Full text of DOI:10.1038/labinvest.3780427

0
023-6837/02/8203-335$03.00/0
LABORATORY INVESTIGATION
Vol. 82, No. 3, p. 335, 2002
Copyright © 2002 by The United States and Canadian Academy of Pathology, Inc.
Printed in U.S.A.
Infantile Dilated X-Linked Cardiomyopathy, G4.5
Mutations, Altered Lipids, and Ultrastructural
Malformations of Mitochondria in Heart, Liver, and
Skeletal Muscle
John J. Bissler, Monica Tsoras, Harald H. H. G o¨ ring, Peter Hug, Gail Chuck,
Esther Tombragel, Catherine McGraw, James Schlotman, Michael A. Ralston, and
George Hug
Department of Pediatrics (JJB, MT, GC, ET, CM, JS, GH), University of Cincinnati, Cincinnati, and Children’s
Medical Center (MAR), Dayton, Ohio; Department of Genetics (HHHG), Southwest Foundation for Biomedical
Research, San Antonio, Texas; and Section on Membrane Structure and Function (PH), Laboratory of Experimental
and Computational Biology, NCI-FCRDC, Frederick, Maryland
SUMMARY: Mutations in the Xq28 gene G4.5 lead to dilated cardiomyopathy (DCM). Differential splicing of G4.5 results in a
family of proteins called “tafazzins” with homology to acyltransferases. These enzymes assemble fatty acids into membrane
lipids. We sequenced G4.5 in two kindreds with X-linked DCM and in two unrelated men, one with idiopathic DCM and the other
with DCM of arrhythmogenic right ventricular dysplasia. We examined the ultrastructure of heart, liver, and muscle biopsy
specimens in these three DCM types; we used gas chromatography to compare fatty acid composition in heart, liver, and muscle
autopsy specimens of two patients of kindred 1 with that of controls. In X-linked DCM, G4.5 had a stop codon (E188X), a
nonsense mutation, in kindred 1 and an amino acid substitution (G240R), a missense mutation, in kindred 2. In the two men with
isolated DCM, G4.5 was not mutated. Ultrastructural mitochondrial malformations were present in the biopsy tissues of the
patients with DCM. Cardiac biopsy specimens of both kindreds with X-linked DCM exhibited greatly enlarged mitochondria with
large bundles of stacked, compacted, disarrayed cristae that differed from those of the two types of isolated DCM. Autopsy tissue
of patients with X-linked DCM had decreased unsaturated and increased saturated fatty acid concentrations. Seven of 13
published G4.5 missense mutations, including the one presented here, occur in acyltransferase motifs. Impaired acyltransferase
function could result in increased fatty acid saturation that would decrease membrane fluidity. Mitochondrial membrane
proliferation may be an attempt to compensate for impaired function of acyltransferase. Cardiac ultrastructure separates X-linked
DCM with G4.5 mutations from the two types of isolated DCM without G4.5 mutations. Electron microscopy of promptly fixed
myocardial biopsy specimens has a role in defining the differential diagnosis of DCM. Mutational analysis of the G4.5 gene also
serves this purpose. (Lab Invest 2002, 82:335–344).
-
linked dilated cardiomyopathy (DCM) has vari-
retardation, 3-methylglutaconic aciduria, and neu-
tropenia whose initial presentations varied from
DCM to isolated neutropenia. Bolhuis et al (1991)
localized a disease locus to distal Xq28, and Bione
et al (1996) identified mutations in the G4.5 gene.
They found that the gene was alternatively spliced,
resulting in multiple proteins they termed “taf-
azzins.” D’Adamo et al (1997) analyzed G4.5 in 11
families with X-linked DCM and suggested that the
type of mutation influenced the disease phenotype.
Johnston et al (1997) analyzed 19 different X-linked
DCM pedigrees and found no relationship between
type or location of G4.5 mutation and severity of
heart disease, neutropenia, or 3-methylglutaconic
aciduria. Neuwald (1997) pointed out that tafazzins
have acyltransferase homology and suggested that
they may be involved in maintaining mitochondrial
morphology and function. Vreken et al (2000) stud-
ied cultured fibroblasts from patients and concluded
that mitochondrial phospholipids may be affected.
X
able presentations. In 1979, Neustein et al re-
ported abnormal mitochondria in cardiac myocytes
obtained from an infant boy with X-linked DCM who
died of heart failure at age 16 months. Barth et al
(
1983) presented an X-linked DCM with skeletal my-
opathy, neutropenia, abnormal mitochondria in heart,
muscle, and neutrophils, and respiratory chain abnor-
malities of mitochondria in muscle and cultured
fibroblasts (Barth et al, 1996). Kelley et al (1991)
described seven boys with X-linked DCM, growth
Received December 12, 2001.
