American Journal of Transplantation 2014; 14: 416–427 Wiley Periodicals Inc.
C
Copyright 2013 The American Society of Transplantation and the American Society of Transplant Surgeons doi: 10.1111/ajt.12588
Human Microbiota Characterization in the Course of Renal Transplantation W. F. Fricke1,*, C. Maddox1, Y. Song1 and J. S. Bromberg2,*
Received 17 July 2013, revised 13 November 2013 and accepted for publication 14 November 2013
1
Institute for Genome Sciences, University of Maryland School of Medicine, Baltimore, MD 2 Department of Surgery and Microbiology and Immunology, University of Maryland School of Medicine, Baltimore, MD Corresponding authors: W. Florian Fricke, [email protected]; Jonathan S. Bromberg, [email protected]
Recent studies demonstrate that the human microbiota, the collection of microorganisms growing on and in individuals, have numerous bidirectional interactions with the host, influencing immunity, resistance to infection, inflammation and metabolism. Little has been done to study the potential associations between microbiota composition and transplant outcome. Here, we investigated the longitudinal changes in the blood, urinary, oral and rectal microbiota of renal allograft recipients before and at 1 and 6 months after transplantation. The results showed major changes in microbiota composition as a result of the transplant episode and associated medications, and these changes persisted over time. The high interindividual variation as well as differences in response to transplantation suggested that it is unlikely that the same specific microbiota members can serve as universal diagnostic markers. Rather, longitudinal changes in each individual’s microbiota have the potential to be indicative of health or disease. Use of sensitive nucleic acid–based testing showed that urine, irrespective of disease states, more often harbors a diverse microbiota than appreciated by conventional culture techniques. These results lay the groundwork to construct more comprehensive future investigations to identify microbiota characteristics that can serve as diagnostic markers for transplant health and to guide intervention strategies to improve transplant outcome. Keywords: Human microbiota, immunosuppression, renal transplantation, transplant outcome, urine microbiota Abbreviations: EDTA, ethylenediaminetetraacetic acid; FDR, false discovery rate; OTU, operational taxonomic unit; PBS, phosphate-buffered saline; PCR, polymerase chain reaction; SD, standard deviation 416
Introduction The successful results of renal transplantation are mainly due to progress in immunosuppressant therapy. Unfortunately, increasing immunosuppression leads to a higher risk of opportunistic infections, which are associated with organ rejection or dysfunction. The clinical outcomes of renal allograft recipients have improved in part because of a decline in infection-related morbidity and mortality. Prior to 1980, 60% of kidney transplant recipients developed at least one serious infection during the first year posttransplant, with mortality rates approaching 50% (1). Although the current infection-related 1-year mortality has been reduced to 90% received Campath. Pediatric and multi-organ recipients were excluded as well as patients enrolled in other interventional studies with investigational drugs. However, patients enrolled in observational studies or studies with FDA-approved drugs were considered for enrollment. All patients received perioperative cefazolin, except for a few patients who were allergic to penicillin and received ciprofloxacin. All patients also received sulfamethoxazoletrimethoprim for 6 months after transplantation, or dapsone if patients were allergic to sulfonamides. Information about infections and rejections after transplantation, and additional medications were collected. A total of 60 patients were enrolled. Samples were obtained just prior to transplantation and at 1 and 6 months after transplantation, and included the following: (a) urine, (b) oral swabs, (c) rectal swabs and (d) blood (from a subset of 23 patients). The demographics of the population were 63% male, 60% Caucasian/40% African-American, age range 30–79 (mean 58, median 58). The study was approved by the University of Maryland Institutional Review Board (#HP-00049438).
16S rRNA gene fragment amplification Hypervariable regions V1–V3 of the 16S rRNA gene fragments were amplified using a sevenfold degenerate primer 27F-YMþ3 (10) and a barcoded 534 reverse primer. An overview of the primers, including barcodes and Roche/454 Life Sciences adaptors, is given in the HMP Consortium 16S 454 sequencing protocol (http://www.hmpdacc.org/doc/ 16S_Sequencing_SOP_4.2.2.pdf). DNA amplification of 16S rRNA genes was performed using AccuPrime Taq DNA polymerase High Fidelity (Invitrogen) and 12 mL of template DNA in a total reaction volume of 25 mL, following the AccuPrime product protocol. Reactions were run in a PTC-100 thermal controller (MJ Research, Waltham, MA) using the following cycling parameters: 3 min of denaturation at 948C, followed by 30 cycles of 30 s at 948C (denaturing), 30 s at 528C (annealing) and 45 s at 688C (elongation), with a final extension at 688C for 5 min. Negative controls (no template DNA) were included for each amplification and barcoded primer pair. The presence of amplicons was confirmed by gel electrophoresis on a 1% agarose gel and staining with ethidium bromide. Polymerase chain reaction (PCR) product concentrations were estimated visually using agarose gel electrophoresis and MassRuler low range DNA ladder (Thermo Scientific, Waltham, MA). Equimolar amounts (75 ng) of the PCR amplicons were mixed in a single tube. Amplification primers and reaction buffer were removed using the AMPure Kit (Agencourt, Brea, CA) and purified amplicon mixtures sequenced by Roche/454 FLX pyrosequencing using 454 Life Sciences primer A by the Genomics Resource Center at the Institute for Genome Sciences, University of Maryland School of Medicine, using protocols recommended by the manufacturer.
Sample collection Immediately after collection, oral (inner cheek) and rectal FLOQSwabs (Copan, Murrieta) were immersed in 2 mL of the nucleic acid-preserving reagent RNAlater (Qiagen, Valencia, CA) and stored at 808C. Urine samples were stored in vacutainers without modification (urine samples) and blood samples were stored in tubes coated with anticoagulants (ethylenediaminetetraacetic acid [EDTA]) (BD, Franklin Lakes, NJ) at 48C. Urine samples were processed within 14 days, and oral and rectal swabs within 4 weeks after collection.
