Protein dynamics associated with failed and rescued learning in the Ts65Dn mouse model of Down syndrome

Deep Unsupervised Active Learning via Matrix Sketching

Most existing unsupervised active learning methods aim at minimizing the data reconstruction loss by using the linear models to choose representative samples for manually labeling in an unsupervised setting. Thus these methods often fail in modelling data with complex non-linear structure. To address this issue, we propose a new deep unsupervised Active Learning method for classification tasks, inspired by the idea of Matrix Sketching, called ALMS. Specifically, ALMS leverages a deep auto-encoder to embed data into a latent space, and then describes all the embedded data with a small size sketch to summarize the major characteristics of the data. In contrast to previous approaches that reconstruct the whole data matrix for selecting the representative samples, ALMS aims to select a representative subset of samples to well approximate the sketch, which can preserve the major information of data meanwhile significantly reducing the number of network parameters. This makes our algorithm alleviate the issue of model overfitting and readily cope with large datasets. Actually, the sketch provides a type of self-supervised signal to guide the learning of the model. Moreover, we propose to construct an auxiliary self-supervised task by classifying real/fake samples, in order to further improve the representation ability of the encoder. We thoroughly evaluate the performance of ALMS on both single-label and multi-label classification tasks, and the results demonstrate its superior performance against the state-of-the-art methods. The code can be found at https://github.com/lrq99/ALMS.

Context Fear Conditioning in Down Syndrome Mouse Models: Effects of Trisomic Gene Content, Age, Sex and Genetic Background

Down syndrome (DS), trisomy of the long arm of human chromosome 21 (Hsa21), is the most common genetic cause of intellectual disability (ID). Currently, there are no effective pharmacotherapies. The success of clinical trials to improve cognition depends in part on the design of preclinical evaluations in mouse models. To broaden understanding of the common limitations of experiments in learning and memory, we report performance in context fear conditioning (CFC) in three mouse models of DS, the Dp(16)1Yey, Dp(17)1Yey and Dp(10)1Yey (abbreviated Dp16, Dp17 and Dp10), separately trisomic for the human Hsa21 orthologs mapping to mouse chromosomes 16, 17 and 10, respectively. We examined female and male mice of the three lines on the standard C57BL/6J background at 3 months of age and Dp17 and Dp10 at 18 months of age. We also examined female and male mice of Dp17 and Dp10 at 3 months of age as F1 hybrids obtained from a cross with the DBA/2J background. Results indicate that genotype, sex, age and genetic background affect CFC performance. These data support the need to use both female and male mice, trisomy of sets of all Hsa21 orthologs, and additional ages and genetic backgrounds to improve the reliability of preclinical evaluations of drugs for ID in DS.

Altered Protein Profiles During Epileptogenesis in the Pilocarpine Mouse Model of Temporal Lobe Epilepsy

Epilepsy is characterized by recurrent, spontaneous seizures and is a major contributor to the global burden of neurological disease. Although epilepsy can result from a variety of brain insults, in many cases the cause is unknown and, in a significant proportion of cases, seizures cannot be controlled by available treatments. Understanding the molecular alterations that underlie or are triggered by epileptogenesis would help to identify therapeutics to prevent or control progression to epilepsy. To this end, the moderate throughput technique of Reverse Phase Protein Arrays (RPPA) was used to profile changes in protein expression in a pilocarpine mouse model of acquired epilepsy. Levels of 54 proteins, comprising phosphorylation-dependent and phosphorylation-independent components of major signaling pathways and cellular complexes, were measured in hippocampus, cortex and cerebellum of mice at six time points, spanning 15 min to 2 weeks after induction of status epilepticus. Results illustrate the time dependence of levels of the commonly studied MTOR pathway component, pS6, and show, for the first time, detailed responses during epileptogenesis of multiple components of the MTOR, MAPK, JAK/STAT and apoptosis pathways, NMDA receptors, and additional cellular complexes. Also noted are time- and brain region- specific changes in correlations among levels of functionally related proteins affecting both neurons and glia. While hippocampus and cortex are primary areas studied in pilocarpine-induced epilepsy, cerebellum also shows significant time-dependent molecular responses.