This work was supported by Grant DK0241801 from the National Insti-
tute of Diabetes and Digestive and Kidney Diseases and by Grant MO1
RR08084 from the National Institutes of Health.
Address reprint requests to: Dr. John J. Bissler, Children’s Hospital Research
Foundation #5, 3333 Burnet Ave, Cincinnati, OH 45229–3039.
E-mail: john.bissler@chmcc.org. Current address for Dr. P. Hug: Master
Chemical Company, 501 W. Boundary, Perrysburg, OH 43551–1200.
Laboratory Investigation • March 2002 • Volume 82 • Number 3 335

Bissler et al
We describe eight boys in two kindreds with
University Medical Center, Pediatric Metabolism Lab-
oratory, Raleigh, North Carolina). These urinary acid
excretions were moderately increased in kindred 1,
Patients IV 1 and 9, but were not increased in kindred
1, Patient IV 2 and kindred 2, Patient V 1. Electron
transport chain complexes I, II, III, and IV in heart
autopsy tissue of kindred 1 Patient IV 10 were normal
as were complexes II, III, and IV in cultured fibroblasts
of kindred 1, Patient IV 9 (Clinical Pharmacology
Laboratory, Department of Veterans Affairs Medical
Center, Cleveland, Ohio).
X-linked DCM, mutations in G4.5, malformed mito-
chondria of heart, liver, and muscle, and unusual fatty
acid composition. The mitochondrial malformations in
these patients differ from those in two unrelated men
with isolated DCM, one with idiopathic DCM (Hug and
Schubert, 1970) and the other with DCM of arrhyth-
mogenic right ventricular dysplasia (Blankenship et al,
1
993). Neither of these two men had a G4.5 mutation.
Results
Kindred 1, Patient IV 8 appeared healthy until age 16
months, when he died after 1 week of an upper
respiratory tract infection without fever or edema. At
autopsy he had cardiomyopathy of unknown etiology.
His brother IV 10 died of DCM at age 23.5 months.
Their cousin IV 1 died at age 4 years after 2 days of a
viral gastroenteritis. He had appeared healthy while
being treated for DCM. Cousin IV 15 died 2 days after
birth with hydrops fetalis. Kindred 1, Patients IV 2 and
9 and kindred 2, Patients V 1 and 3 are being treated
for DCM at ages 10, 16, 9, and 5 years. Currently they
appear clinically healthy and show no symptoms of
cardiac failure.
Patients
DCM was diagnosed in patients based on clinical
examination, chest x-ray, electrocardiogram, echo-
cardiography, and inheritance (Fig. 1, A and B). Kin-
dred 1, Patients IV 1 and 9 had growth retardation
(
Ͻ5% for height); kindred 1, Patient IV 2 and kindred 2,
Patients V 1 and 3 did not. Kindred 1, Patients IV 1, 2,
and 9 and kindred 2, Patients V 1 and 3 could be
examined for symptoms of skeletal myopathy. None
were noted. Neutropenia was not found in any pa-
tients. Four patients were tested for urinary
3-methylglutaconic and 3-methylglutaric acids (Duke
Figure 1.
Eight patients with X-linked dilated cardiomyopathy in two pedigrees. A denotes kindred 1 with the nonsense mutation, and B denotes kindred 2 with the missense
mutation in G4.5.
3
36 Laboratory Investigation • March 2002 • Volume 82 • Number 3

Infantile X-Linked Cardiomyopathy
Figure 2.
Full-length tafazzin protein. Regions that are removed alternatively by differential splicing (exon 5, 6, or 7) are marked as black boxes. Putative acyltransferase motifs
identified by Neuwald (1997) are marked as gray boxes. Missense mutations identified in the literature are labeled by lines extending downward. TM ϭ the putative
transmembrane region of the protein.
Mutations in Gene G4.5
ily members revealed this mutation, and G240 ap-
peared to be conserved even in Drosophila melano-
gaster (AE003821), Caenorhabditis elegans (T28003),
and Arabidopsis thaliana (AC005679) homologues.
Figure 2 is a diagram of the full-length tafazzin
protein. The regions of the alternatively spliced
protein without exons 5, 6, or 7 are marked at the
top of the illustration. These alternative splices
modulate a relatively hydrophobic region of the
protein. Acyltransferase motifs identified by Neu-
wald (1997) are marked as gray boxes and comprise
a quarter of the full-length protein. Of the 13 pub-
We tested whether the cardiomyopathy in kindred 1
cosegregated with the chromosomal region impli-
cated in Barth syndrome, Xq28, by penetrance-
model-based linkage analysis. Assuming an X-linked
recessive disease with full penetrance, and allowing
for mutation to occur at the trait locus, a maximum lod
score of 3.35 at a recombination fraction of 0 was
obtained, providing significant evidence for linkage
between the disease locus and Xq28 (Morton, 1955).
