Biology subjects

Keren, L.

Publications and source records attributed to Keren, L..

15 recordsLinked to original sources

bioRxiv2026

Spatially enforced engagement of PD-1⁺ T cells with AXL⁺ cells reprograms the tumor microenvironment and overcomes anti-PD-1 resistance

Background: AXL overexpression promotes tumor-cell plasticity, local immune suppression, and resistance to PD-1 blockade. Conventional anti-AXL and anti-PD-1 antibodies combinations act independently and do not ensure physical proximity between PD-1+ T cells and AXL+ cells within the immunosuppressive tumor microenvironment. Methods: To address this limitation, we developed a non-tandem, tetravalent IgG bispecific antibody designed to simultaneously engage PD-1 and AXL (BiTE-AXL/PD-1). Its antitumor activity was evaluated using in vitro functional assays and humanized and immunocompetent head and neck squamous cell carcinoma (HNSCC) models and compared with the combination of monospecific anti-AXL and anti-PD-1 antibodies. Spatial multiplex imaging was used to characterize treatment-associated changes in the tumor microenvironment of an anti-PD-1-resistant murine model. Results: BiTE-AXL/PD-1 enhanced T-cell activation, promoted AXL-dependent tumor-T-cell conjugation, and increased tumor-cell killing in vitro, outperforming the combination of the two monospecific antibodies. In humanized and immunocompetent HNSCC models, BiTE treatment significantly delayed tumor growth compared with the antibody combination and was well tolerated, with no overt toxicity. Spatial multiplex imaging revealed a shift from a fibroblast-dominated microenvironment characterized by exhausted CD8+ T cells toward dendritic-cell- and CD8+ T-cell-enriched niches associated with antitumor immune activity. Conclusion: BiTE-AXL/PD-1 promotes physical proximity between PD-1+ T cells and AXL+ cells, enhances antitumor immune responses, and remodels the immunosuppressive tumor microenvironment. These findings support PD-1/AXL bispecific targeting as a promising therapeutic strategy to overcome resistance to PD-1 blockade in HNSCC.

cancer biology↗

bioRxiv2026

Single-section multiplexed imaging enables comprehensive lung cancer diagnosis

Accurate and timely diagnosis is essential for effective lung cancer treatment. However, contemporary workflows rely on sequential immunohistochemistry of small biopsy specimens, which can exhaust tissue, limit biomarker assessment, and delay treatment decisions. Here, we demonstrate that multiplexed imaging addresses these limitations by enabling comprehensive lung cancer diagnosis from a single tissue section. We developed and validated a clinically informed multiplexed antibody panel that integrates tumor classification, predictive biomarker assessment, and immune profiling. In diagnostic biopsies, multiplexed imaging achieved 96% concordance with standard pathology, while enabling accurate automated PD-L1 scoring and rapid detection of clinically approved and emerging actionable targets. Simultaneously measuring dozens of proteins improves standard pathology by incorporating complex multi-protein biomarkers, supporting quantitative computational analysis to streamline diagnosis, and generating spatial data for translational research. By reducing turnaround time and preserving scarce tissue, this workflow has the potential to accelerate treatment decisions, improve patient outcomes and bridge clinical care with translational discovery.

cancer biology↗

bioRxiv2025

CellTune: An integrative software for accurate cell classification in spatial proteomics

Spatial proteomics measures multiple proteins in situ, capturing tissue complexity. However, cell classification in densely packed tissues remains challenging due to the lack of efficient classification algorithms, annotation tools, and high-quality labeled datasets to benchmark computational methods. We introduce CellTune, an integrated software for analysis of large spatial proteomics datasets, which streamlines precise cell classification through an optimized human-in-the-loop active learning workflow. It advances core capabilities across within a unified, intuitive, and code-free interface. To evaluate CellTune, we created CellTuneDepot, a resource of 40k manually-annotated cells and 3.5 million high-quality labeled cells across 60 cell types. CellTune outperforms alternative methods, achieving accuracy comparable to human performance while enabling increased classification resolution and discovery of novel cell types. Together, CellTune and CellTuneDepot provide researchers with a tool for state-of-the-art classification accuracy and resolution at scale to drive biological insights.

