Infections are the second major cause of mortality in patients with kidney disease and accompanying uremia. Both vascular access and non–access-related infections contribute equally to the infection-related deaths in patients with kidney disease. Dialysis is the most common cause of systemic infection by Candida albicans in these patients. C. albicans also reside in the gastrointestinal tract as a commensal fungus. However, the contribution of gut-derived C. albicans in non–access-related infections in kidney disease is unknown. Using a mouse model of kidney disease, we demonstrate that uremic animals showed increased gut barrier permeability, impaired mucosal defense, and dysbiosis. The disturbance in gut homeostasis is sufficient to drive the translocation of microbiota and intestinal pathogen Citrobacter rodentium to extraintestinal sites but not C. albicans. Interestingly, a majority of uremic animals showed fungal translocation only when the gut barrier integrity is disrupted. Our data demonstrate that uremia coupled with gut mucosal damage may aid in the translocation of C. albicans and cause systemic infection in kidney disease. Because most of the individuals with kidney disease suffer from some form of gut mucosal damage, these results have important implications in the risk stratification and control of non–access-related opportunistic fungal infections in these patients.

Kidney disease is a major public health problem in the 21st century (1). Infection is the second leading cause of mortality (20%) in patients with kidney disease (2, 3). Although vascular access–related infections are the primary cause of infection-related deaths in kidney disease, non–access-related infectious complications, although equally important, are often overlooked (4). A study showed that infection-related hospitalizations are not due to vascular access in 77% of identified cases; rather, the most common reason was an infection of unknown source (5). Even though advances in dialysis techniques have lowered the rate of access-related infections over the past decade, prevention of non–access-related infections is still a major problem. In most cases, the source of non–access-related infections in kidney disease remains unidentified, which is a major constraint in controlling infections in these patients.

Several predisposing factors make patients with kidney disease susceptible to infection (68). These include dialysis access, the presence of coexisting illnesses, vaccine hyporesponsiveness, immunosuppressive therapy, and uremia. Uremia is characterized by the retention of ∼900 metabolites in the blood in the absence of kidney function. A subset of these (∼100) have a profound impact on various biological systems and are termed uremic toxins (9). Uremia is implicated in immune dysfunction and is considered an independent risk factor for infections independent of vascular access (7, 10). Although the role of uremia in access-related infections has been extensively studied (11, 12), nothing is known about the impact of uremia in non–access-related infections.

The gastrointestinal (GI) tract harbors trillions of microorganisms referred to as the gut microbiota, which play an important role in digestion and host metabolism (13). The microbiota are also implicated in the development of metabolic, inflammatory, and infectious diseases if gut homeostasis is altered (14, 15). The intestinal epithelial cells maintain the symbiotic relationship between the microbiota and the host by separating them. The mucosal barrier function is regulated by continuous interaction between the microbiota and host immune cells. Such interactions are also critical for maintaining the balance between symbionts and pathobionts and prevent dysbiosis. Hence, altered gut homeostasis aids in the leakage of microorganisms into the circulation, leading to systemic infections and inflammatory diseases.

Accumulating evidence indicates that cross-talk between host and microbiota is pathophysiologically relevant in kidney disease. Uremia impacts both the composition and metabolism of the gut microbiota (16). Additionally, 80% of all uremic patients show GI manifestations, in which the gut epithelial lining is damaged (17, 18). In contrast, many uremic toxins that originated from microbial metabolism cause damage to renal and vascular cells (16, 19). The past decade has seen a major effort in understanding the role of gut-derived uremic toxins in kidney and vascular disease. However, not much emphasis has been placed on investigating whether gut microorganisms can cause non–access-related opportunistic infections in uremia.

Candida albicans causes a severe nosocomial systemic infection known as disseminated candidiasis (20). C. albicans accounts for 79% of all systemic fungal infections in patients with kidney disease (21). Invasive medical procedures are the major sources of disseminated candidiasis in these patients (21). C. albicans is also a commensal fungus in the GI tract of humans (22). In mouse models, the fungus generally resides in the GI tract and does not translocate unless the animals show neutropenia and compromised gut barrier integrity (2325). However, it is unclear whether commensal C. albicans can cause systemic infection in uremia.

In this report, using a well-validated mouse model of kidney disease and associated uremia, we demonstrate that disruption of the gut mucosal layer in uremia may aid in the translocation of C. albicans and cause systemic infection. Collectively, these results highlight the role of commensal C. albicans in causing non–vascular access-related opportunistic infection in a subset of kidney disease patients, in which uremia is coupled with gut mucosal damage.

C57BL/6NTac (wild-type [WT]) male mice were purchased from Taconic Biosciences (Germantown, NY). All mice were housed under specific pathogen-free conditions under the supervision of the Division of Laboratory Animal Resources. All animal experiments were conducted following National Institutes of Health guidelines under protocols approved by the University of Pittsburgh Institutional Animal Care and Use Committee (protocol no. 20087922).

To induce kidney dysfunction, mice were injected i.p. with 10 mg/kg of aristolochic acid I (AAI) (Sigma). Mice in the chemical control group were injected i.p. with 10 mg/kg aristolochic acid II (AAII) (Sigma), and control animals received an i.p. injection of equal volume of PBS. For the pharmacological inhibition study, mice were treated as follows: 1) a single i.p. injection of AAI at 10 mg/kg, 2) the probenecid + AAI group mice were i.p. injected with probenecid at 150 mg/kg bodyweight 30 min before the administration of AAI, and 3) the probenecid-only group mice were i.p. injected with probenecid at 150 mg/kg.

Mice received 2.5% dextran sulfate sodium (DSS) (36,000–50,000 m.w.) (MP Biomedicals) in their drinking water for 5 d. Control animals received autoclaved water for the entire period. Mice were monitored daily for body weight.

