Genetic heterogeneity in primary hyperoxaluria type 1: impact on diagnosis

Molecular Genetics and Metabolism 83 (2004) 38–46 www.elsevier.com/locate/ymgme

Minireview

Genetic heterogeneity in primary hyperoxaluria type 1: impact on diagnosis Marion B. Coulter-Mackiea,¤, Gill Rumsbyb a

Department of Pediatrics, University of British Columbia, Children’s and Women’s Health Centre of B.C. 4500 Oak Street, Vancouver, BC, Canada V6H 3N1 b UCL Hospitals Clinical Biochemistry, 60 WhitWeld Street, London W1T 4EU, UK Received 24 June 2004; received in revised form 18 August 2004; accepted 20 August 2004

Abstract Primary hyperoxaluria type 1 (PH1) is an autosomal recessive disease characterized by progressive kidney failure due to renal deposition of calcium oxalate. The disease is caused by a deWciency of alanine:glyoxylate aminotransferase (AGT) which catalyzes the conversion of glyoxylate to glycine. When AGT is absent, glyoxylate is converted to oxalate which forms insoluble calcium salts that accumulate in the kidney and other organs. In the most common phenotype there is a unique phenomenon wherein AGT is mistargeted to the mitochondria instead of the peroxisomes. The diagnosis of PH1 is complicated by heterogeneity of clinical presentation, course of the disease, biochemical markers, AGT enzymatic activity and genotype. More than 50 mutations and polymorphisms have been reported in the AGT gene; three common mutations accounting for almost 50% of PH1 alleles. The mutations are of all types, with missense making up the largest fraction. There are some mutations with apparent ethnic associations and at least one that appears to be pan-ethnic. Although correlations can in some cases be made between biochemical phenotype and genotype, correlation with clinical phenotype is complicated by the involvement of other genetic and non-genetic factors that aVect disease severity. A number of polymorphisms have been described in the AGT gene some of which cause missense changes and, in some cases, alter enzyme activity. As DNA testing becomes more commonly used for diagnosis it is important to correlate observed sequence changes with previously documented changes as an aid to assessing their potential signiWcance.  2004 Elsevier Inc. All rights reserved. Keywords: PH1; Hyperoxaluria; AGT; AGXT; Alanin:glyoxylate aminotransferase; Mutation; Review

Introduction Primary hyperoxaluria type 1 (PH1) (MIM 259900) is an autosomal recessive metabolic disease caused by a deWciency of a liver peroxisomal enzyme, alanine:glyoxylate aminotransferase (AGT) (EC 2.6.1.44) [1–3]. AGT catalyzes the conversion of glyoxylate to glycine and in its absence, glyoxylate is converted to oxalate which forms insoluble calcium salts. Patients with PH1 present *

Corresponding author. Fax: +1 604 875 2193. E-mail address: [email protected] (M.B. CoulterMackie). 1096-7192/$ - see front matter  2004 Elsevier Inc. All rights reserved. doi:10.1016/j.ymgme.2004.08.009

with calcium oxalate stones, progressive renal insuYciency culminating in renal failure, and a wide spectrum of symptoms related to systemic oxalosis (calcium oxalate deposition). Although the age of onset can vary, the most common pattern is pediatric onset, usually with nephrolithiasis or nephrocalcinosis, decline in renal function, and progression to end-stage renal disease. A unique pathophysiology found in about one-third of patients confers a residual enzyme activity of up to 60% of mean normal [1,2,4]. In many human metabolic disorders, this level of residual activity would be suYcient for normal function. Most of these PH1 patients, however, exhibit a unique protein targeting defect in which an

M.B. Coulter-Mackie, G. Rumsby / Molecular Genetics and Metabolism 83 (2004) 38–46

otherwise functional AGT enzyme is synthesized in adequate amounts but is mis-targeted to mitochondria instead of to peroxisomes where it is normally found and where the glyoxylate substrate remains. Such patients will have PH1 despite the residual enzyme activity. There are numerous reports that pyridoxine (vitamin B6) administration reduces oxalate excretion and plasma oxalate in 25–50% of cases [5,6]. Responders are usually individuals exhibiting signiWcant residual AGT activity, usually those with a documented mis-targeting phenotype [7,8]. Pyridoxine is an essential co-factor for AGT, binding as pyridoxal phosphate (PLP), although the exact mechanism by which pyridoxine works therapeutically is not established. It has been suggested that supplementation may increase the proportion of AGT that has bound co-factor [9]. The incidence of PH1 ranges from 0.4 to 1 per 105 live births depending on the population studied [1,10]. The prevalence is 1–3 per 106 in the European countries depending on the method of ascertainment [11–13]. An increased frequency has been reported in Middle Eastern countries such as Tunisia, Israel, and Iran most likely due to a higher rate of consanguinity [14–16]. A problem with the diagnosis of PH1 is that the clinical signs and symptoms are not unique to this disease. In most patients the Wrst signs of disease are urolithiasis and nephrocalcinosis or end-stage renal disease with, or without, a history of renal stones or nephrocalcinosis. Urinary and plasma oxalate may be elevated but these do not distinguish PH1 from PH2, glyoxylate reductase (GR) deWciency. PH1 diagnosis can be conWrmed by liver biopsy and AGT enzymatic assay. However, this assay is only available in a limited number of labs. A recent survey suggested that in the USA less than half of hyperoxaluria patients had conWrmatory liver biopsy and enzyme testing [17]. In up to 30% of cases the diagnosis was not conWrmed until patients reached end-stage renal failure.

