Reinforcement learning in dendritic structures

The discovery of binary dendritic events such as local NMDA spikes in dendritic subbranches led to the suggestion that dendritic trees could be computationally equivalent to a 2-layer network of point neurons, with a single output unit represented by the soma, and input units represented by the dendritic branches. Although this interpretation endows a neuron with a high computational power, it is functionally not clear why nature would have preferred the dendritic solution with a single but complex neuron, as opposed to the network solution with many but simple units. We show that the dendritic solution has a distinguished advantage over the network solution when considering different learning tasks. Its key property is that the dendritic branches receive an immediate feedback from the somatic output spike, while in the corresponding network architecture the feedback would require additional backpropagating connections to the input units. Assuming a reinforcement learning scenario we formally derive a learning rule for the synaptic contacts on the individual dendritic trees which depends on the presynaptic activity, the local NMDA spikes, the somatic action potential, and a delayed reinforcement signal. We test the model for two scenarios: the learning of binary classifications and of precise spike timings. We show that the immediate feedback represented by the backpropagating action potential supplies the individual dendritic branches with enough information to efficiently adapt their synapses and to speed up the learning process.

Reinforcement learning in dendritic structures

The discovery of binary dendritic events such as local NMDA spikes in dendritic subbranches led to the suggestion that dendritic trees could be computationally equivalent to a 2-layer network of point neurons, with a single output unit represented by the soma, and input units represented by the dendritic branches. Although this interpretation endows a neuron with a high computational power, it is functionally not clear why nature would have preferred the dendritic solution with a single but complex neuron, as opposed to the network solution with many but simple units. We show that the dendritic solution has a distinguished advantage over the network solution when considering different learning tasks. Its key property is that the dendritic branches receive an immediate feedback from the somatic output spike, while in the corresponding network architecture the feedback would require additional backpropagating connections to the input units. Assuming a reinforcement learning scenario we formally derive a learning rule for the synaptic contacts on the individual dendritic trees which depends on the presynaptic activity, the local NMDA spikes, the somatic action potential, and a delayed reinforcement signal. We test the model for two scenarios: the learning of binary classifications and of precise spike timings. We show that the immediate feedback represented by the backpropagating action potential supplies the individual dendritic branches with enough information to efficiently adapt their synapses and to speed up the learning process.

Cerium oxide nanoparticle uptake kinetics from the gas-phase into lung cells in vitro is transport limited

Nowadays, aerosol processes are widely used for the manufacture of nanoparticles (NPs), creating an increased occupational exposure risk of workers, laboratory personnel and scientists to airborne particles. There is evidence that possible adverse effects are linked with the accumulation of NPs in target cells, pointing out the importance of understanding the kinetics of particle internalization. In this context, the uptake kinetics of representative airborne NPs over 30 min and their internalization after 24 h post-exposure were investigated by the use of a recently established exposure system. This system combines the production of aerosolized cerium oxide (CeO(2)) NPs by flame spray synthesis with its simultaneous particle deposition from the gas-phase onto A549 lung cells, cultivated at the air-liquid interface. Particle uptake was quantified by mass spectrometry after several exposure times (0, 5, 10, 20 and 30 min). Over 35% of the deposited mass was found internalized after 10 min exposure, a value that increased to 60% after 30 min exposure. Following an additional 24 h post-incubation, a time span, after which adverse biological effects were observed in previous experiments, over 80% of total CeO(2) could be detected intracellularly. On the ultrastructural level, focal cerium aggregates were present on the apical surface of A549 cells and could also be localized intracellularly in vesicular structures. The uptake behaviour of aerosolized CeO(2) is in line with observations on cerium suspensions, where particle mass transport was identified as the rate-limiting factor for NP internalization.

