Evaluation of Subchronic Toxicity of Pet Food Contaminated with Melamine and Cyanuric Acid in Rats

Abstract

Outbreaks of food-associated renal failure in pets occurred in Asia and the United States of America in 2004 and 2007. They were related to the combined intoxication of cyanuric acid and melamine. Our aims were to investigate cyanuric acid and melamine contamination of pet food and to examine subchronic toxicity in rats. Levels of 10%, 20%, 50%, and 50%–100% (w/w) of contaminated pet food were fed to rats for three months. Analytical results revealed that the tainted food contained significant levels of cyanuric acid and melamine in a ratio of 1:6.8. Rats fed the diet of 50%–100% for three months exhibited elevated serum blood urea nitrogen and creatinine, as well as dose-dependent melamine/cyanuric acid crystal-induced nephrotoxicity. The melamine/cyanuric acid crystals of various sizes were mixed with necrotic cell debris and inflammatory cells, accompanied by tubular dilation and interstitial fibrosis. The immunohistochemistry index of proliferative cellular nuclear antigen and osteopontin in the kidney of the 50%–100% group were elevated, indicating regeneration of renal cells and the formation of crystals. In conclusion, the combination ratio of cyanuric acid to melamine and the acidic urine content were two factors that, upon repeated exposure, determined the severity of the nephrotoxicity.

Introduction

Melamine (2,4,6-triamino-1,3,5-triazine) and cyanuric acid (s-triazine-2,4,6-triol) contamination were found to be the underlying cause for two outbreaks of food-associated renal failure in pets in 2004 and 2007. Melamine is an industrial chemical used in the production of fertilizers and plastics. The acute oral LD50s of melamine are 3200 and 3800 mg/kg in male and female rats, and it is considered to be relatively nontoxic (IARC 1986). Only limited data on the toxicity of cyanuric acid to mammals exist; for instance, sodium cyanurate was fed subchronically to rats and mice at up to 700 and 2200 mg/kg, respectively (Hammond et al. 1986). These chemicals are not normally used in food processing or as an ingredient in powdered milk or feeds, but these products were adulterated with melamine as a nonprotein nitrogen resource to increase the apparent protein levels during chemical analysis (Ingelfinger 2008).

In early 2004, an outbreak of pet food–associated renal failure in dogs and cats occurred in Taiwan and other Asian countries. Blood urea nitrogen, creatinine, and phosphorus levels were severely increased, but sodium and chloride levels were decreased. The aspartate aminotransferase (AST) and alanine aminotransferase (ALT) levels were normal or mildly increased, respectively. Data of serum chemistry indicated severe renal failure and mild hepatic damage. All the clinical signs were first noted for about one month after feeding the dogs with commercial diets (Jeong et al. 2006). The most obvious symptoms were renal malfunction, including lethargy, reduced appetite, weight loss, poor coat condition, and bad breath. In 2004, the Chinese Society of Veterinary Pathology (CSVP) presented the main pathological findings for those pets suspected of being poisoned, including acute to chronic interstitial nephritis and fibrosis and tubular dilation with variously sized green to brown irregular crystals in the kidneys. The Council of Agriculture in Taiwan ordered an emergency recall of the suspected pet food, and a surveillance committee was constituted to investigate.

Epidemiological studies and an analysis of heavy metals, mycotoxins, and feed compositions failed to identify any prominent components that were clearly associated with renal failure. In early 2007, a second outbreak of pet food–associated nephrotoxicosis occurred in North America (Brown et al. 2007; Burns 2007; Puschner et al. 2007; Thompson et al. 2008). More recently, baby milk powder products and cream materials that were contaminated by melamine were suspected of causing kidney stones in babies in China (Lam et al. 2009), and these problems spread rapidly to other parts of the world, causing thousands of milk-associated products to be withdrawn by manufacturers. The World Health Organization is currently assessing thoroughly the effect of the combination of melamine and cyanuric acid (WHO 2008).