Heterogeneity in PH1 PH1 exhibits heterogeneity at a number of levels: age of onset, severity, biochemical metabolite indicators, AGT enzymatic activity, and genotype. The clinical presentation of PH1 is heterogeneous with respect to age of onset and course of the disease [1]. The age of onset is predominantly pediatric but about 10% of patients develop a severe, early onset form of PH1 in infancy. Another 10% may not become symptomatic for 40–50 years. The speciWc mutations in the AGT gene (AGXT) are not good predictors of the age of onset or the clinical course [4]. All cases with normal renal function will have elevated urine oxalate at some point during the course of their disease but data arising from family studies shows that this does not necessarily

39

occur from early life [18]. Similarly there is considerable variation in the presence of glycolicaciduria with approximately one-third of patients not showing this abnormality [19,20]. Heterogeneity of residual AGT activity in PH1 is well documented [1]. About one-third of patient have measurable AGT levels ranging as high as 60% of a mean normal value which overlaps with the normal reference range [4]. It may be diYcult, on the basis of enzyme activity alone, to distinguish a patient with high residual activity from low normal or asymptomatic heterozygotes. In all cases with high residual AGT recognized to date, the enzyme had the G170R mutation and is likely to have been mis-targeted to the mitochondria. Since the substrate is produced in the peroxisome, the patient suVers a functional deWciency of AGT activity. Overall, there is a poor correlation between enzyme levels and clinical severity. AGXT mutations are the basic cause of PH1, but molecular heterogeneity cannot fully account for the clinical and biochemical heterogeneity that is observed. It is recognized that the clinical picture can be modiWed by genetic factors aVecting other enzymes in the overall biochemical pathway, dietary precursors, gastrointestinal absorption, enzyme co-factor (B6) availability and other environmental factors.

The normal AGT gene, AGXT The AGT cDNA was Wrst cloned in 1990 [21]. The genomic structure of the gene and its location on chromosome 2 (2q36–37) were reported in 1991 [22]. The entire genomic sequence can be found between nucleotides 976889 and 988682 in a chromosome 2 contig (GenBank NT 005416). There are two normal haplotypes, the major and minor alleles (AGT-Ma and AGTMi) commonly found among the normal general population [23]. These alleles diVer at several polymorphic sites, 32C ! T (P11L), 1020A ! G (I340M), an intron 1 duplication [24], and an intron 4 VNTR [25], as well as three silent sequence variations [22,23,26,27]; Table 1. AGT-Ma and AGT-Mi occur at about 80 and 20%, respectively, in the normal Caucasian population [2]. AGT-Mi occurs at about 2% in Japanese [25]. The most signiWcant sequence change functionally is P11L; the leucine variant encodes an enzyme with approximately 50% activity of the proline form. The P11L change also generates a weak N-terminal mitochondrial targeting signal [23,28]. There is strong linkage disequilibrium between a number of these polymorphisms such that P11L, I340M, and the 74 bp insertion in intron 1 tend to occur together. However, we are now seeing exceptions to this with dissociation of M340 from the minor allele [26], the absence of the 74 bp insertion from some minor alleles [30] and the identiWcation of the minor allele African

40

M.B. Coulter-Mackie, G. Rumsby / Molecular Genetics and Metabolism 83 (2004) 38–46

Table 1 Documented polymorphisms occurring in AGXT Exon/intron

Nucleotide changea

Major/minor

Amino acid aVected

Typeb

Detectionc

Frequencyd (%)

Reference

E1 I1

32C ! T, CCC ! CTC IVS1 + 74 bp insert IVS1 + 74 bp insert 264C ! T, GCC ! GCT 29/32 bp repeat unit Type I 38 copies Type II 32 copies Type III17 copies Type IV 12 copies 654G ! A, TCG ! TCA 976G ! A, GTA ! ATA 1020A ! G, ATA ! ATG 1220C ! A

Mi Mi MiA

P11L

M Dp Dp Sy VNTR

+StyI PCR PCR ¡HaeIII Southern, PCR

15–20 15–20 19f

[23,26,28] [24,26] [29] [23,26]

60, 19e 7, 6e 33, 31e 0, 5e

[1,25,26]

E2 I4

E6 E10 3⬘UTR

A88A

MiA Mi

S218S V326I I340M

Sy M M

¡AciI* ¡RsaI +AvaII +RsaI

3f 15–20 66

[22,26] [29] [23,26] [26,27]

Nucleotide numbering refers to the cDNA sequence where 1 is the Wrst coding nucleotide. Nomenclature based on recommendations of Human Genome Variation Society (http://www.genome.unimelb.edu.au/mdi/mutnomen/checklist.html). b Mutation type: M, missense; Dp, duplication; Sy, synonomous; VNTR, variable number of tandem repeats. c Restriction enzyme: +, mutation creates new site; ¡, mutation destroys site; *, mismatch primer required. d Frequencies quoted for Caucasians unless otherwise indicated. e Frequencies for Japanese. f Frequencies in South African Blacks. a

