Urobiome composition after renal transplantation: an exploratory study

Abstract

Background

The urobiome of renal transplant recipients is poorly defined. The purpose of this study was to investigate whether there are characteristic changes in the urobiome between pre- to post-transplant states, at varying degrees of post-transplant allograft function, and between those with acute T-cell mediated rejection (TCMR) versus a non-rejector cohort.

Patients and methods

41 patients who consented to have urine stored in our transplant biobank were included in this study: 1) Rejectors (n = 10 pts: 6 borderline, 1 Banff IA, 3 Banff IIA TCMR, mean age: 47.4 ± 12.4 yrs); 2) Women (n = 16 pts; mean age: 49.3 ± 17.3 yrs); 3) Men (n = 15 pts, mean age: 47.5 ± 17.2 yrs). Urine was collected via mid-stream clean-catch technique prior to transplant (n = 21), at the time of TCMR (n = 9; within 1 month of transplant), 1-month (n = 15), and 3-months post-transplant (n = 38). Samples were processed and stored at -80 until 16S rRNA sequencing. Alpha diversity, beta diversity, and differential abundance analysis was performed.

Results

The urobiome was altered post-transplant, with rejectors gaining Corynebacterium and Pseudomonas at time of rejection, and non-rejectors gaining Lactobacillus among other taxa. Within individuals, post-transplant urobiome composition was ~ 75% dissimilar from pre-transplant (p < 0.001). Urobiome composition differed by sex (p = 0.002), but not by age. Differential abundance analysis based on 3-month post-transplant eGFR revealed consistent loss of Lactobacillus with decreased renal function.

Conclusions

Our results suggest that renal transplantation has a strong impact on individual urobiome composition, but not diversity, and microbial imbalance may be associated with acute rejection and post-transplant renal function. Our findings indicate a need for further research into the urobiome during renal transplantation to elucidate its potential as a biomarker of and/or contributor to post-transplant allograft health.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12866-026-05106-4.

Introduction

The symbiotic relationship between microbial communities and the human body is increasingly understood to be a crucial aspect of normal physiologic function. However, disruptions to this delicate balance, referred to as dysbiosis, are often associated with diseased states [1]. The urinary tract, once considered sterile, is now known to possess its own distinct microbiome known as the urobiome [2–4]. Research has unveiled associations between the urobiome and various diseases, including chronic kidney disease (CKD), urolithiasis, and bladder cancer [5–7]. However, to date, the urobiome in renal transplant (RT) recipients remains relatively unexplored.

Urine represents an attractive analytic specimen to gauge the health of RT allografts given the relative ease and non-invasive nature of sample collection and its intimate association with the kidney itself. Several small studies have investigated the urobiome in RT recipients and found that the urobiome is altered post-RT with dysbiosis evident during renal allograft injury/dysfunction [8–12]. Likewise, urobiome diversity and composition has been noted to change with kidney function in both transplant and non-transplant individuals, with as yet undefined significance for the renal transplant population [5, 10, 13, 14]. Although constrained by factors such as small sample size, these early studies suggest potential involvement of the urobiome in RT outcomes, emphasizing the necessity for additional research in this area.

In this study, our goal was to profile the urobiome of adult RT recipients to understand if differences exist between pre- to post-transplant states, at varying degrees of renal function, and between those with acute allograft rejection versus a non-rejector cohort. We hypothesize that characteristic changes will be present in the urobiome following RT and between different post-transplant states.

Materials and methods

Study population and design

This study was conducted within the renal transplant program of British Columbia, Canada, using urine samples collected between 2019–2021. Patients consented to have their urine collected for use in our transplant biobank (IRB: H21-02607). The use of bio-banked urine for this study was approved under the University of British Columbia Clinical Research Ethics Board (IRB: H19-02816).

