Mapping of both autosomal recessive and dominant variants of pseudoxanthoma elasticum to chromosome 16p13.1
Mapping of both autosomal recessive and dominant variants of pseudoxanthoma elasticum to chromosome 16p13.1Berthold Struk1,2,6, Kenneth H. Neldner4, Valluri S. Rao1,3, Pamela St Jean5 and Klaus Lindpaintner1,2,3,6,*
1Cardiovascular Division, Department of Medicine, Brigham and Women's Hospital, Boston, MA 02115, USA, 2Department of Cardiology, Children's Hospital, Boston, MA 02115, USA, 3Department of Medicine, Harvard Medical School, Boston, MA 02115, USA, 4Department of Dermatology, Texas Tech University Health Sciences Center, Lubbock, TX 79430, USA, 5Department of Epidemiology and Biostatistics Case Western Reserve University, Cleveland, OH 44109, USA and 6Max Delbrück Centre for Molecular Medicine, D-13122 Berlin-Buch, Germany
Received February 18, 1997;Revised and Accepted July 28, 1997
Pseudoxanthoma elasticum (PXE) is a classic inherited disorder of the elastic tissue characterized by progressive calcification of elastic fibers with a pathognomonic histological appearance. The clinical manifestations of PXE typically involve the skin, the eye and the cardiovascular system, resulting in skin lesions, decreased vision and vascular disease. Clinically, a more common autosomal recessive and a less common autosomal dominant pattern of inheritance, with high penetrance, have been described; the estimated prevalence of the disease is 1 in 70 000-100 000. Previous failure to link the disease to any of several candidate genes prompted us to conduct a genome-wide screen on a collection of 38 families with two or more affected siblings, using allele sharing algorithms. Excess allele sharing was found on the short arm of chromosome 16 and confirmed by conventional linkage analysis, localizing the disease gene under a recessive model with a maximum two point lod score of 21.27 on chromosome 16p13.1, an area so far devoid of any obvious candidate genes. Under a dominant transmission pattern linkage with a maximum two point lod score of 14.53 was observed to the same region. Linkage heterogeneity analysis predicted the presence of allelic heterogeneity with different variants of a single gene that resides in this chromosomal region accounting for recessive and dominant forms of PXE.
Pseudoxanthoma elasticum (PXE) was first described as an independent disease entity in 1896 by the French dermatologist Jean Darier, who characterized it histologically as a progressive fragmentation of the elastic tissue of the papillary dermis (1 ). Subsequently, PXE was recognized as a multisystem disorder of the elastic tissue leading not only to skin disease but also to ocular (angioid streaks) (2 ) and vascular complications (3 ). The typical cutaneous manifestations of PXE (`cobblestone'-like changes with a predilection for flexural surfaces) (1 ,3 ,4 ) are mostly cosmetic in nature; however, its cardiovascular complications, while rare, can be life threatening (mainly gastrointestinal hemorrhage and coronary disease) (5 -8 ) and its ocular manifestations (fractures of Bruchs' membrane resulting in so-called `angioid streaks' followed by neovascularization) are common, leading to retinal hemorrhage and consecutive central blindness in about half of all affected individuals (2 ,3 ).
PXE represents a paradigm for disorders of heterotopic tissue calcification, with the ultrastructural defect characterized by a centrifugally progressive accumulation of calcium salt deposits (CaCO3 and CaPO4) within elastic fibers, starting in the core and leading, eventually, to their fracture and destruction (9 -11 ). Previous efforts to link the disease in limited numbers of families to several potential candidate genes (such as elastin, fibrillins I and II and lysyl oxidase) were negative or equivocal (12 ); likewise, no biochemical abnormality, except for an accumulation of glycosaminoglycans in the affected skin (13 ), has reproducibly been shown to be associated with the disease. In the majority of cases PXE appears to be inherited as an autosomal recessive trait (horizontal pattern of inheritance) with high penetrance and early onset (on average at 13 years of age), occurring about twice as often in females as in males. In ~10% of pedigrees an autosomal dominant mode of transmission (vertical transmission pattern) appears to be operative. The phenotypic manifestations of PXE show considerable variance, which was found to be unrelated to mode of apparent inheritance (14 ), with regard to system(s) affected and severity of disease (3 ,15 ). Presently, the most widely accepted gold standard for diagnosis of PXE relies on positive von Kossa staining in biopsy material from affected skin, indicating calcification and fracture of elastic fibers.