Characterization of the first FGFRL1 mutation identified in a craniosynostosis patient

Fibroblast growth factor receptor-like 1 (FGFRL1) is a recently discovered transmembrane protein whose functions remain unclear. Since mutations in the related receptors FGFR1-3 cause skeletal malformations, DNA samples from 55 patients suffering from congenital skeletal malformations and 109 controls were searched for mutations in FGFRL1. One patient was identified harboring a frameshift mutation in the intracellular domain of this novel receptor. The patient showed craniosynostosis, radio-ulnar synostosis and genital abnormalities and had previously been diagnosed with Antley-Bixler syndrome. The effect of the FGFRL1 mutation was studied in vitro. In a reporter gene assay, the wild-type as well as the mutant receptor inhibited FGF signaling. However, the mutant protein differed from the wild-type protein in its subcellular localization. Mutant FGFRL1 was mainly found at the plasma membrane where it interacted with FGF ligands, while the wild-type protein was preferentially located in vesicular structures and the Golgi complex. Two motifs from the intracellular domain of FGFRL1 appeared to be responsible for this differential distribution, a tandem tyrosine based motif and a histidine-rich sequence. Deletion of either one led to the preferential redistribution of FGFRL1 to the plasma membrane. It is therefore likely that mutant FGFRL1 contributes to the skeletal malformations of the patient.

Lateralization of dynorphin gene expression in the rat striatum

Hemispheric lateralization is well known in the cerebral cortex, but not in subcortical structures like basal ganglia. The goal of our study was to determine whether lateralization was present in the direct and indirect striatal pathways. We studied gene expression in the striatum of healthy rats, which was divided into two sectors, medial and lateral. Dynorphin (DYN) and enkephalin (ENK) mRNA were analyzed as markers of the direct and indirect striatal pathways, respectively and glutamic acid decarboxylase (GAD) mRNA was analyzed as a marker of all medium spiny neurons. DYN and GAD mRNA expression was higher on the left hemisphere in the medial sector of the striatum, but not in the lateral one. We did not observe any difference between sides with ENK mRNA expression. We suggest the presence of a lateralization in the medial striatum, which is specific for the direct striatal pathway.

Highly efficient in vivo agonist-induced internalization of sst2 receptors in somatostatin target tissues

The successful peptide receptor imaging of tumors, as exemplified for somatostatin receptors, is based on the overexpression of peptide receptors in selected tumors and the high-affinity binding to these tumors of agonist radioligands that are subsequently internalized into the tumor cells in which they accumulate. Although in vitro studies have shown ample evidence that the ligand-receptor complex is internalized, in vivo evidence of agonist-induced internalization of peptide receptors, such as somatostatin receptors, is missing. METHODS: Rats subcutaneously transplanted with the somatostatin receptor subtype 2 (sst(2))-expressing AR42J tumor cells were treated with intravenous injections of various doses of the sst(2) agonist [Tyr(3), Thr(8)]-octreotide (TATE) or of the sst(2) antagonist 1,4,7,10-tetraazacyclododecane-N,N',N'',N''',-tetraacetic acid (DOTA)-Bass and were sacrificed at various times ranging from 2.5 min to 24 h after injection. The tumors and pancreas were then removed from each animal. All tissue samples were processed for sst(2) immunohistochemistry using sst(2)-specific antibodies. RESULTS: Compared with the sst(2) receptors in untreated animals, which localized at the plasma membrane in pancreatic and AR42J tumor cells, the sst(2) receptors in treated animals are detected intracellularly after an intravenous injection of the agonist TATE. Internalization is fast, as the receptors are already internalizing 2.5 min after TATE injection. The process is extremely efficient, as most of the cell surface receptors internalize into the cell and are found in endosomelike structures after TATE injection. The internalization is most likely reversible, because 24 h after injection the receptors are again found at the cell surface. The process is also agonist-dependent, because internalization is seen with high-affinity sst(2) agonists but not with high-affinity sst(2) antagonists. The same internalization properties are seen in pancreatic and AR42J tumor cells. They can further be confirmed in vitro in human embryonic kidney-sst(2) cells, with an immunofluorescence microscopy-based sst(2) internalization assay. CONCLUSION: These animal data strongly indicate that the process of in vivo sst(2) internalization after agonist stimulation is fast, extremely efficient, and fully functional under in vivo conditions in neoplastic and physiologic sst(2) target tissues. This molecular process is, therefore, likely to be responsible for the high and long-lasting uptake of sst(2) radioligands seen in vivo in sst(2)-expressing tumors.

