Research ArticleKidney Disease

Melamine-Induced Renal Toxicity Is Mediated by the Gut Microbiota

See allHide authors and affiliations

Science Translational Medicine  13 Feb 2013:
Vol. 5, Issue 172, pp. 172ra22
DOI: 10.1126/scitranslmed.3005114

This article has a correction. Please see:

Abstract

Melamine poisoning has become widely publicized after a recent occurrence of renal injury in infants and children exposed to melamine-tainted milk in China. This renal damage is believed to result from kidney stones formed from melamine and uric acid or from melamine and its cocrystallizing chemical derivative, cyanuric acid. However, the composition of the stones and the mechanism by which the stones are formed in the renal tubules are unknown. We report that cyanuric acid can be produced in the gut by microbial transformation of melamine and serves as an integral component of the kidney stones responsible for melamine-induced renal toxicity in rats. Melamine-induced toxicity in rats was attenuated and melamine excretion increased after antibiotic suppression of gut microbial activity. We further demonstrated that melamine is converted to cyanuric acid in vitro by bacteria cultured from normal rat feces; Klebsiella was subsequently identified in fecal samples by 16S ribosomal DNA sequencing. In culture, Klebsiella terrigena was shown to convert melamine to cyanuric acid directly. Rats colonized by K. terrigena showed exacerbated melamine-induced nephrotoxicity. Cyanuric acid was detected in the kidneys of rats administered melamine alone, and the concentration after Klebsiella colonization was increased. These findings suggest that the observed toxicity of melamine may be conditional on the exact composition and metabolic activities of the gut microbiota.

Introduction

Melamine has received intense media attention as the cause of renal failure and death in both animals and children (13) because of its addition to pet food and infant formula as a way of boosting the apparent protein content. The low acute toxicity of melamine alone with an oral median lethal dose (LD50) of 3161 mg/kg in rats (4) and a solubility in water of 3.24 g/liter at room temperature (5) has raised questions regarding the mechanism of melamine-induced toxicity. Previous reports (69) suggest that severe renal toxicity in animals is associated with the combination of melamine and its structural analog, cyanuric acid, which can readily self-assemble into supermolecular aggregates through a hydrogen-bonding mechanism, leading to insoluble melamine-cyanurate cocrystals (10, 11). Recent studies have identified melamine-cyanurate crystals in the kidneys of rats, fish, and pigs that consume melamine alone (12, 13); however, the mechanism remains unknown. In contrast to the findings in animals, however, the calculi in infants suffering from melamine-induced renal toxicity were identified as being mostly melamine and uric acid in composition (4, 1416). On the basis of these findings, we set out to explore the mechanism of melamine-induced renal toxicity. Previous studies in our laboratory have shown that melamine-induced toxicity in Wistar rats is dose-dependent, with marked toxicity and metabolic changes observed at a high dose of melamine (600 mg/kg) similar to that seen in rats treated with a low dose of melamine combined with cyanuric acid [melamine (50 mg/kg) and cyanuric acid (50 mg/kg)] (6). We also found substantial fluctuation in urinary metabolite levels such as lowered phenylacetylglycine and elevated trimethylamine-N-oxide and 3-phenylpropionate, which are related to gut microbial-mammalian cometabolism in rats receiving a high dose of melamine compared to those in the control group (6).

Certain bacterial strains, such as Klebsiella terrigena (strain DRS-1), can produce cyanuric acid from melamine through a series of deamination steps (1719). This aerobic strain may be part of the mammalian gut microbiota (20), which prompted us to hypothesize that melamine is partially converted to cyanuric acid through intestinal microbial transformation and that the high renal toxicity in mammals exposed to melamine is a result of melamine-cyanurate or melamine-cyanurate-urate coprecipitation in renal tubules. Here, we perform pharmacological and metabolomic studies to investigate the role of the gut microbiota, and in particular Klebsiella species, in the conversion of melamine to cyanuric acid.

Results

Gut microbiota suppression in rats attenuates melamine-induced toxicity

We performed animal studies to assess the impact of suppressed or nonsuppressed microbiota activity on melamine-induced toxicity in Wistar rats. Microbial suppression was achieved by oral treatment with a broad-spectrum antibiotic agent, imipenem/cilastatin sodium, at a daily dose of 50 mg/kg for 4 days before melamine exposure. Melamine (Sigma-Aldrich, ≥99% purity, without cyanuric acid contamination) was administered orally to rats at a daily dose of 600 mg/kg for 15 days. The experimental rats were divided into four groups: control group, antibiotic group (AB group), melamine group (Mel group), and antibiotic + melamine group (AB + Mel group). The antibiotic used, imipenem/cilastatin, is non-absorbable and is effective on both Gram-positive and Gram-negative bacteria in the gastrointestinal tract without systemic effects (5). To rule out the possibility of exogenous contamination of melamine and cyanuric acid, we analyzed the water, feed, and antibiotic solutions used for the animal experiment using ultra performance liquid chromatography/triple quadrupole tandem mass spectrometry (UPLC/TQMS). Melamine and cyanuric acid were not detected in these materials.

