Evolutionary design of explainable algorithms for biomedical image segmentation.

An unresolved issue in contemporary biomedicine is the overwhelming number and diversity of complex images that require annotation, analysis and interpretation. Recent advances in Deep Learning have revolutionized the field of computer vision, creating algorithms that compete with human experts in image segmentation tasks. However, these frameworks require large human-annotated datasets for training and the resulting "black box" models are difficult to interpret. In this study, we introduce Kartezio, a modular Cartesian Genetic Programming-based computational strategy that generates fully transparent and easily interpretable image processing pipelines by iteratively assembling and parameterizing computer vision functions. The pipelines thus generated exhibit comparable precision to state-of-the-art Deep Learning approaches on instance segmentation tasks, while requiring drastically smaller training datasets. This Few-Shot Learning method confers tremendous flexibility, speed, and functionality to this approach. We then deploy Kartezio to solve a series of semantic and instance segmentation problems, and demonstrate its utility across diverse images ranging from multiplexed tissue histopathology images to high resolution microscopy images. While the flexibility, robustness and practical utility of Kartezio make this fully explicable evolutionary designer a potential game-changer in the field of biomedical image processing, Kartezio remains complementary and potentially auxiliary to mainstream Deep Learning approaches.

Flow cytometry analysis of endothelial cells and subsets of exhausted CD8+ T cells in murine tumor models.

Here, we present a protocol for flow cytometry analysis of endothelial cells (ECs) and CD8+ T cells in murine tumor models, at baseline and after cancer immunotherapy with anti-PD-1/anti-CTLA-4 antibodies. We provide gating strategies for identification of specific cell subsets including ECs from tumor-associated high endothelial venules (TA-HEVs), stem-like, and terminally exhausted CD8+ T cells. This protocol represents a valuable tool for the analysis of rare subsets of tumor ECs and CD8+ T cells with critical roles in antitumor immunity. For complete details on the use and execution of this protocol, please refer to Asrir et al. (2022).

Phenotypic and Histological Distribution Analysis Identify Mast Cell Heterogeneity in Non-Small Cell Lung Cancer.

Mast cells (MCs) are multifaceted innate immune cells often present in the tumor microenvironment (TME). However, MCs have been only barely characterized in studies focusing on global immune infiltrate phenotyping. Consequently, their role in cancer is still poorly understood. Furthermore, their prognosis value is confusing since MCs have been associated with good and bad (or both) prognosis depending on the cancer type. In this pilot study performed on a surgical cohort of 48 patients with Non-Small Cell Lung Cancer (NSCLC), we characterized MC population within the TME and in matching non-lesional lung areas, by multicolor flow cytometry and confocal microscopy. Our results showed that tumor-associated MCs (TAMCs) harbor a distinct phenotype as compared with MCs present in non-lesional counterpart of the lung. Moreover, we found two TAMCs subsets based on the expression of CD103 (also named alphaE integrin). CD103+ TAMCs appeared more mature, more prone to interact with CD4+ T cells, and located closer to cancer cells than their CD103- counterpart. In spite of these characteristics, we did not observe a prognosis advantage of a high frequency of CD103+ TAMCs, while a high frequency of total TAMC correlated with better overall survival and progression free survival. Together, this study reveals that TAMCs constitute a heterogeneous population and indicates that MC subsets should be considered for patients' stratification and management in future research.

Tumor-associated high endothelial venules mediate lymphocyte entry into tumors and predict response to PD-1 plus CTLA-4 combination immunotherapy.

Recruitment of lymphocytes into tumors is critical for anti-tumor immunity and efficacious immunotherapy. We show in murine models that tumor-associated high endothelial venules (TA-HEVs) are major sites of lymphocyte entry into tumors at baseline and upon treatment with anti-PD-1/anti-CTLA-4 immune checkpoint blockade (ICB). TA-HEV endothelial cells (TA-HECs) derive from post-capillary venules, co-express MECA-79+ HEV sialomucins and E/P-selectins, and are associated with homing and infiltration into tumors of various T cell subsets. Intravital microscopy further shows that TA-HEVs are the main sites of lymphocyte arrest and extravasation into ICB-treated tumors. Increasing TA-HEC frequency and maturation increases the proportion of tumor-infiltrating stem-like CD8+ T cells, and ameliorates ICB efficacy. Analysis of tumor biopsies from 93 patients with metastatic melanoma reveals that TA-HEVs are predictive of better response and survival upon treatment with anti-PD-1/anti-CTLA-4 combination. These studies provide critical insights into the mechanisms governing lymphocyte trafficking in cancer immunity and immunotherapy.

