Editor—There are a number of monogenic diseases with complex phenotypes which are clinically distinct but also overlap in phenotype with one or more other syndromes. If mutations in the same gene are responsible for causing the related syndromes, the diseases are allelic. Two diseases linked to Xq28, incontinentia pigmenti (IP, MIM 308310, Bloch-Sulzberger syndrome) and dyskeratosis congenita (DKC, MIM 305000, Zinsser-Cole-Engmann syndrome) show similarities in phenotype, although the modes of expression differ. Whereas IP is X linked dominant with embryonic lethality in males, the major form of DKC is X linked recessive. The gene responsible for causing DKC,DKC1, was recently identified1and maps about 20 kb proximal to the factor VIII gene,F8C. 2 Linkage analyses have provided evidence that the IP gene is located in the telomeric 2 Mb region of Xq28 distal to DXS523 and lod scores of highest significance were found aroundF8C. 4 5 The physical map position of DKC1 and genetic linkage of the IP locus, together with the overlap in the DKC and IP phenotypes (table1), raised the possibility that these two diseases could be allelic.

The IP and DKC phenotypes share abnormalities in ectodermal derivatives, such as nail dystrophy, alopecia, hypodontia, and skin manifestations6 7 (table 1). Both IP and DKC are characterised by the early appearance of reticulate skin pigmentation, although this manifests differently in the two diseases. In IP the clinical signs affecting the skin are initially apparent as an erythematous, inflammatory vesicular rash. The rash later becomes verrucous and streaks of hyperpigmentation follow. The pigmentation then fades in the second decade of life often leaving scarred and atrophic hypopigmented areas. In DKC patients the inflammatory and verrucous stages do not occur and the appearance of hyper- and hypopigmentation is progressive. The overlap in the skin abnormalities is confirmed by microscopic examination of skin biopsies from IP and DKC patients, which show common histological features such as epidermal atrophy and pigment migration.8 In both disorders a defect in the immune system may be causing the skin manifestations. In IP the inflammatory vesicular rash points to an involvement of the immune system and is supported by observations that the rashes are associated with constitutional eosinophilia and may recur during feverish infections. Further, it has been suggested that the skin phenotype in IP resembles that observed in patients with graft versus host (GVH) disease.9 A GVH-like pathogenesis suggestive of an involvement of the immune system in the skin also occurs in some DKC cases.10 11

DKC patients develop progressive pancytopenia of one or more cell lines and bone marrow failure is the main cause of death in the first or second decade of life in 90% of the cases.12 This is accompanied by humoral and cellular disturbances of the immune system.13 Pancytopenia and bone marrow failure are not associated with IP. There have been reports, however, of decreases in lymphocyte number and both neutrophil and lymphocyte dysfunction in IP.14-16 Another abnormality of the peripheral blood system suggesting an involvement of the immune system is the occurrence of leucocytosis with eosinophilia in a substantial proportion of newborn females with IP in the absence of infection. A report on a male IP patient who died postnatally and showed excessive haemorrhaging and haemolysis at birth further indicates a defect in the haematological system.17

Extreme skewing of X chromosome inactivation has been observed in the blood cells of most DKC carrier females18 19 as well as in the skin and haemopoietic cells of affected IP females.20 21 The non-random inactivation of the X chromosome carrying the mutant allele in the skin cells of IP females is responsible for the disappearance of the clinical signs because of a positive selection for cells expressing the normal allele.20 21 It is conceivable that a defect in the haemopoietic system leading to bone marrow failure as is observed in DKC males is not apparent in IP females because of a similar selective pressure favouring cells carrying the active normal X chromosome. The skewed X chromosome inactivation in IP females could in part explain the difference in female presentation and the more severe phenotype observed in hemizyous IP males.

The hypothesis that a different spectrum of mutations in theDKC1 gene causes IP is compatible with the ubiquitous expression pattern of DKC1, its high degree of conservation, and the putative function of the peptide dyskerin in rRNA biogenesis.1 22-24 Seventeen different mutations have been identified in DKC patients of which 82% are missense mutations.25 To date no premature stop codon mutations, frameshifts, or whole gene deletions have been identified. Taken together, these observations strongly suggest an essential function for dyskerin and that complete loss of function mutations would not be viable. It appears likely that a null mutation inDKC1 could explain the prenatal lethality observed in IP males and that the same mutation in an IP female might result in the clinical signs observed.

The genomic structure of the DKC1 gene has been determined.25 The coding sequence is split into 15 exons and the gene extends over 15 kb (accession numbers AJ0101395, AJ0101396). As intronic primers flanking each of the 15 exons had been designed for mutation screening of DKC patients, it was possible to screen the DKC1 gene efficiently for mutations in IP patients. The analysis of a large number of IP patients of different nationalities was possible because of the collaborative efforts of five research groups. Thirteen of these families have been described previously.4 5 21 26 All 15 exons of 23 female IP patients and one spontaneously aborted male fetus carrying the mutant allele5 were subjected to SSCP analyses. SSCP protocols that had previously been shown to be efficient for mutation detection were used and the conditions for each exon were determined to allow good resolution of the two single strands.25 27 No shifts were observed for any of the patients. To exclude point mutations which may have been missed by SSCP, all exons from two spontaneously aborted male patients were PCR amplified and sequenced, but no mutations were found. Furthermore, 18 of the 24 DNA samples analysed by SSCP plus 32 additional IP females and three additional IP males were analysed by Southern hybridisation using the full lengthDKC1 cDNA as a probe. The following restriction enzyme digests were analysed:XbaI, BamHI,EcoRI, PstI,HindIII, SacI,NcoI, BglII, andTaqI. No differences in dosage and no aberrant bands were detected when compared with DNA samples from normal males and females. The results from Southern hybridisations and the fact that all exons were amplifiable for two IP male patients indicate that a partial or whole gene deletion ofDKC1 as a general mechanism for causing IP is unlikely. Moreover, no mutations were identified in the coding region or at the exon-intron boundaries of the two IP male patients. Owing to the difficulty of obtaining sufficient cells with an active IP mutation bearing X chromosome from female patients and because very few IP male patients with a normal XY karyotype exist, no analyses were carried out at the RNA level. It therefore cannot entirely be ruled out that there may be mutations in the promoter region or in the 5′ and 3′ untranslated regions (UTR) which could alter the levels ofDKC1 mRNA directly or alter the stability of the transcript in IP patients. However, we consider this to be a very unlikely possibility and conclude that IP and DKC are not allelic.