Kidney histology (Fig. 1, A to F) showed severe damage from melamine administration (600 mg/kg per day), including crystal formation (Fig. 1D), obvious dilatation of renal tubules (Fig. 1E), and renal interstitial hemorrhage (Fig. 1F). These effects were ameliorated in the animals with gut microbial suppression, as shown by less hemorrhage in the renal interstitium (Fig. 1C). No histological abnormalities were observed in the kidney tissues of control (Fig. 1A) or AB rats (Fig. 1B), suggesting that the short-term administration of antibiotic did not produce histological damage in kidney tissues. At necropsy, the kidneys of melamine-treated rats appeared to have a rough surface (fig. S3), and the relative kidney weight of these rats was higher (P < 0.05) than those of the rats in the control group (Fig. 1G). In contrast, the relative kidney weights of the rats in AB and AB + Mel groups were not significantly different from that of the rats in the control group (P > 0.05). Serum urea nitrogen (Fig. 1H) and creatinine (Fig. 1I) concentrations were slightly higher in the Mel group only (P = 0.08). These results show that microbial inhibition can ameliorate melamine-induced toxicity.

Fig. 1

Evaluation of melamine-induced renal toxicity with or without gut microbiota suppression. (A to F) Representative photographs of histological examination of kidneys in (A) control rats, (B) antibiotic-treated rats (AB group), (C) rats treated with antibiotics and melamine (AB + Mel group), and (D to F) rats treated with melamine alone (Mel group). The red arrow indicates crystals, the white arrows indicate hemorrhage, and the asterisks indicate tubular dilatation. Scale bars, 50 μm. (G) Relative kidney weights (expressed as percentage of the total body weight). Mean values ± SD are plotted. (H and I) Serum urea nitrogen and creatinine concentrations. Mean values ± SD are plotted. (J) Time-dependent trajectories of urinary metabolomic profiles. Control, AB, Mel, and AB + Mel groups are shown across the time course from days 0 to 19. Each arrow represents mean values of the scores from the first (x axis) and the second (y axis) principal components at a certain time point of the corresponding group. The dots represent the pre-dose time points, and the diamonds represent the endpoint at 15 days after melamine or vehicle treatment (day 19). (K) Heat map of identified differential metabolites with a urinary metabolomic profile. Each cell in the heat map represents the fold change of a particular metabolite, which is the ratio of the intensity of each sample in the treatment group to the mean value of the control group [(a) Mel group versus control group; (b) AB + Mel group versus control group; and (c) AB + Mel group versus AB group]. Red indicates increased concentrations, and purple indicates decreased concentrations. (L) The box plot of each metabolite represents the ratio of the mean value of peak intensity in the treatment group to that for the control group (Mel versus control group, AB + Mel group versus control group, and AB + Mel group versus AB group from left to right, respectively). The values on the y axis in the plot of trimethylamine oxide, 5-phenylvaleric acid, and benzaldehyde are shown as log plots. n = 6 to 7 per group. *P < 0.05, one-way analysis of variance (ANOVA).

Metabolomic approaches have potential for deciphering the global phenotypic changes resulting from functional alterations of both host metabolism and the gut microbiome (2123). To explain the metabolic fluctuations induced by melamine with or without microbiota suppression at different time points, we constructed urinary metabolic trajectories (Fig. 1J) using a multivariate statistical method [partial least squares discriminant analysis (PLS-DA)] (24, 25). The trajectory represents the time course of metabolic changes in each group measured at different time points. The four trajectories started at the same spatial position in the upper right quadrant, indicating that the metabolic profiles of the four groups were similar on day 0. As shown in Fig. 1J, the trajectory of the Mel group (red) drastically shifted away from that of the control group (gray) after 1 day of melamine treatment and continued to drift away throughout the experiment, demonstrating that the metabolic profile of the Mel group was largely different because of the melamine intervention. Moreover, the AB + Mel group (pink) showed a different time-dependent trajectory, moving closer to that of the control group, suggesting that the melamine-induced metabolic alterations were attenuated. A metabolic heat map (Fig. 1K) illustrated fluctuation of altered urinary metabolites after microbiota suppression compared to unsuppressed microbiota. The metabolites (table S1) were selected in accordance with the criteria of nonparametric univariate statistics (Kruskal-Wallis, P < 0.05) and multivariate statistics [variable importance in the projection value (VIP) > 1 and the absolute value of correlation coefficients (Pcorr) > 0.5]. Differentially expressed metabolites in the Mel group were attenuated in the AB + Mel group, as indicated by the changes in heat map intensities. Among the identified differential metabolites in the Mel group, several have been previously reported as gut microbiota–related metabolites (Fig. 1L), including phenylpyruvic acid, phenylacetylglycine, indoleacetic acid, trimethylamine oxide, 5-phenylvaleric acid, and benzaldehyde (2628). These metabolites returned to control levels after microbiota suppression. The results suggest that the gut microbiota play an important role in melamine-induced metabolic changes.

