Trung Pham

S1P1 receptor signaling overrides retention mediated by Gαi-coupled receptors to promote T cell egress

The mechanism by which sphingosine-1-phosphate receptor-1 (S1P1) acts to promote lymphocyte egress from lymphoid organs is not defined. Here, we showed that CCR7-deficient T cells left lymph nodes more rapidly than wild-type cells did, whereas CCR7-overexpressing cells were retained for longer. After treatment with FTY720, an agonist that causes downmodulation of lymphocyte S1P1, CCR7-deficient T cells were less effectively retained than wild-type T cells. Moreover, treatment with pertussis toxin to inactivate signaling via G alpha i-protein-coupled receptors restored egress competence to S1P1-deficient lymphocytes. We also found that T cell accumulation in lymph node cortical sinusoids required intrinsic S1P1 expression and was antagonized by CCR7. These findings suggest a model where S1P1 acts in the lymphocyte to promote lymph node egress by overcoming retention signals mediated by CCR7 and additional G alpha i-coupled receptors. Furthermore, by simultaneously upregulating S1P1 and downregulating CCR7, T cells that have divided multiple times switch to a state favoring egress over retention.

Sphingosine-1-phosphate in the plasma compartment regulates basal and inflammation-induced vascular leak in mice

Maintenance of vascular integrity is critical for homeostasis, and temporally and spatially regulated vascular leak is a central feature of inflammation. Sphingosine-1-phosphate (S1P) can regulate endothelial barrier function, but the sources of the S1P that provide this activity in vivo and its importance in modulating different inflammatory responses are unknown. We report here that mutant mice engineered to selectively lack S1P in plasma displayed increased vascular leak and impaired survival after anaphylaxis, administration of platelet-activating factor (PAF) or histamine, and exposure to related inflammatory challenges. Increased leak was associated with increased interendothelial cell gaps in venules and was reversed by transfusion with wild-type erythrocytes (which restored plasma S1P levels) and by acute treatment with an agonist for the S1P receptor 1 (S1pr1). S1pr1 agonist did not protect wild-type mice from PAF-induced leak, consistent with plasma S1P levels being sufficient for S1pr1 activation in wild-type mice. However, an agonist for another endothelial cell Gi-coupled receptor, Par2, did protect wild-type mice from PAF-induced vascular leak, and systemic treatment with pertussis toxin prevented rescue by Par2 agonist and sensitized wild-type mice to leak-inducing stimuli in a manner that resembled the loss of plasma S1P. Our results suggest that the blood communicates with blood vessels via plasma S1P to maintain vascular integrity and regulate vascular leak. This pathway prevents lethal responses to leak-inducing mediators in mouse models.

Cortical sinus probing, S1P1-dependent entry and flow-based capture of egressing T cells.

The cellular dynamics of the egress of lymphocytes from lymph nodes are poorly defined. Here we visualized the branched organization of lymph node cortical sinuses and found that after entry, some T cells were retained, whereas others returned to the parenchyma. T cells deficient in sphingosine 1-phosphate receptor type 1 probed the sinus surface but failed to enter the sinuses. In some sinuses, T cells became rounded and moved unidirectionally. T cells traveled from cortical sinuses into macrophage-rich sinus areas. Many T cells flowed from medullary sinuses into the subcapsular space. We propose a multistep model of lymph node egress in which cortical sinus probing is followed by entry dependent on sphingosine 1-phosphate receptor type 1, capture of cells in a sinus region with flow, and transport to medullary sinuses and the efferent lymph.

Epistasis between mouse Klra and major histocompatibility complex class I loci is associated with a new mechanism of natural killer cell-mediated innate resistance to cytomegalovirus infection

Experimental infection with mouse cytomegalovirus (MCMV) has been used to elucidate the intricate host-pathogen mechanisms that determine innate resistance to infection. Linkage analyses in F2 progeny from MCMV-resistant MA/My (H2 k) and MCMV-susceptible BALB/c (H2 d) and BALB.K (H2 k) mouse strains indicated that only the combination of alleles encoded by a gene in the Klra (also called Ly49) cluster on chromosome 6, and one in the major histocompatibility complex (H2) on chromosome 17, is associated with virus resistance. We found that natural killer cell–activating receptor Ly49P specifically recognized MCMV-infected cells, dependent on the presence of the H2 k haplotype. This binding was blocked using antibodies to H-2Dk but not antibodies to H-2Kk. These results are suggestive of a new natural killer cell mechanism implicated in MCMV resistance, which depends on the functional interaction of the Ly49P receptor and the major histocompatibility complex class I molecule H-2Dk on MCMV-infected cells.

GRK2-Dependent S1PR1 Desensitization Is Required for Lymphocytes to Overcome Their Attraction to Blood

A molecular mechanism that allows lymphocytes to migrate against a chemokine gradient is revealed. Lymphocytes egress from lymphoid organs in response to sphingosine-1-phosphate (S1P); minutes later they migrate from blood into tissue against the S1P gradient. The mechanisms facilitating cell movement against the gradient have not been defined. Here, we show that heterotrimeric guanine nucleotide–binding protein–coupled receptor kinase-2 (GRK2) functions in down-regulation of S1P receptor-1 (S1PR1) on blood-exposed lymphocytes. T and B cell movement from blood into lymph nodes is reduced in the absence of GRK2 but is restored in S1P-deficient mice. In the spleen, B cell movement between the blood-rich marginal zone and follicles is disrupted by GRK2 deficiency and by mutation of an S1PR1 desensitization motif. Moreover, delivery of systemic antigen into follicles is impaired. Thus, GRK2-dependent S1PR1 desensitization allows lymphocytes to escape circulatory fluids and migrate into lymphoid tissues.

