Content uploaded by Kait Al
Author content
All content in this area was uploaded by Kait Al on Sep 22, 2020
Content may be subject to copyright.
Article
Ureteral Stent Microbiota Is Associated with Patient
Comorbidities but Not Antibiotic Exposure
Graphical Abstract
Highlights
dUreteral stents harbor a reproducible and patient-specific
microbiota
dPatient comorbidities (not UTIs or antibiotics) are correlated
with the microbiota
dUrine is not an accurate biomarker of stent microbiota or
encrustation
dStent-specific prophylactic antibiotic administration may
require recalibration
Authors
Kait F. Al, John D. Denstedt,
Brendan A. Daisley, ..., Gregor Reid,
Hassan Razvi, Jeremy P. Burton
Correspondence
jeremy.burton@lawsonresearch.com
In Brief
Ureteral stent-associated bacteria and
their impact on device encrustation are
underexplored. Using microbiota
sequencing and SEM/EDX, Al et al.
identify a reproducible, temporally stable,
patient-specific stent microbiota
associated with patient comorbidities but
not antibiotic exposure.
Al et al., 2020, Cell Reports Medicine 1, 100094
September 22, 2020 ª2020 The Author(s).
https://doi.org/10.1016/j.xcrm.2020.100094 ll
Article
Ureteral Stent Microbiota
Is Associated with Patient Comorbidities
but Not Antibiotic Exposure
Kait F. Al,
1,2
John D. Denstedt,
3
Brendan A. Daisley,
1,2
Jennifer Bjazevic,
3
Blayne K. Welk,
3
Stephen E. Pautler,
3
Gregory B. Gloor,
4
Gregor Reid,
1,2,3
Hassan Razvi,
3
and Jeremy P. Burton
1,2,3,5,6,
*
1
Centre for Human Microbiome and Probiotic Research, Lawson Health Research Institute, London, ON, Canada
2
Department of Microbiology and Immunology, The University of Western Ontario, London, ON, Canada
3
Division of Urology, Department of Surgery, The University of Western Ontario, London, ON, Canada
4
Department of Biochemistry, The University of Western Ontario, London, ON, Canada
5
Twitter: @our_microbiome
6
Lead Contact
*Correspondence: jeremy.burton@lawsonresearch.com
https://doi.org/10.1016/j.xcrm.2020.100094
SUMMARY
Ureteral stents are commonly used to prevent urinary obstruction but can become colonized by bacteria and
encrusted, leading to clinical complications. Despite recent discovery and characterization of the healthy uri-
nary microbiota, stent-associated bacteria and their impact on encrustation are largely underexplored. We
profile the microbiota of patients with typical short-term stents, as well as over 30 atypical cases (all with
paired mid-stream urine) from 241 patients. Indwelling time, age, and various patient comorbidities correlate
with alterations to the stent microbiota composition, whereas antibiotic exposure, urinary tract infection
(UTI), and stent placement method do not. The stent microbiota most likely originates from adhesion of resi-
dent urinary microbes but subsequently diverges to a distinct, reproducible population, thereby negating the
urine as a biomarker for stent encrustation or microbiota. Urological practice should reconsider standalone
prophylactic antibiotics in favor of tailored therapies based on patient comorbidities in efforts to minimize
bacterial burden, encrustation, and complications of ureteral stents.
INTRODUCTION
Ureteral stents are hollow conduits placed in the ureter from the
renal pelvis to the bladder and are commonly used in urological
practice to maintain urine drainage, which can be impeded by
obstruction caused by urolithiasis, stricture, or malignancy.
Due to constant contact with the urine, deposition of urinary
crystals and formation of bacterial biofilms on stents are com-
mon.
1,2
The formation of these encrustations can lead to compli-
cations, including infection, failure of the stent to drain urine,
more frequent device exchanges, and subsequent difficulty
with removal.
1
Indwelling ureteral stents have been associated
with the development of urinary tract infections (UTIs) and, in
more severe cases, pyelonephritis or urosepsis, which may be
related to single species or polymicrobial biofilms attached to
the stent.
3
The urinary tract harbors a unique microbiota that is distinct
from that of the gut in composition and is of much lower abun-
dance.
4,5
Based on recent evidence, it is likely that the different
sites and tissues throughout this system have different microbio-
tas.
5–7
The biofilms that form on urinary devices, such as stents
and catheters, may originate from this microbiota or contamina-
tion during insertion of the device.
8,9
Regardless of their origin,
the development of biofilms on these devices illustrates that
even very low numbers of bacteria can quickly take advantage
of the niche-altering foreign material to expand their populations.
Previous studies in stent patients have identified bacterial colo-
nization rates from 70% to 90%.
10,11
Bacteriuria can be common
in upward of 20% of patients with stents, and Escherichia coli is
often the most commonly cultured and identified organism.
10
Bacterial isolates derived from stent biofilms of clinical origin
often demonstrate resistance to multiple antibiotics, and anti-
biotic prophylaxis or concomitant antibiotic administration
does not appear to reduce the incidence of stent-related symp-
toms or UTI incidence or severity.
12–15
Due to these findings, the
use of antibiotic prophylaxis for stents is controversial, and it is
unclear how these compounds may impact the urinary micro-
biota during stenting.
The purpose of this study was to elucidate how the urinary
microbiota and other host factors impact bacterial colonization
and encrustation of indwelling ureteral stents. Specific interest
was taken in determining whether the urine microbiota accu-
rately recapitulated the adhered stent microbiota, allowing it to
act as a ‘‘biomarker’’ of potential device infection or encrusta-
tion. As such, we utilized 16S rRNA gene sequencing and scan-
ning electron microscopy (SEM) to characterize the urine and
Cell Reports Medicine 1, 100094, September 22, 2020 ª2020 The Author(s). 1
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
ll
OPEN ACCESS
stent microbiota from 241 patients that were sampled from a sin-
gle urology center. The large sample size and complementary
nature of the samples provide a high-resolution insight into bac-
terial attachment to ureteral stents under different clinical
scenarios.
RESULTS
Study Participant Demographics
Two hundred and forty-one participants were recruited
from a single center over approximately 1 year (Figure S1).
Patient demographic characteristics are summarized in Ta-
ble 1. The majority of samples were collected from typical
stenting events, where one double-J stent links between
one of the kidneys and the bladder. However, cases of bilat-
eral (stents between both the left and right kidney to the
bladder that indwell at the same time), longitudinal (multiple
consecutive devices in the same patient over time that were
collected independently), antegrade (placed downward
from the kidney percutaneously rather than upward from the
urethra), uncommonly long indwelling times, and various
encrustation levels were also examined. The majority of study
participants had an indwelling stent placed for treatment
related to stone disease, though in 22 participants, the stents
were necessitated for other reasons, including radiation-
induced ureteral stricture and the presence of retroperitoneal
masses.
The Stent Microbiota Is Dominated by Urinary Bacteria
Microbiota sequencing was performed on 1-cm-long slices of
both proximal and distal ends of the stents and bacterial pellets
from mid-stream urine samples (Figure S1). Stringent bio-
informatic filtering was performed on sequenced reads such
that 711 samples and 43 amplicon sequence variants (SVs)
were maintained for downstream analysis. The most abundant
SVs in stent and urine samples corresponded to the bacterial
genera Staphylococcus,Enterococcus,Lactobacillus, and Es-
cherichia (Table S1). The clinical samples (not including positive
and negative controls) contained an average of 13.5 SVs,
ranging from 3 to 31. There was a positive correlation between
read count and observed SVs (Figure S2A). Urine samples had
significantly more SVs observed (Figure S2B) and higher total
read count compared to stent samples (Figure S2C).
The sequence counts were center log ratio (CLR) transformed,
generating samplewise Aitchison distances.
16
A heatmap repre-
senting the relative abundance of CLR-transformed samples was
generated based on the Aitchison distance average linkage clus-
tering (Figure 1). The differences in microbiota composition at the
genus level were not driven by gender or sample type (urine or
stent). This was confirmed with a Benjamini-Hochberg-corrected
Table 1. Demographic and Clinical Characteristics of Study Participants
Participant Characteristic N = 241 (%)
Age 59.01 ±13.84 (range 22–90)
Gender 122 females (50.6), 119 males (49.4)
Indwelling time 22.79 ±34.61 days (range 2–394)
Body mass index 31.04 ±7.64 (range 17.00–60.00)
Reason for stent placement: urolithiasis 219 (90.9)
stricture 5 (2.1)
mass 10 (4.1)
other 7 (2.9)
Stent placement method: retrograde 219 (90.9)
antegrade 22 (9.1)
Patients with bilateral stents 11
Patients with multiple sequential stents over time 11 (9 patients with 2 devices and 2 patients
with 3 devices)
Time between sequential stent placements 63.5 ±28.6 days (range 25–105)
Use of antibiotics within the last 30 days from stent collection 225 (93.4)
Previous history of UTI 99 (41.1)
UTI within 7 days of stent placement or while indwelling 37 (15.4)
Diabetes 54 (22.4)
Hyperlipidemia 99 (41.1)
Hyperuricemia 16 (6.6)
Hypertension 122 (50.6)
Irritable bowel syndrome 17 (7.1)
Inflammatory bowel disease 28 (11.6)
Crohn’s disease 8 (3.3)
Ulcerative colitis 20 (8.3)
Pulmonary disease 72 (29.9)
2Cell Reports Medicine 1, 100094, September 22, 2020
Article
ll
OPEN ACCESS
Welch’s t test and principal-component analysis (PCA) per-
formed on the log-ratio transformed data at SV level (Figure S3),
where all samples (Figure S3A) and gender subsets (Figures S3B
and S3C) did not separate by sample type. Furthermore, samples
were more similar within participants than between participants
(Figure S4A). These findings demonstrate that the same mi-
crobes dominate both stent and urine samples from a single pa-
tient, and therefore, stent microbiota is likely to be urinary
derived.
