Associations Between Plasma Concentrations of Lenvatinib and Angiopoietin and Clinical Responses to Lenvatinib Therapy in Japanese Patients With Thyroid Cancer.

BACKGROUND/AIM: The purpose of this study was to investigate the relationships between the plasma concentration of Lenvatinib (C0), the levels of angiopoietin (Ang)-1 and Ang-2, and clinical responses to lenvatinib therapy in patients with thyroid cancer. PATIENTS AND METHODS: Lenvatinib C 0 and Ang were measured by high-performance liquid chromatography and enzyme-linked immunosorbent assay, respectively. RESULTS: The median decrease rates of Ang-1 and Ang-2 at 1 month after treatment from baseline were -15.3% and -48.4%, respectively. However, the decrease in the levels of Ang-1 and Ang-2 at 1 month from baseline did not correlate with C0. In patients with partial response (PR) and stable disease, Ang-2 at 1 month was significantly lower than Ang-2 at baseline. The area under the ROC for PR prediction was 0.667, giving the best sensitivity (69.2%) and specificity (73.9%) at a threshold of decrease rate of Ang-2 of -49.83%. CONCLUSION: The decrease in Ang-2 at 1 month of treatment from baseline may be important as a biomarker of the inhibitory effect of lenvatinib on angiogenesis.

Reducing the undesirable odor of barley by cooking with superheated steam.

Superheated steam was used to cook barley and the volatile odor compounds and release of odorants from the steamed barley were analyzed. The main odor compounds in cooked barley were aldehydes (hexanal and (E,E)-2,4-decadienal) and acids (acetic acid and hexanoic acid). Compared to ordinary cooked barley, barley cooked by superheated steam had less odorants, and the release of odorants was reduced by almost half. Sensory evaluation revealed that this barley was preferred to ordinary cooked barley, because it had weaker smell and tasted less sour and less bitter. The steaming process steam distils and eliminates some odor compounds, while some water-soluble compounds (mainly acids) are washed away by water during steaming. Therefore, this steam cooking method, applied to barley for the first time here using a comprehensive analysis, improves the acceptability and palatability of this high-quality food rich in dietary fiber.

Influence of CYP3A4/5 and ABC transporter polymorphisms on lenvatinib plasma trough concentrations in Japanese patients with thyroid cancer.

Lenvatinib is a substrate of cytochrome P450 (CYP) 3A and ATP-binding cassette (ABC) transporters. In this study, we aimed to evaluate how CYP3A4/5 and ABC transporter polymorphisms affected the mean steady-state dose-adjusted plasma trough concentrations (C0) of lenvatinib in a cohort of 40 Japanese patients with thyroid cancer. CYP3A4 20230G > A (*1G), CYP3A5 6986A > G (*3), ABCB1 1236C > T, ABCB1 2677G > T/A, ABCB1 3435C > T, ABCC2 -24C > T, and ABCG2 421C > A genotypes were determined using polymerase chain reaction-restriction fragment length polymorphism. In univariate analysis, there were no significant differences in the mean dose-adjusted C0 values of lenvatinib between the ABCB1, ABCG2, and CYP3A5 genotypes. However, the mean dose-adjusted C0 values of lenvatinib in patients with the CYP3A4*1/*1 genotype and ABCC2 -24T allele were significantly higher than those in patients with the CYP3A4*1G allele and -24C/C genotype, respectively (P = 0.018 and 0.036, respectively). In multivariate analysis, CYP3A4 genotype and total bilirubin were independent factors influencing the dose-adjusted C0 of lenvatinib (P = 0.010 and 0.046, respectively). No significant differences were found in the incidence rates of hypertension, proteinuria, and hand-foot syndrome following treatment with lenvatinib between the genotypes of CYP3A4/5 and ABC transporters. Lenvatinib pharmacokinetics were significantly influenced by the CYP3A4*1G polymorphism. If the target plasma concentration of lenvatinib for efficacy or toxicity is determined, elucidation of the details of the CYP3A4*1G genotype may facilitate decision-making related to the appropriate initial lenvatinib dosage to achieve optimal plasma concentrations.

Association of lenvatinib trough plasma concentrations with lenvatinib-induced toxicities in Japanese patients with thyroid cancer.

