GraphCpG: imputation of single-cell methylomes based on locus-aware neighboring subgraphs.

MOTIVATION: Single-cell DNA methylation sequencing can assay DNA methylation at single-cell resolution. However, incomplete coverage compromises related downstream analyses, outlining the importance of imputation techniques. With a rising number of cell samples in recent large datasets, scalable and efficient imputation models are critical to addressing the sparsity for genome-wide analyses. RESULTS: We proposed a novel graph-based deep learning approach to impute methylation matrices based on locus-aware neighboring subgraphs with locus-aware encoding orienting on one cell type. Merely using the CpGs methylation matrix, the obtained GraphCpG outperforms previous methods on datasets containing more than hundreds of cells and achieves competitive performance on smaller datasets, with subgraphs of predicted sites visualized by retrievable bipartite graphs. Besides better imputation performance with increasing cell number, it significantly reduces computation time and demonstrates improvement in downstream analysis. AVAILABILITY AND IMPLEMENTATION: The source code is freely available at https://github.com/yuzhong-deng/graphcpg.git.

The Absolute Bioavailability and Absorption, Metabolism, and Excretion of Ipatasertib, a Potent and Highly Selective Protein Kinase B (Akt) Inhibitor.

Ipatasertib (GDC-0068) is a potent, highly selective, small-molecule inhibitor of protein kinase B (Akt) being developed by Genentech/Roche as a single agent and in combination with other therapies for the treatment of cancers. To fully understand the absorption, metabolism, and excretion of ipatasertib in humans, an open-label study using 14C-radiolabeled ipatasertib was completed to characterize the absolute bioavailability (period 1) and mass balance and metabolite profiling (period 2). In period 1, subjects were administered a 200 mg oral dose of ipatasertib followed by an 80 µg (800 nCi) intravenous dose of [14C]-ipatasertib. In period 2, subjects received a single oral dose containing approximately 200 mg (100 µCi) [14C]-ipatasertib. In an integrated analytical strategy, accelerator mass spectrometry was applied to measure the 14C microtracer intravenous pharmacokinetics in period 1 and fully profile plasma radioactivity in period 2. The systemic plasma clearance and steady-state volume of distribution were 98.8 L/h and 2530 L, respectively. The terminal half-lives after oral and intravenous administrations were similar (26.7 and 27.4 hours, respectively) and absolute bioavailability of ipatasertib was 34.0%. After a single oral dose of [14C]-ipatasertib, 88.3% of the administered radioactivity was recovered with approximately 69.0% and 19.3% in feces and urine, respectively. Radioactivity in feces and urine was predominantly metabolites with 24.4% and 8.26% of dose as unchanged parent, respectively; indicating that ipatasertib had been extensively absorbed and hepatic metabolism was the major route of clearance. The major metabolic pathway was N-dealkylation mediated by CYP3A, and minor pathways were oxidative by cytochromes P450 and aldehyde oxidase. SIGNIFICANCE STATEMENT: The study provided definitive information regarding the absolute bioavailability and the absorption, metabolism, and excretion pathways of ipatasertib, a potent, novel, and highly selective small-molecule inhibitor of protein kinase B (Akt). An ultrasensitive radioactive counting method, accelerator mass spectrometry was successfully applied for 14C-microtracer absolute bioavailability determination and plasma metabolite profiling.

Single- and multiple-dose pharmacokinetics, potential for CYP3A inhibition, and food effect in patients with cancer and healthy subjects receiving ipatasertib.

