An algorithm based on "first impressions" of immune cells can help predict the onset of disease

First impressions are important – they can set the stage for the entire course of a relationship. The same goes for the impressions that cells form of our immune system when they encounter a new bacterium. Using this information, researchers at the Weizmann Institute of Science have developed an algorithm to predict the onset of diseases such as tuberculosis. The results of this research were published in Nature Communications.

Dr. King Avraham, whose research was conducted by the Institute's Biological Regulatory Department group, explains:

When immune cells and bacteria meet, several consequences can occur. The immune system can kill bacteria. bacteria can defeat immune defenses; or, in the case of diseases such as tuberculosis, the bacteria can remain dormant for years, sometimes causing disease at a later stage and sometimes staying in hibernation for good. We believe that the junction in which one of these paths is chosen takes place early, some 24 to 48 hours after infection. "

The scientists first tested real encounters between immune cells and bacteria – this time between blood samples (containing immune cells) and Salmonella bacteria.

Led by Drs. Noa Bossel Ben Moshe and Shelly Hen-Avivi of Dr. Avraham's group, the team used a method developed in recent years, including at the Weizmann Institute, to sequence gene activity in thousands of cells. In other words, they could see what each cell looked like when they responded to Salmonella and set up their activation profile. This revealed unobserved trends in standard laboratory tests and seemed to confirm their hypothesis: there were indeed differences that allowed them to trace the responses of initial meetings to subsequent results.

This single-cell sequencing, however, remains limited to specialized laboratories. The group therefore asked if there was a way to link their results to real-time blood tests on real patients. They turned to their unicellular databases on Salmonella infection and immune responses to develop an algorithm based on a method called deconvolution, which would then allow them to extract similar information from gambling games. standard data. This algorithm uses the information available from standard blood tests and extrapolates the properties of individual blood cells in the experiments. "The algorithm that we developed," says Bossel Ben Moshe, "allows not only to define all of the immune cells participating in the response, but also to reveal their level of activity and, hence, , their potential strength. "

The first test of the algorithm was performed on blood samples taken from healthy people in the Netherlands. These samples were infected, in a laboratory box, with Salmonella bacteria, and the immune response was recorded. Comparisons with existing methods of genomic badysis have shown that standard methods do not reveal differences between groups, whereas the algorithm developed by the group revealed significant differences related to later variations in destruction of bacteria.

The scientists then wondered if the same algorithm could be used to diagnose the onset of tuberculosis, caused by a bacterium that often chooses the third way – dormancy – and can therefore hide in the body for years. Up to one-third of the world's population carries TB bacteria, although only a small percentage of these actually falls ill. Yet, about two million people die each year from the disease, mainly in underdeveloped areas of China, Russia, and Africa. The researchers turned to another database – a British base that tracked patients and carriers for two years – so that scientists could apply the algorithm to blood test results from both groups, as well as from the – past set of carrier to disease. period.

The researchers found that levels of monocyte activity – a type of immune cell – could be used to predict the onset or progression of the disease. "The algorithm is based on the" first impressions "of immune cells and Salmonella, which cause a type of disease very different from that of Mycobacterium tuberculosis," Hen-Avivi said. "Nevertheless, we were able to predict very early which carriers would develop the active form of the disease."

Once the symptoms of TB have appeared, patients must take three different antibiotics within nine months, and antibiotic resistance has become endemic in these bacteria. "If people at risk of contracting an active disease could be identified when the bacterial burden was smaller, their chances of recovery would be better," says Dr. Avraham. "And state medical systems in countries where TB is endemic could offer a better way to limit suffering and the incidence of the disease while reducing treatment costs."

The researchers intend to continue this path of research by developing their own database of tuberculosis and other pathogens in order to refine the algorithm. They also plan to work on developing tools that could, in the future, be used to predict who will develop a disease in its own right. With the advanced algorithm, other lines of research could lead to methods for predicting the evolution of a number of infectious diseases.