Nucleic acid isolation Oral and rectal swabs were vortexed with the RNAlater in the collection tubes for 10 min at room temperature and 1 mL of the resulting microbial suspension used for the subsequent DNA extraction. Urine samples (2 mL) were centrifuged at 5000g for 8 min and the entire pellet resuspended in 0.6 mL 1 phosphate-buffered saline and used for DNA extraction. Of the blood samples 1 mL was pretreated with the MolYsis Basic kit (Molzym, Germany) following the manufacturer’s recommendations to remove human DNA using their differential lysis protocol. For all microbial suspension types, microbial cell lysis was initiated with two enzymatic incubations, first using 5 mL of lysozyme (10 mg/mL; Amresco, Solon, OH), 13 mL of mutanolysin (11.7 U/mL; Sigma–Aldrich, St. Louis, MO) and 3 mL of lysostaphin (4.5 U/mL; Sigma–Aldrich) for an incubation of 30 min at 378C, and, second, using 10 mL Proteinase K (20 mg/mL; Research Products International, Mount Prospect, IL), 50 mL of 10% SDS and 2 mL RNase (10 mg/mL) for an incubation of 45 min at 568C. After the enzyme treatments, cells were disrupted by bead beating in tubes with Lysing Matrix B (0.1 mm silica spheres; MP Biomedicals, Solon, OH), at 6 m/s for 40 s at room temperature in a FastPrep-24 (MP Biomedicals). The resulting crude lysate was processed using the ZR Fecal DNA mini-prep kit (Zymo, Irvine, CA) according to the manufacturer’s recommendation. The samples were eluted with 0.08 mL of ultra pure water into separate tubes. DNA concentrations in the samples were measured using the Quant-iT PicoGreen dsDNA assay kit (Invitrogen Molecular Probes, Carlsbad, CA).
American Journal of Transplantation 2014; 14: 416–427
Sequencing and analysis 16S rRNA sequence reads were parsed and quality filtered, using the open source software package QIIME version 1.6.0 (http://qiime.sourceforge. net). Sequence reads were trimmed and quality filtered using the QIIME split.library default parameters. The automated CloVR-16S pipeline, which is part of the CloVR software package (11) was used to analyze de-multiplexed and quality filtered samples. Chimeras were removed with UCHIME (12). Each processed 16S rRNA sequence was classified using the RDP Na€ıve Bayesian Classifier (13) with a score filtering threshold of 0.5. All sequence reads were clustered into operational taxonomic units (OTUs) using a similarity threshold of 95%. Rarefaction curves were calculated based on OTU counts using the rarefaction.single routine of the Mothur package (14). Hierarchical clustering was performed using custom R scripts as implemented in the CloVR-16S pipeline. Significant differences in relative abundance of bacterial groups were determined with the Metastats program (15) for binary comparisons and with repeated measures analysis for longitudinal comparisons. For repeated measures analysis, relative abundances of bacterial genera from different sample types (urine, oral and rectal swabs), collected at different time points (A, pretransplantation; B, 1 month posttransplantation; C, 6 months posttransplantation) were analyzed using the PROC MIXED procedure of the SAS statistical software package (version 9.3; SAS Institute, Inc., Cary, NC) with time points as the repeated variable and sample types, time points and interactions of these factors as fixed effects. The output table ‘‘Type III tests of fixed effects’’ was used to evaluate whether among the fixed effects at least one mean was different from another mean. If the combination of sample types and time points showed overall significance (p < 0.05), least square means statements were used for Bonferroni-adjusted comparisons of the relative abundances between all time points, separately for each sample type. Adjusted p-values 4 mg/dL): 2 urine samples High-creatinine (>4 mg/dL): 8 oral samples High-creatinine (>4 mg/dL): 6 rectal samples Patients with rejection episode: 5 pretransplant oral samples
Urine, time point C: 6 samples Oral swabs, time point B: 21 samples Oral swabs, time point C: 9 samples Rectal swabs, time point B: 21 samples Rectal swabs, time point C: 13 samples Low-creatinine (80% relative abundance, in this case Streptococcus as shown by heatmap analysis (Figure 4B). On the phylum level, 31 samples contained >80% Firmicutes, indicating that in most samples different members of the Firmicutes were responsible for a majority of the bacterial species. Bacterial genera with at least 20% relative abundance in any of the oral swab samples were Prevotella (3 samples), Neisseria and Veillonella (each found in 2 samples), Lactobacillus and Gemella (each found in 1 sample) and unknown members of the Micrococcineae (14 samples), Actinomycineae (5 samples) and Lachnospiraceae (1 sample). Rectal swab microbiota Heatmap analysis of the 80 rectal samples showed that none were dominated at >20% by a single bacterial genus (Figure 4C), but members of the phylum Firmicutes accounted for >80% of the bacteria in 31 samples, indicating a diverse microbiota compared to urine and oral samples. Members of the phylum Bacteroides were found at >20% in 24 samples. Other bacterial phyla at >20% in at least one sample included Actinobacteria (eight samples) and Proteobacteria (two samples). Sample cross-contamination To study the extent of cross-contamination, shared OTUs were determined between all sample types (Figure S2). The
B
0
020_B_OS 009_A_OS 040_A_OS 011_A_OS 040_B_OS 051_A_OS 001_C_OS 015_B_OS 008_C_OS 006_A_OS 030_A_OS 023_C_OS 024_A_OS 059_C_OS 033_A_OS 038_A_OS 027_B_OS 011_B_OS 010_A_OS 055_B_OS 017_C_OS_rep2 005_A_OS 012_B_OS 026_A_OS 030_B_OS 034_A_OS 031_B_OS 016_B_OS 025_B_OS 007_A_OS 006_B_OS 048_B_OS 003_A_OS 031_A_OS 032_A_OS 048_A_OS 029_A_OS 021_A_OS 042_A_OS 037_A_OS 007_B_OS 023_A_OS 001_B_OS 022_A_OS 014_A_OS 017_B_OS 017_A_OS 043_A_OS 008_A_OS 025_A_OS 060_A_OS 004_A_OS 027_A_OS 057_C_OS 023_B_OS 041_B_OS 047_A_OS 059_A_OS 050_A_OS 049_A_OS 054_A_OS 017_C_OS_rep1 013_A_OS 002_B_OS 002_A_OS 018_A_OS_rep1 018_A_OS_rep2 004_B_OS 001_A_OS 041_A_OS 044_A_OS 053_A_OS 015_A_OS 012_A_OS_rep2 012_A_OS_rep1 013_B_OS 035_A_OS 006_C_OS_rep1 006_C_OS_rep2 036_A_OS 039_A_OS 056_A_OS 016_A_OS 020_A_OS 052_A_OS 045_A_OS 055_A_OS
