Familial hypomagnesemia with secondary hypocalcemia (HSH) (MIM 307600) was studied in three inbred Bedouin kindreds from Israel. The three kindreds, one extended and two nuclear families, contained 13 affected individuals, 11 males and two females. Assuming that the individuals affected with hypomagnesemia shared a chromosomal region inherited from a common ancestor, we used a DNA pooling strategy in a genome-wide search for loci which show homozygosity for shared alleles in affected individuals. DNA samples from affected individuals within a single kindred were pooled and used as the template for PCR amplification of short tandem repeat polymorphic markers (STRPs). Pooled DNA from unaffected siblings and parents were used as controls. A shift towards homozygosity was observed in the affected DNA pool compared with the control pools with D9S301 (GATA7D12). Genotyping of individual DNA samples with D9S301 and several flanking markers confirmed linkage to chromosome 9 with maximum LOD scores of 3.4 ([theta] = 0.05), 3.7 ([theta] = 0) and 2.3 ([theta] = 0) for the three families. We have identified a 14 cM interval on chromosome 9 (9q12-9q22.2), flanked by proximal marker D9S1874 and distal marker D9S1807, within which all affected individuals from the three kindreds are homozygous for a shared haplotype. The disease segregates with a common affected haplotype in the three families, suggesting that hypomagnesemia is caused by a common ancestral mutation in these families. Although HSH has been previously reported to be X linked, these linkage data demonstrate that the disorder is an autosomal recessive disease in these kindreds. Mapping of a chromosomal breakpoint in a somatic cell line established from a patient with HSH and a balanced X;9 translocation placed the chromosomal breakpoint in a 500 kb region flanked by D9S1844 and D9S273. Identification of the gene responsible for hypomagnesemia will provide insight into the regulation of this essential cation.
Familial hypomagnesemia with secondary hypocalcemia (HSH) (MIM 307600) is described as a rare inherited disorder, primarily characterized by extremely low levels of serum magnesium associated with symptomatic hypocalcemia. The disease is also called hypomagnesemic tetany. Patients with the disease present a number of symptoms resulting from the hypocalcemia, including restlessness, tremor, tetany and overt seizures (1 -3 ) shortly after birth. If left undiagnosed and untreated, the disease is fatal (1 ,3 ). While the cause of the abnormality is not known, it is thought to be due to a defect in the absorption of magnesium (3 ,4 ). Excretion of magnesium in urine and stool are normal. Correction of the magnesium deficiency with high concentrations of magnesium supplementation is required. The excess magnesium is not secreted by the kidneys and is presumed to be eliminated through the bowel (5 ).
HSH is distinct from several other disorders which also have hypomagnesemia as a symptom. Hypomagnesemia due to kidney defects has been implicated in patients with high urinary magnesium excretion (MIM 248250). Conditions causing steatorrhea or severe chronic diarrhea, such as Crohn's disease, ulcerative colitis, celiac disease, Whipple's disease and short bowel syndrome, can result in severe hypomagnesemia (3 ,4 ). Patients with Gitelman syndrome (MIM 263800), known to be caused by mutations in the thiazide-sensitive Na-Cl co-transporter, also display hypomagnesemia as part of their clinical manifestations (6 ). Hypomagnesemia, from renal wasting of magnesium, attributable to mutations in the Na-K co-transporter NKCC2 is also observed in patients with Bartter's syndrome (MIM 241200) (7 ).
Three inbred kindreds from the Bedouin population in Israel with HSH were identified (Fig. 1 ). The subjects of this study belong to one extended kindred and two nuclear Bedouin families in which the parents of all of the affected individuals are consanguineous (Fig. 1 ). The two nuclear families (kindreds B1 and B2) belong to one tribe and have a common ancestry. The extended family (kindred A) belongs to a different tribe. The two tribes are not known to be related to each other. These three kindreds contain 13 affected individuals, 11 males and two females.
Most of the patients first presented at 2-8 weeks of age with restlessness, tremor, neuromuscular hyperexcitability and overt seizures. Two cases were evaluated shortly after birth because of the family history of congenital hypomagnesemia and were discovered to have low serum magnesium concentrations prior to the onset of symptoms. All of the affected patients were born after uneventful pregnancies and normal deliveries. They all developed normally until the first clinical symptoms of hypomagnesemia appeared. None had feeding problems or signs of malabsorption.
