Cystinuria is a heritable disorder of amino acid transport in epithelial cells of the renal proximal tubule and of the intestine. It is characterized by the urinary hyperexcretion of cystine and the dibasic amino acids lysine, ornithine and arginine; the underlying cause of the disease is the defect in reabsorption across the brush border membrane of the epithelial cells of the renal tubule and intestinal tract. Due to its low solubility, cystine precipitates in the distal tubule to form cystine stones that represent approximately 1 to 2% of all urinary nephrolithiasis in adults and 4 to 5% in children1.
All pathological consequences including infections, pyelonephritis, obstructive uropathy, and renal insufficiency of cystinuria are caused by the recurrent urolithiasis. The intestinal defect in the absorption of cystine and dibasic amino acids has no practical consequences.
Two different types of cystinuria can be distinguished based on the urinary cystine concentration pattern of obligate heterozygotes and on their mode of inheritance: Type I follows an autosomal recessive trait, and heterozygotes show a normal aminoaciduria. In contrast, non-type I heterozygotes exhibit moderate or high excretion of cystine and the dibasic amino acids; therefore, an autosomal dominant inheritance with incomplete penetrance can be defined for this type of cystinuria. The homozygous patients (type I/type I, non-type I/non-type I) and mixed heterozygous patients (type I/non-type I) show a similar range of cystine hyperexcretion independent of the disease type.
To date, two cystinuria genes have been identified2,3: the SLC3A1 gene (2p16.3) encodes the heavy chain rbAT of a renal cystine and dibasic amino acid transporter, and the SLC7A9 gene (19q13.1) encodes the light chain b0,+AT. Whereas mutations in the SLC3A1 gene cause type I cystinuria, patients with non-type I cystinuria harbor mutations in the SLC7A9 gene.
Until now it has been assumed that type I patients show a more severe phenotype with earlier manifestation than non-type I patients. Goodyer et al demonstrated that type I/type I patients form stones in their first decade of life, while patients homozygous for non-type I or mixed heterozygotes only show increased urinary cystine values at this age4.
To estimate the proportion by which mutations in SLC3A1 and SLC7A9 contribute to early manifesting forms of cystinuria, we screened 21 patients who were seen at the different contributing centers of pediatric nephrology. Patients were classified by the urinary cystine excretion of their relatives. In addition, 10 unclassified cystinuric patients referred to urological clinics were examined. Apart from determining the detection rate of mutations in these two genes, we looked for the distribution of the mutations in the two patients group; furthermore, we tested for the correlation between phenotype and genotype.
METHODS
Patients
Twenty-one cystinuric children from 16 families were examined. The patients were classified in cystinuria subtypes on the basis of urinary excretion values of cystine and dibasic amino acids in spontaneous urine specimens of their direct relatives according to Langen et al5.
A summary of the phenotypes of the cystinuric children included in this study is given in Table 1. The majority of families were of German or Turkish origin. All but five patients showed stone formation, the age of first manifestation ranged from 9 months to 14 years. One family with increased urinary cystine excretion but without stone formation yet was identified by a neonatal neuroblastoma screening program (Cys 84/85). Another patient and her brother (Cys 93/94) were recruited due to an increased cystine excretion diagnosed in the context of a metabolic analysis. Neither child has shown stone formation so far.
Table 1 - Summary of phenotypes and mutations in the classified cystinuric children and their relatives.
Full table Additionally, 10 patients with a history of recurrent cystine kidney stones and elevated urine cystine level were included in this study Table 2. The composition of stones was determined by radiation diffactrometry; relatives could not be recruited, so these patients remained unclassified.
The appropriate informed consent was obtained from all patients and their relatives.
Classification of probands
The renal transport defect in the pediatric patients and their parents was assessed by urinary amino acid excretion and percentage tubular amino acid reabsorption (%Taa). The pediatric probands were classified according to the renal amino acid excretion in their parents as described previously5.
