The Response of the Urinary Microbiome to Botox

Abstract

Introduction and hypothesis

Our objective was to evaluate if botox alters the urinary microbiome of patients with overactive bladder and whether this alteration is predictive of treatment response.

Methods

This multicenter prospective cohort study included 18–89-year-old patients undergoing treatment for overactive bladder with 100 units of botox. Urine samples were collected by straight catheterization on the day of the procedure (S1) and again 4 weeks later (S2). Participants completed the Patient Global Impression of Improvement form at their second visit for dichotomization into responders and nonresponders. The microbiome was sequenced using 16s rRNA sequencing. Wilcoxon signed rank and Wilcoxon rank sum were used to compare the microbiome, whereas chi-square, Wilcoxon rank sum, and the independent t-test were utilized for clinical data.

Results

Sixty-eight participants were included in the analysis. The mean relative abundance and prevalence of Beauveria bassiana, Xerocomus chrysenteron, Crinipellis zonata, and Micrococcus luteus were all found to increase between S1 and S2 in responders; whereas in nonresponders the mean relative abundance and prevalence of Pseudomonas fragi were found to decrease. The MRA and prevalence of Weissella cibaria, Acinetobacter johnsonii, and Acinetobacter schindleri were found to be greater in responders than nonresponders at the time of S1. Significant UM differences in the S1 of patients who did (n = 5) and did not go on to develop a post-treatment UTI were noted.

Conclusions

Longitudinal urobiome differences may exist between patients who do and do not respond to botox.

Appendices

Appendix 1

Appendix 2

Demographic variables collected

• By Questionnaire from patient:

  • Age

  • Parity

  • Weight

  • Height

  • Race

  • Menopausal status

  • Postmenopausal estrogen use, if applicable

  • A history of medical conditions associated with UTI:

    • Diabetes

    • Immunologic disease

    • Hypertension

    • Coronary artery disease

    • Current smoker

    • History of pelvic radiation

• By chart review:

  • Previous OAB treatment

    • Number of prior OAB treatment(s)

    • Types of prior OAB treatment(s) (anticholinergic medication, beta-3 agonist medication, percutaneous tibial nerve stimulation, BTX injection, sacroneuromodulation)

  • Neurologic disease

  • Chronic pelvic pain

  • Most recent A1c (if applicable)

  • Currently taking potentially confounding medications

    • Opioid analgesics

    • Anticholinergic (non-bladder) medications

  • PVR within 90 days of the BTX procedure

  • Urodynamic parameters within one year of the BTX procedure:

    • Presence of detrusor overactivity

    • Presence of leak with detrusor overactivity

    • Detrusor pressure at max flow

    • Maximum detrusor pressure

    • Maximum flow velocity

Procedural Variables Collected

• Antibiotic prophylaxis given

• Number of injection sites

• Procedure location

• Number of previous BTX

Adverse Event Variables Collected

• Urinary tract infection (within 30 days of procedure)

• Urinary retention (within 30 days of procedure)

Appendix 3

PCR Amplification

Amplifications of the 16S and 5.8S rRNA genes were performed using 16S-515 (5’-GGA CTA CCA GGG TAT CTA ATC CTG- 3’) and 16S-804 (5’-(TCC TAC GGG AGG CAG CAG T-3’) and ITS1 (5’-(TCC GTA GGT GAA CCT GCG G- 3’) and ITS4 (5’-TCC TCC GCT TAT TGA TAT GC- 3’) primers, respectively [24]. The PCR mixture comprised Q5 High-Fidelity Master Mix (New England Bioinformatics) at a 1x concentration, along with a double volume of molecular grade water and 0.05 ul/mM each primer. Undiluted DNA (1.5 μl) was added to each 50 ul reaction. Thermo-cycling conditions consisted of an initial denaturation step (3 min at 98 °C), followed by 30 cycles of denaturation (10 s at 98 °C), annealing (10 s at 55 °C for the 16S primers and 20 s at 58 °C for the ITS primers), extension (10 s at 72 °C), and a final extension step of 3 min at 72 °C. Ten μl of each PCR product were separated using gel electrophoresis on 1.5% agarose gel (containing 7 μg/ml ethidium bromide).

