Human Gut Mycobiome and Fungal Community Interaction: The Unknown Musketeer in the Chemotherapy Response Status in Bladder Cancer

Abstract

Background: Until recently, the properties of microbiome and mycobiome in humans and its relevance to disease have largely been unexplored. While the interest of microbiome and malignancy over the past few years have burgeoned with advent of new technologies, no research describing the composition of mycobiome in bladder cancer has been done. Deciphering of the metagenome and its aggregate genetic information can be used to understand the functional properties and relationships between the bacteria, fungi, and cancer.

Objective: The aim of this project is to characterize the compositional range of the normal versus bladder cancer mycobiome of the gut.

Design setting and participants: An internal transcribed spacer (ITS) survey of 52 fecal samples was performed to evaluate the gut mycobiome differences between noncancer controls and bladder cancer patients.

Outcome measurements and statistical analysis: Our study evaluated the differences in mycobiome among patients with bladder cancer, versus matched controls. Our secondary analysis evaluated compositional differences in the gut as a function of response status with neoadjuvant chemotherapy. Data demultiplexing and classification were performed using the QIIME v.1.1.1.1 platform. The Ion Torrent-generated fungal ITS sequence data were processed using QIIME (v.1.9.1), and the reads were demultiplexed, quality filtered, and clustered into operation taxonomic units using default parameters. Alpha and beta diversity were computed and plotted in Phyloseq, principal coordinate analysis was performed on Bray-Curtis dissimilarity indices, and a one-way permutational multivariate analysis of variance was used to test for significant differences between cohorts. Phylogenetic Investigation of Communities by Reconstruction of Unobserved States (PICRUSt) was applied to infer functional categories associated with taxonomic composition.

Results and limitations: We found distinctive mycobiome differences between control group (n = 32) and bladder cancer (n = 29) gut flora, and identified an increasing abundance of Tremellales, Hypocreales, and Dothideales. Significant differences in alpha and beta diversity were present between the groups (control vs bladder; p = 0.002), noting distinct compositions within each cohort. A subgroup analysis by sex and neoadjuvant chemotherapy status did not show any further differences in mycobiome composition and diversity. Our results indicate that the gut mycobiome may modulate tumor response to preoperative chemotherapy in bladder cancer patients. We propose that patients with a "favorable" mycobiome composition (eg, high diversity, and low abundance of Agaricomycetes and Saccharomycetes) may have enhanced systemic immune response to chemotherapy through antigen presentation.

Conclusions: Our study is the first to characterize the enteric mycobiome in patients with bladder cancer and describe complex ecological network alterations, indicating complex bacteria-fungi interactions, particularly highlighted among patients with complete neoadjuvant chemotherapy response.

Patient summary: Our study has demonstrated that the composition of stool mycobiome (fungal inhabitants of the gastrointestinal tract) in patients with bladder cancer is different from that in noncancer individuals. Furthermore, when evaluating how patients respond to chemotherapy given prior to their surgery, our study noted significant differences between patients who responded and those who did not.

Keywords: Bladder cancer; Microbiome; Mycobiome; Urothelial cancer.

Fig. 1

Alterations of enteric fungi in bladder cancer (BCa). (A) Relative abundance of dominant enteric mycobiome control (n = 29) and BCa (n = 32) (B) Principal component analysis of Bray-Curtis distance showing stratification of BCa from control samples by fungal composition profile. Distinct clustering between BCa and control groups is noted (p = 0.002). Groups were compared using Mann-Whitney U test. (C) Shannon index diversity indices between BCa (n = 29) and control (n = 32) groups. (D) Relative fold change in mycobiome species in bladder cancer patients (n = 29) against a mean normalized relative abundance of control group (n = 32). The greatest difference between Tremellales (log change –2.99, p = 0.0041/p_adj = 0.009), Hypocreales (log change –3.871, p ≤ 0.001/p_adj = 0.003), and Dothideales (log change –2.37, p = 0.002/p_adj = 0.016; Fig. 1D) were seen.

Fig. 2

(A) Differentially abundant KEGG metabolic pathways inferred by LEfSe using KEGG module abundance to identify differentially abundant pathways between controls and bladder cancer patients. PICRUSt was used to predict the functional potential of mycobiome using ITS RNA gene sequence data. The prevalence of top 19 mycobiome-associated KEGG pathways seen in control and healthy gut is presented. (B) Venn diagram analysis showing specific and shared expressed pathways between control and bladder cancer mycobiome KEGG metabolic pathways. Unique KEGG metabolic pathways are identified in the pink (BCa) and blue (controls) regions (p < 0.05). (C) Microbiome and mycobiome interaction network built from OTUs; each node represented a microbe or fungus, and each line presents co-occurrence or interaction (green = negative and red = positive). ITS = internal transcribed spacer; KEGG = Kyoto Encyclopedia of Genes and Genomes; LDA = linear discriminant analysis; LEfSe = LDA effect size; PICRUSt = Phylogenetic Investigation of Communities by Reconstruction of Unobserved States; TCA = tricarboxylic acid.

Fig. 3

Compositional difference in the gut mycobiome associated with neoadjuvant chemotherapy response. (A) Stacked bar plot of phylogenetic composition of common fungal taxa (>0.01% abundance) at the order level in fecal (n = 10) samples by ITS rRNA sequencing. (B) Inverse Simpson diversity scores of the gut mycobiome in the complete response (CR; n = 4) versus no response (NR; n = 3) to gemcitabine-cisplatin neoadjuvant chemotherapy by Mann-Whitney test. The remainder (n = 3) experienced partial response to NAC, with residual low-grade cancer within the specimen. Error bars represent the distribution of diversity scores. (C) Principal coordinate analysis of fecal samples (n = 7), with CR (n = 4), and with NR (n = 3), showing clustering of nonresponders and responders based on mycobiome composition (p = 0.041). (D) LDA scores computed for differentially abundant taxa in the fecal mycobiome of the CR (green) and NR (red) groups. Length indicates the effect size associated with designated taxon (p = 0.05 for the Kruskal-Wallis test; LDA score >3). (E) Cladogram of LEfSe showing differences in fecal taxa across mycobiome in the CR versus NR group. Increased abundance of Agaricomycetes and Sacchaaromycetes is present in the NR group. ITS = internal transcribed spacer; LDA = linear discriminant analysis; LEfSe = LDA effect size; NAC = neoadjuvant chemotherapy.