European Journal of Radiology, 14 (1992)
0 1992 Elsevier Science Publishers
EURRAD
l-10 B.V. All rights reserved.
0720-048X/92/$05.00
00241
State of the art
Collagen, genes and the skeletal dysplasias on the edge of a new era: a review and update Ralph S. Lachman avb, George E. Tiller b, John M. Graham, aDepartment of Radiology, UCLA School of Medicine, Harbor/UCLA
School of Medicine, International Skeletal Dysplasia Registry, Cedars-Sinai
(Accepted
Jr.b and David L. Rimoinb
Medical Center, Torrance, CA and bDivision of Medical Genetics, UCLA
15 October
Medical Center, Los Angeles, Caljfomia USA
1991)
Key words: Skeletal dysplasia, osteogenesis imperfecta; Skeletal dysplasia, achondrogenesis; Skeletal dysplasia, hypochondrogenesis; dysplasia, spondyloepiphyseal dysplasia; Collagen abnormality
Skeletal
Abstract This article reviews the newly described biochemical (type I and II collagen) abnormalities and specific gene defects in the skeletal dysplasias. The model of the collagen molecule is described and how collagen is processed from procollagen, where and how abnormalities occur, and the types of abnormalities produced (quantitative and qualitative). The only known type I collagen defects producing skeletal dysplasias osteogenesis imperfecta, as well as the ‘family’ of established type II collagen disorders - achondrogenesis type II, hypochondrogenesis and spondyloepiphyseal dysplasia congenita are discussed. Finally, using case presentations, the practical approach to these disorders is shqwn. The importance ofthese investigations and the subsequent reevaluation ofthe clinical and radiological findings of specifically delineated skeletal dysplasias are discussed.
Introduction Recently, there has been an explosion in the field of collagen chemistry in relationship to the skeletal dys-* plasias and connective tissue disorders. In the past, we have had to rely on the radiographic and clinical delineation of these disorders alone, but in the face of genetic heterogeneity and intrafamilial phenotypic variability, it was very difficult to know if one was dealing with a single disorder or a group of disorders with the same or different patterns of inheritance. The first step towards a solution to this problem was the examination of chondroosseous morphology (iliac crest, anterior rib or growth plate biopsy/autopsy) to establish criteria for specific diagnoses [ 11. Although many disorders showed rather specific changes, other disorders revealed similar patterns despite obvious radiographic and clinical differences.
Address for reprints: Ralph S. Lachman, M.D., Department of Radiology, Box 27, Harbor/UCLA Medical Center, 1000 W. Carson Street, Torrance, CA 90509, USA.
A concept was proposed by Jurgen Spranger that ‘families’ of skeletal dysplasias, i.e., disorders with similar radiographic and clinical findings could be due to similar genetic defects [2]. We then proposed that skeletal dysplasias with biochemical defects of type II collagen will display abnormalities in tissues where type II collagen is found (i.e., hyaline cartilage, nucleus pulposus and the vitreous of the eye). This ‘family of disorders’ concept dictates which groups of skeletal dysplasias with similar features might share a common underlying defect. For example, those skeletal dysplasias with not only epiphyseal and vertebral abnormalities, but also ocular changes, suggest an underlying defect in type II collagen, the collagen found in these particular areas. A discussion of the known collagen disorders from a skeletal dysplasia point of view must include osteogenesis inperfecta (01) (type I collagen disorders); and achondrogenesis type II hypochondrogenesis (Fig. 1 and 2) and certain spondylepiphyseal dysplasia (SED) cases (known type II collagen disorders). These are of special interest to us as radiologists because of their radiographic importance for diagnosis.
a
b
Fig. 1. Achondrogenesis type II (end of achondrogenesis type II/ hypochondrogenesis - SED congenita spectrum); Radiographs showing good skull ossification, almost no vertebral body ossilication, extremely short ribs, hypoplastic ilia, absent ossification of pubis, ischia, tarsal centers, very short tubular bones with concave ends.
The collagens are the predominant proteins of the extracellular matrix of most tissues. To date, there are in excess of 13 different known human collagens [ 3 J. Type I collagen is the major collagen of skin, bone, tendon and vessel walls. Type II collagen is the pre-
a
dominant collagen of cartilage, nucleus pulposus and vitreous. The distribution of these collagens suggest the tissue sites that should be abnormal in their respective diseases [4]. The long triple helix of the collagen molecule is flanked on each end by a globular amino terminal and carboxy terminal which function in both the assembly of procollagen and its secretion. These peptides are then removed by proteolytic cleavage in the extracellular compartment (Fig. 3). The individual molecules must be normally formed so they can assemble into an ordered libril. A mutation which interferes with : (1) expression of the collagen gene; (2) formation of the triple helix, or (3) with procollagen secretion can have dramatic effects on the structure and function of the fibrils created and potentially result in a severe clinical disorder. The evolutionary relationship of the collagen gene and its protein products provides a unique opportunity to compare mechanisms of mutation in a family of genes and observe the effects of such mutations at the biochemical, histological and phenotypic levels. This review is limited to the skeletal dysplasias with known collagen defects. The type I collagenopathies consists of the various types of osteogenesis imperfecta (01). A modified Sillence classification of osteogenesis imperfecta can be found in the Table. It is based on clinical and radiographic features. The data to date suggest that all types of 01 have primarily autosomal dominant inheritance, but autosomal recessive inheritance is likely in some cases of 01 type II group C and 01 type III. In most cases of type II 01, rare multiple affected offspring of unaffected parents can be explained by parental germline mosaicism for the mutation
b
Fig. 2. (a, b and c) Hypochondrogenesis (middle of spectrum). Skull, trunk and extremities exhibiting more mature ossification than achondrogenesis type II with large defect of posterior skull base; flat vertebral bodies with cervical and lower lumbar hypoplasia, short ribs, ischial but no pubic ossification, and shortened long tubular bones with rounded ends.
Amlno terminal.
AMINO ACID SEQUNCES ( gly - pro- hyp - gly . )
( propeptlde)
.Corboxy terminal (propeptide)
TABLE 1 Classification
of osteogenesis
Type 1 propeptidose
I
u
i
v
COLLAGEN-
Fig. 3. The processing
of procollagen
[ 191
Clinical and radiologic features Mild phenotype. Normal or close to normal stature. Hearing loss in about 50%.
