The Oncology Podcast including The Oncology Journal Club Podcast by Professor Craig Underhill, Dr Kate Clarke and Professor Christopher Jackson; and Supportive Care Matters by Dr. Bogda Koczwara.
Oncology News and Expert Analysis from a unique Australian viewpoint.
Proudly brought to you by The Oncology Network.
The Oncology Network are producers of digital resources that support busy oncology health professionals. For more information visit our website www.oncologynetwork.com.au.
We also invite Healthcare Professionals to subscribe to The Oncology Newsletter and our Oncology Portal for free and exclusive resources at: www.oncologynetwork.com.au
Our Host Professor Craig Underhill meets Dr Ashray Gunjur as they share insights into the gut microbiome's role in predicting how patients will respond to cancer immunotherapies.
Together, they unearth how the tiny worlds within us might hold the key to unlocking personalised medical approaches that are as unique as our microbial fingerprints.
Ashray's expertise in analysing faecal microbiomes at the strain level using shotgun metagenomic sequencing is not just impressive—it's a game-changer for oncologists and patients alike, offering hope for more targeted, effective treatment regimens.
Hello, I'm Rachael Babin from the Oncology Network. Welcome to the OJC Meets sister series of the Oncology Journal Club podcast. In today's episode, Craig Underhill meets Ashray Gunjur to discuss their recently published paper on the microbiome. The paper explored how the microbiome could be the key to identifying individuals who can benefit from combination immunotherapy across various cancers, including rare cancers. Professor Craig Underhill is a medical oncologist practicing in Aubrey Wodonga. He is a renowned champion of telehealth and teletrials. Dr Ashray Gunjur is a medical oncologist from Melbourne currently undertaking a clinical PhD at the Wellcome Sanger Institute in the UK. We hope you enjoy listening. This is Rachael Babin and this is the OJC Meets.
Craig Underhill:
So it's a great pleasure to welcome a long time fan of the podcast, Dr Ashray Gunjur, who is a medical oncologist training in Melbourne, now working in Cambridge doing his PhD in cancer and the microbiome. So fascinating area. Welcome Ash.
Ashray Gunjur:
G'day, g'day, g'day. Yes, it's great to be on. What an honour.
Craig Underhill:
So this sort of fulfills three dreams Ash, we previously talked about. If anyone had an interesting paper that they wanted to put forward, please let us know. So you've done that. It's a blow, your own trumpet paper, and it's from Nature Medicine.
Craig Underhill:
Now one of our favourite clever type journals to have on the show. So it's great to have this paper, Ashray, titled 'A Gut Microbial Signature for Combination Immune Checkpoint Blockade across Cancer T ypes.' Before we get into the paper, just tell me, before this paper was published, what's the kind of summary of where we're at with the interaction between microbiome and cancer and responses to immunotherapy, because there's been some intriguing work before.
Ashray Gunjur:
Yeah, absolutely. I think intriguing is a great word. I think it's a really exciting time to be studying the microbiome, particularly the gut microbiome in cancer. There's a lot of challenges as well, just with understanding the relationship.
Ashray Gunjur:
I think the field really took off back in early 2018 when really pioneering work out of three groups was published currently in the journal Science.
Ashray Gunjur:
So these three groups one in France and two in the US all studied cohorts of patients with metastatic melanoma and they showed that within those cohorts they could find particular species, particular bugs in the gut microbiome that are associated with response antiPD1 therapy.
Ashray Gunjur:
But I think what was really exciting was that all three groups would show that in taking the stool from patients and transplanting it into mice, they could recapitulate those responses.
Ashray Gunjur:
So you could have the stool from a non responder and give it to a gem free mice and that mouse would not respond to anti pd1 therapy and similarly, still from a responder, would actually induce a response. So it really was encouraging of causal relationship in the potential application of microbiome derived therapeutics. But since then there been a lot of studies and what's been really difficult is to actually deduce what the specific bugs across geographies, across studies, across cancers that are predictive, because when we look at the multitude of clinical co-op said in public, since there have been really different findings across the different studies have been some Overlaps but there have been some inconsistencies. In fact, some studies have shown one bug to be good while another is showing that bug be bad. So we really wanted to understand the cause for these inconsistencies and see if we could improve the generalisability of the microbiome and in being something that might be able to predict patient response in this case to get to immunotherapy or that something you might be able to target to enhance patient response.