Variant AGT-MiA, which diVers from AGT-Mi in lacking both P11L and I340M but retaining the insertion in intron 1 [29]. AGT-MiA occurs at about 19% frequency among Blacks from southern Africa [29]; AGT-Mi also segregates in this population although at a lower frequency (3%) than AGT-MiA. Mutations in AGXT Mutations causing PH1 have been identiWed on all three haplotypes; however, the most common, 508G ! A (G170R), occurs on AGT-Mi and is responsible for mis-targeting of AGT to mitochondria [23]. Molecular heterogeneity, i.e., the spectrum of mutations in AGXT that cause PH1 makes a major contribution to the clinical heterogeneity. Functional synergism has been demonstrated between P11L, the most common polymorphism of AGT-Mi, and at least four PH1 mutations including the common mutations, G170R, I244T, and F152I, and the less common, G41R [31]. In vitro studies suggested that L11 exacerbated the eVects of the mutation, whereas on a P11 background, the mutations were milder. There are now more than 50 polymorphisms and mutations identiWed in PH1. These are listed in Tables 1 and 2, respectively. A few mutations (G170R, I244T, F152I, 33_34insC) recur frequently and form the basis of DNA screening panels [4]. A test for AGT-Mi is usually the Wrst step in mutation screening. Another group of mutations has been reported in a few unrelated families. The remainder are apparently family-speciWc private mutations. As data on mutations accumulate it is clear that there are some mutations with

possible ethnic correlations. There may also be potential clustering of mutation in AGXT. Genotype/phenotype correlations A number of attempts have been made to look for genotype/phenotype correlations in PH1 using such categories as age of onset, age of end stage renal failure and pyridoxine-responsiveness. The true age of onset of disease can be diYcult to assess as the presentation can be very variable and non-speciWc including urinary tract infection and metabolic acidosis in addition to the more usual presentations of recurrent kidney stones and nephrocalcinosis. Age of end-stage renal failure is a more easily measured endpoint for the disease. Pyridoxineresponsiveness has been poorly deWned and therefore comparison between patients is diYcult. A protocol has recently been established which deWnes responsiveness [47]; however, it is notable that some of those patients with documented responsiveness have no AGT protein present and thus it suggests that the pyridoxine eVect may be via other enzymes. Homozygotes for severe mutations such as major deletions, and minor deletions or insertions, including 33_34insC,which would be expected to have no immunoreactive protein and no catalytic activity, show a similar spectrum of age of onset. The majority of cases present in childhood, usually less than 10 years of age. However, what is surprising is that in some cases the age of onset can be as late as the fourth decade [4,30,34]. This observation lends support to the role of other genes and/or environmental inXuences in the presentation of the primary hyperoxalurias. As glyoxylate

M.B. Coulter-Mackie, G. Rumsby / Molecular Genetics and Metabolism 83 (2004) 38–46

41

Table 2 Documented PH1 mutations in AGXT Exon/intron

Nucleotide changea

E1

2 T ! C, ATG ! ACG 33_34insC 33delC 121G ! A, GGG ! AGG 122G ! T, GGG ! GTG 130C ! T, CAG ! TAG 198C ! G, TAC ! TAG 227_228insTCACACT 244G ! A, GGA ! AGA 245G ! A, GGA ! GAA 285_286insGAG 322T ! C, TGG ! CGG 331C ! T, CGA ! TGA 336C ! A, GCC ! GAC 346G ! A, GGG ! AGG 454T ! A, TTC ! ATC 466G ! A, GGG ! AGG 508G ! A, GGG ! AGG 518G ! A, TGC ! TAC IVS4-1G ! A (525-1G ! A) 547G ! A, GAT ! AAT 560C ! T, TCC ! TTC 568G ! A, GGG ! AGG 570delG 613T ! C, TCG ! CCG 679_(IVS6+2)delAAGT 697C ! T, CGC ! TGC 698G ! A, CGC ! CAC 731T ! C, ATC ! ACC 738G ! A, TGG ! TGA 744delC 753G ! A, TGG ! TGA IVS7-1G ! C (777-1G ! C) 797_802delCATCAinsACAATCTAG 824_825insAG 844C ! T, CAG ! TAG IVS8+1G ! T (846+1G ! T) IVS8-3C ! G (847-3C ! G) IVS8-1G ! A (847-1G ! C) 865C ! T, CGC ! TGC 886_888delGCG 893T ! C, CTG ! CCG 976delG 983_988delCTGGCT 997A ! T, AGA ! TGA 1007T ! A, GTC ! GAC 1049G ! A, GGC ! GAC 1125_1126delCG 5 kb deletion 7 kb deletion

E2

E4

I4 E5

E6 E6/I6 E7

I7 E8

I8

E9

E10

E11 5⬘UTR to I5 5⬘UTR to I7

Major/minor Amino acid change M1T Ma Ma Mi,Ma Ma Ma Ma Mi Ma Ma Mi Ma MiA Ma Mi Ma Mi

Ma Ma Ma Ma Ma Mi Mi Ma Ma Ma Ma

G41R G41V Q44X Y66X G82R G82E E95_P96insE W108R R111X A112D G116R F152I G156R G170R C173Y D183N S187F G190R S205P R233C R233H I244T W246X W251X

Q282X Ma Ma Ma Mi Mi Ma Ma Ma Mi Ma Ma

R289C 296delA L298P A328-G330delinsD R333X V336D G350D

Typeb Restrictionenzymec Frequencyd Reference M I D M M N N I M M I M N M M M M M M S M M M D M D M M M N D N S I/D I N S S S M D D D N M M D D D

+MwoI*

12%

¡Msp ¡MspI ¡PstI ¡RsaI

R

¡AvaI +NlaIV +BsaMI +MaeIII* +BsaJI +MboI ¡BsaJI ¡MspI* ¡BstNI +EcoRI +BsmI ¡FauI, ¡AciI +MwoI* +SmaI ¡BstXI* ¡AciI ¡AciI +BstX1 +MaeIII ¡HaeIII +DdeI ¡BstN1 +StuI