055_B_UR 008_A_UR 019_A_UR 012_B_UR 059_B_UR 054_B_UR 007_C_UR 043_B_UR 030_B_UR 030_A_UR 006_C_UR 012_C_UR 017_C_UR 024_A_UR 040_B_UR 053_B_UR 060_A_UR 018_B_UR 002_C_UR 003_A_UR 001_A_UR 006_B_UR 013_B_UR 060_B_UR 050_B_UR 044_B_UR 037_B_UR 058_B_UR 045_B_UR 016_A_UR 035_A_UR 051_B_UR 021_A_UR 003_B_UR 003_C_UR 029_A_UR 029_B_UR
0.4
0.8
C
0
0.4
0.8
Porphyromonadaceae_Other Ruminococcaceae Incertae Sedis Rikenellaceae_Alistipes Lachnospiraceae_Coprococcus Erysipelotrichaceae_Coprobacillus Incertae Sedis XI_Other Pasteurellaceae_Haemophilus Veillonellaceae_Dialister Actinomycetales_Propionibacterineae Erysipelotrichaceae_Catenibacterium Leuconostocaceae_Weissella Ruminococcaceae_Subdoligranulum Veillonellaceae_Succinispira Veillonellaceae_Veillonella Veillonellaceae_Megasphaera Incertae Sedis XI_Gallicola Campylobacteraceae_Campylobacter Actinomycetales_Micrococcineae Incertae Sedis XI_Gemella Staphylococcaceae_Staphylococcus Ruminococcaceae_Ruminococcus Verrucomicrobiaceae_Akkermansia Porphyromonadaceae_Parabacteroides Erysipelotrichaceae Incertae Sedis Lachnospiraceae_Dorea Lachnospiraceae_Syntrophococcus Lachnospiraceae_Roseburia Peptostreptococcaceae Incertae Sedis Enterobacteriaceae_Klebsiella Ruminococcaceae_Other Clostridiales_Other Porphyromonadaceae_Porphyromonas Coriobacteriales_Coriobacterineae Bifidobacteriales_Bifidobacteriaceae Lachnospiraceae_Other Ruminococcaceae_Faecalibacterium Enterobacteriaceae_Shigella Clostridiaceae_Clostridiaceae 1 Burkholderiales_Other Enterococcaceae_Enterococcus Actinomycetales_Actinomycineae Lactobacillaceae_Lactobacillus Incertae Sedis XI_Peptoniphilus Prevotellaceae_Prevotella Streptococcaceae_Streptococcus Actinomycetales_Corynebacterineae Incertae Sedis XI_Finegoldia Incertae Sedis XI_Anaerococcus Lachnospiraceae Incertae Sedis Bacteroidaceae_Bacteroides
Flavobacteriaceae_Cloacibacterium Incertae Sedis XI_Finegoldia Enterococcaceae_Enterococcus Campylobacteraceae_Campylobacter Peptostreptococcaceae_Peptostreptococcus Incertae Sedis XIII_Mogibacterium Prevotellaceae_Other Actinomycetales_Other Veillonellaceae_Megasphaera Lachnospiraceae_Moryella Pseudomonadaceae_Pseudomonas Pasteurellaceae_Other SR1_genera_incertae_sedis_Other Comamonadaceae_Caenibacterium Pasteurellaceae_Aggregatibacter Lachnospiraceae_Oribacterium Erysipelotrichaceae_Bulleidia Veillonellaceae_Selenomonas Lactobacillales_Other Lachnospiraceae_Other Coriobacteriales_Coriobacterineae Bifidobacteriales_Bifidobacteriaceae Ruminococcaceae_Ruminococcus Incertae Sedis XI_Parvimonas Incertae Sedis XI_Anaerococcus Peptostreptococcaceae Incertae Sedis Actinomycetales_Corynebacterineae Porphyromonadaceae_Tannerella Fusobacteriaceae_Fusobacterium Clostridiales_Other TM7_genera_incertae_sedis_Other Fusobacteriaceae_Leptotrichia Porphyromonadaceae_Porphyromonas Flavobacteriaceae_Capnocytophaga Streptococcaceae_Lactovum Streptococcaceae_Other Flavobacteriaceae_Other Aerococcaceae_Abiotrophia Carnobacteriaceae_Carnobacteriaceae 2 Pasteurellaceae_Haemophilus Lactobacillaceae_Lactobacillus Lachnospiraceae Incertae Sedis Neisseriaceae_Neisseria Veillonellaceae_Veillonella Prevotellaceae_Prevotella Actinomycetales_Actinomycineae Incertae Sedis XI_Gemella Staphylococcaceae_Staphylococcus Actinomycetales_Micrococcineae Streptococcaceae_Streptococcus
Bifidobacteriales_Incertae sedis 10 Incertae Sedis XI_Helcococcus Enterobacteriaceae_Enterobacter Actinobacteridae_Other Moraxellaceae_Psychrobacter Bacillales_Other Actinomycetales_Propionibacterineae Lachnospiraceae_Moryella Incertae Sedis XI_Parvimonas Porphyromonadaceae_Other Actinomycetales_Micrococcineae Fusobacteriaceae_Sneathia Pseudomonadaceae_Flavimonas Enterobacteriaceae_Citrobacter Enterococcaceae_Melissococcus Incertae Sedis XI_Gallicola Erysipelotrichaceae_Catenibacterium Actinomycetales_Other Veillonellaceae_Veillonella Bacteroidaceae_Bacteroides Lactobacillales_Other Aerococcaceae_Other Veillonellaceae_Megasphaera Enterobacteriaceae_Shigella Aerococcaceae_Facklamia Veillonellaceae_Dialister Veillonellaceae_Anaeroglobus Clostridiales_Other Campylobacteraceae_Campylobacter Mycoplasmataceae_Mycoplasma Aerococcaceae_Aerococcus Porphyromonadaceae_Porphyromonas Staphylococcaceae_Staphylococcus Ruminococcaceae_Other Incertae Sedis XV_Dethiosulfovibrio Lachnospiraceae_Other Mycoplasmataceae_Ureaplasma Coriobacteriales_Coriobacterineae Actinomycetales_Actinomycineae Incertae Sedis XI_Peptoniphilus Incertae Sedis XI_Finegoldia Prevotellaceae_Prevotella Enterobacteriaceae_Escherichia Streptococcaceae_Streptococcus Actinomycetales_Corynebacterineae Incertae Sedis XI_Anaerococcus Pseudomonadaceae_Pseudomonas Enterococcaceae_Enterococcus Bifidobacteriales_Bifidobacteriaceae Lactobacillaceae_Lactobacillus
017_C_RS 039_B_RS 056_A_RS 057_A_RS 003_B_RS 024_A_RS 020_A_RS 050_A_RS 043_C_RS 013_A_RS 015_C_RS 048_A_RS 059_A_RS 009_B_RS 020_B_RS 001_A_RS 047_A_RS 058_A_RS 036_C_RS 027_B_RS 001_B_RS 028_A_RS 060_A_RS 060_C_RS 016_A_RS 023_C_RS 011_A_RS 049_A_RS 052_A_RS 017_A_RS 018_A_RS 009_A_RS 001_C_RS 007_B_RS 004_A_RS 007_A_RS 015_B_RS 027_A_RS 038_A_RS 021_A_RS 008_B_RS 044_A_RS 033_A_RS 031_A_RS 039_A_RS 013_B_RS 029_A_RS 048_B_RS 041_A_RS 007_C_RS_rep2 007_C_RS_rep1 036_A_RS 002_B_RS 019_A_RS_rep1 019_A_RS_rep2 023_A_RS 052_C_RS 033_B_RS 008_C_RS 042_A_RS 026_A_RS 046_A_RS 053_A_RS 005_A_RS 023_B_RS 014_A_RS 022_A_RS 002_A_RS 041_B_RS 012_C_RS 059_B_RS_rep2 059_B_RS_rep1 022_B_RS 054_B_RS 047_C_RS 030_B_RS 032_A_RS 017_B_RS 030_A_RS 035_A_RS
A
which was not associated with the presence of specific bacteria in the urine microbiota. The sample of a patient with infectious complications after removal of a ureteral stent was dominated by Enterococcus species (94%). However, similar relative abundances of Enterococcus were also identified in two patients without reported infections.
