From Cardiac Medicine, Imperial College School of Medicine at the
National Heart and Lung Institute (R.R.K., M.G., E.D., S.M.R., P.A.P.-W.,
N.J.S.), and Royal Brompton Hospital (M.N.S., R.U., D.J.P., K.F., J.P.),
London, England.
Correspondence to Prof N.J. Severs, Cardiac Medicine, Imperial College School of Medicine at the National Heart and Lung Institute, Sydney St, London SW3 6NP, England. E-mail n.severs{at}ic.ac.uk
Methods and Results—Patients with poor ventricular
function and severe coronary artery disease underwent thallium
scanning and MRI to predict regions of normally perfused, reversibly
ischemic, or hibernating myocardium. Twenty-one
patients went on to coronary artery bypass graft surgery,
during which biopsies representative of each of the
above classes were taken. Hibernation was confirmed by improvement in
segmental wall motion at reassessment 6 months after surgery.
Connexin43 was studied by quantitative immunoconfocal laser scanning
microscopy and PC image software. Analysis of en face
projection views of intercalated disks revealed a significant
reduction in relative connexin43 content per unit area in reversibly
ischemic (76.7±34.6%, P<.001) and hibernating
(67.4±24.3%, P<.001) tissue compared with normal
(100±30.3%); ANOVA P<.001. The hibernating regions
were further characterized by loss of the larger gap junctions normally
seen at the disk periphery, reflected by a significant reduction in
mean junctional plaque size in the hibernating tissues (69.5±20.8%)
compared with reversibly ischemic (87.4±31.2%,
P=.012) and normal (100±31.5%, P<.001)
segments; ANOVA P<.001.
Conclusions—These results indicate progressive reduction and
disruption of connexin43 gap junctions in reversible ischemia
and hibernation. Abnormal impulse propagation resulting from such
changes may contribute to the electromechanical dysfunction associated
with hibernation.
Two PET-based studies have shown that the prognosis in medically
treated patients with ischemic heart disease and impaired
ventricular function is poorer when there is additional
evidence of hibernating myocardium, but is significantly
improved by successful
revascularization.9 10 The
majority of deaths in these studies were sudden, possibly implicating
arrhythmias as being a more common feature in the presence of
myocardial hibernation.9 This combination of
segmental wall motion abnormalities and clinically important
ventricular arrhythmias could plausibly result from
severe disruption in conduction of the depolarizing wave front, leading
to asynchronous contraction of individual myocytes and predisposing to
reentry arrhythmias.11 Coordination of
the electrical, and hence mechanical, activity of individual myocytes
depends on both intercellular passive resistivity and active membrane
ionic properties,12 13 14 which together determine
propagation of the action potential and recovery of
excitability.15 Gap junctions, clusters of
transmembrane channels at the intercalated disk, form the passive
low-resistance intercellular pathways.16 17 18 The
constituent proteins of gap junctions are connexins, of which
connexin43 is the predominant isoform in working human
ventricular myocardium. Disruptions of the
distribution and tissue content of connexin43 gap junctions leading to
impaired intercellular conduction have previously been reported to
predispose to reentry arrhythmias in infarct border
zones19 20 21 and in reversibly ischemic
myocardium.22
In this study, we investigated the hypothesis that progressive
disruption in myocardial gap junctional organization during the
transition from normal to reversibly ischemic to hibernating
states may underlie wall motion abnormalities and increasing
arrhythmogenicity. Connexin43 gap junction distribution was compared in
documented hibernating, reversibly ischemic, and normally
perfused segmental biopsies from patients by quantitative
immunoconfocal microscopy.
Surgery was indicated on clinical grounds for each individual patient;
this was either for improvement of symptomatology (dyspnea and/or
angina) or improvement in prognosis.
Imaging
Twenty-one patients with more than two of nine ventricular
regions provisionally classed as hibernating (likely to improve in
function with revascularization) proceeded to CABG.
Per-operative transmural biopsies representing each of the
four classes were taken from each patient with a metal cork-borer of
2-mm ID; the surgeon was directed by a map of the nine-segment left
ventricular model described. This study was approved by the
local Ethics Committee, and all patients gave written consent using a
specifically designed consent form.
Processing of Specimens
Immunofluorescence Microscopy
Techniques for Quantification
Data obtained were finally pooled into the three clinical groups
(normally perfused, reversibly ischemic, and hibernating
myocardium) to obtain a distribution plot of results for
each group and for subsequent statistical analysis using ANOVA.
