ALL is a very genetically diverse disease, it comprises a number of distinct entities each of which has its own constellation of genetic alterations. Some of these are very important for leukaemogenesis and the formation of the initial leukemic clone and some of them are very strongly associated with the risk of treatment failure and the risk of relapse.

In children there are some forms of leukaemia that are associated with gains of whole chromosomes such as high hyperdiploidy where there are a number of whole chromosomal gains that are quite stereotyped - we see the same chromosomes being gained repeatedly. We don’t have a good understanding of how that change contributes to leukaemia formation but that subtype is associated with very good outcome and is also very uncommon in older adults where outcome is worse. Another subtype has a single fusion called ETV6-RUNX1 or TEL-AML1 to use its previous name. That’s a fusion of two different transcription factor genes and it’s also associated with very good outcome and is more common in children than adults.

There are several other subtypes such as rearrangements of MLL, BRC-ABL which is less common in children but more common with older age that activates ABL1, a tyrosine kinase, and that’s been really the first example of a chimeric fusion that can be targeted with a specific therapy either in ALL or in chronic myeloid leukaemia.

So the focus of much of what I spoke about at the meeting related to several new subtypes that are very uniform in terms of their gene expression profile, they have a very distinct pattern of gene expression but they’re more genetically diverse. They have a number of different founding genetic alterations that all result in the same leukemic phenotype. Also many of those alterations were not previously evident by conventional diagnostic approaches such as karyotyping which is used in cytogenetic analysis of leukemic cells. That’s one of the main points to accurately identify these forms of leukaemia because they have prognostic significance. One needs to use more contemporary genetic approaches such as genome sequencing or similar modalities of testing the leukemic cell.

So a couple of key examples are Ph-like leukaemia. We’ve been studying this form of leukaemia for some years now, it’s one of the most exciting new subtypes of B-ALL in that there’s a diverse range now, over fifty different genetic alterations that activate a number of different cytokine receptor genes or tyrosine kinase genes that result in kinase signalling. These alterations cause cell proliferation that’s very druggable with many available tyrosine kinase inhibitors so there’s a great deal of interest in defining the prevalence of this form of leukaemia. There hasn’t been so much work in older adults where prognosis is poor and also implementing the right approaches to identify these patients and treat them prospectively in well-constructed clinical trials.

This was also spoken about in detail by a second talk here by Dr Boer who talked about therapeutic targeting of Ph-like ALL. Two important points that I made in my talk were that in a recently completed study that we performed and published in late 2016 where we studied almost 1,000 adults with ALL we showed that this form of leukaemia was higher than previously suspected in older adults at about 20-25% of cases. So it reinforces the significance of this type of leukaemia right across the age range. Using genome sequencing we showed that there’s an increasing number of kinases and an increasing number of kinase fusion partners that we’re finding rearranged in this type of leukaemia. That underscores the importance of comprehensive genetic analysis at diagnosis and this is challenging for many centres and laboratories but it’s clearly where the field is moving to implement these approaches at initial diagnosis.

There are a number of important issues here but each of them are tractable; you’ve touched on some of them. So one is the clinical question – so what is the desired clinical result that’s going to be acted upon? Is it identifying all patients with Ph-like ALL, the majority of which we think are druggable but not all are at the moment with available drugs? Or is the goal to just identify a subset of the lesions that might be more amenable to a focussed diagnostic approach? For example fluorescence in-situ hybridisation is used very widely, it’s not comprehensive but it can identify many of the key fusions that are the most intense focus of clinical activity at the moment. One is scale, so is the question for a large co-operative group study that’s potentially screening hundreds or thousands of cases and they need to screen all of them or is it a single centre or a single lab thinking about how they employ sequencing or similar approaches? I’ve been involved in both so I work closely with collaborators at St Jude but also with the Children’s Oncology Group so they’re very useful contrasting examples. At the Children’s Oncology Group they have open studies at the moment that are screening thousands of children so they can’t sequence all of them. They’ve implemented a screening approach that uses a very simple standard low density gene expression array that any lab can run to first identify the Ph-like gene expression profile and then uses a series of assays so that sequencing is only needed for a minority of cases eventually.

