Vol. 275, Issue 4, H1225-H1235, October 1998
Are interatrial band myocytes maximally hypertrophied in
normal canine hearts?
Paul C.
Dolber,
Robert P.
Bauman,
Judith C.
Rembert, and
Joseph
C.
Greenfield Jr.
Department of Surgery and Division of Cardiology, Department of
Medicine, and Department of Pathology, Duke University Medical
Center, Durham 27710; and Cardiology Section, Medical Service and
Research and Development Service, Veterans Affairs Medical Center,
Durham, North Carolina 27705
 |
ABSTRACT |
In canine right atrial hypertrophy, the
cross-sectional area
(Axs) of right
atrial myocytes increases, whereas the
Axs of the
broader interatrial band myocytes does not. In the current study,
myocyte reconstructions showed that right atrial myocyte length
increased in proportion to
Axs in right
atrial hypertrophy. On the other hand, mean interatrial band myocyte
length in both normal and right atrial hypertrophy dogs was roughly
inversely proportional to mean
Axs, as expected
if interatrial band myocyte volume was constant. Plotting mean
Axs vs. myocyte
length for individual interatrial band myocytes revealed a distribution
whose border defined a maximal volume curve; many myocytes were well beneath that curve. Mononuclear myocytes (generally diploid) were limited by a 65,000-µm3 curve,
which many binuclear myocytes (generally tetraploid) surpassed; myocyte
ploidy thus constrained myocyte volume. However, because many
mononuclear and binuclear myocytes had lower volumes, their failure to
hypertrophy cannot be attributed to attainment of the maximal volume
possible for their ploidy.
atrium; cardiac muscle; hypertrophy; myocyte size; ploidy
| |
INTRODUCTION |
THE INTERATRIAL BAND is a thickened muscle mass
connecting the two atria, running from the superior vena cava to the
left atrial appendage. Besides forming a mechanical linkage between the
right and left atria that should share the hemodynamic stress of both
chambers, it is the muscle bundle along which interatrial conduction is
most rapid (2).
We found that myocytes of both the right and left atrial portions of
the canine interatrial band have a mean cross-sectional area roughly
twice that of myocytes in the appendage and nonappendage regions of
either atrium (9). Epimyocardial interatrial band myocytes were often
clearly broader than the endomyocardial myocytes and contributed more
to the mean measured breadth in our study because they were more
parallel to the longitudinal axis of the interatrial band and
consequently were more likely to be precisely cross-sectioned and thus
measured.
It is not known why the interatrial band myocytes are so much broader
than noninteratrial band myocytes. One hypothesis is that the stress on
them is greater than the stress elsewhere in the atria, i.e., they
could be physiologically hypertrophied. Their morphology is consistent
with this, in that they contain abundant organized myofilaments (11,
33), but there are no physiological data to support this hypothesis. A
second hypothesis is that interatrial band myocytes are specialized
atrial conduction cells (30) that, like Purkinje cells in the dog
ventricular conduction system, are generally larger than their
unspecialized neighbors. However, the single feature held in common by
all conduction system myocytes is the absence of transverse tubules
(33), whereas canine interatrial band myocytes contain abundant
transverse tubules (9). Finally, the larger size of interatrial band
myocytes might be a consequence of their having a higher ploidy than
noninteratrial band myocytes, given that elevated ploidy (complement of
chromosomes) is associated with an increase in cell size (4, 8).
In a tricuspid regurgitation (TR) model of right atrial hypertrophy in
which right atrial mass more than doubled, myocyte cross-sectional area
was markedly increased in right atrial appendage and nonappendage
regions but was unchanged in either the right or left atrial portion of
the interatrial band (9). The reason for the failure of interatrial
band myocyte breadth to increase in this model is unknown. One
possibility is that the interatrial band myocytes enlarge in length
rather than in breadth. Right atrial myocyte cross-sectional areas went
up by a factor of 1.5-1.8 in TR-induced atrial hypertrophy,
whereas right atrial mass increased by a factor of 2.2, suggesting that
right atrial myocytes and perhaps interatrial band myocytes could be
longer in atrial hypertrophy.
