More information
Description
This project will use a laboratory culture of cancer cells to investigate how small differences between cancer cells cause chemotherapy to fail.Traditional chemotherapy works because cancer cells proliferate faster than normal cells. By carefully adjusting the dose of a toxic chemical, cancer cells can be eliminated. However, since the difference between a dose toxic to cancer and a dose toxic to healthy cells is small, a minor increase in resistance of cancer cells may be sufficient for treatment failure. It is not fully understood why cancer comes back in some patients despite initial positive response to chemotherapy, and why its success rate varies greatly among different types of cancer. One possible reason is that cancer cells differ in their response to a chemotherapeutic drug even in a single person. This can be caused by mutations. Alternatively, non-genetic variations that naturally occur in most cells in the level of the protein could be responsible. To distinguish between these two possibilities, we will perform experiments in which cancer cells are cultured in a small transparent container, which makes it possible to observe the behaviour of cells using a special, automated microscope. We will thus be able to see how the effect of the drug depends on how the cells behaved prior to exposure to the drug. Cells will also be screened for mutations; this needs to be done at the end of the experiment because it involves destroying the cells in order to extract their DNA. These experiments will help to determine how much of chemotherapy failure can be attributed to genetic, and how much to non-genetic causes. This is very important for the future of chemotherapy. If non-genetic effects are very strong, sequencing DNA from cancerous tumours will not give us full information about what therapy would be best for a specific patient. Instead, it may be necessary to expose cells from a patient biopsy to a range of drugs and select the ones with best response.
Summary of project results
In this project we investigated how individual glioblastoma cancer cells responded to chemotherapy, and how the response correlated with their behaviour prior to treatment. Glioblastoma is an aggressive form of brain cancer which is very difficult to treat. A combination of surgery, radiation, and chemotherapy is often used but in most cases the cancer comes back.
We were hoping to better understand the reasons behind chemotherapy failure. Our hypothesis was that, in any glioblastoma cancer, there are cells with different levels of resistance to chemotherapy: most cells are sensitive but there is a small fraction of highly-resistant cells. These resistant cells do not respond to treatment and, while sensitive cells are killed by the drug, the resistant ones continue to proliferate and are responsible for tumour regrowth.
We posited that the resistant cells emerge due to natural variations in the level of certain proteins produced by the cells. One of them, a protein called MGMT, had been previously linked to resistance to a certain class of drugs called alkylating drugs. We wanted to check if the level of this protein inside cancer cells determined how the cells reacted when exposed to an anti-cancer drug.
To test our hypothesis, it was required to see what individual cancer cells were doing. Since observing individual cancer cells is not possible in a patient undergoing chemotherapy, we created a simple “in vitro” (“in a test tube”) model of glioblastoma treatment. We incubated glioblastoma cancer cells in a special plastic plate which could be inserted into a computer-controlled microscope and imaged for many days.
After the cells settled and started multiplying, we added the drug temozolomide – an alkylating drug used to treat glioblastoma. We then followed the fate of individual cells for a few days, counted how often they divided, how many of them died, and how much of the “resistance factor” MGMT was present in each cell. Because MGMT is not normally visible, we created genetically modified glioblastoma cells in which MGMT fluoresced green when blue light was shining on the cells. We could then obtain the amount of MGMT by measuring how much green light the cells were emitting.
The number of resistant cells was expected to be low, therefore we had to image thousands of cells under the microscope to find at least a few resistant ones. Tracking so many cells would have been a very time-consuming and exhausting task for a human. We therefore used neural networks (machine-learning algorithms) to automatically detect cells in tens of thousands microscopic images. We then analysed the behaviour of those cells using automated computer algorithms.
We found that different cells responded differently to chemotherapy: some stopped dividing very soon after the addition of the drug, while other cells continued to multiply. Non-growing cells started to move around much more when exposed to the drug than before the drug. We think this “migratory behaviour” could be a contributing factor to glioblastoma spreading to other parts of the brain during chemotherapy.
Contrary to our initial hypothesis, we did not find MGMT-rich cells to behave differently to cells with low levels of MGMT, that is there was no correlation between MGMT and drug resistance. We therefore concluded that there must be other factors responsible for differences in resistance of individual cells that are more important than differences in MGMT level. While we weren’t able to find the exact nature of these factors, a mathematical model of glioblastoma treatment which we created suggested that these factors probably involve damage to other proteins inside the cell.
These findings may be of interest to other scientists researching the resistance to chemotherapy, because they show that MGMT may not be the only factor relevant for resistance as it has been thought. MGMT is also used as a diagnostic marker to predict how well a patient will respond to chemotherapy. Our research shows that there may be other important factors and markers that medical doctors could use to better predict therapeutic outcome. This however requires more research to identify these factors.
Besides the scientific outcomes, our project has also led to the creation of a computer game in which players can treat a “virtual patient” suffering from glioblastoma. We presented this game to the public at the European Picnic organized by the German Embassy in Warsaw, 11 May 2024, to which we had been invited to. The game turned out to be very popular and attracted a large number of adults and children alike.
We also presented the project to the general public during the 2023 Warsaw Science Fair.