Changes in the Stool Virome and Fungome of Children Undergoing Bone Marrow Transplant

Abstract

Background Bacterial dysbiosis is associated with acute graft versus host disease (GVHD) and blood stream infections (BSI) after bone marrow transplant (BMT). Intestinal viral and fungal changes after BMT are unknown in children. Methods Metagenomic shotgun paired reads sequencing was performed for 56 patients who underwent a BMT at Cincinnati Children's Hospital Medical Center. Stool samples were collected pre BMT and at least one post BMT time point (range day+1-day+35). Metagenomics data were analyzed using DIAMOND / MEGAN6 CE and NCBI RefSeq database. Results Demographics are shown in Figure 1. A total of 180 stool samples were analyzed. Twelve of 56 patients had stool samples analyzed at 2 time points (pre BMT and post BMT) while 44 patients had stool analyzed at > 2 time points with one being a pre BMT sample. All patients were on fungal and viral prophylaxis. Fungome Fungi were detected in 19 of 56 patients' stool samples (Fig 2A). The most frequent fungi isolated from stool were Candida sp. Nine of 56 patients had evidence of clinical fungal infection after BMT (Candida N=7, Malassezia N=1, Rhizopus N=1). One patient with clinical Candida infection had Candida detected from stool pre BMT but the remaining patients did not have a fungal infection similar to the fungi isolated from stool. BSI occurred in 7/19 in patients with fungi isolated from stool compared to 11/37 in patients without fungi in stool. GI GVHD occurred in 7/16 (43%) in allogeneic BMT patients with fungi in stool compared to 5/35 (14%) patients without fungi isolated in stool. likely reflecting antibiotic exposure and resultant dysbiosis. Virome Viruses were identified in 31 of 56 patients' stool samples (Fig 2B). Caudalovirales (bacteriophages) were the most frequent organisms seen in stool (n=20). Interestingly Caudalovirales in stool were detected in 18 patients who were on human milk as their nutritional source. Mastadenovirus sp. (adenovirus in clinical practice) was observed in 5 patients. Mastadenoviruses in stool were seen with ATG (n=1) or alemtuzumab (n=4) exposure. Ten of 56 patients had clinical adenoviral reactivation. Four of the 10 patients with clinical adenoviral reactivation had adenovirus isolated in the stool prior to onset of clinical viremia (pre BMT n=1; day+1 n=2; day+13 n=1). Conclusion We observed that fungi were not frequently detected from stool, and presence of fungi did not always imply clinical fungal infection with the same organism but could reflect antibiotic exposure and contribute to intestinal dysbiosis leading to acute GVHD. Presence of bacteriophages in stool is a novel observation, which could be reflective of a diet of human milk. We have previously shown that human milk lowers intestinal inflammation after BMT and these findings will be explored in independent studies. Mastadenovirus was detected in stool before or early after BMT and frequently heralded clinical viral reactivation.