Student talks

Taylor Newton

Shedding light on the cellular origins of voltage-sensitive dye imaging: an in silico study​

We report on an in silico implementation of voltage-sensitive dye imaging (VSDI) for validating and exploring mesoscopic neural activity in a biologically detailed digital reconstruction of rat somatosensory cortex. This model comprises a network of 31,346 morphologically detailed neurons arranged in a 462 x 400 x 2082μm columnar microcircuit. We evaluated the behavior of our in silico VSDI model against several findings reported in literature, including the relative contributions of different cell morphologies and cortical layers to the VSD signal, the fraction of the VSD signal due to spiking versus subthreshold activity, and the phase velocity of a stimulus driven VSDI wavefront. Using simulated VSDI measurements, we confirm that our reconstructed microcircuit exhibits behavior that is qualitatively similar to experimental findings. Furthermore, we show that contrary to widely held assumptions, dendritic projections within L2/3 but emanating from L5 pyramidal cells contribute significantly to the VSD signal, which may influence the interpretation of future VSDI-based studies.

Anna Lena Biel

Phase coupling between posterior EEG theta and gamma as a signature of memory matching

Our visual perception is strongly influenced by our expectancies about incoming sensory information. It is assumed that mental templates of expected sensory input are created that are compared to actual sensory input, which can be matching or not. In cases where such mental templates have to be held in short-term memory, such as in visual attention or search tasks, cross-frequency synchronization between theta and gamma band EEG oscillations has been proposed to serve matching processes between prediction and sensation. In this study, we investigated how matching between sensory input and mental templates from working memory is affected by the certainty about which activated template must be matched. In a visual search paradigm, we compared cross-frequency phase coupling for conditions where participants had to keep either one or multiple templates in mind for successful search. We find that memory matching appeared as a transient posterior phase-synchronization between EEG theta and gamma oscillations in an early time window after search display presentation, around 150 ms. Our results suggest a stronger transient phase-synchronization of theta and gamma over posterior sites contralateral to target presentation for conditions where one mental template was required than for multiple templates. This is understood in line with previous theoretical accounts, lending promising support for such transient phase coupling between posterior theta and gamma as a neuronal correlate of matching of incoming sensory information with memory contents from working memory.

Nidhi Sharma

Differential Expression Profile of NLRs for Glioma prognosis

Gliomas are the most common and challenging type of primary brain tumor with low prognosis and survival rate. Glioma accounts for ~80% of primary malignant brain tumors. Current therapeutic regimes include radiology, chemotherapy and surgery, providing a median survival of less than 15 months. The tumor microenvironment is heavily infiltrated by the innate immune cells and therefore, a greater understanding of the innate immune gene pool is essential for disrupting tumor growth and recurrence. Despite of the dual tumor-promoting and -inhibitory roles of NLRs, the function of NLRs in glioma remain unexplored. Our study reports for the first time, the regulation of highly conserved innate immune receptors, nucleotide-binding domain, and leucine-rich repeat containing (NLR) genes in glioma pathology. Here, we utilized an multi-omics data approaches to identify NLRs and NLR-associated gene expression across different grades of glioma. We found significant differential methylation and expression of NLRs and genes associated with inflammation, cell death and DNA repair in glioblastoma, a high grade glioma. The strong association of differentially expressed genes with patient survival highlights their prognostic significance in high and lower grade gliomas. NLRP12 emerged as promising prognostic biomarker for glioblastoma, showing significant association with patient survival and cellular proliferation assessed in-vitro. Our findings provide novel insights into differential regulation of NLRs and NLR-associated genes in lower and high grade glioma, and clinical importance of innate immune signaling pathways in glioma pathogenesis.