Gut microbiota suppression increases urinary excretion of melamine in rats

We also observed that with the suppression of the gut microbiota, urinary excretion of melamine was substantially increased, presumably due to the reduced microbial conversion of melamine to cyanuric acid in vivo. We quantitatively analyzed urinary melamine from the rats in the Mel group and the AB + Mel group at 15 days after melamine intervention. Only 11.7 ± 1.8% of the melamine dose (600 mg/kg) was excreted in urine of the rats in the Mel group, whereas 23.8 ± 8.4% of the melamine dose was detected in urine of the rats in the AB + Mel group, representing a twofold increase (P < 0.01) (Fig. 2A). In addition, as melamine administration was increased to 100, 300, and 600 mg/kg in a parallel study performed in our laboratory, renal toxicity increased and the urinary excretion of melamine decreased significantly at a ratio of about 6:2:1 relative to the melamine dose (P < 0.001) (Fig. 2B). The urinary excretion of both melamine and cyanuric acid was the lowest at the low dose of melamine combined with cyanuric acid (11.4% of melamine dose shown in Fig. 2B, and 4.5% of cyanuric acid dose shown in Fig. 2C). The histological findings of renal toxicity in this group were similar to those observed in the Mel group rats. The excretion rates of melamine and cyanuric acid are shown in tables S3 and S4. These results suggest that a substantial portion of the “inert” melamine is somehow converted to other chemical forms in vivo and that gut microbiota suppression inhibits such a conversion.

Fig. 2

The urinary excretion of melamine and cyanuric acid. (A) Urinary excretion of melamine in rats from the high-dose Mel and AB + Mel groups. Mean values ± SD are plotted. (B) Urinary excretion of melamine administered at different doses: low-dose melamine (100 mg/kg), mid-dose melamine (300 mg/kg), high-dose melamine (600 mg/kg), and low-dose melamine combined with cyanuric acid [melamine (50 mg/kg) and cyanuric acid (50 mg/kg)]. Mean values ± SD are plotted. (C) Urinary excretion of cyanuric acid in the low-dose cyanuric acid (100 mg/kg) group and low-dose melamine and cyanuric acid group [melamine (50 mg/kg) and cyanuric acid (50 mg/kg)]. Mean values ± SD are plotted. n = 6 to 7 per group. **P < 0.01, ***P < 0.001, one-way ANOVA.

Gut microbiota can convert melamine to cyanuric acid

We further performed in vitro studies to confirm the existence of intestinal microbes in experimental rats that can convert melamine to cyanuric acid. Fecal specimens collected from Wistar rats were incubated in nutrient broth medium supplemented with melamine (1000 μg/ml) under aerobic conditions. The concentration of melamine in the culture continuously decreased to 635.12 μg/ml after 36 hours of cultivation (Fig. 3A). The concentration of cyanuric acid in the medium after 24 hours of cultivation was 0.19 μg/ml (Fig. 3A). In controls without fecal specimens, we were not able to detect cyanuric acid in the melamine-supplemented medium. These results confirmed the microbial conversion of melamine to cyanuric acid in vitro by bacteria derived from rat feces.

Fig. 3

The conversion of melamine to cyanuric acid. (A) Concentrations of melamine, ammeline, ammelide, cyanuric acid, biuret, and urea in fecal cultures containing different microbes and in cultures containing K. terrigena alone at different time points. Mean values ± SD are plotted. n = 3. (B) Several genera of gut microbes with high abundance in the mammalian gut are listed including Bacteroides, Clostridium, Lactobacillus, Bifidobacterium, Escherichia, and Klebsiella. Seven Klebsiella species were identified and are listed with similarity hits (shown as percentages) after each species name.

