Computer algorithm found a new treatment for childhood cancer
Computer algorithms are becoming increasingly important in research on new cancer drugs and can give unexpected angles.
Computer algorithms that weigh together large amounts of genetic and pharmacological information are increasingly useful tools in drug development. The researchers at Uppsala have found a new promising treatment for childhood cancer neuroblastoma. This is done with the help of a smart computer algorithm. The Uppsala researchers did the study together with colleagues at the Karolinska Institute, Lund University, and the Chalmers University of Technology. The study has been published in Nature communications.
– We were very surprised when the computer proposed completely new treatment ideas that no one has discussed in these contexts, says Sven Nelander, associate professor at the Department of Immunology, Genetics, and Pathology at Uppsala University and responsible for the study.
Computer algorithm weighs data together
Neuroblastoma is a rare form of cancer that affects young children. In Sweden, 15-20 children a year get sick. Most are for two years.
Neuroblastoma originates from cells in nerve tissue in the sympathetic nervous system, and the tumors can, therefore, occur at different points in the body. The most common place is one of the adrenal glands. Neuroblastoma can be anything from a mild illness that grows away by itself to a life-threatening and difficult-to-treat disease. In about half of the severe cases, effective treatment is lacking.
In a research project supported by the Child Cancer Foundation, the researchers have developed a computer algorithm that they call Target translator. They use it instead of traditional methods for drug development and thus quickly find new possible cancer drugs.
In the current project, they used the algorithm to search drug candidates for the most severe forms of neuroblastoma. They entered large amounts of neuroblastoma data from, among other things, tumor biobanks and drug databases at US and European hospitals and universities.
The algorithm weighed all the data together and found over 80 possible molecular targets for new tailor-made treatment and proposed new treatments aimed at these.
The researchers selected about ten possible treatments that they tested using tumor tissue samples from patients and animal models (zebrafish and mice). Many were found to have measurable effects, and one of the tested substances produced particularly exciting results. The substance stimulates a receptor protein in the nervous system called CNR2. This treatment impaired tumor cell survival and decreased tumor growth in experiments with zebrafish as models for human neuroblastoma.
The researchers have received continued funding from the Children’s Cancer Foundation and will proceed with both the neuroblastoma treatment and the development of the Target Translator. The purpose is to be able to use it to find treatments for other cancers as well.
– Smart algorithms will become increasingly important in cancer research in the coming years because they can help us researchers find unexpected angles. We have already started a more substantial project here in Uppsala, where several types of cancer in children and adults will be investigated in this way. We hope this can lead to more unexpected treatment options, says Sven Nelander.