J Am Soc Nephrol 12: 2072–2079, 2001 AGXT Mutations and Their Influence on Clinical Heterogeneity of Type 1 Primary Hyperoxaluria

ANTONIO AMOROSO,*† DOROTI PIRULLI,* FIORELLA FLORIAN,‡ DANIELA PUZZER,† MICHELE BONIOTTO,* SERGIO CROVELLA,* SILVIA ZEZLINA,† ANDREA SPANO` ,† GINA MAZZOLA,§ SILVANA SAVOLDI,࿣ CRISTINA FERRETTINI,¶ SILVIA BERUTTI,¶ MICHELE PETRARULO,¶ and MARTINO MARANGELLA¶ *Section of Genetics, Department of Reproductive and Developmental Science, University of Trieste, Trieste, Italy; †Medical Genetics Service, IRCCS Burlo Garofolo, Trieste, Italy; ‡Department of Biology, University of Trieste, Trieste, Italy; §Transplant Immunology Service, Ospedale S. Giovanni Battista di Torino, Torino, Italy; ࿣Division of Nephrology and Dialysis, Azienda Ospedaliera Triestina, Trieste, Italy; and ¶Renal Stones Center, Ospedale Mauriziano Umberto I, Torino, Italy.

Abstract. Primary hyperoxaluria type 1 (PH1) is an autosomal were performed. Both mutant alleles were found in 21 out of recessive disorder that is caused by a deficiency of alanine: 23 patients, and 13 different mutations were recognized in glyoxylate aminotransferase (AGT), which is encoded by a exons 1, 2, 4, and 10. Normalized AGT activity was lower in single copy gene (AGXT). Molecular diagnosis was used in the severe form than in the adult form (P Ͻ 0.05). Double conjunction with clinical, biochemical, and enzymological data heterozygous patients presented a lower age at the onset of the to evaluate genotype-phenotype correlation. Twenty-three un- disease (P ϭ 0.025), and they were more frequent in group A related, Italian PH1 patients were studied, 20 of which were (75%) than in the group B (14%; P ϭ 0.0406). The T444C grouped according to severe form of PH1 (group A), adult mutation was more frequent in the severe form (P Ͻ 0.05), and form (group B), and mild to moderate decrease in renal func- the opposite was observed for G630A (P Ͻ 0.05). G630A tion (group C). All 23 patients were analyzed by using the mutation homozygotes had a higher AGT residual activity (P single-strand conformation polymorphism technique followed ϭ 0.00001). This study confirms the allelic heterogeneity of by the sequencing of the 11 AGXT exons. Relevant chemistries, the AGXT, which could to some extent be responsible for the including plasma, urine and dialyzate oxalate and glycolate phenotypic heterogeneity in PH1. assays, liver AGT activity, and pyridoxine responsiveness,