The studies described in this paper were initiated to determine whether
right atrial appendage myocytes and interatrial band myocytes increased
in length in TR-induced right atrial hypertrophy.
| |
MATERIALS AND METHODS |
The same dog hearts and, except for left ventricular epimyocardium, the
same blocks were used for the reconstruction studies as for the
previously reported morphological studies (9). Five of the ten dogs
studied had tricuspid regurgitation initiated by cutting the chordae
tendineae one year before they were killed. Right atrial pressure rose
from 0.8 ± 1.5 mmHg before intervention to 8.1 ± 1.5 mmHg at 1 yr, and right atrial-to-body weight ratio increased from 0.31 ± 0.02 g/kg in normal dogs to 0.69 ± 0.05 g/kg in TR dogs at 1 yr.
The five normal dogs were acquired at the time of study. All dogs were
studied according to a protocol approved by the Animal Care and Use
Committees of both Duke University and the Durham Veterans Affairs
Medical Centers.
Each dog was killed with 40 mg/kg thiamylal sodium. The heart and great
vessels were exposed via a left thoracotomy. All of the hearts were
fixed in situ using Langendorff perfusion with a perfusion pressure of
40 mmHg. First, a room temperature oxygenated cardioplegic solution was
perfused to clear the cardiac vasculature and to relax both the
vasculature and the myocardium. The composition of the cardioplegic
solution was (in mM) 12 NaCl, 7 KCl, 2 MgCl2, 11 dextrose, 234 fructose,
10 HEPES, 7 procaine, 25 mg/l heparin, 2 IU/l insulin, 4%
polyvinylpyrrolidone (mol wt 40,000), and 0.01% methylene blue, pH 7.4 (after Ref. 15). After 5-10 min of cardioplegic perfusion, the
heart was perfused for 15 min with room temperature neutral buffered
4% formaldehyde. The heart was excised and left in fixative for at
least 2 wk. For the studies reported here, one sample per animal was
taken from each of three regions: right atrial appendage roof, right
atrial portion of the interatrial band, and left ventricular lateral
free wall approximately midway between apex and base. The samples were
processed routinely and embedded in paraffin, and sections were cut at
7 µm from blocks kept at room temperature.
Sections were stained with fluorescein-conjugated wheat germ agglutinin
(WGA-FITC) (9, 10). Sections were deparaffinized and rehydrated, rinsed
5× 3 min in phosphate-buffered saline (PBS; in mM: 8.5 Na2HPO4,
1.5 KH2PO4,
150 NaCl, pH 7.2), incubated for 60 min in 50 mg/ml WGA-FITC (Vector
Laboratories, Burlingame, CA) in PBS with 1 mM
CaCl2, and rinsed in PBS (10). In
some later experiments, sections were subsequently incubated in 50 mg/ml propidium iodide in PBS for 60 min to stain the nuclei, then
briefly rinsed with PBS. The slides were mounted in 25 g/l NaI in 1:1
glycerol:PBS. Sections were examined on an Olympus BH-2 or IX70
epifluorescence microscope.
Myocyte reconstructions.
Myocyte length determinations were made by serial section
reconstructions. Whereas the use of isolated cardiomyocytes yields quicker results, there were two pragmatic reasons for resorting to
reconstructions. 1) The dog hearts with right atrial
hypertrophy were used to measure regional blood flow with radioactive
microspheres (3), requiring perfusion fixation and thereby ruling out
myocyte isolation. 2) In our prior morphometric study (9), it
was usually only epimyocardial and not endomyocardial interatrial band
myocytes that were cut in cross section and thus included in the
morphometric study. Because epimyocardial myocytes appear to be larger
than endomyocardial myocytes, and because it may be that only the
larger myocytes fail to enlarge with TR-induced right atrial
hypertrophy, we limited the myocyte length study to the epimyocardial
myocytes. It was therefore necessary to cleanly separate the epi- from
the endomyocardium, which is more precisely done by microscopic
examination of the serial sections than by gross dissection of the
irregularly shaped interatrial band before myocyte isolation.
Estimation of myocyte length by serial section reconstruction is
conceptually simple. Because intercalated disks are captured en face in
myocyte cross sections and are clearly visible with WGA-FITC staining
(9), the extreme ends of the myocyte are easily found. Assuming that
the intercalated disk is on average found midway through the section,
it is necessary only to count the number of sections between the
intercalated disks at the extreme ends and multiply times section
thickness to obtain myocyte length.