PURPOSE: To examine the single- and multiple-dose pharmacokinetics (PK), CYP3A inhibition potential of ipatasertib, and effect of food on PK of ipatasertib in patients with refractory solid tumors and a dedicated food effect assessment in healthy subjects. METHODS: The Phase I dose-escalation study enrolled patients with solid tumors in a standard 3 + 3 design with a 1 week washout after the first dose, followed by once-daily dosing on a 3-week-on/1-week-off schedule. In the expansion cohort, the effect of ipatasertib on CYP3A substrate (midazolam) was assessed by examining the change in midazolam exposure when dosed in the absence and presence of steady-state ipatasertib at 600 mg. The effect of food on ipatasertib PK was studied with ipatasertib administered in fed or fasted state (6 patients from Phase I patient study and 18 healthy subjects from the dedicated food effect study). RESULTS: Ipatasertib was generally well tolerated at doses up to 600 mg given daily for 21 days. Ipatasertib showed rapid absorption (tmax, 0.5-3 h), was dose-proportional over a range of 200-800 mg, had a median half-life (range) of 45.0 h (27.8-66.9 h), and had approximately two-fold accumulation following once-daily dosing. Midazolam exposure (AUC0-∞) increased by 2.2-fold in the presence of ipatasertib. PK was comparable in subjects administered ipatasertib in a fed or fasted state. CONCLUSION: Ipatasertib exhibited rapid absorption and was dose-proportional over a broad dose range. Ipatasertib appeared to be a moderate CYP3A inhibitor when administered at 600 mg and could be administered with or without food in clinical studies. TRAIL REGISTRATION: NCT01090960 (registered March 23, 2010); NCT02536391 (registered August 31, 2015).

Ocular phenotypes in a mouse model of impaired glucocerebrosidase activity.

Mutations in the GBA1 gene encoding glucocerebrosidase (GCase) are linked to Gaucher (GD) and Parkinson's Disease (PD). Since some GD and PD patients develop ocular phenotypes, we determined whether ocular phenotypes might result from impaired GCase activity and the corresponding accumulation of glucosylceramide (GluCer) and glucosylsphingosine (GluSph) in the Gba1D409V/D409V knock-in (Gba KI/KI; "KI") mouse. Gba KI mice developed age-dependent pupil dilation deficits to an anti-muscarinic agent; histologically, the iris covered the anterior part of the lens with adhesions between the iris and the anterior surface of the lens (posterior synechia). This may prevent pupil dilation in general, beyond an un-responsiveness of the iris to anti-muscarinics. Gba KI mice displayed atrophy and pigment dispersion of the iris, and occlusion of the iridocorneal angle by pigment-laden cells, reminiscent of secondary open angle glaucoma. Gba KI mice showed progressive thinning of the retina consistent with retinal degeneration. GluSph levels were increased in the anterior and posterior segments of the eye, suggesting that accumulation of lipids in the eye may contribute to degeneration in this compartment. We conclude that the Gba KI model provides robust and reproducible eye phenotypes which may be used to test for efficacy and establish biomarkers for GBA1-related therapies.

Effect of Hepatic Impairment on Cobimetinib Pharmacokinetics: The Complex Interplay Between Physiological Changes and Drug Characteristics.

Cobimetinib is a kinase inhibitor indicated for use in combination with vemurafenib for treatment of unresectable/metastatic melanoma with specific BRAF mutations. Cobimetinib is extensively metabolized in liver; thus, patients with hepatic impairment (HI) might have increased cobimetinib exposure. In this study, we investigated the impact of HI on the pharmacokinetics (PK) and safety of cobimetinib. Subjects with normal hepatic function and mild to severe HI were enrolled. All subjects received a single oral dose of 10 mg cobimetinib, and serial blood samples were collected at specified times. Cobimetinib PK in subjects with mild and moderate HI was similar to that in those with normal liver function. However, subjects with severe HI, on average, showed ∼30% lower total AUC0-∞ and ∼2-fold higher unbound AUC0-∞ compared with those with normal hepatic function. These exposure differences can be explained by lower albumin levels observed in subjects with severe HI, the strong correlation between albumin level and the unbound fraction and the general PK variability of cobimetinib. In addition, previous studies with cobimetinib showed a lack of an exposure-response relationship for efficacy and safety. Therefore, collectively, our results suggest that the starting dose for patients with hepatic impairment can be the same as that for those with normal hepatic function.

Catalytic Cleavage of Disulfide Bonds in Small Molecules and Linkers of Antibody-Drug Conjugates.