PROCOLLAGEN propepildase
imperfecta
Group IA
Blue sclerae. Normal teeth.
Group IB
Dentinogenesis
Imperfecta
(DI).
w
to collagen.
(some cell lines have the mutation and some do not), resulting in a 5-77; recurrence risk. The 01s are actually a group of disorders in which at least 95% of the cases are due to mutation affecting the synthesis and/or structure of type I collagen [5-l 11. Two types of defects have been found, quantitative defects characterized by decreased production of normal type I collagen and qualitative defects in which there are populations of both abnormal type I molecules as well as structurally normal molecules. The quantitative defects appear to cause the milder forms of the disease and are found primarily in 01 type I. On the other hand, qualitative defects have been discovered in 01 types II, III and IV, and can be found in 85-90% of affected individuals tested thus far. Most of these defects are point mutations that result in substitutions for glycine in the triple helix of one of the type I collagen chains (Fig. 3). This disruption of the normal amino acid sequence slows the assembly of the molecule creating a less stable structure. Consequently, decreased secretion and/or intracellular degradation of procollagen can occur. Quantitative defects which limit the amount of type I collagen produced result in decreased deposition of normal collagen in the extracellular matrix and a decrease in mineralization while qualitative defects, in which abnormal collagen is incorporated into the matrix, severely disrupt mineralization. The site and nature of the amino acid substitution within the triple helical region can often (but not consistently) predict the biochemical and clinical phenotype [ 12-141. Case studies Patient A was referred in the first year of life in order to substantiate the diagnosis of osteogenesis imperfecta, and for the parents to receive genetic counseling and discuss prognostication of her condition. Her vaginal delivery at term was complicated by several rib fractures
Type II
Clinical and radiologic features Most severe phenotype. Blue sclerae. Hydrops 15%. Almost no calvarial ossification. Short wide femurs. Beaded ribs. Multiple rib fractures. Group IIA
Accordian-like Beaded ribs.
Group IIB
Broad crumpled long bones. Rib fractures - no beading.
Group IIC
Thin short fractured long bones. Thin ribs - fractured or mildly beaded.
femurs.
Type III
Clinical and radiologic features Blue sclera at birth changes to normal later, dentinogenesis imperfecta is common Progressive deformed skull; wormian bones or severely deossitied. Slightly short thin angulated long bones. ‘Popcorn’ calcifications in metaphyses. ‘Codfish’ vertebrae.
Type IV
Clinical and radiologic features Similar to type I. Normal sclera. Less deafness. Shorter stature than type I. Group IVA Normal teeth. Group IVB DI.
both clinically and radiographically. Her height, weight, and head circumference were all at the 50th percentile for age. Parents gave no family history of 01 and were both clinically normal. There was no consanguinity noted. Physical findings included blue sclerae and frontal bossing. The radiographs revealed thin ribs, several rib fractures and a clavicular fracture in the newborn
4
period, generalized osteoporosis and over modelled tubular long bones with thin cortices (Figs. 4-5). Radiographically and clinically, this patient appeared most likely to have 01 type IA. Collagen analysis from cultured skin fibroblasts of the patient and both parents revealed normal migration on electrophoresis. However, the patient’s cells synthesized a reduced quantity of type I collagen, whereas her parents’ synthesized collagen in normal amounts. This finding of a quantitative defect suggested 01 type I as the result of a new dominant mutation and only a small risk of recurrence in this family, confirming the clinical impression prior to the collagen analysis. The family was also advised about the 50 y0 chance of deafness, as well as other potential orthopedic management problems. Patient B was a newborn male delivered by caesarean section at term because of polyhydramnios and prolonged labor. Both parents were normal and not consanguineous. The diagnosis of 01 was suggested on clinical grounds and conlirmed by the radiographs. The lower extremities were bowed and distorted. The sclerae were blue and a soft skull was present. The newborn radiographs revealed generalized osteoporosis, shortening of the long tubular bones (especially short, widened femurs and tibias with very thin cortices, angulation and fractures), ribs containing many acute and healing rib fractures, and very poor calvarial ossification
4
(Fig. 6). Radiographically, this patient appeared to represent a case of 01 type IIB on the milder end of the spectrum of type II and not exhibiting the classic ‘accordion’ femurs and beaded ribs of the more common type IIA. Prior to the patient’s death in the second week of life, a skin biopsy was obtained. Collagen biochemical analysis revealed both normal and abnormal populations of type I procollagen molecules. This qualitative defect appeared to be an autosomal dominant new mutation with a recurrence risk of 5-7%. Although the recurrence risk is small, future pregnancies in this case can be monitored either by fetal ultrasound or by chorionic villus biopsy (CM) for collagen analysis since type I collagen is produced by villus tissue. Serial fetal ultrasound with no known morbidity is the first test of choice because of: (a) the time required to culture and label cells, and to perform the biochemical analysis and (b) the small but significant risk of pregnancy loss due to the CVS procedure [ 151. Patient C was the full term male product of unrelated parents delivered by caesarean section for breech presentation and a suspected skeletal dysplasia. Prenatal ultrasound was performed at 15 weeks gestation (with serial repetition), delineating short and thickened bones in the lower extremities (Fig. 7). At birth, the patient had blue gray sclerae, inguinal hernias and short, curved, long bones (especially of the lower extremities). Newborn radiographs revealed generalized osteoporo-
6
5
Fig. 4. Patient A, 01 type IA, newborn; chest radiograph Fig. 5. Patient A, 01 type IA, 6 months old; Lower extremities
revealing thin ribs, clavicular fracture and generalized revealing overtubulated
osteoporosis.
long bones with thin cortices and osteoporosis.
Fig. 6. Patient B, 01 type IIB, f&term newborn; Full body radiograph showing generalized osteoporosis, deficient calvariai ossification, especially short widened and bent femurs and tibias with thin cortices and fractures.
7
9
8
Fig. 7. Patient C, 01 type IIC-III,
fetal ultrasound
at 15 weeks gestation
showing a short widened femur.