bioinformatics↗

bioRxiv2025

Context-dependent spatial multicellular network motifs for single-cell spatial biology

The clinical state of diseased tissue is caused by complex intercellular processes that go beyond pairwise cell-cell interactions and are difficult to infer due to the combinatorial explosion of such high-dimensionality. We present context-dependent identification of spatial motifs (CISM), a two-step method to identify local cell structures associated with a disease state in single cell spatial data. First, for each tissue, CISM enumerates structures of enriched reoccurring multicellular patterns that define modular motifs in the multicellular network. Second, discriminative motifs are selected according to the context - their presence in patients at different clinical disease states. By applying CISM, we show that modular structures composed of as little as 3-5 cells and their relative spatial arrangement can encode differences in clinical disease states in cohorts of triple-negative breast cancer (TNBC) and melanoma patients. Machine learning validation indicated that discriminative motifs outperform state-of-the-art methods for disease state prediction while enabling interpretation of which interactions in what spatial context are associated with these predictions. CISM-derived discriminative motifs may define an intermediate spatial scale of abstraction and modularity in multicellular organization and function with broad applicability in the domain of spatial single cell omics and beyond.

systems biology↗

bioRxiv2024

Immune organization in sentinel lymph nodes of melanoma patients is prognostic of distant metastases

Sentinel lymph node (sLN) biopsy is part of melanoma staging, as involved LNs indicate a higher risk of recurrence. However, how the sLN is shaped by the tumor and reciprocally affects metastatic progression is poorly understood. Here, we mapped immune organization in involved and non-involved sLNs of 69 melanoma patients using high-resolution spatial proteomics, spatial transcriptomics and deep learning, leveraging the data for prognostic evaluation. In patients with involved LNs, a robust T cell response correlated with absence of recurrence. In non-involved LNs, protection from metastases was linked to expanded sinuses colocalized with plasmablasts whereas CCR7+ cells and Tregs correlated with future development of distant metastases. We trained a model that predicts development of distant metastases with 79% and 93% AUC for non-metastatic and metastatic LNs, respectively. Our findings reveal conserved immune patterns in sLNs that prospectively identify patients at risk for metastatic disease and may aid in therapeutic decisions.

cancer biology↗

bioRxiv2024

Generalized cell phenotyping for spatial proteomics with language-informed vision models

We present DeepCell Types, a novel approach to cell phenotyping for spatial proteomics that addresses the challenge of generalization across diverse datasets with varying marker panels collected across different platforms. Our approach utilizes a transformer with channel-wise attention to create a language-informed vision model; this models semantic understanding of the underlying marker panel enables it to learn from and adapt to heterogeneous datasets. Leveraging a curated, diverse dataset named Expanded TissueNet with cell type labels spanning the literature and the NIH Human BioMolecular Atlas Program (HuBMAP) consortium, our model demonstrates robust performance across various cell types, tissues, and imaging modalities. Comprehensive benchmarking shows that our method outperforms existing approaches on cell-type prediction and, from the same model, predicts marker positivity competitively with a dedicated specialist; it further matches manual expert gating and adapts to new data with modest fine-tuning, well past what baselines reach when trained from scratch. This work equips the spatial proteomics community with a single, continuously improvable phenotyping model that generalizes to new marker panels and can be fine-tuned efficiently when needed. We release both DeepCell Types and Expanded TissueNet as open-source resources.

bioinformatics↗

bioRxiv2024

A spatial atlas of human gastro-intestinal acute GVHD reveals epithelial and immune dynamics underlying disease pathophysiology

