17th Congress of EHA, Amsterdam, 14-17 June 2012

Whole genome sequence analysis for acute lymphoblastic leukaemia

Dr Charles Mullighan – St Jude Children’s Research Hospital, Memphis, USA

Thanks for joining us, Dr Mullighan, on ecancer.tv here. You have the difficult job of investigating the whole genome sequence in acute lymphoblastic leukaemia, first of all could you tell me why you’re looking at the whole sequence because I know you’ve got a lot of information already about individual genetic differences between patients with ALL and those without, so why did you go for the whole genome?

At its heart we know that ALL is a genetic disease and we’ve known many of those genetic alterations from other approaches that are whole genome sequencing but those other approaches haven’t been able to identify the full nature of the genetic changes that underlie ALL. Also, they haven’t been able to tell us and predict which patients are likely to be cured with therapy and which ones will relapse. So there’s been a strong interest in using next generation sequencing, including whole genome sequencing, to find all genetic changes that drive leukaemia.

Now, could I ask you briefly about a paper you’re co-author on, on the 22 MLL rearrangements in ALL; you were looking for mutations, weren’t you in that?

Yes, that’s one sub-type of ALL that we’ve studied.

An aggressive form of disease.

It’s an aggressive form of leukaemia often found in very young children and infants that’s characterised by rearrangement of a gene called MLL. This was a study that sought to see if there were other genetic changes that co-operated with that alteration and very few were found.

What’s the significance of that?

It suggests that in this particular type of leukaemia that very few genetic changes are needed to drive the aggressive nature of the disease. That’s actually in contrast to many other sub-types of leukaemia that are really constellations of multiple genetic changes that act together to drive the disease.

So that is both bad and good because if you can find out what it is then there is a possibility of intervening?

In terms of infant leukaemia, that change was known about already so I guess the question is what other genetic information can be used to help influence therapy. Now in some other types of leukaemia the mutations we find from whole genome sequencing may be actionable and there are some other examples of that. In the case of MLL rearranged leukaemia it may be that additional genetic changes aren’t going to identify new therapy, it may be other approaches, perhaps epigenetic approaches, that may be therapeutically beneficial and that’s been shown also in MLL leukaemia from other studies.

Now I know particularly you’ve been looking at sub-types of leukaemia and you’re very interested in BCR-ABL like leukaemias. What’s going on there and what are the hopes?

That’s a sub-type of leukaemia that has some similarity to BCR-ABL positive leukaemia which is another aggressive form of ALL which harbours the Philadelphia chromosome that encodes BCR-ABL and is now treated with tyrosine kinase inhibitors. BCR-ABL like leukaemias are BCR-ABL negative but have some similarity and our hope is that we can find genetic changes that may also be druggable in that form of leukaemia.

Do you think that the same sort of approach, a TKI, might work in those cases?

It may but time will tell.

So you’re looking also for genetic changes that could give you predictive information. Now I gather that some of your work has been looking at some of the diseases that are already understood in genetic terms but which you’re now finding having additional genetic effects which are complicating that or giving the disease a different course. Can you explain that part of your work to me?

Yes, we’ve known also for a long period of time that there are some well-established genetic changes in leukaemia that strongly influence the risk of relapse or treatment failure. So one has been mentioned, MLL, there are others as well such as hypodiploid in which there are losses of multiple whole chromosomes, but there are many other patients that fail therapy where the genetic lesions responsible for failure are unknown. So over the last few years we and others have performed quite detailed genetic analysis, not just whole genome sequencing but other approaches as well, that have identified some genetic changes that are associated with a substantial variation in the risk of treatment failure and relapse. One example is mutation of a gene called ikaros, or IKZF1, which if present in B-cell ALL, the commoner form of ALL, can be associated with up to a tripling in the risk of treatment failure and relapse. How exactly it’s doing that we don’t know yet, there are probably multiple ways that it influences the course of the disease, one is that it makes it a more primitive, more aggressive form of leukaemia. But clinically that’s important because it suggests that if one screens patients at diagnosis for that genetic alteration, one might perhaps consider adjusting their therapy to something that’s either more aggressive or use a novel therapeutic approach and that’s one example that has been found.

So for the practising clinician, what are the pay-offs coming from the whole genome approach and the more detailed understanding of translocations, deletions and all of the genetic processes happening?

I’d say overall from a clinical perspective we’re still relatively early, we’re still in a process of cataloguing the genetic landscape in ALL. We are finding a lot of new genetic changes, many of which are biologically exceedingly interesting, they’re going to give us important insights into what drives the development of leukaemia. Some are likely to be clinically important, like ikaros, and others may not be, at least not immediately. What I would say is watch this space; I think over the next couple of years we will see comprehensive cataloguing of the leukaemia genome and its genetic changes. Then the next phase of work will begin to see which ones of those can be put into the clinic as diagnostic tests that might help us better stratify or diagnose patients when they first present with leukaemia and then further down the track future work will determine which of those are going to be actionable, whether it’s with a tyrosine kinase inhibitor or some other new approach to treat the leukaemia cells.

So all very promising, some way off in some respects, but you’re talking about getting predictive information. Could you give me an example of the sorts of choices you might be able to make, choices of therapy, as a result of the predicted information of understanding these genetic events?

I guess there are a few different options, one is that there are different regimens that are used to treat ALL that are of different intensity. So if you were to have a predictive marker that you know is associated with an increased risk of relapse, one might immediately put a patient on a more intensive regimen or monitor them more closely expecting that they might have a less than optimal response to initial therapy. Maybe those patients need to be then transferred to a regimen that has a novel agent or incorporates transplantation because of a poor treatment outcome. In specific examples they may have something actionable, whether it’s a tyrosine kinase mutation, genus kinase mutation or other, and those studies that are investigating the role of targeted therapies such as JAK inhibitors and others are already under way.

And what are the other tools, then, that are needed that can be provided by a better understanding of the genetics of the whole situation that would empower the doctor?

Better tools would include better diagnostic tools, so at the moment our diagnosis at a genetic level is limited to carrier typing, which is cytogenetic analysis, and a fairly small panel of targeted molecular tests. I would expect that over time that would become more sophisticated, for example, there will be broader based testing of new genetic changes to detect individual changes and ultimately it’s highly likely that next generation sequencing itself will become more of a routine diagnostic tool. That’s being very actively pursued by many places, it’s not widely done at the moment for various reasons, including costs and complexity and the challenges of data analysis and how to report this information back to the treating doctor. But again, at the moment we’re very much in a discovery research phase and it’s likely that many of those approaches will become diagnostically more routine.

And finally, briefly, there’s a whole raft of genetic information coming out here in Amsterdam, how would you sum up the one or two points that the doctors should remember from all of this?

I guess I’m not familiar with all of the genetic studies but there are several genetic changes that are being found in multiple neoplasms, both myeloid and lymphoid as well. I guess for the clinicians it’s to see those results, note them but also to wait and see which ones prove their clinical worth in subsequent studies over time to see which ones are clinically important.

Charles Mullighan, thank you very much indeed.

You’re welcome.