Chun Hao Wong

Characterisation of ARPP21 as a novel amyotrophic lateral sclerosis gene

Amyotrophic lateral sclerosis (ALS) is a neurodegenerative disease caused by multiple rare strong-effect variants whereby each gene is mutated in only a few individuals, resulting in a common endpoint of progressive degeneration of motor neurons. Here, we report a new candidate gene ARPP21 identified through whole-exome sequencing in a cohort of familial ALS cases. Two novel variants are shared by several unrelated index cases for which currently known genes have yet to be accounted for. To understand the contribution of ARPP21 in ALS, we have screened the gene extensively and modelled the identified variants in cellular models. Two novel variants (p.P563L; p.P747L) shared by several unrelated index cases are identified in a replication sporadic ALS cohort. The same p.Proline563 is found to be substituted to glutamine in two sporadic cases of UK and Italian origin. ARPP21 encodes a predominantly cytoplasmic phosphoprotein and is highly enriched in the neuronal population. Cellular modelling of the ALS-associated mutants recapitulated pathological hallmarks of ALS, including detergent insoluble aggregates, proteasome dysregulation, TDP-43 pathology, and decreased neurite outgrowth. By combining the genetics and follow-up functional assays, we have identified ARPP21 as a novel ALS gene.

Dominique Fernandes

Disrupted AMPA receptor function and synaptic plasticity upon genetic- or antibody-mediated loss of autism-associated Caspr2

Mutations in the CNTNAP2 gene have been recurrently implicated in neuropsychiatric disorders such as autism, schizophrenia and intellectual disability, and autoantibodies targeting the CNTNAP2-encoded protein CASPR2 have also been found in association with autoimmune encephalitis. However, the pathogenic mechanisms ensuing from perturbations in CASPR2 function are still unclear. Herein, we describe a novel role for Caspr2 in the regulation of glutamatergic function and synaptic plasticity. We show that Caspr2 is expressed in cortical excitatory synapses, and that it interacts with the GluA1 subunit of AMPA receptors. Silencing Caspr2 in vitro not only decreases basal synaptic content of GluA1-containing AMPARs in cortical neurons, but also hinders homeostatic synaptic scaling of AMPARs following prolonged neuronal inactivity. Caspr2 is further required for experience-dependent plasticity in vivo, since its loss in the mouse visual cortex prevents the scaling of AMPAR-mediated currents following paradigms of chronic visual deprivation. Finally, we observe that CASPR2 antibodies purified from a patient with CASPR2-encephalitis significantly alter dendritic levels of Caspr2 and synaptic GluA1-AMPAR content in cortical neurons, and perturb excitatory transmission in the visual cortex. Our results indicate that genetic- and antibody-mediated loss of Caspr2 impair AMPAR function and excitatory synaptic transmission in the cortex, and reveal a requirement for Caspr2 in the regulation of homeostatic and experience-dependent synaptic plasticity. Overall, these findings suggest that disruption of such mechanisms may underlie the pathogenesis of CASPR2-related disorders.

Konstanze Simbriger

The role of MMP-9-dependent regulation of gene-expression in Fragile X syndrome

Autism spectrum disorders (ASDs) are neurodevelopmental disorders that clinically present with characteristic behavioural and anatomical phenotypes. No cure or effective treatment is currently available for ASDs. Fragile X syndrome (FXS) is the leading known genetic cause of autism, accounting for 1-5% of ASD cases. In FXS, mutations in the Fmr1 gene lead to loss of fragile X mental retardation protein (FMRP) expression. FMRP is an mRNA binding protein and translational inhibitor that binds to and regulates a subset of mRNAs at synapses. Loss of FMRP results in exaggerated global translation, aberrant dendritic spine morphology and altered synaptic transmission and plasticity. One FMRP-regulated mRNA is matrix metalloproteinase 9 (MMP-9), encoding a protease active in the extracellular matrix (ECM), that can influence cell signalling. MMP-9 overexpression has previously been shown in FXS patients and model mice, regulating synaptic remodelling and plasticity. Mice overexpressing MMP9 display behavioural, anatomical and cellular phenotypes reminiscent of FXS. Recently, it was hypothesised, that Mmp-9 could affect signalling upstream of protein synthesis. We aimed to elucidate the role of Mmp-9 in regulating gene-expression in the brain. We used ribosome profiling to quantitatively measure changes in brain gene expression in mice overexpressing MMP-9. We find that MMP-9 overexpression regulates translation and transcription of a subset of mRNAs in the brain and affects major signalling pathways upstream of translation. We further show that these pathways are also dysregulated in FXS. Thus, our work provides evidence about common pathways for MMP-9 in FXS, which may be exploited to design novel FXS therapeutics.