Figure 4: Hierarchical clustering of urine (A), oral swab (B) and rectal swab (C) samples based on genus level microbiota compositions. Dendrograms show sample similarities based on similar patterns in relative abundance of microbiota members based on Euclidean distance calculations. Rows are also clustered hierarchically based on similar relative abundances across samples (dendrogram not shown). Only the 50 most abundant microbiota members are shown for each sample type.
American Journal of Transplantation 2014; 14: 416–427
421
Fricke et al
results showed that urine and rectal and oral and rectal samples, but not urine and oral samples, showed similar overlap (219, 193 and 16 OTUs, respectively). As contamination would not be expected to occur between oral and rectal swab samples, these results suggest that OTU overlap results from related bacteria colonizing different body sites rather than contamination. To further address potential cross-contamination of urine and rectal samples and of both sample types with skin bacteria, the ratios of relative abundances of all bacterial genera shared between urine and rectal sample pairs were calculated (Figure 5). For this analysis, intrinsic urine microbiota members were expected to always be present at high concentrations in urine and at low concentrations in rectal samples, whereas intrinsic intestinal inhabitants were expected to show the opposite effect. The underlying assumption was supported by the identification of the known urinary tract inhabitant Ureaplasma (16) and other bacterial genera previously identified in urine, such as Megasphaera and Mycoplasma (17) among the predicted intrinsic urine microbiota. Predicted intrinsic intestinal inhabitants included among others known intestinal bacteria from the genera Anaerotruncus, Bacteroides and Veillonella. Further it was assumed that skin contaminants would only be present at high concentrations in a few urine and rectal
samples, rectal contaminants would mostly show high abundance in rectal samples and occasionally in urine samples, and urine contaminants mostly in urine samples and occasionally in rectal samples. Based on this assumption, predicted contaminants from the skin included, among others, the genera Staphylococcus and Streptococcus that include important skin commensals and pathogens (18). Predicted urine contaminants originating from the intestinal microbiota included Dorea, Escherichia, Finegoldia and an unknown member of the Lachnospiraceae and predicted rectal swab contaminants originating from urine included Sneathia, Lactobacillus and an unknown member of the Bifidobacteriaceae. Overall, these results demonstrate not only that different body sites are likely to nurture specific microbial communities but also that there is frequent exchange of individual members of these communities between body sites, which is likely to play a role for infectious processes.
Longitudinal microbiota analysis Principal coordinate analysis was used to visualize sample relationships based on phylogenetic distance. Overall, urine, oral and rectal samples showed substantial diversity when compared cross-sectionally between individuals and longitudinally between pre- and posttransplant time points (Figure 3). No significant changes in overall phylogenetic
Figure 5: Cross-contamination between urine and rectal samples. The relative abundance ratios of genera shared between urine and rectal samples from the same patients were plotted using either rectal (left plot) or urine (right plot) concentrations as the denominator. Red genus names: predicted intrinsic urine microbiota members; dark blue genus names: predicted intrinsic intestinal microbiota members; green genus names: skin contaminants in either urine or rectal samples; light blue genus names: rectal contaminants in urine samples; orange genus names: urine contaminants in rectal samples.
