Retinoblastoma, an embryonic neoplasm of retinal origin (38), is the most common malignant intraocular tumor in childhood, with an average incidence of one case in every 18,000 live births. Most of the clinical phenotypes of retinoblastoma can be explained by the mutational inactivation of the two alleles (18) in the retinoblastoma tumor suppressor gene, RB1 (4), which maps on chromosome band 13q14 (12).
In the hereditary form of the disease—40% of the cases in the classic statistical studies by Knudson (18) —the initial predisposing mutation is in the patient's germline, either as the result of a new mutation (sporadic presentation) or through germline transmission (positive family history), and has been distributed to nearly all the cells in the body. Because the second tumorigenic event, the somatic inactivating mutation of RB1 in retinoblasts, occurs spontaneously at a relatively frequent rate (1:1,000 (24)), the prevalent phenotype in this group of patients is mainly bilateral, multifocal retinoblastomas with an early onset (mean age at diagnosis, 1 year) and a lifetime predisposition to other RB1-dependent tumors, such as osteosarcoma. The predisposing germline mutation is transmitted as a highly penetrating (90%) autosomal dominant tract, resulting in a 45% risk for offspring of patients with hereditary retinoblastoma. In the familial presentation, which accounts for nearly one quarter of the hereditary forms, the risk for siblings is equally high (45%). In the nonhereditary (somatic) form of the disease, both inactivating events occur during somatic development of a single retinal cell, resulting in the relatively late onset of a single tumor in one eye (i.e., unilateral disease with a median age of 2 years at diagnosis). Unfortunately, the clinical phenotype (unilateral vs. bilateral) is not a perfect risk predictor because nearly 15% of unilaterally affected patients have germline RB1 mutations, representing a 45% risk for offspring, and these patients cannot be distinguished clinically from patients with true somatic, unilateral retinoblastoma, who present a negligible risk for siblings and offspring. These masked hereditary unilateral cases raise the share of hereditary retinoblastoma (unilateral and bilateral) with 50% of patients, according to recent statistics (22,24,25).
A predictive test that could distinguish persons prone to developing retinoblastoma from those who are not, especially with regard to familial retinoblastoma, is essential for genetic counseling and early ophthalmologic diagnosis, and would probably reduce the cost of patient management (14,26). Direct mutation analysis is the best approach to characterizing most of the gene defects in retinoblastoma, but its usefulness as a screening method to approach the familial form is hampered by the fact that mutations are scattered among 27 exons and the promoter region of the RB1 gene, and no single hotspot has been found (19). Moreover, nearly 15% of the mutations in hereditary retinoblastoma consist of microscopic or large deletions (24) that cannot be monitored by exon-by-exon mutation analysis. Thus, linkage analysis remains a useful tool for the assessment of individuals at risk in families with retinoblastoma. In the past, allelic segregation was investigated using intragenic deoxyribonucleic acid (DNA) probes, isolated from within the genomic RB1 sequence, to recognize high-frequency restriction fragment length polymorphisms (13,27,28,30). Although accurate, this procedure has the intrinsic drawbacks of Southern blot analysis and conventional radioactive labeling techniques, and it is frequently necessary to include many probes in the study to obtain definitive results. A combination of PCR amplification and restriction enzyme analysis has also been used for linkage studies in familial retinoblastoma. Although this approach is simpler and less labor-intensive than conventional Southern blot methodologies, it has given noninformative results in a large number of cases (23,35).
The availability of multiple microsatellite loci for polymerase chain reaction (PCR) amplification (29) and automated fluorescent DNA sizing technology (37) has provided a rapid and precise genotyping procedure for the assessment of individuals at risk in families with genetic diseases such as hemophilia A (33) or Duchenne's and Becker's muscular dystrophies (6), among others. Microsatellite analysis has also been used for linkage studies in familial retinoblastoma (21,23), but no definite protocol has been described until now.
In the current study we describe a simple predictive test that screens for familiar susceptibility to retinoblastoma. It employs two intragenic and two flanking highly polymorphic microsatellite loci and the automated fluorescent DNA sizing technology of PCR products. The method can also substitute Southern blot and restriction fragment length polymorphism methodologies in the detection of gross RB1 deletions in patients with hereditary retinoblastoma and can determine tumoral allelic losses of prognostic value in osteosarcoma patients at the RB1 locus (11). The method is rapid, simple, and can be implemented at low cost in any clinical laboratory with normal molecular biology equipment.