Library Preparation and Sequencing

The amplicon library was cleaned and barcoded followed by emulsion PCR using Ion Torrent S5 Prime workflow according to the manufacturer’s instructions (ThermoFisher). Equal volumes of bacterial 16S rRNA and fungal ITS amplicons were pooled, cleaned with AMPure XP beads (Beckman Coulter, CA, USA) to remove unused primers, and then exposed to end repair enzyme for 20 min at room temperature. After an additional AMPure clean up, ligation was performed at 25 °C for 30 min using Ion Torrent P1 and a unique barcoded 'A' adaptor per pooled sample. After AMPure removal of residual adaptors, samples were concentrated to 1/4 volume for 1 h using Labconco Vacuum under heat. All separate barcoded samples were then pooled in equal amounts (10 μl) and size selected for the anticipated 16S and ITS range (200–800 bp) using Pippin Prep (Sage Bioscience). The library was amplified for seven cycles and quantitated on StepOne qPCR instrument ahead of proper dilution to 300 pM going into IonSphere templating reaction on the Ion Chef. Library sequencing was completed on an Ion Torrent S5 sequencer (ThermoFisher Scientific) and barcode-sorted samples analyzed in our custom pipeline based on Greengenes V13_8 and UNITE database V7.2 designed for the taxonomic classification of 16SrRNA and ITS sequences, respectively. Sequencing reads were clustered into operational taxonomic units (99% homology), described by community metrics, and taxonomically classified within the Qiime 1.8 bioinformatics pipeline.

Appendix 4

Pre-Processing

De-multiplexing of the sequencing output was performed in Python 2.7 with the input equal to adapter-trimmed output from the ITS platform. A map file which links the sample ID to the barcodes, adapted to the amplicons in the ligation phase of the library preparation, was created. The de-multiplexed data was then parsed for quality (Q20) and sequence lengths sequestered to 200–400/400–800 bp, for 16s/ITS, respectively. Finally, the sequences were de-noised and chimeras removed.

The operational taxonomic units were generated using de novo clustering by a defined similarity threshold of 0.97. Sequences that were similar at or above the accepted threshold level represented the presence of a taxonomic unit (e.g., a species similarity threshold is set at 0.97) in the sequence collection. Taxonomy was assigned in our custom pipeline based on Greengenes V13_8 and UNITE database V7.2 taxonomic classification of 16s and ITS sequences, respectively. Blastn was used for alignment with an error threshold of 0.001 for selection. Data was summarized, and prepared reads were represented by absolute count per ID. Statistical analysis was performed in R using Packages.

Data Preparation

Raw 16S and ITS absolute count matrices were obtained from the pre-processing step described previously, and sample annotation was loaded to R version 4.0.3. Using R package microbiome 1.12.0 and phyloseq 1.34.0, a phyloseq object containing a read count matrix (species level identification OTUs), taxonomic table (kingdom down to species) and sample annotation was constructed. Initially the 16S and ITS are handled separately until after the read counts are normalized and transformed to relative abundance.

Data Cleaning and QC

OTU identifiers were used to construct the taxonomic table. The table was cleaned by removing OTUs that were annotated as “unknown/unidentified” at a phyla, species or genus level. This step ensured that only species that were identifiable were being investigated in the analysis. Quality of the samples was assessed using the total read count after the taxonomic table had been cleaned and aggregated to a species level. A minimum of 500 16S read counts was used as the lower cut-off. Finally, samples with a read count <500 were removed, and the data was normalized and transformed to relative abundance.

Data exploration

Composition bar graphs were generated using 16S and ITS relative abundance data separately, after aggregating the data to the phyla level. Filtering on detection limit (0.001% minimum) and prevalence (10% among all samples) was used. These figures allow us to have a simple overview of the phyla abundances in the samples. Ordinate analysis/PCA was performed on 16S and ITS data separately, using the species level relative abundances. Filtering on detection limit (0.00001% minimum) and prevalence (50% among all samples) was done. Ordinate analysis was done using the “ordinate()” function with method “MDS” (multidimensional scaling) and “Bray-Curtis” distance, from R package microbiome version 1.12.0. Alpha diversity measures are calculated using the “alpha()” function with “Shannon index” as the diversity measure, from R package microbiome version 1.12.0.

Data Analysis

Given that we had Pre & Post timepoints for the same samples, we used the Wilcoxon-rank-sum test “wilcox.test()” with option “paired = TRUE” from R package stats version 4.0.3 is used on comparisons from the same response group (PGII Response Yes/No). For comparisons between timepoints for the same response group we perform the prior test; however, with option “paired = FALSE.” For comparisons between Responders & Nonresponders groups, we use “wilcox.test()” with option “paired = FALSE.” A significance level of ≤ 0.05 was accepted as statistically significant. Foldchange between groups was calculated using mean relative abundance (MRA) of the taxa using the “foldchange()” function from R package gtools version 3.9.2. All figures related to this data were generated using ggplot2 and ggpubr R packages versions 3.3.5 & 0.4.0, respectively.