In-depth Microbiological Characterization of Urine From Subjects With Type 2 Diabetes Authors: Marco Calvigioni, Edoardo Biancalana, Diletta Mazzantini, Francesco Celandroni, Chiara Rossi, Alessandro Mengozzi, Emilia Ghelardi, Anna Solini Published: 2024 ABSTRACT AbstractContextLower urinary tract symptoms (LUTS) are common in type 2 diabetes (T2D), affecting quality of life and potentially leading to medication discontinuation. Among various factors contributing to LUTS, recent observations suggest a critical role of the urinary microbiota. Research on urinary dysbiosis in T2D remains underexplored.ObjectiveWe conducted a pilot study to investigate differences in the urinary microbiota between patients with T2D and healthy individuals and its potential indirect association with LUTS risk.MethodsThis case–control study included 50 patients with T2D and no LUTS, and 25 healthy controls. Microbial DNAs were extracted from urinary sediments and bacterial populations quantified by real-time quantitative polymerase chain reaction (PCR) and qualitatively investigated by 16S rRNA gene sequencing. Validation experiments with digital PCR were also performed.ResultsIn patients with T2D a higher total bacterial load and an increased abundance of Bacillota were found. After stratification by gender, these results were observed only in women. However, no significant quantitative differences were observed at the genus level. Alpha diversity analysis showed no significant differences between T2D and control groups, or by gender. At the species level, a substantial qualitative and often gender-dependent shift was present in individuals with T2D.ConclusionThe urinary microbiome of subjects with T2D was found to be different from that of healthy controls. Specifically, patients with T2D displayed higher total bacterial load and Bacillota levels, as well as qualitative changes in bacterial species. These changes suggested a dysbiotic condition of the urinary microbiota of subjects with T2D, with some gender-related differences. Although causality cannot be inferred, these findings highlight the impact of T2D on the urinary microbiota and its potential relevance in developing LUTS and, from a broader perspective, metabolic abnormalities.Keywords: urinary microbioma, type 2 diabetes, Bacillota Lower urinary tract symptoms (LUTS) are one of the most common complications of type 2 diabetes (T2D) (1, 2). T2D-associated LUTS often present as recurrent diseases and affect the patients’ quality of life (3, 4), sometimes leading to drug discontinuation, as in the case of SGLT2 inhibitors (SGLT2i) (5), and to a lack of trust in the medical community. Several risk factors for LUTS have been defined, such as urinary tract anatomy, diet, metabolic comorbidities (1), and impaired glucose control per se (3). However, prevention and treatment of LUTS remain a challenge (4), and this is particularly true for urine culture–negative LUTS (cnLUTS) whose pathogenesis is still unknown (2). The recent identification of a urinary microbial community (6), which should not be confused with the microbiota residing in the lower urinary tract that can contaminate outgoing urines, has disproved the dogma of urinary sterility and opened new avenues for LUTS management (7). Even at relatively low biomass, the urinary microbiota may play a critical protective role in maintaining the correct physiological and functional homeostasis of the urinary tract and host health. As for other body sites, the urinary microbiota could be influenced by the development of host disease, particularly metabolic diseases (8). However, only few studies have investigated whether the urinary microbiota was influenced by the presence of T2D and whether this could predispose to the future development of potentially related comorbidities, such as LUTS (9), with controversial findings (2, 10). Therefore, we conducted a pilot study to comprehensively investigate global and sex-related differences between the urinary microbiota of patients with T2D and healthy controls, and whether these changes could be potentially associated with an increased risk of LUTS. Materials and Methods Study Population In a case–control setting, 50 subjects with T2D were enrolled among those consecutively referred to the diabetes outpatient clinic at the University Hospital of Pisa. Inclusion criteria were age ≥18 and ≤75 years, both sexes, and Hb1Ac ≤8.0%. Exclusion criteria were (1) systemic diseases (including cancer) requiring anti-inflammatory or immunosuppressive therapies; (2) treatment