MATERIALS AND METHODS
The studies presented in this paper were carried out in 17 families (N = 65) with hereditary retinoblastoma treated in the Ophthalmic and Hemato-Oncology Units, Hospital Infantil La Paz, Madrid. Diagnosis of retinoblastoma was established by standard ophthalmologic and histologic criteria. Patients with osteosarcoma were diagnosed at the Hemato-Oncology Unit, Hospital Infantil La Paz, Madrid.
Whole blood from patients and relatives was collected in standard 10-mL ethylenediamine tetraacetic acid blood collection tubes, and the DNA was extracted from two 200-μL aliquots using a commercial kit according to the manufacturer's protocol (QIAamp blood kit; Qiagen Inc., Valencia, CA, USA). For DNA extraction from osteosarcoma tumors, specimens were immediately snap-frozen in dry ice after surgery, powdered in a mortar that held liquid nitrogen, and then DNA was extracted using a commercial kit (Qiagen genomic-tips; Qiagen Inc.). The purity and quantity of the DNA preparation were ascertained by spectrophotometric and agarose electrophoretic criteria.
DNA was PCR amplified using high performance liquid chromatography (HPLC)-purified and fluorochrome-labeled primers (Cruachem Ltd., Glasgow, UK) described in Table 1. Standard PCRs were carried out in 0.5-mL tubes with 100 ng DNA, 1× PCR buffer (Perkin–Elmer, Norwalk, CT, USA), 1.5 mM MgCl2, 100 mM each deoxynucleotide triphosphates, 0.1 μM each set of primers, and 0.125 U Taq polymerase (Perkin–Elmer) in a final volume of 25 μL. PCR conditions were an initial denaturing step of 2 minutes at 94°C, 10 cycles composed of a 1-minute denaturing step at 94°C, a 30-second annealing step at 55°C, a 45-second extension step at 72°C, and 30 cycles composed of 30 seconds at 94°C, 30 seconds at 55°C, and 45 seconds at 94°C, followed by a final step of 10 minutes at 72°C. PCR was done in a programmable thermal controller model PTC-100 (MJ Research Inc., Watertown, MA, USA).
PCR-amplified microsatellites from each DNA sample were pooled at a fixed ratio of 2:1:1:2 (D13S262:Rbi2: RB1.20:D13S284). These ratios were found to give similar fluorescent intensities in previous optimization trials. However, and as consequence of differences in yield during the manufacturer's labeling procedure, these ratios should be optimized with each new set of primers. Two microliters of the mixture was mixed with 0.5 μL carboxytetramethylrhodamine 500 marker (Perkin–Elmer) and 25 μL deionized formamide, and was stored at −20°C until analysis. Before analysis, samples were denatured for 5 minutes at 95°C and cooled immediately on ice. Samples were run in the fluorescent automatic sequencer model ABI310 (Applied Biosystems, Foster City, CA, USA) and analyzed using Genescan software V2.1 (Perkin–Elmer). Conditions for electrophoresis were matrix C, 15 kV, 10 μA, 60°C, and a run time of 22 minutes.
Loss of heterozygosity (LOH) in osteosarcoma tumors was evaluated according to the method described by Cawkwell et al. (5). Briefly, the ratio R = T1 × N2/T2 × N1 was calculated for each normal (blood) and tumor sample, where T1 and N1 are the area of the short allele product peak for the tumor and normal sample respectively, and T2 and N2 are the area of the long allele product peak for the tumoral and normal sample respectively. If R is less than 0.6 or more than 1.67, then one of the alleles had decreased by more than 40%, resulting in LOH.
Microsatellite markers used in this study were chosen on the basis of their reported heterozygosity and close linkage to the RB1 gene. The two intragenic microsatellite loci—Rbi2 (D13S153), a CA repeat located in intron 2, and RB1.20, a CTTT(T) repeat in intron 20—have been proved to be highly informative (31,34). Although the rate of intragenic recombinations within or near RB1 is low (30), we have also included two closely flanking microsatellite loci (D13S262 and D1S3284) to evaluate the possible occurrence of crossover between the intragenic markers. The sex-averaged genetic distance between the flanking markers D13S262 and D13S284 was 0.1 cM, according to the latest version of the Genethon map (7).
