Nature Biotechnology published a thought-provoking piece addressing critical questions in the development of microbiome-based solutions focused on disease interception. Co-authored by over a dozen key opinion leaders in the microbiome field, the article summarizes the main outcomes from Microbiome Futures, an effort to create a roadmap for translation in the microbiome space. The Microbiome Futures project was spearheaded by Gaspar Taroncher-Oldenburg, who currently serves as a consultant with Janssen Research & Development, LLC and its Janssen Human Microbiome Institute.
This BioVox article was guest authored by Dirk Gevers, Global Head of the Janssen Human Microbiome Institute, Janssen Research & Development, LLC.
As stated in the Nature article: the potential for innovative disease interception in the microbiome field is matched only by the difficulties faced in trying to develop them. Although the interest in microbiome products is massive, no Phase III clinical trials for microbiome therapeutics have yet been completed.
According to Microbiome Futures, among the main factors contributing to this dearth of products on the market are the novelty of this modality—it takes time for a new product type to pave a path through unknown territory—and an increasing need to determine mechanisms of action (MoAs) before going into the clinic, a controversial aspect given different definitions of what constitutes sufficient evidence for translation. It is within this framework that the authors outline a roadmap for microbiome translation with the following action points:
Defining host-microbiome interactions to understand if, how and when to intervene.
Thinking outside the gut: expanding the focus to the lungs, oral cavity and the skin, despite accessibility and/or density challenges associated with these areas.
Looking at microbial communities more broadly: in addition to bacteria, we need to consider the virome, mycobiome, and others.
Increasing efforts to integrate data, to gain a more holistic view of the microbiome.
Generating robust translational animal models for microbiome research.
Developing clear pharmacokinetic and pharmacodynamic metrics for microbiome-based therapeutics.
To me, the question of mechanism is indeed a central one in the microbiome field when it comes to understanding microbiome dynamics and microbiome-host interactions. However, I also believe that when it comes to the development of actual microbiome-based products, identifying robust in human insights that are reproducible and predictable is equally if not more valuable and sufficient to initiate translation, even without a clear MoA. Especially in human intervention studies, targeting the microbiome can provide a wealth of information, that provides data that then can be mined for its active component. In certain cases, this could be brought back all the way to a specific metabolite, in other cases it will remain more difficult to reduce the complex host-microbe interactions to the conventional silver bullet paradigm.
The novelty of the field also raises challenges when it comes to regulating new drugs: What are the endpoints and how do we measure them? How do we adapt ‘traditional’ measures of drug performance, such as pharmacokinetics and pharmacodynamics, to live bacterial product-based interventions? Along with others in the field, we are thinking very hard about tricky questions such as these, and how best we can help inform the regulatory process to work with microbiome solutions while still maintaining the current high standards of safety and efficacy. Over the past year, for example, we have developed and rolled out three community challenges designed to foster collaboration among academia, industry, and government organizations to advance translation in the microbiome field. The challenges aimed at driving innovation and provide a platform to compare the performance of microbiome tools and protocols globally, particularly for the purpose of how we best measure drug performance.
Modelling future solutions
Concerning the lack of good models for pre-clinical trials, I feel it’s true that a lot of the extant animal models fall short when attempting to address microbiome-related diseases. The vast majority of animal models are imperfect for studying diseases in general, let alone for working through the complex microbe and host interactions that need to be identified in microbiome research.
I agree with the authors that better models can and should be developed, but in the meanwhile my Janssen colleagues and I think we should maximally leverage in human studies. Conducting microbiome research with longitudinal studies, with the initial experimental interventions done in human, can be considered as more valuable to progress in our knowledge of microbiome function in a clinically relevant way. What’s more: we are able to conduct these types of trials safely, as most microbiome interventions use microbes already a part of a healthy human system.
Data integration, in particular vertical and multi-omic association analyses, is another area I am keenly interested in. As the authors say: understanding the function of our microbiome, not just the composition, is key to developing targeted microbiome solutions to healthcare issues. To do this, we need to engage analysts, modelers and biomathematicians in the early design stages of research projects.
In my opinion, for a Microbiome Future to materialize, stakeholders—the biopharmaceutical industry, startups and academics—will need to work together to create collaborative platforms. When researchers can develop, improve, and share data and methods we will be able to facilitate the translation of information into valuable insights and progress microbiome science. Ultimately, it is only through collaboration that we will be able to advance transformational approaches to treat, prevent, and even eliminate diseases in the future.
Read the Nature Biotech ‘Translating Microbiome Futures’ article here.
Image provided by Global Engage.