Novel method for selecting the best drug therapy for cancer

Choosing the right drug for each cancer patient is key to successful treatment, but physicians currently have few reliable pointers to guide them in designing treatment protocols.

Researchers at the Weizmann Institute of Science and the Broad Institute of MIT and Harvard have now developed a new method for selecting the best drug therapy for a given tumor based on assigning scores to the cells' internal messaging activities.

In addition to helping physicians choose from a list of existing treatments, the method can help identify new molecular targets for the development of future drugs.

In fact, the researchers have already used it to single out a gene that can be targeted for effectively treating breast cancers with a BRCA mutation. The study was recently published in Nature Communications.

The most common molecular way of matching drugs to a tumor is to look for particular mutations in the tumor's cells.

Unfortunately, the presence of such mutations is no guarantee that a drug will work and, in any event, many drugs are not aimed at mutations in the first place.

There have also been attempts to predict a drug's effectiveness by analyzing the expression of certain genes in a tumor; however, the expression level of very few genes was found to be helpful in guiding physicians in making treatment decisions.

In the new study, two labs – one headed by Dr. Ravid Straussman of Weizmann's Department of Molecular Cell Biology, the other by Prof. Gad Getz of the Broad Institute – joined forces to develop a more effective approach, basing their research on the enormous cancer datasets that have become available in the past few years.

This approach does not rely on mutations or on individual genes but, rather, on signaling pathways: chains of biochemical signals that convey crucial cellular messages – for example, whether a cell should divide or grow, or in what way its metabolism should be altered.

Numerous genes are expressed in the cell in order to transmit the message in each pathway; thus, sophisticated methods are needed to uncover the activity in these chains.

Weizmann postdoctoral fellow Dr. Rotem Ben-Hamo analyzed vast international datasets containing information on the expression of all genes in some 460 cancer cell lines – that is, different models of cancer – from 10 different cancer types.

Using an advanced bioinformatics tool, PathOlogist, developed by Prof. Sol Efroni of Bar-Ilan University, the researchers assigned to each pathway an activity score that takes into account not only gene expression levels, but also prior knowledge about the structure of each pathway, the interactions of the genes within it, and whether a given gene blocks or stimulates the pathway's message.

The team then correlated these scores with datasets containing information on the sensitivity of a variety of cancer cells to nearly 500 different cancer drugs.

The scientists found that the activity scores of some of the pathways enabled them to predict whether a specific cancer would be sensitive to a particular drug.

In other words, the researchers created a profile for the cancerous tissue that could direct clinicians to the best drugs for eradicating the tumor.

Overall, they were able to make such predictions for more than 30 existing drugs.

For example, when certain lung cancer cells had a high score for a pathway triggering apoptosis (a form of cell suicide), the cells were likely to be killed by a class of drugs known as microtubule inhibitors.

Next, the group showed that they could use the pathway knowledge to not only predict, but to alter the cells' response to a drug.

They obtained, from one patient, lung cancer tissue in which, their analysis indicated, the apoptosis-triggering pathway was not particularly active.

In test-tube experiments, this lung cancer tissue was resistant to microtubule-inhibitor drugs – as expected.

But when, in addition to microtubule inhibitors, the researchers simultaneously applied a substance that increased activity in the apoptosis pathway, the combination of drugs effectively killed the cancer cells.

In further analysis of the datasets, the researchers correlated the activity scores of signaling pathways with yet another type of information: which genes play such an essential role in various tumors that silencing, or blocking, these genes can kill the tumor.

They found that here, too, the pathway scores helped them identify such "sensitive" genes in a variety of tumors.

For example, the scientists found that breast tumors with specific activity in the BRCA pathway – correlating with the presence of a BRCA mutation – were extremely dependent on the activity of a gene called MAD2L1.

The bioinformatics analysis predicted that silencing this gene can result in the death of tumor cells in patients with a BRCA mutation.

This prediction can serve as a starting point in the search for drugs, both new and existing, to treat this devastating form of breast cancer.

Overall, the study's findings suggest that signaling pathways can serve as predictive biological markers in the personalized medicine of the future, helping physicians determine in advance which patients will best respond to which drug.

Moreover, the pathways can help researchers identify the Achilles heel in tumors, thus providing a direction for drug development.

Journal reference:

Ben-Hamo, R., et al. (2020) Predicting and affecting response to cancer therapy based on pathway-level biomarkers. Nature Communications.


The opinions expressed here are the views of the writer and do not necessarily reflect the views and opinions of News Medical.
You might also like... ×
Cancer cells hibernate to survive harsh environment created by chemotherapy