Fig. 8. Patient C, 01 type IIC-III, full-term newborn; trunk and extremity radiograph revealing generalized osteoporosis; shortened angulated long bones with thin cortices (especially significant femoral involvement), and slightly thin ribs. Radiographic features closest to 01 type IIC. Fig. 9. Patient C, early infancy; CT revealing signiticant generalized
sis, a brachycephalic osteoporotic skull with high orbital roofs, dentinogenesis imperfecta, shortened angulated long bones with thin cortices, and marked shortening and widening of the femurs (Fig. 8). Radiologically and clinically, he appeared to represent 01 type IIC as an infant but with the passage of time resembled those in the severe end of the type III spectrum. The patient developed hydrocephalus without clinical signs of increased intracranial pressure (Fig. 9). During the ftrst year of life, frequent apparently spontaneous fractures occurred, with profuse callus formation. Collagen analysis from cultured skin f’ibroblasts showed a qualitative defect with both normal and abnormal populations of type I collagen. Cyanogen bromide peptide analysis revealed overmodification of the abnormal alpha chains confined to the amino end of the molecule. Collagen studies of the parents were normal. This condition in this patient appeared to represent a new autosomal dominant disorder with a low likelihood of a recurrence due to parental germline mosaicism (5-7x). Since early changes were present by fetal ultrasound in the proband, one can recommend serial fetal ultrasonography and/or CVS to monitor future pregnancies. The type II collagen defects that affect certain of the skeletal dysplasias have very similar abnormalities to the collagen I defects of osteogenesis imperfecta. They also exhibit quantitative and qualitative abnormalities. Patient D is now a 24-year-old male with a diagnosis
ventricular
dilatation.
of SED made at about a year of age and followed since age 6 years by our clinic. He was the full-term product of a 34-year-old gravida 3 following an uncomplicated pregnancy, labor and delivery. At birth, clinical examination was normal except for short stature. There was no history of consanguinity in the family. He was first evaluated at about 6-8 months because of inability to sit and crawl and radiographs at that time were consistent with a diagnosis of SED congenita. At 5 years of age, significant myopia developed. Cervical spine fusion for excess atlanto-axial mobility was performed at age 9 years as well as some other orthopedic procedures. On physical examination, he has remained disproportionately short (at age 15 years he was at the 50th percentile for a 9 year old), with a normal skull. He exhibits widening at all the major joints and diminished motion at the hips and elbows. Radiographic examinations at ages 6, 7 and 16 years revealed very significant generalized epiphyseal ossification delay, as well as significant thoracic platyspondyly with irregularity suggesting the radiographic diagnosis of some form of spondyloepiphyseal dysplasia (Figs. 10-14). The electrophoretic analysis of type II collagen extracted from an iliac crest biopsy revealed both normal and abnormally migrating a-chains. Cyanogen bromide peptide analysis showed overmodification of all major peptides. DNA sequence analysis of the type II collagen gene revealed a 45 base-pair dupli-
b
b
a
Fig. 10. (a) Patient D spondyloepiphyseal dysplasia (SED) at age 7 years; pelvis and hips revealing small capital femoral epiphyseal centers with flocculent calcification, vertical ischia and down slanted pubic bones. (b) Patient D, at 16years of age; The femoral heads are completely ossified and the epiphyseal plate is fused.
12b
12a
Fig. 11. Patient D, at age 7 years; knees revealing marked epiphyseal ossification epiphyseal bone density.
delay, apparent
epiphyseal
irregularity
and decreased
Fig. 12. (a and b) Patient D, 6 years old; thoracic and lumbar spine showing lack of interpediculate flare of lower lumbar vertebrae, platyspondyly of thoracic vertebrae with anterior pointing and rounding; but normal height of lumbar vertebrae. Fig. 13. Patient D, at 16 years of age; Showing progression
of the platyspondyly spine.
cation in exon 48 of the triple helical coding region. The mutation arose in the maternally derived allele, most likely by an unequal crossover during meiosis. This resulted in the addition of 15 amino acids to the triple
and end plate irregularity
and sclerosis of the thoracic
helical portion of the a-chain, and likely disrupted the rate of assembly of molecules incorporating one or two such chains [ 161. The patient was advised of a 50% recurrence risk in his offspring and that the molecular
a
b
a
b
Fig. 14. (a and b) Patient D, flexion films at 7 and 16 years of age delineating excess mobility of Cl, C2 before posterior fusion and the resultant immobility following surgical fusion.
definition of his disease should allow for prenatal diagnosis by CVS at approximately 10 weeks gestation. Patient E was the product of a normal spontaneous delivery at 33 weeks gestation, with a birth weight of 1710 g, and a length of 13.5 inches. The pregnancy was complicated by the discovery of short limbs and a small chest on prenatal ultrasound at 19 weeks. Physical examination at birth showed disproportionate extremities and an abnormal bell-shaped thorax. The baby required intubation in the delivery room, and chronic respirator support. The prenatal ultrasound findings suggested the possibility of Ellis Van Crefeld, asphyxiating thoracic dysplasia or one of the short rib polydactyly syndromes as the most likely diagnosis. Postnatally the diagnosis of Ellis Van Crefeld (chondroectodermal dysplasia) was made clinically and radiographically by the referring physicians primarily because of a finding of unilateral polydactyly. Upon referral to the International Skeletal Dysplasia Registry, we felt that the radiographic findings more closely fit the achondrogenesis-hypochondrogenesis-SED spectrum. The early radiographs were those of SED congenita, but repeat examination at 3 months of age showed changes closer to the hypochondrogenesis (a more severe, usually lethal) portion of that spectrum (Figs. 15-2 1). The infant was removed from respiratory support after appropriate discussion with the parents, and the baby expired. Chondroosseous morphology showed hypervascular cartilage, friable matrix, foamy areas around lacunae,
Fig. 15. Patient E, hypochondrogenesis - SED congenita, 33 week gestation newborn: spine and pelvis revealing rounded platyspondyly; absent pubic bones, hypoplastic vertical ischia, and rounded iliac wings.