422
American Journal of Transplantation 2014; 14: 416–427
Microbiota Changes During Renal Transplantation
distance (weighted and unweighted UniFrac analysis) were measured in pair-wise comparisons of samples from any two time points for either oral or rectal swab samples (Figure S3). Since most urine samples were dominated by a single highly abundant bacterial taxon, similar comparisons were not carried out for this sample type. Similarly, with the exception of rectal swabs, which showed a significant drop in microbial diversity as measured by Shannon diversity index 1 month after transplantation compared to before transplantation (p ¼ 0.01; Wilcoxon rank sum test), no significant change in diversity was found over time in any of the compared sample types or for any of the time points (Figure S1). In order to identify specific microbiota members that exhibited significant changes in relative abundance over time, microbiota compositions from urine, oral and rectal sample types were compared longitudinally over the three sample time points using repeated measures analysis (Table 4). In general, more significant changes were observed between pretransplant and 1-month posttransplant time points, than between 1-month and 6-month posttransplant time points, suggesting that the transplant episode and associated changes in organ function or exposure to high dose immunosuppression and perioperative antibiotics resulted in major changes in the microbiota structure. No common trends were identified between longitudinal changes of the different samples types. In urine and oral swab samples, the majority of bacterial genera with significant changes between pre- and 1-month postransplantation time points belonged to the phylum Proteobacteria, and in rectal swabs to the Firmicutes. Microbiota compositions and serum creatinine levels To determine whether renal function and creatinine levels were reflected in microbiota compositions, and to try to determine large group differences, samples collected at posttransplant time points and corresponding to high (>4 mg/dL) and low (
C
Copyright 2013 The American Society of Transplantation and the American Society of Transplant Surgeons doi: 10.1111/ajt.12588
Human Microbiota Characterization in the Course of Renal Transplantation W. F. Fricke1,*, C. Maddox1, Y. Song1 and J. S. Bromberg2,*
Received 17 July 2013, revised 13 November 2013 and accepted for publication 14 November 2013
1
Institute for Genome Sciences, University of Maryland School of Medicine, Baltimore, MD 2 Department of Surgery and Microbiology and Immunology, University of Maryland School of Medicine, Baltimore, MD Corresponding authors: W. Florian Fricke, [email protected]; Jonathan S. Bromberg, [email protected]
Recent studies demonstrate that the human microbiota, the collection of microorganisms growing on and in individuals, have numerous bidirectional interactions with the host, influencing immunity, resistance to infection, inflammation and metabolism. Little has been done to study the potential associations between microbiota composition and transplant outcome. Here, we investigated the longitudinal changes in the blood, urinary, oral and rectal microbiota of renal allograft recipients before and at 1 and 6 months after transplantation. The results showed major changes in microbiota composition as a result of the transplant episode and associated medications, and these changes persisted over time. The high interindividual variation as well as differences in response to transplantation suggested that it is unlikely that the same specific microbiota members can serve as universal diagnostic markers. Rather, longitudinal changes in each individual’s microbiota have the potential to be indicative of health or disease. Use of sensitive nucleic acid–based testing showed that urine, irrespective of disease states, more often harbors a diverse microbiota than appreciated by conventional culture techniques. These results lay the groundwork to construct more comprehensive future investigations to identify microbiota characteristics that can serve as diagnostic markers for transplant health and to guide intervention strategies to improve transplant outcome. Keywords: Human microbiota, immunosuppression, renal transplantation, transplant outcome, urine microbiota Abbreviations: EDTA, ethylenediaminetetraacetic acid; FDR, false discovery rate; OTU, operational taxonomic unit; PBS, phosphate-buffered saline; PCR, polymerase chain reaction; SD, standard deviation 416
Introduction The successful results of renal transplantation are mainly due to progress in immunosuppressant therapy. Unfortunately, increasing immunosuppression leads to a higher risk of opportunistic infections, which are associated with organ rejection or dysfunction. The clinical outcomes of renal allograft recipients have improved in part because of a decline in infection-related morbidity and mortality. Prior to 1980, 60% of kidney transplant recipients developed at least one serious infection during the first year posttransplant, with mortality rates approaching 50% (1). Although the current infection-related 1-year mortality has been reduced to 90% received Campath. Pediatric and multi-organ recipients were excluded as well as patients enrolled in other interventional studies with investigational drugs. However, patients enrolled in observational studies or studies with FDA-approved drugs were considered for enrollment. All patients received perioperative cefazolin, except for a few patients who were allergic to penicillin and received ciprofloxacin. All patients also received sulfamethoxazoletrimethoprim for 6 months after transplantation, or dapsone if patients were allergic to sulfonamides. Information about infections and rejections after transplantation, and additional medications were collected. A total of 60 patients were enrolled. Samples were obtained just prior to transplantation and at 1 and 6 months after transplantation, and included the following: (a) urine, (b) oral swabs, (c) rectal swabs and (d) blood (from a subset of 23 patients). The demographics of the population were 63% male, 60% Caucasian/40% African-American, age range 30–79 (mean 58, median 58). The study was approved by the University of Maryland Institutional Review Board (#HP-00049438).
16S rRNA gene fragment amplification Hypervariable regions V1–V3 of the 16S rRNA gene fragments were amplified using a sevenfold degenerate primer 27F-YMþ3 (10) and a barcoded 534 reverse primer. An overview of the primers, including barcodes and Roche/454 Life Sciences adaptors, is given in the HMP Consortium 16S 454 sequencing protocol (http://www.hmpdacc.org/doc/ 16S_Sequencing_SOP_4.2.2.pdf). DNA amplification of 16S rRNA genes was performed using AccuPrime Taq DNA polymerase High Fidelity (Invitrogen) and 12 mL of template DNA in a total reaction volume of 25 mL, following the AccuPrime product protocol. Reactions were run in a PTC-100 thermal controller (MJ Research, Waltham, MA) using the following cycling parameters: 3 min of denaturation at 948C, followed by 30 cycles of 30 s at 948C (denaturing), 30 s at 528C (annealing) and 45 s at 688C (elongation), with a final extension at 688C for 5 min. Negative controls (no template DNA) were included for each amplification and barcoded primer pair. The presence of amplicons was confirmed by gel electrophoresis on a 1% agarose gel and staining with ethidium bromide. Polymerase chain reaction (PCR) product concentrations were estimated visually using agarose gel electrophoresis and MassRuler low range DNA ladder (Thermo Scientific, Waltham, MA). Equimolar amounts (75 ng) of the PCR amplicons were mixed in a single tube. Amplification primers and reaction buffer were removed using the AMPure Kit (Agencourt, Brea, CA) and purified amplicon mixtures sequenced by Roche/454 FLX pyrosequencing using 454 Life Sciences primer A by the Genomics Resource Center at the Institute for Genome Sciences, University of Maryland School of Medicine, using protocols recommended by the manufacturer.
Sample collection Immediately after collection, oral (inner cheek) and rectal FLOQSwabs (Copan, Murrieta) were immersed in 2 mL of the nucleic acid-preserving reagent RNAlater (Qiagen, Valencia, CA) and stored at 808C. Urine samples were stored in vacutainers without modification (urine samples) and blood samples were stored in tubes coated with anticoagulants (ethylenediaminetetraacetic acid [EDTA]) (BD, Franklin Lakes, NJ) at 48C. Urine samples were processed within 14 days, and oral and rectal swabs within 4 weeks after collection.