In this study we report mapping of the recessive and dominant forms of PXE to the same chromosomal region, 16p13.1, an area without any apparent candidate gene for the disease.
Column two indicates the number of families included in the initial genome screen analysis using SAGE, carried out in all families showing apparent recessive inheritance that were available at the initiation of molecular genetic studies (at this time the family with five affected siblings was considered autosomal recessive; this was later revised to autosomal dominant upon availability of additional information) and using only affected sib pairs (no other family members). Columns 3 and 4 present information on pedigrees selected for parametric linkage analysis according to apparent mode of inheritance and under inclusion of all available family members regardless of affection status. Numbers of affected sib pairs in a family with n affected sibs are calculated as n(n - 1)/2. n/a, not applicable.
. Results of allele sharing analysis, using all available family members and high density markers for chromosome 16p13
Marker
APM
SIBPAL
/TR>
T statistic
P value
Mean [pi] (SE)
Z value
P value
TR>
D16S406
3.93
0.00004
0.73(0.03)
8.40
<0.0001
D16S407
3.24
0.00061
0.70 (0.03)
6.58
<0.0001
D16S748
6.90
<0.00001
0.77 (0.03)
9.75
<0.0001
D16S3114
7.71
<0.00001
0.84 (0.02)
14.45
<0.0001
D16S3069
4.52
<0.00001
0.76 (0.03)
9.55
<0.0001
D16S500
5.36
<0.00001
0.83 (0.02)
13.68
<0.0001
D16S405
4.73
<0.00001
0.77 (0.02)
11.54
<0.0001
D16S3079
4.31
0.00001
0.80 (0.03)
11.84
<0.0001
D16S3103
4.80
<0.00001
0.76 (0.03)
9.62
<0.0001
D16S499
4.81
<0.00001
0.73 (0.03)
9.09
<0.0001
D16S3036
4.36
0.00001
0.74 (0.03)
9.75
<0.0001
D16S403
2.69
0.00354
0.69 (0.03)
5.82
<0.0001
Columns 2 and 3 indicate the T statistics and associated P values from the APM program. In this analysis 45 families with 113 affected members were used. The estimated proportion of alleles shared identical-by-descent, mean [pi], among 79 affected sib pairs obtained from SIBPAL is given in column 4, with the corresponding Z statistic and P value in columns 5 and 6 respectively.
Most of the patients and families investigated in the present study had originally been recruited for a previous clinical study of PXE performed at the University of Colorado Health Science Center and at Texas Tech University Health Sciences Center over a 20 year period from 1973 to 1993 (3 ). Others were obtained from the database of the National Association for PXE. Only biopsy-proven diagnoses with positive von Kossa staining were considered as PXE cases. Except for eight families, all showed a strictly horizontal pattern of disease prevalence, suggestive of autosomal recessive inheritance. In keeping with previous studies, we observed a female-to-male ratio of ~2:1 among our patients.
We followed a two stage strategy to map the causative gene, designated PXE (the gene name PXE has been approved by the HUGO Nomenclature Committee). First, a genome-wide screen using a panel of tri- and tetranucleotide single sequence length polymorphisms (16 -18 ) was performed on 73 affected sib pairs belonging to 38 families (see Table 1 ). Data were analyzed by non-parametric allele sharing methods (SIBPAL) (19 -21 ). Three markers on the short arm of chromosome 16, D16S2619, D16S748 and D16S403, were found to show significant excess allele sharing compared with chance (61 versus 50%, P = 0.0005, 66 versus 50%, P = 0.0002, and 67 versus 50%, P = 0.0003, respectively).
Based on this information, we proceeded to genotype all 264 affected and unaffected subjects belonging to the 42 pedigrees with an apparently recessive inheritance pattern, as well as those belonging to the eight families with an apparently dominant inheritance pattern, using a set of high resolution, highly informative markers (PIC > 0.7). Twelve microsatellite markers fulfilling these conditions were tested. They were selected from the collections assembled by the Cooperative Human Linkage Center (CHLC; Extended Généthon V2 Recombination-Minimization Sex Averaged Map, v.4.0) (16 -18 ) and by Généthon (22 ). All were localized in the area of interest and had an average spacing of 2.9 cM and an average polymorphism information content of 0.77.