Acute graft-versus-host disease (aGVHD) is a significant complication of allogeneic hematopoietic stem cell transplantation (aHSCT), driven by alloreactive donor T cells in the gut. However, the roles of additional donor and host cells in this process are not fully understood. We conducted multiplexed imaging on 59 biopsies from patients with gastrointestinal GVHD and 10 healthy controls, revealing key pathological changes, including fibrosis, crypt alterations, loss of Paneth cells, accumulation of endocrine cells, and disrupted immune organization, particularly a reduction in IgA-secreting plasma cells. Interestingly, CD8T cells were enriched only in a subset of patients, while others exhibited non-canonical enrichments of macrophages and neutrophils. Post-transplantation time significantly influenced immune composition, with host cells dominating plasma and T cell compartments long after transplantation. This spatial atlas of healthy duodenum and GVHD uncovers non-canonical immune dynamics, offering insights into disease pathophysiology and potential clinical applications in GVHD and other inflammatory bowel diseases.

immunology↗

bioRxiv2024

Data-modeling the interplay between single cell shape, single cell protein expression, and tissue state

Changes in cell shape are fundamentally involved in signaling, intracellular organization, function, and intercellular interactions within tissues, in health and disease. Investigating the interplay between cell shape and protein expression was limited, until recently, by the number of proteins that can be imaged simultaneously or by population averaging. We combined spatial multiplexed single cell imaging and machine learning to systematically investigate the intricate relationships between cell shape and protein expression in the context of heterogeneous human cells in their native state in human tissue samples in situ. Our analysis established a universal bi-directional link between the cells shape and its protein expression across different cell types, diseases, and disease states in human tissues, enabling new applications. Machine learning interpretability showed that the contribution of shape features to a prediction can potentially infer new protein functions. Unbiased screening of the links between all pairs consisting of one protein and one cell type identified a subpopulation of large p53-positive tumor cells across two cancers. Ultimately, inclusion of single cell shape properties enhanced Graph Neural Network disease state prediction. Our results open the door to unraveling the intricate connections between protein expression at the single cell level, cell shape, tissue organization, and tissue state in a physiological context.

cell biology↗

bioRxiv2023

Escalating High-dimensional Imaging using Combinatorial Channel Multiplexing and Deep Learning

Understanding tissue structure and function requires tools that quantify the expression of multiple proteins at single-cell resolution while preserving spatial information. Current imaging technologies use a separate channel for each individual protein, inherently limiting their throughput and scalability. Here, we present CombPlex (COMBinatorial multiPLEXing), a combinatorial staining platform coupled with an algorithmic framework to exponentially increase the number of proteins that can be measured from C up to 2c - 1. In CombPlex, every protein can be imaged in several channels, and every channel contains agglomerated images of several proteins. These combinatorically-compressed images are then decompressed to individual protein-images using deep learning. We achieve accurate reconstruction when compressing the stains of twenty-two proteins to five imaging channels and demonstrate that the approach works in both fluorescence microscopy and in mass-based imaging. Combinatorial staining coupled with deep-learning decompression can escalate the number of proteins measured using any imaging modality, without the need for specialized instrumentation. Coupling CombPlex with instruments for high-dimensional imaging could pave the way to image hundreds of proteins at single-cell resolution in intact tissue sections.

systems biology↗

bioRxiv2022

CellSighter - A neural network to classify cells in highly multiplexed images

Multiplexed imaging enables measurement of multiple proteins in situ, offering an unprecedented opportunity to chart various cell types and states in tissues. However, cell classification, the task of identifying the type of individual cells, remains challenging, labor-intensive, and limiting to throughput. Here, we present CellSighter, a deep-learning based pipeline to accelerate cell classification in multiplexed images. Given a small training set of expert-labeled images, CellSighter outputs the label probabilities for all cells in new images. CellSighter achieves over 80% accuracy for major cell types across imaging platforms, which approaches inter-observer concordance. Ablation studies and simulations show that CellSighter is able to generalize its training data and learn features of protein expression levels, as well as spatial features such as subcellular expression patterns. CellSighters design reduces overfitting, and it can be trained with only thousands or even hundreds of labeled examples. CellSighter also outputs a prediction confidence, allowing downstream experts control over the results. Altogether, CellSighter drastically reduces hands-on time for cell classification in multiplexed images, while improving accuracy and consistency across datasets.