This indicates that the cardiomyopathy segregating in
this pedigree is associated with mutations in the
tafazzin gene implicated in Barth syndrome.
When absence of mutation at the trait locus was
assumed, the maximum lod score was only 0.02,
obtained at recombination fraction 0.32. The prob-
able explanation for the difference in the linkage
signal is that Patient I 1 carries a novel mutation in
the tafazzin gene. It is possible to pinpoint the likely
origin of the mutation in this pedigree, because
Patient I 1 received marker allele 2 from her unaf-
fected father, which is the marker allele that coseg-
regates with the disease in the lower generations of
the pedigree. (Note that the pedigree drawing in Fig.
1
A shows only part of the pedigree used for linkage
analysis. Neither the parents of Patient I 1 nor
observed marker genotypes are indicated.) If no
mutations are allowed for in the analysis, obligatory
recombination events are inferred, thus deflating the
lod score dramatically. Although it is not unex-
pected from population genetics theory to find
evidence for new mutations for X-linked lethal dis-
eases, this pedigree serves as a good example of
how linkage may be easily overlooked if the possi-
bility of mutation is not taken into account.
We sequenced the tafazzin gene cDNA and por-
tions of the genomic DNA. In the first kindred, we
found a transversion G850T that resulted in a 50%
truncation in the predicted protein size (E188X), and
we attribute the disease phenotype to this nonsense
mutation because it always cosegregated with the
disease phenotype. This mutation always cosegre-
gated with a downstream A866T. Both mutations
were found only in patients or carriers of kindred 1.
The second kindred exhibited a single transition
G1006A. This missense mutation resulted in the
amino acid substitution G240R. No unaffected fam-
Figure 3.
Mismatch annealing mutation analysis from kindred 2. The complementary
oligonucleotide (GGCCAAGAATCCAGAAGGCAGC) was used in conjunction
with the oligonucleotide (CACAGAAAATCACTGTGCTCG) for the wild-type allele
and (CACAGAAAATCACTGTGCTGATCA) for the mutant allele. Because these
last two oligonucleotides differ at their 3' nucleotide, the mutant and wild-type
allele can be identified by the production of a PCR product.
Laboratory Investigation • March 2002 • Volume 82 • Number 3 337

Bissler et al
Figure 4.
A, Heart, kindred 2, Patient V 1, bar: 1 ␮m. An aggregate of mitochondria is shown. Many mitochondria are normal (n) in size and structure; others are enlarged
(
e). They contain closely stacked cristae. Intracristal space and matrix space between cristae are narrow and vary in width. Black particles between mitochondria are
normal ␤-glycogen particles in the sarcoplasm.
lished missense mutations, 7 occur in these motifs.
Mutations V183G and G197E have been found re-
peatedly (Bleyl et al, 1997; Cantlay et al, 1999;
D’Adamo et al, 1997; Ichida et al, 2001; Johnston et
al, 1997; Neuwald, 1997). The significant clustering
of over one-half the reported disease-causing mu-
Histologic Examination
Light Microscopy
Liver and muscle did not show diagnostic abnormali-
ties. Cardiac autopsy specimens of kindred 1, Patients
IV 1, 8, and 10 showed myocyte hypertrophy, myofiber
disarray, focal interstitial fibrosis, and mild endocardial
thickening.
tations in putative acyltransferase motifs (p Ͻ 0.022,
2
␹
test) may reveal regions of the tafazzin protein
critical for its function. Mutations were analyzed in
family members using a mismatch amplification
mutation analysis technique. An example for kin-
dred 2 is shown in Figure 3 and can be correlated
with Figure 1B.
Electron Microscopy
Figure 4, A to E, is representative of the findings in the
biopsied tissue of the current patients. Figure 4F is
3
38 Laboratory Investigation • March 2002 • Volume 82 • Number 3

Infantile X-Linked Cardiomyopathy
Figure 4.
B, Heart, kindred 1, Patient IV 10, bar: 1 ␮m. Several normal mitochondria (n) are adjacent to a greatly enlarged mitochondrion with tightly stacked cristae. These
are in disarray. The intracristal space and matrix space between cristae are narrow or absent. Cristae extend into electron-dense material. In gaps of the outer and
inner membranes of the enlarged mitochondrion, one sees ␤-glycogen particles that extend from sarcoplasm to matrix (arrow).
representative of heart findings of the two men with
isolated DCM and normal G4.5 gene. The findings in
these two men have been described previously in
more detail (Blankenship et al, 1993; Hug and Schu-
bert, 1970).
cristae, others of just a few. A minority of mitochondria
were large or very large. Their cristae were closely
packed in straight or circular patterns. These cristae
appeared disarrayed, disorderly, wavy, and contorted
(Fig. 4A).