In cells, catalytic disulfide cleavage is an essential mechanism in protein folding and synthesis. However, detailed enzymatic catalytic mechanism relating cleavage of disulfide bonds in xenobiotics is not well understood. This study reports an enzymatic mechanism of cleavage of disulfide bonds in xenobiotic small molecules and antibody conjugate (ADC) linkers. The chemically stable disulfide bonds in substituted disulfide-containing pyrrolobenzodiazepine (PBD, pyrrolo[2,1-c][1,4]benzodiazepine) monomer prodrugs in presence of glutathione or cysteine were found to be unstable in incubations in whole blood of humans and rats. It was shown the enzymes involved were thioredoxin (TRX) and glutaredoxin (GRX). For a diverse set of drug-linker conjugates, we determined that TRX in the presence of TRX-reductase and NADPH generated the cleaved products that are consistent with catalytic disulfide cleavage and linker immolation. GRX was less rigorously studied; in the set of compounds studied, its role in the catalytic cleavage was also confirmed. Collectively, these in vitro experiments demonstrate that TRX as well as GRX can catalyze the cleavage of disulfide bonds in both small molecules and linkers of ADCs.

Exposure-Effect Relationships in Established Rat Adjuvant-Induced and Collagen-Induced Arthritis: A Translational Pharmacokinetic-Pharmacodynamic Analysis.

The ability of rodent immune-mediated arthritis models to quantitatively predict therapeutic activity of antiarthritis agents is poorly understood. Two commonly used preclinical models of arthritis are adjuvant-induced arthritis (AIA) and collagen-induced arthritis (CIA) in rats. The objective of the current study is to investigate the relationship between efficacy in AIA and CIA in rats, and clinical efficacy in rheumatoid arthritis patients using translational pharmacokinetic-pharmacodynamic (PK-PD) analysis. A range of doses of indomethacin (a nonsteroidal anti-inflammatory drug), and three disease-modifying antirheumatic drugs (DMARDs), methotrexate, etanercept, and tofacitinib, were evaluated in AIA and CIA rats. Dexamethasone was included in this study as a positive control. The area under the ankle diameter-time profile (AUCankle) and ankle histopathology summed scores (AHSS) were used as efficacy endpoints for activity against disease symptoms (joint inflammation) and disease progression (joint damage), respectively. Translational PK-PD analysis was performed to rank order preclinical efficacy endpoints at clinically relevant concentrations. For each drug tested, inhibition of AUCankle and AHSS scores was generally comparable in both magnitude and rank order. Overall, based on both AUCankle and the AHSS inhibition, the rank ordering of preclinical activity for the DMARDs evaluated was tofacitinib > etanercept ≥ methotrexate. This ranking of preclinical efficacy was consistent with reported clinical efficacy. Of interest, indomethacin showed equal or often better efficacy than the three DMARDs evaluated on inhibiting AHSS despite having limited ability to prevent joint damage clinically in patients. The translational value of performing PK-PD analysis of arthritis models in rats is discussed.

An Integrated Analysis of Solid Form Change Impact on Solubility and Permeability: Case Study of Oral Exposure in Rats of an RAR Related Orphan Receptor C Inhibitor.

It is well acknowledged that the oral absorption of a drug can be influenced by its solubility, which is usually associated with its solid form properties. G1032 is a retinoic acid-related orphan receptor inverse agonist. Crystalline solid (form A) was identified with an aqueous solubility of 130 µg/mL. This form was used in an oral dose escalation study in rodents up to 300 mg/kg and achieved good exposures. Later on, a more stable crystalline hydrate (form B) was identified and the aqueous solubility was reduced to 55 µg/mL. A modeling exercise suggested that this solubility change would cause a 2-fold decrease in exposure at tested doses; however, the actual reduction was far larger than the model predicted. At high dose, exposure was found to be reduced by almost 10-fold. A parameter sensitivity analysis suggested that such a drop in exposure could be associated with permeability reduction as well. More in vitro permeability experiments were performed, indicating G1032 was an efflux transporter substrate. This finding was integrated into the modeling and the design for in vivo studies. Data obtained from those studies allowed us to better understand the causes of the higher-than-expected exposure change and enabled decision-making.

Exploration of Pyrrolobenzodiazepine (PBD)-Dimers Containing Disulfide-Based Prodrugs as Payloads for Antibody-Drug Conjugates.

