Mapping of a second locus for lamellar ichthyosis to ...

 1996 Oxford University Press

Human Molecular Genetics, 1996, Vol. 5, No. 4

555–559

Mapping of a second locus for lamellar ichthyosis to chromosome 2q33–35 Laurent Parmentier, Hakima Lakhdar1, Claudine Blanchet-Bardon2, Sylvie Marchand, Louis Dubertret2 and Jean Weissenbach* CNRS URA 1922, Généthon, 1 rue de l′Internationale, 91000 Evry, France, 1Service de Dermatologie, Centre Hospitalier Universitaire Averoes, Casablanca, Morocco and 2Institut de Recherche sur la Peau, Hôpital Saint-Louis, 1 avenue Claude Vellefeaux, 75010 Paris, France Received December 1, 1995; Revised and Accepted January 18, 1996

INTRODUCTION The term ichthyosis designates a group of skin diseases characterized by a significant and incapacitating scaling of the skin. Most forms are congenital and display different modes of inheritance (1). Among them, lamellar ichthyosis (LI; MIM number 242100) is transmitted with an autosomal recessive pattern and occurs in 1 in 200 000–300 000 individuals. The diagnosis depends on a characteristic appearance of the skin which is covered by large scales encompassing all the body surface. No extracutaneous symptoms are observed. Although ichthyosis is constant, some patients present variable additional cutaneous signs. Most of them present a ‘collodion-baby’ syndrome at birth: the newborn is embedded in a hyperkeratotic skin, a life-threatening condition sometimes responsible for major dehydration or systemic infections. Spontaneous shedding of the membrane leaves residual ichthyosis and may lead to grave sequelae, such as joint contractures, ectropion and secondary corneal involvement. During adulthood, scarring alopecia, palmoplantar keratoderma

*To whom correspondence should be addressed

and associated erythroderma are sometimes observed, and are therefore features of phenotypic variability (1–3). Positive linkage of LI to chromosome 14q and to the transglutaminase 1 gene (TGM1) (4) was demonstrated. Deleterious mutations in TGM1 were identified (5–7) and it was unambiguously reported as an LI-causing gene. TGM1 is expressed exclusively in the keratinocyte and codes for an enzyme which catalyses cross-linking between different proteins, mostly involucrin, loricrin and small proline-rich proteins (SPRR) (8). The resulting macromolecular complex is linearly deposited on the inner side of the plasma membrane and is known as the cornified envelope (CE). The precise effects of each TGM1 mutation on the protein structure and enzymatic activity are unknown to date. However, absence of linkage between LI and TGM1 in 13 families from a cohort of 23 clearly demonstrated the existence of one or more other disease-causing gene(s) (7). Mutations in TGM1 were also recently excluded in two patients presenting with LI (9). Considering the key role of TGM1 in the formation of the CE, other genes involved in cornification could be suspected. These are especially other transglutaminases expressed in the epidermis (transglutaminases 2 and 3, which map to chromosome 20q11.2–q12) and the precursor proteins of the CE (involucrin, loricrin and SPRR, which are clustered on chromosome 1q21) (8,10,11). We present a linkage study on three unrelated LI consanguinous families from Morocco. All were selected on the basis of their common physical appearance and geographical origin, and all were first demonstrated to be unlinked to TGM1. We failed to reveal any significant linkage to candidate regions on chromosomes 1 and 20, but demonstrated strong linkage to markers mapped to chromosome 2q33–35. However, a genetic study of additional consanguinous families excluded linkage either to TGM1 or to 2q33–35. RESULTS Clinical data Three large Moroccan consanguinous families were selected. All members of these families underwent physical examination in the Department of Dermatology (Hospital Averoes ibn Rochd,

Downloaded from http://hmg.oxfordjournals.org/ by guest on June 2, 2013

Lamellar ichthyosis (LI) is an inherited autosomal recessive disorder of cornification. It was recently demonstrated to result from deleterious mutations in the transglutaminase 1 (TGM1) gene. However, the disease was shown to be genetically heterogeneous, since some families were found to be unlinked to TGM1. Homozygosity mapping on three consanguinous families originating from Morocco shows (i) absence of linkage with TGM1 and other regions of the genome containing genes involved in cornification, and (ii) location of a second disease-causing gene on chromosome 2q33–35. A maximum two-point lodscore of 7.60 was obtained with D2S157 for θ = 0. The analysis of recombination events places the gene within a 7–8 cM interval. Additional consanguinous pedigrees were also demonstrated to be unlinked both to TGM1 and to 2q33–35, suggesting the existence of at least a third disease-causing gene.