and extension of chondrocytes into trabecular areas. These findings are also those of mild hypochondrogenesis or severe SED. Electrophoretic analysis of type II collagen extracted from cartilage revealed both normal and abnormally migrating a-chains. Cyanogen bromide peptide analysis showed overmodification of most peptides. Amino acid sequencing demonstrated a substitution of glutamic acid for glycine at position 853 of the triple helical region. DNA sequencing confirmed the single nucleotide substitution that would account for this amino acid substitution. This disrupted the development of the normal triple helix in this region, and likely interfered with the rate of assembly, thus resulting in overmodification of the procollagen molecule [ 17-181. The patient was heterozygous for a gene defect and restriction fragment length polymorphism(RFLP), which enabled us to determine that the mutation arose in the paternally derived allele of that collagen (COL2AI) gene. Since the molecular studies showed that this was a dominant trait, the family was advised that their recurrence risk was low. Analysis of DNA from the father’s sperm showed no evidence for gorminal mosaicism, thus lowering the parent’s recurrence risk for another child with this form of skeletal dysplasia to that of the general population. In conclusion, it is clear that we have come a long way from the radiographic definition (as the solitary
8 16
17
Fig. 16. Patient E, newborn;
showing thin ribs, elongated clavicles, deficient ossification
Fig. 17. Patient E, 33 week gestation newborn; demonstrating
no talus or calcaneal ossification; margins.
tool) for the skeletal dysplasias through an era of chondromorphologic (both light and ultrastructural) evaluation and into this present era of biochemical and gene
of the cervical spine and shortened
humeri.
and shortened tibiae with smooth metaphyseal
delineation. All of this information is very important in order to not only determine a distinct entity, but to achieve the best possible genetic counseling and in-
18b
16a
Fig. 18. (a and b) Patient E, at 3 months of age; skull revealing elevated orbital roofs, midface hypoplasia occipital region. Fig. 19. Patient E, at 3 months;
and an ossification
Later spine film showing some vertebral growth but still characteristic
defect in the
rounded platyspondyly.
20b
Fig. 20. Patient E, at 3 months;
demonstrating
2la
21b
short anteriorly flared ribs and small thoracic volume.
Fig. 21. (a and b) Patient E, at 3 months: hips and lower extremities and wrist revealing the development sclerosis and still lack of ankle ossification.
trauterine diagnosis for the family. Now that we can determine the biochemical and gene defect in certain of these disorders, the radiologist can carefully examine biochemically proven cases with these disorders to define the variability of expression of each mutation. In this review article, we have tried to clearly delineate these recent biochemical and genetic breakthroughs in the chondrodysplasias, as well as present specific cases which show the practical value of these investigations. Acknowledgements Special thanks to Elise Spears and Savona Huching for technical and artistic assistance for this manuscript. This study was supported in part by a USPHSNIH Program Project Grant (HD22657).
6
7
8
9
10 11
References Sillence DO, Horten WA, Rimoin DL. Morphologic studies in the skeletal dysplasias. A Review. Am J Path01 1979; 96: 81 l-870. Spranger J. Pattern recognition in bone dysplasias in: Endocrine Genetics and Genetics of Growth. C.J. Papadatos and C.S. Bartocas, eds. New York: Liss, 1985; 315-342. Kuivaniemi H, Tromp G, Prockop DJ: Mutations in collagen genes: causes of rare and some common diseases in humans. FASEB J 1991; 5: 2052-2060. Cohn DH, Byers PH. Clinical screening for collagen defects in connective tissue diseases. Clin Perinatol 1990; 17: 793-809. Byers PH, Wallis GA, Willing MC. Osteogenesis imperfecta:
12
13
14
of diffuse metaphyseal
cupping and
translation of mutation to phenotype. J Med Genet 1991; 28: 433-442. Cohn DH, Byers PH. Osteogenesis imperfecta and other inherited disorders of the structure and synthesis of type I collagen. Path01 Immunopathol Res 1988; 7: 132-138. Cohn DH, Starman BJ, Blumberg B, Byers PH. Recurrence of lethal osteogenesis imperfecta due to parental mosaicism for a dominant mutation in a human type I collagen gene (COLlAl). Am J Hum Genet 1990; 46: 591-601. Cole WG, Campbell PE, Rogers JG, Bateman JF. The clinical features of osteogenesis imperfecta resulting from a nonfunctional carboxy terminal pro I (I) propeptide of type I procollagen and a severe deficiency of normal type I collagen in tissues. J Med Genet 1991; 27: 545-551. Cole WG, Chow CW, Rogers JG, Bateman JF. The Clinical features of three babies with osteogenesis imperfecta resulting from the substitution of glycine by arginine in the pro alpha 1 (I) chain of type I procollagen. J Med Genet 1990; 27: 228-235. Edwards MJ, Graham JM. Studies of type I collagen in osteogenesis imperfecta. J Pediatr 1990; 117: 67-72. Wallis GA, Starman BJ, Zinn AB, Byers PH. Variable expression of osteogenesis imperfecta in a nuclear family is explained by somatic mosaicism for a lethal point mutation in the I( 1) gene (COLlAl) oftype I collagen in a parent. Am J Hum Genet 1990; 46: 1034-1040. Constantinou CD, Pack M, Young SB, Prockop DJ. Phenotypic heterogeneity in osteogenesis imperfecta: The mildly affected mother of a proband with a lethal variant has the same mutation substituting cysteine for 1 - glycine 904 in a type I procollagen gene (COLlAl). Am J Hum Genet 1990; 47: 670-679. Wenstrup RJ, Willing MC, Starman BJ, Byers PH. Distinct biochemical phenotypes predict clinical severity in non-lethal variants of osteogenesis imperfecta. Am J Hum Genet 1990; 46: 975-982. Superti-Furga A, Pistone F, Roman0 C, Steinmann B. Clinical variability of osteogenesis imperfecta linked to COLlA2 and
10 associated with a structural defect in the type I collagen molecule. J Med Genet 1989; 26: 358-362. 15 Munoz C, Filly RA, Golbus MS. Osteogenesis imperfecta type II: Prenatal sonographic diagnosis. Radiology 1990; 174: 181-185. 16 Tiller GE, Rimoin DL, Murray LA, Cohn DH. Tandem duplication within a type II collagen gene (COLZAI) exon in an individual with spondyloepiphyseal dysplasia. Proc Nat1 Acad Sci USA 1990; 87: 3889-3893.
17 Lee B, Vissing H, Ramirez F, Rogers D, Rimoin DL. Identilication of the molecular defect in a family with spondyloepiphyseal dysplasia. Science 1990; 244: 978-980. 18 Murray L, Bautista J, James PL, Rimoin DL. Type II collagen defects in the chondrodysplasias 1. spondyloepiphyseal dysplasias. Am J Hum Genet 1989; 45: 5-15. 19 Taybi H, Lachman RS. The radiology of syndromes, metabolic disorders and skeletal dysplasias. Yearbook Publishers, 3rd Edn, Chicago, 1990.
