Joint regression and association modeling of longitudinal ordinal data

We propose models for longitudinal, or otherwise clustered, ordinal data. The association between subunit responses is characterized by dependence ratios (Ekholm, Smith, and McDonald, 1995, Biometrika 82, 847-854), which are extended from the binary to the multicategory case. The joint probabilities of the subunit responses are expressed as explicit functions of the marginal means and the dependence ratios of all orders, obtaining a computational advantage for likelihood-based inference. Equal emphasis is put on finding regression models for the univariate cumulative probabilities, and on deriving the dependence ratios from meaningful association-generating mechanisms. A data set on the effects of treatment with Fluvoxamine, which has been analyzed in parts before (Molenberghs, Kenward, and Lesaffre, 1997, Biometrika 84, 33-44), is analyzed in its entirety. Selection models are used for studying the sensitivity of the results to drop-out.     

Coenzyme Q10: absorption, antioxidative properties, determinants, and plasma levels

The purpose of this article is to summarise our studies, in which the main determinants and absorption of plasma coenzyme Q10 (Q10, ubiquinone) have been assessed, and the effects of moderate dose oral Q10 supplementation on plasma antioxidative capacity, lipoprotein oxidation resistance and on plasma lipid peroxidation investigated. All the supplementation trials carried out have been blinded and placebo-controlled clinical studies. Of the determinants of Q10, serum cholesterol, serum triglycerides, male gender, alcohol consumption and age were found to be associated positively with plasma Q10 concentration. A single dose of 30 mg of Q10, which is the maximum daily dose recommended by Q10 producers, had only a marginal elevating effect on plasma Q10 levels in non-Q10-deficient subjects. Following supplementation, a dose-dependent increase in plasma Q10 levels was observed up to a daily dose of 200 mg, which resulted in a 6.1-fold increase in plasma Q10 levels. However, simultaneous supplementation with vitamin E resulted in lower plasma Q10 levels. Of the lipid peroxidation measurements, Q10 supplementation did not increase LDL TRAP, plasma TRAP, VLDL+LDL oxidation resistance nor did it decrease LDL oxidation susceptibility ex vivo. Q10 with minor vitamin E dose neither decreased exercise-induced lipid peroxidation ex vivo nor muscular damage. Q10 supplementation might, however, decrease plasma lipid peroxidation in vivo , as assessed by the increased proportion of plasma ubiquinol (reduced form, Q10H 2 ) of total Q10. High dose vitamin E supplementation decreased this proportion, which suggests in vivo regeneration of tocopheryl radicals by ubiquinol.     

Impact of polymer surface affinity of novel antifouling agents

In a previous study we found two agents, the alpha(2)-agonist medetomidine ((+/-)-4-[1-(2,3-dimethylphenyl)ethyl]-1H-imidazole) and the alpha(2)-agonist clonidine (2-(2,6-dichloroanilino)-2-imidazoline), that specifically and efficiently impede settlement of the barnacle Balanus improvisus, one of the most serious biofouling organisms in Swedish waters. Medetomidine, but not clonidine, is known to adsorb to solid polystyrene (PS) surfaces in the presence of salt, a feature that is of particular interest in attempts to develop an efficient antifouling surface. We show that medetomidine, but not clonidine, has a significant ability to adsorb to untreated (hydrophobic) PS in two different incubation media: filtered seawater (FSW) and deionized water (mQ). At negatively charged (hydrophilic) PS, medetomidine displays a strong interaction with the surface in both incubation media. At the hydrophilic PS, clonidine also displays a significant interaction with the surface when incubated in mQ and a weaker, but not significant, interaction when incubated in FSW. By studying the effects of time, incubation media, and pH on the adsorption of medetomidine and clonidine, we suggest that medetomidine is associated to hydrophobic PS by means of hydrophobic interactions, while the adsorption of medetomidine and clonidine to hydrophilic PS contains elements of electrostatic interaction. Using time-of-flight secondary ion mass spectroscopy (TOF-SIMS) we detected only weak signals from medetomidine on the hydrophobic PS surfaces, while strong medetomidine signals were observed on hydrophilic PS. This suggests that the adsorbed medetomidine, to a greater extent, desorbed from the hydrophobic rather than from the hydrophilic PS surfaces during exposure to vacuum. The strong surface affinity of medetomidine on both types of surfaces and the preserved antifouling activity are valuable features in designing a marine coating.     

