Understanding Both Cause and Effect of Microbial Perturbations After Allogeneic Transplantation

Patients undergoing allogeneic stem cell transplantation (allo-SCT) are exposed to many perturbing stimuli, including chemotherapy, radiation, and a multitude of new pharmacologic agents. These result in some predictable consequences: mucosal damage, immune compromise, infection, and, in up to 50 percent of patients, graft-versus-host disease (GVHD). Furthermore, it is now accepted that alongside planned treatments for patients undergoing transplantation (and their well-established consequences) there is a marked disruption to the intestinal microbiome in the early weeks after allo-SCT.¹  One important disruptor of microbial communities is broad-spectrum antibiotic therapy, which is administered in response to the very common febrile episodes experienced by this patient population. Typically, antimicrobial treatment with piperacillin/tazobactam or cefepime is used in the first instance, followed by escalation to a carbapenem antibiotic (e.g., meropenem) in the case of persistent fever or clinical decompensation.

Many in the field are now focused on trying to move beyond associative studies to understand exactly why the microbiome is so disrupted by the process of allo-SCT, and why this disruption seems to effect so many different post-transplant outcomes. In their recent study published in Cell, Dr. Eiko Hayase and colleagues from The University of Texas MD Anderson Cancer Center identified an association between meropenem treatment and the incidence of acute GVHD of the intestinal tract by analyzing patient samples. They hypothesized that this phenomenon was microbiome-dependent, setting out to explore this hypothesis via carefully designed mouse experiments. As expected, meropenem induced marked changes in the intestinal microbiome of transplanted mice — most strikingly, a loss of members of the Clostridia class and expansion of Bacteroides at the genus level. Clostridia are known producers of short-chain fatty acids (e.g., butyrate and propionate), which are known to play a role in maintaining intestinal epithelial health in GVHD and are among the most frequently proposed mediators of microbiome-derived immunomodulation.² Bacteroides, conversely, are a genus of Gram-negative anaerobes with a reputation for causing opportunistic infection when they escape the intestinal tract.³ 

The complexity of microbial communities necessitates the use of summary measures to describe the microbiome (e.g., alpha and beta-diversity) and compare microbial features between patients, or mice, with different outcomes. In addition to the summary measures, we often lump bacteria together at the class, order, family, or genus level, comparing the abundances of these phylogenetically similar but functionally complex groups to try and understand how these bacteria may influence disease outcomes.

In this study, Dr. Hayase and colleagues took their investigation to the next level and identified a specific strain of a specific species of Bacteroides (Bacteroides thetaiotomicron; B. theta) that was markedly expanded in meropenem-treated mice. The identification of this strain of bacteria allowed detailed mechanistic studies to be performed, and because B. theta is known for a capacity to degrade the intestinal mucus layer and dietary polysaccharides, the authors examined the colonic mucus layer in mice with meropenem-exacerbated GVHD. As predicted, meropenem-driven B. theta expansion led to thinning of the colonic mucus layer, but surprisingly, the meropenem treatment seemed to supercharge the mucus-degrading capacity of B. theta by upregulating the expression of mucus-degrading enzymes. This unexpected finding highlights how dynamic these systems can be and how challenging it is to define microbial function in vivo. Because B. theta is known to consume dietary polysaccharides as well as mucin glycans, the authors performed a series of experiments to ask whether delivery of excess polysaccharides to the gut could perhaps “spare” the mucus layer. Indeed, this was the case with dietary xylose, and when meropenem-treated mice were supplemented with dietary xylose, they were spared from the excess GVHD-related mortality seen in the setting of meropenem alone.