Primary hyperoxaluria type 1 (PH1; OMIM: 259900) is a rare progressive storage of CaOx crystals, causing a syndrome autosomal recessive disorder that is characterized by the im- referred to as systemic oxalosis (5). paired hepatic detoxification of glyoxylate. It is caused by a The clinical setting of the disease is highly heterogeneous deficiency of alanine:glyoxylate aminotransferase (AGT; EC with respect to age at onset, type of presentation, severity of 2.6.1.44), which catalyzes the transamination of glyoxylate to hyperoxaluria, residual enzymatic activity, and progression to glycine (1). This disorder leads to the endogenous overproduc- renal insufficiency (6). Most patients suffered from recurrent tion of oxalate and glycolate, resulting in oxalic and glycolic episodes of nephrolithiasis in childhood or adolescence; the hyperacidurias, which are the hallmarks of the disease (2,3). infantile form of oxalosis is diagnosed only in few cases, a The pathophysiology of hyperoxaluria is a result of the low factor that often leads to death for renal failure during the first degree of solubility of the calcium oxalate (CaOx) that pro- months of life. An increasing number of patients are diagnosed duces urolithiasis and/or nephrocalcinosis, often leading to only in adulthood, usually after a long-standing history of renal failure. The loss of the renal function generally increases recurrent nephrolithiasis and sometimes after starting dialysis the oxalate plasma levels, and it induces calcium oxalate over- treatment or after a kidney transplant (6). saturation in the body fluids (4). In many tissues, this causes a The AGT is encoded by a single copy gene (AGXT), consisting of 11 exons, ranging from 65 bp to 407 bp, and spanning over a 10 Kb DNA segment in the 2q37.3 human Received April 21, 2000. Accepted April 26, 2001. region (7). AGT is a 392-amino acid protein with a molecular Correspondence to Dr. Antonio Amoroso, Servizio di Genetica—IRCCS Burlo weight of 43 kD (8). From the cytosol, where it has a homo- Garofolo, Via dell’Istria 65/1 - 34137 Trieste - Italy. Phone: ϩ39 040 3785275; Fax: ϩ39 040 3785210; E-mail: [email protected] dimeric structure, it is then imported into the peroxisomes, 1046-6673/1210-2072 where it detoxifies the glyoxylate by using pyridoxal-5-phos- Journal of the American Society of Nephrology phate as a cofactor (1). Copyright © 2001 by the American Society of Nephrology Although the enzyme is normally located in the peroxi- J Am Soc Nephrol 12: 2072–2079, 2001 Genotype-Phenotype in Primary Hyperoxaluria Type 1 2073 somes, a small AGT amount (5%) can also be found in the removal of the same chemistries were measured in the other patients mitochondria (9), where the enzyme does not function (6). on renal replacement therapy. So far, seven polymorphisms and 35 mutations have been The results belonging to patients who were suspected of having identified in the AGXT gene by using several technical ap- PH1 (in any of the above groups) were compared with reference proaches (10–22). values belonging to normal individuals or patients undergoing dialysis for oxalosis-unrelated nephropathies. PH1 was, therefore, diagnosed The disease is caused by homozygous point mutations or by when both the levels of oxalate and of glycolate were higher than the compound heterozygous in two different point mutations. highest reference range (Table 1). Polymorphic sequences have also been proven to combine with We then analyzed the response of 20 patients to pyridoxine by two haplotypes that were identified in the normal white pop- repeating the same assays after a 10 to 30 mg/kg per d pyridoxine ulation. The major (80% frequency) and the minor (20% fre- treatment that lasted 1 month. We arbitrarily defined the positive quency) haplotype present a combination of three polymor- response as a normalization or a 50% decrease in both oxalate and phisms: 74 bp duplication within the first intron (12,18) and glycolate levels in plasma and urine. C154T and A1142G point mutations, which specify Pro11Leu Liver biopsies were performed to test the AGT and gamma-glu- and Ile340Met amino acid substitutions, respectively (9). The tamyl (GGT) activity. At the time of biopsy, the patients normal minor haplotype is responsible for a 5% mistargeting of had been off the pyridoxine treatment for an entire month. Clinical the AGT to the mitochondria; the mistargeting then rises to and biochemical details are summarized in Table 2. 90% when the minor haplotype is associated to G630A (Gly170Arg amino acid substitution) (19,23). Biochemical Procedures In this article, we report the clinical findings and the AGXT The levels of oxalate and glycolate in plasma, urine, and dialysis mutations of 23 unrelated Italian PH1 patients. To establish fluids were determined by HPLC procedures, as specified elsewhere genotype-phenotype correlations, we used a molecular diagno- (24). Specimens from hepatic biopsies were analyzed for AGT and sis associated to clinical, biochemical, and enzymological data. GGT by means of a HPLC-based microassay (25). The GFR mea- surements, relating to patients with maintained renal function, were Materials and Methods carried out contemporaneously with the plasma and urine assaying for oxalate, glycolate, and L-glycerate during the last visit to our center. Patients As outlined previously, the reference ranges derive from the values of In this study, we considered 23 unrelated patients from 11 different normal individuals or patients undergoing dialysis for oxalosis-unre- Italian regions; 14 were men, and 9 were women. Ages ranged from lated nephropathies; these findings were published elsewhere and are Ϯ 1 to 50 yr (mean, 25.3 16.6 yr). Some of the relatives of 10 patients summarized in Table 1 (26,27). were also analyzed for the genetic study. Eight of the 23 patients had a normal or only mildly reduced renal function, 7 were on regular dialysis treatment, 3 received a kidney graft, and 5 had undergone a Molecular Approach combined liver-kidney transplant. The DNA extraction and the in vitro amplification were performed Based on the age at onset, 1 patient had an infantile oxalosis, 7 had as described previously (28). Initially, the PCR products were ana- an adult form, and the disease was diagnosed during adolescence in 15 lyzed by using the technique of the single-strand conformation poly- patients. According to the clinical data, 4 patients had renal failure morphism (SSCP) under standard conditions (29). The samples show- associated to nephrocalcinosis, whereas 19 had recurrent nephrolithi- ing an abnormal electrophoretic pattern were analyzed by direct asis. When analyzing the renal function among our 23 patients, we sequencing of both strands using the BigDye Terminator Cycle Se- saw that 8 patients were diagnosed with PH1 after they had started quencing Kit (PE Biosystems, Foster City, CA) on an automated ABI dialysis and that 2 were diagnosed after a kidney transplantation. For PRISM 310 Sequencer (PE Biosystems) (10). For the unrecognized eight patients, those with maintained renal function, the classification mutation, we performed a mutation analysis using direct sequencing of PH1 in terms of clinical data and diagnosis was made by assaying together with SSCP. For patient 20 and patient 23, we performed the plasma and the urine for oxalate, glycolate, and L-glycerate in at direct sequencing of all 11 AGXT exons. The co-inheritance of the least two separate sample collections. Plasma levels and dialysis mutations with the major or minor haplotype was checked by ampli-

Table 1. Means and reference ranges for normal individuals or patients on regular hemodialysis treatment for oxalosis- unrelated renal failure (25,26)

POX PGLY UOX UGLY Parameter (␮mol/l) (␮mol/l) (␮mol/mmol Cr) (␮mol/mmol Cr)

Normal individuals mean (ϮSD) 2.4 Ϯ 0.6 7.9 Ϯ 2.4 30 Ϯ 16 42 Ϯ 14 range 1–43–10 12–55 25–76 Oxalosis-unrelated renal failure mean (ϮSD) 47.4 Ϯ 11.4 6.6 Ϯ 2.8 ND ND range 22–60 4–10 ND ND

POX, plasma oxalate; PGLY, plasma glycolate; UOX, urine oxalate ␮mol/mmol creatinine; UGLY, urine glycolate ␮mol/mmol creatinine; ND, not determined. 2074 Journal of the American Society of Nephrology J Am Soc Nephrol 12: 2072–2079, 2001

Table 2. Clinical and biochemical findings of PH1 patients, subdivided according to the different phenotypic groups