Nucleic acid isolation Oral and rectal swabs were vortexed with the RNAlater in the collection tubes for 10 min at room temperature and 1 mL of the resulting microbial suspension used for the subsequent DNA extraction. Urine samples (2 mL) were centrifuged at 5000g for 8 min and the entire pellet resuspended in 0.6 mL 1 phosphate-buffered saline and used for DNA extraction. Of the blood samples 1 mL was pretreated with the MolYsis Basic kit (Molzym, Germany) following the manufacturer’s recommendations to remove human DNA using their differential lysis protocol. For all microbial suspension types, microbial cell lysis was initiated with two enzymatic incubations, first using 5 mL of lysozyme (10 mg/mL; Amresco, Solon, OH), 13 mL of mutanolysin (11.7 U/mL; Sigma–Aldrich, St. Louis, MO) and 3 mL of lysostaphin (4.5 U/mL; Sigma–Aldrich) for an incubation of 30 min at 378C, and, second, using 10 mL Proteinase K (20 mg/mL; Research Products International, Mount Prospect, IL), 50 mL of 10% SDS and 2 mL RNase (10 mg/mL) for an incubation of 45 min at 568C. After the enzyme treatments, cells were disrupted by bead beating in tubes with Lysing Matrix B (0.1 mm silica spheres; MP Biomedicals, Solon, OH), at 6 m/s for 40 s at room temperature in a FastPrep-24 (MP Biomedicals). The resulting crude lysate was processed using the ZR Fecal DNA mini-prep kit (Zymo, Irvine, CA) according to the manufacturer’s recommendation. The samples were eluted with 0.08 mL of ultra pure water into separate tubes. DNA concentrations in the samples were measured using the Quant-iT PicoGreen dsDNA assay kit (Invitrogen Molecular Probes, Carlsbad, CA).
American Journal of Transplantation 2014; 14: 416–427
Sequencing and analysis 16S rRNA sequence reads were parsed and quality filtered, using the open source software package QIIME version 1.6.0 (http://qiime.sourceforge. net). Sequence reads were trimmed and quality filtered using the QIIME split.library default parameters. The automated CloVR-16S pipeline, which is part of the CloVR software package (11) was used to analyze de-multiplexed and quality filtered samples. Chimeras were removed with UCHIME (12). Each processed 16S rRNA sequence was classified using the RDP Na€ıve Bayesian Classifier (13) with a score filtering threshold of 0.5. All sequence reads were clustered into operational taxonomic units (OTUs) using a similarity threshold of 95%. Rarefaction curves were calculated based on OTU counts using the rarefaction.single routine of the Mothur package (14). Hierarchical clustering was performed using custom R scripts as implemented in the CloVR-16S pipeline. Significant differences in relative abundance of bacterial groups were determined with the Metastats program (15) for binary comparisons and with repeated measures analysis for longitudinal comparisons. For repeated measures analysis, relative abundances of bacterial genera from different sample types (urine, oral and rectal swabs), collected at different time points (A, pretransplantation; B, 1 month posttransplantation; C, 6 months posttransplantation) were analyzed using the PROC MIXED procedure of the SAS statistical software package (version 9.3; SAS Institute, Inc., Cary, NC) with time points as the repeated variable and sample types, time points and interactions of these factors as fixed effects. The output table ‘‘Type III tests of fixed effects’’ was used to evaluate whether among the fixed effects at least one mean was different from another mean. If the combination of sample types and time points showed overall significance (p < 0.05), least square means statements were used for Bonferroni-adjusted comparisons of the relative abundances between all time points, separately for each sample type. Adjusted p-values 4 mg/dL): 2 urine samples High-creatinine (>4 mg/dL): 8 oral samples High-creatinine (>4 mg/dL): 6 rectal samples Patients with rejection episode: 5 pretransplant oral samples
Urine, time point C: 6 samples Oral swabs, time point B: 21 samples Oral swabs, time point C: 9 samples Rectal swabs, time point B: 21 samples Rectal swabs, time point C: 13 samples Low-creatinine (80% relative abundance, in this case Streptococcus as shown by heatmap analysis (Figure 4B). On the phylum level, 31 samples contained >80% Firmicutes, indicating that in most samples different members of the Firmicutes were responsible for a majority of the bacterial species. Bacterial genera with at least 20% relative abundance in any of the oral swab samples were Prevotella (3 samples), Neisseria and Veillonella (each found in 2 samples), Lactobacillus and Gemella (each found in 1 sample) and unknown members of the Micrococcineae (14 samples), Actinomycineae (5 samples) and Lachnospiraceae (1 sample). Rectal swab microbiota Heatmap analysis of the 80 rectal samples showed that none were dominated at >20% by a single bacterial genus (Figure 4C), but members of the phylum Firmicutes accounted for >80% of the bacteria in 31 samples, indicating a diverse microbiota compared to urine and oral samples. Members of the phylum Bacteroides were found at >20% in 24 samples. Other bacterial phyla at >20% in at least one sample included Actinobacteria (eight samples) and Proteobacteria (two samples). Sample cross-contamination To study the extent of cross-contamination, shared OTUs were determined between all sample types (Figure S2). The
B
0
020_B_OS 009_A_OS 040_A_OS 011_A_OS 040_B_OS 051_A_OS 001_C_OS 015_B_OS 008_C_OS 006_A_OS 030_A_OS 023_C_OS 024_A_OS 059_C_OS 033_A_OS 038_A_OS 027_B_OS 011_B_OS 010_A_OS 055_B_OS 017_C_OS_rep2 005_A_OS 012_B_OS 026_A_OS 030_B_OS 034_A_OS 031_B_OS 016_B_OS 025_B_OS 007_A_OS 006_B_OS 048_B_OS 003_A_OS 031_A_OS 032_A_OS 048_A_OS 029_A_OS 021_A_OS 042_A_OS 037_A_OS 007_B_OS 023_A_OS 001_B_OS 022_A_OS 014_A_OS 017_B_OS 017_A_OS 043_A_OS 008_A_OS 025_A_OS 060_A_OS 004_A_OS 027_A_OS 057_C_OS 023_B_OS 041_B_OS 047_A_OS 059_A_OS 050_A_OS 049_A_OS 054_A_OS 017_C_OS_rep1 013_A_OS 002_B_OS 002_A_OS 018_A_OS_rep1 018_A_OS_rep2 004_B_OS 001_A_OS 041_A_OS 044_A_OS 053_A_OS 015_A_OS 012_A_OS_rep2 012_A_OS_rep1 013_B_OS 035_A_OS 006_C_OS_rep1 006_C_OS_rep2 036_A_OS 039_A_OS 056_A_OS 016_A_OS 020_A_OS 052_A_OS 045_A_OS 055_A_OS
