After con-ventional karyotyping, some samples were selected for CGH analysis, either because the patient phenotype was highly suggestive of a chromosome abnormality even though the karyo
Trang 1American Journal of Medical Genetics 113:125 – 136 (2002)
Usefulness of High-Resolution Comparative
Genomic Hybridization (CGH) for Detecting
and Characterizing Constitutional Chromosome
Abnormalities
Gro Oddveig Ness, Helle Lybæk, and Gunnar Houge*
Center for Medical Genetics and Molecular Medicine, Haukeland University Hospital, Bergen, Norway
Comparative genomic hybridization (CGH)
is a technique for detection of chromosomal
imbalances in a genomic DNA sample We
here report the application of the recently
developed method of high-resolution CGH
on DNA samples from 66 children having
various degrees of delayed psychomotor
development with or without clear
dysmor-phic features and congenital malformations
In 5 of 50 patients with apparently normal
karyotypes, a deletion or duplication was
revealed by CGH Only one of these cases had
a subtelomeric rearrangement In one of
se-ven cases with a de novo apparently balanced
translocation, deletions were found In all
nine cases where the origin of a marker
chro-mosome or additional chromosomal
mate-rial was difficult to determine, CGH gave a
precise identification The following
find-ings were from cases having a deletion or
duplication as the sole chromosomal
imbal-ance; dup(2)(p16p21), del(4)(q21q21), del(6)
(q14q15), del(6)(p12p12), dup(6)(q24qter),
and dup(15)(q11q13) One case had dup(9)
(p11pter) combined with a very small
sub-telomeric deletion on 6q In our hands, CGH
is highly useful not only for identifying
known chromosomal imbalances, but also
for finding elusive deletions or duplications
in the large group of children with
develop-mental delay with or without congenital
abnormalities In such cases, the diagnostic
yield of CGH appears to be higher than what
has been reported from subtelomeric FISH
screening ß 2002 Wiley-Liss, Inc.
KEY WORDS: dysmorphology;
cytogene-tics; FISH; deletions; dupli-cations
INTRODUCTION The complex relationship between genotype and phenotype, the plethora of established and possible syn-dromes, and the sometimes highly variable phenotypic expression of a given gene or chromosomal defect con-strain our ability to find a diagnosis or an explanation for
an abnormal child’s condition in the dysmorphology clinic Our ability to accurately predict a mutation in a specific gene or a given chromosomal abnormality based
on phenotype investigations is even more limited When lacking phenotypic handles pointing to specific syn-dromes or a family history that gives indicative clues, the finding of a genetic cause heavily depends on the sensitivity of available screening methods If there is no reason to suspect a metabolic disease, the only screening method for genetic defects in general use is conventional chromosome analysis by G- or R-banding Because our clinical diagnostic skills are likely to remain insufficient also in the future, better molecular screening methods would be extremely useful
Comparative genomic hybridization (CGH) is the only DNA-based screening method that can detect chromo-somal imbalances (e.g., deletions and duplications) in one experiment All other DNA-based methods (e.g., specific gene tests, or fluorescence in situ hybridization [FISH] tests for chromosomal imbalances such as the microdeletion syndromes) return only the result that is asked for An additional advantage of CGH is the inde-pendence of living cells from the tissue to be investi-gated The usefulness of CGH in tumor cytogenetics for identifying chromosomal gains and losses has been well documented [James, 1999; Gray and Collins, 2000] When searching for constitutional chromosome ab-normalities, CGH analysis has so far been less useful than conventional G-banding This is partly because balanced aberrations will not be found, and partly because the sensitivity for detecting imbalances has
*Correspondence to: Gunnar Houge, Center for Medical
Genetics and Molecular Medicine, Haukeland University Hospital,
N-5021 Bergen, Norway E-mail: gunnar.houge@haukeland.no
Received 10 September 2001; Accepted 14 April 2002
DOI 10.1002/ajmg.10593
ß2002 Wiley-Liss, Inc
Trang 2been on the same level as routine G-banding [detection
of gains and losses of at least 10 Mb; Kallioniemi et al.,
1994; Bentz et al., 1998] Therefore, CGH has been most
widely used for identifying or confirming the presence of
aberrations that already have been observed by
conven-tional G-banding analysis [Erdel et al., 1997; Boceno
et al., 1998; Levy et al., 1998; Breen et al., 1999; Weimer
et al., 2000; Rigola et al., 2001]
Recently, a method of analyzing CGH data based on
dynamic standard reference intervals instead of fixed
intervals was reported [Kirchhoff et al., 1998], resulting
in a two-to-threefold improvement of resolution, i.e.,