Michael Whitehead

Development of a novel gene therapy for diabetic retinopathy

Diabetic retinopathy (DR) is the leading cause of vision loss in the working age population of the developed world with around 13 million individuals thought to be affected. Clinically, DR is characterised by damage to retinal blood vessels that may cause leakage of fluid and protein from retinal vessels and/or retinal neovascularisation leading to progressive visual loss. Currently available treatments for DR include laser photocoagulation, glucocorticoids, vitrectomy and anti-VEGF agents.

Currently available treatments for DR include laser photocoagulation, glucocorticoids, vitrectomy and anti-VEGF agents. Considering the significant limitations to currently available treatments, we have conceptualised a novel, polycistronic gene therapy-based therapeutic paradigm for DR. Our approach seeks to target multiple disease pathways - vascular endothelial growth factor (VEGF) signalling and angiopoietin-4 (Ang4)/Tie2 ligand/receptor interactions - in order to elicit a synergistic and diffuse therapeutic effect through our proprietary gene therapy construct.

We present herein preliminary ELISA data describing the potency of our VEGF-ablating construct, and report 10-fold reductions in VEGF levels in our in-vitro assays compared to GFP-only controls (p<0.01). The Tie2-stimulating effect of our Ang4 construct is also demonstrated through western blotting and immunohistochemistry, in which 4-fold increases in Tie2 phosphorylation are evident compared to GFP-only controls (p<0.05). Moreover, we report data suggestive of a novel mechanism-of-action of anti-VEGF agents, wherein VEGF ablation leads to a 4-fold elevation in Tie2 phosphorylation, thereby mimicking the effects of angiopoietin-mediated receptor agonism (p<0.001). Finally, we purport effective viral 2A sequence-mediated polypeptide production of anti-VEGF and GFP proteins from a single plasmid, via live cell visualisation of GFP expression and corroborative western blotting analysis.

In conclusion, our data suggest our construct is an effective mediator of DR-associated pathological pathways, and warrants further investigation in vivo.

Amanda Almacellas

One mouse, many (scientific) problems

Psychiatric diseases present a big challenge for neuroscience research. They are genetically complex syndromes with substantial inter-subject variability in multiple domains. That is why the research on psychiatric disorders is consistently focused on the endophenotypes that are recurrent in several of them. Social behavior impairment is an endophenotype characteristic of a number of disorders including schizophrenia, autism, and depression. Kidins220 is a scaffold transmembrane protein that has been correlated with psychiatric disorders and is the cause of the recently described SINO syndrome, characterized by spastic paraplegia, intellectual disability, syntagmus and obesity. My work is focused on the function of Kidins220 throughout the characterization of a conditional knockout (cKO) mouse model of this protein, in which the CaMKII promoter drives protein deletion specifically in the postnatal forebrain. The Kidins220 cKO mice display alterations in anxiety levels and social behavior, with a clear impairment in social memory. At the morphological level, cortical and hippocampal neurons are characterized by reduced dendritic branching in the absence of cell death, while at the molecular level, neuronal response to brain-derived neurotrophic factor (BDNF) stimulation is blunted, as well as mitogen activated protein kinase (MAPK) pathway activation. Given the behavioral profile of these animals, we are currently investigating the link between Kidins220, neurotrophins and oxytocin. Thus, the study of Kidins220 paths an interesting approach towards a deeper understanding of the link between the high heritability of some psychiatric diseases, the characteristic defects on neuronal physiology and the behavioral symptomatology.

