Unexplained aplastic anaemia, immunodeficiency, and cerebellar hypoplasia (Hoyeraal-Hreidarsson syndrome) due to mutations in the dyskeratosis congenita gene, DKC1


Dr Inderjeet Dokal, Department of Haematology, Imperial College School of Medicine, Hammersmith Campus, DuCane Road, London W12 ONN. e-mail: i.dokal@rpms.ac.uk


Hoyeraal-Hreidarsson (HH) syndrome is a multisystem disorder affecting boys characterized by aplastic anaemia (AA), immunodeficiency, microcephaly, cerebellar-hypoplasia and growth retardation. Its pathogenesis is unknown. X-linked dyskeratosis congenita (DC) is an inherited bone-marrow-failure syndrome characterized by skin pigmentation, nail dystrophy and leucoplakia which usually develop towards the end of the first decade of life. AA occurs in >90% of cases of DC. We speculated that mutations in the gene responsible for X-linked DC (DKC1) may account for the HH syndrome, due to the phenotypic similarities between the disease in respect of AA and gender bias. We therefore analysed the DKC1 gene in two HH families. In one family a nucleotide change at position 361(A → G) in exon 5 was found in both affected brothers; in the other family a nucleotide change at position 146(C → T) in exon 3 was found in the affected boys. The finding of these two novel missense DKC1 mutations demonstrates that HH is a severe variant of DC. They also show that mutations in DKC1 can give rise to a very wide clinical spectrum of manifestations. Boys with unexplained AA or immunodeficiency should be tested for mutations in DKC1 even though they may lack diagnostic features of DC.

Hoyeraal-Hreidarsson (HH) syndrome (MIM 600545) is a severe multisystem disorder affecting boys characterized by microcephaly, cerebellar hypoplasia, growth retardation of prenatal onset, and aplastic anaemia (AA) (Hoyeraal et al, 1970; Hreidarsson et al, 1988; Berthet et al, 1994; Aalfs et al, 1995; Ohga et al, 1997; Nespoli et al, 1997). Recently, progressive combined immunodeficiency has become a recognized feature of the syndrome since immunoglobulin deficiency has been documented in 4/7 published cases with 3/4 patients investigated also manifesting lymphocyte abnormalities (lymphopenia, B-cell depletion, T-cell dysfunction) (Berthet et al, 1994; Aalfs et al, 1995; Ohga et al, 1997; Nespoli et al, 1997). The pathogenesis of HH is unknown.

X-linked recessive DC (MIM 305000) is an inherited bone marrow failure syndrome characterized by the mucocutaneous triad of skin pigmentation, nail dystrophy and leucoplakia which usually develops towards the end of the first decade of life (Drachtman & Alter, 1995; Dokal, 1996). Aplastic anaemia occurs in >90% of cases and is the principal cause of early mortality. DC is clinically very heterogenous and a given individual may also have a variable number of other somatic (dental, gastrointestinal, genitourinary, neurological, ophthalmic, pulmonary, skeletal) abnormalities (Knight et al, 1998). Recently we have shown that X-linked DC is caused by mutations in a highly conserved gene (DKC1) encoding a nucleolar protein (dyskerin) with a putative role in ribosome biogenesis (Heiss et al, 1998). The observation that HH has been reported only in boys and the presence of aplastic anaemia in these patients raised the possibility that HH may be due to mutations in DKC1.



To test whether HH was allelic to DC we analysed the DKC1 gene in two HH families, one of which (family A) is the HH family described by Aalfs et al (1995) (Table I, case 1). Since that publication (Aalfs et al, 1995) the original patient (case 1) has died from complications of Staphylococcus aureus infection aged 67 months. At post mortem he was found to have pneumonia, laryngitis, foci of demyelination throughout the cerebral hemispheres and atrophy of the cerebellum. His younger brother (case 2) has also developed all the features of HH. He was born in November 1995. He had intrauterine growth retardation (IUGR; birth weight 2.7 kg), microcephaly, developmental delay, hypoplasia of the cerebellar vermis on CT scan of head, and hypocellular bone marrow with an abnormal blood count (Hb 6.2 g/dl, WBC 5.4 × 109/l, platelets 74 × 109/l, high haemoglobin F; 11%). In December 1998 (aged 37 months), his immunoglobulin levels were normal and metabolic studies including plasma amino acids, lactate and pyruvate were within the normal range. Fresh blood samples for this study were obtained from case 2 and both parents. Frozen skin fibroblasts were used as a source of DNA for case 1 (Table I). We were unable to obtain patient material from any of the other published cases (Table I) (Hoyeraal et al, 1970; Hreidarsson et al, 1988; Berthet et al, 1994; Ohga et al, 1997; Nespoli et al, 1997) of HH syndrome.

