Diversity of the midstream urine microbiome in adults with chronic kidney disease

Diversity of the midstream urine microbiome in adults with chronic kidney disease Purpose To examine the characteristics of the midstream urine microbiome in adults with stage 3–5 non-dialysis-dependent chronic kidney disease (CKD). Methods Patients with non-dialysis-dependent CKD (estimated glomerular filtration rate [eGFR] < 60 ml/min/1.73 m ) and diuretic use were recruited from outpatient nephrology clinics. Midstream voided urine specimens were collected using the clean-catch method. The bacterial composition was determined by sequencing the hypervariable (V4) region of the bacterial 16S ribosomal RNA gene. Extraction negative controls (no urine) were included to assess the contribution of extraneous DNA from possible sources of contamination. Midstream urine microbiome diversity was assessed with the inverse Simpson, Chao and Shannon indices. The diversity measures were further examined by demographic characteristics and by comorbidities. Results The cohort of 41 women and 36 men with detectable bacterial DNA in their urine samples had a mean age of 71.5 years (standard deviation [SD] 7.9) years (range 60–91 years). The majority were white (68.0%) and a substantial minority were African-American (29.3%) The mean eGFR was 27.2 (SD 13.6) ml/min/1.73 m . Most men (72.2%) were circumcised and 16.6% reported a remote history of prostate cancer. Many midstream voided urine specimens were dominated (> 50% reads) by the genera Corynebacterium (n = 11), Staphylococcus (n = 9), Streptococcus (n = 7), Lactobacillus (n = 7), Gardnerella (n = 7), Prevotella (n = 4), Escherichia_Shigella (n = 3), and Enterobacteriaceae (n = 2); the rest lacked a domi- nant genus. The samples had high levels of diversity, as measured by the inverse Simpson [7.24 (95% CI 6.76, 7.81)], Chao [558.24 (95% CI 381.70, 879.35)], and Shannon indices [2.60 (95% CI 2.51, 2.69)]. Diversity measures were generally higher in participants with urgency urinary incontinence and higher estimated glomerular filtration rate (eGFR). After controlling for demographics and diabetes status, microbiome diversity was significantly associated with estimated eGFR ( P < 0.05). Conclusions The midstream voided urine microbiome of older adults with stage 3–5 non-dialysis-dependent CKD is diverse. Greater microbiome diversity is associated with higher eGFR. Keywords Urinary microbiome · Chronic kidney disease · Microbiome diversity · Urinary symptoms * Michael J. Zilliox Urology, Stritch School of Medicine, Loyola University mizilliox@luc.edu Chicago, Maywood, IL, USA Hines VA Medical Center, Hines, IL, USA Department of Public Health Sciences, Loyola University Chicago, 2160 S. First Avenue, Maywood, IL 60153, USA Present Address: Evy Health, 325 Sharon Park Dr., Suite 522, Menlo Park, CA, USA Medicine, Division of Nephrology and Hypertension, Loyola University Chicago, Maywood, IL, USA Department of Reproductive Medicine, Division of Female Pelvic Medicine and Reconstructive Surgery, University Departments of Microbiology and Immunology, Loyola of California San Diego, La Jolla, CA, USA University Chicago, Maywood, IL, USA Obstetrics and Gynecology, Loyola University Chicago, Maywood, IL, USA Vol.:(0123456789) 1 3 1124 International Urology and Nephrology (2018) 50:1123–1130 absence of a clinical UTI would be diverse and that this Introduction diversity would be associated with eGFR, comorbidities, and urinary symptoms. Bacterial communities in the bladders of men and women without clinical urinary tract infections (UTI) have been discovered and emerging research suggests that the urinary microbiome may influence bladder health [1 –6]. Although Methods the vast majority of genitourinary bacteria, which includes bacteria from the urethra, vagina, vulva, and bladder, can- Study population not be cultured by standard clinical microbiology urine culture methods [7, 8], they can be identified using high Details of the study have been previously published [24]. throughput 16S rRNA sequencing or new enhanced urine The Loyola University Chicago Health Sciences Division culture techniques [2, 4, 8–10]. The presence of urinary Institutional Review Board approved the study, and all par- bacteria does not indicate UTI, as the bacteria that com- ticipants provided written informed consent. Participants prise the resident urinary microbiome differ from those were recruited from Loyola Outpatient Center Nephrology associated with clinical UTIs [5, 6, 11, 12]. While few clinics during a routine outpatient clinic visit from Novem- studies have examined the diversity of the urine microbi- ber 1, 2014, to June 30, 2015. Adults were invited to par- ome, current evidence links microbial diversity in cath- ticipate if they were aged ≥ 60 years with an eGFR < 60 ml/ eterized urine specimens with urgency urinary inconti- min/1.73  m based on the Modification of Diet in Renal nence (UUI) and response to treatment for lower urinary Disease [25] formula within four weeks of study enroll- tract symptoms, especially those associated with UUI [13, ment and were taking diuretics (thiazide or loop diuretics, 14]. Midstream voided urine microbiome diversity also potassium sparing diuretics, or aldosterone blockade medi- has been associated with lower urinary symptoms, as well cations) either solely or in addition to other medications for as with hormonal status and body mass index [15]. These hypertension management. Diuretic use was an inclusion preliminary findings are consistent with other non-urine criterion because the study was specifically designed to microbiome studies showing that human microbiome examine urinary symptoms in adults with CKD receiving diversity may influence health status. For example, low diuretics. Patients were excluded from the study if they were gut microbiome diversity is associated with inflammatory receiving any immunosuppression medications, had a his- bowel disease [16, 17], while high vaginal microbiome tory of a neobladder or ileal conduit, had used antibiotics diversity is associated with bacterial vaginosis [18]. within the past 4 weeks or had any active cancer treatment Chronic kidney disease (CKD) is a disease generally or urinary instrumentation within the past 6 months. A total associated with advanced age and comorbidities, such as of 135 patient participants were asked to enroll in the study, diabetes and obesity. The older age of adults with CKD 103 provided consent, and 99 contributed completed urinary along with accompanying comorbidities may influence symptom questionnaires. One participant withdrew consent midstream urine microbiome diversity and could influence from the study after completing the questionnaires, and five urinary symptoms and bladder health. Diabetes leads to participants did not provide a urine sample. An additional high ambient urinary glucose levels, which may increase 11 samples were contaminated, and 5 had too few sequence bacterial growth and influence urinary microbiome diver - reads. Thus, the analyzed cohort with detectable urine DNA sity. Severe reductions in estimated glomerular filtration included 77 participants (41 females and 36 males). rate (eGFR) also may influence bacterial growth via its effects on production of uromodulin, produced exclusively by the renal tubules [19–22] and promotes urinary excre- Midstream urinary microbiome tion of bacteria [23]. Urinary symptoms are troublesome for older adults with CKD, and we have previously dem- Midstream urines were obtained by the clean-catch method onstrated a high burden of urinary symptoms among older with a special emphasis on proper clinical collection pro- adults with CKD receiving diuretics [24]. cedures. The urine specimens were collected in 60 ml ster- This single-center pilot study examined the presence ile urine cups containing AssayAssure (Sierra Molecular, and diversity of the midstream voided urinary microbiome Incline Village, NV), a preservative that maintains bacte- among older adults with CKD who require diuretics for riostasis by retarding the reproduction and lysis of bacte- disease management. It further examined the association ria. The preservative also minimizes freeze/thaw damage to of the observed microbiome diversity with eGFR, comor- nucleic acids and acts like a chemical refrigerator by pre- bidities and urinary symptoms. We hypothesized that the serving specimens without refrigeration or freezing from midstream urine microbiome in adults with CKD in the 7 to 45 days. 1 