with SGLT2i or DPP4i in the previous 4 weeks, to avoid the risk of qualitative/quantitative variations of urinary bacterial populations; (3) irritative and/or obstructive urinary or genital symptoms; (4) current or 4 weeks prior antibiotic treatment for any reason; (5) anatomical or functional abnormalities of the urinary tract (eg, incontinence, neurological bladder, bladder prolapse). Twenty-five age-matched subjects without diabetes and relatively healthy (ie, no systemic inflammatory diseases, no neoplasms, no ongoing therapies with nonsteroidal anti-inhibitory drugs or steroids) were enrolled as controls. The study was conducted in accordance with the local ethics committee (CEAVNO Pisa protocol number 17606) and the Declaration of Helsinki. All participants provided written informed consent. Study Protocol The day of the study, after an overnight fast, each patient underwent a full clinical assessment, and urine and venous blood samples were collected. Weight and height were measured, and body mass index was calculated. Blood pressure values were measured with a sphygmomanometer after resting for 5 minutes. For urine collection, voided, clean-catch and mid–late-stream urine samples were collected from all participants after appropriate instruction in sterile containers for routine urine analysis and culture sensitivity. All the urine samples were promptly verified to be negative for urine cultures using the standard culture procedure. For urine culture, samples were incubated at 37 °C for 48 hours on blood agar and MacConkey agar plates (Meus, Italy). Biochemical Characterization Plasma glucose concentration was measured by an enzymatic method using glucose oxidase and peroxidase on an automated analyzer (Hitachi 912, Roche, Switzerland). The HbA1c level was measured by high-performance liquid chromatography standardized to the diabetes control and complications trial assay. Total and high-density lipoprotein cholesterol, and triglycerides were measured by routine methods; low-density lipoprotein cholesterol was calculated according to Friedwald formula. Serum creatinine was measured by the modified Jaffe method, and estimated glomerular filtration rate was calculated from the CKD-EPI equation. DNA Extraction From Urinary Sediment Microbial genomic DNAs were extracted from urinary sediments using the QIAamp DNA Mini Kit (QIAGEN, Germany) according to the manufacturer's protocol. Extracted DNAs were subsequently quantified using the NanoDrop Lite Spectrophotometer (Thermo Fisher Scientific, USA) and normalized to a standard concentration of 2 ng DNA/µL. Absolute Quantification of Bacterial Populations by Real-Time Quantitative Polymerase Chain reaction Absolute abundances of total bacterial load and of the main phyla (ie, Actinomycetota, Bacillota, Bacteroidota, and Pseudomonadota) and genera (ie, Aerococcus, Corynebacterium, Gardnerella, Lactobacillus, Prevotella, and Streptococcus) were assessed on extracted DNAs by 16S rRNA gene-targeting quantitative polymerase chain reactions (qPCRs). Primer pairs targeting phylum- or genus-specific 16S rDNA regions were used. To evaluate the total bacterial abundance, a primer pair targeting a sequence of the 16S rRNA gene conserved in all bacteria was used (listed in Table S1). qPCR reactions were performed using the CFX96 Real-Time System (BioRad, USA) and CFX Maestro Software (version 2.3, BioRad). All reactions were carried out in duplicate in a 96-well plate with a final volume of 20 µL per well, including 8 µL of sterile water, 0.5 µL of each primer (10 µM), 10 µL of Luna Universal qPCR Master Mix (New England BioLabs, USA), and 1 µL of 2 ng DNA/µL template DNA. The amplification protocol was as follows: an initial denaturation step at 95 °C for 1 minute followed by 45 cycles comprising a denaturation step at 95 °C for 15 seconds, an annealing step at each primer set specific temperature (listed in Table S1 and Table S2 (11)) for 30 seconds, and an extension step at 72 °C for 10 seconds. Absolute quantifications were performed by comparing with calibration curves generated using external standards with known concentrations subjected to serial 10-fold dilutions ranging from 102 to 1010. For each standard curve, the R2 coefficient was higher than 0.98. Absolute quantification of total bacterial load was also confirmed by digital PCR in a subgroup of samples. The same primer pair targeting a sequence of the 16S rRNA gene conserved in all bacteria was used (see Table S1 (11)) on QIAcuity instrument (Qiagen, Germany) following the manufacturer's protocol; briefly, reactions were carried out in a final volume of 40 µL per well, using 0.8 µL of each primer (20 µM), 13.3 µL of Qiacuity Eva Green master mix, and 0.1 ng DNA as template. Results were generated by the QIAcuity Software Suite 2.1.8.23. 