The animals that were affected in the more recent North American outbreak died or were euthanized because of severe uremia. Unique polarizable crystals with striations were present in distal tubules or collecting ducts in all examined animals. A chronic pattern of histological changes, characterized by interstitial fibrosis and inflammation, were observed in some sick animals. One notable discovery was that melamine and cyanuric acid were both present in renal tissues in both outbreaks (Brown et al. 2007). Cyanuric acid and melamine in the offending food have since been studied, and the results demonstrated that the combination of melamine and cyanuric acid cause acute renal failure that is similar to previously observed pet food toxicosis in cats (Puschner et al. 2007).

Numerous commercial pet food companies have recalled potentially contaminated products (Weise and Schmit 2007). Cianciolo et al. (2008) investigated a total of seventy cats that were exposed to dietary melamine and cyanuric acid in commercially prepared cans or pouches of wet cat food; the most consistent clinical and pathologic abnormalities in cats fed with contaminated pet food were associated with the urinary tract, and specifically tubular necrosis and crystalluria. Analyses identifying melamine, cyanuric acid, and ammelide in deparaffinized, formalin-fixed dog kidney samples have been performed in Korea (Yhee et al. 2009). Little information on the pathological findings of animals sickened during the 2004 outbreak in Taiwan has been reported (Thompson et al. 2008). The crystals were identified as smooth and platelike; staining characteristics and IR spectroscopy and SEM/EDXA results revealed calcium oxalate crystals.

Although several investigations have shown that the outbreak of pet food–associated nephrotoxicosis was caused by the combination of melamine and cyanuric acid, a comprehensive toxicological evaluation of the contaminated pet food in experimental rats has not been conducted, mainly because most of the suspected batches of feed were recalled and destroyed, making such a study difficult. Therefore, to elucidate the pathogenesis associated with the renal toxicity of the suspected contaminated pet food, this study evaluates the renal failure caused by the pet food in experimental animals (rats) and the toxicological changes in animals fed with the lot of contaminated pet food that caused the 2004 outbreak.

Materials and Methods

Mycotoxin and Chemical Analysis

The commercial dog food was L107 (lot number 107, March 9, 2003, Thailand), which was suspected of being contaminated. A lot produced in Australia after the outbreak (PB03, lot number AP72240-55407-72240/A/03, 15-kg package, beef, Master Foods, Australia) was used as a blank control pet food. The suspected pet food products were stored at 4^°C after they were collected from the pet owners during the outbreak. The residues of citrinin and ochratoxin in feed were entrusted to the Taiwan Animal Technology Institute (Maiole, Taiwan) and the Food Industry Research and Development Institution (Hsinchu, Taiwan), respectively, and were analyzed by high-performance liquid chromatography (HPLC). Standards of melamine (2,4,6-triamino-1,3,5-triazine, 99% purity Sigma-Aldrich, St. Louis, MO, USA), cyanuric acid (s-triazine-2,4,6-triol, Acros, 98% purity, Acros, Merck, Darmstadt, Germany), and their extracts from pet food were quantified using the reverse phase–HPLC-UV method. The separation was carried out using a Phenomenex Gemini-NX C18 column (4.6 × 250 mm, 5 μm), using a pump system (L-2130, Hitachi, Japan) that was interfaced with an autosampler (L-2200), a column oven (L-2300), and a UV detector (L-2400). The UV detection wavelength was set to 220 nm (Ehling et al. 2007). The mobile phase was 1% acetonitrile (Ah2301 72ec, Echo, HPLC grade) and 99% 0.1 M sodium phosphate buffer solution, at pH 8.0 and a flow rate of 0.5 mL/min. The standard calibration curves of cyanuric acid ranged from 2 to 200 ppm and those of melamine ranged from 2 to 100 ppm. Each 0.1 g of pet food was extracted using 10 mL of 0.1 M sodium hydroxide solution, and then centrifuged at 1,788 × g for ten minutes before the supernatant was filtered through a 0.45-mm membrane (Millipore, Millex-HN) for HPLC injection.