A contrasting example would be at St Jude Children’s Research Hospital where we feel that comprehensive sequencing at diagnosis has many advantages. It provides the most comprehensive information across all subtypes of leukaemia and other tumours as well, not just restricted to Ph-like, as well as identifying the key lesions that we think are targetable. That’s the approach in the next study of ALL at St Jude is that all patients will have whole genome, exome and RNA sequencing at diagnosis.

The two issues that that then leads to are cost and feasibility. The cost of next generation sequencing is not trivial but it’s not exorbitant when you consider the existing costs of diagnostics. If you consider current cytogenetics, molecular testing, flow cytometry and all of those approaches add up to several thousands of dollars. The consumable cost at the moment for that kind of comprehensive next gen sequencing is less than US$10,000. So from a cost perspective it is feasible and the costs are only going to decline further. The most important issues are speed and analysis. The sequencing itself is fairly standard now in many centres but the challenge is getting the material to the site to be sequenced quickly and also having the right laboratory structure that’s accredited with the appropriate regulations to perform the sequencing and then doing the analysis of the data in a timeframe that provides the right information to clinicians quickly. One example is we need a subset of the information within fifteen days and that’s also quite tractable but it does pose challenges for many centres to deliver sequencing that quickly.

If we look at the genomic landscape of Ph-like ALL alone about 90% of cases have been found to have druggable kinase activating lesions. Not all of them have a drug in the market at the moment for ALL but there are two major subclasses that comprise at least two-thirds of cases where there are two drugs that we know are active in preclinical models. There’s a subset of cases where there’s a very clear oncogenic kinase activating driver and there are interesting drugs that can target them but they’re not yet in the clinic. That’s another area of active development is getting these targeted agents that will potentially treat only a small subset of patients into the clinic.

Now, that aside, as you suggest, there are many mutations that are leukaemogenic or they’re markers of high risk disease that aren’t kinase activating lesions, that don’t have an intuitively druggable landscape but perturbing them somehow might be therapeutically beneficial. There are a number of examples of that sort of work that’s ongoing. Some are very focussed on mechanism and some are focussed on broad based drug discoveries. So, for example, you set up an engineered cell line or isogeneic lines with or without mutation and then you can do high throughput drug screens or synthetic lethal screens where you have, for example, a drug line with a mutation that’s resistant to existing chemotherapy. Then you look for active agents that restore the activity of the existing chemotherapeutic approaches. That’s been very widely used; the newer functional genomic screens are also being used to that end as well – previously RNA interference and now CRISPR screening so that you can perform those screens in the right models and perhaps identify genes or pathways that might then lead to an alternative or synthetic lethal approach. Maybe unrelated to that direct mutation itself but it’s activating or working through another pathway that can bypass and restore sensitivity to drug action.

There are also other pathways that aren’t just signalling that are potentially amenable to therapeutic intervention. There’s one that we and several other groups have worked on for one class of alteration and that’s alterations in lymphoid transcription factor genes which are a hallmark of both B- and T-cell ALL. Some of those alterations are very strongly associated with drug resistance. One mechanism that has been elucidated is that they can act by deregulating adhesion and they cause leukemic cells to adhere to each other and also to the bone marrow vasculature and mis-localise in the bone marrow microenvironment and then they become resistant to therapy. The approach that several groups have taken there is to target that adhesion so you’re not actually directly acting on the mutation itself but you are acting on the downstream pathway that is the key mediator of drug resistance.

So there are multiple ways to try and act on these different mutations. We clearly have an active signalling cascade and that’s logical that you can turn that off but there are other ways, either in a very focussed approach or in a synthetic lethal or functional genomic approach that one can consider finding new ways to overcome resistance.