We suspected that melamine is degraded by intestinal microbes through a mechanism of nitrogen consumption by environmental aerobic bacteria as previously reported (17, 19, 29). This type of deamination process is highly efficient if melamine is used as the sole nitrogen source, but melamine degradation may be suppressed if there are other nitrogen sources present. To test this hypothesis, we monitored melamine utilization efficiency in high–nitrogen content tryptic soy broth (30) with the same concentration of melamine supplementation. We found that melamine degradation in nitrogen-rich tryptic soy broth was markedly hindered and that cyanuric acid was not detectable throughout 36 hours of cultivation.

Other intermediate products of melamine metabolism, including ammeline, ammelide, and biuret, which are a result of successive deaminations of triazines, were detected in the fecal samples supplemented with melamine (Fig. 3A). Ammelide reached a peak concentration (0.47 ± 0.12 μg/ml) after 8 hours of cultivation, and cyanuric acid and its downstream product biuret were detected after 24 hours of cultivation.

The Klebsiella species are responsible for deamination of melamine

We next sought to determine which aerobic bacteria were responsible for the biotransformation of melamine. Three Pseudomonas isolates as well as one Rhodococcus isolate were previously reported to have a high capacity for triazine conversion (29, 31). Klebsiella seemed to be a strong candidate because the K. terrigena strain has been shown to convert melamine to cyanuric acid (19). Two other strains, Klebsiella pneumonia and Klebsiella planticola DSZ, were also reported to have a capacity for degrading triazines (29, 32). We performed 16S ribosomal DNA (rDNA) sequencing analysis of cultivated rat feces and identified the presence of seven species of Klebsiella genus, including oxytoca, terrigena, pneumoniae, planticola, singaporensis, ornithinolytica, and variicola (Fig. 3B). The species of Klebsiella genus as well as other bacterial species were identified in the cultivated rat feces using the BLASTN program available on the National Center for Biotechnology Information (NCBI) nucleotide sequence database Web site (http://blast.ncbi.nlm.nih.gov) based on 98 to 97% similarity with the hits (Fig. 3B). Among the seven species, three (K. terrigena, K. pneumonia, and K. planticola) were previously reported to have the capability of converting melamine to cyanuric acid (19, 29, 32).

We further cultivated K. terrigena in medium containing melamine (1000 μg/ml). Cyanuric acid was immediately detected within 1 hour of cultivation (Fig. 3A), indicating the strong capacity of K. terrigena for melamine biotransformation. After 10 hours of cultivation, the concentration of melamine decreased by about 14% of its original level in the Klebsiella culture, and cyanuric acid reached a peak concentration of 0.57 ± 0.03 μg/ml, remaining stable for the entire experimental period. Other melamine derivatives, including ammeline, ammelide, biuret, and urea, were all detected within 36 hours of cultivation (Fig. 3A).

Together, these findings show that aerobic bacteria of the Klebsiella genus exist in rodents and that microbial transformation of melamine to cyanuric acid can occur in the mammalian gastrointestinal tract. Intestinal aerobic microbes of the Klebsiella genus, in particular the K. terrigena strain, may be key players in the conversion of melamine.

Colonization of rat guts with Klebsiella exacerbates melamine-induced toxicity

To confirm the impact of Klebsiella species on melamine-induced toxicity in mammals, we colonized the guts of Wistar rats with K. terrigena. K. terrigena bacterial cells were harvested from cultures and were diluted to a final concentration of 109 colony-forming units/ml. Colonization of rat guts was achieved by oral administration of 500 μl of bacterial suspension for four consecutive days. The rats with or without Klebsiella colonization were then orally administered melamine (600 mg/kg per day) for 15 days (Mel group and K + Mel group, respectively). Two control groups (an untreated control group and a group receiving Klebsiella but no melamine) were also used in parallel in the study.

At necropsy, the kidneys of rats receiving melamine with or without Klebsiella colonization were edematous, and the relative kidney weight was significantly higher than that of the rats in the two control groups (P < 0.05) (Fig. 4A). The kidneys of rats in the K + Mel group seemed to have more histological lesions and were pale in appearance compared to those from the other three groups (fig. S4).

Fig. 4

Evaluation of melamine-induced renal toxicity in rats with or without Klebsiella colonization. (A) Relative kidney weights (expressed as percentage of the total body weight). Mean values ± SD are plotted. (B and C) Serum urea nitrogen and creatinine concentrations. Mean values ± SD are plotted. (D to F) Representative photographs of wet-mount analysis of rat kidney tissue analyzed by polarized light microscopy. The crystals observed in the kidneys from rats administered melamine without Klebsiella colonization (Mel group) and rats administered melamine with Klebsiella colonization (K + Mel group) were scattered in the kidneys and were characteristically birefringent. (G to L) Representative photographs of the histological sections of kidneys in control group (G), Klebsiella group (H), Mel group (I), and K + Mel group (J to L). The red arrow (J) indicates crystals, the white arrows (K) indicate hemorrhage, the asterisks indicate tubular dilatation, and the black arrows indicate inflammatory cell infiltrates. Scale bars, 50 μm. n = 6 to 8 per group. *P < 0.05, **P < 0.01, one-way ANOVA.

