07 Mar Sanger Institute Developing Microbiome-Based Biomarkers To Predict Response to Cancer Immunotherapy

MedicalResearch.com Interview with:

Dr. Gunjur

Dr Ashray Gunjur
MBBS (Hons), B. Med Sci, MPHTM FRACP
Clinical Research Training Fellow
Melbourne, Australia

MedicalResearch.com: What is the background for this study?

Response: As background, the last ~5 years have seen a surge of interest in the relationship between gut microbiota and cancer response to immune checkpoint blockade (ICB). We know that though a fraction of many different cancer types will respond to these therapies, it is currently very hard to predict who that will be- so ‘microbiome’ based biomarkers to select patients, or even strategies to change a patient’s microbiome to enhance their chance of responding, are very attractive.

A key challenge, however, has been a lack of consistency in the microbes associated with response or non-response across different studies from different regions. While geographic, methodological, and technical variation likely contribute to this, most studies examined the gut microbiome at a genus- or species- taxonomic rank level, while we know there is significant intra-species (strain-level) diversity. As such, one of our key research questions was whether we could improve the reproducibility of microbial ‘signatures’ of response across cohorts using higher resolution approaches- with our hypothesis being that strain-resolution signatures would outperform species- or lower resolution signatures.

We obtained our signature by analysing baseline faecal samples from the CA209-538 clinical trial, a wonderful investigator-initiated study sponsored by the Olivia Newton-John Cancer Research Institute (Melbourne, Australia). I was fortunate enough to work on this trial as a clinical investigator while training to be a medical oncologist.

MedicalResearch.com: What are the main findings?

Response:  The CA209-538 clinical trial recruited patients with very diverse, rare cancer types, with very different treatment histories, to all receive combination ipilimumab (anti-CTLA-4) and nivolumab (anti-PD-1). Amazingly, 25% responded in all 3 pre-specified tumour sub-groups, and these responses were very durable, translating into prolonged overall survival for patients.

I am now doing a PhD at the Wellcome Sanger Institute, a world-leading genomics research centre in the UK. As part of my PhD, we used deep shotgun metagenomic sequencing of these faecal samples to precisely quantify the relative abundance of different microbes, down to the intra-species (strain) level- including for many microbes that have never been cultivated (and some, not even named) before. We then trained machine learning models to learn a ‘signature’ of response or progression from microbiome abundance data, and found that, indeed, strain-level signatures outperformed species- or lower resolution microbiome signatures, and cross-validated across cancer types, validating our initial hypothesis.

However, on assessing the external validity of the CA209-538 strain-response signature using published cohorts of patients with melanoma treated with ICB, we found that while performance was good/fair for 3 of 4 cohorts receiving ‘combination ICB’ (as for CA209-538 patients), performance was universally poor in all cohorts treated with anti-PD-1 ICB alone. The reverse was true, with signatures trained using gut microbiome data from anti-PD-1 recipients extrapolating better to other anti-PD-1 recipient cohorts, and not combination ICB cohorts. On reflection, we feel this reflects the fundamentally different mechanism of action of anti-PD-1 versus combination anti-PD-1 and anti-CTLA-4 immunotherapy. Interestingly, when we examine the CA209-538 strain-response signature, the ‘beneficial’ taxa (whose abundance is associated with response predictions) were in line with those reported for anti-PD-1 monotherapy previously, while the ‘deleterious’ taxa were novel, which may imply that the distinction lies in these negative taxa.

MedicalResearch.com: What should readers take away from your report?

Response: Firstly, I think it highlights the importance of intra-species, strain-level variation when considering the immunomodulatory properties of bacteria; a phenomenon well known for pathogens, but less explored for ‘commensals’. We provide a roadmap to incorporate study-specific, strain-level analyses from shotgun metagenomic sequencing data that others could apply in the future.

Secondly, the consistent signatures within the CA209-538 trial (where patients had very different tumour types, different prior therapies, etc) but distinct signatures of combination vs. monotherapy ICB suggests that future development of microbiome-based biomarkers or adjunctive therapies may be better tailored to immunotherapy type, rather than cancer of origin. 

MedicalResearch.com: What recommendations do you have for future research as a results of this study?

Response: I think it is important that future work considers the ICB regimen of cancer patients rather than lumping together very different treatments (for example, single agent immunotherapy, combination immunotherapy or combination chemo-immunotherapy). Larger works will be important to validate our finding!

Furthermore, I think it is very important to appreciate strain-level variation and isolate and study exact strains implicated- difficult work but very important to get down to the mechanism for their influence on immunotherapy response!

For full author disclosures, please see the publication.

Citation: Gunjur, A., Shao, Y., Rozday, T., et al.: A gut microbial signature for combination immune checkpoint blockade across cancer types. Nat Med (2024). DOI: 10.1038/s41591-024-02823-z, https://www.nature.com/articles/s41591-024-02823-z

The information on MedicalResearch.com is provided for educational purposes only, and is in no way intended to diagnose, cure, or treat any medical or other condition.

Some links may be sponsored. Products are not endorsed.

Always seek the advice of your physician or other qualified health and ask your doctor any questions you may have regarding a medical condition. In addition to all other limitations and disclaimers in this agreement, service provider and its third party providers disclaim any liability or loss in connection with the content provided on this website.

Last Updated on March 7, 2024 by Marie Benz MD FAAD