16S rRNA Gene Sequencing and Metagenomic Analysis 16S rRNA gene sequencing and subsequent data processing were performed by Novogene (Beijing, China). 16S rRNA gene regions V3-V4 were amplified with primers 341F (5′-CCTAYGGGRBGCASCAG-3′) and 806R (5′-GGACTACNNGGGTATCTAAT-3′). PCR products were detected with a 2% agarose gel electrophoresis and purified with the QIAGEN Gel Extraction Kit (QIAGEN). Sequencing libraries were generated using the NEBNext Ultra DNA Library Prep Kit for Illumina (New England BioLabs). Their quality was evaluated with the Qubit@2.0 Fluorometer (Thermo Fisher Scientific) and the BioAnalyzer 2100 System (Agilent Technologies, USA). Libraries were sequenced on the HiSeq Illumina platform, and 250 bp reads were generated. Raw data were filtered using QIIME2. Operational taxonomic units (OTUs) were clustered with a ≥97% similarity cut-off using Uparse. Representative sequences of each OTU were then analyzed using the GreenGene database based on the RDP classifier algorithm. Phylogenetic relations between OTUs were assessed with MUSCLE, and α-diversity analysis was performed using QIIME2 and R. Statistical Analysis All statistical analyses were performed using GraphPad Prism (version 9.5.0, Dotmatics, USA) and JMPPro 13.0 software (SAS Institute Inc., Cary, NC). Statistical significance was set at P < .05. Data are expressed as mean ± SD or median (interquartile range) for normally and not normally distributed variables. Group differences for continuous and categorical variables were tested using the Mann–Whitney test or Fisher exact test, respectively. Student's t-tests for unpaired data were applied to compare results obtained from qPCR and α-diversity indexes between healthy controls and subjects with T2D. Fisher's exact test was applied for each analyzed bacterial species to infer significant differences in the species prevalence in healthy controls and patients with T2D. Results The clinical and biochemical characteristics of subjects with T2D are shown in Table 1. Our population reflected a quite common phenotype, being overweight with mild hypertension and satisfactory metabolic control; estimated glomerular filtration rate was preserved. Clinical characteristics of healthy controls are reported elsewhere (Table S3 (11)). Table 1.Clinical and biochemical characteristics of the whole study cohort Malen = 34, 68%Femalen = 16, 32% P valueAge, years65 ± 964 ± 10nsDiabetes duration, years7 [2; 11]5 [3; 11]nsBMI, kg/m228.4 ± 5.328.3 ± 4.0nsSBP, mmHg142 ± 16144 ± 18nsDBP, mmHg78 ± 1085 ± 14nsFasting glucose, mg/dL130 ± 30148 ± 42nsHbA1c, %6.8 ± 0.77.1 ± 1.2nsSerum creatinine, mg/dL0.97 ± 0.240.75 ± 0.300.0104eGFR, mL/min/1.73 m287.3 ± 16.789.7 ± 25.0nsTotal cholesterol, mg/dL153 ± 40163 ± 36nsHDL cholesterol, mg/dL47 ± 1354 ± 15nsLDL cholesterol, mg/dL83 ± 3285 ± 32nsTriglycerides, mg/dL115 ± 54122 ± 65nsTreatment for diabetes, n (%)30, 88.2%14, 87.5%ns Metformin only21, 70%10, 71.4%ns GLP-1 RA5, 16.7%2, 14.3%ns Insulin4, 13.3%2, 14.3%nsAbbreviations: BMI, body mass index; DBP, diastolic blood pressure; eGFR, estimated glomerular filtration rate; HDL, high-density lipoprotein cholesterol; LDL, low-density lipoprotein cholesterol; ns, not significant; SBP, systolic blood pressure. Clinical and biochemical characteristics of the whole study cohort Abbreviations: BMI, body mass index; DBP, diastolic blood pressure; eGFR, estimated glomerular filtration rate; HDL, high-density lipoprotein cholesterol; LDL, low-density lipoprotein cholesterol; ns, not significant; SBP, systolic blood pressure. Absolute Quantification of Bacterial Populations by qPCR To assess the potential link between T2D and the development of distinct microbial communities in the upper urinary tract, genomic DNAs extracted from urinary sediments of healthy individuals and those with T2D were analyzed using 16S rRNA gene-targeted qPCR techniques. This aimed to gather quantitative data on the composition of the urinary microbiota. The total bacterial load (P = .0473) and the abundance of Bacillota (P = .0463) were found to be elevated in patients with diabetes compared with the healthy control group (Fig. 1A). Despite lacking statistical significance, there were observable upward trends for Actinomycetota, Bacteroidota, and Pseudomonadota (P > .05, Fig. 1A), implying