+MaeIII* +Hpy188I* ¡DraII ¡NarI ¡SacII M ¡RsaI Hpy188I* ¡GsuI +MaeIII

R

R R 11%e R R 1,6.8% 11%e 23–27%

R 2.5% R R

6–9

R

R

11%e

R

[32] [4,33–36] [33] [15,37] [33,36] [33,36] [24] [38] [39] [40,41] [33,36] [30,36,42] [34] [29] [33,36] [4,37] [15,33,36] [23,28] [42] [38] [38] [43] [42] [34] [44] [34,41] [27] [27] [4,27] [27] [34] [34] [34] [45] [32] [32] [34] [34,38] [38] [15] [42] [15] [33,36] [34] [15] [39] [42] [41] [41] [46]

Nucleotide numbering refers to the cDNA sequence where 1 is the Wrst coding nucleotide. Nomenclature based on recommendations of Human Genome Variation Society (http://www.genome.unimelb.edu.au/mdi/mutnomen/checklist.html). b Mutation type: M, missense; N, nonsense; D, deletion; I, insertion; S, splice junction. c Restriction enzyme: +, mutation creates new site; ¡, mutation destroys site; *, mismatch primer required. d Frequency: frequencies quoted where justiWed by numbers of cases studied; R D recurrent, at least 2 unrelated individuals. e These frequencies are quoted for an Italian population. a

appears to be the primary precursor of oxalate, such diVerences may result from individual variation in the activity of other glyoxylate-using enzymes, such as glyoxylate reductase (GR) or glutamate : glyoxylate

aminotransferase which may compensate for the absence of AGT. For example, we know that there is at least a four-fold variation in GR enzyme activity between individuals as determined from the reference

42

M.B. Coulter-Mackie, G. Rumsby / Molecular Genetics and Metabolism 83 (2004) 38–46

range [48]. Alternatively, there may be diVerences in dietary intake of glyoxylate precursors or periods of dehydration that may precipitate disease onset in an individual genetically predisposed. The 508G ! A (G170R) mutation always occurs on the background of the minor allele. In the normal situation, the AGT protein encoded by the minor allele dimerises rapidly and the resulting dimer is unavailable for uptake into the mitochondria, which require unfolded or only loosely folded proteins for import. The G170R mutation acts in concert with P11L to slow dimerization, and the cryptic N-terminal mitochondrial targeting sequence generated by P11L leads to uptake of the protein into the mitochondria [31,49,50]. In vitro, the mutant protein has little residual enzyme activity when expressed in Escherichia coli [31] but the mutant protein does have quite variable activity in human liver biopsies [4]. This Wnding may possibly reXect the mitochondrial content of an individual biopsy and recently we have shown that activity can be within the normal reference range in patients homozygous for this mutation [4]. Again, there is a very wide spectrum of age of onset in patients with this mutation [4,30], not dissimilar to that seen with the other severe mutations. A large number of patients are compound heterozygotes for two diVerent mutations. In these cases, when both mutations encode missense mutations, the AGT dimer will be formed of two diVerent proteins. It is not known what eVect this has and whether one mutation might rescue another if present in a mixed dimer form. One study [30] found that compound heterozygotes had a signiWcantly earlier age of onset than homozygotes, suggesting that this was not the case. It is however, diYcult to draw conclusions from this limited study as some of the mutations would not have produced a protein product anyway. Types of mutations and their expected biochemical eVects Mutations in the human AGT gene are of all types: single nucleotide changes, including missense, nonsense

and splice junction mutations; frameshifts including small insertions, small deletions, and an insertion/deletion; and large deletions. Fig. 1 illustrates the locations of the PH1 mutations on the cDNA map. To date, no mutations have been reported in exon 3 and only one in each of exons 6 and 11. Some clustering of mutations is apparent in exons 5 and 7, and at the 3⬘ end of exon 2. The fraction of total AGXT mutations located within exons 2, 5, and 7 is at least 30% higher than that expected based on exon size and chance alone. Single nucleotide mutations Missense changes comprise the largest single group of mutations, making up about 44% of published mutations. Few of these are associated with speciWc genotype/ phenotype changes although the recent publication of the crystal structure [51] helps to rationalize the eVects of selected amino acid sequence changes. Many of the remaining missense mutations have not had speciWc biochemical phenotypes associated with them except for loss of enzymatic activity. Many will result in mis-folding of the protein product and subsequent elimination by intracellular quality control processes [52]. The AGT amino acid sequence is highly conserved especially among mammalian species. Comparison of seven mammalian protein sequences (GenBank: PX53414, NM 016702, X06357, M84414, Y10727, M84647, X75923) showed 76–89% pairwise sequence identity after adjusting for an extended N-terminal mitochondrial targeting sequence in some species. Seventeen of the 20 amino acids aVected by missense changes are absolutely conserved among the seven mammalian species. Two particularly well-conserved areas are exons 4 and 6. Exon 4, has amino acid sequence identity of 83% and biochemical property conservation of 94% across the species compared. Exon 6 spans the PLP co-factor binding site consensus sequence (amino acids 201–221) common to aminotransferases [53] and critical to the catalytic site. Crystallization studies conWrmed that the lysine at

Fig. 1. cDNA map of AGXT showing documented PH1 mutations. Point mutations are shown above the cDNA. Deletions and insertions are shown below. Clustering of mutations is apparent in exons 2, 5, and 7.

