As you may recall, we actually used to call, name, colorectal cancer colorectal cancer which very much now is a wrong name because over the years we’ve started to realise that colorectal cancer is not one disease, it’s actually different diseases and we can very much divide it into subgroups of cancer. It’s differences of biology. That is also important because the biology is different and also now we have different treatment options that we can use against each subgroup.
One of the very important subgroups of colorectal cancer is what we call an MSI high tumour, or mismatch repair deficient tumour. That’s obviously a very important subgroup because of the role of immunotherapy in the metastatic setting. We know now that these tumours, most of them respond very well to a checkpoint inhibitor. But among those MSI high or mismatch repair deficient colorectal cancer tumours there is a subgroup who also carry a BRAF-V600E mutation, so you have both coexistence of MSI high plus BRAF-V600E. Again, that’s important because for BRAF-V600E now we have targeted therapies we can use against this tumour which are anti-EGFR plus anti-BRAF or BRAF inhibitor drugs.
So the question in our minds was how BRAF-V600E MSI high tumours compare to BRAF wildtype MSI high tumours in terms of the immune microenvironment. Trying to figure out if they are the same or different and trying to see if there is any hint that perhaps one may respond to immunotherapy or may be more immunogenic than the other.
In this study we looked almost at 459 patients with tumours with colorectal cancer. All of them had MSI high or MMR tumours, so all tumours were MSI highs, and BRAF status, BRCA mutation status, was also known. All these tumours were also RAS wildtype. At the end what we found was that about 336 tumours were MSI high BRAF wildtype but there were about 123 tumours that were MSI high and BRAF mutated which was very much the cohort we are now focussed on.
In terms of clinical characteristics, in the BRAF mutant cohort the median age was about 76 which was older than the group with BRAF wildtype which was about 62. Also 84% of patients were women in the BRAF mutation group.
The first observation was to see that the co-mutations, the mutations that happen with BRAF mutation V600E are a little bit different from those with wildtype, particularly in MSH6, P2M, ATM and MSH2 were more frequently seen in BRAF mutated MSI high tumours compared to those BRAF wildtype MSI high tumours. But when we looked at the TMB high, which in our cohort was defined at a level of 10, actually 100%, all of the BRAF mutant MSI high tumours, carried high TMB while 95% of the BRAF wildtype carried high TMB. So, yes, the majority of them but not 100% like the mutated arm. But the neoantigen mutation burden was the same in both groups, it was 15% exactly in the BRAF wildtype and 15% in the BRAF mutated group.
When we looked at the microenvironment, the immune microenvironment, and compared the BRAF mutant versus BRAF wildtype, the proportion of natural killer cells was significantly higher in the BRAF mutated cohort, it was 20%, 21%, versus 15%. But there was no difference, or no significant difference, observed in terms of CD4 and CD8 expression; CD4 and CD8 expression was not different. Then also we observed that the cancer stem cell pathway was significantly downregulated in the BRAF mutated tumours. On the other hand, there was significant upregulation of five pathways that were associated with growth and metabolic reprogramming of BRAF mutated. When we looked at the genetic analysis of immune pathways there was limited difference in the tumour inflammation between the BRAF mutant and BRAF wildtype MSI high cohorts.
So the picture that we were able to observe is that it appears that among those tumours with MSI high status, whether they carry a BRAF mutation or they have the wildtype status, there is no significant difference in the immune microenvironment which suggests that all of them should be equally sensitive and equally respond to checkpoint inhibitors. This was very much what was seen in the CheckMate 177 when they did the forest plot analysis, subgroup analysis. They found that both wildtype and mutated tumours, tumours with BRAF, were very much on that side of the graph, indicating that both responded.
This was a different story from what we actually observed within the RAS mutations. So, as you know, about 50% of colorectal cancer tumours carry a KRAS or RAS mutation. We did very much the same analysis on a similar group with MSI high tumours and looking at KRAS mutant versus KRAS wildtype. The neoantigen mutation burden was significantly higher in the RAS wild tumour compared to those with RAS mutant tumours. It was 16 compared to 12, so it was different and this was statistically significant. But there was no difference actually in the TMB high status among RAS mutant and RAS wildtype MSI high tumours.
There was also some suggest that the tumour inflammation was downregulated in the RAS mutant tumour compared to RAS wildtype. This is important because, yes, it appears that MSI high tumours that carry also KRAS mutation they do respond to checkpoint inhibitors but at least from our data it appears that perhaps RAS wild tumour, MSI high, have more immunogenic phenomena compared to RAS mutations. So RAS mutated tumours still respond to immunotherapy but sound like the immune microenvironment is not as those ones with RAS wildtype.
This was very exciting but this is very much scratching the surface of what we really need to learn about this disease and how can one continue to better understand the biology and differences among those tumours. But also, more importantly, try to come up with therapy to target each subgroup of patients and avoid and overcome the primary and secondary resistance.