Some subtypes of gastric/oesophageal adenocarcinoma responsive to immunotherapy
Interview with Oncode Investigator Sarah Derks
This interview was first published in the magazine Immunoncologie.
Author: Marten Dooper
Gastric and oesophageal adenocarcinomas are a heterogeneous group, so it should come as no surprise that not all subtypes of gastric and oesophageal adenocarcinoma respond equally well to immunotherapy. Tumours that test positive for Epstein-Barr virus and tumours with microsatellite instability are particularly responsive to immune checkpoint inhibitors. Medical oncologist and Oncode Investigator Dr Sarah Derks (Amsterdam UMC, VUmc site) explains.
Like colon cancer, gastric/oesophageal adenocarcinomas can be classified into molecular subtypes, Derks begins. ‘The Cancer Genome Atlas Network identified the precise molecular properties of just under 300 primary gastric/oesophageal adenocarcinomas in 2014. Based on that data, the Network concluded that these tumours can be classified into four molecular subgroups: tumours positive for Epstein-Barr virus (EBV), tumours with microsatellite instability (MSI-high), genome-stable tumours, and tumours with chromosomal instability.1 This classification is important when considering the responsiveness of gastric/oesophageal adenocarcinomas to different types of treatment, including immunotherapy.’
To get straight to the point, the first two subtypes (EBV-positive and MSI-high tumours) are very immunogenic, in other words predisposed to be responsive to immunotherapy, whereas the other two are not. ‘MSI-high tumours account for about fifteen percent of primary gastric and oesophageal adenocarcinomas and about five percent of metastasised gastric and oesophageal adenocarcinomas. In a metastasised setting, a few clinical studies have now been conducted using PD1/PD-L1 checkpoint inhibitors. The Phase 3 KEYNOTE-062 study, the results of which were recently presented at the ASCO 2019 Conference and the ESMO 2019 Conference, is particularly important: it shows not only that front-line treatment of patients with metastasised or locally advanced gastric or oesophageal adenocarcinoma using the PD1 inhibitor pembrolizumab is non-inferior to chemotherapy, but also that pembrolizumab treatment in patients with an MSI-high subtype tumour even results in improved survival.2,3 Conversely, these MSI-high tumours respond less well to chemotherapy on average.
For that reason, since the results of the KEYNOTE-062 were published, there have been strong calls to determine the MSI status of all gastric and oesophageal adenocarcinomas. Front-line treatment of patients with metastasised gastric or oesophageal adenocarcinoma using pembrolizumab has not yet been approved by the FDA and EMA, however. As a result, for the time being we can only provide this treatment in clinical practice as part of a clinical study (the DRUP study in the Netherlands, for example, but at a later stage of treatment).’
Full of TILs
In addition, EBV-positive gastric and oesophageal adenocarcinomas – about ten percent of primary gastric and oesophageal adenocarcinomas – also respond very well to pembrolizumab treatment, as shown by a Phase 2 study in which 61 Korean patients with metastasised gastric or oesophageal adenocarcinoma were treated with pembrolizumab in a second or third-line setting. The total response rate among the patients with an MSI-high tumour was 85.7%, and even 100% among the patients with an EBV-positive tumour.4 Although the number of patients was small in this study, the results are promising.
It is not strange that these two subtypes respond so well to treatment with a checkpoint inhibitor, Derks explains. ‘We know from our own research that these two tumour subtypes are chock full of TILs, tumour-infiltrating lymphocytes. The presence of these cells is known to be correlated with a good response to immunotherapy. MSI-high tumours have a high tumour mutational load, resulting in the production of large numbers of neoantigens, which are recognised by cytotoxic cells. In EBV-positive tumours the presence of the virus in the tumour cells causes cytotoxic T cells to be attracted: these recognise the virus and therefore attack the tumour cells. So in this case the immune system does not target a neoantigen produced by the tumour but a ‘standard’ antigen produced by a virus. We showed a few years ago that these EBV-positive tumours, without exception, have very high PDL-1 expression. EBV-positive tumours also have a high interferon gamma signature. Both the high PDL-1 expression and the high interferon gamma signature are correlated with a good response to treatment with a PD1/PD-L1 checkpoint inhibitor.’5
Despite this good development, these two subtypes represent only a small proportion of gastric and oesophageal adenocarcinomas as a whole, about a quarter of the primary tumours. Derks goes on: ‘The majority of gastric and oesophageal adenocarcinomas fall into the chromosomally unstable or genome-stable subtypes, which respond poorly if at all to treatment with checkpoint inhibitors. The strange thing, however, is that these tumours develop in an environment of chronic inflammation, for example Barrett’s oesophagus or gastritis – in the presence of very large numbers of immune cells, in other words. Consequently, these tumour cells have probably developed mechanisms to keep these immune cells at bay.’6
Rationale for immunotherapy
Derks was granted a Veni grant from the Dutch Research Council in 2017 to find out more about these tumours and the mechanisms that they use to counter the immune response. ‘We see that these tumour subtypes are also heterogeneous. Some do contain TILs, whereas others are complete immunological deserts. We also find large numbers of M2 type macrophages in these tumours, in other words macrophages that suppress the immune response. In addition, we showed a few years ago that certainly in the oesophageal adenocarcinomas there is high PD-L2 expression. That is probably due to the presence of large quantities of interleukin 4 and interleukin 13, cytokines that are characteristic of the presence of an immunosuppressive infiltrate in the tumour.6 We’re now looking for ways of influencing that immunosuppressive infiltrate to make it immunoactivating, so as to restore the cytotoxic response to the tumour cells – for example by targeting the M2 macrophages specifically. It also appears that the tumour cells themselves give off signals that make the environment immunosuppressive. If we can find out more about this, it may provide pointers to treatment that produces an immunostimulating environment. So, to sum up, there is a rationale for immunotherapy with these subtypes of gastric and oesophageal adenocarcinoma too. This will require a different approach, however, not just blocking the PD1/PD-L1 checkpoint.’
The role of immunotherapy in treating gastric and oesophageal adenocarcinomas in clinical practice (in the Netherlands) is very limited for the time being, Derks concludes. ‘Given the absence of EMA approval, we cannot treat the MSI and EBV-positive group with immunotherapy outside a trial setting. Let us hope that the EMA will soon make a decision, followed by approval for the Netherlands by the BOM committee on oncological drugs.’
- TCGA Network. Nature 2014, 513: 202-209
- Tabernero J, et al. ASCO 2019 : abstr LBA4007
- Shitara K, et al. ESMO 2019: abstr LBA44
- Kim ST, et al. Nat Med 2018; 24: 1449-1458.
- Derks S, et al. Oncotarget 2016; 7: 32925-32932.