Heart, current patients. Mitochondria were plentiful
and frequently present in large aggregates between
myofibrils (Fig. 4, A to C). The majority of mitochondria
were normal in size and structure, and contained
normal cristae. Many mitochondria had bundles of
packed cristae. Some bundles consisted of numerous
The largest cardiomyocytic mitochondria (Fig. 4, B
and C) had tubular and tightly packed lamellar cristae.
Intracristal space and matrix space between adjacent
lamellar cristae were narrow and sometimes not de-
tectable. The width of both these spaces varied from
crista to crista and along the course of the same two
Laboratory Investigation • March 2002 • Volume 82 • Number 3 339

Bissler et al
Figure 4.
C, Heart, kindred 2, Patient V 1, bar: 1 ␮m. This large mitochondrion is similar to that in B except one sees no gaps in the outer or inner membrane and no glycogen
particles in the matrix.
adjacent cristae. Some cristae extended into finely
lamellar, feathery or fluffy, electron-dense material.
Some of the largest mitochondria had gaps in the
inner and outer membranes with ␤-glycogen particles
extending from cytoplasm to matrix space.
Liver, current patients. Mitochondria were numerous
and generally not enlarged (Fig. 4D). Some had bun-
dles of cristae in straight, circular, or angular patterns.
Skeletal muscle, current patients. Mitochondria
were numerous, small, and had dense granules in-
creased in number and size (Fig. 4E). Some mitochon-
dria contained a few bundled cristae in straight or
circular patterns.
Heart, man with idiopathic DCM heart. Mitochondria
were plentiful (Fig. 4F). Some were enlarged. The
majority had abundant, closely stacked cristae. Even
in the largest mitochondria these cristae were orderly
and not in disarray. They were positioned parallel to
each other in straight or circular patterns. The width of
the intracristal space was constant, as was that of the
3
40 Laboratory Investigation • March 2002 • Volume 82 • Number 3

Infantile X-Linked Cardiomyopathy
Figure 4.
D, Liver, kindred 1, Patient IV 9, bar: 1 ␮m. Many hepatocytic mitochondria are normal (not shown). These two mitochondria are typical of the changes seen in the
patients. Both mitochondria are normal in size but have straight and circular bundles of closely aligned cristae. The black particles are normal ␣-glycogen particles.
matrix space between adjacent lamellar cristae. The
matrix was often floccular. Cristae were absent in
floccular areas. The findings in Figure 4F are not
distinguishable from those of the man with arrhythmo-
genic right ventricular dysplasia. We did not encounter
any large mitochondria similar to those of the current
patients in these two men with isolated DCM. The
mitochondria of the current patients were separable
from those of the two men. Neutrophils were exam-
ined in one patient from each kindred and appeared
normal.
(Cantlay et al, 1999; D’Adamo et al, 1997; Johnston et
al, 1997) and remains difficult today.
Mitochondrial malformations were present in all
biopsy specimens. The malformations were pro-
nounced in heart but less marked in liver and muscle.
These lesions were similar for the same organ in both
kindred. We conclude that DCM with G4.5 mutations
exhibits mitochondrial malformations that can be used
in defining the differential diagnosis of DCM.
This conclusion is strengthened by the differently
appearing stacked cristae of cardiomyocytic mito-
chondria found in the two unrelated males with iso-
lated DCM. One was a 6-month-old, adopted boy
Fatty Acid Analysis
(
Hug and Schubert, 1970) who 30 years later still had
In Table 1, fatty acid concentrations in autopsy tissues
of heart, liver, and muscle of kindred 1, Patients IV 1
and 10 are listed that differ from those of controls (p Ͻ
idiopathic DCM with malformed mitochondria. The
other man died at the age of 59 years of arrhythmo-
genic right ventricular dysplasia (Blankenship et al,
0.05). Fatty acids of chain length C12 to C22 exhibited
1
993). Both patients’ cardiac mitochondrial malforma-
concentrations that were different in patients and
controls. In the patients, the concentrations of four
unsaturated fatty acids were increased and those of
tions appeared similar and are inseparable from each
other (Blankenship et al, 1993; Hug and Schubert,
1
970; and Fig. 4F), but they are substantially different
and separable from those of the current patients (Fig.
, A to C). We analyzed G4.5 in both men with isolated
17 were decreased; whereas, the concentrations of 13
saturated fatty acids were increased and none were
decreased. In brief, patient tissues contained de-
creased unsaturated and increased saturated fatty
acids.
4
DCM and found no mutations.
Ichida et al (2001) and Bleyl et al (1997) reported
G4.5 mutations in some of their patients with left
ventricular noncompaction (LVNC). One could hy-
pothesize that LVNC patients with G4.5 mutations
would have the mitochondrial changes of our cur-
rent patients, whereas LVNC patients lacking G4.5
mutations would not. In any event, the evaluation of
endomyocardial biopsy specimens of DCM should
include electron microscopy. Useful ultrastructural
information can be obtained when the specimens
are placed in the fixative less than 2 minutes after
the biopsy.