Alessandro Soloperto

The bacterial mechanosensor MscL as mean for studying neuron mechanobiology and neuromodulation

The study of cell mechanobiology has recently provided growing evidences that mechanical cues play a key role in the development and physiology of the nervous system. The mechanical properties of the cells and/or of the local microenvironment influence cell differentiation and migration, and the alteration of such properties could represent the onset or the signature of a pathological state. In this regards, the development of novel approaches to stimulate neuronal circuits is crucial to understand the physiology of neuronal networks, and to provide new strategies to treat neurological disorders. Nowadays, chemical, electrical or optical approaches are the main exploited strategies to interrogate and dissect neuronal circuit functions. Alternative approaches to circumvent surgical implantation of probes include transcranial electrical, thermal, magnetic, and ultrasound stimulation. The use of magnetic and ultrasound fields represents the most promising vector to remotely convey information to the brain tissue. Herein, we report the use of the exclusively-mechanosensitive bacterial MscL channel to obtain an experimental model of mechano-sensitized mammalian neuronal networks, which could be used to study and understand the mechanism of ultrasonic cell stimulation, and thus pave the way for the engineering and development of a cell type specific ultrasonic neuro-stimulation approach. In addition, we verified the effective development of in-vitro neuronal networks expressing the engineered MscL channel in terms of channel mechanosensitivity, cell survival, number of synaptic puncta, and spontaneous network activity. Overall, our data demonstrate the successful development of a mechano-sensitized neuronal network model to allow reliable investigation, testing and calibration for the stimulation of excitable circuits through the use of remotely-generated ultrasound pressure waves. Moreover, considering the ease engineering of MscL channel properties, the mammalian-engineered MscL may represent the ideal starting point to develop a mechanogenetic approach analogous to the optogenetic method.

Mariagrazia Popeo

Functional connectivity in the newborn's brain: a fNIRS source-based data analysis approach

The organization of functional connectivity in the adult brain has been extensively explored. The study of the inception and development of functional connectivity in the newborn brain, on the other hand, remains a challenge. To this end, Functional Near Infrared Spectroscopy (fNIRS) represents a promising technique to study changes in cortical activation across development and with atypical population. Considering its technical advantages and the low degree of tolerance needed from infants, we chose it to investigate the inception of functional connectivity in the newborn brain. However, the lack of anatomical reference and of standard procedure for the arrangement of source and detectors onto the scalp are critical drawbacks of the technique. Typically, fNIRS signals collected by the sensors distributed on the scalp are analyzed directly, resulting in coarse-grained distribution of activity in the sensor space. In this study, we developed a tool to reconstruct patterns of spatially distributed functional signals on an anatomical template. This method enables assessment of the anatomical location of fNIRS signals, and provides a means to optimize probe design for the investigation of specific anatomical districts. We validated our tool for preterm and term newborns using a 4D dedicated atlas. Synthetic correlation patterns were generated simulating light propagation in brain tissues. We simulated signals with different noise levels to test the accuracy of reconstruction under realistic SNR conditions, and to assess different probe geometries. This work aims to investigate the applicability of fNIRS to functional connectivity studies in infants. Evaluation of sources localization accuracy confirms the goodness of the method to reconstruct reliable spatial patterns. Moreover, the tool represents a powerful means for probe design and provides a first step towards a source space connectivity analysis of real newborn’s data.

