I focussed my presentation on MLL-rearranged leukaemia, specifically in infants. This is a developmental disease and it remains an outlier of all the B-ALLs. This means that it is probably the infant leukaemia with the worst clinical outcome – the overall survival and the event free survival of these kids is very poor – so we need more investigation, more research and new drugs and new therapeutic approaches.

What I disclosed was, first, how this disease evolves because it evolves prenatally in utero during pregnancy. From the theory point of view we know that it is linked to exposure to topo II inhibitors; this is similar to what happens in adults with therapy related AML that have received previously etoposide or other topo II inhibitors. Then I commented on the cell of origin. This disease occurs prenatally, as has been beautifully shown by Mel Greaves using monozygotic twins and cord bloods and [??]. The cell of original, we know it’s a prenatal cell, a prenatal haematopoietic stem cell very early in development; it’s a cell, we call it a pre-VDJ cell, this means that it’s a cell before the BCL receptor gets rearranged, and it has to be prenatal rather than neonatal. This has been beautifully shown now by transcriptomics, by PCR sequencing as well. Then we demonstrated that we can generate a translocation itself, per se, in vivo, in these CD34 positive haematopoietic stem cells using CRISPR Cas9 genome editing. We learned a lot with these experiments, now we know that the translocation can occur in these cells and these cells can be immortalised in vitro. However, these cells don’t transfer in vivo. This means that the cells, just by having the translocation, the MLL-A4 translocation, these cells cannot initiate leukaemia in mice.  So this suggests that we are missing secondary cooperating oncogenic events.

So we have then, together with INTERFANT, with our colleagues Ronald Stam, Rob Pieters and so on in the Netherlands, in Princess Máxima, we have been doing a whole genome sequencing on many patients and we discovered that this is the infant with the lowest mutational burden. I would say that this leukaemia is silent, mutationally silent. The only pathway that is recurrently mutated is KRAS and the PI3 kinase pathway but all these mutations are sub-clonal, indicating that we have so many sub-clones that these mutations in KRAS are not initiating events, they are probably passenger events. So treating these patients with drugs tackling these mutations is not probably going to be a solution.

Then we went on and I showed that at relapse the scenario is different. We have a clone selection at relapse and there are many more mutations and some of the mutations present at diagnosis are lost at relapse and there are also new mutations at relapse that there weren’t at diagnosis. So this even complicates more the scenario.

We are doing also whole genome epigenetics. So far all this stuff is still in progress. What we know is that MLL rearranged leukaemias, infant leukaemias, they cluster away from normal BC progenitors, B-cell progenitors, and also normal B-cells. Among the different subtypes of MLL rearranged infant leukaemias there is a subtype that is the most common and the one with the worst prognosis is the MLL-A4 subtype that really clusters away from the other subtypes. So we are now selecting some putative candidates that may have a contribution to the pathology but this is still in progress. We probably need a few months to finish this story.

That’s on the therapeutic intervention, new therapeutic approaches. I’ve reviewed the role of NG2, NG2 stands for neuron-glial antigen-2, this is a molecule that is expressed only in MLL rearranged leukaemia. We have shown that NG2 does not enrich for leukaemia initiating cells but it does select the cells migrating to the CNS, to the central nervous system, and to the extramedullary sites. NG2 correlates with a worse prognosis, with a more immature phenotype, with more aggressive leukaemia and we have shown beautifully in xenograft models that if we combine induction treatment like VXL – vincristine, asparaginase and dexamethasone – with anti-NG2 antibodies we can both reduce relapse, or I would say delay relapse, and also we increase the rates of complete remission.

So, overall, we are slow but we are really understanding more this disease. All this talk together with other speeches we saw from other colleagues, I think we are really advancing the biology and the clinics of this poor outcome dismal disease.