I’m going to talk about a BCL2 mutation which we detected in patients treated with venetoclax. As you probably all know, venetoclax is a highly potent and selective BCL2 inhibitor; it counts itself as one of the most effective medications for CLL. It’s typically given in monotherapy at the moment although it’s currently being trialled in combination. Despite its efficacy the majority of patients will relapse on venetoclax and up until now we didn’t really have a good idea about why people were relapsing.

You can relapse with CLL either with a low grade CLL progression or you can relapse with transformation to diffuse large B-cell lymphoma. So what we looked at was 67 patients who were treated in the early phase clinical trials. These were some of the first patients treated in the world with venetoclax and we took 21 of those patients who had progressed whilst taking venetoclax single agent therapy. We did next generation sequencing on their tumour samples, both before and after progression. What we found was in four patients we found a single nucleotide variation, the G101V, in all four patients and in all four patients we were able to show that this wasn’t present in their cells before they received venetoclax.

So what does all that genetic information mean? This is the BCL2 protein here, so this is the business end of the BCL2 protein, it’s where all the pro-survival molecules bind and this is also where venetoclax binds to and this is venetoclax here in orange. Our mutation occurs here, so this is a blown up version of the alpha-2 helix and you can see here that venetoclax is sitting right next to this amino acid. This amino acid basically changes from a small glycine amino acid to a bulky valine amino acid so you can imagine it pushing into that groove and potentially stopping venetoclax binding, or at least that’s what we thought.

Importantly, you never see this mutation in cancer otherwise so it hasn’t been described in the entire COSMIC database in over 50,000 cancers. We sequenced 400 patients with BCL malignancy in our laboratory and didn’t find it in anyone who hadn’t been exposed to venetoclax. This is the time course of its development in these patients so in four of the patients where it was at significant level you can see here that essentially it begins to appear from 20 months of therapy onwards so it’s quite a late mutation but then in everyone where it’s detectable the patients inevitably progressed to overt clinical relapse and had to either come off therapy or be treated with something else.

So we took this discovery into the laboratory and worked out that the patients’ cells with venetoclax have up to a thirty- to fiftyfold decrease in sensitivity to venetoclax. This is as close as we came to recapitulating what was happening in the cell dish. We took a cell line, we gave it the G101V mutant and we put it in the context of sub-lethal doses of venetoclax and you can see here that you gradually get an increasing dose of this mutation over time. This is over a three week period but you can imagine this is what’s happening in a patient, the gradual increase of this G101V mutant in the context of this selection pressure of venetoclax.

This is just the experiment showing how it works and basically this is a protein binding experiment showing that it decreases the affinity of venetoclax for that group by 180-fold compared to wildtype, so a 180-fold reduction in the efficiency of venetoclax binding to BCL2 which makes it a very good candidate variant for being a resistance mechanism.
Finally, importantly, it’s not the only mechanism that’s present in patients. So in all the patients we detected it we didn’t detect it in all the cells. The most we detected it was was in 75% of cells which means 25% of the cells which are resistant to venetoclax are not driven by G101V. In this patient we managed to find upregulation of a molecule that’s like BCL2 which is Bcl-xL which was driving venetoclax in those cells. So there’s not only one type of resistance in these cells but this is certainly an important one.

So, in summary, this is the first BCL2 mutation that’s been described in patients with CLL and mediates resistance to venetoclax. It seems to work by markedly reducing the affinity of venetoclax for BCL2. It’s sub-clonal and we were very surprised about the frequency with which it was observed – we observed it in 50% of the patients we tested and the fact that there was only just this one mutation occurring and not a bunch of mutations occurring in BCL2, it’s just this one. So basically we think this is a strong rationale for using combination therapies in the future; this is what happens when you give a targeted agent to a genetically complex leukaemia. Second of all it’s a rationale for giving time-limited therapies, as Tait was saying, that limiting the exposure to venetoclax will limit the development of this resistance.