisfactory. A combined fitting of all data was done assuming two different kinds of SC depots, one undergoing fast absorption, the other undergoing a slower absorption. The split between these absorption depots differed across administration sites, with absorption from "fast depot" in inguinal area > flank > interscapular area > behind the ear. Comparisons with clinical data show that tocilizumab PK after SC administration behind the ear translates best to humans, considering both bioavailability and rate of absorption. Whether this translation from minipigs to humans is prototypic for other mAb remains to be confirmed.

Preliminary pharmacokinetics of tramadol hydrochloride after administration via different routes in male and female B6 mice.

OBJECTIVE: 1) To determine the pharmacokinetics of tramadol hydrochloride and its active metabolite, O-desmethyltramadol (M1), after administration through different routes in female and male C57Bl/6 mice; 2) to evaluate the stability of tramadol solutions; and 3) to identify a suitable dose regimen for prospective clinical analgesia in B6 mice. STUDY DESIGN: Prospective, randomized, blinded, parallel design. ANIMALS: A total of 18 male and 18 female C57Bl/6 mice (20-30 g). METHODS: Mice were administered 25 mg kg-1 tramadol as a bolus [intravenously (IV), intraperitoneally (IP), subcutaneously (SQ), orally per gavage (OSgavage)] over 25 hours [orally in drinking water (OSwater) or Syrspend SF (OSSyrsp)]. Venous blood was sampled at six predetermined time points over 4 to 31 hours, depending on administration route, to determine tramadol and M1 plasma concentrations (liquid chromatography and tandem mass spectrometry detection). Pharmacokinetic parameters were described using a noncompartmental model. The stability of tramadol in water (acidified and untreated) and Syrspend SF (0.20 mg mL-1) at ambient conditions for 1 week was evaluated. RESULTS: After all administration routes, Cmax was >100 ng mL-1 for tramadol and >40 ng mL-1 for M1 (reported analgesic ranges in man) followed by short half-lives (2-6 hours). The mean tramadol plasma concentration after self-administration remained >100 ng mL-1 throughout consumption time. M1 was found in the OSSyrs group only at 7 hours, whereas it was detectable in OSwater throughout administration. Tramadol had low oral bioavailability (26%). Short-lasting side effects were observed only after IV administration. Water and Syrspend SF solutions were stable for 1 week. CONCLUSIONS AND CLINICAL RELEVANCE: 1) At the dose administered, high plasma concentrations of tramadol and M1 were obtained, with half-life depending on the administration route. 2) Plasma levels were stable over self-consumption time. 3) Solutions were stable for 1 week at ambient conditions.

Extensive metabolism and route-dependent pharmacokinetics of bisphenol A (BPA) in neonatal mice following oral or subcutaneous administration.

Orally administered bisphenol A (BPA) undergoes efficient first-pass metabolism to produce the inactive conjugates BPA-glucuronide (BPA-G) and BPA-sulfate (BPA-S). This study was conducted to evaluate the pharmacokinetics of BPA, BPA-G and BPA-S in neonatal mice following the administration of a single oral or subcutaneous (SC) dose. This study consisted of 3 phases: (1) mass-balance phase in which effective dose delivery procedures for oral or SC administration of (3)H-BPA to postnatal day three (PND3) mice were developed; (2) pharmacokinetic phase during which systemic exposure to total (3)H-BPA-derived radioactivity in female PND3 mice was established; and (3) metabolite profiling phase in which 50 female PND3 pups received either a single oral or SC dose of (3)H-BPA. Blood was collected from 5 pups/route/time-point at various times post-dosing, the blood plasma samples were pooled by group, and time-point and samples were profiled by HPLC with fraction collection. Fractions were analyzed for total radioactivity and data used to reconstruct radiochromatograms and to integrate individual peaks. The identity of the BPA, BPA-G, and BPA-S peaks was confirmed using authentic standards and LC-MS/MS analysis. The result of this study revealed that female PND3 mice have the capacity to metabolize BPA to BPA-G, BPA-S and other metabolites after both routes of administration. Systemic exposure to free BPA is route-dependent as the plasma concentrations were lower following oral administration compared to SC injection.

Lack of neuropathological changes in rats administered tedizolid phosphate for nine months.