Human Molecular Genetics, 1996, Vol. 5, No. 4

556

Casablanca). Eight members were found to be affected by pure LI. Diagnosis was not ambiguous and all patients fulfil the diagnostic criteria for LI: collodion baby phenotype at birth, generalized ichthyosis with involvement of the large folds, palmoplantar keratoderma and absence of skin erythema. None of the patients had associated extracutaneous disorders, such as spastic paraplegia, polyneuritis, cerebellar ataxia or hepatosplenomegaly, which characterize Sjögren–Larsson syndrome, Refsum’s disease or Dorffman–Chamarin syndrome. The phenotype was consistent within each of the three families. Mild semiologic differences in the same family were observed in the extent of the scales and the consequently mechanical ectropion. No major differences in phenotypes between the families A, B and C were noted (results are summarized in Table 1). We studied an additional eight families of diverse geographical origin (France, Guadeloupe, Portugal, Tunisia and Turkey). All presented with an autosomal recessive generalized ichthyosis, and all except one had a typical history of collodion baby at birth. Nevertheless, phenotypic variability was observed: gravity of the ichthyosis, intensity of palmoplantar keratoderma and existence of an underlying erythroderma were not similar in all the families. They were therefore considered as a heterogeneous subgroup.

Linkage analysis

Table 1. Summary of the clinical features observed in the three families Family

A

B

C

Number of patients

3

4

1

Collodion baby at birth

3

4

1

Generalized ichthyosis

3

4

1

Severed

1

1

0

Moderate

2

3

1

Mild

0

0

0

Erythema

0

0

0

Palmoplantar keratoderma

3

4

1

Ectropion

3

4

1

Moderate

0

3

0

Mild

3

1

1

Table 2. Lodscore table. For each marker, two-point lodscore is reported according to the following order: family A, B and C

Two-point lodscores

Cumulative lodscores

θ

0.00

0.01

0.05

0.1

0.2

0.3

0.00

D2S325 D2S355 D2S322 D2S154 D2S157 D2S371 D2S317 D2S128 D2S334 D2S143 D2S137 D2S301

0.39/∞/0.31 0.92/1.84/0.79 2.99/3.30/0.11 3.07/3.17/–0.05 3.07/3.23/1.30 1.42/3.00/0.16 0.68/2.72/0.74 3.07/2.73/0.30 0.95/3.29/–0.15 2.49/2.92/1.37 ∞/3.28/0.12 ∞/3.23/0.53

0.37/1.24/0.30 0.90/1.80/0.76 2.92/3.24/0.11 3.01/3.11/–0.06 3.01/3.17/1.26 1.39/2.94/0.15 0.67/2.66/0.72 3.01/2.67/0.29 0.93/3.23/–0.16 2.43/2.86/1.34 0.83/3.17/0.51 0.46/3.22/0.11

0.32/1.71/0.25 0.82/1.66/0.64 2.66/2.99/0.09 2.75/2.87/–0.06 2.75/2.92/1.13 1.27/2.70/0.12 0.62/2.43/0.63 2.74/2.45/0.25 0.85/2.98/–0.15 2.19/2.62/1.20 1.29/2.92/0.45 0.93/2.99/0.08

0.26/1.72/0.20 0.72/1.48/0.51 2.34/2.68/0.07 2.42/2.61/–0.07 2.42/2.61/0.96 1.13/2.40/0.10 0.54/2.14/0.53 2.42/2.16/0.20 0.76/2.67/–0.13 1.89/2.32/1.03 0.96/2.69/0.02 1.29/2.61/0.37

0.14/1.42/0.11 0.52/1.10/0.28 1.69/2.02/0.03 1.76/1.91/–0.04 1.76/1.96/0.64 0.82/1.77/0.05 0.39/1.55/0.34 1.76/1.57/0.11 0.55/2.01/–0.08 1.31/1.71/0.70 0.72/2.06/0.02 1.00/1.96/0.23

0.06/0.98/0.05 0.31/0.72/0.13 1.06/1.35/0.01 1.13/1.25/–0.02 1.13/1.29/0.64 0.52/1.13/0.02 0.24/0.95/0.18 1.12/0.97/0.06 0.33/1.34/–0.03 0.75/1.08/0.40 0.41/1.39/0.01 0.63/1.29/0.12

