Membrane marker selection for segmenting single cell spatial proteomics data

Using adjusted local assortativity with Molecular Pixelation unveils colocalization of membrane proteins with immunological significance

Advances in spatial proteomics and protein colocalization are a driving force in the understanding of cellular mechanisms and their influence on biological processes. New methods in the field of spatial proteomics call for the development of algorithms and open up new avenues of research. The newly introduced Molecular Pixelation (MPX) provides spatial information on surface proteins and their relationship with each other in single cells. This allows for in silico representation of neighborhoods of membrane proteins as graphs. In order to analyze this new data modality, we adapted local assortativity in networks of MPX single-cell graphs and created a method that is able to capture detailed information on the spatial relationships of proteins. The introduced method can evaluate the pairwise colocalization of proteins and access higher-order similarity to investigate the colocalization of multiple proteins at the same time. We evaluated the method using publicly available MPX datasets where T cells were treated with a chemokine to study uropod formation. We demonstrate that adjusted local assortativity detects the effects of the stimuli at both single- and multiple-marker levels, which enhances our understanding of the uropod formation. We also applied our method to treating cancerous B-cell lines using a therapeutic antibody. With the adjusted local assortativity, we recapitulated the effect of rituximab on the polarity of CD20. Our computational method together with MPX improves our understanding of not only the formation of cell polarity and protein colocalization under stimuli but also advancing the overall insight into immune reaction and reorganization of cell surface proteins, which in turn allows the design of novel therapies. We foresee its applicability to other types of biological spatial data when represented as undirected graphs.

Optimizing deep learning-based segmentation of densely packed cells using cell surface markers

BACKGROUND

Spatial molecular profiling depends on accurate cell segmentation. Identification and quantitation of individual cells in dense tissues, e.g. highly inflamed tissue caused by viral infection or immune reaction, remains a challenge.

METHODS

We first assess the performance of 18 deep learning-based cell segmentation models, either pre-trained or trained by us using two public image sets, on a set of immunofluorescence images stained with immune cell surface markers in skin tissue obtained during human herpes simplex virus (HSV) infection. We then further train eight of these models using up to 10,000+ training instances from the current image set. Finally, we seek to improve performance by tuning parameters of the most successful method from the previous step.

RESULTS

The best model before fine-tuning achieves a mean Average Precision (mAP) of 0.516. Prediction performance improves substantially after training. The best model is the cyto model from Cellpose. After training, it achieves an mAP of 0.694; with further parameter tuning, the mAP reaches 0.711.

CONCLUSION

Selecting the best model among the existing approaches and further training the model with images of interest produce the most gain in prediction performance. The performance of the resulting model compares favorably to human performance. The imperfection of the final model performance can be attributed to the moderate signal-to-noise ratio in the imageset.

Tissue biomarkers of immune checkpoint inhibitor therapy

Cancer immunotherapy has been rejuvenated by the growing understanding of the immune system's role in tumor activity over the past two decades. During cancer initiation and progression, tumor cells employ various mechanisms that resemble peripheral immune tolerance to evade the antitumor responses of the immune system. Immune checkpoint molecules are the major mechanism of immune resistance that are exploited by tumor cells to inhibit T-cell activation and suppress immune responses. The targeting of immune checkpoint pathways has led to substantial improvements in survival rates in a number of solid cancers. However, a lack of understanding of the heterogeneity of the tumor microenvironment (TME) has resulted in inefficient therapy responses. A greater understanding of the TME is needed to identify patients likely to respond, and those that will have resistance to immune checkpoint inhibitors (ICIs). Advancement in spatial single-cell technologies has allowed deeper insight into the phenotypic and functional diversities of cells in the TME. In this review, we provide an overview of ICI biomarkers and highlight how high-dimensional spatially resolved, single-cell approaches provide deep molecular insights into the TME and allow for the discovery of biomarkers of clinical benefit.

Tunable PhenoCycler imaging of the murine pre-clinical tumour microenvironments

BACKGROUND

The tumour microenvironment (TME) consists of tumour-supportive immune cells, endothelial cells, and fibroblasts. PhenoCycler, a high-plex single cell spatial biology imaging platform, is used to characterize the complexity of the TME. Researchers worldwide harvest and bank tissues from mouse models which are employed to model a plethora of human disease. With the explosion of interest in spatial biology, these panoplies of archival tissues provide a valuable resource to answer new questions. Here, we describe our protocols for developing tunable PhenoCycler multiplexed imaging panels and describe our open-source data analysis pipeline. Using these protocols, we used PhenoCycler to spatially resolve the TME of 8 routinely employed pre-clinical models of lymphoma, breast cancer, and melanoma preserved as FFPE.

RESULTS

Our data reveal distinct TMEs in the different cancer models that were imaged and show that cell-cell contacts differ depending on the tumour type examined. For instance, we found that the immune infiltration in a murine model of melanoma is altered in cellular organization in melanomas that become resistant to αPD-1 therapy, with depletions in a number of cell-cell interactions.

CONCLUSIONS

This work presents a valuable resource study seamlessly adaptable to any field of research involving murine models. The methodology described allows researchers to address newly formed hypotheses using archival materials, bypassing the new to perform new mouse studies.

