Statistics from Altmetric.com
Editor—Retinoblastoma (RB) is an intraocular eye tumour with a newborn incidence of 1 in 15 000-25 000.1Of all retinoblastoma patients, 10-15% represent familial cases with an autosomal dominantly inherited predisposition for tumour development. On the basis of epidemiological data, Knudson2 hypothesised that retinoblastoma is a cancer caused by two mutational events. DNA analysis by Cavenee et al 3confirmed this hypothesis. In the familial form, the first mutation is inherited via a germ cell and is, therefore, present in all somatic cells. The second mutation occurs in a somatic cell. In the non-hereditary form, both mutations occur in one and the same somatic cell. From this “two hit” model, it can be predicted that statistically a fraction of carriers of a mutation predisposing to retinoblastoma will never become affected. Indeed, in approximately 10% of families with hereditary retinoblastoma, obligate asymptomatic carriers of a mutant allele of the retinoblastoma gene (RB1) can be identified.1Asymptomatic carriers of a predisposing RB1mutation are, however, not randomly distributed over families with hereditary retinoblastoma. Some families have been identified in which the majority of mutation carriers are either unaffected or only unilaterally affected. In a small number (fewer than 2%) of mutation carriers, clinically benign retinal lesions, or “retinomas”, occur.4 Most likely, retinoma arises when the second mutation occurs in an almost terminally differentiated retinal cell.5 Patients with retinoma can only be detected on ophthalmological inspection. They may, therefore, usually pass for being unaffected, but transmit the mutation predisposing to retinoblastoma to their offspring. Thus, they resemble true asymptomatic carriers of an RB1 germline mutation. DNA analysis of markers within and flanking theRB1 gene allows the identification within retinoblastoma families of healthy subjects in whom the mutation is non-penetrant.6 It has been shown that the non-random distribution of asymptomatic carriers, the predominant occurrence of unilaterally affected subjects, and the high incidence of retinomas in particular families is caused by partially defectiveRB1 alleles because of mutations in the promoter region,7 8 particular missense mutations,9 a genomic in frame deletion,10 a large genomic deletion encompassing RB1exons 24 and 25,11 or a splice site mutation leading to in frame exon skipping.12 From our own mutation screening in retinoblastoma families, we present two novel specific mutations apparently associated with incomplete penetrance.
One of the families described was referred to us directly, the other by the Departments of Clinical Genetics and Ophthalmology, Free University Hospital, Amsterdam. DNA and RNA were isolated from peripheral blood using standard methods. The analysis of the marker RB1.20 was performed as previously described.13
For initial mutation scanning, six multiplex PCR reactions were performed. In each reaction, three or fourRB1 exons were amplified simultaneously. In order to obtain PCR bands of the same intensity, it is often necessary to vary the amount of primer in a reaction, since some primers amplify better than others under standard conditions. Optimised primer sequences, amounts of primer per reaction, and product lengths are listed in table 1. For control experiments and sequencing, each exon was amplified separately using the same reaction conditions. Using a primer design program (Scientific and Educational Software, Durham, USA), primers located in the flanking introns were selected from the human retinoblastoma sequence.14 Amplification of genomic DNA was performed as previously described with the addition of MgCl2 to a final concentration of 3.0 mmol/l to reaction 5.
The multiplex PCR products were analysed by means of the SSCP technique. To improve the resolution of the SSCP analysis, restriction cleavage of reactions 1, 5, and 6 was carried out withRsaI, PstI, andTaqI, respectively. To increase the detection rate of SSCP variant bands, each set of reactions was analysed using three gel conditions: (1) denaturing 6% (w/v) polyacrylamide gel (19/1 (w/w) acrylamide/bisacrylamide) containing 7% (w/v) urea for analysis of small length alterations, (2) non-denaturing 6% (w/v) polyacrylamide gel (19/1 (w/w) acrylamide/bisacrylamide) containing 10% (v/v) glycerol, and (3) 0.5 × MDE gel (FMC-Bioproducts, Rockland, Maine, USA), according to the manufacturer's recommendations. Samples were denatured and electrophoresed for four hours at 20°C and 2000 V in a Macrophor electrophoresis apparatus (Pharmacia, Uppsala, Sweden) at room temperature. After electrophoresis, gels were dried and exposed tox ray film. RB1exon 14 was analysed by direct sequencing.
RNA was isolated using the RNAZOL-B kit (Tel-Test Inc, Friendswood, Texas, USA). Subsequently, cDNA was synthesised with the Ready To Go first strand beads (Pharmacia, Uppsala, Sweden) by using a specific 3′ primer: 5′-CAGACAGAAGGCGTTCACAA-3′. RT-PCR was performed using this primer and the 5′-primer: 5′-GAAACACAGAGAACACCACG-3′. Nested PCR was performed using a primer located in exon 11 and a primer located in exon 15 of the RB1 gene: 5′-CCAGTT AGGACTGTTATGAA-3′ and 5′-GGATAATCGTTC TTCTTCTG-3′. PCR products were analysed on an agarose gel.
