A Bayesian neural network approach for modelling censored data with an application to prognosis after surgery for breast cancer

A Bayesian framework is introduced to carry out Automatic Relevance Determination (ARD) in feedforward neural networks to model censored data. A procedure to identify and interpret the prognostic group allocation is also described. These methodologies are applied to 1616 records routinely collected at Christie Hospital, in a monthly cohort study with 5-year follow-up. Two cohort studies are presented, for low- and high-risk patients allocated by standard clinical staging. The results of contrasting the Partial Logistic Artificial Neural Network (PLANN)-ARD model with the proportional hazards model are that the two are consistent, but the neural network may be more specific in the allocation of patients into prognostic groups. With automatic model selection, the regularised neural network is more conservative than the default stepwise forward selection procedure implemented by SPSS with the Akaike Information Criterion.

An Update on Avian Influenza in Hong Kong 2002

An outbreak of highly pathogenic avian influenza caused by multiple genotypes of H5N1 virus occurred in Hong Kong, commencing in January 2002. Infection in local chicken farms was preceded by the detection of virus in multiple retail markets and the main poultry wholesale market. The first case of this disease on a local farm was detected on February 1, 2002. By February 9, 2002, 15 farms were infected, and by late March a total of 22 infected farms had been identified. Three main clusters of infected farms were seen, suggesting multiple incursions of virus, and subsequent limited lateral spread to neighboring firms. Control of this disease has been effected through a combination of quarantine, tightening of biosecurity measures, and depopulation of infected and contact farms. About 950,000 birds have been destroyed. Vaccination using a killed H5 vaccine was introduced in April 2002 to farms in one zone where infection has persisted. None of the viruses isolated contained the internal genes found in the 1997 H5N1 virus.

A Bayesian neural network approach for modelling censored data with an application to prognosis after surgery for breast cancer

A Bayesian framework is introduced to carry out Automatic Relevance Determination (ARD) in feedforward neural networks to model censored data. A procedure to identify and interpret the prognostic group allocation is also described. These methodologies are applied to 1616 records routinely collected at Christie Hospital, in a monthly cohort study with 5-year follow-up. Two cohort studies are presented, for low- and high-risk patients allocated by standard clinical staging. The results of contrasting the Partial Logistic Artificial Neural Network (PLANN)-ARD model with the proportional hazards model are that the two are consistent, but the neural network may be more specific in the allocation of patients into prognostic groups. With automatic model selection, the regularised neural network is more conservative than the default stepwise forward selection procedure implemented by SPSS with the Akaike Information Criterion.

Early dialysis in a neonate with intrauterine lisinopril exposure

In general, angiotensin converting enzyme (ACE) inhibitors should be discontinued in pregnancy, as they can induce an ACE fetopathy. For the treatment of the latter, early peritoneal dialysis is recommended for in utero exposure to captopril and enalapril, although the outcome is poor. Early peritoneal dialysis has not previously been reported for lisinopril induced multiorgan failure. A case is reported in which treatment was given on postnatal day 3. The patient recovered from oligoanuria to almost normal renal function, and heart, brain, and musculoskeletal injury was reversible. This is despite relatively poor clearance of the drug through peritoneal dialysis. Analysis of the pharmacokinetic data suggests that haemodialysis or haemofiltration would be more efficacious for removal of the drug, and these treatments should be performed if available.

Expression of the neuropeptide Y Y1 receptor in the CNS of rat and of wild-type and Y1 receptor knock-out mice. Focus on immunohistochemical localization

