Skip to main content
Device Global Header

Subscribe
or Renew

Advanced Search

Original Article

Location on Chromosome 15 of the Gene Defect Causing Marfan Syndrome

List of authors.
  • Katariina Kainulainen, Cand.Med.,
  • Leena Pulkkinen, M.Sci.,
  • Aslak Savolainen, M.D.,
  • Ilkka Kaitila, M.D., Ph.D.,
  • and Leena Peltonen, M.D., Ph.D.

Abstract

Background.

Marfan syndrome, "the founding member" of the heritable disorders of connective tissue, is a common autosomal dominant disorder with highly variable clinical manifestations in the skeletal, ocular, and cardiovascular systems. The fundamental defect leading to this disease has escaped definition despite decades of research efforts by several groups of investigators.

Methods and Results.

Using linkage analyses with polymorphic markers of the human genome, we mapped the genetic defect to chromosome 15 in five families with Marfan syndrome. With three polymorphic markers we obtained definitive proof of linkage in these families (lod score = 3.92, θ = 0.0±0.11). The most probable location of the gene for the disease is currently D15S45 (lod score = 3.32, θ = 0.0±0.12).

Conclusions.

The chromosomal localization of the mutation in Marfan syndrome is a first step toward the isolation and characterization of the defective gene and serves as a diagnostic test in families in which cosegregation of these markers with the disease has been confirmed. (N Engl J Med 1990; 323:935–9.)

Introduction

MARFAN syndrome is one of the most common inherited connective-tissue disorders, with an estimated prevalence of 40 to 60 cases per million population.1 It is inherited in an autosomal dominant fashion, although it is sporadic in 15 percent of cases.1 The most prominent clinical manifestations of the disorder occur in the skeletal, ocular, and cardiovascular systems.1 Diagnosis has been problematic because of the extreme variability of clinical expression. The current diagnostic criteria were established at the Seventh International Congress on Human Genetics in 1986 and defined at the First International Symposium on Marfan Syndrome in 1988.2 , 3 These criteria include "more specific manifestations," which are ectopia lentis, aortic-root dilatation, aortic dissection, and durai ectasia; and "other manifestations" in the musculoskeletal, cardiovascular, ocular, integumentary, pulmonary, and central nervous systems. The penetrance of the disease is considered complete, but variable expression of its clinical manifestations is the rule even within a single family.4

Despite intensive research carried out in various laboratories, nothing is known about the genetic defect leading to Marfan syndrome. Several abnormalities of connective-tissue proteins have been observed in patients,5 6 7 8 9 10 11 but their role as the primary defect in the disease is unclear. In fact, linkage analyses have allowed many of the genes coding for these proteins to be excluded as the defective gene in the syndrome. The excluded genes include genes coding for Type I, Type II, and Type III collagens and fibronectin, as well as genes coding for one polypeptide chain of Type V and Type VI collagens.12 13 14 15

Over the past four years we have been collecting information on Finnish families with Marfan syndrome. We studied eight three-generation families, each with a minimum of three living members who were affected. In these families, linkage analysis of several candidate genes resulted in their exclusion.15 These findings have been included in an international collaborative study by nine laboratories in the United States, Denmark, the United Kingdom, Finland, and France, which recently produced a preliminary exclusion map for Marfan syndrome.16 This map was constructed by combining genetic data from linkage analyses of Marfan syndrome and 75 informative loci on 18 autosomes, and excludes 75 percent of loci in the genome as likely locations of the defective gene for the syndrome.16 The map suggests that the gene is located on chromosome 5p, 6q, 8, 9p, 10p, 13, 15, 17p, 20p, 21, or 22. Encouraged by this study, we used the polymorphic markers of these "candidate chromosomes" to evaluate eight Finnish families with Marfan syndrome. Here we report evidence of the location of the gene for this common connective-tissue disorder in the human genome, presenting a definitive linkage between the disease and three polymorphic markers in chromosome 15.

