3D Spatial Proteomics Provides View Of Diseases At Early Stages

February 2, 2023

By Deborah Borfitz 

February 2, 2023 | Researchers in Germany have developed a series of innovative solutions for probing the body more broadly and deeply for early pathological signs of diseases ranging from metabolic disorders and neurodegenerative diseases to cancer. Their novel technology, dubbed DISCO-MS, does spatial proteomics in three-dimensional (3D) intact specimens using a combination of whole-organ and whole-organism clearing and imaging, deep-learning-based image analysis, robotic tissue extraction, and ultrahigh-sensitivity mass spectrometry, according to Harsharan Singh Bhatia, Ph.D., spatial-omics team leader at Helmholtz Munich. 

In mouse models, DISCO-MS has been used to investigate microglia activation in mild traumatic brain injury (mTBI), revealing otherwise-missed molecules involved in the immune system’s coordinated response to disease-associated neuroinflammation, he notes. It has also characterized the amyloid-beta (Aβ) plaques in early- and late-stage Alzheimer’s, providing an unbiased and large-scale study of the culprit proteins and identifying new potential disease biomarkers. 

With the aid of a bot, DISCO-MS allowed the team to study the regional heterogeneity of immune cells in whole-mouse bodies as well as aortic plaques in an intact human heart, adds Bhatia. This identified plaque-related dysregulated molecular pathways in heart cells that could be targets for treatment. 

Specifics about these imaging feats, and the enabling technology, were all recently reported in Cell (DOI: 10.1016/j.cell.2022.11.021). In short, DISCO-MS combines methods to turn mouse and human tissues transparent and then uses the latest robotics and proteomics tools to determine their molecular makeup. 

The Technique 

DISCO-MS optically clears organs using an organic solvent, which effectively dehydrates them so the pathology of cells deep in the tissue can be observed through a light sheet microscope, explains Bhatia. Another solvent, dichloromethane, is also utilized to remove lipids. Prior to all this, a paraformaldehyde agent is used to keep all the proteins in the tissue unscathed. 

The tissue is imaged with the light sheet microscope in a refractive index matching solution designed to improve the reliability of the optical signals coming from the cells, made visible by labeling with various colored fluorescent dyes. Mass spectrometry is then used to read those proteins, typically after a couple of days, he continues. 

Two new solutions were required to extract a small number of cells (40 to 60) from a whole organ—the so-called regions of interest (ROI), such as neurons or microglia from a mouse brain, says Bhatia. One is known as “laser capture microdissection,” which uses an ultraviolet laser to precisely cut out the ROI, and the other a technical rehydration process to make the sample soft again before cryosectioning at minus 20 degrees Celsius. The sample is then cut into thin slices roughly 10 to 20 micrometers thick. 

The research team knows which cells to take because they automated registration of the two-dimensional small tissue regions back onto the compete 3D image stack of the whole organ. Mass spectrometry is then performed on the extracted cells. 

To study cancer metastasis on the whole mouse level, where stray cancer cells might be dispersed throughout the body, the robotic extraction system—named DISCO-bot—comes in handy, says Bhatia. The microenvironment of those cells is of intense scientific interest and, with that in mind, researchers attached a needle biopsy at the end of the robotic arm to remove color-coded cells in the ROI to examine their proteins and how they’re dysregulated.    

Currently, the DISCO-MS technology is limited in terms of how big a chunk of tissue can be robotically extracted and from where, says Bhatia. The method provides molecular details on no fewer than about 60 cells, which is sufficient to study plaque heterogeneity but still far from the single-cell level.   

This is partly a limitation of the needle size/depth (~400 to 4,000 cells), since 18- to 25-gauge biopsy needles are being used, he explains. It’s a practical issue related to the fact that smaller-sized needles have too much bend and are thus not precise enough for the job. With heart tissue, the problem is that it can’t be rehydrated, and the hardness makes it difficult for smaller needles to penetrate. 

New Frontiers 

For the latest study, Bhatia and his team have used DISCO-MS in a mouse model of Alzheimer’s disease, a devastating condition where misfolded Aβ proteins are suspected of generating the plaque that kills off brain cells, he says. Clinical trials of drugs targeted that toxic plaques have nonetheless repeatedly failed. 

“We hypothesized this is due to not getting the information early enough,” he continues. The exact molecular changes driving progression of the Aβ peptides at the whole-organ level is also unknown. 

Most preclinical Alzheimer’s studies involve taking a biopsy from the hippocampus, the memory region of the brain, notes Bhatia. “We forget about the other regions.” 

Using DISCO-MS, the study team detected Aβ at six weeks rather than six months, as in other mouse studies, he points out. They also used a registration and biomarker process, and deep learning analysis, to calculate the presence and size of the plaques across the whole brain to find many of the identified proteins were uniquely distributed across two distinct subregions—a potential new disease progression clue. 

The technology was similarly used to look at mTBI, which in human patients is known to cause dementia-like changes and has been associated with the development of Parkinson’s disease as well as Alzheimer’s disease later in life, says Bhatia. When the same sort of mTBI endured by individuals playing sports like football and soccer were mimicked in mice, the animals were all fine in terms of their learning, memory, and walking behavior after eight weeks. 

However, whole-brain imaging and spatial proteomics revealed subtle changes to microglia cells—the primary innate immune cells of the brain—in certain regions and new molecules related to microglia activation around the Aβ plaques were discovered. Eight proteins were uniquely detected in mTBI, including metadherin previously associated with the progression of multiple types of human malignancies.  

The DISCO-bot-aided DISCO-MS approach next aided clinical investigation on the postmortem hearts of people who died of coronary artery disease, Bhatia continues. Dozens of new molecules were found in atherosclerotic plaque regions in the coronary artery, including six that bestow rigidity. “It is important to study these molecules early enough and more unbiasedly, so we know where they are located and how they are changing the disease early [on].” 

Clinical Utility 

DISCO-MS protocols have all been published and posted on the web, and have been applied by outside groups, says Bhatia. The research team is happy to field incoming questions about use of the technique on any organ of interest. 

Bhatia foresees the technology first being used clinically by physicians interested in learning more from whole biopsies, such as 3D information on intercellular crosstalk or the volume of Aβ plaque or proliferating glial cells in the brain. They would then have a better idea which drug to administer based on the pathology and progression of the disease in question.  

As will be further demonstrated for policymaking purposes, the clinical benefits of the DISCO-MS approach outweigh safety concerns about working with the chemicals, says Bhatia. Inevitably, the cost and time of the required instrumentation will also come down. 

The utility of DISCO-MS extends to the design and development of therapeutics at the level of both proteins and messenger RNA, he adds. If preclinical trials start capturing more useful information, drugs also may start costing something less than the oft-cited $1 billion over 10 years and have better odds of succeeding in the marketplace.