Animals

Sprague-Dawley rats (four weeks old) were purchased from the National Laboratory Animal Center (Taipei, Taiwan) and kept in a temperature-controlled (20–22 °C) room with 12 h of light daily. Purina Rat Chow (Ralston Purina Co, St. Louis, MO, USA) and reverse-osmosis water were available ad libitum. This study was approved by the Institutional Animal Care and Use Committee of National Chung-Hsing University (IACUC: 94–63). All animals were cared for following the guidelines in the Guidebook for the Care and Use of Laboratory Animals (Yu et al. 2005).

Experimental Design

Rats were divided into control (Group 1, fed with Purina Rat Chow diet) and suspected contaminated pet food (L107) groups (Table 1), with five males and five females each in a group. According to the manufacturer’s labeling, the crude protein was no less than 23% and the fiber contents no more than 6% in the Purina Rat Chow. The L107 pet food was mixed well with Purina Rat Chow at three dietary levels (10%, 20%, and 50%, w/w, as Groups 2, 3, and 4) using a stainless mixing machine, and the rats were fed with this mixture for three months. To mimic the consumption of the contaminated food in pets, an additional group (Group 5) was fed a 50%–100% diet for eight weeks and then fed a 100% L107 diet for the following four weeks. Rats were observed for clinical signs of toxicity. Body weight and water and food consumption were recorded weekly. All surviving rats were sacrificed at the end of the twelve weeks.

Urinalysis

Urine was collected over approximately eighteen hours using metabolic cages before the animals were sacrificed and necropsy was performed. Urinalysis was performed on fresh urine samples from all surviving animals, and the following parameters were determined: specific gravity, bilirubin, urobilirubin, protein, pH, glucose, ketone, nitrite, occult blood, and leukocyte contents. Total volume and color were recorded, and the uric creatinine was identified using a urine chemistry analyzer (Clinitex 100, Miles Inc, IN., USA). Urinary sediments including red blood cells, white blood cells, epithelial cells, casts, and crystals were also counted under an optical microscope (BX50, Olympus Corporation, Tokyo, Japan).

Hematological and Biochemical Examination

Rats were fasted overnight and anesthetized with 2% isoflurane. Blood was removed from the abdominal aorta and transferred to EDTA-containing anticouagulative tubes (K3 EDTA syringes, Vacutainer, NJ, USA). Complete blood counts were determined using an automated hematology analyzer (Sysmex K-4500, Toa Medical Electronics Co.). The differential leukocyte count was obtained by blood smear and staining using Weigert’s iron hematoxylin stain kit (A.J.P. Scientific Inc, NJ, USA). Sera were analyzed by enzymatic methods using an automatic analyzer (Chiron Diagnostics Corporation, Oberlin, OH, USA). Albumin, alkaline phosphatase, alanine aminotransferase (ALT), amylase, aspartate aminotransferase (AST), γ-glutamyl transpeptidase (GGT), creatine kinase (CK), lactate dehydrogenase (LDH), creatinine, blood urea nitrogen (BUN), total protein, total bilirubin, glucose, total cholesterol, triglyceride, and uric acid were determined. Serum chloride, phosphate, calcium, and magnesium concentrations were measured using an AVL Model 9130 automatic electrolyte analyzer (AVL Scientific Corporation, Georgia, USA).

Pathological Examination

All surviving animals were subjected to complete necropsy. All organs were observed macroscopically, and selected organs (brain, heart, kidney, liver, spleen, testes/ovary, and thymus) were excised and weighed. Organs were fixed in 10% buffered formaldehyde solution for one week, routinely processed, and embedded in paraffin wax. Two-micrometer sections were stained with hematoxylin and eosin. For semiquantitative grading, kidney was stained histochemically with Masson’s trichrome to evaluate the degree of interstitial fibrosis. The severity of the lesions was graded using the method of Shackelford et al. (2002), with five grades of severity of lesions specified: 1 = minimal (<1%); 2 = slight (1%–25%); 3 = moderate (26%–50%); 4 = moderate/severe (51%–75%); and 5 = severe (76%–100%). Immunohistochemical examination included staining the kidneys with antibodies against proliferative cellular nuclear antigen (PCNA; PC10, sc-56, Santa Cruz Biotechnology, Inc., 1:800 dilution) to assess renal tubular cell proliferation, as well as staining with osteopontin (OPN; ab36125, Abcam, 1:500 dilution) to locate the crystals in the tubules of kidneys (Khan 2004). The number of PCNA-positive cells was calculated using forty fields in the cortex and twenty fields in the medulla under 200× magnification. Levels of OPN expression in rat kidneys were evaluated using the following scores: 0 = absent, 1 = <5% tubules affected, 2 = 5%–10% tubules affected, and 3 = >10% tubules affected in each kidney section.

