slc7a9-deficient mice develop cystinuria non-i and cystine urolithiasis

Slc7a9-deficient mice develop cystinuria non-Iand cystine urolithiasis

Lıdia Feliubadalo1,2, Marıa Lourdes Arbones1,{, Sandra Manas1, Josep Chillaron2,

Joana Visa3, Margot Rodes4, Ferran Rousaud5, Antonio Zorzano2,3,

Manuel Palacın2,3 and Virginia Nunes1,*

1Medical and Molecular Genetics Center, Institut de Recerca Oncologica, L’Hospitalet de Llobregat, 08907 Barcelona,

Spain, 2Department of Biochemistry and Molecular Biology, Universitat de Barcelona, 08028 Barcelona, Spain, 3Parc

Cientıfic de Barcelona, 08028 Barcelona, Spain, 4Institut de Bioquımica Clınica, Corporacio Sanitaria Clınic, 08028

Barcelona, Spain and 5Servei de Nefrologia, Fundacio Puigvert, 08025 Barcelona, Spain

Received March 28, 2003; Revised and Accepted July 5, 2003

Cystinuria is a common recessive disorder of renal reabsorption of cystine and dibasic amino acids thatresults in urolithiasis of cystine. Cystinuria is caused by defects in the amino acid transport system b0,þ (i.e. therBAT/b0,þAT heteromeric complex). Mutations in SLC3A1, encoding rBAT, cause cystinuria type A,characterized by a silent phenotype in heterozygotes (phenotype I). Mutations in SLC7A9, encoding b0,þAT,cause cystinuria type B, in which heterozygotes in most cases hyperexcrete cystine and dibasic amino acids(phenotype non-I). To facilitate in vivo investigation of b0,þAT in cystinuria, Slc7a9 knockout mice have beengenerated. Expression of b0,þAT protein is completely abolished in the kidney of Slc7a9�/�mice (‘Stones’). Incontrast, Stones expressed significant amounts of rBAT protein, which is covalently linked to unidentified lightsubunit(s). Stones mice present a dramatic hyperexcretion of cystine and dibasic amino acids, while Slc7a9þ/�

mice show moderate but significant hyperexcretion of these amino acids (phenotype non-I). Forty-two per centof Stones mice develop cystine calculi in the urinary system. Calculi develop during the first month of life andgrow throughout the life span of the animals. Histopathology in kidney reveals typical changes for urolithiasis(tubular and pelvic dilatation, tubular necrosis, tubular hyaline droplets and chronic interstitial nephritis). Thefact that some Stones mice, generated in a mixed genetic background, develop cystine calculi from an earlyage, while others do not develop them in their first year of life, suggests the involvement of modifier genes inthe lithiasis phenotype. Thus, Stones provide a valid model of cystinuria which can be used in the study ofgenetic, pharmacological and environmental factors involved in cystine urolithiasis.

INTRODUCTION

Cystinuria (OMIM 220200) is an autosomal-recessive diseaseof renal reabsorption and intestinal absorption of cystine anddibasic amino acids. Cystine precipitates in the urinary systemto form calculi that produce obstruction, infection and,ultimately, renal insufficiency (1). Two cystinuria phenotypesare distinguished on the basis of the cystine and dibasic aminoaciduria of the obligate heterozygotes: (i) normal urinaryexcretion in phenotype I carriers; and (ii) moderate to highexcess of urinary excretion in phenotype non-I carriers.

Mutations in SLC3A1, located on chromosome 2p16.3–21and encoding rBAT, cause phenotype I cystinuria (2,3). Thegene causing phenotype non-I cystinuria was assigned bylinkage to chromosome 19q12–13.1 (4,5), and was identified asSLC7A9, encoding b0,þAT (6). Mutations in SLC7A9 causephenotype non-I and also some phenotype I cases (7–9). In thecohort of patients studied by the International CystinuriaConsortium, mutations in SLC3A1 have been found in 74% ofphenotype I alleles, and mutations in SLC7A9 have been foundin 84% of phenotype non-I alleles and in 10% of phenotype Ialleles (9). Consequently, a genotypic classification of

*To whom correspondence should be addressed at: Centre de Genetica Medica i Molecular, Institut de Recerca Oncologica, Gran Via de Les CortsCatalanes s/n km 2,7, L’Hospitalet de Llobregat, Barcelona 08907, Spain. Email: [email protected]{Present address:Genes and Disease Program, Centre de Regulacio Genomica, 08003 Barcelona, Spain.

Human Molecular Genetics, 2003, Vol. 12, No. 17 2097–2108DOI: 10.1093/hmg/ddg228

Human Molecular Genetics, Vol. 12, No. 17 # Oxford University Press 2003; all rights reserved

Downloaded from https://academic.oup.com/hmg/article-abstract/12/17/2097/692137by gueston 12 February 2018

cystinuria has been proposed: type A, due to SLC3A1mutations, and type B, due to SLC7A9 mutations (9).

rBAT and b0,þAT form a heterodimeric complex in the brush-border membranes of the epithelial cells of the renal proximaltubule (10). Expression of the rBAT/b0,þAT heterodimericcomplex in cultured cells resulted in the induction of systemb0,þ amino acid transport activity (11–13). System b0,þ is anobligatory exchanger that mediates influx of dibasic aminoacids and cystine, and efflux of neutral amino acids (for areview see 14). The involvement of the rBAT/b0,þATheterodimeric complex in cystinuria indicates that systemb0,þ is the main apical reabsorption system for cystine inkidney (10).

The only proven clinical manifestation of cystinuria isurolithiasis. Hexagonal crystals appear in the urine andradiopaque cystine stones develop repeatedly in most cystinuricindividuals. Cystinuria is diagnosed by demonstrating selectivehyperexcretion of cystine and dibasic amino acids in urine.Urolithiasis is prevented by high fluid intake and alkalinizationof urine to maximize cystine solubility. Patients whopersistently develop stones are sometimes treated with oralsulfhydryl agents like D-penicillamine and a-mercaptopropio-nylglycine (15,16). Although quite effective, these agents havemultiple side-effects that often cause discontinuation oftreatment (17,18).

Natural models for cystinuria have been described.Cystinuria has been found in a cat (19), in several breeds ofdog (20) and in wolves (21). The Newfoundland dogs form agroup with more homogeneous biochemical parameters andknown etiology. All of those analyzed have a nonsensemutation in canine Slc3a1 (22). These species are, however,suboptimal for biochemical and genetic studies.

Reported here is the development of a mouse model forcystinuria type B. Homologous recombination was used togenerate mice with a disruption of Slc7a9 in a mixed geneticbackground. All homozygous mutant mice show massiveurinary hyperexcretion of cystine and dibasic amino acids,whereas heterozygotes show lower but clear hyperexcretion ofthese amino acids (phenotype non-I). About 40% of thehomozygous mutants present cystine calculi (i.e. cystinestones) in the urinary system (bladder, renal pelvis and/orureter). This strongly suggests that the lithiasic phenotype isaffected by modifier genes.

