Painting the brain: multiplex immunohistochemistry of the human olfactory bulb in normal ageing and disease
The human olfactory bulb has a disorganised laminar structure and is affected very early in the progression of neurodegenerative diseases. Here we present a comprehensive neuroanatomical map of the olfactory bulb using a multiplex immunohistochemistry and high-content image analysis pipeline.
Traditional neuroanatomy studies use fluorescent immunohistochemistry to label brain tissue with up to three antibodies simultaneously and thus require sequential tissue sections to investigate more proteins of interest. In our paper recently published in Communications Biology, we present a versatile, cost-effective, and efficient multiplexed fluorescence-based immunohistochemistry approach to label up to 100 antibodies on formalin-fixed paraffin-embedded human brain sections. We further describe a spatial protein analysis pipeline for unsupervised assessment of this high content labelling (Figure 1).
Painting the brain with antibodies
The efficiency, consistency, and ultimate insights of neuroanatomical studies can be substantially improved using multiplex immunofluorescence labelling on a single tissue section. When combined with whole slide imaging, multiplex labelling approaches generate spatial maps of large numbers of proteins that can facilitate discoveries in neuroanatomy and neurodegenerative diseases (Figure 2).
Multiplex immunofluorescence technologies have not yet become routine in neuroanatomy despite being popular for immunology studies of tumour cell phenotyping with spatial context. Current technologies that use readily available antibodies and standard microscopy are very accessible and relatively low cost but involve slow iterative cycles of 1-5 antibodies and require weeks to capture 10-40 markers. Multiplexing using readily available antibodies and standard microscopy such as MxIF1, OPAL2, array tomography3, CycIF4and 4i5 are very accessible and relatively low cost but involve slow iterative cycles of 1-5 antibodies and require weeks to capture 10-40 markers. Alternatively, DNA barcoding protocols such as CODEX6or SABER7are high throughput with 10+ antibodies labelled per cycle but involve considerable preparation and cost as the primary antibodies must be directly conjugated to the DNA strands.
In our paper, we present an expanded and efficient multiplexed fluorescence-based immunohistochemistry approach that maintains throughput by labelling up to 10 antibodies per cycle, with no limitation on the number of cycles, and maintains versatility and accessibility by using readily available commercial reagents and standard epifluorescence microscopy imaging8.
We demonstrate this approach by labelling 81 antibodies on sections of human olfactory bulb from neurologically normal, Alzheimer’s, and Parkinson’s disease patients. This brain region has a complex and disorganised laminar structure and is involved early in the symptomology and pathophysiology of neurodegenerative diseases. Our paper presents a comprehensive neurochemical atlas of the human olfactory bulb that identifies key structures, layers, and cell populations.
The production of such high-content anatomical information necessitates new image alignment and analysis tools. We developed an unsupervised spatial protein analysis pipeline to investigate the complex spatial arrangement of different tissue features and account for the different labelling patterns of various antibodies (cytoplasmic, nuclear, membranous). This approach draws on current spatial transcriptomics and single-cell genomics pipelines and uses the Seurat R package.
We split each image into pixel bins (31x31 pixels corresponding to a 10x10 μm image area), and for each bin, we determined the fluorescence intensity of each antibody. The measurements from each bin were combined into a ‘bin count matrix’ analogous to single-cell genomics analysis. We subsequently performed an unsupervised clustering analysis to identify bins with similar label signatures that correspond to distinct tissue layers or structures. We also investigated the differential labelling of individual antibodies across the tissue sections and between the disease groups.
In summary, we demonstrate an efficient, versatile, robust, and accessible multiplex immunohistochemistry approach on human brain tissue from neurologically normal aged cases and disease cases. By combining this approach with whole slide imaging and an unsupervised analysis pipeline, we show the potential for this technology to enhance neuroanatomical studies. We believe this technology will be helpful for a wide range of researchers and applicable to various biological questions, from broad studies of cell phenotyping or tissue atlases to specific interrogations of disease pathology.
- Gerdes, M. J. et al. Highly multiplexed single-cell analysis of formalin-fixed, paraffin-embedded cancer tissue. Proceedings of the National Academy of Sciences of the United States of America 110, 11982–7 (2013).
- Stack, E. C., Wang, C., Roman, K. A. & Hoyt, C. C. Multiplexed immunohistochemistry, imaging, and quantitation: A review, with an assessment of Tyramide signal amplification, multispectral imaging and multiplex analysis. Methods 70, 46–58 (2014).
- Micheva, K. D. & Smith, S. J. Array tomography: a new tool for imaging the molecular architecture and ultrastructure of neural circuits. Neuron 55, 25–36 (2007).
- Lin, J.-R. et al. Highly multiplexed immunofluorescence imaging of human tissues and tumors using t-CyCIF and conventional optical microscopes. eLife 7, (2018).
- Gut, G., Herrmann, M. D. & Pelkmans, L. Multiplexed protein maps link subcellular organization to cellular states. Science (New York, N.Y.) 361, (2018).
- Goltsev, Y. et al. Deep Profiling of Mouse Splenic Architecture with CODEX Multiplexed Imaging. Cell 174, 968-981.e15 (2018).
- Saka, S. K. et al. Immuno-SABER enables highly multiplexed and amplified protein imaging in tissues. Nature biotechnology 37, 1080–1090 (2019).
- Maric, D. et al. Whole-brain tissue mapping toolkit using large-scale highly multiplexed immunofluorescence imaging and deep neural networks. Nature communications 12, 1550 (2021).