Effector and stem-like memory cell fates are imprinted in distinct lymph node niches directed by CXCR3 ligands.

T cells dynamically interact with multiple, distinct cellular subsets to determine effector and memory differentiation. Here, we developed a platform to quantify cell location in three dimensions to determine the spatial requirements that direct T cell fate. After viral infection, we demonstrated that CD8+ effector T cell differentiation is associated with positioning at the lymph node periphery. This was instructed by CXCR3 signaling since, in its absence, T cells are confined to the lymph node center and alternatively differentiate into stem-like memory cell precursors. By mapping the cellular sources of CXCR3 ligands, we demonstrated that CXCL9 and CXCL10 are expressed by spatially distinct dendritic and stromal cell subsets. Unlike effector cells, retention of stem-like memory precursors in the paracortex is associated with CCR7 expression. Finally, we demonstrated that T cell location can be tuned, through deficiency in CXCL10 or type I interferon signaling, to promote effector or stem-like memory fates.

Stochastic asymmetric repartition of lytic machinery in dividing CD8+ T cells generates heterogeneous killing behavior.

Cytotoxic immune cells are endowed with a high degree of heterogeneity in their lytic function, but how this heterogeneity is generated is still an open question. We therefore investigated if human CD8+ T cells could segregate their lytic components during telophase, using imaging flow cytometry, confocal microscopy, and live-cell imaging. We show that CD107a+-intracellular vesicles, perforin, and granzyme B unevenly segregate in a constant fraction of telophasic cells during each division round. Mathematical modeling posits that unequal lytic molecule inheritance by daughter cells results from the random distribution of lytic granules on the two sides of the cleavage furrow. Finally, we establish that the level of lytic compartment in individual cytotoxic T lymphocyte (CTL) dictates CTL killing capacity.

Sequential adjustment of cytotoxic T lymphocyte densities improves efficacy in controlling tumor growth.

Understanding the human cytotoxic T lymphocyte (CTL) biology is crucial to develop novel strategies aiming at maximizing their lytic capacity against cancer cells. Here we introduce an agent-based model, calibrated on population-scale experimental data that allows quantifying human CTL per capita killing. Our model highlights higher individual CTL killing capacity at lower CTL densities and fits experimental data of human melanoma cell killing. The model allows extending the analysis over prolonged time frames, difficult to investigate experimentally, and reveals that initial high CTL densities hamper efficacy to control melanoma growth. Computational analysis forecasts that sequential addition of fresh CTL cohorts improves tumor growth control. In vivo experimental data, obtained in a mouse melanoma model, confirm this prediction. Taken together, our results unveil the impact that sequential adjustment of cellular densities has on enhancing CTL efficacy over long-term confrontation with tumor cells. In perspective, they can be instrumental to refine CTL-based therapeutic strategies aiming at controlling tumor growth.

Context-Dependent Role for T-bet in T Follicular Helper Differentiation and Germinal Center Function following Viral Infection.

Following infection, inflammatory cues upregulate core transcriptional programs to establish pathogen-specific protection. In viral infections, T follicular helper (TFH) cells express the prototypical T helper 1 transcription factor T-bet. Several studies have demonstrated essential but conflicting roles for T-bet in TFH biology. Understanding the basis of this controversy is crucial, as modulation of T-bet expression instructs TFH differentiation and ultimately protective antibody responses. Comparing influenza and LCMV viral infections, we demonstrate that the role of T-bet is contingent on the environmental setting of TFH differentiation, IL-2 signaling, and T cell competition. Furthermore, we demonstrate that T-bet expression by either TFH or GC B cells independently drives antibody isotype class switching. Specifically, T cell-specific loss of T-bet promotes IgG1, whereas B cell-specific loss of T-bet inhibits IgG2a/c switching. Combined, this work highlights that the context-dependent induction of T-bet instructs the development of protective, neutralizing antibodies following viral infection or vaccination.