RESULTS

Targeted disruption of the Slc7a9 gene and generationof Slc7a9�/� mice

A targeting vector that replaces 6.1 kb of Slc7a9 genomicsequence, including exons 3–9, with a neo1 cassette orientedin the opposite direction to the endogenous Slc7a9 has beenconstructed (Fig. 1A). Of 570 G418-gancyclovir-resistant EScell clones analyzed, two demonstrated the appropriaterecombination as confirmed by Southern blot analysis. Figure1B shows the analysis of the three initial positives, one ofwhich (clone 155) resulted from recombination only in the lefthomology arm. Clones 164 and 291 were used to generatechimeric founder mice. Heterozygotic mice from the F1

generation were identified by PCR analysis and were crossedin order to obtain b0,þAT-deficient mice. An example of the PCRused to genotype all the generations is shown in Figure 1C.Of 280 F2 mice genotyped, 59 were wild-type, 146 wereheterozygous, and 75 were homozygous mutants, in accor-dance with the expected Mendelian ratios (w2

¼ 2.34; d.f.¼ 2).

Genetic and biochemical analysis

To assess the correct splicing of the recombined alleles in targettissues, total RNA was isolated from kidney cortex andmedulla, and from jejunum, of adult male Slc7a9þ/þ,Slc7a9þ/� and Slc7a9�/� mice. Northern blot analysis (Fig. 2)showed two Slc7a9 mRNA species, the wild-type messenger of�1.9 kb and a new messenger of �0.9 kb generated from therecombinant allele by splicing between exons 2 and 10 (datanot shown). Slc7a9þ/� mice showed reduced expression of thewild-type mRNA in both tissues. Slc7a9�/� mice showed noexpression of this messenger, as expected. Conversely, therecombined mRNA appeared in Slc7a9þ/� mice at half thedose found in Slc7a9�/� mice.

Elimination of exons 3–9 in the targeted allele (as confirmedby the sequencing of the RT–PCR product, data not shown)leads to a frameshift after amino acid residue 29 of b0,þAT, andthe addition of 12 missense amino acids before the first stopcodon. This would yield a truncated protein without any of the12 putative transmembrane domains. Since b0,þAT forms adisulfide bound heterodimer with rBAT (10,13), the expressionof both proteins in the presence and absence of DTT in micefrom the three genotypes was studied. Western blot analysis ofb0,þAT and rBAT in renal brush-border membranes is shown inFigure 3. In reducing conditions, b0,þAT appeared as twoprotein bands of �40 and �80 kDa in Slc7a9þ/þ and Slc7a9þ/

� mice, which correspond to monomeric and homodimericb0,þAT (10). Both protein bands showed lower expression inSlc7a9þ/� mice and were completely abolished in Slc7a9�/�

mice. Indeed, densitometric analysis corrected by proteinloading as determined by Ponceau S staining showed that theexpression of b0,þAT protein in Slc7a9þ/� mice was 24% ofthat in Slc7a9þ/þ mice (20 and 28% in two independentdeterminations). In addition a band of �90 kDa of unknownnature was detected, both in the presence and absence of DTT.This band, which is not immunoprecipitated by the anti-b0,þATserum (10), is not revealed with the pre-immune serum and itsexpression is not affected by the Slc7a9þ/� genotype (data notshown). In the same conditions, rBAT is revealed as a proteinband of �94 kDa, which corresponds to the mature N-glycosylated monomeric form of the protein (11). Theexpression of rBAT was only moderately reduced in Slc7a9þ/

� mice (78 and 57% of that in Slc7a9þ/þ mice, as describedabove). In contrast to b0,þAT, the rBAT protein was expressed,although clearly lower, in the renal brush-border membranes ofSlc7a9�/� mice (35 and 48% of that in Slc7a9þ/þ mice). Innon-reducing conditions, b0,þAT and rBAT showed two bandsof �135 and �250 kDa, as expected. The former bandcorresponds to the rBAT/b0,þAT heterodimer, and the latterprobably represents a dimer of heterodimers (10). The nature ofthe b0,þAT band of �170 kDa is unknown. The level ofexpression of b0,þAT and rBAT bands in non-reducingconditions reflects the situation already described for the three

2098 Human Molecular Genetics, 2003, Vol. 12, No. 17


Figure 1. Generation of Slc7a9-knockout mice by homologous recombination in ES cells. (A) Schematic representation of the targeting strategy. The vector isrepresented by the second line from the top, while the wild-type and targeted Slc7a9 alleles are indicated by the top and bottom lines, respectively. TheBamHI sites are indicated by B and the KpnI sites by K. Exons are represented by solid rectangles. The homology arms are represented by gray boxes. The back-bone of the vector is denoted by checked boxes. In the vector and the mutant allele, the PGK-tk and PGK-neo gene cassettes are denoted by open boxes labeledaccordingly. The two external probes (SE50 and SE30) are represented as solid lines at the bottom, and the PCR primers (BreF, NorR and RecR) used to genotypethe mice as solid triangles. (B) Screening and confirmation of recombinant clones by Southern blot. DNA was digested with BamHI (left) or KpnI (right) andprobed with SE50 and SE30 probes, respectively. The wild-type allele generates a 4.5 kb BamHI and a 7.0 kb KpnI fragment, whereas the targeted allele producesa 4.0 kb BamHI and a 10.6 kb KpnI fragment. Note that clone 155 had undergone recombination only in the 50 homology region, yielding the insertion of the wholevector. (C) Genotyping of mice by PCR from tail biopsy DNA. The combination of the three primers shown in (A) allows the discrimination between wild-type (þ/þ),heterozygous (þ/�) and homozygous mutant (�/�) mice, containing respectively only the wild-type (410 bp), the wild-type and the targeted (336 bp) and only thetargeted allele.

Human Molecular Genetics, 2003, Vol. 12, No. 17 2099


genotypes in reducing conditions. The most striking result isthat rBAT shows the characteristic mobility of the heterodimerin the absence of b0,þAT in Slc7a9�/� mice.

Urinary phenotype of Slc7a9 knockout mice

The diagnostic trait of cystinuria in humans is the urinaryhyperexcretion of cystine, lysine, arginine and ornithine. Toassess the correspondence of this phenotype in our mice, 24 hurine from the three mouse genotypes was analyzed. Slc7a9�/�

mice showed massive urine hyperexcretion of cystine anddibasic amino acids (Fig. 4). Thus, urine excretion of cystine,arginine, lysine and ornithine was between 76- and 267-foldhigher in Slc7a9�/� than in Slc7a9þ/þ mice (Fig. 4B andTable 1). The urine excretion of other amino acids was onlyslightly affected (if at all) in Slc7a9�/� mice (Fig. 4A, Table 1and data not shown). Thus, the urine excretion of glutamine is2-fold higher in Slc7a9�/� mice than in Slc7a9þ/� andSlc7a9þ/þ mice. Slc7a9þ/� mice also presented hyperexcretionof cystine and dibasic amino acids, but to a much lesser extentthan Slc7a9�/� mice (2- to 14-fold higher than in Slc7a9þ/þ

mice; Fig. 4B and Table 1). This pattern of urine hyperexcre-tion of amino acids is reminiscent of that reported in humanclassic cystinuria due to severe human SLC7A9 mutations (i.e.

phenotype non-I cystinuria) (6,8,9). The urine excretion ofcystine and dibasic amino acids in females showed a tendencyto be higher (140–200%) than in males for the threephenotypes. This difference is significant for arginine excretionin Slc7a9þ/þ and for the excretion of the three dibasic aminoacids in Slc7a9�/� mice.