Methods

Families

Figure 1. Figure 1. Pedigrees of the Two Families Most Informative for Marfan Syndrome.

The respective lod scores for families 1 and 2 were 1.82 and 1.05 for D15S45 (θ = 0.0), and 1.95 and 1.10 for D15S29 (θ = 0.0). A+ and A− indicate the presence and absence, respectively, of the 3.7-kb and 4.0-kb alleles of locus D15S45; B+ and B−, the presence and absence of the 2.4-kb and 2.5-kb alleles of locus D15S25; and C+ and C−, the presence and absence of the 1.1-kb and 2.5-kb alleles of locus D15S29. The clinical manifestations in each member evaluated are indicated as the numbers of major (Ma) and minor (Mi) features observed during evaluation (see Methods).

Squares denote male subjects; circles, female subjects; symbols with diagonals, dead subjects; open symbols, unaffected subjects; solid symbols, subjects with the syndrome; and half-solid symbols, obligate carriers. ND denotes not determined.

The results described below are based on studies of five families with Marfan syndrome that were informative with respect to the polymorphic markers used and that could be evaluated by linkage analysis. The probands were referred to the Genetics Clinic or Cardiovascular Clinic of Helsinki University Central Hospital for the evaluation of possible Marfan syndrome and for genetic counseling. Diagnosis relied on a detailed physical examination; cardiovascular ultrasonograms; radiographs, including those obtained for determination of the metacarpal index; and a slit-lamp ophthalmologic evaluation. All children under the age of 16 in these families were carefully examined with aortic ultrasonography for any evidence of dilatation of the aortic root. The diagnostic criteria were those defined at the First International Symposium on Marfan Syndrome in 1988.2 , 3 The features found in each patient were classified as more specific manifestations (Fig. 1, "major features") or other manifestations ("minor features") observed during evaluation.2 The identification of the healthy members of the families was based on personal examination or on history and physical measurements given by the members.

DNA Analysis

DNA was isolated from fresh and frozen peripheral blood according to the technique described by Vandenplas et al.17 Five micrograms of DNA was digested with RsaI, HindIII, TaqI, and EcoRI restriction enzymes. The resulting fragments were separated on 0.8 percent agarose gels and transferred to Hybond-N membranes (Amersham) according to the method of Southern18 with slight modifications. Prehybridization and hybridization with probes Radio-labeled to a specific activity of 1 to 2×109 dpm per microgram by means of the primer-extension reaction (Random Primed DNA labeling kit, Boehringer–Mannheim) were performed according to standard procedures.17 Autoradiography was carried out on Kodak X-Omat films with intensifying screens for one to seven days at −80°C. The probes used were established polymorphic markers of chromosome 15: pTHH114 (D15S25),19 pEFZ33 (D15S45),20 pEFD49.3 (D15S29),21 and pEFD85.7 (D15S37)22 (kindly provided by Dr. R. White).

Linkage Analysis

Linkage analysis was carried out with the LINKAGE package of computer programs (version 5.03,23 updated by Dr. J. Ott). The programs calculate the odds of the observed inheritance pattern for markers linked at hypothetical distances on the chromosome map and for unlinked, randomly assorting markers. An odds ratio of more than 1000:1 in favor of linkage (expressed on a logarithmic scale as a lod score of more than 3) is considered a statistically significant demonstration of linkage in humans. On the other hand, an odds ratio of less than 1:100 (lod score less than −2) is agreed to indicate the exclusion of linkage. Approximate confidence intervals (indicated by ±) are obtained in the usual manner (maximal lod score minus 1).

The linkage analysis was carried out with different recombination fractions for male and female subjects. Since no significant differences were obtained, the results are expressed in terms of the recombination fractions for the male subjects. Because clinical signs of Marfan syndrome can appear later in childhood or in early adulthood,1 we adopted three age-dependent categories of penetrance for the disease phenotype: penetrance of 40 percent among children under the age of 10, penetrance of 60 percent among children from 10 to 16 years of age, and penetrance of 90 percent among all subjects over 16 years of age (the presence of clinical manifestations and a positive family history are considered to permit the diagnosis to be made in 90 percent of cases).1 , 15 , 24 We did not allow for a new mutation in the linkage analyses.