This study encompassed urine samples from a post-hoc selected cohort of 41 patients distributed across 3 distinct groups: a "rejection" group consisting of 10 patients displaying varying degrees of early cell-mediated inflammation uncovered during for-cause biopsy (acute allograft dysfunction drove decision to biopsy and subsequently treat), and 2 non-rejection groups with stable post-transplant allograft function: 1) females (n = 16), 2) males (n = 15). The rejector cohort was selected based on the outcome of interest. Non-rejector patients were selected based on the availability of samples at the pre-defined study timepoints. Due to variable sample availability in our biobank, not all patients had complete samples across all timepoints. Patients diagnosed with antibody-mediated rejection were excluded. Urine was collected via mid-stream clean-catch technique at various timepoints: prior to transplant (n = 21), at time of acute transplant rejection (n = 9; within 1 month of transplant), 1-month post-transplant (n = 15), and 3-months post-transplant (n = 38), for a total of 83 urine samples. Analysis based on eGFR at 3-months post-transplant was performed with 2 groups: CKD stages 1–2 (n = 18) and CKD stages 3–5 (n = 20). Samples were processed and stored at −80 °C, then batched and sent to Cleveland Clinic (Cleveland, Ohio) for 16S rRNA sequencing and bioinformatics analysis.

A subset of this cohort was included in our prior study using shotgun metagenomic sequencing, which was limited by cost and sample size (all 18 rejector samples and 14 samples from five non-rejector patients)[12]. The rejector cohort represents a high-value subset due to the uniqueness of the early acute rejection endpoint, enabling focused analyses of urobiome changes associated with this clinically important outcome. The current study employs 16S rRNA sequencing, allowing expansion of the non-rejector cohort by 51 samples, which facilitates validation of temporal urobiome changes and exploratory analyses of sex, age, rejection status, and post-transplant allograft function. The current analysis focuses on urobiome changes in kidney transplantation, an area with limited prior study.

DNA extraction

1 mL of urine was centrifuged at 14000RPM for 2 min to pellet microorganisms. A 500 L aliquot of pellet was stored at −80ºC prior to DNA extraction on a KingFisher™ Duo Prime system (ThermoFisher) following the manufacturer’s protocol for urine with two extraction controls (sterile PBS and extraction reagents) and a PCR negative control. Positive controls included a commercially available DNA mixture of known origin (ZymoBIOMICS Microbial Community DNA standard, Zymo Research) and three previously sequenced urine DNA samples. DNA extraction protocol included mechanical lysis and proteinase K chemical lysis, followed by removal with magnetic beads and elution in buffer.

16S rRNA sequencing

Extracted DNA was submitted to the Microbial Sequencing and Analytics Core at Cleveland Clinic for 16S rRNA sequencing on an Illumina MiSeq platform after amplification of the V4 hypervariable region of the 16S rRNA gene with 515 F and 806R primers. DNA quality was assessed using a Qubit fluorometer (ThermoFisher) and normalized prior to library preparation with the Nextera XT kit (Illumina) for paired 150 bp sequences.

Bioinformatics and statistical analysis

Raw sequencing data was filtered for low quality and bimeric sequences using default parameters in the DADA2 package in R [15, 16]. High-quality sequences were assigned to amplicon sequence variants (ASVs) using a dereplicated 16S rRNA database that combined 16S rRNA sequences from the NCBI, Greengenes, and SILVA databases. ASVs were aligned in multiple sequence alignment (MSA) and maximum likelihood phylogeny was generated using the phangorn package [17]. The mapping file, ASV table, taxonomy table, and phylogenetic tree were combined as a PhyloSeq object for further analysis. Rarefaction analysis was performed using the Vegan package to determine if sequencing depth adequately captured urobiome diversity [18]. For each sample, random subsamples were drawn and ASVs recorded at 100-read increments. An inclusion threshold was defined as the depth at which > 90% of samples had a slope of the rarefaction curve of < 0.1, and reads for all samples that exceeded this threshold were retained for subsequent analysis. Contaminant sequences were assessed and removed with the Decontam package using negative controls as a source of contamination [19]. Sequences assigned to Chloroplasts, Mitochondria, or Eukaryotes were removed.