Tedizolid, a novel oxazolidinone antibacterial, was administered to Long Evans rats by oral gavage once daily for up to 9 months at doses near the maximum tolerated dose (MTD) to evaluate for potential neurotoxicity. Mean plasma exposures of tedizolid at the low-, medium-, and high-dose levels (7.5, 15, and 30 mg/kg of body weight/day for males; 2.5, 5, and 10 mg/kg/day for females) were similar between males and females and were 1.8-, 3.9-, and 8.0-fold greater than exposures in patients at the therapeutic dose (200 mg once daily). Evaluated endpoints included survival, clinical observations, body weight, and food consumption. At 1, 3, 6, and 9 months, ophthalmic examinations, functional observational batteries, and locomotor activity measures were conducted, brain weights/sizes were recorded, and perfusion-fixed tissues were collected from 12 rats/sex/group/time point. A detailed morphological assessment was conducted on brain, eyes, optic nerve/tract, spinal cord, peripheral nerves (includes sciatic, sural, tibial, peroneal, trigeminal), and skeletal muscle. At the end of 9 months, less body weight gain was seen in high-dose males (-6.7%) and females (-5.8%) compared with that seen in controls. There were no tedizolid-related adverse neurobehavioral effects or tedizolid-related histopathologic changes in the central/peripheral nervous systems, including the optic nerve. Results of this study indicate that tedizolid was not neurotoxic when administered long term to pigmented rats at doses near the MTD, which were up to 8-fold higher than the human therapeutic exposure.

High-performance liquid chromatography analysis of N-acyl homoserine lactone hydrolysis by paraoxonases.

Mammalian paraoxonases (PONs) are a unique, highly conserved family of calcium-dependent esterases consisting of PON1, PON2, and PON3. The PONs can hydrolyze the lactone ring of a range of N-acyl-L: -homoserine lactone (AHL) quorum sensing signaling molecules, rendering them inactive. This chapter describes a method that utilizes high-performance liquid chromatography analysis with UV detection for determining the rate of AHL hydrolysis in cell lysates, tissue homogenates, serum, and with purified proteins. Also described are the techniques used to prepare cell culture lysates and tissue homogenates for analysis and the use of class-specific enzyme inhibitors to determine the contribution of PONs to AHL hydrolysis in the samples.

LTBP2 and CYP1B1 mutations and associated ocular phenotypes in the Roma/Gypsy founder population.

Primary congenital glaucoma (PCG) is a genetically heterogeneous autosomal recessive disorder, which is an important cause of blindness in childhood. The first known gene, CYP1B1, accounts for a variable proportion of cases in most populations. A second gene, LTBP2, was recently reported in association with a syndrome, in which glaucoma is secondary to lens dislocation. We report on the molecular and clinical profile of 34 families diagnosed as PCG, all originating from the Roma/Gypsy founder population. Comprehensive sequencing analysis revealed a level of heterogeneity unusual for this population, with five CYP1B1 and one ancestral LTBP2 mutation accounting for ∼70% of patients (25 out of 37) and the remainder still unexplained. Homozygosity for the founder LTBP2 p.R299X mutation resulted in a more severe clinical phenotype and poorer outcome despite a markedly higher number of surgical interventions. The genetically homogeneous group of p.R299X homozygotes showed variable phenotypes (presumably also underlying pathogenetic mechanisms), wherein PCG proper with primary dysgenesis of the trabecular meshwork, and Marfan syndrome-like zonular disease with ectopia lentis and later onset secondary glaucoma are two extremes. The spectrum manifestations may occur in different combinations and have a different evolution even within the same sibship or a single patient. Preliminary observations on compounds with mutations in both CYP1B1-LTBP2 suggest that the observed combinations are of no clinical significance and digenic inheritance is unlikely. We provide a population genetics perspective to explain the allelic heterogeneity, comparing the history and geographic distribution of the two major founder mutations--p.R299X/LTBP2 and p.E387K/CYP1B1.

PON1 and oxidative stress in human sepsis and an animal model of sepsis.