0 1992 Elsevier Science Publishers
EURRAD
l-10 B.V. All rights reserved.
0720-048X/92/$05.00
00241
State of the art
Collagen, genes and the skeletal dysplasias on the edge of a new era: a review and update Ralph S. Lachman avb, George E. Tiller b, John M. Graham, aDepartment of Radiology, UCLA School of Medicine, Harbor/UCLA
School of Medicine, International Skeletal Dysplasia Registry, Cedars-Sinai
(Accepted
Jr.b and David L. Rimoinb
Medical Center, Torrance, CA and bDivision of Medical Genetics, UCLA
15 October
Medical Center, Los Angeles, Caljfomia USA
1991)
Key words: Skeletal dysplasia, osteogenesis imperfecta; Skeletal dysplasia, achondrogenesis; Skeletal dysplasia, hypochondrogenesis; dysplasia, spondyloepiphyseal dysplasia; Collagen abnormality
Skeletal
Abstract This article reviews the newly described biochemical (type I and II collagen) abnormalities and specific gene defects in the skeletal dysplasias. The model of the collagen molecule is described and how collagen is processed from procollagen, where and how abnormalities occur, and the types of abnormalities produced (quantitative and qualitative). The only known type I collagen defects producing skeletal dysplasias osteogenesis imperfecta, as well as the ‘family’ of established type II collagen disorders - achondrogenesis type II, hypochondrogenesis and spondyloepiphyseal dysplasia congenita are discussed. Finally, using case presentations, the practical approach to these disorders is shqwn. The importance ofthese investigations and the subsequent reevaluation ofthe clinical and radiological findings of specifically delineated skeletal dysplasias are discussed.
Introduction Recently, there has been an explosion in the field of collagen chemistry in relationship to the skeletal dys-* plasias and connective tissue disorders. In the past, we have had to rely on the radiographic and clinical delineation of these disorders alone, but in the face of genetic heterogeneity and intrafamilial phenotypic variability, it was very difficult to know if one was dealing with a single disorder or a group of disorders with the same or different patterns of inheritance. The first step towards a solution to this problem was the examination of chondroosseous morphology (iliac crest, anterior rib or growth plate biopsy/autopsy) to establish criteria for specific diagnoses [ 11. Although many disorders showed rather specific changes, other disorders revealed similar patterns despite obvious radiographic and clinical differences.
Address for reprints: Ralph S. Lachman, M.D., Department of Radiology, Box 27, Harbor/UCLA Medical Center, 1000 W. Carson Street, Torrance, CA 90509, USA.
A concept was proposed by Jurgen Spranger that ‘families’ of skeletal dysplasias, i.e., disorders with similar radiographic and clinical findings could be due to similar genetic defects [2]. We then proposed that skeletal dysplasias with biochemical defects of type II collagen will display abnormalities in tissues where type II collagen is found (i.e., hyaline cartilage, nucleus pulposus and the vitreous of the eye). This ‘family of disorders’ concept dictates which groups of skeletal dysplasias with similar features might share a common underlying defect. For example, those skeletal dysplasias with not only epiphyseal and vertebral abnormalities, but also ocular changes, suggest an underlying defect in type II collagen, the collagen found in these particular areas. A discussion of the known collagen disorders from a skeletal dysplasia point of view must include osteogenesis inperfecta (01) (type I collagen disorders); and achondrogenesis type II hypochondrogenesis (Fig. 1 and 2) and certain spondylepiphyseal dysplasia (SED) cases (known type II collagen disorders). These are of special interest to us as radiologists because of their radiographic importance for diagnosis.
a
b
Fig. 1. Achondrogenesis type II (end of achondrogenesis type II/ hypochondrogenesis - SED congenita spectrum); Radiographs showing good skull ossification, almost no vertebral body ossilication, extremely short ribs, hypoplastic ilia, absent ossification of pubis, ischia, tarsal centers, very short tubular bones with concave ends.
The collagens are the predominant proteins of the extracellular matrix of most tissues. To date, there are in excess of 13 different known human collagens [ 3 J. Type I collagen is the major collagen of skin, bone, tendon and vessel walls. Type II collagen is the pre-
a
dominant collagen of cartilage, nucleus pulposus and vitreous. The distribution of these collagens suggest the tissue sites that should be abnormal in their respective diseases [4]. The long triple helix of the collagen molecule is flanked on each end by a globular amino terminal and carboxy terminal which function in both the assembly of procollagen and its secretion. These peptides are then removed by proteolytic cleavage in the extracellular compartment (Fig. 3). The individual molecules must be normally formed so they can assemble into an ordered libril. A mutation which interferes with : (1) expression of the collagen gene; (2) formation of the triple helix, or (3) with procollagen secretion can have dramatic effects on the structure and function of the fibrils created and potentially result in a severe clinical disorder. The evolutionary relationship of the collagen gene and its protein products provides a unique opportunity to compare mechanisms of mutation in a family of genes and observe the effects of such mutations at the biochemical, histological and phenotypic levels. This review is limited to the skeletal dysplasias with known collagen defects. The type I collagenopathies consists of the various types of osteogenesis imperfecta (01). A modified Sillence classification of osteogenesis imperfecta can be found in the Table. It is based on clinical and radiographic features. The data to date suggest that all types of 01 have primarily autosomal dominant inheritance, but autosomal recessive inheritance is likely in some cases of 01 type II group C and 01 type III. In most cases of type II 01, rare multiple affected offspring of unaffected parents can be explained by parental germline mosaicism for the mutation
b
Fig. 2. (a, b and c) Hypochondrogenesis (middle of spectrum). Skull, trunk and extremities exhibiting more mature ossification than achondrogenesis type II with large defect of posterior skull base; flat vertebral bodies with cervical and lower lumbar hypoplasia, short ribs, ischial but no pubic ossification, and shortened long tubular bones with rounded ends.
Amlno terminal.
AMINO ACID SEQUNCES ( gly - pro- hyp - gly . )
( propeptlde)
.Corboxy terminal (propeptide)
TABLE 1 Classification
of osteogenesis
Type 1 propeptidose
I
u
i
v
COLLAGEN-
Fig. 3. The processing
of procollagen
[ 191
Clinical and radiologic features Mild phenotype. Normal or close to normal stature. Hearing loss in about 50%.
