Ataxia-Telangiectasia

A number sign (#) is used with this entry because ataxia-telangiectasia (AT) is caused by homozygous or compound heterozygous mutation in the ATM gene (607585) on chromosome 11q22.

Ataxia-telangiectasia (AT) is an autosomal recessive disorder characterized by cerebellar ataxia, telangiectases, immune defects, and a predisposition to malignancy. Chromosomal breakage is a feature. AT cells are abnormally sensitive to killing by ionizing radiation (IR), and abnormally resistant to inhibition of DNA synthesis by ionizing radiation. The latter trait has been used to identify complementation groups for the classic form of the disease (Jaspers et al., 1988). At least 4 of these (A, C, D, and E) map to chromosome 11q23 (Sanal et al., 1990) and are associated with mutations in the ATM gene.

Homozygotes

Patients present in early childhood with progressive cerebellar ataxia and later develop conjunctival telangiectases, other progressive neurologic degeneration, sinopulmonary infection, and malignancies. Telangiectases typically develop between 3 and 5 years of age. The earlier ataxia can be misdiagnosed as ataxic cerebral palsy before the appearance of oculocutaneous telangiectases. Gatti et al. (1991) contended that oculocutaneous telangiectases eventually occur in all patients, while Maserati et al. (1988) wrote that patients without telangiectases are not uncommon. A characteristic oculomotor apraxia, i.e., difficulty in the initiation of voluntary eye movements, frequently precedes the development of telangiectases.

Gonadal dysfunction in ataxia-telangiectasia was discussed by Miller and Chatten (1967), Zadik et al. (1978), and others. Thibaut et al. (1994) reviewed cases of necrobiosis lipoidica in association with ataxia-telangiectasia.

According to Boder (1985), the oldest known AT patients were a man who died in November 1978 at age 52 years and his sister who died in July 1979 at the age of almost 49 years. The sister was the subject of the report by Saxon et al. (1979) on T-cell leukemia in AT. The possibility of heteroalleles at the ataxia-telangiectasia loci might be suggested.

Neurologic Manifestations

AT may be the most common syndromic progressive cerebellar ataxia of early childhood. Truncal ataxia precedes appendicular ataxia. Oculomotor apraxia is progressive and opticokinetic nystagmus is absent. Choreoathetosis and/or dystonia occur in 90% of patients and can be severe. Deep tendon reflexes become diminished or absent by age 8 and patients later develop diminished large-fiber sensation. Gatti et al. (1991) pointed out that 'a significant proportion of older patients in their twenties and early thirties develop progressive spinal muscular atrophy, affecting mostly hands and feet, and dystonia.' Interosseous muscular atrophy in the hands in combination with the early-onset dystonic posturing leads to striking combined flexion-extension contractures of the fingers, which they illustrated. Mental retardation is not a feature of AT, although some older patients have a severe loss of short-term memory.

Neurologic dysfunction is a clinically invariable feature in homozygotes. Woods and Taylor (1992) studied 70 affected persons in the British Isles, 29 females and 41 males with an age range of 2 to 42 years. Most presented by 3 years of age with truncal ataxia. All had ataxia, ocular motor apraxia, an impassive face, and dysarthria, although clinical immune deficiency was present only in 43 of 70 patients. Ocular telangiectases was seen in all but one. All 60 tested showed increased sensitivity to ionizing radiation, 43 of 48 had an elevated alpha-fetoprotein level, and 14 of 21 had an immunoglobulin deficiency.

Malignancy

Patients with AT have a strong predisposition to malignancy. Hecht et al. (1966) observed lymphocytic leukemia in patients with AT. A nonleukemic sib and 2 unrelated patients with AT had multiple chromosomal breaks and impaired responsiveness to phytohemagglutinin. This was the first report of chromosomal breakage in AT. Leukemia and chromosomal abnormalities occur in at least 2 other mendelian disorders--Fanconi pancytopenia (FA; 227650) and Bloom syndrome (BS; 210900).

Saxon et al. (1979) demonstrated thymic origin of the neoplastic cells in a 48-year-old woman with AT and chronic lymphatic leukemia. The neoplastic cells had the specific 14q+ translocation and showed both helper and suppressor function, suggesting that the malignant transformation had occurred in an uncommitted T-lymphocyte precursor that was capable of differentiation. This is a situation comparable to chronic myeloid leukemia in which the Philadelphia chromosome occurs in a stem cell progenitor of both polymorphs and megakaryocytes.

In general, lymphomas in AT patients tend to be of B-cell origin (B-CLL), whereas the leukemias tend to be of the T-CLL type. Rosen and Harris (1987) discussed the case of a 30-year-old man with AT who developed a malignant lymphoma of B-cell type involving the tonsil and lungs.

Haerer et al. (1969) described a black sibship of 12, of whom 5 had ataxia-telangiectasia; 2 of those affected died of mucinous adenocarcinoma of the stomach at ages 21 and 19 years. Bigbee et al. (1989) demonstrated an increased frequency of somatic cell mutation in vivo in individuals with AT. Obligate heterozygotes for the disease did not appear to have a significantly increased frequency of such mutations. The authors speculated that the predisposition to somatic cell mutation may be related to the increased susceptibility to cancer in AT homozygotes. Other solid tumors, including medulloblastomas and gliomas, occur with increased frequency in AT (Gatti et al., 1991).