Forty five of the 50 families had affected relative pairs other than parent-child pairs; they underwent non-parametric analysis by both the APM (23 ) and SIBPAL (21 ) methods. Results from these analyses were consistent with the original findings of linkage to the region, i.e. to chromosome 16p13, and with each other (Table 2 ). Both analyses provide significant evidence of linkage for markers spanning the region D16S406-D16S403, with a peak around D16S500.
Intermarker distances and orders were estimated using the family data and ILINK (24 ) and are shown in Table 3 . Parametric two point maximum lod scores (MLS) were then calculated for all 50 pedigrees assuming either autosomal recessive or autosomal dominant models of transmission respectively (Table 3 ). An MLS of 21.27 was found for D16S500 at [theta] = 0.045 assuming the recessive model and an MLS of 14.53 for the same marker at [theta] = 0.011 under an assumption of dominant inheritance. Model-based multipoint linkage analysis results using GENEHUNTER (25 ) are shown in Figure 1 . Under the recessive model (Fig. 1 , graph AR) an MLS of 21.8 was observed between D16S3079 and D16S3103. Allowing for heterogeneity, this lod score increased to 23.7 with an estimated 88% of families linked. The MLS obtained under the dominant model (Fig. 1 , graph AD) was 16.9. Non-parametric multipoint linkage analysis using GENEHUNTER NPL revealed similarly significant results (Fig. 1 , graph NPL), with a maximum NPL score of 9.81 (P = 2.47 * 10-25) close to D16S405. The information content of the map of chromosome 16 markers was >= 80% across the entire region. Thus both the model-based and model-free analyses provide highly significant evidence of linkage between PXE and markers in this region on chromosome 16.
Our present data provide robust evidence that PXE maps to chromosome 16p13.1, within a 4.8 cM region between D16S500 and D16S3103. A review of integrated map data available for chromosome 16 (28 ) reveals that the region of interest (16p13.1) has been saturated with ordered mega-YAC contigs as well as with cosmid and flow-sorted mini-YAC contigs. So far only two genes, MRP and MHY11, which code for a multiple drug resistance-associated protein and for smooth muscle myosin heavy chain-11 respectively, have been mapped to this region, which also represents the breakpoint for a pericentric inversion, in(16)(p13q22), associated with acute myeloid leukemia. The homologous region in the mouse, which maps to mouse chromosome 16, likewise does not contain any apparent candidate genes for PXE. Publicly accessible EST databases have also so far failed to point to a candidate disease gene.
Whereas our analyses provide strong evidence either for locus heterogeneity with two closely linked, but distinct, disease loci or for allelic heterogeneity, they do not presently allow a distinction between the two (although the results tend to favor the latter) and they may even be compatible with both. If, indeed, we are witnessing allelic heterogeneity with both recessive and dominant forms of PXE resulting from molecular variations of the same gene, then this raises intriguing questions about the nature of the gene product and of its respective mutations. By analogy to the hypothesis advanced for rhodopsin-linked retinitis pigmentosa, one of the few similar examples of allelic heterogeneity giving rise to dominant and recessive variants of the same disease (29 ,30 ), one might suspect the presence of unstable or non-translated `null' mutants in the recessive and of structurally abnormal (dominant negative acting) gene products in the dominant form of the disease.
Our localization of the causative gene is sufficiently precise to allow genetic diagnosis of affection or carrier status by linkage analysis in families with at least one affected member for both dominant and recessive forms of PXE. We expect that a continued search for additional informative meioses, coupled with ongoing efforts at targeted marker development and positional cloning, will lead us to the discovery of the causative gene and, in due course, open up specific approaches to prevention or therapy of PXE and its complications. Perhaps even more importantly, the status of PXE as a model for disorders of aberrant tissue calcification may give the present results potentially broader implications for our understanding of a number of similar disease states characterized by abnormal tissue calcification.
Recruitment of patients and their family members was done by K.H.N. at the Clinical Coordinating Center in Lubbock, based on this investigator's clinical database. Clinical data and blood samples were collected over a 2 year period from families with two or more affected individuals. Affection status was confirmed based on positive von Kossa staining in biopsy material in all patients. Only individuals older than 30 years who showed no evidence of cutaneous and/or retinal lesions were considered as unaffected family members for linkage analyses. All participants in the study provided written, informed consent using a form that was approved by the Institutional Review Board at Texas Tech University Health Sciences Center.