∞ 3.54 6.40 6.19 7.60 4.58 4.14 6.10 4.08 6.78 ∞ ∞


As a first step, we focused our linkage study on the TGM1 locus, the loricrin–involucrin–SPRR locus on chromosome 1 and the TGM2 and TGM3 loci on chromosome 20q (see Materials and Methods for details). No region of homozygosity by descent was observed and haplotype analysis showed no evidence for linkage (data not shown). Subsequently, we undertook a genome-wide search with polymorphic microsatellite markers chosen from the GENETHON panel (12). Markers were selected on the basis of their high heterozygosity (>70%) and were on average spaced every 20 cM. A region of homozygosity was observed for all patients with markers from chromosome 2 (Fig. 1). Evidence of significant linkage was first observed with locus D2S143, which provides a cumulative pairwise lodscore of 6.78 at θ = 0. This positive finding was confirmed by testing neighbouring markers (Table 2). The maximum two-point lodscore (7.60 at θ = 0) was observed with locus D2S157. Multipoint linkage analysis was not feasible with the number of markers included in this interval and was considered computationally too complex. We preferred to establish the smallest co-segregating region in the three families using haplotype analysis. The haplotypes were constructed assuming the most parsimonious linkage phase. The telomeric limit is defined by marker D2S137 in family A. The centromeric limit can either be defined by markers D2S325 or D2S355 in family B, as the first marker is not polymorphic in the mother III-2 (see Fig. 1). Assuming these data, the interval may be either 7 or 8 cM (Fig. 2). We have mapped the gene to 2q33–35, since some markers flanking the LI locus have been mapped to this band by cytogenetic studies (13). We failed to observe a common haplotype in the three pedigrees and to demonstrate linkage disequilibrium between one or several markers and the disease (see Materials and Methods). Finally, we extended our linkage study to eight other consanguinous LI pedigrees, previously demonstrated to be unlinked to TGM1. No significant region of homozygosity was observed with the chromosome 2 marker set (not shown). For four of them, the locus was clearly unlinked, either by including or ignoring the consanguinity loops.

557 Human Acids Molecular Genetics, Vol.No. 5, No. Nucleic Research, 1994,1996, Vol. 22, 1 4

557

DISCUSSION The combination of linkage data and the recombinant analysis allowed us to position a second LI gene on chromosome 2q33–35. The current interval is 7–8 cM and is flanked by D2S137 and D2S325/D2S355. We also demonstrated the existence of at least a third LI locus, as other LI families were excluded both in this region and in the TGM1 locus. In terms of a physiological role, no obvious candidate gene maps to chromosome 2q33–35 (13). Genes coding for CRYG (responsible for hereditary cataracts) map precisely in the LI interval (13) but no association between these diseases and ichthyosis has been reported. As human 2q33–35 shows strong synteny with the mouse chromosome 1, we also searched for murine mutants mapping to this region (14). Tight-skin 2 (Tsk2), a murine cutaneous mutation, maps to the syntenic region. The phenotype of Tsk2 is very similar to tight-skin 1 (the mouse model for human scleroderma) which does not occur together with ichthyosis. We also did not retain the ichthyosis (ic) mutation, a possible model for human ichthyosis, as its position on chromosome 1 is too far from the mouse region of synteny with human 2q33–35. We therefore consider that no murine

model known to date will be useful in our search for the second LI gene. Concerning the physiopathology of ichthyosis, searching for new functional candidate genes can be approached in two main ways. On the one hand, we could postulate that the mechanism of LI is restricted to a disorder of the formation of CE. No known gene coding for a CE component maps to chromosome 2q. Nevertheless, it is to be expected that the CE involves other protein components which have not yet been described (15). We cannot exclude that the gene encoding one of them maps to our candidate interval. On the other hand, the disease mechanism could also involve a defect in the lipid envelope linked to the outer side of the CE (8). This envelope is composed for the most part of ceramides, the concentrations of which have been found to be modified in cutaneous scales from patients with LI (16,17). These lipids are important factors in the function of the water reservoir of the stratum corneum and are also involved in cell proliferation, differentiation and programmed cell death (18). The second LI gene could therefore be an enzyme or a co-factor involved in ceramide metabolism or in the covalent linkage of these molecules to the CE.


