Collagen type VI gene expression in the skin of trisomy 21 fetuses

OBSTETRICS& GYNECO

Volume 91

March 1998

Number

3

Collagen Type VI Gene Expression in the Skin of Trisomy 21 Fetuses C. S. VQN KAISENBERG, MD, B. BRAND-SABERI, PhD, B. CHRIST, MD, S. VALLIAN, PhD, F. FARZANEH, PhD, AND K. H. NICOLAIDES, MD Objective: To determine whether the mechanism for the retention of interstitial fluid in trisomy 21 fetuses presenting with nuchal translucency at lo-14 weeks’ gestation is an alteration in the composition of collagen type VI, which is normally a triple helix formed of three single chains, (~1, (~2, and (~3. The genes responsible for the al and (~2 chains, COLMl and COL6A2, are located on chromosome 21 and therefore may be overexpressed in trisomy 21, whereas COL6A3 is located in chromosome 2. Methods: Skin tissue was obtained after termination of pregnancy at 11-16 weeks’ gestation in five fetuses with trisomy 21 and five normal controls. Total RNA was extracted and the steady-state levels of COL6Al and COL6A3 mRNA expression of the gene transcripts were determined. Additionally, the distribution of collagen type VI in the skin of trisomy 21 and normal fetuses was analyzed using an immunohistochemical method. Results: The ratio of the normalized densitometric scores for the mRNA expression of COLMZ to COL6A3 in the skin of trisomy 21 fetuses was twice as high as in normal fetuses.

From the Harris Birthright Research Centre for Fetal Medicine, and the Department of Molecular Medicine, Kings College Hospital Medical School, London, United Kingdom; and the Institute of Anatomy, University of Freiburg, Freiburg, Germany. This study was supported by a grant from the Deutsche Forschungs Gemeinschaft (Number Ku 1136/l-11 and the Fetal Medicine Foundation.

VOL.Yl,NO.3,MARCH

1998

Immunohistochemistry demonstrated that in trisomy 21 fetuses collagen type VI formed a dense network extending from the epidermal basement membrane to the subcutis, whereas in normal fetuses dense staining was confined to the upper region of the dermis. Conclusion: The distribution for collagen type VI is different from normal in the skin of trisomy 21 fetuses, and there is overexpression of COL6AZ compared with COL6A3. (Obstet Gynecol 1998;91:319-23. 0 1998 by The American College of Obstetricians and Gynecologists.)

At lo-14 weeks‘ gestation about 80% of fetuses with chromosomal defects have abnormal accumulation of subcutaneous fluid in the nuchal region that is visualized by ultrasonography as nuchal translucency.1-3 A possible cause for this translucency is cardiac dysfunction due to the associated defects in the heart and great arteries.4 An alternative mechanism for increased nuchal translucency in trisomic fetuses is an alteration in the composition of the extracellular matrix of the skin leading to local retention of fluid. The collagen type VI molecule is a heterotrimer composed of three polypeptide chains ~1, cr2, and cu3.The genes encoding for al and 012, COL6Al and COL6A2, are located on chromosome 21, whereas COL6A3 is in chromosome 2. (Weil D,

OOZY-7844/98/$19.00 PI1 SOOZY-7844(97)00697-2

319

Mattei MG, Passage E, Van Cong N, Pribula-Conway D, Mann K, et al. Assignment of the three genes coding for the different chains of type VI collagen [COL6AZ, COL6A2, COL6A3] [abstract]. Cytogenet Cell Genet 1987;46:713.)5 It is possible that in trisomy 21 there is overexpression of COL6Al and COL6A2, with consequent alterations in the quantity and physiologic properties of collagen type VI. The aim of the present study is to investigate this hypothesis by examining the expression of COMA1 and COL6A3 genes and the morphologic distribution of collagen type VI in the skin of trisomy 21 and normal fetuses.