Neuwald et al (1997) identified six regions of the
tafazzin protein with acyltransferase homology. He
proposed that tafazzin proteins might be involved in
phospholipid biosynthesis and remodeling and that
this process, if defective because of mutations in
Discussion
Kindred 1 had a nonsense mutation (E188X) and
kindred 2 a missense mutation (G240R) in G4.5. The
two different mutations produced indistinguishable
mitochondrial malformations. The eight patients were
being treated similarly for X-linked cardiomyopathy.
Four patients who had the nonsense mutation died.
None of our patients had skeletal myopathy or neu-
tropenia. Some patients of both kindred had growth
retardation or increased 3-methylglutaconic aciduria;
others did not. Respiratory chain abnormalities, re-
ported in other patients with X-linked DCM (Barth et al,
1983, 1996), were not found in our patients. Relating
genotype to phenotype has been difficult previously
Laboratory Investigation • March 2002 • Volume 82 • Number 3 341

Bissler et al
Figure 4.
E, Skeletal muscle, kindred 1, Patient IV 10, bar: 1 ␮m. The seven identifiable mitochondria are small; three have bundles of circular, closely aligned cristae and five
have between 5 and 10 prominent dense granules. F, Heart biopsy specimen from the man with idiopathic dilated cardiomyopathy (DCM) (Hug and Schubert, 1970).
Findings do not differ and are not separable from those in the heart of the man with arrhythmogenic right ventricular dysplasia (Blankenship et al, 1993). Bar: 1 ␮m.
Circular and straight bundles of aligned, parallel cristae maintain the width of intracristal space and of matrix space in between adjacent cristae. These changes affect
the majority of mitochondria and are different and separable from those of the current patients.
G4.5, might lead to malformed mitochondrial mem-
branes (Neuwald, 1997). The unusual fatty acid com-
position in autopsy tissues of kindred 1, Patients IV 1
and 10 may be consistent with this proposal. Patient
tissue had more saturated and less unsaturated fatty
acids than did controls (Table 1). In skeletal and
cardiac muscle, a large fraction of esterified fatty acids
is derived from phospholipid. The acyl moieties of
phospholipid molecules determine many of the phys-
ical properties of the membrane. The degree of unsat-
uration is important in determining membrane fluidity,
a measure of the ease of movement of molecules
within the membrane bilayer. Membrane fluidity at a
given temperature is inversely proportional to the
fraction of acyl chain moieties comprised of saturated
fatty acids.
Patients and Methods
Patients
In the current generation of kindred 1 (Fig. 1A), six
boys had DCM (Patients IV 1, 2, 8, 9, 10, and 15); and
of kindred 2 (Fig. 1B), two boys had DCM (Patients V
1 and 3). Tissue specimens of heart (by endomyocar-
dial biopsy), muscle, and liver (by open or needle
biopsy) were obtained for diagnostic studies from
kindred 1, Patients IV 1, 9, and 10, and kindred 2,
Patient V 1. Autopsy tissue (frozen or fixed in formalin
or glutaraldehyde) was available from kindred 1, Pa-
tients IV 1, 8, 10, and 15. DNA isolation from leuko-
cytes and/or tissue specimens was performed in the
eight patients and in potential carrier females.
Cellular membrane fluidities are maintained within a
narrow range by acyltransferase-mediated remodeling
of phospholipid within the membrane. If the tafazzin
gene product is an acyltransferase present within the
mitochondrial membrane, these membranes in af-
fected patients may have reduced fluidity, possibly
reducing the normal function of integral membrane
proteins. The activity of proteins such as lecithin:cho-
lesterol acyltransferase is affected by changes in
membrane fluidity (Parks et al, 2000).
Proliferation of inner mitochondrial membrane may
be in response to reduced mitochondrial unit effi-
ciency. The proliferation may boost mitochondrial
function to near normal. The proliferation could result
in enlarged mitochondria, in overabundant, dissolving
cristae, and in mitochondrial membrane disintegra-
tion. These are the observed mitochondrial changes in
the cardiomyocytes of our patients. With mitochon-
drial function only marginally sufficient, the clinical
outcome may depend in part on environmental factors
such as dietary fatty acids, carnitine (Barth et al, 1983;
Ostman-Smith et al, 1994), or pantothenic acid
Genetic Analysis
Linkage Analysis. Because of similarities of our
patients’ disease with Barth syndrome (Barth et al,
1983), we performed penetrance-model-based link-
age analysis with DXS52, a marker ~1 MB from G4.5,
using the LINKAGE software package (Lathrop et al,
1984). The disease was assumed to be an X-linked
fully penetrant recessive disease without pheno-
copies. The frequency of the disease-causing allele
was set to 0.0001. A mutation rate of 1/3 of this allele
frequency was assumed (Haldane, 1935). Individuals
in whom X-linked cardiomyopathy was suspected but
not confirmed as the cause of death were coded as
having unknown disease status. Published allele fre-
quency estimates were used for genotype marker
DXS52 (Richards et al, 1991).