Several difficulties must be recognized in making the reconstructions.
1) True cross sections must be used
because recognition of narrow intercalated disks would be difficult in
oblique sections. Departures from true cross sections are easily
recognized in WGA-FITC-stained myocardium (9).
2) Taking the center section of a
set of serial sections as a starting point for reconstruction, it is
necessary for the total number of sections to be large enough for any
myocyte crossing (or even ending on) that center section to be
completely enclosed within the set of sections. As many as 120 sections
were needed for some interatrial band sites. To obtain such long sets of virtually flawless serial paraffin sections is not trivial; thus
only a portion of the myocytes that are truly cross-sectioned can be
reconstructed. 3) It is very difficult to follow given myocytes
from section to section at the microscope because the appearance of the
myocytes changes at every intercalated disk. Accordingly, when taking
pictures for a reconstruction, it was necessary to use landmarks such
as blood vessels and connective tissue septa to find the same site for
photography in successive sections. Because such landmarks are not
uniformly distributed throughout the tissue, the sites that are sampled
may not be representative of the tissue as a whole. 4) The
method is so time-consuming that it is not feasible to study a large
number of sites from each block.
Micrographs were taken using a 40× objective, with which
virtually all intercalated disks throughout the thickness of 7-µm paraffin sections are simultaneously in reasonable focus, and printed
at a final magnification of ~875×.
Estimation of myocyte cross-sectional area from reconstructions.
Myocyte cross-sectional area was estimated in two different ways. The
first way was to measure the mean cross-sectional area of myocyte
profiles on selected photographs. The mean cross-sectional area of the
myocyte profiles corresponds to the mean cross-sectional area of the
average myocyte provided that the myocytes photographed are
representative of the average myocyte, all myocyte profiles are
included, and each myocyte is represented by a single profile per
photograph, i.e., there is no myocyte branching.
The second way to estimate myocyte cross-sectional area used the serial
sections for each of the reconstructed myocytes. For each myocyte, its
cross-sectional area on each section was measured. For branched
portions of myocytes, the areas of the branch profiles were summed to
yield the cross-sectional area for the myocyte on that section. The
mean cross-sectional area of the entire myocyte was then determined as
the mean of the values recorded for each of the sections. The mean of
the mean values of a group of myocytes would yield the mean
cross-sectional area of the average myocyte provided only that the
myocytes reconstructed are representative of the average myocyte.
Comparison of the two methods for a given site reveals how successful
the first method is in relation to this more time-consuming method.
For either method, photographs were mounted on a SummaSketch digitizer
with a resolution of 40 lines/mm (Summagraphics, Fairfield, CT)
connected to a 386SX-based personal computer, and the myocyte cross-sectional areas were determined with the program Easydij, version
5.0 (Geocomp, Golden, CO).
Myocyte isolation.
Using three additional dogs, we obtained isolated myocytes from the
left ventricular epimyocardium of one dog and the interatrial band of
two others as follows. After anesthetization, the heart was excised and
retrogradely perfused by the Langendorff method for 5 min with ice-cold
Normosol-R (in mM: 90 NaCl, 27 sodium acetate, 23 sodium gluconate, 5 KCl, and 3 MgCl2) containing 30 mM 2,3-butanedione. Myocardial samples were cut into chunks of about
2-4 mm on a side. These chunks were put into a chamber for combined enzymatic and mechanical disaggregation (17), and the chamber
was put into a 37° incubator gassed with 5%
CO2. With the disaggregator
turning at 7 rpm, the chunks were treated first for 7 min with 0.025%
trypsin (Sigma type XI) in modified saline 1 (in mM: 90 NaCl, 30 KCl, 5.5 dextrose, 42 sucrose, 2 NaHCO3, and 2 HEPES, pH 7.4). The
trypsin was drawn off and replaced with 0.1% collagenase (Worthington
type II) and 0.004% DNase II (Sigma type IV) in modified
saline 1. Every 30 min, the fluid and
disaggregated myocytes were drawn off and examined. Digests with good
yields of disaggregated myocytes were allowed to settle for 10 min,
then successively suspended and allowed to settle in
1) calcium-free modified
Krebs-Henseleit saline (KH) (in mM: 140 NaCl, 9.6 KCl, 2.4 MgSO4, 4 NaHCO3, 1.2 NaH2PO4,
