Am. J. Hum. Genet. 50:15-28, 1992

Mechanisms of Ring Chromosome Formation in I I Cases of Human Ring Chromosome 21 M. J. McGinniss,* H. H. Kazazian, Jr.,* G. Stetten,*T M. B. Petersen,* H. Boman,t E. Engel,§ F. Greenberg,!1 J. M. Hertz,# A. Johnson,** Z. Laca,tt M. Mikkelsenjt S. R. Patil,§§ A. A. Schinzel,IIII L. Tranebjaerg,$4 and S. E. Antonarakis* *Center for Medical Genetics, Department of Pediatrics and Medicine, and tDepartment of Obstetrics and Gynecology, Johns Hopkins University School of Medicine, Baltimore; $Department of Medical Genetics, University of Bergen, Bergen; §Institute of Medical Genetics, University of Geneva, Geneva; Illnstitute for Molecular Genetics, Baylor College of Medicine, Houston; #Institute of Human Genetics, University of Aarhus, Aarhus, Denmark; * *Division of Medical Genetics, Jefferson Medical College, Philadelphia; t tCenter for Medical Genetics, Institute for Mental Health, Belgrade; t$Department of Medical Genetics, John F. Kennedy Institute, Glostrup, Denmark; §§Department of Pediatrics, University of Iowa, Iowa City; and II Ilnstitute of Medical Genetics, University of Zurich, Zurich

Summary We studied the mechanism of ring chromosome 21 (r(21)) formation in 13 patients (11 unique r(21)s), consisting of 7 from five families with familial r(21) and 6 with de novo r(21). The copy number of chromosome 21 sequences in the rings of these patients was determined by quantitative dosage analyses for 13 loci on 21q. Nine of 11 r(21)s, including the 5 familial r(21)s, showed no evidence for duplication of 21q sequences but did show molecular evidence of partial deletion of 21q. These data were consistent with the breakage and reunion of short- and long-arm regions to form the r(21), resulting in deletion of varying amounts of 21q22.1 to 21qter. The data from one individual who had a Down syndrome phenotype were consistent with asymmetric breakage and reunion of 21q sequences from an intermediate isochromosome or Robertsonian translocation chromosome as reported by Wong et al. Another patient, who also exhibited Down syndrome, showed evidence of a third mechanism of ring formation. The likely initial event was breakage and reunion of the short and long arms, resulting in a small r(21), followed by a sister-chromatid exchange resulting in a double-sized and symmetrically dicentric r(21). The phenotype of patients correlated well with the extent of deletion or duplication of chromosome 21 sequences. These data demonstrate three mechanisms of r(21) formation and show that the phenotype of r(21) patients varies with the extent of chromosome 21 monosomy or trisomy. Introduction

The molecular basis of ring chromosome formation has not been well studied, despite the hundreds of case reports of human ring chromosomes (Schinzel 1984). Ring chromosomes are seen with all human chromosomes (Therman 1986), and their frequency is 1 in 25,000 recognized conceptions (Jacob 1981). Almost one-half of reported ring autosomes are from the acrocentric chromosomes (Kosztolanyi 1987). It has been Received June 20, 1991; revision received August 27, 1991. Address for correspondence and reprints: Matthew J. McGinniss, Ph.D., Center for Medical Genetics CMSC 1004, The Johns Hopkins Hospital, 600 North Wolfe Street, Baltimore, MD 21205. o 1992 by The American Society of Human Genetics. All rights reserved. 0002-9297/92/5001-0003$02.00

assumed that ring chromosomes arise from breaks in each chromosome arm, followed by' end-to-end reunion (Lejeune 1968; Therman 1986; Kosztolanyi 1987; Gardner and Sutherland 1989). However, this mechanism remains unproved. The resulting phenotype, at least with acrocentric rings, would be similar to that seen with terminal deletions of the long arm of the particular chromosome. We recently studied the molecular mechanism for the formation of a human ring chromosome 21 (r(21)) (Wong et al. 1989). In this isolated case, an isochromosome of one maternal chromosome 21 (i(21 q)) was the likely structure that preceded ring formation. Asymmetric breaks and reunion of the long arms of this intermediate chromosome gave rise to a ring that contained duplicated regions of the proximal long 15