Immune Disorders

Defects of the immune mechanism and hypoplasia of the thymus have been demonstrated. Serum IgG2 or IgA levels are diminished or absent in 80% and 60% of patients, respectively (Gatti et al., 1991). IgE levels can be diminished, IgM levels diminished or normal. Peripheral lymphopenia as well as decreased cellular immunity to intradermally injected test antigens can be seen early in the disorder. Sinopulmonary infections are frequent, but their severity cannot be simply correlated with the degree of immunodeficiency.

Carbonari et al. (1990) found that patients with AT have more circulating T cells bearing gamma/delta receptors characteristic of immature cells than alpha/beta receptors typical of mature cells. Normal ratios were found in the patients with other immune deficits, except for 1 child with a primary T-cell defect. Peterson and Funkhouser (1990) proposed that these findings are consistent with a defect in genetic recombination leading to the switch from gamma/delta to alpha/beta. There may also be a defect in DNA ligation or some other aspect of DNA repair. Elucidation of the molecular abnormalities of lymphocytes may demonstrate fundamental molecular mechanisms for cellular differentiation not only of lymphocytes but of other cell systems such as the nervous system.

Variant Ataxia-Telangiectasia (Atypical)

Ying and Decoteau (1981) described a family in which a brother and sister may have had an allelic (and milder) form of AT. The proband, a 58-year-old male of Saskatchewan Mennonite origin, had spinocerebellar degeneration associated with choreiform movements beginning at about age 10 years. Despite considerable physical handicap, he was able to work as a delivery man in the family store. No telangiectases were found at age 44 (they were carefully sought because of typical AT in a niece) or on later examinations. He showed total absence of IgA in serum and concentrated saliva and low IgE in serum. He was anergic on skin testing. Glucose tolerance was markedly decreased. Serum alpha-fetoprotein was 840 ng per ml (normal, less than 10 ng per ml). Lymphocyte response to phytohemagglutinin was blunted. He died of lymphoma at age 58. He showed cytogenetic abnormalities typical of AT; 4 abnormal clones were identified, all involving chromosome 14 in some way. The proband had 4 brothers and 2 sisters. A brother died of leukemia at age 16. A sister was likewise diagnosed as having spinocerebellar degeneration with choreiform movements at age 46; she died at age 55 of breast cancer. The proband's niece with typical AT had telangiectases of the bulbar conjunctivae and earlobes noted at age 3, when she began to have recurrent and severe sinopulmonary infections. She died at age 20 of staphylococcal pneumonia superimposed on bronchiectasis. The brother and sister who died in their 50s may have been genetic compounds. Their parents denied consanguinity.

Taylor et al. (1987) described 3 patients who were atypical in terms of clinical features and cellular features as observed in vitro. One of the patients was a 45-year-old woman with onset of neurologic manifestations in her early twenties. Maserati et al. (1988) described 2 sisters, aged 9 and 11 years, with a progressive neurologic disorder similar to AT, chromosome instability with rearrangements involving chromosomes 7 and 14, but no telangiectases or immunologic anomalies typical of AT. Byrne et al. (1984) reported similar cases of ataxia without telangiectases with selective IgE deficiency but normal IgA and alpha-fetoprotein. Ziv et al. (1989) described 2 Turkish sibs with an atypically prolonged course and atypical behavior of cultured fibroblasts. See 208910 and 208920 for AT-like syndromes.

Rare cases of AT patients with milder manifestations of the clinical or cellular characteristics of the disease have been reported and have been designated 'AT variants.' Gilad et al. (1998) quantified ATM protein levels in 6 patients with an AT variant and searched their ATM genes for mutations. Cell lines from these patients exhibited considerable variability in radiosensitivity while showing the typical radioresistant DNA synthesis of AT cells. Unlike classic AT patients, however, these patients exhibited 1 to 17% of the normal level of ATM. The underlying genotypes were either homozygous for mutations expected to produce mild phenotypes or compound heterozygous for a mild and a severe mutation. In an attempt to determine whether the AT(Fresno) variation correlated with ATM mutations and levels of ATM protein expression, Gilad et al. (1998) searched for ATM mutations in a cell line derived from one of the sisters studied by Curry et al. (1989). This cell line was found to be devoid of the ATM protein and homozygous for a severe ATM mutation. Gilad et al. (1998) concluded that certain AT variant phenotypes, including some of those without telangiectasia, represent ATM mutations.