Group PT S/A C.D ESRF GFR POX PGLY UOX UGLY %AGT GGT PyrR

Group A 1 6/16 NL 15 TxLK 116 597 ND ND 10 0.72 No 3 6/10 NC 10 TxLK 78 377 ND ND 8 0.77 No 5 8/19 NL 16 TxLK 253 420 ND ND 9 1 No 10 1/20 NL 9 TxLK 37 327 ND ND 0 0.65 No 11 1/6 NC 1 TxLK 260 297 ND ND 0 ND ND 16 1/15 NL 7 HD 222 388 ND ND 0 0.54 No 18 4/4 NC 4 HD 167 498 ND ND 0 0.21 No 20 1/10 NL 9 HD 169 201 ND ND 2 0.3 ND Group B 2 16/52 NL 35 HD 165 41.3 ND ND 12 0.84 Yes 6 20/46 NL 30 TxK 16 43 ND ND 34 1.02 Yes 13 31/39 NL 31 HD 166 63 ND ND 7 0.9 ND 14 40/54 NL 45 TxK 63 145 ND ND 50 1.7 Yes 17 49/55 NL 50 HD 227 581 ND ND 4 0.84 Yes 19 22/50 NC 49 HD 147 163 ND ND 34 0.55 Yes 21 13/33 NL 29 TxK 80 177 365 195 10 0.4 Yes Group C 4 4/20 NL — 80 8 138 169 478 5 2.05 No 9 6/22 NL — 40 8 103 240 258 10 0.6 Partial 12 20/30 NL — 100 21 36 92 179 40 0.7 Yes 22 8/29 NC — 77 22 56 161 32 3.6 0.72 Yes 23 12/22 NL — 54 9.3 157 144 76 ND ND Yes Group D 7 4/13 NL — 80 15 159 319 320 3 0.69 No 8 1/5 NL — 80 7 58 534 536 58 1.4 Yes 15 8/12 NL — 70 15.3 186 161 458 15 1.24 Partial

PH1, Primary hyperoxaluria type 1; PT, patient; S/A, age at onset/present age (yr); ESRF, age at onset of end stage renal failure; CD, clinical details; NL, nephrolitiasis; NC, nephrocalcinosis; GFR, glomerular filtration rate (ml/min/1.73m2); TxK, kidney transplant; TxLK, kidney-liver transplant; HD, hemodialysis; POX, plasma oxalate (␮mol/l); PGLY, plasma glycolate (␮mol/l); UOX, urine oxalate ␮mol/ mmol creatinine; UGLY, urine glycolate ␮mol/mmol creatinine; %AGT, normalized alanine:glyoxylate aminotransferase activity; GGT, gamma-glutamyl transferase activity (normal value, 0.7–1.2 ␮mol/h per mg protein); PyrR, response to pyridoxine; ND, not determined; Group A, onset before 10 yr and ESRF before 20 yr; Group B, onset after 10 yr and ESRF after 20 yr; Group C, early or late onset with maintained renal function over 20 yr; Group D, patients not included in the groups A, B, or C because of short follow-up.

fication and restriction analysis for C154T and A1142G polymor- Statistical Analyses phisms and by amplification and agarose gel electrophoresis for the Results were analyzed by the ␹2 test, Fisher’s exact test, t test for 74bp duplication in intron 1, as previously reported (12,23). unpaired data, and simple regression analysis where appropriate. Twenty-two PH1 patients and 25 unrelated controls were analyzed Gene frequencies were calculated by direct gene counting, and for the AGXT flanking markers D2S2285 and D2S125, which are 2.5 differences in frequencies were analyzed by the ␹2 test or Fisher’s cM distant, and for D2S140, which is located far away from the exact test using 2 ϫ 2 contingency tables. former two (5.5 and 3.0 cM, respectively). Microsatellites D2S2285 Frequency of haplotypes and linkage disequilibria from population (using 5'6fam-AGGACCACCTCGTTGC and 5'ATGGCTGTGAAT- were calculated using the Arlequin 2.0 software. This software al- GCCTG primers) and D2S140 (using 5'6fam-GCTACAATGATTTC- lowed us to calculate haplotype frequencies when gametic phases CAAAGTC and 5'GTTGTCCCATACTGATCTTACC) were ampli- were unknown, using the expectation maximization algorithm (30). fied in a 50-␮l final volume by using 50 ng of DNA, 5 pmol of each primer, and1UofAmpliTaq GOLD DNA polymerase (PE Biosys- Results tems) in the GeneAmp PCR System 2400 (PE Biosystems). D2S125 Correlation between Phenotype and Biochemical Data was amplified by using Linkage mapping set version 2 (PE Biosys- We attempted a phenotypic classification of the patients tem) in a 7.5-␮l final volume. We ran the amplified products on an according to clinical course and biochemical parameters; only automated sequencer (ABI PRISM 310, PE Biosystems) and analyzed eight patients with residual renal function could be analyzed them by using Genescan 2.0.2 software and Genescan-400 Rox as size for biochemical metabolites (the others were on renal replace- standard (PE Biosystems). ment therapy). When analyzing the residual AGT activity and J Am Soc Nephrol 12: 2072–2079, 2001 Genotype-Phenotype in Primary Hyperoxaluria Type 1 2075 pyridoxine responsiveness, apparently nonrelated to renal doxine; two had Ͻ10% AGT activity (one responded to pyri- function, we obtained a broader classification of the clinical doxine, and the test was not performed for the other). In setting. patients with conserved renal function (group C), the behavior By resorting to this latter classification, 20 out of 23 patients varied according to cases: one had residual AGT activity and were grouped as follows: group A, 8 patients presenting the an excellent response to pyridoxine, two had a very low AGT most severe form with early onset (before the age of 10 yr) and (one with and the other without a response to pyridoxine), and progression to end-stage renal failure (ESRF) before the age of one had an intermediate AGT activity with a partial response to 20 yr; group B, 7 patients presenting milder course with onset pyridoxine. after 10 yr of age and progression to ESRF after the age of 20 In groups A, B, and C, the mean values (ϮSD) of normal- yr; and group C, 5 patients aged over 20 yr with maintained ized AGT activity were 3.6 Ϯ 4.5, 21.6 Ϯ 17.6, and 14.6 Ϯ renal function (Table 2). The three patients excluded from this 17.1, respectively. The median values (range) were 1 (0–10), classification were young individuals (one with an actual age 12 (4–50), and 7.5 (4–40), respectively. Ͻ10 yr and two aged Ͻ20 yr) whose onset of PH1 occurred The mean AGT value was significantly lower in group A before the age of 10 yr but whose renal function was preserved. than in the other groups (P Ͻ 0.05). All patients in group B, These patients cannot be included in the previous groups three of group C, and none of group A responded to the because of the shorter follow-up. These three patients were pyridoxine test. The response was associated to some residual analyzed only for genetic studies and were not included in AGT activity (except for patient 17). The mean values (ϮSD) genotype-phenotype correlation. of normalized AGT activity were significantly higher in pa- The patients with stable renal function had a mean GFR of tients who were responsive to pyridoxine than in the non- 72.6 Ϯ 18.3 ml/min per 1.73 m2 body surface area (range, 40 responsive ones (23.1 Ϯ 19.1 versus 3.7 Ϯ 4.0; P ϭ 0.005). to 100) over 11.3 Ϯ 6.0 yr of follow-up (range, 4 to 21 yr). The patients with severe form (group A) had poor or no Mutation Typing of Italian PH1 Patients residual AGT activity (six out of eight were tested with pyri- PH1 was investigated at a molecular level by means of in doxine, and none were responsive). Among patients with less vitro amplification, SSCP analysis, and DNA sequencing. The severe form of PH1 (group B), five had a significant residual mutations found in the unrelated Italian PH1 patients are sum- AGT activity (Ͼ10%) and gave a positive response to pyri- marized in Table 3. We successfully characterized both mutant