Molecular Analysis. RNA was isolated from normal
tissue and from patient heart and liver using the
Qiagen RNeasy total RNA kit (Qiagen, Valencia, Cali-
fornia). Reverse transcription was carried out using the
Perkin Elmer Thermostable rTth Reverse Transcrip-
tase RNA PCR kit and primers as described by Bione
et al (1996).
(
Ostman-Smith et al, 1994).
3
42 Laboratory Investigation • March 2002 • Volume 82 • Number 3

Infantile X-Linked Cardiomyopathy
Table 1. Tissue Fatty Acids
Esterified
Control
Free
Chain
length
Lipid name
Left ventricle
⌬
Patient
p Value
Patient
Control
p Value
Stearic
18
20
20
22
22
0
3
5
5
6
20.42 Ϯ 4.80 8.52 Ϯ 7.39
0.00 0.25 Ϯ 0.55
0.76 Ϯ 0.08 0.01 Ϯ 0.04
0.00 0.28 Ϯ 0.32
0.037
0.039
0.001
0.001
0.632
17.98 Ϯ 7.74
0.00
0.84 Ϯ 0.56
0.00
7.09 Ϯ 4.46
0.13 Ϯ 0.24
0.023 Ϯ 0.08
0.24 Ϯ 0.28
1.47 Ϯ 0.63
0.004
0.018
0.289
0.001
0.001
Eicosatrienoic
Eicosapentaenoic
Docosapentaenoic
Docosahexaenoic
Right ventricle
Lauric
Myristic
Palmitic
Stearic
Eicosatrienoic
Docosapentaenoic
Docosahexaenoic
Liver
1.36 Ϯ 0.28 1.74 Ϯ 1.09
0.95 Ϯ 0.01
12
14
16
18
20
22
22
0
0
0
0
3
5
6
4.45 Ϯ 1.04 0.48 Ϯ 1.09
5.30 Ϯ 6.77 2.45 Ϯ 5.94
40.42 Ϯ 8.30 25.83 Ϯ 49.49
24.72 Ϯ 1.28 8.52 Ϯ 7.39
0.001
0.52
0.687
0.006
0.039
0.001
0.900
2.14 Ϯ 0.44
7.40 Ϯ 1.171
46.19 Ϯ 1.61
25.27 Ϯ 2.57
0.00
0.00
0.50 Ϯ 0.68
0.70 Ϯ 1.03
1.81 Ϯ 3.11
16.62 Ϯ 19.05
7.09 Ϯ 4.46
0.13 Ϯ 0.24
0.24 Ϯ 0.28
1.47 Ϯ 0.63
0.064
0.021
0.021
0.001
0.018
0.001
0.044
0.00
0.00
0.25 Ϯ 0.55
0.28 Ϯ 0.32
1.85 Ϯ 0.97 1.74 Ϯ 1.09
Myristic
Palmitic
Stearic
Docosapentaenoic
Oleic
Arachidonic
Muscle
14
16
18
22
18
20
0
0
0
5
1
4
14.01 Ϯ 7.52 3.29 Ϯ 7.25
136.72 Ϯ 17.63 38.49 Ϯ 70.62
45.06 Ϯ 15.15 12.19 Ϯ 13.52
0.024
13.78 Ϯ 4.56
3.17 Ϯ 5.45
0.004
0.001
0.001
0.014
0.025
0.041
0.026 135.51 Ϯ 27.73 30.98 Ϯ 47.89
0.001
0.008
0.139 127.65 Ϯ 47.35 35.48 Ϯ 64.38
0.17
42.42 Ϯ 2.1
0.00
9.94 Ϯ 10.44
0.14 Ϯ 0.26
0.00
0.17 Ϯ 0.28
149.58 Ϯ 58.11 49.58 Ϯ 110.16
20.23 Ϯ 5.96 10.72 Ϯ 11.33
15.46 Ϯ 5.22
7.00 Ϯ 6.52
Linoleic
Linolenic
Eicosadienoic
Eicosatrienoic
18
18
20
20
2
3
2
3
16.6 Ϯ 2.97 42.82 Ϯ 29.16
0.003
0.022
0.027
0.006
10.7 Ϯ 8.10
0.17 Ϯ 0.29
0.31 Ϯ 0.54
0.00
29.65 Ϯ 22.97
1.56 Ϯ 2.06
0.035 Ϯ 0.14
0.15 Ϯ 0.25
0.184
0.021
0.471
0.031
0.00
0.68 Ϯ 0.63 0.09 Ϯ 0.35
0.0 0.43 Ϯ 0.52
1.4 Ϯ 2.14
In patient autopsy tissues, saturated fatty acids are increased and unsaturated fatty acids are decreased. List, by name and number of unsaturated bonds (⌬) in
means Ϯ 1 SD ␮g/mg tissue protein, of the free or esterified tissue fatty acids (FA) of chain length C12 to C22 that differ between patients and controls (p Ͻ 0.05).