The distribution of neuropeptide Y (NPY) Y1 receptor-like immunoreactivity (Y1R-LI) has been studied in detail in the CNS of rat using a rabbit polyclonal antibody against the C-terminal 13 amino acids of the rat receptor protein. The indirect immunofluorescence technique with tyramide signal amplification has been employed. For specificity and comparative reasons Y1 knock-out mice and wild-type controls were analyzed. The distribution of Y1R mRNA was also studied using in situ hybridization. A limited comparison between Y1R-LI and NPY-LI was carried out.A widespread and abundant distribution of Y1R-LI, predominantly in processes but also in cell bodies, was observed. In fact, Y1R-LI was found in most regions of the CNS with a similar distribution pattern between rat and wild-type mouse. This staining was specific in the sense that it was absent in adjacent sections following preadsorption of the antibody with 10(-5) M of the antigenic peptide, and that it could not be observed in sections of the Y1 KO mouse. In contrast, the staining obtained with an N-terminally directed Y1R antiserum did not disappear, strongly suggesting unspecificity. In brief, very high levels of Y1R-LI were seen in the islands of Calleja, the anterior olfactory nucleus, the molecular layer of the dentate gyrus, parts of the habenula, the interpeduncular nucleus, the mammillary body, the spinal nucleus of the trigeminal, caudal part, the paratrigeminal nucleus, and superficial layers of the dorsal horn. High levels were found in most cortical areas, many thalamic nuclei, some subnuclei of the amygdaloid complex, the hypothalamus and the nucleus of the stria terminalis, the nucleus of the solitary tract, the parabrachial nucleus, and the inferior olive. Moderate levels of Y1R-LI were detected in the cornu Ammonis and the subicular complex, many septal, some thalamic and many brainstem regions. Y1R staining of processes, often fiber and/or dot-like, and occasional cell bodies was also seen in tracts, such as the lateral lemniscus, the rubrospinal tract and the spinal tract of the trigeminal. There was in general a good overlap between Y1R-LI and NPY-LI, but some exceptions were found. Thus, some areas had NPY innervation but apparently lacked Y1Rs, whereas in other regions Y1R-LI, but no or only few NPY-positive nerve endings could be detected. Our results demonstrate that NPY signalling through the Y1R is common in the rat (and mouse) CNS. Mostly the Y1R is postsynaptic but there are also presynaptic Y1Rs. Mostly there is a good match between NPY-releasing nerve endings and Y1Rs, but 'volume transmission' may be 'needed' in some regions. Finally, the importance of using proper control experiments for immunohistochemical studies on seven-transmembrane receptors is stressed.

CCK(B)/gastrin receptor mediates synergistic stimulation of DNA synthesis and cyclin D1, D3, and E expression in Swiss 3T3 cells

In order to develop a model system for identifying signaling pathways and cell cycle events involved in gastrin-mediated mitogenesis, we have used high efficiency retroviral-mediated transfection of cholecystokinin (CCK)(B)/gastrin receptor into Swiss 3T3 cells. The retrovirally-transfected CCK(B)/gastrin receptor binds 125I-CCK-8 with high affinity (Kd = 1.1 nM) and is functionally coupled to intracellular signaling pathways including rapid and transient increase in Ca2+ fluxes, protein kinase C-dependent protein kinase D activation, and MEK-dependent ERK1/2 activation. In the presence of insulin, CCK-8 or gastrin induced a 66.5 +/- 8.8-fold (mean +/- SEM, n = 24 in eight independent experiments) increase in cellular DNA synthesis, reaching a level similar to that achieved by stimulation with a saturating concentration of fresh serum, and much greater than the response to each agonist added alone. CCK-8 also induced a striking increase in the expression of cyclins D1, D3, and E and hyperphosphorylation of Rb acting synergistically with insulin. Similar effects were observed when CCK(B)/gastrin receptor was activated in the presence of EGF or bombesin. Our results demonstrate that activation of CCK(B)/gastrin receptor retrovirally-transfected into Swiss 3T3 induces a potent synergistic effect on DNA synthesis, accumulation of cyclins D1, D3, and E and hyperphosphorylation of Rb in combination with insulin, EGF, or bombesin. Thus, the CCK(B)/gastrin receptor transfected into Swiss 3T3 cells provides a novel model system to elucidate mitogenic signal transduction pathways and cell cycle events activated via this receptor.

Synthesis and antitumor activity of novel C-7 paclitaxel ethers: discovery of BMS-184476

The preparation of C-7 paclitaxel ethers is described. Various substituted ethers were prepared via activation of the corresponding methylthiomethyl ether followed by alcohol addition. Variation of the C-7 ether group as well the 3' side chain position led to the discovery of a novel taxane, BMS-184476 (4), with preclinical antitumor activity superior to paclitaxel.