High-quality, decontaminated data were normalized with DESeq2 to minimize sequencing depth biases [20]. Normalized data was used to calculate alpha diversity using Margalef’s species richness, which quantifies the number of unique ASVs present, and beta-diversity, using a weighted UniFrac dissimilarity analysis which looks at presence/absence and relative abundance of phylogenetic clades. Alpha diversity data was statistically analyzed with a paired t-test and Holm’s correction where applicable. Statistical analysis of beta-diversity was conducted with a PERMANOVA after 999 permutations. The DESeq2 algorithm was used for differential abundance analysis.

Longitudinal within-patient microbiome dynamics were evaluated using paired analyses across sampling timepoints. For beta diversity, weighted UniFrac distances were calculated and each subject’s distance from their baseline sample was computed to quantify community compositional change over time. These within-subject distances were analyzed using linear mixed-effects models with timepoint, rejection status, and their interaction as fixed effects and patient as a random effect to account for repeated sampling. Alpha diversity (species richness) was analyzed using an analogous mixed-effects model framework to test whether within-patient diversity trajectories differed over time or between groups. All longitudinal analyses therefore evaluated microbiome changes relative to each patient’s baseline while accounting for within-subject correlation across repeated measurements.

Results

Patient demographics

Patient demographics for each group are outlined in Table 1. In the rejection group, 6 patients were diagnosed with borderline changes suspicious for cell-mediated rejection, 1 patient had Banff criteria IA cellular rejection, and 3 patients had Banff criteria IIA cellular rejection. The Banff 2019 criteria was used to diagnose rejection. All patients in the rejection group were treated with pulse intravenous methylprednisolone. Patients received pre-operative cephazolin, trimethoprim/sulfamethoxazole daily for life for Pneumocystis jirovecci pneumonia prophylaxis, and ciprofloxacin at time of stent removal 6–8 weeks post-transplant as per our centre’s protocol. Patients in the rejector group had decreased renal function compared to the non-rejector group at 1-, 2-, and 3-months post-transplant. The rejection group had a higher proportion of living donors (70% vs. 38%), which corresponded with shorter cold ischemia times compared to the non-rejection group (5.1 ± 2.8 h vs. 7.6 ± 4.6 h). 1 patient in the rejection group had culture-proven urinary tract infection (UTI) within 3-months post-transplant, compared to 6 patients in the non-rejection group.

Table 1.

Patient characteristics by rejection status

	Non-rejectors	Rejectors	p-value (a = 0.05)
Number of patients	31	10	NA
Patient characteristics
Mean age (SD) (years)	48.7 (16.7)	47.4 (12.4)	0.79
Sex, n (%)
Male	15 (48.4)	9 (90)	0.05
Male > 50 years	7 (22.6)	4 (40)	0.5
Male < 50 years	8 (25.8)	5 (50)	0.30
Female	16 (51.6)	1 (10)	0.05
Female > 50 years	8 (25.8)	0 (0)	0.18
Female < 50 years	8 (25.8)	1 (10)	0.54
Blood type, n (%)
A	12 (38.7)	5 (50)	0.79
B	7 (22.6)	0 (0)	0.24
AB	0 (0)	1 (10)	0.55
O	13 (38.7)	4 (40)	0.32
Diabetes mellitus, n (%)	17 (54.8)	3 (30)	0.32
Kidney disease characteristics
Dialysis status, n (%)
Pre-dialysis	4 (12.9)	4 (40)	0.16
Hemodialysis	17 (51.6)	3 (30)	0.32
Peritoneal dialysis	11 (35.5)	3 (30)	1
Mean dialysis duration (SD) (years)	2.1 (1.8)	1.5 (2.0)	0.40
Cause of ESRD, n (%)
Diabetes mellitus	12 (38.7)	2 (20)	0.48
Hypertension	4 (12.9)	0 (0)	0.56
IgA nephropathy	3 (9.7)	3 (30)	0.29
Polycystic kidneys	2 (6.45)	0 (0)	1
Other*	10 (32.3)	5 (50)	0.53
Pre-transplant serum creatinine (SD) mol/L)	697 (251)	649 (132)	0.43
Donor characteristics
Mean donor age (SD) (years)	47.5 (13.4)	50.5 (12.2)	0.52
Donor type, n (%)
Living donor	12 (38.7)	7 (70)	0.17
Deceased donor	19 (61.3)	3 (30)	0.17
DBD	16 (51.6)	3 (30)	0.41
DCD	3 (9.7)	0 (0)	0.75
SCD	13 (41.9)	2 (20)	0.38
ECD	6 (19.4)	1 (10)	0.84
Transplant characteristics
DGF, n (%)	6 (19.4)	0 (0)	0.32
Cold ischemia time (SD) (hr)	7.6 (4.6)	5.1 (2.8)	0.047
Induction immunosuppression, n (%)
Basiliximab	19 (61.3)	8 (80)	0.48
Anti-thymocyte globulin (ATG)	12 (38.7)	2 (20)	0.48
Maintenance immunosuppression, n (%)
Tacrolimus, MMF, Prednisone	20 (61.3)	10 (100)	0.07
Tacrolimus, MMF	10 (32.3)	0 (0)	0.10
Tacrolimus, Azathioprine, Prednisone	2 (6.5)	0 (0)	1
Post-transplant mean serum creatinine (SD), (mol/L)
1-week post-transplant	213 (159)	275 (117)	0.20
1-month post-transplant	133 (57)	180 (59)	0.04
2-months post-transplant	121 (42)	147 (18)	0.01
3-months post-transplant	118 (40)	150 (34)	0.03
Rejection type, n (%)
Borderline TCMR	-	6 (60)	NA
TCMR Type IA	-	1 (10)	NA
TCMR Type IIA	-	3 (30)	NA
Rejection treatment, n (%)
Pulse steroids	-	10 (100)	NA
Culture-confirmed urinary tract infection within 3 months post-transplant, n (%)	6 (19.4)	1 (10)	1

