What I presented is actually the teamwork of approximately eighty people that at the beginning of 2011 started to harness the preclinical model of patient derived xenograft, that is putting pieces of patient tumour into mice, in order to understand why in colon cancer the majority of patients are resistant to what at the time was the standard treatment with targeted therapy, i.e. cetuximab which is a monoclonal antibody against the EGFR gene. EGFR is the epithelial growth factor receptor, it’s one of the most important genes for the survival of normal and tumour cells.

The funny thing in colon cancer is that this gene is never mutated or amplified or translocated therefore it’s not an onco-driver but still was the mainstay of biological therapy. What we did was in this huge colony of patient derived tumours which are really avatars of real patients, and we have a thousand now of them, the reason we wanted so many is because we were interested in finding molecular aberrations that may be present in the population at very low frequency, like 1%, 2%. Because precision medicine is this, precision medicine stands on exception so we wanted to see the exception. We used cetuximab before doing a trial to identify the patients that were not responding to cetuximab in the hope that they had some of these new aberrations. Lo and behold, the first one we found was HER2.

HER2 is a cousin of EGFR and it’s a well-known oncogene because it’s found in breast cancer, in approximately 15-20% of breast cancers, and in approximately 10% of gastric cancer. In colon it’s much less, it’s 3%, but still 3% of a large, one of the second most frequent tumours, it’s a lot of patients. Just to give you an idea, in 2030 there are going to be 2.5 million patients with colon cancer, 78,000 of them will have HER2 so it’s still a large number.

So we found that HER2 was amplified, meaning that there are many copies of the gene. We also found that to really be addicted for the tumour, that if you shut down the HER2 the tumour will shrink at least in the PDX in the avatars, you need to have many copies. We eventually in the clinic found out that the threshold for being addicted is around ten copies of the gene compared to one, obviously. So once we found that what we did then was to do the preclinical trial in the seven avatars in which we found the HER2 amplification. The interesting thing which we would had missed if we hadn’t gone into the preclinical area was that in order to shut down HER2 and induce tumour shrinkage response in colon you need to do a combination of lapatinib, which is a small molecule that inhibits both HER2 then EGFR, which is also close to HER1, so the two cousins, and trastuzumab which is a monoclonal antibody against HER2. That was unusual because in breast you just need one of the two; it works better if there are two together but they still work as a monotherapy, not in colon. This is interesting, this is what we are finding with targeted therapy, that the same aberration, molecular aberration, a mutation or an amplification, the same one in the same gene behaves differently in different contexts. What happens in breast doesn’t necessarily happen in colon or in gastric, even if the same gene is involved. So this underlines the beauty of using, as we did, the mice avatars as potential patients.

After having observed that we then went to the clinic. First we established a specific diagnostic criteria for defining HER2 positivity in colon cancer and then we treated patients that were positive with a combination of lapatinib and trastuzumab. We treated 33 patients, had 70% of patients, 23, that had a clinical benefit with a tumour shrinkage of some kind for a certain time of observation. Ten of these patients, i.e. 30%, had an objective response by RECIST meaning that their tumour dropped more than 30%. Actually in the majority of cases it almost dropped to zero; in two cases we had complete response. In one of these two cases this lady is still alive and without evidence of disease 36 months after having started treatment. This is very exciting in colon cancer because, just to give you a comparison, if you use chemotherapy in these patients that were at the fourth, fifth, even eleventh re-treatment you get less than 5% and we got 30%. If you use immunotherapy in another niche of patients which are the patients that have microsatellite instability, which again only 3% have the HER2, you get again 35-40% response.
So suddenly we have found something; nobody has ever treated colon cancer with targeted treatment alone and we have established that you can do that and you can have really exciting results.