A news and analysis article entitled ‘Biomarker search illuminates cancer immune biology,’ published in the July 2016 issue of the Nature Reviews Drug Discovery, emphasises that checkpoint inhibitors are transforming care in oncology, however in an area of such progress, an ideal biomarker to predict response is proving elusive.
While tumour PD-L1 expression was an obvious first candidate, it was found to be an imperfect biomarker. The search has now shifted to other immune players, the tumour microenvironment, genetic factors, etc. The search for better immunotherapy biomarkers is also forcing the reconsideration of how to classify cancers by integrating an immune component, with implications for future combination studies.
The article highlights comments from some prominent researchers in the field of onco-immunology. Prof. Suzanne Topalian emphasised a need to move away from the traditional concept of biomarkers towards a concept of immune system dynamics.
Although responses to checkpoint inhibitors can be durable in melanoma, only about one third of patients respond. Response rates are even lower in non-small-cell lung cancer (NSCLC) and renal cell cancer (approximately 20-25%). Autoimmune side effects associated with checkpoint inhibitors are another reason why careful patient selection is needed.
PD-L1 levels, as measured by immunohistochemistry (IHC), can identify groups of patients that are more likely to respond to PD-1-PD-L1 blockade. However, some patients with high PD-L1 levels do not respond to treatment, and a subset of patients who test negative for PD-L1 expression can derive a substantial benefit. PD-L1 IHC is approved as a companion diagnostic only for pembrolizumab in NSCLC. FDA approved it as a complimentary diagnostic in melanoma and for nivolumab in NSCLC.
The value of this biomarker might be limited by technical pitfalls such as irregular expression levels throughout the tumour and the lack of a single, standardised IHC test. However, a more fundamental limitation is that tumoural expression of PD-L1 does not tell the whole story. According to Prof. Antoni Ribas, focusing on PD-L1 and not looking for T cells is equal to looking at one side of the equation.
Biomarker candidates
Researchers capture information not just on T-cell activity, but also on metabolic and microbiome-related factors, as well as the mutational load of a tumour. Mutations cause tumours to express more neoantigens for the immune system to recognise, which can then recruit a greater pool of activated T cells to the tumours. In a recent phase II study with pembrolizumab in colorectal cancer, patients with mutations causing mismatch repair deficiency had a response rate of 40%, compared with 0% in patients without such mutations. Cancer types with mismatch repair deficiency, including 2-4% of certain gastrointestinal, gynaecological and prostate tumours, may be similarly susceptible to this treatment.
Merck & Co. is looking for RNA signature biomarkers, Working together with NanoString Technologies, they defined a gene expression signature that involves a panel of six immune-related genes, such as interferon-γ, that might identify patients across cancer subtypes who are likely to respond to pembrolizumab.
Pfizer, which is partnering with Merck KGaA on avelumab, is collaborating with Adaptive Biotechnologies to use high-throughput screening to profile patients’ immune cell repertoire and thereby predict response to immunotherapy. They are all pointing to particular aspects of the immune microenvironment.
Immune classification of cancers
Ribas and co-authors from the Netherlands Cancer Institute have recently proposed a ‘cancer immunogram’ that scores patients on seven tumour- or immune-related characteristics to enable a personalised approach to immunotherapy selection.
Recent concept of a ‘tumour immunity continuum’ builds on studies of tumour immune phenotypes as characterised by patient biopsies. Tumours with inflamed phenotype and pre-existing immunity, including high levels of interferon-γ and CD8+ T cells, tend to respond favourably to checkpoint inhibitors. In the middle of the continuum, tumours exhibit inflammatory markers but restrict T cells to the edges of the tumour. At the other end of the spectrum are immunologically ignorant tumours that are not inflamed and lack CD8+ T cells. These tumours, typified by paediatric cancers, are entirely resistant to checkpoint inhibitors. Gene signatures associated with the resistant phenotypes are beginning to emerge.
The academic community has now started to look into immune as well as tumour biology in the classification criteria for cancer; independent of histology and the organ.
Emerging combination opportunities
A partner drug might be selected to first stimulate inflammation and induce expression of interferons in the tumour microenvironment. This puts T cells in place so that they can attack as soon as checkpoint inhibitors are applied.
Merck & Co. is partnering with numerous other companies in more than 170 combination studies involving pembrolizumab. One of the most advanced is a phase III trial in melanoma to test the addition of Amgen’s oncolytic virus therapy T-VEC. Another combines pembrolizumab with Incyte’s investigational IDO-targeted immunotherapy epacadostat.
Chemotherapy agents, whose immunomodulatory properties are becoming increasingly appreciated, are also having a renaissance as partner drugs, increasing the urgent need for suitable biomarkers.