12.0 HEPES), 2) calcium-free KH with
1% bovine serum albumin (BSA), 3)
KH containing 25 µM calcium and 1% BSA,
4) KH containing 0.5 mM calcium
saline and 1% BSA, and 5) medium
199 (GIBCO BRL). Myocytes were next put onto Nunc Lab-Tek two-well
chamber slides that had been treated with 20 mg/ml natural mouse
laminin (GIBCO BRL) and allowed to attach overnight. The culture medium
was withdrawn and replaced with 4% neutral buffered paraformaldehyde,
in which the myocytes remained for ~2 wk before examination using an
Olympus IX70 inverted microscope with Hoffman modulation optics. Images
of the first 100 nonoverlapping myocytes were captured using a Dage-MTI
CCD72 video camera connected through a Scion LG-3 framegrabber to a Macintosh 7100/80 PowerPC. Myocyte lengths were measured using the
public domain NIH Image program, version 1.59 (developed at the
National Institutes of Health and available from the Internet by
anonymous FTP from zippy.nimh.nih.gov).
Ploidy determination.
Ploidy of isolated myocytes was estimated by video-based densitometry
after Feulgen staining. Paraformaldehyde-fixed isolated myocytes
attached to laminin-coated slides as described above and chicken red
blood cells (Sigma R-0504) attached to Alcian blue-coated slides (32)
were rinsed in distilled water, washed overnight in 95% ethanol to
eliminate the plasmal reaction, rinsed in running water, and hydrolyzed
for 60 min at room temperature in 5 N HCl. After a 5-min running water
rinse, the myocytes were stained for 60 min with pararosaniline-based
Schiff reagent (Fisher), rinsed three times in 0.5% sodium
metabisulfite for 2 min each, rinsed in distilled water, then
dehydrated, cleared, and mounted. Preparations were viewed with a
560-nm interference filter with a half-peak bandwidth of 10 nm (Oriel,
Stratford, CT) on an Olympus BH-2 microscope. Video images were
collected as described above. As each image was collected, image
ratioing was used to correct for camera and lighting nonuniformities.
Density calibration was accomplished using neutral density filters
whose density was established by spectrophotometry. Density slicing was
used to define a region of interest (ROI) around each nucleus, and the
area and mean density of the nucleus were then measured. Next, the ROI
was moved to the cytoplasm adjacent to the nucleus, and the mean
density was determined there. Integrated gray value was calculated as
nuclear area times the difference in the mean gray value between the
nucleus and the cytoplasm. This is converted to optical density using the density calibration and then converted to DNA content by comparison with nuclei of chicken red blood cells, which have ~30% of the DNA
content of mammalian diploid myocytes.
Statistical and graphical analysis.
As described above, it was necessary
1) to search for sites with
cross-sectioned myocytes that were flawless over a large number of
sections, 2) to select myocytes that
were near recognizable tissue landmarks to reliably locate the same
site in successive sections, and 3)
to limit the study to relatively few reconstructions. In consequence,
the myocytes cannot be regarded as randomly selected. Note, however,
that myocyte selection had no dependence on myocyte length, which could
not be estimated until the reconstruction was completed. Thus, although
it is possible that the selected myocytes are not representative of the
whole population of myocytes, there was no possibility of selection
bias as regards myocyte lengths. One-tailed
t-tests were performed of the
hypotheses that myocytes from different regions of hypertrophied atria
were larger than those of control atria and that binuclear myocytes
were larger than mononuclear myocytes.
To visually assess differences among distributions of myocyte lengths
from different sites, we used the kernel density estimator method to
produce smoothed histograms with estimated probability density plotted
against myocyte cross-sectional area (23, 31). The area under each
curve, roughly equal to the computational step size (4 µm2) times the sum over all
computational steps of the probability density calculated at each step,
is approximately equal to unity. The procedure of Lexell and Taylor
(23) was closely followed, using the same kernel and the same method of
choosing window width.
| |
RESULTS |
Determination of myocyte length by reconstructions depends on the
recognition of intercalated disks in cross-sectioned myocytes stained
with WGA-FITC. Figure 1 shows four serial
sections stained with WGA-FITC in which two interatrial band myocytes,
m1 and
m2, are connected end-to-end by
intercalated disks (marked with stars). The end-to-end connection of
the two myocytes is illustrated schematically in Fig.