Results

Table 1. Table 1. Lod Scores for Marfan Syndrome, According to Two-Point Linkage Analyses with Polymorphic Markers on Chromosome 15.

Using the preliminary exclusion map of Marfan syndrome produced by international collaboration,16 we began testing of the study families for genetic linkage to polymorphic markers on chromosome 15. We chose three markers at different locations on the chromosome — i.e., D15S25 (nearest the centromere), D15S45, and D15S37 (nearest the telomere); the assigned distances between them are 24 cM (centimorgans) and 35 cM, respectively.19 We could exclude D15S37 as the site of the Marfan syndrome mutation since the lod score was −2.33 (θ = 0.1) (Table 1). In contrast, with the use of two other markers, D15S45 and D15S25, as well as an additional marker in the immediate vicinity, D15S29, we obtained strongly positive values for the lod score and were able to demonstrate a linkage of Marfan syndrome to this area of chromosome 15.

Figure 2. Figure 2. Location Map Summarizing Lod Scores Calculated for Marfan Syndrome in Five Families, at Various Positions in a Fixed-Marker Map.

The relative genetic position of D15S45 was arbitrarily placed at 0 cM. The distances of the three markers from D15S1, which has been physically mapped to chromosome 15,25 are also shown.

In two-point linkage analyses the maximal lod score obtained with the D15S45 marker20 was 3.32 (θ = 0.0±0.12); that with the D15S29 marker,21 3.24 (θ = 0.0±0.14); and that with the D15S25 marker,19 1.40 (θ = 0.14) (Table 1). When the multipoint likelihood was calculated in relation to all three markers with the LINKMAP program (part of the LINKAGE package), the maximal lod score was 3.92 (θ = 0.0±0.11). The multipoint-likelihood calculations are summarized in the location map shown in Figure 2. The published order of the three loci demonstrating a linkage to Marfan syndrome19 is cen—D15S25—D15S45—D15S29—ter (θ =0.18 for the distance between D15S25 and D15S45 and θ = 0.01 for the distance between D15S45 and D15S29). However, the odds favoring this order over the order cen—D15S25—D15S29—D15S45—ter are not significant.19 Our data suggested the latter order, and when we used it in our linkage analyses, the lod score obtained was even higher, 4.24 (θ = 0.0±0.11). The data provide statistically significant evidence of a close genetic linkage of Marfan syndrome to these markers on chromosome 15.

The linkage map for Marfan syndrome shown in Figure 2 is based on data on the families studied and spans a distance of about 50 cM, corresponding to 50 million nucleotides of chromosome 15 on the physical scale. The most probable location of the Marfan syndrome mutation in this chromosomal area is currently D15S45. In the published linkage map of chromosome 15,19 which gives reference points for 146 cM, this marker is located about 60 cM from the marker nearest the centromere, D15S24. Thus, the tentative gene for the disorder should be located about 60 cM from the centromere and 86 cM from the marker nearest the telomere in the established linkage map of chromosome 15.19 (The recombination fractions between the loci were transformed into map distances with the use of Haldane's formula.26) The absence of recombinations in the meiotic events studied thus far prevents a more precise determination of the genetic distance of the Marfan syndrome mutation from D15S45. The actual distance can be calculated only after a larger number of families with this disease have been studied.

Discussion

Linkage analysis with polymorphic markers, milestones of the human genome, can be used effectively to pinpoint the locus of mutations in mendelian disorders.27 , 28 However, the study of Marfan syndrome according to this approach has been complicated by at least two factors: difficulty in identifying extended pedigrees with affected members over several generations, because of the shortened life expectancy of patients; and difficulty in establishing a reliable diagnosis in young children. Also, the possibility of incomplete penetrance, suggested by early studies, and the wide variability of clinical manifestations among affected family members make reliable linkage analyses difficult.4 In view of these problems, we adopted three age-dependent categories of penetrance for our linkage analyses (see Methods). Although such categorization reduces the power of linkage analyses, it is a necessity in studying a disease with a high degree of heterogeneity in its clinical features, such as Marfan syndrome. Furthermore, we combined extensive clinical examination of the family members with maximal efforts at diagnosis in the children. Taken together, these steps allow a cautious and careful search for the gene responsible for the syndrome.