Sepsis is the leading cause of death in critically ill patients. The pathophysiological mechanisms implicated in the development of sepsis and organ failure are complex and involve activation of systemic inflammatory response and coagulation together with endothelial dysfunction. Oxidative stress is a major promoter and mediator of the systemic inflammatory response. Serum PON1 has been demonstrated in multiple clinical and animal studies to protect against oxidative stress, but also to undergo inactivation upon that condition. We found decreased plasma PON1 activity in patients with sepsis compared to healthy controls or critically ill patients without sepsis; furthermore, in sepsis patients PON1 activity was lower and remained lower in the course of sepsis in the non-survivors compared to the survivors. Plasma PON1 activity was positively correlated with high-density lipoprotein cholesterol and negatively correlated with markers of lipid peroxidation. In an experimental animal model of sepsis, murine cecal ligation and puncture, the time course of plasma PON1 activity was very similar to that found in sepsis patients. Persistently low PON1 activity in plasma was associated with lethal outcome in human and murine sepsis.

Lactonases with organophosphatase activity: structural and evolutionary perspectives.

Serum paraoxonase (PON1) is well recognized for its ability to hydrolyze arylesters, toxic oxon metabolites of organophosphate insecticides and nerve agents. PON1 is a member of gene family including also PON2 and PON3; however, the later two enzymes have very limited arylesterase and practically no organophosphatase activity. We have established that all three PONs are lactonases/lactonyzing enzymes with overlapping, but also distinct substrate specificity. Dihydrocoumarin (DHC), long chain fatty acid lactones and acylhomoserine lactones (AHLs) are hydrolyzed by all three PONs and likely represent their natural substrates. The 3D structure of PON1 is a six-bladed beta-propeller containing two Ca(2+) ions necessary for the enzyme stability and enzymatic activity. Senescence marker protein (SMP30), another putative six-bladed beta-propeller, hydrolyzes DFP, sarin and soman in the presence of Mg(2+) or Mn(2+). More recently, SMP30 was characterized as a gluconolactonase with a role in vitamin C metabolism. Bacterial phosphotriesterases (PTEs) are members of the amidohydrolase superfamily and differ in their structure from the eukaryotic organophosphatases; PTEs are (beta/alpha)(8) barrels with an active site containing two transition metal ions such as Co(2+), Mn(2+) or Zn(2+). PTE from Pseudomonas diminuta hydrolyzes paraoxon extremely efficiently; this enzyme was shown to hydrolyze also DHC and other lactones. At least 3 more bacterial lactonases, dubbed PTE-like lactonases (or PLL), have been identified to possess both lactonase and organophosphatase activities. Lactones are natural compounds, many of them with high biological activity, while organophosphates are human-made chemicals introduced in the 20th century. Thus, it is plausible that lactonase is the primary activity for which the enzymes discussed here evolved for, while the organophosphatase activity arose as a promiscuous activity during their evolution. Laboratory (directed) evolution studies provided mechanisms for their catalytic versatility and demonstrated experimentally the evolvability of promiscuous enzyme functions.

Dominant role of paraoxonases in inactivation of the Pseudomonas aeruginosa quorum-sensing signal N-(3-oxododecanoyl)-L-homoserine lactone.

The pathogenic bacterium Pseudomonas aeruginosa causes serious infections in immunocompromised patients. N-(3-oxododecanoyl)-L-homoserine lactone (3OC12-HSL) is a key component of P. aeruginosa's quorum-sensing system and regulates the expression of many virulence factors. 3OC12-HSL was previously shown to be hydrolytically inactivated by the paraoxonase (PON) family of calcium-dependent esterases, consisting of PON1, PON2, and PON3. Here we determined the specific activities of purified human PONs for 3OC12-HSL hydrolysis, including the common PON1 polymorphic forms, and found they were in the following order: PON2 >> PON1(192R) > PON1(192Q) > PON3. PON2 exhibited a high specific activity of 7.6 +/- 0.4 micromols/min/mg at 10 microM 3OC12-HSL, making it the best PON2 substrate identified to date. By use of class-specific inhibitors, approximately 85 and 95% of the 3OC12-HSL lactonase activity were attributable to PON1 in mouse and human sera, respectively. In mouse liver homogenates, the activity was metal dependent, with magnesium- and manganese-dependent lactonase activities comprising 10 to 15% of the calcium-dependent activity. In mouse lung homogenates, all of the activity was calcium dependent. The calcium-dependent activities were irreversibly inhibited by extended EDTA treatment, implicating PONs as the major enzymes inactivating 3OC12-HSL. In human HepG2 and EA.hy 926 cell lysates, the 3OC12-HSL lactonase activity closely paralleled the PON2 protein levels after PON2 knockdown by small interfering RNA treatment of the cells. These findings suggest that PONs, particularly PON2, could be an important mechanism by which 3OC12-HSL is inactivated in mammals.