Fgf9 signaling regulates inner ear morphogenesis through epithelial-mesenchymal interactions

The mammalian inner ear comprises the cochleovestibular labyrinth, derived from the ectodermal otic placode, and the encasing bony labyrinth of the temporal bone. Epithelial-mesenchymal interactions are thought to control inner ear development, but the modes and the molecules involved are largely unresolved. We show here that, during the precartilage and cartilage stages, Fgf9 is expressed in specific nonsensory domains of the otic epithelium and its receptors, Fgfr1(IIIc) and Fgfr2(IIIc), widely in the surrounding mesenchyme. To address the role of Fgf9 signaling, we analyzed the inner ears of mice homozygous for Fgf9 null alleles. Fgf9 inactivation leads to a hypoplastic vestibular component of the otic capsule and to the absence of the epithelial semicircular ducts. Reduced proliferation of the prechondrogenic mesenchyme was found to underlie capsular hypoplasticity. Semicircular duct development is blocked at the initial stages, since fusion plates do not form. Our results show that the mesenchyme directs fusion plate formation and they give direct evidence for the existence of reciprocal epithelial-mesenchymal interactions in the developing inner ear. In addition to the vestibule, in the cochlea, Fgf9 mutation caused defects in the interactions between the Reissner's membrane and the mesenchymal cells, leading to a malformed scala vestibuli. Together, these data show that Fgf9 signaling is required for inner ear morphogenesis.     

Selection for increased brood size in historical human populations

Human twinning rates are considered to either reflect the direct fitness effects of twinning in variable environments, or to be a maladaptive by-product of selection for other maternal reproductive traits (e.g., polyovulation). We used historical data (1710-1890) of Sami populations from Northern Scandinavia to contrast these alternative hypotheses. We found that women who produced twins started their reproduction younger, ceased it later, had higher lifetime fecundity, raised more offspring to adulthood, and had higher fitness (individual lambda) than mothers of singletons in all populations studied. For example, an average of 1.2 offspring survived to adulthood from a twin delivery, irrespective of its sex ratio, whereas only 0.8 offspring survived to adulthood from a singleton delivery. Only if mothers started reproduction at very late age (> 37 yr), or had a very long reproductive life span (> 20 yr), was it more beneficial to produce only singletons. These findings suggest that twin deliveries among Sami could not be explained as a maladaptive by-product of selection for other maternal reproductive traits. In contrast, our results suggest that twinning was under natural selection, although the strength of selection was likely to have been context dependent.     

Coexistence of the social types: genetic population structure in the ant Formica exsecta

The ant Formica exsecta has two types of colonies that exist in sympatry but usually as separate subpopulations: colonies with simple social organization and single queens (M type) or colonial networks with multiple queens (P type). We used both nuclear (DNA microsatellites) and mitochondrial markers to study the transition between the social types, and the contribution of males and females in gene flow within and between the types. Our results showed that the social types had different spatial genetic structures. The M subpopulations formed a fairly uniform population, whereas the P subpopulations were, on average, more differentiated from each other than from the nearby M subpopulations and could have been locally established from the M-type colonies, followed by philopatric behavior and restricted emigration of females. Thus, the relationship between the two social types resembles that of source (M type) and sink (P type) populations. The comparison of mitochondrial (phiST) and nuclear (FST) differentiation indicates that the dispersal rate of males is four to five times larger than that of females both among the P-type subpopulations and between the social types. Our results suggest that evolution toward complex social organization can have an important effect on genetic population structure through changes in dispersal behavior associated with different sociogenetic organizations.     

Prediction of biomechanical properties of articular cartilage with quantitative magnetic resonance imaging