Saviozzi et al. (2002) noted that milder cases of AT, termed 'AT variants,' comprise a heterogeneous group characterized by later onset of clinical symptoms, slower progression, extended life span compared to most AT patients, and decreased levels of chromosomal instability and cellular radiosensitivity. In these patients, telangiectasia and/or immunodeficiency may be absent, while the neurologic features are present. The genotype of AT variants is most often compound heterozygous for a severe mutation together with a mild or leaky mutation, which expresses some ATM protein with residual function. In 2 sisters with variant AT with onset of ataxia at 27 years, polyneuropathy, choreoathetosis, and absence of telangiectasia, immunodeficiency, and cancer, Saviozzi et al. (2002) identified compound heterozygosity in the ATM gene for a missense (607585.0028) and a frameshift (607585.0029) mutation. Western blot analysis showed a low level of ATM protein with residual phosphorylation activity, which the authors suggested contributed to the milder phenotype.

Hiel et al. (2006) reported 3 brothers and an unrelated woman with late-onset AT. All 4 were ambulatory and ranged in age from 37 to 43 years; unsteady gait developed approximately 10 years earlier. Cerebellar signs were mild, but all had striking distal muscular atrophy and weakness, decreased or absent ankle reflexes, and normal or borderline delayed motor conduction velocities with markedly decreased compound muscle action potentials. Muscle biopsies showed neurogenic changes. The patients had normal sensation and normal sensory studies. Other features included severe resting tremor, slight intention tremor, and mild dysarthria. ATM phosphorylation activity was only slightly decreased, suggesting that other factors were involved in damage to anterior horn neurons.

Verhagen et al. (2009) provided a retrospective analysis of 13 adult patients with variant AT from 9 families and 6 unrelated patients with classic AT. All patients were from the Netherlands; 2 of the patients with variant AT had been reported by Hiel et al. (2006). All patients with classic AT were diagnosed in childhood, presented with ataxic gait, and were wheelchair-bound by age 11 years. Five of the 6 died between ages 21 and 27. Those with variant AT were only correctly diagnosed in adulthood, although 7 presented with slowly progressive chorea-athetosis from early childhood. Five with variant AT presented with resting tremor between age 12 to 34, and the remaining patient with variant AT presented with distal muscle weakness of the lower extremities at age 6. Five patients with variant AT became wheelchair-bound between ages 15 and 43, and 2 had died of malignancy at ages 51 and 23 years, respectively. All variant AT patients had dysarthria by adulthood, 9 had choreoathetosis, 8 had resting tremor, 7 had oculomotor apraxia, and 5 had nystagmus. Eight patients had normal cerebellum on MRI, whereas 4 had cerebellar atrophy. Only 7 of 13 had ocular telangiectasia, but all had increased serum alpha-fetoprotein. Six with variant AT had polyneuropathy. Four developed a malignancy, including ALL, pituitary tumor, and breast cancer. Only 1 had slightly decreased IgG levels. Chromosomal instability was found in 8 variant AT patients tested. Those with the mildest form of the disorder had residual ATM protein expression with kinase activity.

Saunders-Pullman et al. (2012) reported 13 patients from 3 Canadian Mennonite families with variant AT due to a homozygous missense mutation in the ATM gene (A2067D; 607585.0033). The patients had onset of dystonia in the first 2 decades (range, 1-20 years). Dystonia mostly affected the neck, face, tongue, and limbs, and became generalized in 60% of patients. Dysarthria was very common. Additional features in some patients included myoclonus, facial choreiform movements, and irregular tremor. Some patients had clumsy gait, and although none had overt ataxia, 2 patients had ataxia in childhood that spontaneously resolved. None had prominent telangiectases. Postmortem examination showed mild loss of cerebellar Purkinje cells in 1 patient, but cerebellar atrophy was not a prominent finding in any of the patients. Cells from 2 mutation carriers showed increased radiosensitivity and only trace amounts of ATM protein. Heterozygous mutation carriers did not have dystonia. Family history revealed that 2 homozygous mutation carriers in 1 family had died of malignancy in adulthood.

Cancer Risk in Heterozygotes

Welshimer and Swift (1982) studied families of homozygotes for AT, Fanconi anemia (FA), and xeroderma pigmentosum (XP; see 278700) to test the hypothesis that heterozygotes may be predisposed to some of the same congenital malformations and developmental disabilities that are common among homozygotes. Among XP relatives, 11 of 1,100 had unexplained mental retardation, whereas only 3 of 1,439 relatives of FA and AT homozygotes showed mental retardation. Four XP relatives but no FA or AT relatives had microcephaly. Idiopathic scoliosis and vertebral anomalies occurred in excess in AT relatives, while genitourinary and distal limb malformations were found in FA families.

Swift (1980) defended, from the viewpoint of not causing anxiety, the usefulness and safety of cancer risk counseling of heterozygotes for AT. Swift et al. (1987) examined the cancer risk of heterozygotes for AT in 128 families, including 4 of Amish ancestry, 110 white non-Amish families, and 14 black families. They measured documented cancer incidence rather than cancer mortality based solely on death certificates and compared the cancer incidence in adult blood relatives of probands directly with that in spouse controls. The incidence rates in AT relatives were significantly elevated over those in spouse controls. In persons heterozygous for AT, the relative risk of cancer was estimated to be 2.3 for men and 3.1 for women. Breast cancer in women was the cancer most clearly associated with heterozygosity for AT. Swift et al. (1987) estimated that 8 to 18% of patients with breast cancer in the U.S. white population would be heterozygous for AT. Pippard et al. (1988) reported an excess of breast cancer deaths in British mothers of AT patients (significant at the 5% level), but no excess mortality from malignant neoplasms in the grandparents.