The PDZ protein MPP2 interacts with c-Src in epithelial cells

c-Src is a non-receptor tyrosine kinase involved in regulating cell proliferation, cell migration and cell invasion and is tightly controlled by reversible phosphorylation on regulatory sites and through protein-protein interactions. The interaction of c-Src with PDZ proteins was recently identified as novel mechanism to restrict c-Src function. The objective of this study was to identify and characterise PDZ proteins that interact with c-Src to control its activity. By PDZ domain array screen, we identified the interaction of c-Src with the PDZ protein Membrane Protein Palmitoylated 2 (MPP2), a member of the Membrane-Associated Guanylate Kinase (MAGUK) family, to which also the Discs large (Dlg) tumour suppressor protein belongs. The function of MPP2 has not been established and the functional significance of the MPP2 c-Src interaction is not known. We found that in non-transformed breast epithelial MCF-10A cells, endogenous MPP2 associated with the cytoskeleton in filamentous structures, which partially co-localised with microtubules and c-Src. MPP2 and c-Src interacted in cells, where c-Src kinase activity promoted increased interaction of c-Src with MPP2. We furthermore found that MPP2 was able to negatively regulate c-Src kinase activity in cells, suggesting that the functional significance of the MPP2-c-Src interaction is to restrict Src activity. Consequently, the c-Src-dependent disorganisation of the cortical actin cytoskeleton of epithelial cells expressing c-Src was suppressed by MPP2. In conclusion we demonstrate here that MPP2 interacts with c-Src in cells to control c-Src activity and morphological function.

Supramolecular Organization of Heptapyrenotide Oligomers—An in Depth Investigation by Molecular Dynamics Simulations

We present a molecular modeling study based on ab initio and classical molecular dynamics calculations, for the investigation of the tridimensional structure and supramolecular assembly formation of heptapyrenotide oligomers in water solution. Our calculations show that free oligomers self-assemble in helical structures characterized by an inner core formed by π- stacked pyrene units, and external grooves formed by the linker moieties. The coiling of the linkers has high ordering, dominated by hydrogen-bond interactions among the phosphate and amide groups. Our models support a mechanism of longitudinal supramolecular oligomerization based on interstrand pyrene intercalation. Only a minimal number of pyrene units intercalate at one end, favoring formation of very extended longitudinal chains, as also detected by AFM experiment. Our results provide a structural explanation of the mechanism of chirality amplification in 1:1 mixtures of standard heptapyrenotides and modified oligomers with covalently linked deoxycytidine, based on selective molecular recognition and binding of the nucleotide to the groove of the left-wound helix.

Importance of vegetation, topography and flow paths for water transit times of base flow in alpine headwater catchments

The mean transit time (MTT) of water in a catchment gives information about storage, flow paths, sources of water and thus also about retention and release of solutes in a catchment. To our knowledge there are only a few catchment studies on the influence of vegetation cover changes on base flow MTTs. The main changes in vegetation cover in the Swiss Alps are massive shrub encroachment and forest expansion into formerly open habitats. Four small and relatively steep headwater catchments in the Swiss Alps (Ursern Valley) were investigated to relate different vegetation cover to water transit times. Time series of water stable isotopes were used to calculate MTTs. The high temporal variation of the stable isotope signals in precipitation was strongly dampened in stream base flow samples. MTTs of the four catchments were 70 to 102 weeks. The strong dampening of the stable isotope input signal as well as stream water geochemistry points to deeper flow paths and mixing of waters of different ages at the catchments' outlets. MTTs were neither related to topographic indices nor vegetation cover. The major part of the quickly infiltrating precipitation likely percolates through fractured and partially karstified deeper rock zones, which increases the control of bedrock flow paths on MTT. Snow accumulation and the timing of its melt play an important role for stable isotope dynamics during spring and early summer. We conclude that, in mountainous headwater catchments with relatively shallow soil layers, the hydrogeological and geochemical patterns (i.e. geochemistry, porosity and hydraulic conductivity of rocks) and snow dynamics influence storage, mixing and release of water in a stronger way than vegetation cover or topography do.

