Urinary microbiota is associated to clinicopathological features in benign prostatic hyperplasia

The urinary microbiota of patients with benign prostatic hyperplasia (BPH) has been associated with lower urinary tract symptoms (LUTS), however, little is known about urinary microbiota correlations with clinicopathological parameters associated with BPH. Here, we investigate associations between the urinary microbiota and clinical parameters of patients with BPH undergoing surgery.


| INTRODUCTION
Sterile urine is a paradigm that is being widely refuted.The urinary system has a set of microorganisms that cannot be identified by traditional culture methods, but novel strategies involving genetic sequencing of urinary lavage have been changing this scenario. 1ese findings have expanded the frontiers of research on the contribution of the urinary microbiota to the maintenance of a healthy urinary environment, as well as on how its dysbiosis can contribute to the development of various pathologies, ranging from urinary incontinence 2 and bladder cancer 3 to lower urinary tract symptoms (LUTS) in men and women. 4,5nign prostatic hyperplasia (BPH) is a common urological disease that affects elderly men, causing LUTS.These symptoms can be divided into storage, micturition, and post-micturition symptoms.LUTS relate to bladder outlet obstruction caused by BPH due to histological changes in the prostate gland. 6α−1 adrenergic receptor antagonists (also called alpha-blockers) and 5-α-reductase inhibitors are the drugs of choice for BPH, but many patients require surgical treatment. 7,8The absolute indications for BPH prostate surgery are persistent macroscopic hematuria, LUTS refractory to drug treatment, recurrent urinary infections, renal failure, hydronephrosis, and acute urinary retention. 9,10e pathophysiologic mechanisms of BPH are poorly understood, 11 however, inflammation has been associated with disease progression 12,13 and bacterial infections are possible inducers of prostatic inflammation. 14In this scenario and given the anatomical proximity between the prostate gland and the urinary system, a possible link between the urinary microbiota and BPH has been suggested, [15][16][17][18] but more studies are needed to assess the impact of urinary microbes in the different pathophysiological aspects of this condition.
In this exploratory study, we collected catheterized urine samples from BPH patients undergoing surgery and analyzed their urinary microbiota by sequencing the 16S rRNA gene.We searched for associations between the urinary microbiota and clinical parameters relevant to BPH.

| Sample collection
We included in this prospective study patients with BPH undergoing transurethral resection of the prostate (TURP) who had at least one surgical indication for BPH.All included patients signed an informed consent form for the study, which was approved by the institutional ethics committees (#4,053,513).
From each patient, a urine sample was collected during a sterile surgical procedure for TURP performed at Hospital Sírio-Libanês or Hospital Nossa Senhora das Graças for the treatment of BPH.The samples were collected using a urinary catheter immediately after the start of TURP for the treatment of BPH and always before antibiotic prophylaxis.All urine samples were placed in sterile 80 ml collection tubes and stored at −80°C until DNA extraction.

| DNA extraction and 16S rRNA amplicon sequencing
We extracted bacterial DNA from urinary samples using the QIAamp DNA Microbiome kit (Qiagen, Germany) and prepared sequencing libraries using the QIAseq.16S/ITS Region Panel kit (Qiagen).We chose to amplify and sequence the 16S gene hypervariable regions V1V2, as recommended for male urinary samples. 19Libraries with a final concentration <0.4 nM were deemed to have undetectable microbiota and were not sequenced.Libraries with detectable microbes were sequenced at a final concentration of 10 pM in the Illumina MiSeq System using the MiSeq Reagent Kit v3.DNA extraction negative controls were also prepared and sequenced as described previously. 19

| Bioinformatic analysis
Read processing was carried out in QIIME2. 20DADA2 was used to generate amplicon sequence variants (ASVs) and chimeric ASVs were removed using VSEARCH. 21,22After taxonomic classification of ASVs using a naive-Bayes-based classifier and the SILVA database, nonbacterial ASVs were removed. 23Finally, contaminant ASVs were identified in negative controls and removed as described previously. 19

| Microbiome and statistical analysis
The data set was normalized by scaling with ranked subsampling, 24 where we established 905 as the minimum number of reads needed to characterize the microbiota.Samples with <905 reads were also considered to have undetectable microbiota.Next, we calculated microbiota alpha-(ASV richness, Gini-Simpson, Shannon, and Faith phylogenetic diversity) and beta-diversity metrics (weighted UniFrac).
Patients were segregated into two groups, according to each clinical parameter.Mann-Whitney was used to compare the alphadiversity between groups.The difference in composition between groups and the influence of covariates on the bacterial compositions were evaluated by PERMANOVA.Differences in genera abundances between groups were assessed using the MaAsLin2 method. 25| RESULTS

| Characteristics of the study patients
Forty-one patients were recruited between March 2019 and May 2022.
All patients had at least one clinical indication for TURP, whether due to LUTS, paradoxical incontinence, acute urinary retention, recurrent urinary tract infection, hematuria, urinary flow abnormality, voiding residue, or structural changes in the bladder.None of the patients had hydronephrosis, uremia, lithiasis, or urological cancer.The clinical and demographic data are shown in Table 1.We observed a high rate of changes in the bladder structure, such as trabeculations, diverticula, and/or detrusor hypertrophy, affecting 63.4% of the patients.The mean prostate-specific antigen (PSA) was 3.63 (0.37-22.96).The prostate, assessed by digital rectal examination, had a mean weight of 56.25 g (25-110 g).Five patients needed to have an indwelling urinary catheter due to acute urinary retention.More than a third of the patients (36.7%) did not use drug therapy before TURP.