Plasma amino acid levels for the three genotypes are shown inTable 1. The only clear difference between the three genotypeswas found in Slc7a9�/� mice, which showed a �30% decreasein plasmatic cystine and dibasic amino acids compared toSlc7a9þ/þ and Slc7a9þ/�mice. This decrease was significant forcystine and lysine. No differences were found for the otheramino acids (Table 1 and data not shown). The hyperexcretion ofcystine and dibasic amino acids with low to normal levels inplasma demonstrates the renal reabsorption defect for theseamino acids in Slc7a9�/� and Slc7a9þ/� mice.

Lithiasic phenotype of Slc7a9 knockout mice

Similarly to human classic cystinuria (for review, see 1), cystinecrystalluria (Fig. 5A) was observed in � 82% of Slc7a9�/�

Figure 2. Northern blot analysis of kidney and small intestine RNA from thethree genotypes (top). The expected wild-type 1.9 kb transcript is detected fromSlc7a9þ/þ and (at a lower intensity) Slc7a9þ/� mice, and absent in Slc7a9�/�

mice. The 0.9 kb targeted transcript appears in Slc7a9þ/� and Slc7a9�/� mice.Ethidium bromide staining (bottom) of the blot reveals equal amounts of RNA.

Figure 3. Western blot analysis of renal membranes from the three genotypesusing antibodies directed against b0,þAT (left panels) and rBAT (right panels).Fifty micrograms of protein renal brush-border membranes from adult malemice was subjected to SDS–PAGE in the absence (�DTT; bottom panels) orin the presence of 100 mM DTT (þDTT; top panels). The genotype correspond-ing to each lane is indicated between top and bottom panels. In reducing con-ditions (þDTT), b0,þAT is revealed as two protein bands: (i) b0,þAT monomer(solid arrowhead; �40 kDa); and (ii) b0,þAT dimer (open arrowhead;�80 kDa). In reducing conditions, rBAT appeared as a prominent unique band(open arrowhead; �94 kDa). In non-reducing conditions, b0,þAT and rBATappear as heterodimer bands (solid arrow; �125 kDa) and as high molecular-weight complexes. The most conspicuous complex (open arrow; �250 kDa)probably corresponds to a dimer of heterodimers.



mice (n¼ 22) but not in Slc7a9þ/þ (n¼ 9) or Slc7a9þ/� mice(n¼ 9).

Most cystinuric patients develop cystine calculi in the urinarysystem (9). Slc7a9�/� mice presented lithiasis in the urinarybladder, renal pelvis and ureter (Fig. 5B–E). In contrast,lithiasis was not observed in adult Slc7a9þ/þ (n¼ 20) orSlc7a9þ/� (n¼ 24) mice. Infrared spectroscopy revealed thatcalculi in the Slc7a9�/� mice were composed of pure cystine(Fig. 5C). Cystine lithiasis in the urinary bladder is a fairlycommon phenotype in Slc7a9�/� mice. Thus, 41 out of 98Slc7a9�/� mice analyzed (�42%) developed cystine calculi inthe bladder. No sex differences were observed (lithiasis in thebladder was observed in 44 and 41% male and female,respectively; w2

¼ 0.26; d.f.¼ 1). Cystine lithiasis in the urinarybladder of Slc7a9�/� mice was observed one month after birthand the size of the calculi increased with age (Fig. 6). Theproportion of animals with lithiasis was conserved throughoutthe life span of the animals (Fig. 6A). Cystine lithiasis in therenal pelvis, ureter and male urethra is a less commonphenotype: it was observed in six mice (four females andtwo males), one female and two males, respectively, out of the41 lithiasic Slc7a9�/� mice.

Crystalluria and cystine urolithiasis are not directly related inSlc7a9�/� mice. Thus, crystalluria was observed in eight out of11 non-lithiasic Slc7a9�/� mice and in 10 out of 11 lithiasicSlc7a9�/� mice.

The presence of calculi in the urinary system can producerenal obstruction (Fig. 5D and E). Eight out of the 41 lithiasicSlc7a9�/� mice analyzed suffered hydronephrosis, hyperure-mic shock or sepsis, produced by renal obstruction due tolithiasis in the urinary system between 2 and 13 months of age.

Histopathological analysis of the kidneys of 11 Slc7a9�/� (withand without calculi) and three Slc7a9þ/þ mice older than 4months showed the following (Fig. 7A–H): the four lithiasicSlc7a9�/� mice presented either hydronephrosis or kidneycalculi, all of them had severe tubular and pelvic dilatation, threepresented tubular necrosis and chronic interstitial nephritis, andtwo presented tubular hyaline droplets. Of the seven non-lithiasic Slc7a9�/�mice, three presented light chronic interstitialnephritis, two moderate pelvic dilatation, and one moderatetubular dilatation. None of them presented tubular necrosis ortubular hyaline droplets. None of the three Slc7a9þ/þ micestudied presented any of these lesions. Several organs from 17mice of 5–7 weeks of age of the three genotypes were alsostudied histopathologically. Skin, spleen, thymus, skeletalmuscle, heart, lung, stomach, intestine, liver, pancreas, brainand eye from the 17 young animals were unaltered.

In cystine lithiasis the bladder increases its size (Fig. 5D–F).Thus, the weight of the bladder (mean� SEM; n¼ 7) was33� 6 mg in Slc7a9þ/þ mice, 42� 9 mg in non-lithiasicSlc7a9�/� mice and 211� 32 mg in lithiasic Slc7a9�/� mice(�6-fold higher than in the two other groups; P< 0.01).Preliminary comparison of the urinary bladder of Slc7a9�/�

mice with Slc7a9þ/þ mice showed urothelial hyperplasia andchronic inflammation (data not shown).

Some lithiasic Slc7a9�/� mice suffered sporadic weight loss,attributable to pain induced by obstruction and friction ofstones. When weight loss and abnormal behavior weredetected, mice were treated with analgesics until recovery orsacrificed. However, the weight progression of Slc7a9�/� micewas not significantly different from that of Slc7a9þ/� andSlc7a9þ/þ mice (data not shown).