Serum magnesium levels at initial presentation ranged from 0.2 to 0.8 mg% (normal 1.5-2.6 mg%) and serum calcium levels of 3.7-7.6 mg% (normal 8.4-10.8 mg%). Serum calcium concentration in the presymptomatically diagnosed newborns was within the normal range. Laboratory evaluation for intestinal malabsorption was negative in all cases. Supplementary treatment with 20-90 mg elementary Mg/kg/day eliminated the clinical symptoms in all the patients and restored serum magnesium concentration to near normal (1.1-1.5 mg%) and serum calcium concentration to normal levels.
Biochemical profiles of the 13 patients receiving treatment with supplementary magnesium showed mild hypomagnesemia and normal calcemia. Parathyroid hormone and 1,25-dihydroxyvitamin D concentrations were normal in all tested individuals. Urine fractional excretory magnesium was extremely low in all patients. Some patients had mildly to moderately delayed bone age and short stature. Weight was normal in all patients. Renal ultrasound examinations were normal in all but one patient, who was found to have mild nephrolithiasis. An abnormal EEG was found in four cases, all of whom were not symptomatic at the time of examination. EKG was normal in all patients. Most patients revealed mild osteoporosis on skeletal X-rays; however, bone density was normal.
Initial diagnosis of hypomagnesemia was delayed in two cases. These patients suffered prolonged, intractable seizures which were unresponsive to various anticonvulsants. Thereafter the patients continued to suffer from a chronic convulsive disorder in spite of normalized serum magnesium and calcium concentrations. Both of them are also mentally retarded, presumably as a result of the prolonged, intractable seizures. One of them is institutionalized. All other patients show normal psychomotor development and have normal IQs on formal testing.
By using a DNA pooling method, we searched for short tandem repeat polymorphic markers (STRPs) linked to the HSH phenotype. Pooled DNA from affected and unaffected individuals were PCR amplified with STRP primers and the amplification products were compared by electrophoresis on denaturing polyacrylamide gels. If a disease gene has been inherited by the affected individuals from a common ancestor, the banding pattern of an STRP that is linked to the disease phenotype will show a shift in intensity, toward a predominant allele in the affected pool, compared with the unaffected pool. In the case of a recessive gene inherited from a common ancestor individuals will share a single homozygous allele at the disease locus.
A genome-wide screen using affected and unaffected pools from two of the three Bedouin kindreds (kindreds A and B1) (Fig. 1 ) was performed with 350 tri- and tetranucleotide STRPs developed by the Cooperative Human Linkage Center (11 ). The STRP markers used were separated by an average distance of 8 cM. A reduction in the number of alleles in the affected DNA pool compared with the unaffected pool was observed with marker D9S301 (Fig. 2 ). The most prominent allele within the affected DNA pool for kindred A was also noted to be the most prominent allele within the affected DNA pool for kindred B1.
Chromosome 9 STRPs from CHLC and Genethon and STSs from Whitehead/MIT maps were amplified with DNA purified from a HSH somatic cell line. A combination of PCR and cytogenetic analysis indicated that the cell line had retained the derivative X and had lost the normal X, normal 9 and derivative 9 chromosomes. The markers detectable by PCR indicated that the p arm of chromosome 9 was absent, but a large portion of the q arm had been retained, consistent with the karyotype and earlier reports (9 ,10 ). A single allele was observed for each marker, further substantiating retention of the derivative chromosome and loss of the normal chromosome 9 in the cell line. Markers spanning a distance of 40 cM from D9S165 to D9S1974 were tested. The marker data with the HSH somatic cell DNA is summarized in Figure 3 . D9S1777, D9S1787, D9S1800, D9S1844, D9S1862 and D9S1879 are not resolved on current genetic maps. The presence of D9S1777, D9S1787, D9S1800, D9S1862 and D9S1879 were detected in the HSH somatic cell hybrid DNA, while D9S1844 was not. D9S1844 is possibly located more centromeric to the other five clustered markers and is the proximal flanking marker of the breakpoint. D9S273 is the distal flanking marker. The breakpoint lies in the 500 kb interval that is defined by D9S1844 and D9S273 (Fig. 3 ). This region is contained within the disease interval identified in the affected individuals of the Bedouin kindreds.
