Probabilistic spatial analysis in quantitative microscopy with uncertainty-aware cell detection using deep Bayesian regression

The investigation of biological systems with three-dimensional microscopy demands automatic cell identification methods that not only are accurate but also can imply the uncertainty in their predictions. The use of deep learning to regress density maps is a popular successful approach for extracting cell coordinates from local peaks in a postprocessing step, which then, however, hinders any meaningful probabilistic output. We propose a framework that can operate on large microscopy images and output probabilistic predictions (i) by integrating deep Bayesian learning for the regression of uncertainty-aware density maps, where peak detection algorithms generate cell proposals, and (ii) by learning a mapping from prediction proposals to a probabilistic space that accurately represents the chances of a successful prediction. Using these calibrated predictions, we propose a probabilistic spatial analysis with Monte Carlo sampling. We demonstrate this in a bone marrow dataset, where our proposed methods reveal spatial patterns that are otherwise undetectable.

Chronic viral infections persistently alter marrow stroma and impair hematopoietic stem cell fitness

Chronic viral infections are associated with hematopoietic suppression, bone marrow (BM) failure, and hematopoietic stem cell (HSC) exhaustion. However, how persistent viral challenge and inflammatory responses target BM tissues and perturb hematopoietic competence remains poorly understood. Here, we combine functional analyses with advanced 3D microscopy to demonstrate that chronic infection with lymphocytic choriomeningitis virus leads to (1) long-lasting decimation of the BM stromal network of mesenchymal CXCL12-abundant reticular cells, (2) proinflammatory transcriptional remodeling of remaining components of this key niche subset, and (3) durable functional defects and decreased competitive fitness in HSCs. Mechanistically, BM immunopathology is elicited by virus-specific, activated CD8 T cells, which accumulate in the BM via interferon-dependent mechanisms. Combined antibody-mediated inhibition of type I and II IFN pathways completely preempts degeneration of CARc and protects HSCs from chronic dysfunction. Hence, viral infections and ensuing immune reactions durably impact BM homeostasis by persistently decreasing the competitive fitness of HSCs and disrupting essential stromal-derived, hematopoietic-supporting cues.

Modality Attention and Sampling Enables Deep Learning with Heterogeneous Marker Combinations in Fluorescence Microscopy

Fluorescence microscopy allows for a detailed inspection of cells, cellular networks, and anatomical landmarks by staining with a variety of carefully-selected markers visualized as color channels. Quantitative characterization of structures in acquired images often relies on automatic image analysis methods. Despite the success of deep learning methods in other vision applications, their potential for fluorescence image analysis remains underexploited. One reason lies in the considerable workload required to train accurate models, which are normally specific for a given combination of markers, and therefore applicable to a very restricted number of experimental settings. We herein propose Marker Sampling and Excite - a neural network approach with a modality sampling strategy and a novel attention module that together enable (i) flexible training with heterogeneous datasets with combinations of markers and (ii) successful utility of learned models on arbitrary subsets of markers prospectively. We show that our single neural network solution performs comparably to an upper bound scenario where an ensemble of many networks is naïvely trained for each possible marker combination separately. In addition, we demonstrate the feasibility of this framework in high-throughput biological analysis by revising a recent quantitative characterization of bone marrow vasculature in 3D confocal microscopy datasets and further confirm the validity of our approach on an additional, significantly different dataset of microvessels in fetal liver tissues. Not only can our work substantially ameliorate the use of deep learning in fluorescence microscopy analysis, but it can also be utilized in other fields with incomplete data acquisitions and missing modalities.

Multiparametric imaging reveals that mitochondria-rich intercalated cells in the kidney collecting duct have a very high glycolytic capacity

Alpha intercalated cells (αICs) in the kidney collecting duct (CD) belong to a family of mitochondria rich cells (MRCs) and have a crucial role in acidifying the urine via apical V-ATPase pumps. The nature of metabolism in αICs and its relationship to transport was not well-understood. Here, using multiphoton live cell imaging in mouse kidney tissue, FIB-SEM, and other complementary techniques, we provide new insights into mitochondrial structure and function in αICs. We show that αIC mitochondria have a rounded structure and are not located in close proximity to V-ATPase containing vesicles. They display a bright NAD(P)H fluorescence signal and low uptake of voltage-dependent dyes, but are energized by a pH gradient. However, expression of complex V (ATP synthase) is relatively low in αICs, even when stimulated by metabolic acidosis. In contrast, anaerobic glycolytic capacity is surprisingly high, and sufficient to maintain intracellular calcium homeostasis in the presence of complete aerobic inhibition. Moreover, glycolysis is essential for V-ATPase-mediated proton pumping. Key findings were replicated in narrow/clear cells in the epididymis, also part of the MRC family. In summary, using a range of cutting-edge techniques to investigate αIC metabolism in situ, we have discovered that these mitochondria dense cells have a high glycolytic capacity.

Imaging and spatial analysis of hematopoietic stem cell niches

Hematopoietic stem cells (HSCs) have been long proposed to reside in defined anatomical locations within bone marrow (BM) tissues in direct contact or close proximity to nurturing cell types. Imaging techniques that allow the simultaneous mapping of HSCs and interacting cell types have been central to the discovery of basic principles of these so-called HSC niches. Despite major progress in the field, a quantitative and comprehensive model of the cellular and molecular components that define these specialized microenvironments is lacking to date, and uncertainties remain on the preferential localization of HSCs in the context of complex BM tissue landscapes. Recent technological breakthroughs currently allow for the quantitative spatial analysis of BM cellular components with extraordinary precision. Here, we critically discuss essential technical aspects related to imaging approaches, image processing tools, and spatial statistics, which constitute the three basic elements of rigorous quantitative spatial analyses of HSC niches in the BM microenvironment.