3 International Urology and Nephrology (2018) 50:1123–1130 1125 reads (Illumina, San Diego, CA). Sequencing data was Demographic variables, medication use and medical history deposited in SRA accession SRP127363. All participants completed self-administered questionnaires Data analysis that queried cancer history (including prostate), and previ- ous surgeries (including prostatectomy or hysterectomy), Raw sequencing reads were processed using the Mothur software package (v. 1.31.2) [29] to remove low quality circumcision status (men only). Medication use, diabetes status (physician diagnosis and/or use of glucose lowering and chimeric sequences. The ribosomal database project classifier (RDP) [30] was used to generate taxonomic clas- medication), body mass index (BMI) and history of prostate cancer and benign prostatic hypertrophy were obtained from sifications of the sequencing reads from phylum to genus level. Samples were subsampled at a depth of 2000 sequenc- the electronic medical record. Stage of CKD was defined as stage 3 if eGFR was between 59 and 30 ml/min/1.73 m , ing reads; five samples were eliminated from downstream analysis due to not meeting this threshold. The taxonomic stage 4 if eGFR was between 15 and 29 ml/min/1.73 m and stage 5 if eGFR was < 15 ml/min/1.73 m [26]. Presence information was then used to split the sequences into bins using a 97% similarity threshold to assign operational taxo- of urgency urinary incontinence (UUI) and nocturia were assessed using the National Health and Nutrition Examina- nomic units (OTUs), using the clustering algorithm built into the Mothur software; this resulted in 8658 OTUs. The tion Surveys (NHANES) 2013 kidney condition items that assesses urinary leakage, amount, and frequency over the relative abundance of taxonomic classifications was calcu- lated at the genus level for each sample using the R pack- current month and over the previous 12 months [27]. Pres- ence of UUI was defined as responding “Yes” to the question age Phyloseq (v. 1.19.1). Taxa that did not constitute 1% of the total sequencing reads and/or greater than 40% of “During the past 12 months have you leaked or lost control of even a small amount of urine with an urge or pressure to a single sample’s sequencing reads comprise the taxa cat- egory “Other.” Samples were clustered using only the iden- urinate and you couldn’t get to the toilet fast enough?” Noc- turia was defined as reporting awakening 2 or more times tified taxa to assess variance using distance matrices and the Bray–Curtis index in the R package vegan (v. 2.4.3), per night to urinate. producing a dendrogram in which the short branches link similar samples and longer branches link more dissimilar DNA extraction samples; this segregates the dendrogram into distinct groups or clades. Unclassified sequencing reads and reads that were Genomic DNA was extracted from midstream voided urine samples using a validated mixture of lysozyme and grouped into the “Other” category were not used for clus- tering. The relative abundance graph was then aligned to mutanolysin [28] and the DNeasy Blood and Tissue Kit (Qiagen, Valencia, CA). All steps were performed in a UV the dendrogram to delineate the clades of the tree by the predominant organism, termed “urotype,” above a defined irradiated, HEPA-filtered PCR workstation to minimize potential contaminants. The amount of DNA in each sample threshold. We also plotted the proportion of sequences per genus across all patient samples to help visualize the total was quantified using the Qubit 2.0 Flurometer (Life Technol - ogies, Carlsbad, CA). The microbial composition was deter- number of genera represented in the midstream voided urine samples. Alpha diversity was assessed using three types of mined by sequencing the hypervariable (V4) region of the bacterial 16S ribosomal RNA gene, as previously described measurements. The total number of unique taxa was esti- mated by the CHAO measure, and the species richness was [2, 14]. The V4 region is about 250 base-pairs long and can be used to classify most bacteria to the genus level. measured by the inverse of the Simpson index. The Shannon index was used to assess both the richness and evenness or Sequencing-ready libraries were generated using a 2-step polymerase chain reaction (PCR). The V4 region was ampli- equality of representation of taxa within an environment. Higher values of these indices indicate higher diversity. fied using modified universal primers 515F and 806R. In a limited cycle PCR, Illumina sequencing adapters and dual- Statistical analysis index barcodes were added to the amplified targets. Extrac- tion negative controls (no urine) were included to assess the Descriptive statistics of the 77 study participants were exam- contribution of extraneous DNA from possible sources of contamination. The final PCR products were purified using ined by sex. The midstream urine microbiome diversity measures were examined by demographic characteristics and the Agencourt AMPure XP system (Beckman Coulter, Brea, CA) and pooled together in equimolar concentrations to cre- by comorbidities (diabetes, obesity, UUI and CKD stage) in the total sample. Due to a non-normal distribution, the ate a final library of barcoded fragments ready for sequenc- ing. The samples were sequenced on the Illumina MiSeq diversity measures were log-transformed for analyses. To determine significance of differences in diversity measures, bench-top sequencer rendering 250 base-pair paired-end 1 3 1126 International Urology and Nephrology (2018) 50:1123–1130 the unpaired t test was used to compare log-transformed Table 1 Characteristics of the study participants values by a given characteristic and geometric mean values Male (n = 36) Female (n = 41) were reported. The ANOVA test was used to examine sig- Age (years) 76.6 (7.6) 70.5 (8.2) nificant differences in log-transformed diversity measures Race across CKD stages 3, 4 and 5. If the overall test was signifi-  White 69.4% 63.4% cant (P < 0.05), then log-transformed diversity measures for  Black 25.0% 31.7% CKD stages 4 and 5 were compared to CKD stage 3 using  Other race 5.6% 4.9% an unpaired t test. Linear regression was used to examine Hispanic ethnicity 7.7% 2.4% the adjusted association of demographic characteristics and BMI (kg/m ) 30.8 (6.9) 32.0 (5.9) comorbidities with log-transformed diversity measures as Obesity 50.0% 58.5% the dependent variable. Models include age, BMI and eGFR Diabetes 69.4% 70.0% as continuous variables and male sex, diabetes, and white eGFR (ml/min/1.73 m ) 29.5 (6.2) 27.5 (13.5) race as dichotomous variables. Analyses were completed Stage 3 CKD 44.4% 41.5% using STATA v13. Stage 4 CKD 27.8% 36.0% Stage 5 CKD 27.8% 21.9% Urgency-UI 41.7% 51.2% Results Nocturia 77.8% 51.2% Circumcised 72.2% – The cohort of 41 women and 36 men with detectable bac- Prostate cancer history 16.6% – terial DNA in their midstream voided urine sample had a Benign prostatic hyperplasia 30.6% – mean age of 71.5 years (standard deviation [SD] 7.9) (range 60–91 years); race/ethnicity was reported as white (68.0%), UI urinary incontinence, eGFR estimated glomerular filtration rate, African-American (29.3%) or other race/ethnicity (2.7%). CKD chronic kidney disease; stage 3, 4 and 5 defined as eGFR 59–30, 29–15 and < 15  ml/min/1.73  m but not on dialysis, respec- Mean eGFR was 27.2 (SD 13.5) ml/min/1.73 m ; nocturia tively and UUI were reported by 63.6% and 46.8%, respectively. Nocturia defined as waking up at least twice per night to urinate Table 1 shows participant characteristics by sex. More than half (69.4% of men and 70.0% of women) had diabetes. Uri- a dominant urotype (> 50% of sequences) contained several nary symptoms were common in men and women. Nocturia was reported by 77.8 and 51.2% of men and women, respec- other prominent genera. In most samples, each genus was detected as a low proportion of all sequences. Table 2 shows tively, whereas UUI was reported by 41.7 and 51.2% of men and women, respectively. Most men (72.2%) were circum- the midstream urine microbiome diversity measures (Inverse Simpson, CHAO and Shannon indices) by demographic char- cised and 16.6% reported a remote history of prostate cancer. Figure  1 shows bacterial profiles in terms of relative acteristics and comorbidities. No significant difference was noted relative to age, sex, ethnicity or either diabetes or obesity abundance, clustered by Bray–Curtis dissimilarity and depicted as a dendrogram. Many samples were dominated status. Overall, diversity