Materials and Methods During a 5-month period (April 1995 to September 1995), fetal skin tissue from the nuchal region, scalp, or thigh was obtained after termination of pregnancy in five pregnancies with fetal trisomy 21 and five normal controls. In the trisomic group, the mean maternal age was 39 years (range 36-43); in the control group the mean maternal age was 20 years (range 17-22). Gestational age was determined by ultrasound measurement of crown-rump length, and a regular heart beat was demonstrated in all pregnancies prior to termination. The gestational age range for the trisomic group was 12-14 weeks (mean 13), and for the controls it was 11-16 weeks (mean 13). Written informed consent was obtained from the patients. The study was approved by the hospital ethical committee, and tissue collection was made in accordance with the Polkinghorne guidelines on the research use of fetal material (Polkinghorne J. Review of the guidance on the research use of fetuses and fetal material. Her Majesty’s Stationery Office, London. 1989; Cm7627). In the trisomic group, prenatal diagnosis was made by chorion villous sampling, which was carried out at the request of the parents because an ultrasound scan at lo-14 weeks’ gestation demonstrated increased fetal nuchal translucency thickness3 The normal control tissues were obtained from pregnancies terminated for psychosocial reasons. Skin tissue was dissected, frozen immediately in RNAse-free polypropylene tubes in liquid nitrogen and stored at -70C. Total RNA extraction was performed, RNA was separated in a gel and transferred onto nitrocellulose membranes (Northern blot). These were stripped and reprobed as previously described.6 The probe for COLGAZ (Genbank accession number M20776) was the 2 kb EcoRl fragment excised from the ~18 plasmid obtained from American Type Culture Collection No. 61313 (Rockville, MD). This was amplified using Escherichia coli DH5, and DNA extraction was done using a large scale preparation of DNA (max-

320

von

Kaisenberg

et al

Collagen Type VI in Fetal Skin

iprep). The probe for COL6A3 (Genebank accession number M20778), was a 407 bp polymerase chain reaction amplified fragment using a human liver cDNA library as a template, cloned into a TA vector (Invitrogen, Carlsbad, CA) and confirmed by sequence analysis (Central Molecular Biological Services, King’s College School of Medicine and Dentistry, London, UK). Glyceraldehyde-3-phosphate dehydrogenase was used as a control for loading and transfer. The glyceraldehyde-3phosphate dehydrogenase probe was the Pstl 360 bp fragment of human glyceraldehyde-3-phosphate dehydrogenase (Genbank accession number M33197). MRC5 human fibroblast cells were used as a positive control for collagen type VI gene expression. MRSCVl cells (SV-40 transformed MRC5 cells) are known to have reduced expression of extracellular matrix factors, thus serving as a negative control for collagen gene expression. Autoradiography was then carried out, and the intensity of hybridization signals from Northern blots were quantified using a densitometer (GDS 2000; Mitsubishi, Tokyo, Japan). The densitometric scores of COL6Al and COL6A3 were normalized to the signal obtained for glyceraldehyde-3-phosphate dehydrogenase by dividing the densitometric values of the target gene by the glyceraldehyde-3-phosphate dehydrogenase values, thus correcting for any uneven loading of the RNA samples. Since the gene for glyceraldehyde-3phosphate dehydrogenase has been mapped to 12~13, no differences in glyceraldehyde-3-phosphate dehydrogenase mRNA gene expression between fetuses with trisomy 21 and normal controls would be expected.sr9 Cryosections of nuchal skin from five trisomy 21 and five normal fetuses were stained with polyclonal antibodies against collagen type VI obtained from Hey1 (Berlin, Germany) using affinity-purified goat-antirabbit IgG conjugated with fluorochrome Cy3 (Dianova, Hamburg, Germany) for detection, Frozen sections of 20 km were prepared from unfixed tissues that were obtained immediately after termination of pregnancy and embedded in Tissue Tek (OCT-compound; Leica, Bensheim, Germany). Sections were collected on chrome-alum-gelatin-coated slides. After being dried, they were blocked with 1% bovine serum albumin in 0.1 M potassium phosphate buffer (pH 7.6) for 10 minutes and incubated with primary antibodies for 60 minutes at room temperature. After being rinsed with phosphate buffer, the sections were incubated with the second antibodies also for 60 minutes at room temperature. Controls consisted of cryostat sections from fetuses with trisomy 21 and normals in which the primary antibody was replaced by preimmune rabbit serum before the application of the goat-anti-rabbit IgG to assess the degree of nonspecific staining. The sections

Obstetrics b Gynecology

ColG a3

SSkb

CAPDH

Figure 1. Northern blot of RNA extracted from skin tissue of five trisomy 21 fetuses (columns l-5), five normal controls (columns 6-lo), and MRC5 and MRSCVl human fibroblast cells (columns 11 and 12, respectively). Transcripts of approximately 4.2 kb for COLGAZ, 8.5 kb for COL6A3, and 1.6 kb for glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were detected. The skin was from the nuchal region in columns 1, 2, 3, 6, and 7, from the thigh in columns 4, 5, and 8, and from the scalp in columns 9 and 10.

were examined by epifluorescence microscopy and photographed using Ektachrome 400 color film (Kodak). Mann-Whitney test was used to compare normalized densitometric scores for COL6A3 between normal and trisomy 21 and to compare COL6AZ and COL6A3 in the normal and trisomic group, respectively.