The Rac-GEF Tiam1 Promotes Dendrite and Synapse Stabilization of Dentate Granule Cells and Restricts Hippocampal-Dependent Memory Functions

The dentate gyrus (DG) controls information flow into the hippocampus and is critical for learning, memory, pattern separation, and spatial coding, while DG dysfunction is associated with neuropsychiatric disorders. Despite its importance, the molecular mechanisms regulating DG neural circuit assembly and function remain unclear. Here, we identify the Rac-GEF Tiam1 as an important regulator of DG development and associated memory processes. In the hippocampus, Tiam1 is predominantly expressed in the DG throughout life. Global deletion of Tiam1 in male mice results in DG granule cells with simplified dendritic arbors, reduced dendritic spine density, and diminished excitatory synaptic transmission. Notably, DG granule cell dendrites and synapses develop normally in Tiam1 KO mice, resembling WT mice at postnatal day 21 (P21), but fail to stabilize, leading to dendrite and synapse loss by P42. These results indicate that Tiam1 promotes DG granule cell dendrite and synapse stabilization late in development. Tiam1 loss also increases the survival, but not the production, of adult-born DG granule cells, possibly because of greater circuit integration as a result of decreased competition with mature granule cells for synaptic inputs. Strikingly, both male and female mice lacking Tiam1 exhibit enhanced contextual fear memory and context discrimination. Together, these results suggest that Tiam1 is a key regulator of DG granule cell stabilization and function within hippocampal circuits. Moreover, based on the enhanced memory phenotype of Tiam1 KO mice, Tiam1 may be a potential target for the treatment of disorders involving memory impairments.SIGNIFICANCE STATEMENT The dentate gyrus (DG) is important for learning, memory, pattern separation, and spatial navigation, and its dysfunction is associated with neuropsychiatric disorders. However, the molecular mechanisms controlling DG formation and function remain elusive. By characterizing mice lacking the Rac-GEF Tiam1, we demonstrate that Tiam1 promotes the stabilization of DG granule cell dendritic arbors, spines, and synapses, whereas it restricts the survival of adult-born DG granule cells, which compete with mature granule cells for synaptic integration. Notably, mice lacking Tiam1 also exhibit enhanced contextual fear memory and context discrimination. These findings establish Tiam1 as an essential regulator of DG granule cell development, and identify it as a possible therapeutic target for memory enhancement.

Bioinformatics Investigation and Contribution of Other Chromosomes Besides Chromosome 21 in the Risk of Down Syndrome Development

INTRODUCTION

Down syndrome as a genetic disorder is a popular research topic in molecular studies. One way to study Down syndrome is via bioinformatics.

METHODS

In this study, a gene expression profile from a microarray study was selected for more investigation.

RESULTS

The study of Down syndrome patients shows specific genes with differential expression and network centrality properties. These genes are introduced as RHOA, FGF2, FYN, and CD44, and their level of expression is high in these patients.

CONCLUSION

This study suggests that besides chromosomes 21, there are additional contributing chromosomes to the risk of Down syndrome development. Besides, these genes could be used for clinical studies after more analysis.

A data-driven dimensionality-reduction algorithm for the exploration of patterns in biomedical data

Dimensionality reduction is widely used in the visualization, compression, exploration and classification of data. Yet a generally applicable solution remains unavailable. Here, we report an accurate and broadly applicable data-driven algorithm for dimensionality reduction. The algorithm, which we named 'feature-augmented embedding machine' (FEM), first learns the structure of the data and the inherent characteristics of the data components (such as central tendency and dispersion), denoises the data, increases the separation of the components, and then projects the data onto a lower number of dimensions. We show that the technique is effective at revealing the underlying dominant trends in datasets of protein expression and single-cell RNA sequencing, computed tomography, electroencephalography and wearable physiological sensors.