approaching 3 Mb In their hands, such high-resolution
CGH analysis proved useful for detecting aberrations
in patients with an apparently balanced karyotype
[Kirchhoff et al., 2000, 2001] In the present work, we
have explored the usefulness of this novel CGH analysis
software in our routine diagnostic service In addition
to being the most rational and precise way to identify
chromosomal material of unknown nature, the novel
CGH analysis software has sufficient sensitivity to
detect deletions and duplications missed by previous
good-quality G-banding Some of our presented cases,
with a small deletion or duplication as the only finding,
might add to our understanding of genotype/phenotype
correlation In conclusion, we find that CGH is a highly
useful tool also in the routine cytogenetic laboratory,
providing an explanation for an aberrant phenotype in
approximately 10% of previously undiagnosed cases
MATERIALS AND METHODS
Patient Samples Patient samples were received for routine cytogenetic
analysis at Center for Medical Genetics and Molecular
Medicine at Haukeland University Hospital After
con-ventional karyotyping, some samples were selected for
CGH analysis, either because the patient phenotype
was highly suggestive of a chromosome abnormality
(even though the karyogram appeared to be normal), or
because observed aberrations were difficult to classify or
investigate thoroughly Most cases with aberrations
detected by CGH have later been referred to us for
phenotypic evaluation of the child and genetic
counsel-ing of the parents Brief information about the patients’
phenotypes is given in Table I
Cytogenetic Analysis Metaphases from peripheral blood lymphocytes were
prepared according to standard procedures, using
phy-tohemagglutinin for stimulation of the lymphocytes,
and methotrexate for synchronization of the cell cycle
Metaphase chromosomes of good quality were
karyo-typed by conventional G-banding, giving a band level of
approximately 500
CGH Slides with normal lymphocyte metaphase
chromo-somes for CGH analysis were postfixed in 1%
formalde-hyde for 5 min at 48C, dehydrated in an ethanol series
(70%, 85%, and 100%) and stored at208C before hybri-dization CGH was performed essentially as described
by Kallioniemi et al [1994] Normal male or female DNA was used as reference DNA after labeling with Texas Red-5-dUTP (NEN Life Science Products, Zaventem, Belgium) using nick translation Patient DNA was labeled with fluorescein isothiocyanate (FITC)-12-dUTP (NEN Life Science Products) Genomic DNA was digested to fragment lengths of 0.3–2 kb Labeled patient DNA (800 ng) and reference DNA (800 ng) together with excess unlabeled Cot-1 DNA (Life Tech-nologies Ltd., Gibco BRL, Paisley, UK) were hybridized
to normal lymphocyte metaphase chromosomes Slides were counterstained with 4,6-diamidine-2-phenylindole
in an anti-fade solution (Vector Laboratories, Burlin-game, CA) Each sample was hybridized twice, with both sex-matched and mismatched reference DNA
Digital Image Analysis The hybridizations were analyzed using the CytoVi-sion System with the VerCytoVi-sion 2.7 High Resolution CGH analysis software (Applied Imaging, Newcastle, UK) Fifteen to 20 metaphases were collected using a Nikon Eclipse E800 epifluorescence microscope mounted with
a CCD camera interfaced to a CytoVision Station The green (patient DNA) to red (reference DNA) fluorescence ratio along the length of the chromosomes was calcu-lated The CGH profiles were compared to a dynamic standard reference interval based on an average of normal cases, as described by Kirchhoff et al [1998] The dynamic standard reference intervals are wide at regions known to produce unreliable CGH profiles The intervals were scaled automatically to fit the test case The mean ratio profile of each case with 99.5% con-fidence intervals was compared to the average ratio profile of the normal cases with similar confidence intervals To increase sensitivity in cases of mosaicism, 95% confidence intervals were initially used Positive findings were those where the confidence intervals of the patient profile and normal averaged profile did not overlap
FISH Whole chromosome painting probes were from Vysis Ltd (Downers Grove, IL) or Oncor (Appligene Oncor, Illeirch, France), whereas the telomer probes were from Cytocell Ltd (Oxfordshire, U.K.) BAC probes were selected from a human genomic library at the Julie R Korenberg Lab at Cedars-Sinai Medical Center, and were delivered by Research Genetics (Paisley, U.K.) Metaphase slides were postfixed in 1% formaldehyde and dehydrated in an ethanol series (70%, 85%, and 100%) before denaturation followed by hybridization FISH with whole chromosome painting probes or telomer probes were performed according to the manu-facturer’s protocol The BAC clone was labeled with biotin-16-dUTP (Roche Diagnostics GmbH, Mannheim, Germany) by standard nick translation The probe was resuspended in 10 mL hybridization buffer (50% formamide, 2xSSC, 10% dextran sulfate) After
