Oral Microbiome Features Correlate with Mucositis in Pediatric Patients Receiving Chemotherapy

Background

Mucositis is a common, burdensome, and morbid toxicity of chemotherapy; understanding and preventing mucositis is an oncology research priority. The pathobiology of mucositis invokes a complex interplay between inflammation and oral commensals. Based on this model, the oral microbiome may be a potentially modifiable factor in mucositis development.

Methods

We conducted a pilot study to confirm the feasibility of oral microbiome sample collection in pediatric patients receiving chemotherapy. All patients were inpatient and received daily chlorhexidine mouth washes. Gingival swabs and saliva samples were collected at the start of chemotherapy (baseline) and day 7-14 post chemotherapy, when maximal mucositis is expected (mucositis period).

After DNA extraction from samples, shotgun libraries were generated and sequenced. Bacterial taxa were quantified by kraken2. Previously sequenced tongue swab samples [PMID:32822584] were used as age-similar healthy controls. Microbiome measurements were compared across swab versus saliva, chemotherapy patients versus healthy controls, and mucositis versus no mucositis.

Results

Nineteen patients were enrolled in this study [median age 15y, range (4-21y)]. Nine (47%) developed mucositis.

All patients provided samples at both time points. Taxonomic makeup in swab and saliva samples were qualitatively similar. There were no differences in alpha or beta diversity at baseline. Exact binomial tests identified relative taxa abundance differences in only 4/29 matched pair sample-timepoint comparisons.

Shannon and richness diversity measures were similar between chemotherapy patients and healthy controls (p=0.25 and p=0.9, respectively), but groups separated by beta diversity measurements (p=0.001), figure 1A. Chemotherapy patient samples were enriched for Schaalia odontolytica and had fewer Streptococcus, Veillonella and Haemophilus spp (q<0.05). Differences compared to controls were generally larger at the mucositis risk timepoint than baseline.

There were differences in patients undergoing chemo who developed mucositis versus those who did not. After mucositis onset, samples from patients with mucositis clustered by Bray-Curtis dissimilarity (p=0.018) and not by Jaccard distance (p=0.5), suggesting community differences based on proportions rather than taxa presence/absence. After mucositis onset, patients with mucositis demonstrated fewer Actinomyces spp. (q=0.03), and were enriched for Staphylococcus epidermidis (q = 0.03), figure 1B. Patients with and without mucositis were similar at baseline (beta and alpha diversity metrics p>0.05).

Conclusions

Oral microbiome sample collection is feasible in pediatric patients receiving chemotherapy, including as young as 4y and in those with mucositis. These samples demonstrate substantial differences, even at baseline, from healthy age-similar controls. Bacterial taxonomic differences that are seen between patients with and without mucositis require further investigation to understand the potential as a therapeutic target to mitigate mucositis-risk.