Secondary analysis was performed between the three groups by
use of Bonferroni's correction to Student's t test. Data
are expressed as mean±SD; secondary t test significance is
expressed as P'.
General Histology and Ultrastructure
Distribution of Connexin Label
Transmission electron microscopy of representative
intercalated disks from each tissue class revealed preservation of the
basic ultrastructure of the disk; desmosomes and fasciae adherentes
occupied the transverse portions of the disks, with gap junctions
situated in the zones parallel to the long axis of the cells (Fig 6
Quantification
Size of Immunolabeled Gap Junctions
In a study of this type, extrapolations of the nature of regional
tissue structure are made from assessment of only a small amount of
tissue taken as a biopsy. Because pathological changes resulting from
ischemic heart disease are known to be
inhomogeneous, there will be an unavoidable sampling error
from the examination of such biopsy material. In the present study,
such errors were minimized in three ways. First, we were able to
examine relatively large samples by use of a metal cork-borer, which
provides 2-mm-diameter transmural cylindrical biopsies, significantly
larger than a specimen from a needle biopsy. Second, concurrent tissue
from an individual patient was taken, whenever possible, from
hibernating, reversibly ischemic, and normally perfused
segments. This allowed direct comparison of diseased tissue with
internal controls of "normal" tissue collected in an identical
fashion. Third, all tissue examination was performed with the observer
blinded to the tissue origin of each biopsy. The consistent
patterns in immunodetectable connexin43 demonstrated under these
conditions indicate that the data reflect true differences between the
control and disease groups.
Earlier studies emphasized myofibril depletion as a hallmark of
myocytes in hibernation.3 However, separate
assessment of tissue histology and ultrastructure from the present
patient series has shown no significant differences in frequency of
myofibril-depleted myocytes between the groups of biopsies, including
the control tissues,6 suggesting that alternative
pathophysiological mechanisms are required to
account for the wall motion abnormalities characteristic of
hibernation. Alterations to gap junctions affecting intercellular
coupling are one such candidate mechanism. The present study, in
accordance with previous work, revealed that in reversibly
ischemic myocardium, areas distant from a zone of
infarction show a loss in connexin43 gap junction content but with
maintenance of their overall tissue distribution. A key finding
is that progressive decrease in the density of connexin43 gap junctions
and loss of the large gap junctions at the disk periphery occurred in
hibernating segments. No significant changes in other connexin isoforms
were apparent in working myocytes in our tissue.
A relationship between abundance of gap junctions and intercellular
conduction is suggested by studies of the functional role of the gap
junction in cardiac muscle,31 the absence of
conduction in vitro between cells without gap
junctions,32 and the findings of sparse junctions
in areas of heart with low conductance
velocities.17 18 33 A relationship between gap
junctional area and conductance has also been demonstrated in newly
forming connections between isolated myocytes in
vitro,34 as well as by the finding of concomitant
increases in conduction velocity and connexin expression in dibutyryl
cAMP–treated cultured cardiac myocytes.35
Although the precise in vivo relationship between absolute number of
gap junctions and conduction velocity is not known, significant cardiac
conduction defects are reported in heterozygotes for a connexin43 null
mutation with an
The mechanisms by which reduced intercellular coupling predisposes to
arrhythmia may involve the interaction of the characteristics
of the action potential of individual myocytes and the degree of
intercellular coupling. Human and canine myocytes from endocardial,
epicardial, and the deep subepicardial (M-cell layer) regions of normal
and diseased ventricular myocardium have been
shown to have different action potential durations, suggested to be due
to a variable transient outward K+ current
(Ito).41 42 43 44 45 The
coupling of these cells to their neighbors in vivo masks such
differences by modulating action potential duration of each cell to
reach a homogeneous depolarization.34
A moderate reduction in such communication by disruption of gap
junctions could result in an unmasking of these variable action
potential characteristics and hence to slowing and inhomogeneities in
the conduction of the depolarization wave.15 This
may, in turn, be manifest at the clinical level by the establishment of
potential reentrant circuits and hence ventricular
arrhythmias.14
A common feature of hibernating myocardium is the ability
to respond to inotropic stimulation with dobutamine. It is
becoming increasingly clear, however, that certain myocardial segments
are resistant to dobutamine stimulation but
eventually show recovery of function after
revascularization and hence are defined as
hibernating.46 47 A variable severity of gap
junctional disruption could provide a mechanism for this phenomenon,
with the dobutamine-resistant tissue being the more
severely affected.