In silico identification of critical proteins associated with learning process and immune system for Down syndrome

Understanding expression levels of proteins and their interactions is a key factor to diagnose and explain the Down syndrome which can be considered as the most prevalent reason of intellectual disability in human beings. In the previous studies, the expression levels of 77 proteins obtained from normal genotype control mice and from trisomic Ts65Dn mice have been analyzed after training in contextual fear conditioning with and without injection of the memantine drug using statistical methods and machine learning techniques. Recent studies have also pointed out that there may be a linkage between the Down syndrome and the immune system. Thus, the research presented in this paper aim at in silico identification of proteins which are significant to the learning process and the immune system and to derive the most accurate model for classification of mice. In this paper, the features are selected by implementing forward feature selection method after preprocessing step of the dataset. Later, deep neural network, gradient boosting tree, support vector machine and random forest classification methods are implemented to identify the accuracy. It is observed that the selected feature subsets not only yield higher accuracy classification results but also are composed of protein responses which are important for the learning and memory process and the immune system.

Rodent models in Down syndrome research: impact and future opportunities

Down syndrome is caused by trisomy of chromosome 21. To date, a multiplicity of mouse models with Down-syndrome-related features has been developed to understand this complex human chromosomal disorder. These mouse models have been important for determining genotype-phenotype relationships and identification of dosage-sensitive genes involved in the pathophysiology of the condition, and in exploring the impact of the additional chromosome on the whole genome. Mouse models of Down syndrome have also been used to test therapeutic strategies. Here, we provide an overview of research in the last 15 years dedicated to the development and application of rodent models for Down syndrome. We also speculate on possible and probable future directions of research in this fast-moving field. As our understanding of the syndrome improves and genome engineering technologies evolve, it is necessary to coordinate efforts to make all Down syndrome models available to the community, to test therapeutics in models that replicate the whole trisomy and design new animal models to promote further discovery of potential therapeutic targets.

Summary: Mouse models have boosted therapeutic options for Down syndrome, and improved models are being developed to better understand the pathophysiology of this genetic condition.

Exploring patterns enriched in a dataset with contrastive principal component analysis

Visualization and exploration of high-dimensional data is a ubiquitous challenge across disciplines. Widely used techniques such as principal component analysis (PCA) aim to identify dominant trends in one dataset. However, in many settings we have datasets collected under different conditions, e.g., a treatment and a control experiment, and we are interested in visualizing and exploring patterns that are specific to one dataset. This paper proposes a method, contrastive principal component analysis (cPCA), which identifies low-dimensional structures that are enriched in a dataset relative to comparison data. In a wide variety of experiments, we demonstrate that cPCA with a background dataset enables us to visualize dataset-specific patterns missed by PCA and other standard methods. We further provide a geometric interpretation of cPCA and strong mathematical guarantees. An implementation of cPCA is publicly available, and can be used for exploratory data analysis in many applications where PCA is currently used.

Dimensionality reduction and visualization methods lack a principled way of comparing multiple datasets. Here, Abid et al. introduce contrastive PCA, which identifies low-dimensional structures enriched in one dataset compared to another and enables visualization of dataset-specific patterns.

microRNA 155 up regulation in the CNS is strongly correlated to Down's syndrome dementia

This study examined the molecular correlates of Down's dementia. qRTPCR for chromosome 21 microRNAs was correlated with in situ hybridization, immunohistochemistry for microRNA targets, mRNAs located on chromosome 21, and neurofibrillary tangles in human and the Ts65 dn mouse Down's model. qRTPCR for the microRNAs on the triplicated chromosome showed miR-155 dominance in brain tissues (14.3 fold increase, human and 24.2 fold increase, mouse model) that co-expressed with hyperphosphorylated tau protein. miR-155 was not elevated in Alzheimer's disease or neonates with Downs' syndrome. Chromosome 21 genes APP/BA-42, DYRK1a and BACH1 were not correlated to pathologic changes in Down's dementia. Validated CNS targets of miR-155 that were present in controls and Alzheimer's disease but lacking in Down's dementia brains included BACH1, CoREST1, bcl6, BIM, bcl10, cyclin D, and SAPK4. It is concluded that Down's dementia strongly correlated with overexpression of chromosome 21 microRNA 155 with concomitant reduction of multiple CNS-functional targets. This study highlights the need for anatomic pathologists to determine the specific and diverse pathways cells may take to form neurofibrillary tangles in the different dementias.

