Conventional myelosuppressive chemotherapy for non-haematological malignancy disrupts the intestinal microbiome

Abstract

Background: The gut microbiota influences many aspects of host physiology, including immune regulation, and is predictive of outcomes in cancer patients. However, whether conventional myelosuppressive chemotherapy affects the gut microbiota in humans with non-haematological malignancy, independent of antibiotic exposure, is unknown.

Methods: Faecal samples from 19 participants with non-haematological malignancy, who were receiving conventional chemotherapy regimens but not antibiotics, were examined prior to chemotherapy, 7-12 days after chemotherapy, and at the end of the first cycle of treatment. Gut microbiota diversity and composition was determined by 16S rRNA gene amplicon sequencing.

Results: Compared to pre-chemotherapy samples, samples collected 7-12 days following chemotherapy exhibited increased richness (mean 120 observed species ± SD 38 vs 134 ± 40; p = 0.007) and diversity (Shannon diversity: mean 6.4 ± 0.43 vs 6.6 ± 0.41; p = 0.02). Composition was significantly altered, with a significant decrease in the relative abundance of gram-positive bacteria in the phylum Firmicutes (pre-chemotherapy median relative abundance [IQR] 0.78 [0.11] vs 0.75 [0.11]; p = 0.003), and an increase in the relative abundance of gram-negative bacteria (Bacteroidetes: median [IQR] 0.16 [0.13] vs 0.21 [0.13]; p = 0.01 and Proteobacteria: 0.015 [0.018] vs 0.03 [0.03]; p = 0.02). Differences in microbiota characteristics from baseline were no longer significant at the end of the chemotherapy cycle.

Conclusions: Conventional chemotherapy results in significant changes in gut microbiota characteristics during the period of predicted myelosuppression post-chemotherapy. Further study is indicated to link microbiome changes during chemotherapy to clinical outcomes.

Keywords: Cancer; Chemotherapy; Microbiome.

Fig. 1

Paired sample α-diversity changes during chemotherapy. a Observed species as a measure of bacterial richness and b Shannon diversity index as a measure of bacterial diversity. Pre-chemo: baseline samples (prior to chemotherapy), Post-1: 7–12 days post start of chemotherapy, Post-2 at the end of one chemotherapy cycle (median 21 days after chemotherapy). * = p < 0.05; ** = p < 0.01, performed Wilcoxon matched pairs signed rank test of 19 paired subject samples

Fig. 2

Non-metric multi-dimensional scaling (nMDS) plot showing paired-sample changes to microbiota composition following 7–12 days of chemotherapy (Post-1). Each colour represents an individual participant, with the pre-chemo sample (outline, lighter shade) linked to the post-chemotherapy sample (no outline, solid shade) by a line. Samples are shown to cluster by participant rather than by sampling time point, with no significant difference between the pre-chemo and post-1 groups PERMANOVA; p = 0.99

Fig. 3

The box plot figure depicts the median, IQR and range of the degree of similarity of the microbiomes in groups of samples using the Bray Curtis dissimilarity index where 0 indicates sample composition is identical and 1 indicates there are no shared species. The degree of similarly in samples from different participants in the cohort before chemotherapy (pre-chemo, unpaired) and 7–12 days following chemotherapy (post-1, unpaired) is depicted on the left. This shows that individual participant’s microbiomes were very different from each other before chemotherapy and remained very different (with no significant change in the degree of dissimilarity) following chemotherapy. On the right the degree of similarity between paired samples from the same participants before and 7–12 days after chemotherapy (chemo, paired) or healthy participants (healthy, paired) at matching sampling intervals are depicted. This shows that participant microbiomes were more similar to their own matched sample than to unrelated samples, but that the degree of difference in within-participant microbiomes before and after chemotherapy was significantly greater than that of paired samples from healthy participants. Significant comparisons are indicated by stars (**** = p < 0.0001; one-way ANOVA)

Fig. 4

Effect of chemotherapy on microbiome composition: Phyla relative abundance. The relative abundance of the four most abundant bacterial phyla (Firmicutes, Bacteroidetes, Proteobacteria and Actinobacteria) representing 97% of bacteria in the samples, were analysed. Pre-chemotherapy faecal microbiome composition (Pre-chemo) was compared to 7–12 days post-chemotherapy faecal microbiome composition (Post-1) and to faecal microbiome composition at the end of a chemotherapy cycle (median 21 days post-chemotherapy, Post-2) in 19 participants. Box and whiskers depict median ± interquartile range with bars representing minimum and maximum values. All significant comparisons are indicated by stars (* = p < 0.05; ** = p < 0.01; *** = p < 0.001; Wilcoxon matched-pairs signed rank test)

Fig. 5

Absolute abundance of E. coli bacteria determined using E. coli specific qPCR. Total E. coli equivalent colony forming per gram of stool (CFU/gram stool) of each participant’s stool sample was assessed at three time points before chemotherapy (Pre-chemo), 7–10 days post chemotherapy (Post-1) and at the end of one chemotherapy cycle (Post-2). Dots represent individual values of 19 paired subject samples. The increase in E. coli absolute abundance following chemotherapy was not significant (p = 0.18; paired t-test)