Mice were orally gavaged with 2 × 108 CFU of C. albicans (strain SC5314). For establishment of C. albicans intestinal colonization, mice were supplemented with ampicillin (1 mg/ml) in drinking water 2 d prior to oral C. albicans gavage. Thereafter, mice were maintained on ampicillin-supplemented drinking water throughout the experiment. C. albicans colonization in the GI tract was enumerated by culturing fecal contents on yeast extract peptone dextrose (YPD) agar supplemented with 0.010 mg/ml vancomycin and 0.100 mg/ml gentamicin. Liver and spleen tissues were homogenized and plated on YPD agar for the determination of systemic dissemination of C. albicans.

Mice were infected by oral gavage with 2 × 109 CFUs of C.rodentium (strain DBS100). Mice were weighed daily and monitored for signs of illness or distress. Bacterial counts were determined in freshly collected fecal pellets or aseptically collected liver and spleens by homogenization in PBS, followed by plating of serial dilutions of the homogenate on MacConkey agar plates.

A unilateral ureteral obstruction (UUO) kidney disease model was induced in 8-wk-old WT male mice by left ureteral ligation.

Serum blood urea nitrogen (BUN) was measured using BUN Enzymatic Kit (Bioo Scientific), and creatinine levels were measured with QuantiChrom Creatinine Assay Kit (BioAssay Systems).

Food was withheld from mice for 4 h prior to FITC–dextran gavage. FITC–dextran (4 kDa; Sigma) was resuspended in PBS at a concentration of 100 mg/ml and orally gavaged to each mouse at a dose of 40 mg/100 g body weight. Four hours later, mice were euthanized, and blood was collected immediately by retro-orbital bleeding in tubes containing anticoagulant. Plasma was isolated from blood samples. One hundred microliters of plasma was added to a black 96-well microplate in duplicate. The concentration of FITC in plasma was determined by spectrophotometry fluorometry with an excitation of 485 nm (20 nm bandwidth) and an emission wavelength of 528 nm (20 nm bandwidth) using standard serially diluted FITC–dextran. Plasma from mice gavaged with PBS was used to determine background.

Mice were euthanized, and aseptically collected liver and spleen samples were homogenized in PBS. Tissue homogenates were plated on brain heart infusion agar and YPD agar plates analysis of bacterial and fungal translocation, respectively.

Spleen and mesenteric lymph nodes (MLN) were harvested from mice and subjected to mechanical dissociation to prepare single cell suspensions, followed by RBC lysis by ammonium–chloride–potassium lysing buffer (Thermo Fisher Scientific). Small intestinal lamina propria (SILP) leukocytes were isolated. Briefly, Peyer’s patches were carefully excised from a 10-cm piece of a terminal part of small intestine (SI). Intestinal tissues were opened longitudinally and washed with HBSS to remove luminal contents. Tissues were incubated with 20 ml of HBSS containing 5 mM EDTA for 20 min at 37°C in a shaking incubator to remove epithelial cells. Tissues were cut into small pieces and incubated in 10 ml of RPMI 1640 containing 0.3 mg/ml collagenase D and 0.1 mg/ml DNase I for 20 min at 37°C in a shaking water bath. Then, 10% FBS was added to stop the activity of digesting enzymes, and the tissue suspension was passed through a 70-microns cell strainer. Cell suspension was washed twice with RPMI 1640 containing 10% FBS and passed through a 40-μm cell strainer for removal of tissue debris.

RNA was extracted from intestinal tissues using RNeasy kits (QIAGEN, Valencia, CA). cDNA was synthesized by SuperScript III First Strand Kits (Thermo Fisher Scientific, Waltham, MA). Quantitative real-time PCR was performed with the PerfeCTa SYBR Green FastMix (Quanta BioSciences, Beverly, MA) and analyzed on an ABI 7300 real-time instrument. Primers were from obtained QuantiTect Primer Assays (QIAGEN). The expression of each gene was normalized to that of GAPDH.

Fecal DNA was isolated using QIAamp stool DNA extraction kit. Microbial community analysis used PCR amplification of the of 16S rRNA followed by sequencing on an Illumina MiSeq. Sequences from the Illumina MiSeq were deconvolved and then processed through the Center for Medicine and the Microbiome in-house sequence quality control pipeline, which includes dust low complexity filtering, quality value trimming, trimming of primers used for 16S rRNA gene amplification, and minimum read length filtering. Forward and reverse reads were merged into contigs and then processed through the Center for Medicine and the Microbiome’s Mothur-based 16S clustering and sequence annotation pipeline. Sequence taxonomic classifications was performed with the Ribosomal Database Project and Naive Bayesian classifier with the Silva reference database. Microbiota profiles were statistically quantified and analyzed using three distinct methods at the genus taxonomic level: distance-based methods (intersample difference) used the Manhattan distance metric, distribution-based methods (e.g., diversity) used the Tail statistic and Shannon index, and abundance-based methods used the additive-log ratio transformation. Sample time points (baseline and 10 d post–AAI injection) from the same subject were paired, and linear models were fit with the paired differences as the response (paired analyses).

The terminal ileum and colon tissues was fixed with 10% buffered formaldehyde and embedded in paraffin. Slices 4-μm thick were made, stained with H&E, and then observed by a pathologist blinded to this study design with light microscopy.

Sections from frozen SI and colon were fixed in 4% paraformaldehyde and stained with ZO-1 primary Ab (Invitrogen). The secondary Ab used was goat anti-rabbit Cy3. Sections were subsequently stained with DAPI nuclear stain. Images were acquired with EVOS FL Auto microscope (Life Technologies).

All data are expressed as mean ± SD. Statistical analyses were performed using the ANOVA, Mann–Whitney, or unpaired Student t test through GraphPad Prism 7 program: *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001. All experiments were performed a minimum of twice in independently performed replicates.