126 Ness et al
Trang 3preannealing with Cot-1 and salmon sperm DNA, the
probes were hybridized overnight at 378C in a HYBrite
machine (Vysis) Hybridization was detected by
FITC-anti DIG (Appligene Oncor) Chromosomes were
coun-terstained with 4,6-diamidine-2-phenylindole in an
antifade solution (Vector Laboratories)
RESULTS The clinical features of patients, whose exact
karyo-types were determined by CGH analysis, are listed in
Table I The initial karyotypes after G-banding, CGH
results and final karyotypes are shown in Table II In
all cases the CGH findings were confirmed, either by
FISH (Cases 1, 4, 6–14) or careful retrospective analysis
of G-banded chromosomes (Cases 2, 3, 5, 15; see Figs 1
and 4–6, and Table II)
Cases with Normal Karyotypes After Routine Chromosome Analysis Cases 1–5 had negative (normal) results on routine chromosome analyses, but were selected for CGH be-cause of clinical features suggesting a chromoso-mal aberration (Case 1–3), our special interest in autism (Case 4), or for further investigation of what we considered to be a variant chromosome 15 satellite (Case 5)
In Case 1, we found an unbalanced terminal 10;20-translocation (Fig 1) We still do not know if one of the parents carries a translocation [approximately 50% do; see Knight et al., 1999] The translocation was impos-sible to see on G-banded chromosomes (analyzed in 1995 and 1997), even when knowing what to look for (Fig 1) Whole chromosome paint probes for chromosome 10 and
TABLE I Clinical Features of Cases With Genomic Imbalances Found by CGH Analysis Case no Sex/YoB Genomic imbalance Available information on patient phenotype
1 F/1995 del 10p15
dup 20q13.3
Short stature, mental retardation, cri de chat–like cry, dysmorphic features (hypertelorism, epicanthus, hypognatia)
2 F/1988 del 6q14-6q15 Short stature, severe mental retardation with autistic-like features, hypotonia,
atrial septal defect, dysmorphic features (frontal bossing, protruding jaw, hypoplastic midface, deep-seated eyes, small mouth)
3 F/1990 del 4q21.1-4q21.3 Short stature, severe mental retardation, epilepsy (infantile spasms), atrial septal
defect, dysmorphic features (frontal bossing, synophrys, mongoloid slant, short philtrum, depressed angles of the mouth, small hands and feet)
4 M/1990 dup 15q11-15q13 Mental retardation with autistic-like features
5 M/1992 del 6p12 Attention deficit disorder (without hyperactivity), subnormal auditive perception
with delayed language development, motorically clumssy
6 M/1987 mosaic dup
4p12-4p13
Attention deficit and hyperactivity disorder, poor social adaptation, rigid personality, delayed development of language and motoric skills, bilateral cryptorchism, congenital hip dysplasia, unequal lower extremity length (3 cm difference), scoliosis
7 F/1984 mosaic dup
8p11-8p12
Short stature, mental retardation, motorically clumsy
8 F/2001 dup 2p16-2p21 Delayed development of motoric functions, no obvious mental retardation
(at age 4 months), dysmorphic features (anteverted nostrils, frontal bossing, small hands)
9 M/1987 del 3p25-3pter
dup 3q26-3qter
Moderate mental retardation, minimal dysmorphic features (slight synophrys, anteverted nostrils), normal stature, chronic obstipation
10 M/1998 dup 6q24-6qter Short stature, neonatal hypotonia, moderate mental retardation, neonatal transient
hyperglycemia, prominent joint contractures, left-sided club foot, left-sided inguinal hernia, many dysmorphic features (telecanthus with prominent eyes, small mandible, low-set ears, webbed and short neck)
11 M/2000 dup 9p11-9pter Neonatal hypotonia with feeding problems, atrial septal defect, mild dysmorphic
features (transverse palmar crease, small mandible, five-finger clinodactyly, broad-based bulbous nose, antimongoloid eye slant, low-set cupped ears), still no obvious mental retardation (age 5 months)
12 M/2000 del 5p14-5pter
dup 1q42-1qter
Multiple congenital abnormalities, severe respiratory problems requiring assisted ventilation, dysmorphic features (cleft lip/palate, preauricular tags, increased intermammary distance), early neonatal death
13 M/2000 del 11q24-11qter
dup 4q26-4qter
Hypoplastic left ventricle, dysmorphic features, bilateral cryptorchism, curved penis, early neonatal death
14 F/1996 del 10q21
del 11q22
Short stature, mental retardation with moderately delayed psychomotor and language development, bilateral club feet, congenital hip dislocations, some dysmorphic features (mongoloid eye slant, deep set eye, posteriorly rotated low-set ears)
15 F/1987 mosaic del
21q22.2-21qter
Mild mental retardation with normal motor but late language development, bilateral congenital hip dysplasia, club feet and inguinal hernia, epilepsy (onset 12 years), dysmorphic features (short philtrum, short and broad neck, very curly hair)
YoB, year of birth.