Results In the Northern blot of total RNA extracted from the skin tissue of trisomy 21 and normal fetuses and from MRC5 and MRSCVl cells, single transcripts of approximately 4.2 kb for COL6A1, 8.5 kb for COL6A3, and 1.6 kb for glyceraldehyde-3-phosphate dehydrogenase were detected (Figure 11, which are the same size as previously reported.5 These findings were reproduced in several blots. In the skin of trisomy 21 fetuses compared with that of the normal controls (Table l), the normalized densitometric score for COL6AI was significantly higher (n = 5, z = -2.6, P < .Ol). The score for

Figure 2. Immunofluorescence detection of collagen type VI in a section through nuchal skin of a fetus with trisomy 21 (a) and in a normal control fetus (b), both at 14 weeks’ gestation. In trisomy 21, there is intense staining of the extracellular matrix extending down to the dermis-subcutis junction, whereas in the normal fetus intense staining is restricted to the upper dermis immediately underlying the basement membrane. ep = epidermis; d = dermis; bm = basement membrane. Bar = 0.1 mm.

Trisomy

21

Nuchal

14

2.88

1.04

2.76

Trisomy Trisomy Trisomy Normal

21 21 21

Nuchal Thigh Thigh Nuchal

14 13 12 11

2.28 1.73 1.55 1.45

0.92 0.37 0.91 1.02

2.47 4.67 1.70 1.42

Normal

Nuchal

14

0.88

0.74

1.18

COL6A3 in trisomies was not significantly different from normal (n = 5, 2 = 0.94, P = .35). Morphologic studies have shown that the texture of the connective tissue in the skin of fetuses with trisomy 21 differs from that in normal controls. In trisomy 21, the collagen lattice appears to be less regular, and immunohistochemical studies demonstrate that collagen type VI formed a dense network extending from the epidermal basement membrane to the subcutis (Figure 2a). In normal controls, intense staining and densely packed fibres were confined to the upper region of the dermis, immediately underlying the epidermal basement membrane (Figure 2b).

Normal Normal Normal

Thigh Scalp Scalp

14 16 14

0.65 1.51 0.91

0.49 1.36 1.37

1.32 1.11 0.66

Discussion

0.45 0.08

1.38 0.121

0.72 0.66

Table

1. Normalized

Densitometric Expression Trisomy 21 Fetuses

COL6A3 Gene From

Scores for in Skin and and Normal Normalized

Trisomy

21

Skin site Nuchal

Gestation (wk) 13

MRC5 MRSCVl

VOL.

91, NO.

3, MARCH

1998

COL6Al 2.15

COL6Al

and Their Ratios Controls

densitometric score COL6A3 0.69

COLGAl/ COL6A3 3.11

The data demonstrate that COL6AZ and COL6A3 genes are expressed in the skin from the nuchal and other

von

Kaisenberg

et al

Collagen

Type

VI in Fetal

Skin

321

regions of fetuses at 11-16 weeks’ gestation. Furthermore, the results suggest that in trisomy 21 the COL6AI gene is overexpressed by comparison with the COL6A3 gene, which is located on chromosome 2. Although there are no reports on the expression of the various collagen type VI genes in adults with Down syndrome, Langdon Down, who first described the phenotypic expression of the syndrome, noted that the skin of affected individuals was lacking in elasticity (Down JLH. Observations on an ethnic classification of idiots. Clin Lecture Rep London Hosp 1866;3:259). Previous studies in trisomy 21 have documented overexpression of several other genes found in chromosome 21 with consequent effects on the phenotype. For example, in trisomy 21 overexpression of the amyloid precursor protein gene results in deposition of amyloid in the brain and early-onset of Alzheimer disease.*’ Additionally, transgenic mice that carry the Ets2 gene, a proto-oncogene and transcription factor that is overexpressed in human trisomy 21, develop cranial and cervical skeletal abnormalities that are similar to those in Down syndrome? Collagens are a large number of heterotrimeric or homotrimeric triple-helical proteins that are present in the extracellular matrix. Collagen type VI has a low sequence homology to fiber-forming and basementmembrane collagens, and its triple helix is specifically adjusted for the formation of microfibrils in tissues? Although normally collagen type VI is a heterotrimer composed of the three polypeptide chains orl, a2, and 01312,13both dimeric and tetrameric forms have been described.i4,15 Additionally, when there is abundance in the expression of one of the collagen type VI genes, homotrimeric collagen type VI molecules may form, but the biophysical properties of the various types of collagen VI likely are differentI’ On the basis of our findings it is possible that in trisomy 21 the composition and consequently the properties of collagen type VI are altered leading to the accumulation of subcutaneous edema. The immunohistochemical studies found that in trisomy 21, collagen type VI formed a dense network reaching from the epidermal basement membrane to the subcutis, whereas in the skin of normal controls dense staining was restricted to the upper region of the epidermis. These findings are compatible with previous reports that in trisomy 21 fetuses collagen type VI was more irregularly arranged than in normal fetuses.16’17 Collagen type VI is known to bind to hyaluronan, a hygroscopic glycosaminoglycan of the extracellular matrix.l’ Hyaluronan is also abundant in the skin of trisomy 21 fetuses, which is not the case in the skin of normal fetuses.16,r7 This could be explained by overexpression of super-oxide-dismutase of which the gene

