The 2012 CTRC-AACR San Antonio Breast Cancer Symposium, 4-8 December

Diverse genetic alterations found in triple-negative breast cancers after neoadjuvant chemo

Dr Justin Balko – Vanderbilt-Ingram Cancer Center, Nashville, USA

We’re interested in the 70% of patients that receive neoadjuvant chemotherapy but lack a pathological complete response. So neoadjuvant chemotherapy is used almost routinely now in more advanced stage triple negative breast cancer patients; about 30% of patients have a pathological complete response. Those patients that have that path CR, as we call it, meaning that there is no pathological evidence of disease left in the tumour at the time of surgery following chemotherapy, have a very good prognosis. So those patients, those 30%, have a very good prognosis. However, the majority of patients have some residual disease left at the time of resection and those patients have a very poor prognosis, many of which recur in a matter of years. We don’t have clinically any approved targeted therapies to treat those patients so we know that they’re likely to recur in several years but we don’t offer them routinely more chemotherapy; we already know that the chemotherapy was not 100% effective. So what we did was we performed molecular analyses using state of the art techniques, using both next generation sequencing and digital gene expression analysis on the residual disease specifically of patients who had failed neoadjuvant chemotherapy and by failed we mean they had residual disease at the time of surgery. What we were hoping to find were molecular alterations that would help guide rationally designed adjuvant studies for those patients with residual disease after neoadjuvant chemotherapy.

In summary, we found a very diverse population of lesions and genetic alterations within that residual disease. However, that diversity can be simplified by organising those mutations and alterations, including copy number changes, into pathways that could be targetable in a single clinical trial with a single inhibitor. We have found a number of interesting biological pieces of data from that dataset and one of the most interesting findings, I think, was that we detected previously unreported JAG2 amplifications. The JAG2 gene is involved in interleukin 6, which is a cytokine, through JAG2, through STAT3 signalling that has been shown to be important in something we call cancer stem cell-like behaviour. The 10% of patients that had JAG2 amplifications in their tumour genome performed very poorly so they were almost all deceased in two years after surgery. This is clinically important because we didn’t know these existed before and there are JAG2 inhibitors which are currently in clinical trials right now, several of which have been approved by the FDA for inflammatory diseases. So we would hypothesise that this may be a good clinical area to test those JAG2 inhibitors in triple negative breast cancer for patients that have JAG2 amplifications.

Could this be a way of subdividing patients into different groups?

I think it’s a challenge because of the diversity of the lesions that we identified but, again, part of our goal was to bin or group some of these alterations into single pathways. So, for instance, PI3 kinase mutations or PI3 kinase amplifications or AKT mutations or amplifications can all be organised into a single pathway, the PI3 kinase pathway, and conceivably we would hypothesise that we could group those patients in the course of a clinical study into a single arm of a trial testing PI3 kinase inhibitors in those patients. So I think when we organise them into pathways in groups they can be targeted in the context of a single clinical study that there’s a possibility of offering a mechanism to perform those studies.

Did most of the patients fall into one category?

Definitely a good percentage of the patients unfortunately had aberrations in multiple pathways. Probably about 30-40%, I don’t have those numbers in my head so I don’t want to misquote myself, but there was a significant percentage, maybe I should say that, of patients that did have alterations in multiple pathways. I think that the majority of, or the highest recurring, pathways that we saw altered were the PI3 kinase pathway which was present in about 30-40% of tumours and the cell cycle pathway, so amplifications of cell cycle genes such as cyclin D1, D2, D3, cyclin E1, which may or may not all be targetable with, for instance, CDK 4/6 inhibitors. So those were the two most commonly altered pathways followed by the Ras-MAP kinase pathway, the DNA repair pathway which would include BRCA1 and 2 mutations. Then we also noted a significant group of patients that had focally amplifications of growth factor receptors, so EGFR, c-Kit, MET. So sort of old growth factor receptors that we’ve known have been involved in resistance and oncogenesis in a number of different tumour types but they were so sporadic that it would be difficult to group all of those patients into a single clinical trial. So that will be a bigger challenge, I think.

Did you analyse the tumours before treatment?

We did not present that data here but we are in the process of taking a subset of the 81 patients that we looked at by next generation sequencing and sequencing also the pre-treatment core biopsy, so the diagnostic biopsy. We have some of that data back in a percentage of patients. At this time we have not noticed any distinct discordance between the pre-treatment and the post-treatment but we have not, importantly, quantitatively analysed those, so looking at allele frequencies and selection of the mutant alleles. It’s a very rich dataset and I think we’re going to learn a lot from it so this is the first glance.