Our data are consistent with the hypothesis that a reduction in
gap junction coupling is involved in the pathophysiology of
hibernation, possibly by unmasking local inhomogeneities in individual
cell action potential durations, hence disrupting wave-front
propagation, slowing conduction, and leading to loss of the local
coordination in myocyte contraction. The changes at the cellular level
would be manifested by nonspecific changes at the surface ECG. The
disruption in intercellular conduction and investigation of
excitation-contraction coupling needs further documentation by cellular
electrophysiological studies. A more
complete understanding of the role of gap junctions in hibernation
would be the first step for subsequent studies of these channels as
targets for therapeutic modulation.
Received July 3, 1997;
revision received October 23, 1997;
accepted October 29, 1997.
© 1998 American Heart Association, Inc.
Clinical Investigation and Reports
Downregulation of Immunodetectable Connexin43 and Decreased Gap Junction Size in the Pathogenesis of Chronic Hibernation in the Human Left Ventricle
Abstract
Top
Abstract
Introduction
Methods
Results
Discussion
References
Background—The regional wall motion
impairment and predisposition to arrhythmias in human
ventricular hibernation may plausibly result from abnormal
intercellular propagation of the depolarizing wave front. This study
investigated the hypothesis that altered patterns of expression of
connexin43, the principal gap junctional protein responsible for
passive conduction of the cardiac action potential, contribute to the
pathogenesis of hibernation.
Key Words: junctions, gap • myocardial contraction • hibernation
Introduction
Top
Abstract
Introduction
Methods
Results
Discussion
References
The term
"hibernating myocardium" describes a state of
persistently impaired myocardial function at rest due to severely
reduced coronary blood flow that is reversible after successful
revascularization.1 2
Previous studies on pathophysiological mechanisms
in hibernating human tissue have focused on the identification of
pathognomonic myocardial structural changes.
Histological and electron microscopic examination of
biopsy material obtained from hibernating regions reveal a variable
degree of fibrous tissue replacement, with a proportion of the myocytes
showing depleted contractile material, numerous and small mitochondria,
and irregular nuclear envelopes.3 4 Although
early studies suggested that such altered myocytes characterized the
hibernating state3 and could explain the delay in
recovery in contractile function after successful
revascularization,5 more
recent reports have described them in equal frequency in both
hibernating and nonhibernating ventricular
segments.4 6 Similar cellular features have also
been described in ventricular specimens from patients with
dilated, nonischemic
cardiomyopathy7 as well as in
rat cardiac myocytes after a short period of
unloading.8 It can therefore be hypothesized that
these are secondary features of cells contracting against a low
afterload and are not characteristic of hibernation.
Methods
Top
Abstract
Introduction
Methods
Results
Discussion
References
Patient Selection
Principal entry criteria for the study included a history of
previous myocardial infarction and subsequent established left
ventricular dysfunction with dyspnea as the predominant
symptom for at least 6 months. All patients had angiographic
three-vessel coronary artery disease (defined as >70% luminal
diameter stenosis) and left ventricular ejection
fraction <35%. Patients with significant valvular disease,
uncontrolled atrial fibrillation, permanent pacemaker in situ, or
previous coronary artery bypass graft surgery (CABG) were
excluded.
Before surgery, all patients underwent stress/redistribution and
separate-day rest/redistribution 201Tl single
photon emission CT. The stress procedure was performed by infusion of
adenosine 140 µg · kg-1
· min-1 coupled with 25 to 75 W of ergometer
exercise for 6 minutes. For both rest and stress studies, 80 MBq of
thallium was injected, and images were acquired immediately and again
after 4 hours of redistribution with an IGE Optima dual-headed gamma
camera. Data were processed with Ramp-Hanning filtering, and transaxial
tomograms were reoriented into the vertical and horizontal long-axis
and short-axis planes. Scanning was repeated after surgery to confirm
successful revascularization. MRI was performed
with a 1.5-T system (HPQ, Picker International
Inc).23 Cine gradient echo images were obtained
in the vertical and horizontal long-axis planes and in the basal and
apical short-axis planes. Preoperative images were acquired at rest and
during infusion of dobutamine 5 to 10 µg ·
kg-1 · min-1.