Using a mouse model of AAI nephropathy (AAN) (26), we tested whether uremia negatively impacts gut homeostasis and causes disseminated candidiasis. In this study, mice injected with AAI show kidney dysfunction and uremia (Fig. 1A). Control animals received PBS. AAII was used as a nonnephrotoxic control. We first measured the impact of uremia on intestinal barrier function in mice with AAN. Uremic mice showed a 10-fold increase in the gut barrier permeability than controls (Fig. 1B). The increase in barrier permeability was highest around 6–10 d post–AAI injection, the time point at which BUN level peaked in the uremic group, as shown before (26) (Fig. 1C, 1D). Consequently, uremic animals showed a reduced expression of ZO-1 tight junction protein in the epithelial lining of the SI and colon (Fig. 1E). There was an increase in the expression of a few tight junction protein genes (ZO1, occludin, and claudin 4), signifying the onset of repair responses in dysfunctional gut epithelium, as previously described (27) (Fig. 1F). The loss of barrier function in uremia was not due to mucosal damage of the SI and colon (Fig. 1G).

FIGURE 1.

Increased gut barrier permeability in AAI-injected mice.

C57BL/6 (WT) mice (n = 6–8) were either injected with AAI, PBS (control), or AAII. (A) Serum BUN and creatinine levels (n = 5) were measured at day 10 post–AAI injection. At (B) day 10 (n = 6–8) and (C) indicated time points post–AAI injection (n = 3–6), mice were gavaged with FITC–dextran and assessed for FITC–dextran concentration in the plasma. (D) Correlation between gut barrier permeability and BUN level (n = 8). (E) SI and colon sections were stained for ZO-1 expression. (F) Transcript expression of tight junction protein genes were evaluated by quantitative real-time PCR (n = 6–7). (G) H&E staining of SI and colon of uremic and control mice. Images from one of three mice/group for (E and G). Original magnification ×200. Data pooled from at least two independent experiments for (A–D and F) and expressed as mean ± SD (A, B, C, and F). Statistical analyses by Pearson correlation (D) and one-way ANOVA (A–C and F). **p < 0.01, ***p < 0.001, ****p < 0.0001.

FIGURE 1.

Increased gut barrier permeability in AAI-injected mice.

C57BL/6 (WT) mice (n = 6–8) were either injected with AAI, PBS (control), or AAII. (A) Serum BUN and creatinine levels (n = 5) were measured at day 10 post–AAI injection. At (B) day 10 (n = 6–8) and (C) indicated time points post–AAI injection (n = 3–6), mice were gavaged with FITC–dextran and assessed for FITC–dextran concentration in the plasma. (D) Correlation between gut barrier permeability and BUN level (n = 8). (E) SI and colon sections were stained for ZO-1 expression. (F) Transcript expression of tight junction protein genes were evaluated by quantitative real-time PCR (n = 6–7). (G) H&E staining of SI and colon of uremic and control mice. Images from one of three mice/group for (E and G). Original magnification ×200. Data pooled from at least two independent experiments for (A–D and F) and expressed as mean ± SD (A, B, C, and F). Statistical analyses by Pearson correlation (D) and one-way ANOVA (A–C and F). **p < 0.01, ***p < 0.001, ****p < 0.0001.

Close modal

We next determined whether kidney damage is responsible for the alteration in gut barrier permeability independent of uremia. To test this issue, we adapted an UUO model of renal fibrosis, in which one of the ureters is ligated below the kidney, whereas the contralateral ureter is left intact, allowing for normal function by the nonligated kidney (28). Consequently, UUO causes unilateral kidney damage but not uremia (Fig. 2A–C). Mice with UUO showed no changes in the gut barrier permeability in comparison with non-UUO control (Fig. 2D), indicating that the increase in gut permeability is due to uremia and not kidney damage per se. Moreover, to determine whether increased permeability is due to a direct effect of AAI on the intestinal epithelial cells, AAI-injected mice were treated with probenecid, an organic anion transporter inhibitor that prevents kidney dysfunction and uremia without neutralizing AAI (29) (Fig. 2E–G). Uremic mice treated with probenecid showed reduced FITC–dextran in the plasma, indicating that AAI does not exert any direct effect on the barrier permeability (Fig. 2H). These data suggest that uremia drives increased barrier permeability in mice with kidney disease (12).

FIGURE 2.

Uremia drives increased gut permeability.

(A) Schematic diagram of the experimental design. WT mice (n = 5–6) were subjected to UUO. At day 7 postsurgery, UUO and non-UUO control mice were gavaged with FITC–dextran and assessed for barrier permeability. (B) Kidney histopathology following Masson trichome staining, (C) serum BUN level, and (D) plasma FITC–dextran concentration were measured at day 7 postsurgery. (E) Schematic diagram of the experimental design. Groups of uremic mice (n = 6–7) were either treated with probenecid (AAI+PRB) or left untreated (AAI). Control mice received probenecid only (PRB only). Mice were evaluated for (F) kidney fibrosis, (G) serum BUN level, and (H) gut barrier permeability. Images from one of three mice/group for (B and F). Original magnification ×200. The data are pooled from at least two independent experiments for (C, D, G, and H) and expressed as mean ± SD (C, D, G, and H). Statistical analyses by Student t test (C and D) and one-way ANOVA (G and H). ns, statistically not significant. ****p < 0.0001.

FIGURE 2.

Uremia drives increased gut permeability.

(A) Schematic diagram of the experimental design. WT mice (n = 5–6) were subjected to UUO. At day 7 postsurgery, UUO and non-UUO control mice were gavaged with FITC–dextran and assessed for barrier permeability. (B) Kidney histopathology following Masson trichome staining, (C) serum BUN level, and (D) plasma FITC–dextran concentration were measured at day 7 postsurgery. (E) Schematic diagram of the experimental design. Groups of uremic mice (n = 6–7) were either treated with probenecid (AAI+PRB) or left untreated (AAI). Control mice received probenecid only (PRB only). Mice were evaluated for (F) kidney fibrosis, (G) serum BUN level, and (H) gut barrier permeability. Images from one of three mice/group for (B and F). Original magnification ×200. The data are pooled from at least two independent experiments for (C, D, G, and H) and expressed as mean ± SD (C, D, G, and H). Statistical analyses by Student t test (C and D) and one-way ANOVA (G and H). ns, statistically not significant. ****p < 0.0001.