Morrell et al. (1990) reported cancer incidence measured retrospectively in 574 close blood relatives of AT patients and 213 spouse controls in 44 previously unreported families. For heterozygous carriers of the AT gene, the relative risk of cancer was estimated to be 6.1 as compared with non-heterozygotes. The most frequent cancer site in the blood relatives was the female breast, with 9 cancers observed. Gatti et al. (1991) provided a review in which they noted the possibly high frequency of breast cancer in AT heterozygotes.

Swift et al. (1991) reported the results of a prospective study of 1,599 adult blood relatives of patients with AT and 821 of their spouses distributed in 161 families. Cancer rates were significantly higher among the blood relatives than in their spouses, specifically in the subgroup of 294 blood relatives who were known to be heterozygous for the AT gene. The estimated risk of cancer of all types among heterozygotes as compared with noncarriers was 3.8 in men and 3.5 in women, and that for breast cancer in carrier women was 5.1. Among the blood relatives, women with breast cancer were more likely to have been exposed to selected sources of ionizing radiation than controls without cancer. Male and female blood relatives also had 3-fold and 2.6-fold excess mortality from all causes, respectively, from the ages of 20 through 59 years. Swift et al. (1991) suggested that diagnostic or occupational exposure to ionizing radiation increases the risk of breast cancer in women heterozygous for AT. The work of Swift et al. (1991) on the frequency of breast cancer in AT was critiqued by numerous authors, including Bridges and Arlett (1992).

Since the genes responsible for most cases of AT are located on 11q, Wooster et al. (1993) typed 5 DNA markers in the AT region in 16 breast cancer families. They found no evidence for linkage between breast cancer and these markers and concluded that the contribution of AT to familial breast cancer is likely to be minimal.

Athma et al. (1996) determined the AT gene carrier status of 776 blood relatives in 99 AT families by tracing the ATM gene in each family through tightly linked flanking DNA markers. There were 33 women with breast cancer who could be genotyped; 25 of these were AT heterozygotes, compared to an expected 14.9. For 21 breast cancers with onset before age 60, the odds ratio was 2.9 and for 12 cases with onset at age 60 or older, the odds ratio was 6.4. Thus, the breast cancer risk for AT heterozygous women is not limited to young women but appeared to be even higher at older ages. Athma et al. (1996) estimated that, of all breast cancers in the U.S., 6.6% may occur in women who are AT heterozygotes. This proportion is several times greater than the estimated proportion of carriers of BRCA1 mutations (113705) in breast cancer cases with onset at any age.

The reported increased risk for breast cancer for AT family members has been most evident among younger women, leading to an age-specific relative risk model predicting that 8% of breast cancer in women under age 40 arises in AT carriers, compared with 2% of cases between 40 and 59 years (Easton, 1994). To test this hypothesis, FitzGerald et al. (1997) undertook a germline mutational analysis of the ATM gene in a population of women with early onset of breast cancer, using a protein truncation (PTT) assay to detect chain-terminating mutations, which account for 90% of mutations identified in children with AT. They detected a heterozygous ATM mutation in 2 of 202 (1%) controls, consistent with the frequency of AT carriers predicted from epidemiologic studies. ATM mutations were present in only 2 of 401 (0.5%) women with early onset of breast cancer (P = 0.6). FitzGerald et al. (1997) concluded that heterozygous ATM mutations do not confer genetic predisposition to early onset of breast cancer.

The results of FitzGerald et al. (1997) are discrepant with those of Athma et al. (1996), who conducted a study 'from the other direction' by following identified AT mutations through the families of those with clinically recognized AT. Analysis of DNA markers flanking the AT gene allowed them to identify precisely which female relatives with breast cancer carried the AT mutation. On the basis of the genetic relationship between each case and the AT proband, the a priori probability that these 2 share the AT mutation was calculated. This led to an estimated relative risk of 3.8 as compared to noncarriers. This result was similar to that found by Easton (1994), who reanalyzed the previous studies of breast cancer risk in mothers (and other close relatives) of AT cases. Bishop and Hopper (1997) analyzed these 2 studies and suggested that they may not be discrepant. Indeed, they estimated that the study of FitzGerald et al. (1997) yielded an upper limit of the 95% confidence interval for the proportion of early onset breast cancer occurring in AT heterozygotes as 2.4% (assuming that their assay identified 75% of all mutations).