322

von Kaisenberg

et al

Collagen Type VI in Fetal Skin

lies on chromosome 21,19,” which has been shown to stabilize the hyaluronic acid molecule against free radicals, therefore reducing the degradation of hyaluronan21 Collagen type VI bound hyaluronan may play a crucial role in the process of water binding in the skin leading to increased nuchal translucency. An alternative mechanism for a link between possible alterations in the composition of collagen type VI and increased translucency in trisomy 21 fetuses is impairment in cardiac function or structure. A previous study has demonstrated COL6Al and COL6A2 gene expression in the heart of human fetuses at lo-14 weeks’ gestation.22 Therefore, in the heart of trisomy 21 fetuses, the expression of these genes may be increased in relation to the COL6A3 gene, as observed in the skin. Such a potential imbalance in the expression of the various genes involved in the production of heterotrimerit collagen type VI could result in impairment in contractility of the myocardium and could interfere with the normal development of the heart. Collagen type VI is a powerful substrate for cell adhesion,23,24 and it has been suggested that the atrioventricularseptal defects, commonly observed in trisomy 21, are the consequence of failure of endocardial cushion fusion due to increased adhesiveness of fibroblasts.25,26 A similar imbalance in the expression of genes encoding for the component proteins of collagens also may be present in trisomy 13 and account for the increased nuchal translucency observed in this condition.27 Interestingly, the genes COL4AI and COL4A2 for the cwland cu2 chains of collagen type IV are located on chromosome 13,28,29whereas the COL4A4 gene is located on chromosome 2.27

References 1. Pandya PI’, Brizot ML, Kuhn I’, Snijders RJM, Nicolaides KH. First-trimester fetal nuchal translucency thickness and risk for trisomies. Obstet Gynecol 1994;84:420-3. 2. Nicolaides KH, Brizot ML, Snijders RJM. Fetal nuchal translucency: Ultrasound screening for fetal trisomy in the first trimester of pregnancy. Br J Obstet Gynaecol 1994;101:782-6. 3. Snijders RJM, Johnson S, Sebire NJ, Noble FL, Nicolaides KH. First-trimester ultrasound screening for chromosomal defects. Ultrasound Obstet Gynecol 1996;7:216-26. 4. Hyett JA, Moscoso G, Nicolaides KH. Abnormalities of the heart and great arteries in the first trimester chromosomally abnormal fetuses. Am J Med Genet 1997;69:207-16. 5. Weil D, Mattei M-G, Passage E, Van Cong N, l’ribula-Conway C, Mann K, et al. Cloning and chromosomal localization of human genes encoding the three chains of type VI collagen. Am J Hum Genet 1988;42:435-45. 6. Brizot ML, Jauniaux E, Mckie AT, Farzaneh F, Nicolaides KH. Placental expression of a and b subunits of human chorionic gonadotropin in early pregnancies with Down’s syndrome. Mol Human Reprod 1995;10:2506-9. 7. Sambrook J, Fritsch EF, Mania& T. Large scale preparation of