Type I interferon induces CXCL13 to support ectopic germinal center formation.

Ectopic lymphoid structures form in a wide range of inflammatory conditions, including infection, autoimmune disease, and cancer. In the context of infection, this response can be beneficial for the host: influenza A virus infection-induced pulmonary ectopic germinal centers give rise to more broadly cross-reactive antibody responses, thereby generating cross-strain protection. However, despite the ubiquity of ectopic lymphoid structures and their role in both health and disease, little is known about the mechanisms by which inflammation is able to convert a peripheral tissue into one that resembles a secondary lymphoid organ. Here, we show that type I IFN produced after viral infection can induce CXCL13 expression in a phenotypically distinct population of lung fibroblasts, driving CXCR5-dependent recruitment of B cells and initiating ectopic germinal center formation. This identifies type I IFN as a novel inducer of CXCL13, which, in combination with other stimuli, can promote lung remodeling, converting a nonlymphoid tissue into one permissive to functional tertiary lymphoid structure formation.

A Task Force Against Local Inflammation and Cancer: Lymphocyte Trafficking to and Within the Skin.

The skin represents a specialized site for immune surveillance consisting of resident, inflammatory and memory populations of lymphocytes. The entry and retention of T cells, B cells, and ILCs is tightly regulated to facilitate detection of pathogens, inflammation and tumors cells. Loss of individual or multiple populations in the skin may break tolerance or increase susceptibility to tumor growth and spread. Studies have significantly advanced our understanding of the role of skin T cells and ILCs at steady state and in inflammatory settings such as viral challenge, atopy, and autoimmune inflammation. The knowledge raised by these studies can benefit to our understanding of immune cell trafficking in primary melanoma, shedding light on the mechanisms of tumor immune surveillance and to improve immunotherapy. This review will focus on the T cells, B cells, and ILCs of the skin at steady state, in inflammatory context and in melanoma. In particular, we will detail the core chemokine and adhesion molecules that regulate cell trafficking to and within the skin, which may provide therapeutic avenues to promote tumor homing for a team of lymphocytes.

Plasmacytoid dendritic cell heterogeneity is defined by CXCL10 expression following TLR7 stimulation.

Plasmacytoid dendritic cells (pDCs) play a critical role in bridging the innate and adaptive immune systems. pDCs are specialized type I interferon (IFN) producers, which has implicated them as initiators of autoimmune pathogenesis. However, little is known about the downstream effectors of type I IFN signaling that amplify autoimmune responses. Here, we have used a chemokine reporter mouse to determine the CXCR3 ligand responses in DCs subsets. Following TLR7 stimulation, conventional type 1 and type 2 DCs (cDC1 and cDC2, respectively) uniformly upregulate CXCL10. By contrast, the proportion of chemokine positive pDCs was significantly less, and stable CXCL10+ and CXCL10- populations could be distinguished. CXCL9 expression was induced in all cDC1s, in half of the cDC2 but not by pDCs. The requirement for IFNAR signaling for chemokine reporter expression was interrogated by receptor blocking and deficiency and shown to be critical for CXCR3 ligand expression in Flt3-ligand-derived DCs. Chemokine-producing potential was not concordant with the previously identified markers of pDC heterogeneity. Finally, we show that CXCL10+ and CXCL10- populations are transcriptionally distinct, expressing unique transcriptional regulators, IFN signaling molecules, chemokines, cytokines, and cell surface markers. This work highlights CXCL10 as a downstream effector of type I IFN signaling and suggests a division of labor in pDCs subtypes that likely impacts their function as effectors of viral responses and as drivers of inflammation.

L-selectin controls trafficking of chronic lymphocytic leukemia cells in lymph node high endothelial venules in vivo.