Table 1. Urine and plasma amino acid levels. Amino acid level is given as [nmol/(24 h � g body weight)]. Each value is the average (�standard error of the mean) of12 animals of each genotype for urine (top) and eight animals of each genotype for plasma (bottom). Significant differences (P� 0.05) between groups areasterisked in the ratios columns

Amino acid Ratios

�/� (1) þ/� (2) þ/þ (3) (1)/(3) (1)/(2) (2)/(3)

Urine values [nmol/(24 h � g body weight)]Cystine 334.4� 43.7 63.2� 13.2 4.4� 0.6 76.0* 5.3* 14.4*Lysine 2334.0� 243.2 28.1� 6.3 10.8� 1.6 216.1* 83.1* 2.6*Arginine 1096.0� 120.3 8.8� 1.6 4.1� 0.6 267.3* 123.9* 2.1*Ornithine 409.3� 37.9 7.0� 1.5 3.4� 0.3 120.4* 58.1* 2.1Histidine 8.9� 1.4 7.3� 0.8 6.8� 0.6 1.3 1.2 1.1Glycine 28.5� 2.5 26.5� 3.2 25.2� 2.4 1.1 1.1 1.1Alanine 21.0� 3.3 17.2� 2.5 14.5� 1.7 1.4 1.2 1.2Serine 8.1� 1.0 8.2� 1.4 6.4� 0.7 1.3 1.0 1.3Asparagine 9.8� 1.3 10.3� 1.5 7.8� 1.1 1.3 0.9 1.3Glutamine 32.1� 4.6 16.1� 1.6 15.7� 2.7 2.0* 2.0* 1.0

Plasma values (mmol/l)Cystine 29� 3 42� 3 43� 4 0.7* 0.7* 1.0Lysine 145� 18 222� 11 211� 19 0.7* 0.7 1.1Arginine 64� 8 87� 8 80� 5 0.8 0.7 1.1Ornithine 30� 3 36� 5 43� 7 0.7 0.8 0.8Histidine 48� 4 41� 2 44� 2 1.1 1.2 0.9Glycine 171� 15 149� 18 143� 12 1.2 1.1 1.0Alanine 256� 47 204� 34 230� 26 1.1 1.3 0.9Serine 113� 12 99� 10 97� 6 1.2 1.1 1.0Asparagine 37� 5 29� 2 30� 2 1.2 1.3 0.9Glutamine 528� 34 443� 27 444� 25 1.2 1.2 1.0



DISCUSSION

Animal models of human diseases have proven their value forobtaining insight into pathogenic mechanisms and for testingtherapeutic drugs and innovative treatment protocols. With bothapplications in mind we planned to generate a mouse model ofcystinuria by targeted disruption of Slc7a9. Evidence thatStones, the knockout mouse described in this study, is a validmodel of human cystinuria is based on the disruption of Slc7a9gene, the deficiency in b0,þAT protein, impairment of the renalreabsorption of cystine and dibasic amino acids, cystinecrystalluria and cystine urolithiasis.

Analysis of genomic DNA in Stones mice demonstrates theabsence of the normal gene. The normal messenger is not

present in any of the target tissues of Slc7a9 (i.e. kidney andsmall intestine). Instead, there is a smaller messenger, whosesize and sequence correspond to an aberrant Slc7a9 splicedvariant that would render a soluble protein of 41 amino acidresidues (the first N-terminal 29 amino acid residues plus 12missense amino acid residues). This is consistent with thefinding that renal brush border membranes of Stones mice aretotally deficient in b0,þAT protein.

The physiological proof of the impairment of cystine anddibasic amino acid reabsorption in the kidney of all Stonesmice is the massive hyperexcretion of cystine and dibasicamino acids with subnormal levels of these amino acids inplasma. This hyperexcretion in urine may explain the low levelsin plasma. Reduced levels of the four amino acids have alsobeen reported in human cystinuria (23,24). Moreover, Stonesmice present cystine crystalluria, as a result of the high urineexcretion of cystine combined with its low solubility. Finally,Stones mice develop cystine calculi in the urinary system. Mostcalculi are located in the urinary bladder, where urineaccumulates. Less commonly, calculi can also appear in therenal pelvis, ureter and male urethra. This lithiasic phenotyperesembles that of human classic cystinuria. Indeed, the firstcystine stone discovered was found in the urinary bladder (25),and the most common location for calculi in patients withcystinuria is pelvi-calyceal. In humans, when cystinuria issymptomatic, cystine lithiasis occurs early in life (9). Thus,most symptomatic patients have their first calculus identifiedbefore the age of 20 (83%). This corresponds well with theearly appearance, in the first 3 months of life, of calculidetected by radiography in Stones mice.

Lithiasis in Stones mice produces hydronephrosis, thecharacteristic consequence in humans (26,27) and animals(28,29) of urinary outflow obstruction. This includes tubularand pelvic dilatation, tubular necrosis, tubular hyaline dropletsand chronic interstitial nephritis. Some Stones mice died fromcomplications from cystine lithiasis, like hyperuremic shock orsepsis.

The murine cystinuria described here strongly resembleshuman classic cystinuria type B with phenotype non-I (9). Theurinary excretion of the affected amino acids in type B patients(i.e. those with two mutations in SLC7A9) is, compared withcontrol individuals: 33-fold (cystine); 35-fold (lysine); 175-fold(arginine); and 74-fold (ornithine) (9). Similarly, Stones miceexcrete 76-fold (cystine), 216-fold (lysine), 267-fold (arginine)and 120-fold (ornithine) that of Slc7a9þ/þ mice The higherhyperexcretion in cystinuric mice might be the reflection of ahigher amino acid renal load. System b0,þ is believed to be themain renal reabsorption system of cystine, whereas othertransport systems should play a role in the reabsorption ofdibasic amino acids, as discussed elsewhere (10). Lysine showsthe largest difference in amino acid hyperexcretion betweencystinuric mice and humans. This suggests a more relevant roleof system b0,þ in the renal reabsorption of lysine in mice thanin humans. Most type B carriers show an increase in theexcretion of the four amino acids (phenotype non-I): 8-fold(cystine), 9-fold (lysine), 6-fold (ornithine) and 6-fold(arginine) that of control individuals (8). Similarly, Slc7a9þ/�

mice excrete 14-fold (cystine), 3-fold (lysine), 2-fold (arginine)and 2-fold (ornithine) that of Slc7a9þ/þ mice. In contrast,phenotype I human carriers do not hyperexcrete these amino

Figure 4. Aminoaciduria of cystine and dibasic amino acids in homozygous(�/�) and heterozygous (þ/�) mutant mice. (A) Representative amino acidchromatograms of urine of a homozygous mutant (�/�) mouse and a wild-type(þ/þ) mouse. The homozygous mutant (�/�) mice show large peaks forcystine, lysine, arginine and ornithine. All other amino acids showed peaksof similar size in homozygous mutant (�/�) and wild-type (þ/þ) mice. (B)Hyperexcretion of cystine and dibasic amino acids in the urine of homozygousmutant (�/�) and heterozygous mutant (þ/�) mice. Urine was collected from12 animals (six males and six females) in each group for 24 h and processed foramino acid level quantification by HPLC. Amino acid levels [nmol/(24 h�gbody weight)] are shown on a logarithmic scale. Homozygous mutant (�/�)mice showed dramatic hyperexcretion of cystine and dibasic amino acids.Heterozygous mutant (þ/�) mice showed moderate hyperexcretion of theseamino acids, which was significant for arginine, lysine and cystine. A moredetailed description of the urine amino acid levels of these mice is given inTable 1.



