• Research

    Minnesota Microscopy

Ever since Antonie van Leeuwenhoek spotted "animalcules" with his first crude microscope lens, microscopy has been a mainstay of the science of small. In the 19th century, optical microscopes were instrumental in early microbial research and became a coveted medical tool. The founder of Mayo Clinic, William Worrall Mayo, M.D., deemed a good microscope so essential to his practice that he and his wife mortgaged their house to buy one.

While accuracy has increased exponentially with the advent of electron microscopes, so has the cost. A single piece of equipment can reach well into seven figures. Yet, because the answers to medical questions increasingly lie amid the smallest bits of biology known, microscopy remains the essential tool to understand the invisible.

Today at Mayo Clinic, Jeffrey Salisbury, Ph.D., is the person to talk to about microscopes. He has used them extensively in cancer tumor biochemistry research, and he is the head of Mayo Clinic's Microscopy and Cell Analysis Core Facility.

Dr. Salisbury is the head of Mayo Clinic's Microscopy and Cell Analysis Core Facility. In 2012, he authored an article in Discovery's Edge called, "Scientific Vision: Microscopes at Mayo Clinic." Read it in an archived PDF, below.

"Our goal is to ensure researchers have access to the highest-quality technology for their biomedical research," he says. "It's the same mission as in Dr. Mayo's day: meeting the needs of our patients."

To that end, Dr. Salisbury and colleagues have been helped by the Minnesota Partnership for Biotechnology and Medical Genomics. In addition to awarding annual research grants supported by the state, the partnership also awards infrastructure grants. Key to the success of these efforts, has been a long-standing and productive scientific relationship between Dr. Salisbury and Mark Sanders, Ph.D., director of the University of Minnesota Imaging Centers. In 2010, together with Timothy Ebner, M.D., Ph.D., of the University of Minnesota, they received a grant that paid for two state-of-the-science microscopes. One, a 2-photon optical microscope housed at the university, allows researchers to use high spatial resolution to study living cells and organs within an animal. The other, at Mayo, is a "superresolution" microscope capable of visualizing at twice the resolution of standard devices.

The power of 2-photon microscopy allows deep tissue penetration in living animals or tissue, as illustrated in this 3D architecture of pyramidal cells of a mouse's brain cortex.

These tools are highly valuable in neurodegenerative disease, diabetes and cancer preclinical research. Despite where they reside, the partnership requires that equipment is for the use of staff at Mayo and the university, thus aiding the two major medical research centers in Minnesota and helping defray the rising expense of these tools.

In 2016, Drs. Salisbury and Sanders, and Claudia Neuhauser, Ph.D., of the University of Minnesota (now at the University of Houston), received a Minnesota Partnership grant for a "3D nanoscale-resolution microscopy for understanding human disease processes." That translates as a scanning electron microscope, along with a "cyberinfrastructure" ― the computer capability to manage, store and analyze the large data files it produces.

Installed in late 2017, this machine has been invaluable in capturing images that fall into the "mesoscale" ― that area between the ultrathin sectors of material and the tissues of some structures. In other words, it can explore at the cellular, molecular and tissue levels, such as interactions between membranes of organelles; activity of neurons and synapses; and the architecture of tissues. Currently, over 60 research projects are underway, including many 3D visualizations. Three years later, the march to evolve this technology continued, with a partnership-funded addition of artificial intelligence (AI). The computational advancement expanded and sped up analysis of disease imaging.

Electron microscopy can give 1,000 times higher resolution, compared to optical microscopy, and through Minnesota Partnership infrastructure support, investigators can study large-volume 3D architecture of cells and tissues, as seen in this example of kidney tubules.

Drs. Salisbury and Sanders joined with Jeffrey McDonald, Ph.D., of The Hormel Institute, and Thomas Pengo, Ph.D. and Lucy Fortsom of the University of Minnesota to use this digital technology to analyze 2D and 3D images of cells and tissues. Until recently, this has been uncharted territory in the invisible realm. How do diseases interact with the structure of the biology, and how do they connect or enter the system? More so, how can they be blocked from doing so? Once developed, the AI analytical tools have to be oriented or trained. Just as a robot is programmed to respond to situations or commands, these algorithms need to learn the ropes.

But how do you train algorithms? With a lot of help, it turns out.

This Minnesota Microscopy Infrastructure Partnership reached out to the volunteer scientific workforce harnessed by Zooniverse and The Francis Crick Institute Ltd. for a hand. These platforms connect projects in need of help with people who volunteer their time and take advantage of the so-called wisdom of crowds. The team launched a Zooniverse platform early in 2021 to get citizen scientists from around the world to view and manually highlight specific cell components to help train the AI algorithms that will mimic human intervention in microscopic image analysis.

This pair of images shows lipid droplets, mitochondria and cellular membranes in a single electron microscope image, as shown at left, and a 3D reconstruction from 20 images, as shown at right. What becomes clear in the 3D reconstruction are the spatial relationships among these cellular components that are difficult to appreciate in a single section. These relationships are important for understanding many metabolic processes.

The response has been remarkable, according to Dr. Salisbury. During the first week alone, the platform received image annotations from over 20,000 citizen scientists. This is crowdsourcing taken to a high-tech level.

"This really paves the way for meeting the goals of our program," says Dr. Salisbury. "These folks are looking for, and circling, lipids they see in the form of fat droplets within cells. The more that they identify, the greater the basis for judgment the algorithms will have for their future analysis. Their work helps establish the relationship between the lipids, their size and the other structures in the cell."

But why lipids? These fats and oils are essential to the work of cells, cell membranes and cell signaling. They store and release energy, when needed. If that function is disrupted or lipid levels go out of balance, it can result in diabetes or heart disease. Correspondingly, muscles can weaken and nerves can degenerate, sometimes leading to death. Thus, the ability to accurately analyze what's happening in a disease process for a person could eventually save lives and perhaps reverse the decline.

Given the prevalence of diabetes mellitus and related heart disease in the nation, the work of this small cadre of microscopy experts in Minnesota and the large auxiliary of international volunteers is critically important. Collaboration, innovation and the latest technology, combined with state support, create the key formula to help save lives.

― Bob Nellis, April 2021