High molecular weight genomic DNA was extracted from 1 ml EDTA-anticoagulated whole blood using a commercially available, adsorption-based method (QIAmp; Qiagen), and stored in multiple aliquots in 0.1* TE at -80oC.
A subset of 169 CHLC microsatellite tri- and tetranucleotide markers (Weber 6a) with an average heterozygosity of 78% and an average spacing of 24.2 cM was purchased from Research Genetics (Huntsville, AL). Sequence information for additional microsatellites used for high resolution mapping of chromosome 16p was retrieved from either The Genome Data Base (GDB) (http://gdbwww.gdb.org/) or the CHLC (http://www.chlc.org/) database. Primers were designed such that amplification products for up to four markers could be co-electrophoresed without interference, to optimize the genotyping program. Primers were designed using the Oligotm program (31 ).
Twenty nanograms of genomic DNA were amplified by PCR in a final volume of 10 [mu]l containing 100 nM each primer, 200 [mu]M dNTPs, 1.5 mM MgCl2, 50 mM KCl, 10 mM Tris-HCl, 0.1% Triton X-100 and 0.025 U Taq DNA polymerase. Forward primers were labeled with [[gamma]-32P]ATP (3000 Ci/mmol; Dupont/NEN) using T4 polynucleotide kinase (NEB). Reactions were processed on a MJR Thermal Cycler (PTC 100; MJ Research, Watertown, MA) using an initial denaturation at 96oC for 3 min, followed by 35 cycles of annealing for 1 min, extension at 72oC for 1 min and denaturation at 96oC for 15 s, with a final extension step at 72oC for 7 min. Optimized annealing temperatures ranged from 55 to 65oC. PCR products were multiplex-loaded in groups of up to four markers and size fractionated on 6% polyacrylamide sequencing gels containing 37.5% formamide, 8 M urea, 90 mM Tris-borate and 2 mM EDTA (Sequageltm; National Diagnostics). Gels were run at 60 W for 3.5-6 h, depending on the PCR product size, and exposed to Kodak XAR-5 film at -80oC for autoradiography. All markers were scored independently by two observers who were blind as to diagnosis and family structure.
Prior to embarking on the genome-wide screen we carried out extensive simulations using the computer program SLINK (32 ), under the assumption of recessive inheritance and (near) complete penetrance. Results suggested an expected lod score of >4.0 for markers with a polymorphism information content (PIC) of >= 0.7, given the number of families and the average marker spacing we planned to use. For the genome-wide screen all 169 genotyped markers in 38 families were tested for linkage to affection status using a qualitative affected sib pair method as implemented in SIBPAL, in v.2.7 of SAGE (21 ). SIBPAL provides an estimate of the proportion of alleles shared identical-by-descent in affected sib pairs.
After excess sharing on chromosome 16 was revealed, an additional 12 families were collected (a total of 50 families) and 12 chromosome 16 microsatellites were genotyped and used in subsequent linkage analyses. Of the 50 families available, 45 contained affected relative pairs other than parent-child and there were a total of 79 affected sib pairs. APM (23 ) and SIBPAL were used to investigate excess allele sharing on chromosome 16 among affected relative pairs (APM) or affected sib pairs (SIBPAL). The APM method makes use of identity-by-state sharing and a weighting function for marker allele matches that takes into account that it is more striking for affected relatives to share rare alleles. The weighting function f(p) = 1/[radic]p, where p is the allele frequency, was used, as Weeks and Lange (23 ) have suggested that it is the best compromise between maintaining approximate normality with respect to the test statistic, T, and maximizing the power to detect linkage.
Two different disease models were utilized for the parametric analysis. For the autosomal recessive model we assumed a disease allele frequency of 2.5 * 10-3 with 90% penetrance and a phenocopy rate of 10-5, while for the autosomal dominant model we assumed a disease allele frequency of 10-5 with 50% penetrance and a phenocopy rate of 5 * 10-6. Two point linkage analysis was conducted using ILINK (24 ) with the FASTLINK program (33 ) for all 50 families under the two different models. Parametric and non-parametric multipoint linkage analysis was carried out using both GENEHUNTER (25 ) and VITESSE (27 ). For the multipoint analyses intermarker distances and marker order were estimated using ILINK, based on genotype data from all available family members. Analyses for homogeneity were conducted using the HOMOG programs (26 ), as detailed in Results, to test for general locus heterogeneity, for locus heterogeneity constrained to the region of interest and for allelic heterogeneity.