Table 3. Patients genotypes for AGXT mutations, minor (m) and major (M) haplotypes, and AGXT-linked markers

Patient AGXT Mutations (exons) Haplotype D2S2285 D2S125 D2S140

1 GAG408ins(2) /T576A(4) Mm ND 98/100 154/154 2 G244T(1) /C252T(1) MM 250/252 90/92 150/154 3 T444C(2) /G1098del(10) Mm 252/252 90/94 150/154 4 G1098del(10) /G1098del(10) MM 252/258 92/100 154/154 5 G588A(4) /G588A(4) MM ND ND ND 6 G630A(4) /G630A(4) mm 250/252 90/92 148/148 7 C155del(1) /C155del(1) MM 252/256 86/86 154/154 8 G630A(4) /G630A(4) mm 252/252 92/92 148/148 9 G588A(4) /G588A(4) MM 252/252 ND 156/156 10 G468A(2) /C156ins(1) MM* 250/254 92/94 154/154 11 G630A(4) /G640A(4) Mm 252/258 88/90 150/150 12 G630A(4) /G630A(4) mm 252/252 90/92 148/148 13 G243A(1) /G243A(1) MM 258/258 96/96 150/150 14 G630A(4) /G630A(4) mm 252/252 ND 148/148 15 C156ins(1) /C156ins(1) MM 250/252 ND ND 16 C156ins(1) /T444C(2) Mm 252/258 90/94 150/154 17 G1098del(10) /G1098del(10) MM 250/250 88/88 ND 18 T444C(2) /T444C(2) mm 252/252 90/90 ND 19 G630A(4) /G630A(4) mm ND 92/92 148/148 20 G468A(2) /X° Mm 252/258 ND 156/158 21 C156ins(1) /C156ins(1) MM 252/252 96/96 154/154 22 C252T(1) /X Mm 252/256 88/94 150/150 23 T444C(2) /G588A(4) Mm 250/252 94/100 148/158

ND, not determined; M*, major haplotype with an unusual combination of G1142 polymorphism with C154T and without the insertion of 74bp in the first intron; X°, undetected AGXT mutation. 2076 Journal of the American Society of Nephrology J Am Soc Nephrol 12: 2072–2079, 2001

Table 4. Frequencies of AGXT mutations in patients with different PH1 phenotype

Gene Frequenciesa Mutation of AGXT Gene Group A (n ϭ 8) Group B (n ϭ 7) Group C (n ϭ 5) Total (n ϭ 23)

T444C 0.25b 0.06 0.108 G468A 0.125 0.043 C156ins 0.125 0.142 0.13 G588A 0.125 0.187 0.108 T576A 0.06 0.0217 G640A 0.06 0.0217 GAG408ins 0.06 0.0217 G630A 0.06c 0.428 0.125 0.239 G1098del 0.06 0.143 0.125 0.108 G243A 0.143 0.043 G244T 0.071 0.0217 C252T 0.071 0.0625 0.043 C155del 0.043 Xd 0.06 0.0625 0.043