potential involvement of these major bacterial phyla to the overall increase in total bacterial load. When participants were categorized by gender, the total bacterial load (P = .0361) and Bacillota (P = .0075) were elevated only for females subjects with diabetes (Fig. 1B). Men with T2D showed higher but not significant levels than their respective controls (P > .05, Fig. 1B). At the genus level, no differences were highlighted between healthy controls and people with diabetes for the tested genera, due to marked interindividual variabilities (Fig. 1C and 1D). Absolute abundances (DNA copies/mL of urine) of the main bacterial phyla and genera constituting the urinary microbiota obtained by real-time quantitative PCR. (A) Comparison of total bacterial load, Actinomycetota, Bacillota, Bacteroidota, and Pseudomonadota levels between healthy controls and subjects with type 2 diabetes (T2D). (B) Comparison of total bacterial load, Actinomycetota, Bacillota, Bacteroidota, and Pseudomonadota levels between healthy controls and subjects with T2D differentiating on the basis of gender (M: male; F: female). (C) Comparison of Aerococcus, Corynebacterium, Gardnerella, Lactobacillus, Prevotella, and Streptococcus levels between healthy controls and subjects with T2D. (D) Comparison of Aerococcus, Corynebacterium, Gardnerella, Lactobacillus, Prevotella, and Streptococcus levels between healthy controls and subjects with T2D differentiating on the basis of gender. *P < .05, **P < .01. The absolute quantification of total bacteria load repeated in a subgroup of samples using digital PCR confirmed the results obtained in qPCR (P = .14 by paired samples t-test) (Table S4 (11)). Qualitative Analysis of Bacterial Populations by 16S rDNA Gene Sequencing 16S rRNA-gene sequencing of the DNA extracted from urinary sediments showed that the α-diversity indexes (ie, Chao1 and Shannon) and the number of observed species were not significantly different between healthy controls and patients with T2D (Fig. 2A, 2D, and 2G). This indicates a similarity in the biodiversity and abundance of bacterial populations in both conditions. Gender-based stratification led to the same result (Fig. 2B, 2C, 2E, 2F, 2H, and 2I). α-Diversity indexes obtained from 16S-based metagenomic analysis: (A) Chao1 index considering all individuals and comparing (B) control and males with T2D and (C) control and females with T2D. (D) Shannon index considering all individuals and comparing (E) control and males with T2D and (F) control and females with T2D. (G) number of observed species considering all individuals and comparing (H) control and males with T2D and (I) control and females with T2D. Relative abundances obtained from the metagenomic analysis provided a wider overview of the composition of the urinary microbial consortia and confirmed the results obtained by qPCR for phyla (Fig. 3A and 3C) and genera (Fig. 3B and 3D). Relative abundances obtained from 16S-based metagenomic analysis. (A) Main phyla compared between healthy controls and subjects with T2D. (B) Main genera compared between healthy controls and subjects with T2D. (C) Main phyla compared between healthy controls and subjects with T2D differentiating on the basis of gender. (D) Main genera compared between healthy controls and subjects with T2D differentiating on the basis of gender. Table 2 shows the prevalence of the most abundant and relevant species identified in the 2 groups of subjects. Actinobaculum massiliense (P = .0286), Actinotignum urinale (P = .0032), Aerococcus christensenii (P = .0034), Anaerococcus hydrogenalis (P = .0118), Campylobacter ureolyticus (P = .0015), Corynebacterium aurimucosum (P = .0118), Corynebacterium coyleae (P < .0001), Corynebacterium glucuronolyticum (P < .0001), Enterococcus faecalis (P < .0001), Escherichia coli (P < .0001), Ezakiella massiliensis (P = .0032), Gardnerella vaginalis (P = .0003), Lactobacillus iners (P < .0001), Mobiluncus curtisii (P = .0004), Peptoniphilus grossensis (P < .0001), Porphyromonas uenonis (P = .0239), Prevotella amnii (P = .0125), Prevotella buccalis (P = .0008), Veillonella atypica (P = .0004), and Veillonella montpellierensis (P = .0034) were significantly more prevalent in patients with T2D than in healthy controls (Table 2). Interestingly, A. urinale, A. christensenii, C. coyleae, E. faecalis, E. coli, E. massiliensis, G. vaginalis, P. grossensis, P. amnii, and V. montpellierensis were only detectable in subjects with T2D and not present in the healthy controls. In contrast, Actinotignum schaalii (P < .0001), Bifidobacterium breve (P < .0001), Bifidobacterium longum (P = .0400), Bifidobacterium scardovii (P = .0104), Corynebacterium pyruviciproducens (P < .0001), Dialister propionifaciens (P = .0497), Facklamia hominis (P = .0005), Lactobacillus gasseri (P < .0001), Negativicoccus succinicivorans (P = .0060), Peptoniphilus lacrimalis (P = .0025), Schaalia radingae (P < .0001), Streptococcus agalactiae (P < .0001), Streptococcus anginosus (P = .0251), and Winkia neuii (P = .0154) were found to be significantly more prevalent in healthy controls than in people with diabetes (Table 2). Bifidobacterium scardovii was the only species not found in patients with T2D, while its prevalence in healthy controls was relatively limited (16.7%). Streptococcus anginosus was detected in all healthy controls, whereas Corynebacterium tuberculostearicum and Prevotella timonensis were detected in all patients with T2D. The other identified species were equally distributed among participants regardless of their healthy status (Table 2). Table 2.Prevalence of the main bacterial species in individuals with T2D and controls ControlT2D Bacterial speciesOut of 25%Out of 50% P valueSignificance Actinobaculum massiliense 833.33060.7.0286* Actinomyces urogenitalis 28.31325.0757ns Actinotignum schaalii 135047.1<.0001**** Actinotignum urinale 001428.6.0032** Aerococcus christensenii 001325.0034** Aerococcus urinae 2183.34692.9.4286ns Alloscardovia omnicolens 19753060.7.2051ns Anaerococcus hydrogenalis 416.72346.4.0118* Anaerococcus lactolyticus 625917.9.5534ns Anaerococcus vaginalis 625714.3.3377ns Atopobium vaginae 28.3714.3.7900ns Bifidobacterium breve 2391.723.6< .0001**** Bifidobacterium dentium 28.31325.0757ns Bifidobacterium longum 2183.33060.7.0400* Bifidobacterium scardovii 416.700.0104* Brevibacterium ravenspurgense 28.31121.4.1980ns Campylobacter ureolyticus 1558.34692.9.0015** Corynebacterium amycolatum 625917.9.5430ns Corynebacterium aurimucosum 416.72346.4.0118* Corynebacterium coyleae 002957.1< .0001**** Corynebacterium glucuronolyticum 6254589.3< .0001**** Corynebacterium pyruviciproducens 2183.31428.6< .0001**** Corynebacterium riegelii 416.71835.7.1065ns Corynebacterium tuberculostearicum 2391.750100.1081ns Dialister microaerophilus 1041.71121.4.1122ns Dialister propionifaciens 1766.72142.9.0497* Enterococcus faecalis 002142.9< .0001**** Escherichia coli 002550< .0001**** Ezakiella massiliensis 001428.6.0032** Facklamia hominis 19751632.1.0005*** Facklamia ignava 416.723.6.0910ns Gardnerella vaginalis 001835.7.0003*** Lactobacillus gasseri 2183.31428.6< .0001**** Lactobacillus iners 28.33467.9< .0001**** Limosilactobacillus fermentum 416.7510.7.4704ns Mobiluncus curtisii 416.73060.7.0004*** Negativicoccus succinicivorans 1766.71632.1.0060** Peptoniphilus duerdenii 1558.33060.7> .9999ns Peptoniphilus lacrimalis 2183.32346.4.0025** Peptoniphilus grossensis 002753.6< .0001**** Porphyromonas uenonis 28.31632.1.0239* Prevotella amnii 001121.4.0125* Prevotella bivia 13501835.7.2190ns Prevotella buccalis 28.32346.4.0008*** Prevotella colorans 416.71428.6.3902ns Prevotella corporis 416.7917.9> .9999ns Prevotella timonensis 2391.750100.1081ns Schaalia radingae 19751325< .0001**** Sneathia amnii 28.347.1> .9999ns Streptococcus agalactiae 2183.31632.1< .0001**** Streptococcus anginosus 251004182.1.0251* Streptococcus salivarius 19754182.1.5534ns Veillonella atypica 416.73060.7.0004*** Veillonella montpellierensis 001325.0034** Winkia neuii 19752346.4.0154*Abbreviation: ns, not significant. * P < .05, **P < .01, ***P < .001, ****P < .0001. Prevalence of the main bacterial species in individuals with T2D and controls Abbreviation: ns, not significant. * P < .05, **P < .01, ***P < .001, ****P < .0001. When the sequencing data were categorized by gender, differences in the prevalence of numerous species were observed between men and women (Table 3). Prevalences of A. schaalii (P = .0002), B. scardovii (P = .0476), F. hominis (P = .0084), N. succinicivorans (P = .0033), and P. lacrimalis (P = .0271) were reduced in men only, whereas A. christensenii (P = .0177), A. hydrogenalis (P = .0233), Brevibacterium ravenspurgense (P = .0024), C. aurimucosum (P = .0290), G. vaginalis (P = .0008), M. curtisii (P = .0010), P. buccalis (P = .0068), Prevotella colorans (P = .0177), and V. montpellierensis (P = .0177) increased in women only (Table 3). Facklamia ignava (P = .0434) and W. neuii (P = .0038) were the unique species reduced in women only. Bifidobacterium breve (P < .0001), C. ureolyticus (M: P = .0367; F: P = .0373), C. coyleae (M: P = .0023; F: P = .0002), C. glucuronolyticum (M: P < .0001; F: P = .0002), C. pyroviciproducens (M: P = .0001; F: P = .0038), E. faecalis (M: P = .0183; F: P = .0068), E. coli (M: P = .0092; F: P = .0002), L. gasseri (M: P = .0006; F: P = .0002), L. iners (M: P < .0001; F: P = .0233), P. grossensis (M: P = .0037; F: P = .0002), S. radingae (M: P = .0270; F: P = .0002), S. agalactiae (M: P = .0270; F: P = .0010), and V. atypica (M: P = .0027; F: P = .0121) were found to be significantly altered in people with diabetes compared with controls irrespective of gender (Table 3). Table 3.Prevalence of the main bacterial species in the control group and the T2D group stratified by gender Control maleT2D male Control femaleT2D female Bacterial speciesOut of 10%Out of 34% P valueSignificanceOut of 15%Out of 16% P valueSignificance Actinobaculum massiliense 4402058.8.4716ns426.71062.5.0732ns Actinomyces urogenitalis 00617.6.3102ns213.3743.8.1134ns Actinotignum schaalii 55000.0002***853.3425.1489ns Actinotignum urinale 001029.4.0853ns00425.0.1012ns Aerococcus christensenii 00720.6.1771ns00637.5.0177 * Aerococcus urinae 7703088.2.3223ns1493.316100.4839ns Alloscardovia omnicolens 6602161.8> .9999ns1386.7956.3.1134ns Anaerococcus hydrogenalis 2201441.2.2834ns213.3956.3.0233 * Anaerococcus lactolyticus 330823.5.6922ns320.016.3.3326ns Anaerococcus vaginalis 220411.8.6058ns426.7318.8.6851ns Atopobium vaginae 00411.8.0867ns213318.8> .9999ns Bifidobacterium breve 1010025.9< .0001****1386.700< .0001**** Bifidobacterium dentium 001029.4.0853ns213318.8> .9999ns Bifidobacterium longum 8801852.9.1607ns13871275.0.6539ns Bifidobacterium scardovii 22000.0476 * 213.300.2258ns Brevibacterium ravenspurgense 22038.8.3167ns00850.0.0024** Campylobacter ureolyticus 6603191.2.0367 * 9601593.8.0373 * Corynebacterium amycolatum 00411.8.5592ns640531.3.7160ns Corynebacterium aurimucosum 1101338.2.1322ns320.01062.5.0290 * Corynebacterium coyleae 001955.9.0023**001062.5.0002*** Corynebacterium glucuronolyticum 002985.3<.0001****64016100.0002*** Corynebacterium pyruviciproducens 101001132.4.0001***1173.3318.8.0038** Corynebacterium riegelii 00823.5.1666ns426.71062.5.0732ns Corynebacterium tuberculostearicum 99034100.2273ns1493.316100.4839ns Dialister microaerophilus 660823.5.0514ns426.7318.8.6851ns Dialister propionifaciens 8801544.1.0725ns960637.5.2890ns Enterococcus faecalis 001441.2.0183 * 00743.8.0068** Escherichia coli 001544.1.0092**001062.5.0002*** Ezakiella massiliensis 001029.4.0853ns00425.0.1012ns Facklamia hominis 8801029.4.0084**1173.3637.5.0732ns Facklamia ignava 0025.9> .9999ns426.700.0434 * Gardnerella vaginalis 00926.5.0923ns00956.3.0008*** Lactobacillus gasseri 101001338.2.0006***1173.316.3.0002*** Lactobacillus iners 002573.5< .0001****213.3956.3.0233 * Limosilactobacillus fermentum 22025.9.2176ns213.3318.8> .9999ns Mobiluncus curtisii 2201852.9.0834ns213.31275.0.0010*** Negativicoccus succinicivorans 9901235.3.0033**853.3425.0.1489ns Peptoniphilus duerdenii 6602367.6.7138ns960743.8.7152ns Peptoniphilus lacrimalis 9901647.1.0271 * 1280743.8.4795ns Peptoniphilus grossensis 001750.0.0037**001062.5.0002*** Porphyromonas uenonis 2201235.3.4618ns00425.0.1012ns Prevotella amnii 00720.6.1771ns00425.0.1012ns Prevotella bivia 5501441.2.7234ns853.3425.0.1489ns Prevotella buccalis 2201647.1.1607ns00743.8.0068** Prevotella colorans 440823.5.4218ns00637.5.0177* Prevotella corporis 440617.6.1992ns00318.8.2258ns Prevotella timonensis 1010034100> .9999ns1386.716100.2258ns Schaalia radingae 8801235.3.0270 * 1173.316.3.0002*** Sneathia amnii 0025.9> .9999ns213.3212.5> .9999ns Streptococcus agalactiae 8801235.3.0270 * 1386.7425.0.0010*** Streptococcus anginosus 101002676.5.1666ns151001593.8> .9999ns Streptococcus salivarius 8802676.5> .9999ns1173.31593.8.1719ns Veillonella atypica 001852.9.0027**4271275.0.0121* Veillonella montpellierensis 00720.6.1771ns00637.5.0177* Winkia neuii 7701955.9.4889ns1280425.0.0038**Abbreviation: ns: not significant. * P < .05, **P < .01, ***P < .001, ****P < .0001. Prevalence of the main bacterial species in the control group and the T2D group stratified by gender Abbreviation: ns: not significant. * P < .05, **P < .01, ***P < .001, ****P < .0001. Although male-associated bacterial species were generally less prevalent in male subjects with diabetes than in male healthy controls, the female-associated species were more prevalent in the T2D group than in controls (Table 3). Discussion In this proof of concept case–control study, we found that (1) total urinary bacterial load and Bacillota were higher in