Ana Bottura de Barros

Interrogating Retrosplenial Cortex function with Optogenetics

Numerous studies implicate the hippocampus and neocortex in the formation of associative memories. Still, it is unclear how sensory and contextual information is stored in these regions. The retrosplenial cortex (RSC) is known to reciprocally connect to the hippocampal formation and to sensory areas. Although RSC has been previously described as important for processing spatial information, this connectivity suggests that the RSC might be involved in hippocampus-neocortex interplay and more generally required in the formation and storage of sensory associative memory. This notion is supported by RSC-lesion and -inactivation studies that produced impairments in associative memory. Here, we aim to develop an associative memory task in mice and perform precise optogenetic manipulations of RSC activity to understand its role in facilitating the association of neutral sensory stimuli. We develop an aversive preconditioning behavioural paradigm in mice which allows us to test neutral stimuli associations. We used the red-shifted opsin, Jaws, to inactivate the RSC during different phases of the task to investigate the role of this structure in sensory associative learning. Our results show that control mice successfully learn the preconditioned association in our task. Remarkably, optogenetic silencing of the RSC at the time when the mice form associations between neutral stimuli reduces performance to chance levels. In summary, we establish a new sensory preconditioning task for mice allowing us to probe associative memory of neutral stimuli. Optogenetic silencing of RSC enables us to demonstrate the necessity of the RSC in this type of memory.

Aamir Abbasi

A sensorimotor brain-machine interface to probe the role of somatotopy in artificial sensory feedback

Motor brain-machine interfaces (BMI) which use neuronal activity to control prostheses are making fast progress towards the long-term goal of restoring motor abilities to impaired patients To this day, tetraplegic patients can perform simple motor actions using BMI. However, unlike natural limbs which channel rich somatosensory and proprioceptive feedback to the brain, current BMIs are devoid of force and touch sensors. Therefore, patients have to rely only on visual and auditory feedback. This leads to cumbersome and crude control of the BMI. To address this issue we have developed a BMI where firing rate of multiple neurons in the whisker primary motor cortex drives a 1-D movement of a cursor, and simultaneous spatially structured photo-stimulation is delivered on the whisker primary somatosensory cortex, based on a mapping obtained by intrinsic imaging. We are using this BMI to study different cortical feedback schemes applied using optogenetics in awake head-fixed transgenic mice expressing channelrodopsin-2 in excitatory neurons. By our experiments, we are evaluating the impact of somatotopic-like cortical feedback on learning a BMI as compared to non-somatotopic feedback. The results of this study could be applied towards future human BMI applications.

Charline Peylo

Theta:gamma phase coupling is associated with the fidelity of mental templates in visual perception

Transient theta:gamma phase coupling around 150 ms post-stimulus in parieto-occipital cortices has been proposed as brain oscillatory correlate of matching sensation and prediction in visual perception, that is bottom-up visual information and top-down mental templates held in working memory. Here, we investigated whether theta:gamma phase coupling changes as a function of template fidelity. To this end, we recorded EEG during a target identification task, in which participants learned to form increasingly concrete predictions about (i.e., mental templates of) an upcoming target symbol from a preceding cue sequence, which were to be compared against a test symbol (matching vs. nonmatching). At the end of the experiment participants were asked to sketch the learned targets in an unannounced free recall drawing task. Our results indicate that participants significantly improved in the correct identification of test symbols from the first to the second half of the target identification task. Larger performance increments were associated with higher template fidelity in the following drawing test. Against our expectations, this behavioral improvement was not accompanied by stronger matching-related theta:gamma phase coupling in the second test half. However, stronger matching-related theta:gamma phase coupling in the second half around 150-300 ms post-stimulus localized to visual area V2 was significantly correlated with higher template fidelity (as quantified in the quality of drawings in the subsequent recall). These results suggest that theta:gamma phase coupling may serve as a general mechanism for template-to-input matching processes and that this signature reflects the fidelity of mental templates in visual perception.

Davide Spalla

Can grid cell ensembles represent multiple spaces?