055_B_UR 008_A_UR 019_A_UR 012_B_UR 059_B_UR 054_B_UR 007_C_UR 043_B_UR 030_B_UR 030_A_UR 006_C_UR 012_C_UR 017_C_UR 024_A_UR 040_B_UR 053_B_UR 060_A_UR 018_B_UR 002_C_UR 003_A_UR 001_A_UR 006_B_UR 013_B_UR 060_B_UR 050_B_UR 044_B_UR 037_B_UR 058_B_UR 045_B_UR 016_A_UR 035_A_UR 051_B_UR 021_A_UR 003_B_UR 003_C_UR 029_A_UR 029_B_UR
0.4
0.8
C
0
0.4
0.8
Porphyromonadaceae_Other Ruminococcaceae Incertae Sedis Rikenellaceae_Alistipes Lachnospiraceae_Coprococcus Erysipelotrichaceae_Coprobacillus Incertae Sedis XI_Other Pasteurellaceae_Haemophilus Veillonellaceae_Dialister Actinomycetales_Propionibacterineae Erysipelotrichaceae_Catenibacterium Leuconostocaceae_Weissella Ruminococcaceae_Subdoligranulum Veillonellaceae_Succinispira Veillonellaceae_Veillonella Veillonellaceae_Megasphaera Incertae Sedis XI_Gallicola Campylobacteraceae_Campylobacter Actinomycetales_Micrococcineae Incertae Sedis XI_Gemella Staphylococcaceae_Staphylococcus Ruminococcaceae_Ruminococcus Verrucomicrobiaceae_Akkermansia Porphyromonadaceae_Parabacteroides Erysipelotrichaceae Incertae Sedis Lachnospiraceae_Dorea Lachnospiraceae_Syntrophococcus Lachnospiraceae_Roseburia Peptostreptococcaceae Incertae Sedis Enterobacteriaceae_Klebsiella Ruminococcaceae_Other Clostridiales_Other Porphyromonadaceae_Porphyromonas Coriobacteriales_Coriobacterineae Bifidobacteriales_Bifidobacteriaceae Lachnospiraceae_Other Ruminococcaceae_Faecalibacterium Enterobacteriaceae_Shigella Clostridiaceae_Clostridiaceae 1 Burkholderiales_Other Enterococcaceae_Enterococcus Actinomycetales_Actinomycineae Lactobacillaceae_Lactobacillus Incertae Sedis XI_Peptoniphilus Prevotellaceae_Prevotella Streptococcaceae_Streptococcus Actinomycetales_Corynebacterineae Incertae Sedis XI_Finegoldia Incertae Sedis XI_Anaerococcus Lachnospiraceae Incertae Sedis Bacteroidaceae_Bacteroides
Flavobacteriaceae_Cloacibacterium Incertae Sedis XI_Finegoldia Enterococcaceae_Enterococcus Campylobacteraceae_Campylobacter Peptostreptococcaceae_Peptostreptococcus Incertae Sedis XIII_Mogibacterium Prevotellaceae_Other Actinomycetales_Other Veillonellaceae_Megasphaera Lachnospiraceae_Moryella Pseudomonadaceae_Pseudomonas Pasteurellaceae_Other SR1_genera_incertae_sedis_Other Comamonadaceae_Caenibacterium Pasteurellaceae_Aggregatibacter Lachnospiraceae_Oribacterium Erysipelotrichaceae_Bulleidia Veillonellaceae_Selenomonas Lactobacillales_Other Lachnospiraceae_Other Coriobacteriales_Coriobacterineae Bifidobacteriales_Bifidobacteriaceae Ruminococcaceae_Ruminococcus Incertae Sedis XI_Parvimonas Incertae Sedis XI_Anaerococcus Peptostreptococcaceae Incertae Sedis Actinomycetales_Corynebacterineae Porphyromonadaceae_Tannerella Fusobacteriaceae_Fusobacterium Clostridiales_Other TM7_genera_incertae_sedis_Other Fusobacteriaceae_Leptotrichia Porphyromonadaceae_Porphyromonas Flavobacteriaceae_Capnocytophaga Streptococcaceae_Lactovum Streptococcaceae_Other Flavobacteriaceae_Other Aerococcaceae_Abiotrophia Carnobacteriaceae_Carnobacteriaceae 2 Pasteurellaceae_Haemophilus Lactobacillaceae_Lactobacillus Lachnospiraceae Incertae Sedis Neisseriaceae_Neisseria Veillonellaceae_Veillonella Prevotellaceae_Prevotella Actinomycetales_Actinomycineae Incertae Sedis XI_Gemella Staphylococcaceae_Staphylococcus Actinomycetales_Micrococcineae Streptococcaceae_Streptococcus
Bifidobacteriales_Incertae sedis 10 Incertae Sedis XI_Helcococcus Enterobacteriaceae_Enterobacter Actinobacteridae_Other Moraxellaceae_Psychrobacter Bacillales_Other Actinomycetales_Propionibacterineae Lachnospiraceae_Moryella Incertae Sedis XI_Parvimonas Porphyromonadaceae_Other Actinomycetales_Micrococcineae Fusobacteriaceae_Sneathia Pseudomonadaceae_Flavimonas Enterobacteriaceae_Citrobacter Enterococcaceae_Melissococcus Incertae Sedis XI_Gallicola Erysipelotrichaceae_Catenibacterium Actinomycetales_Other Veillonellaceae_Veillonella Bacteroidaceae_Bacteroides Lactobacillales_Other Aerococcaceae_Other Veillonellaceae_Megasphaera Enterobacteriaceae_Shigella Aerococcaceae_Facklamia Veillonellaceae_Dialister Veillonellaceae_Anaeroglobus Clostridiales_Other Campylobacteraceae_Campylobacter Mycoplasmataceae_Mycoplasma Aerococcaceae_Aerococcus Porphyromonadaceae_Porphyromonas Staphylococcaceae_Staphylococcus Ruminococcaceae_Other Incertae Sedis XV_Dethiosulfovibrio Lachnospiraceae_Other Mycoplasmataceae_Ureaplasma Coriobacteriales_Coriobacterineae Actinomycetales_Actinomycineae Incertae Sedis XI_Peptoniphilus Incertae Sedis XI_Finegoldia Prevotellaceae_Prevotella Enterobacteriaceae_Escherichia Streptococcaceae_Streptococcus Actinomycetales_Corynebacterineae Incertae Sedis XI_Anaerococcus Pseudomonadaceae_Pseudomonas Enterococcaceae_Enterococcus Bifidobacteriales_Bifidobacteriaceae Lactobacillaceae_Lactobacillus
017_C_RS 039_B_RS 056_A_RS 057_A_RS 003_B_RS 024_A_RS 020_A_RS 050_A_RS 043_C_RS 013_A_RS 015_C_RS 048_A_RS 059_A_RS 009_B_RS 020_B_RS 001_A_RS 047_A_RS 058_A_RS 036_C_RS 027_B_RS 001_B_RS 028_A_RS 060_A_RS 060_C_RS 016_A_RS 023_C_RS 011_A_RS 049_A_RS 052_A_RS 017_A_RS 018_A_RS 009_A_RS 001_C_RS 007_B_RS 004_A_RS 007_A_RS 015_B_RS 027_A_RS 038_A_RS 021_A_RS 008_B_RS 044_A_RS 033_A_RS 031_A_RS 039_A_RS 013_B_RS 029_A_RS 048_B_RS 041_A_RS 007_C_RS_rep2 007_C_RS_rep1 036_A_RS 002_B_RS 019_A_RS_rep1 019_A_RS_rep2 023_A_RS 052_C_RS 033_B_RS 008_C_RS 042_A_RS 026_A_RS 046_A_RS 053_A_RS 005_A_RS 023_B_RS 014_A_RS 022_A_RS 002_A_RS 041_B_RS 012_C_RS 059_B_RS_rep2 059_B_RS_rep1 022_B_RS 054_B_RS 047_C_RS 030_B_RS 032_A_RS 017_B_RS 030_A_RS 035_A_RS
A
which was not associated with the presence of specific bacteria in the urine microbiota. The sample of a patient with infectious complications after removal of a ureteral stent was dominated by Enterococcus species (94%). However, similar relative abundances of Enterococcus were also identified in two patients without reported infections.