The aim of this study was to examine the association of lenvatinib-induced adverse events with the trough plasma concentration (C0) in Japanese patients with thyroid cancer. Patients received lenvatinib 24 mg as an initial dose, and sequential dose reductions were conducted based on the grade of each side effect. Assessment of adverse events, assay of lenvatinib C0, and analysis of clinical laboratory tests were performed at the same time of day and were retrospectively analyzed. There were no significant differences in lenvatinib C0 among grades of hypertension, proteinuria, hand-foot syndrome, and diarrhea. However, levels of aspartate transaminase, alanine transaminase, and total bilirubin were significantly higher in lenvatinib C0 quartile (Q) 4 (≥ 88 ng/mL) than in Q1 (< 42 ng/mL) and Q2-3 (42-88 ng/mL). Additionally, platelet counts were highest in the lowest Q1 group. The median dose of lenvatinib in patients with UGT1A1*6/*6 or *6/*28 (poor metabolizers [PMs]) was significantly lower than that in patients with UGT1A1*1/*1 (10 and 14 mg, respectively), whereas the median bilirubin levels were significant higher in UGT1A1 PMs (0.9 and 0.5 mg/dL, respectively). There were no significant differences in median lenvatinib C0 values between patients with UGT1A1*1/*1 and PMs (58.0 and 50.0 ng/mL, respectively). The threshold between the C0 and toxicity of lenvatinib may be more than 88 ng/mL. Therefore, the dose of lenvatinib could be controlled to maintain a lower C0 of less than 88 ng/mL. The target C0 for lenvatinib as the threshold between the C0 and optimal response may be in the range from 42 to 88 ng/mL; however, further studies are necessary.

Dynamics of Learning in MLP: Natural Gradient and Singularity Revisited.

The dynamics of supervised learning play a main role in deep learning, which takes place in the parameter space of a multilayer perceptron (MLP). We review the history of supervised stochastic gradient learning, focusing on its singular structure and natural gradient. The parameter space includes singular regions in which parameters are not identifiable. One of our results is a full exploration of the dynamical behaviors of stochastic gradient learning in an elementary singular network. The bad news is its pathological nature, in which part of the singular region becomes an attractor and another part a repulser at the same time, forming a Milnor attractor. A learning trajectory is attracted by the attractor region, staying in it for a long time, before it escapes the singular region through the repulser region. This is typical of plateau phenomena in learning. We demonstrate the strange topology of a singular region by introducing blow-down coordinates, which are useful for analyzing the natural gradient dynamics. We confirm that the natural gradient dynamics are free of critical slowdown. The second main result is the good news: the interactions of elementary singular networks eliminate the attractor part and the Milnor-type attractors disappear. This explains why large-scale networks do not suffer from serious critical slowdowns due to singularities. We finally show that the unit-wise natural gradient is effective for learning in spite of its low computational cost.

Characterization and Clinical Implication of Th1/Th2/Th17 Cytokines Produced from Three-Dimensionally Cultured Tumor Tissues Resected from Breast Cancer Patients.

OBJECTIVES: Several cytokines secreted from breast cancer tissues are suggested to be related to disease prognosis. We examined Th1/Th2/Th17 cytokines produced from three-dimensionally cultured breast cancer tissues and related them with patient clinical profiles. METHODS: 21 tumor tissues and 9 normal tissues surgically resected from breast cancer patients were cultured in thermoreversible gelatin polymer-containing medium. Tissue growth and Th1/Th2/Th17 cytokine concentrations in the culture medium were analyzed and were related with hormone receptor expressions and patient clinical profiles. RESULTS: IL-6 and IL-10 were expressed highly in culture medium of both cancer and normal tissues. However, IFN-γ, TNF-α, IL-2, and IL-17A were not detected in the supernatant of the three-dimensionally cultured normal mammary gland and are seemed to be specific to breast cancer tissues. The growth abilities of hormone receptor-negative cancer tissues were significantly higher than those of receptor-positive tissues (P=0.0383). Cancer tissues of stage ≥IIB patients expressed significantly higher TNF-α levels as compared with those of patients with stage

Neurodegenerative disorder FTDP-17-related tau intron 10 +16C → T mutation increases tau exon 10 splicing and causes tauopathy in transgenic mice.

