Abstract
Abstract. Primary hyperoxaluria type 1 is an autosomal recessive inherited metabolic disease in which excessive oxalates are formed by the liver and excreted by the kidneys, causing a wide spectrum of phenotypes ranging from renal failure in infancy to mere renal stones in late adulthood. Mutations in the AGXT gene, encoding the liver-specific enzyme alanine:glyoxylate aminotransferase, are responsible for the disease. Seven mutations were detected in eight families in Israel. Four of these mutations are novel and three occur in children living in single-clan villages. The mutations are scattered along various exons (1, 4, 5, 7, 9, 10), and on different alleles comprissing at least five different haplotypes. All but one of the mutations are in a homozygous pattern, reflecting the high rate of consanguinity in our patient population. Two affected brothers are homozygous for two different mutations expressed on the same allele. The patients comprise a distinct ethnic group (Israeli Arabs) residing in a confined geographic area. These results, which are supported by previous data, suggest for the first time that the phenomenon of multiple mutations in a relatively closed isolate is common and almost exclusive to the Israeli-Arab population. Potential mechanisms including selective advantage to heterozygotes, digenic inheritance, and the recent emergence of multiple mutations are discussed.
Primary hyperoxaluria type 1 (PH1) is a rare autosomal recessive inborn error of peroxisomal metabolism (1). It is caused by the deficiency of the liver-specific enzyme alanine:glyoxylate aminotransferase (AGT). Detoxification of glyoxylate is thus impaired, and overproduction of oxalate ensues. Oxalates are physiologically excreted by the kidneys. The elevated urinary oxalate concentration leads to formation of calciumoxalate crystals and subsequently to renal stones, nephrocalcinosis, and in severe cases to end-stage renal disease (ESRD). The oxalates are then deposited in many other tissues, including bone, bone marrow, retina, cornea and heart, leading to the clinical entity known as oxalosis. The incidence and severity of PH1 varies in different geographic regions (2,3). It is much more prevalent in Mediterranean countries. In Tunisia, for example, it accounts for 13.5% of cases of ESRD in children compared with only 0.7% in North America (4).
The gene encoding the AGT enzyme (AGXT) has been identified and mapped to chromosome 2q36-37 (5). The AGXT gene is organized in 11 exons, and the cDNA consists of approximately 1600 bp. Seventeen AGXT mutations and five intragenic polymorphic sites have been identified. These mutations account for only a small fraction (25%) of the affected individuals studied (6). The prevalent phenotype in our group of patients is ESRD within the first decade of life rendering prenatal diagnosis very important.
The aim of our study was to analyze the specific mutations causing PH1 in our patient population. Identification of a given mutation could provide an accurate tool for prenatal diagnosis in the affected families, allowing for genetic counseling and detection of presymptomatic individuals for timely medical management.
Materials and Methods
Patients
The study included 11 children from eight heavily inbred families diagnosed with PH1 who are followed in our service. All patients are of Arab origin except for one family of Jewish-Libyan descent. Most children are offspring of consanguineous matings (7 of 8) with intermarriage of first-degree cousins being the most common pattern. Four patients presented in the first year of life, and five reached ESRD requiring renal replacement therapy before age 11 (Table 1). None of the patients was pyridoxine-sensitive. Four patients succumbed to the disease while awaiting a combined kidney and liver transplantation. The diagnosis of PH1 was confirmed by the detection of decreased enzymatic activity of AGT in liver biopsy specimens (Table 1). In families with more than one affected child, diagnosis was confirmed by measuring the enzymatic activity of AGT in only one member. Our control group consisted of healthy unrelated adult volunteers of Israeli-Arab descent with no family history of renal stones. The study was approved by the Ethics Committee of the Shaare Zedek Medical Center.
Clinical and enzymatic characteristicsa
AGT Assay
The specific catalytic activity of AGT was measured in liver tissue (7). In several instances, immunoblots were performed to detect cross-reactive material (CRM) using polyclonal antibody raised against purified human AGT (8).
Genetic Analysis
Blood samples from all affected individuals and their first-degree relatives (parents and siblings) were collected. Genomic DNA was extracted from lymphocytes using standard molecular biology techniques. PCR was used to amplify individual exons using primers based on published information regarding intron-exon boundaries (5,9,10). PCR primers and programs are shown in Table 2.