Close modal

The innate and adaptive immune cells have a distinct yet complementary role in maintaining gut homeostasis and mucosal immune defense (30). In the SILP, the percentages of macrophage and dendritic cell subsets were comparable between the uremic and control mice (Fig. 3A, 3B). Strikingly, we observed a reduction in the percentage of neutrophils in the SILP of uremic mice rather than in controls (Fig. 3C). Uremic mice also demonstrated a reduction in the percentage of Th17 but not Th1 cells (Fig. 3D). Markedly, the percentage of IgA+CD11b+ plasmablasts in the SILP of uremic mice was reduced compared to the percentage in control animals (Fig. 3E). There was a trend toward diminished IgA+CD11b plasmablasts in uremia, although the difference between the groups did not achieve statistical significance. However, the frequency of regulatory T (Treg) cells was similar between the groups (Fig. 3F). We observed no difference in the frequencies of Th17, Th1, and Treg cells in the MLN between uremic and nonuremic groups (Fig. 3G, 3H). These results indicate that uremia negatively impacts the number of neutrophils, Th17, and IgA-producing plasmablasts in the SILP.

FIGURE 3.

Compromised gut mucosal immunity and dysbiosis in uremia.

At day 10 post–AAI injection, SILP (n = 5–19) were evaluated for the frequency of (A) macrophages (liveCD45+CD11b+ F4/80+CX3CR1+CD11c+; liveCD45+CD11b+F4/80+CX3CR1+CD11c), (B) dendritic cells (liveCD45+CD11b+CD103+CD11c+, liveCD45+CD11bCD103+CD11c+; liveCD45+CD11b+ CD103CD11c+), (C) neutrophils (liveCD45+CD11b+Ly6G+), (D) Th17 (liveCD45+CD4+IL-17+) and Th1 (liveCD45+CD4+IFN-γ+), (E) IgA-producing plasmablasts (liveCD45+CD11b+IgA+; liveCD45+CD11bIgA+), and (F) Treg cells (liveCD4+Foxp3+) by FACS at day 10 post–AAI injection. (G) Percentages of Th17 and Th1 cells in the MLN (n = 8–12) were determined by intracellular cytokine staining following in vitro stimulation with PMA/ionomycin. (H) Frequency of Treg cells was determined in the MLN by FACS. (I) At day 10 post–AAI injection, fecal pellets from uremic and control (n = 5) mice were subjected to targeted 16S rRNA sequencing. Data pooled from at least two independent experiments for (A–H) and expressed as mean ± SD (A–H). Statistical analyses by one-way ANOVA (A–H) and pair-wise using Wilcoxon rank sum test (I). *p < 0.05, **p < 0.01, ***p < 0.001.

FIGURE 3.

Compromised gut mucosal immunity and dysbiosis in uremia.

At day 10 post–AAI injection, SILP (n = 5–19) were evaluated for the frequency of (A) macrophages (liveCD45+CD11b+ F4/80+CX3CR1+CD11c+; liveCD45+CD11b+F4/80+CX3CR1+CD11c), (B) dendritic cells (liveCD45+CD11b+CD103+CD11c+, liveCD45+CD11bCD103+CD11c+; liveCD45+CD11b+ CD103CD11c+), (C) neutrophils (liveCD45+CD11b+Ly6G+), (D) Th17 (liveCD45+CD4+IL-17+) and Th1 (liveCD45+CD4+IFN-γ+), (E) IgA-producing plasmablasts (liveCD45+CD11b+IgA+; liveCD45+CD11bIgA+), and (F) Treg cells (liveCD4+Foxp3+) by FACS at day 10 post–AAI injection. (G) Percentages of Th17 and Th1 cells in the MLN (n = 8–12) were determined by intracellular cytokine staining following in vitro stimulation with PMA/ionomycin. (H) Frequency of Treg cells was determined in the MLN by FACS. (I) At day 10 post–AAI injection, fecal pellets from uremic and control (n = 5) mice were subjected to targeted 16S rRNA sequencing. Data pooled from at least two independent experiments for (A–H) and expressed as mean ± SD (A–H). Statistical analyses by one-way ANOVA (A–H) and pair-wise using Wilcoxon rank sum test (I). *p < 0.05, **p < 0.01, ***p < 0.001.

Close modal

Next, we assessed the relative change in the diversity and composition of gut microbiota in uremic and control animals between days 0 (baseline) and 10 post–AAI injection. There was no significant difference in the diversity of microbiota between uremic and control groups at day 10 post–AAI injection (Fig. 3I). Both control and AAII-injected animals demonstrated altered microbiota composition between days 0 and 10. Interestingly, uremic mice showed significantly greater change (p = 0.0094) in the overall microbiota composition than control animals at day 10 postinjection. Whereas the control group showed a reduction in the abundance of Tannerellaceae and Lactobacillus, uremic animals exhibited an increase in the Tannerellaceae. Collectively, these results show a modest impact of uremia on the composition of gut microbiota in uremic mice.

We assessed whether the uremia-driven disturbance in gut hemostasis can drive the translocation of gut microbiota. Although we were unable to detect any bacterial colonies in the control or AAII mice, 30–50% of uremic animals showed translocation of microbiota in the liver and spleen at day 10 post–AAI injection (Fig. 4A). Moreover, probenecid-treated uremic mice did not show any bacterial colony in the liver and spleen (Fig. 4B). The translocation of microbiota caused the activation of T cells in the MLN and spleen of uremic mice, as shown before (Fig. 4C, 4D) (12).

FIGURE 4.

Uremic mice exhibit translocation of gut microbiota.