Figure 1. Pedigrees of the three families linked to chromosome 2. Affected individuals are represented by black symbols and unaffected family members by open symbols. Haplotypes from individual III-1 in family B were deduced by analysing his descendants and are therefore indicated in parentheses. Disease-bearing chromosomes are surrounded in each family.

558

Human Molecular Genetics, 1996, Vol. 5, No. 4 positive results were obtained by analysing large consanguinous families by homozygosity mapping and by restricting the focus of the study to clinically homogeneous subgroups. We are currently pursuing the same approach to localize other LI-causing gene(s). MATERIALS AND METHODS Pedigree assessment Each family member was examined by one of the authors (H.L. and C.B.B.) to confirm status, and all living participants received counselling and signed informed consent forms. Blood for DNA extraction was collected from 11 family members of pedigree A, from eight family members of pedigree B and from five family members of pedigree C. DNA was extracted from whole blood using standard procedures. No blood or tissue sample could be obtained from individual III-1 of pedigree B. Genotyping

In the present situation, cloning of the responsible gene remains difficult due to the large size of the candidate interval. The localization of expressed sequence tags in the interval should rapidly provide new candidate genes. However, refining the localization of the LI gene will be of great help in reducing the number of candidate transcripts to study. We looked for linkage disequilibrium since this method provides a powerful tool for fine gene mapping (19) and since the common geographical origin of these families might suggest a founder effect. We failed to demonstrate any common disease-sharing haplotype and no allele displayed linkage disequilibrium with the LI gene. Nevertheless, the small number of chromosomes studied and the presence of 2 cM intervals devoid of polymorphic markers does not permit firm exclusion of linkage disequilibrium in this region. An alternate hypothesis would be the presence of different mutations in the three families. Analysing new families linked to chromosome 2 could provide observations of new recombination event(s) and would also be of great help in refining the actual LI interval. We are therefore currently collecting other pedigrees and searching for more genetic markers in this region. The evidence for at least a third LI gene is now established. Genetic heterogeneity is not so surprising as LI is well known to present with a consistent degree of phenotypic variability (1–3). This last point, added to the low prevalence of the disease in the general population, renders the linkage studies difficult. In this work,

Linkage analysis Lodscores were calculated using the LINKAGE 5.1 package assuming an autosomal recessive inheritance, complete penetrance and a disease gene frequency of 0.33×10–5. For the random search, each polymorphism was assumed to have six alleles of equal frequency. Calculations with linked markers used the current allele frequencies determined from analysis of eight CEPH families. Two point calculations were done using MLINK (21). Linkage disequilibrium was tested using DISLAMB from the DISEQUILIBRIUM package (22). Taking into account the consanguinity in the three pedigrees, we compared three diseasesharing with six disease-free chromosomes. ACKNOWLEDGEMENTS We wish to acknowledge the essential contributions of Annette Bullot and Rolande Chaumeron for their help in the family collection, and Cécile Fizames for her support in all the computer analyses. We are also especially grateful to Susan Cure for her help in writing this manuscript. This study was supported by the Association Française contre les Myopathies. L.P. was supported by a fellowship from the Fondation pour la Recherche Médicale.


Figure 2. Abridged map of the chromosome 2q33–35 interval containing the disease-causing gene. Marker D2S153 (previously co-localized with D2S157, D2S154 and D2S322) was excluded from the study because of its ambiguous position on the GENETHON map. All genetic distances are as previously described (11) except for those between locus D2S137 and clustered loci D2S128, D2S334 and D2S143, where the recombination fraction was set to 0.01 because of a recombination event in patient IV-6 of family A. The centromeric limit can be defined either by marker D2S325 or D2S355, in patient IV-6 of family B. A possible crossing-over between these two markers would lead to a recombination fraction of 0.01 as shown on this map. The 7–8 cM interval containing the LI locus is shaded and is flanked by markers D2S137 and D2S325 or D2S355.

Polymerase chain reactions (PCR) were performed as previously described (20). Several amplification products from the same DNA sample, generated with different primer sets, were pooled and/or co-precipitated and co-migrated in a single lane of a 6% polyacrylamide gel. After transfer to HybondN+ (Amersham) membranes, hybridization was performed either with peroxidaselabelled PCR primers, or with a poly(AC) probe (20). TGM1 locus markers were those previously reported (7). For TGM2 and TGM3, we studied a set of microsatellite markers from the long arm of chromosome 20, spaced at 10 cM on average (assuming low likelihood of a double recombination event in this interval). As the loricrin–involucrin–SPRR cluster includes the profilaggrin gene (8), we analysed eight CEPH families with an intragenic polymorphic probe. We could therefore define the following order: D1S442–D1S498–profilaggrin–D1S305– D1S303 (total length 6 cM). No region of homozygosity was observed in the affected probands with these four microsatellite markers.

