Thank you very much. These are my disclosures. The intestinal microbiome harbours the highest density of bacteria in any location on the planet, that is the mammalian gut. We have as many bacterial cells living in and on our body as we have human cells so we are actually ecosystems walking around. The surface of the GI tract, if you were to spread it out, is actually the size of two tennis courts and the intestine is also the largest lymphoid organ in the body by raw numbers of lymphocytes. So we have this incredible meeting of the immune system and enormous numbers of bacteria separated by just one thin layer of epithelium. So our immune systems are really maintaining homeostasis or an equilibrium balance with the bacteria that we coevolved with. You can interpret many immune reactions as perturbations from that equilibrium.

After bone marrow transplantation for leukaemia or lymphoma there are four main causes of death – relapse, graft versus host disease where the donor immune system attacks the patient’s body, infections and organ toxicities.  Each of these outcomes have been associated with the composition of the gut microbiota. Here’s a phylogenetic tree of different gut bacteria that we have found in allotransplant patients and those that are associated with good outcomes are colour coded in blue and those associated with negative outcomes are colour coded in red. We’ve found associations in the past with overall survival, with GVHD, with infections, with organ toxicity and even with relapse. This represents ten years of work approximately in both mouse models and human studies.

Today I’d like to tell you about a different time of measuring or characterising the contents of the gut microbiota which is just before the transplant. Most of this work that I’m showing you here in the last ten years was done studying the microbiota in the few weeks after the transplantation. This project was done in close collaboration with colleagues at Duke University in North Carolina, Hokkaido in Japan and Regensburg in Germany and Memorial Sloan Kettering, our institution in New York.

Here, this is a Tisney projection of several thousand stool samples. Each dot is a stool sample and its composition, the bacteria that are found in it, are represented by where it is in this cluster of dots. So similar samples are close together and dissimilar samples are far apart. These are pre-transplant samples collected from patients at these four centres and they’re colour-coded by which institution they came from. What’s striking here is that the dots are all mixed in together, it’s not as if, say, all the green dots are in one corner and all the red dots are in another. So this indicates that no matter where you live around the world, when you come for a bone marrow transplant the composition of your gut flora is comparable from centre to centre.

Now, what happens during the transplantation? This is the diversity which counts basically how many different bugs you have in your gut. During the transplant at all four centres there is a striking drop in diversity. This is a major insult to the health of the flora that’s not seen in many other clinical scenarios. In the past we’ve shown that low diversity here at this time when the stem cells are engrafting is associated with poor outcomes after transplantation; we’ve shown that in a single centre study.

We asked here, what if we look earlier in the transplant course, before the stem cells are administered, before the pre-transplant chemotherapy is administered, at the time when treatment decisions are being made. What we found is that even at this early time point where patients are in this diversity plot can predict their outcome so that patients with low diversity pre-transplant have a poorer overall survival than patients with a high diversity after transplantation. The implication is if we could come up with a way to remediate microbiome injury there might be time to implement it before the transplant.

Another way to measure a gut microbiome injury besides calculating the diversity is looking at a phenomenon that we find in our patients that we term monodomination where instead of hundreds of different bacteria growing in the gut a third, or in some cases 95% , of the bacteria in an individual’s intestine are all the exact same strain. Here we define monodomination as any time that one bacterium accounts for 30% or more of the gut composition which is a very unusual scenario in people walking around, eating a normal diet and not being exposed to broad spectrum antibiotics. We can see that the incidence and the prevalence of this monodomination phenotype is strikingly similar at all four centres. Even though each of these centres uses different antibiotic strategies – the people in Japan have a diet that’s different to the people in the United States – but the patterns of microbiome injury are nearly superimposable. This here shows which bacteria account for the different monodomination events.

I’ll point out to you that studies from animals suggest that the T-cells which cause graft versus host disease that have just been infused migrate to the gut as early as day two or three after the transplant. By that time already 50% of patients have experienced a monodomination event in their gut so the milieu, the environment, that the donor immune system is encountering as it comes into the recipient patients is already seeing a damaged gut.

What about the post-transplant setting? Here we find that also, as in the pre-transplant setting, after the patient has recovered from the first phase of the bone marrow transplant and the donor cells have set up shop and started making their own new white blood cells, we call this the peri-neutrophil engraftment period, again the diversity at the peri-neutrophil engraftment period is predictive of survival. Here we see this reproducibly at three different centres.

So where can we go from here? There are different strategies to try to remediate or prevent microbiome damage. One strategy might be a probiotic and that could be something that’s available on the shelf or perhaps a rationally designed probiotic where we make something with just the right strains that we figure out are the right ones. You could try a prebiotic approach where you might give a patient a food that promotes the growth of a good bacterium. Or perhaps you could use a post-biotic approach where if we can discern which metabolites the bacteria produce that are giving a salutary effect perhaps we could just give that metabolite which may be an attractive strategy in an immunocompromised patient. Finally we can think about different antibiotic strategies, to use or not use different types of antibiotics at different times in a rational way.

So the bottom line is these are different strategies that are in development, this is where many fields are going and we’re trying to take bone marrow transplantation in this direction as well. The abstract that we’re presenting at this conference indicates that you could try to deploy these strategies either after neutrophil engraftment or before the transplant in a clinical trial.

So, again in conclusion, microbiota injury, including loss of diversity and Enterococcal domination occur across geography. The association of diversity with overall survival is reproducible over time and across geography. Microbiota diversity is predictive of overall survival when sampled either pre-haematopoietic cell transplant or at the time of neutrophil engraftment. As I mentioned, this work was done in collaboration with colleagues at Duke, Regensburg and Hokkaido and my mentor Marcel van den Brink and my computational biologist co-author Antonio Gomes. Thank you very much.