Statistical Analysis

Data listed in tables and figures were evaluated statistically and are expressed as mean ± SD. One-way analysis of variance and Duncan’s least-significant difference tests (Statisca, StatSoft, Inc.) were used to compare the parameters among the various groups. The difference between the control and treated groups was regarded as statistically significant when p < .05.

Results

Mycotoxins, Cyanuric Acid, and Melamine in Suspected Pet Food

Neither citrinin nor ochratoxin was detected in the contaminated (lot L107) or control pet food (lot PB03; data not shown). The concentrations of cyanuric acid and melamine in lot L107 were 909 and 6191 mg/kg, respectively. Neither cyanuric acid nor melamine peaks were obtained from lot PB03.

Clinical Observation, Body Weight, and Water and Food Consumption

During the experimental period, one female and one male rat in the 50% group were found dead in weeks 5 and 9, respectively. These rats had exhibited lethargy, anorexia, and loss of body weight before death, and the bedding was wet. At necropsy, autolysis and multiple brown spots were noted on the surface of the kidneys (data not shown). Rats fed with the L107 diet at levels of Group 1, 2, and 3 for twelve weeks did not vary in water consumption, regardless of sex (data not shown). Rats in the Group 5 for eight weeks exhibited no significant change in body weight or food consumption until the diet was switched to 100% L107. A decline in the body weight of male rats was noted in weeks 9–11, and a significant decrease in body weight was not evident until week 12 (control 515.3 ± 53.0 g, treated 431.2 ± 48.3 g body weight; p < .05; Figure 1A). Female rats showed less body weight loss than male rats. A significant decrease in food consumption of treated male rats was observed from weeks 10 to 12 (p < .05). No significant effect was observed in female rats that were fed the L107 diet below 50% (Group 4) for twelve weeks, except for a small change in food consumption in week 12 (control 22.0 ± 1.3 g, treated 17.0 ± 2.3 g food consumption per day; Figure 1B).

As presented in Table 2, the daily intake of cyanuric acid in Groups 2, 3, and 4 ranged from 6.0 to 29.4 mg/kg body weight in male rats and 7.2 to 35.2 mg/kg body weight in female rats. For rats fed with 100% L107 for four weeks, the daily intakes of cyanuric acid for males and females were 38.3 and 60.3 mg/kg body weight. The daily intake of melamine in Groups 2, 3, and 4 also depended on dose, ranging from 40.6 to 200.4 mg/kg body weight for male rats and 48.9 to 240.0 mg/kg body weight for female rats. For rats fed with 100% L107 (Group 5) for four weeks, the daily intake of melamine by male and female rats was 260.9–410.7 mg/kg body weight. The daily intake of cyanuric acid and melamine in male rats was less than that in female rats, especially in the high-dose group. No significant difference was noted in cyanuric acid and melamine intake between male and female rats at each dosage. The daily intake ratio of cyanuric acid and melamine was about 1:6.8.

Hematology and Serum Biochemistry

Rats fed with 10%, 20%, and 50% L107 diets experienced only small changes in hematological parameters; however, significantly elevated white blood cell counts, mainly in segmented neutrophils in the white blood cell differentiation (p < .05), were detected in both male and female rats in Group 5. Mean cell volume and mean cell hemoglobin values significantly declined in Groups 4 and 5 (data not shown). No significant change occurred in the levels of aspartate aminotransferase, alanine aminotransferase, or other examined enzymatic parameters (data not shown), except for significant increases in blood urea nitrogen and creatinine levels (ten-and eight-fold, respectively) in male rats. Moreover, a significant decline in albumin level and increases in total cholesterol, triglycerides, GGT, creatine kinase, and phosphorus levels were noted. All of these occurred in Group 5, whereas only a few of these changes were noted in Group 4. Most importantly, neither blood urea nitrogen nor creatinine level changed at this second highest dosage (Table 3).