To simplify the PCR conditions and product detection, primer sequences were modified with respect to those reported previously (2,7,31). As shown in Table 1, the newly designed primers had a similar melting temperature, compatible with a single PCR assay condition. The other criterion in the design of the primers was directed to obtaining PCR products in two size ranges so that the proper combination of fragment length and fluorochrome would make it possible to analyze the four markers in a single run. All four microsatellite markers were amplified efficiently and specifically with the PCR conditions described in Materials and Methods (Fig. 1). In the case of marker RB1.20, we observed the presence of two additional bands that migrated more slowly in agarose electrophoresis, but this did not interfere with the fluorescent fragment analysis. As shown in Table 2, the observed heterozygosity in 37 unrelated individuals ranged from 77% in D13S262 to 89% in RB1.20, with average values that were very close to those previously reported (2,7). The number of observed alleles was higher than that previously reported (see Table 2), a fact that probably represents differences within the population under study.
We studied five pedigrees of familiar retinoblastoma, including four bilateral cases (family nos. 8, 16, 24, and 26) and one unilateral case (family no. 8). In the case illustrated in Figure 2, the mother and her oldest daughter had unilateral disease with a late age at diagnosis, suggesting low-penetrance retinoblastoma (20). Linkage analysis showed that the youngest daughter did not carry the allele that segregated with the disease, indicating that this child was not at risk for developing retinoblastoma. Supporting this prognosis, she is free of disease to date. In another case of familiar retinoblastoma (not shown), a perinatal study was performed of the brother of an affected child 4 days after birth. Linkage analysis showed that the haplotype segregating with the disease was also inherited by the newborn and, as predicted from the genotype, the child developed bilateral disease 3 weeks later. In the remaining three families, linkage analysis provided definitive information for genetic counseling (data not shown).
Detection of Germline Deletions and LOH at the RB1 Locus
Our method can also be used as an alternative to Southern blot for the analysis of germinal deletions affecting the RB1 gene. In fact, using the described methodology we were able to detect the germline loss of an RB1 allele in two of 12 patients with sporadic, bilateral hereditary retinoblastoma. In patient no. JA 153, the region deleted spanned at least the intragenic markers Rbi2 and RB1.20 from the paternal allele. The patient retained both alleles for the D13S284 marker and was homozygous for marker D13S262 (Fig. 3). In another case (patient no. JA 203, not shown), the four markers from the paternal allele were lost, indicating the presence of a deletion larger than the RB1 locus. A broad deletion was suggested by the patient's clinical record of mental and physical retardation and was confirmed by a cytogenetically visible deletion, 13(q12.3-q22). It was also noted that in the two germinal deletions reported in the current study, the paternal allele of the RB1 gene is the lost allele—an observation that agrees with earlier reports showing that germline mutations in familial retinoblastoma preferentially involve paternal alleles (8,36).
In view of these results, we tested the suitability of this method for analyzing LOH at the RB1 locus in patients with osteosarcoma, a childhood tumor related closely to retinoblastoma (32). Gross deletions in the RB1 locus have been associated with a poor prognosis in osteosarcoma patients, so this analysis could aid in the management of patients with this aggressive tumor (11). In one of the three patients analyzed, LOH for three RB1 markers (D13S284, Rbi2, and RB1.20 ) was evident in the tumor DNA (Fig. 4), and the other marker (D13S262) was noninformative, indicating that our method can be applied successfully to the assessment of LOH in osteosarcoma.
The method described herein presents several advantages with respect to conventional Southern blot or PCR restriction fragment length polymorphism analysis for linkage studies. First, our technique is comparatively faster and simpler than restriction fragment length polymorphism analysis, which requires high-molecular weight DNA isolation and Southern blot hybridization with radioactive probes. In our procedure, DNA appropriate for PCR-based analysis can be purified in only 2 hours using commercially available kits, which, in addition, avoid the use of hazardous components such as phenol. Amplification of target microsatellites takes less than 4 hours, including the agarose electrophoretic control. The last step, capillary electrophoresis and fragment analysis, takes less than 2 hours. Second, and as a consequence of the marked reduction in labor, the described procedure is less expensive than other protocols. Third, as shown in this article, the combination of four highly informative microsatellite markers is suitable for most genotyping studies involving the RB1 locus. Indeed, with the exclusion of one patient from a family with a high consanguinity rate, three markers were found to be informative in all cases, and all four markers were informative in 69% of cases. Thus, in most of the cases, complete genotype and accurate genetic counseling can be provided within 2 to 3 days after receipt of blood samples.