(A) Mice (n = 6) were subjected to AAN and evaluated for the translocation of gut microbiota in the liver and spleen at day 10 post–AAI injection. Images from one of six mice/group. (B) Uremic mice (n = 6) were either treated with probenecid (AAI+PRB) or left untreated (AAI) and assessed for microbiota translocation in the liver and spleen. Uremic and control groups (n = 6–9) were evaluated for the activation of T cells in the (C) MLN (liveCD4+CD44hi) and (D) spleen (liveCD4+CD62LloCD44hi and liveCD8+CD62LloCD44hi) by FACS. Pooled data from at least two experiments for (A–D) and expressed as mean ± SD (C and D). Statistical analyses by one-way ANOVA (C and D). *p < 0.05, **p < 0.01, ****p < 0.0001.

FIGURE 4.

Uremic mice exhibit translocation of gut microbiota.

(A) Mice (n = 6) were subjected to AAN and evaluated for the translocation of gut microbiota in the liver and spleen at day 10 post–AAI injection. Images from one of six mice/group. (B) Uremic mice (n = 6) were either treated with probenecid (AAI+PRB) or left untreated (AAI) and assessed for microbiota translocation in the liver and spleen. Uremic and control groups (n = 6–9) were evaluated for the activation of T cells in the (C) MLN (liveCD4+CD44hi) and (D) spleen (liveCD4+CD62LloCD44hi and liveCD8+CD62LloCD44hi) by FACS. Pooled data from at least two experiments for (A–D) and expressed as mean ± SD (C and D). Statistical analyses by one-way ANOVA (C and D). *p < 0.05, **p < 0.01, ****p < 0.0001.

Close modal

We next evaluated the ability of harmful bacteria C. rodentium to cause systemic infection in uremic animals (Fig. 5A). C. rodentium is a murine intestinal pathogen that is closely related to human pathogens, including enteropathogenic Escherichia coli and enterohemorrhagic E. coli (31). Following oral infection, we observed more C. rodentium in the fecal pellet and cecal contents of uremic mice (Fig. 5B). Few control mice exhibited very low colony counts of C. rodentium in the liver and spleen, as shown before (32) (Fig. 5C). In contrast, 60–70% of the uremic animals showed higher C. rodentium burden in the liver and spleen. Our data suggest that gut microbiota and C. rodentium can translocate from the gut and cause systemic infection in uremia.

FIGURE 5.

Uremic mice exhibit translocation of C. rodentium.

(A) Schematic diagram of the experimental design. AAI, control, and AAII-injected mice (n = 12) were gavaged with C. rodentium at day 3 post–AAI injection. At day 7, C. rodentium burden in the (B) fecal pellet and cecal content and (C) liver and spleen were measured. Pooled data from at least two experiments for (B and C) and expressed as mean ± SD (B and C). Statistical analyses by one-way ANOVA (B and C). *p < 0.05.

FIGURE 5.

Uremic mice exhibit translocation of C. rodentium.

(A) Schematic diagram of the experimental design. AAI, control, and AAII-injected mice (n = 12) were gavaged with C. rodentium at day 3 post–AAI injection. At day 7, C. rodentium burden in the (B) fecal pellet and cecal content and (C) liver and spleen were measured. Pooled data from at least two experiments for (B and C) and expressed as mean ± SD (B and C). Statistical analyses by one-way ANOVA (B and C). *p < 0.05.

Close modal

The role of gut-derived C. albicans in non–access-related infections in kidney disease is unknown. Based on our data (Figs. 4, 5), we hypothesize that C. albicans could translocate and cause disseminated candidiasis in uremia. C. albicans is not a commensal fungus in mice. Hence, colonization of the fungus requires alteration in gut homeostasis (33). Following oral gavage, uremic mice showed a modest increase in fungal colonization in the gut at day 7 postinjection (Fig. 6A, 6B). We were unable to detect any fungal colony in the liver and spleen of uremic animals at this time point (Fig. 6C). We also treated uremic and control mice with oral antibiotic to aid in robust colonization (34) (Fig. 6D, 6E). However, uremic animals showed no fungal dissemination in the organs (Fig. 6F). Thus, unlike C. rodentium, C. albicans cannot translocate from the gut in uremia.

FIGURE 6.

Uremic mice show fungal translocation following DSS treatment.

(A) Schematic representation of the experimental design. Mice (n = 10–15) were gavaged with C. albicans at day 2 post–AAI injection. At day 8, C. albicans and microbiota burden in the (B) fecal pellet and (C) liver and spleen were evaluated. (D) Schematic diagram of the experimental plan. Mice (n = 6–13) were fed with ampicillin in the drinking water throughout the experiment. At day 10 post–oral fungal infection, mice were either injected with AAI or PBS. Fungal and bacterial CFU in the (E) fecal pellet and (F) liver and spleen were determined. (G) Schematic representation of the experimental plan. Oral antibiotic–treated animals were fed with 2.5% DSS in water 3 d after AAI injection. (H) Survival (n = 4–8) was evaluated for 9 d post–AAI injection. (I) Mice were evaluated for the translocation of C. albicans in the liver. Pooled data from at least two experiments for (B, C, E, F, H, and I) and expressed as mean ± SD (B and E). Statistical analyses by one-way ANOVA (B), Mann–Whitney t test (E), and log-rank test (H). *p < 0.05, **p < 0.01.

FIGURE 6.

Uremic mice show fungal translocation following DSS treatment.

(A) Schematic representation of the experimental design. Mice (n = 10–15) were gavaged with C. albicans at day 2 post–AAI injection. At day 8, C. albicans and microbiota burden in the (B) fecal pellet and (C) liver and spleen were evaluated. (D) Schematic diagram of the experimental plan. Mice (n = 6–13) were fed with ampicillin in the drinking water throughout the experiment. At day 10 post–oral fungal infection, mice were either injected with AAI or PBS. Fungal and bacterial CFU in the (E) fecal pellet and (F) liver and spleen were determined. (G) Schematic representation of the experimental plan. Oral antibiotic–treated animals were fed with 2.5% DSS in water 3 d after AAI injection. (H) Survival (n = 4–8) was evaluated for 9 d post–AAI injection. (I) Mice were evaluated for the translocation of C. albicans in the liver. Pooled data from at least two experiments for (B, C, E, F, H, and I) and expressed as mean ± SD (B and E). Statistical analyses by one-way ANOVA (B), Mann–Whitney t test (E), and log-rank test (H). *p < 0.05, **p < 0.01.