PROCOLLAGEN propepildase
imperfecta
Group IA
Blue sclerae. Normal teeth.
Group IB
Dentinogenesis
Imperfecta
(DI).
w
to collagen.
(some cell lines have the mutation and some do not), resulting in a 5-77; recurrence risk. The 01s are actually a group of disorders in which at least 95% of the cases are due to mutation affecting the synthesis and/or structure of type I collagen [5-l 11. Two types of defects have been found, quantitative defects characterized by decreased production of normal type I collagen and qualitative defects in which there are populations of both abnormal type I molecules as well as structurally normal molecules. The quantitative defects appear to cause the milder forms of the disease and are found primarily in 01 type I. On the other hand, qualitative defects have been discovered in 01 types II, III and IV, and can be found in 85-90% of affected individuals tested thus far. Most of these defects are point mutations that result in substitutions for glycine in the triple helix of one of the type I collagen chains (Fig. 3). This disruption of the normal amino acid sequence slows the assembly of the molecule creating a less stable structure. Consequently, decreased secretion and/or intracellular degradation of procollagen can occur. Quantitative defects which limit the amount of type I collagen produced result in decreased deposition of normal collagen in the extracellular matrix and a decrease in mineralization while qualitative defects, in which abnormal collagen is incorporated into the matrix, severely disrupt mineralization. The site and nature of the amino acid substitution within the triple helical region can often (but not consistently) predict the biochemical and clinical phenotype [ 12-141. Case studies Patient A was referred in the first year of life in order to substantiate the diagnosis of osteogenesis imperfecta, and for the parents to receive genetic counseling and discuss prognostication of her condition. Her vaginal delivery at term was complicated by several rib fractures
Type II
Clinical and radiologic features Most severe phenotype. Blue sclerae. Hydrops 15%. Almost no calvarial ossification. Short wide femurs. Beaded ribs. Multiple rib fractures. Group IIA
Accordian-like Beaded ribs.
Group IIB
Broad crumpled long bones. Rib fractures - no beading.
Group IIC
Thin short fractured long bones. Thin ribs - fractured or mildly beaded.
femurs.
Type III
Clinical and radiologic features Blue sclera at birth changes to normal later, dentinogenesis imperfecta is common Progressive deformed skull; wormian bones or severely deossitied. Slightly short thin angulated long bones. ‘Popcorn’ calcifications in metaphyses. ‘Codfish’ vertebrae.
Type IV
Clinical and radiologic features Similar to type I. Normal sclera. Less deafness. Shorter stature than type I. Group IVA Normal teeth. Group IVB DI.
both clinically and radiographically. Her height, weight, and head circumference were all at the 50th percentile for age. Parents gave no family history of 01 and were both clinically normal. There was no consanguinity noted. Physical findings included blue sclerae and frontal bossing. The radiographs revealed thin ribs, several rib fractures and a clavicular fracture in the newborn
4
period, generalized osteoporosis and over modelled tubular long bones with thin cortices (Figs. 4-5). Radiographically and clinically, this patient appeared most likely to have 01 type IA. Collagen analysis from cultured skin fibroblasts of the patient and both parents revealed normal migration on electrophoresis. However, the patient’s cells synthesized a reduced quantity of type I collagen, whereas her parents’ synthesized collagen in normal amounts. This finding of a quantitative defect suggested 01 type I as the result of a new dominant mutation and only a small risk of recurrence in this family, confirming the clinical impression prior to the collagen analysis. The family was also advised about the 50 y0 chance of deafness, as well as other potential orthopedic management problems. Patient B was a newborn male delivered by caesarean section at term because of polyhydramnios and prolonged labor. Both parents were normal and not consanguineous. The diagnosis of 01 was suggested on clinical grounds and conlirmed by the radiographs. The lower extremities were bowed and distorted. The sclerae were blue and a soft skull was present. The newborn radiographs revealed generalized osteoporosis, shortening of the long tubular bones (especially short, widened femurs and tibias with very thin cortices, angulation and fractures), ribs containing many acute and healing rib fractures, and very poor calvarial ossification
4
(Fig. 6). Radiographically, this patient appeared to represent a case of 01 type IIB on the milder end of the spectrum of type II and not exhibiting the classic ‘accordion’ femurs and beaded ribs of the more common type IIA. Prior to the patient’s death in the second week of life, a skin biopsy was obtained. Collagen biochemical analysis revealed both normal and abnormal populations of type I procollagen molecules. This qualitative defect appeared to be an autosomal dominant new mutation with a recurrence risk of 5-7%. Although the recurrence risk is small, future pregnancies in this case can be monitored either by fetal ultrasound or by chorionic villus biopsy (CM) for collagen analysis since type I collagen is produced by villus tissue. Serial fetal ultrasound with no known morbidity is the first test of choice because of: (a) the time required to culture and label cells, and to perform the biochemical analysis and (b) the small but significant risk of pregnancy loss due to the CVS procedure [ 151. Patient C was the full term male product of unrelated parents delivered by caesarean section for breech presentation and a suspected skeletal dysplasia. Prenatal ultrasound was performed at 15 weeks gestation (with serial repetition), delineating short and thickened bones in the lower extremities (Fig. 7). At birth, the patient had blue gray sclerae, inguinal hernias and short, curved, long bones (especially of the lower extremities). Newborn radiographs revealed generalized osteoporo-
6
5
Fig. 4. Patient A, 01 type IA, newborn; chest radiograph Fig. 5. Patient A, 01 type IA, 6 months old; Lower extremities
revealing thin ribs, clavicular fracture and generalized revealing overtubulated
osteoporosis.
long bones with thin cortices and osteoporosis.
Fig. 6. Patient B, 01 type IIB, f&term newborn; Full body radiograph showing generalized osteoporosis, deficient calvariai ossification, especially short widened and bent femurs and tibias with thin cortices and fractures.
7
9
8
Fig. 7. Patient C, 01 type IIC-III,
fetal ultrasound
at 15 weeks gestation
showing a short widened femur.