a Gene frequency calculated by direct gene counting; b group A versus B ϩ C(P Ͻ 0.05); c group A versus B ϩ C(P Ͻ 0.05); d uncharacterized mutation. alleles in 21 patients, and only one mutation was identified in Genotype-Phenotype Correlation two patients. During this study, we identified 13 different In group A, 75% of patients (six out of eight) were double mutations in exons 1, 2, 4, and 10, which had been partially heterozygous for two different mutations that were carried in described in a previous report (10). the majority of them (five out of six) by two different exons. The most frequent mutation carried by our patients was the Homozygous genotypes were found only in two patients: in G630A with a gene frequency equal to 0.239. The second most patient 5 (carrying mutation G588A) and in patient 18 (homo- frequent mutation was C156 ins (gene frequency, 0.13). The zygous for the T444C mutation). It is worth noticing that mutations G1098del, T444C, and G588A showed an 11% gene another patient (patient 9) presented the same genotype as frequency. Finally, the mutations G468A and C252T were patient 5, but with a milder course. Five out of six heterozy- present twice in a heterozygous condition. All the others were gous patients carried mutation in different exons. found only in one patient in homozygous (G243A and In group B, 86% (six out of seven) of the patients were C155del) or heterozygous condition (T576A, G640A, G244T, homozygous for AGXT mutations. Only patient 2 was double GAG408ins). Family studies were carried out to confirm that heterozygous for mutations G244T and C252T that were in- the parents were heterozygous for the mutations present in their terestingly located in the same exon. The frequency of homo- offspring. zygous mutations was significantly lower (P Fisher ϭ 0.0406) Twenty-two patients were also typed for minor and major in group A than in group B. The genotype carrying both AGXT haplotypes as well as for DS2285, D2S125, and D2S140 mi- mutations affecting a single exon was significantly more fre- crosatellites, as reported in Table 3. quent in group B than in group A patients (100% versus 37.5%; The G630A and T444C mutations were always inherited P Fisher ϭ 0.0256). The age at onset of PH1 was significantly with the minor haplotype, which was also co-inherited with the Ϯ mutations T576A in patient 1. Both patient 20 and patient 23, lower in nine heterozygous patients (5.8 5.4 yr) than in the Ϯ ϭ heterozygotes for the unknown mutations, were also heterozy- 14 homozygous ones (16.4 14.8 yr; P 0.025). gous for the minus haplotypes. The overall frequency of the Microsatellite analysis showed that patients and controls minor haplotype was 43.4%. were not significantly different for the homozygosity rate oc- No preferential gametic association was found between curring at loci D2S2285 (52% and 38%, respectively), D2S125 AGXT mutations and polymorphisms at D2S2285, D2S125, (47% and 23%, respectively), and D2S140 (50% and 28%, and D2S140 loci. No significant differences were found among respectively). Almost half (47.6%) of the patients and almost patients and controls for allelic and haplotype frequencies. one third (28.6%) of the controls were homozygotes for at least Frequency of the allele 252 at locus D2S2285 was slightly two of these loci. However, when patients were subdivided higher in patients than in controls (0.548 versus 30.9; P ϭ according to AGXT genotype, 83% (10 of 12) of the AGXT 0.047), but the opposite was true for the allele 94 at D2S125 homozygous patients and none of the AGXT heterozygous (0.132 versus 0.357; P ϭ 0.0386). These differences, however, patients were also homozygous for at least two of the three were no longer statistically significant after the correction of linked microsatellites (P Fisher ϭ 0.0002). the P values for the number of comparisons made. Mutations T576A, G640A, G468A, and GAG408ins were J Am Soc Nephrol 12: 2072–2079, 2001 Genotype-Phenotype in Primary Hyperoxaluria Type 1 2077 typical only in patients with the most severe form, whereas homozygous rate (32), which became 22% in our study. Nev- G243A, G244T, and C252T were exclusive of the mild form. ertheless, rare mutations have already been described in the Mutations G630A, C156ins, G1098del, T444C, and G588A literature only in homozygous form (15,33). Therefore, to were distributed throughout the different groups. The allele assess the validity of our results, excluding a recently reported frequencies of the AGXT mutations in groups A, B, and C are rare deletion (22), we checked the patients’ parents and rela- reported in Table 4. The gene frequency of T444C mutation tives in 10 families. This analysis confirmed that they were was significantly higher in group A patients (25%) than in the heterozygous for the mutation carried by patients. group B (4.2%; P ϭ 0.05). Instead, the G630A was signifi- The great molecular heterogeneity of PH1 may be the reason cantly more frequent in groups B and C than in group A (P Ͻ why it is so difficult to correlate phenotype, biochemistry, and 0.05). Comparing the normalized AGT activity and the AGXT residual AGT activity to clinical severity. There are also other mutations, a residual AGT activity Ͼ30% was found only in factors, not of genetic origin, that might influence the clinical the five patients homozygous for the G630A mutation and course of the disease, including diet, infections, and climate. never in patients with the other genotypes (P Fisher ϭ Moreover, there are other genetic factors that might control the 0.00001). catalytic activity and the subcellular targeting of glyoxylate Worthy of