16

arm, single-copy regions of a portion of the distal long arm, and deletion of terminal regions of the long arm. The child with this r(21) was mentally retarded and had some dysmorphic features (Stetten et al. 1984). This mechanism of ring formation could in other instances result in an array of phenotypes ranging from normal to severely abnormal, depending on what DNA sequences are duplicated in or lost from the ring chromosome. The objectives of the present study were (1) to determine the mechanism of formation of the r(21) in additional patients and (2) to compare the clinical phenotype with the chromosome 21 rearrangement. Using quantitative Southern blot analysis of RFLP markers and assessment of dosage of alleles at PCR markers, we inferred the copy number of DNA sequences in 11 unique r(21)s. We hypothesized that other r(21)s may also be derived from an intermediate (i(21q)) or a Robertsonian translocation chromosome (21 q21 q). However, results from 9 of 11 cases indicated that the ring was formed by breakage and reunion of the short and long arms, resulting in partial monosomy of distal 21q. Material and Methods Cytogenetics

The patient population consisted of 13 Caucasians with cytogenetically verified r(21) who were collected from several centers. Clinical diagnosis of r(21) was made from cytogenetic analyses of blood lymphocytes, amniocytes, or skin fibroblasts, performed in the individual laboratories. Parents of most patients were also karyotyped. In addition, karyotypes of the lymphoblastoid cell lines were prepared to confirm the presence of the r(21) in transformed cell lines. In some cases the constitutive heterochromatin near centromeres was studied by C-banding (Sumner 1972; Sahakian et al. 1990), and the nucleolus organizer region was visualized by Ag-NOR staining (Goodpasture and Bloom 1975). DNA Analysis

Genomic DNA was isolated from whole blood or cells from r(21) individuals and family members. Control genomic DNA samples were also obtained from unrelated normal individuals and from nonmosaic trisomy 21 individuals. Methods for restriction-enzyme digestion, agarose-gel electrophoresis, Southern transfer, 32P-labeling by random priming, hybridization,

McGinniss et al.

and autoradiography were performed according to standard procedures (Sambrook et al. 1989). Probes.-The probe-enzyme combinations used in quantitative Southern blot analyses were (1) D21K9 of locus D21S13 that recognizes a TaqI DNA polymorphism, (2) p21-4U of locus D21S110 that recognizes an MspI polymorphic site, (3) Fr8-77 of locus D21S82 that detects a VNTR polymorphism with at least three alleles, (4) pPW518-lR of locus D21S55 that recognizes a polymorphic insertion/deletion with XbaI, (5) pBSM132-8b of locus MX1 that recognizes a polymorphic site with MspI, (6) SF-43 of locus D21S42 that detects a TaqI DNA polymorphism, (7) CRI-L427 of the VNTR polymorphic locus D21S1 12, (8) ML18 of locus COL6A1 that detects a VNTR polymorphic system, and (9) pHS100/2.2 of the S1OOB locus, the most distal and nonpolymorphic marker (references for these probes and polymorphisms are found in Kidd et al. 1989; Petersen et al. 1991b; McGinniss et al., in press). These chromosome 21 loci were chosen because they are highly informative and because they map from the most proximal (21q11.2) to the most distal (21q22.3) regions of the long arm. These loci are listed in order from centromere (e.g., D21S13) to telomere (e.g., S1OOB), as determined from genetic (Petersen et al. 1991b) and physical maps (Gardiner et al. 1990; Burmeister et al. 1991). The reference probe for quantification of chromosome 21 sequences was a 1.8-kb BamHI fragment from the 5' region of the beta-globin locus on chromosome 11. PCR Markers. -Polymorphic markers due to short sequence repeats at four chromosome 21 loci (D21S120, 21-GT12, D21S156, and HMG14) were also used after PCR amplification, to determine genomic copy number of chromosome 21 sequences in r(21) individuals. Polymorphisms at these loci are either due to (GT). dinucleotide repeats or due to variation in the poly(A) tail of an Alu repeat. These polymorphisms are detected using PCR amplification of genomic DNA by radiolabeled PCR primers. The polymorphic markers studied here were (1) a (GT)n repeat of locus D21S120 (Burmeister et al. 1990), (2) a (GT). repeat at 21-GT12 (A. C. Warren, personal communication), (3) a (GT). repeat of locus D21S156 (Lewis et al. 1990), and (4) two (GT)n repeats and variation of the poly(A) tract of an Alu repeat within introns of the HMG14 gene (Petersen et al. 1990, 1991a). Methods for primer radiolabeling, PCR amplification, electrophoresis, and autoradiography have been described (Petersen et al. 1990, 1991a). The PCR markers