Quantitative magnetic resonance imaging (MRI) is the most potential non-invasive means for revealing the structure, composition and pathology of articular cartilage. Here we hypothesize that cartilage mechanical properties as determined by the macromolecular framework and their interactions can be accessed by quantitative MRI. To test this, adjacent cartilage disk pairs (n=32) were prepared from bovine proximal humerus and patellofemoral surfaces. For one sample, the tissue Young's modulus, aggregate modulus, dynamic modulus and Poisson's ratio were determined in unconfined compression. The adjacent disk was studied at 9.4T to determine the tissue T(2) relaxation time, sensitive to the integrity of the collagen network, and T(1) relaxation time in the presence of Gd-DTPA, a technique developed for the estimation of cartilage proteoglycan (PG) content. Quantitative MRI parameters were able to explain up to 87% of the variations in certain biomechanical parameters. Correlations were further improved when data from the proximal humerus was assessed separately. MRI parameters revealed a topographical variation similar to that of mechanical parameters. Linear regression analysis revealed that Young's modulus of cartilage may be characterized more completely by combining both collagen- and PG-sensitive MRI parameters. The present results suggest that quantitative MRI can provide important information on the mechanical properties of articular cartilage. The results are encouraging with respect to functional imaging of cartilage, although in vivo applicability may be limited by the inferior resolution of clinical MRI instruments.     

Graduality and innovation in the evolution of complex phenotypes: insights from development

The neo-Darwinian paradigm benefits from the assumption that phenotypic variation is gradual and that phenotype and genotype have a relatively simple relationship. These assumptions are historically inherited from the times of the neo-Darwinian synthesis and, consequently, do not include present understanding about development. In this study, understanding about the dynamics of pattern formation is used to explore to that extent phenotypic variation can be expected to be gradual and simply related to molecular variation. Variation in simple phenotypes seems to fit neo-Darwinian assumptions but variation in complex phenotypes does not. Instead, variation in complex phenotypes would have a tendency to relatively less gradual evolution, even at microevolutionary time scales, that would make phylogenetic reconstructions more difficult. In addition, they will have a tendency to exhibit specific trends in innovation rates over group radiations with early accelerations and late decelerations. This work also explores further consequences of these results in our understanding of phenotypic evolution.     

Crystal structure of Streptococcus suis Dps-like peroxide resistance protein Dpr: implications for iron incorporation

The Dps-like peroxide resistance protein (Dpr) is an aerotolerance and hydrogen peroxide resistance agent found in the meningitis-associated pathogen Streptococcus suis. Dpr is believed to act by binding free intracellular iron to prevent Fenton chemistry-catalysed formation of toxic hydroxyl radicals. The crystal structure of Dpr has been determined to 1.95 A resolution. The final model has an Rcyst value of 18.5% (Rfree = 22.4%) and consists of 12 identical monomers (each of them comprising a four alpha-helix bundle) that form a hollow sphere obeying 23 symmetry. Structural features show that Dpr belongs to the Dps family of bacterial proteins. Twelve putative ferroxidase centers, each formed at the interface of neighboring monomer pairs, were identified in the Dpr structure with structural similarities to those found in other Dps family members. Dpr was crystallized in the absence of iron, hence no bound iron was found in the structure in contrast to other Dps family members. A novel metal-binding site approximately 6A from the ferroxidase centre was identified and assigned to a bound calcium ion. Two residues from the ferroxidase centre (Asp63 and Asp74) were found to be involved in calcium binding. Structural comparison with other family members revealed that Asp63 and Asp74 adopt different conformation in the Dpr structure. The structure of Dpr presented here shows potential local conformational changes that may occur during iron incorporation. A role for the metal-binding site in iron uptake is proposed.     

Does mercury promote lipid peroxidation?

In order to explore the observed association among mercury, atherosclerosis, and coronary heart disease, the effects of mercury, copper, and iron on the peroxidation of low-density lipoprotein (LDL) and on the enzymatic activities of glutathione peroxidase and myeloperoxidase were investigated in vitro. On the basis of our nuclear magnetic resonance (NMR) experiments, we conclude that mercury does not promote the direct nonenzymatic peroxidation of LDL, like copper and iron. In our enzyme measurements, mercury inhibited slightly myeloperoxidase, although not significantly in presence of LDL. Instead, inorganic mercury, but not methylmercury chloride, inhibited glutathione peroxidase effectively and copper event at 10 μmol/L, below physiological concentrations, doubled the inhibition rate. Copper and iron had no direct effect on glutathione peroxidase, but they both seem to activate production of HOCl by myeloperoxidase. We conclude here that, first, mercury and methylmercury do not promote direct lipid peroxidation, but that, second, a simultaneous exposure to high inorganic mercury, copper, and iron and low selenium concentrations can lead to a condition in which mercury promotes lipid peroxidations. This mechanism provides a plausible molecular-level explanation for the observed association between high body mercury content and atherosclerosis.