Fragments amplified by PCR showing an aberrant SSCP pattern were subjected to direct sequencing by an automated sequencer (ALF, Pharmacia, Uppsala, Sweden). Reverse primers were extended by an M13 sequence (CGACGTTGTAAAACGACGGCCAGT) to enable the use of an identical M13 primer in all sequence reactions.
In the low penetrance family RB43, shown in fig 1A, a splice donor site mutation was identified. An A to G transition occurred at the last but one nucleotide of exon 13 (c.1331A>G mutation; cf sequence analysis in fig 1C), altering the splice donor site. By RT-PCR and subsequent sequencing we could show that this mutation results in loss of exon 13 in the RB1 mRNA. Skipping of exon 13, which consists of 117 nucleotides, does not alter the reading frame. In addition to the two unilaterally affected subjects in this pedigree, six asymptomatic carriers of the c.1331A>G mutation were identified. Both the affected subjects and the six asymptomatic carriers shared an identical allele of the intragenic marker RB1.20 (fig 1A). Ophthalmological investigation did not show any indication of retinoma, nor was there any indication of a second primary tumour in any of these subjects. A similar mutation (c.1332G>A, a G to A transition of the last nucleotide of this exon) occurred in the non-familial bilateral case RB163, also leading to loss of exon 13 from the aberrantRB1 transcript (data not shown). The ratios of the amount of normal length and aberrant length RT-PCR products from transcripts were identical for both mutations (data not shown), as well as the ratios of normal length and aberrant length RT-PCR products from transcripts in affected subjects compared to asymptomatic carriers in family RB43 (fig 1B). Apparently, more normal than aberrant transcript was produced, raising the possibility that the aberrant transcript is unstable. Since no other family members could be studied in family RB163, we do not know whether this mutation was already asymptomatically present in one of the parents, nor can we estimate the degree of penetrance in this family.
In family RB-K, one of the incomplete penetrance families in which we had previously identified an asymptomatic carrier by linkage analysis,6 a missense mutation in exon 17 was identified (fig 2). Codon 563 for tryptophan changes into a leucine by a G to T transversion (W563L). The mutation abolishes anNlaIII site. The mutation was present in all bilaterally affected subjects as well as the asymptomatic subject previously identified by linkage analysis (fig 2). In the other incomplete penetrance retinoblastoma family previously described, RB-E, sequence analysis of the complete RB1 coding region including the exon/intron junctions (except exon 1) failed to detect a mutation in the coding sequence and flanking intronic sequences. Most likely, in this family, there is a mutation inRB1 exon 1, in the promoter region or in an intronic sequence that has not been analysed.
In the low penetrance family RB43, a splice donor site mutation c.1331A>G was identified, leading to loss of exon 13 in theRB1 mRNA. Because exon 13 consists of 117 nucleotides and hence the reading frame of the coding sequence remains intact in the mutant RB1 allele segregating in this family, we hypothesise that the lack ofRB1 coded for by exon 13 has only a mild effect on cellular function of the RB protein. This mutation resembles the low penetrance mutation identified by Schubertet al,12 which causes skipping of exon 21 from the RB1 transcript owing to a G to A transition at the last nucleotide of this exon. This mutation did not cause a frameshift either. The reduced amount of aberrant transcript is probably caused by its instability.
The cellular effect of splice site mutations has to be evaluated carefully. Not all splice site mutations result in a shift of the reading frame. Consequently, they may result in a less severe clinical presentation or even lack clinical symptoms in a proportion of mutation carriers in a particular family. Such a pedigree would thus present with incomplete penetrance. In some cases it is difficult to predict what the effect of a splice site mutation will be at the mRNA level, that is, which exons will be skipped, or whether alternative cryptic splice sites will be used. Ideally, the effect of any splice site mutation should be studied at the mRNA level.
The W563L mutation detected in family RB-K represents a missense mutation. At the protein level, this mutation is located within the oncogenic binding pocket domain A, a region of the retinoblastoma protein with binding capacity for the SV40 large T antigen or the adenovirus E1A protein. Out of 74 differentRB1 mutations identified in a cohort of familial/bilateral retinoblastoma patients (initially identified by SSCP, later by an updated DGGE approach; manuscript in preparation), only four are missense mutations, including the W563L mutation and also a R661W mutation, previously also described in a reduced penetrance family.9 The other two missense mutations identified in our cohort were an L657P mutation in a familial case, without any indication of reduced penetrance in the pedigree, and an M704R mutation in a sporadic bilateral case. From this we conclude that in families in which RB1 missense mutations segregate, the likelihood of the occurrence of asymptomatic carriers may be increased, dependent, however, on the particular mutation identified.
We would like to thank the retinoblastoma families and their clinicians for collaboration in this study.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.