In a family with multiple cancers, Bay et al. (1999) described heterozygosity for a mutant allele of ATM that caused skipping of exon 61 in the mRNA (607585.0020) and was associated with a previously undescribed polymorphism in intron 61. The mutation was inherited by 2 sisters, one of whom developed breast cancer at age 39 years and the second at age 44 years, from their mother, who developed kidney cancer at age 67 years. Studies of irradiated lymphocytes from both sisters revealed elevated numbers of chromatid breaks, typical of AT heterozygotes. In the breast tumor of the older sister, loss of heterozygosity (LOH) was found in the ATM region of 11q23.1, indicating that the normal ATM allele was lost in the breast tumor. LOH was not seen at the BRCA1 (113705) or BRCA2 (600185) loci. BRCA2 was considered an unlikely cancer-predisposing gene in this family because each sister inherited different chromosomes 13 from each parent. The findings suggested that haploinsufficiency at ATM may promote tumorigenesis, even though LOH at the ATM locus supported a more classic 2-hit tumor suppressor gene model.

The finding that ATM heterozygotes have an increased relative risk for breast cancer had been supported by some studies but not confirmed by others. Broeks et al. (2000) analyzed germline mutations of the ATM gene in a group of Dutch patients with breast cancer using normal blood lymphocytes and the protein truncation test followed by genomic sequence analysis. A high percentage of ATM germline mutations was demonstrated among patients with sporadic breast cancer. The 82 patients included in this study had developed breast cancer before the age of 45 years and had survived 5 years or more (mean, 15 years), and in 33 (40%) of the patients a contralateral breast tumor had been diagnosed. Among these patients, 7 (8.5%) had germline mutations of the ATM gene, of which 5 were distinct. One splice site mutation, IVS10-6T-G (607585.0021), was detected 3 times in this series. Four heterozygous carriers had bilateral breast cancer. Broeks et al. (2000) concluded that ATM heterozygotes have an approximately 9-fold increased risk of developing a type of breast cancer characterized by frequent bilateral occurrence, early age at onset, and long-term survival. They suggested that the characteristics of this population of patients may explain why such a high frequency was found here and not in other series.

Olsen et al. (2005) reported on an extended and enlarged follow-up study of cancer incidence in blood relatives of 75 patients with verified AT from 66 Nordic families. When 7 mothers of probands were excluded, no clear relationship was observed between the allocated mutation carrier probability of each family member and the extent of breast cancer risk. They concluded that the increased risk for female breast cancer seen in 66 Nordic AT families appeared to be restricted to women under the age of 55 years and was due mainly to a very high risk in the group of mothers. Olsen et al. (2005) concluded that the findings of breast cancer risk in mothers, but not in other likely mutation carriers, in this and other studies raised questions about the hypothesis of a simple causal relationship with ATM heterozygosity.

Although the defining characteristic of recessive diseases is the absence of a phenotype in heterozygous carriers, Watts et al. (2002) suggested that expression profiling by microarray techniques might reveal subtle manifestations. Individual carriers of AT cannot be identified; as a group, however, carriers of a mutant AT allele have a phenotype that distinguishes them from normal control individuals: increased radiosensitivity and risk of cancer. Watts et al. (2002) showed that the phenotype was also detectable, in lymphoblastoid cells from AT carriers, as changes in expression level of many genes. The differences were manifested both in baseline expression levels and in response to ionizing radiation. The findings showed that carriers of the recessive disease may have an 'expression phenotype,' which suggested a new approach to the identification of carriers and enhanced understanding of their increased cancer risk.

Renwick et al. (2006) screened individuals from 443 familial breast cancer pedigrees and 521 controls for ATM sequence variants and identified 12 mutations in affected individuals and 2 in controls (p = 0.0047). Their results demonstrated that ATM mutations that cause ataxia-telangiectasia in biallelic carriers are breast cancer susceptibility alleles in monoallelic carriers, with an estimated relative risk of 2.37 (95% CI = 1.57-3.78, p = 0.0003).

Waldmann and McIntire (1972) showed raised alpha-fetoprotein in the blood of patients with AT. This, they felt, suggests immaturity of the liver and is consistent with the view that the primary defect is in tissue differentiation, specifically, a defect in the interaction necessary for differentiation of gut-associated organs such as the thymus and liver. Ishiguro et al. (1986) concluded that the elevated alpha-fetoprotein in patients with AT probably originates in the liver.

On the circulating monocytes of AT patients, Bar et al. (1978) demonstrated an 80 to 85% decrease in insulin receptor affinity. This decrease was not observed in the cultured fibroblasts of AT patients or in the monocytes and fibroblasts of relatives of these patients. In addition, they found that whole plasma and immunoglobulin-enriched fractions of plasma from AT patients inhibited the normal binding of insulin to its receptors on cultured human lymphocytes and on human placental membranes. This suggested the presence of antireceptor immunoglobulins. AT and type B acanthosis nigricans have several features in common that suggest the possibility of similar causes for the insulin resistance each demonstrates.

Shaham and Becker (1981) showed that the AT clastogenic (chromosome breaking) factor present in plasma of AT patients and in the culture medium of AT skin fibroblasts is a peptide with a molecular weight in the range of 500 to 1000. No clastogenic activity could be demonstrated in extracts of cultured AT fibroblasts.