Dark boxes list the 10 FA that differ in either the free or esterified form but not in both. Four unsaturated FA are increased and 17 are decreased; 13 saturated FA
are increased and none are decreased.
The sequencing of cDNA from one patient in kindred
yielded a sequence that lacked exon 7 (nucleotides
55–954; GenBank accession number X92764) in con-
with the wild-type oligonucleotide. Male patients with
a mutant allele yielded a product only with the mutant
oligonucleotide, whereas carriers with both alleles
would yield a product with both the mutant and the
wild-type oligonucleotide.
1
6
trast to cDNAs from normal controls. Therefore, se-
quencing of genomic DNA including splice junctions
of exon 7 was performed on patient DNA. Sequencing
cDNA from a patient in kindred 2 was successful in
identifying a missense mutation. Nonaffected family
members were analyzed for this mutation, and a Blast
search analysis (NCBI) was undertaken to determine
the conservation of the amino acid involved in the
missense mutation. Previously described primers
were used in the terminator cycle sequencing as
described in the Perkin-Elmer protocol (no. 402116).
After sequencing identified the mutation sites in the
two kindreds, we used a mismatch annealing mutation
amplification technique to confirm the mutation iden-
tification and to screen other family members (Cha et
al, 1992). Briefly, three PCR primers were used. One of
them was a complementary oligonucleotide, and the
two others were designed so that one contained the
wild-type base and the other contained the mutant
base in the 3' location. These two primers were used
in separate PCR reactions with the complementary
oligonucleotide. Individuals with only the wild type, or
normal, allele yielded an amplification product only
Histologic Examination
Specimens of autopsy and biopsy of heart, liver, and
muscle were embedded for light and electron micro-
scopic evaluation. For the latter we placed biopsy
specimens in glutaraldehyde within 1 to 2 minutes
after removal from the body. This is critical because
delayed fixation or the use of autopsy tissue speci-
mens can result in significant artifacts. Cardiac find-
ings in the current patients were compared with those
in two unrelated men who had isolated DCM without
G4.5 mutations (Blankenship et al, 1993; Hug and
Schubert, 1970).
Fatty Acid Analysis
Free and esterified fatty acids were determined by gas
chromatography and expressed in micrograms per
milligram of tissue protein. The Hewlett Packard GC
system (HP 5890A) had a flame ionization detector
Laboratory Investigation • March 2002 • Volume 82 • Number 3 343

Bissler et al
and an HP Ultra2 (cross-linked 5% Ph Me Silicone
Haldane J (1935). The rate of spontaneous mutations of a
human gene. J Genetics 31:317–326.
25 m ϫ 0.32 mm ϫ 0.25 ␮m film) capillary column
capable of separating the methyl ester derivatives of
Hug G and WK Schubert (1970). Idiopathic cardiomyopathy.
Mitochondrial and cytoplasmic alterations in heart and liver.
Lab Invest 22:541–552.
6
1
2
:0, 8:0, 10:0, 12:0, 14:0, 16:0, 16:1c, 16:1t, 18:0,
8:1c, 18:1t, 18:2, 18:3, 20:0, 20:1, 20:2, 20:3, 20:4,
0:5, 22:0, 22:1, 22:4, 22:5, 22:6, 24:0, 24:1, 26:0
Ichida F, Tsubata S, Bowles KR, Haneda N, Uese K,
Miyawaki T, Dreyer WJ, Messina J, Li H, Bowles NE, and
Towbin JA (2001). Novel Gene Mutations in Patients With Left
Ventricular Noncompaction or Barth Syndrome. Circulation
103:1256–1263.
(
Ackman et al, 1972; Russell et al, 1974). The assay
was calibrated with a methyl ester mixture (GLC-30,
Supelco); methyl heptadecanoate was used as an
internal standard. Results in autopsy tissues of kin-
dred 1, Patients IV 1 and 10 were compared by
independent group t test with those of patients with-
out defective lipid metabolism.
Johnston J, Kelley RI, Feigenbaum A, Cox GF, Iyer GS,
Funanage VL, and Proujansky R (1997). Mutation character-
ization and genotype-phenotype correlation in Barth syn-
drome. Am J Hum Genet 61:1053–1058.
References
Kelley RI, Cheatham JP, Clark BJ, Nigro MA, Powell BR,
Sherwood GW, Sladky JT, and Swisher WP (1991). X-linked
dilated cardiomyopathy with neutropenia, growth retarda-
tion, and 3-methylglutaconic aciduria. J Pediatr 119:738–
747.