M.B. Coulter-Mackie, G. Rumsby / Molecular Genetics and Metabolism 83 (2004) 38–46

codon 209 is the actual site of the SchiV base with PLP [51]. Interestingly, only one mutation, S205P, has been reported in this highly conserved sequence motif. Nonsense mutations make up another 14.6% of reported mutations. In these cases we would expect there to be either a nonfunctional truncated protein that is degraded or more likely, no protein product translated due to nonsense-mediated mRNA decay [54] depending on the location of the nonsense codon. The splice junction mutations, except for 679_(IVS6+2)delAAGT at the exon 6/intron6 junction, are also point mutations and comprise 8.3% of mutations. Many splice junction mutations result in improperly spliced and unstable mRNA [55]. CpG dinucleotides are recognized as potential mutation hotspots due to deamination of 5mC to form T [55]. The CpG dinucleotide is associated with a high frequency of CT ! TG or CG ! CA changes and can be expected to be involved in 10–50% of mutational events. Of the 38 single base change mutations in the AGT gene, including missense, nonsense, and splice junction mutations, 11 or 29% involve mutations in CpG dinucleotides and nine of these are of the CT ! TG or CG ! CA type. In this regard, mutations in AGT are typical. 508G ! A is by far the most common PH1 mutation among Caucasians [4], involves a CpG dinucleotide and to date has always been found associated with the AGTMi. Its high frequency among Caucasians reXects the frequency of the minor allele haplotype in this group and it is possibly a very old mutation. Given the substantial partial activity associated with 508G ! A and the demonstrated activity of artiWcial constructs of 508G ! A on AGT-Ma [31], we cannot exclude the possibility that the mutation may have occurred more than once without causing disease if on the major allele where mis-targeting and functional synergism are not issues.

43

The two large deletions encompassing major portions of the 5⬘ end of the gene and upstream 5⬘UTR [41,46] would prevent transcription. Ethnic associations The world map in Fig. 2 shows the geographic origin of mutations that may have possible ethnic associations. Only those mutations that, to date, have been reported in at least two unrelated individuals and not in other ethnic groups, are indicated. Those representing a single case may simply be family-speciWc mutations. Some apparent associations may be due to chance in the small group examined and will disappear if additional occurrences are found outside the original population group or if a larger cohort is studied. This data is slow to accumulate due to the rarity of the disease and the problems of diagnosis. No mutations are shown mapped speciWcally to North America since the population is ethnically diverse and mutations have not been documented in the aboriginal populations of North America. The reported frequency of individual mutations is very variable and inXuenced by the size of the study. The 508G ! A mis-targeting mutation is by far the most common in all studies to date ranging from 23 to 27% [4,26,30] among Caucasians. The 731T ! C (I244T) mutation originally described as having a gene frequency of 9% [27] has now fallen to around 6% in a wider patient group [4] but of course will be much higher in populations of Spanish/North African extraction [56] where the mutation appears to have a founder eVect. There are several mutations that appear to recur with signiWcant frequency in the Italian population. The 33_34insC insertion, the most common PH1 mutation on the major allele, is prevalent in all populations studied and has a reported frequency from 12 to 13% [4,30,35].

Deletions and insertions Small deletions and insertions make up 27% of the mutations in the AGT gene. Most of these result in frameshifts and a mis-placed termination codon. We expect a result similar to that seen for nonsense mutations. All the small insertions and deletions occur in regions of multiple repeats of the aVected nucleotide sequence. Such sequences have an elevated frequency of deletions or insertions presumably due to slippage and mis-pairing [55]. The second most common PH1 mutation, 33_34insC occurs in a string of 8C’s and has been found in a variety of ethnic and racial groups [33–35]. A deletion, 33delC, has also been reported at the same site [33]. The potential instability of this repeat sequence may result in multiple mutation events over time so that the mutation is more likely to occur in diVerent populations.

Mutations and diagnosis The large number of mutations in the AGXT gene has meant that up to now the initial diagnosis of PH1 has been made by measurement of AGT enzyme activity in a liver biopsy from symptomatic patients. Molecular genetics, using either mutation or linkage analysis, could then be used to make a diagnosis in other siblings or prenatally. Recently, the eVectiveness of screening for three common mutations (508G ! A, 731T ! C, 33_34insC) in AGXT was evaluated in over 300 patients with liver biopsy-proven disease [4]. The data found that in one-third of patients a molecular diagnosis was possible as they were homozygous or compound heterozygous for one of the three mutations. In a further one-third, only one of the mutations was found and thus there was a high index of suspicion of

44

M.B. Coulter-Mackie, G. Rumsby / Molecular Genetics and Metabolism 83 (2004) 38–46

Fig. 2. World map showing geographic locations of PH1 mutations. Mutations listed are, so far, unique to the geographic area, and are reported in at least two unrelated patients. I244T likely represents a founder eVect in the Spanish/North African population.

PH1 as the patients were all symptomatic for the disease, but a liver biopsy would be required to conWrm diagnosis if liver transplantation were to be considered. In the Wnal one-third, none of the three mutations were present and therefore a liver biopsy was essential to make the diagnosis. This approach has a number of factors to commend it. First, an invasive liver biopsy can be put oV and may be avoided altogether; second, the diagnosis of 508A homozygotes is made at the outset without the possibility of misdiagnosis due to high, but misleading, in vitro activity. Balanced against this is the additional delay that may occur in making a Wnal diagnosis and the attendant anxiety for the family. The alternative strategy, that of sequencing the entire gene, does not always yield complete information in all patients [35]. As mentioned earlier, the eVect of any novel mutation found would ideally need in vitro conWrmation, expressed on the background of the relevant polymorphic allele. This is particularly true for missense changes. In those families where a mutation is known, direct detection of the mutation can be used for diagnosis of

other family members or for prenatal diagnosis [57]. Linkage analysis using polymorphic markers within and close to the AGT gene has more general applicability and is particularly useful for prenatal diagnosis where the family mutation has not been identiWed. Two polymorphisms which occur within the AGXT gene itself are quite useful. One of these is a 74 bp insertion in intron 1 [24] and the other is a variable number of tandem repeat (VNTR) in intron 4 [25] both of which can be detected by PCR [24,41]. Other microsatellite markers outside the AGXT gene on chromosome 2 are more polymorphic [58] and have been used for prenatal diagnosis [57] and looking for common founder haplotypes [30].