Close modal

Most kidney disease patients suffer from some form of GI manifestations, in which the gut mucosa is damaged (17). These include perforation of intestine, intestinal ulcer, uremic colitis, ischemic colitis, diverticular disease, and intestinal hemorrhage. Hence, we tested whether C. albicans can cause systemic infection when the gut mucosa is disrupted in uremia. Accordingly, uremic and control mice were fed with antibiotic to favor colonization followed by oral gavage with 2.5% DSS to damage the epithelial lining (23) (Fig. 6G). AAI-injected mice treated with DSS showed reduced survival (30%) in comparison with AAI only, control only, and control + DSS animals (100%) (Fig. 6H). Control animals with DSS showed no fungal dissemination in the liver, as shown before (23) (Fig. 6I). There was no fungal translocation in uremic mice not receiving DSS. Interestingly, 65% of the uremic animals fed with DSS showed fungal translocation in the liver. These data indicate that C. albicans can translocate and cause systemic infection only when the mucosal barrier is damaged in the majority of uremic animals. Thus, it is reasonable to speculate that uremic individuals with GI manifestations may have a higher risk of systemic infection by gut-derived C. albicans.

Prevention of infection is one of the few avenues to reduce hospitalizations, control costs, and improve quality of life for patients with kidney disease. In the past decade, technological advances in the dialysis procedure have lowered the incidence of access-related infections. However, it has not been successful in preventing infection-related hospitalizations in these patients (2, 3). This is partly due to an alarming rise in the prevalence of previously unappreciated non–access-related infections. Currently, our knowledge about the risk and source of non–access-related infections in patients with kidney disease is surprisingly rudimentary. Using a clinically relevant mouse model of kidney disease and associated uremia, we show that gut commensal C. albicans causes systemic infection when uremia is coupled with mucosal damage. These results identify gut-derived commensal C. albicans as a source of non–access-related systemic infection in patients with kidney disease and showing GI damage.

Our data show that the loss of barrier function in uremic mice is due to an increase in the gut barrier permeability and mucosal damage. The level of uremia correlates with the loss of barrier function, indicating that uremia is directly implicated in the increased gut permeability. Although our study provided evidence of leaky gut in 100% of the uremic mice, only 30–50% of the animals consistently showed translocation of microbiota. These data imply that factors other than an increase in gut permeability play a major role in bacterial translocation in uremia. Moreover, when microbiota and C. rodentium showed translocation, C. albicans failed to do so. The fungal yeast (2–10 μm) and hyphae (>10 μm) are bigger in size than bacteria (0.5–5 μm), which may act as an impediment for the fungal yeast and hyphal form to pass through the tight junctions of GI epithelial cells. Additionally, control or uremic mice do not show any fungal translocation without mucosal damage. These data argue against the fact that C. albicans can translocate by causing damage to GI epithelial cells as proposed by others (35).

Dysbiosis observed in this study may be due to iatrogenic causes or uremia per se. Loss of kidney function leads to diffusion of urea in the GI tract. Subsequent hydrolysis of urea by urease expressed by some gut microbes results in the formation of large quantities of ammonia, which could affect the growth of commensal bacteria (16, 19). We observed a modest change in the gut microbiota in our mouse model of AAN. This is in contrast to chronic kidney disease patients, in which a markedly altered change in terms of quantity and quality of microbiota is evident (16, 19). There may be several reasons for this discrepancy. First, AAN is an acute kidney injury model, in which the gut bacteria are exposed to uremic toxins for a relatively short period of time (i.e., 7–10 d). Second, mice and human microbiota differ considerably, and their susceptibility to uremia may reflect the difference in alterations in gut bacterial content (36). Finally, uremic toxins differ between mice and humans, making it difficult to compare their impact on the gut microbiota (37).

We observed a decrease in the percentages of innate and adaptive immune cells in the SILP of uremic mice. Uremia-induced neutrophil apoptosis may account for the reduced number of neutrophils in the uremic gut (38). This observation is in line with a previous report showing that immunosuppressive treatment–induced neutropenia is required for the fungus to translocate and cause systemic infection in nonuremic animals (23). Additionally, IgA-producing plasmablast deficiency in uremic mice can be simultaneously mediated by increased B cell apoptosis and reduced expression of BAFF-R, as demonstrated before (39). The role of IgA in C. albicans colonization and translocation is poorly understood. Interestingly, we did not see any change in Treg cells in the gut, a hallmark of chronic kidney disease patients (40).

The GI symptoms are reported in up to 80% of kidney disease patients (17). Some of these symptoms, including intestinal necrosis, spontaneous colonic perforation, uremic colitis, gastric ulcer, GI hemorrhage, and acute diverticulitis, result in the damage of mucosa and development of sepsis. Thus, it is reasonable to speculate that uremic individuals with GI manifestations may have a higher risk of systemic infection by commensal C. albicans. To date, no studies have looked at the prevalence of disseminated candidiasis in uremic patients with GI manifestations. Additionally, these findings have compelling implications in the risk stratification and clinical management of infection control and prevention in patients with kidney disease.

We thank Drs. Sarah Gaffen, Tim Hand, and Mandy McGeachy for suggestions.

This work was supported by National Institutes of Health Grants AI142354, DK104680, and R21AI45242 to P.S.B.

Abbreviations used in this article:

     
  • AAI

    aristolochic acid I

  •  
  • AAII

    aristolochic acid II

  •  
  • AAN

    AAI nephropathy

  •  
  • BUN

    blood urea nitrogen

  •  
  • DSS

    dextran sulfate sodium

  •  
  • GI

    gastrointestinal

  •  
  • MLN

    mesenteric lymph node

  •  
  • SI

    small intestine

  •  
  • SILP

    small intestinal lamina propria

  •  
  • Treg

    regulatory T

  •  
  • UUO

    unilateral ureteral obstruction

  •  
  • WT

    wild-type

  •  
  • YPD

    yeast extract peptone dextrose.