Craig Underhill:
Fantastic and it kind of makes sense biologically. We know that microbiome is part of the main immune system and there's cross talk with the adaptive immune system. People have really kind of let the head go and started talking about giving patients antibiotics or not giving them probiotics, or vehicle microbiome transplants. So there's a lot of kind of work and expiration. So then tell us about your study and what we learned.
Ashray Gunjur:
Yeah, absolutely. I guess the setting was this kind of challenge and defining really consistent Got microbial signatures. We call them signatures because it's kind of like a pattern of my crows that in this case is predictive of immune checkpoint inhibitor response. So the challenge of finding generalisable signatures of response from the gut microbiome, given those, those inconsistencies, but knowing that there was promise from those Preclinical studies, is my study showing that you can actually improve response and they started to actually be some human data, vehicle microbial transplants like that and of course the retrospective evidence around antibiotics.
Ashray Gunjur:
So one of our hypothesis was that maybe one of the problems, in addition to the problems that had been previously covered about differences between the gut microbiome across different communities, across different countries, maybe one of the other problems was that actually we weren't looking at a deep enough resolution at what the bugs where, because conventionally we look at bugs at the species level or even a higher taxonomic rank, that the generous level. So we lump all bugs together that are part of the same species in bugs that prokaryotes. So we lump species based on a, I guess, generic threshold, based on how similar the genomes are. So you might have lots of different bugs that we call the same species that do very different things, and we actually know that from the work on pathogens. So when you studied E Coli, you know that some strains of E Coli are completely benign, while others can cause horrendous disease, and the difference might just be one percent of the genome.
Ashray Gunjur:
So we thought, well, maybe that's the case for commensal bugs as well. Maybe the ability to have these consistent signatures would be enhanced if we could look down at a sub species level of strain levels. So that's what we set up to do. So we started off with the great rare cancer study, which you know well, craig. It was a study run out of five different sites across Victoria and New South Wales, including Albury-Wodonga, that's right.
Craig Underhill:
it was one of the first first tele trials done in Australia using the teletrial methodology, so close to our heart for many reasons, gave patients in regional Victoria access to immunotherapy at a time when it wasn't readily available. I'm sorry to interrupt you, but keep going.
Ashray Gunjur:
No, no, yeah, absolutely. It was super exciting. And it was also patients with really rare cancers who are historically underserved, who often lack standard of care treatment because their diseases are just not well studied enough. And it was a basket design. So various different cancers, cancer types, patients with advanced cancers of the gynaecological tract that were rare, neuroendocrine neoplasms or upper GI or biliary tract neoplasms, were all treated with epilimumab and nevolumab in a pretty conventional fashion and actually, amazingly, 25% of patients responded across those three pre-specified tumour cohorts, which was just fantastic, and we saw that those responses were durable. So it was a fantastic study. But I think it was especially fantastic just because of the translational research opportunities that were built in and full credit to you, the other PIs and, of course, Oliver Klein and Jonathan Cebon, the coordinating PIs, who had the foresight of collecting all these samples.
Ashray Gunjur:
So one of the biospecimens that was collected from the majority of patients was a pre-treatment stool sample, so that's what we used to assess the gut microbiome or, in this case, the faecal microbiome. So we extracted the DNA from those samples and we used something called shotgun metagenomic sequencing. So this is where it's kind of the equivalent of whole genome sequencing, of the metagenomes, because there's lots of different species, lots of different genomes in that stool sample but you have a really deep understanding of the whole genomes, of what is present. And then we used some bioinformatics to not only stitch together what the bugs were at the strain level but then also map the reads to those genomes to really get pretty accurate quantifications of how much of the bug was present in each of those pre-treatment stool samples. And then we could actually use it to predict response or progression across these very, very different cancer types. Because we had that information in the follow up and what we found was, when we used the strain level information, we actually saw an improvement in the predictive performance when we cross validated across this very diverse rare cancers cohort. Then when we use them all conventional species or genus level information. So I think it to us really affirmed that hypothesis that having a higher resolution, higher granularity of the microbiome does actually make an improvement and it actually allowed us to cross validate against very different cancer types. So when we train the model just on two of those pre-specified tumour cohorts, we could actually quite reliably predict it on the left out tumour cohort. So this was really exciting to us that not only could we see an improvement in that cross validated performance with the strain resolution. But we could actually see generalisability across cancer types.