B-cell chronic lymphocytic leukemia (CLL) is the most common leukemia in adults. Lymph nodes (LNs) are sites of malignant proliferation and LN enlargement is associated with poor prognosis in the clinics. The LN microenvironment is believed to favor disease progression by promoting CLL cell growth and drug resistance. A better understanding of the mechanisms regulating trafficking of CLL cells to LNs is thus urgently needed. Here, we studied the first step of CLL cell migration to LNs, their interaction with high endothelial venules (HEVs), specialized blood vessels for lymphocyte extravasation in lymphoid organs. We observed that the density of HEV blood vessels was increased in CLL LNs and that CD20(+) CLL cells accumulated within HEV pockets, suggesting intense trafficking. We used intravital imaging to visualize the behavior of human CLL cells within the mouse LN microcirculation, and discovered that CLL cells bind to HEVs in vivo via a multistep adhesion cascade, which involves rolling, sticking, and crawling of the leukemic cells on the endothelium. Functional analyses revealed that the lymphocyte homing receptor L-selectin (CD62L) is the key factor controlling the binding of CLL cells to HEV walls in vivo. Interestingly, L-selectin expression was decreased on CLL cells from patients treated with idelalisib, a phosphoinositide-3-kinase δ inhibitor recently approved for CLL therapy. Interference with L-selectin-mediated trafficking in HEVs could represent a novel strategy to block dissemination of CLL cells to LNs and increase the efficacy of conventional therapy.

CIP4 controls CCL19-driven cell steering and chemotaxis in chronic lymphocytic leukemia.

Solid tumor dissemination relies on the reprogramming of molecular pathways controlling chemotaxis. Whether the motility of nonsolid tumors such as leukemia depends on the deregulated expression of molecules decoding chemotactic signals remains an open question. We identify here the membrane remodeling F-BAR adapter protein Cdc42-interacting protein 4 (CIP4) as a key regulator of chemotaxis in chronic lymphocytic leukemia (CLL). CIP4 is expressed at abnormally high levels in CLL cells, where it is required for CCL19-induced chemotaxis. Upon CCL19 stimulation of CLL cells, CIP4 associates with GTP-bound Cdc42 and is recruited to the rear of the lamellipodium and along microspikes radiating through the lamellipodium. Consistent with its cellular distribution, CIP4 removal impairs both the assembly of the polarized lamellipodium and directional migration along a diffusible CCL19 gradient. Furthermore, CIP4 depletion results in decreased activation of WASP, but increased activation of PAK1 and p38 mitogen-activated protein kinase (MAPK). Notably, p38 MAPK inhibition results in impaired lamellipodium assembly and loss of directional migration. This suggests that CIP4 modulates both the WASP and p38 MAPK pathways to promote lamellipodium assembly and chemotaxis. Overall, our study reveals a critical role of CIP4 in mediating chemotaxis of CLL cells by controlling the dynamics of microspike-containing protrusions and cell steering.

Blood leukocytes and macrophages of various phenotypes have distinct abilities to form podosomes and to migrate in 3D environments.

Leukocytes migrate through most tissues in the body, a process which takes place in 3D environments. We have previously shown that macrophages use the amoeboid migration mode in porous matrices such as fibrillar collagen I and the mesenchymal mode involving podosomes and matrix proteolysis in dense matrices such as Matrigel. Whether such a plasticity may apply to other leukocytes and to all subsets of macrophages is unknown. Here, we therefore provide a comparative analysis of the in vitro 3D migration modes adopted by primary human leukocytes. Blood-derived monocytes, neutrophils and T lymphocytes were found to use the amoeboid mode in a porous fibrillar collagen I matrix but were unable to infiltrate dense Matrigel and to form podosomes. M2-polarized macrophages and elicited peritoneal macrophages formed podosome rosettes, degraded the ECM and infiltrated both matrices. In contrast, M1 macrophages were motionless in 2D and 3D environments, whilst resident macrophages, devoid of podosomes, were only able to use the amoeboid mode. Thus, we conclude that whereas all leukocytes use the amoeboid mode to migrate through porous matrices, it is only certain macrophages that can adopt the mesenchymal mode that permits migration through dense matrices. Interestingly, the acquisition of mesenchymal migration capacity by macrophages correlates with the presence of podosomes and with their capacity to organize those as rosettes, which appears to be modulated by their differentiation and polarization states. As a perspective, specific control of the mesenchymal migration would be a potential target for therapeutic approaches aiming at decreasing macrophage tissue infiltration.

Wiskott-Aldrich syndrome protein controls antigen-presenting cell-driven CD4+ T-cell motility by regulating adhesion to intercellular adhesion molecule-1.