DBD donation after brain death, DCD donation after circulatory death, SCD standard criteria donor, ECD extended criteria donor, DGF delayed graft function, MMF mycophenolate mofetil, TCMR T-cell mediated rejection

^*Other causes of ESRD included obstructive nephropathy, lupus nephritis, etc.

Samples

100% of samples exceeded the sequencing threshold on rarefaction analysis (Supplemental Fig. 1), indicating adequate sequencing depth to capture all microbial diversity present and were hence included in further downstream analyses. When comparing patient samples to positive and negative controls (Supplemental Fig. 2), bacterial composition (beta diversity) was significantly different between groups (p = 0.003), but species richness (alpha diversity) was not (p = 0.605).

Fig. 1.

Urobiome profiles of rejector (n = 10) and non-rejector (n = 31) patients at the phylum (A) and genus (B) levels showing the relative abundance of bacterial taxa using pooled samples from all timepoints (rejector: pre-transplant, time of rejection, and post-rejection/3-months post-transplant; non-rejector: pre-transplant, 1-month post-transplant, and 3-months post-transplant)

Fig. 2.

Urobiome cohort analysis of rejection patients. A Species richness compared between pre-transplant (n = 5), rejection (n = 9), and post-rejection (n = 7) samples. B Beta diversity analysis of urobiome composition between pre-transplant (n = 5) and post-transplant (rejection and 3-month post-transplant samples, n = 16) timepoints

Group comparisons

Urobiome profiles of both rejector and non-rejector patients, generated from pooled samples from all timepoints, were dominated by the phyla Pseudomonadota, Bacillota, and Actinobacteriota (Fig. 1A). At the genus level, there were high levels of Escherichia, Lactobacillus, Bacillus, Enterococcus, Klebsiella, and Ureaplasma (Fig. 1B). There were no significant differences in genus-level relative abundances between rejection and non-rejection groups (Supplemental Table 1).

To investigate whether baseline urobiome characteristics predicted rejection, we compared the pre-transplant samples of rejectors and non-rejectors. At both the phylum (Supplemental Fig. 3 A) and genus (Supplemental Fig. 3B) level, urobiome profiles appeared quite different, with a greater abundance of the phylum Bacillota and the genera Propionibacterium and Ureaplasma in rejectors. However, neither pre-transplant alpha nor beta diversity differed significantly by rejector status at the group level (Supplemental Fig. 3 C and D).

Fig. 3.