To avoid bacterial breakdown of amino acids, urine samples were passed through a sterile filter (0.2 
m pore size) and immediately stored at -20°C. Blood samples were deproteinized within 30 minutes after collection to prevent cystine from binding to plasma protein. Amino acids were determined by ion exchange chromatography on a Biotronic amino acid analyzer LC3000. Creatinine was determined by a kinetic Jaffé reaction. Urinary amino acids were corrected per gram of creatinine. The determination of %Taa was performed with creatinine as marker of glomerular filtration rate as described previously5.
DNA analysis
Genomic DNA was isolated from peripheral lymphocytes by a salting-out procedure6. The whole coding region and the intron-exon boundaries of the SLC3A1 and SLC7A9 genes were screened for variations in genomic DNA by single strand conformation polymorphism analysis (SSCP) in all patients. Fragments sizes were up to 250 bp at maximum. Primer sequences for SLC3A1 were taken from the literature2,7,8,9. The International Cystinuria Consortium kindly provided information on primers in the SLC7A9 gene. In case of exons 2, 4, and 12, SSCP analysis was carried out by investigating two overlapping fragments. Details on these primers as well as polymerase chain reaction (PCR) conditions may be obtained directly from the authors. PCR was carried out in a 25 L volume, containing 80 ng genomic DNA, 50 pmol of each primer, 50 mmol/L KCl, 10 mmol/L Tris-HCl pH 8.3, 1.5 mmol/L MgCl2, 0.01% gelatin, 200 mol/L of each deoxynucleoside triphosphate (dNTP), and 1 U Taq Polymerase (Gibco BRL, Eggenstein, Germany).
For SSCP analysis, 5 L of the PCR product were mixed with 5 L of denaturing solution containing 80% deionized formamide, 0.0125% bromophenol blue, and 0.75% Ficoll 400 and denatured for five minutes at 94°C. Samples were subsequently chilled on ice and then loaded on a 10% polyacrylmide gel (acrylamide/bisacrylamide = 49:1; 110 
120 1.0 mm, Multigel-Long/Biometra) with and without 5% glycerol, containing 0.5 Tris-buffered eosin (TBE). Gels were allowed to run for 16 to 18 hours at 7 V/cm, both at room temperature and at +4°C. Bands were detected by silver staining. To enhance the sensitivity of SSCP, all PCR fragments were electrophoresed under four conditions: at 4°C and at room temperature, with and without glycerol. Nevertheless, there remains the possibility that we have missed a mutation by relying on SSCP as a single mutation screening method.
Unusual SSCP or restriction patterns were characterized by direct sequencing of PCR products using the BigDyeTerminationCycleSequencingSystem (ABI, Foster City, CA, USA). Due to the presence of several polymorphisms and mutations in exon 4 of SLC7A9, interpretation of SSCP data patterns was impossible. Therefore, this exon was sequenced in all patients.
Additionally, PCR-based restriction fragment length polymorphism assays for the known variants T216M, S217R, R270X, R270L, L346P, R362C, R365W, R365L, S547W, M467T, M467K, E483X, c.1500+1G>T in SLC3A1 as well as G105R, T123M, F140S, A182T, c.747delG, 
E244 and R333W in SLC7A9 were used to confirm the results obtained by SSCP. Analyses were performed as described previously2,3,7,8,10,11,12.
In the case of the new mutations P508A in SLC3A1 and A331V in SLC7A9, we established restriction assays as listed in Table 3.
Table 3 - Description of new mutations identified in the SLC3A1 and the SLC7A9 genes, primers for amplification and strategies for mutation detection.
Full table RESULTS
In this study, 21 children with cystinuria from 16 families were investigated. The patients were classified according to the renal amino acid excretion of their parents. Following the classification of urinary amino acid excretion published by Langen et al5, the patients and their parents were typed as follows Table 1: type I homozygotes (I/I), non-type I homozygotes (non-I/non-I), mixed heterozygotes (I/non-I) as well as patients with classical cystinuria, but only one parent available for classification (I/unclassified or non-I/unclassified) excreted the affected amino acids in similar amounts with very high excretion rates of cystine (1.00 to 4.49 mmol/g creatinine), ornithine (1.10 to 3.96 mmol/g creatinine), lysine (4.57 to 16.71 mmol/g creatinine) and arginine (0.89 to 1.43 mmol/g creatinine). The obligate type I heterozygotes showed normal excretion of cystine, ornithine, lysine and arginine, while non-type I heterozygotes showed intermediate values Table 1.