Within Patients, the Stent Microbiota Is Stable and
Reproducible
Although sample types were dominated by similar organisms,
stent samples were further compared based on curl position to
determine whether the two curls (proximal curl in the kidney
and distal in the bladder) had a distinct microbial profile
compared to that of the patient’s urine (Figure S4B). Specifically,
beta diversity was measured by Aitchison distance to evaluate
the distance between proximal and distal curls from each stent,
between the urine and the stent curls of the same patient, and
finally between the urine and stent curls from all other patients
(Figure S4B). Within participants, stent curls had significantly
shorter distances between proximal and distal curls versus the
distance between stent curls to the urine, although distance be-
tween urine and stent curls from other participants was the
greatest. Thus, microbiota composition of the stent curls was
more similar to each other than either curl to the urine, indicating
the presence of a patient- and stent-specific microbiota that
does not directly reflect the composition of the urine but likely
derives from it.
From the devices recovered from participants with multiple
sequentially placed stents, many of the same organisms were
detected within the same individual over time (Figures 2 and
S5). Upon PCA, samples from the same individual generally
Figure 1. The Stent Microbiota Is Dominated by Urinary Bacteria
Samples are plotted left to right and ordered by the dendrogram. The dendrogram was generated from CLR-transformed read counts grouped by genera, based
on the average linkage clustering of per-sample Aitchison distance. Branches of the dendrogram are colored by sample type (stents are navy; urine is orange).
The heatmap represents the relative abundance of genera within samples (more abundant genera are lighter in color). Color coding below the heatmap cor-
responds to patient gender (females are pink; males are blue). An excerpt from the fourteen leftmost branches of the tree illustrates that, in general, samples from
the same individual group nearby on the dendrogram (see also Figure S4A). n = 667; 213 urine and 454 stent samples from 241 patients.
Cell Reports Medicine 1, 100094, September 22, 2020 3
Article
ll
OPEN ACCESS
cluster together (Figure 2A). Distance between samples from the
same participants at different time points was shorter than be-
tween samples from different participants (Figure 2B). There
were no significant effects of visit number on the samples (Ben-
jamini-Hochberg-corrected Wilcoxon rank sum test; data not
shown). Thus, on a per-patient basis, the stent microbiota is a
reproducible community over time, even over the course of up
to 150 days.
The microbiota of bilateral stents did not differ significantly, as
determined from eleven subjects (Figures 3 and S6). Within pa-
tients, both proximal and distal ends of bilateral stents clustered
separately from the urine (Figure 3A). Intraindividual samples
were closer together than interindividual samples (Figure 3B).
There was greatest spread between stent and urine samples
from the same individual, and the distance between stent sam-
ples was the shortest (Figure 3B), indicating the presence of a
distinct stent-specific microbiota.
Microbiota Variation of Ureteral Stents Correlates with
Patient Attributes
To determine whether patient and sample attributes (metadata)
correlated with microbiota variation, CLR-transformed sample-
wise Aitchison distances were evaluated.
17
With this approach,
several metadata factors were determined to be microbiota con-
founders, including stent indwelling time and patient comorbid-
ities (Table S2). These confounders were subsequently adjusted
for, and several statistically significant associations of micro-
biota variation remained between metadata characteristics and
taxonomic features as determined using a general linear model,
including patient age, body mass index, stent indwelling time,
pulmonary disease, hypertension, diabetes, irritable bowel syn-
drome (IBS), inflammatory bowel disease (IBD), and hyperlipid-
emia (Table 2).
18
To determine whether the degree of encrustation correlated
with microbial composition, stents were categorized based on
visible encrustation level (Table S3). There was a correlation be-
tween the degree of stent encrustation and the amount of time
stents were indwelling (Figure S7A). Shannon’s index of alpha di-
versity was negatively correlated with degree of stent encrusta-
tion (Figure S7B) and was lower in grade-3 encrusted stents
compared to grade 0 (Figure S7C). This suggests that the longer
a stent is indwelling, the more likely it will be to become en-
crusted and dominated by a less diverse microbial community.
Ten study participants had stents indwelling for greater than
2 months; these participants were determined to be outliers,
having stents significantly longer than the average indwelling
time of 23 days (ROUT method of outlier detection; Q =
0.1%).
19
The microbiota of these patients was not significantly
different when compared to all other samples or to samples
from the ten participants with the shortest indwelling durations
(Figure S8). However, as indwelling time increased, relative
abundance of the genera Finegoldia and Porphyromonas
increased, whereas Enterococcus and Escherichia decreased
(Table 2).
Antibiotic exposure was widespread among participants;
about 93% had been exposed to antibiotics within 30 days of
sample collection (Table 1). However, the microbiota of the few
participants without recent antibiotic exposure was not signifi-
cantly different than the majority (Table 2;Figure 4A). These par-
ticipants also did not differ by encrustation grade or Shannon’s
index of alpha diversity (Figures 4B and 4C).
Stents were evaluated based on their placement method. The
majority (90.9%; Table 1) of stents were placed in a retrograde
manner; however, the microbiota of the stents placed antegrade
(i.e., during percutaneous nephrolithotomy or nephroscopy)
was not significantly different than those placed retrograde
AB
C
Figure 2. Principal-Component Analysis of
Longitudinal Samples
(A) PCA was performed on CLR-transformed
Aitchison distances of longitudinally collected
samples. Each colored point represents a sample.
Distance between samples on the plot represents
differences in microbial community composition,
with 24.9% of total variance being explained by the
first two components shown. Strength and associ-
ation for genera (sequence variants) are depicted by
the length and direction of the gray arrows,
respectively. Points are colored by participant and
shaped by visit number. n = 64.
(B) Aitchison distance was greater between inter-
individual samples of the same type (n = 1,254) than
between samples from the same participant of the
same type at different visits (n = 37; Bonferroni-
corrected Mann-Whitney U test; p < 0.0001). Box-
plot whiskers represent minimum and maximum.
(C) Representative relative abundance bar plot of
three longitudinal stent patients. Each vertical bar
represents the relative SV abundance within a single
sample. Samples are grouped by participant.
Relative abundance of SVs is colored by genera,
with common genera shown in the legend. Days
between sample collections are listed in the green
visit code.
4Cell Reports Medicine 1, 100094, September 22, 2020
Article
ll
OPEN ACCESS
(Table 2;Figure 4D). These participants also did not differ by
encrustation grade or Shannon’s index of alpha diversity (Figures
4E and 4F).
About 15% of patients had culture-confirmed UTIs within
7 days of stent placement or throughout the stent indwelling
period (Table 1). The microbiota and degree of stent encrustation
in these patients was not significantly different than those
without UTIs (Figures 4G and 4H); however, Shannon’s index
of alpha diversity was lower for patients with UTIs (Figure 4I).
SEM/EDX Confirm the Presence of Urinary Crystals and
Bacterial Biofilms
SEM imaging of stent samples revealed characteristic crystal
phases and the presence of bacterial biofilms (Figure 5). Where
bacteria-like structures were visualized, their morphology
showed concordance with the genera that were present in the
sample based on microbiota sequencing (Figure 5, samples
014 and 195). The predominant substances on the stent surfaces
consisted of organic deposits and crystals. X-ray diffraction of
crystalline structures confirmed the presence of calcium oxalate
monohydrate in oval and multiple-twinning morphologies, cal-
cium oxalate dihydrate, calcium phosphate (Figure 5), uric
acid, and struvite (not shown).
DISCUSSION
This study characterized the urinary and device-adhered micro-
biota of ureteral stent patients. Importantly, we identified several
patient factors and comorbidities that correlated with stent mi-
crobiota composition and demonstrated that gender, antibiotic
exposure, and stent placement method did not have any signif-
icant associations with the urinary or stent microbiota. Our find-
ings also demonstrate consistency in stent microbial community
over time in patients with multiple stent placements and in both
Figure 3. Microbial Communities of Bilateral Stents
(A) PCA was performed on CLR-transformed Aitchison distances of samples from patients with bilateral indwelling stents. Each colored point represents a
sample. Distance between samples on the plot represents differences in microbial community composition, with 29.3% of total variance being explained by the
first two components shown. Strength and association for genera (sequence variants) are depicted by the length and direction of the gray arrows, respectively.