Trang 420 used retrospectively were also unable to detect any
abnormality (data not shown) We have found no reports
on the phenotype of small terminal 10p deletions, but it
appears that the phenotype is quite unlike the DiGeorge
syndrome-like phenotype of more proximal 10p
dele-tions [Daw et al., 1996] Likewise, we have not found
any published cases with similar terminal duplications
of 20q
In Case 2, it was fairly easy to confirm the aberration
on G-banded chromosomes when they were examined
retrospectively for the deletion (Fig 1) The aberration
had previously been missed twice by routine analysis of
G-banded chromosomes This region of chromosome 6
has a banding pattern that easily obscures the presence
of abnormalities Only 12 children with proximal
interstitial 6q deletions have so far been described
[Roland et al., 1993; Kumar et al., 1997; Passarge, 2000]
The phenotype of our case (Table I, Fig 2), a severely
retarded 13-year-old girl, was only slightly reminiscent
of the previously published proximal interstitial 6q
deletion syndrome [Kumar et al., 1997] Our patient lacked the dolichocephaly (she is in fact brachycephalic) and long philtrum (her philtrum is short) often found in this syndrome More in line with previously described stigmata was neonatal hypotonia, a cardiac abnormality (atrial septal defect), large ears, short nose with a broad nasal tip, and a thin upper lip (Fig 2) [Kumar et al., 1997]
In Case 3, the aberration had been missed twice by routine G-banding Even after informing the cytogenetic technicians that a deletion was present, it was still very hard to detect on high-quality G-banded chromosomes (Fig 1) This case has been through many rounds of evaluation by well-known experts on dysmorphology Brachmann-de Lange syndrome was initially suggested because of the presence of synophrys, increased body hair, and severe mental retardation, but later was discarded The patient phenotype was slightly reminis-cent of the previously described deletion 4q21/4q22-syndrome with frontal bossing, hypotonia, and short
TABLE II Initial Karyotype, CGH Results, Confirmatory Studies, and Final Adjusted Karyotype
Case no.
Karyotype after initial G-banding
CGH result:
.rev isha
CGH result confirmation
Karyotype after CGH analysis and follow-up studiesb,c
enh(20q13.3) tel20px3 (tel10p-,tel20px3)
2 46,XX dim(6)(q14q15) G-banding 46,XX,del(6)(q14q15) de novo
6 47,XY,þmar[19]/46,XY[11] enh(4)(p12p13)d wcp4þ 47,XY,þder(4)del(4)(p13)del(4)(q10)[19]/
46,XY[11]
7 47,XX,þmar[18]/46,XX[12] enh(8)(p11p12)d cen8þ 47,XX,þder(8)del(8)(p12)del(8)(q10)[18]/
46,XX[12]
8 46,XX,dup(2)(p?) enh(2)(p16p21) G-banding wcp2þ 46,XX,dup(2)(p21p16)
9 46,XY,add(3)(p25) enh(3)(q26qter) wcp3þ 46,XY,der(3)t(3;3)(p25;q26.2)del(3)(p25)
dim(3)(p25pter) tel3qx3 dup(3)(q26.2qter)
10 46,XY,add(6)(q25) enh(6)(q24qter) wcp6þ 46,XY,dup(6)(q27q24.3)
G-banding
11 46,XY,add(6)(q25) enh(9)(p11pter) tel9px3 46,XY,der(6)t(6;9)(q27;p12)
tel6q-12 46,XY,der(5)t(5;?)(p14;?) enh(1)(q42qter) wcp5- 46,XX,der(5)t(1;5)(q42;p14)
dim(5)(p14pter) tel1qx3
13 46,XY,der(11)t(11;?)(q24;?) enh(4)(q26qter) wcp4þ 46,XY,der(11)t(4;11)(q26;q24)
dim(11)(q24qter) d
wcp11-14 46,XX,t(10;11)(q21;q22) dim(10q21) d BAC11q22.3- 46,XX,t(10;11)(q21;q22)del(10)(q21q21)
15 46,XX,r(21)[17]/46,XX[13] dim(21q22) d G-banding 46,XX,r(21)(p11?q22.1)([17]/46,XX[13]
a rev ish, reverse in situ hybridization; dim, diminished (signal intensity), enh, enhanced.
b Nomenclature follows the 1995 revision of the International System for human Cytogenetic Nomenclature.
c For simplicity, the ish karyotype is only included if an aberration is impossible to detect, also in retrospect, on high-quality G-banded chromosomes.
d Not observed using fixed threshold 0.8–1.2.
Fig 1 CGH and follow-up studies for cases where no aberration was
initially found on the G-banded karyotype (Cases 1–5, see Table I) Case 1:
CGH indicating a deletion of terminal 10p and duplication of terminal 20q.
FISH analysis with subtelomeric probes confirmed the suspected presence of
an unbalanced 10;20 translocation (two normal chromosome 20, one normal
chromosome 10 and one der(10) with terminal 20q replacing terminal 10p).
Case 2: CGH showing a deletion on 6q, confirmed by G-banding analysis.
Case 3: CGH found a 4q21 deletion, confirmed by G-banding analysis Case
4: CGH indicated a duplication of proximal 15q to be present This could not
be seen on G-banded chromosomes, but FISH with the SNRPN probe
confirmed the duplication Case 5: The observed deletion of 6p12 in the CGH
profiles could retrospectively also be seen on long G-banded chromosomes.
