Combining machine learning and cell engineering for cancer therapies

Combining machine learning and cell engineering for cancer therapies
© iStock/KENGKAT

Scientists are tackling the ‘Holy Grail’ of oncology by combing machine learning and cell engineering to create ‘living medicines’ that precisely target cancer tumours.

Scientists at UC San Francisco (UCSF) and Princeton University have presented new strategies to address the ‘Holy Grail’ of oncology – finding medicines that can kill cancer without killing healthy cells. The teams have proposed using smart cell therapies – or ‘living’ medicines that are activated by combinations of proteins that are only ever present together in cancerous cells. This is achieved by using cutting-edge therapeutic cell engineering together with advanced machine learning methods.

Harnessing computational methods

Members of Wendell Lim’s lab at UCSF, who has been exploring this area of research for a number of years, joined together with computer scientists at Princeton’s Lewis-Sigler Institute for Integrative Genomics and the Simons Foundation’s Flatiron Institute to use machine learning to analyse massive databases of thousands of proteins found in both cancer and normal cells.

The paper, published in Cell Systems, collates millions of possible protein combinations into a catalogue that could help with the precise targeting of cancer cells only. A further paper published by the team shows how this data can be used to design sensitive cell therapies which could precisely treat cancers.

Lim, professor and chair of cellular and molecular pharmacology and a member of the UCSF Helen Diller Family Comprehensive Cancer Center, said: “Currently, most cancer treatments, including cell therapies, are told ‘block this,’ or ‘kill this’. We want to increase the nuance and sophistication of the decisions that a therapeutic cell makes.”

Programming CAR T cells

Chimeric antigen receptor (CAR) T cell therapy is a very efficient way to treat certain types of cancers such as blood cancers, however, it is not as beneficial for treating solid, tumour type cancers as it risks damaging healthy tissues along with the cancer cells. Lim says that a more “complex product” is needed for targeting solid tumours.

To do this, the Lim lab researchers harnessed the power of machine learning techniques to explore public databases to examine the gene expression profile of more than 2,300 genes in normal and tumour cells to see what antigens could help discriminate one from the other.

Based on this gene expression analysis, Lim, graduate student Troyanskaya, and colleagues applied Boolean logic to antigen combinations to determine if they could significantly improve how T cells recognise tumours while ignoring normal tissue. For example, using the Booleans AND, OR, or NOT, tumour cells might be differentiated from normal tissue using markers “A” OR “B,” but NOT “C,” where “C” is an antigen found only in normal tissue.

To improve how T cells recognise tumours, the team used Boolean logic, then programmed these instructions into T cells using a system known as ‘synNotch’ – a customisable molecular sensor developed by the Lim Lab that allows for the fine-tuning of cell programming.. The experiment was successful in killing kidney cancer cells.

Troyanskaya said: “The field of big data analysis of cancer and the field of cell engineering have both exploded in the last few years, but these advances have not been brought together. The computing capabilities of therapeutic cells combined with machine learning approaches enable actionable use of the increasingly available rich genomic and proteomic data on cancers.”

“This work is essentially a cell engineering manual that provides us with blueprints for how to build different classes of therapeutic T cells that could recognise almost any possible type of combinatorial antigen pattern that could exist on a cancer cell,” said Lim.

“You’re not just looking for one magic-bullet target. You’re trying to use all the data. We need to comb through all of the available cancer data to find unambiguous combinatorial signatures of cancer. If we can do this, then it could launch the use of these smarter cells that really harness the computational sophistication of biology and have a real impact on fighting cancer.”

The research team is now exploring how these circuits could be used in CAR T cells to treat glioblastoma.

Subscribe to our newsletter


Please enter your comment!
Please enter your name here