After surgery, only resting images were acquired. Scans were conducted
<3 months before CABG and between 3 and 6 months (mean, 4.8 months)
after surgery. Images were analyzed by two independent
observers for the thallium studies and by two others for the MRI.
Single photon emission CT images were graded for thallium uptake on a
5-point scale (4=normal, 3=slightly reduced, 2=moderately reduced,
1=severely reduced, 0=absent). Wall motion on cine-MRI was also graded
on a 5-point scale (3=normal, 2=mild hypokinesia, 1=severe hypokinesia,
0=akinesia, -1=dyskinesia). Coronary territories were assigned
as anterior, septum and apex to the left anterior descending
coronary artery, lateral to the circumflex artery, and
inferior to either the right or circumflex artery depending
on dominance. A nine-segment left ventricular model (basal
and apical parts of the septum, anterior, lateral, and
inferior walls, together with the apex) was used for
segmental classification as follows: The normally perfused
classification was defined as preoperative MRI wall motion grade 
2
and thallium uptake grade 4. Reversibly ischemic was
preoperative MRI wall motion grade 2 and improvement in thallium
uptake by at least one grade on stress/redistribution images. These
segments did not show any improvement in contractile characteristics
after successful revascularization. Hibernating was
defined as preoperative MRI wall motion grade 
1 and improvement in
contraction of at least one grade recorded with postoperative MRI.
(Late rest/redistribution thallium uptake grade 2 and/or improvement
by at least one wall motion grade with low-dose dobutamine
infusion was used to predict hibernation before surgery, directing
biopsy sites). Infarcted myocardium showed preoperative MRI
wall motion grade 1 and late rest/redistribution thallium uptake 1,
with no recovery of wall motion after surgery.
All biopsies were immediately fixed in freshly prepared
Zamboni's fixative on removal in the operating
theater.24 Biopsy orientation was recorded
before further division into epicardial and endocardial samples. For
histology and confocal microscopy, tissue was fixed for 2 to 6 hours,
washed overnight, and embedded in paraffin. Sections were routinely
stained with hematoxylin-eosin, periodic acid–Schiff reagent, and
phosphotungstic acid–hematoxylin. For electron microscopy and
high-resolution light microscopy, tissue divided from the original
biopsy was fixed in 2.5% glutaraldehyde in 0.1 mol/L
sodium cacodylate buffer at pH 7.3, postfixed in 2% osmium tetroxide,
stained en bloc in saturated uranyl acetate in 50% ethanol, and
embedded in epoxy resin. Semithin sections for light microscopy were
stained with toluidine blue. Ultrathin sections for electron microscopy
were stained with uranyl acetate and lead citrate before
examination.
Sections (10 µm) on polylysine-coated slides were dewaxed
with xylene and rehydrated. Microwave treatments in different
buffers25 26 and 0.1%
trypsin27 for retrieval of antigenic sites masked
by fixation were compared. Optimal antigen detection was found with 7
minutes of microwave oven treatment in 0.01% sodium bicarbonate at pH
7.4. Sections were blocked with 0.1 mmol/L L-lysine in
PBS containing 0.1% triton X-100 for 45 minutes before overnight
incubation with mouse monoclonal anti-connexin43 antibody (Chemicon
International Ltd) at dilution 1:1000 in 0.5% BSA in PBS at room
temperature. The detection system used was monoclonal sheep anti-mouse
biotinylated antibody/streptavidin Texas red (Amersham Life Sciences),
both at 1:250 dilution in PBS containing 0.5% BSA for 1 hour. The
sections were washed for 20 minutes and mounted in 50% glycerol
containing para-phenylenediamine as an
antibleaching agent. Sections from 20 blocks were processed
simultaneously in each immunolabeling run; controls in
which the primary antibody was omitted were routinely run in parallel.
In addition to connexin43, trials were conducted to test for expression
of two more connexin types, connexin40 and connexin45.
Affinity-purified polyclonal antisera against synthetic peptides
(residues 254 to 268 of rat connexin40 sequence; residues 354 to 367 of
human connexin45) raised in rabbit were used for this purpose. The
procedure was similar to that for connexin43, except that 0.1% trypsin
was used for antigen rescue. Connexin40 antibody was diluted to 1:1000
and connexin45 to 1:50; secondary detection was with anti-rabbit Cy3
conjugate (Chemicon International Ltd). Immunostained
tissue was coded and examined blind with conventional
epifluorescence and subsequently confocal microscopy with a
Leica TCS 4D equipped with an argon/krypton laser with the appropriate
filter blocks for the fluorescent signal. All images were
recorded within 48 hours of labeling with single-channel
scanning.