The nitrogen cycle of tropical montane forest in Ecuador turns inorganic under environmental change

Water-bound nitrogen (N) cycling in temperate terrestrial ecosystems of the Northern Hemisphere is today mainly inorganic because of anthropogenic release of reactive N to the environment. In little-industrialized and remote areas, in contrast, a larger part of N cycling occurs as dissolved organic N (DON). In a north Andean tropical montane forest in Ecuador, the N cycle changed markedly during 1998–2010 along with increasing N deposition and reduced soil moisture. The DON concentrations and the fractional contribution of DON to total N significantly decreased in rainfall, throughfall, and soil solutions. This inorganic turn of the N cycle was most pronounced in rainfall and became weaker along the flow path of water through the system until it disappeared in stream water. Decreasing organic contributions to N cycling were caused not only by increasing inorganic N input but also by reduced DON production and/or enhanced DON decomposition. Accelerated DON decomposition might be attributable to less waterlogging and higher nutrient availability. Significantly increasing NO3-N concentrations and NO3-N/NH4-N concentration ratios in throughfall and litter leachate below the thick organic layers indicated increasing nitrification. In mineral soil solutions, in contrast, NH4-N concentrations increased and NO3-N/NH4-N concentration ratios decreased significantly, suggesting increasing net ammonification. Our results demonstrate that the remote tropical montane forests on the rim of the Amazon basin experienced a pronounced change of the N cycle in only one decade. This change likely parallels a similar change which followed industrialization in the temperate zone of the Northern Hemisphere more than a century ago.

Ar diffusion in K-feldspar: present and absent

The mechanisms of Ar release from K-feldspar samples in laboratory experiments and during their geological history are assessed here. Modern petrology clearly established that the chemical and isotopic record of minerals is normally dominated by aqueous recrystallization. The laboratory critique is trickier, which explains why so many conflicting approaches have been able to survive long past their expiration date. Current models are evaluated for self-consistency; especially Arrhenian non-linearity leads to paradoxes. The models’ testable geological predictions suggest that temperature-based downslope extrapolations often overestimate observed geological Ar mobility substantially. An updated interpretation is based on the unrelatedness of geological behaviour to laboratory experiments. The isotopic record of K-feldspar in geological samples is not a unique function of temperature, as recrystallisation promoted by aqueous fluids is the predominant mechanism controlling isotope transport. K-feldspar should therefore be viewed as a hygrochronometer. Laboratory degassing proceeds from structural rearrangements and phase transitions such as are observed in situ at high temperature in Na and Pb feldspars. These effects violate the mathematics of an inert Fick’s Law matrix and preclude downslope extrapolation. The similar upward-concave, non-linear shapes of Arrhenius trajectories of many silicates, hydrous and anhydrous, are likely common manifestations of structural rearrangements in silicate structures.

Exposure of silver-nanoparticles and silver-ions to lung cells in vitro at the air-liquid interface

BACKGROUND: Due to its antibacterial properties, silver (Ag) has been used in more consumer products than any other nanomaterial so far. Despite the promising advantages posed by using Ag-nanoparticles (NPs), their interaction with mammalian systems is currently not fully understood. An exposure route via inhalation is of primary concern for humans in an occupational setting. Aim of this study was therefore to investigate the potential adverse effects of aerosolised Ag-NPs using a human epithelial airway barrier model composed of A549, monocyte derived macrophage and dendritic cells cultured in vitro at the air-liquid interface. Cell cultures were exposed to 20 nm citrate-coated Ag-NPs with a deposition of 30 and 278 ng/cm2 respectively and incubated for 4 h and 24 h. To elucidate whether any effects of Ag-NPs are due to ionic effects, Ag-Nitrate (AgNO3) solutions were aerosolised at the same molecular mass concentrations. RESULTS: Agglomerates of Ag-NPs were detected at 24 h post exposure in vesicular structures inside cells but the cellular integrity was not impaired upon Ag-NP exposures. Minimal cytotoxicity, by measuring the release of lactate dehydrogenase, could only be detected following a higher concentrated AgNO3-solution. A release of pro-inflammatory markers TNF-alpha and IL-8 was neither observed upon Ag-NP and AgNO3 exposures as well as was not affected when cells were pre-stimulated with lipopolysaccharide (LPS). Also, an induction of mRNA expression of TNF-alpha and IL-8, could only be observed for the highest AgNO3 concentration alone or even significantly increased when pre-stimulated with LPS after 4 h. However, this effect disappeared after 24 h. Furthermore, oxidative stress markers (HMOX-1, SOD-1) were expressed after 4 h in a concentration dependent manner following AgNO3 exposures only. CONCLUSIONS: With an experimental setup reflecting physiological exposure conditions in the human lung more realistic, the present study indicates that Ag-NPs do not cause adverse effects and cells were only sensitive to high Ag-ion concentrations. Chronic exposure scenarios however, are needed to reveal further insight into the fate of Ag-NPs after deposition and cell interactions.