Table 1. Table I. Characteristics of patients with Hoyeraal-Hreidarsson syndrome. +, Present; −, absent; ?, not described or not investigated; AA, aplastic anaemia; H, haemorrhage; I, infection.* (a) Hoyeraal et al, 1970; (b) Hreidarsson et al, 1988; (c) Berthet et al, 1994; (d) Ohga et al, 1997; (e) Nespoli et al, 1997.† Case 1 is the HH patient published previously by Aalfs et al (1995).Thumbnail image of

The second family (family B) is a Caucasian (English) family in which there were three affected boys: case 3, case 4, and the index case 5 (II.1, III.2 and III.1 respectively in Fig 2b). Cases 4 and 5 were first cousins and case 3 was a maternal cousin of the mothers of cases 4 and 5. Case 3 was born in August 1961 and had IUGR (birth weight 2.3 kg). He had repeated infections, persistent diarrhoea and failure to thrive (FTT). At 35 months of age he was transferred to the Hospital for Sick Children at Great Ormond Street (London), with pneumonia and was found to have hypogammaglobulinaemia and pancytopenia (Hb 8.5 g/dl, WBC 2.5 × 109/l, and platelets 35 × 109/l). Despite treatment with blood products and antibiotics he died aged 36 months. Post-mortem revealed bronchopneumonia and an atrophic thymus but no definitive diagnosis was made. Case 4 presented with FTT, recurrent infections and chronic diarrhoea with enterocolitis requiring a colectomy. At 18 months of age he was noted to have hypogammaglobulinaemia with absent IgG and IgA and low IgM (0.3 g/l, normal range 0.5–1.6) associated with lymphopenia and low B cells (total lymphocyte count 0.6 × 109/l; >96% CD3 positive, <2% B cells). He died in 1984 from pulmonary infection aged 32 months; no post-mortem studies were done and no definitive diagnosis was made. Case 5, who is the index case in this family, presented aged 7 months with respiratory symptoms and was found to have Pneumocystis carinii pneumonia (PCP). He had a history of severe IUGR (birth weight 1.5 kg at 35 weeks gestation), FTT, recurrent diarrhoea and microcephaly. He was found to have thrombocytopenia (Hb 10.6 g/dl, WBC 3.5 × 109/l, platelets 43 × 109/l), and evidence of severe immunodeficiency: lymphopenia (0.38 × 109/l) with very low B cells and NK (natural killer) cells (CD3 92%, CD4 83%, CD8 16%, CD16 <0.5%, CD19 1%), poor response to phytohaemagglutinin (PHA) and low immunoglobulin (IgG, IgA and IgM) levels. Following this illness his lymphocyte count and PHA response improved but NK and B cells remained very low. In view of his family history known X-linked immunodeficiencies were excluded [X-linked agammaglobulinaemia (XLA), Wiskott-Aldrich syndrome (WAS), hyper IgM (HIM) and X-linked severe combined immunodeficiency (X-SCID)] by gene linkage studies and mutation analysis. From 12 months of age he developed severe enterocolitis and recurrent gastrointestinal bleeding requiring parenteral nutrition (TPN). In light of his continuing enteropathy and family history, he was treated with an unrelated bone marrow transplant (BMT) in 1996 aged 23 months. Post transplant, he has had numerous complications including infections, recurrent haemorrhagic and thrombotic episodes, delayed marrow reconstitution, severe gastro-intestinal problems (severe oral ulceration, reflux oesophagitis, oesophageal stricture, and ongoing enterocolitis) requiring chronic steroid therapy, long-term TPN and jejunostomy feeding. Developmental delay and cerebellar hypoplasia (on CT scan of head) were first recognized at the age of 25 months. He is now 60 months old, has global developmental delay (performance skills to the level of a 24-month-old), is just beginning to walk and has signs of nail dystrophy (one of the hallmarks of DC). It was the identification of case 5, with all the classic features and the recent recognition of HH as a separate entity (Aalfs et al, 1995), that led to the diagnosis of HH in family B. Furthermore, recurrent infections and severe enteropathy reflecting underlying immunodeficiency (severe combined immunodeficiency with hypogammaglobulinaemia, low/absent B cells and NK cells and evidence of T-cell dysfunction despite normal T-cell numbers) appear to have been very prominent features in all three affected members. Peripheral blood was used as source of DNA for several members of this family. No DNA source was available for case 3; fibroblasts were used as a source of DNA for cases 4 and 5.

Figure 2.

. Segregation of DKC1 mutations in families A and B. (a) PCR products from exon 5 of the DKC1 gene cut with Hph I. (b) PCR products from exon 3 of DKC1 gene cut with Nla III. Members of each family are drawn over the appropriate lanes. Unavailable family members are drawn to the left of the gel. M, size marker (pEMBL8 cut with TaqI and Pvu II ); U, uncut PCR products; C, digestion of control subjects; I–III, family generations 1–3. Symbols represent: □, normal male; ▪, affected male; ○, normal female; ⊙, carrier female; ▪/, deceased affected male.

