Pseudoautosomal linkage in chronic lymphocytic leukaemia.

Abstract

The X and Y chromosomes share two homologous pseudoautosomal regions (PARs) that recombine at meiosis, PAR1 and PAR2 at the tips of the short arms and long arms respectively. PAR genes exhibit a unique segregation pattern with affected siblings tending to be same sex. This segregation pattern has been reported in familial Hodgkin's Disease (HD), suggesting the disease is linked to PAR (Horwitz & Wiernik, 1999). There is an association with HD in familial B-cell chronic lymphocytic leukaemia (CLL) (Gunz & Veale, 1969; Potolosky et al, 1971; Eriksson & Bergstrom, 1987) and a common clonal origin is seen between CLL and Hodgkin/Reed–Sternberg cells (Kanzler et al, 2000). We have therefore examined CLL siblings for linkage to PAR. CLL families with affected siblings were identified from the series ascertained by the Institute of Cancer Research (ICR) and from the pedigrees in the literature, searched using MEDLINE and BIDS EMBASE. The observed distribution of affected sibling pairs by sex was compared with the random distribution predicted on the basis of the observed prevalence of male and female cases using the chi-squared test. To test formally for linkage between PAR and CLL, non-parametric analysis was undertaken using the program mapmakersibs (Kruglyak & Lander, 1995). Non-parametric linkage is based on testing the departure in allele sharing probabilities between affected individuals from that predicted on the basis of random sharing. No assumption is made about the underlying genetic model, gene frequency or penetrance of the disease allele. Pedigrees were coded for affection status and numbered alleles corresponding to the X and Y chromosomes: 1–2 for males and 1–1 for females. Table I shows sex concordance in the CLL sibships studied. The overall prevalence of CLL is greater for males than for females, 0·61 compared with 0·39. The distribution of affected sibling pairs by sex shows evidence of an excess of sex-concordant pairs when compared with the random distribution predicted on the basis of the observed prevalence of male and female cases (χ2 = 7·7; 1 df, P < 0·01). The fact that CLL affects more males than females at all ages is unlikely to be responsible for the distortion in sex concordance because the expected frequencies were calculated on the basis of the prevalence of male and female cases. The logarithm of the odds ratio (LOD) score determined from analysis of all 75 families was 1·51. Restricting linkage analysis only to families with three or more affected siblings, the LOD score was 0·21. Although the overall LOD from analysis of all families lends some support for the notion of linkage of CLL to the pseudoautosomal region, it does not attain the established threshold for significance. The issue of bias was an important concern in this study as it was highly dependent on published families. When the analysis of the data was restricted to families with three or more affected siblings, support for pseudoautosomal linkage was substantially reduced. Such families with a large number of affected individuals are most likely to have a genetic basis. Sources of bias include the possibility that there may be reporting bias in favour of same sex cases in families with affected siblings, so favouring sex concordance in this type of family, whereas families with multiply affected individuals are likely to be reported irrespective of sex composition. The findings of our study do not lend strong support to the involvement of the pseudoautosomal region of the genome in CLL, suggesting that the molecular basis of CLL is distinct from Hodgkin's disease. This work was supported by the Leukaemia Research Fund.