CCK(B)/gastrin receptor mediates synergistic stimulation of DNA synthesis and cyclin D1, D3, and E expression in Swiss 3T3 cells

In order to develop a model system for identifying signaling pathways and cell cycle events involved in gastrin-mediated mitogenesis, we have used high efficiency retroviral-mediated transfection of cholecystokinin (CCK)(B)/gastrin receptor into Swiss 3T3 cells. The retrovirally-transfected CCK(B)/gastrin receptor binds 125I-CCK-8 with high affinity (Kd = 1.1 nM) and is functionally coupled to intracellular signaling pathways including rapid and transient increase in Ca2+ fluxes, protein kinase C-dependent protein kinase D activation, and MEK-dependent ERK1/2 activation. In the presence of insulin, CCK-8 or gastrin induced a 66.5 +/- 8.8-fold (mean +/- SEM, n = 24 in eight independent experiments) increase in cellular DNA synthesis, reaching a level similar to that achieved by stimulation with a saturating concentration of fresh serum, and much greater than the response to each agonist added alone. CCK-8 also induced a striking increase in the expression of cyclins D1, D3, and E and hyperphosphorylation of Rb acting synergistically with insulin. Similar effects were observed when CCK(B)/gastrin receptor was activated in the presence of EGF or bombesin. Our results demonstrate that activation of CCK(B)/gastrin receptor retrovirally-transfected into Swiss 3T3 induces a potent synergistic effect on DNA synthesis, accumulation of cyclins D1, D3, and E and hyperphosphorylation of Rb in combination with insulin, EGF, or bombesin. Thus, the CCK(B)/gastrin receptor transfected into Swiss 3T3 cells provides a novel model system to elucidate mitogenic signal transduction pathways and cell cycle events activated via this receptor.

Liquid chromatography-mass spectrometry assay of a thiadiazole derivative in mice: application to pharmacokinetic studies

Modern atmospheric pressure ionization (API) ion-trap mass spectrometry in connection with fast chromatographic separations using a short narrow-bore C8 column was developed to determine 5-phenyl-3-thioureido-1,2,4-thiadiazole (301029), a novel virus inhibitor in serum. Both 301029 and an internal standard (I.S.) were separated from serum samples by acetonitrile deproteinization and extraction without time-consuming reconstitution. The chromatographic separation was achieved on a C8 reversed-phase narrow-bore column using acetonitrile-water-acetic acid (90:10:0.01, v/v/v) as a mobile phase. The mass spectrometric analysis was performed by atmospheric pressure chemical ionization (APCI) mode with positive ion detection. Single ion monitoring (SIM) scan mode of m/z 237 and 158 was used to quantitatively determine 301029 and I.S., respectively. The low limit of quantitation was 25 ng/ml. The assay exhibited a linear range of 25-2500 ng/ml. Recovery from serum proved to be 100-113%. The precision (C.V.) and accuracy (RE) of the method were 2-12% and 94-112%, respectively. The present method was applied to determine the pharmacokinetic parameters of 301029 following oral administration of the agent to mice at 5 g/kg. The results revealed that the elimination half-life of 301029 was 413 min and the area under serum concentration-time curve was 354 microg/ml/min.

In vitro and in vivo assessment of cellular permeability and pharmacodynamics of S-nitrosylated captopril, a nitric oxide donor

1. The present studies were aimed at testing the hypothesis that S-nitrosylated captopril (CapNO), a novel crystalline nitric oxide (NO) donor, readily permeates both in vitro and in vivo endothelial monolayers, resulting in its pharmacodynamic effects. 2. CapNO and Captopril (Cap) were added to apical side of endothelial monolayers formed on microporous membranes, and the permeated drugs were collected from basolateral side and detected by a HPLC method. The permeability coefficient (P(app); cm sec(-1)) of CapNO across the endothelial monolayers was 6.0 x 10(-5), higher than that of Cap (3.13 x 10(-5)), indicating the enhancement effect of the attached NO group in CapNO on cellular permeability. The P(app) of CapNO and Cap across Caco-2 cells were 3.15 x 10(-5) and 1.53 x 10(-5), respectively. The low P(app) of CapNO to Caco-2 cells may be attributed to the high membrane resistance of Caco-2 cells. 3. A bolus injection of CapNO to epicardial coronary artery of chronically-instrumented awake dogs caused significant increases in coronary blood flow and coronary diameters dose-dependently without significant changes in aortic pressure. In contrast, the equimolar doses of Cap did not produce haemodynamic responses. 4. Intravenous CapNO caused an instant increase in the regional cerebral blood flow determined by H(2)-clearance, whereas the equimolar doses of Cap did not enhance the cerebral blood flow. 5. These results conclude that the NO group, an active component of CapNO, enhances both in vitro and in vivo endothelial permeability to the entire compound, resulting in instant increases in blood flow and vascular diameters. In contrast, the equimolar Cap does not have the instant vascular effects.