Points are colored by participant and shaped by sample type, which include both proximal and distal stent ends (n = 55).
(B) Aitchison distance was compared between interindividual samples and intraindividual samples. S, distance between stent samples from the same participant
(n = 154); U versus S, distance between urine and stent samples from the same participant (n = 39); all, all samples from a single participant (n = 234). All in-
traindividual comparisons had significantly shorter distances than the distance between samples from different individuals (n = 2,736; Bonferroni-corrected
Dunn’s tests; p < 0.0001). In intraindividual comparisons, the distance was shortest between stent samples and furthest from urine to stent samples (p = 0.022).
Boxplot whiskers represent minimum and maximum.
(C) Representative relative abundance bar plot of three bilateral stent patients. Each vertical bar represents the relative SV abundance within a single sample.
Samples are grouped by participant. Relative abundance of SVs is colored by genera, with common genera shown in the legend. Sample type is color coded.
Stents from the left side are denoted by ‘‘L’’ and from the right side by ‘‘R’’; urine is denoted by ‘‘U.’’
Cell Reports Medicine 1, 100094, September 22, 2020 5
Article
ll
OPEN ACCESS
left and right sides during bilateral stent placement, solidifying
the true presence of a reproducible stent microbiota and corrob-
orating previous findings from urinary catheters.
8
Interestingly,
intraindividual microbiotas of proximal and distal stent ends
were more similar than either stent end compared to the urine,
indicating that, although the same genera may be present in
the urine, it is not proportionally representative of the bacterial
community colonizing the stent.
Previous culture-based studies have shown that removed
stents are frequently culture positive despite patients exhibiting
a culture-negative urine profile.
1,20
Corroborating these findings,
we demonstrated that urinary and stent microbiotas were domi-
nated by similar bacterial genera; however, when investigating
the patterns on a per-patient basis, both proximal and distal
ends of the stent, as well as left and right bilateral stents, were
more similar to each other in microbiota composition than to
the urine. Additionally, culture-confirmed UTI was not associated
with increased encrustation level in this cohort, although a
caveat to this analysis was that only 15% of patients had
confirmed UTI. These findings illustrate that urine is not an accu-
rate biomarker of stent encrustation or representative of the
stent-adhered microbiota. Instead, the degree of stent encrusta-
tion was positively correlated with indwelling time and negatively
correlated with microbial diversity, indicating that the longer a
stent is indwelling, the greater the likelihood of it becoming en-
crusted and colonized with a less-diverse microbial community.
The urinary microbiome may extend as far as the renal collect-
ing system. This renal microbiota may contribute to the microbial
community of the proximal stent curl, or bacteria residing in the
bladder could adhere to the proximal stent curl during retrograde
insertion.
6,21
Bacteria are also thought to ascend from the distal
curl during movement of the stent while indwelling or by utilizing
active motility, a process that can occur quite rapidly.
11,22–26
Our
findings support these previous studies and suggest that the
stent-associated microbiota is derived from the urinary bladder,
based on the fact that no difference was observed in microbial
composition between antegrade or retrograde placement
method (though only 9% of stents were inserted in an antegrade
fashion). Further validation is provided by the stent-associated
microbiota being dominated by common urinary bacteria, which
is unlikely to have originated from skin or gut contamination dur-
ing placement.
27
Taken together, our findings suggest that,
although the urinary microbiota may originally seed onto the
stent, the stent microbial community is shaped and enriched
for competitively adherent bacteria and eventually diverges
significantly from the urine.
A previous microbiota study of stent encrustations demon-
strated a lack of association between ‘‘urotype’’ and patient con-
ditions, including age, gender, BMI, diabetes, urinary crystals,
and other factors.
28
The current study differed by utilizing a
non-partisan analysis method, whereby arbitrary community
groups or ‘‘urotypes’’ were not used and instead the entire data-
set was tested against all patient and sample characteristics.
With this approach, confounders were adjusted for and signifi-
cant associations between eight metadata features and genus-
level microbiota changes were established.
In concordance with previous studies, age was determined to
be significantly associated with increased Veillonella spp. and
decreased Lactobacillus spp.
29,30
In humans, Veillonella spp.
are commensals of the oral cavity and gastrointestinal and uro-
genital tracts, with the potential to cause opportunistic infec-
tions, including UTI.
31–35
Veillonella spp. are also commonly
associated with a more-diverse urinary microbiota, an attribute
often accompanied with urological disorders.
36,37
In contrast,
Lactobacillus spp. are commensals, with a robust body of evi-
dence detailing their beneficial effects in the healthy urinary tract
of both men and women.
4,27,38
It is unclear what effect the aging
process does to alter the urinary microbiota; however, the
observed decrease in protective urinary lactobacilli may account
for common stent-associated UTI and encrustation in older
populations.
39,40
Patients with IBS and IBD had increased stent and urinary
presence of Prevotella and Veillonella species and decreased
lactobacilli. These findings are consistent with previous literature
on the gut microbiota in these conditions.
41–43
These genera
have also been implicated in urogenital infections and disorders,
such as pelvic inflammatory disease.
44,45
Our findings add
further credence to the hypothesis that the gut microbiota is a
reservoir for the genito-urinary microbiota.
46,47
In the same
manner that gut colonization with uropathogenic Escherichia
coli (UPEC) increases the risk of UPEC UTI, the concurrence of
inflammatory urinary tract symptoms in patients with IBS may
Table 2. Significant Correlations between Metadata Attributes
and the Microbiota after Adjusting for Cofounders
Metadata Genus Coefficient
a
FDR
Age Campylobacter 0.250 0.034
Lactobacilli 0.370 0.002
Veillonella 0.457 0.002
Body mass index Actinotignum 0.499 0.048
Morganella 0.546 0.049
Stent indwelling time Enterococcus 0.292 0.008
Escherichia 0.309 0.034
Finegoldia 0.202 0.034
Porphyromonas 0.245 0.001
Pulmonary disease Campylobacter 0.627 0.004
Ezakiella 0.498 0.034
Hypertension Campylobacter 0.595 0.013
Klebsiella 0.592 0.034
Moryella 0.349 0.035
Diabetes Citrobacter 1.653 0.002
Enterococcus 1.086 0.034
IBS Prevotella 0.046 0.001
Veillonella 0.041 0.020
Crohn’s disease Lactobacillus 0.112 0.023
Staphylococcus 0.009 0.061
Ulcerative colitis Veillonella 0.036 0.020
Hyperlipidemia Aerococcus 0.424 0.049
Ureaplasma 0.816 0.003
FDR, false discovery rate.
a
Coefficients of association >0 are correlated with higher and <0 with
lower relative abundance of the specified genus.
6Cell Reports Medicine 1, 100094, September 22, 2020
Article
ll
OPEN ACCESS
Figure 4. Stent Microbiota and Encrustation Are Unchanged by Antibiotic Exposure, Device Placement Method, and UTI
PCA was performed on CLR-transformed Aitchison distances. Each colored point represents a sample. Distance between samples on the plot represents
differences in microbial community composition, with 20.5% of total variance being explained by the first two components shown. Strength and association for
genera (sequence variants) are depicted by the length and direction of the gray arrows, respectively.
(A, D, and G) Samples are colored based on (A) whether the study participant had exposure to antibiotics within the last 30 days prior to sample collection (blue) or
not (pink), (D) whether the stents were placed in a retrograde (purple) or antegrade (orange) manner, and (G) whether the participant had a UTI within 7 days of
stent placement or throughout the indwelling period (orange) or not (green). Ellipses represent the 95% confidence interval.
(legend continued on next page)
Cell Reports Medicine 1, 100094, September 22, 2020 7
Article
ll
OPEN ACCESS
be explained.
46,48–51
A limitation of the current study was that
lower urinary tract symptoms and extended quantitative urine
culture were not evaluated in the stent patient population.
52
Future studies should look to correlate the urinary and stent mi-
crobiota with device encrustation, patient outcomes, and further
serum and urinary parameters throughout the indwelling period if
urological patients with IBS/IBD experience increased stent-
associated complications in addition to the documented urinary
tract symptoms.
Of the various comorbidities thatsignificantly correlated with the
stent microbiota, it was notable that they originated from distant
sites (pancreas, respiratory tract, liver, and gastrointestinal tract),
suggesting some common physiological denominator. Potentially,
it is the gastrointestinal tract that is altered by these conditions,
with systemic consequences of bacterial translocation. For this
reason, it is feasible that microbiota-based treatment, including
oral consumption of probiotic lactobacilli, or even fecal microbiota
transplantation could be of therapeutic potential to stent patients.
(B, E, and H) The degree of stent encrustation was compared between groups of interest. Groups were not significantly different by two-tailed Mann-Whitney U
test.
(C, F, and I) Shannon’s index of alpha diversity was not significantly different between antibiotic (C) or placement (F) groups, but patients with a UTI had lower
diversity than those without (I; two-tailed Mann-Whitney U test; p = 0.002). Boxplot whiskers represent minimum and maximum.