17

Human Ring Chromosome 21 2

D21S210, D21S156, and HMG14 have been positioned on the map of chromosome 21 (Burmeister et al. 1990; Petersen et al. 1991b). The position of 21-GT12 was based on unpublished results from our laboratory.

b

1P*7

Gene Dosage Analysis

Genomic copy number of chromosome 21 sequences was determined as described elsewhere (McCormick et al. 1989). In brief, densitometric scans (UltraScan XL; LKB Bromma) of autoradiograms were made, and the ratio (chromosome 21 probe/reference probe for the beta-globin locus on chromosome 11) of the r(21) patient was compared with the average ratio determined from several unrelated normal control samples. This ratio, after comparison with that of normals, should equal 0.5 if one copy of the genomic region is present (monosomy), 1.0 for normal gene copy (disomy), and 1.5 if the genomic region is present in three copies (trisomy). Alternatively, when two alleles of polymorphic loci were identified, a 1:1 ratio of allele intensities on autoradiograms was interpreted as indicating two copies of a sequence, while a 1:2 ratio of allele intensities was interpreted as indicating three copies. The determination of gene dosages by using PCR amplification of regions of genomic DNA containing short sequence repeats has been described elsewhere (Petersen et al. 1991a). A normal copy number of two was inferred when individuals showed two alleles that were approximately equal in intensity. Trisomy for particular DNA sequences was interpreted from the autoradiograms only when an individual showed three different allelic fragments, or dosage was evident when two fragments showed a 1:2 ratio of allele intensities. Individual DNA samples that revealed single fragments on the autoradiograms were scored as noninformative at that locus, since we could not distinguish between one, two, or three copies of the same allele. Results Clinical Phenotypes

The 13 r(21) patients in the present study were diagnosed at various ages from birth to 45 years. Their phenotypes varied from clinically normal to moderately retarded with dysmorphic features. All seven individuals from the five families with inherited r(21)s (fig. 1) were clinically normal on the basis of detailed physical examinations. These seven individuals showed

3

5

4

TN. Partial pedigrees of five familial cases of ring chroFigure I mosome 21. Blackened symbols denote ring chromosome 21 individuals; unblackened symbols denote individuals with a normal karyotype. The unblackened diamonds represent spontaneous abortions of unknown sex. The arrows indicate the ring chromosome 21 individuals studied here.

normal intelligence, but none were formally tested for the present study. Four of the five familial cases of r(21) were ascertained by cytogenetic analyses for prenatal diagnosis. The fifth familial case (number 5 in fig. 1) was ascertained because the son (not studied here) had bilateral cryptorchidism and delayed onset of puberty at age 15 years (Hertz 1987). In two of the families the r(21) segregated in three generations. The six isolated cases of r(21) were ascertained either because of the presence of dysmorphic features at birth or because developmental delay during the first few years of life. The clinical features of the isolated cases are listed in table 1, and photographs of the six isolated cases are shown in figure 2. All of the isolated cases had an abnormal phenotype, ranging from slight mental retardation with few dysmorphic features to the presence of multiple dysmorphic features and moderate mental retardation. Mental retardation and developmental delay were the most frequent features (6/6) associated with the isolated r(21) patients, while other features such as short stature (4/6), microcephaly (4/6), bilateral epicanthal folds (4/6), short neck (3 / 6), and small or simplified ears (3 / 6) were less frequent. Two of the six isolated cases showed a Down syndrome phenotype (see patients 10 and 11 in table 1), suggesting that chromosome 21 sequences in these two individuals might be duplicated in the r(21). For these two patients a clinical questionnaire of features usually associated with Down syndrome (Jackson et al. 1976) was completed by the referring physician. Some of the abnormalities associated with Down syndrome and noted in both patients included short stat-

:>

0

E > ON-

r

P.