measures were higher among men versus women, but the differences did not meet statistical (> 50% reads) by a singe highly prevalent taxon, most often the genus Corynebacterium (n = 11), followed by the genera significance. In contrast, significant differences in diversity measures were observed for those reporting UUI versus those Staphylococcus (n = 9), Lactobacillus (n = 7), Gardnerella (n = 7), Streptococcus (n = 7), Prevotella (n = 4), and Esher- without UUI. In the total study group, diversity measures also dier ff ed significantly with respect to CKD stage (Table  3). The ichia_Shigella (n = 3), as well as the family Enterobacte- riaceae (n = 2). The remaining 27 samples had no dominant highest diversity measures were associated with CKD stage 3, whereas the lowest diversity measures were generally noted or highly prevalent taxon. In the total sample, the inverse Simpson and Chao1 indices, which measure richness (abun- with CKD stage 5. After controlling for demographics and diabetes status, only eGFR remained significantly and consist- dance), were 7.24 (95% CI 6.76, 7.81) and 558.24 (95% CI 381.70, 879.35), respectively. The Shannon index, which ently associated with diversity measures with higher eGFR associated with higher measures of diversity (Table 4). accounts for both richness and evenness (distribution) of present species, was 2.60 (95% CI 2.51, 2.69). This diversity was also observed in the across sample dis- tribution, which plots the proportion of sequences per genus Discussion across all 77 patient samples (Fig. 2). The total number of genera represented in the samples was 19, as represented by This study shows that the midstream voided urine micro- biomes of adults with CKD stage 3–5 are diverse. In the the x-axis. Very few urine samples were overwhelmingly dominated by a single genus. Thus, most urine samples with total sample and among women, the diversity measures were 1 3 International Urology and Nephrology (2018) 50:1123–1130 1127 Fig. 1 Genus level relative abundance. Each vertical bar represents ilarity index is calculated to generate the dendrogram and the thick the midstream urine of a study participant with percent of total clas- black line shows the urotype cut-off sified reads to the genus level on the y-axis. The Bray–Curtis dissim- dramatically higher than those from another study that eval- [13]. In that study, compared to women with less diverse uated midstream voided urine samples obtained by the clean urine microbiomes, those with more diverse microbiomes catch method in women with stress urinary incontinence either required a higher dose of the anticholinergic for the [15]. In that prior study of midstream voided urine samples same response or did not respond to the anticholinergic at in women without CKD, they reported an inverse Simpson all [13]. Another smaller study of 10 women with UUI and index of only 1.86 and a Chao index of 124.08, values sev- 10 women with normal bladder function found that lower eral fold lower (3.9 and 4.5 fold lower, respectively), than urine microbiome diversity from transurethral catheterized values reported in this study of adults with CKD. However, specimens was associated with increased UUI severity [31]. information on midstream urine microbiome diversity and Others have also reported lower midstream urine microbi- potential influencing factors remain very limited. ome diversity in women with interstitial cystitis compared In the current study, we also noted higher diversity meas- to women without urinary symptoms [32, 33]. Additionally, ures in the midstream urine microbiome from adults with one previous small study of midstream urine specimens from UUI versus those without UUI. The association of urine 16 persons age 20–70 + years showed that midstream urine microbiome diversity with urinary health measures remains microbiome diversity generally decreased with age with the poorly explored. One study analyzed urine obtained by lowest diversity noted in adults aged ≥ 70 years [34]. transurethral catheter and found that a more diverse urinary Our finding that the midstream urine microbiome diver - microbiome for women correlates with less robust treat- sity measures were generally lower with reduced eGFR, ment response for UUI with an anticholinergic treatment even after adjustment for demographics and diabetes status, 1 3 1128 International Urology and Nephrology (2018) 50:1123–1130 shows a potentially intriguing link between eGFR and the midstream urine microbiome diversity. This association will require additional study to determine contributing factors such as alterations in urinary antimicrobial peptides. In the renal tubules, the tacky uromodulin proteins stick to bacte- ria; this leads to formation of larger particles that are more readily excreted by the kidney [19, 23] and reduces the risk of a UTI [19–22]. It is possible that kidney function decline alters the secretion of uromodulin and other antimicrobial peptides into the urine, which may influence urinary micro- biome diversity, but this remains an untested hypothesis. While uromodulin concentrations generally decrease as GFR declines [35], large inter-individual variability in uromodu- lin urine concentrations exists. Other unmeasured factors, such as vaginal and urethral bacterial growth, may also be influenced by kidney disease and alter the urine microbiome diversity. Microbial assessments obtained using voided samples may include microbial contributions from adjacent pelvic niches, especially in women. Despite these limitations, the Fig. 2 The across sample distribution is shown for the genera found at > 1% abundance in at least one of the 77 samples. Note that very use of voided samples avoids the participant burden of uri- few genera constitute > 50% of the proportion of sequences in a sam- nary catheterization and the potential for alteration of the ple, highlighting the microbiome diversity in these patients urinary microbial community. The microbes detected in this Table 2 Geometric means of diversity measures by patient characteristics Inverse Simpson CHAO Shannon Age ≥ 75 years Age < 75 years Age ≥ 75 years Age < 75 years Age ≥ 75 years Age < 75 years 7.33 (6.84, 7.88) 7.20 (6.72, 7.77) 555.57 (380.82, 882.42) 558.21 (382.18, 877.70) 2.65 (2.55, 2.74) 2.57 (2.47, 2.67) Male Female Male Female Male Female 7.24 (6.76, 7.81) 6.09 (5.71, 6.52) 595.83 (424.68, 892.45) 527.19 (347.57, 868.0.1) 2.99 (2.90, 3.08) 2.30 (2.21, 2.40)* White race Non-white race White race Non-white race White race Non-white race 8.29 (7.73, 8.93) 5.67 (5.30, 6.11) 552.76 (379.60, 867.19) 560.72 (383.00, 883.74) 2.70 (2.61, 2.80) 2.43 (2.33, 2.52) Obese Non-obese Obese Non-obese Obese Non-obese 6.51 (6.08, 7.00) 8.24 (7.67, 8.90) 571.12 (378.98, 929.71) 543.16 (385.00, 822.51) 2.44 (2.34, 2.53) 2.81 (2.72, 2.90) Diabetic Non-diabetic Diabetic Non-diabetic Diabetic Non-diabetic 7.32 (6.83, 7.89) 7.61 (7.11, 8.20) 555.33 (383.40, 866.03) 560.93 (376.05, 901.96) 2.70 (2.61, 2.80) 2.41 (2.31, 2.51) UUI+ UUI− UUI+ UUI− *UUI+ UUI− 8.40 (7.81, 9.09) 6.36 (5.96, 6.83) 627.97 (420.06, 1009.73) 503.42 (350.92, 778.82) 2.83 (2.74, 2.95) 2.41 (2.31, 2.51) Nocturia+ Nocturia− Nocturia+ Nocturia− Nocturia+ Nocturia− 6.87 (6.39, 7.43) 8.09 (7.59, 8.67) 519.51 (366.83, 791.96) 627.20 (407.41, 1041.23) 2.59 (2.50, 2.69) 2.62 (2.52, 2.71) *P < 0.05 Table 3 Geometric means of diversity measures by chronic kidney disease (CKD) stage Diversity measure CKD stage 3 (n = 33) CKD stage 4 (n = 25) CKD stage 5 (n = 19) Overall P value Inverse Simpson index 10.48 (7.54, 14.44) 5.53 (3.86, 8.00)* 5.47 (3.42, 8.67)* 0.01 CHAO 614.00 (533.79, 699.24) 555.57 (464.05, 658.52)* 478.19 (507.76, 561.16)* 0.04 Shannon 2.99 (2.91, 3.08) 2.37 (2.27, 2.46)* 2.31 (2.20, 2.41) 0.4 *P < 0.05 versus CKD stage 3 1 3 International Urology and Nephrology (2018) 50:1123–1130 1129 Table 4 Results of multivariable linear regression analyses of log-transformed diversity measures Inverse Simpson+ CHAO+ Shannon index Beta (95% CI) P value Beta (95% CI) P value Beta (95% CI) P value Age 0.01 (−0.02, 0.04) 0.62 0.004 (−0.01, 0.02) 0.45 −0.01 (−0.02, 0.01) 0.9 Male 0.40 (0.30, 0.83) 0.07 0.13 (−0.05, 0.31) 0.14 0.28 (0.06, 0.51) 0.01 White race (vs. non-white) 0.35 (−0.12, 0.82) 0.15 −0.01 (−0.20, 0.18) 0.90 0.06 (−0.18, 0.30) 0.62 BMI 0.02 (−0.02, 0.05) 0.4 0.02 (0.001, 0.03) 0.04 0.001 (−0.017, 0.02) 0.88 eGFR 0.02 (0.01, 0.04) 0.008 0.01 (0.001. 