Figure 5. SEM Confirms the Presence of Uri-
nary Crystals and Bacterial Biofilms
Representative scanning electron micrographs of
stent encrustations illustrating typical bacterial bio-
films and crystal morphologies. Based on micro-
biota sequencing, bacteria visible (white arrow-
heads) likely correspond to the genera (014)
Lactobacillus and (195) Enterococcus. X-ray
diffraction of crystalline microstructures (white ar-
rows) correspond to calcium oxalate dihydrate (010
and 019a), calcium oxalate monohydrate in oval
(022) and multiple-twinning (095) morphologies, and
calcium phosphate (019b and 195). Scale bars
represent 20 mm.
In addition to many other maladies, these
treatments have shown efficacy against
IBS/IBD symptoms, urogenital infections in
the elderly, and recurrent UTI.
53–56
The majority of stents imaged by SEM
revealed encrustations composed of
organic material and urinary crystals,
although bacteria were only visualized in
a small number of cases. This was ex-
pected due to the low bacterial load pre-
sent in urinary samples, as well as the
high proportion of urolithiasis patients
among the study participants.
57,58
If these
organisms were involved in crystal deposi-
tion on the biomaterial, the urologist should
ensure device removal within 3 weeks,
given the positive correlation between
indwelling time and stent encrustation.
Due to the low bacterial biomass nature
of the samples collected, this study utilized
stringent pre-sequencing processing methods in addition to the
application of conservative bioinformatic cutoffs and analysis
tools in order to minimize contamination effects.
59,60
In future
studies, quantification of total 16S rRNA gene copies by qPCR
or the use of extended quantitative urine culture may comple-
ment and validate microbiota analysis of urinary and ureteral
stent samples.
28,52
Nevertheless, the detection of reproducible,
patient-specific, stent microbiota signatures provides confi-
dence that our findings are not due to contamination.
In summary, this study has characterized the urinary and stent
microbiota of ureteral stent patients from a single center over a 1-
year period, uncovering the importance of patient characteristics
in explaining microbiota variation. Actions taken by the physi-
cian, such as antibiotic exposure and stent placement method,
had no association with the microbiota in these samples but co-
morbidities and patient age did. The stent microbiota appears to
originate from patient-specific adhesion of urinary microbes and
subsequently diverges to a distinct reproducible population,
8Cell Reports Medicine 1, 100094, September 22, 2020
Article
ll
OPEN ACCESS
thereby negating the urine as an accurate biomarker for stent
encrustation or microbiota status. These findings suggest that
timely stent removal is likely the most important action to be
taken by the treating urologist in preventing encrustation and
that stent-specific antibiotic administration practices need
recalibration. Elderly patients or those diagnosed with pulmo-
nary disease, hypertension, diabetes, or IBS/IBD may need
closer evaluation to minimize stent- and microbiota-associated
complications.
Limitations of Study
These data were derived at a single center from a heterogeneous
patient population. Thus, the identified metadata factors associ-
ated with microbiota variability in this study may be cohort spe-
cific. For this reason, the subgroup analyses should be
confirmed in a larger study population. Additionally, no standard-
ized method exists for determining stent encrustation; the
approach taken in this study, which was developed and vali-
dated internally, should be taken into consideration when
comparing this work with future studies of the ureteral stent
microbiota.
STAR+METHODS
Detailed methods are provided in the online version of this paper
and include the following:
dKEY RESOURCES TABLE
dRESOURCE AVAILABILITY
BLead Contact
BMaterials Availability
BData and Code Availability
dEXPERIMENTAL MODEL AND SUBJECT DETAILS
dMETHOD DETAILS
BSample processing and DNA extraction
B16S rRNA gene sequencing
BSEM and X-ray diffraction spectroscopy
dQUANTIFICATION AND STATISTICAL ANALYSIS
SUPPLEMENTAL INFORMATION
Supplemental Information can be found online at https://doi.org/10.1016/j.
xcrm.2020.100094.
ACKNOWLEDGMENTS
We thank Dr. Todd Simpson from the Western University Nanofabrication fa-
cility for SEM and X-ray diffraction spectroscopy analysis. We also thank Linda
Nott and Dr. Patricia Rosas-Arellano for logistical support and sample collec-
tion. K.F.A. was supported by an Ontario Graduate Scholarship. This work was
supported by the W. Garfield Weston Foundation.
AUTHOR CONTRIBUTIONS
J.P.B. and H.R. designed the study. J.D.D., B.K.W., S.E.P., and H.R. acquired
clinical samples. K.F.A. performed experiments. K.F.A., B.A.D., and G.B.G.
performed data analysis. K.F.A. and J.P.B. wrote the paper. All authors
contributed to the editing and revision of the paper and approved the final
manuscript.
DECLARATION OF INTERESTS
The authors declare no competing interests.
Received: February 28, 2020
Revised: May 29, 2020
Accepted: August 21, 2020
Published: September 22, 2020
REFERENCES
1. Riedl, C.R., Plas, E., H€
ubner, W.A., Zimmerl, H., Ulrich, W., and Pfl€
uger, H.
(1999). Bacterial colonization of ureteral stents. Eur. Urol. 36, 53–59.
2. Zumstein, V., Betschart, P., Albrich, W.C., Buhmann, M.T., Ren, Q.,
Schmid, H.P., and Abt, D. (2017). Biofilm formation on ureteral stents -
Incidence, clinical impact, and prevention. Swiss Med. Wkly. 147,
w14408.
3. Lange, D., Bidnur, S., Hoag, N., and Chew, B.H. (2015). Ureteral stent-
associated complications–where we are and where we are going. Nat.
Rev. Urol. 12, 17–25.
4. Whiteside, S.A., Razvi, H., Dave, S., Reid, G., and Burton, J.P. (2015). The
microbiome of the urinary tract–a role beyond infection. Nat. Rev. Urol. 12,
81–90.
5. Wolfe, A.J., Toh, E., Shibata, N., Rong, R., Kenton, K., Fitzgerald, M., Mu-
eller, E.R., Schreckenberger, P., Dong, Q., Nelson, D.E., and Brubaker, L.
(2012). Evidence of uncultivated bacteria in the adult female bladder.
J. Clin. Microbiol. 50, 1376–1383.
6. Barr-Beare, E., Saxena, V., Hilt, E.E., Thomas-White, K., Schober, M., Li,
B., Becknell, B., Hains, D.S., Wolfe, A.J., and Schwaderer, A.L. (2015).
The interaction between enterobacteriaceae and calcium oxalate de-
posits. PLoS ONE 10, e0139575.
7. Cavarretta, I., Ferrarese, R., Cazzaniga, W., Saita, D., Luciano
`, R., Cere-
sola, E.R., Locatelli, I., Visconti, L., Lavorgna, G., Briganti, A., et al.
(2017). The microbiome of the prostate tumor microenvironment. Eur.
Urol. 72, 625–631.
8. Bossa, L., Kline, K., McDougald, D., Lee, B.B., and Rice, S.A. (2017). Uri-
nary catheter-associated microbiota change in accordance with treatment
and infection status. PLoS ONE 12, e0177633.
9. Frank, D.N., Wilson, S.S., St Amand, A.L., and Pace, N.R. (2009). Culture-
independent microbiological analysis of foley urinary catheter biofilms.
PLoS ONE 4, e7811.
10. Rahman, M.A., Alam, M.M., Shamsuzzaman, S.M., and Haque, M.E.
(2010). Evaluation of bacterial colonization and bacteriuria secondary to
internal ureteral stent. Mymensingh Med. J. 19, 366–371.
11. Reid, G., Denstedt, J.D., Kang, Y.S., Lam, D., and Nause, C. (1992). Micro-
bial adhesion and biofilm formation on ureteral stents in vitro and in vivo.
J. Urol. 148, 1592–1594.
12. Chatterjee, S., Maiti, P., Dey, R., Kundu, A., and Dey, R. (2014). Biofilms on
indwelling urologic devices: microbes and antimicrobial management
prospect. Ann. Med. Health Sci. Res. 4, 100–104.
13. Chew, B.H., and Lange, D. (2009). Ureteral stent symptoms and associ-
ated infections: a biomaterials perspective. Nat. Rev. Urol. 6, 440–448.
14. Moltzahn, F., Haeni, K., Birkha
¨user, F.D., Roth, B., Thalmann, G.N., and
Zehnder, P. (2013). Peri-interventional antibiotic prophylaxis only vs
continuous low-dose antibiotic treatment in patients with JJ stents: a pro-
spective randomised trial analysing the effect on urinary tract infections
and stent-related symptoms. BJU Int. 111, 289–295.
15. Paz, A., Amiel, G.E., Pick, N., Moskovitz, B., Nativ, O., and Potasman, I.
(2005). Febrile complications following insertion of 100 double-J ureteral
stents. J. Endourol. 19, 147–150.
16. Gloor, G.B., Macklaim, J.M., Pawlowsky-Glahn, V., and Egozcue, J.J.
(2017). Microbiome datasets are compositional: and this is not optional.