We thank the members of the PXE families for their participation in this study. This work was supported by Project Grant 695-0209 from the March of Dimes Birth Defects Foundation (White Plains, NY) and by a Pilot and Feasibility Grant from the Harvard Skin Disease Research Center at Brigham and Women's Hospital. K.L. is the recipient of a Research Career Development Award (K04-HL03138-01) from the National Heart, Lung, and Blood Institute. Some of the results were obtained using the program package SAGE which is supported by US PHS resource grant RR03655 from the Division of Research Resources. A preliminary report of this study was presented at the Fifth International Workshop on Human Chromosome 16, Toronto, Canada, March 3-4, 1997.
4 Eddy,D.D. and Farber,E.M. (1962) Pseudoxanthoma elasticum. Internal manifestations: a report of cases and a review of the literature. Arch. Dermatol., 86, 729-740.
5 Nishida,H., Endo,M., Koyanagi,H., Ichihara,T., Takao,A. and Maruyama,M. (1990) Coronary artery bypass in a 15-year-old girl with pseudoxanthoma elasticum. Annls Thoracic Surg., 49, 483-485.
6 McCreedy,C.A., Zimmerman,T.J. and Webster,S.F. (1989) Management of upper gastrointestinal hemorrhage in patients with pseudoxanthoma elasticum. Surgery, 105, 170-174.MEDLINE Abstract
7 Keim,H.J., Holtermuller,K.H., Walter,U., Bork,K., Rothmund,M. and Gabbert,H. (1980) Massive upper gastrointestinal bleeding as first clinical manifestation of pseudoxanthoma elasticum. Zeitschr. Gastroenterol., 18, 20-29.
8 Lebwohl,M., Halperin,J. and Phelps,R.G. (1993) Brief report: occult pseudoxanthoma elasticum in patients with premature cardiovascular disease. New Engl. J. Med., 329, 1237-1239.MEDLINE Abstract
9 Huang,S.N., Steele,H.D., Kumar,G. and Parker,J.O. (1967) Ultrastructural changes of elastic fibers in pseudoxanthoma elasticum. A study of histogenesis. Arch. Pathol., 83, 108-113.MEDLINE Abstract
10 Otkjaer-Nielsen,A., Johnson,E., Hentzer,B., Danielsen,L. and Carlsen,F. (1977) Apatite crystals in pseudoxanthoma elasticum: a combined study using electron microscopy and selected area diffraction analysis. J. Invest. Dermatol., 69, 376-378.MEDLINE Abstract
11 Calap,J., Mascaro,J.M. and Alonso,J. (1977) Scanning electron microscopic and X ray microanalysis study of the mineral deposits in pseudoxanthoma elasticum. Arch. Dermatol. Res., 259, 91-96.
12 Christiano,A.M., Lebwohl,M.G., Boyd,C.D. and Uitto,J. (1992) Workshop on pseudoxanthoma elasticum: molecular biology and pathology of the elastic fibers. Jefferson Medical College, Philadelphia, Pennsylvania, June 10, 1992. J. Invest. Dermatol., 99, 660-663.MEDLINE Abstract
13 Smith,J.G.J., Davidson,E.A. and Clark,R.D. (1962) Dermal elastin in actinic elastosis and pseudoxanthoma elasticum. Nature, 195, 716-717.
14 Lebwohl,M., Neldner,K., Pope,F.M., de Paepe,A., Christiano,A.M., Boyd,C.D., Uitto,J. and McKusick,V.A. (1994) Classification of pseudoxanthoma elasticum: report of a consensus conference. J. Am. Acad. Dermatol., 30, 103-107.MEDLINE Abstract
15 McKusick,V.A. (1972) In McKusick,V.A. (ed.), Heritable Disorders of Connective Tissue, 4th Edn. C.V.Mosby Press, St Louis, MO, pp. 475-520.