In case of one family (Cys 77), the classification of the mother was ambiguous due to a mild general hyperaminoaciduria of unknown origin.
The classification was confirmed by the patterns of tubular amino acid reabsorption (data not shown): in the cystinuric patients (I/I, non-I/non-I, I/unclassified, non-I/unclassified) %Taa was severely reduced for cystine and the dibasic amino acids; in non-type I heterozygotes %Taa was moderately reduced for cystine and lysine, while ornithine and arginine were reabsorbed normally. Type I heterozygotes removed cystine and dibasic amino acids nearly completely from the tubular lumen.
Based on these biochemical criteria, the 16 unrelated patients were classified as follows: four patients were homozygous for type I cystinuria, two showed homozygosity for non-type I cystinuria, and amino acid excretion in a further two cases revealed heterozygosity for non-type I cystinuria Table 1. The latter (Cys 84/85, Cys 93/94) were identified by a neuroblastoma screening program and a metabolic analysis, respectively. Since single paternal blood and urinary samples were not available, only one cystinuria chromosome could be typed in five patients, in three further patients specimens of both parents were not available. Therefore, 11 of the 30 cystinuria chromosomes were type I, 8 non-type I, and 11 were unclassified (type I or non-type I). The age of first stone formation in the affected children ranged from 9 months to 14 years. Eight of the 16 patients with stone formation have not developed additional stones, 4 have suffered from less than 5 recurrent stones, three from more than 5 stones. In patients without recurrent stone formation, the difference between first stone formation and last examination was at least one year.
In the group of unclassified patients (N = 10), the age of first stone manifestation ranged from 6 to 24 years. All 10 patients were ascertained due to recurrent kidney stone formation.
The whole coding regions and all intron/exon boundaries of the SLC3A1 and the SLC7A9 genes were investigated in the cystinuric children and 10 unclassified patients. In case of patients showing apparent homozygosity, we could not distinguish between homozygosity and hemizygosity, since only small fragments with SSCP and restriction assays were analyzed; large deletions as described for SLC3A1, spanning more than one exon, could not be detected. However, both situations revealed two mutations in the same gene, either homozygosity for the same mutation or compound heterozygosity for the mutation and a deletion including the respective exon.
In total, two novel mutations were identified in SLC3A1 and three in SLC7A9 (Tables 1 to 3). Three were missense mutations and were not found in 50 healthy controls. Additionally we identified one new frame shift mutation in each gene.
Missense mutations
In the SLC3A1 gene, a C to G transition at cDNA position 1523 (exon 9) leading to a substitution from proline to alanine at position 508 (P508A) was detected in heterozygous state in a German patient (Cys 65).
In SLC7A9, a C to T substitution at nucleotide c.856 (exon 6), which changed an alanine codon to a valine codon at position 224 (A224V), was identified in a heterozygous state in two German sibs (Cys 93/94).
A Turkish patient (Cys 70) showed homo- or hemizygosity for the amino acid substitution A331V, caused by a C to T transition at nucleotide c.1177 in exon 10.
To allow rapid genotyping in the future, we developed PCR-based restriction assays for two of these mutations Table 3.
Frame shift mutations
In the SLC3A1 gene, we identified a deletion in exon 10 at position c.1767, resulting in a frame shift mutation (c.1749_1751del). The German patient (Cys 109) was heterozygous for this genomic variation. In an unclassified German cystinuric patient, homo/hemizygosity for a deletion in exon 8 of the SLC7A9 gene (c.969_971del) was found.
Furthermore, a possible splice site mutation was detected in one child of a German sibship (Cys 107): A deletion at nucleotide IVS6+2T (splice site exon 6/intron 6) or IVS6+3T in SLC3A1 was found in heterozygous state.