Changes in Urinary and Sediment Parameters

Although no significant increase in daily water consumption was observed (data not shown), the urine volume in Group 5 male and female rats was higher than control volumes (p < .05). Urinary ketone, pH, protein, urobilirubin, and uric creatinine levels were clearly lower in Group 5 (p < .05; Table 4). Many aggregated green to brownish and mostly round, plate-like crystals with radiating striations, identified as MC crystals, were clearly observed (Figure 2).

Organ Weight and Pathological Findings

The kidney weight in Group 5 was significantly elevated in both male and female rats (two to three times that of the control group). An increase in brain weight and a decrease in thymus weight of male rats were also noted in Group 5. Decreased spleen weight in the male rats of Group 2 and reduced testes weight in the male rats in Group 3 were also noted (Table 5).

At necropsy, gross changes of enlarged or shrunken kidneys, pale or brown kidneys with hemorrhagic plaques, irregularly shaped kidneys, and kidneys with rough borders were variously found in Group 4 and 5. On the middle section of the kidneys, renal pelvises were dilated and variable fine birefringent MC crystals were located in the cortex and medulla (Figure 2B). The birefringence or basophilic MC crystals in the renal tubules were easily observed under a polarized microscope (Figure 3A). Numerous MC crystals of various sizes were mixed with the necrotic cell debris in both proximal and distal renal tubules. Slight to severe inflammatory cell infiltration was accompanied by renal tubular dilation and epithelial cell regeneration with interstitial fibrosis (Figure 3B). We found no significant change of glomeruli but noted slight dilation of the glomerular space. Additionally, slight to moderate focal interstitial fibrosis and MC crystals were defined by Masson’s trichrome staining (data not shown).

The PCNA and OPN index in the cortex and medulla tubules in Group 4 increased slightly, and PCNA and OPN (Figure 3C and 3D) indices in Group 5 were substantially elevated = <5% tubules affected, 2 = 5%–10% tubules affected, and 3 = >10% tubules affected in each kidney section. *Significant difference from the control group at p < .05. (Figure 4), indicating regeneration of renal epithelial cells and the formation of crystals. The severity of histopathological change in the kidney in Groups 4 and 5 exceeded those of the lower dietary level groups (Table 6), and no significant lesions were observed in male and female rats in Groups 2 and 3.

Discussion

We conducted a toxicological evaluation of pet food–associated renal failure in rats that had consumed the contaminated pet food product (lot L107) collected during the 2004 outbreak (CSVP 2004). The renal toxicity observed in this study was related to the combination of melamine and cyanuric acid. The nephrotoxicities were characterized by chemical analysis, crystal morphology, histopathological features, and dose response, and our findings revealed that the kidney was a selective target organ of toxicity.

In a clinical study, azotemia and hyperphosphatemia were associated the most with tubular injury by increasing blood urea nitrogen, creatinine, and phosphorous in pets (Cianciolo et al. 2008). In this study, we also found elevation of these parameters in the serum of rats fed cyanuric acid and melamine–contaminated feed. Our results suggest that renal cell necrosis and intrarenal obstruction by MC crystals are possible reasons for the development of acute nephropathy.

γ-Glutamyl transpeptidase (GGT) is an enzyme that catalyzes the transfer of the γ-glutamyl group from glutathione to accepter amino acids. The kidney has the highest GGT activity, with expression in the proximal tubules. Measurement of urinary and serum GGT levels is a means by which proximal tubular disease can be diagnosed during its developmental stages (Melo et al. 2006). In this study, rats had high GGT in serum, indicating that renal injury by melamine and cyanuric acid was located not only in the distal tubular cells, but also in the proximal tubular cells.

