Hypothesis / aims of study
The current definition of overactive bladder (OAB) is based on the absence of infection. The bladder has historically been regarded as a sterile environment until researchers tried to identify microorganisms that could have a causative association with OAB in order to question not only the dogma of urinary sterility but also the syndrome’s definition. Using 16s rRNA sequencing technology, bacterial DNA and live bacteria have been detected in human urine in the absence of clinical infection and an association between OAB and an altered URM. There have been criticisms of the 16s rRNA urinary studies and contamination of the urine during collection. The once thought unidirectional effect of pathogenic bacteria on the host, is now believed to be a symbiotic interaction influencing both healthy and pathogenic disease. Some Lactobacilli spp, predominant in the vaginal habitat, are thought to be protective against OAB. Overall, the presence of these in the urinary tract suggests partial colonization with vaginal microbiota. However, no study has examined the microbiota migration between the vaginal and urinary tract habitats. Also, this study aims to get samples of the bladder wall without contamination through a cystoscope in an operating room environment through a sheath. Our aim was to assess the relationship between the urinary, urothelial and vaginal microbiomes in patients suffering from OAB.
Study design, materials and methods
From a tertiary Urogynaecology centre, over a two-year period, we recruited women who suffered from a variety of OAB symptoms and who were undergoing cystoscopy. After informed consent, women completed the 12 item ICIQ-FLUTS questionnaire assessed by recall over the previous one-month period. We collected a bladder biopsy for the analysis of the urothelial microbiome (UTM), a urine sample for the analysis of the URM and a high vaginal swab for the analysis of the vaginal microbiome (VGM). Samples were stored at -80ºC prior to extraction. DNA from the samples were extracted using QiAMP DNA minikit, and they were sent to Research and Testing Labs in Texas for 16S rRNA sequencing. The microbiomes were first assessed individually against the variables urinary incontinence (UI), relative abundance, urotype and vaginal class, and then compared with each other. UI was assessed on question 9a from the ICIQ-FLUTS questionnaire “does urine leak before you get to the toilet?”. Patients who scored more than two were labelled as cases, less than two were controls. Patients were also split into two cohorts, high and low abundance, depending on the microbiome. We classified the VGM into Lactobacillus dominant or Diverse (no dominant species) and the URM into three urotypes: Lactobacillus, Diverse and Other. Data were analysed using STAMP statistical software. For variables where patients were split into two cohorts, we used t-test (equal variance) to account for multiple hypothesis testing. For variables where patients were split into more than two cohorts, Tukey-Kramer post-hoc test was performed. Both tests were added a Benjamini-Hochberg FDR correction.
We recruited 103 women with a mean age of 50 years (SD 16.4) and median parity of 2 (range 0-5). All participants had a bladder biopsy, from 99 women we obtained a urine from 68 women a high vaginal specimen. From a total of 964, 405 identified species were defined as “unclassified” at different taxonomic levels and have been assigned a number for inclusion in the statistics. Contaminated samples were excluded from our analyses. Table 1 summarises all the significant bacteria species found in the individual microbiomes.
We successfully sequenced 65 pairs of urine and vaginal samples, 99 pairs of biopsy and urine samples, and 68 pairs of vaginal and biopsy samples. These analyses showed that patients suffering from UI had significantly more Aerococcus christensenii, Microbacterium spp (p=0.042) and Unclassified 313 (p=0.036) in their URM. Figure 1 shows all species in the UTM depending on their relative abundance of URM bacteria. Low abundance VGM patients had increased Parvimonas sps (p=0.027) in their URM and patients with a low abundance URM had increased Escherichia coli (p=0.05) in their VGM. Lactobacillus dominant urotypes were associated with Lactobacillus crispatus dominated VGM (p=0.044) and increased amounts of Actinomyces spp (p=2.74e-3) in the UTM. Other urotypes had increased Prevotella bivia (p=7.51e-3) and Bacteroides vulgatus (p=0.04) in the VGM and increased Burkholderia pseudomallei (p=8.6e-3) in the UTM. Lactobacillus spp dominant VGM had significantly more Prevotella bivia (p=0.032) and Acinetobacter iwoffii (p=0.047) in their URM. Lactobacillus spp dominant VGM had increased Lactobacillus crispatus (p=2.42e-3) in their UTM, diverse VGM patients had significantly more Anaerococcus spp (p=0.023) in their UTM.
Interpretation of results
To our knowledge, our study is the first to examine the microbiome from urothelial biopsy samples. UTM analyses allows identification of bacteria that a simple microscopic culture does not detect. Hence, we confirmed that Prevotella spp plays a protective role for UI and uncovered Propionmicrobium lymphophilum as a possible contributor to the development of OAB. The latter’s presence indicates that these patients may have previously suffered from clinical infection and that the immune system was unable to completely remove the bacteria. We also found that there is a link between the three microbiomes, with regards to Lactobacillus spp in general and Lactobacillus crispatus in particular, and that some vaginal species predict the urine composition depending on abundance and diversity. Our results also confirmed that uropathogenic organisms, such as Aerococcus christenseni arise from the VGM.