0.02) 0.02 0.01 (.0002, 0.02) 0.04 Diabetes status −0.02 (−0.48, 0.45) 0.94 0.01 (−0.18, 0.20) 0.9 0.01 (−0.11, 0.38) 0.27 analysis cannot be confirmed to originate from the bladder References itself. However, the characteristics of the midstream urine 1. Brubaker L, Nager CW, Richter HE et al (2014) Urinary bacteria microbiome in women have similarities to urine microbi- in adult women with urgency urinary incontinence. Int Urogy- omes in catheterized urine specimens [4, 14, 15]. Few stud- necol J 25(9):1179–1184 ies have examined the voided urine microbiome character- 2. Hilt E, McKnley K, Pearce M et al (2014) Urine is not sterile: use istics in men [36, 37]. of enhanced urine culture techniques to detect resident bacterial flora in the adult female bladder. J Clin Microbiol 52(3):871–876 While these pilot data are hypothesis generating, we can- 3. Nienhouse V, Gao X, Dong Q et al (2014) Interplay between blad- not determine whether kidney disease influences the mid - der microbiota and urinary antimicrobial peptides: mechanisms stream urinary microbiome or vice versa. Because partici- for human urinary tract infection risk and symptom severity. PLoS pants contributed a single urine sample, we cannot comment ONE 9(12):e114185 4. Wolfe AJ, Toh E, Shibata N et al (2012) Evidence of unculti- on microbial stability over time. Future studies that examine vated bacteria in the adult female bladder. J Clin Microbiol the urine microbiome at several time points can determine 50(4):1376–1383 intra-individual variability of the urine microbiome and its 5. Brubaker L, Wolfe AJ (2017) The female urinary microbiota, change over time. urinary health and common urinary disorders. Ann Transl Med 5(2):34 In summary, this study shows that the midstream urine 6. Wolfe AJ, Brubaker L (2015) “Sterile urine” and the presence of microbiomes are diverse in adults with CKD and that this bacteria. European Urology Published online March 13, 2015 diversity appears to be lower with more advanced CKD. 7. Oliver JD (2005) The viable but nonculturable state in bacteria. J These findings should be confirmed in other populations Microbiol 43(Spec No):93–100 8. Price TK, Dune T, Hilt EE et al (2016) The clinical urine cul- with CKD. Future studies should also examine the charac- ture: enhanced techniques improve detection of clinically relevant teristics of the urinary microbiome using catheterized speci- microorganisms. J Clin Microbiol 54(5):1216–1222 mens and examine the association of the urinary microbiome 9. Siddiqui H, Nederbragt AJ, Lagesen K, Jeansson SL, Jakobsen KS characteristics with lower urinary tract symptoms in adults (2011) Assessing diversity of the female urine microbiota by high throughput sequencing of 16S rDNA amplicons. BMC Microbiol with CKD. 11:244 10. Fouts DE, Pieper R, Szpakowski S et al (2012) Integrated next- Funding NIDDK DK104718 (awarded to AJW and LB) and from an generation sequencing of 16S rDNA and metaproteomics differen- institutional Grant from Loyola University Chicago. tiate the healthy urine microbiome from asymptomatic bacteriuria in neuropathic bladder associated with spinal cord injury. J Transl Compliance with ethical standards Med 10:174 11. Brubaker L, Wolfe AJ (2017) Microbiota in 2016: associating infection and incontinence with the female urinary microbiota. Conflict of interest The authors declare that they have no conflict of Nat Rev Urol 14(2):72–74 interest. 12. Brubaker L, Wolfe A (2016) The urinary microbiota: a para- digm shift for bladder disorders? Curr Opin Obstet Gynecol Open Access This article is distributed under the terms of the Crea- 28(5):407–412 tive Commons Attribution 4.0 International License (http://creat iveco 13. Thomas-White KJ, Hilt EE, Fok C et  al (2016) Incontinence mmons.or g/licenses/b y/4.0/), which permits unrestricted use, distribu- medication response relates to the female urinary microbiota. Int tion, and reproduction in any medium, provided you give appropriate Urogynecol J 27(5):723–733 credit to the original author(s) and the source, provide a link to the 14. Pearce MM, Hilt EE, Rosenfeld AB et  al (2014) The female Creative Commons license, and indicate if changes were made. urinary microbiome: a comparison of women with and without urgency urinary incontinence. MBio 5(4):e01283-14 15. Thomas-White KJ, Kliethermes S, Rickey L et al (2017) Evalu- ation of the urinary microbiota of women with uncomplicated stress urinary incontinence. Am J Obstet Gynecol 216(1):55. e1–55.e16 1 3 1130 International Urology and Nephrology (2018) 50:1123–1130 16. Gevers D, Kugathasan S, Knights D, Kostic AD, Knight R, Xavier 27. National Center for Health Statistics 2013–2014 National Health RJ (2017) A microbiome foundation for the study of Crohn’s dis- and Nutrition Examination Survey (NHANES). Kidney Inventory ease. Cell Host Microbe 21(3):301–304 Questionnaire. Centers for disease control and prevention. http:// 17. Gevers D, Kugathasan S, Denson LA et al (2014) The treatment- www .cdc.go v/nc hs/data/nhane s/nhane s_13_14/KIQ_C API_H.pdf. naive microbiome in new-onset Crohn’s disease. Cell Host August 6, 2015. Cited September 2014 Microbe 15(3):382–392 28. Yuan S, Cohen DB, Ravel J, Abdo Z, Forney LJ (2012) Evaluation 18. Fredricks DN, Fiedler TL, Marrazzo JM (2005) Molecular iden- of methods for the extraction and purification of DNA from the tification of bacteria associated with bacterial vaginosis. N Engl human microbiome. PLoS ONE 7(3):e33865 J Med 353(18):1899–1911 29. Schloss PD (2013) Secondary structure improves OTU assign- 19. Ghirotto S, Tassi F, Barbujani G et al (2016) The uromodulin ments of 16S rRNA gene sequences. ISME J 7(3):457–460 gene locus shows evidence of pathogen adaptation through human 30. Wang Q, Garrity GM, Tiedje JM, Cole JR (2007) Naive Bayesian evolution. J Am Soc Nephrol 27(10):2983–2996 classifier for rapid assignment of rRNA sequences into the new 20. Raffi HS, Bates JM Jr, Laszik Z, Kumar S (2009) Tamm-horsfall bacterial taxonomy. Appl Environ Microbiol 73(16):5261–5267 protein protects against urinary tract infection by proteus mirabi- 31. Karstens L, Asquith M, Davin S et al (2016) Does the urinary lis. J Urol 181(5):2332–2338 microbiome play a role in urgency urinary incontinence and its 21. Mo L, Zhu XH, Huang HY, Shapiro E, Hasty DL, Wu XR (2004) severity? Front Cell Infect Microbiol 6:78 Ablation of the Tamm–Horsfall protein gene increases susceptibil- 32. Siddiqui H, Lagesen K, Nederbragt AJ, Jeansson SL, Jakobsen ity of mice to bladder colonization by type 1-fimbriated Escheri- KS (2012) Alterations of microbiota in urine from women with chia coli. Am J Physiol Renal Physiol 286(4):F795–F802 interstitial cystitis. BMC Microbiol 12:205 22. Bates JM, Raffi HM, Prasadan K et al (2004) Tamm–Horsfall 33. Abernethy MG, Rosenfeld A, White JR, Mueller MG, Lewicky- protein knockout mice are more prone to urinary tract infection: Gaupp C, Kenton K (2017) Urinary microbiome and cytokine rapid communication. Kidney Int 65(3):791–797 levels in women with interstitial cystitis. Obstet Gynecol 23. Anders HJ, Schaefer L (2014) Beyond tissue injury-damage- 129(3):500–506 associated molecular patterns, toll-like receptors, and inflam- 34. Lewis DA, Brown R, Williams J et al (2013) The human urinary masomes also drive regeneration and fibrosis. J Am Soc Nephrol microbiome; bacterial DNA in voided urine of asymptomatic 25(7):1387–1400 adults. Front Cell Infect Microbiol 3:41 24. Patel M, Vellanki K, Leehey DJ et al (2016) Urinary incontinence 35. Garimella PS, Biggs ML, Katz R et al (2015) Urinary uromodu- and diuretic avoidance among adults with chronic kidney disease. lin, kidney function, and cardiovascular disease in elderly adults. Int Urol Nephrol 48(8):1321–1326 Kidney Int 88(5):1126–1134 25. Levey AS, Stevens LA, Schmid CH et al (2009) A new equa- 36. Dong Q, Nelson DE, Toh E et al (2011) The microbial communi- tion to estimate glomerular filtration rate. Ann Intern Med ties in male first catch urine are highly similar to those in paired 150(9):604–612 urethral swab specimens. PLoS ONE 6(5):e19709 26. Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D 37. Nelson DE, Van Der Pol B, Dong Q et al (2010) Characteristic (1999) A more accurate method to estimate glomerular filtration male urine microbiomes associate with asymptomatic sexually rate from serum creatinine: a new prediction equation. Modifi- transmitted infection. PLoS ONE 5(11):e14116 cation of Diet in Renal Disease Study Group. Ann Intern Med 130(6):461–470 1 3 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png International Urology and Nephrology Springer Journals
Loading next page...
