Hirschsprung disease (HSCR) is a frequent congenital disorder (with an incidence of approximately 1 in every 5,000 live births) appearing during the first years of life and characterized by the absence of parasympathetic intrinsic ganglion cells in the hindgut. A major gene for HSCR with incompletely penetrant autosomal dominant mode of inheritance was mapped to the proximal long arm of chromosome 10 [1, 2] and subsequently restricted by genetic and physical mapping to the RET locus [3], in which point mutations as well as deletion and insertion mutations were identified [4, 5]. Among these mutations, a proportion varying between 21 and 26% [6–9] can be estimated to be de novo, which is in keeping with the high incidence of sporadic cases (80–90%) observed among HSCR patients.

HSCR mutations are distributed among 18 of the 21 exons of RET [9]. On the contrary, mutations causing multiple endocrine neoplasia type 2A (MEN 2A) are exclusively found in exons 10 and 11 [10, 11], those causing familial medullary thyroid carcinoma (FMTC) mostly in exons 10, 11 and rarely in exons 13 and 14 [12, 13], while a unique mutation in codon 918 (Met→Thr) of exon 16 is found in almost all multiple endocrine neoplasia type 2B (MEN 2B) patients [14, 15]. The latter one frequently arises as a de novo event almost exclusively on the paternal chromosome [16, 17], with skewed segregation in subsequent generations, a fact arguing in favor of genomic imprinting [18]. In addition a recent study indicates that up to 9% of MEN 2A and FMTC patients have de novo mutations which occur exclusively on the paternal allele [19]. In our series of 121 HSCR patients, we identified 23 RET mutations, 6 in familial and 17 in sporadic cases [6, 9]. Of the latter ones, in 6 cases a parent carried the mutation, in 6 the mutation occurred de novo, while investigations of the other 5 were inconclusive because of lack of parental DNA. Thus between 35% (6/17) and 65% (11/17) of RET mutations in sporadic HSCR patients occur de novo, which represents a prevalence higher than that previously calculated [6–9].

Table 1 summarizes investigations of the 6 de novo mutations. In patients 1 and 2 the RET exon containing the mutation [6] also carried an informative polymorphism [20], allowing the parental origin to be established. Parental origin in patients 3 and 4 [6, 9] was established by isolating the normal and abnormal chromosomes in somatic cell hybrids [21] and typing for informative polymorphisms [22–24] both parents and the affected child, together with hybrid clones (fig. 1). Patient 5 [25] had a de novo deletion of the entire RET gene, with loss of paternal allele at the linked marker D10S196. No conclusion could be reached in patient 6 because of lack of informative polymorphisms.

Fig. 1
figure 1

Parental origin analysis for the del3118(CTGG) de novo mutation. Normal and mutated RET homologues on chromosome 10 were segregated into distinct somatic cell hybrids after fusion of the patient’s lymphoblasts with a Chinese hamster ovary cell line (YH21). Hybrid cells containing chromosome 10 were selected for the expression of the gene coding for the beta subunit of the fibronectin receptor (FNRB), which maps in 10p11.2, using a monoclonal antibody against FNRB [21]. Hybrids 1 and 2 retain the normal and the deleted alleles of RET, respectively. A Pattern (on a 6% Polyacrylamide gel) of the PCR product of exon 19 of RET amplified from genomic DNA of the patient and his parents, as well as from 2 hybrids. The lower band represents the deleted allele, while the upper band represents the normal allele. B Typing obtained with the RET-INT5 microsatellite [23]. The genotypes of the father, of the affected son and of the mother are 4/6, 4/5 and 5/5, respectively. Hybrid 1 retains the normal RET and allele 5 of RET-INT5 derived from the mother, while hybrid 2 retains the deleted RET and allele 4 of RET-INT5 derived from the father.

Full size image
Table 1 Parental origin of de novo mutations in HSCR
Full size table

In conclusion, de novo mutations of RET occur in both maternal (3/5) and paternal (2/5) alleles of HSCR patients, in contrast with MEN 2B [16, 17], MEN 2A and FMTC [19], in which de novo RET mutations arise almost exclusively on paternal chromosomes. It remains to be established on a larger series of de novo HSCR mutations (a) whether the observed breakdown (point mutations of maternal and deletions of paternal origin) is due to chance or has any biological meaning, and (b) how often the somatic mosaicism already reported in 1 HSCR patient [8] is present in patients carrying de novo mutations.