Marfan syndrome is phenotypically heterogeneous. This has led to speculations that mutations at different loci produce the same set of clinical manifestations in different families. Five of the eight families that we studied had positive lod scores for the markers D15S45, D15S29, and D15S25, whereas the other three remained uninformative. None of the families had negative lod scores for these markers in a pairwise analysis performed with the MLINK program (part of the LINKAGE package) (data not shown). We conclude that in these families, one or several mutations in one locus of chromosome 15 result in Marfan syndrome. We suggest that this tentative locus be designated as MFS1. However, the possibility of other mutated loci in this or other populations cannot be excluded. Further analysis of families in other populations are in progress and may resolve questions about the genetic heterogeneity of Marfan syndrome.

Chromosome 15 contains some interesting genes with respect to connective-tissue disorders. The genes coding for Type I collagen receptor,29 chondroitin sulfate proteoglycan I core protein,30 and cardiac muscle α-actin31 are all located in chromosome 15 and are also candidate genes for the mutation in Marfan syndrome. These and other genes mapped to this chromosomal area deserve further investigation.

The linkage of Marfan syndrome to chromosome 15 presented here will allow diagnostic testing for the disorder in persons at risk, at least in families in which cosegregation between the disease and the markers described above can be demonstrated. Our study is also a step toward the isolation and characterization of the responsible mutation. Marfan syndrome is an especially fascinating mendelian disorder, since its tissue manifestations are highly variable and represent a spectrum of symptoms also found in more common connective-tissue disorders, including those causing mitral-valve prolapse and aortic aneurysm. Elucidation of the molecular mechanisms behind Marfan syndrome should also provide insights into these more common phenomena of tear and wear in connective tissue.

Funding and Disclosures

Supported by grants from the Academy of Finland, the Paulo Foundation, the Meilahti Foundation, the Research Foundation of Farmos, and the Orion Medical Company.

We are indebted to Dr. Victor McKusick for his help in the critical evaluation of the manuscript; to Dr. Ray White for providing the probes used in the study; to Dr. Jaakko Leisti for clinical evaluations of four members of the study families; to Dr. Jürg Ott for the computer programs and advice about linkage analyses; to Professor Kari Kivirikko for never-failing support and guidance; to Dr. Aarno Palotie and Mr. Jari Kainulainen for constant help and guidance in the linkage analyses; and to Ms. Tuula Manninen and Ms. Lea Puhakka for sophisticated technical help.

Author Affiliations

From the Laboratory of Molecular Genetics, National Public Health Institute, Helsinki (K.K., L. Peltonen); the Department of Clinical Genetics, Kuopio University Central Hospital, Kuopio, Finland (L. Pulkkinen); the First Department of Medicine, Helsinki University Central Hospital, Helsinki (A.S.); and the Department of Medical Genetics, University of Helsinki, Helsinki (I.K.). Address reprint requests to Dr. Peltonen at the Laboratory of Molecular Genetics, National Public Health Institute, Mannerheimintie 166, SF-00300 Helsinki, Finland.

References (31)

  1. 1. Pyeritz RE, McKusick VA. . The Marfan syndrome: diagnosis and management . N Engl J Med 1979; 300:772–7.

  2. 2. Beighton P, de Paepe A, Danks D, et al. . International Nosology of Heritable Disorders of Connective Tissue, Berlin, 1986 . Am J Med Genet 1988; 29:581–94.

  3. 3. Pyeritz RE. . Pleiotropy revisited: molecular explanations of a classic concept . Am J Med Genet 1989; 34:124–34.