Quantification of connexin43 content was performed blind on
tissue finally classified as normally perfused, reversibly
ischemic, and hibernating by measurement of the
fluorescent label in individual intercalated disks from six
fields selected by random scanning across each tissue section. En face
disks seen in transversely sectioned tissue were used, because in this
view, the overlap of individual gap junctions is minimized. The x63
objective with a detector pinhole of 90 µm gave an optical section
thickness of 0.5 µm on theoretical grounds. A zoom factor of
2.96 with a 512x512-pixel image was selected, yielding a pixel size of
0.1x0.1 µm, allowing proportionate representation of
the smallest junctions (1 µm) detectable with this
technique.28 Because all of the
fluorescence associated with an intercalated disk was rarely
seen completely in a single optical section, a series of confocal
images through the depth of tissue containing the complete disk was
recorded. Images were taken at 0.5-µm intervals to ensure
acquisition of all gap junction label with minimal overlap between
adjacent optical sections. Each image was signal-averaged during
acquisition to improve image quality, and the entire series was
projected as a single composite image by superimposition (Fig 1
). This final image underwent high-pass
filtering, which increased image contrast, and then was transferred in
binary format for analysis by PC image software (Foster Findlay
Associates). Individual complete disks in each image were outlined by
hand for separate quantification. Thresholds for detection on a
255-point gray scale were adjusted to reduce any background cell
outlines, and the image was then edited by hand to eliminate spots of
easily identifiable autofluorescent lipofuscin. The binary
images of the labeled disks were inspected and compared with the
corresponding projection images to ensure that clearly separated
juxtaposed junctions were not contiguous. If such artifactual merging
occurred between proximate junctions, the binary image was further
edited to separate them. The total remaining "on" pixels were then
automatically counted to yield the total number of gap junctions within
the disk, the sizes of these individual junctions, and the total area
of the disk itself. The mean gap junction size in square
micrometers (total area of junctions/number of junctions in
disk) and the disk content of junctions (total area of junctions/disk
area) were calculated. Because there was variability in the absolute
label between immunofluorescence runs, results were
expressed as a percentage relative to the normally perfused tissue
acting as internal normal controls for each patient.
View larger version (40K):
[in a new window]
Figure 1. Diagrammatic representation of image
acquisition by immunoconfocal microscopy, which allows high-resolution,
low-background optical sections with 0.5-µm depth of field to be
recorded. Multiple images taken as a series at set intervals
through thickness of specimen are required to scan all spots of label
in a single disk, from which a single composite projection image of
entire disk can be reconstructed.
Results
Top
Abstract
Introduction
Methods
Results
Discussion
References
Patients
Twenty-one patients were selected for the study: 91% male; median
age, 61 years (range, 40 to 71 years); mean NYHA grade, 2.8; mean
ejection fraction, 23%. Of the 21 patients initially biopsied, full
follow-up investigations were completed in 15, allowing final
classification of tissue. Four patients did not survive to follow-up,
and the remaining 2 withdrew from the study. The demographic and
clinical details of the 15 patients are shown in Table 1.
View this table:
[in a new window]
Table 1. Patient Details
Ventricular biopsies from all four clinical groups
showed variable degrees of fibrosis, occurring both in the
interstitial spaces and as separate islands of connective
tissue representing local infarction. Tissue classified as
infarcted showed extensive fibrous tissue replacement. The myocytes
themselves varied from a normal appearance at light microscopy to
pathological cells severely depleted of myofibrillar proteins (Fig 2). These cells showed evidence of
glycogen accumulation by periodic acid–Schiff staining of
histological sections (not shown). On transmission
electron microscopy, the pathological cells revealed consistent
features of reduced myofibrillar content, glycogen accumulation,
numerous and small mitochondria, and irregular nuclear envelopes (Fig 3).
View larger version (194K):
[in a new window]
Figure 2. Light microscopy of biopsies from two patients. A,
Histological section from a reversibly ischemic
segment stained with phosphotungstic acid–hematoxylin demonstrating
interstitial fibrosis (F). Bar=100 µm. B and C,
Toluidine blue–stained semithin section of resin-embedded tissue from
a hibernating segment showing cellular myofibril loss (asterisks); one
cell in B shows severe changes with only a thin rim of striated
myofibrillar protein at cell periphery (large asterisk). Bars=50
µm.