General comment to Figures 1 and 4–6: In each case, the CGH profiles
(with a red arrow pointing to the aberration if visible), and FISH pictures (if any) on the right All the CGH profiles, except in Cases 6 and 7 (99%), are shown with confidence intervals of 99.5% The CGH profiles include ideograms of the involved chromosomes indicating the gains (green bar) or losses (red bar) detected by CGH On the CGH plots the average (pink lines), the 99.5% confidence intervals (yellow lines) and the 99.5% standard reference intervals (black lines) can be seen An aberration is recorded when the intervals are nonoverlapping n indicates the number of chromosomes analyzed The FISH-probes used are denoted wcp for whole chromosome paints, tel for telomeric probes, and cen for centromeric probe The name of the probe is written with the same color as the probes FISH signal [Color figure can be viewed in the online issue, which is available at www.inter-science.wiley.com.]
128 Ness et al
Trang 5hands and feet (between the 10th and 25th centile), but
she did not have the macrocephaly reported to be
common, and her facial appearance was quite different
(Fig 3) [Nowaczyk et al., 1997] Her head circumference
was between the 50th and 75th centile, her length a few
centimeters below the 2.5 centile
Case 4 was screened for a proximal duplication on chromosome 15 because several reports had linked autistic-like features to maternal duplications or tripli-cations of the Prader-Willi/Angelman critical region [e.g., Cook et al., 1997], sometimes as a part of the so-called inv dup(15)-syndrome [Battaglia et al., 1997]
Trang 6CGH analysis detected a duplication of band 15q12
(Fig 1) This duplication was not evident on
good-quality G-banded chromosomes, although a few
chromo-some pairs appeared suspicious FISH with a probe for
the SNRPN locus showed that this region indeed was
duplicated (Fig 1)
Case 5 was selected for CGH in order to analyze a large
and strange-looking satellite on 15p, which later turned
out to be an unusual normal variant that the boy had
inherited from his mother However, we discovered an
unexpected deletion of band 6p12.1 (Fig 1) The patient’s
phenotype was quite mild (some learning difficulties,
ADHD-like behavior, motorically clumsy) The presence
of a small deletion could be confirmed by careful analysis
of good quality metaphases None of his parents had a
deletion of the same band Further analyses using BAC
and YAC probes are in progress in order to define this
deletion better To our knowledge, a similar case has not
previously been published
Cases with Marker Chromosome Mosaicisms
In Case 6, a small marker chromosome was observed
in 60% of lymphocyte metaphases By using a confidence interval of 99% (and not 99.5%) to increase sensitivity, the marker could be identified by CGH as a chromo-some 4 derivative, containing 4p12-4p13 DNA (Fig 4)
A painting probe as well as a centromer probe for chro-mosome 4 confirmed this finding (Fig 4, Table II)
In Case 7 too, it was necessary to compare profiles with less stringent confidence intervals (99% instead of 99.5%) in order to identify the marker, which also in this case was present in 60% of the lymphocyte metaphases The marker was a der(8), containing 8p11-8p12 derived DNA The finding was confirmed by a FISH centromer 8 probe (Fig 4)
Cases with Additional Chromosome Material of Unknown Origin
In Case 8, it was obvious that additional material was present on 2p FISH paint 2 suggested a duplication, but
it was not possible to determine which part of chromo-some 2 was duplicated The clinical features at age
4 months appeared to be mild (not investigated by us), and of little help in defining the exact duplication be-cause of the lack of dysmorphic handles pointing to a specific region CGH identified the duplicated segment
as 2p16-2p21, and the banding pattern was most con-sistent with an inverse duplication (Fig 5, Table II) Approximately 20 patients with internal duplications in 2p have so far been described [Aviram-Goldring et al., 2000], but none of these patients were reported to have a duplication similar to our patient’s
The difficulty in defining the exact origin of a suspected duplication is also well illustrated by Case 9
In this case, the cytogenetic laboratory initially (in 1995) interpreted the G-banded karyotype as an inverse ter-minal duplication of 3p (Fig 5) Only recently the family was referred for genetic counseling, and CGH was done
to confirm the presence of a distal partial trisomy 3p
Fig 2 Case 2, girl born 1988, at age 12.