1E.
View larger version (48K):
[in this window]
[in a new window]
|
Fig. 1.
Detecting myocyte ends.
A-D:
4 serial sections that show end-to-end connections of 2 interatrial
band myocytes (m1 and
m2);  , intercalated disk.
Bar, 10 µm. E: schematic
reconstruction of joined ends of myocytes
m1 and
m2, using 4 serial sections of
A-D, as marked on
right.
|
|
Because the left ventricle is the only site in the dog heart for which
published values are available for myocyte length, our first
reconstructions were performed with left ventricular free wall
epimyocardium. Figure 2 shows a full
reconstruction of one left ventricular epimyocardial myocyte 161 µm
in length, which is close to the mean value for these myocytes. The
intercalated disks, which are generally strongly labeled, are outlined
in ink. The intercalated disks are located mostly, albeit not
exclusively, toward the ends of the myocyte. The shape of the myocyte
changes at every intercalated disk and also changes somewhat even well away from the intercalated disks, a phenomenon that is so common as to
be the rule. This may be due in part to deformations of the paraffin
sections but probably is due more to true changes in myocyte geometry.
Transverse tubules are numerous and conspicuous in every myocyte
profile other than those of intercalated disks.
View larger version (127K):
[in this window]
[in a new window]
|
Fig. 2.
Reconstruction of left ventricular (LV) epimyocardial myocyte.
Twenty-four serial sections through the myocyte are shown and numbered.
Intercalated disks along this myocyte are outlined in ink and marked
with arrows. Numerous intercalated disks nearing ends of myocyte would
lend "staircase" appearance to a longitudinal section through
reconstructed myocyte. Myocyte length = section thickness (7 µm) × difference between first and last section numbers (i.e., 23) = 161 µm. Bar in panel 33, 10 µm.
|
|
Fifty myocytes were reconstructed from within 500 µm of the
epicardium of the left ventricular lateral free wall midway between apex and base of the hearts from each of four dogs (thus 200 myocytes total). The distribution of myocyte lengths in the four dogs is shown
in Fig.
3A, in
which each smoothed histogram represents 50 myocytes. The distributions
were very similar, as were the mean lengths for each group (mean ± SD: 149 ± 26, 153 ± 34, 158 ± 34, and 164 ± 29 µm). The average of these values, 156 µm, is in excellent
agreement with our measurements of left ventricular epimyocardial
isolated myocyte length (157 ± 27 µm) and similar to published
values of 138-142 µm for isolated myocytes of adult dog left
ventricle (14, 36), validating the reconstruction technique as a means
of finding myocyte length.
View larger version (36K):
[in this window]
[in a new window]
|
Fig. 3.
Smoothed histograms of myocyte lengths (µm).
A: normal LV epimyocardium (LV epi).
B: normal (con) and tricuspid
regurgitation (TR) right atrial (RA) appendage.
C: normal right interatrial band
(IAB). Dotted curves 1 and
2 represent numbered sites in Fig. 4.
D: TR interatrial band. pde,
probability density estimate.
|
|
Mean myocyte cross-sectional area was estimated in two ways. First,
mean myocyte cross-sectional area was estimated for each block as the
mean area of the nonintercalated disk myocyte profiles on the
micrograph of the center section. The overall mean from the four
ventricles was 202 µm2, which
falls within the range found for canine left ventricular epimyocardium
using transmission electron microscopy of plastic-embedded specimens
(13). Second, the areas of every cross section through 25 myocytes were
determined for one dog, and their individual mean cross-sectional areas
thus determined. The mean value of the 25 individual myocyte mean
cross-sectional areas was 211 ± 30 µm2; this close correspondence
shows that the second method is indeed a good predictor of the first.
Right atrial myocyte reconstructions.