16 Sheffield,V.C. et al. (1995) A collection of tri- and tetranucleotide repeat markers used to generate high quality, high resolution human genome-wide linkage maps. Hum. Mol. Genet., 4, 1837-1844.MEDLINE Abstract
17 Gastier,J.M., Pulido,J.C., Sunden,S., Brody,T., Buetow,K.H., Murray,J.C., Weber,J.L., Hudson,T.J., Sheffield,V.C. and Duyk,G.M. (1995) Survey of trinucleotide repeats in the human genome: assessment of their utility as genetic markers. Hum. Mol. Genet., 4, 1829-1836.MEDLINE Abstract
18 Murray,J.C. et al. (1994) A comprehensive human linkage map with centimorgan density. Cooperative Human Linkage Center (CHLC). Science, 265, 2049-2054.MEDLINE Abstract
19 Bishop,D.T. and Williamson,J.A. (1990) The power of identity-by-state methods for linkage analysis. Am. J. Hum. Genet., 46, 254-265.MEDLINE Abstract
20 Suarez,B.K. and Van Eerdewegh,P. (1984) A comparison of three affected-sib-pair scoring methods to detect HLA-linked disease susceptibility genes. Am. J. Med. Genet., 18, 135-146.MEDLINE Abstract
21 SAGE (1996) Statistical Analysis for Genetic Epidemiology, release 2.2, SIBPAL v.2.7. Lousiana State University Center, New Orleans, LA.
22 Dib,C. et al. (1996) A comprehensive genetic map of the human genome based on 5,264 microsatellites. Nature, 380, 152-154.MEDLINE Abstract
23 Weeks,D.E. and Lange,K. (1988) The affected-pedigree-member method of linkage analysis. Am. J. Hum. Genet., 42, 315-326.MEDLINE Abstract
24 Lathrop,G.M. and Lalouel,J.M. (1984) Easy calculations of lod scores and genetic risks on small computers. Am. J. Hum. Genet., 36, 460-465.MEDLINE Abstract
25 Kruglyak,L., Daly,M.J., Reeve-Daly,M.P. and Lander,E.S. (1996) Parametric and non-parametric linkage analysis: a unified multipoint approach. Am. J. Hum. Genet., 58, 1347-1363.MEDLINE Abstract
26 Ott,J. (1991) Analysis of Human Genetic Linkage, 2nd Edn. Johns Hopkins University Press, Baltimore, MD, p. 203.
27 O'Connell,J.R. and Weeks,D.E. (1995) The VITESSE algorithm for rapid exact multilocus linkage analysis via genotype set-recoding and fuzzy inheritance. Nature Genet., 11, 402-408.MEDLINE Abstract
28 Doggett,N.A. et al. (1995) An integrated physical map of human chromosome 16. Nature, 377, 335-365.MEDLINE Abstract
29 Dryja,T.P., Hahn,L.B., Cowley,G.S., McGee,T.L. and Berson,E.L. (1991) Mutation spectrum of the rhodopsin gene among patients with autosomal dominant retinitis pigmentosa. Proc. Natl. Acad. Sci. USA, 88, 9370-9374.MEDLINE Abstract
30 Rosenfeld,P.J., Cowley,G.S., McGee,T.L., Sandberg,M.A., Berson,E.L. and Dryja,T.P. (1992) A null mutation in the rhodopsin gene causes rod photoreceptor dysfunction and autosomal recessive retinitis pigmentosa. Nature Genet., 1, 209-213.MEDLINE Abstract
31 Rychlik,W. and Rhoads,R.E. (1989) A computer program for choosing optimal oligonucleotides for filter hybridization, sequencing and in vitro amplification of DNA. Nucleic Acids Res., 17, 8543-8551.MEDLINE Abstract
32 Weeks,D.E., Ott,J. and Lathrop,G.M. (1990) A general simulation program for linkage analysis. Am. J. Hum. Genet., 47, A204.
33 Cottingham,R.W.,Jr, Idury,R.M. and Schaffer,A.A. (1993) Faster sequential genetic linkage computations. Am. J. Hum. Genet., 53, 252-263.MEDLINE Abstract
*To whom correspondence should be addressed at: Cardiovascular Division, Brigham and Women's Hospital, 75 Francis Street-Thorn 1103, Boston, MA 02115-6195, USA. Tel: +1 617 732 8173; Fax: +1 617 264 6830; Email: kl@calvin.bwh.harvard.edu