CCK(B)/gastrin receptor mediates synergistic stimulation of DNA synthesis and cyclin D1, D3, and E expression in Swiss 3T3 cells

In order to develop a model system for identifying signaling pathways and cell cycle events involved in gastrin-mediated mitogenesis, we have used high efficiency retroviral-mediated transfection of cholecystokinin (CCK)(B)/gastrin receptor into Swiss 3T3 cells. The retrovirally-transfected CCK(B)/gastrin receptor binds 125I-CCK-8 with high affinity (Kd = 1.1 nM) and is functionally coupled to intracellular signaling pathways including rapid and transient increase in Ca2+ fluxes, protein kinase C-dependent protein kinase D activation, and MEK-dependent ERK1/2 activation. In the presence of insulin, CCK-8 or gastrin induced a 66.5 +/- 8.8-fold (mean +/- SEM, n = 24 in eight independent experiments) increase in cellular DNA synthesis, reaching a level similar to that achieved by stimulation with a saturating concentration of fresh serum, and much greater than the response to each agonist added alone. CCK-8 also induced a striking increase in the expression of cyclins D1, D3, and E and hyperphosphorylation of Rb acting synergistically with insulin. Similar effects were observed when CCK(B)/gastrin receptor was activated in the presence of EGF or bombesin. Our results demonstrate that activation of CCK(B)/gastrin receptor retrovirally-transfected into Swiss 3T3 induces a potent synergistic effect on DNA synthesis, accumulation of cyclins D1, D3, and E and hyperphosphorylation of Rb in combination with insulin, EGF, or bombesin. Thus, the CCK(B)/gastrin receptor transfected into Swiss 3T3 cells provides a novel model system to elucidate mitogenic signal transduction pathways and cell cycle events activated via this receptor.

Lipopolysaccharide-stimulated or granulocyte-macrophage colony-stimulating factor-stimulated monocytes rapidly express biologically active IL-15 on their cell surface independent of new protein synthesis

Although IL-15 shares many of the biological activities of IL-2, IL-2 expression is primarily under transcriptional regulation, while the mechanisms involved in the regulation of IL-15 are complex and not completely understood. In the current study, we found that CD14(+) monocytes constitutively exhibit both IL-15 mRNA and protein. IL-15 protein was found stored intracellularly and stimulation of CD14(+) monocytes with either LPS or GM-CSF resulted in mobilization of IL-15 stores to the plasma membrane. This rapidly induced surface expression was the result of a translocation of preformed stores, confirming that posttranslational regulatory stages limit IL-15, because it was not accompanied by an increase in IL-15 mRNA and occurred independent of de novo protein synthesis. After fixation, activated monocytes, but not resting monocytes, were found to support T cell proliferation, and this effect was abrogated by the addition of an IL-15-neutralizing Ab. The presence of preformed IL-15 stores and the ability of stimulated monocytes to mobilize these stores to their surface in an active form is a novel mechanism of regulation for IL-15.

Multiwatt mid-IR output from a Nd:YALO laser pumped intracavity KTA OPO

We have achieved 4.1W of 3.5-micron output from a non-critically phasematched (NCPM), type II, KTiOAsO4 (KTA) optical parametric oscillator (OPO) pumped within the cavity of a Q-switched diode-pumped Nd: YALO laser operating at 10kHz. We adopted the simplest configuration with a compact diode-pumped Nd: YALO module pumping the singly resonant KTA OPO. Besides 4.1W of 3.5um, 10.9W of 1.5 micron and 11.3W of 1-micron radiation were obtained simultaneously.

Dose-dependent pharmacokinetics and metabolism of valproic acid in newborn lambs and adult sheep