Front. Microbiol. 8, 2224.
Cell Reports Medicine 1, 100094, September 22, 2020 9
Article
ll
OPEN ACCESS
17. Oksanen, J., Blanchet, F.G., Friendly, M., Kindt, R., Legendre, P., McGlinn,
D., Minchin, P.R., O’Hara, R.B., Simpson, G.L., Solymos, P., et al. (2019).
Vegan: Community ecology package. R package version 2.5-6. https://
cran.r-project.org/web/packages/vegan/index.html.
18. Morgan, X.C., Tickle, T.L., Sokol, H., Gevers, D., Devaney, K.L., Ward,
D.V., Reyes, J.A., Shah, S.A., LeLeiko, N., Snapper, S.B., et al. (2012).
Dysfunction of the intestinal microbiome in inflammatory bowel disease
and treatment. Genome Biol. 13, R79.
19. Motulsky, H.J., and Brown, R.E. (2006). Detecting outliers when fitting data
with nonlinear regression - a new method based on robust nonlinear
regression and the false discovery rate. BMC Bioinformatics 7, 123.
20. Kehinde, E.O., Rotimi, V.O., Al-Hunayan, A., Abdul-Halim, H., Boland, F.,
and Al-Awadi, K.A. (2004). Bacteriology of urinary tract infection associ-
ated with indwelling J ureteral stents. J. Endourol. 18, 891–896.
21. Tavichakorntrakool, R., Boonsiri, P., Prasongwatana, V., Lulitanond, A.,
Wongkham, C., and Thongboonkerd, V. (2017). Differential colony size,
cell length, and cellular proteome of Escherichia coli isolated from urine
vs. stone nidus of kidney stone patients. Clin. Chim. Acta 466, 112–119.
22. Chew, B.H., Knudsen, B.E., Nott, L., Pautler, S.E., Razvi, H., Amann, J.,
and Denstedt, J.D. (2007). Pilot study of ureteral movement in stented pa-
tients: first step in understanding dynamic ureteral anatomy to improve
stent comfort. J. Endourol. 21, 1069–1075.
23. Lane, M.C., Alteri, C.J., Smith, S.N., and Mobley, H.L. (2007). Expression
of flagella is coincident with uropathogenic Escherichia coli ascension to
the upper urinary tract. Proc. Natl. Acad. Sci. USA 104, 16669–16674.
24. Nickel, J.C., Downey, J., and Costerton, J.W. (1992). Movement of pseu-
domonas aeruginosa along catheter surfaces. A mechanism in pathogen-
esis of catheter-associated infection. Urology 39, 93–98.
25. Siitonen, A., and Nurminen, M. (1992). Bacterial motility is a colonization
factor in experimental urinary tract infection. Infect. Immun. 60, 3918–
3920.
26. Tenke, P., Ko
¨ves, B., Nagy, K., Hultgren, S.J., Mendling, W., Wullt, B.,
Grabe, M., Wagenlehner, F.M., Cek, M., Pickard, R., et al. (2012). Update
on biofilm infections in the urinary tract. World J. Urol. 30, 51–57.
27. Gottschick, C., Deng, Z.L., Vital, M., Masur, C., Abels, C., Pieper, D.H., and
Wagner-Do
¨bler, I. (2017). The urinary microbiota of men and women and
its changes in women during bacterial vaginosis and antibiotic treatment.
Microbiome 5,99.
28. Buhmann, M.T., Abt, D., Nolte, O., Neu, T.R., Strempel, S., Albrich, W.C.,
Betschart, P., Zumstein, V., Neels, A., Maniura-Weber, K., and Ren, Q.
(2019). Encrustations on ureteral stents from patients without urinary tract
infection reveal distinct urotypes and a low bacterial load. Microbiome 7,
60.
29. Liu, F., Ling, Z., Xiao, Y., Yang, Q., Zheng, L., Jiang, P., Li, L., and Wang, W.
(2017). Characterization of the urinary microbiota of elderly women and the
effects of type 2 diabetes and urinary tract infections on the microbiota.
Oncotarget 8, 100678–100690.
30. Rowe, T.A., and Juthani-Mehta, M. (2013). Urinary tract infection in older
adults. Aging Health 9.https://doi.org/10.2217/ahe.13.38.
31. Aujoulat, F., Bouvet, P., Jumas-Bilak, E., Jean-Pierre, H., and Marchandin,
H. (2014). Veillonella seminalis sp. nov., a novel anaerobic Gram-stain-
negative coccus from human clinical samples, and emended description
of the genus Veillonella. Int. J. Syst. Evol. Microbiol. 64, 3526–3531.
32. Berenger, B.M., Chui, L., Borkent, A., and Lee, M.C. (2015). Anaerobic uri-
nary tract infection caused by Veillonella parvula identified using cystine-
lactose-electrolyte deficient media and matrix-assisted laser desorption
ionization-time of flight mass spectrometry. IDCases 2, 44–46.
33. Mashima, I., and Nakazawa, F. (2014). The influence of oral Veillonella spe-
cies on biofilms formed by Streptococcus species. Anaerobe 28, 54–61.
34. Scheiman, J., Luber, J.M., Chavkin, T.A., MacDonald, T., Tung, A., Pham,
L.D., Wibowo, M.C., Wurth, R.C., Punthambaker, S., Tierney, B.T., et al.
(2019). Meta-omics analysis of elite athletes identifies a performance-
enhancing microbe that functions via lactate metabolism. Nat. Med. 25,
1104–1109.
35. van de Wijgert, J.H., Borgdorff, H., Verhelst, R., Crucitti, T., Francis, S.,
Verstraelen, H., and Jespers, V. (2014). The vaginal microbiota: what
have we learned after a decade of molecular characterization? PLoS
ONE 9, e105998.
36. Pearce, M.M., Zilliox, M.J., Rosenfeld, A.B., Thomas-White, K.J., Richter,
H.E., Nager, C.W., Visco, A.G., Nygaard, I.E., Barber, M.D., Schaffer, J.,
et al. (2015). The female urinary microbiome in urgency urinary inconti-
nence. Am. J. Obstet. Gynecol. 213, 347.e1–347.e11.
37. Thomas-White, K.J., Kliethermes, S., Rickey, L., Lukacz, E.S., Richter,
H.E., Moalli, P., Zimmern, P., Norton, P., Kusek, J.W., Wolfe, A.J., et al.
(2017). Evaluation of the urinary microbiota of women with uncomplicated
stress urinary incontinence. Am. J. Obstet. Gynecol. 216, 55.e1–55.e16.
38. Arago
´n, I.M., Herrera-Imbroda, B., Queipo-Ortun
˜o, M.I., Castillo, E., Del
Moral, J.S., Go
´mez-Milla
´n, J., Yucel, G., and Lara, M.F. (2018). The urinary
tract microbiome in health and disease. Eur. Urol. Focus 4, 128–138.
39. Akay, A.F., Aflay, U., Gedik, A., Sahin, H., and Bircan, M.K. (2007). Risk
factors for lower urinary tract infection and bacterial stent colonization in
patients with a double J ureteral stent. Int. Urol. Nephrol. 39, 95–98.
40. Altunal, N., Willke, A., and Hamzao
glu, O. (2017). Ureteral stent infections:
a prospective study. Braz. J. Infect. Dis. 21, 361–364.
41. Lee, K.J., and Tack, J. (2010). Altered intestinal microbiota in irritable
bowel syndrome. Neurogastroenterol. Motil. 22, 493–498.
42. Sha, S., Xu, B., Wang, X., Zhang, Y., Wang, H., Kong, X., Zhu, H., and Wu,
K. (2013). The biodiversity and composition of the dominant fecal micro-
biota in patients with inflammatory bowel disease. Diagn. Microbiol. Infect.
Dis. 75, 245–251.
43. Shankar, V., Agans, R., Holmes, B., Raymer, M., and Paliy, O. (2013). Do
gut microbial communities differ in pediatric IBS and health? Gut Microbes
4, 347–352.
44. Brook, I. (2004). Urinary tract and genito-urinary suppurative infections
due to anaerobic bacteria. Int. J. Urol. 11, 133–141.
45. Haggerty, C.L., and Taylor, B.D. (2011). Mycoplasma genitalium: an
emerging cause of pelvic inflammatory disease. Infect. Dis. Obstet. Gyne-
col. 2011, 959816.
46. Magruder, M., Sholi, A.N., Gong, C., Zhang, L., Edusei, E., Huang, J., Al-
bakry, S., Satlin, M.J., Westblade, L.F., Crawford, C., et al. (2019). Gut ur-
opathogen abundance is a risk factor for development of bacteriuria and
urinary tract infection. Nat. Commun. 10, 5521.
47. Yamamoto, S., Tsukamoto, T., Terai, A., Kurazono, H., Takeda, Y., and
Yoshida, O. (1997). Genetic evidence supporting the fecal-perineal-ure-
thral hypothesis in cystitis caused by Escherichia coli. J. Urol. 157,
1127–1129.