acids in urine (9). The fact that hyperexcretion of dibasic aminoacids in Slc7a9þ/� mice is lower than in human type Bheterozygotes suggests that mice are gifted with an excess ofsystem b0,þ function. Thus, 24% expression of b0, þ AT

results only in a 2-fold excretion of these amino acids. Stonesmice provide a useful model to study the biology of the rBAT/b0,þAT heterodimeric complex. On the one hand, it has beenshown that b0,þAT increases the stability of rBAT when

Figure 5. Cystine lithiasis in Slc7a9�/�mice. (A) Crystalluria in Slc7a9�/�mice. The picture shows cystine hexagonal crystals in the urine of an Slc7a9�/� mouse.Bar, 10 mm. (B) Radiograph of an Slc7a9�/�mouse, showing cystine calculi in the kidneys (arrowheads) and the urinary bladder (arrow). (C) Infrared spectrometry of aurinary bladder calculus from an Slc7a9�/�mouse. This spectrum matches that of pure cystine. (D) Morphological aspect of the urinary system in an Slc7a9�/�mousecompared with that of a normal individual (Slc7a9þ/þ). Scale in mm. (E) Radiograph of the urinary systems in (B). Arrowheads point to calculi in both kidneys. Arrowspoint to calculi in the urinary bladder and in the right ureter. Genotypes are indicated as in (D). (F) Cystine calculi and macroscopic changes in kidney and urinarybladder bearing a calculus removed in the necropsy of the Slc7a9�/� mouse shown in (D)–(F). The urinary bladder calculus is shown in the top right-hand cornerof the figure. The renal pelvis calculus is shown at the bottom of the figure. Notice the dramatic increase in size in the urinary bladder and kidney in the Slc7a9�/�

mouse compared with the wild-type. Urinary bladders (top) and kidneys (bottom) are shown excised in two halves. Genotypes are shown at the top. Scale in mm.



overexpressed in cultured cells (11,13). In agreement with this,Stones mice, which are fully deficient in b0,þAT, showedreduced expression of rBAT in renal brush-border membranes.On the other hand, co-immunopurification studies showed thatall b0,þAT in human and mouse renal brush-border membranesheterodimerizes with rBAT, whereas part of the rBAT proteinheterodimerizes with unidentified light subunit(s) (10). Ourfindings in Stones mice fully support these results, because inthe absence of b0,þAT the remaining rBAT protein has theelectrophoretic mobility of a heterodimer in renal brush-bordermembranes.

Stones mice also provide a useful model to study themechanisms underlying phenotypes I and non-I in cystinuria.Mutations in either of the two subunits that form system b0,þ

(i.e. the rBAT/b0,þAT heterodimeric complex) probably explainall cases of human classic cystinuria (9). All the mutationsidentified in rBAT, including mild and severe mutations withlarge SLC3A1 deletions, cause phenotype I in cystinuria (i.e.normal excretion of cystine and dibasic amino acids inheterozygotes) (9). In contrast, most mutations of b0,þATcause phenotype non-I (i.e. hyperexcretion of cystine anddibasic amino acids in heterozygotes), whereas mild b0,þATmutations (i.e. with significant residual transport activity) causephenotype I (8,9). Very recently it has been shown that b0,þATis the ‘catalytic subunit’ of system b0,þAT (11). This indicatesthat severe mutations in the ‘catalytic subunit’ in heterozygosishave a greater impact on renal reabsorption than mutations inrBAT. Interestingly, patients homozygous for rBAT (type Acystinuria) or b0,þAT (type B cystinuria) show the same levelsof hyperexcretion of cystine and dibasic amino acids (9). In thisscenario, Slc7a9þ/� mice provide a new clue: the absence ofb0,þAT results in phenotype non-I. In all, these data support thefollowing hypothesis for a mechanism of physiopathology incystinuria: (i) the amount of b0,þAT controls the expression ofthe functional rBAT/b0,þAT heterodimeric complex; (ii) therBAT protein is produced in excess in kidney, and therefore anrBAT mutation in heterozygosis does not lead to hyperexcre-tion of amino acids; (iii) interaction with b0,þAT stabilizesrBAT, and the excess of rBAT is degraded; and (iv) a half-doseof b0,þAT, results in half-expression of rBAT/b0,þAT hetero-dimer, and therefore of the system b0,þ reabsorption activitythat causes hyperexcretion of cystine and dibasic amino acids.

Stones mice may show that modifier genes are involved in thedevelopment of cystine calculi. This possibility in humancystinuria is suggested by certain clues: (i) marked differencesbetween siblings sharing the same mutations (i.e. one siblingcan have very aggressive lithiasic phenotype while the otherdoes not develop calculi) (30); (ii) in the cohort of patients ofthe International Cystinuria Consortium, 12 (three of themolder than 40) out of 224 patients with cystinuria did notdevelop renal calculi (9); and (iii) even when the severity ofSLC7A9 mutations can be correlated with the urinary excretionof amino acids in carriers (8), renal stone formation cannot bedirectly correlated with amino acid urinary excretion in patients(9). It therefore appears that, although mutations in the twoknown cystinuria genes are related to amino acid excretion,once a patient’s urine becomes saturated with cystine, there areother factors, both environmental and genetic, that play adecisive role in calculi formation. A number of factors affectingrenal calculi formation and evolution have been described forcystine and for other urolithiases (e.g. calcium oxalatelithiasis). These include the electrolyte urinary content, thepresence of proteins like bikunin and other inter-alphainhibitors (31) and osteopontin (32). Analysis of Stones micecan help us to disclose environmental and genetic factors. AllStones mice analyzed showed hyperexcretion of cystine inurine, but only 82% developed crystalluria and only 42%developed cystine lithiasis. The reason why some Stones miceproduce cystine calculi while others did not is uncertain, butthe mixed genetic background (strains 129P2 and C57BL/6) ofthe mice could be contributory. We are currently providing

Figure 6. Frequency and size of cystine lithiasis in the bladder of Slc7a9�/�

mice. (A) Number of Slc7a9�/� mice with (solid bars) or without (open bars)lithiasis in the urinary bladder distributed by age in months. The 98 Slc7a9�/�

mice studied correspond to F2 and F3 generations in a mixed genetic back-ground of 129P2/OlaHsd and C57BL/6J. Mice were not radiographed duringthe first month of life. The proportion of mice bearing cystine calculi in theurinary bladder is 42%, which is maintained after the first month of life. (B)Estimation of the volume of the cystine calculi in the bladder of Slc7a9�/� micedistributed by age in months. The volume of the cystine calculi was estimatedfrom dorso-ventral and latero-lateral radiographs of the mice, by multiplyingtheir greatest diameter in both radiographs (y-axis) by their lowest diameterin each of the radiographs (x- and z- axes). Calculi tend to grow throughoutthe life of lithiasic Slc7a9�/� mice.



knockouts with a pure genetic background, by back-crossingheterozygous mice to C57BL/6J. When an appropriate genera-tion is achieved, we will study the presence of calculi in theurinary system. A change in the frequency of lithiasis inSlc7a9�/�mice with different genetic backgrounds will confirm

whether modifier genes are involved in the cystine lithiasisphenotype. Mouse genetics will help to identify these genes.