+++

ro o1

W3

3 Q

+ +

._

+ + +

a

U. + N .0 (A

U

(A

O

Sv

+

I+ +

)

rq

N k0) v

en

Us

-~

N

C

.-

.-

4.

+ + + +

+

H

0

z

+

4en

+

e

t,

H w

z

0 H-

fn

+ +

I+ + +

+ +

+

-4

CL 0

4a

0 c

0. 0.

0

N-

N rq

E

+

.0 c I.

49 I I

0~~~~+*

r-r0 N

+

El

+

VI.

*

*

_

.

.

El

AI.)L 'I ._

U

X

* 4.

0

s

.~~~~~~0

V) .0

I18

.

.

.

_

.D.

6o ° I>

+

ww

CZ

+

I I I

+

,

Xw

.

0( +I3

C4

'-CC4)

Ce :1

.-

bo cus 4)4)

+

I

+

+ I

+

I

+ +

-

U'

0 (0

4.

o-j

O Cd

0

I

1+r.C0

I +

I

I +

I I

co 4-

Cu

Cu

'CuU *1 Co W ^u -

CU 4)Cu#

+

+

.1

0 4)

I

#

-

C

4)4~

2 o +

+

0

_

Cu

+II

C

I r4 .4 4-

D> 4)

CZ 4)

bt CZ

4)4 Cu

*

.0

.

*4

..)

*C

..u

.

.

Cr

.

.u

.

Cu

4)j

.

*

*

*

.

.,

v

(A *

*

*m

Cu*

.

Y

4) 0 4) 4)

* *

4)

&)

VJ

.

*

.

4.,

*

.

.

Cu

0.4

19

McGinniss et al.

20

D

Figure 2 Photographs of six individuals (A-F) with r(21) and abnormal phenotype. A, patient 6 at age 12 years. B, Patient 7 at age 3.5 years. C, Patient 8 at age 7 years. D, Patient 9 at age 5 years. E, Patient 10 at age 12 years. F, Patient 11 at age 1.5 years. These patients all exhibited mental retardation and dysmorphic features. (See table 1 for the listing of individual phenotypes. Patient numbers are as listed in table 2.)

ure, mental retardation, microcephaly, flat occiput, bilateral epicanthal folds, short neck, and a short fifth finger (table 1). These patients did not have Brushfield spots, heart defects, or duodenal atresia. Cytogenetic Analyses Patient karyotypes are listed in table 2. Fibroblast cultures were established on patients 4-6 and 9. Lymphoblastoid cell lines were established on patients 1A, 1B, 6-8 and 10. The presence of r(21) was confirmed in lymphoblasts from patients 1 A, 1 B, and 6-8, where the DNA analyses were performed exclusively with DNA isolated from lymphoblasts. In all the pa-

tients except one, r(21) was present in 90% or more of the cells examined. In patient 9 only 16% of blood lymphocytes sampled during the neonatal period contained an r(21) (table 2). However, at 3 years of age, 92% of fibroblasts from this patient contained the r(21). High-resolution banding analyses indicated breakpoints in the r(21), at p11 and q22.3 in patients 1A, 1B, and 5 and at p12 and q22.3 in patient 6. C-banding results on three of these individuals (1B, 5, and 6) showed one C-band-positive region in the r(21), suggesting that these r(21)s were monocentric (chromosomes from patient 1A were not C-banded). The

(A -

-0

0 0 0 o 0 0 0 0 0 0000 0 w \0 r~m tn N t tn N 0 0 n 00

n t

N .

r.

z

I-

:s

V)

to

V

S

-4 44

.

0

N

a 0 O~ 0

W0

>

N

,

w +

ce

oo

N Cl

44*

N

-

11

1.0

%

S +

''-

N +

4

>

00 0

+

A'

INI - 0

0 .-

N N ~