130 Ness et al
Trang 7Such patients can have psychomotor retardation
with-out dysmorphic features [Smeets et al., 2001], well in
line with our patient’s phenotype (Table I) However,
instead of a 3p duplication, CGH revealed a duplication
of terminal 3q combined with a small deletion of
ter-minal 3p (Fig 5, Table II) The CGH analysis was
confirmed by subtelomer 3q FISH (Fig 5) It thus turned
out that the patient had a duplication of the proposed
critical region of the dup(3q)-syndrome [Aqua et al.,
1995], which superficially resembles Brachmann-de
Lange syndrome His phenotype was consistent with a
mild variant of the dup(3q)-syndrome: moderate mental
retardation, slight synophrys, and anteverted nares,
but no hirsutism or malformations of internal organs
[Rizzu et al., 1997]
Case 10 is one of the few examples where the patient’s
features are highly suggestive of a specific chromosome
abnormality (Table I) The transient neonatal
hypergly-cemia combined with dysmorphic features indicated in
itself of a paternal duplication of terminal 6q [Cave et al.,
2000], and the boy’s joint contractures were typical
for patients with such duplications [Schinzel, 1983]
As a newborn, he appeared severely mentally retarded,
and he was expected to live a short life He was
resus-citated once, at age 4 months With time, however, he
improved, and he now seems to have a moderate mental
retardation Routine chromosome analysis showed extra
material on the terminal part of chromosome 6q,
identi-fied as a terminal duplication of 6q25-qter by CGH This
finding was confirmed by FISH paint chromosome 6 (Fig 5)
Case 11 appeared to have a G-banded karyotype identical to Case 10 (Fig 5), but the clinical features were completely different (Table 1) CGH identified the material to correspond to almost the whole of 9p, which was confirmed by subtelomer 9p FISH (Fig 5) Because this is a 6;9 translocation, a small deletion of terminal 6q was also suspected, and this was confirmed by 6q subtelomer FISH (Fig 5) The deletion was not detected
by the CGH analysis software, indicating that it was smaller than 3 Mb Retrospectively, the patient features were recognized as typical for a dup(9p)-syndrome: downslanting eyes, broad-based nose with globous tip, asymmetric upper lip, cup-shaped ears, and no congeni-tal malformations [Schinzel, 1983]
In Cases 12 and 13, CGH was used to classify a sus-pected unbalanced translocation Both cases had lethal congenital malformations and obvious dysmorphic features (Table I) In Case 12, chromosome 5 had an aberrant banding pattern on the p-arm (Fig 5) CGH showed that almost the entire 5p was deleted and replaced by a duplicated terminal 1q, thus revealing
an unbalanced 1;5 translocation The finding was con-firmed by FISH subtelomer 1q and FISH paint chro-mosome 5 (Fig 5) The 5q deletion and terminal 1q duplication are by themselves reported compatible with life [Schinzel, 1983], but their combination is unlikely
to survive until birth [Gardner and Sutherland, 1996],
as in this case In Case 13, the additional material on chromosome 4 was identified as the terminal half of 4q, revealing an unbalanced 4;11 translocation (Fig 5) Again, each abnormality has previously been reported
to be compatible with life [Schinzel, 1983], but the combination is expected to be lethal, even though this case shows that survival until birth is possible
Cases with Chromosomal Breaks Possibly Associated with Deletions Case 14 had a de novo apparently balanced 10;11 translocation (Fig 6), but because the patient’s diverse phenotypic features were more compatible with a dele-tion than a break in one or two critical genes, we used CGH to investigate whether a deletion could be found This was indeed the case, and CGH suggested a deletion
in both breakpoints (10q21 and 11q22) Only the dele-tion corresponding to band 11q22 has so far been verified
by FISH, using a BAC probe recognizing 11q22.3 (Fig 6)
A few children with deletions of 10q21 and 11q14-11q22 have previously been described [Wakazono et al., 1992; Doheny et al., 1997], but in these cases the deletions were larger The shape of the nose of our patient resem-bles the small nose with broad nasal root and tendency
to anteverted nostrils shown in a patient with del(10) (q21.2q22.1) [Wakazono et al., 1992], but otherwise the phenotype is quite different [Doheny et al., 1997] Case 15 had mosaicism for a ring chromosome 21 (Fig 6), and we wanted to test the usefulness of CGH
in finding the extent of a deletion in such a case, with a marker being present in 57% of lymphocyte metaphases CGH picked up the deletion even at high stringency
Fig 4 CGH and supplementary FISH studies for cases with marker
chromosomes (Cases 6 and 7, see Table I) Case 6: CGH identified the 60%
mosaic marker as a der(4) with the 4p12-p13 region, confirmed by cen4
FISH Case 7: Here the 60% mosaic marker was from 8p11-p12, and a cen8
FISH confirmed chromosome 8 origin Mar, marker chromosome, also called
ESAC or SMC For further explanations, see legend to Figure 1 [Color figure
can be viewed in the online issue, which is available at www.interscience.
wiley.com.]
Trang 8Fig 5 CGH and FISH confirmatory studies for cases with an aberration
of unknown origin found by G-banding (Cases 8–13, see Table I) Case 8:
CGH indicated that an elongated 2p contained a duplication of 2p16-p21,
consistent with the wcp2 FISH results Case 9: CGH showed a duplication of
terminal 3q and deletion of terminal 3p The extra material on terminal 3p
(arrow) is thus a duplicated 3q26-qter region, replacing terminal 3p
sequence as in an unbalanced interchromosomal 3q;3p translocation, as
also shown by tel3q FISH Case 10: The addition to 6q (arrow) represented an
inverse terminal 6q duplication, consistent with wcp6 FISH Case 11: Here,
CGH showed the addition to 6q to be 9p11-pter, confirmed by tel9p FISH.
der(6) having most of 9p replacing the very distal part of 6q The latter deletion was not revealed by CGH Case 12: CGH showed an aberrant 5p (arrow) to result from an unbalanced 1;5 translocation, having a der(5) with
a large part of 5p deleted and terminal 1q duplicated, confirmed by wcp5 and tel1 FISH Case 13: CGH identified an aberrant chromosome 11 (arrow) to result from an unbalanced 4;11 translocation, having a der(11) with terminal 11q deleted and half of 4q duplicated, confirmed by wcp4 and wcp11 FISH For further explanations, see legend to Figure 1 [Color figure can be viewed in the online issue, which is available at www.interscience.-wiley.com.]