Ackman RG, Hooper SN, and Hansen RP (1972). Some
monomethyl-branched fatty acids from ruminant fats: open-
tubular GLC separations and indications of substitution on
even-numbered carbon. Lipids 7:683–691.
Barth PG, Scholte HR, Berden JA, Van der Klei-Van Moorsel
JM, Luyt-Houwen IE, Van’t Veer-Korthof ET, Van der Harten
JJ, and Sobotka-Plojhar MA (1983). An X-linked mitochon-
drial disease affecting cardiac muscle, skeletal muscle and
neutrophil leucocytes. J Neurol Sci 62:327–355.
Lathrop GM, Lalouel JM, Julier C, and Ott J (1984). Strategies
for multilocus linkage analysis in humans. Proc Natl Acad Sci
USA 81:3443–3446.
Morton NE (1955). Sequential tests for the detection of
linkage. Am J Human Genet 7:277–318.
Barth PG, Van den Bogert C, Bolhuis PA, Scholte HR, van
Gennip AH, Schutgens RB, and Ketel AG (1996). X-linked
cardioskeletal myopathy and neutropenia (Barth syndrome):
respiratory-chain abnormalities in cultured fibroblasts. J In-
herit Metab Dis 19:157–160.
Neustein HB, Lurie PR, Dahms B, and Takahashi M (1979).
An X-linked recessive cardiomyopathy with abnormal mito-
chondria. Pediatrics 64:24–29.
Neuwald AF (1997). Barth syndrome may be due to an
acyltransferase deficiency [letter]. Curr Biol 7:R465–R466.
Bione S, D’Adamo P, Maestrini E, Gedeon AK, Bolhuis PA,
and Toniolo D (1996). A novel X-linked gene, G4.5. is respon-
sible for Barth syndrome. Nat Genet 12:385–389.
Ostman-Smith I, Brown G, Johnson A, and Land JM (1994).
Dilated cardiomyopathy due to type II X-linked
3
-methylglutaconic aciduria: successful treatment with pan-
Blankenship DC, Hug G, Balko G, van der Bel-Kann J, Coith
RL Jr, and Engel PJ (1993). Hemodynamic and myocyte
mitochondrial ultrastructural abnormalities in arrhythmogenic
right ventricular dysplasia. Am Heart J 126:989–995.
tothenic acid. Br Heart J 72:349–353.
Parks JS, Huggins KW, Gebre AK, and Burleson ER (2000).
Phosphatidylcholine fluidity and structure affect lecithin:
cholesterol acyltransferase activity. J Lipid Res 41:546–553.
Bleyl SB, Mumford BR, Thompson V, Carey JC, Pysher TJ,
Chin TK, and Ward K (1997). Neonatal, lethal noncompaction
of the left ventricular myocardium is allelic with Barth syn-
drome. Am J Hum Genet 61:868–872.
Richards RI, Shen Y, Holman K, Kozman H, Hyland VJ,
Mulley JC, and Sutherland GR (1991). Fragile X syndrome:
diagnosis using highly polymorphic microsatellite markers.
Am J Hum Genet 48:1051–1057.
Bolhuis PA, Hensels GW, Hulsebos TJ, Baas F, and Barth PG
(
1991). Mapping of the locus for X-linked cardioskeletal
myopathy with neutropenia and abnormal mitochondria
Barth syndrome) to Xq28. Am J Hum Genet 48:481–485.
Russell PT, Miller WJ, and McLain CR (1974). Palmitic acid
content of amniotic fluid lecithin as an index to fetal lung
maturity. Clin Chem 20:1431–1434.
(
Cantlay AM, Shokrollahi K, Allen JT, Lunt PW, Newbury-Ecob
RA, and Steward CG (1999). Genetic analysis of the G4.5
gene in families with suspected Barth syndrome. J Pediatr
Vreken P, Valianpour F, Nijtmans LG, Grivell LA, Plecko B,
Wanders RJ, and Barth PG (2000). Defective remodeling of
cardiolipin and phosphatidylglycerol in Barth syndrome.”
Biochem Biophys Res Commun 279:378–382.
135:311–315.
Cha RS, Zarbl H, Keohavong P, and Thilly WG (1992).
Mismatch amplification mutation assay (MAMA): application
to the c-H- ras gene. PCR Methods Appl 2:14–20.
D’Adamo P, Fassone L, Gedeon A, Janssen EA, Bione S,
Bolhuis PA, Barth PG, Wilson M, Haan E, Orstavik KH, Patton
MA, Green AJ, Zammarchi E, Donati MA, and Toniolo D
(
1997). The X-linked gene G4.5 is responsible for different
infantile dilated cardiomyopathies. Am J Hum Genet 61:862–
67.
8
3
44 Laboratory Investigation • March 2002 • Volume 82 • Number 3