Human cystine calculi are removed by surgery and, in thecase of recently formed calculi, by extracorporeal shock wavelithotripsy. Conservative treatment in patients with cystinuria

Figure 7. Histopathological study of kidney sections of Slc7a9�/� and Slc7a9þ/þ mice (periodic acid-Schiff stained). (A) Dilatation of the renal pelvis caused bythe presence of a calculus in a Slc7a9�/� mouse (10� ). (B) Tubular dilatation with flattening of renal tubular epithelial cells in the renal cortex of an Slc7a9�/�

mouse (100� ). (C) Tubular necrosis and tubular hyaline droplets (rounded and densely fuchsin-stained, pointed by arrowheads), both sign of atrophy of the func-tional tissue in the renal medulla of an Slc7a9�/� mouse (400� ). (D) Chronic interstitial nephritis characterised by interstitial inflammatory infiltrate and tubularloss in the renal cortex of an Slc7a9�/� mouse. (E)–(H) Equivalent kidney sections from Slc7a9þ/þ mice. pd, pelvis dilatation; m, medulla; c, cortex; t, tubule; gl,glomerulus; icn, interstitial chronic nephritis.



combines increased oral fluid intake, reduction of sodium in thediet, moderate reduction of protein intake and urine alkaliniza-tion. Pharmacological treatment with thiols, although quiteeffective, is reserved for the most severe cases, due to thenumber and severity of their secondary effects (18,33).Lithiasic Stones mice will hopefully be useful for testing newtherapeutic and metaphylaxis protocols for cystinuria.

MATERIALS AND METHODS

Generation of Slc7a9�/�mice

15.4 kb of mouse Slc7a9 genomic DNA was isolated byscreening a 129/SvJ library (Stratagene) using mouse Slc7a9cDNA as a probe. Using this genomic clone and the pSP72vector (Promega), a replacement vector was constructed(Fig. 1A). This has a 2.8 kb left homology arm containingexons 1–2, a 1.6 kb PGK-neo cassette (in antisense orienta-tion), a 2.9 kb right homology arm containing exons 10–11,and a 2.7 kb PGK-tk cassette. E14-1 embryonic stem (ES) cells(34) were electroporated (Biorad Gene Pulser at 500 mF, 240 V)with the PvuI-linearized targeting vector, and selected in amedium containing 170 mg/ml G418 (Life Technologies) and2 mM gancyclovir (Syntex). Resistant ES clones were pickedafter 7–9 days of double selection, and cell clones wereestablished. To screen for recombinant events, ES cell genomicDNA was digested with BamHI and analyzed by Southernblotting using a BamHI–KpnI probe upstream the left arm ofhomology. KpnI-digested DNA and a 30-external probe (SE3 inFig. 1A) were used for further analysis of positive recombinantclones. Slc7a9 heterozygous ES cells from two targeted cloneswere microinjected into C57BL/6J blastocysts, which weretransferred into uteri of pseudopregnant CD1 females.Chimeric male progeny were crossed to C57BL/6J females(Charles River). Germline transmission of the disrupted allelewas detected in the agouti progeny by PCR from tail-biopsyspecimens. The same PCR was used to genotype subsequentgenerations. The forward primer (BreF: 50-CTGTTCTGTTCTGACCAACTGAGGGGCA-30) is on the left homologyarm. One reverse primer specific for the wild-type allele (NorR:50-GATACAGATGCCACTGAGAAGACCCACC-30) annealsto the deleted region, and another reverse primer specific forthe recombined allele (RecR: 50-ATTCGCAGCGCATCGCCTTCTATCGCC-30) anneals to the Neo gene. The expectedsizes of wild-type and targeted alleles are shown in Figure 1C.Only clone 291 was successful in germline transmission.Analysis thus far has been carried out on the hybrid C57BL/6J-129P2/OlaHsd background. All animals were housed in theAnimal Care Unit of the Institut de Recerca Oncologica (IRO;Barcelona) in accordance with animal care guidelines. Allprocedures were approved by the IRO Animal Use and CareCommittee.

mRNA analysis

Mouse kidney cortex and medulla, and jejunum RNA wereisolated from adult male mice using the Tripure IsolationReagent (Roche), according to the manufacturer’s directions.Aliquots of RNA were separated on a formaldehyde agarose gel

and transferred to nylon, and Slc7a9 mRNA was detected witha 32P-labeled probe. The probe was a PCR product from theIntegrated Molecular Analysis of Genomes and theirExpression (IMAGE) clone 578502. A portion of Slc7a9cDNA (NM_021291) between positions 1337 and 1811 wasamplified using primers P6oli12F (50-GCTTAAAGTGCTCT-CCTACATC-30) and P6oli1R (50-GGGCTACGAGTGAT-GGACCTT-30). RT–PCR analysis of the recombinant kidneymRNA was performed on Slc7a9�/� mice. Reverse transcrip-tion was carried out with the First-strand cDNA Synthesis Kit(Amersham), according to the manufacturer’s instructions. ThecDNA was amplified with several primers, all of them yieldingthe expected sizes. The forward primers were: P6oli16F.2(50-GTCTTTCTATGTACCCCAAT-30) and P6oli1F (50-ATG-AGAAATCCACCCACAGTA-30), and the reverse primerswere P6oli1R (50-GGGCTACGAGTGATGGACCTT-30),P6oli2R (50-ACAATGATAGGGATGAAGAGG-30) andP6oli3R (50-CCAGGGATGATGTAAATGATG-30). PCR pro-ducts were purified and sequenced with the same primers.

Protein analysis

Brush-border membrane vesicles of mouse kidney wereprepared by the Ca2þ precipitation method (35). N-ethylmalei-mide at 5 mM was present in all buffers used (except in theresuspension buffer) following Wang and Tate (36), to preventartifactual reduction/shuffling of disulfides. The membraneswere kept at �80�C until use. Brush-border membrane vesicleswere obtained from the cortex and medulla of adult male miceof the three genotypes from the F2 generation. The proteincontent of the membrane preparations was measured by themethod of Bradford (37) using g-globulin as a standard.Proteins were boiled in the presence or absence of 100 mM

dithiothreitol (DTT), separated using SDS–PAGE (20 mgprotein per lane), transferred and probed with a rabbitpolyclonal anti-mouse b0,þAT antibody (10) and with a rabbitpolyclonal anti-rabbit rBAT antibody MANR-X (38), followedby enhanced chemiluminiscence detection (Amersham) asdescribed (10,38). As a loading control, the membranes werePonceau S stained [0.5% (w/v) Ponceau S in 1% (v/v) aceticacid].