Grid cells are neurons in the Medial Entorhinal Cortex of mammals whose firing activity is tuned to the position of the animal in space, exhibiting multiple peaks of activity that span an hexagonal grid. Early work investigating the behavior of these cells in representing multiple spaces [1] showed that mutual relationships between grids of different cells where conserved between environments, suggesting that Grid Cells ensembles encode a low-dimensional, "universal" map. However, this conservation was found between spaces that where both flat and geometrically similar. Moreover, recent studies [2,3] in environments with complex geometries showed distortion of the grid patterns in response to environmental features like shape or the presence of walls. These results shake the view of a universal map, together with the finding that the hippocampal-entorhinal circuit can encode abstract or conceptual spaces [4] which could in principle require a complex geometrical and topological description. In our work we estimate, with both numerical simulations and analytical calculations, the capacity of an ensemble of grid cells, modeled as a recurrent neural network, to store and retrieve multiple maps: a feature that would be crucial for a system of this kind to cope with complex geometrical features. Results show a striking enhancement of the storage capacity in the case of hexagonal patterns with respect to square ones, suggesting that the former satisfy optimality criteria for stability and memory capacity. Moreover, the high value of storage capacity for hexagonal grids suggests that a grid-like mechanism could be effective in representing multiple complex geometrical environments. References: [1] Fyhn M. et al., Nature 446, 190–194 (08 March 2007) [2] Krupic J. et al., Nature 518, 232–235 (12 February 2015) [3] Stensola T. et al., Nature 518, 207–212 (12 February 2015) [4] Aronov D. et al., Nature 543, 719–722 (30 March 2017)

Pau Aceituno

Synaptic Time-Dependent Plasticity Increases Signal Transmission Speed

Synaptic time-dependent plasticity (STDP) is a biological mechanism which changes the strength of the connections between two neurons depending on the timing of the spikes in the pre- and postsynaptic neurons. STDP is closely associated to the development of input selectivity and temporal coding. Our work focuses on the property of STDP to reduce latencies, which has only been partially addressed (Song & Abbot, Neuron 2001). As a trivial example, suppose three presynaptic neurons consistently trigger a postsynaptic spike; by STDP, their strengths increase to the point where only two synapses are necessary. Since the first two synapses always arrive before the third, the postsynaptic spike is triggered earlier. For the simple case of a leaky integrate-and-fire neuron, and under the assumption that all inputs increase the membrane potential, we provide an equation describing the reduction in latency. For the more complicated case of a network of neurons, we show how different network structures and learning properties change the evolution of the latency. Our work relates a system-level goal (processing delays) to a mechanistic, neuron level rule. All the models are simple in nature and should be easy to test experimentally.

Paulina Dabrowska

Comparison of experimental resting state data with large scale neural network simulations

The analysis of massively parallel spiking activity during behavior from monkey motor cortex reveals interesting patterns (e.g. Torre et al. 2016). We aim for a better understanding of the network mechanisms leading to such features based on simulated neural networks. Existing large scale models, however, do not account for behavior, but represent the ‘resting state’. Therefore we recorded spiking activity with a Utah array (100 electrodes, 4x4 mm^2, Blackrock Microsystems) when the monkey was not subject to any task or stimulus. It forms the basis for the comparison between experiment and model aiming to improve the simulation until the experimental statistics can be quantitatively reproduced. We started out with a layered model of a cortical microcircuit (Potjans & Diesmann 2014), subsequently extended to 4x4 mm^2 corresponding to the cortical patch we record from. For the comparison, model data are subsampled to 140 neurons (layer 5), and the experimental data are separated into inhibitory (INH) and excitatory (EXC) neurons based on their spike width (Dehghani et al. 2016). Applying the same analysis workflow and software (Elephant, http://neuralensemble.org/elephant/), we derive distributions of firing rates (FR), regularity, pairwise fine temporal correlations (CC), and rate covariances. In both data sets the average covariance and its standard deviation are significantly larger for INH than EXC population. The same holds for the experimental CC. In the model, however, the average CC is higher for EXC neurons. Also the mean FR differs between populations in experiment, but not in the simulation. Since our model is currently based on connectivity parameters from various cortices and species, the deviations between the data sets were to be expected. Next, we aim to adjust the model towards motor cortex. Support: DFG GR 1753/4-2 Priority Program (SPP 1665), Helmholtz Portfolio Theme SMHB, EU Grant 720270 (HBP). Network simulations are performed in NEST.