1
Nugent
R. A.
,
S. F.
Fathima
,
A. B.
Feigl
,
D.
Chyung
.
2011
.
The burden of chronic kidney disease on developing nations: a 21st century challenge in global health.
Nephron Clin. Pract.
118
:
c269
c277
.
2
Sarnak
M. J.
,
B. L.
Jaber
.
2000
.
Mortality caused by sepsis in patients with end-stage renal disease compared with the general population.
Kidney Int.
58
:
1758
1764
.
3
Dalrymple
L. S.
,
K. L.
Johansen
,
G. M.
Chertow
,
S. C.
Cheng
,
B.
Grimes
,
E. B.
Gold
,
G. A.
Kaysen
.
2010
.
Infection-related hospitalizations in older patients with ESRD.
Am. J. Kidney Dis.
56
:
522
530
.
4
Collins
A. J.
,
R. N.
Foley
,
C.
Herzog
,
B. M.
Chavers
,
D.
Gilbertson
,
A.
Ishani
,
B. L.
Kasiske
,
J.
Liu
,
L.-W.
Mau
,
M.
McBean
, et al
.
2010
.
Excerpts from the US renal data system 2009 annual data report.
Am. J. Kidney Dis.
55
:
S1
S420, A6–A7
.
5
Allon
M.
,
T. A.
Depner
,
M.
Radeva
,
J.
Bailey
,
S.
Beddhu
,
D.
Butterly
,
D. W.
Coyne
,
J. J.
Gassman
,
A. M.
Kaufman
,
G. A.
Kaysen
, et al;
HEMO Study Group
.
2003
.
Impact of dialysis dose and membrane on infection-related hospitalization and death: results of the HEMO Study.
J. Am. Soc. Nephrol.
14
:
1863
1870
.
6
Pastan
S.
,
J. M.
Soucie
,
W. M.
McClellan
.
2002
.
Vascular access and increased risk of death among hemodialysis patients.
Kidney Int.
62
:
620
626
.
7
Vaziri
N. D.
,
M. V.
Pahl
,
A.
Crum
,
K.
Norris
.
2012
.
Effect of uremia on structure and function of immune system.
J. Ren. Nutr.
22
:
149
156
.
8
Syed-Ahmed
M.
,
M.
Narayanan
.
2019
.
Immune dysfunction and risk of infection in chronic kidney disease.
Adv. Chronic Kidney Dis.
26
:
8
15
.
9
Dobre
M.
,
T. W.
Meyer
,
T. H.
Hostetter
.
2013
.
Searching for uremic toxins.
Clin. J. Am. Soc. Nephrol.
8
:
322
327
.
10
Cohen
G.
,
W. H.
Hörl
.
2012
.
Immune dysfunction in uremia—an update.
Toxins (Basel)
4
:
962
990
.
11
Andersen
K.
,
M. S.
Kesper
,
J. A.
Marschner
,
L.
Konrad
,
M.
Ryu
,
S.
Kumar Vr
,
O. P.
Kulkarni
,
S. R.
Mulay
,
S.
Romoli
,
J.
Demleitner
, et al
.
2017
.
Intestinal dysbiosis, barrier dysfunction, and bacterial translocation account for CKD-related systemic inflammation.
J. Am. Soc. Nephrol.
28
:
76
83
.
12
Anders
H. J.
,
K.
Andersen
,
B.
Stecher
.
2013
.
The intestinal microbiota, a leaky gut, and abnormal immunity in kidney disease.
Kidney Int.
83
:
1010
1016
.
13
Amato
K. R.
,
R.
Martinez-Mota
,
N.
Righini
,
M.
Raguet-Schofield
,
F. P.
Corcione
,
E.
Marini
,
G.
Humphrey
,
G.
Gogul
,
J.
Gaffney
,
E.
Lovelace
, et al
.
2016
.
Phylogenetic and ecological factors impact the gut microbiota of two neotropical primate species.
Oecologia
180
:
717
733
.
14
Rooks
M. G.
,
W. S.
Garrett
.
2016
.
Gut microbiota, metabolites and host immunity.
Nat. Rev. Immunol.
16
:
341
352
.
15
Tilg
H.
,
N.
Zmora
,
T. E.
Adolph
,
E.
Elinav
.
2020
.
The intestinal microbiota fuelling metabolic inflammation.
Nat. Rev. Immunol.
20
:
40
54
.
16
Meijers
B.
,
P.
Evenepoel
,
H. J.
Anders
.
2019
.
Intestinal microbiome and fitness in kidney disease.
Nat. Rev. Nephrol.
15
:
531
545
.
17
Ala-Kaila
K.
,
A.
Pasternack
.
1989
.
Gastrointestinal complications in chronic renal failure.
Dig. Dis.
7
:
230
242
.
18
Carrera-Jiménez
D.
,
P.
Miranda-Alatriste
,
X.
Atilano-Carsi
,
R.
Correa-Rotter
,
Á.
Espinosa-Cuevas
.
2018
.
Relationship between nutritional status and gastrointestinal symptoms in geriatric patients with end-stage renal disease on dialysis.
Nutrients
10
:
425
.
19
Ramezani
A.
,
D. S.
Raj
.
2014
.
The gut microbiome, kidney disease, and targeted interventions.
J. Am. Soc. Nephrol.
25
:
657
670
.
20
Lionakis
M. S.
,
I. D.
Iliev
,
T. M.
Hohl
.
2017
.
Immunity against fungi.
JCI Insight
2
: e93156.
21
Gandhi
B. V.
,
M. M.
Bahadur
,
H.
Dodeja
,
V.
Aggrwal
,
A.
Thamba
,
M.
Mali
.
2005
.
Systemic fungal infections in renal diseases.
J. Postgrad. Med.
51
(
Suppl. 1
):
S30
S36
.
22
Limon
J. J.
,
J. H.
Skalski
,
D. M.
Underhill
.
2017
.
Commensal fungi in health and disease.
Cell Host Microbe