A number of cytotoxic pyrrolobenzodiazepine (PBD) monomers containing various disulfide-based prodrugs were evaluated for their ability to undergo activation (disulfide cleavage) in vitro in the presence of either glutathione (GSH) or cysteine (Cys). A good correlation was observed between in vitro GSH stability and in vitro cytotoxicity toward tumor cell lines. The prodrug-containing compounds were typically more potent against cells with relatively high intracellular GSH levels (e.g., KPL-4 cells). Several antibody-drug conjugates (ADCs) were subsequently constructed from PBD dimers that incorporated selected disulfide-based prodrugs. Such HER2 conjugates exhibited potent antiproliferation activity against KPL-4 cells in vitro in an antigen-dependent manner. However, the disulfide prodrugs contained in the majority of such entities were surprisingly unstable toward whole blood from various species. One HER2-targeting conjugate that contained a thiophenol-derived disulfide prodrug was an exception to this stability trend. It exhibited potent activity in a KPL-4 in vivo efficacy model that was approximately three-fold weaker than that displayed by the corresponding parent ADC. The same prodrug-containing conjugate demonstrated a three-fold improvement in mouse tolerability properties in vivo relative to the parent ADC, which did not contain the prodrug.

Correction to: Strategies of Drug Transporter Quantitation by LC-MS: Importance of Peptide Selection and Digestion Efficiency.

Liling Liu was not noted as the co-first author in the original article. Buyun Chen and Liling Liu contributed equally to the article.

Practical strategies when using a stable isotope labeled microtracer for absolute bioavailability assessment: A case study of a high oral dose clinical candidate GDC-0810.

The pH labile metabolite, hydrophobicity, high oral dose and systematic exposure of GDC-0810 posed tremendous challenges to develop a LC-MS method for a stable isotope labeled aBA study. In this study, we explored practical solutions to balance stability and sensitivity and to cope with the impact of high Cp.o. to Ci.v. ratio on the labeling selection and assay dynamic range. A [13C9] GDC-0810 was synthesized to minimize the isotopic interference between PO dose, internal standard and I.V. microtracer. A highly sensitive LC-MS assay was validated for quantitation of [13C9] GDC-0810 from 5 to 1250 pg/mL. The optimized method was applied to a proof of concept cynomolgus monkey aBA study and the bioavailability calculated using microtracer dosing and regular dosing were similar to each other.

Strategies of Drug Transporter Quantitation by LC-MS: Importance of Peptide Selection and Digestion Efficiency.

Huge variation of drug transporter abundance was seen in the literature, making PBPK prediction difficult when transporters play a major role. Among multiple factors such as membrane fraction, digestion, and peptide selection that contributed to such variation, peptide selection is the least discussed. Herein, a strategy was established by using a small amount of purified protein standard to select a peptide with near 100% digestion efficiency for quantitation of a transporter protein MDR1. The impact of native membrane protein's tertiary structure on the digestion efficiency of surrogate peptides of MDR1 was investigated. Peptides in more solvent accessible regions are found to be digested much more efficiently than those in large stretches of helical structures. The concentration of peptide EALDESIPPVSFWR(EAL) in the most solvent accessible linker region of MDR1 was found closest to the true protein concentration. When using EAL for MDR1 quantitation, the abundance is over 10 times higher than previously reported, indicating the importance of peptide selection for transporter quantitation. In addition, this study also proposes a screening strategy to select peptides appropriate for relative quantitation for in vitro-in vivo extrapolation in the absence of any protein standard.

Absorption, metabolism and excretion of cobimetinib, an oral MEK inhibitor, in rats and dogs.

1. The absorption, metabolism and excretion of cobimetinib, an allosteric inhibitor of MEK1/2, was characterized in mass balance studies following single oral administration of radiolabeled (14C) cobimetinib to Sprague-Dawley rats (30 mg/kg) and Beagle dogs (5 mg/kg). 2. The oral dose of cobimetinib was well absorbed (81% and 71% in rats and dogs, respectively). The maximal plasma concentrations for cobimetinib and total radioactivity were reached at 2-3 h post-dose. Drug-derived radioactivity was fully recovered (∼90% of the administered dose) with the majority eliminated in feces via biliary excretion (78% of the dose for rats and 65% for dogs). The recoveries were nearly complete after the first 48 h following dosing. 3. The metabolic profiles indicated extensive metabolism of cobimetinib prior to its elimination. For rats, the predominant metabolic pathway was hydroxylation at the aromatic core. Lower exposures for cobimetinib and total radioactivity were observed in male rats compared with female rats, which was consistent to in vitro higher clearance of cobimetinib for male rats. For dogs, sequential oxidative reactions occurred at the aliphatic portion of the molecule. Though rat metabolism was well-predicted in vitro with liver microsomes, dog metabolism was not. 4. Rats and dogs were exposed to the two major human circulating Phase II metabolites, which provided relevant metabolite safety assessment. In general, the extensive sequential oxidative metabolism in dogs, and not the aromatic hydroxylation in rats, was more indicative of the metabolism of cobimetinib in humans.

