Periodic Reporting for period 1 - EVOMET (The rise and fall of metastatic clones under immune attack)
Reporting period: 2019-05-01 to 2021-04-30
Charles Darwin proposed more than 100 years ago that species evolved by means of natural selection. But selective forces not only act on organisms, they also govern the evolution of the cells in our body. One of these forces is the immune system that, besides protecting us from viruses and bacteria, recognises our own faulty somatic cells and ultimately shapes the emergence of cancer in a process called immunoediting. However, to what extent the genome of our cells undergoes immunoediting, and what are the determinants that predispose us to develop cancer and metastasis remain poorly understood.
Over the last decade, next-generation sequencing (NGS) technologies have allowed us to explore the landscape of the genetic causes of cancer. However, no consistent drivers of metastasis have been identified so far and cancer is still a disease of concern. In parallel, the immune system has gained relevance in the clinical management of cancer thanks to the incredible results obtained in patients treated with immune checkpoint inhibitors (ICIs). ICIs unmask the patient own's immune system to act against malignant cells leading to complete or partial regressions in a substantial number of patients. Nonetheless, many patients fail to respond to the treatment and, to date, we lack scientific knowledge to understand the mechanisms of this resistance. In my Marie-Curie project, I proposed a simple mechanism of evolutionary adaptation driven by the immune-mediated negative selection against growing tumors harbouring immunogenic mutations. This process is intrinsically linked to the primary resistance to treatment in metastatic tumors. Tumors that have been subjected to the selective pressures of a healthy immune system can either be immune-adapted, by depleting all immunogenic mutations, or immune-escaped, by acquiring mechanisms of immune evasion. Therefore, understanding the interaction between the immune system and cancer progression could help us to differentiate between individuals which are sensitive or resistant to immunotherapies, ultimately reducing the overall burden of the current standard of care in the standard setting of clinical practice.
I developed a mathematical model to theoretically demonstrate the process of immunoediting during tumor growth, and implemented a computational method to determine the levels of immune selection acting in primary and metastatic tumors. I demonstrated how an evolutionary metric could differentiate between patients that are going to benefit from immunotherapeutic agents based on the personalized genomic data.
Over the last decade, next-generation sequencing (NGS) technologies have allowed us to explore the landscape of the genetic causes of cancer. However, no consistent drivers of metastasis have been identified so far and cancer is still a disease of concern. In parallel, the immune system has gained relevance in the clinical management of cancer thanks to the incredible results obtained in patients treated with immune checkpoint inhibitors (ICIs). ICIs unmask the patient own's immune system to act against malignant cells leading to complete or partial regressions in a substantial number of patients. Nonetheless, many patients fail to respond to the treatment and, to date, we lack scientific knowledge to understand the mechanisms of this resistance. In my Marie-Curie project, I proposed a simple mechanism of evolutionary adaptation driven by the immune-mediated negative selection against growing tumors harbouring immunogenic mutations. This process is intrinsically linked to the primary resistance to treatment in metastatic tumors. Tumors that have been subjected to the selective pressures of a healthy immune system can either be immune-adapted, by depleting all immunogenic mutations, or immune-escaped, by acquiring mechanisms of immune evasion. Therefore, understanding the interaction between the immune system and cancer progression could help us to differentiate between individuals which are sensitive or resistant to immunotherapies, ultimately reducing the overall burden of the current standard of care in the standard setting of clinical practice.
I developed a mathematical model to theoretically demonstrate the process of immunoediting during tumor growth, and implemented a computational method to determine the levels of immune selection acting in primary and metastatic tumors. I demonstrated how an evolutionary metric could differentiate between patients that are going to benefit from immunotherapeutic agents based on the personalized genomic data.
Here, we developed a stochastic model of cancer evolution that simulates the temporal dynamics of immune dN/dS. Our model predicts that in the absence of immune escape, surviving tumours must adapt to their local immune environment by depleting the pool of nonsynonymous mutations in immune-exposed loci, resulting in immune dN/dS < 1. If an immune-escape mechanism evolves, neoantigens can then accumulate neutrally, asymptotically pushing immune dN/dS towards 1. These insights could have clinical application because predict that immune-adapted tumours will have low immunogenicity and will not respond to ICIs, whereas immune escaped tumours will have accumulated immunogenic variants, i.e. neoantigens, that provide the substrate for an effective immunotherapeutic treatment. To test these predictions, we assessed dN/dS in 8543 untreated primary tumours, as well as in 356 metastatic cancers that received immunotherapy. We demonstrate that immune dN/dS correlates with lymphocytic infiltration in primary tumours that have not evolved immune escape, and that immune dN/dS in combination with escape status can predict immune checkpoint inhibitor response better than tumour mutation burden in metastatic tumours.
In conclusion, our study provides a theoretical framework to understand the evolutionary dynamics of immunoediting during tumour evolution using a simple metric, dN/dS. We demonstrated that immune dN/dS quantifies immune-mediated negative selection and provided theoretical explanations for the lack of immune selection signals observed in current studies. Among these explanations is the improper mixture of tumours originated through different evolutionary paths, especially between two possible outcomes of immunoediting, adaptation or escape. By using immune dN/dS, we conclude that immune adapted patients will not benefit from immunotherapies as they lack neoantigens. In the future, we believe that immune dN/dS, an evolutionary-based metric based solely on genomic data, can be used as a read-out of tumour immunogenicity, and that, in combination with other genomic metrics such as escape status and mutation burden, can be used to predict response to immunotherapy.