PCR primers and annealing temperature
Screening for the previously published mutations that alter recognition sites in a given exon was performed using the relevant restriction enzymes and ethidium-stained agarose electrophoresis. Exonspecific PCR products were separated on agarose gel followed by extraction and DNA purification (QIAgen, Hilden, Germany). Direct cycle sequencing of the entire coding region in each affected individual, using 32P-γATP-labeled primers, was implemented (Sequitherm Excell DNA, sequencing kit; Epicentre Technologies, Madison, WI).
Results
Mutations
Four novel point mutations causing PH1 were detected in five affected individuals from three unrelated families (Table 3). Several measures were taken to infer the disease-causing nature of the mutations. First, direct sequencing of genomic DNA from patients failed to disclose any other mutation in the entire coding region of AGXT. Second, both parents were found to carry only the specific point mutation in a heterozygous pattern. Finally, the screening of 100 alleles from healthy unrelated individuals from the same ethnic background did not detect the given mutation.
Mutations, haplotypes, and restriction site changesa
In seven families, the affected individuals are homozygous for a given mutation and only one is a compound heterozygote. A compound heterozygote patient, with infantile PH1, was detected in family A in which the parents are nonconsanguineous. The maternal mutation occurs in a well-conserved region of exon 4, where G 588 is substituted by A and leads to the substitution of arginine instead of glycine at amino acid position 156 (Table 3). The paternal mutation has yet to be determined. However, the paternal mutated allele is associated with the intragenic polymorphic marker C154T (Pro11Leu), which is absent in the maternal allele, and enables the screening of related individuals. Two of the siblings carry the maternal mutation and one carries the paternal mutated allele. The affected child is the only individual who carries both mutated alleles (Figure 1).
PCR amplification of genomic DNA derived from patient A, M and her family. (Top Panel) The 150-bp exon 4 digested by the restriction enzyme BsaJI. The mutation G588A abolishes a restriction site for the enzyme BsaJI creating a fragment that measures 150 bp, whereas the normal allele results in two fragments of 77 and 73 bp. The mother and three of her children carry one undigested allele. (Bottom Panel) Exon 1 digested by the restriction enzyme StyI. The C154T polymorphism abolishes a restriction site resulting in a 280-bp fragment, whereas the normal allele results in a 106-bp fragment. The C154T polymorphism cosegregates with the paternal mutated allele. It is carried by the father and two offspring. The affected individual is the only member of this family who carries both mutated alleles.
Family Z carries a point mutation in a well-conserved region in exon 10, where A1119 is substituted by T, resulting in a premature termination codon (TGA) replacing a codon for amino acid 333. (Figure 2, Table 3). This family resides in a single-clan Arab village in Lower Galilee. In the extended family, there were several infants who died undiagnosed, from a disease that is highly suggestive of PH1.
Exon 10 amplification of genomic DNA derived from family Z digested by the restriction enzyme GsuI. The uncut exon (U) measures 240 bp. The healthy control (H) results in 155- and 85-bp fragments. The parents (P1 and P2) are heterozygotes for the mutation, which abolishes the restriction site. The affected child (A) is homozygous for the mutation. MM denotes a molecular size marker.
Family H resides in the Gaza Strip and has two affected children (H, A, and M) suffering from neonatal PH1. Interestingly, two homozygous point mutations were detected by direct sequencing of both strands in exon 9 and are only 28 nucleotides apart (Figure 3, Table 3). Both mutations were found in the parents in a heterozygous pattern and were absent in other PH1 patients and in unrelated individuals from the same ethnic group. The first mutation C987T predicts the substitution Arg289Cys. The second mutation is the T1015C substitution, which leads to Leu298Pro.
DNA sequencing gel of exon 9 derived from patient H, A. Two mutations C987T and T1015C, 28 bp apart, are demonstrated on the same allele. The affected child (designated A) is homozygous for both mutations, whereas the parents (P) are shown to be heterozygous for both mutations. A healthy control (H) is shown in the far right lanes.
We identified the previously published mutation G243A predicting Gly41Arg (11), for the first time, in a homozygous pattern in families J1 and J2. It was formerly reported to occur in compound heterozygotes, and in association with the Pro11Leu polymorphic marker (together with intron A duplication and A1142G, intragenic polymorphisms comprise the “minor” allele [(9)]). Our studies indicate that the mutation is always carried by the other (“major”) allele. We detected this homozygous mutation in four children with recurrent nephrolithiasis from two related nuclear families residing in an Arab village whose 5500 inhabitants can be traced to a single ancestor. The mutation is carried by the corresponding parents and segregates in the families, resulting in three additional siblings who are heterozygotes. The affected individuals are also homozygotes for the C1342A variant in the 3′ untranslated region.