132 Ness et al
Trang 9testing (99.5% confidence intervals), indicating that
the percentage of metaphases containing a ring
chromo-some might be biased because of poorer growth of
r(21)-containing cells Ring chromosome 21 gives a highly
variable phenotype (from normal development to severe
mental retardation with multiple congenital
malforma-tions), dependent on the presence and extent of
dele-tions and sometimes even duplicadele-tions [Schinzel, 1983;
McGinniss et al., 1992]
DISCUSSION
In our cytogenetic laboratory, CGH has become an
important supplement to the routine diagnostic
proce-dures for two main reasons First, CGH is a more
infor-mative and often faster way of identifying chromosome
pieces of unknown origin (additions, marker
chromo-somes) than targeted FISH investigations [Erdel et al.,
1997; Boceno et al., 1998; Levy et al., 1998; Breen et al.,
1999; Rigola et al., 2001] The guesswork concerning
which FISH probes to use is eliminated (e.g., Cases 6–7,
11–13), a single experiment (and not consecutive
hybri-dizations) returns an answer, and not only the
chromo-some but also the chromosomal region, domain, and
sometimes even the subdomain is identified (e.g., Cases
6–11) In addition, one gets a clue for predicting if a
marker chromosome is likely to cause a phenotypic
abnormality or not It is often difficult to determine if a
small G-banded marker chromosome contains
gene-containing euchromatin or only heterochromatin with
structural/repetitive DNA (e.g., Cases 6 and 7) Because this latter type of DNA is excluded from the CGH analysis (blocked by Cot-1 DNA), a positive CGH result indicates that genes, with a potentially dose-dependent influence on normal embryonic and brain development, might be present in the extra chromosomal material The novel CGH software is also sufficiently sensitive to
be used for identification of mosaicisms, at least when the chromosomal abnormality (usually a marker chro-mosome) is present in more than 50% of the metaphases (Cases 6 and 7) Cases 6, 7, and 15 all had marker chromosomes that could be characterized only by the high-resolution CGH analysis and not by applying fixed thresholds on the CGH profiles
The second important point is that the high-resolution CGH software with its two- to threefold improved sen-sitivity [Kirchhoff et al., 1998, 1999], has made CGH
a method of choice for finding cryptic chromosomal aberrations [Cases 1–5 and 14–15) [Kirchhoff et al.,
2000, 2001] Alternatively, screening for cryptic aberra-tions can be done with subtelomeric probe kits [Knight
et al., 1999], multicolor FISH techniques involving differential chromosome painting (M-FISH) [Uhrig
et al., 1999; Jalal et al., 2001], spectral karyotyping [Schrock et al., 1997; Haddad et al., 1998], or multi-subtelomer FISH [Brown et al., 2001] Multicolor FISH techniques are useful for deciphering complex chromo-somal rearrangements difficult to resolve by G-banding [Schrock et al., 1997; Eils et al., 1998; Haddad et al., 1998; Uhrig et al., 1999; Jalal et al., 2001], and may
Fig 6 CGH and the FISH follow-up studies for cases with chromosomal breaks (Cases 14 and 15, see Table I) Case 14: CGH revealed deletions in both breakpoints (arrows) in an apparently balanced 10;11 translocation (G-banding) FISH with a BAC probe toward 11q22.3 confirmed the deletion on chromosome 11 Case 15: CGH identified a deletion of terminal 21q in a ring 21 chromosome (arrow) For further explanations, see legend to Figure 1 [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]
Trang 10also detect subtle interchromosomal rearrangements
[Azofeifa et al., 2000; Bezrookove et al., 2000], but for
several reasons (workload, complexity, cost, level of
sensitivity, methodological problems) the method is not
well suited when screening for cryptic aberrations
[Lee et al., 2001] We have not applied M-FISH on our
cases, but noticed that whole chromosome painting
probes used retrospectively did not reveal the
unba-lanced translocation in Case 1 Probe sets for subtelomer
FISH screening appears to be more useful and has been
claimed to find aberrations in 7.4% of a selected group of
patients (children having moderate to severe mental
retardation, congenital malformations, and dysmorphic
features) [Knight et al., 1999] The technique is more
sensitive than CGH for small (<3 Mb) subtelomeric
deletions, but usually less sensitive for duplications
Our Case 11 illustrates the limited sensitivity of CGH
for small subtelomeric rearrangements Like CGH,
sub-telomer FISH screening is quite laborious [Knight et al.,
1999; Kleefstra et al., 2000; Brown et al., 2001; de Vries
et al., 2001], and typically the work involved with
scre-ening one case takes a cytogenetic technician 5–7 hours
[Kjetil Solland, personal communication]
Compared to subtelomer FISH screening, the greatest
advantage of CGH is that imbalances other than those of
the subtelomeric regions are detected It appears that