Conclusions Although three mutations account for more then 45% of PH1 mutations [4], the remainder recur rarely or are family speciWc. The eVects of nonsense and frameshift mutations will usually be obvious, although the conse-

M.B. Coulter-Mackie, G. Rumsby / Molecular Genetics and Metabolism 83 (2004) 38–46

quences of missense mutations may be more complex. As diagnosis turns more to DNA testing, either by speciWc mutation panel analysis or by high throughput DNA sequencing it is important to be able to correlate observed sequence changes with previously identiWed changes as an aid to assessing their potential signiWcance. To this end, a database of all mutations and their functional consequences is under construction and will aid rapid comparison.

References [1] C.J. Danpure, Primary hyperoxaluria, in: C.R. Scriver, A.L. Beaudet, W.S. Sly, D. Valle (Eds.), The Metabolic and Molecular Bases of Inherited Disease, McGraw-Hill, New York, 2001, pp. 3323–3367. [2] C.J. Danpure, P.R. Jennings, P. Fryer, P.E. Purdue, J. Allsop, Primary hyperoxaluria type 1: genotypic and phenotypic heterogeneity, J. Inher. Metab. Dis. 17 (1994) 487–499. [3] M.B. Coulter-Mackie, C. White, R.M. Hurley, Primary hyperoxaluria type 1, in: /GeneReviews/ at GeneTests: Medical Information Resource [database online]. Available at www.genetests.org. (2004 update). [4] G. Rumsby, E. Williams, M. Coulter-Mackie, Evaluation of mutation screening as a Wrst line test for the diagnosis of the primary hyperoxalurias, Kidney Int. 66 (2004) 959–963. [5] C. Toussaint, Pyridoxine-responsive PH1: treatment, J. Nephrol. 11 (1998) 49–50. [6] M. Marangella, Transplantation strategies in type 1 primary hyperoxaluria: the issue of pyridoxine responsiveness, Nephrol. Dial. Transplant. 14 (1999) 301–303. [7] J.I. Scheinman, Primary hyperoxaluria: therapeutic strategies for the 90’s, Kidney Int. 40 (1991) 389–399. [8] E.H. Nielsen, M.T. Severinsen, P. Jensen, T. Ring, Primary hyperoxaluria type 1 detected by liver biopsy, Ugeskr Laeger 161 (1999) 1601–1602. [9] C.J. Danpure, G. Rumsby, Molecular aetiology of primary hyperoxaluria and its implications for clinical management, Exp. Rev. Mol. Med. 6 (2004) 1–15. [10] D.A. Applegarth, J.R. Toone, R.B. Lowry, Incidence of inborn errors of metabolism in British Columbia, 1969–1996, Pediatrics 105 (2000) e10(11–16). [11] C.S. van Woerden, J.W. GroothoV, R.J.A. Wanders, J.C. Davin, F.A. Wijburg, Primary hyperoxaluria type 1 in the Netherlands: prevalence and outcome, Nephrol. Dial. Transplant. 18 (2003) 273–279. [12] P. Cochat, A. Deloraine, M. Rotily, F. Olive, F. Liponski, N. Deries, Epidemiology of primary hyperoxaluria type 1, Nephrol. Dial. Transplant. 10 (Suppl. 8) (1995) 3–7. [13] N. Kopp, E. Leumann, Changing pattern of primary hyperoxaluria in Switzerland, Nephrol. Dial. Transplant. 10 (1995) 2224– 2227. [14] A. Kamoun, R. Lakhoua, End-stage renal disease of the Tunisian child: epidemiology, etiologies, and outcome, Pediatr. Nephrol. 10 (1996) 479–482. [15] C. Rinat, R.J.A. Wanders, A. Drukker, D. Halle, Y. Frishberg, Primary hyperoxaluria type: a model for multiple mutations in a monogenic disease within a distinct ethnic group, J. Am. Soc. Nephrol. 10 (1999) 2352–2358. [16] K. Madani, H. Otoukesh, A. Rastegar, S. Van Why, Chronic renal failure in Iranian children, Pediatr. Nephrol. 16 (2001) 140–144. [17] B. Hoppe, C. Langman, A United States survey on diagnosis, treatment, and outcome of primary hyperoxaluria, Pediatr. Nephrol. 18 (2003) 986–991.

