AI-Powered Whole-Body Maps Expose Obesity’s Systemic Toll

Globally, obesity affects more than one billion people, which increases their risk for cardiometabolic diseases, stroke, some cancers, and neuropathies.¹ This underscores the systemic effects of obesity, which leaves few organs untouched.

“It's not happening in an isolated tissue in isolated organ,” said Doris Kaltenecker, a postdoctoral researcher in neuroscientist Ali Ertürk’s lab at Helmholtz Munich, where his group develops technologies to map organs at cellular level. “We were looking for approaches and also analysis pipelines that are able to actually capture the impact of a systemic disease on the whole-body level but also at very fine resolution at single cell resolution,” explained Kaltenecker.

Now, by combining advanced fluorescence microscopy with AI-based tools, Kaltenecker, Ertürk, and their team mapped obesity-associated changes throughout the mouse body at cellular resolution.² Their pipeline, published in Nature, offers an approach for viewing and measuring the systemic effects of diseases.

Obtaining a Window into the Obese Body

For their experiments, Kaltenecker and her colleagues used mice that had fluorescently labeled nerves and immune cells. They exposed these animals to a high fat diet to induce obesity or kept them on a normal diet as a control group. Before they could observe the effects of this condition, the researchers had to overcome a problem preventing them from visualizing tissues in whole animals.

“The tissue is opaque because within the tissue, you have different components like proteins, water, and lipids, and all of those have a different refractive index,” said Kaltenecker. To remove these obscuring elements, the team carried out tissue clearing to turn the animals transparent while keeping the fluorescent labeling intact. This involved treating the tissues with certain chemicals to render the different biomolecules with a similar refractive index, turning them transparent to certain wavelengths, Kaltenecker explained.

According to Kaltenecker, just this part of the protocol took them seven days. After this, they used light-sheet fluorescence microscopy that uses a plane of laser light to illuminate the samples. They scanned each mouse dorsally and ventrally—or from the top and the bottom—and merged the scans to obtain a picture of the whole body.

But this process was not simple to do in whole, obese animals, with scanning times as long as seven hours per mouse. However, this wait was still an upgrade because, “When I started originally [in 2021] and scanned my first mouse, it took seven days,” said Kaltenecker.

But the time and arduous labor paid off when the light-sheet microscopy approach resulted in images of intact mice with tissues glowing thanks to the fluorescent labels. Kaltenecker recalled seeing the first images show up on the screen with her colleague, “We were really sitting there with our mouths open.”

One of the things that stood out to Kaltenecker was the transparency that the team achieved for the light-sheet microscopy. “We were really just amazed that it worked with these very obese mice,” she said.

Mapping Obesity-Driven Systemic Damage

Based on this work, the researchers developed MouseMapper, a deep-learning framework for three-dimensional analysis of whole-body images of the nervous and immune systems. Kaltenecker and her colleagues used this model to analyze their microscopy data and compared this in lean and obese mice. This helped them determine where obesity-induced inflammation and damage occurred across fat, muscle, liver, and peripheral nervous tissues.

MouseMapper revealed obesity-induced structural changes to facial nerves in mice. Consistent with this finding, behavioral experiments indicated that obese animals had diminished responses to facial stimulation compared to lean mice. Visualizing the peripheral nervous system in this manner was “striking,” said Kaltenecker. “This was always something for me, personally, where I always had to smile when I saw the data.”

The researchers also applied MouseMapper to characterize the immune cell compositions in tissues affected by obesity-induced inflammation. They observed increased immune cell infiltration in adipose tissues and the muscle, stomach and abdominal wall of obese mice, signifying inflammatory activity.

Kaltenecker said that it remains to be seen whether these obesity-driven inflammatory and structural changes are reversible. The team plans to follow up on this question by investigating whether putting the mice back on a normal diet would undo the changes that a high-fat diet induced.

According to Kaltenecker, there is a power in visualizing entire animals. “[The tool] has a big advantage if you're able to really see the whole system and you're not bound to a specific organ,” she said. “With this kind of approach, you're actually able to identify new parts which you didn't consider before.”