subjects with T2D than in healthy controls, but the amount of tested genera was not different; (2) bacterial species qualitatively differed in patients with T2D and healthy controls although no differences in α-diversity were evidenced between the 2 groups; (3) the dysbiotic shift in subjects with T2D is consistent with that observed in other body sites (ie, intestine) and includes species potentially involved in glucose intolerance and urinary tract infections; (4) profound gender differences exist consistently in both control and subjects with T2D. In the last decade, the importance of the microbiota and its interactions with the host has been shown to be one of the critical factors linking environmental stressors and metabolic diseases (12). Although most attention has been focused on the gut microbial communities, recent observations have suggested that the microbiota profile in different biological sites and body fluids, such as urine, may also be relevant in the pathogenesis of metabolic diseases and related comorbidities (13, 14). In this pilot observation, we report that the urinary microbiota of patients with T2D in good metabolic control, and no evidence of LUTS, is quantitatively and qualitatively different from that of healthy controls. We found that the total urinary bacterial load and Bacillota were higher in subjects with T2D. Bacillota are one of the signature phyla of the urinary microbiota (9) and their increase has been observed in subjects with metabolic diseases (15, 16). Except for members of Bifidobacterium, Lactobacillus, Veillonella, and a few other genera, several microbes residing in the urinary tract act as commensals but in predisposing altered conditions, including diabetes, may become opportunistic pathogens, thus shifting from a state of commensalism to a potential infection. In fact, while healthy individuals most commonly harbored beneficial bacteria in their urinary tract, such as bifidobacteria, lactobacilli, and streptococci (6), urines of patients with T2D displayed a higher prevalence of bacteria that are more likely to cause infections and discomforts in the genitourinary tract in both genders. Bifidobacteria, especially B. breve, strongly decrease in T2D, thus losing the protective effect of commensal bifidobacteria in the urinary tract. Conversely, the prevalence of pathogenic C. ureolyticus (17, 18), E. faecalis (19-21), and E. coli (22, 23) drastically increased in concomitance with diabetes. This shift towards a higher prevalence of potentially infectious species in patients with T2D could concur to the development of T2D-associated LUTS. Gender differences in the human microbiota are implicitly due to differences in anatomical structures and hormone levels (24). In particular, higher levels of Lactobacillus spp. have been observed in the urinary tract of aging women with T2D (10), confirming our finding of more lactobacilli in slightly younger females with T2D. The increase of lactobacilli could be mainly addressed to L. iners (10) prevailing in T2D in our study; conversely, the prevalence of L. gasseri, a probiotic bacterium also inhabiting the vagina of some women, was drastically reduced in T2D. In previous studies, L. gasseri showed anti-obesity and anti-insulin resistance effects (25, 26). Its reduction in the urinary tract in concomitance with T2D could be one of the factors leading to the increase of insulin resistance, typical of T2D. Prevalences of F. hominis and N. succinicivorans were selectively reduced in males with T2D. A few cases of genitourinary and urinary tract–derived blood infections due to these 2 species have been reported in recent years, including pyelonephritis, balanoposthitis, and interstitial cystitis (27-32). Although their role in human health still needs to be explored, the lowering of such bacteria in T2D may thus correlate with a reduced risk of infection in males. In parallel, in females with T2D W. neuii was found to be less prevalent than in healthy women. Winkia neuii is a common colonizer of the skin and female urinary tract, and rarely can determine infections in humans (33-37). As for F. hominis and N. succinicivorans, the reduction of such a potential pathogen in T2D could represent a positive aspect but, at the same time, a symptom of alteration of the vaginal microflora. Conversely, B. ravenspurgense, G. vaginalis, and M. curtisii were more present in women with T2D than in healthy individuals. While little is known about the role of B. ravenspurgense in humans (38), G. vaginalis and M. curtisii have been much