Quantitative image analysis of intracellular protein translocation in 3-dimensional tissues for pharmacodynamic studies of immunogenic cell death

A recent development in cancer chemotherapy is to use cytotoxics to induce tumor-specific immune response through immunogenic cell death (ICD). In ICD, calreticulin is translocated from endoplasmic reticulum to cell membrane (ecto-CRT) which serves as the 'eat-me-signal' to antigen-presenting cells. Ecto-CRT measurements, e.g., by ecto-CRT immunostaining plus flow cytometry, can be used to study the pharmacodynamics of ICD in single cells, whereas ICD studies in intact 3-dimensional tissues such as human tumors require different approaches. The present study described a method that used (a) immunostaining with fluorescent antibodies followed by confocal microscopy to obtain the spatial locations of two molecules-of-interest (CRT and a marker protein WGA), and (b) machine-learning (trainable WEKA segmentation) and additional image processing tools to locate the target molecules, remove the interfering signals in the nucleus, cytosol and extracellular space, enable the distinction of the inner and outer edges of the cell membrane and thereby identify the cells with ecto-CRT. This method, when applied to 3-dimensional human bladder cancer cell spheroids, yielded drug-induced ecto-CRT measurements that were qualitatively comparable to the flow cytometry results obtained with single cells disaggregated from spheroids. This new method was applied to study drug-induced ICD in short-term cultures of surgical specimens of human patient bladder tumors.

SCS: cell segmentation for high-resolution spatial transcriptomics

Spatial transcriptomics promises to greatly improve our understanding of tissue organization and cell-cell interactions. While most current platforms for spatial transcriptomics only offer multi-cellular resolution, with 10-15 cells per spot, recent technologies provide a much denser spot placement leading to subcellular resolution. A key challenge for these newer methods is cell segmentation and the assignment of spots to cells. Traditional image-based segmentation methods are limited and do not make full use of the information profiled by spatial transcriptomics. Here we present subcellular spatial transcriptomics cell segmentation (SCS), which combines imaging data with sequencing data to improve cell segmentation accuracy. SCS assigns spots to cells by adaptively learning the position of each spot relative to the center of its cell using a transformer neural network. SCS was tested on two new subcellular spatial transcriptomics technologies and outperformed traditional image-based segmentation methods. SCS achieved better accuracy, identified more cells and provided more realistic cell size estimation. Subcellular analysis of RNAs using SCS spot assignments provides information on RNA localization and further supports the segmentation results.

Spatial mapping of cellular senescence: emerging challenges and opportunities

Cellular senescence is a well-established driver of aging and age-related diseases. There are many challenges to mapping senescent cells in tissues such as the absence of specific markers and their relatively low abundance and vast heterogeneity. Single-cell technologies have allowed unprecedented characterization of senescence; however, many methodologies fail to provide spatial insights. The spatial component is essential, as senescent cells communicate with neighboring cells, impacting their function and the composition of extracellular space. The Cellular Senescence Network (SenNet), a National Institutes of Health (NIH) Common Fund initiative, aims to map senescent cells across the lifespan of humans and mice. Here, we provide a comprehensive review of the existing and emerging methodologies for spatial imaging and their application toward mapping senescent cells. Moreover, we discuss the limitations and challenges inherent to each technology. We argue that the development of spatially resolved methods is essential toward the goal of attaining an atlas of senescent cells.

Application of Machine Learning in Spatial Proteomics

Spatial proteomics is an interdisciplinary field that investigates the localization and dynamics of proteins, and it has gained extensive attention in recent years, especially the subcellular proteomics. Numerous evidence indicate that the subcellular localization of proteins is associated with various cellular processes and disease progression. Mass spectrometry (MS)-based and imaging-based experimental approaches have been developed to acquire large-scale spatial proteomic data. To allow the reliable analysis of increasingly complex spatial proteomics data, machine learning (ML) methods have been widely used in both MS-based and imaging-based spatial proteomic data analysis pipelines. Here, we comprehensively survey the applications of ML in spatial proteomics from following aspects: (1) data resources for spatial proteome are comprehensively introduced; (2) the roles of different ML algorithms in data analysis pipelines are elaborated; (3) successful applications of spatial proteomics and several analytical tools integrating ML methods are presented; (4) challenges existing in modern ML-based spatial proteomics studies are discussed. This review provides guidelines for researchers seeking to apply ML methods to analyze spatial proteomic data and can facilitate insightful understanding of cell biology as well as the future research in medical and drug discovery communities.

The fast image segmentation algorithms using dynamic programming for modals of image histograms

A machine learning method is now considered capable of accurately segmenting images. However, one significant disadvantage of this strategy is that it requires a lengthy training phase and an extensive training dataset. This article uses an image segmentation by histogram thresholding approach that does not require training to overcome this difficulty. This article proposes straightforward and time-optimal algorithms, which are guaranteed by mathematical proofs. Furthermore, we experiment with the proposed algorithms using 100 images from a standard database. The results show that, while their performances are not significantly different, the two proposed methods are roughly 10 and 20 times faster than the most simple and optimal method, Brute Force. They also show that the proposed algorithms can deal with bimodal images and images with various shapes of the image histogram. Because our proposed algorithms are the most efficient and effective. As a result, they can be used for real-time segmentations and as a pre-processing approach for multiple object segmentation.