Familial hypomagnesemia with secondary hypocalcemia, HSH (MIM 307600), has previously been reported to be inherited as an X-linked recessive disease based on the predominance of male patients and the report of a single female patient with HSH and a balanced X;9 chromosomal translocation, which suggested that the disease maps to Xp22.2 (9 ,10 ). The kindreds used in this study show that the disorder is inherited as an autosomal recessive disorder. A predominance of affected males (11 ) to females (2 ) exists in these kindreds, consistent with previous reports. However, the disease predominance in males is at least partially explained by a substantial predominance of male offspring (31) compared with female offspring (13 ) in these kindreds. Non-penetrance was not observed in any of the unaffected females in these pedigrees.
Linkage mapping using a genome-wide search identified a region on the long arm of chromosome 9 which is linked to the disease. Analysis of multiple chromosome 9 STRPs with the individual DNA samples revealed a shared disease haplotype between all three kindreds, demonstrating a common ancestral gene for the three families. By using markers within the disease region on chromosome 9, a region of homozygosity was observed. The disease interval, as defined by the shared region of homozygosity in all affected individuals in all three kindreds, consists of a 14 cM region between D9S1874 and D9S1807. The observed maximum LOD scores for chromosome 9 markers in this region is consistent with linkage (Table 2 ).
Two unaffected individuals from kindred A (VI-16 and VI-22) contain the complete affected haplotype on one chromosome and show observed recombinant events on the homologous chromosome. These individuals are asymptomatic and have been shown to have normal serum magnesium and calcium levels. The assignment of unaffected status to these patients and the use of the two recombination events narrow the disease region to the ~2.8 cM region between D9S1115 and D9S1807.
We have characterized the derivative X chromosome in a single somatic cell hybrid established from a female HSH patient with a balanced chromosomal translocation, 46,X,t(X;9) (p22.3;q12). It was hypothesized by Chery et al. (10 ) that HSH is caused by the interruption or inactivation of a gene at Xp22. Alternatively, our data, demonstrating linkage of HSH to chromosome 9, suggest that it is interruption of the function of a gene on 9q that is responsible for the disease. Using markers from chromosome 9 we have identified a 500 kb interval from D9S1844 to D9S273 which contains the translocation breakpoint.
The breakpoint region identified using the somatic cell hybrid DNA from the translocation patient with HSH is contained within the conservative disease interval based on the shared haplotypes of the affected family members shown in Figure 1 . An attractive hypothesis is that the HSH gene is interrupted by the translocation breakpoint. Should this be the case, the somatic cell line containing the translocation breakpoint is a valuable system for identification of the HSH gene. This scenario would require that the recombinant unaffected individuals VI-22 and VI-16 be non-penetrant for the disease phenotype, because these two individuals appear to narrow the disease region to an interval not containing the breakpoint. Non-penetrance in recessive disorders is rare. Alternatively, the HSH gene may be located outside the breakpoint interval, as suggested by the observed recombination in unaffected individual VI-22, which narrows the genetic interval to the 2.8 cM region between D9S1115 and D9S1807. It is known that chromosomal position can affect the expression of genes some distance from actual chromosomal breakpoints due to positional variegation effects (13 -15 ). A note of caution is warranted regarding the position of the translocation breakpoint mapped using the somatic cell hybrid. Because of the unavailability of viable cells from the patient, we cannot totally exclude the possibility that the derivative X chromosome in the somatic cell hybrid differs from the derivative X chromosome in the patient. Our search for the gene responsible for HSH will focus first on the 500 kb breakpoint interval and then, if necessary, on the 2.8 cM interval between D9S1115 and D9S1807.
Candidate genes for HSH are suggested by the clinical findings of the affected individuals. A likely candidate gene would be a receptor or ion channel involved in the absorption of intestinal magnesium. Clinical evaluation of the patients described in this paper revealed that the low levels of serum magnesium are likely due to a defect in the intestinal reabsorption of magnesium. Renal magnesium secretion is normal. Normal values are observed for general intestinal absorption of other molecules. In addition to hypomagnesemia, these patients developed hypocalcemia. This observation is expected, because hypocalcemia has been known to develop in instances of severe hypomagnesemic states, as a result of unresponsiveness to parathyroid hormone (16 ). Restoration of serum magnesium levels to the low normal values with high doses, up to 20 times the normal daily requirement, can overcome the defect in absorption and restore serum calcium levels to normal. The Ca2+-Mg2+ ATPases, primary transporters of Mg2+ ions, could be good candidates. Several genes coding Ca2+-Mg2+ ATPases have been characterized; however, none map to chromosome 9 (17 -19 ).