DOCK8 is essential for T-cell survival and the maintenance of CD8+ T-cell memory.

Deficiency in the guanine nucleotide exchange factor dedicator of cytokinesis 8 (DOCK8) causes a human immunodeficiency syndrome associated with recurrent sinopulmonary and viral infections. We have recently identified a DOCK8‐deficient mouse strain, carrying an ethylnitrosourea‐induced splice‐site mutation that shows a failure to mature a humoral immune response due to the loss of germinal centre B cells. In this study, we turned to T‐cell immunity to investigate further the human immunodeficiency syndrome and its association with decreased peripheral CD4+ and CD8+ T cells. Characterisation of the DOCK8‐deficient mouse revealed T‐cell lymphopenia, with increased T‐cell turnover and decreased survival. Egress of mature CD4+ thymocytes was reduced with increased migration of these cells to the chemokine CXCL12. However, despite the two‐fold reduction in peripheral naïve T cells, the DOCK8‐deficient mice generated a normal primary CD8+ immune response and were able to survive acute influenza virus infection. The limiting effect of DOCK8 was in the normal survival of CD8+ memory T cells after infection. These findings help to explain why DOCK8‐deficient patients are susceptible to recurrent infections and provide new insights into how T‐cell memory is sustained.

The Random Cluster Model for Robust Geometric Fitting

Random hypothesis generation is central to robust geometric model fitting in computer vision. The predominant technique is to randomly sample minimal or elemental subsets of the data, and hypothesize the geometric model from the selected subsets. While taking minimal subsets increases the chance of simultaneously “hitting” inliers in a sample, it amplifies the noise of the underlying model, and hypotheses fitted on minimal subsets may be severely biased even if they contain purely inliers. In this paper we propose to use Random Cluster Models, a technique used to simulate coupled spin systems, to conduct hypothesis generation using subsets larger than minimal. We show how large clusters of data from genuine instances of the geometric model can be efficiently harvested to produce more accurate hypotheses. To take advantage of our hypothesis generator, we construct a simple annealing method based on graph cuts to fit multiple instances of the geometric model in the data. Experimental results show clear improvements in efficiency over other methods based on minimal subset samplers.

Interacting Geometric Priors For Robust Multimodel Fitting

Recent works on multimodel fitting are often formulated as an energy minimization task, where the energy function includes fitting error and regularization terms, such as low-level spatial smoothness and model complexity. In this paper, we introduce a novel energy with high-level geometric priors that consider interactions between geometric models, such that certain preferred model configurations may be induced. We argue that in many applications, such prior geometric properties are available and should be fruitfully exploited. For example, in surface fitting to point clouds, the building walls are usually either orthogonal or parallel to each other. Our proposed energy function is useful in dealing with unknown distributions of data errors and outliers, which are often the factors leading to biased estimation. Furthermore, the energy can be efficiently minimized using the expansion move method. We evaluate the performance on several vision applications using real data sets. Experimental results show that our method outperforms the state-of-the-art methods without significant increase in computation.

Geometrically consistent plane extraction for dense indoor 3D maps segmentation

Modern SLAM systems with a depth sensor are able to reliably reconstruct dense 3D geometric maps of indoor scenes. Representing these maps in terms of meaningful entities is a step towards building semantic maps for autonomous robots. One approach is to segment the 3D maps into semantic objects using Conditional Random Fields (CRF), which requires large 3D ground truth datasets to train the classification model. Additionally, the CRF inference is often computationally expensive. In this paper, we present an unsupervised geometric-based approach for the segmentation of 3D point clouds into objects and meaningful scene structures. We approximate an input point cloud by an adjacency graph over surface patches, whose edges are then classified as being either on or off. We devise an effective classifier which utilises both global planar surfaces and local surface convexities for edge classification. More importantly, we propose a novel global plane extraction algorithm for robustly discovering the underlying planes in the scene. Our algorithm is able to enforce the extracted planes to be mutually orthogonal or parallel which conforms usually with human-made indoor environments. We reconstruct 654 3D indoor scenes from NYUv2 sequences to validate the efficiency and effectiveness of our segmentation method.

Hierarchical Higher-Order Regression Forest Fields: An Application to 3D Indoor Scene Labelling

This paper addresses the problem of semantic segmentation of 3D indoor scenes reconstructed from RGB-D images. Traditionally label prediction for 3D points is tackled by employing graphical models that capture scene features and complex relations between different class labels. However, the existing work is restricted to pairwise conditional random fields, which are insufficient when encoding rich scene context. In this work we propose models with higher-order potentials to describe complex relational information from the 3D scenes. Specifically, we relax the labelling problem to a regression, and generalize the higher-order associative P n Potts model to a new family of arbitrary higher-order models based on regression forests. We show that these models, like the robust P n models, can still be decomposed into the sum of pairwise terms by introducing auxiliary variables. Moreover, our proposed higher-order models also permit extension to hierarchical random fields, which allows for the integration of scene context and features computed at different scales. Our potential functions are constructed based on regression forests encoding Gaussian densities that admit efficient inference. The parameters of our model are learned from training data using a structured learning approach. Results on two datasets show clear improvements over current state-of-the-art methods.