Frontotemporal dementia and parkinsonism linked to chromosome 17 (FTDP-17) is a neurodegenerative disorder caused by mutations in the tau gene. Many mutations identified in FTDP-17 have been shown to affect tau exon 10 splicing in vitro, which presumably causes pathologic imbalances in exon 10(-) [3-repeat (3R)] and exon 10(+) [4-repeat (4R)] tau expression and leads to intracellular inclusions of hyperphosphorylated tau in patient brains. However, no reports have investigated this theory using model mice with a tau intronic mutation. Herein, we generated new transgenic mice harboring the tau intron 10 +16C → T mutation. We prepared a transgene construct containing intronic sequences required for exon 10 splicing in the longest tau isoform cDNA. Although mice bearing the construct without the intronic mutation showed normal developmental changes of the tau isoform from 3R tau to equal amounts of 3R and 4R tau, mice with the mutation showed much higher levels of 4R tau at the adult stage. 4R tau was selectively recovered in insoluble brain fractions in their old age. Furthermore, these mice displayed abnormal tau phosphorylation, synapse loss and dysfunction, memory impairment, glial activation, tangle formation, and neuronal loss in an age-dependent manner. These findings provide the first evidence in a mouse model that a tau intronic mutation-induced imbalance of 3R and 4R tau could be a cause of tauopathy.

Relationship of dietary intake and lifestyle factors to health-related quality of life in the community-dwelling elderly.

Quality of life (QOL) encompasses a broad notion of health and is increasingly used to evaluate the effectiveness of health care services. The purpose of this cross-sectional study was to assess health-related quality of life (HR-QOL) in the community-dwelling elderly (mean age, 72.7±0.3 y; female 46.4%, n=179) in Japan and to explore diminished factors. We assessed HR-QOL by the 36-Item Short-Form Health Survey (SF-36), maximum walking speed, step counts, Tokyo Metropolitan Institute of Gerontology (TMIG) index, dietary intake and blood biochemistry. Two summary scores of the SF-36 were calculated (physical component score: PCS, mental component score: MCS). We divided participants into two groups based on the standard values of PCS and MCS classified by age (high and low group), which was further stratified into a high PCS and high MCS (BH) group and a low PCS and low MCS (BL) group. Mean maximum walking speed and daily step counts were 207.7±2.8 cm/s and 7,008±289. Eighty-one percent of the participants had full scores in the TMIG index. Daily intake of energy and protein were 2,081±33 kcal and 74.5±0.9 g. The maximum walking speed, TMIG index, alcohol consumption and chewing ability were significantly higher in the BH group than those in the BL group, but not dietary intake. Stepwise logistic regression analysis indicated that maintaining maximum walking speed might be associated with sustaining HR-QOL in the healthy elderly.

Dynamics of learning in multilayer perceptrons near singularities.

The dynamical behavior of learning is known to be very slow for the multilayer perceptron, being often trapped in the "plateau." It has been recently understood that this is due to the singularity in the parameter space of perceptrons, in which trajectories of learning are drawn. The space is Riemannian from the point of view of information geometry and contains singular regions where the Riemannian metric or the Fisher information matrix degenerates. This paper analyzes the dynamics of learning in a neighborhood of the singular regions when the true teacher machine lies at the singularity. We give explicit asymptotic analytical solutions (trajectories) both for the standard gradient (SGD) and natural gradient (NGD) methods. It is clearly shown, in the case of the SGD method, that the plateau phenomenon appears in a neighborhood of the critical regions, where the dynamical behavior is extremely slow. The analysis of the NGD method is much more difficult, because the inverse of the Fisher information matrix diverges. We conquer the difficulty by introducing the "blow-down" technique used in algebraic geometry. The NGD method works efficiently, and the state converges directly to the true parameters very quickly while it staggers in the case of the SGD method. The analytical results are compared with computer simulations, showing good agreement. The effects of singularities on learning are thus qualitatively clarified for both standard and NGD methods.

Dynamics of learning near singularities in layered networks.

We explicitly analyze the trajectories of learning near singularities in hierarchical networks, such as multilayer perceptrons and radial basis function networks, which include permutation symmetry of hidden nodes, and show their general properties. Such symmetry induces singularities in their parameter space, where the Fisher information matrix degenerates and odd learning behaviors, especially the existence of plateaus in gradient descent learning, arise due to the geometric structure of singularity. We plot dynamic vector fields to demonstrate the universal trajectories of learning near singularities. The singularity induces two types of plateaus, the on-singularity plateau and the near-singularity plateau, depending on the stability of the singularity and the initial parameters of learning. The results presented in this letter are universally applicable to a wide class of hierarchical models. Detailed stability analysis of the dynamics of learning in radial basis function networks and multilayer perceptrons will be presented in separate work.

CYP2D6-Mediated metabolism of a novel acyl coenzyme A:cholesterol acyltransferase inhibitor, pactimibe, and its unique plasma metabolite, R-125528.