notice is that mutation C155del, which belonged metabolism (D-amino acid oxidase, glycolate oxidase, GGT, exclusively to patient 7 in homozygous form, dramatically GR, LDH) (33,34,35). changed the gene function by shifting the open reading frame Early recognition with a thorough diagnostic work-up, close and creating a stop codon after amino acid 166. What is medical follow-up, and effective prevention of stone events remarkable is that this patient had a GFR of 80 ml/min at 13 yr can influence the outcome of PH1 (36). Patients with adult of age and after a 9-yr follow-up. form (group B) had unfortunately been found to be affected Identical genotypes occurred in two pairs of patients: two late in the course of the disease. All those who could be tested were homozygous for the C156ins (patient 15 and patient 21) for pyridoxine exhibited a clear-cut response to this drug, and and two for the G588A (patient 5 and patient 9); despite it seems particularly reasonable to hypothesize that if PH1 is sharing very similar residual AGT activities, response to pyr- managed early enough and adequately, it would avoid progres- idoxine and clinical course were quite different. sion to end-stage renal disease for many. Within the same mutation, the progression itself can mani- Discussion fest quite differently, as, for example, happened in patient 5 of SSCP analysis and DNA sequencing are both functional group A and patient 9 of group C, who had quite different approaches to the detection of mutations involved in PH1. Both progressions despite the same genotype (G588A/G588A) and mutated alleles were identified in 21 out of 23 unrelated Italian quite similar amounts of residual AGT. It is tempting to sug- PH1 patients. We found a range of 13 different mutations, gest that this occurred in patient 5, who progressed rapidly to confirming the high degree of genetic heterogeneity in this ESRF, because he was diagnosed late after the onset of renal disease. insufficiency and had virtually no medical follow-up. Instead, With a gene frequency of 23.9%, the G630A mutation was in patient 9, in whom renal function was preserved until the age confirmed as the most frequent in Italian patients, as previously of 22, the disease had been recognized at an early stage and had reported in other populations (5,31). This may also explain the been treated with medications, including long-term follow-ups, high frequency found in the Italian patients of the minor pyridoxine supplementation, potassium citrate, and high fluid haplotype because of its well-known linkage disequilibrium intake. with the G630A mutation. The C156ins ranked second, with an This contention actually applies to some of the patients in allele frequency of 13% (Table 4). The G630A and G1098del group C (patient 4 and patient 22) who, despite having neither mutations were the only ones common to groups A, B, and C. residual AGT nor response to pyridoxine, were doing fairly Although only the parents of patient 7 were consanguineous well, and their renal function was preserved over a Ͼ15-yr (i.e. first cousins), we observed a very high frequency of follow-up. homozygous mutations. Sixty-one percent of the patients were It will also be of interest to compare the long-term outcomes homozygous. The microsatellite analysis confirmed that these under medical management of the three patients aged Ͻ20 yr, patients were frequently homozygous for the enlarged genomic presenting different genotypes, residual AGT, and response to region encompassing D2S2285 and D2S140 loci. pyridoxine. For instance, will patient 15 progress to ESRF in a One explanation of this phenomenon could be the small way that is similar to patient 21 because they had the same number of the analyzed sample and/or a high inbreeding rate in genotype? the populations to which these families belong. Most of these The highest values of catalytic activity were found in pa- patients had been re-rooted to us from different Italian centers, tients with the G630A mutation, which causes a Gly170Arg and the information on their origin seems to exclude their substitution that, if combined with the C154T polymorphism belonging to close genetic isolates. Haplotype analysis failed to (Pro11Leu substitution), is responsible for the mistargeting of show a founder effect for Italian PH1 patients, in fact no the AGT enzyme from the peroxisomes to the mitochondria, preferential microsatellite haplotypes were found for the AGXT where the enzyme cannot work properly (9). mutations. The preservation of residual activity with less severe clinical The most frequent mutation (G630A) had a reported 10% manifestations, in the patients with very early protein trunca- 2078 Journal of the American Society of Nephrology J Am Soc Nephrol 12: 2072–2079, 2001 tion, may depend on the mechanism of skipping constitutive References exons. Nonsense mutations are supposed to either reduce the 1. Danpure CJ, Jennings PR: Peroxisomal alanine:glyoxylate ami- amount of mutant allele transcripts (because of the process of notransferase deficiency in primary hyperoxaluria type 1. FEBS nonsense-mediated mRNA decay) or to generate a peptide Lett 201: 20–24, 1986 truncation at the carboxyl end. A mature mRNA deleted in the 2. Danpure CJ: Recent advances in the understanding, diagnosis and treatment of primary hyperoxaluria type 1. J Inherit Metab entire exon carrying the mutation but in frame was observed in ␦ Dis 12: 210–224, 1989 several mutant (such as fibrillin gene, ornithine -ami- 3. Danpure CJ, Rumsby G: Enzymology and molecular genetics of ␤ notransferase, coagulation factor VIII gene, -globin gene, and primary