In addition to the identification of new mutations, six previously described mutations in SLC3A1 and one in SLC7A9 were detected Tables 1 and 2. Out of these, the mutation M467T in SLC3A1 was the most frequent one found, as it was identified in two of the children and four of the unclassified patients. The mutations R365W and R270X in SLC3A1 were detected twice each in the unclassified patients, T216M and S547W were found once each in children, and R362C in an unclassified patient. G105R in SLC7A9 was detected in two children and in one unclassified patient.
Combining these mutations and the new variants, we characterized 12 chromosomes in the pediatric patients (30 chromosomes). In detail, mutations in SLC3A1 were characterized in 54% of the type I chromosomes in children (6 of 11), and in the case of SLC7A9 25% of the non-type I chromosomes (2 of 8) were identified. The combined detection rate was 42.1%. In the remaining 11 unclassified chromosomes, four mutations were detected in SLC7A9 and two in SLC3A1. An overall detection rate of 46.6% (14 of 30 chromosomes) for mutations in both genes was delineated in all of our pediatric patients.
In the 10 unclassified patients, we detected 70% of mutations (14 of 20 chromosomes) in SLC3A1 and SLC7A9: 55% showed mutations in SLC3A1 (11 of 20), 15% in SLC7A9 (3 of 20).
DISCUSSION
In our study, 21 cystinuric children from 16 families as well as 10 unclassified patients were screened for genomic variants in the coding regions of the SLC3A1 and the SLC7A9 genes. We identified two and three novel mutations in each gene, respectively. For SLC3A1, the number of mutations described to date increases to 64, and for SLC7A9 the number increases to 39 [reviewed in10,11,13. Two of the new mutations are frame shift mutations, and the others are missense mutations.
The two new mutations in SLC3A1 (P508A, c.1766_1768del) are localized within the glycosidase-like extracellular domain of the protein. The variant c.1749_1751del leads to a truncation of the C-terminus of the protein, which is thought to be necessary for interaction with the light subunit b0,+AT and may affect transport properties of the transporter complex13. The effect of the missense mutations on transport properties cannot be predicted based on the current knowledge for the rbAT protein structure. Most mutations described and analyzed in this region, however, affect trafficking of the rbAT protein in the membrane (A. Albers et al, personal communication)13. Patients carrying these SLC3A1 mutations show a classical cystinuria, as described for all patients homozygous or compound heterozygous for mutations in this gene.
Heterozygosity for A224V in SLC7A9 was detected in a German sibship (Cys 93/94). The mutation is localized in a highly conserved transmembrane stretch of the b0,+AT subunit. The mutation A331V in SLC7A9 is localized in an intracellular loop. It was detected in a Turkish child whose heterozygous mother also shows a moderate increase of cystine excretion; since we could not distinguish between homo- or hemizygosity of A331V in our patient, we cannot estimate whether homozygosity of this mutation leads to a severe abnormal excretion pattern or whether this pattern is caused by a putative deletion.
The deletion c.969_971del in exon 8 causes the frame shift mutation L262fsX263 with a premature stop codon after position 295.
Moreover, we analyzed the distribution of frequencies of mutations detected in our patients and described in the literature. M467T in SLC3A1 has been described as the most preponderant mutation in German and Southwest European patients: in our pediatric patients, it accounted for 2 of 11 type I chromosomes, in unclassified patients it was detectable in 9 of 56 cystinuria chromosomes, including those published previously by Schmidt et al Tables 1 and 212. Three other mutated alleles also were detected more than once (T216M, R270X, R365W), out of which T216M seems to be preponderant in Greece9 and Turkey. The mutation R270X was detected in a German patient as well as in homo-/hemizygous status in an Israeli. All other mutations in SLC3A1 were detected only once. Thus, in good agreement with the literature, it can be assumed that mutations in SLC3A1 are private or population specific.
In the gene SLC7A9, G105R has been established as the most frequent mutation Tables 1 and 23,11. Here, it was detected in a German and an Italian child suffering from cystinuria, and in two of the unclassified patients.