A recently published mutation in a well-conserved region of exon 5 was found in a patient (S,S) who presented at the age of 4 yr with nephrolithiasis and at the age of 10 yr still has normal kidney function. G690 is substituted by A, leading to Gly190Arg (Figure 4, Table 3) (12).
DNA sequencing gel of exon 5 from patient S, S shown in the left lanes (designated A) and the healthy control (designated H) in the right lanes. The G690A substitution is demonstrated leading to Gly190 Arg.
The seventh mutation was detected in a homozygous pattern in children from two unrelated families who reached ESRD before the age of 11 yr. One is an Arab family (EA2) from the Gaza Strip and the other is a Jewish-Libyan family (M) residing in a village whose inhabitants, all carrying the same surname, emigrated from a closed community. This is a missense mutation T853C in exon 7 resulting in the Ile244Thr substitution (Table 3). The same mutation was recently reported (13), and accounts for 9% of the PH1-causing mutated alleles in the population described.
Polymorphisms and Haplotypes
Our studies confirm the association between the following three intragenic polymorphisms: C154T, A1142G, and intron A 74-bp duplication, previously designated the “minor” allele (9). C154T, along with the intronic duplication, in the absence of A1142G, was detected in only one out of 100 alleles derived from healthy unrelated individuals. The frequency of the “minor” allele in our unaffected population is 20%, and is similar to that reported in Caucasians (14). In several families, the intronic duplication could serve as a marker for a mutated or a normal allele. Additional intragenic polymorphic markers include C386T, G776A, and C1342A. Altogether, we have defined five different haplotypes using these six intragenic polymorphisms in the affected children (Table 3).
Discussion
We have identified seven point mutations in the AGXT gene that cause PH1 among eight families in Israel: seven of Arab origin and one of Jewish-Libyan extraction. Four mutations are novel, an additional one was detected for the first time in a homozygous pattern, and only two have been published previously. All of our patients tested demonstrated minimal immunoreactive material along with negligible enzymatic activity (Table 1), thus excluding the possibility of mitochondrial mistargeting (15).
The segregation of the different mutations with disease in our group of patients points to their causal role in PH1. The majority of the mutations occur in regions of evolutionary conservation common to humans, marmosets, rats, and rabbits (16), which suggests that these regions are of functional significance.
In light of the minimal immunoreactive material detected, any attempt to determine the expected alterations in the protein structure attributed to the different mutations is irrelevant. The paucity of immunoreactive material may point to either mRNA or protein instability. The ultimate proof of the effect of these mutations requires in vitro expression studies of the mutated gene.
We detected two mutations in a single allele (C987T and T1015C) in two affected siblings. The finding of multiple mutations in one allele is rare. A similar phenomenon has been noted in only a few diseases, including Hurler syndrome (17) and Dejerine-Sottas (18). Both mutations lead to an amino acid substitution, and neither was found in the unaffected population. This may indicate that either one of them or both affect the gene-product structure and function, and only expression studies would provide a definitive clarification of this issue. Both mutations occur in exon 9, in which no other mutations or polymorphisms have been described. This argues against the possibility of this being a “hot spot,” and the exact mechanism remains obscure.
Our studies demonstrate a phenomenon best described by the following characteristics: the existence of multiple mutations in the same gene within a small geographic area comprising a seemingly homogeneous population. Most rare autosomal recessive diseases have one common mutation and several rarer ones (19). A single common mutation also has been observed in diseases prevalent in highly consanguineous communities, including maple syrup urine disease among the Mennonites (20), phenylketonuria among Jews of Yemenite origin (21), as well as Tay-Sachs disease and familial hypercholesterolemia among French Canadians (22,23). This is in contrast to several diseases described in Israel including thalassemia (24), Hurler syndrome (17), metachromatic leukodystrophy (MLD) (25), and glutaric aciduria type 1 (26). An example outside our region is limb-girdle muscular dystrophy type 2A described in La Reunion Isle, a small island east of Madagascar (27). The major point of our findings is an extension of the above observations of this latter group. It suggests that this phenomenon is common and almost exclusive to the Israeli-Arab population. This is particularly relevant when considering any pathogenetic theory.