Pactimibe sulfate is a novel acyl coenzyme A:cholesterol acyltransferase inhibitor. We conducted metabolic studies of pactimibe and its plasma metabolite, R-125528. Pactimibe had multiple metabolic pathways including indolin oxidation to form R-125528, omega-1 oxidation, N-dealkylation, and glucuronidation. Among them, the indolin oxidation and the omega-1 oxidation were dominant and were mainly catalyzed by CYP3A4 and CYP2D6, respectively. The intrinsic clearance (CL(int)) values for these pathways in human hepatic microsomes were 0.63 and 0.76 microl/min/mg-protein, respectively. On the other hand, the metabolic reaction for R-125528 was restricted. It was demonstrated that omega-1 oxidation was the only pathway that could eliminate R-125528 from the systemic circulation. To our surprise, only CYP2D6-expressing microsomes could catalyze the reaction, and omega-1 oxidation was strongly correlated with the CYP2D6 marker reaction, dextromethorphan O-demethylation (r(2) = 0.90), in human hepatic microsomes. Although R-125528 is an atypical substrate for CYP2D6 because of its acidity, the K(m) value was 1.8 microM for the reaction in human hepatic microsomes and the CL(int) value was as high as 75.0 microl/min/mg-protein. These results suggested that the systemic clearance of R-125528 was highly dependent on CYP2D6 activity and that several studies with CYP2D6 including drug-drug interaction and polymorphism sensitivity should be performed during development from the viewpoint of metabolite safety assessment. The finding that R-125528, an acidic compound devoid of basic nitrogen, was a good substrate for CYP2D6 raised a question about previously reported CYP2D6 models based on a critical electrostatic interaction with Asp(301) and/or Glu(216).

Indoline derivatives II: synthesis and factor Xa (FXa) inhibitory activities.

Factor Xa (FXa) is well known to play a pivotal role in blood coagulation, so FXa inhibitor is a promising drug candidate for prophylaxis and treatment of thromboembolic diseases. In the course of our research, we have found that (R)-5-[1-(acetimidoyl)piperidin-4-yloxy]-2-(7-amidinonaphthalen-2-yl)-1-(ethanesulfonyl)indoline ((R)-1) showed potent FXa inhibitory activity in vitro. However, single oral administation (RS)-1 showed high toxicity in mice. Among newly synthesized compounds, ({(RS)-5-[1-(acetimidoyl)piperidin-4-yloxy]-2-(7-amidinonaphthalen-2-yl)indolin-1-yl}sulfonyl)acetic acid ((RS)-11d) showed more potent FXa inhibitory activity and higher safety than (RS)-1. The R-isoform of compound 11d ((R)-11d) exhibited potent in vitro anticoagulant activity in human and hamster plasma. Orally administered (R)-11d also showed dose-dependent potent anticoagulant activity in hamsters, marmosets and cynomolgus monkeys. Compound (R)-11d with potent anticoagulant activity and high safety is therefore favorable as a novel oral FXa inhibitor.

Singularities affect dynamics of learning in neuromanifolds.

The parameter spaces of hierarchical systems such as multilayer perceptrons include singularities due to the symmetry and degeneration of hidden units. A parameter space forms a geometrical manifold, called the neuromanifold in the case of neural networks. Such a model is identified with a statistical model, and a Riemannian metric is given by the Fisher information matrix. However, the matrix degenerates at singularities. Such a singular structure is ubiquitous not only in multilayer perceptrons but also in the gaussian mixture probability densities, ARMA time-series model, and many other cases. The standard statistical paradigm of the Cramér-Rao theorem does not hold, and the singularity gives rise to strange behaviors in parameter estimation, hypothesis testing, Bayesian inference, model selection, and in particular, the dynamics of learning from examples. Prevailing theories so far have not paid much attention to the problem caused by singularity, relying only on ordinary statistical theories developed for regular (nonsingular) models. Only recently have researchers remarked on the effects of singularity, and theories are now being developed. This article gives an overview of the phenomena caused by the singularities of statistical manifolds related to multilayer perceptrons and gaussian mixtures. We demonstrate our recent results on these problems. Simple toy models are also used to show explicit solutions. We explain that the maximum likelihood estimator is no longer subject to the gaussian distribution even asymptotically, because the Fisher information matrix degenerates, that the model selection criteria such as AIC, BIC, and MDL fail to hold in these models, that a smooth Bayesian prior becomes singular in such models, and that the trajectories of dynamics of learning are strongly affected by the singularity, causing plateaus or slow manifolds in the parameter space. The natural gradient method is shown to perform well because it takes the singular geometrical structure into account. The generalization error and the training error are studied in some examples.