more investigated over the years. Gardnerella vaginalis is part of the vaginal microbiota but it can frequently cause bacterial vaginosis in concomitance with a local dysbiosis (39-43). Mobiluncus curtisii is also commonly associated with recurrent bacterial vaginoses (41, 44) and some reports indicated the species as causative agent of bacteriemia and sepsis in adults (45, 46). Their increased prevalence in T2D should be regarded as a potentially alarming signal, due to their well-known pathogenic role in the female genitourinary tract. Our findings are also consistent with the few studies investigating the urinary microbiota in the context of T2D (2, 47), suggesting that T2D, as expected, is capable of inducing meaningful changes in the local urinary microenvironment leading to signature-specific dysbiosis. Our study shows several limitations. First, the case–control design implicitly limits the ability to infer causality. Furthermore, although we observed a peculiar dysbiosis of the microbiota in subjects with T2D, we cannot associate this with a direct risk of developing LUTS, due to the cross-sectional nature of this observation, and a follow-up longitudinal observation would be advisable. Second, the lack of information on lifestyle hinders the adjustment for confounders, such as low physical activity or poor dietary habits, all influencing the microbiota, at least in other major body sites (48, 49). Third, the prevalence of men in our population may have confounded the appreciation of the smaller effect in women, although this reinforces the female-specific signature that we observed. Finally, the small sample size might have hindered the appreciation of smaller effects. However, to the best of our knowledge this is the first comprehensive comparison of urine microbiota in males and females subjects with T2D, and consecutive recruitment from the outpatient clinic ensures a good representativity of a real-life cohort of individuals with T2D. If confirmed in larger studies, these findings might also represent relevant information to better tailor SGLT2i therapy, whose administration might sometimes be limited in compliance due to the occurrence of LUTS. In conclusion, the urinary microbiota results were radically altered in subjects with T2D and no urinary infections. Interestingly, these changes show that several species are involved in glucose dysregulation and pathogenesis of urinary tract infections and display a strong gender signature. Although further longitudinal studies could elucidate the mechanisms underlying this qualitative shift and, from a broader perspective, the consequences for the host phenotype, our pilot study supports the exploitation of the urinary microbiota in T2D as a potential target for LUTS prevention. Acknowledgments We thank Dr. Francesco Raggi and Dr. Mariarosaria Distaso for the invaluable laboratory help. Abbreviations lower urinary tract symptom operational taxonomic unit quantitative polymerase chain reaction type 2 diabetes Contributor Information Marco Calvigioni, Department of Translational Research and New Technologies in Medicine and Surgery, University of Pisa, I-56126 Pisa, Italy. Edoardo Biancalana, Department of Clinical and Experimental Medicine, University of Pisa, I-56126 Pisa, Italy. Diletta Mazzantini, Department of Translational Research and New Technologies in Medicine and Surgery, University of Pisa, I-56126 Pisa, Italy. Francesco Celandroni, Department of Translational Research and New Technologies in Medicine and Surgery, University of Pisa, I-56126 Pisa, Italy. Chiara Rossi, Department of Surgical, Medical, Molecular and Critical Area Pathology, University of Pisa, I-56126 Pisa, Italy. Alessandro Mengozzi, Department of Clinical and Experimental Medicine, University of Pisa, I-56126 Pisa, Italy; Center for Translational and Experimental Cardiology (CTEC), Department of Cardiology, Zurich University Hospital and University of Zurich, 8952 Zurich, Switzerland. Emilia Ghelardi, Department of Translational Research and New Technologies in Medicine and Surgery, University of Pisa, I-56126 Pisa, Italy. Anna Solini, Department of Surgical, Medical, Molecular and Critical Area Pathology, University of Pisa, I-56126 Pisa, Italy. Disclosures The authors have nothing to disclose. Data Availability Datasets generated during and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request. References Associated Data Data Citations Data Availability Statement Datasets generated during and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request.