Kidney tissue slides were also prepared by wet-mount procedure and short-term fixation with formalin followed by hematoxylin and eosin (H&E) staining. Microscopically, wet-mount sections from frozen kidneys revealed the presence of golden green crystalline deposits in the kidneys of all rats in the K + Mel group, whereas the same type of kidney stones was observed only in three rats of the Mel group (table S5). The kidney stones in the Mel and K + Mel groups appeared to be birefringent crystals based on polarized light optical microphotography (Fig. 4, D to F), which is consistent with previous reports (8, 3335). Marked tubular lesions were observed in the H&E-stained sections from the K + Mel group, including crystal formation (Fig. 4J), hemorrhage, dilatation of renal tubules, and inflammatory cell infiltration (Fig. 4, K and L). In the Mel group, inflammatory cell infiltrates were observed (Fig. 4I), but no histological abnormalities were observed in the kidneys of control (Fig. 4G) and Klebsiella alone (Fig. 4H) groups. In addition, biochemical analysis showed that serum urea nitrogen (Fig. 4B) and creatinine (Fig. 4C) concentrations were significantly (P < 0.05) increased in the rats of Mel and K + Mel groups and that these two markers were substantially higher in the K + Mel group than in the Mel alone group. These results confirm aggravated renal toxicity in the rats with Klebsiella colonization.

Colonization of rat guts with Klebsiella increases cyanuric acid concentrations

To determine the composition of kidney stones in rats, we extracted and dissolved the kidney tissues using a mixture of acetonitrile/water/diethylamine at a ratio of 5:4:1 and analyzed the contents by UPLC/TQMS (table S6). Melamine was detected in the rats from the Mel and K + Mel groups at concentrations of 16.58 ± 6.79 nmol/g and 89.68 ± 19.14 nmol/g, respectively (Fig. 5A). In these two groups, cyanuric acid was also detected at concentrations of 0.45 ± 0.14 nmol/g and 2.32 ± 0.73 nmol/g, respectively (Fig. 5B). There was a fivefold increase (P < 0.01) in the concentration of cyanuric acid in the kidneys of rats with Klebsiella colonization.

Fig. 5

Assessment of crystals in the rat kidney. (A) Melamine concentrations in the kidneys of rats given melamine without Klebsiella colonization (Mel group) and rats given melamine with Klebsiella colonization (K + Mel group). (B) Cyanuric acid concentrations in the kidneys of rats from the Mel and K + Mel groups. (C) Uric acid concentrations in the kidneys of rats in the control, Klebsiella, Mel, and K + Mel groups. Mean values ± SD are plotted. n = 6 to 8 per group. **P < 0.01, one-way ANOVA.

In addition to melamine and cyanuric acid, uric acid was also detected in the kidneys from all groups. The concentrations of uric acid in the kidneys of Mel and K + Mel groups were much higher than those in the control and Klebsiella groups (Fig. 5C). We also performed in vitro chemical precipitation assays by mixing the compounds together in the following pairs: melamine–uric acid, melamine–cyanuric acid, uric acid–cyanuric acid, and melamine–uric acid–cyanuric acid. We observed that crystals readily formed (within ~24 hours) when cyanuric acid was present, but melamine–uric acid did not precipitate, even after 96 hours.

Discussion

It is well established that the mammalian gut microbiota interact extensively with the host through metabolic exchange and cometabolism of substrates. Although the mechanisms are poorly understood, they have been suggested to play a role in the etiology of many human diseases as well as adverse drug effects (36). Recent studies have shown that melamine-cyanurate crystals can form in the kidneys of rats, fish, or pigs that consume only melamine (8, 12, 13). Although some reports have speculated that the source of cyanuric acid was contaminated food (37), no cyanuric acid was detected in water, feed, or antibiotic solutions used in our animal studies. Therefore, our results suggest that the cyanuric acid detected in the kidneys of rats was derived from the gut microbial conversion of melamine. Because the concentration of cyanuric acid was much lower than that of melamine in the kidneys of melamine-dosed rats, we suspect that cyanuric acid may serve as a nidus for crystal formation by melamine-cyanurate and melamine-urate, which constitute the chemical composition of kidney stones reported in renal tubules of mammals exposed to melamine.