48. Matsumoto, S., Hashizume, K., Wada, N., Hori, J., Tamaki, G., Kita, M.,
Iwata, T., and Kakizaki, H. (2013). Relationship between overactive
bladder and irritable bowel syndrome: a large-scale internet survey in
Japan using the overactive bladder symptom score and Rome III criteria.
BJU Int. 111, 647–652.
49. Moreno, E., Andreu, A., Pe
´rez, T., Sabate
´, M., Johnson, J.R., and Prats, G.
(2006). Relationship between Escherichia coli strains causing urinary tract
infection in women and the dominant faecal flora of the same hosts. Epi-
demiol. Infect. 134, 1015–1023.
50. Moreno, E., Andreu, A., Pigrau, C., Kuskowski, M.A., Johnson, J.R., and
Prats, G. (2008). Relationship between Escherichia coli strains causing
acute cystitis in women and the fecal E. coli population of the host.
J. Clin. Microbiol. 46, 2529–2534.
51. Zingone, F., Iovino, P., Santonicola, A., Gallotta, S., and Ciacci, C. (2017).
High risk of lower urinary tract symptoms in patients with irritable bowel
syndrome. Tech. Coloproctol. 21, 433–438.
52. Hilt, E.E., McKinley, K., Pearce, M.M., Rosenfeld, A.B., Zilliox, M.J., Muel-
ler, E.R., Brubaker, L., Gai, X., Wolfe, A.J., and Schreckenberger, P.C.
(2014). Urine is not sterile: use of enhanced urine culture techniques to
10 Cell Reports Medicine 1, 100094, September 22, 2020
Article
ll
OPEN ACCESS
detect resident bacterial flora in the adult female bladder. J. Clin. Micro-
biol. 52, 871–876.
53. Ducrotte
´, P., Sawant, P., and Jayanthi, V. (2012). Clinical trial: Lactoba-
cillus plantarum 299v (DSM 9843) improves symptoms of irritable bowel
syndrome. World J. Gastroenterol. 18, 4012–4018.
54. Hocquart, M., Pham, T., Kuete, E., Tomei, E., Lagier, J.C., and Raoult, D.
(2019). Successful fecal microbiota transplantation in a patient suffering
from irritable bowel syndrome and recurrent urinary tract infections.
Open Forum Infect. Dis. 6, ofz398.
55. Kim, J.M., and Park, Y.J. (2017). Probiotics in the prevention and treatment
of postmenopausal vaginal infections: Review article. J. Menopausal Med.
23, 139–145.
56. Tariq, R., Pardi, D.S., Tosh, P.K., Walker, R.C., Razonable, R.R., and
Khanna, S. (2017). Fecal microbiota transplantation for recurrent clos-
tridium difficile infection reduces recurrent urinary tract infection fre-
quency. Clin. Infect. Dis. 65, 1745–1747.
57. Dyer, R., and Nordin, B.E. (1967). Urinary crystals and their relation to
stone formation. Nature 215, 751–752.
58. Lewis, D.A., Brown, R., Williams, J., White, P., Jacobson, S.K., Marchesi,
J.R., and Drake, M.J. (2013). The human urinary microbiome; bacterial
DNA in voided urine of asymptomatic adults. Front. Cell. Infect. Microbiol.
3,41.
59. Karstens, L., Asquith, M., Davin, S., Fair, D., Gregory, W.T., Wolfe, A.J.,
Braun, J., and McWeeney, S. (2019). Controlling for contaminants in
low-biomass 16s rrna gene sequencing experiments. mSystems 4,
e00290-19.
60. Minich, J.J., Sanders, J.G., Amir, A., Humphrey, G., Gilbert, J.A., and
Knight, R. (2019). Quantifying and understanding well-to-well contamina-
tion in microbiome research. mSystems 4, e00186-19.
61. Caporaso, J., et al. (2012). Ultra-high-throughput microbial community
analysis on the Illumina HiSeq and MiSeq platforms. ISME J 6, 1621–1624.
62. Callahan, B.J., McMurdie, P.J., Rosen, M.J., Han, A.W., Johnson, A.J.,
and Holmes, S.P. (2016). DADA2: High-resolution sample inference from
Illumina amplicon data. Nat. Methods 13, 581–583.
63. Quast, C., Pruesse, E., Yilmaz, P., Gerken, J., Schweer, T., Yarza, P., Pe-
plies, J., and Glo
¨ckner, F.O. (2013). The SILVA ribosomal RNA gene data-
base project: improved data processing and web-based tools. Nucleic
Acids Res. 41, D590–D596.
64. Gloor, G.B., and Reid, G. (2016). Compositional analysis: a valid approa ch
to analyze microbiome high-throughput sequencing data. Can. J. Micro-
biol. 62, 692–703.
65. Fernandes, A.D., Macklaim, J.M., Linn, T.G., Reid, G., and Gloor, G.B.
(2013). ANOVA-like differential expression (ALDEx) analysis for mixed
population RNA-seq. PLoS ONE 8, e67019.
66. Mallick, H., McIver, L.J., Rahnavard, A., Ma, S., Zhang, Y., Nguyen, L.H.,
Tickle, T.L., Weingart, G., Ren, B., Schwager, E., et al. (2020). Maaslin2:
Multivariable association in population-scale meta-omics studies. http://
huttenhower.sph.harvard.edu/maaslin.
67. Parada, A.E., Needham, D.M., and Fuhrman, J.A. (2016). Every base mat-
ters: assessing small subunit rRNA primers for marine microbiomes with
mock communities, time series and global field samples. Environ. Micro-
biol. 18, 1403–1414.
68. R Core Team (2019). R: A language and environment for statistical
computing (R Foundation for Statistical Computing).
69. Salonen, A., Lahti, L., Saloja
¨rvi, J., Holtrop, G., Korpela, K., Duncan, S.H.,
Date, P., Farquharson, F., Johnstone, A.M., Lobley, G.E., et al. (2014).
Impact of diet and individual variation on intestinal microbiota composition
and fermentation products in obese men. ISME J. 8, 2218–2230.
Cell Reports Medicine 1, 100094, September 22, 2020 11
Article
ll
OPEN ACCESS
STAR+METHODS
KEY RESOURCES TABLE
RESOURCE AVAILABILITY
Lead Contact
Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Jeremy
Burton (Jeremy.Burton@lawsonresearch.com).
Materials Availability
This study did not generate new unique reagents.
Data and Code Availability
16S rRNA gene sequencing data generated in this study is available through the NCBI Sequence Read Archive:BioProject ID
#PRJNA601180.
REAGENT or RESOURCE SOURCE IDENTIFIER
Bacterial Strains
Escherichia coli DH5aATCC ATCC 68233
Staphylococcus aureus Newman ATCC ATCC 25904
Biological Samples
Healthy adult human urine (5 – 50 mL) This paper N/A
Healthy adult ureteral stents This paper N/A
Chemicals, Peptide, and Recombinant Proteins
LB broth BD Difco Catalog No. B244601
RNase AWAY ThermoScientific Catalog No. 7003PK
Nuclease free water Ambion Catalog No. AM9932
GoTaq hot start colorless Master Mix Promega Catalog No. M5133
Critical Commercial Assays
DNEasy PowerSoil HTP 96 kit QIAGEN Catalog No. 12955-4
Quanti-iT PicoGreen dsDNA assay Invitrogen Catalog No. P11496
QIAquick PCR Purification kit QIAGEN Catalog No. 28106
MiSeq Reagent Kit v3 (600-cycle) Illumina Catalog No. MS-102-3003
Deposited Data
Raw data This paper 16S rRNA sequence data (NCBI) BioProject ID
#PRJNA601180
Oligonucleotides
16S rRNA forward primer 515F:
GTGCCAGCMGCCGCGGTAA
(Caporaso et al.
61
) N/A
16S rRNA reverse primer 806R:
GGACTACHVGGGTWTCTAAT
(Caporaso et al.
61
) N/A
Software and Algorithms
DADA2 v1.14 (Callahan et al.
62
)https://benjjneb.github.io/dada2/
SILVA Database v132 (Quast et al.
63
)https://www.arb-silva.de
R v3.6.1 R Core Team https://www.r-project.org
CoDaSeq v0.99.4 (Gloor and Reid
64
)https://github.com/ggloor/CoDaSeq
ALDEx2 v1.11.0 (Fernandes et al.
65
)https://bioconductor.org/packages/release/
bioc/html/ALDEx2.html
Vegan v2.5-6 (Oksanen et al.
17
)https://cran.r-project.org/web/packages/
vegan/vegan.pdf
MaAsLin2 v1.1.1 (Mallick et al.