T-cell scanning for antigen-presenting cells (APC) is a finely tuned process. Whereas non-cognate APC trigger T-cell motility via chemokines and intercellular adhesion molecule-1 (ICAM-1), cognate APC deliver a stop signal resulting from antigen recognition. We tested in vitro the contribution of the actin cytoskeleton regulator Wiskott-Aldrich syndrome protein (WASP) to the scanning activity of primary human CD4(+)  T cells. WASP knock-down resulted in increased T-cell motility upon encounter with non-cognate dendritic cells or B cells and reduced capacity to stop following antigen recognition. The high motility of WASP-deficient T cells was accompanied by a diminished ability to round up and to stabilize pauses. WASP-deficient T cells migrated in a normal proportion towards CXCL12, CCL19 and CCL21, but displayed an increased adhesion and elongation on ICAM-1. The elongated morphology of WASP-deficient T cells was related to a reduced confinement of high-affinity lymphocyte function-associated antigen 1 to the mid-cell zone. Our data therefore indicate that WASP controls CD4(+) T-cell motility upon APC encounter by regulating lymphocyte function-associated antigen 1 spatial distribution.

The Wiskott-Aldrich syndrome protein permits assembly of a focused immunological synapse enabling sustained T-cell receptor signaling.

BACKGROUND: T-cell activation relies on the assembly of the immunological synapse, a structure tightly regulated by the actin cytoskeleton. The precise role of the Wiskott-Aldrich syndrome protein, an actin cytoskeleton regulator, in linking immunological synapse structure to downstream signaling remains to be clarified. DESIGN AND METHODS: To address this point, CD4(+) T cells from patients with Wiskott-Aldrich syndrome were stimulated with antigen-presenting cells. The structure and dynamics of the immunological synapse were studied by confocal and video-microscopy. RESULTS: Upon stimulation by antigen-presenting cells, Wiskott-Aldrich syndrome protein-deficient T cells displayed reduced cytokine production and proliferation. Although Wiskott-Aldrich syndrome T cells formed conjugates with antigen-presenting cells at normal frequency and exhibited normal T-cell receptor down-regulation, they emitted actin-rich protrusions away from the immunological synapse area and their microtubule organizing center failed to polarize fully towards the center of the immunological synapse. In parallel, abnormally dispersed phosphotyrosine staining revealed unfocused synaptic signaling in Wiskott-Aldrich syndrome T cells. Time-lapse microscopy confirmed the anomalous morphology of Wiskott-Aldrich syndrome T-cell immunological synapses and showed erratic calcium mobilization at the single-cell level. CONCLUSIONS: Taken together, our data show that the Wiskott-Aldrich syndrome protein is required for the assembly of focused immunological synapse structures allowing optimal signal integration and sustained calcium signaling.

Revertant T lymphocytes in a patient with Wiskott-Aldrich syndrome: analysis of function and distribution in lymphoid organs.

BACKGROUND: The Wiskott-Aldrich syndrome (WAS) is a rare genetic disease characterized by thrombocytopenia, immunodeficiency, autoimmunity, and hematologic malignancies. Secondary mutations leading to re-expression of WAS protein (WASP) are relatively frequent in patients with WAS. OBJECTIVE: The tissue distribution and function of revertant cells were investigated in a novel case of WAS gene secondary mutation. METHODS: A vast combination of approaches was used to characterize the second-site mutation, to investigate revertant cell function, and to track their distribution over a 18-year clinical follow-up. RESULTS: The WAS gene secondary mutation was a 4-nucleotide insertion, 4 nucleotides downstream of the original deletion. This somatic mutation allowed the T-cell-restricted expression of a stable, full-length WASP with a 3-amino acid change compared with the wild-type protein. WASP(+) T cells appeared early in the spleen (age 10 years) and were highly enriched in a mesenteric lymph node at a later time (age 23 years). Revertant T cells had a diversified T-cell-receptor repertoire and displayed in vitro and in vivo selective advantage. They proliferated and produced cytokines normally on T-cell-receptor stimulation. Consistently, the revertant WASP correctly localized to the immunologic synapse and to the leading edge of migrating T cells. CONCLUSION: Despite the high proportion of functional revertant T cells, the patient still has severe infections and autoimmune disorders, suggesting that re-expression of WASP in T cells is not sufficient to normalize immune functions fully in patients with WAS.