The concentrations of uric acid in the kidneys of melamine-dosed animals were much higher than those in the animals of control groups, suggesting that uric acid is an important component of melamine-induced kidney stones. Melamine and cyanuric acid are able to form self-assembling complexes through organized intramolecular networks of hydrogen bonds and pi-pi aromatic ring stacking (38). Uric acid has imide groups known to interact with melamine in self-associating complexes (39). Minimum solubility for the complex of melamine-cyanurate was observed at a pH of 4.5 to 7, whereas melamine-urate exhibited tighter binding under acidic conditions (pH 4) (38). At neutral pH, melamine tends to bind to cyanuric acid with an affinity that is 29-fold greater than that of melamine-urate (38). At the pH of urine, which is about 6.0, melamine should combine with cyanuric acid more easily than uric acid. Together, these results suggest that melamine-induced renal toxicity occurs as a result of a small amount of cyanuric acid being produced by the gut microbial conversion of melamine, which leads to the development of melamine-cyanurate crystals in the kidneys. The crystal sedimentation of melamine-cyanurate facilitates further conglomeration and crystallization of melamine-urate in renal tubules, thereby causing acute or chronic kidney failure.

These data from animal studies are relevant to melamine-induced nephrotoxicity in humans. K. terrigena was previously detected in the stool specimens of 0.9% of 5377 healthy subjects (40). A recent study of children in rural China who consumed melamine-contaminated dairy products reported that the overall prevalence of urinary tract abnormalities among 7933 exposed children was 0.61% (3). These clinical data suggest that the incidence of melamine-induced toxicity in humans is similar to the incidence of K. terrigena colonization in humans, suggesting that the population that is susceptible to melamine adulteration may correlate with high levels of gut Klebsiella, although this correlation has not yet been formally established. A direct linkage to melamine-induced renal toxicity would require genotyping of the gut microbiota present in the fecal samples of infants suffering from adulterated milk powder–induced nephrotoxicity. The genotyping would reveal whether infants with nephrotoxicity had a higher abundance of the Klebsiella species. However, this type of study would be very challenging because 4 years has passed since the major incident of melamine poisoning in Chinese infants, which occurred in late 2008. Our study demonstrates that gut microbial activities can affect the metabolism and toxicity of food contaminants and pollutants and, therefore, should be taken into consideration in measuring the impact of human environmental exposure events.

Materials and Methods

Animal study 1: Melamine-induced toxicity in rats with gut microbiota suppression

The protocols of gut microbiota suppression and induction of melamine toxicity were based on our previous reports (6, 26). After 1 week of acclimation, a total of 26 rats were randomly divided into four groups: control group (n = 6), which received the same volume of water as the other three groups from days 1 to 4 and then 1% CMC-Na (vehicle of melamine) from days 5 to 19; antibiotic group (AB group) (n = 6), which received antibiotic solution at an oral dose of 50 mg/kg per day from days 1 to 4 and then 1% CMC-Na from days 5 to 19; antibiotic + melamine group (AB + Mel group) (n = 7), which received antibiotic solution at a daily dose of 50 mg/kg from days 1 to 4 and then melamine (600 mg/kg) from days 5 to 19 [this was equivalent to the high-dose melamine group in our previous study (6)]; melamine group (Mel group) (n = 7), which received water from days 1 to 4 and then melamine at a daily dose of 600 mg/kg from days 5 to 19. Animals were sacrificed on the 21st day.

Biochemical assays

Blood was collected from the ocular orbit of rats before sacrifice. The blood was clotted at room temperature for 30 min, and serum was separated by centrifugation at 3000g for 10 min for biochemical assessment. Serum biochemical markers, including urea nitrogen and creatinine, were measured with a Hitachi 7600 Automatic Analyzer (Hitachi Co.).

Wet-mount analysis

At necropsy, the left and right kidneys from each rat were weighed and then sectioned longitudinally. One half of each kidney was placed in 10% neutral buffered formalin for 12 hours. The other half of kidney was flash-frozen. The left kidney was used for wet-mount preparation, and the right kidney was used for crystal analysis.

The wet-mount analysis was conducted by pressing about 2-mm-thick sections of fresh tissue between two microscope slides and then observing the tissue under bright field and polarized light microscopy (Leica DM LP, Hitachi Co.). On the basis of a method described in a previous report (35), the presence of crystals was scored from 0 to 5 as shown in table S5.

Histological analysis

Half of one kidney was fixed in formalin and embedded in paraffin wax. Histological sections (3 μm thick) of the paraffin-embedded tissue were stained with H&E and prepared for light microscopy using Leica DMRE Microsystems equipment with a SPOT Flex Microscope Digital Camera (Diagnostic Instruments Inc.).