Intra-individual change in urobiome composition (beta diversity) over time, measured as the weighted UniFrac distance between pre- to 3-months post-transplant samples in non-rejectors (n = 16 paired samples) and from pre-transplant to rejection samples in rejectors (n = 5 paired samples). When comparing non-rejectors to rejectors, p = 0.524, when comparing both non-rejectors and rejectors from pre- to post-transplant, p = 2.2 × 10^–16

Next, we investigated whether transplant itself changed the urobiome. In the rejection group, there were no significant differences in alpha diversity across timepoints (Fig. 2A). However, bacterial composition differed significantly between pre-transplant and grouped post-transplant samples (i.e. including both rejection and post-rejection timepoints, p = 0.018, Fig. 2B). In the non-rejection group, we limited our comparison to the pre-transplant and 3-months post-transplant timepoints due to the fewer number of samples available at 1-month post-transplant. There were no significant differences in either alpha or beta diversity from pre- to post-transplant in the non-rejector group (Supplemental Fig. 4).

Fig. 4.

Differential abundance analysis of bacterial taxa based on the DESeq2 algorithm. Only significantly different taxa are shown (false discovery rate = 0.05). A Rejector (n = 5) analysis comparing paired samples from pre-transplant and time of rejection. Positive log2 fold change values indicate taxa more abundant pre-transplant and negative log2 fold change values indicate taxa more abundant at time of rejection. B Non-rejector (n = 16) analysis comparing paired samples from pre-transplant and 3-months post-transplant. Positive log2 fold change values indicate taxa more abundant pre-transplant and negative log2 fold change values indicate taxa more abundant at 3-months post-transplant

Focusing on non-rejectors, we then analyzed the pre-transplant and 3-months post-transplant urobiomes by age group (over/under 50 years of age) and sex (male/female) (Supplemental Fig. 5). There were no significant differences between age groups in either alpha or beta diversity, nor was there an interaction with the effect of transplant. While there was similarly no significant effect of sex on alpha diversity, there was a clear independent association of urobiome composition (beta diversity) with sex (p = 0.002). Sex did not interact with the effect of transplant on urobiome composition. We did not perform these analyses in rejectors as the rejection group was smaller and consisted almost entirely of males.

Fig. 5.

Differential abundance analysis of the 3-month post-transplant urobiomes of patients stratified by 3-month post-transplant eGFR into CKD stages 1–2 (n = 18) vs CKD stages 3–5 (n = 20). Positive log2foldchange values indicate taxa with higher abundance in the CKD 3–5 group and negative log2foldchange values indicate higher abundance in the CKD 1–2 group. Only significantly different taxa are shown (false discovery rate = 0.05)

Individual comparisons

To remove the effect of inter-individual urobiome variability, we performed intra-individual analyses to investigate the change in urobiome over time at the individual level. To do so, we examined both the change in species richness (alpha diversity) and urobiome composition (beta diversity) between pre- and post-transplant samples within individuals. In rejectors, pre-transplant samples were compared with those taken at time of rejection. In non-rejectors, pre-transplant samples were compared with those taken at 3-months post-transplant. Change in alpha diversity did not differ between rejectors and non-rejectors (data not shown), nor did it significantly differ from zero for either group, indicating that transplant did not significantly affect individual urobiome species richness. Likewise, the change in urobiome composition from pre- to post-transplant was not significantly different between rejectors and non-rejectors (Fig. 3). However, the change in urobiome composition from pre- to post-transplant when rejector and non-rejector samples were combined was significantly different from zero, as based on a one-sample t-test (p = 2.2 × 10^–16). This indicates that urobiome composition changed significantly from pre- to post-transplant across all patients. Weighted UniFrac dissimilarity values for beta diversity range from 0–1, where 0 indicates that the presence and abundance of taxa between two samples are identical while 1 represents no overlap between two samples. Across all patients, weighted UniFrac dissimilarity values were approximately 0.75, indicating a post-transplant urobiome ~ 75% dissimilar from pre-transplant.