Mohamed et al. (1987) found marked reduction of topoisomerase II (126430) in some but not all AT cell lines. DNA topoisomerases I and II are enzymes that introduce transient single- and double-strand breaks into DNA and thus are capable of interconverting various DNA conformations. The isolation of mutants of the 2 enzymes in yeast and the increased levels of DNA topoisomerase II in cells undergoing DNA synthesis provide evidence for the role of these enzymes in DNA replication and in chromosome segregation and organization.

In a study of 47 families ascertained throughout the United Kingdom, Woods et al. (1990) found a low parental consanguinity rate; no parents were first cousins or more closely related, whereas 10% had been expected. Furthermore, the incidence of the disorder in 79 sibs of index cases was 1 in 7, rather than the expected 1 in 4.

The presence of early-onset ataxia with oculocutaneous telangiectases permits diagnosis of AT. The clinical diagnosis of AT can be problematic before the appearance of telangiectases. Oculomotor apraxia is a useful aid to early clinical diagnosis. Early-onset cerebellar ataxia and oculomotor apraxia are also typical of X-linked Pelizaeus-Merzbacher disease (312080) and can be seen in Joubert syndrome (213300). These disorders can be distinguished by leukoencephalopathy in the former, and by profound cerebellar hypoplasia in the latter. See also 257550. Elevated levels of alpha-fetoprotein (126430) and carcinoembryonic antigen are the most useful readily available markers for confirmation of the diagnosis of AT (Gatti et al., 1991). Dysgammaglobulinemia, decreased cellular immune responses, and peripheral lymphopenia are supportive findings but are not invariable.

Henderson et al. (1985) devised a rapid diagnostic method based on the hypersensitivity of AT lymphocytes to killing by gamma irradiation. Similar studies in fibroblasts require skin biopsy and a prolonged culture time. Llerena et al. (1989) concluded that in chorionic villus sampling, gamma radiation is a reliable way of discriminating between unaffected fetuses and those with AT. The reliability of this approach is in question, however. Painter and Young (1980) suggested that the radiosensitivity of AT cells may be caused by their failure to respond to DNA damage with a delay in DNA synthesis that could give time for repair to take place.

Shiloh et al. (1989) presented evidence that the extent of chromatid damage induced in the G2 phase of the cell cycle by moderate dosage of x-rays is markedly higher in AT heterozygous cells than in normal controls. They used this as a test of heterozygosity.

Rosin and Ochs (1986) applied the exfoliated cell micronucleus test to the question of in vivo chromosomal instability in AT. This test is performed on exfoliated cells from the oral cavity collected by swabbing the mucosa with a moistened tongue depressor and also on urinary bladder cells obtained by centrifugation of freshly voided urine specimens. Micronuclei in these cells result from fragmentation of chromosomes in the dividing cells from the epithelium, resulting in acentric fragments which are excluded from the main nucleus when the cell divides. These fragments form their own membrane and can be identified as extranuclear Feulgen-positive bodies in daughter cells which migrate up through the epithelium to be exfoliated. Rosin and Ochs (1986) found that AT homozygotes had a 5- to 14-fold increase in the frequency of exfoliated cell micronuclei. Heterozygotes could be reliably identified by this method (Rosin et al., 1989).

Using X-radiation with 1 Gy on G2-phase lymphocytes from 7 AT patients, 13 obligate AT heterozygotes, and 14 normal controls, Tchirkov et al. (1997) found that both AT homozygotes and heterozygotes showed significantly increased levels of radiation-induced chromatid damage relative to that of normal controls.

Patients with AT and their cultured cells are unusually sensitive to x-ray just as patients and cells with xeroderma pigmentosum are sensitive to ultraviolet. Treatment of malignancy with conventional dosages of radiation can be fatal to AT patients.

Oxford et al. (1975) found that chromosome 14 was often involved in rearrangements in AT and that band 14q12 was a highly specific exchange point. In addition to the changes in chromosome 14, a pericentric inversion of chromosome 7 is characteristic. McCaw et al. (1975) described t(14;14)(q11;q32) translocation in T-cell malignancies of patients with AT. T cells show a t(14;14)q12q32 rearrangement in about 10% of AT patients.

Croce et al. (1985) assigned the alpha subunit of the T-cell antigen receptor (TCRA; see 186880) to the region of one of the common breakpoints in AT (14q11.2) and suggested that the oncogene TCL1 (186960) is located in the region of the other breakpoint (14q32.3). It is thought that the TCL1 gene may be activated by chromosome inversion or translocation, either of which results in juxtaposition of the TCL1 gene and the TCRA gene. In AT, circulating lymphocytes show characteristic rearrangements involving the site of the T-cell receptor gamma gene (7p15) (TCRG; see 186970), T-cell receptor beta genes (7q35) (TCRB; see 186930), T-cell receptor alpha genes (14q11), and immunoglobulin heavy chain genes (14q32) (IGHG1; 147100) (McFarlin et al., 1972; Ying and Decoteau, 1981).