In addition to linkage analysis, the method is also appropriate for the detection of large RB1 deletions in hereditary retinoblastoma. Because most of the large RB1 deletions involve introns 2 and/or 20 (1,3,9,15–17,30), a similar proportion should be detected with our microsatellite methodology, provided a parental blood sample is available. In fact, the proportion of deletions detected in our series (two of 12 hereditary cases, or 17%) was close to the 16% germinal deletions reported in a larger series using Southern blot analysis with several DNA probes (17). In that study, abnormal restriction fragments were observed in only 20% of the deletions and, therefore, a densitometric analysis was required to evaluate most of the Southern blot results. In this respect our method also improves on the Southern blot-based assays because it avoids the uncertainties inherent in densitometric scanning. The same arguments justify the use of our methodology for LOH detection in RB1-dependent tumors, such as osteosarcoma, breast carcinoma, small cell lung carcinoma, and others (10).
In summary, our microsatellite methodology can be applied to the molecular diagnostic procedure for analyzing RB1-related diseases. Genotyping is essential for identification of carriers and genetic counseling in familiar retinoblastoma, and this linkage analysis also scans for large mutations. In these patients, further mutation analysis (PCR and single-strand conformation polymorphism, direct sequencing) would be needed only for the detection of a founder mutation in a pedigree. In sporadic (hereditary nonfamilial) retinoblastoma, genotyping is irrelevant but the microsatellite analysis is important to the detection of large deletions, and functions as prescreening in the search of mutations that help to determine the risk in offspring. Microsatellite analysis can also be applied to screen for prognostic or pathogenic RB1 deletions in tumors such as osteosarcoma, breast carcinoma, or small cell lung tumors. In all these cases, the method described in this study provides a rapid and reliable alternative to Southern blot-based methodologies. Moreover, the assay conditions and technology involved can be implemented at a low cost in any clinical laboratory with normal molecular biology equipment.
The authors are grateful for the able technical assistance of Ms Paloma Nebreda. J.A. held postdoctoral fellowship from Rhône Poulenç Rourer and Comunidad de Madrid. M.M. held a graduate student fellowship from the Universidad Autónoma de Madrid.
1. Blanquet V, Créau–Goldberg N, de Grouchy J, et al. Molecular detection of constitutional deletions in patients with retinoblastoma
. Am J Med Genet 1991; 39:355–61.
2. Brandt B, Greger V, Yandell D, et al. A simple and nonradioactive method for detecting the Rb1.20 DNA polymorphism in the retinoblastoma
gene. Am J Hum Genet 1992; 51:1450–1.
3. Bremner R, Chan Du D, Connolly–Wilson MJ, et al. Deletion
of RB exons 24 and 25 causes low-penetrance retinoblastoma
. Am J Hum Genet 1997; 61:556–70.
4. Cavenee WK, Dryja TP, Phillips RA, et al. Expression of recessive alleles by chromosomal mechanism in retinoblastoma
. Nature 1983; 305:779–84.
5. Cawkwell L, Bell SM, Lewis FA, et al. Rapid detection of allele loss in colorectal tumors using microsatellites and fluorescent DNA technology. Br J Cancer 1993; 67:1262–7.
6. Clemens PR, Fenwick RG, Chamberlain JS, et al. Carrier detection and prenatal diagnosis in Duchenne and Becker muscular dystrophy families, using dinucleotide repeat polymorphisms. Am J Hum Genet 1991; 49:951–60.
7. Did C, Faure S, Pizames C, et al. A comprehensive genetic map of the human genome based on 5,264 microsatellites. Nature
1996;380:152–4. Also available from the Genethon database, http://www.genethon.fr.
8. Dryja TD, Mukai S, Petersen R, et al. Parental origin of mutations in the retinoblastoma
gene. Nature 1989; 339:556–8.
9. Dryja TD, Rapaport JM, Joyce JM, et al. Molecular detection of deletions involving band q14 of chromosome 13 in retinoblastomas. Proc Natl Acad Sci U S A 1986; 83:7391–4.
10. Fearon ER. Tumor suppressor genes. In: Vogelstein B, Kinzler KW, eds. The genetic basis of human cancer. New York: McGraw–Hill, 1998:229–36.
11. Feugeas O, Guriec N, Babin–Boilletot A, et al. Loss of heterozygosity of the RB gene is a poor prognostic factor in patients with osteosarcoma
. J Clin Oncol 1996; 14:467–72.
12. Friend SH, Bernards R, Rogelj S, et al. A human DNA segment with properties of the gene that predisposes to retinoblastoma
. Nature 1986; 323:643–6.
13. Greger V, Kerst S, Messmer E, et al. Application of linkage analysis to genetic counselling in families with hereditary retinoblastoma
. J Med Genet 1988; 25:217–21.