45

[18] B. Hoppe, C. Danpure, G. Rumsby, P. Fryer, P.R. Jennings, N. Blau, G. Schubiger, T. Neuhaus, E. Leumann, A vertical (pseudodominant) pattern of inheritance in the autosomal recessive disease primary hyperoxaluria type 1: lack of relationship between genotype, enzymic phenotype, and disease severity, Am. J. Kidney Dis. 29 (1997) 36–44. [19] G. Rumsby, Biochemical and genetic diagnosis of the primary hyperoxalurias: a review, Mol. Urol. 4 (2000) 349–353. [20] C. Danpure, Molecular and clinical heterogeneity in primary hyperoxaluria type 1, Am. J. Kidney Dis. 17 (1991) 366–369. [21] Y. Takada, M. Kaneko, H. Esumi, P.E. Purdue, C.J. Danpure, Human peroxisomal L-alanine:glyoxylate aminotransferase. Evolutionary loss of a mitochondrial targeting signal by point mutation of the initiation codon, Biochem. J. 268 (1990) 517–520. [22] P.E. Purdue, M.J. Lumb, M. Fox, G. GriVo, C. Hamon-Benais, S. Povey, C.J. Danpure, Characterization and chromosomal mapping of a genomic clone encoding human alanine:glyoxylate aminotransferase, Genomics 10 (1991) 34–42. [23] P.E. Purdue, Y. Takada, C.J. Danpure, IdentiWcation of mutations associated with peroxisome-to-mitochondrion mistargeting of alanine/glyoxylate aminotransferase in primary hyperoxaluria type 1, J. Cell Biol. 111 (1990) 2341–2351. [24] P.E. Purdue, M.J. Lumb, J. Allsop, C.J. Danpure, An intronic duplication in the alanine:glyoxylate aminotransferase gene facilitates identiWcation of mutations in compound heterozygote patients with primary hyperoxaluria type 1, Hum. Genet. 87 (1991) 394–396. [25] C.J. Danpure, G.M. Birdsey, G. Rumsby, M.J. Lumb, P.E. Purdue, J. Allsop, Molecular characterization and clinical use of a polymorphic tandem repeat in an intron of the human alanine:glyoxylate aminotransferase gene, Hum. Genet. 94 (1994) 55–64. [26] A.C. Tarn, C. Von Schnakenburg, G. Rumsby, Primary hyperoxaluria type 1: diagnostic relevance of mutations and polymorphisms in the alanine:glyoxylate aminotransferase gene (AGXT), J. Inher. Metab. Dis. 20 (1997) 689–696. [27] C. von Schnakenburg, G. Rumsby, Primary hyperoxaluria type 1: a cluster of new mutations in exon 7 of the AGXT gene, J. Med. Genet. 34 (1997) 489–492. [28] P.E. Purdue, J. Allsop, G. Isaya, L.E. Rosenberg, C.J. Danpure, Mistargeting of peroxisomal L-alanine:glyoxylate aminotransferase to mitochondria in primary hyperoxaluria patients depends upon activation of a cryptic mitochondrial targeting sequence by a point mutation, Proc. Natl. Acad. Sci. USA 88 (1991) 10900–10904. [29] M.B. Coulter-Mackie, A. Tung, H.E. Henderson, J.R. Toone, D.A. Applegarth, The AGT gene in Africa: a distinctive minor allele haplotype, a polymorphism (V326I), and a novel PH1 mutation (A112D) in black Africans, Mol. Genet. Metab. 78 (2003) 44–50. [30] A. Amoroso, D. Pirulli, D. Puzzer, M. Boniotta, S. Crovella, S. Zezlina, A. Spano, G. Mazzola, S. Savoldi, C. Ferrettini, S. Berutti, M. Petrarulo, M. Marangella, AGXT gene mutations and their inXuence on clinical heterogeneity of type 1 primary hyperoxaluria, J. Am. Soc. Nephrol. 12 (2001) 2072–2079. [31] M.J. Lumb, C.J. Danpure, Functional synergism between the most common polymorphism in human alanine: glyoxylate aminotransferase and four of the most common disease-causing mutations, J. Biol. Chem. 275 (2000) 36415–36422. [32] Y.P. Yuen, C.K. Lai, G.M. Tong, P.M. Wong, F.K. Wong, K.Y. Lo, a.K. Wong, C.W. Lam, Novel mutations of the AGXT gene in Chinese patients with childhood and adult -onset primary hyperoxaluria type 1, J. Inher. Metab. Dis. 26 (Suppl. 2) (2003) 196. [33] D. Pirulli, D. Puzzer, L. Ferri, S. Crovella, A. Amoroso, C. Ferretini, M. Marangella, G. Mazzola, F. Florian, Molecular analysis of hyperoxaluria type 1 in Italian patients reveals eight new mutations in the alanine:glyoxylate aminotransferase gene, Hum. Genet. 104 (1999) 523–525. [34] M.B. Coulter-Mackie, D. Applegarth, J.R. Toone, H. Henderson, The major allele of the alanine:glyoxylate aminotransferase gene:

46

[35]

[36]

[37]

[38]

[39]

[40]

[41]

[42]

[43]

[44]

[45]