PCR, SSCP (single-strand conformation polymorphism) analysis and direct sequencing of PCR products

Each of the 15 DKC1 exons was amplified by PCR using primers based on the DKC1 genomic sequence (Knight et al, 1999). In case 2 of family A the purified PCR products from each exon were sequenced directly using an ABI automated sequencer. In the index case (case 5) of family B the amplified PCR products were first screened for sequence changes by SSCP analysis as described elsewhere (Knight et al, 1999).

Family segregation analyses

In family A, since the mutation did not create a restriction enzyme site, we designed an ACRES (amplification-created restriction enzyme site) primer (5′AACCAGTCACCTTGGGATCCAGAGTATCAC3′). This contains a mismatch (underlined) which in the presence of the 361G mutation, creates an Hph I site when used in PCR amplification. The forward primer used in this reaction was 5′GATTTGTTGTTTCACTGGAGC3′. In family B the presence of the mutation creates an Nla III site and the segregation analysis was performed by digestion with this enzyme after amplification with the primer pair 5′AAAGGCATACATTT CCATGG3′ and 5′CAAGGATGCCAGCAGTAAG3′. Restriction fragments were analysed on a 2.5% agarose gels. The mutations in families A and B were tested for in the normal population by the amplification and SSCP analysis of genomic DNA from 50 Caucasian women.


Direct sequencing of case 2 in family A revealed a nucleotide change at position 361 (A → G) in exon 5 of the DKC1 gene (Fig 1a); this results in a codon change of AGT → GGT leading to an amino acid substitution of Ser → Gly at position 121 in dyskerin. A family study was undertaken using an amplification-created Hph I site for the 361G allele, and this showed segregation of the mutation with the disease (Fig 2a). In addition, a screen of 50 normal caucasian women failed to detect this sequence change, which excluded the possiblity that this mutation represented a rare polymorphism.

Figure 1.

. Missense mutations in the DKC1 gene in families A and B. Sequence analysis of the patient (case 2) (a) and his mother (b) from family A for exon 5 showing the anti-sense strand with the T → C mutation boxed. Sequence analysis of patient (case 5) from family B (c) and normal control (d) for exon 3 showing part of the sense strand of the sequence trace with the C → T mutation boxed.

In case 5 of family B, SSCP analysis showed a shift in exon 3 of the DKC1 gene which was found to be due to a C → T change at position 146 (Fig 1c); this results in a codon change of ACG → ATG leading to an amino acid substitution of Thr → Met at postion 49. This mutation creates an Nla III site; amplification of exon 3 followed by Nla III digestion showed segregation of the mutation with the disease (Fig 2b). In addition, X-chromosome inactivation patterns (XCIPs) were studied in peripheral blood DNA of the women from this family using a methylation-sensitive Hpa II site in the polymorphic human androgen receptor locus at Xq11.2-Xq12 (HUMARA). All heterozygotes for the mutation show non-random XCIPs (data not shown) as observed previously (Vulliamy et al, 1997) in carriers of X-linked DC.


The finding of these two novel missense mutations in two families with all the features of HH, one of which is the family described previously by Aalfs et al (1995), demonstrates that HH is a severe variant of X-linked DC. It also shows that mutations in DKC1 can give rise to a much more heterogenous clinical phenotype than previously thought, including cerebellar hypoplasia and severe immunodeficiency, which have not previously been regarded as an integral part of the DC phenotype. The severe developmental problems and cerebellar hypoplasia in this and other HH families emphasize an important role of dyskerin in brain development. The wide range of abnormalities seen in patients with HH and classic DC suggest that dyskerin has a functional role in many tissues and this is supported by the ubiquitous expression of the DKC1 gene (Heiss et al, 1998). What remains unclear is how different mutations in the same gene can give rise to such a variable combination of clinical phenotypes. Further studies are needed to determine functionally important residues of dyskerin, and how different mutations affect its activity. It is noteworthy the AGT  → GGT mutation in exon 5 in family A represents the first DKC1 mutation within the TruB domain, which, based on studies on the dyskerin yeast homologue Cbf5p, is predicted to have a function in pseudouridine synthesis in pre-rRNA (Heiss et al, 1998; Lafontaine et al, 1998).

This study suggests that DKC1 mutations may contribute to a very wide clinical spectrum of patients. This may not have been recognized previously because clinical syndromes due to mutations in DKC1 that resulted in death (for example fatal infections because of severe immunodeficiency/aplastic anaemia) below the age of 10 years (as in case 3 and case 4 in family B) would not have been categorized as DC or HH since these patients did not live long enough to develop the diagnostic mucocutaneous features of classic DC or have all the features of HH. This suggests that children with severe clinical phenotypes who have some features of DC or HH (e.g. unexplained/idiopathic aplastic anaemia or immunodeficiency), but who lack the skin and nail changes, should now be screened for mutations in the DKC1 gene. These novel mutations may provide important clues to the function of the different protein domains of dyskerin, leading to the emergence of genotype–phenotype correlations of clinical/prognostic significance.


We thank the families for providing blood samples and all medical and nursing staff involved in the clinical management of these patients. We also thank the Wellcome Trust and Action Research for their financial support.