Obstetrics & Gynecology

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

DNA. In: Molecular cloning, a laboratory manual. Cold Spring Harbour, New York: Cold Spring Harbour Laboratory, 1989;1.33. Tokunaga K, Nakamura Y, Sakata K, Fujimori K, Ohkubo M, Sawada K, et al. Enhanced expression of a glyceraldehyde-3phosphate dehydrogenase gene in human lung cancers. Cancer Res 1987;47:5616-9. Ercolani L, Florence B, Denaro M, Alexander M. Isolation and complete sequence of a functional human glyceraldehyde3phosphate dehydrogenase gene. J Biol Chem 1988;263:15335-41. Golaz J, Charnay Y, Vallet I’, Bouras C. Alzheimer’s disease and Down’s syndrome. Some recent etiopathogenic data. Encephale 1991;17:29-31. Sumarsono SH, Wilson TJ, Tymms MJ, Venter DJ, Corrick CM, Kola R, et al. Down’s syndrome-like skeletal abnormalities in Ets2 transgenic mice. Nature 1996;379:534-7. Chu M-L, Conway D, Pan T-C, Baldwin C, Mann K, Deutzmann R, et al. Amino acid sequence of the triple-helical domain of human collagen type VI. J Biol Chem 1988;263:18601-6. Timpl R, Engel J, Collagen type VI. In: Mayne R, Burgeson RE, eds. Structure and function of collagen types: Biological and extracellular matrix. New York: Academic Press, 1987:105-43. Engel J, Furthmayr H, Odermatt E, von der Mark H, Aumailley M, Fleischmajer R, et al. Structure and macromolecular organization of type VI collagen. Ann NY Acad Sci 1985;460:25-37. Furthmayr H, Wiedemann H, Timpl R, Odermatt E, Engel J. Electron-microscopical approach to a structural model of intima collagen. Biochem J 1983;211:303-11. Brand-Saberi B, Floe1 H, Christ B, Schulte-Vallentin M, Schindler H. Alterations of the fetal extracellular matrix in the nuchal oedema of Down’s syndrome. Ann Anat 1994;176:539-47. Brand-Saberi B, Epperlein HH, Romanos GE, Christ B. Distribution of extracellular matrix components in nuchal skin from fetuses carrying trisomy 18 and trisomy 21. Cell Tissue Res 1994;277:465-

75. 18. McDevitt CA, Marcelino J, Tucker L. Interaction of intact type VI collagen with hyaluronan. FEBS Lett 1991;294:167-70. 19. Cox DR, Epstein LB, Epstein CJ. Genes coding for sensitivity to interferon (IfRec) and soluble superoxide dismutase (SOD-l) are linked in mouse and man and map to mouse chromosome 16. Proc Nat1 Acad Sci USA 1980;77:2168-72. 20. Aliakbar S, Brown PR, Bidwell D, Nicolaides KH. Human erythrocyte superoxide dismutase in adults, neonates and normal, hypoxaemic, anaemic and chromosomally abnormal fetuses. Clin Biochem 1993;26:109-15. 21. Prehm I’. Release of hyaluronate from eukaryotic cells. Biochem J 1990;267:185-9. 22. Duff K, Williamson R, Richards S-J. Expression of genes encoding

VOL.

91, NO.

3, MARCH

1998

23. 24. 25.

26. 27.

two chains of the collagen type VI molecule during human fetal heart development. Int J Cardiol 1990;27:128-9. Carter WG. The cooperative role of the transformation sensitive glycoproteins GP140 and fibronectin in cell attachment and spreading. J Biol Chem 1982;257:3249-57. Aumailley M, Mann K, von der Mark H, Timpl R. Cell attachment properties of collagen type VI and arg-gly-asp dependent binding to its ct2(VI) and c~3(V1) chains. Exp Cell Res 1989;181:463-74. Kurnit DM, Aldridge JF, Matsouka R, Matthyse S. Increased adhesiveness of trisomy 21 cells and atrioventricular canal malformations in Down syndrome: A stochastic model. Am J Med Genet 1985;30:385-99. Kurnit DM, Layton WM, Matthyse S. Genetics, chance and morphogenesis. In: Epstein CJ, ed. Morphogenesis of Down syndrome. New York: Wiley Liss, 1987:19-41. Kamagata Y, Mattei MG, Ninomiya Y. Isolation and sequencing of cDNAs and genomic DNAs encoding the alpha4 chain of basement membrane collagen type IV and assignment of the gene to the distal long arm of human chromosome 2. J Biol Chem 1992;

267123753-8. 28. Bowcock AM,

29.

Hebert JM, Christian0 AM, Wijsman E, CavalliSforza LL, Boyd CD. The pro alpha 1 (IV) collagen gene is linked to the D13S3 locus at the distal end of human chromosome 13. Cytogenet Cell Genet 1987;45:234-6. Solomon E, Hiorns LR, Spurr N. Chromosomal assignments of the genes coding for human types II, III and IV collagen: A dispersed gene family. Proc Nat1 Acad Sci USA 1985;82:3330-4.

Address reprint requests to: K. H. Nicolaides, MD Harris Birtkvigkt Research Center for Fetal Medicirze Kings College Hospital Medical Center Denmark Hill London SE5 SRX United Kingdom

Received March 24, 1997. Received in revised form August 29, 1997 Accepted November 24, 1997.

Copyright 0 1998 by The American Gynecologists. Published by Elsevier

von Kaisenberg

et al

College of Obstetricians Science Inc.

Collagen Type VI in Fetal Skin

and

323