Although there were no effects on AST, ALT and hepatocyte morphology, significantly higher total cholesterol and tri-glycerol were noted in serum, implying that feed contaminated with cyanuric acid and melamine might interfere with lipid metabolism (Jeong et al. 2006; Lee et al. 2007). Creatine kinas, LDH, and their isoenzymes in serum have been proven to leak from muscles owing to an increase in the permeability of cell membranes associated with myotoxicity (Gupta et al. 2002). An increase of creatine kinase activity was noted in Group 5 (51.2 ± 17.7 vs. 145.2 ± 52.1 U/L; Table 2), whereas no significant change was found in LDH activity (144.6 ± 69.4 vs. 370.6 ± 201.8 U/L; data not shown). Although we did not find significant histopathological lesions related to myotoxicity or cardiomyopathy, more detailed studies on cyanuric acid and melamine contamination is warranted.

Optimal nutrition for dogs is thought to differ from that of rats. The crude protein and fat in dog diets are 25% and 9% (Laboratory Rat Diet, Purina Rat Chow 5006), respectively, which are slightly higher than the 23% and 6% in rat diets (Laboratory Rat Diet, Purina Rat Chow 5001). Along with the work on contaminated pet food, a full-diet (100%) study was conducted for twenty-eight days using the control lot of pet food (PB03). The nutritional quality of the dogs’ diet was comparable to that of the rats’ diet. The twenty-eight–day feeding toxicity study of rats revealed no significant abnormality in male or female rats, indicating that the control lot of pet food was not contaminated (data not shown).

The toxic renal lesions formed in rats in this study appeared to be related to MC and were similar to those associated with the renal failure observed previously in dogs and cats (Brown et al. 2007; Jeong et al. 2006; Puschner et al. 2007; Thompson et al. 2008). Mycotoxins have been implicated as renal tubular toxicants; however, neither citrinin nor ochratoxin was detected in the L107 lot. Therefore, mycotoxicosis of the dogs and cats was unlikely to be a cause of the 2004 outbreak.

Although insufficient data are available for humans, the recommended maximum tolerable daily intake is 0.2 mg/kg for melamine and 1.5 mg/kg for cyanuric acid (WHO 2008). Unlike melamine, urinary cyanuric acid levels were not significantly different in affected and unaffected babies. Pathophysiological findings from feeding animals MC may not be directly applicable to humans (Lam et al. 2009), but the combination of melamine and cyanuric acid is clearly responsible for acute renal failure in pets. The lowest administered doses of MC of 32 mg/kg caused acute renal failure in cats (Puschner et al. 2007). In this study, no mycotoxins were found, but large amounts of cyanuric acid (909 mg/kg) and melamine (6191 mg/kg) were detected in lot L107, but not in lot PB03, excluding the possibility that major toxicological effects might have been caused by other triazines in the pet food.

One potential important factor in the development of renal toxicity identified in this study was the ratio of cyanuric acid to melamine. The daily intakes of cyanuric acid and melamine in the 20% group were 11.8 and 80.4 mg/kg for male rats and 14.6 and 99.7 mg/kg for female rats, representing a ratio of approximately 1:6.8 for both males and females. In fact, the lower cyanuric acid with higher melamine diet for twelve weeks in rats did not cause acute renal toxicity. This result is in contrast with the results of previous studies, in which 32 mg/kg of cyanuric acid and melamine (1:1) in cats (Puschner et al. 2007) or a 1:1 mixture of melamine and cyanuric acid (400/400 mg/kg/day) caused acute nephrotoxicity or even death in rats within three days (Dobson et al. 2008). We found that rats fed with 20% or less L107 for three months did not develop renal failure, indicating that the amount of L107 ingested was safe. Renal toxicity began to be evident only at higher doses of L107, such as in Group 4 (for which the daily intakes of cyanuric acid and melamine were 29.4 and 200.4 mg/kg in male rats and 35.2 and 240.0 mg/kg in female rats), and significant intoxication was not observed until the amount of L107 in the diet was increased to 100% for four weeks (when the daily intakes were estimated to be 38.3 and 260.9 mg/kg for male rats and 60.3 and 410.7 mg/kg for female rats). Therefore, it is tempting to speculate that both the ratio and the total amount of cyanuric acid and melamine ingested can contribute to renal intoxication, although the ratio and dose threshold of these compounds seemed to determine critically the severity of nephrotoxicity.