Selective advantage to the heterozygote is the mechanism most likely responsible for the occurrence of multiple mutations that result in a high incidence of a given disease within a distinct ethnic group (28). Thus far, no evidence for this has been found in carriers of PH1, MLD, glutaric aciduria, or Hurler syndrome. Over-representation of heterozygotes for mutated alleles in the affected population can provide a clue to such an advantage. However, it should be noted that social influences including consanguinity, intraracial marriages, and population migration may skew the trends of natural selection as is exemplified by the increased prevalence of sickle-cell anemia among the African-American community outside the region of selective pressure. A wide population screening is warranted, yet its results may be biased due to the reasons noted above. Our cohort of affected families is too small to draw conclusions regarding the proportion of heterozygous siblings.
Two additional mechanisms have recently been suggested as alternative explanations for this phenomenon. Both assume a homogeneous population. The first one is the digenic model that is exemplified by the “Reunion paradox” (27,29). In this model, mutations in two unlinked genes are required to result in a phenotype. It postulates a suppressor, compensatory, or a modifier gene (27) active in the general population, but inactive in the isolated population because of a mutation of ancestral origin in this second site. The digenic model requires a similar distribution of mutated alleles in the affected families as in the general population and the existence of asymptomatic homozygotes. The screening of 100 alleles from healthy Israeli-Arabs for each of the mutations detected in our study revealed no carriers. Additionally, none of the asymptomatic siblings of the affected individuals was found to be homozygous for the relevant mutation. The digenic model further predicts an increased prevalence of compound heterozygotes due to the relative frequent occurrence of each mutation in the general population. This was not noted in our study, yet that may be due to a high rate of consanguinity leading to a genetic drift. We are searching for a unifying explanation applicable to the aforementioned diseases among the Israeli-Arab population. A common hypothetical suppressor/modifier gene would have to counterbalance the deleterious effect of each one of the different disease-causing mutations in this population. The existence of such a redundant gene-product is unlikely.
The second mechanism (30) ascribes the recent emergence of multiple mutations in a given gene within a confined homogeneous population to a high mutation rate. The prevalence of consanguineous mating in the Arab population in Israel reaches 45%, with nearly half being first-degree cousins (31). Given these circumstances, it was suggested that a newly occurring recessive mutation will lead to a homozygous affected individual within four to five generations (30). A high rate of consanguinity in families with a large number of offspring is responsible only for the expression, rather than the emergence, of multiple mutations. Marriages within the Israeli-Arab population occur preferentially within the same village, creating isolated communities residing in small villages. A high incidence of new mutations would therefore result in many local founders. This raises disturbing questions regarding the nature of the high mutagenic activity affecting a limited number of genes in this particular population. Unlike achondroplasia, in which a high de novo mutation rate results from a unique nucleotide substitution (32), the multiple mutations in the affected genes described above are scattered along various exons and do not recur or share an identical sequence, thus making “hot spots” a less likely explanation for this phenomenon.
However, it may be argued that the population studied is not homogeneous and consists of offspring of several founders who have lately emigrated from neighboring countries and have settled in separate villages. In this case, the mutations would have been ancient but the end point would be similar: local founders rather than a common ancestor for the entire population. There are no accurate data concerning the origin of the Arab population in Israel. Political events may have resulted in population shifts, such that people who currently live in a confined region originated in fact from remote sites. Due to genetic drift, the mutation distribution here may be different from that in the countries of origin. Therefore, one cannot draw conclusions regarding the origins of this population from mutation analyses of the same diseases in neighboring communities. This option overlooks the fact that multiple mutations recur in a finite number of genes within the Israeli-Arab population and is therefore unlikely.
In summary, we describe the genetic analysis of PH1 patients from eight families in Israel. Our results show that almost every family carries a different point mutation. Our findings, supported by previous observations in Israel, lead us to conclude that the phenomenon of multiple mutations in a single gene within a confined region is a frequent occurrence in the Israeli-Arab population and therefore points to a common genetic mechanism. Although our data cast doubt on various mechanisms mentioned previously, selective advantage to the heterozygotes is more plausible. The factor(s) inducing selective advantage have yet to be determined. It is disturbing that in a confined population selective advantage to the heterozygotes recurs and gives rise to rare recessive diseases.
Acknowledgments
Acknowledgments
The work was supported in part by the Joint Research Fund of the Hebrew University and Shaare Zedek Medical Center, the Teva Medical Research Fund, and the Mirsky Foundation. The authors are indebted to Dr. Orly Reiner for continual support and helpful discussions and to Drs. Karl Skorecki and Annick Rothschild for the critical review of the manuscript. The advice of Dr. Pinchas Renbaum is appreciated.
Footnotes
-
American Society of Nephrology
- © 1999 American Society of Nephrology