Fig. 8. Patient C, 01 type IIC-III, full-term newborn; trunk and extremity radiograph revealing generalized osteoporosis; shortened angulated long bones with thin cortices (especially significant femoral involvement), and slightly thin ribs. Radiographic features closest to 01 type IIC. Fig. 9. Patient C, early infancy; CT revealing signiticant generalized
sis, a brachycephalic osteoporotic skull with high orbital roofs, dentinogenesis imperfecta, shortened angulated long bones with thin cortices, and marked shortening and widening of the femurs (Fig. 8). Radiologically and clinically, he appeared to represent 01 type IIC as an infant but with the passage of time resembled those in the severe end of the type III spectrum. The patient developed hydrocephalus without clinical signs of increased intracranial pressure (Fig. 9). During the ftrst year of life, frequent apparently spontaneous fractures occurred, with profuse callus formation. Collagen analysis from cultured skin f’ibroblasts showed a qualitative defect with both normal and abnormal populations of type I collagen. Cyanogen bromide peptide analysis revealed overmodification of the abnormal alpha chains confined to the amino end of the molecule. Collagen studies of the parents were normal. This condition in this patient appeared to represent a new autosomal dominant disorder with a low likelihood of a recurrence due to parental germline mosaicism (5-7x). Since early changes were present by fetal ultrasound in the proband, one can recommend serial fetal ultrasonography and/or CVS to monitor future pregnancies. The type II collagen defects that affect certain of the skeletal dysplasias have very similar abnormalities to the collagen I defects of osteogenesis imperfecta. They also exhibit quantitative and qualitative abnormalities. Patient D is now a 24-year-old male with a diagnosis
ventricular
dilatation.
of SED made at about a year of age and followed since age 6 years by our clinic. He was the full-term product of a 34-year-old gravida 3 following an uncomplicated pregnancy, labor and delivery. At birth, clinical examination was normal except for short stature. There was no history of consanguinity in the family. He was first evaluated at about 6-8 months because of inability to sit and crawl and radiographs at that time were consistent with a diagnosis of SED congenita. At 5 years of age, significant myopia developed. Cervical spine fusion for excess atlanto-axial mobility was performed at age 9 years as well as some other orthopedic procedures. On physical examination, he has remained disproportionately short (at age 15 years he was at the 50th percentile for a 9 year old), with a normal skull. He exhibits widening at all the major joints and diminished motion at the hips and elbows. Radiographic examinations at ages 6, 7 and 16 years revealed very significant generalized epiphyseal ossification delay, as well as significant thoracic platyspondyly with irregularity suggesting the radiographic diagnosis of some form of spondyloepiphyseal dysplasia (Figs. 10-14). The electrophoretic analysis of type II collagen extracted from an iliac crest biopsy revealed both normal and abnormally migrating a-chains. Cyanogen bromide peptide analysis showed overmodification of all major peptides. DNA sequence analysis of the type II collagen gene revealed a 45 base-pair dupli-
b
b
a
Fig. 10. (a) Patient D spondyloepiphyseal dysplasia (SED) at age 7 years; pelvis and hips revealing small capital femoral epiphyseal centers with flocculent calcification, vertical ischia and down slanted pubic bones. (b) Patient D, at 16years of age; The femoral heads are completely ossified and the epiphyseal plate is fused.
12b
12a
Fig. 11. Patient D, at age 7 years; knees revealing marked epiphyseal ossification epiphyseal bone density.
delay, apparent
epiphyseal
irregularity
and decreased
Fig. 12. (a and b) Patient D, 6 years old; thoracic and lumbar spine showing lack of interpediculate flare of lower lumbar vertebrae, platyspondyly of thoracic vertebrae with anterior pointing and rounding; but normal height of lumbar vertebrae. Fig. 13. Patient D, at 16 years of age; Showing progression
of the platyspondyly spine.
cation in exon 48 of the triple helical coding region. The mutation arose in the maternally derived allele, most likely by an unequal crossover during meiosis. This resulted in the addition of 15 amino acids to the triple
and end plate irregularity
and sclerosis of the thoracic
helical portion of the a-chain, and likely disrupted the rate of assembly of molecules incorporating one or two such chains [ 161. The patient was advised of a 50% recurrence risk in his offspring and that the molecular
a
b
a
b
Fig. 14. (a and b) Patient D, flexion films at 7 and 16 years of age delineating excess mobility of Cl, C2 before posterior fusion and the resultant immobility following surgical fusion.
definition of his disease should allow for prenatal diagnosis by CVS at approximately 10 weeks gestation. Patient E was the product of a normal spontaneous delivery at 33 weeks gestation, with a birth weight of 1710 g, and a length of 13.5 inches. The pregnancy was complicated by the discovery of short limbs and a small chest on prenatal ultrasound at 19 weeks. Physical examination at birth showed disproportionate extremities and an abnormal bell-shaped thorax. The baby required intubation in the delivery room, and chronic respirator support. The prenatal ultrasound findings suggested the possibility of Ellis Van Crefeld, asphyxiating thoracic dysplasia or one of the short rib polydactyly syndromes as the most likely diagnosis. Postnatally the diagnosis of Ellis Van Crefeld (chondroectodermal dysplasia) was made clinically and radiographically by the referring physicians primarily because of a finding of unilateral polydactyly. Upon referral to the International Skeletal Dysplasia Registry, we felt that the radiographic findings more closely fit the achondrogenesis-hypochondrogenesis-SED spectrum. The early radiographs were those of SED congenita, but repeat examination at 3 months of age showed changes closer to the hypochondrogenesis (a more severe, usually lethal) portion of that spectrum (Figs. 15-2 1). The infant was removed from respiratory support after appropriate discussion with the parents, and the baby expired. Chondroosseous morphology showed hypervascular cartilage, friable matrix, foamy areas around lacunae,
Fig. 15. Patient E, hypochondrogenesis - SED congenita, 33 week gestation newborn: spine and pelvis revealing rounded platyspondyly; absent pubic bones, hypoplastic vertical ischia, and rounded iliac wings.