66
)http://huttenhower.sph.harvard.edu/maaslin
GraphPad Prism v8.3.1 Graph Pad Software N/A
e1 Cell Reports Medicine 1, 100094, September 22, 2020
Article
ll
OPEN ACCESS
EXPERIMENTAL MODEL AND SUBJECT DETAILS
Two hundred and forty-one ureteral double-J stent patients (122 females, 119 males) were recruited from the Urology Department at
St. Joseph’s Hospital in London, Ontario. Patients ranged in age from 22-90 years (average 59). Ethical approval for the study was
granted by the Health Sciences Research Ethics Board at the University of Western Ontario (REB #107941) in London, Ontario. Writ-
ten consent was obtained from all the study participants at the time of study inclusion and the methods were carried out in accor-
dance with the approved guidelines. Inclusion and exclusion criteria for the participants are provided in Table S4. All patients that met
the inclusion criteria were recruited to the study during regularly scheduled clinic appointments.
METHOD DETAILS
Sample processing and DNA extraction
Upon recruitment, patients were asked about relevant demographic and medical history including antibiotic usage and their history of
urinary tract infections. Following enrolment, participants provided a mid-stream urine sample. Stents were collected during cystos-
copy (either in-clinic or OR) and placed by the surgeon into a sterile urine collection cup (Figure S1). Urine samples were processed
within 6-hours of their collection. The entire volume of urine (5-50 mL) was centrifuged for 10 minutes at 5,000 x g, after which the
supernatant was decanted off and the pellet was stored dry at 20C until DNA extraction. The urine volume that resulted in the pellet
for 16S rRNA gene sequencing was recorded to identify confounding factors in the downstream sequencing analysis associated with
processing conditions.
Within 6 hours of their collection, stents were frozen at 20C and stored until DNA extraction. Bacterial culture was not undertaken
on stent encrustations in order to preserve as much of the biomaterial as possible for 16S rRNA gene sequencing and SEM/EDX.
Instead, a qualitative grade for the degree of device encrustation was determined (Table S3). Both proximal and distal ends of
each stent were graded by a single evaluator twice: prior to and following frozen storage. After frozen storage, the evaluator was
blinded to the grading from the first evaluation. Grades at both time points were identical for all samples.
On the day of DNA extraction, the stents were thawed and processed in a sterile biosafety hood. A scalpel sterilized with RNase
AWAY was utilized to slice two x 1 cm segments from both the proximal and distal curls of the stents (Figure S1). One segment from
each curl was reserved for 16S rRNA gene sequencing. Potential external contamination which may have occurred during device
removal or frozen storage was mitigated by rinsing: tweezers sterilized with RNase AWAY were used to hold the stent segment
over a sterile reservoir while 1 mL of nuclease free water was gently rinsed over the external surface. The rinsed segment was
then sliced open lengthwise to expose the interior lumen and directly transferred into the bead plate of the DNeasy PowerSoil
HTP 96 Kit utilized for DNA extraction. The second 1 cm cut segment was reserved for SEM: both the internal lumen and exterior
of the stent were of interest for imaging so the stent cut was sliced lengthwise and both halves were transferred to separate sterile
1.5 mL Eppendorf tubes for SEM preparation.
For DNA extraction, frozen urine pellets were thawed and suspended in 100 uL of nuclease-free water, then pipetted into individual
wells of the PowerSoil HTP bead plate with PCR-grade filter tips. Two wells in every plate were left empty and acted as negative con-
trols. Two positive controls, or spikes, were added to each plate and were 100 mL of pure bacterial culture: Spike 1 was Escherichia
coli strain DH5a, and Spike 2 was Staphylococcus aureus strain Newman. For preparation of the spikes, a single colony of the bac-
teria was inoculated into 10 mL of Luria-Bertani (LB) broth and grown overnight at 37C. One hundred 100 mL aliquots of the overnight
cultures were portioned into 1.5 mL Eppendorf tubes and frozen at 80C. For each DNA extraction plate, a single tube of both spikes
was thawed and pipetted directly into the PowerSoil HTP bead plate with PCR-grade filter tips. DNA was isolated from urine and stent
samples using the DNeasy PowerSoil HTP 96 Kit according to the manufacturer’s instructions. DNA was stored at 20C until PCR
amplification.
16S rRNA gene sequencing
PCR amplification was completed using the Earth Microbiome universal primers, 515F and 806R, which are specific for the V4 var-
iable region of the 16S rRNA gene.
67
Primers and barcode sequences are listed in Table S5. PCR reagent set-up was performed using
a Biomek3000 Laboratory Automation Workstation (Beckman-Coulter, Mississauga, ON, CAN). Ten mL of each left and right- bar-
coded primers (3.2 pMole/mL) were arrayed in 96-well plates such that each well contained a unique combination of left- and right-
barcodes (up to a maximum of 576 unique combinations). Two mL of DNA template was added to the primer plate, followed by 20 mL
of Promega GoTaq hot-start colorless master mix. The reaction was briefly mixed by pipetting, then plates were sealed with foil plate
covers and centrifuged for 2 minutes at room temperature at 2250 x g.
Amplification was carried out using an Eppendorf thermal cycler (Eppendorf, Mississauga, ON, CAN), where the lid temperature
was maintained at 104C. An initial warm-up of 95C for 4 minutes was utilized to activate the GoTaq, followed by 25 cycles of 1 min-
ute each of 95C, 52C, and 72C.
Due to the total number of samples exceeding the number of unique barcode combinations, two Illumina MiSeq runs were
completed to accommodate the sequencing of all the samples (Illumina Inc., San Diego, CA, USA). In order to identify potential batch
effects between the two sequencing runs, several samples and controls were sequenced on both runs. In total, accounting for doubly
sequenced samples, 822 samples were sequenced across 9 PCR plates (5 396-well plates containing 438 samples on the first
Cell Reports Medicine 1, 100094, September 22, 2020 e2
Article
ll
OPEN ACCESS
MiSeq run, 4 plates containing 384 samples on the second). Sequencing was carried out at the London Regional Genomics Centre
(http://www.lrgc.ca; London, ON, CAN). Amplicons were quantified using pico green and pooled at equimolar concentrations before
cleanup. Using the 600-cycle MiSeq Reagent Kit, paired-end sequencing was carried out as 2 3260 cycles with the addition of 5%
ɸX-174 at a cluster density of 1100. Data was exported as raw fastq files (uploaded to NCBI Sequence Read Archive, BioProject ID
#PRJNA601180).
From the two sequencing runs, run 1 contained 438 samples and yielded a total of 16,211,576 reads, ranging from 419 to 358,493
reads per sample. Run 2, containing 384 samples, yielded a total of 10,424,180 reads, ranging from 168 to 400,010 reads per sample.
An average of 20.8% and 18.8% of reads were removed from each sample in Runs 1 and 2, respectively, following quality filtering
performed utilizing the DADA2 pipeline.
62
The remaining filtered reads from the two runs (14,477,624 and 9,697,990) were then
merged by amplicon sequence variants (SVs). SVs that were only detected in one of the two runs were removed. Samples and
SVs were then further pruned such that the final dataset utilized in all downstream analyses retained samples with greater than
1,000 filtered reads, SVs present at 1% relative abundance in any sample, and SVs with greater than 10,000 total reads across all
samples in both runs. This cleaning reduced the dimensions of the dataset from 460 SVs and 822 samples down to 43 SVs and
710 samples. The remaining 43 SVs were assigned taxonomy with the SILVA (v132) training set, and a further 5 SVs were removed
due to their alignment to human mitochondrial sequences.
63
SEM and X-ray diffraction spectroscopy
One-centimeter stent cuts were cut open lengthwise with a sterile razor and mounted upon aluminum stubs such that one half
exposed the inner lumen, and the other half exposed the external surface. They were then gently rinsed with DI water to remove
salt precipitation prior to SEM and X-ray diffraction spectroscopy analysis.
QUANTIFICATION AND STATISTICAL ANALYSIS
Raw 16S rRNA gene sequencing reads were demultiplexed and quality filtered utilizing the DADA2 pipeline,
62
and assigned taxon-
omy with the SILVA (v132) training set.
63
Downstream analysis including PCA was performed conservatively in agreement with stan-
dards in the field, using CoDaSeq, ALDEx2, MaAsLin2, Vegan and core R packages.
16–18,64–66,68
For subgroup comparisons (Figures
2,3, and 4), all pairwise distances were incorporated in the analysis in an effort to avoid artificially minimized data variance through
averaging, and the appropriate false-discovery rate corrections were employed.
64,69
P values, sample numbers, and names of sta-
tistical tests are provided in the main text and figure legends for Figures 2,3,4, and S2–S4. Determination of data stratification
and statistical tests were performed in GraphPad Prism (v8.3.1) and R (Method Details). All tests of statistical significance used a
p value %0.05 as a cut-off.
e3 Cell Reports Medicine 1, 100094, September 22, 2020
Article
ll
OPEN ACCESS
Cell Reports Medicine, Volume 1
Supplemental Information
Ureteral Stent Microbiota
Is Associated with Patient Comorbidities
but Not Antibiotic Exposure
Kait F. Al, John D. Denstedt, Brendan A. Daisley, Jennifer Bjazevic, Blayne K.
Welk, Stephen E. Pautler, Gregory B. Gloor, Gregor Reid, Hassan Razvi, and Jeremy P.