In a complementary analysis, we assessed alpha and beta diversity over time while controlling for the effects of within-patient correlation. Alpha diversity did not differ over time nor between rejectors and non-rejectors (Supplemental Fig. 6 A). In contrast, beta diversity changed significantly from baseline over time, but no significant differences were observed between rejectors and non-rejectors (Supplemental Fig. 6B).

Differential abundance analysis

To investigate whether specific taxa were associated with rejection, we performed a differential abundance analysis using the DESeq2 algorithm to look for changes in taxa between pre- and post-transplant samples in rejectors (Fig. 4A) and non-rejectors (Fig. 4B). Only non-rejectors with both a pre-transplant and 3-months post-transplant sample and rejectors with a pre-transplant and rejection sample were included in these analyses. There were more differentially abundant taxa overall in the non-rejection group, likely because there were more samples available for this analysis (16 non-rejectors vs. 5 rejectors). In non-rejectors, the abundance of Lactobacillus, Corynebacterium, Acinetobacter, and Campylobacter increased post-transplant, with a decrease in Bifidobacterium, Gardnerella, and Peptoniphilus. In the rejection group, the taxa most associated with the rejection period was Corynebacterium, which was much less abundant at pre-transplant. Pseudomonas was also enriched at rejection, while Faecalibaculum and Veillonella were decreased.

Renal function analysis

To investigate the relationship between the urobiome and allograft function, we analyzed all patients with an available 3-month post-transplant sample, including rejectors (post-rejection, n = 7) and non-rejectors (n = 31). Patients were grouped into either CKD stages 1–2 (n = 2 rejectors, n = 16 non-rejectors), or CKD stages 3–5 (n = 5 rejectors, n = 15 non-rejectors) based on their 3-month post-transplant eGFR. There were no significant differences in alpha or beta diversity between CKD groups across all patients nor within rejector and non-rejector sub-groups (data not shown). However, differential abundance analysis revealed bacterial genera whose abundance differed significantly between CKD groups irrespective of rejection status (Fig. 5). Lactobacillus, Gardnerella, Streptococcus, and Staphylococcus were enriched in the CKD 1–2 group. Conversely, CKD 3–5 patients were characterized by enriched Campylobacter, Prevotella, Corynebacterium, and Ureaplasma. Differential abundance analysis within rejection (Supplemental Fig. 7 A) and non-rejection subgroups (Supplemental Fig. 7B) revealed consistent Lactobacillus enrichment in the CKD 1–2 group. There were no consistent taxa enriched in the CKD 3–5 group, although rejectors were characterized by enriched Corynebacterium and non-rejectors were characterized by enriched Campylobacter.

Discussion

In this exploratory study, we identified temporal changes in the urobiome over time and differences by rejection status and post-transplant allograft function. Consistent with past studies, we observed significant sex differences in urobiome composition [21]. Conversely, while other studies have observed changes in the urobiome with age [22], we found no significant differences in community-level measures of either urobiome diversity or composition between age groups (non-rejectors: over vs. under 50 years). This may be in part due to small sample size but may also reflect our arbitrary age cutoff of 50 years used to delineate subgroups in lieu of an accepted literature consensus beyond which age-related changes are expected to occur in the urobiome.

Our main finding was a significant intra-individual shift in urobiome composition from pre- to post-transplant across all patients on paired analysis. Indeed, our results show that, when comparing within individuals, the post-transplant urobiome was approximately 75% dissimilar from pre-transplant when accounting for both presence/absence of taxa as well as their relative abundance. Past clinical urobiome studies that have undertaken repeated sampling of the same individuals have found dissimilarities of approximately 0.25, whereas between-individuals dissimilarities have been approximately 0.5 [23–26]. As such, our results suggest the presence of an almost entirely new urobiome post-transplant. Our observed shift in urobiome composition across transplant is in line with the findings of Fricke et al. (2014), who observed major changes in urobiome composition from pre-transplant to 1-month post-transplant [8]. Their findings, in conjunction with our data, strongly suggest that transplant surgery and associated factors such as perioperative antibiotics and immunosuppressive treatment dramatically alter the urobiome.