| Urinary microbiota detectability association with clinical/demographic data
Each patient provided a single catheterized urine sample during TURP.Among the 41 samples analyzed, 31 (76%) were considered to have detectable microbiota (see Section 2) and had their urinary microbiota characterized.
We did not find a significant association between urinary microbiota detectability and clinical/demographic data, such as age, smoking status, PSA, and residual urinary volume (Fisher's exact test, p > 0.05; Supplementary Table 1).However, a trend towards higher detectability among patients with greater prostate weight (≥50 g) was noticed (Fisher's exact test, p = 0.13).

| Urinary microbiota alpha-diversity association with clinicopathological parameters
Considering different metrics (see Section 2), we searched for associations between the alpha-diversity of the urinary microbiota and clinicopathological parameters associated with BPH.We found that the urinary microbiota of BPH patients with higher PSA showed greater Faith's phylogenetic diversity and higher ASV richness (Figure 1A), indicating that in this group the urinary microbiota was composed of a higher number of clades and that these clades were more phylogenetically diverse.In addition, patients with greater prostate weight showed higher Faith's phylogenetic diversity (Figure 1B).Importantly, all associations remain significant after removal of patients with recent (past 3 months) history of antibiotics use for urinary tract infection (Mann-Whitney test, p = 0.030 for ASV richness vs. PSA, p = 0.006 for Faith's PD vs. PSA, and p = 0.033 for Faith's PD vs. prostate weight).
Other parameters, such as age and presence of structural changes of the bladder, were not associated with urinary microbiota alpha-diversity (Supporting Information S1: Figure 1).

| Urinary microbiota beta-diversity association with clinicopathological parameters
Next, we evaluated associations between the composition of the urinary microbiota and clinicopathological parameters.In multivariate PERMA-NOVA including all parameters of interest, we found that the prior use of 5-α-reductase inhibitors was the only parameter significantly associated with the composition of the urinary microbiota, but it explained only ~7% of the variance observed across the whole cohort (Figure 2A).Principal coordinate analysis with groups stratified by previous use of 5-αreductase inhibitors and univariate PERMANOVA test also indicated a significant difference in composition between groups (Figure 2B).

| Urinary microbiota taxonomic composition association with clinicopathological parameters
The most relevant bacterial genera in the urinary microbiota of the population evaluated in terms of median relative abundance were Corynebacterium (14.4%),Lactobacillus (8.0%), Variovorax (7.9%), Staphylococcus (7.5%), and Cutibacterium (6.5%).However, the overall taxonomic composition of the urinary microbiota varied considerably T A B L E 1 Clinical and demographic data of study patients.between BPH patients (Figure 3A).For instance, among the four patients with detectable microbiota that were treated with antibiotics for a urinary tract infection at most 3 months before sample collection, #009, #012, and #035 showed a clearly less diverse urinary environment (Figure 3A).Nevertheless, some genera showed high prevalence, such as Anaerococcus, Corynebacterium, Cutibacterium, Finegoldia, Lactobacillus, Peptoniphilus, SN8, Staphylococcus, and Streptococcus, which were present (relative abundance ≥1%) in at least a fourth of patients with detectable microbiota.
Focusing on the most prevalent genera mentioned above, we searched for associations between the urinary microbiota taxonomic composition and clinicopathological parameters of the BPH context using the MaAsLin2 tool (Figure 3B).After correcting for the number of genera tested, we found, for example, that the relative abundance of Streptococcus was significantly lower in patients with greater prostate weight and that the relative abundance of Lactobacillus was significantly higher in patients with higher PSA levels.We also found that BPH patients who had previously used 5-α-reductase inhibitors showed higher relative abundances of Corynebacterium and Anaerococcus, which partly explains the significantly different urinary microbiota composition observed previously for this group.