 
/lp/springer_journal/diversity-of-the-midstream-urine-microbiome-in-adults-with-chronic-9ObsXcx1dy
Publisher
Springer Netherlands
Copyright
Copyright © 2018 by The Author(s)
Subject
Medicine & Public Health; Nephrology; Urology
ISSN
0301-1623
eISSN
1573-2584
D.O.I.
10.1007/s11255-018-1860-7
Publisher site
See Article on Publisher Site

Abstract

Purpose To examine the characteristics of the midstream urine microbiome in adults with stage 3–5 non-dialysis-dependent chronic kidney disease (CKD). Methods Patients with non-dialysis-dependent CKD (estimated glomerular filtration rate [eGFR] < 60 ml/min/1.73 m ) and diuretic use were recruited from outpatient nephrology clinics. Midstream voided urine specimens were collected using the clean-catch method. The bacterial composition was determined by sequencing the hypervariable (V4) region of the bacterial 16S ribosomal RNA gene. Extraction negative controls (no urine) were included to assess the contribution of extraneous DNA from possible sources of contamination. Midstream urine microbiome diversity was assessed with the inverse Simpson, Chao and Shannon indices. The diversity measures were further examined by demographic characteristics and by comorbidities. Results The cohort of 41 women and 36 men with detectable bacterial DNA in their urine samples had a mean age of 71.5 years (standard deviation [SD] 7.9) years (range 60–91 years). The majority were white (68.0%) and a substantial minority were African-American (29.3%) The mean eGFR was 27.2 (SD 13.6) ml/min/1.73 m . Most men (72.2%) were circumcised and 16.6% reported a remote history of prostate cancer. Many midstream voided urine specimens were dominated (> 50% reads) by the genera Corynebacterium (n = 11), Staphylococcus (n = 9), Streptococcus (n = 7), Lactobacillus (n = 7), Gardnerella (n = 7), Prevotella (n = 4), Escherichia_Shigella (n = 3), and Enterobacteriaceae (n = 2); the rest lacked a domi- nant genus. The samples had high levels of diversity, as measured by the inverse Simpson [7.24 (95% CI 6.76, 7.81)], Chao [558.24 (95% CI 381.70, 879.35)], and Shannon indices [2.60 (95% CI 2.51, 2.69)]. Diversity measures were generally higher in participants with urgency urinary incontinence and higher estimated glomerular filtration rate (eGFR). After controlling for demographics and diabetes status, microbiome diversity was significantly associated with estimated eGFR ( P < 0.05). Conclusions The midstream voided urine microbiome of older adults with stage 3–5 non-dialysis-dependent CKD is diverse. Greater microbiome diversity is associated with higher eGFR. Keywords Urinary microbiome · Chronic kidney disease · Microbiome diversity · Urinary symptoms * Michael J. Zilliox Urology, Stritch School of Medicine, Loyola University mizilliox@luc.edu Chicago, Maywood, IL, USA Hines VA Medical Center, Hines, IL, USA Department of Public Health Sciences, Loyola University Chicago, 2160 S. First Avenue, Maywood, IL 60153, USA Present Address: Evy Health, 325 Sharon Park Dr., Suite 522, Menlo Park, CA, USA Medicine, Division of Nephrology and Hypertension, Loyola University Chicago, Maywood, IL, USA Department of Reproductive Medicine, Division of Female Pelvic Medicine and Reconstructive Surgery, University Departments of Microbiology and Immunology, Loyola of California San Diego, La Jolla, CA, USA University Chicago, Maywood, IL, USA Obstetrics and Gynecology, Loyola University Chicago, Maywood, IL, USA Vol.:(0123456789) 1 3 1124 International Urology and Nephrology (2018) 50:1123–1130 absence of a clinical UTI would be diverse and that this Introduction diversity would be associated with eGFR, comorbidities, and urinary symptoms. Bacterial communities in the bladders of men and women without clinical urinary tract infections (UTI) have been discovered and emerging research suggests that the urinary microbiome may influence bladder health [1 –6]. Although Methods the vast majority of genitourinary bacteria, which includes bacteria from the urethra, vagina, vulva, and bladder, can- Study population not be cultured by standard clinical microbiology urine culture methods [7, 8], they can be identified using high Details of the study have been previously published [24]. throughput 16S rRNA sequencing or new enhanced urine The Loyola University Chicago Health Sciences Division culture techniques [2, 4, 8–10]. The presence of urinary Institutional Review Board approved the study, and all par- bacteria does not indicate UTI, as the bacteria that com- ticipants provided written informed consent. Participants prise the resident urinary microbiome differ from those were recruited from Loyola Outpatient Center Nephrology associated with clinical UTIs [5, 6, 11, 12]. While few clinics during a routine outpatient clinic visit from Novem- studies have examined the diversity of the urine microbi- ber 1, 2014, to June 30, 2015. Adults were invited to par- ome, current evidence links microbial diversity in cath- ticipate if they were aged ≥ 60 years with an eGFR < 60 ml/ eterized urine specimens with urgency urinary inconti- min/1.73  m based on the Modification of Diet in Renal nence (UUI) and response to treatment for lower urinary Disease [25] formula within four weeks of study enroll- tract symptoms, especially those associated with UUI [13, ment and were taking diuretics (thiazide or loop diuretics, 14]. Midstream voided urine microbiome diversity also potassium sparing diuretics, or aldosterone blockade medi- has been associated with lower urinary symptoms, as well cations) either solely or in addition to other medications for as with hormonal status and body mass index [15]. These hypertension management. Diuretic use was an inclusion preliminary findings are consistent with other non-urine criterion because the study was specifically designed to microbiome studies showing that human microbiome examine urinary symptoms in adults with CKD receiving diversity may influence health status. For example, low diuretics. Patients were excluded from the study if they were gut microbiome diversity is associated with inflammatory receiving any immunosuppression medications, had a his- bowel disease [16, 17], while high vaginal microbiome tory of a neobladder or ileal conduit, had used antibiotics diversity is associated with bacterial vaginosis [18]. within the past 4 weeks or had any active cancer treatment Chronic kidney disease (CKD) is a disease generally or urinary instrumentation within the past 6 months. A total associated with advanced age and comorbidities, such as of 135 patient participants were asked to enroll in the study, diabetes and obesity. The older age of adults with CKD 103 provided consent, and 99 contributed completed urinary along with accompanying comorbidities may influence symptom questionnaires. One participant withdrew consent midstream urine microbiome diversity and could influence from the study after completing the questionnaires, and five urinary symptoms and bladder health. Diabetes leads to participants did not provide a urine sample. An additional high ambient urinary glucose levels, which may increase 11 samples were contaminated, and 5 had too few sequence bacterial growth and influence urinary microbiome diver - reads. Thus, the analyzed cohort with detectable urine DNA sity. Severe reductions in estimated glomerular filtration included 77 participants (41 females and 36 males). rate (eGFR) also may influence bacterial growth via its effects on production of uromodulin, produced exclusively by the renal tubules [19–22] and promotes urinary excre- Midstream urinary microbiome tion of bacteria [23]. Urinary symptoms are troublesome for older adults with CKD, and we have previously dem- Midstream urines were obtained by the clean-catch method onstrated a high burden of urinary symptoms among older with a special emphasis on proper clinical collection pro- adults with CKD receiving diuretics [24]. cedures. The urine specimens were collected in 60 ml ster- This single-center pilot study examined the presence ile urine cups containing AssayAssure (Sierra Molecular, and diversity of the midstream voided urinary microbiome Incline Village, NV), a preservative that maintains bacte- among older adults with CKD who require diuretics for riostasis by retarding the reproduction and lysis of bacte- disease management. It further examined the association ria. The preservative also minimizes freeze/thaw damage to of the observed microbiome diversity with eGFR, comor- nucleic acids and acts like a chemical refrigerator by pre- bidities and urinary symptoms. We hypothesized