Wilson disease (WD) is an autosomal recessive copper transport disorder caused by mutations of ATP7B, a liver cell copper-transporting P-type ATPase (Cuthbert, 1998). Over 200 different mutations of ATP7B have been described in WD (Kenney & Cox, 2000). Although most patients are identified because of liver disease and/or neurological complications, a small minority of affected individuals can present with haematological, renal, cardiovascular, musculoskeletal or endocrine abnormalities. A cardinal feature of WD is that it is a lethal condition in the absence of adequate treatment. Thus, identifying WD at a presymptomatic or early clinical stage has enormous implications for the outcome of patients with this condition (Cuthbert, 1998).
We describe a 19-year-old woman with haemolytic anaemia and thrombocytopenia as the initial manifestation of Wilson disease (WD). There are two reasons for reporting such an improbable case. First, it emphasizes the importance of recognizing atypical clinical presentations of potentially lethal recessive traits for which therapy is available. Second, it shows that, even in a monogenic disorder like WD, the phenotype cannot be extrapolated from the mutated genotype in a simple fashion; this patient had a relatively late-onset form of WD despite homozygosity for a genetic lesion leading to an apparent complete loss of function of the WD copper transporter.
A 19-year-old pharmacy student from Catania, Sicily, born to apparently healthy parents who were first cousins, was admitted to a local hospital in August 1998 because of a 2-week history of weakness, menometrorrhagia and dark urine. Her temperature was 37·3°C, she was icteric and had a painful right hypochondrium. A petechial rash was present on her legs. Laboratory testing showed haemolytic anaemia (Hb 8·6 g/dl, reticulocytes 272 × 109/l, indirect bilirubin 118 μmol/l, lactate dehydrogenase (LDH) 520 U/l, haptoglobin < 0·32 g/l), and normal liver and renal function. The platelets were 180 × 109/l. Haemoglobin electrophoresis showed a normal pattern, the Coombs test was negative, as were studies for erythrocyte enzymopathies. A few days later, the patient became more anaemic (Hb 6·4 g/dl) and thrombocytopenic (65 × 109/l), and schistocytes were noticed on a blood film. A diagnosis of thrombotic thrombocytopenic purpura was made and she was treated for the next 35 d by daily plasma exchange.
In October 1998, she was transferred to another hospital. Several haematomas were seen on the lower limbs. The anaemia (Hb 10·7 g/dl) and platelet count (100 × 109/l) had improved. The reticulocyte count was 137 × 109/l and the Coombs test was negative. A bone marrow aspirate showed normal cellularity with increased numbers of erythroid precursors and megakaryocytes. As bilirubin and LDH levels were repeatedly in the normal range and schizocytes were undetectable on blood films, it was felt that the diagnosis of thrombotic thrombocytopenic purpura was improbable. Also, paroxysmal nocturnal haemoglobinuria (negative Ham test) and various viral infections were ruled out. A computerized tomographic scan of the abdomen showed that the largest diameter of the spleen was 16 cm. At this stage, the patient was considered to have autoimmune haemolytic anaemia and thrombocytopenia.
On November 9, 1998, the patient consulted at the Centre Hospitalier Universitaire Vaudois. She was extremely weak. The tip of the spleen was palpable. Hb was 12·8 g/dl and the platelet count was 100 × 109/l. The Coombs test was negative. The expression of the glycosyl–phosphatidylinositol-anchored receptors CD55, CD58 and CD59 on the patient erythrocytes and lymphocytes was within the normal range. While being evaluated, she had an attack of biliary colic. Abdominal ultrasonography showed a dilated gallbladder and the presence of biliary sludge. Liver tests showed evidence of cholestatic hepatitis and a prolonged prothrombin time. At this point, she was investigated for WD. This diagnosis was established by the presence of low serum ceruloplasmin (0·05 g/l; normal 0·20–0·60 g/l), increased urinary copper excretion (26·5 μmol/24 h; normal < 0·8 μmol/24 h) and Kayser–Fleischer rings. The patient was immediately started on the copper chelator trientine. In March 1999, she had resumed a normal lifestyle although thrombocytopenia and liver function abnormalities were still detectable. These abnormalities improved progressively under continuous trientine therapy. In June 2000, her clinical state and performance status were excellent, Kayser–Fleischer rings had disappeared, and haematological values and liver tests were within the normal range.
Genetic analysis and discussion
The patient, her parents and her sister were evaluated for mutations in ATP7B. Sequencing of exons 5, 6, 8, 10, 12, 13, 14, 15, 18 and 19 showed that the patient was homozygous for C→T at position 3955 of ATP7B (Arg1319Stop), while her parents were heterozygous for this mutation and her sister had a wild-type genotype (Fig 1). No other mutation was detected in the 10 exons of ATP7B that were sequenced. C3955T in ATP7B has been observed before in a few WD patients with compound heterozygosity, but not in a homozygous individual (Kenney & Cox, 2000). This mutation results in premature termination of ATP7B synthesis at Arg1319, a residue that is located just before transmembrane helix Tm7 (Melchers et al, 1999). The structure and function of the protein domains homologous to the Tm7 and Tm8 transmembrane helices of WD-ATPase have been extensively studied in Na+/K+- and Ca2+-ATPases (Vilsen et al, 1997). These studies and the recently published structure of sarcoplasmic reticulum Ca2+-ATPase (Toyoshima et al, 2000) show that the cation binding sites of these P-type ATPases is largely made of these two segments. From this data and the assumption that ATP7B has a cation transport mechanism closely related to that of the Na+/K+-and Ca2+-ATPases, it can be predicted that deletion of Tm7 and Tm8 in ATP7B will inevitably lead to a protein unable of cation translocation, i.e. that C3955T in ATP7B is a null mutation.
This case illustrates the important notion that young adults with Coombs-negative haemolytic anaemia of unclear origin should be investigated for WD (Dacie, 1995). Haemolysis seems to result from erythrocyte membrane oxidative damage (Forman et al, 1980). This mechanism may also be responsible for thrombocytopenia. In the patient described here, haemolysis and thrombocytopenia were initially interpreted as evidence of thrombotic thrombocytopenic purpura. Initial therapy during her first hospitalization included intensive plasmapheresis. This intervention improved haemolysis, as shown by an increase in Hb level from 6·4 to 10·7 g/dl. This effect is probably a consequence of copper level reduction (Sarles et al, 1992). However, it is critical to emphasize that lifelong copper chelation is the sole long-term therapy of WD that can prevent irreversible tissue damage (Cuthbert, 1998).
This patient had a relatively late-onset form of WD. She became clinically ill at the age of 19·4 years, while most patients are already symptomatic during mid-adolescence, some of them even during childhood. However, as seen with several other Mendelian traits (Hubbard & Lewontin, 1996; Scriver & Waters, 1999), genotype–phenotype correlation studies in WD have indicated that phenotypic manifestations are often not as predictable from the genotype as would be expected. Indeed, suggestion has been made that, apart from WD-ATPase, additional factors (genetic, epigenetic, environmental) have important roles in normal copper homeostasis (Cuthbert, 1998; Schiefermeier et al, 2000). The observation that relatively late-onset WD can be observed in the context of homozygosity for a null mutation of ATP7B provides strong support for this notion. Clearly, it is interesting to note that this genetic lesion was not accompanied by childhood liver failure, the phenotype usually encountered in severe WD, and that haemolytic anaemia and thrombocytopenia was the initial clinical manifestation.