Figure 4: Hierarchical clustering of urine (A), oral swab (B) and rectal swab (C) samples based on genus level microbiota compositions. Dendrograms show sample similarities based on similar patterns in relative abundance of microbiota members based on Euclidean distance calculations. Rows are also clustered hierarchically based on similar relative abundances across samples (dendrogram not shown). Only the 50 most abundant microbiota members are shown for each sample type.
American Journal of Transplantation 2014; 14: 416–427
421
Fricke et al
results showed that urine and rectal and oral and rectal samples, but not urine and oral samples, showed similar overlap (219, 193 and 16 OTUs, respectively). As contamination would not be expected to occur between oral and rectal swab samples, these results suggest that OTU overlap results from related bacteria colonizing different body sites rather than contamination. To further address potential cross-contamination of urine and rectal samples and of both sample types with skin bacteria, the ratios of relative abundances of all bacterial genera shared between urine and rectal sample pairs were calculated (Figure 5). For this analysis, intrinsic urine microbiota members were expected to always be present at high concentrations in urine and at low concentrations in rectal samples, whereas intrinsic intestinal inhabitants were expected to show the opposite effect. The underlying assumption was supported by the identification of the known urinary tract inhabitant Ureaplasma (16) and other bacterial genera previously identified in urine, such as Megasphaera and Mycoplasma (17) among the predicted intrinsic urine microbiota. Predicted intrinsic intestinal inhabitants included among others known intestinal bacteria from the genera Anaerotruncus, Bacteroides and Veillonella. Further it was assumed that skin contaminants would only be present at high concentrations in a few urine and rectal
samples, rectal contaminants would mostly show high abundance in rectal samples and occasionally in urine samples, and urine contaminants mostly in urine samples and occasionally in rectal samples. Based on this assumption, predicted contaminants from the skin included, among others, the genera Staphylococcus and Streptococcus that include important skin commensals and pathogens (18). Predicted urine contaminants originating from the intestinal microbiota included Dorea, Escherichia, Finegoldia and an unknown member of the Lachnospiraceae and predicted rectal swab contaminants originating from urine included Sneathia, Lactobacillus and an unknown member of the Bifidobacteriaceae. Overall, these results demonstrate not only that different body sites are likely to nurture specific microbial communities but also that there is frequent exchange of individual members of these communities between body sites, which is likely to play a role for infectious processes.
Longitudinal microbiota analysis Principal coordinate analysis was used to visualize sample relationships based on phylogenetic distance. Overall, urine, oral and rectal samples showed substantial diversity when compared cross-sectionally between individuals and longitudinally between pre- and posttransplant time points (Figure 3). No significant changes in overall phylogenetic
Figure 5: Cross-contamination between urine and rectal samples. The relative abundance ratios of genera shared between urine and rectal samples from the same patients were plotted using either rectal (left plot) or urine (right plot) concentrations as the denominator. Red genus names: predicted intrinsic urine microbiota members; dark blue genus names: predicted intrinsic intestinal microbiota members; green genus names: skin contaminants in either urine or rectal samples; light blue genus names: rectal contaminants in urine samples; orange genus names: urine contaminants in rectal samples.
422
American Journal of Transplantation 2014; 14: 416–427
Microbiota Changes During Renal Transplantation
distance (weighted and unweighted UniFrac analysis) were measured in pair-wise comparisons of samples from any two time points for either oral or rectal swab samples (Figure S3). Since most urine samples were dominated by a single highly abundant bacterial taxon, similar comparisons were not carried out for this sample type. Similarly, with the exception of rectal swabs, which showed a significant drop in microbial diversity as measured by Shannon diversity index 1 month after transplantation compared to before transplantation (p ¼ 0.01; Wilcoxon rank sum test), no significant change in diversity was found over time in any of the compared sample types or for any of the time points (Figure S1). In order to identify specific microbiota members that exhibited significant changes in relative abundance over time, microbiota compositions from urine, oral and rectal sample types were compared longitudinally over the three sample time points using repeated measures analysis (Table 4). In general, more significant changes were observed between pretransplant and 1-month posttransplant time points, than between 1-month and 6-month posttransplant time points, suggesting that the transplant episode and associated changes in organ function or exposure to high dose immunosuppression and perioperative antibiotics resulted in major changes in the microbiota structure. No common trends were identified between longitudinal changes of the different samples types. In urine and oral swab samples, the majority of bacterial genera with significant changes between pre- and 1-month postransplantation time points belonged to the phylum Proteobacteria, and in rectal swabs to the Firmicutes. Microbiota compositions and serum creatinine levels To determine whether renal function and creatinine levels were reflected in microbiota compositions, and to try to determine large group differences, samples collected at posttransplant time points and corresponding to high (>4 mg/dL) and low (