Urine and plasma collection and analysis

From each experimental animal, urine was collected in a mousemetabolic cage (Tecniplast) after a 2-day adaptation period.Twenty-four-hour urine from three consecutive days wascollected and immediately frozen. Thawed urine was mixedwith an equal part of 0.4 M homoarginine in 0.1 M HCl (as astandard for the HPLC analysis) and deproteinized byultrafiltration through a 10 000 nominal molecular weight limitregenerated cellulose membrane (Millipore). For cystineanalysis, urine samples of Slc7a9�/� mice were diluted 20-fold to solubilize possible cystine crystals prior to the additionof the homoarginine standard. Blood was collected by cardiacpuncture of anesthetized animals with a heparinized syringe.After a 5 min centrifugation at 1500g, plasma was recovered,mixed with the standard and deproteinized like the urine. Thiswas done immediately after blood collection to preventcystine binding to proteins. Quantitative analysis of amino



acids was performed by pre-column derivatization withphenylisothiocyanate (PITC) followed by reverse-phasehigh-performance liquid chromatography (HPLC) in a Watersapparatus with a digital computer according to the WatersPICO-TAG procedure.

Calculi detection and analysis

Calculi in the urinary system were either detected by X-rayradiography (at 28 kV, 16 mA � s) or detected and recoveredduring necropsy. Calculi composition was analyzed with aninfrared spectrometer (Brucker), after mixing a fragment withKBr, pulverizing it in an agate mortar and pressing the mixtureat 10 atm for 10 min.

Crystalluria

Crystalluria was detected in urine collected on a Petri dish afterspontaneous voiding. Crystals were observed under anOlympus BX60 microscope, at a magnification of �1000.

Histology

Animals were sacrificed by CO2 and tissues were fixed with 4%paraformaldehyde O/N at 4�C. Paraffin sections of 4 mm werestained with hematoxylin and eosin or with periodic acid-Schiff, examined and photographed with a light microscopeNikon Eclipse E-800 or a lens Leica MZ125.

Statistical analysis

Data are expressed as mean� SEM. One-way ANOVA,Bonferroni and Dunnett T3 tests were used to compare valuesbetween the different genotypes. Expected and experimentalfrequencies were compared by chi-squared test. The level ofsignificance was set at P< 0.05.

ACKNOWLEDGEMENTS

We thank Marta Oset from the Transgenic Unit AnimalResearch Center (Parc Cientıfic de Barcelona) for themicroinjection of the clones, and David Solanes and co-workers at the Animal Care Unit of the IRO for their supportwith the mice. We thank Robin Rycroft for editorial help. Thisstudy was supported in part by the Spanish Ministry of Scienceand Technology PM99-017-CO-01/02 (M.P., V.N.), FundacioLa Marato de TV3 ref. 98/1930 (V.N.), BIOMED BMH4CT98-3514 (V.N., M.P.), the support of the Comissionat per aUniversitats i Recerca (M.P.) and grants 2001SGR00118(A.Z., M.P.) and 2001SGR00399 (V.N.) de la Generalitat deCatalunya (Spain), and by Instituto de Salud Carlos IIInetworks C03/07 (V.N.) and G03/054 (M.P., V.N.).

REFERENCES

1. Palacın, M., Goodyer, P., Nunes, V. and Gasparini, P. (2002) Cystinuria. InThe Metabolic and Molecular Bases of Inherited Disease. McGraw-Hill,New York. Available at: www.genetics.accessmedicine.com.

2. Calonge, M.J., Volpini, V., Bisceglia, L., Rousaud, F., de Sanctis, L.,Beccia, E., Zelante, L., Testar, X., Zorzano, A., Estivill, X. et al. (1995)Genetic heterogeneity in cystinuria: the SLC3A1 gene is linked to type I butnot to type III cystinuria. Proc. Natl Acad. Sci. USA, 92, 9667–9671.

3. Calonge, M.J., Gasparini, P., Chillaron, J., Chillon, M., Gallucci, M.,Rousaud, F., Zelante, L., Testar, X., Dallapiccola, B., Di Silverio, F. et al.(1994) Cystinuria caused by mutations in rBAT, a gene involved in thetransport of cystine. Nat. Genet., 6, 420–425.

4. Bisceglia, L., Calonge, M.J., Totaro, A., Feliubadalo, L., Melchionda, S.,Garcia, J., Testar, X., Gallucci, M., Ponzone, A., Zelante, L. et al. (1997)Localization, by linkage analysis, of the cystinuria type III gene tochromosome 19q13.1. Am. J. Hum. Genet., 60, 611–616.

5. Wartenfeld, R., Golomb, E., Katz, G., Bale, S.J., Goldman, B., Pras, M.,Kastner, D.L. and Pras, E. (1997) Molecular analysis of cystinuria inLibyan Jews: exclusion of the SLC3A1 gene and mapping of a new locuson 19q. Am. J. Hum. Genet., 60, 617–624.

6. Feliubadalo, L., Font, M., Purroy, J., Rousaud, F., Estivill, X., Nunes, V.,Golomb, E., Centola, M., Aksentijevich, I., Kreiss, Y. et al. (1999) Non-type I cystinuria caused by mutations in SLC7A9, encoding a subunit(b0,þAT) of rBAT. International Cystinuria Consortium. Nat. Genet., 23,52–57.

7. Leclerc, D., Boutros, M., Suh, D., Wu, Q., Palacın, M., Ellis, J.R.,Goodyer, P. and Rozen, R. (2002) SLC7A9 mutations in all three cystinuriasubtypes. Kidney Int., 62, 1550–1559.

8. Font, M., Feliubadalo, L., Estivill, X., Nunes, V., Golomb, E., Kreiss, Y.,Pras, E., Bisceglia, L., d’Adamo, A.P., Zelante, L. et al. (2001) Functionalanalysis of mutations in SLC7A9, and genotype–phenotype correlation innon-type I cystinuria. Hum. Mol. Genet., 10, 305–316.

9. Dello Strologo, L., Pras, E., Pontesilli, C., Beccia, E., Ricci-Barbini, V.,De Sanctis, L., Ponzone, A., Gallucci, M., Bisceglia, L., Zelante, L. et al.(2002) Comparison between SLC3A1 and SLC7A9 cystinuria patientsand carriers: a need for a new classification. J. Am. Soc. Nephrol., 13,2547–2553.