14. Harbour JW. Overview of RB gene mutations in patients with retinoblastoma
. Ophthalmology 1998; 105:1442–7.
15. Hashimoto T, Takahashi R, Yandell DW, et al. Characterization of intragenic deletions in two sporadic germinal mutation cases of retinoblastoma
resulting in abnormal gene expression. Oncogene 1991; 6:463–9.
16. Horsthemke B, Greger V, Barnert HJ, et al. Detection of submicroscopic deletions and a DNA polymorphism at the retinoblastoma
locus. Hum Genet 1987; 76:257–61.
17. Kloss K, Währisch P, Greger V, et al. Characterization of deletions at the retinoblastoma
locus in patients with bilateral retinoblastoma
. Am J Med Genet 1991; 39:196–200.
18. Knudson AG. Mutation and cancer. Statistical study of retinoblastoma
. Proc Natl Acad Sci U S A 1971; 68:820–23.
19. Lohmann DR. RB1 gene mutations in retinoblastoma
. Hum Mut 1999; 14:283–8.
20. Lohmann DR, Brandt B, Höpping W, et al. Distinct RB1 gene mutations with low penetrance in hereditary retinoblastoma
. Hum Genet 1994; 94:349–54.
21. Lohmann DR, Brandt B, Oehlschlänger U, et al. Molecular analysis and predictive testing in retinoblastoma
. Ophthalmic Genet 1995; 4:135–42.
22. Lohmann DR, Gerick M, Brandt B, et al. Constitutional RB1-gene mutations in patients with isolated unilateral retinoblastoma
. Am J Hum Genet 1997; 61:282–94.
23. Mateu E, Sánchez F, Nájera C, et al. Genetics of retinoblastoma
: a study. Cancer Genet Cytogenet 1997; 95:40–50.
24. Murphree AL. Molecular genetics of retinoblastoma
. Adv Ophthal Pathol 1995; 8:155–66.
25. Naumova A, Sapienza C. The genetics of retinoblastoma
, revisited. Am J Hum Genet 1994; 54:264–73.
26. Noorani HZ, Khan HN, Gallie BL, et al. Cost comparison of molecular versus conventional screening of relatives at risk for retinoblastoma
. Am J Hum Genet 1996; 59:301–7.
27. Onadim ZO, Hykin PG, Hungerford JL, et al. Genetic counselling in retinoblastoma
: importance of ocular fundus examination of first degree relatives and linkage analysis. Br J Ophthalmol 1991; 75:147–50.
28. Onadim ZO, Mitchell CD, Rutland PC, et al. Application of intragenic DNA probes in prenatal screening for retinoblastoma
gene carriers in the United Kingdom. Arch Dis Child 1990; 65:651–6.
29. Reed P, Davies J, Copeman J, et al. Chromosome-specific microsatellite
sets for fluorescent-based, semi-automated genome mapping. Nat Genet 1994; 7:390–5.
30. Scheffer H, Meermen GJ, Kruize YCM, et al. Linkage analysis of families with hereditary retinoblastoma
: nonpenetrance of mutation, revealed by combined use of markers within and flanking the RB1 gene. Am J Med Genet 1989; 45:252–60.
31. Toguchida J, McGee TL, Paterson JC, et al. Complete genomic sequence of the human retinoblastoma
susceptibility gene. Genomics 1993; 17:535–43.
32. Wadapause BI, Togudrida J, Shimizu T. Mutation spectrum of the retinoblastoma
gene in osteosarcomas. Cancer Res 1994; 54:3042–8.
33. Windsor S, Taylor SAM, Lilligrap D. Multiplex analysis of two intragenic microsatellite
repeat in the diagnosis of haemophilia A. Br J Haematol 1994; 86:810–5.
34. Yandell DW, Dryja TD. Detection of DNA sequence polymorphisms by enzymatic amplification and direct genomic sequencing. Am J Hum Genet 1989; 45:547–55.
35. Yazici H, Kir N, Ayan I, et al. Investigation of the loss of heterozygosity and familial segregation by PCR in retinoblastoma
. Clin Biochem 1996; 29:595–8.
36. Zhu X, Dunn JM, Phillips RA, et al. Preferential germline mutation of the paternal allele in retinoblastoma
. Nature 1989; 340:312–3.
37. Ziegle JS, Su Y, Corcoran KP, et al. Application of automated DNA sizing technology for genotyping microsatellite
loci. Genomics 1992; 14:1026–31.
38. Zucker JM, Desjardins L, Doz F. Retinoblastoma
. Eur J Cancer 1998; 34:1045–9.