  4. 4. Pyeritz RE, Murphy EA, McKusick VA. Clinical variability in the Marfan syndrome(s). In: O'Donnell JJ, Hall BO, eds. Penetrance and variability in malformation syndromes. Vol. 15. No. 5B of Birth defects original article series. New York: Alan R. Liss, 1979:155–78.

  5. 5. Appel A, Horwitz AL, Dorfman A. . Cell-free synthesis of hyaluronic acid in Marfan syndrome . J Biol Chem 1979; 254:12199–203.

  6. 6. Scheck M, Siegel RC, Parker J, Chang Y-H, Fu JC. . Aortic aneurysm in Marfan's syndrome: changes in the ultrastructure and composition of collagen . J Anat 1979; 129:645–57.

  7. 7. Boucek RJ, Noble NL, Gunja-Smith Z, Butler WT. . The Marfan syndrome: a deficiency in chemically stable collagen cross-links . N Engl J Med 1981; 305:988–91.

  8. 8. Byers PH, Siegel RC, Peterson KE. et al. . Marfan syndrome: abnormal α2 chain in type I collagen . Proc Natl Acad Sci U S A 1981; 78:7745–9.

  9. 9. Abraham PA, Perejda AJ, Carnes WH, Uitto J. . Marfan syndrome: demonstration of abnormal elastin in aorta . J Clin Invest 1982; 70:1245–52.

  10. 10. Godfrey M, Olson S, Burgio RG, et al. . Unilateral microfibrillar abnormalities in a case of asymmetric Marfan syndrome . Am J Hum Genet 1990; 46:661–71.

  11. 11. Godfrey M, Menashe V, Weleber RG, et al. . Cosegregation of elastin-associated microfibrillar abnormalities with the Marfan phenotype in families . Am J Hum Genet 1990; 46:652–60.

  12. 12. Tsipouras P, Borresen A-L, Bamforth S, Harper PS, Berg K. . Marfan syndrome: exclusion of genetic linkage to the COL1A2 gene . Clin Genet 1986; 30:428–32.

  13. 13. Dalgleish R, Hawkins JR, Keston M. . Exclusion of the α2(I) and α1(III) collagen genes as the mutant loci in a Marfan syndrome family . J Med Genet 1987; 24:148–51.

  14. 14. Ogilvie DJ, Wordsworth BP, Priestley LM, et al. . Segregation of all four major fibrillar collagen genes in the Marfan syndrome . Am J Hum Genet 1987; 41:1071–82.

  15. 15. Kainulainen K, Savolainen A, Palotie A, Kaitila I, Rosenbloom J, Peltonen L. . Marfan syndrome: exclusion of genetic linkage to five genes coding for connective tissue components in the long arm of chromosome 2 . Hum Genet 1990; 84:233–6.

  16. 16. Blanton SH, Sarfarazi M, Eiberg H, et al. . An exclusion map of Marfan syndrome . J Med Genet 1990; 27:73–7.

  17. 17. Vandenplas S, Wiid I, Grobler-Rabie A, et al. . Blot hybridisation analysis of genomic DNA . J Med Genet 1984; 21:164–72.

  18. 18. Southern EM. . Detection of specific sequences among DNA fragments separated by gel electrophoresis . J Mol Biol 1975; 98:503–17.

  19. 19. Nakamura Y, Lathrop M, O'Connell P, Leppert M, Lalouel J-M, White R. . A mapped set of DNA markers for human chromosome 15 . Genomics 1988; 3:342–6.

  20. 20. Fujimoto E, Nakamura Y, O'Connell P, et al. . Isolation and mapping of a polymorphic DNA sequence (pEFZ33) on chromosome 15 [D15S45] . Nucleic Acids Res 1988; 16:10946.

  21. 21. Fujimoto. . Isolation and mapping of a polymorphic DNA sequence (pEFD49.3) on chromosome 15 [D15S29] . Nucleic Acids Res 1988; 16:10944.

