DUBLIN – Has the era of truly personalized cancer immunotherapy begun? Biontech AG certainly thinks so – and the German biotech firm has published in vivo proof-of-concept data in the April 23, 2015, issue of Nature to back up its claims.
The Mainz-based company has obtained what it described as "potent tumor control and complete rejection of established aggressively growing tumors" in two murine cancer models, using RNA vaccines encoding 10 immunogenic neo-epitopes identified by surveying the mutational landscape – or mutanome – of each individual cancer.
A phase I study based on the same concept is already under way in 15 melanoma patients. "At the moment, seven patients have received treatment," Biontech co-founder and CEO Ugur Sahin, who is lead author on the Nature paper, told BioWorld Today. "We will have completed treatment by the end of this year."
Sahin, who also holds an academic post at the Johannes Gutenberg University, in Mainz, has been developing this approach for some time. His group published an influential study three years ago, in the March 1, 2012, issue of Cancer Research, which set out basic contours of the approach.
"We were the first group worldwide suggesting this approach, using next-generation sequencing and vaccination," he said.
It involves whole-exome sequencing of tumor biopsies and samples taken from healthy tissue, to identify tumor-specific mutations. The company has developed a bioinformatics-based method for identifying the likely immunogenic mutations within a given cancer based on two parameters – their binding affinity for major histocompatibility complex (MHC) class II molecules and their level of expression in the tumor microenvironment, based on mRNA expression profiling.
"When we started this approach, most people were extremely skeptical," Sahin said. But his stance has been vindicated by accumulating evidence from other groups, which links the presence of a pre-existing T-cell response to a likely response to checkpoint inhibitor blockade. "It is very clear that about 20 percent of patients respond," Sahin said. "The question is what are the rules? Why do 80 percent of patients not respond and 20 percent do respond?"
The current paper contains two significant findings that challenge existing dogma on the immunogenicity of tumor mutations. It has generally been thought, Sahin said, that only about one in 200 cancer mutations is immunogenic, a frequency that would limit Biontech's approach to a narrow set of cancers with a high frequency of mutations, such as melanoma or smoking-associated lung cancers.
"Our paper shows that the rate of immunogenicity is much higher," Sahin said. His group estimated that about one in 20 mutations elicit an immune response, meaning that the approach could cover more than 75 percent of cancer types.
"The main reason for the underestimate is almost all researchers looked for spontaneous immune responses," he said. Many may have been masked by tumor-induced immune suppression. "We asked questions about vaccine-relevant immunogenicity."
The new research also shows that those tumor-associated mutations that are immunogenic generally drive a CD4 helper T-cell response, rather than a CD8 killer T-cell response. "If you take 10 immunogenic mutations, nine are recognized by CD4 T-cells – only one is recognized by CD8 T-cells," Sahin said. The latter have generally received more attention up to now, he said. "Most tumors are infiltrated by both CD4 and CD8 cells but the number of CD8 cells is usually higher," he said. However, this neglects the biological importance of CD4 T-cells as orchestrators of the immune response. "You cannot determine the relevance just by counting them."
The CD4 T-cell response alters the tumor microenvironment, leading to CD8 T-cell infiltration, the elimination or reduction of immunosuppressive regulatory T-cells. The approach was successful in the B16F10 melanoma model, which is considered very aggressive, Sahin said. "Most checkpoint inhibitor approaches fail in this cell line." It also worked in the CT26 model of colon cancer with lung metastasis, which usually results in death 20 days after inoculation.
Translating this approach into a clinical context is already under way – an interim data analysis will provide the first clues about whether it's working. Optimizing the technology for scale-up is a work in progress, but developing individualized RNA-based vaccines is not the same challenge as delivering autologous cell therapies, Sahin sad. Biontech only needs a sample from a standard cancer biopsy and a blood sample.
"We do not need to build a complex infrastructure at the study site," he said.
Nor does the company need to treat every single vaccine as a single molecule, from a regulatory perspective. "We are not seeking approval for 100 molecules," Sahin said. The company is also engaged in dialogue with regulators. Sahin, together with several other academic and industry researchers and German regulators, published a position paper – titled "The regulatory landscape for actively personalized cancer immunotherapies" – in the October 2013 issue of Nature Biotechnology, which set out a development strategy based on the existing regulatory framework. The approach is similar to that used in regulating cell therapy.
"You keep the process stable and do some sort of quality control on the product – that's what we are doing," he said.
Several others are developing personalized cancer vaccines based on slightly different approaches. For example, Beatriz Carreno, of Washington University in St. Louis, and colleagues, recently reported on the treatment of three advanced melanoma patients with an autologous dendritic cell vaccine, which had been exposed ex vivo to seven immunogenic peptides identified through cancer genomics. (See BioWorld Today, April 2, 2014.)
The Nature paper included co-authors from TRON Translational Oncology at the University Medical Center of Johannes Gutenberg University, the Research Center for Immunotherapy, in Mainz, and La Jolla Institute for Allergy and Immunology, in La Jolla, Calif.