and extension of chondrocytes into trabecular areas. These findings are also those of mild hypochondrogenesis or severe SED. Electrophoretic analysis of type II collagen extracted from cartilage revealed both normal and abnormally migrating a-chains. Cyanogen bromide peptide analysis showed overmodification of most peptides. Amino acid sequencing demonstrated a substitution of glutamic acid for glycine at position 853 of the triple helical region. DNA sequencing confirmed the single nucleotide substitution that would account for this amino acid substitution. This disrupted the development of the normal triple helix in this region, and likely interfered with the rate of assembly, thus resulting in overmodification of the procollagen molecule [ 17-181. The patient was heterozygous for a gene defect and restriction fragment length polymorphism(RFLP), which enabled us to determine that the mutation arose in the paternally derived allele of that collagen (COL2AI) gene. Since the molecular studies showed that this was a dominant trait, the family was advised that their recurrence risk was low. Analysis of DNA from the father’s sperm showed no evidence for gorminal mosaicism, thus lowering the parent’s recurrence risk for another child with this form of skeletal dysplasia to that of the general population. In conclusion, it is clear that we have come a long way from the radiographic definition (as the solitary
8 16
17
Fig. 16. Patient E, newborn;
showing thin ribs, elongated clavicles, deficient ossification
Fig. 17. Patient E, 33 week gestation newborn; demonstrating
no talus or calcaneal ossification; margins.
tool) for the skeletal dysplasias through an era of chondromorphologic (both light and ultrastructural) evaluation and into this present era of biochemical and gene
of the cervical spine and shortened
humeri.
and shortened tibiae with smooth metaphyseal
delineation. All of this information is very important in order to not only determine a distinct entity, but to achieve the best possible genetic counseling and in-
18b
16a
Fig. 18. (a and b) Patient E, at 3 months of age; skull revealing elevated orbital roofs, midface hypoplasia occipital region. Fig. 19. Patient E, at 3 months;
and an ossification
Later spine film showing some vertebral growth but still characteristic
defect in the
rounded platyspondyly.
20b
Fig. 20. Patient E, at 3 months;
demonstrating
2la
21b
short anteriorly flared ribs and small thoracic volume.
Fig. 21. (a and b) Patient E, at 3 months: hips and lower extremities and wrist revealing the development sclerosis and still lack of ankle ossification.
trauterine diagnosis for the family. Now that we can determine the biochemical and gene defect in certain of these disorders, the radiologist can carefully examine biochemically proven cases with these disorders to define the variability of expression of each mutation. In this review article, we have tried to clearly delineate these recent biochemical and genetic breakthroughs in the chondrodysplasias, as well as present specific cases which show the practical value of these investigations. Acknowledgements Special thanks to Elise Spears and Savona Huching for technical and artistic assistance for this manuscript. This study was supported in part by a USPHSNIH Program Project Grant (HD22657).
6
7
8
9
10 11
References Sillence DO, Horten WA, Rimoin DL. Morphologic studies in the skeletal dysplasias. A Review. Am J Path01 1979; 96: 81 l-870. Spranger J. Pattern recognition in bone dysplasias in: Endocrine Genetics and Genetics of Growth. C.J. Papadatos and C.S. Bartocas, eds. New York: Liss, 1985; 315-342. Kuivaniemi H, Tromp G, Prockop DJ: Mutations in collagen genes: causes of rare and some common diseases in humans. FASEB J 1991; 5: 2052-2060. Cohn DH, Byers PH. Clinical screening for collagen defects in connective tissue diseases. Clin Perinatol 1990; 17: 793-809. Byers PH, Wallis GA, Willing MC. Osteogenesis imperfecta:
12
13
14
of diffuse metaphyseal
cupping and
translation of mutation to phenotype. J Med Genet 1991; 28: 433-442. Cohn DH, Byers PH. Osteogenesis imperfecta and other inherited disorders of the structure and synthesis of type I collagen. Path01 Immunopathol Res 1988; 7: 132-138. Cohn DH, Starman BJ, Blumberg B, Byers PH. Recurrence of lethal osteogenesis imperfecta due to parental mosaicism for a dominant mutation in a human type I collagen gene (COLlAl). Am J Hum Genet 1990; 46: 591-601. Cole WG, Campbell PE, Rogers JG, Bateman JF. The clinical features of osteogenesis imperfecta resulting from a nonfunctional carboxy terminal pro I (I) propeptide of type I procollagen and a severe deficiency of normal type I collagen in tissues. J Med Genet 1991; 27: 545-551. Cole WG, Chow CW, Rogers JG, Bateman JF. The Clinical features of three babies with osteogenesis imperfecta resulting from the substitution of glycine by arginine in the pro alpha 1 (I) chain of type I procollagen. J Med Genet 1990; 27: 228-235. Edwards MJ, Graham JM. Studies of type I collagen in osteogenesis imperfecta. J Pediatr 1990; 117: 67-72. Wallis GA, Starman BJ, Zinn AB, Byers PH. Variable expression of osteogenesis imperfecta in a nuclear family is explained by somatic mosaicism for a lethal point mutation in the I( 1) gene (COLlAl) oftype I collagen in a parent. Am J Hum Genet 1990; 46: 1034-1040. Constantinou CD, Pack M, Young SB, Prockop DJ. Phenotypic heterogeneity in osteogenesis imperfecta: The mildly affected mother of a proband with a lethal variant has the same mutation substituting cysteine for 1 - glycine 904 in a type I procollagen gene (COLlAl). Am J Hum Genet 1990; 47: 670-679. Wenstrup RJ, Willing MC, Starman BJ, Byers PH. Distinct biochemical phenotypes predict clinical severity in non-lethal variants of osteogenesis imperfecta. Am J Hum Genet 1990; 46: 975-982. Superti-Furga A, Pistone F, Roman0 C, Steinmann B. Clinical variability of osteogenesis imperfecta linked to COLlA2 and
10 associated with a structural defect in the type I collagen molecule. J Med Genet 1989; 26: 358-362. 15 Munoz C, Filly RA, Golbus MS. Osteogenesis imperfecta type II: Prenatal sonographic diagnosis. Radiology 1990; 174: 181-185. 16 Tiller GE, Rimoin DL, Murray LA, Cohn DH. Tandem duplication within a type II collagen gene (COLZAI) exon in an individual with spondyloepiphyseal dysplasia. Proc Nat1 Acad Sci USA 1990; 87: 3889-3893.
17 Lee B, Vissing H, Ramirez F, Rogers D, Rimoin DL. Identilication of the molecular defect in a family with spondyloepiphyseal dysplasia. Science 1990; 244: 978-980. 18 Murray L, Bautista J, James PL, Rimoin DL. Type II collagen defects in the chondrodysplasias 1. spondyloepiphyseal dysplasias. Am J Hum Genet 1989; 45: 5-15. 19 Taybi H, Lachman RS. The radiology of syndromes, metabolic disorders and skeletal dysplasias. Yearbook Publishers, 3rd Edn, Chicago, 1990.