View larger version (190K):
[in a new window]
Figure 3. Electron micrograph showing characteristic
appearances of "pathological" myocytes. Note depletion of
contractile material, with myofibrils confined to periphery of cells
(arrowhead), empty regions containing glycogen (g), minimitochondria
(m), and irregular nuclei (n). Cells of this morphology were found
throughout poorly functioning ventricles, regardless of segmental
classification. Example from hibernating region. Bar=10
µm.
Low-power views of connexin43 immunolabeling revealed a normal
distribution of label in the tissue away from zones of scarring, the
fluorescence being confined to typical intercalated disks (Fig 4). In infarcted tissue, there was
disruption of this normal pattern, as reported
previously,20 with degenerative myocytes
partially embedded in scar tissue having gap junctions distributed
widely over their surface membranes. Reversibly ischemic,
hibernating, and normally perfused (acting as internal normal control
tissue) biopsies were examined in greater detail at higher
magnification in tissue distant from such disrupted zones to determine
connexin43 distribution in individual intercalated disks. Disks seen en
face in transversely sectioned areas of tissue showed that in the
control tissues, there was a normal distribution of gap junction
label,16 17 with small central gap junctions
surrounded by extensive larger spots of label in the disk periphery
(Fig 5A). Disks from reversibly
ischemic biopsies had a similar pattern of connexin43 labeling
but demonstrated a decrease in the amount of label compared with the
control disks in the same staining run (Fig 5B). The hibernating disks
also showed a diminution of label, although in this case there was a
striking loss of the larger peripheral gap junctions (Fig 5C). Endocardial and epicardial areas of clearly oriented
full-thickness biopsies were compared to investigate any differences
that might reflect the recognized variation in perfusion and degree of
scarring present in these two areas, but none were apparent. There
were no immunodetectable changes in other connexin isoforms in the
samples examined; connexin40, present in arteriolar
endothelium, was not detected in myocytes, and
connexin45, which is expressed only in very low amounts in the heart,
was found not to be detectable in fixed specimens with the antibody
used. Label was absent in the controls in which the primary antibody
was omitted. Tissue autofluorescence was low after microwave
treatment, and highly fluorescent lipofuscin granules
ubiquitous in human tissue could easily be distinguished from specific
label.
View larger version (63K):
[in a new window]
Figure 4. Low-power views of connexin43-labeled
myocardium (longitudinal section) showing normal
distribution of label confined to intercalated disks in hibernating
(left) and reversibly ischemic (right) segments. Strong
fluorescence from lipofuscin (arrowhead) is easily
distinguished from punctate labeling of gap junctions (arrow).
Bar=25 µm (A and B).
View larger version (33K):
[in a new window]
Figure 5. High-power en face view of control (A), reversibly
ischemic (B), and hibernating (C) intercalated disks. Normal
pattern of large peripheral and small central gap junctions
are seen in control tissue. This distribution is maintained in biopsies
from reversibly ischemic regions, with an impression of
decrease in overall intensity of labeling of disk. Hibernating
intercalated disks show loss of normal pattern of labeling, with
absence of large peripheral junctions making disk borders
harder to define. Bar=25 µm (A, B, and C). ). This structure was maintained
regardless of the degree of loss of contractile material in a given
myocyte. There was no evidence of gap junction internalization at the
disk margin or of isolated junctions in nondisk membrane. These
findings were consistent with the described pattern of
connexin43 immunolabeling being confined to the intercalated disks in
areas distant from infarcts.
View larger version (164K):
[in a new window]
Figure 6. Electron micrograph showing
representative intercalated disk from hibernating
region of myocardium. Preservation of positions of three
membrane specializations in this zone of contact are shown: desmosomes
(d) and adherens junctions (a) are seen in transverse portion of disk,
with gap junctions (g) in longitudinal parts of disk. Inset, High-power
view of area indicated by large arrow showing gap junction and
desmosome. Bar=5 µm; inset, magnification x45 000.