Metabolomic studies

The metabolomic study was performed following our previously published protocol (6). Details regarding sample preparation, instrumental analysis, and compound annotation can be obtained in the Supplementary Materials.

Quantitative analysis of urinary melamine and cyanuric acid excretion

Melamine and cyanuric acid in each 300 μl of urine sample were prepared by adding 300 μl of mix solvent (water/diethylamine at a ratio of 9:1). The preparation procedures used were the same as those described for the metabolomic study. Melamine and cyanuric acid were analyzed by UPLC/quadrupole time-of-flight MS (UPLC/QTOFMS) with optimized chromatography and MS conditions (table S2). Melamine and cyanuric acid stock standard solutions were prepared at a concentration of 1 mg/ml for quantitation. To eliminate the sample matrix effect, we diluted the stock solutions with control urine samples to concentration series of 0.1, 0.5, 1, 5, 10, 25, 50, 100, or 500 μg/ml, and then we prepared each standard solution using the same procedures as the sample preparation before analysis. The calibration curve was constructed by plotting the peak area of each standard versus concentration. Melamine and cyanuric acid excretion percentages were the portions of the administered amounts that appeared in the urine.

Biotransformation of melamine to cyanuric acid

Feces samples were collected from 6-week-old male Wistar rats. Nutrient broth was used as culture medium in this study. Fresh feces pellets (about 5 g) were placed into 100 ml of sterile medium. Fecal microbes were propagated in a shaking incubator (Labnet 311DS) at a speed of 120 rpm at 37°C for 36 hours until the medium was obviously turbid. Each 500 μl of the propagated fecal culture was then transferred to 4.5 ml of sterile nutrient broth supplemented with melamine (1 mg/ml) for cultivation. During the cultivation, samples were taken out at 0, 1, 2, 3, 4, 6, 8, 12, 24, and 36 hours and centrifuged at 12000g for 20 min. The collected supernatants were stored at −80°C until analysis. For analysis, an aliquot of 500 μl supernatant was mixed with 500 μl of mix solution (methanol/water/diethylamine at a ratio of 5:4:1) and then centrifuged at 12000g for 10 min. The supernatant was filtered through a syringe filter (0.22 μm) for UPLC/QTOFMS analysis. The standard solutions of melamine, cyanuric acid, ammeline, ammelide, biuret, and urea for quantitation were dissolved in nutrient broth and prepared with the same procedures as those for sample preparation before analysis. The calibration curves were constructed with the same procedure described above.

16S rDNA sequence analysis of Klebsiella in gut microbiota

The Klebsiella sp. strain was previously reported to have the capability of converting melamine to cyanuric acid (19). To identify whether Klebsiella strains exist in mammalian gut microbiota, we collected fresh feces from 6-week-old Wistar male rats. The feces was propagated in nutrient broth for 24 hours at a rotation speed of 120 rpm to facilitate bacterial growth. The procedures of DNA extraction, primer design and synthesis, polymerase chain reaction, and sequence analysis can be obtained in the Supplementary Materials. K. terrigena [American Type Culture Collection (ATCC) 700372] was used as a positive control, and Escherichia coli was used as a negative control.

Melamine biotransformation to cyanuric acid by a Klebsiella sp. strain

K. terrigena was used to confirm the Klebsiella strain–mediated biotransformation because this strain has been shown to be one of the most important of the genus (19). Freeze-dried K. terrigena powder was propagated following instructions from ATCC in tryptic soy broth and a shaking incubator at 26°C for 36 hours. Each 500 μl of K. terrigena liquid culture was inoculated into 4.5 ml of tryptic soy broth supplemented with melamine (1 mg/ml) or without melamine (negative control). The standard solutions of melamine, cyanuric acid, ammeline, ammelide, biuret, and urea for quantitation were dissolved in tryptic soy broth. The calibration curve was constructed with the same procedure described above.

Animal study 2: Melamine-induced toxicity in rats colonized with Klebsiella

Klebsiella colonization was performed according to a previously described method (41). The preparation of Klebsiella bacterial cell suspension can be obtained in the Supplementary Materials. After 1 week of accommodation, a total of 28 rats were randomly divided into four groups: control group (n = 6), which received 0.5 ml of saline as the other three groups from days 1 to 4 and 1% CMC-Na (vehicle of melamine) from days 5 to 19; Klebsiella group (n = 6), which received 0.5 ml of Klebsiella bacterial cell suspension from days 1 to 4 and 1% CMC-Na from days 5 to 19; melamine group (Mel group) (n = 8), which received 0.5 ml of saline from days 1 to 4 and then administered with melamine at a daily dose of 600 mg/kg from days 5 to 19; Klebsiella and melamine group (K + Mel group) (n = 8), which was dosed with 0.5 ml of Klebsiella bacterial cell suspension from days 1 to 4 and then with melamine (600 mg/kg) from days 5 to 19 (an equivalent dose of the high-dose melamine group). Animals were sacrificed on the 21st day.