The HSH somatic cell line, established from a female patient with hypomagnesemia, mental retardation, dysmorphic features and the karyotype 46,X,t(X;9) (p22.3;q12), was previously described (10 ). Lymphoblastoid cells derived from the patient were fused with hamster RSK88 cells (HPRT-) and cultured in RPMI growth medium supplemented with 15% fetal calf serum, HAT growth medium (10-4 M hypoxanthine, 1.6 * 10-5 M thymidine, 4 * 10-4 M aminopterin), 100 U/ml penicillin and 100 [mu]g/ml streptomycin. DNA was purified from a T75 flask containing a confluent monolayer of cells using a high salt lysis method (20 ).
DNA was isolated from peripheral blood samples by standard methods (21 ). The DNA concentrations were calculated from spectrophotometric measurements at OD260. The DNAs were diluted to 20 ng/[mu]l, based on absorbance at OD260. The individual DNA samples were tested in PCR assays to demonstrate that equivalent quantities of DNA would yield equivalent quantities of PCR product. The DNAs were then pooled for the genome-wide search for linkage.
From the three families described in Figure 1 , DNA samples from 13 affected and 33 unaffected individuals were characterized. The DNA pools from kindred A consisted of samples from seven affected individuals (VI-4, VI-11, VI-12, VI-14, VI-17 and VI-18) diluted to a final total DNA concentration of 20 ng/[mu]l. An unaffected DNA pool was made from kindred A with DNA samples from seven unaffected sibs (VI-1, VI-2, VI-7, VI-8, VI-10, VI-15 and VI-16). A pool of parental DNA samples (V-1, V-2, V-3, V-4, V-5 and V-6) was also tested. In a similar manner, an affected DNA pool from kindred B1 was made from five individuals (IV-7, IV-8, IV-9, IV-10 and IV-11) and an unaffected sib pool from five samples (IV-2, IV-3, IV-4, IV-5 and IV-6) diluted to a final total DNA concentration of 20 ng/[mu]l.
Primers used in the genome-wide search consisted of tri- and tetranucleotide STRP markers developed by the Cooperative Human Linkage Center (11 ). Each PCR reaction contained 40 ng DNA, 1.25 [mu]l PCR buffer (100 mM Tris-HCl, pH 8.8, 500 mM KCl, 15 mM MgCl2, 0.01% w/v gelatin), 200 [mu]M each dATP, dCTP, dTTP and dGTP, 2.5 pmol each primer and 0.25 U Taq DNA polymerase in a final volume of 8.35 [mu]l. Samples were cycled through 35 cycles of 94oC for 30 s, 55oC for 30 s and 72oC for 30 s. Amplification products were analyzed following electrophoresis on 6% denaturing polyacrylamide gels and visualization by silver staining according to the method of Bassam (22 ).
Linkage analyses were performed using the LOD score method and the MENDEL program (23 -26 ). The genetic maps used for the analyses were generated by the information contained within the CHLC, Genethon and Whitehead databases.
We are grateful to the patients and their families for participating in this study. We thank Dr Adam Kanis for help with the LOD calculations. We also thank Tatiana Rokhlina, Nancy Kramer and Kristin Thompson for excellent technical assistance. This work was supported by NIH grants HL42266 and HG00457 and the Roy J.Carver Charitable Trust.
Human Molecular Genetics
Pages
Introduction
Results
Clinical findings
Genetic linkage analysis
Mapping the breakpoint in a HSH somatic hybrid DNA
Discussion
Materials And Methods
Cell line
Genotyping
Linkage analyses
Acknowledgements
References
REFERENCES
This page is maintained by OUP admin. Last updated Wed Aug 13 15:52:16 BST 1997. Part of the OUP Journals World Wide Web service. Copyright Oxford University Press, 1996