Relative Abundance Analysis of the Oral and Gastrointestinal Microbiome during Autologous Transplantation for Multiple Myeloma: Results of a Prospective Pilot Study and Association with Transplant Outcomes

Background

The human microbiome has been associated with allogeneic hematopoietic cell transplantation (HCT) outcomes. To date, there are no studies describing the longitudinal changes in the oral or gastrointestinal microbiome in the setting of autologous HCT. We conducted a prospective study to determine the longitudinal microbial profile in patients undergoing HCT for multiple myeloma (MM), and whether changes in microbial abundance correlate with HCT outcomes and/or toxicities.

Methods

Samples were collected from 15 MM patients on admission (baseline, T-2), during marrow aplasia (T+7) and after engraftment (T+30) (Table 1- summarizes baseline characteristics).

We evaluated the bacterial and fungal microbiome of 15 patients using Ion-Torrent PGM workflow. The amplicons generated from the 16s rRNA and the ITS genes were sequenced for bacterial and fungal identification, respectively. Sequencing reads were clustered into operational taxonomic units (OTUs, 3% distance) and taxonomically classified via Qiime bioinformatics pipeline.

Statistical analysis was performed using the statistical programming language R. Multivariate distance based association between communities and outcome was performed using the Adonis function as implemented in the R package vegan using Bray-curtis dissimilarities distances with and without presence/absence standardization (BrayPA). Non-parametric Spearman correlation and wilcoxon rank-sum test were used for association with continuous outcome and binary outcome respectively.

Results

Relative abundance was determined at the phylum and genus levels across each time point, in the oral and fecal microbiome, both bacterial and fungal. At the bacterial phylum level, the oral bacterial community composition significantly changed at T+30 compared to baseline (Bray-curtis, p=0.046) and T+7 (Bray-curtis, p=0.025). When examining the changes in the mycobiome composition, test for dissimilarity showed a significant difference in the fecal fungal genus between baseline and T+30 (BrayPA, p=0.025). No other statistically significant differences were noted.

Next, we correlated the microbial community (Table 2) and individual composition with outcome and transplant related toxicity, the later will be reported here. In fecal samples, the bacterial phylum Bacteriodetes present at T+7 was associated with the development and severity of diarrhea, such that patients with higher abundance of Bacteroidetes experienced lower gastrointestinal toxicity (pAdj=0.03). Additionally, the 2 genera Blautia and Ruminococcus, both belonging to the Firmicutes phylum and Clostridium Class, when detected at T+7 were associated with a higher development and severity of vomiting after exposure to high dose melphan (pAdj=0.05 for both respectively). In oral samples, the presence of the genus Glomerella at T+7 was negatively associated with rates of neutrophil engraftment (pAdj=0.03).

Conclusion and Future Directions

While acknowledging the limitation inherent in the small sample size, our results show that oral and fecal microbiota undergo dynamic changes in their diversity (Abstract ID 120038), composition and relative abundance during the course of transplantation. This change is likely multifactorial owing to the conditioning regimen, antimicrobial exposure and immune dysregulation. Our data suggest the bacterial and fungal microbiota present specifically at count nadir could be important players in this population of patients. Interestingly, the relative abundance of particularly the anaerobic bacterial phyla, Bacteroidetes and Firciumtes (Blautia and Ruminococcus) during marrow aplasia, correlated with transplant related toxicities in our cohort. This could further contribute to our knowledge of the role anaerobic organisms play in HCT outcomes, in accordance to what has been described in the allogeneic HCT literature.

Amifostine was used as a cytoprotectant before high dose melphalan for our patients. The effect of this organic thiophosphate on the microbiota is unclear and warrants future investigations. Further studies conducted on a larger scale and incorporating metabolomics and proteomics will help elucidate the interactions between the host and the microbiome and their effect on short term and long term transplantation outcomes as well as toxicities.

View largeDownload slide

View largeDownload slide

Close modal