10. Fernandez, E., Carrascal, M., Rousaud, F., Abian, J., Zorzano, A.,Palacın, M. and Chillaron, J. (2002) rBAT-b(0,þ)AT heterodimer is themain apical reabsorption system for cystine in the kidney. Am. J. Physiol.Renal Physiol., 283, F540–F548.

11. Reig, N., Chillaron, J., Bartoccioni, P., Fernandez, E., Bendahan, A.,Zorzano, A., Kanner, B., Palacın, M. and Bertran, J. (2002) The lightsubunit of system b(o,þ) is fully functional in the absence of the heavysubunit. EMBO J., 21, 4906–4914.

12. Chairoungdua, A., Segawa, H., Kim, J.Y., Miyamoto, K., Haga, H.,Fukui, Y., Mizoguchi, K., Ito, H., Takeda, E., Endou, H. et al. (1999)Identification of an amino acid transporter associated with the cystinuria-related type II membrane glycoprotein. J. Biol. Chem., 274, 28845–28848.

13. Pfeiffer, R., Loffing, J., Rossier, G., Bauch, C., Meier, C., Eggermann, T.,Loffing-Cueni, D., Kuhn, L.C. and Verrey, F. (1999) Luminalheterodimeric amino acid transporter defective in cystinuria. Mol. Biol.Cell, 10, 4135–4147.

14. Chillaron, J., Roca, R., Valencia, A., Zorzano, A. and Palacın, M. (2001)Heteromeric amino acid transporters: biochemistry, genetics, andphysiology. Am. J. Physiol. Renal Physiol., 281, F995–F1018.

15. Chow, G.K. and Streem, S.B. (1996) Medical treatment of cystinuria:results of contemporary clinical practice. J. Urol., 156, 1576–1578.

16. Barbey, F., Joly, D., Rieu, P., Mejean, A., Daudon, M. and Jungers, P. (2000)Medical treatment of cystinuria: critical reappraisal of long-term results.J. Urol., 163, 1419–1423.

17. Crawhall, J.C. and Watts, R.W. (1968) Some complications observed in thetreatment of cystinuria with D-penicillamine and N-acetyl-D-penicillamine.Postgrad. Med. J. (suppl.), 8–11.

18. Dahlberg, P.J., van den, B., Kurtz, S.B., Wilson, D.M. and Smith, L.H. (1977)Clinical features and management of cystinuria. Mayo Clin. Proc., 52, 533–542.

19. DiBartola, S.P., Chew, D.J. and Horton, M.L. (1991) Cystinuria in a cat.J. Am. Vet. Med. Assoc., 198, 102–104.

20. Hoppe, A., Denneberg, T., Jeppsson, J.O. and Kagedal, B. (1993) Urinaryexcretion of amino acids in normal and cystinuric dogs. Br. Vet. J., 149,253–268.

21. Bovee, K.C., Bush, M., Dietz, J., Jezyk, P. and Segal, S. (1981) Cystinuriain the maned wolf of South America. Science, 212, 919–920.

22. Henthorn, P.S., Liu, J., Gidalevich, T., Fang, J., Casal, M.L., Patterson, D.F.and Giger, U. (2000) Canine cystinuria: polymorphism in the canineSLC3A1 gene and identification of a nonsense mutation in cystinuricNewfoundland dogs. Hum. Genet., 107, 295–303.



23. Morin, C.L., Thompson, M.W., Jackson, S.H. and Sass-Kortsak, A.(1971) Biochemical and genetic studies in cystinuria: observationson double heterozygotes of genotype I–II. J. Clin. Invest., 50, 1961–1976.

24. Dent, C.E., Senior, B. and Walshe, J.M. (1954) Pathogenesis of cystinuria.II. Polarographic studies of metabolism of sulfur containing amino acids.J. Clin. Invest., 33, 1216–1226.

25. Wollaston, W.H. (1810) On cystic oxide, a new species of urinary calculus.Phil. Trans. Roy. Soc. London, 100, 223–230.

26. Moller, J.C. and Skriver, E. (1985) Quantitative ultrastructure of humanproximal tubules and cortical interstitium in chronic renal disease(hydronephrosis). Virchows Arch. A. Pathol. Anat. Histopathol., 406,389–406.

27. Hill, G.S. (1989) Basic physiology and morphology of hydronephrosis.In Uropathology. Churchill Livingstone, New York, p. 467.

28. Greaves, P. (2000) Histopathology of Preclinical Toxicity Studies. Elsevier,Oxford, pp. 545–596.

29. Ward, J.M., Mahler, J.F., Macronpo, R.R., Sunberg, J.P. andFrederikson, R.M. (2000). Pathology of Genetically Engineered Mice. IowaState University Press, Iowa, pp. 347–351.

30. Dello Strologo, L., Carbonari, D., Gallucci, M., Gasparini, P., Bisceglia, L.,Zelante, L., Rousaud, F., Nunes, V., Palacin, M. and Rizzoni, G. (1997)Inter and intrafamilial clinical variability in patients with cystinuria type Iand identified mutation. J. Am. Soc. Nephrol., 8, 388A.

31. Iida, S., Peck, A.B., Johnson-Tardieu, J., Moriyama, M., Glenton, P.A.,Byer, K.J. and Khan, S.R. (1999) Temporal changes in mRNA expressionfor bikunin in the kidneys of rats during calcium oxalate nephrolithiasis.J. Am. Soc. Nephrol., 10, 986–996.

32. Xie, Y., Sakatsume, M., Nishi, S., Narita, I., Arakawa, M. and Gejyo, F.(2001) Expression, roles, receptors, and regulation of osteopontin in thekidney. Kidney Int., 60, 1645–1657.

33. Crawhall, J.C. (1981) Experience with penicillamine in the treatment ofcystinuria. J. Rheumatol. Suppl., 7, 100–102.

34. Kuhn, R., Rajewsky, K. and Muller, W. (1991) Generation and analysis ofinterleukin-4 deficient mice. Science, 254, 707–710.

35. Malathi, P., Preiser, H., Fairclough, P., Mallett, P. and Crane, R.K. (1979) Arapid method for the isolation of kidney brush border membranes. Biochim.Biophys. Acta, 554, 259–263.

36. Wang, Y. and Tate, S.S. (1995) Oligomeric structure of a renal cystinetransporter: implications in cystinuria. FEBS Lett., 368, 389–392.

37. Bradford, M.M. (1976) A rapid and sensitive method for the quantitation ofmicrogram quantities of protein utilizing the principle of protein-dyebinding. Anal. Biochem., 72, 248–254.

38. Furriols, M., Chillaron, J., Mora, C., Castello, A., Bertran, J., Camps, M.,Testar, X., Vilaro, S., Zorzano, A. and Palacın, M. (1993) rBAT, related toL-cysteine transport, is localized to the microvilli of proximal straighttubules, and its expression is regulated in kidney by development. J. Biol.Chem., 268, 27060–27068.



slc7a9-deficient mice develop cystinuria non-i and cystine urolithiasis

Documents