Dose-dependent pharmacokinetics and metabolism of valproic acid (VPA) were studied in newborn and adult sheep to assess age-related differences in plasma protein binding and metabolic elimination. Newborn lambs received either a 10- (n = 8), 50- (n = 5), 100- (n = 4), or 250-mg/kg (n = 4) VPA i.v. bolus. Individual adult sheep (n = 5) received all four doses in a random order with an appropriate washout period between experiments. Unbound or metabolic clearance of VPA was significantly higher in adult sheep at the two lower doses when compared with lambs, and similar to the lambs at the two higher doses. Plasma protein binding was nonlinear at all doses. Estimates of binding capacity (B(max1)) at the saturable site were higher in adults (91.8 microg/ml) when compared with lambs (44.9 microg/ml), whereas the opposite trend was observed for binding affinity [K(d1) = 9.6 microg/ml (adult) versus 3.2 microg/ml (lambs)]. Characterization of developmental differences in overall VPA metabolic elimination involved fitting of unbound VPA plasma concentration data to a two-compartment model with Michaelis-Menten elimination. This resulted in similar in vivo estimates of apparent V(max) [445.0 microg/min/kg (adult) versus 429.9 microg/min/kg (lambs)]. However, apparent K(m) estimates appeared to be higher in lambs [30.0 microg/ml (adult) versus 69.6 microg/ml (lambs)]. Similar findings were obtained from in vivo estimates of V(max) and K(m) for VPA glucuronidation obtained from VPA-glucuronide metabolite urinary excretion data. Thus, it appears that age-related differences in metabolic clearance may be related to differences in the apparent in vivo K(m) as opposed to V(max) of VPA glucuronidation.

In vivo interactions of apoA-II, apoA-I, and hepatic lipase contributing to HDL structure and antiatherogenic functions

Studies with mice have revealed that increased expression of apolipoprotein A-II (apoA-II) results in elevations in high density lipoprotein (HDL), the formation of larger HDL, and the development of early atherosclerosis. We now show that the increased size of HDL results in part from an inhibition of the ability of hepatic lipase (HL) to hydrolyze phospholipids and triglycerides in the HDL and that the ratio of apoA-I to apoA-II determines HDL functional and antiatherogenic properties. HDL from apoA-II transgenic mice was relatively resistant to the action of HL in vitro. To test whether HL and apoA-II influence HDL size independently, combined apoA-II transgenic/HL knockout (HLko) mice were examined. These mice had HDL similar in size to apoA-II transgenic mice and HLko mice, suggesting that they do not increase HDL side by independent mechanisms. Overexpression of apoA-I from a transgene reversed many of the effects of apoA-II overexpression, including the ability of HDL to serve as a substrate for HL. Combined apoA-I/apoA-II transgenic mice exhibited significantly less atherosclerotic lesion formation than did apoA-II transgenic mice. These results were paralleled by the effects of the transgenes on the ability of HDL to protect against the proinflammatory effects of oxidized low density lipoprotein (LDL). Whereas nontransgenic HDL protected against oxidized LDL induction of adhesion molecules in endothelial cells, HDL from apoA-II transgenic mice was proinflammatory. HDL from combined apoA-I/apoA-II transgenic mice was equally as protective as HDL from nontransgenic mice. Our data suggest that as the ratio of apoA-II to apoA-I is increased, the HDL become larger because of inhibition of HL, and lose their antiatherogenic properties.

[Protective effects of organic acids on human vascular endothelial cells]

OBJECTIVES

In order to examine if organic acids (OA), a kind of natural food components, will prevent from atherosclerosis formation, the effects of four OA, namely chlorogenic acid (CHA), ascorbic acid (AA), citric acid (CA), and malic acid (MA) on human vascular endothelial cells were studied.

METHODS

Primary human umbilical vein endothelial cells (EC) were cultured, and the changes of cell morphology, cell growth, lactate dehydrogenase (LDH) released from cells were measured when EC was incubated with oxidized low density lipoprotein (oxLDL) in the presence or not of four OA added at three different conditions, i.e. before 4 hours oxLDL added, at the same time as oxLDL added, and after 3 hours oxLDL added, as well as at the indicated doses (10, 20 or 40 mg/L).

RESULTS

The EC survival rate of the oxLDL control group was lower (P < 0.01) and LDH release rate was higher (P < 0.01) than those of normal control group. CHA, AA, CA and MA protected EC from damage caused by oxLDL when they were added before and/or at the same time of oxLDL, but the same beneficial effects of these four OA added after oxLDL were not observed. The EC survival rates of CHA + oxLDL and AA + oxLDL groups in three doses were higher (P < 0.05) and the LDH release rates were lower (P < 0.05) in a dose-effect manner when compared with those of oxLDL control group, respectively. Similar results were found in the CA + oxLDL and MA + oxLDL of the high dose groups.

CONCLUSION

Four tested OA could effectively prevent EC from damage induced by oxLDL. CHA and AA had the strongest effects, the next was CA, followed by MA.