hyperoxaluria type 1. In: Consequences for Clinical distrophin gene) (37,38,39). The RNA messenger lacking ex- Management in Calcium Oxalate in Biological Systems, edited on(s) containing the nonsense codon normally terminate the by Khan SR, Boca Raton, CRC Press, 1995, pp 189–205 translation and are more stable than the properly spliced 4. Marangella M, Cosseddu D, Petrarulo M, Vitale C, Linari F: mRNA with the nonsense codon. Thresholds of serum calcium oxalate supersaturation in relation Our studies suggest that this phenomenon may occur in the to renal function in patients with or without primary hyperox- case of frame-shift mutations that cause a termination codon, aluria: Nephrol Dial Transplant 8: 1333–1337, 1993 5. Watts RWE: The clinical spectrum of the primary hyperoxalurias such as the C155del, C156ins, and G1098del. The homozygous and their treatment. J Nephrol 11[Suppl 1]: S4–S7, 1998 condition for these mutations could be responsible for a smaller 6. Danpure CJ, Jennings PR, Fryer P, Purdue PE, Allsop J: Primary but functional protein, which leads to milder clinical manifes- hyperoxaluria type 1: Genotypic and phenotypic heterogeneity. tations. We could hypothesize that the clinical symptoms ap- J Inherit Metab Dis 17: 487–499, 1994 pear to be more severe when a mutated subunit protein dimer- 7. Lu-Kuo J, Ward DC, Spritz RA: Fluorescence in situ hybridiza- izes with a subunit that bears a different mutation, especially tion mapping of 25 markers on distal human 2q when codified by a different exon. However, other studies, surrounding the human Waardenburg Syndrome, type I (WS1) such as mRNA length experiments, should be conducted to locus (PAX3 gene). Genomics 16: 173–179, 1993 8. Nishiyama K, Funai T, Yokota S, Ichiyama A: ATP-dependent support such a statement. degradation of a mutant serine:pyruvate/alanine:glyoxylate ami- Several molecular aspects of PH1 are still unresolved, in- notransferase in primary hyperoxaluria type 1 case. J Cell Biol cluding the high degree of homozygous mutations and the 123: 1237–1248, 1993 mechanisms responsible for reduced catalytic activity of the 9. Purdue PE, Takada Y, Danpure CJ: Identification of mutation enzyme. This study allowed us to identify the AGXT mutations associated with peroxisome-to-mitochondrion mistargeting of that appeared in the majority of analyzed patients, thereby alanine:glyoxylate aminotransferase in primary hyperoxaluria providing a targeted prenatal diagnosis using DNA analysis in type 1. J Cell Biol 111: 2341–2351, 1990 chorionic biopsies or amniotic fluid. This approach eliminates 10. Pirulli D, Puzzer D, Ferri L, Crovella S, Amoroso A, Ferrettini C, the need for linkage analysis with polymorphic markers, both Petrarulo M, Marangella M, Florian F: Molecular analysis of Hyperoxaluria type 1 Italian patients reveals eight new mutations intragenic or located near the gene (40,41). A better under- in the alanine:glyoxylate aminotransferase gene. Hum Genet 104: standing of the molecular basis of PH1 and its impact on 523–525, 1999 phenotypic presentation could be achieved by the creation of a 11. Nishiyama K, Funai T, Katafuchi R, Hattori F, Onoyama K, European Registry designed to collect data from several cen- Ichiyama A: Primary hyperoxaluria type 1 due to point mutation ters. It is reasonable to foresee that this will improve the of T to C in the coding region of the serine:pyruvate aminotrans- clinical management of PH1 patients, the diagnostic tools, and ferase gene. Biochem Biophys Res Commun 176: 1093–1099, medical intervention. 1991 12. Purdue PE, Lumb MJ, Allsop J, Danpure CJ: An intronic dupli- cation in the alanine:glyoxylate aminotransferase gene facilitates Acknowledgments identification of mutations in compound heterozygote patients with primary hyperoxaluria type 1. Hum Genet 87: 394–396, This study was partly supported by 40% and 60% projects of 1991 MURST and of the Ministry of Health (grant RC: 19/99). Doroti 13. Minatogawa Y, Tone, Allsop J, Purdue PE, Takada Y, Danpure Pirulli is the recipient of a long-term fellowship from University of CJ, Kido R: A serine-to-phenylalanine substitution leads to loss Trieste. Daniela Puzzer, Michele Boniotto, Silvia Zezlina, and Andrea of alanine:glyoxylate aminotransferase catalytic activity and im- Spano`are the recipients of a fellowship from IRCCS Burlo Garofolo, munoreactivity in a patient with primary hyperoxaluria type 1. Trieste. Hum Mol Genet 1: 643–644, 1992 We are indebted to the following Italian Nephrology Centers that 14. Purdue PE, Lumb MJ, Allsop J, Minatogawa Y, Danpure CJ: A referred PH1 patients: Ospedale Regionale, Aosta; Regina Margherita glycine-to-glutamate substitution abolishes alanine:glyoxylate and San Giovanni Battista, Torino; Gaslini, Genova; University Hos- aminotransferase catalytic activity in a subset of patients with pital, Padova; Provinciale, Trento; Pediatric Division, Bambin Gesu`, primary hyperoxaluria type 1. Genomics 13: 215–218, 1992 Roma; Ospedale Acquaviva delle Fonti; Ospedale dei Bambini G. di 15. Danpure CJ, Purdue PE, Fryer P, Griffiths S, Allsop J, Lumb MJ, Cristina, Palermo; Regina Margherita, Messina. Guttridge KM, Jennings PR, Scheinman JI, Mauer SM, Davidson The authors are greatly indebted to Nicole Bolzicco for her collab- NO: Enzymological and mutational analysis of a complex pri- oration in the preparation of this manuscript and for the linguistic mary hyperoxaluria type 1 phenotype involving alanine:glyoxy- reviews. late aminotransferase peroxisome-to-mitochondrion mistargeting J Am Soc Nephrol 12: 2072–2079, 2001 Genotype-Phenotype in Primary Hyperoxaluria Type 1 2079

