Dear colleagues, my name is Pablo Valdez and I'm currently a PGY-6 neurosurgery resident at Brigham and Women's Hospital in Boston and a research scientist at MIT. Today I will describe a new technology called protein-decoding expansion microscopy or DEXM, in which we chemically modify tissues, allowing us to separate proteins apart, making room for antibodies to stain previously inaccessible protein epitopes. We show its potential for discovery of new and previously unidentified structures and cell populations in both low- and high-grade gliomas. I have no disclosures for this work. Immunohistochemistry or immunostaining has facilitated discoveries on protein localization in essentially all fields of biology and has been integral as a routine diagnostic process in diverse clinical settings. The physical basis of immunostaining is the binding event between the antibody molecule and the target molecule, which we call the epitope. As such, immunostaining has a fundamental limitation, what we call the protein crowding problem of immunostaining, where proteins that are tightly packed in space cannot be accessed by antibodies, thus limiting the protein epitopes that can be visualized. This is particularly noticeable at nanoscale super-resolution scales of less than 250 nanometers, where we see gaps in staining and the inability of conventional antibodies to reach their targets. Seen here are microtubules, endosomes, and membrane receptors where we use conventional antibodies and then smaller nanobodies or aptamer probes, which can partially address this problem because they are much, much smaller. Recently, a group developed a set of technologies called EXM or expansion microscopy, where we perform conventional immunostaining on tissues, followed by chemical modification with preservation of fluorescent tags and subsequent expansion of tissues, turning smaller structures into much, much larger structures, enabling nanoscale imaging using conventional diffraction-limited microscopes like confocal microscopes. Briefly, we treat tissue with a chemical anchor which attaches to proteins. Then we infuse tissue with a monomer solution and catalyze a polymerization reaction of these monomers to create a polymer gel embedded throughout the tissue, which in turn is then connected to the anchor, which is connected to the proteins. Finally, we treat tissue with a harsh proteinase digestion that will break apart tissue and proteins, meanwhile retaining the fluorescent tags, so we can subsequently expand tissue the same in every direction, so we retain the structural relationship between tags. This capability raises the question, can we preserve proteins, which we currently cannot with EXM, and if we can preserve proteins, can we immunostain post-expansion? And if we can introduce antibodies post-expansion, that is when proteins have been decrowded or spaced apart in a post-decrowding manner, can we potentially gain access to previously inaccessible epitopes, improving our immunostaining? To this end, we present a novel epitope preservation tissue expansion protocol, decrowding EXM, in which clinical samples undergo chemical modification to anchor proteins onto a hydrogel network, shown in B, followed by protein preservation tissue denaturation treatment, shown in C, which enables tissue expansion and decrowding of proteins, shown in D, and thus immunostaining post-decrowding for visualization of previously inaccessible epitopes, shown in E, with sequential rounds of immunostaining for multiplexed imaging, if needed, shown in F. To validate our technology, we tested clinical samples of human hippocampus, top row, and glioblastoma, bottom row, and found that before-treatment images on the left and post-treatment images on the right retained their three-dimensional integrity without any significant distortion that would be caused by inadequate chemical treatments, with errors of less than 5%, shown in our quantitative analysis on the right side. We then asked, does post-decrowding immunostaining, as enabled by DEXM, indeed allow antibody molecules to detect previously inaccessible protein epitopes and thus improve the quality of immunohistochemistry? We perform immunostaining of normal human hippocampus, top row, and glioblastoma, middle and bottom rows. Antibodies are first introduced prior to any tissue expansion using conventional tissue immunostaining, and are shown here with pre-expansion images on the left. Next, the same tissues who have only had pre-expansion conventional immunostaining undergo expansion to about four times the size to produce the pre-decrowding images shown in the middle. Finally, the same expanded tissues in which proteins are now spread apart and decrowded now undergo post-expansion or post-decrowding staining to produce our post-decrowding images shown on the right. We demonstrate a qualitative improved visualization of neuronal structures with the dendritic marker MAP2 in cyan, and of astrocyte processes with GFAP in magenta, shown in panel C. We see additional vimentin-positive cell populations in tumor tissue with vimentin, shown in cyan, shown in panel H, and also a dense network of cellular processes extending between tumor cells and also at the perivascular space stained for GFAP in cyan, shown in panel M. Finally, we also observed a significant quantitative difference in fluorescent staining comparing our pre- and post-decrowding images shown by the graphs on the right side, which confirm our qualitative assessment across all the markers discussed. We devised a DXM antibody stripping protocol that removes antibodies introduced in previous routes of immunostaining with no residual signal post-stripping, as shown in the images and graphs in the top row. We also demonstrate that we retain protein antigenicity with no change in signal intensity after multiple rounds of immunostaining, as shown in the images and graph in the middle row. Finally, we demonstrate that we can retain structural integrity of the sample without significant distortion after multiple rounds of immunostaining to less than 5% distortion, as shown in the images and graph in the bottom row. This data demonstrates that DXM can achieve sequential rounds of immunostaining, which would permit detection of more antigen targets than spectrally permitted within the same tissue sample. We next performed multiple rounds of immunostaining on the same clinical specimen and were able to demonstrate up to 12 different markers localized within the same specimen, including neurofilament markers, organelle markers, synaptic markers, and various cell populations within the brain, such as oligodendrocytes, astrocytes, and neurons. Similarly, we performed multiplexed DEXM immunostaining in human GBM to visualize the perivascular network beyond the conventional four markers. This demonstrates the ability for highly multiplexed post-decroding staining imaging of clinical samples. Finally, we discovered a radical difference in kind in post-decroding immunostaining in human low-grade gliomas compared to pre-expansion, pre-decroding staining. First, we observed new tubular structures, denoted by the white arrows, that were not previously seen between distinct cells, as well as a continuous, rather than puncted pattern of staining, when staining for intermediate filaments by mentin, a mesenchymal marker, and GFAP, a marker of astrocytic lineage cells. Second, we observed a significantly greater proportion of cells showing a two- to four-fold greater colocalization of proteins than seen with conventional immunostaining. DEXM-mediated post-decroding staining of GFAP and vimentin showed colocalized filaments and cells demonstrating a network of cellular projections which were not observed under standard immunostaining in a greater proportion of cells with a mesenchymal phenotype. Post-decroding staining of IBA1 and vimentin colocalization also demonstrated a larger proportion of activated microglia in our sample than previously seen. And finally, post-decroding staining of IBA1 and GFAP colocalization also demonstrated a larger proportion of astrocytic lineage tumor cells that have acquired a phagocytic phenotype. In summary, our colocalization studies revealed a quantitatively significant difference in colocalization of proteins showing novel structures and cell populations that could point to a more aggressive phenotype in the samples studied. In summary, we present a novel technology, protein-decroding EXM, to address a fundamental limitation of immunostaining, that is, the protein-crowding problem. We demonstrate its potential in human clinical samples, showing a radical difference in kind in post-decroding immunostaining of human low- and high-grade gliomas compared to pre-expansion and pre-decroding staining. Thank you very much for this opportunity to share this work.

Pablo Valdez

protein-decoding expansion microscopy

DEXM

nanoscale imaging

decrowding EXM