Development of an immunoblot assay with infrared fluorescence to quantify paraoxonase 1 in serum and plasma.

Paraoxonase 1 (PON1) requires calcium for activity and is inactivated in the presence of EDTA. Because of this, studies to date have used serum or heparinized plasma for both activity and mass assays of PON1. Whole serum and EDTA plasma were analyzed by SDS-electrophoresis and Western blot using anti-PON1 monoclonal antibody 4C10. Because PON1 has one disulfide and one free cysteine residue, the samples were reduced with dithiothreitol before electrophoresis. Western blot identified a major PON1 band with a molecular mass of approximately 45 kDa and two minor bands of approximately 40 and 35 kDa in both serum and EDTA plasma. This established that PON1 is inactive, but structurally intact, in EDTA plasma and suggested that a mass assay could be developed based on SDS-electrophoresis and Western blot. Linearity was established for plasma and for a PON1 standard. Quantification was based on the major PON1 band at 45 kDa. The correlation between serum and plasma PON1 mass was 0.9553. The between-run variation was determined with a serum pool to be 7.8%. The mass of PON1 in serum was significantly correlated with arylesterase activity (r = 0.85). Thus, we have demonstrated the feasibility of measuring PON1 mass in either serum or EDTA plasma.

Estrogen esters as substrates for human paraoxonases.

Mammalian paraoxonases (PONs 1, 2 and 3) are a highly conserved family of esterases, with uncertain physiological functions and natural substrates. Here we characterize the ability of purified recombinant human PONs to hydrolyze estrogen esters, a class of compounds previously not known to be PON substrates. PONs hydrolyzed estrogen mono- and diesters at position 3 of the steroid A-ring. Diesters were better substrates for the PONs and were very efficiently hydrolyzed, particularly by PON3. Esters at position 17 were not cleaved by the PONs unless an adjacent double bound was present. Purified human serum butyryl cholinesterase also hydrolyzed estrogen esters, however it preferably hydrolyzed the mono-esters. The ability of the PONs' to effectively hydrolyze a variety of estrogen esters provides further insight into the structure of their active sites and suggests that natural compounds with aromatic ester groups might be relevant substrates for the PONs.

Human paraoxonases (PON1, PON2, and PON3) are lactonases with overlapping and distinct substrate specificities.

The paraoxonase (PON) gene family in humans has three members, PON1, PON2, and PON3. Their physiological role(s) and natural substrates are uncertain. We developed a baculovirus-mediated expression system, suitable for all three human PONs, and optimized procedures for their purification. The recombinant PONs are glycosylated with high-mannose-type sugars, which are important for protein stability but are not essential for their enzymatic activities. Enzymatic characterization of the purified PONs has revealed them to be lactonases/lactonizing enzymes, with some overlapping substrates (e.g., aromatic lactones), but also to have distinctive substrate specificities. All three PONs metabolized very efficiently 5-hydroxy-eicosatetraenoic acid 1,5-lactone and 4-hydroxy-docosahexaenoic acid, which are products of both enzymatic and nonenzymatic oxidation of arachidonic acid and docosahexaenoic acid, respectively, and may represent the PONs' endogenous substrates. Organophosphates are hydrolyzed almost exclusively by PON1, whereas bulky drug substrates such as lovastatin and spironolactone are hydrolyzed only by PON3. Of special interest is the ability of the human PONs, especially PON2, to hydrolyze and thereby inactivate N-acyl-homoserine lactones, which are quorum-sensing signals of pathogenic bacteria. None of the recombinant PONs protected low density lipoprotein against copper-induced oxidation in vitro.

Paraoxonase-1 does not reduce or modify oxidation of phospholipids by peroxynitrite.