In the unclassified patients including those 18 published previously12, 62.5% of mutations in SLC3A1 and SLC7A9 were detected: 50% showed mutations in SLC3A1, 12.5% in SLC7A9. This result in an ethnically mixed population agrees with a study in 27 untyped North American patients in which mutations in SLC3A1 and SLC7A9 could be detected only in 56% of chromosomes.
Interestingly, the mutation detection rate of 46.6% in the classified cystinuric children is lower than that of the aforementioned 62.5% in the unclassified patients, although the first group is defined more precisely. In particular, we could explain only 25% of non-type I cystinuria chromosomes by mutations in SLC7A9, which differs from a detection rate of 79% in non-type I chromosomes published by the International Cystinuria Consortium11. In 16 pediatric patients, four thus far unknown mutations, two in the SLC3A1 gene and two in the SLC7A9 gene, have been detected Table 1. In the cohort consisting of 10 unclassified patients Table 2, one new mutation in SLC3A1 and three in SLC7A9 were identified. While the search for mutations in SLC7A9 is just beginning, the spectrum of mutations in SLC3A1 is well established. Thus, it can be hypothesized that children with cystinuria, that is, early manifestation, show in part a different spectrum of mutations from those patients with late-onset. Since we screened the whole coding region by SSCP, these undetected mutations should be localized in genomic regions not analyzed so far or should consist of large deletions that are not detectable with the methods used here. Additionally, further factors/genes seem to be responsible for the disease in children.
Furthermore, the fact that the detection rates for mutations in the two cystinuria genes published by different groups never reach 100% as well as immunohistological investigations14 strongly suggest that other genes encoding subunits for renal amino acids transporters may exist and may carry mutations11. Indeed, Leclerc et al suggested recently that mutations in the asc light chain subunit (SLC7A10) also could be responsible for cystinuria non-type I15. This gene is localized in 19q13, close to SLC7A9. However, it is thought that the asc transporter is expressed basolaterally in polarized epithelia such as the proximal tubule or small intestine and interacts with the heavy chain 4F2hc [reviewed in13. Therefore, it is not clear at present whether SLC7A10 represents a third cystinuria gene.
Variability of age at first stone formation between sibs with the same genotype was observed in some of our patients (Cys 63/64, Cys9/13 in12). Of course, whether a patient is mildly or severely affected is influenced by exogenous factors such as nutrition, drinking volume and therapy. Nevertheless, it can be assumed that sibs grow up under similar circumstances and treatment, and therefore show similar ages of manifestation and severity of the disease, as observed in other genetic diseases such as cystic fibrosis. Thus, the finding of different phenotypes in one sibship points to further modulating factors.
Goodyer et al observed a preponderance of an age of first stone formation in the first decade of life in type I-homozygous patients, while non-type I-homozygous as well as mixed type patients developed stones later in life4. In contrast, we could not find a correlation between age of first stone manifestation and affected gene/type of cystinuria in our study population. Apart from patients homozygous for type I cystinuria and mutations in SLC3A1, we detected patients with early stone formation showing two mutations in SLC7A9 as well as patients being classified as homo-/heterozygous for non-type I cystinuria.
Two patients showed homo-/hemi- or heterozygosity for G105R (Cys 106, Cys 110/112), respectively. The early manifestation of the disease in these patients confirms the observation of the International Consortium of Cystinuria of a severe increase of cystine excretion with heterozygous carriers of the variant G105R11.
The observation of intrafamilial variation of the disease severity is in agreement with the finding of Dahlberg et al that not all homozygous cystinuric patients develop urinary calculi16. They reported that 55% of their series of 89 cystinuric patients formed documented cystine stones, while an additional 13% formed primarily calcium-containing stones. Certain individuals may develop massive staghorn calculi in their early years, while others may form stones only in the second or third decade of life. Clearly, elevated urine cystine levels alone cannot account for stone formation. Other metabolic factors possibly contributing to stone formation are hypocitraturia, hyperuricemia, hyperuricosuria, and hypercalciuria. Further studies are necessary to elucidate the influence of these factors, which also will have hereditary components.
