ASH 2015

How can functional drug screening guide personalised cancer therapy?

Prof Jeff Tyner - Oregon Health & Science University, Portland, USA

You’re addressing the meeting on functional screening to guide personalised cancer therapy. Everybody wants personalised cancer therapy, what is this topic and why is it so crucially important?

Yes, a great question. Almost when everyone thinks about personalised cancer therapy it’s almost implicit that people think about genomics and mutant genes and how that can drive new therapies and understanding of better treatments. That certainly is a very powerful paradigm but unfortunately there are a lot of mutations that aren’t drugged now and so there are a lot of these mutational reports for which they’re really not useful, currently, for clinicians to prescribe therapies because we don’t have therapies that can target some of these mutational events. So my work, and the work of a lot of other people in the field, both in the United States and in Europe, has been taking a different approach which is to ask the cells themselves. Let’s take the cancer cells out of a patient, let’s test these cells in an ex vivo, in vitro setting in the laboratory, let’s test them against a panel of different drugs that could be useful in the clinic and ask the cells to tell us which drugs they are sensitive to.

How fascinating.

Absolutely.

And what actually would you do? What kinds of tests do you carry out? You have a whole battery of drugs, any of which could be used for your cancer and you can then choose one, is that it?

Absolutely. So in my lab we have a drug panel that has over a hundred different drugs, mostly targeted therapies, a lot of them are kinase inhibitors, some of them we’ve really branched out to diversify, including BH3 mimetics, bromodomain inhibitors, proteasome inhibitors, so on and so forth. We can test all of these drugs simultaneously against individual patient samples; we’ve tested over a thousand key malignancy patient samples against this panel of drugs and we see that about 90% of patient samples have hypersensitivity to at least one of the drugs on the panel. We can say confidently that almost none of these drugs are currently standard of care for patients so there’s a real opportunity, our data would argue there’s a real opportunity to repurpose these drugs and use them much more broadly on the basis of a functional screen. Even if we don’t understand immediately the mechanism, even if we don’t understand immediately the mutation that’s causing hypersensitivity to that drug, the functional assay can tell us that that drug could be useful and that patient might benefit from that drug.

You’ve got a pretty big sample there with a thousand or so patients giving samples, are the different approaches or the different sensitivities that you’re discovering widely different from the common methods of selecting drugs or just a little bit different?

I think the idea of using functional screening to drive therapeutic choices is an idea that has been around for some time. Earlier, back in the ‘80s and ‘90s, when this was done people were culturing cell lines which were not necessarily indicative. After they had been in culture for a few months the cell lines were no longer representative of the tumour and therefore there were some disappointing results in those early stages of functional testing. Now we’re not culturing cell lines, we’re taking the cells immediately out of the patient, putting it into the culture, doing three day assay, in my lab it’s a three day assay, and so there’s really no opportunity for the cells to deviate from their original tumour characteristics. We’re finding that there’s also much higher fidelity to clinical responses. It’s still an ongoing process to gather more and more examples of clinical responses relative to the ex vivo assay but the data thus far indicate a very high correlation rate. About 80% of the time our assay is successful in predicting the clinical response of a patient to these targeted therapies.

You might select one type of cell but it might be a complex system where other cells are very valid and you might go off on one track and fail to follow another. Isn’t that a danger?

Absolutely, these are complex diseases, particularly the acute leukaemias, and nobody expects that a single agent therapy is going to be effective at curing AML or ALL. We know that we’re going to need combinations and we are starting to test combinations. This is an assay that can also deliver us information about which combinations could be most effective. You also bring up another good point, right now we’re doing fairly basic cell viability assays in order to look more at the heterogeneity of the sample we need to expand into more imaging, single cell based imaging assays that will allow us to be more sophisticated in how we understand these drug responses.

And wouldn’t you have to use this method serially one time after another because of the clonal evolution of the cells?

Absolutely, and we have some examples of that. We’ve actually published a case report in a paper in Cancer Research a few years ago where a patient with very aggressive AML was on three or four lines of cytotoxic therapies, refractory disease in every case. We tested, we found sensitivity to a number of different targeted agents, was treated with one of them, had a nice response, relapse, tested again, sequentially treated with a different one, had another response. So we have done some of this. My argument would be I don’t want to let the patient’s disease evolve to that stage, I would rather do comprehensive testing up front, give a combination of different drugs that hopefully will put the patient into a long term remission so therefore we don’t have to deal with these very complex relapsed refractory cases.

Get ahead of the game, so to speak.

Exactly.

But individual doctors may well think, ‘Well, what a wonderful idea but how can I be involved soon?’ What would you say to them?

There’s a lot of interest in this technology and there are a lot of centres, not just in my lab, there are a lot of groups doing great work in Europe and across the United States, mostly at academic medical centres at this point, who are doing this type of functional testing. So we’re always receptive to people sending us samples, we actually have collaborators all across the country who are sending us their patient samples and we’re doing this testing and we can report back some results. There are also some start-up companies that are beginning to evolve and hopefully in the very near future will be offering this as a service, a commercial service.

This could be highly cost-effective.

I would think so.

So would you sum up what is the bottom line for doctors coming out of this practically?

My summary would be that while precision medicine can be powered by genomics, and that’s a very powerful strategy, I would argue that we should also include functional testing to power precision medicine and at least in the interim while we are still learning the biology of all these complex mutations and combinations of mutations this functional testing can actually be a very powerful tool to drive precision medicine.