Burton
Supplementary Figure 1. Schematic of sample collection and processing, related to STAR Methods
A) A stent patient’s clinical course was unchanged by study participation. Patients were recruited, and clinical samples
collected, during regularly scheduled ureteral stent removal appointments. All participants had stents indwelling when
the urine sample was provided. Urine samples were pelleted by centrifugation prior to frozen storage. B) Stents were
collected by the surgeon and placed into a sterile urine collection cup, then frozen. Upon DNA extraction, a sterile
scalpel was utilized to slice two x 1 cm portions from both the proximal and distal curls of the stents. One 1 cm slice
from each curl was utilized for SEM imaging, the other for 16S rRNA gene sequencing. The portion for 16S sequencing
was gently rinsed externally with nuclease-free water over a sterile reservoir and added directly to the DNA extraction
plate. Image templates from Servier Medical Art by Servier were used and modified under the Creative Commons
Attribution 3.0 Unported License.
A
B
Supplementary Figure 2. Assessment of sample read count and microbiota richness as measured by the
number of observed sequence variants, related to Figure 1
A) Sample read count was positively correlated with the number of observed SVs, as calculated by the Spearman
correlation coefficient (r = 0.44, P < 0.0001). R2was calculated as the least-squares fit of the semilog line. n= 665.
B) The number of SVs observed was higher in urine samples (n = 212) compared to stents (n = 453) (Mann-
Whitney U test, P = 0.0063). C) The total read count was higher in urine samples compared to stents (Mann-
Whitney U test, P < 0.0001). Box plot whiskers represent minimum and maximum.
A B
C
Female Male
A
BC
Supplementary Figure 3. Principal component analysis of all samples, related to Figure 1
PCA was performed on CLR-transformed Aitchison distances. Each coloured point represents a sample. Distance
between samples on the plot represents differences in microbial community composition, with 20.5% of total
variance being explained by the first two components shown. Strength and association for genera (sequence variants)
are depicted by the length and direction of the gray arrows, respectively. Points are coloured by sample type (stents
are navy, urine are orange); ellipses represent the 95% confidence interval of the sample types. Benjamini-Hochberg
corrected Welch’s t-test between sample types determined no statistically differential sequence variants. All samples
are shown in A) and shaped by participant sex (females are circles, males are triangles); female patients are shown in
B), and males are shown in C). n = 665, 337 females, 318 males.
Supplementary Figure 4. Microbiota similarity between samples assessed with Beta diversity, related to Figure 1
A) Aitchison distance was greater between samples from different participants than within samples from the same
participant (Bonferroni corrected Mann-Whitney U test, P < 0.0001). B) Aitchison distance was greater from stent
samples to urine within (W) the same participant than between proximal to distal stent curls of the same stent
(Bonferroni corrected Mann-Whitney U test, P < 0.0001). Aitchison distance was greatest between (B) urine of one
participant to stent samples from the other participants (Bonferroni corrected Mann-Whitney U tests, P < 0.0001). Box
plot whiskers represent minimum and maximum.
AB
Supplementary Figure 5. Microbial communities in longitudinally collected samples, related to Figure 2
Each vertical bar represents the relative SV abundance within a single sample. Samples are grouped by participant. Relative abundance of SVs is coloured by
genera, with common genera shown in the legend. Sample and participant attributes are described in the legend and coloured accordingly (participant sex, grade
of stent encrustation, sample type, and the visit number). Stents from the left side are denoted by “L” and from the right side by “R”, while urine are denoted by
“U”. Days between sample collections are listed in the green visit code.
Supplementary Figure 6. Microbial communities of bilateral stents, related to Figure 3
Each vertical bar represents the relative SV abundance within a single sample. Samples are grouped by participant. Relative abundance of
SVs is coloured by genus, with common genera shown in the legend. Sample and participant attributes are described in the legend and
coloured accordingly (participant sex, grade of stent encrustation, sample type). Stents from the left side are denoted by “L” and from the
right side by “R”; Urine are denoted by “U”.
A B
Supplementary Figure 7. Relationship of stent encrustation to indwelling time and alpha diversity, related
to Figure 1
A) Indwelling time was positively correlated with the degree of stent encrustation, as calculated by the Spearman
correlation coefficient (r = 0.3221, P < 0.0001). R2was calculated from ordinary linear regression. B) Shannon’s
index of alpha diversity was negatively correlated with the degree of stent encrustation, as calculated by the
Spearman correlation coefficient (r = -0.1378, P = 0.0005). R2was calculated from ordinary linear regression. n =
450. C) Shannon’s index was lower for grade-3 encrusted stents compared to grade-0 (Kruskall-Wallis test with
Dunn’s multiple comparisons, P = 0.027). n = 210 urine, 65 Grade-0, 255 Grade-1, 117 Grade-2, and 13 Grade-3.
C
Supplementary Figure 8. Microbial communities of long-term stents, related to Table 2
Each vertical bar represents the relative SV abundance within a single sample. Samples are grouped by participant. Relative abundance of SVs is coloured by
genus, with common genera shown in the legend. Sample and participant attributes are described in the legend and coloured accordingly (participant sex, grade
of stent encrustation, sample type). Stents from the left side are denoted by “L” and from the right side by “R”, while urine are denoted by “U”. Stent
indwelling time is stated under each participant (from 62 to 394 days).
Supplementary Table 1. Ten most abundant sequence variants in urine and stent samples,
related to Figure 1
Urine samples
Rank
Sequence Variant
Corresponding Genus
Average abundance (%)
1
SV_603
Staphylococcus
18.09
2
SV_705
Enterococcus
17.53
3
SV_709
Lactobacillus
11.75
4
SV_213
Escherichia
11.61
5
SV_695
Lactobacillus
6.13
6
SV_713
Lactobacillus
5.20
7
SV_108
Prevotella
2.42
8
SV_60
Pseudomonas
2.28
9
SV_40
Ureaplasma
2.10
10
SV_208
Citrobacter
1.63
Stent samples
Rank
Sequence Variant
Corresponding Genus
Average abundance (%)
1
SV_705
Enterococcus
21.43
2
SV_603
Staphylococcus
18.73
3
SV_213
Escherichia
16.21
4
SV_709
Lactobacillus
9.15
5
SV_713
Lactobacillus
4.98
6
SV_695
Lactobacillus
4.24
7
SV_60
Pseudomonas
3.67
8
SV_208
Citrobacter
2.34
9
SV_40
Ureaplasma
1.90
10
SV_648
Veillonella
1.66
Supplementary Table 2. Significant covariates of microbiota variation at genus level PCA
ordination, related to Table 2
Metadata
P- value
Stent indwelling time
0.008
Pulmonary disease
0.001
Hypertension
0.001
Diabetes
0.001
IBS
0.002
Procedure (e.g. ureteroscopy, extracorporeal
shock wave lithotripsy)
0.007
Supplementary Table 3. Classification of stent encrustation, related to STAR Methods
Grade of encrustation
Visual characteristics
Number
0
Like-new
84
1
Discolouration only
308
2
Mild encrustation (≤1 mm thick)
135
3
Heavy encrustation (>1 mm thick)
10
Supplementary Table 4. Inclusion and exclusion criteria for study participation, related to
STAR methods
Inclusion Criteria
Exclusion Criteria
Male or Female
In the opinion of the treating urologist, it is
not in the patient’s best interest to participate
At least 18 years of age
Has a ureteric stent scheduled for removal
Able and willing to provide informed consent
Supplementary Table 5. Primer and barcode sequences, related to STAR methods
Name
Sequence (5’ – 3’)
Left Illumina adapter
ACACTCTTTCCCTACACGACGCTCTTCCGATCT
Right Illumina adapter
CGGTCTCGGCATTCCTGCTGAACCGCTCTTCCGATCT
Left primer (515F)
GTGCCAGCMGCCGCGGTAA
Right primer (806R)
GGACTACHVGGGTWTCTAAT
Barcode #1
TGCATACACTGG
Barcode #2
ACTCACAGGAAT
Barcode #3
GTAGGTGCTTAC
Barcode #4
CAGTCGTTAAGA
Barcode #5
CACTACGCTAGA
Barcode #6
GCTCGAAGATTC
Barcode #7
TGAACGTTGGAT
Barcode #8
ATGGTTCACCCG
Barcode #9
CGAGGGAAAGTC
Barcode #10
TACTACGTGGCC
Barcode #11
GTTCCTCCATTA
Barcode #12
ACGATATGGTCA
Barcode #13
TATCGACACAAG
Barcode #14
AGCATGTCCCGT
Barcode #15
CCAGATATAGCA
Barcode #16
GTGTCCGGATTC
Barcode #17
ATCGCACAGTAA
Barcode #18
CAGCTCATCAGC
Barcode #19
GCATATGCACTG
Barcode #20
TGTAGGTGTGCT
Barcode #21
ACGAGACTGATT
Barcode #22
CATCAGTACGCC
Barcode #23
GTATCTGCGCGT
Barcode #24
TGCGTCAGCTAC