that the serving specimens without refrigeration or freezing from midstream urine microbiome in adults with CKD in the 7 to 45 days. 1 3 International Urology and Nephrology (2018) 50:1123–1130 1125 reads (Illumina, San Diego, CA). Sequencing data was Demographic variables, medication use and medical history deposited in SRA accession SRP127363. All participants completed self-administered questionnaires Data analysis that queried cancer history (including prostate), and previ- ous surgeries (including prostatectomy or hysterectomy), Raw sequencing reads were processed using the Mothur software package (v. 1.31.2) [29] to remove low quality circumcision status (men only). Medication use, diabetes status (physician diagnosis and/or use of glucose lowering and chimeric sequences. The ribosomal database project classifier (RDP) [30] was used to generate taxonomic clas- medication), body mass index (BMI) and history of prostate cancer and benign prostatic hypertrophy were obtained from sifications of the sequencing reads from phylum to genus level. Samples were subsampled at a depth of 2000 sequenc- the electronic medical record. Stage of CKD was defined as stage 3 if eGFR was between 59 and 30 ml/min/1.73 m , ing reads; five samples were eliminated from downstream analysis due to not meeting this threshold. The taxonomic stage 4 if eGFR was between 15 and 29 ml/min/1.73 m and stage 5 if eGFR was < 15 ml/min/1.73 m [26]. Presence information was then used to split the sequences into bins using a 97% similarity threshold to assign operational taxo- of urgency urinary incontinence (UUI) and nocturia were assessed using the National Health and Nutrition Examina- nomic units (OTUs), using the clustering algorithm built into the Mothur software; this resulted in 8658 OTUs. The tion Surveys (NHANES) 2013 kidney condition items that assesses urinary leakage, amount, and frequency over the relative abundance of taxonomic classifications was calcu- lated at the genus level for each sample using the R pack- current month and over the previous 12 months [27]. Pres- ence of UUI was defined as responding “Yes” to the question age Phyloseq (v. 1.19.1). Taxa that did not constitute 1% of the total sequencing reads and/or greater than 40% of “During the past 12 months have you leaked or lost control of even a small amount of urine with an urge or pressure to a single sample’s sequencing reads comprise the taxa cat- egory “Other.” Samples were clustered using only the iden- urinate and you couldn’t get to the toilet fast enough?” Noc- turia was defined as reporting awakening 2 or more times tified taxa to assess variance using distance matrices and the Bray–Curtis index in the R package vegan (v. 2.4.3), per night to urinate. producing a dendrogram in which the short branches link similar samples and longer branches link more dissimilar DNA extraction samples; this segregates the dendrogram into distinct groups or clades. Unclassified sequencing reads and reads that were Genomic DNA was extracted from midstream voided urine samples using a validated mixture of lysozyme and grouped into the “Other” category were not used for clus- tering. The relative abundance graph was then aligned to mutanolysin [28] and the DNeasy Blood and Tissue Kit (Qiagen, Valencia, CA). All steps were performed in a UV the dendrogram to delineate the clades of the tree by the predominant organism, termed “urotype,” above a defined irradiated, HEPA-filtered PCR workstation to minimize potential contaminants. The amount of DNA in each sample threshold. We also plotted the proportion of sequences per genus across all patient samples to help visualize the total was quantified using the Qubit 2.0 Flurometer (Life Technol - ogies, Carlsbad, CA). The microbial composition was deter- number of genera represented in the midstream voided urine samples. Alpha diversity was assessed using three types of mined by sequencing the hypervariable (V4) region of the bacterial 16S ribosomal RNA gene, as previously described measurements. The total number of unique taxa was esti- mated by the CHAO measure, and the species richness was [2, 14]. The V4 region is about 250 base-pairs long and can be used to classify most bacteria to the genus level. measured by the inverse of the Simpson index. The Shannon index was used to assess both the richness and evenness or Sequencing-ready libraries were generated using a 2-step polymerase chain reaction (PCR). The V4 region was ampli- equality of representation of taxa within an environment. Higher values of these indices indicate higher diversity. fied using modified universal primers 515F and 806R. In a limited cycle PCR, Illumina sequencing adapters and dual- Statistical analysis index barcodes were added to the amplified targets. Extrac- tion negative controls (no urine) were included to assess the Descriptive statistics of the 77 study participants were exam- contribution of extraneous DNA from possible sources of contamination. The final PCR products were purified using ined by sex. The midstream urine microbiome diversity measures were examined by demographic characteristics and the Agencourt AMPure XP system (Beckman Coulter, Brea, CA) and pooled together in equimolar concentrations to cre- by comorbidities (diabetes, obesity, UUI and CKD stage) in the total sample. Due to a non-normal distribution, the ate a final library of barcoded fragments ready for sequenc- ing. The samples were sequenced on the Illumina MiSeq diversity measures were log-transformed for analyses. To determine significance of differences in diversity measures, bench-top sequencer rendering 250 base-pair paired-end 1 3 1126 International Urology and Nephrology (2018) 50:1123–1130 the unpaired t test was used to compare log-transformed Table 1 Characteristics of the study participants values by a given characteristic and geometric mean values Male (n = 36) Female (n = 41) were reported. The ANOVA test was used to examine sig- Age (years) 76.6 (7.6) 70.5 (8.2) nificant differences in log-transformed diversity measures Race across CKD stages 3, 4 and 5. If the overall test was signifi-  White 69.4% 63.4% cant (P < 0.05), then log-transformed diversity measures for  Black 25.0% 31.7% CKD stages 4 and 5 were compared to CKD stage 3 using  Other race 5.6% 4.9% an unpaired t test. Linear regression was used to examine Hispanic ethnicity 7.7% 2.4% the adjusted association of demographic characteristics and BMI (kg/m ) 30.8 (6.9) 32.0 (5.9) comorbidities with log-transformed diversity measures as Obesity 50.0% 58.5% the dependent variable. Models include age, BMI and eGFR Diabetes 69.4% 70.0% as continuous variables and male sex, diabetes, and white eGFR (ml/min/1.73 m ) 29.5 (6.2) 27.5 (13.5) race as dichotomous variables. Analyses were completed Stage 3 CKD 44.4% 41.5% using STATA v13. Stage 4 CKD 27.8% 36.0% Stage 5 CKD 27.8% 21.9% Urgency-UI 41.7% 51.2% Results Nocturia 77.8% 51.2% Circumcised 72.2% – The cohort of 41 women and 36 men with detectable bac- Prostate cancer history 16.6% – terial DNA in their midstream voided urine sample had a Benign prostatic hyperplasia 30.6% – mean age of 71.5 years (standard deviation [SD] 7.9) (range 60–91 years); race/ethnicity was reported as white (68.0%), UI urinary incontinence, eGFR estimated glomerular filtration rate, African-American (29.3%) or other race/ethnicity (2.7%). CKD chronic kidney disease; stage 3, 4 and 5 defined as eGFR 59–30, 29–15 and < 15  ml/min/1.73  m but not on dialysis, respec- Mean eGFR was 27.2 (SD 13.5) ml/min/1.73 m ; nocturia tively and UUI were reported by 63.6% and 46.8%, respectively. Nocturia defined as waking up at least twice per night to urinate Table 1 shows participant characteristics by sex. More than half (69.4% of men and 70.0% of women) had diabetes. Uri- a dominant urotype (> 50% of sequences) contained several nary symptoms were common in men and women. Nocturia was reported by 77.8 and 51.2% of men and women, respec- other prominent genera. In most samples, each genus was detected as a low proportion of all sequences. Table 2 shows tively, whereas UUI was reported by 41.7 and 51.2% of men and women, respectively. Most men (72.2%) were circum- the midstream urine microbiome diversity measures (Inverse Simpson, CHAO and Shannon indices) by demographic char- cised and 16.6% reported a remote history of prostate cancer. Figure  1 shows bacterial profiles in terms of relative acteristics and comorbidities. No significant difference was noted relative to age, sex, ethnicity or either diabetes or obesity abundance, clustered by Bray–Curtis