| DISCUSSION
6][17][18] In this study, we analyzed catheterized urine microbiota from patients with BPH undergoing surgery and investigated the association between urinary microbiota detectability, diversity, and composition with clinical parameters.| 289 We showed that greater prostate weight and higher PSA levels are both associated with higher urinary microbiota alpha-diversity in BPH patients.We also found associations between clinical parameters and the abundance of specific genera, such as a higher abundance of Streptococcus in patients with greater prostatic weight and a higher abundance of Lactobacillus in patients with higher PSA levels.The latter is in line with a recent study showing a higher abundance of Lactobacillus in the urinary microbiota of BPH patients in comparison with controls. 18The aforementioned study further described an association between Haemophilus abundance and PSA levels that we did not find, but this is possibly due to their adoption of midstream urine collection as sampling strategy.
Many studies have shown that catheterized urine is more appropriate than voided urine samples for urinary microbiota analyses, since catheterization avoids urine sample contamination with distal urinary tract and skin bacteria. 16,26One such study also investigated BPH patients. 16They showed an association between the detectability of microbiota in catheterized urine samples, but not in the midstream voided ones, with the severity of LUTS.Of note, detectability was considerably higher in our cohort (76% vs. 27%), likely reflecting the optimized protocol we adopted. 193][14] Although the mechanisms inducing inflammation are not yet completely understood, recent studies suggest a role of the urinary microbiota in this process. 4,27,28sed on our findings, we speculate that BPH generates an inflammatory milieu in the genitourinary tract, which alters the physicochemical properties of urine (pH, osmolarity, and the concentration of nitrite and proteins), allowing urine colonization by non-commensal bacteria and resulting in increased microbiota diversity and dysbiosis.The presence of non-commensal bacteria could, in turn, potentiate inflammation of the genitourinary tract and contribute to BPH progression.Although high microbiota diversity has been associated with health in complex microbial communities, such as the gut and oral microbiotas, 29,30 in simpler microbial communities, such as the skin and vagina microbiotas, high diversity is frequently associated with disease. 31,32is hypothesis aligns with the observed association between the use of 5-α-reductase inhibitor with a higher abundance of urinary Corynebacterium and Anaerococcus, which have been reported to be part of the healthy human urinary microbiota.The use of alphablockers and/or 5-α-reductase inhibitors is the gold standard firstline therapy for patients with BPH, reducing inflammation and preventing disease progression. 33Although causality cannot be evaluated, we can hypothesize that 5-α-reductase inhibitors reduce the inflammatory milieu and microbiota dysbiosis.Future studies to explore this hypothesis should include associations with inflammatory biomarkers Of note, due to the small sample size, we considered the use of 5-α-reductase inhibitors as an independent variable for statistical analysis, so that we could not evaluate interactions between drugs (e.g., 5-α-reductase inhibitor + alpha-blocker).Besides, all patients analyzed in our study underwent surgery.Thus, it is possible that the microbiota differences between groups could be even sharper if patients clinically responsive to the drugs-and, consequently, that had not had to undergo TURP-were considered in our analysis.
Other limitations are the heterogeneous patient population regarding symptoms that prompted the TURP and the use of 16S rRNA gene amplicon-sequencing, which has a low taxonomic resolution and does not provide microbiota functional information.Further studies using shotgun metagenomic sequencing in a larger and more homogeneous patient cohort are needed to validate our findings and establish causal associations between urinary microbiota and BPH pathophysiology.Ideally, these studies should also incorporate microbiota analysis of prostate tissues and associations with inflammatory markers to elucidate whether microbiota dysbiosis causes symptoms or drives prostate growth.
In conclusion, we analyzed the catheterized urine microbiota from patients with BPH undergoing surgery and investigated the association between the urinary microbiota detectability, diversity, and composition with clinicopathological parameters.Urinary microbiota diversity and composition were associated with clinical parameters in BPH, such as the link between the abundance of commensal urinary bacterial genera and the use of 5-α-reductase inhibitors.Future studies investigating those associations in more depth may pave the way for new therapeutic strategies for BPH.

F I G U R E 1
Association between urinary microbiota alpha-diversity and prostate-specific antigen (PSA) (A) or prostate weight (B).Different alpha-diversity metrics were evaluated in each subplot.Mann-Whitney test was used.[Color figure can be viewed at wileyonlinelibrary.com]F I G U R E 2 (A) Association between urinary microbiota composition and clinicopathological parameters.Multivariate PERMANOVA test was used.(B) Principal coordinate analysis (PCoA) representing taxonomic compositions among patients that used or did not use 5-α-reductase inhibitors.Compositional distances were calculated using weighted UniFrac.Ellipsoids represent 95% confidence intervals.PERMANOVA test was used.[Color figure can be viewed at wileyonlinelibrary.com]

F
I G U R E 3 (A) Taxonomic compositions of the urinary microbiota of each patient with detectable microbiota included in the study.Only genera present at relative abundance ≥5% in at least two patients are depicted.(B) Multivariate association between highly prevalent genera (relative abundance ≥5% in ≥25% of patients) and clinicopathological parameters.The MaAsLin2 tool was used and the linear model coefficients are depicted as a surrogate of effect size.+: positive significant association; −: negative significant association.[Color figure can be viewed at wileyonlinelibrary.com]MARIOTTI ET AL.