Drug development is lengthy, expensive, and often unsuccessful. MechPath is a unique collaboration among various parties seeking innovative ways to improve drug development by determining how a patient can best be represented by a cell model. This is crucial for the use of animal-free methods, as without proper validation, many animal-free innovations remain stuck in the experimental phase.

TNO, Galapagos, Leiden University (LACDR), and the Dutch Society for the Replacement of Animal Testing joined forces to develop a method that better predicts how new medicines will work in different patients. They achieved this by leveraging data, building AI infrastructure, and applying knowledge of cell models. Along the way, they made a surprising discovery: cells remember whether they are sick.

Why MechPath is so important: smarter tests with AI

A recent report by Innovation Quarter concludes that Life Sciences & Health is one of the fields where AI can really make a difference. This is exactly what MechPath focuses on. Currently, many drugs that seem promising in a laboratory do not work well enough in practice for patients. For example, the majority of drugs that test positively on animals do not have the desired effect on humans. And even drugs that do work in humans are not perfect: they often only work for a limited number of patients. This is partly because test methods do not take into account the differences between patients.

MechPath wants to address this problem by analyzing all available patient data with artificial intelligence and translating it into disease mechanisms, and using these to distinguish different groups of patients. By doing this, it will ultimately be possible to better predict how new drugs will work in different groups of patients.

The MechPath approach: bringing together and analyzing data

The MechPath team has set to work using a special approach they call MECHanistic PATHology understanding. In this approach, various technologies and knowledge sources come together:

  1. Analyzing large amounts of patient data with advanced computational methods. 
  2. Developing better laboratory models that more accurately mimic the situation in the human body. 
  3. Using artificial intelligence to discover patterns in data and make predictions.

An important goal of the project is to improve so-called ‘in vitro models’, cell models that use human cells or tissues. MechPath has investigated how these models can be improved to mimic the differences between patients.


What makes MechPath unique: validation as the key to success

What’s innovative about MechPath is that tools are being developed within this project to validate in vitro models. In other words: to determine whether a model reliably measures what you want to know, namely the biological process that is disturbed in the patient. This is extremely important, because without proper validation, many animal-free innovations remain stuck in the experimental phase. They are then not widely accepted and used. And that’s a shame: because an animal-free model gathering dust somewhere on a shelf doesn’t replace any animal testing.

By working on validation tools, MechPath aims to ensure that new animal-free methods are approved and used more quickly in research for new drugs. This is not only good news for laboratory animals, but also for patients who can benefit from new treatments sooner.

What MechPath has already achieved: cells with memory

The researchers have determined that human cells used in the lab are much more variable than we originally thought. This insight can help to better mimic the differences between patients in cell models. Surprising was the discovery that cells from patients react differently than cells from healthy people, even after prolonged culture in the lab. This points to a kind of ‘disease memory’ in the cells: cells remember that they are sick, even when they have long been disconnected from the donor’s body.

Initially, the researchers mainly focused on fibrosis, the formation of scar tissue in organs. Because there is a lot of variation in the patient population with fibrosis, this condition lent itself well to this research. Among other things, the researchers saw that oxygen deficiency plays an important role in the development of fibrosis. By adding this to cell models, the situation in patients can be better mimicked.

Future perspective: developing drugs faster, cheaper, and animal-free

The knowledge and methods that have come out of the MechPath project will be applied in research on muscle weakening at TNO in the short term. In addition, the researchers are publishing their findings in three articles, so that other researchers working on more effective and animal-friendly ways of drug development can also build on their discoveries.

The fact that a pharmaceutical company is also involved in the collaboration is extremely important. By developing models in collaboration with the companies that will use them, you can be sure that they are tailored to the needs these companies have in practice and that they will actually be used. In this way, collaborations like MechPath can make an important contribution to the way we develop drugs: faster, cheaper, animal-free, and better tailored to specific patients.

Want to read more about this unique collaboration? We spoke with biochemist Roeland Hanemaaijer and data scientist Lars Verschuren, both from TNO and part of this collaboration.