and intraperoxisomal aggregation. Am J Hum Genet 53: 417– 29. Sheffield VC, Beck J, Kwitek AE, Sandstrom DW, Stone EM: 432, 1993 The sensitivity of single-strand conformation polymorphism 16. von Schnakenburg C, Rumsby G: Primary hyperoxaluria type 1: analysis for the detection of single base substitutions. Genomics A cluster of new mutations in exon 7 of AGXT gene. J Med Genet 16: 325–332, 1993 34: 489–492, 1997 30. Dempster A, Laird N, Rubin D: Maximum likelihood estimation 17. von Schnakenburg C, Weir T, Rumsby G: Identification of new from incomplete data via the EM algorithm. J Roy Statist Soc 39: mutations in primary hyperoxaluria type 1 (PH1). J Nephrol 1–38, 1977 11[Suppl 1]: S15–S17, 1998 31. Danpure CJ: The molecular basis of alanine:glyoxylate amino- 18. Danpure CJ, Birdsey GM, Rumsby G, Lumb MJ, Purdue PE, transferase mistargeting: the most common single cause of pri- Allsop J: Molecular characterization and clinical use of polymor- mary hyperoxaluria type 1. J Nephrol 11[Suppl 1]: 8–12, 1998 phic tandem repeat in an intron of the human alanine:glyoxylate 32. Lhotta K, Rumsby G, Vogel W, Pernthaler H, Feichtinger H, aminotransferase gene. Hum Genet 94: 55–64, 1994 Koning P: Primary hyperoxaluria type 1 caused by peroxisome- 19. Purdue PE, Allsop J, Isaya G, Rosenberg LE, Danpure CJ: to-mitochondrion mistargeting of alanine:glyoxylate aminotrans- Mistargeting of peroxisomal L- alanine:glyoxylate aminotrans- ferase. Nephrol Dial Transplant 11: 2296–2298, 1996 ferase to mitochondria in primary hyperoxaluria patients depends upon activation of a cryptic mitochondrial targeting sequence by 33. von Schnakenburg C, Hulton SA, Milford DV, Roper HP, a point mutation. Proc Natl Acad Sci USA 88: 10900–10904, Rumsby G: Variable presentation of primary hyperoxaluria type 1991 1 in 2 patients homozygous for a novel combined deletion and 20. Rinat C, Wanders RJ, Drukker A, Halle D, Frishberg Y: Primary insertion mutation in exon 8 of the AGXT gene. Nephron 78: hyperoxaluria type I: A model for multiple mutations in a mo- 485–488, 1998 nogenic disease within a distinct ethnic group. J Am Soc Nephrol. 34. Danpure CJ, Smith LH: The primary hyperoxalurias. In: Kidney 10: 2352–2358, 1999 Stones: Medical and Surgical Management, edited by Coe FL, 21. Basmaison O, Rolland MO, Cochat P, Bozon D: Identification of Favus MJ, Pak CYC, Parks JH, Preminger GM, Philadelphia, 5 novel mutations in the AGXT gene. Hum Mutat 15: 577, 2000 Lippincott-Raven, 1996, pp 859–881 22. Nogueira PK, Vuong TS, Bouton O, Maillard A, Marchand M, 35. Hoppe B, Danpure CJ, Rumsby G Fryer P, Jennings PR, Blau N, Rolland MO, Cochat P, Bozon D: Partial deletion of the AGXT Schublger G, Neuhaus T, Leumann E: A vertical (pseudodomi- gene (EX1_EX7del): A new genotype in hyperoxaluria type 1. nant) pattern of inheritance in the autosomal recessive disease Hum Mutat 15: 384–385, 2000 primary hyperoxaluria type 1: Lack of relationship between 23. Danpure CJ: Primary hyperoxaluria type 1 and peroxisome-to- genotype, enzymic phenotype and disease severity. Am J Kidney mitochondrion mistargeting of alanine:glyoxylate aminotransfer- Dis 29: 36–44, 1997 ase. Biochimie 75: 309–315, 1993 36. Scheinman JI: Primary hyperoxaluria: Therapeutic strategies for 24. Petrarulo M, Vitale C, Facchini P, Marangella M: Biochemical the 90’s. Kidney Int 40: 389–99, 1991 approach to diagnosis and differentiation of primary hyperoxalu- 37. Dietz HC, Valle D, Francomano CA, Kendzior RJ, Pyeritz RE, ria type 1: An update. J Nephrol 11[Suppl 1]: S23–S28, 1998 Cutting GR: The skipping of costitutive exons in vivo induced by 25. Petrarulo M, Pellegrino S, Marangella M, Cosseddu D, Linari F: nonsense mutations. Science 259: 680–683, 1993 High-performance liquid chromatographic microassay for L-ala- 38. Maquat LE: Defects in RNA splicing and the consequence of nine:glyoxylate aminotransferase activity in human liver. Clin shortened translational reading frames. Am J Hum Genet 59: Chim Acta 208: 183–192, 1992 279–286, 1996 26. Marangella M, Petrarulo M, Bianco O, Vitale C, Finocchiaro P, Linari F: Glycolate determination detects type I primary hyper- 39. Nagy E, Maquat LE: A rule for termination-codon position oxaluria in dialysis patients. Kidney Int 39: 149–154, 1991 within intron-containing genes: When nonsense affects RNA 27. Marangella M, Petrarulo M, Vitale C, Cosseddu D, Linari F: abundance. Trends Biochem Sci 23: 198–199, 1998 Plasma and urine glycolate assays for differentiating hyperox- 40. Rumsby G: Experience in prenatal diagnosis of primary hyper- aluria syndromes. J Urol 148: 986–989, 1992 oxaluria type 1. J Nephrol 11[Suppl 1]: S13–S14, 1998 28. Ferrettini C, Pirulli D, Cosseddu D, Marangella M, Petrarulo M, 41. von Schnakenburg C, Weir T, Rumsby G: Linkage of microsat- Mazzola G, Vatta S, Amoroso A: Molecular analysis of the ellites to AGXT gene on the chromosome 2q37.3 and their role in AGXT gene in Italian patients with primary hyperoxaluria type 1. the prenatal diagnosis of primary hyperoxaluria type 1. Ann Hum J Nephrol 11[Suppl 1]: S18–S22, 1998 Genet 61: 365–368, 1997