Hyperplexed immunofluorescence (HIFI) spatial biology of a mouse model of breast-to-brain metastasis

User Story

Brain Tumor Microenvironment Spatial Biology with High-Throughput Hyperplexed Imaging

byZEISS Microscopy 26 June 2024 3 min read

User Story

Brain Tumor Microenvironment Spatial Changes Uncovered by High-Throughput Hyperplexed Immunofluorescence Imaging

A New Hyper-Multiplex Workflow using ZEISS Axioscan Enables High-Throughput Imaging of Fragile Tissues with 40-50 Markers over Large Regions and High Magnifications

The tumor microenvironment is comprised of cellular and non-cellular components that can engage with cancerous cells and impact tumor behavior. Professor Johanna A. Joyce leads a lab at the University of Lausanne, Switzerland, with the goal of understanding the complex cellular interactions and spatial biology of the tumor microenvironment, including in primary and metastatic brain tumors. Her team is working to find new ways to exploit the immune and stromal cell populations to develop new treatments to improve the lives of patients.

Dr. Spencer S. Watson, a Research Fellow in Professor Joyce’s lab, published an article investigating how the brain tumor microenvironment responds to radiotherapy, not just in terms of changes in different cell populations, but also regarding how the cells spatially organize themselves. In order to do this, the team worked to create a new multiplex workflow, called Hyperplexed Immunofluorescence Imaging (HIFI), to study the spatial biology of tumors before and after treatment. Their workflow, which utilizes the ZEISS Axioscan digital slide scanner, allowed them to achieve 40-50 multiplexed markers with extremely fragile tissue over large regions and high magnifications, and revealed interesting differences in spatial reorganizations post-radiotherapy treatment.

We believe that identifying consistent changes in the spatial organization of cellular superstructures after treatment can give us clues as to how tumors resist therapies. However, many cell types and phenotypes must be simultaneously assessed over many samples. Our approach for whole slide, high-dimensional imaging scales to high-throughput to derive critical, single-cell data from these images.

Dr. Spencer S. Watson (right) with Dr. Benoit Duc
Dr. Watson is a Research Fellow and first author, Dr. Duc is a MD, PhD Student, and second author, Prof. Johanna A. Joyce’s Lab, Ludwig Institute for Cancer Research, University of Lausanne, Switzerland

Hyperplexed immunofluorescence (HIFI) spatial biology of a mouse model of breast-to-brain metastasis, animated to show both the fluorescent image and the digital pathology single cell annotation image. Tumor cells are shown in green, endothelial cells in yellow, immune populations in red, basement membranes in magenta, and all other nuclei in blue. Imaged with ZEISS Axioscan.

Hyperplexed Imaging (HIFI)

High-Throughput, Multiplex Spatial Imaging

HIFI is based on cyclic immunofluorescence to achieve 40-50 multiplexed markers, but extensively modified and optimized to work with extremely fragile tissue types over large region sizes and high magnification. ZEISS Axioscan scales the approach to high throughput with its capability to image up to 100 slides in a single experiment. Combining the ZEISS Colibri 7 light source for low-light exposure with a gentle marker elution, the team was able to repeatedly stain and image fragile samples in a non-destructive manner. Deep-learning models automatically detect individual cells and extracellular structures in the images, and orthogonal spatial analytical approaches find consistent cellular organization features following treatment, and at the point of recurrence.

Spatial Organization of Brain Tumor Microenvironments

Imaged with ZEISS Axioscan

Hyperplexed multiplex spatial biology image of a mouse brain section bearing a high-grade glioblastoma consisting of 45 markers acquired with ZEISS Axioscan high-throughput digital scanner. The image depicts neuronal cells in magenta, astrocytes in cyan, tumor cells in green, collagen in red, and all other nuclei in blue.

Hyperplexed multiplex spatial biology image of a mouse brain section bearing a high-grade glioblastoma consisting of 45 markers acquired with ZEISS Axioscan high-throughput digital scanner. The image depicts neuronal cells in magenta, astrocytes in cyan, tumor cells in green, collagen in red, and all other nuclei in blue.
Download image (921 KB)
Hyperplexed multiplex (HIFI) image exploring the spatial biology of a mouse brain glioblastoma focused on the dynamic interface between the tumor and the surrounding normal brain. The image depicts tumor cells in green, vasculature in yellow, astrocytes in magenta, axons in cyan, neuronal cells in red, and all other nuclei in blue. Imaged with ZEISS Axioscan.

Hyperplexed multiplex (HIFI) image exploring the spatial biology of a mouse brain glioblastoma focused on the dynamic interface between the tumor and the surrounding normal brain. The image depicts tumor cells in green, vasculature in yellow, astrocytes in magenta, axons in cyan, neuronal cells in red, and all other nuclei in blue. Imaged with ZEISS Axioscan.
Download image (1 MB)

An XY projection plot of single cells derived from HIFI data, animated to move between their spatial localization in the image and their location in a dimensional reduction UMAP plot. UMAP plots cells based on how similar their expression of every marker is, clustering different cell types together in feature-space. Moving between 2-dimensional plots and high-dimensional clustering shows the spatial context of cellular heterogeneity, allowing for easier interpretation of unique cell clusters. This plot depicts tumor cells in green, fibroblast-like cells in red, neurons in magenta, astrocytes in cyan, endothelial cells in yellow, and other cell types in white.

Reorganization of Brain Tumors Explored by HIFI Imaging

Using the HIFI workflow, Dr. Watson and his colleagues analyzed how the spatial biology of two different brain tumor models, a primary glioblastoma model and a brain metastasis model, responded to radiotherapy. Interestingly, these models showed completely different reactions to treatment.

The brain metastasis model did not substantially reorganize the microenvironment after treatment, rather, it shut down cancer cell proliferation.

Conversely, the glioma model dramatically regressed in size and the cellular organization was completely restructured. This cellular reorganization created a spatial superstructure of hypertrophic tumor cells surrounded by fibrosis and tumor-associated immune cells. This fibrotic niche is very similar to the Joyce lab's other research showing a highly localized, pro-survival environment that promotes treatment resistance in tumor cells.

Read the full article.

We will delve deeper into this post-treatment fibrotic niche that occurs in glioblastoma following multiple treatment interventions. We find that inhibiting this response improves response to therapy. We will combine our HIFI approach using ZEISS Axioscan with spatial transcriptomics, creating a rich data source for spatial multi-omics analysis.

Dr. Spencer S. Watson
Research Fellow, Prof. Johanna A. Joyce’s Lab, Ludwig Institute for Cancer Research, University of Lausanne, Switzerland