The mechanism by which cyanuric acid and melamine in combination form crystals in the kidneys is remain unclear. The possible mechanisms of melamine nephrotoxicity were recently suggested by Bhalla et al. (2009). However, unlike melamine-caused nephrotoxicity in humans (Bhalla et al. 2009), pets suffered more severe nephrotoxicity that caused numerous deaths during the outbreak of melamine and cyanuric acid co-contamination in pet food. We proposed that these two events bear different toxic pathways. The other possible explanation is that the two compounds have different values of pKa (6.9 for cyanuric acid, 5 for melamine) and re-establish a crystalline structure when they coexist in the kidney. Because melamine is minimally cytotoxic in canine kidney cells, Dobson et al. (2008) proposed that the initial adverse effect was a physical blockage. However, the cytotoxicity of cyanuric acid and melamine at different ratios and toward various cells and species needs to be further investigated.

The histomorphological characteristics and chemical composition of the crystals associated with nephrotoxicosis in dogs were only recently described, and a second type of crystal (calcium oxalate crystals) has been identified (Thompson et al. 2008). In the present study, male rats appeared to be more sensitive and had more severe renal lesions than female rats, possibly reflecting differences in metabolism (Dobson et al. 2008). The formation of MC crystals in unspecified ratios in kidneys has been proposed. The presence of both melamine and cyanuric acid networks reflects the fact that NH···O hydrogen bonds are stronger than NH···N bonds, resulting in the formation of hydrogen bonds between molecules of melamine and cyanuric acid (Dobson et al. 2008; Xu et al. 2007).

Most crystals can evoke an inflammatory response, leading to fibrosis, loss of nephrons, and eventually, chronic renal failure. The most common examples of such crystals in kidneys include calcium oxalate (CaOx), calcium phosphate (CaP), and uric acid or urate crystals (Khan 2004). In the 50%–100% group, MC crystals were observed in the urine and kidneys in affected rats that had suffered severe nephrotoxicity, in which the urine pH level was significantly lower (pH 6.5) than that in the control (pH 7.7) male rats. These findings suggest that a lower pH may promote the formation of MC crystals and nephrotoxicity. In vitro studies of crystal formation, mixing different ratios of cyanuric acid and melamine at different pH levels, are warranted.

Various forms of renal injury not only promote crystal formation, but also may up-regulate OPN expression (Evan et al. 2005). Osteopontin may be one of the crystal-associated urine proteins involved in the formation of the organic layers of plaque particles in MC crystals. The higher OPN expression in Groups 4 and 5 may be related to the severity of lesions in the kidneys. The exposure of renal epithelial cells to CaOx crystals promotes the synthesis of OPN, which is known to participate in inflammatory processes and in the production of extracellular matrix. Deposition of CaOx crystals in rat kidneys also activates the rennin–angiotensin system (Khan 2004), potentially explaining the increase in urine volume in affected rats. Proliferative cellular nuclear antigen is well known as a cell proliferation marker and/or expressed during cell regeneration and has been shown to be involved in DNA repair synthesis (Mehta 1995). The higher PCNA index in the rat kidneys may also be related to tubular regeneration after injury of renal tubules, suggesting that MC crystals may participate in inflammatory processes and induce the injury of the kidneys. The precipitation of these crystals contributes to obstructive renal failure (Puschner et al. 2007). The shapes of the crystals may also be important. Needle-like crystals formed spontaneously from a combination of cyanuric acid and melamine following vacuum drying. The needle-like crystals may be the main cause of acute renal failure, like that of acute uric acid nephropathy, in the early stage (Reimschuessel et al. 2008). However, the pathogenesis of cyanuric acid and melamine combination still needs to be elucidated.

This study demonstrated that rats fed with contaminated pet food developed significant nephrotoxicity in a dose-related manner. The ratio of cyanuric acid to melamine and the level of acidic urine are two factors that determine, upon repeated exposure, the severity of nephrotoxicity. Male rats are more sensitive than female rats, suggesting a sex difference in toxicity.