intrachromosomal deletions or duplications are a more
common finding than unbalanced subtelomeric
rear-rangements in this group of patients without a family
history of mental retardation [Kirchhoff et al., 2000]
This is also the case in our group of patients Only one of
the cryptic aberrations we have found (Case 1) would
have been detected by subtelomer FISH screening In
the evaluation of CGH results, it should be borne in mind
that CGH is expected to be somewhat more sensitive for
finding deletions (1:2 comparisons) than duplications
(2:3 comparisons)
The terminal deletions of chromosome 11q (Case 13)
and 21q (Case 15), as well as the
translocation-associated deletions in domains 10q21 and 11q22 in
Case 14, were detected only by high-resolution CGH
analysis (Table II) The laboratory that developed the
high-resolution CGH software has recently published
the detection of imbalances in seven cases with
pre-viously normal karyotypes [Kirchhoff et al., 2000] To
our knowledge, our report is the first study confirming
the usefulness of CGH for finding such cryptic
abnorm-alities We have so far studied 50 cases with apparently
normal karyotypes, and in 5 cases (10%) we identified
an imbalance Kirchhoff et al found 20 abnormalities
among 165 cases with apparently normal karyotypes,
i.e., a cryptic imbalance was found in 12% of their cases
[Kirchhoff et al., 2001] If it eventually turns out that
imbalances can be detected in approximately 10% of
cases suggestive of a chromosome abnormality, CGH
has a diagnostic yield that is higher than for subtelomer
FISH screening [Knight et al., 1999], at least for patients
having no close relatives with mental retardation [de
Vries et al., 2001]
Finally, there are a few points that deserve to be
discussed concerning the phenotype of our cases First,
the phenotype caused by a deletion or duplication may
be very mild and not very indicative of a chromosome abnormality It is quite likely that small duplications
or deletions in certain domains might pass unnoticed, the phenotype being within normal limits Even larger deletions have been reported to have no observable phenotypic effect if the deletion occurs in certain dark G-banded domains [Gardner and Sutherland, 1996]
An example of the latter is a 15-year-old girl with a del(4)(q12q21.1) of at least 15 Mb, having no mental retardation (she learned to read when she was 5), that were karyotyped only because she had moderate growth retardation (G Houge, unpublished observation) On this background, we do not think that our Case 5 is a rarity It just represents a case that is unlikely to be ascertained as a chromosomal abnormality Second, it is usually very difficult to suspect the presence of a certain chromosomal aberration from the patient’s phenotype (e.g., most of our cases), but exceptions do exist (e.g., our Case 10) Case 2, which has a deletion in the critical region for the so-called proximal 6q deletion syndrome, and Case 3, with a deletion in the 4q21/4q22-syndrome region, have both phenotypes that are quite different from the published syndrome descriptions [Kumar et al., 1997; Nowaczyk et al., 1997] It should be noticed that our Case 3 indicates that a deletion in band 4q21 may give a Cornelia de Lange–like syndrome (Table I, Fig 3) Third, a de novo translocation may be associated with cryptic deletions (Case 14) Such deletions are more likely to cause an abnormal phenotype than a single gene disruption Imbalances of this kind may be found
by mapping with YAC or BAC probes, or by searching for hemizygosity for microsatellite markers in the region, but these are not routine procedures Thus, it
is probably more sensible to do a CGH analysis before undertaking to clone a gene suspected to be destroyed by
a breakpoint, than to start with laborious mapping using domain-specific YAC probes [Kumar et al., 1998; Wirth
et al., 1999]
Several of the cases in this study have been through repeated rounds of standard chromosome analysis only with negative findings We have shown that high-resolution CGH analysis is a reliable and relatively quick way to identify aberrations that either would not have been detected by conventional G-banding, or that would have been difficult to unravel by conventional FISH It should be considered to establish CGH as a routine analysis for screening patients with unexpla-ined mental retardation and normal karyotypes after G/R-banding In the future, a similar but more sensitive CGH-based method is expected to be established on DNA matrices (microchips), where BAC clones can cover the hole genome with less than 1-Mb intervals
ACKNOWLEDGMENTS This work had been impossible without the aid of clinicians supplying us with patient samples and phenotypic information We are especially grateful to
Dr J Høgseth (Case 1), Dr G Aas Herder (Case 2),
Dr M Mork (Case 3), Dr N Breivik (Case 4), Dr E Czarnecke (Case 5), Dr N Mjellem and Dr L Bindoff (Case 6), Dr V Lu¨tcherath, and Dr N Øyen (Case 7),
134 Ness et al