Proteomic analysis of six- and twelve-month hippocampus and cerebellum in a murine Down syndrome model

This study was designed to investigate the brain proteome of the Ts65Dn mouse model of Down syndrome. We profiled the cerebellum and hippocampus proteomes of 6- and 12-month-old trisomic and disomic mice by difference gel electrophoresis. We quantified levels of 2082 protein spots and identified 272 (170 unique UniProt accessions) by mass spectrometry. Four identified proteins are encoded by genes trisomic in the Ts65Dn mouse. Three of these (CRYZL11, EZR, and SOD1) were elevated with p-value <0.05, and 2 proteins encoded by disomic genes (MAPRE3 and PHB) were reduced. Intergel comparisons based on age (6 vs. 12 months) and brain region (cerebellum vs. hippocampus) revealed numerous differences. Specifically, 132 identified proteins were different between age groups, and 141 identified proteins were different between the 2 brain regions. Our results suggest that compensatory mechanisms exist, which ameliorate the effect of trisomy in the Ts65Dn mice. Differences observed during aging may play a role in the accelerated deterioration of learning and memory seen in Ts65Dn mice.

Age exacerbates abnormal protein expression in a mouse model of Down syndrome

The Ts65Dn is a popular mouse model of Down syndrome (DS). It displays DS-relevant features of learning/memory deficits and age-related loss of functional markers in basal forebrain cholinergic neurons. Here we describe protein expression abnormalities in brain regions of 12-month-old male Ts65Dn mice. We show that the magnitudes of abnormalities of human chromosome 21 and non-human chromosome 21 orthologous proteins are greater at 12 months than at ∼6 months. Age-related exacerbations involve the number of components affected in the mechanistic target of rapamycin pathway, the levels of components of the mitogen-activated protein kinase pathway, and proteins associated with Alzheimer's disease. Among brain regions, the number of abnormalities in cerebellum decreased while the number in cortex greatly increased with age. The Ts65Dn is being used in preclinical evaluations of drugs for cognition in DS. Most commonly, drug evaluations are tested in ∼4- to 6-month-old mice. Data on age-related changes in magnitude and specificity of protein perturbations can be used to understand the molecular basis of changes in cognitive ability and to predict potential age-related specificities in drug efficacies.

A Comprehensive Diverse '-omics' Approach to Better Understanding the Molecular Pathomechanisms of Down Syndrome

Diverse ‘-omics’ technologies permit the comprehensive quantitative profiling of a variety of biological molecules. Comparative ‘-omics’ analyses, such as transcriptomics and proteomics, are powerful and useful tools for unraveling the molecular pathomechanisms of various diseases. As enhanced oxidative stress has been demonstrated in humans and mice with Down syndrome (DS), a redox proteomic analysis is useful for understanding how enhanced oxidative stress aggravates the state of individuals with oxidative stress-related disorders. In this review, ‘-omics’ analyses in humans with DS and mouse models of DS are summarized, and the molecular dissection of this syndrome is discussed.

Sex differences in protein expression in the mouse brain and their perturbations in a model of Down syndrome

Background

While many sex differences in structure and function of the mammalian brain have been described, the molecular correlates of these differences are not broadly known. Also unknown is how sex differences at the protein level are perturbed by mutations that lead to intellectual disability (ID). Down syndrome (DS) is the most common genetic cause of ID and is due to trisomy of human chromosome 21 (Hsa21) and the resulting increased expression of Hsa21-encoded genes. The Dp(10)1Yey mouse model (Dp10) of DS is trisomic for orthologs of 39 Hsa21 protein-coding genes that map to mouse chromosome 10 (Mmu10), including four genes with known sex differences in functional properties. How these genes contribute to the DS cognitive phenotype is not known.

Methods

Using reverse phase protein arrays, levels of ~100 proteins/protein modifications were measured in the hippocampus, cerebellum, and cortex of female and male controls and their trisomic Dp10 littermates. Proteins were chosen for their known roles in learning/memory and synaptic plasticity and include components of the MAPK, MTOR, and apoptosis pathways, immediate early genes, and subunits of ionotropic glutamate receptors. Protein levels were compared between genotypes, sexes, and brain regions using a three-level mixed effects model and the Benjamini-Hochberg correction for multiple testing.