Aurias et al. (1986) described a possible 'new' type of chromosome rearrangement, namely, telomere-centromere translocation (tct) followed by double duplication. This type of rearrangement was found between chromosomes 7 and 14 in cases of AT. Gatti et al. (1985) and Aurias and Dutrillaux (1986) found that the sites of breaks in rearrangements (7p14, 7q35, 14q12, 14qter, 2p11, 2p12, and 22q11-q12) are those where members of the immunoglobulin superfamily are located: IGK, IGH, IGL, TCRA, TCRB, TCRG. The somatic gene rearrangement must precede expression of these genes.

Kennaugh et al. (1986) studied a patient with an inversion of 14q which had been present for many years in T cells. It was found that the breakpoint in 14q32 lay outside the IgH locus and proximal to it. The constant region gene of the T-cell receptor alpha chain (TCRA) locus was translocated to the 14q32 position. Johnson et al. (1986) found that the 14q32 breakpoint in the 14/14 translocation found in T-CLL cells and in an AT patient occurred within the immunoglobulin gene cluster. The AT patient had the characteristic chromosome 14 tandem translocation in 100% of karyotyped T cells 10 years before her death from T-cell leukemia. (This was the same patient described earlier by Saxon et al. (1979).) Stern et al. (1988) used in situ chromosomal hybridization to map the TCRA gene in 3 different nonmalignant T-cell clones derived from patients with AT. The constant region was translocated in each clone; the variable region remained in its original position in 2 clones and was deleted in 1 which lost the derivative chromosome 14.

Stern et al. (1988) mapped the 14q32.1 recurrent breakpoint of AT clones by in situ hybridization. They found that the breakpoint lay between D14S1 (107750) and PI (107400). In a t(14;14) clone they found an interstitial duplication including D14S1 and a part of the IGH locus. Studying the chromosomes by R-banding, Zhang et al. (1988) concluded that the distal breakpoint in the chromosome 14 inversion in an AT clone was different from that in the chromosome 14 inversion in a malignant T-cell line; specifically, in AT, the breakpoint was centromeric to both the immunoglobulin heavy chain locus and the D14S1 anonymous locus (107750). They suggested that this finding favors the existence of an unknown oncogene in band 14q32.1.

Russo et al. (1989) presented evidence for a cluster of breakpoints in the 14q32.1 region, the site of the putative oncogene TCL1, in cases of ataxia-telangiectasia with chronic lymphocytic leukemia. The 14q32.1 breakpoint is at least 10,000 kb centromeric to the immunoglobulin heavy chain locus. In a cell line with a translocation t(14;14)(q11;q32) from an AT patient with T-cell chronic lymphocytic leukemia, Russo et al. (1989) showed that a J(alpha) sequence from the TCRA locus was involved. This was again the patient first reported by Saxon et al. (1979). Humphreys et al. (1989) found some rearrangements involving chromosomes 7 and 14 at the usual 4 sites associated with AT--7p14, 7q35, 14q12, and 14q32--all sites of T-cell receptor genes.

Kojis et al. (1989) suggested that the very high frequency of lymphocyte-associated rearrangements (LARs) in peripheral blood chromosome preparations is a diagnostic criterion of the disease. They pointed out a striking difference in the types of rearrangements observed in lymphocytes and fibroblasts. LARs are not commonly observed in fibroblasts, despite the increased but random instability of chromosomes from these cells relative to lymphocytes. The region of location of the AT gene, 11q22-q23, is not involved in site-specific rearrangements in either lymphocytes or fibroblasts.

Lipkowitz et al. (1990) showed that an abnormal V(D)J recombination, joining V segments of the T-cell receptor gamma gene (186970) with J segments of the T-cell receptor beta gene (186930), occurs in peripheral blood lymphocytes of AT patients at a frequency 50- to 100-fold higher than normal. This frequency is roughly the same as the increase in the risk for lymphoid malignancy in these individuals. There is also an increase in the frequency of the lymphocyte-specific cytogenetic abnormalities thought to be due to interlocus recombination in non-AT patients with non-Hodgkin lymphoma, further suggesting a relationship between these translocations and lymphoid malignancies. Agriculture workers occupationally exposed to pesticides used in the production and storage of grain have a high frequency of cytogenetic abnormalities in peripheral blood lymphocytes in a pattern reminiscent of those in AT patients. Furthermore, these agriculture workers have an increased risk of developing T- and D-lymphoid malignancies. Lipkowitz et al. (1992) used a PCR-based assay developed for the study of AT patients to demonstrate a 10- to 20-fold increased frequency of hybrid antigen-receptor genes in peripheral blood lymphocytes of agriculture workers with chemical exposure.

By linkage to RFLP markers, Gatti et al. (1988) localized the AT gene to 11q22-q23. They had previously excluded 171 markers, comprising approximately 35% of the genome. The most promising marker in a large Amish pedigree was found to be THY1 (188230), which is located at 11q22.3; it showed linkage with maximum lod = 1.8 at theta = 0.00. When data from the other 4 informative group A AT families were added, the maximum lod score rose to 3.63 with no observed recombinants. The maximum lod score for all 31 families studied for linkage of AT to THY1 was 4.34 at theta = 0.10. The large Amish pedigree diagrammed in their Figure 1 is the kindred reported by McKusick and Cross (1966), Ginter and Tallapragada (1975), and Rary et al. (1975). By further mapping with a panel of 10 markers, Sanal et al. (1990) concluded that the AT locus is in band 11q23.