Discovery of oxa-sultams as RORc inverse agonists showing reduced lipophilicity, improved selectivity and favorable ADME properties.

Modification of the δ-sultam ring of RORc inverse agonist 2 led to the discovery of more polar oxa-sultam 65. The less lipophilic inverse agonist (65) displayed high potency in a biochemical assay, which translated into inhibition of IL-17 production in human peripheral blood mononuclear cells. The successful reduction of lipophilicity of this new analog gave rise to additional improvements in ROR selectivity and aqueous kinetic solubility, as well as reduction in plasma protein binding, while maintaining high cellular permeability.

Simultaneous determination of itraconazole, hydroxy itraconazole, keto itraconazole and N-desalkyl itraconazole concentration in human plasma using liquid chromatography with tandem mass spectrometry.

A high-performance liquid chromatography tandem mass spectrometry (LC-MS/MS) assay was developed and validated for simultaneous determination of itraconazole (ITZ), hydroxy-itraconazole (OH-ITZ), keto-itraconazole (keto-ITZ) and N-desalkyl itraconazole (ND-ITZ) concentration in human plasma. One hundred and fifty microliters of human plasma were extracted using a solid-supported liquid extraction (SLE) method and the final extracts were analyzed using reverse-phase chromatography and positive electrospray ionization mass spectrometry. The standard curve range is 5-2500 ng/mL for ITZ and OH-ITZ and 0.4-200 ng/mL for keto-ITZ and ND-ITZ. The curve was fitted to a 1/x(2) weighted linear regression model for all analytes. The precision and accuracy of the LC-MS/MS assay based on the five analytical quality control (QC) levels were well within the acceptance criteria from both FDA and EMA guidance for bioanalytical method validation. Average extraction recovery was 97.4% for ITZ, 112.9% for OH-ITZ, 103.4% for keto-ITZ, and 102.3% for ND-ITZ across their respective curve range. Matrix factor was close to 1.0 at both high and low QC levels of all 4 analytes, which indicates minimal ion suppression or enhancement in our validated assay. Itraconazole and all three metabolites are stable in human plasma for 145 days stored at -70 °C freezers. The validated assay was successfully applied to a clinical study, which has a drug-drug interaction (DDI) arm using ITZ as a cytochrome P450, family 3, subfamily A (CYP3A) inhibitor.

Surrogate analyte approach for quantitation of endogenous NAD(+) in human acidified blood samples using liquid chromatography coupled with electrospray ionization tandem mass spectrometry.

A high-performance liquid chromatography tandem mass spectrometry (LC-MS/MS) assay for the quantitative determination of NAD(+) in human whole blood using a surrogate analyte approach was developed and validated. Human whole blood was acidified using 0.5N perchloric acid at a ratio of 1:3 (v:v, blood:perchloric acid) during sample collection. 25µL of acidified blood was extracted using a protein precipitation method and the resulting extracts were analyzed using reverse-phase chromatography and positive electrospray ionization mass spectrometry. (13)C5-NAD(+) was used as the surrogate analyte for authentic analyte, NAD(+). The standard curve ranging from 0.250 to 25.0µg/mL in acidified human blood for (13)C5-NAD(+) was fitted to a 1/x(2) weighted linear regression model. The LC-MS/MS response between surrogate analyte and authentic analyte at the same concentration was obtained before and after the batch run. This response factor was not applied when determining the NAD(+) concentration from the (13)C5-NAD(+) standard curve since the percent difference was less than 5%. The precision and accuracy of the LC-MS/MS assay based on the five analytical QC levels were well within the acceptance criteria from both FDA and EMA guidance for bioanalytical method validation. Average extraction recovery of (13)C5-NAD(+) was 94.6% across the curve range. Matrix factor was 0.99 for both high and low QC indicating minimal ion suppression or enhancement. The validated assay was used to measure the baseline level of NAD(+) in 29 male and 21 female human subjects. This assay was also used to study the circadian effect of endogenous level of NAD(+) in 10 human subjects.

