Novel therapeutic approaches in B-progenitor ALL

Bookmark and Share
Published: 23 Jan 2020
Views: 642
Dr Charles Mullighan - St. Jude's Children's Research Hospital, Memphis, USA

Dr Charles Mullighan speaks to ecancer at the 2020 ALL Assembly meeting in Frankfurt about the treatment landscape of B-progenitor acute lymphoblastic leukaemia (B-ALL), including the use of novel approaches.

He describes the challenges of treating Ph-like acute lymphoblastic leukaemia (ALL), in which patients with the CRF2-rearranged subtype respond poorly to the JAK/STAT inhibitor ruxolitinib.

To target this subtype differently, Dr Mullighan explores the effectiveness of using targeted protein degradation to directly inhibit the JAK/STAT signalling pathway.

He also discusses the mechanisms of resistance to blinatumomab, which is a widely used agent that is increasingly used in the front-line.

ecancer's filming has been kindly supported by Amgen through the ecancer Global Foundation. ecancer is editorially independent and there is no influence over content.

My presentation at the meeting focussed on two areas related to novel therapeutic approaches in B-progenitor acute lymphoblastic leukaemia. The first is related to Ph-like or BCR-ABL-like ALL. This is a very high risk subtype of leukaemia that remains a clinical challenge and lacks effective therapy. Many of the patients that have Ph-like leukaemia have an alteration of a gene called CRLF2, a rearrangement of this gene that results in activation of the JAK-STAT signalling pathway. For that reason people are now testing the activity of JAK inhibitors such as ruxolitinib, the clinically approved inhibitor, in patients with CRLF2 rearranged ALL. However, it’s been seen anecdotally in relapse studies and also experimentally that these leukaemias tend to respond poorly to ruxolitinib in contrast to other subtypes of leukaemia.

So in this part of the talk I present some recent data where we’ve been taking a different approach to try and target CRLF2 rearranged ALL. That is, rather than trying to use small molecule inhibitors of the JAK-STAT signalling pathway, to use targeted protein degradation to directly degrade the Janus kinase family and to shut off signalling through this pathway. This is a very interesting and highly disruptive technology called PROTACs, or proteolysis-targeting chimeras, where the ubiquitination machinery of the cell can be harnessed to target new substrates, in this case the JAK family, for protein degradation. We’ve shown that this approach is highly effective in targeting the JAK family and also killing these leukemic cells.

The second part of the presentation also focuses on a novel therapy but a different area; it’s looking at mechanisms of resistance to blinatumomab. This is a widely used agent now that’s increasingly being brought frontline to treat patients with ALL. But like most novel agents and targeted agents some patients respond and some patients don’t and the reasons for that are poorly understood. So also in a study that we’ve conducted in the last several years we’ve taken a cohort of patients at different stages of disease, most of whom are relapsed and refractory and many of whom have failed multiple lines of therapy, that have samples at baseline, pre-blinatumomab, after response and at treatment failure. We’ve used a variety of different approaches including genomic analysis of the tumour cells and interrogation of the tumour marker environment to gain insight into what are the factors that dictate or influence whether patients respond or not.

The findings have been very striking. Firstly they show that tumour genotype, so the presence of sentinel rearrangements or genetic changes, is not the principal dictator of whether patients respond or not. We do see, however, that some of the highest risk leukaemias, including CRLF2 rearranged leukaemia, actually tend to respond quite well. There is a gene expression signature of immune activation of the tumour cell that’s associated with favourable response.

In a second part of the study we’ve looked at the notion of escape due to loss of CD19. This has been reported in CAR T-cell therapy because CD19 CAR T as well as blinatumomab are both directed against CD19, the antigen that’s expressed on B-cells. We show that there are multiple types of genetic alteration of CD19 that are important. The first is sequence mutation of CD19 which occurs in a majority of patients who lose expression of CD19. We go on and look at that pathway in some detail; we show that the mutations are often a shower of mutations that affect the epitope for blinatumomab. Importantly, they’re always acquired, they are never present at initial diagnosis or before treatment with blinatumomab commences, they are always induced by exposure to the drug.

In another very interesting and exciting finding we looked very carefully at splicing variants of CD19, particularly looking at the region of CD19 that forms the epitope that’s recognised by blinatumomab. This had been examined for CAR T-cell therapy and it had been shown that one can have splicing that removes exon 2 in its entirety and results in a splice isoform that is now not recognised adequately by CAR T. However, in this study we found no evidence of expression of that isoform that has been seen in CAR T-cell therapy. However, what we do see consistently is a strong association with expression of a different splice isoform that removes part of exon 2, an internal part of exon 2. This same splice isoform is exactly the same that’s seen in every patient, it’s very consistently expressed, and the level of expression is strongly correlated with blinatumomab resistance. So if it’s expressed highly before treatment patients will fail or if its level increases during treatment then that’s associated with treatment failure. So this is a new finding that potentially is of clinical relevance. It’s mechanistically important, it gives insight into why blinatumomab may fail but also it provides a novel biomarker that can be used before or during treatment to understand whether a patient is responding or not.