559 Human Acids Molecular Genetics, Vol.No. 5, No. Nucleic Research, 1994,1996, Vol. 22, 1 4 REFERENCES 1. Traupe, H. (1989) In The Ichthyosis—The Lamellar Ichthyosis. Springer-Verlag, Heidelberg, pp. 111–118. 2. Williams, M.L. and Elias, P.M. (1987) In Apler, J.L. (ed.), Dermatologic Clinics—Disorders of Cornification. Saunders, Philadelphia, PA pp. 155–178. 3. Williams, M.L. (1992) Pediatr. Dermatol., 9, 365–368. 4. Russell, L.J., DiGiovanna, J.J., Nashem, N., Compton, J.G. and Bale, S.J. (1994) Am. J. Hum. Genet., 55, 1146–1152. 5. Huber, M., Rettler, I., Bernasconi, K., Frenk, E., Lavrijsen, S.P.M., Ponec, M., Lautenschlager, S., Schorderet, D.F. and Hohl, D. (1995) Science, 267, 525–538. 6. Russell, L.J., DiGiovanna, J.J., Rogers, G.R., Steinert, P.M., Nashem, N., Compton, J.G. and Bale, S.J. (1995) Nature Genet., 9, 279–283. 7. Parmentier, L., Blanchet-Bardon, C., Nguyen, S., Prud′homme, J.F., Dubertret, L. and Weissenbach, J. (1995) Hum. Mol. Genet., 4, 1391–1395. 8. Reichnert, U., Michel, S. and Schmidt, R. (1993) In Darmon, M. and Blumenberg, M. (eds), Molecular Biology of the Skin—The Keratinocyte, Academic Press, Inc., San Diego, CA, pp. 107–150. 9. Huber, M., Rettler, I., Bernasconi, R., Wyss, M. and Hohl, D. (1995) J. Invest. Dermatol., 105, 653–654. 10. Wang, M., Kim, I.G., Steinert, P.M. and Mc Bride, O.W. (1994) Genomics, 23, 721–722.

559

11. Gentile, V., Davies, P.J.A. and Baldini, A. (1994) Genomics, 20, 295–297. 12. Gyapay, G., Morissette, J., Vignal, A., Dib, C., Fizames, C., Millasseau, P., Marc, S., Bernardi, G., Lathrop, M. and Weissenbach, J. (1994) Nature Genet., 7 special issue, 246–339. 13. Spurr, N.K., Kruse, T. and Naylor, S. (1994) Cytogenet. Cell Genet., 67, 215–244. 14. Mouse Genome Informatic, The Jackson Laboratory, Bar Harbor, Maine http://www.jax.org. 15. Steinert, P.M. and Marekov, L.N. (1995) J. Biol. Chem., 270, 17702–17711. 16. Paige, D.G., Morse-Fisher, N. and Harper, J.I. (1993) Br. J. Dermatol., 129, 380–383. 17. Paige, D.G., Morse-Fisher, N. and Harper, J.I. (1994) Br. J. Dermatol., 131, 23–27. 18. Hannun, Y.A. (1994) J. Biol. Chem., 269, 3125–3128. 19. Hästbacka, J., De la Chapelle, A., Kaitila, I., Sistonen, P., Weaver, A. and Lander, E. (1992) Nature Genet., 2, 204–211. 20. Vignal, A., Gyapay, G., Hazan, J., Nguyen, S., Dupraz, C., Cheron, N., Becuwe, N., Tranchant, M. and Weissenbach, J. (1993) In Adolph, K.W. (ed.), Gene and Chromosome Analysis–Part A. Academic Press, Inc., San Diego, CA,Vol.1, pp. 211–221. 21. Lathrop, G.M., Lalouel, J.M., Julier, C. and Ott, J. (1985) Am. J. Hum. Genet., 37, 482–498. 22. Terwilliger, J.D. (1995) Am. J. Hum. Genet., 56, 777–787. Downloaded from http://hmg.oxfordjournals.org/ by guest on June 2, 2013