While further work is needed to understand the contributions of these factors to the urobiome and transplant outcomes, another possibility that we can hypothesize, but not confirm, is that there may be transfer of the donor urobiome during RT. This could contribute to the high microbial turnover within patients, and it is highly plausible that a dramatic change in the microbial ecosystem induces host immune responses in some individuals with implications for allograft health. Outside of the urobiome, Kim et al. [27] found that gut microbiome similarity in living kidney donors and their recipients was predictive of 6-month allograft function [27]. Taking all this together, it will be important for future studies to elucidate the causes of post-transplant urobiome composition changes and to specifically look at the relationship between the donor and recipient urobiomes. It will also be important to determine whether shifts in the urobiome across RT are merely byproducts of transplant factors or if they play an active role in allograft health or rejection.

As part of the urobiome shift observed from pre- to post-transplant, rejectors and non-rejectors gained and lost different taxa that may be of relevance to allograft health. In rejectors, the urobiome at time of rejection was characterized by a major increase in Corynebacterium, with a noticeable increase in Pseudomonas as well. Bacteria from these genera can cause opportunistic infections and have been associated with urological diseases, with Corynebacterium urealyticum having been implicated in acute cystitis and pyelonephritis [28], while Pseudomonas aeruginosa is the most common cause of nosocomial catheter-associated UTI [29]. Additionally, Wu et al. found an increase in Corynebacterium spp. in the urine of both males and females with chronic renal allograft dysfunction [10]. Although our 16S rRNA sequencing data does not provide species-level resolution, our findings, taken together with those of Wu et al., suggest that an increase in urinary Corynebacterium – and potentially Pseudomonas – even at levels below the threshold for UTI, may play a pathogenic role in or be a marker of allograft dysfunction. Testing this hypothesis and investigating the relationship between Corynebacterium, Pseudomonas, and allograft function should be a focus of future research. However, it should also be noted that there was a decrease in the abundance of several taxa at time of rejection, including most notably the short-chain fatty acid (SCFA) producers Faecalibaculum and Veillonella [30, 31]. SCFAs have been shown to be reno-protective via anti-inflammatory and anti-oxidative properties [32]. Moreover, a decrease in the specific SCFAs butyrate and propionate, which are produced by Faecalibaculum and Veillonella, has been linked to CKD progression [33]. While we were unable to perform a urinary metabolite analysis in this study to confirm, it could be hypothesized that some of the bacteria lost at time of rejection may have been exerting a positive/protective effect on the urobiome and/or host tissue. Thus, it may be that the gain of potential pathogens in Corynebacterium and Pseudomonas, in conjunction with a decrease in other potentially beneficial bacteria, represents a urobiome dysbiosis that may be associated with rejection.

In non-rejectors, notable taxa that increased post-transplant included Lactobacillus, Corynebacterium, Acinetobacter and Anaerococcus. The increase in Corynebacterium in this context is particularly interesting and seemingly contradictory, as we have discussed above its presence in our rejection cohort. On the other hand, the presence of Lactobacillus is less surprising, as it is a well-studied probiotic with anti-inflammatory and anti-microbial properties that has been shown to positively regulate the microbiome through the production of lactic acid and anti-microbial compounds [34]. While interpreting the relevance of these taxa as a whole is difficult and limited by the observational nature of our study, it could be hypothesized that the balance of bacteria in non-rejectors post-transplant favors a neutral or protective state. Taken as a whole, our findings suggest that it is the interactions between microbes and their products within the urobiome, and not just the presence/absence of individual taxa, that determine the net effect of the microbial community on the host. This interpretation would be in line with contemporary perspectives on the human microbiome and would reconcile the seemingly contradictory presence of Corynebacterium in both rejectors and non-rejectors.