systems biology↗

bioRxiv2021

Spatio-temporal coordination at the maternal-fetal interface promotes trophoblast invasion and vascular remodeling in the first half of human pregnancy

Beginning in the first trimester, fetally derived extravillous trophoblasts (EVTs) invade the uterus and remodel its spiral arteries, transforming them into large, dilated blood vessels left with a thin, discontinuous smooth muscle layer and partially lined with EVTs. Several mechanisms have been proposed to explain how EVTs coordinate with the maternal decidua to promote a tissue microenvironment conducive to spiral artery remodeling (SAR). However, it remains a matter of debate which immune and stromal cell types participate in these interactions and how this process evolves with respect to gestational age. Here, we used a multiomic approach that combined the strengths of spatial proteomics and transcriptomics to construct the first spatiotemporal atlas of the human maternal-fetal interface in the first half of pregnancy. We used multiplexed ion beam imaging by time of flight (MIBI-TOF) and a 37-plex antibody panel to analyze [~]500,000 cells and 588 spiral arteries within intact decidua from 66 patients between 6-20 weeks of gestation, integrating this with coregistered transcriptomic profiles. Gestational age substantially influenced the frequency of many maternal immune and stromal cells, with tolerogenic subsets expressing CD206, CD163, TIM-3, Galectin-9, and IDO-1 increasingly enriched and colocalized at later time points. In contrast, SAR progression preferentially correlated with EVT invasion and was transcriptionally defined by 78 gene ontology pathways exhibiting unique monotonic and biphasic trends. Lastly, we developed an integrated model of SAR supporting an intravasation mechanism where invasion is accompanied by upregulation of pro-angiogenic, immunoregulatory EVT programs that promote interactions with vascular endothelium while avoiding activation of immune cells in circulating maternal blood. Taken together, these results support a coordinated model of decidualization in which increasing gestational age drives a transition in maternal decidua towards a tolerogenic niche conducive to locally regulated, EVT-dependent SAR.

systems biology↗

bioRxiv2021

Whole-cell segmentation of tissue images with human-level performance using large-scale data annotation and deep learning

Understanding the spatial organization of tissues is of critical importance for both basic and translational research. While recent advances in tissue imaging are opening an exciting new window into the biology of human tissues, interpreting the data that they create is a significant computational challenge. Cell segmentation, the task of uniquely identifying each cell in an image, remains a substantial barrier for tissue imaging, as existing approaches are inaccurate or require a substantial amount of manual curation to yield useful results. Here, we addressed the problem of cell segmentation in tissue imaging data through large-scale data annotation and deep learning. We constructed TissueNet, an image dataset containing >1 million paired whole-cell and nuclear annotations for tissue images from nine organs and six imaging platforms. We created Mesmer, a deep learning-enabled segmentation algorithm trained on TissueNet that performs nuclear and whole-cell segmentation in tissue imaging data. We demonstrated that Mesmer has better speed and accuracy than previous methods, generalizes to the full diversity of tissue types and imaging platforms in TissueNet, and achieves human-level performance for whole-cell segmentation. Mesmer enabled the automated extraction of key cellular features, such as subcellular localization of protein signal, which was challenging with previous approaches. We further showed that Mesmer could be adapted to harness cell lineage information present in highly multiplexed datasets. We used this enhanced version to quantify cell morphology changes during human gestation. All underlying code and models are released with permissive licenses as a community resource.

bioinformatics↗

bioRxiv2021

Multiplexed Imaging Analysis of the Tumor-Immune Microenvironment Reveals Predictors of Outcome in Triple-Negative Breast Cancer