Crystal composition assessment by UPLC/TQMS

The extraction method was conducted based on a previously described method with minor improvements (42). Half of one kidney, including the capsule, cortex, and medulla, was weighted into a 15-ml tube. Each 100-mg tissue was homogenized with 1 ml of acetonitrile/water/diethylamine (5:4:1) for 5 min with a BioSpec Mini-Beadbeater and centrifuged at 3000g for 20 min. A 2-ml aliquot of extraction was transferred into a new 15-ml tube, 4 ml of acetonitrile was added, and the sample was vortexed briefly. The tube was then centrifuged for 20 min at 4000g, and a 2-ml aliquot of supernatant was filtered through a 0.8-μm syringe filter. The supernatant was then evaporated dry under nitrogen with a nitrogen evaporator (Organomation Associates Inc.). The dried extract was reconstituted in 400 μl of water/acetonitrile at a ratio of 4:1 for UPLC/TQMS analysis. The settings of UPLC/TQMS can be obtained in the Supplementary Materials and table S6.

Standard solutions were diluted with water/acetonitrile (4:1) in seven concentrations. To eliminate the sample matrix effect, we added a 10-μl standard solution to 390 μl of control sample. These standard solutions were used to generate calibration curves for quantitation. Sample assessment was performed with a Waters Acquity UPLC system coupled to an AB Sciex Triple Quad 5500 mass spectrometer with Analyst 1.5.1 software. The optimized instrumental parameters are listed in the Supplementary Materials.

Statistical analysis

All data are shown as means ± SD. The statistical significance was evaluated by one-way ANOVA with Bonferroni posttest. Metabolomic data were analyzed with Simca-P+ 12.0 software package (Umetrics). PLS-DA was used to determine metabolomic patterns (fig. S1), and a perturbation trajectory was then constructed by plotting score parameters (Fig. 1J). The differential metabolites selected should meet the following three requirements: nonparametric Kruskal-Wallis test (P < 0.05), multivariate statistical analysis–orthogonal projections to latent structures discriminant analysis (OPLS-DA) (Pcorr > 0.5), and OPLS-DA VIP > 1. The fold change was the ratio of mean intensity of each metabolite in the treatment group to that in the control group. Heat maps were created in MATLAB 7.1 (The MathWorks Inc.). The DNA sequences were analyzed with the BLASTN program from the NCBI Web site.

Supplementary Materials

www.sciencetranslationalmedicine.org/cgi/content/full/5/172/172ra22/DC1

Materials and Methods

Fig. S1. PLS-DA scores plot derived from urinary metabolites.

Fig. S2. Daily physiological records.

Fig. S3. The gross appearance of kidneys of rats in control, AB, AB + Mel, and Mel groups.

Fig. S4. The gross appearance of kidneys of rats in control, Klebsiella, Mel, and K + Mel groups.

Table S1. List of urinary differential metabolites relevant to melamine-induced toxicity.

Table S2. Optimized UPLC/QTOFMS conditions for melamine and cyanuric acid quantitation.

Table S3. The absolute excretion rate of melamine.

Table S4. The absolute excretion rate of cyanuric acid.

Table S5. The crystal intensities in rat kidneys.

Table S6. Optimized UPLC/TQMS conditions for melamine, cyanuric acid, and uric acid quantitation.

References and Notes

  1. Funding: This study was supported by the National Basic Research Program of China (2007CB914700) and the National Natural Science Foundation of China Program (81170760). Author contributions: Wei Jia and A.Z. designed the study. X.Z., A.Z., Y.C., L.Z., C.W., and Y.B. performed the animal, in vitro bacterial, and metabolomic studies. X.Z. and H.L. performed the gene expression analysis. X.Z. and M.S. analyzed the data. G.X., H.L., Weiping Jia, and M.L. helped prepare the manuscript. X.Z., A.Z., and Wei Jia wrote the manuscript. Wei Jia and J.K.N. contributed to the overall metabolomic design. Competing interests: The authors declare that they have no competing interests. Data and materials availability: 16S rDNA sequence data have been deposited in GenBank under accession numbers KC287217 (cultivated rat gut microbiome) and KC287218 (K. terrigena DRS-1).
View Abstract

Stay Connected to Science Translational Medicine

Navigate This Article