Absorption, Metabolism, Excretion, and the Contribution of Intestinal Metabolism to the Oral Disposition of [14C]Cobimetinib, a MEK Inhibitor, in Humans.

The pharmacokinetics, metabolism, and excretion of cobimetinib, a MEK inhibitor, were characterized in healthy male subjects (n = 6) following a single 20 mg (200 µCi) oral dose. Unchanged cobimetinib and M16 (glycine conjugate of hydrolyzed cobimetinib) were the major circulating species, accounting for 20.5% and 18.3% of the drug-related material in plasma up to 48 hours postdose, respectively. Other circulating metabolites were minor, accounting for less than 10% of drug-related material in plasma. The total recovery of the administered radioactivity was 94.3% (±1.6%, S.D.) with 76.5% (±2.3%) in feces and 17.8% (±2.5%) in urine. Metabolite profiling indicated that cobimetinib had been extensively metabolized with only 1.6% and 6.6% of the dose remaining as unchanged drug in urine and feces, respectively. In vitro phenotyping experiments indicated that CYP3A4 was predominantly responsible for metabolizing cobimetinib. From this study, we concluded that cobimetinib had been well absorbed (fraction absorbed, Fa = 0.88). Given this good absorption and the previously determined low hepatic clearance, the systemic exposures were lower than expected (bioavailability, F = 0.28). We hypothesized that intestinal metabolism had strongly attenuated the oral bioavailability of cobimetinib. Supporting this hypothesis, the fraction escaping gut wall elimination (Fg) was estimated to be 0.37 based on F and Fa from this study and the fraction escaping hepatic elimination (Fh) from the absolute bioavailability study (F = Fa × Fh × Fg). Physiologically based pharmacokinetics modeling also showed that intestinal clearance had to be included to adequately describe the oral profile. These collective data suggested that cobimetinib was well absorbed following oral administration and extensively metabolized with intestinal first-pass metabolism contributing to its disposition.

Identification of N-sulfonyl-tetrahydroquinolines as RORc inverse agonists.

A high-throughput screen of the Genentech/Roche compound collection using a retinoic acid receptor-related orphan receptor C (RORc, RORγ, or NR1F3) biochemical assay revealed a N-sulfonyl-tetrahydroquinoline hit. Herein, we describe the hit-to-lead optimization and structure-activity relationships of these tetrahydroquinoline RORc inverse agonists. Through iterative synthesis and analog design, we identified compounds with improved biochemical RORc inverse agonist activity and RORc cellular potencies. These improved N-sulfonyl-tetrahydroquinoline compounds also exhibited selectivity for RORc over other nuclear receptors.

Discovery of 1-{4-[3-fluoro-4-((3s,6r)-3-methyl-1,1-dioxo-6-phenyl-[1,2]thiazinan-2-ylmethyl)-phenyl]-piperazin-1-yl}-ethanone (GNE-3500): a potent, selective, and orally bioavailable retinoic acid receptor-related orphan receptor C (RORc or RORγ) inverse agonist.

Retinoic acid receptor-related orphan receptor C (RORc, RORγ, or NR1F3) is a nuclear receptor that plays a major role in the production of interleukin (IL)-17. Considerable efforts have been directed toward the discovery of selective RORc inverse agonists as potential treatments of inflammatory diseases such as psoriasis and rheumatoid arthritis. Using the previously reported tertiary sulfonamide 1 as a starting point, we engineered structural modifications that significantly improved human and rat metabolic stabilities while maintaining a potent and highly selective RORc inverse agonist profile. The most advanced δ-sultam compound, GNE-3500 (27, 1-{4-[3-fluoro-4-((3S,6R)-3-methyl-1,1-dioxo-6-phenyl-[1,2]thiazinan-2-ylmethyl)-phenyl]-piperazin-1-yl}-ethanone), possessed favorable RORc cellular potency with 75-fold selectivity for RORc over other ROR family members and >200-fold selectivity over 25 additional nuclear receptors in a cell assay panel. The favorable potency, selectivity, in vitro ADME properties, in vivo PK, and dose-dependent inhibition of IL-17 in a PK/PD model support the evaluation of 27 in preclinical studies.