Triple-negative breast cancer, the poorest-prognosis breast cancer subtype, lacks clinically approved biomarkers for patient risk stratification and treatment management. Prior literature has shown that interrogation of the tumor-immune microenvironment may be a promising approach for the discovery of methods to fill these gaps. Recently developed high-dimensional tissue imaging technology, such as multiplexed ion beam imaging, provide spatial context to protein expression in the microenvironment, allowing in-depth characterization of cellular processes. We demonstrate that profiling the functional proteins involved in cell-to-cell interactions in the microenvironment can predict recurrence and overall survival. We highlight the immunological relevance of the immunoregulatory proteins PD-1, PD-L1, IDO, and Lag3 by tying interactions involving them to recurrence and survival. Multivariate analysis reveals that our methods provide additional prognostic information compared to clinical variables. In this work, we present a computational pipeline for the examination of the tumor-immune microenvironment using multiplexed ion-beam imaging that produces interpretable results, and is generalizable to other cancer types.

cancer biology↗

bioRxiv2021

Transition to invasive breast cancer is associated with progressive changes in the structure and composition of tumor stroma

Ductal carcinoma in situ (DCIS) is a pre-invasive lesion that is thought to be a precursor to invasive breast cancer (IBC). To understand how the tumor microenvironment (TME) changes with transition to IBC, we used Multiplexed Ion Beam Imaging by time of flight (MIBI-TOF) and a 37-plex antibody staining panel to analyze 140 clinically annotated surgical resections covering the full spectrum of breast cancer progression. We compared normal, DCIS, and IBC tissues using machine learning tools for multiplexed cell segmentation, pixel-based clustering, and object morphometrics. Transition from DCIS to IBC was found to occur along a trajectory marked by coordinated shifts in location and function of myoepithelium, fibroblasts, and infiltrating immune cells in the surrounding stroma. Taken together, this comprehensive study within the HTAN Breast PreCancer Atlas offers insight into the etiologies of DCIS, its transition to IBC, and emphasizes the importance of the TME stroma in promoting these processes.

cancer biology↗

bioRxiv2020

Multiplexed imaging of human tuberculosis granulomas uncovers immunoregulatory features conserved across tissue and blood

Tuberculosis (TB) is an infectious disease caused by Mycobacterium tuberculosis that is distinctly characterized by granuloma formation within infected tissues. Granulomas are dynamic and organized immune cell aggregates that limit dissemination, but can also hinder bacterial clearance. Consequently, outcome in TB is influenced by how granuloma structure and composition shift the balance between these two functions. To date, our understanding of what factors drive granuloma function in humans is limited. With this in mind, we used Multiplexed Ion Beam Imaging by Time-of-Flight (MIBI-TOF) to profile 37 proteins in tissues from thirteen patients with active TB disease from the U.S. and South Africa. With this dataset, we constructed a comprehensive tissue atlas where the lineage, functional state, and spatial distribution of 19 unique cell subsets were mapped onto eight phenotypically-distinct granuloma microenvironments. This work revealed an immunosuppressed microenvironment specific to TB granulomas with spatially coordinated co-expression of IDO1 and PD-L1 by myeloid cells and proliferating regulatory T cells. Interestingly, this microenvironment lacked markers consistent with T-cell activation, supporting a myeloid-mediated mechanism of immune suppression. We observed similar trends in gene expression of immunoregulatory proteins in a confirmatory transcriptomic analysis of peripheral blood collected from over 1500 individuals with latent or active TB infection and healthy controls across 29 cohorts spanning 14 countries. Notably, PD-L1 gene expression was found to correlate with TB progression and treatment response, supporting its potential use as a blood-based biomarker. Taken together, this study serves as a framework for leveraging independent cohorts and complementary methodologies to understand how local and systemic immune responses are linked in human health and disease.

immunology↗