The site of the AT1 gene (11q22-q23) is the same as or adjacent to the region occupied by the CD3 (186790), THY1, and NCAM (116930) genes, all of which are members of the immunoglobulin-gene superfamily and therefore may be subject to the same defect that afflicts the T-cell receptor and immunoglobulin molecules in AT. Concannon et al. (1990) excluded the AT1 gene from a region extending 15 cM to either side of ETS1 (164720), which maps to 11q24. According to Gatti (1990), the gene in families from complementation groups A, C, and D, representing approximately 97% of all families, has been mapped to 11q23. Thus, a single gene may exist with various intragenic defects permitting complementation.

In studies of 35 consecutively obtained families in the British Isles, McConville et al. (1990) found support for linkage with THY1 at zero recombination. They found evidence suggesting a second AT locus on 11q, centromeric to the site previously postulated. With 3 exceptions, the families had not been assigned to complementation groups. The series of families included the only group E family described to date. They quoted Jaspers et al. (1988) as giving the proportion of group A, group C, and group D cases as approximately 56%, 28%, and 14%, respectively.

By linkage studies in a Jewish-Moroccan family with AT of the group C type, Ziv et al. (1991) found that the disorder was linked to the same region (11q22-q23) as found in group A families. McConville et al. (1990) located the AT1 gene to a 5-cM region in 11q22-q23, flanked by NCAM and DRD2 (126450) on one side and STMY1 (185250) on the other.

On the basis of an 18-point map of the 11q23 region of 11q, derived from linkage analysis of 40 CEPH families, Foroud et al. (1991) analyzed 111 AT families from Turkey, Israel, England, Italy, and the United States, localizing the gene to an 8-cM sex-averaged interval between the markers STMY1 and D11S132/NCAM. Ziv et al. (1992) obtained results from linkage study indicating that the ATA gene in 3 large Arab families was located in 11q23. However, in a Druze family unassigned to a specific complementation group, several recombinants between AT and the same markers were observed.

Sobel et al. (1992) pointed to linkage evidence suggesting that there are 2 AT loci on 11q and that group D AT may be located distal to the site of groups A and C in the 11q23 region.

In linkage studies of 14 Turkish families, 12 of which were consanguineous, Sanal et al. (1992) obtained results indicating that the most likely location for a single AT locus is within a 6-cM sex-averaged interval defined by STMY and the marker CJ77. However, it appeared that there are at least 2 distinct AT loci (ATA and ATD) at 11q22-q23, with perhaps a third locus, ATC, located very near the ATA gene.

Hernandez et al. (1993) described a large inbred family in which 2 adult cousins had AT with a somewhat milder clinical course than usual. Since genetic linkage analysis did 'not provide any evidence that the gene for AT in this family is located at 11q22-23,' further locus heterogeneity was suggested.

In 2 families clinically diagnosed with AT and previously reported by Hernandez et al. (1993) and Klein et al. (1996), respectively, Stewart et al. (1999) identified mutations in the MRE11A gene (600814). Consistent with the clinical outcome of these mutations, cells established from the affected individuals within the 2 families exhibited many of the features characteristic of both AT and Nijmegen breakage syndrome (251260), including chromosomal instability, increased sensitivity to ionizing radiation, defective induction of stress-activated signal transduction pathways, and radioresistant DNA synthesis. The authors designated the disorder ATLD, for AT-like disorder (604391). Because the MRE11A gene maps to 11q21 and the ATM gene maps to 11q23, Stewart et al. (1999) concluded that only a very detailed linkage analysis would separate ATLD from AT purely on the basis of genetic data. Assuming that the mutation rate is proportional to the length of the coding sequences of the 2 genes, they suggested that approximately 6% of AT cases might be expected to have MRE11A mutations.

Gatti et al. (1993) reported prenatal genotyping in this disorder. They pointed out that although at least 5 complementation groups have been defined, linkage studies of more than 160 families from various parts of the world have failed to show linkage heterogeneity. All but 2 families were linked to a 6-cM (sex-averaged) region at 11q22.3 defined by the markers STMY1 and D11S385. A further analysis of 50 British families narrowed the localization to a 4-cM (sex-averaged) region defined by D11S611 and D11S535. The demonstrated complementation groups may represent different intragenic mutations or separate ataxia-telangiectasia genes clustered within the 11q22.3 region, neither of which would challenge the validity of linkage or haplotyping studies. A possible reinterpretation of the complementation data is that the radiosensitivity of AT fibroblasts can be complemented by many genes besides the AT gene or genes. Gatti et al. (1993) used the flanking markers to show that the haplotypes in a fetus were identical to those in a previously born affected child. The parents chose to continue the pregnancy.

Complementation Groups

On the basis of complementation studies of DNA repair in cultured fibroblasts, Paterson et al. (1977) suggested