Serum paraoxonase (PON1) is a high-density lipoprotein (HDL)-associated esterase/lactonase implicated to play a role in protection against atherosclerosis. However, the exact mechanism(s) and substrates for PON1 are still uncertain. In this article, we review some of the evidence for PON1's antioxidant activity, as well as our efforts to identify the actual substrates and products for this activity. We originally reported that PON1 had phospholipase activity toward oxidized phosphatidylcholine (J. Biol. Chem. 276:24473-24481; 2001). Subsequently, Marathe et al. (J. Biol. Chem. 278:3937-3947; 2003) reported that this activity was due to a contaminating lipase. However, that article did not replicate the conditions used in our previous study. To address this controversy, we purified serum PON1 by a modified method that separates the paraoxonase activity from an activity detectable as platelet-activating factor acetyl hydrolase (PAF-AH) (Teiber et al., J. Lipid. Res. 2004; Epub ahead of print, PMID 15342686) and reexamined the oxidation of phosphatidylcholine by peroxynitrite using 3-morpholinosydnonimine as a peroxynitrite generator and apolipoprotein AI-phosphatidylcholine- PON1 complexes. The phosphatidylcholines were studied by electrospray ionization tandem mass spectrometry. PON1 preparations free of PAF-AH activity showed no phospholipase activity when reconstituted into apolipoprotein AI-phosphatidylcholine complexes. We conclude that PON1 does not affect the accumulation of phosphatidylcholine oxidation products. Further, we have no evidence that PON1 has an intrinsic phospholipase A2 activity toward oxidized phospholipids.

Purified human serum PON1 does not protect LDL against oxidation in the in vitro assays initiated with copper or AAPH.

Purified serum paraoxonase (PON1) had been shown to attenuate the oxidation of LDL in vitro. We critically reevaluated the antioxidant properties of serum PON1 in the in vitro assays initiated with copper or the free radical generator 2,2'-azobis-2-amidinopropane hydrochloride (AAPH). The antioxidant activity of different purified PON1 preparations did not correlate with their arylesterase (AE), lactonase, or phospholipase A2 activities or with the amounts of detergent or protein. Dialysis of three of these preparations resulted in a 30-40% loss of their AE activities but in a complete loss of their antioxidant activities. We also followed the distribution of the antioxidant activity during human serum PON1 purification by two purification methods. The antioxidant activity of the anion-exchange chromatography fractions did not copurify with PON1 using either method and could largely be accounted for by the "antioxidant" activity of the detergent present. In conclusion, using the copper or AAPH in vitro assays, no PON1-mediated antioxidant activity was detected, suggesting that the removal of PON1 from its natural environment may impair its antioxidative activity and that this assay with highly purified PON1 may be an inappropriate method with which to study the antioxidative properties of the enzyme.

The role of hsp90 in heme-dependent activation of apo-neuronal nitric-oxide synthase.

Like other nitric-oxide synthase (NOS) enzymes, neuronal NOS (nNOS) turnover and activity are regulated by the ubiquitous protein chaperone hsp90. We have shown previously that nNOS expressed in Sf9 cells where endogenous heme levels are low is activated from the apo- to the holo-enzyme by addition of exogenous heme to the culture medium, and this activation is inhibited by radicicol, a specific inhibitor of hsp90 (Billecke, S. S., Bender, A. T., Kanelakis, K. C., Murphy, P. J. M., Lowe, E. R., Kamada, Y., Pratt, W. B., and Osawa, Y. (2002) J. Biol. Chem. 278, 15465-15468). In this work, we examine heme binding by apo-nNOS to form the active enzyme in a cell-free system. We show that cytosol from Sf9 cells facilitates heme-dependent apo-nNOS activation by promoting functional heme insertion into the enzyme. Sf9 cytosol also converts the glucocorticoid receptor (GR) to a state where the hydrophobic ligand binding cleft is open to access by steroid. Both cell-free heme activation of purified nNOS and activation of steroid binding activity of the immunopurified GR are inhibited by radicicol treatment of Sf9 cells prior to cytosol preparation, and addition of purified hsp90 to cytosol partially overcomes this inhibition. Although there is an hsp90-dependent machinery in Sf9 cytosol that facilitates heme binding by apo-nNOS, it is clearly different from the machinery that facilitates steroid binding by the GR. hsp90 regulation of apo-nNOS heme activation is very dynamic and requires higher concentrations of radicicol for its inhibition, whereas GR steroid binding is determined by assembly of stable GR.hsp90 heterocomplexes that are formed by a purified five-chaperone machinery that does not activate apo-nNOS.