  22. 22. Fujimoto E, Nakamura Y, Kanamori M, et al. . Isolation and mapping of a polymorphic DNA sequence (pEFD85.7) on chromosome 15 (D15S37) . Nucleic Acids Res 1988; 16:6256.

  23. 23. Lathrop GM, Lalouel JM, Julier C, Ott J. . Strategies for multilocus linkage analysis in humans . Proc Natl Acad Sci U S A 1984; 81:3443–6.

  24. 24. Tsipouras P. . Marfan syndrome: light at the end of the tunnel? Am J Hum Genet 1990; 46:643–5.

  25. 25. Brissenden JE, Page DC, de Martinville B, Trowsdale J, Botstein D, Francke U. . Regional assignments of three polymorphic DNA segments on human chromosome 15 . Genet Epidemiol 1986; 3:231–9.

  26. 26. Haidane JBS. . The combination of linkage values, and the calculation of distances between the loci of linked factors . J Genet 1919; 8:299–309.

  27. 27. Gusella JF, Wexler NS, Conneally PM, et al. . A polymorphic DNA marker genetically linked to Huntington's disease . Nature 1983; 306:234–8.

  28. 28. Barker D, Wright E, Nguyen K, et al. . Gene for von Recklinghausen neurofibromatosis is in the pericentromeric region of chromosome 17 . Science 1987; 236:1100–2.

  29. 29. Pignatelli M, Bodmer WF. . Genetics and biochemistry of collagen binding-triggered glandular differentiation in a human colon carcinoma cell line . Proc Natl Acad Sci U S A 1988; 85:5561–5.

  30. 30. Rettig WJ, Real FX, Spengler BA, Biedler JL, Old LJ. . Human melanoma proteoglycan: expression in hybrids controlled by intrinsic and extrinsic signals . Science 1986; 231:1281–4.

  31. 31. Gunning P, Ponte P, Kedes L, Eddy R, Shows T. . Chromosomal location of the co-expressed human skeletal and cardiac actin genes . Proc Natl Acad Sci U S A 1984; 81:1813–7.

Citing Articles (274)

    Figures/Media

    1. Figure 1. Pedigrees of the Two Families Most Informative for Marfan Syndrome.
      Figure 1. Pedigrees of the Two Families Most Informative for Marfan Syndrome.

      The respective lod scores for families 1 and 2 were 1.82 and 1.05 for D15S45 (θ = 0.0), and 1.95 and 1.10 for D15S29 (θ = 0.0). A+ and A− indicate the presence and absence, respectively, of the 3.7-kb and 4.0-kb alleles of locus D15S45; B+ and B−, the presence and absence of the 2.4-kb and 2.5-kb alleles of locus D15S25; and C+ and C−, the presence and absence of the 1.1-kb and 2.5-kb alleles of locus D15S29. The clinical manifestations in each member evaluated are indicated as the numbers of major (Ma) and minor (Mi) features observed during evaluation (see Methods).

      Squares denote male subjects; circles, female subjects; symbols with diagonals, dead subjects; open symbols, unaffected subjects; solid symbols, subjects with the syndrome; and half-solid symbols, obligate carriers. ND denotes not determined.

    2. Table 1. Lod Scores for Marfan Syndrome, According to Two-Point Linkage Analyses with Polymorphic Markers on Chromosome 15.
      Table 1. Lod Scores for Marfan Syndrome, According to Two-Point Linkage Analyses with Polymorphic Markers on Chromosome 15.
    3. Figure 2. Location Map Summarizing Lod Scores Calculated for Marfan Syndrome in Five Families, at Various Positions in a Fixed-Marker Map.
      Figure 2. Location Map Summarizing Lod Scores Calculated for Marfan Syndrome in Five Families, at Various Positions in a Fixed-Marker Map.

      The relative genetic position of D15S45 was arbitrarily placed at 0 cM. The distances of the three markers from D15S1, which has been physically mapped to chromosome 15,25 are also shown.

      More Research