dissimilarity and depicted as a dendrogram. Many samples were dominated status. Overall, diversity measures were higher among men versus women, but the differences did not meet statistical (> 50% reads) by a singe highly prevalent taxon, most often the genus Corynebacterium (n = 11), followed by the genera significance. In contrast, significant differences in diversity measures were observed for those reporting UUI versus those Staphylococcus (n = 9), Lactobacillus (n = 7), Gardnerella (n = 7), Streptococcus (n = 7), Prevotella (n = 4), and Esher- without UUI. In the total study group, diversity measures also dier ff ed significantly with respect to CKD stage (Table  3). The ichia_Shigella (n = 3), as well as the family Enterobacte- riaceae (n = 2). The remaining 27 samples had no dominant highest diversity measures were associated with CKD stage 3, whereas the lowest diversity measures were generally noted or highly prevalent taxon. In the total sample, the inverse Simpson and Chao1 indices, which measure richness (abun- with CKD stage 5. After controlling for demographics and diabetes status, only eGFR remained significantly and consist- dance), were 7.24 (95% CI 6.76, 7.81) and 558.24 (95% CI 381.70, 879.35), respectively. The Shannon index, which ently associated with diversity measures with higher eGFR associated with higher measures of diversity (Table 4). accounts for both richness and evenness (distribution) of present species, was 2.60 (95% CI 2.51, 2.69). This diversity was also observed in the across sample dis- tribution, which plots the proportion of sequences per genus Discussion across all 77 patient samples (Fig. 2). The total number of genera represented in the samples was 19, as represented by This study shows that the midstream voided urine micro- biomes of adults with CKD stage 3–5 are diverse. In the the x-axis. Very few urine samples were overwhelmingly dominated by a single genus. Thus, most urine samples with total sample and among women, the diversity measures were 1 3 International Urology and Nephrology (2018) 50:1123–1130 1127 Fig. 1 Genus level relative abundance. Each vertical bar represents ilarity index is calculated to generate the dendrogram and the thick the midstream urine of a study participant with percent of total clas- black line shows the urotype cut-off sified reads to the genus level on the y-axis. The Bray–Curtis dissim- dramatically higher than those from another study that eval- [13]. In that study, compared to women with less diverse uated midstream voided urine samples obtained by the clean urine microbiomes, those with more diverse microbiomes catch method in women with stress urinary incontinence either required a higher dose of the anticholinergic for the [15]. In that prior study of midstream voided urine samples same response or did not respond to the anticholinergic at in women without CKD, they reported an inverse Simpson all [13]. Another smaller study of 10 women with UUI and index of only 1.86 and a Chao index of 124.08, values sev- 10 women with normal bladder function found that lower eral fold lower (3.9 and 4.5 fold lower, respectively), than urine microbiome diversity from transurethral catheterized values reported in this study of adults with CKD. However, specimens was associated with increased UUI severity [31]. information on midstream urine microbiome diversity and Others have also reported lower midstream urine microbi- potential influencing factors remain very limited. ome diversity in women with interstitial cystitis compared In the current study, we also noted higher diversity meas- to women without urinary symptoms [32, 33]. Additionally, ures in the midstream urine microbiome from adults with one previous small study of midstream urine specimens from UUI versus those without UUI. The association of urine 16 persons age 20–70 + years showed that midstream urine microbiome diversity with urinary health measures remains microbiome diversity generally decreased with age with the poorly explored. One study analyzed urine obtained by lowest diversity noted in adults aged ≥ 70 years [34]. transurethral catheter and found that a more diverse urinary Our finding that the midstream urine microbiome diver - microbiome for women correlates with less robust treat- sity measures were generally lower with reduced eGFR, ment response for UUI with an anticholinergic treatment even after adjustment for demographics and diabetes status, 1 3 1128 International Urology and Nephrology (2018) 50:1123–1130 shows a potentially intriguing link between eGFR and the midstream urine microbiome diversity. This association will require additional study to determine contributing factors such as alterations in urinary antimicrobial peptides. In the renal tubules, the tacky uromodulin proteins stick to bacte- ria; this leads to formation of larger particles that are more readily excreted by the kidney [19, 23] and reduces the risk of a UTI [19–22]. It is possible that kidney function decline alters the secretion of uromodulin and other antimicrobial peptides into the urine, which may influence urinary micro- biome diversity, but this remains an untested hypothesis. While uromodulin concentrations generally decrease as GFR declines [35], large inter-individual variability in uromodu- lin urine concentrations exists. Other unmeasured factors, such as vaginal and urethral bacterial growth, may also be influenced by kidney disease and alter the urine microbiome diversity. Microbial assessments obtained using voided samples may include microbial contributions from adjacent pelvic niches, especially in women. Despite these limitations, the Fig. 2 The across sample distribution is shown for the genera found at > 1% abundance in at least one of the 77 samples. Note that very use of voided samples avoids the participant burden of uri- few genera constitute > 50% of the proportion of sequences in a sam- nary catheterization and the potential for alteration of the ple, highlighting the microbiome diversity in these patients urinary microbial community. The microbes detected in this Table 2 Geometric means of diversity measures by patient characteristics Inverse Simpson CHAO Shannon Age ≥ 75 years Age < 75 years Age ≥ 75 years Age < 75 years Age ≥ 75 years Age < 75 years 7.33 (6.84, 7.88) 7.20 (6.72, 7.77) 555.57 (380.82, 882.42) 558.21 (382.18, 877.70) 2.65 (2.55, 2.74) 2.57 (2.47, 2.67) Male Female Male Female Male Female 7.24 (6.76, 7.81) 6.09 (5.71, 6.52) 595.83 (424.68, 892.45) 527.19 (347.57, 868.0.1) 2.99 (2.90, 3.08) 2.30 (2.21, 2.40)* White race Non-white race White race Non-white race White race Non-white race 8.29 (7.73, 8.93) 5.67 (5.30, 6.11) 552.76 (379.60, 867.19) 560.72 (383.00, 883.74) 2.70 (2.61, 2.80) 2.43 (2.33, 2.52) Obese Non-obese Obese Non-obese Obese Non-obese 6.51 (6.08, 7.00) 8.24 (7.67, 8.90) 571.12 (378.98, 929.71) 543.16 (385.00, 822.51) 2.44 (2.34, 2.53) 2.81 (2.72, 2.90) Diabetic Non-diabetic Diabetic Non-diabetic Diabetic Non-diabetic 7.32 (6.83, 7.89) 7.61 (7.11, 8.20) 555.33 (383.40, 866.03) 560.93 (376.05, 901.96) 2.70 (2.61, 2.80) 2.41 (2.31, 2.51) UUI+ UUI− UUI+ UUI− *UUI+ UUI− 8.40 (7.81, 9.09) 6.36 (5.96, 6.83) 627.97 (420.06, 1009.73) 503.42 (350.92, 778.82) 2.83 (2.74, 2.95) 2.41 (2.31, 2.51) Nocturia+ Nocturia− Nocturia+ Nocturia− Nocturia+ Nocturia− 6.87 (6.39, 7.43) 8.09 (7.59, 8.67) 519.51 (366.83, 791.96) 627.20 (407.41, 1041.23) 2.59 (2.50, 2.69) 2.62 (2.52, 2.71) *P < 0.05 Table 3 Geometric means of diversity measures by chronic kidney disease (CKD) stage Diversity measure CKD stage 3 (n = 33) CKD stage 4 (n = 25) CKD stage 5 (n = 19) Overall P value Inverse Simpson index 10.48 (7.54, 14.44) 5.53 (3.86, 8.00)* 5.47 (3.42, 8.67)* 0.01 CHAO 614.00 (533.79, 699.24) 555.57 (464.05, 658.52)* 478.19 (507.76, 561.16)* 0.04 Shannon 2.99 (2.91, 3.08) 2.37 (2.27, 2.46)* 2.31 (2.20, 2.41) 0.4 *P < 0.05 versus CKD stage 3 1 3 International Urology and Nephrology (2018) 50:1123–1130 1129 Table 4 Results of multivariable linear regression analyses of log-transformed diversity measures Inverse Simpson+ CHAO+ Shannon index Beta (95% CI) P value Beta (95% CI) P value Beta (95% CI) P value Age 0.01 (−0.02, 0.04) 0.62 0.004 (−0.01, 0.02) 0.45 −0.01 (−0.02, 0.01) 0.9 Male 0.40 (0.30, 0.83) 0.07 0.13 (−0.05, 0.31) 0.14 0.28 (0.06, 0.51) 0.01 White race (vs. non-white) 0.35 (−0.12, 0.82) 0.15 −0.01 (−0.20, 0.18) 0.90 0.06 (−0.18, 0.30) 0.62 BMI 0.02 (−0.02, 0.05) 0.4 0.02 (0.001, 0.03) 0.04 0.001 (−0.017, 0.02) 0.88 eGFR 0.02 (0.01, 0.04) 0.008 0.01 (0.001. 