Disk Gap Junctional Density
Differences in the extent of labeling were observed between
individual immunolabeling runs, with values of percentage gap junction
area per unit area of disk in the control biopsies varying between
10.4% and 13.0% (Table 2). Overall,
however, there were consistent patterns in the measurements
between the three groups. When expressed as relative values, mean gap
junction density per unit area of disk was significantly reduced
compared with control (100±30.3%) in both reversibly ischemic
(76.7±34.6%; P'<.001) and hibernating (67.4±24.3%;
P'<.001; ANOVA P<.001) tissues. These last
values were not significantly different from one another on secondary
analysis (P'=.069) (Fig 7). The findings confirmed the visual
impressions described above.
View this table:
[in a new window]
Table 2. Results of Quantification
View larger version (22K):
[in a new window]
Figure 7. Histograms showing results of quantification of
disk gap junction density and gap junction size expressed relative to
values obtained from control segments. Reversibly ischemic and
hibernating tissues show significant reductions in gap junctional
density in individual disks (ANOVA P<.001). Hibernating
tissue contains significantly smaller gap junctions than other groups
(ANOVA P<.001). Bars=SD. *P'<.001;
**P'=.012.
The mean gap junction size of control tissue was 0.21±0.06
µm2 (Table 2
), yielding a mean gap junctional
diameter of 0.53 µm (assuming a circular structure to a single
gap junction). If mean junctional size of the diseased tissues was
expressed as percentage relative to control, there was a significant
reduction in size to 87.4±31.2% (P'<.001) in reversibly
ischemic tissue, and a further significant reduction in
hibernation to 69.5±20.8% (P'=.012; ANOVA<.001). This
last value reflects the loss of peripheral large junctions
observed on examination of the disks from the hibernating tissue, with
maintenance of the frequency distributions of area measurements
of the smaller-size individual junctions from the different tissue
classes (Fig 8).
View larger version (36K):
[in a new window]
Figure 8. Frequency distributions of gap junction long-axis
measurements in confocal images from (A) normal segments and (B)
hibernating tissue, showing loss of larger junctions (highlighted bars
represent junctions >3.5 µm in diameter) in hibernating
tissue, with similar distribution of smaller junctions.
Discussion
Top
Abstract
Introduction
Methods
Results
Discussion
References
Evidence implicating alterations in connexin43 gap junctions in
arrhythmogenesis in different models of human heart disease has
steadily accumulated.20 21 22 26 29 30 The novel
findings of the present study are that (1) a reduction in
connexin43 gap junctions occurs in hibernating myocardium
beyond that seen in reversible ischemia, and (2) a specific
feature characterizing hibernating myocardium is a loss of
the population of large gap junctions at the disk periphery. These
changes seen by laser scanning immunoconfocal microscopy were
subsequently confirmed by quantification with PC image
analysis. The use of quantitative
immunofluorescence for measurement of gap junction
size has been previously validated with a polyclonal anti-connexin43
antibody in rat left ventricular tissue by comparison of
measurements from immunoconfocal (0.53 µm) and freeze-fracture
(0.57 µm) electron microscopy.28 In the
present study, quantification of images obtained with a different
primary antibody yielded a mean gap junctional diameter of 0.53
µm in control tissue, in accordance with the published data,
indicating that the technique reflects true structural measurements. We
elected to study connexin43 content in individual intercalated disks.
Previous work has shown a lack of significant changes in intercalated
disk number per myocyte in ischemic human
myocardium,22 indicating that the
density of gap junctions in the existing disks closely reflects the
tissue content of connexin43 in the ventricle as a whole. 
50% reduction in gap junction
content.36 Our findings of 23% and 33%
reduction of disk gap junction density in reversibly ischemic
and hibernating segments, respectively, may therefore plausibly cause
significant disruption of intercellular propagation of depolarization.
The specific loss of the peripheral large gap junctions in
the disks of hibernating myocardium is potentially
critical. Recent findings suggest that tetrodotoxin-insensitive
voltage-dependent sodium channels, essential for the rapid early
upstroke of the cardiac action potential,37 have
been localized to the lateral sarcolemma38 and
may be preferentially concentrated around the intercalated
disks.39 The peripheral intercalated
disk conductance, determined by these large junctional plaques, may be
of particular importance because of this spatial relationship. On
theoretical grounds, intercellular conduction depends on the reciprocal
relationship between intercalated disk resistance, determined by the
absolute number of connexin channels, and cell size. Although cell
volume was not directly assessed in our study, intercalated disk area,
which reflects the cross-sectional area of the corresponding myocyte,
provides an indirect measure of cell size. The finding of progressively
decreasing density of labeling in the disks in reversible
ischemia and hibernation partly reflects larger disk sizes in
these groups. Disruption in passive intercellular
communication22 from such changes may, in turn,
predispose to the increasing frequency of arrhythmia and sudden
death associated with reversible
ischemia40 and
hibernation.9 Furthermore, regional contractile
dysfunction in hibernating areas may be partly accounted for by local
incoordination in contraction of individual myocytes secondary to
fragmentation of the depolarizing wave front.
