��˾��ֱ��

Researchers at the University of Toronto’s Donnelly Centre for Cellular and Biomolecular Research have received a funding boost to help realize their vision of using tiny robots controlled by artificial intelligence to one day find and capture rare stem cells from brain tissue for therapy. 

Working with Mike Shaw, a machine learning expert at University College London, ��˾��ֱ��'s Aaron Wheeler and Cindi Morshead will receive more than $1 million from the new Canada-UK Artificial Intelligence Initiative.

Supported by the two countries’ federal governments, the initiative seeks to harness AI for societal benefit by bringing together experts from diverse disciplines.

“We have previously developed microrobots for manipulating individual cells in a dish,” says Wheeler, a professor in ��˾��ֱ��’s department of chemistry in the Faculty of Arts & Science and the Institute of Biomaterials and Biomedical Engineering in the Faculty of Applied Science & Engineering.

“Now we want to take it to the next level to design robots that can isolate single cells from a crowded environment such as brain tissue and make the system fully automated.”

A total of 10 international teams shared approximately $5 million and £5 million over three years, by Navdeep Bains, Canada's minister of innovation, science and industry, and British High Commissioner to Canada Susan le Jeune d'Allegeershecque. Other projects funded through the program, a collaboration between Canada's three research funding agencies and four UK research councils, seek to harness AI across different sectors, from countering abusive online language to improving labour market equality and monitoring global disease outbreaks.

“Artificial intelligence is transforming all industries and sectors, opening up more opportunities for Canadians,” Bains said in a statement. “Today we take one step further toward ensuring that AI innovation and growth builds competitive and resilient economies, and maximizes the social and health benefits in both Canada and the UK.”

Stem cells hold promise for regenerative medicine thanks to their ability to self-renew and turn into specialized cells in the body. Scientists around the world are exploring how resident stem cells in the brain can be harnessed to treat neurodegenerative diseases or repair injury.

Morshead, who is chair of anatomy in the department of surgery in the Faculty of Medicine and a stem cell scientist, and her team previously showed that brain stem cells can be directed to repair stroke injury in miceand they continue to investigate how to make the repair more efficient.

Optoelectronic microrobots designed by ��˾��ֱ�� researchers Shuailong Zhang and Aaron Wheeler can load, transport and deliver cellular material (photo courtesy of Shuailong Zhang)

The clues likely lie in the stem cells’ tissue microenvironment, where they are influenced by molecular signals released by neighbouring cells. Scientists are keen to map out this cellular cross-talk, which remains largely unexplored. A tool that can reproducibly pick out defined and intact cells from a complex mix of cells in brain tissue would be a huge asset. And tiny robots, working at the sub-millimetre scale, could be up for the task.

“Having very methodical repetitive dissections will allow us to feel confident that the behaviours of cells will be similar across samples, which is important for stem cell biology and regenerative medicine,” says Morshead.

With the help of their UK collaborators, the ��˾��ֱ�� researchers aim to teach the microrobots how to distinguish stem cells and their neighbours from microscopy images of brain tissue through AI and image-recognition algorithms.

A more immediate goal is to pair AI with the existing microrobotic platform developed by Wheeler and Morshead’s teams for manipulating individual stem cells in the dish to gain insight into their molecular makeup and behavior. They previously demonstrated how cog-wheel shaped microrobots can scoop up and move the cells about. With AI’s help, it should be possible to teach the microrobots how to recognize different types of cells based on their appearance and deliver them to various pipelines for molecular profiling.

“In the long term, we would like to have one platform that can start with a slab of tissue and go to collecting the cells of interest,” says Wheeler. “We will end up with a tool that's useful for lots of folks in the life sciences who are trying to streamline and reproducibly collect interesting cells for further analysis.”