References
| 1. | Stapleton FB, McKay CP & Noe HN. Urolithiasis in children: The role of hypercalciuria. Pediatr Ann 1987; 16: 980−992. | PubMed | ISI | ChemPort | |
| 2. | Calonge MJ, Gasparini P & Chilllaron J et al. Cystinuria caused by mutations in rBAT, a gene involved in the transport of cystine. Nat Genet 1994; 6: 420−425. | Article | PubMed | ISI | ChemPort | |
| 3. | International Consortium of Cystinuria. Non-type I cystinuria caused by mutations in SLC7A9, encoding a subunit (b0,+)AT of rBAT. Nat Genet 1999; 23: 52−57. | ISI | |
| 4. | Goodyer P, Saadi I & Ong P et al. Cystinuria subtype and the risk of nephrolithiasis. Kidney Int 1998; 54: 56−61 10.1046/j.1523-1755.1998.00957.x. | Article | PubMed | ISI | ChemPort | |
| 5. | Langen H, von Kietzell D & Byrd D et al. Renal polyamine excretion, tubular amino acid reabsorption and molecular genetics in cystinuria. Pediatr Nephrol 2000; 14: 376−384 10.1007/s004670050778. | Article | PubMed | ISI | ChemPort | |
| 6. | Miller SA, Dykes DD & Polesky HF. A simple salting out procedure for extracting DNA from human nucleated cells. Nucleic Acids Res 1988; 16: 1215. | PubMed | ISI | ChemPort | |
| 7. | Pras E, Raben N & Golomb E et al. Mutations of the SLC3A1 transporter gene in cystinuria. Am J Hum Genet 1995; 56: 1297−1303. | PubMed | ISI | ChemPort | |
| 8. | Bisceglia L, Calonge MJ & Dello Strologo L et al. Molecular analysis of the cystinuria disease gene: Identification of four new mutations, one large deletion, and one polymorphism. Hum Genet 1996; 98: 447−451. | Article | PubMed | ISI | ChemPort | |
| 9. | Albers A, Lahme S & Wagner C et al. Mutations in the SLC3A1 gene in cystinuric patients: Frequencies and identification of a novel mutation. Genet Test 1999; 3: 227−231. | PubMed | ISI | ChemPort | |
| 10. | Bisceglia L, Purroy J & Jimenez-Vidal M et al. Cystinuria type I: Identification of eight new mutations in SLC3A1. Kidney Int 2001; 59: 1250−1256. | Article | PubMed | ISI | ChemPort | |
| 11. | International Consortium of Cystinuria. Functional analysis of mutations in SLC7A9, and genotype-phenotype correlation in non-type I cystinuria. Hum Mol Genet 2001; 10: 305−316 10.1093/hmg/10.4.305. | Article | PubMed | ISI | ChemPort | |
| 12. | Schmidt C, Albers A & Tomiuk J et al. Analysis of the gene SLC7A9 and SLC3A1 in unclassified cystinurics: Mutation detection rates and association between variants in SLC7A9 and the disease. Clin Nephrol 2002; 57: 342−348. | PubMed | ISI | ChemPort | |
| 13. | Wagner CA, Lang F & Broer S. Function and structure of heterodimeric amino acid transporters. Am J Physiol (Cell Physiol) 2001; 281: C1077−C1093. | PubMed | ISI | ChemPort | |
| 14. | Pfeiffer R, Loffing J & Rossier G et al. Luminal heterodimeric amino acid transporter defective in cystinuria. Mol Biol Cell 1999; 10: 4135−4147. | PubMed | ISI | ChemPort | |
| 15. | Leclerc D, Wu Q & Ellis JR et al. Is the SLC7A10 gene on chromosome 19 a candidate locus for cystinuria? Mol Genet Metab 2001; 73: 333−339 10.1006/mgme.2001.3209. | Article | PubMed | ISI | ChemPort | |
| 16. | Dahlberg PM, Van den Borg CJ & Kurtz SB et al. Clinical features and managements of cystinuria. Mayo Clin Proc 1977; 52: 533−542. | PubMed | ISI | ChemPort | |
Acknowledgments
We thank all of the families involved for their cooperation. We are grateful to all the clinicians who provided us with blood and urinary samples.