Ashray Gunjur:
So then we wanted to take it a bit further and look at public data. So there is fortunately a lot of published data from predominantly patients with metathetic melanoma their pretreatment, gut microbiome subjected to shotgun metagenomic sequencing, so stool samples, shotgun metagenomic, so similar design. Obviously some differences in terms of the specific kits that we use and the sequences that we use. That do introduce some batch effects. But we were able to reanalyse that data in the same way.
Ashray Gunjur:
And we found another interesting thing. We found that the signature from the rare cancers trial, where all patients were treated with IPI-NIVO, so anti-CTLA4, anti-pd1, seemed to only perform well on external cohorts where the patients had received IPI-NIVO, so anti-CTLA4, anti-pd1. So it didn't seem to perform very well. In fact it performed like a coin toss when we tested on patients who just received anti-PD1 therapy. And this was a little bit surprising because we always thought of anti-PD1 as kind of the where-course it's carrying the combination of mean checkpoint blockade efficacy.
Ashray Gunjur:
But I guess when we thought more about it it seemed like there was already evidence that the mechanism of action of combination anti-PD1, anti-c2a4 is quite different and that the addition of anti-C2A4 probably does make a difference, especially in the interaction with the gut microbiome. There's evidence that it can change the permeability of the gut. Obviously, colitis is a much bigger phenomenon with anti-C2A4. In addition, and there was evidence from other sources of data, like immunotranscriptomic data, that the immune responses were quite different to this treatment. So it suggests to us that there is a common signature of response to immune checkpoint blockade across cancer types, but there might be different microbial signatures to different regimens, which we thought was interesting and we wanted to share.
Craig Underhill:
So, in summary, I think it's really important information so that signature of response is predicated on the treatment regimen rather than cancer type, which is the conclusion in the paper, and so it serves an important knowledge base for further trials looking at microbiome diagnostics or any interventions. So is that a valid conclusion, do you think?
Ashray Gunjur:
that? Yeah, absolutely, I think you're exactly right. I think this is a foundation, but maybe future studies could unpick this further and focus on that split in different immune checkpoint blockade regimens, because historically we were sort of lumping them together Antipyton monotherapy and combo-antipyton, anti-c2a4, they were both just immune checkpoint blockade and now obviously we're going to be studying combination IO chemo for a lot of cancer patients, so that's possibly a different kettle of fish as well. So I think that this regimen is a really important covariate. Obviously, cancer type is too, but we were actually reassured that there were quite similar signatures across different cancers in terms of what defined or what predicted a future responder versus someone who unfortunately wouldn't respond Fantastic.
Craig Underhill:
So if you look really intriguing. We've had other papers over the years about microbiome. I think it's a really fascinating area. It might be even worth an episode down the track summarizing the whole area, but I think intriguing is the word of the day. Let's hope that this knowledge will then lead to tools that we can predict people's response and maybe alter the microbiome if they are unlikely to be responders. A really fascinating area, ash. You made it sound really simple, straightforward and it's not like the technical stuff in the paper is really quite high level. So a fantastic paper Well done. Congratulations on all the work you're doing.
Ashray Gunjur:
It was a real team effort and we couldn't have done it without you and all the sites you're creating to this great trial. So thank you.
Craig Underhill:
Yeah, thank you, ash. All the best for finishing your PhD and let's hope we see you back in Australia in the near future.
Ashray Gunjur:
Yeah, I hope so too. Thanks so much, Craig. Thank you so much for having me on.
Rachael Babin:
You've been listening to the OJC Meets proudly produced by the Oncology Network. For regular news and podcast updates, we invite healthcare professionals to join us at the Oncology Network website, oncologynetwork. com. au. Your free registration includes a free subscription to our weekly publication, The Oncology Newsletter, and it's a fantastic way to support the OJC. This is Rachael Babin and this is the OJC Meets podcast.