Results

In control mice, levels of approximately one half of the proteins differ significantly between females and males in at least one brain region; in the hippocampus alone, levels of 40 % of the proteins are significantly higher in females. Trisomy of the Mmu10 segment differentially affects female and male profiles, perturbing protein levels most in the cerebellum of female Dp10 and most in the hippocampus of male Dp10. Cortex is minimally affected by sex and genotype. Diverse pathways and processes are implicated in both sex and genotype differences.

Conclusions

The extensive sex differences in control mice in levels of proteins involved in learning/memory illustrate the molecular complexity underlying sex differences in normal neurological processes. The sex-specific abnormalities in the Dp10 suggest the possibility of sex-specific phenotypic features in DS and reinforce the need to use female as well as male mice, in particular in preclinical evaluations of drug responses.

Electronic supplementary material

The online version of this article (doi:10.1186/s13293-015-0043-9) contains supplementary material, which is available to authorized users.

Timing of therapies for Down syndrome: the sooner, the better

Intellectual disability (ID) is the unavoidable hallmark of Down syndrome (DS), with a heavy impact on public health. Accumulating evidence shows that DS is characterized by numerous neurodevelopmental alterations among which the reduction of neurogenesis, dendritic hypotrophy and connectivity alterations appear to play a particularly prominent role. Although the mechanisms whereby gene triplication impairs brain development in DS have not been fully clarified, it is theoretically possible to correct trisomy-dependent defects with targeted pharmacotherapies. This review summarizes what we know about the effects of pharmacotherapies during different life stages in mouse models of DS. Since brain alterations in DS start to be present prenatally, the prenatal period represents an optimum window of opportunity for therapeutic interventions. Importantly, recent studies clearly show that treatment during the prenatal period can rescue overall brain development and behavior and that this effect outlasts treatment cessation. Although late therapies are unlikely to exert drastic changes in the brain, they may have an impact on the hippocampus, a brain region where neurogenesis continues throughout life. Indeed, treatment at adult life stages improves or even rescues hippocampal neurogenesis and connectivity and hippocampal-dependent learning and memory, although the duration of these effects still remains, in the majority of cases, a matter of investigation. The exciting discovery that trisomy-linked brain abnormalities can be prevented with early interventions gives us reason to believe that treatments during pregnancy may rescue brain development in fetuses with DS. For this reason we deem it extremely important to expedite the discovery of additional therapies practicable in humans in order to identify the best treatment/s in terms of efficacy and paucity of side effects. Prompt achievement of this goal is the big challenge for the scientific community of researchers interested in DS.

Self-Organizing Feature Maps Identify Proteins Critical to Learning in a Mouse Model of Down Syndrome

Down syndrome (DS) is a chromosomal abnormality (trisomy of human chromosome 21) associated with intellectual disability and affecting approximately one in 1000 live births worldwide. The overexpression of genes encoded by the extra copy of a normal chromosome in DS is believed to be sufficient to perturb normal pathways and normal responses to stimulation, causing learning and memory deficits. In this work, we have designed a strategy based on the unsupervised clustering method, Self Organizing Maps (SOM), to identify biologically important differences in protein levels in mice exposed to context fear conditioning (CFC). We analyzed expression levels of 77 proteins obtained from normal genotype control mice and from their trisomic littermates (Ts65Dn) both with and without treatment with the drug memantine. Control mice learn successfully while the trisomic mice fail, unless they are first treated with the drug, which rescues their learning ability. The SOM approach identified reduced subsets of proteins predicted to make the most critical contributions to normal learning, to failed learning and rescued learning, and provides a visual representation of the data that allows the user to extract patterns that may underlie novel biological responses to the different kinds of learning and the response to memantine. Results suggest that the application of SOM to new experimental data sets of complex protein profiles can be used to identify common critical protein responses, which in turn may aid in identifying potentially more effective drug targets.