We were also interested in whether the urobiome was associated with post-transplant renal function. The microbiome is known to play a role in renal function outside of the transplant context, with gut dysbiosis contributing to CKD/ESRD via the production of uremic toxins and the so-called gut-kidney axis [35]. There is comparatively less data on the urobiome as it relates to renal function, however, available evidence underscores its relevance in kidney disease. Supporting this idea, Kramer et al. [5] found that increased urobiome diversity was associated with improved renal function in a cohort of CKD non-dialysis patients, and Yang et al. found a positive correlation between urinary Lactobacillus with eGFR in a cohort of diabetic kidney disease patients [5, 36]. In our analysis of the 3-month post-transplant urobiome, we grouped patients by preserved or impaired renal function based on eGFR but did not observe differences based on community-level metrics of alpha or beta diversity. However, differential abundance analysis revealed consistently enriched Lactobacillus in patients with preserved renal function in both rejection and non-rejection subgroups. As such, our results are in line with those of Yang et al. and it may be that, like in our discussion of rejection above, urobiome dysbiosis may be associated with reduced renal allograft function. Future studies with larger sample sizes will be needed to explore this hypothesis.

Our study has several limitations. First, specimens were collected via clean-catch technique then processed for storage, and despite attempts to minimize sample contamination, it is possible that contaminants may have been introduced during collection or handling and may influence our results. Specifically, Corynebacterium was found to be significant on our differential abundance analysis, and given that it is a known skin commensal [37], it is possible that it may have been at least in part introduced in our samples via contamination. Next, our sample size was relatively small due to limited available specimens in our biobank and compounded by the fact that 10 patients were anuric/oliguric prior to transplant. Furthermore, our rejection cohort represents a heterogeneous group with varying degrees of kidney inflammation; it may be that each form and severity of rejection may reveal differing results. Similarly, cause of ESRD across patients was variable, and it is possible that underlying disease etiology may contribute uniquely to urobiome profiles that were not able to be investigated in the current study. Please note, all patients followed our centre protocols for antibiotic prophylaxis and immunosuppression, which may limit the generalizability of our findings. Moreover, the rejection cohort had limited sex diversity which, given the known urobiome differences between males and females, may confound our results. In addition, several patients (1 rejector, 6 non-rejectors) experienced UTI within 3-months post-transplant. While there were no active infections at the time of sample collection, it is possible that prior UTI may have influenced the urobiome of these patients. Finally, our study relied on 16S rRNA sequencing and was hence unable to achieve bacterial taxonomic resolution beyond the genus level. This is of particular relevance given interspecies differences in bacterial characteristics and pathogenicity that could impact host-microbe interactions. 16S rRNA sequencing is also understood to be unable to discriminate between live/viable and non-viable microorganisms in a given sample/niche, such that future studies utilizing such an approach may benefit from complementary urine culture analysis that we were unable to perform in the current study.

Conclusions

Our results highlight a major shift in urobiome composition post-RT and hint at increased pathogenic and decreased commensal bacteria during allograft rejection and impaired post-transplant renal function. Our work supports the need for future longitudinal studies with larger sample sizes to build upon our findings to understand how transplant-associated changes to urobiome characteristics relate to pathologic states. Our hope is that investigating the urobiome represents an initial step towards developing new non-invasive tools/treatments to facilitate early identification of allograft distress with tailored microbiome interventions to restore kidney health.

Abbreviations

ASV

Amplicon Sequence Variant

CFU

Colony-Forming Units

RT

Renal Transplantation

ESRD

End-Stage Renal Disease

UTI

Urinary Tract Infection

Authors’ contributions

DH and DL participated in the research design. KS aided with specimen acquisition. AM and TD performed the sequencing and data analysis. DH, AN and DL wrote the manuscript. All authors participated in the performance of the research and reviewing the manuscript.

Funding

This work was supported by the Canadian Donation and Transplant Research Program/Transplant Research Foundation of British Columbia Research Innovation Grant and the Vancouver Coastal Health Research Institute Mentored Clinician Scientist Award.

Data availability

The datasets generated and analysed during the current study are available in the National Center for Biotechnology Information (NCBI) Sequence Read Archive (SRA) database under the accession number PRJNA1420816.

Declarations

Ethics approval and consent to participate

Informed consent was obtained from all patients included in this study. Patients consented to have their urine collected for use in our transplant biobank (IRB: H21-02607). The use of bio-banked urine for this study was approved under the University of British Columbia Clinical Research Ethics Board (IRB: H19-02816). All procedures were performed in accordance with the principles of the Declaration of Helsinki.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.