0.02) 0.02 0.01 (.0002, 0.02) 0.04 Diabetes status −0.02 (−0.48, 0.45) 0.94 0.01 (−0.18, 0.20) 0.9 0.01 (−0.11, 0.38) 0.27 analysis cannot be confirmed to originate from the bladder References itself. However, the characteristics of the midstream urine 1. Brubaker L, Nager CW, Richter HE et al (2014) Urinary bacteria microbiome in women have similarities to urine microbi- in adult women with urgency urinary incontinence. Int Urogy- omes in catheterized urine specimens [4, 14, 15]. Few stud- necol J 25(9):1179–1184 ies have examined the voided urine microbiome character- 2. Hilt E, McKnley K, Pearce M et al (2014) Urine is not sterile: use istics in men [36, 37]. of enhanced urine culture techniques to detect resident bacterial flora in the adult female bladder. J Clin Microbiol 52(3):871–876 While these pilot data are hypothesis generating, we can- 3. Nienhouse V, Gao X, Dong Q et al (2014) Interplay between blad- not determine whether kidney disease influences the mid - der microbiota and urinary antimicrobial peptides: mechanisms stream urinary microbiome or vice versa. Because partici- for human urinary tract infection risk and symptom severity. PLoS pants contributed a single urine sample, we cannot comment ONE 9(12):e114185 4. Wolfe AJ, Toh E, Shibata N et al (2012) Evidence of unculti- on microbial stability over time. Future studies that examine vated bacteria in the adult female bladder. J Clin Microbiol the urine microbiome at several time points can determine 50(4):1376–1383 intra-individual variability of the urine microbiome and its 5. Brubaker L, Wolfe AJ (2017) The female urinary microbiota, change over time. urinary health and common urinary disorders. Ann Transl Med 5(2):34 In summary, this study shows that the midstream urine 6. Wolfe AJ, Brubaker L (2015) “Sterile urine” and the presence of microbiomes are diverse in adults with CKD and that this bacteria. European Urology Published online March 13, 2015 diversity appears to be lower with more advanced CKD. 7. Oliver JD (2005) The viable but nonculturable state in bacteria. J These findings should be confirmed in other populations Microbiol 43(Spec No):93–100 8. Price TK, Dune T, Hilt EE et al (2016) The clinical urine cul- with CKD. Future studies should also examine the charac- ture: enhanced techniques improve detection of clinically relevant teristics of the urinary microbiome using catheterized speci- microorganisms. J Clin Microbiol 54(5):1216–1222 mens and examine the association of the urinary microbiome 9. Siddiqui H, Nederbragt AJ, Lagesen K, Jeansson SL, Jakobsen KS characteristics with lower urinary tract symptoms in adults (2011) Assessing diversity of the female urine microbiota by high throughput sequencing of 16S rDNA amplicons. BMC Microbiol with CKD. 11:244 10. Fouts DE, Pieper R, Szpakowski S et al (2012) Integrated next- Funding NIDDK DK104718 (awarded to AJW and LB) and from an generation sequencing of 16S rDNA and metaproteomics differen- institutional Grant from Loyola University Chicago. tiate the healthy urine microbiome from asymptomatic bacteriuria in neuropathic bladder associated with spinal cord injury. J Transl Compliance with ethical standards Med 10:174 11. Brubaker L, Wolfe AJ (2017) Microbiota in 2016: associating infection and incontinence with the female urinary microbiota. Conflict of interest The authors declare that they have no conflict of Nat Rev Urol 14(2):72–74 interest. 12. Brubaker L, Wolfe A (2016) The urinary microbiota: a para- digm shift for bladder disorders? Curr Opin Obstet Gynecol Open Access This article is distributed under the terms of the Crea- 28(5):407–412 tive Commons Attribution 4.0 International License (http://creat iveco 13. Thomas-White KJ, Hilt EE, Fok C et  al (2016) Incontinence mmons.or g/licenses/b y/4.0/), which permits unrestricted use, distribu- medication response relates to the female urinary microbiota. Int tion, and reproduction in any medium, provided you give appropriate Urogynecol J 27(5):723–733 credit to the original author(s) and the source, provide a link to the 14. Pearce MM, Hilt EE, Rosenfeld AB et  al (2014) The female Creative Commons license, and indicate if changes were made. urinary microbiome: a comparison of women with and without urgency urinary incontinence. MBio 5(4):e01283-14 15. Thomas-White KJ, Kliethermes S, Rickey L et al (2017) Evalu- ation of the urinary microbiota of women with uncomplicated stress urinary incontinence. Am J Obstet Gynecol 216(1):55. e1–55.e16 1 3 1130 International Urology and Nephrology (2018) 50:1123–1130 16. Gevers D, Kugathasan S, Knights D, Kostic AD, Knight R, Xavier 27. National Center for Health Statistics 2013–2014 National Health RJ (2017) A microbiome foundation for the study of Crohn’s dis- and Nutrition Examination Survey (NHANES). Kidney Inventory ease. Cell Host Microbe 21(3):301–304 Questionnaire. Centers for disease control and prevention. http:// 17. Gevers D, Kugathasan S, Denson LA et al (2014) The treatment- www .cdc.go v/nc hs/data/nhane s/nhane s_13_14/KIQ_C API_H.pdf. naive microbiome in new-onset Crohn’s disease. Cell Host August 6, 2015. Cited September 2014 Microbe 15(3):382–392 28. Yuan S, Cohen DB, Ravel J, Abdo Z, Forney LJ (2012) Evaluation 18. Fredricks DN, Fiedler TL, Marrazzo JM (2005) Molecular iden- of methods for the extraction and purification of DNA from the tification of bacteria associated with bacterial vaginosis. N Engl human microbiome. PLoS ONE 7(3):e33865 J Med 353(18):1899–1911 29. Schloss PD (2013) Secondary structure improves OTU assign- 19. Ghirotto S, Tassi F, Barbujani G et al (2016) The uromodulin ments of 16S rRNA gene sequences. ISME J 7(3):457–460 gene locus shows evidence of pathogen adaptation through human 30. Wang Q, Garrity GM, Tiedje JM, Cole JR (2007) Naive Bayesian evolution. J Am Soc Nephrol 27(10):2983–2996 classifier for rapid assignment of rRNA sequences into the new 20. Raffi HS, Bates JM Jr, Laszik Z, Kumar S (2009) Tamm-horsfall bacterial taxonomy. Appl Environ Microbiol 73(16):5261–5267 protein protects against urinary tract infection by proteus mirabi- 31. Karstens L, Asquith M, Davin S et al (2016) Does the urinary lis. J Urol 181(5):2332–2338 microbiome play a role in urgency urinary incontinence and its 21. Mo L, Zhu XH, Huang HY, Shapiro E, Hasty DL, Wu XR (2004) severity? Front Cell Infect Microbiol 6:78 Ablation of the Tamm–Horsfall protein gene increases susceptibil- 32. Siddiqui H, Lagesen K, Nederbragt AJ, Jeansson SL, Jakobsen ity of mice to bladder colonization by type 1-fimbriated Escheri- KS (2012) Alterations of microbiota in urine from women with chia coli. Am J Physiol Renal Physiol 286(4):F795–F802 interstitial cystitis. BMC Microbiol 12:205 22. Bates JM, Raffi HM, Prasadan K et al (2004) Tamm–Horsfall 33. Abernethy MG, Rosenfeld A, White JR, Mueller MG, Lewicky- protein knockout mice are more prone to urinary tract infection: Gaupp C, Kenton K (2017) Urinary microbiome and cytokine rapid communication. Kidney Int 65(3):791–797 levels in women with interstitial cystitis. Obstet Gynecol 23. Anders HJ, Schaefer L (2014) Beyond tissue injury-damage- 129(3):500–506 associated molecular patterns, toll-like receptors, and inflam- 34. Lewis DA, Brown R, Williams J et al (2013) The human urinary masomes also drive regeneration and fibrosis. J Am Soc Nephrol microbiome; bacterial DNA in voided urine of asymptomatic 25(7):1387–1400 adults. Front Cell Infect Microbiol 3:41 24. Patel M, Vellanki K, Leehey DJ et al (2016) Urinary incontinence 35. Garimella PS, Biggs ML, Katz R et al (2015) Urinary uromodu- and diuretic avoidance among adults with chronic kidney disease. lin, kidney function, and cardiovascular disease in elderly adults. Int Urol Nephrol 48(8):1321–1326 Kidney Int 88(5):1126–1134 25. Levey AS, Stevens LA, Schmid CH et al (2009) A new equa- 36. Dong Q, Nelson DE, Toh E et al (2011) The microbial communi- tion to estimate glomerular filtration rate. Ann Intern Med ties in male first catch urine are highly similar to those in paired 150(9):604–612 urethral swab specimens. PLoS ONE 6(5):e19709 26. Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D 37. Nelson DE, Van Der Pol B, Dong Q et al (2010) Characteristic (1999) A more accurate method to estimate glomerular filtration male urine microbiomes associate with asymptomatic sexually rate from serum creatinine: a new prediction equation. Modifi- transmitted infection. PLoS ONE 5(11):e14116 cation of Diet in Renal Disease Study Group. Ann Intern Med 130(6):461–470 1 3

Journal

International Urology and NephrologySpringer Journals

Published: Apr 12, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off