Acknowledgments
This work was supported by British Heart Foundation Junior
Fellowship FS/94073, the Coronary Artery Disease Research
Association, and also in part by project grants from the Wellcome
Trust (grant 046218/2/95) and the British Heart Foundation (grant
PG93136).
References
Top
Abstract
Introduction
Methods
Results
Discussion
References
This article has been cited by other articles:
 |
|
 |
|
  N. J. Severs, A. F. Bruce, E. Dupont, and S. Rothery Remodelling of gap junctions and connexin expression in diseased myocardium Cardiovasc Res, June 17, 2008; (2008) cvn133v2. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Sato, T. Ohkusa, H. Honjo, S. Suzuki, M.-a. Yoshida, Y. S. Ishiguro, H. Nakagawa, M. Yamazaki, M. Yano, I. Kodama, et al. Altered expression of connexin43 contributes to the arrhythmogenic substrate during the development of heart failure in cardiomyopathic hamster Am J Physiol Heart Circ Physiol, March 1, 2008; 294(3): H1164 - H1173. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. F. Bruce, S. Rothery, E. Dupont, and N. J. Severs Gap junction remodelling in human heart failure is associated with increased interaction of connexin43 with ZO-1 Cardiovasc Res, March 1, 2008; 77(4): 757 - 765. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. L. Wit and H. S. Duffy Drug development for treatment of cardiac arrhythmias: targeting the gap junctions Am J Physiol Heart Circ Physiol, January 1, 2008; 294(1): H16 - H18. [Full Text] [PDF]
|
|
|
|
|
|
R. J.P. Musters Atrial gap junction remodeling: Looking for lost gaps and orphaned connexins in three dimensions Cardiovasc Res, October 1, 2006; 72(1): 5 - 6. [Full Text] [PDF]
|
|
|
|
|
|
C. Rucker-Martin, P. Milliez, S. Tan, X. Decrouy, M. Recouvreur, R. Vranckx, C. Delcayre, J.-F. Renaud, I. Dunia, D. Segretain, et al. Chronic hemodynamic overload of the atria is an important factor for gap junction remodeling in human and rat hearts Cardiovasc Res, October 1, 2006; 72(1): 69 - 79. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Li, V. V. Patel, and G. L. Radice Dysregulation of cell adhesion proteins and cardiac arrhythmogenesis. Clin. Med. Res., March 1, 2006; 4(1): 42 - 52. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Betsuyaku, N. S. Nnebe, R. Sundset, S. Patibandla, C. M. Krueger, and K. A. Yamada Overexpression of cardiac connexin45 increases susceptibility to ventricular tachyarrhythmias in vivo Am J Physiol Heart Circ Physiol, January 1, 2006; 290(1): H163 - H171. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. H. Lamers and S. E. Kohler How much intimacy is compatible with survival for a cardiomyocyte? Cardiovasc Res, October 1, 2005; 68(1): 1 - 2. [Full Text] [PDF]
|
|
|
|
|
|
D. E. Gutstein, S. B. Danik, S. Lewitton, D. France, F. Liu, F. L. Chen, J. Zhang, N. Ghodsi, G. E. Morley, and G. I. Fishman Focal gap junction uncoupling and spontaneous ventricular ectopy Am J Physiol Heart Circ Physiol, September 1, 2005; 289(3): H1091 - H1098. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Yamada, K. G. Green, A. M. Samarel, and J. E. Saffitz Distinct Pathways Regulate Expression of Cardiac Electrical and Mechanical Junction Proteins in Response to Stretch Circ. Res., August 19, 2005; 97(4): 346 - 353. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. J. Gard, K. Yamada, K. G. Green, B. C. Eloff, D. S. Rosenbaum, X. Wang, J. Robbins, R. B. Schuessler, K. A. Yamada, and J. E. Saffitz Remodeling of gap junctions and slow conduction in a mouse model of desmin-related cardiomyopathy Cardiovasc Res, August 15, 2005; 67(3): 539 - 547. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 < |