22
:
156
165
.
23
Koh
A. Y.
,
J. R.
Köhler
,
K. T.
Coggshall
,
N.
Van Rooijen
,
G. B.
Pier
.
2008
.
Mucosal damage and neutropenia are required for Candida albicans dissemination.
PLoS Pathog.
4
: e35.
24
Kobayashi-Sakamoto
M.
,
R.
Tamai
,
E.
Isogai
,
Y.
Kiyoura
.
2018
.
Gastrointestinal colonisation and systemic spread of Candida albicans in mice treated with antibiotics and prednisolone.
Microb. Pathog.
117
:
191
199
.
25
Fan
D.
,
L. A.
Coughlin
,
M. M.
Neubauer
,
J.
Kim
,
M. S.
Kim
,
X.
Zhan
,
T. R.
Simms-Waldrip
,
Y.
Xie
,
L. V.
Hooper
,
A. Y.
Koh
.
2015
.
Activation of HIF-1α and LL-37 by commensal bacteria inhibits Candida albicans colonization.
Nat. Med.
21
:
808
814
.
26
Jawale
C. V.
,
K.
Ramani
,
D.-D.
Li
,
B. M.
Coleman
,
R. S.
Oberoi
,
S.
Kupul
,
L.
Lin
,
J. V.
Desai
,
G. M.
Delgoffe
,
M. S.
Lionakis
, et al
.
2020
.
Restoring glucose uptake rescues neutrophil dysfunction and protects against systemic fungal infection in mouse models of kidney disease.
Sci. Transl. Med.
12
: eaay5691.
27
Kigerl
K. A.
,
J. C.
Hall
,
L.
Wang
,
X.
Mo
,
Z.
Yu
,
P. G.
Popovich
.
2016
.
Gut dysbiosis impairs recovery after spinal cord injury.
J. Exp. Med.
213
:
2603
2620
.
28
Ramani
K.
,
R. J.
Tan
,
D.
Zhou
,
B. M.
Coleman
,
C. V.
Jawale
,
Y.
Liu
,
P. S.
Biswas
.
2018
.
IL-17 receptor signaling negatively regulates the development of tubulointerstitial fibrosis in the kidney.
Mediators Inflamm.
2018
: 5103672.
29
Baudoux
T. E.
,
A. A.
Pozdzik
,
V. M.
Arlt
,
E. G.
De Prez
,
M. H.
Antoine
,
N.
Quellard
,
J. M.
Goujon
,
J. L.
Nortier
.
2012
.
Probenecid prevents acute tubular necrosis in a mouse model of aristolochic acid nephropathy.
Kidney Int.
82
:
1105
1113
.
30
Belkaid
Y.
,
T. W.
Hand
.
2014
.
Role of the microbiota in immunity and inflammation.
Cell
157
:
121
141
.
31
Bouladoux
N.
,
O. J.
Harrison
,
Y.
Belkaid
.
2017
.
The mouse model of infection with Citrobacter rodentium.
Curr. Protoc. Immunol.
119
:
19.15.1
19.15.25
.
32
Vallance
B. A.
,
W.
Deng
,
K.
Jacobson
,
B. B.
Finlay
.
2003
.
Host susceptibility to the attaching and effacing bacterial pathogen Citrobacter rodentium.
Infect. Immun.
71
:
3443
3453
.
33
Koh
A. Y.
2013
.
Murine models of Candida gastrointestinal colonization and dissemination.
Eukaryot. Cell
12
:
1416
1422
.
34
Shao
T. Y.
,
W. X. G.
Ang
,
T. T.
Jiang
,
F. S.
Huang
,
H.
Andersen
,
J. M.
Kinder
,
G.
Pham
,
A. R.
Burg
,
B.
Ruff
,
T.
Gonzalez
, et al
.
2019
.
Commensal Candida albicans positively calibrates systemic Th17 immunological responses.
Cell Host Microbe
25
:
404
417.e6
.
35
Allert
S.
,
T. M.
Förster
,
C.-M.
Svensson
,
J. P.
Richardson
,
T.
Pawlik
,
B.
Hebecker
,
S.
Rudolphi
,
M.
Juraschitz
,
M.
Schaller
,
M.
Blagojevic
, et al
.
2018
.
Candida albicans-induced epithelial damage mediates translocation through intestinal barriers.
mBio
9
: e00915-18.
36
Hugenholtz
F.
,
W. M.
de Vos
.
2018
.
Mouse models for human intestinal microbiota research: a critical evaluation.
Cell. Mol. Life Sci.
75
:
149
160
.
37
Itoh
Y.
,
A.
Ezawa
,
K.
Kikuchi
,
Y.
Tsuruta
,
T.
Niwa
.
2013
.
Correlation between serum levels of protein-bound uremic toxins in hemodialysis patients measured by LC/MS/MS.
Mass Spectrom. (Tokyo)
2
:
S0017
.
38
Majewska
E.
,
Z.
Baj
,
Z.
Sulowska
,
J.
Rysz
,
M.
Luciak
.
2003
.
Effects of uraemia and haemodialysis on neutrophil apoptosis and expression of apoptosis-related proteins.
Nephrol. Dial. Transplant.
18
:
2582
2588
.
39
Pahl
M. V.
,
S.
Gollapudi
,
L.
Sepassi
,
P.
Gollapudi
,
R.
Elahimehr
,
N. D.
Vaziri
.
2010
.
Effect of end-stage renal disease on B-lymphocyte subpopulations, IL-7, BAFF and BAFF receptor expression.
Nephrol. Dial. Transplant.
25
:
205
212
.
40
Meier
P.
2009
.
FOXP3+ regulatory T-cells in chronic kidney disease: molecular pathways and clinical implications.
Adv. Exp. Med. Biol.
665
:
163
170
.

The authors have no financial conflicts of interest.

This article is distributed under the terms of the CC BY-NC 4.0 Unported license.