M.B. Coulter-Mackie, G. Rumsby / Molecular Genetics and Metabolism 83 (2004) 38–46 seven novel mutations causing primary hyperoxaluria type 1, Mol. Genet. Metab. 82 (2004) 64–68. D. Milosevic, C. Rinat, D. Batinic, Y. Frishberg, Genetic analysis a diagnostic tool for primary hyperoxaluria type 1, Pediatr. Nephrol. 17 (2002) 896–898. D. Pirulli, M. Marangella, A. Amorosa, Primary hyperoxaluria: genotype-phenotype correlation, J. Nephrol. 16 (2003) 297–309. C.J. Danpure, P.E. Purdue, P. Fryer, S. GriYths, J. Allsop, M.J. Lumb, K.M. Guttridge, P.R. Jennings, J.I. Scheinman, S.M. Mauer, N.O. Davidson, Enzymological and mutational analysis of a complex primary hyperoxaluria type 1 phenotype involving alanine:glyoxylate aminotransferase peroxisome-to-mitochondrion mis-targeting and intraperoxisomal aggregation, Am. J. Hum. Genet. 53 (1993) 417–432. O. Basmaison, M. Rolland, P. Cochat, D. Bozon, IdentiWcation of5 novel mutations in the AGXT gene, Hum. Mutat. (2000) Online #327. C.S. van Woerden, J.W. GroothoV, F.A. Wijburg, C. Annink, R.J.A. Wanders, H.R. Waterham, Clinical implications of mutation analysis in primary hyperoxaluria type 1, Kidney Int. 66 (2004) 746–752. P.E. Purdue, M.J. Lumb, J. Allsop, Y. Minatogawa, C.J. Danpure, A glycine-to-glutamate substitution abolishes alanine:glyoxylate aminotransferase catalytic activity in a subset of patients with primary hyperoxaluria type 1, Genomics 13 (1992) 215–218. M.B. Coulter-Mackie, G. Rumsby, D.A. Applegarth, J.R. Toone, Three novel deletions in the alanine:glyoxylate aminotransferase gene of three patients with type 1 hyperoxaluria, Mol. Genet. Metab. 74 (2001) 314–321. C. Von Schnakenburg, G. Rumsby, IdentiWcation of new mutations in primary hyperoxaluria type 1 (PH1), J. Nephrol. 11 (S-1) (1998) 15–17. Y. Minatogawa, S. Tone, J. Allsop, P.E. Purdue, Y. Takada, C.J. Danpure, R. Kido, A serine-to-phenylalanine substitution leads to loss of alanine:glyoxylate aminotransferase catalytic activity and immunoreactivity in a patient with primary hyperoxaluria type 1, Hum. Mol. Genet. 1 (1992) 643–644. K. Nishiyama, T. Funai, R. Katafuchi, F. Hattori, K. Onoyama, A. Ichiyama, Primary hyperoxaluria type 1 due to a point mutation of T to C in the coding region of the serine:pyruvate aminotransferase gene, Biochem. Biophys. Res. Commun. 176 (1991) 1093– 1099. C. Von Schnakenburg, S.A. Hulton, D. MIlford, H.P. Roper, G. Rumsby, Variable presentation of primary hyperoxaluria type 1 in 2 patients homozygous for a novel combined deletion and

[46]

[47] [48]

[49]

[50]

[51]

[52]

[53]

[54] [55]

[56]

[57] [58]

insertion mutation in exon 8 of the AGXT gene, Nephron 778 (1998) 485–488. P.K. Nogueira, T.S. Vuong, O. Bouton, A. Maillard, M. Marchand, M.O. Rolland, P. Cochat, D. Bozon, Partial deletion of the AGXT gene (EX1_EX7del): a new genotype in hyperoxaluria type 1, Hum. Mutat. (2000) Online #311. K. Latta, Pyridoxine sensitivity in primary hyperoxaluria type 1, Nephrol. Dial. Transplant. 14 (1999) 2784–2789. C.F. GiaW, G. Rumsby, Kinetic analysis and tissue distribution of human D-glycerate dehydrogenase/glyoxylate reductase and its relevance to the diagnosis of primary hyperoxaluria type 2, Ann. Clin. Biochem. 35 (1998) 104–109. M. Lumb, A. Drake, C. Danpure, EVect of N-terminal
helix formation on the dimerization and intracellular targeting of alanine:glyoxylate aminotransferase, J. Biol. Chem. 274 (1999) 20587–20596. J.M. Leiper, P.B. Oatey, C.J. Danpure, Inhibition of alanine:glyoxylate aminotransferase 1 dimerization is a prerequisite for its peroxisome-to-mitchondrion mistargeting in primary hyperoxaluria type 1, J. Cell Biol. 135 (1996) 939–951. X. Zhang, S.M. Roe, Y. Hou, M. Bartlam, Z. Rao, L.H. Pearl, C.J. Danpure, Crystal structure of alanine:glyoxylate aminotransferase and the relationship between genotype and enzymatic phenotype in primary hyperoxaluria type 1, J. Mol. Biol. 331 (2003) 643–652. P.J. Waters, Degradation of mutant proteins, underlying “loss of function” phenotypes, plays a major role in genetic disease, Curr. Issues Mol. Biol. 3 (2001) 57–65. C. Ouzounis, C. Sander, Homology of the NifS family of proteins to a new class of pyridoxal phosphate-dependent enzymes, FEBS Lett. 330 (1993) 241–242. P.A. Frischmeyer, H. Dietz, Nonsense-mediated mRNA decay in health and disease, Hum. Mol. Genet. 8 (1999) 1893–1900. S.E. Antonarakis, M. Krawczak, D.N. Cooper, The nature and mechanisms of human gene mutation, in: C.R. Scriver, A.L. Beaudet, W.S. Sly, D. Valle (Eds.), The Metabolic and Molecular Bases of Inherited Disease, McGraw-Hill, New York, 2001, pp. 343–344. A. Santana, E. Salido, A. Torres, L.J. Shapiro, Primary hyperoxaluria type 1 in the Canary Islands: a conformational disease due to I244T mutation in the P11L-containing alanine:glyoxylate aminotransferase, Proc. Natl. Acad. Sci. USA 100 (2003) 7277–7282. C.J. Danpure, G. Rumsby, Strategies for the prenatal diagnosis of primary hyperoxaluria type 1, Prenat. Diagn. 16 (1996) 587–598. C. von Schnakenburg, T. Weir, G. Rumsby, Linkage of microsatellites to the AGXT gene on chromosome 2q37.3 and their role in prenatal diagnosis of primary hyperoxaluria type 1, Ann. Hum. Genet. 61 (1997) 365–368.