Reconstructions were made of right atrial free wall myocytes in sites
from three normal and three TR hearts. The smoothed histograms are
shown in Fig. 3B, with normal curves
drawn with solid lines and TR curves drawn with dotted lines. It is
clear that myocyte length is considerably increased for the right
atrial myocytes in the TR dogs, with the mean lengths being 113, 114, and 134 µm in the normal hearts and 161, 179, and 192 µm for the TR
hearts. The overall mean lengths of 120 and 177 µm were significantly different (P = 0.007). The mean
cross-sectional areas measured on the center sections from the six
sites ranged from 129-152 µm2 for the normal hearts to
188-345 µm2 for the TR
hearts, and the mean values of 144 and 276 µm2 are significantly different
(P = 0.05). These values are in
substantial agreement with the mean values from our prior morphometric
study of these hearts: 140 µm2
in the normal hearts and 220 µm2
in the TR hearts. These results indicate that the right atrial myocytes
are both broader and longer in the TR hearts.
Right interatrial band myocyte reconstructions.
Inspection of cross sections of the right atrial portion of the
interatrial band often revealed considerable regional variation in
myocyte cross-sectional areas. An example is shown in Fig. 4A, where
two subepicardial regions separated laterally by ~1.4 mm contained
myocytes with noticeably different cross-sectional areas. At higher
magnification (Fig. 4, B1 and
B2), the myocytes in both regions
are similar in appearance and contain plentiful transverse tubules,
ruling out the possibility that one of the regions contains specialized
conduction system myocytes [which lack transverse tubules
(33)]. For all three of the normal interatrial band blocks that
were studied and for one of the three TR interatrial band blocks, two
sites were reconstructed and, where there were conspicuous
interregional differences, one set of reconstructions was made in each
region.
View larger version (160K):
[in this window]
[in a new window]
|
Fig. 4.
Local variations in myocyte cross-sectional area in interatrial band.
A: interatrial band epimyocardium.
Regions 1 and
2, regions from which myocytes were
reconstructed; E, epicardium; N, nerve.
B1: narrow myocytes from
region 1; mean cross-sectional area of
71 myocytes measured on center section of reconstruction was 242 µm2.
B2: broad myocytes from
region 2; mean cross-sectional area of
31 reconstructed myocytes on center section was 394 µm2. Bar in
A, 250 µm. Bar in B1, 20 µm.
|
|
A full reconstruction of a 329-µm-long interatrial band myocyte,
taken from region 1 of Fig.
4A, is shown in Fig.
5. As with the ventricular reconstruction
shown in Fig. 2, the intercalated disks are found mostly, but not
exclusively, toward the ends of the myocyte, and shape changes occur
along the whole length of the myocyte. Interatrial band myocytes
sometimes, as here, showed myocyte branching near their ends (as
depicted by the separate a and
b profiles in
sections 42-44 of Fig. 5).
Transverse tubules are often numerous and conspicuous in the larger
cross sections near the middle of the myocyte but are frequently rare
or absent in smaller cross sections. Thus the absence of transverse
tubules from any given myocyte cross section is no indication of the
presence of transverse tubules in the myocyte as a whole.
View larger version (152K):
[in this window]
[in a new window]
|
Fig. 5.
Serial sections showing a complete interatrial band myocyte 329 µm
long. Intercalated disks are outlined in ink and marked with arrows.
Transverse tubules, visible as lines roughly perpendicular to myocyte
surface, are present only in larger profiles. Near one end, myocyte is
branched over 21 µm (sections
42-44). Myocyte branching indicated by
a and
b. Bar in first frame, 10 µm.
|
|
The smoothed histograms of the lengths of the interatrial band myocytes
are shown in Fig. 3, C (normal) and
D (TR). Two striking features of these
curves are the great variability from site to site and the often very
great mean myocyte length. In Fig. 3C, dotted curves 1 and
2 correspond to the two regions so
labeled in Fig. 4A. The mean lengths
of the myocytes for these two curves were 223 and 161 µm,
respectively.
For these two curves, the cross-sectional area of the myocytes on the
center section of the reconstruction was greater for the shorter
myocytes of region 2 (427 µm2) than for the longer
myocytes of region 1 (268 µm2). This is the opposite of
the trend for the right atrial free wall myocytes, where the broader
myocytes tended to be longer. To see how general this trend was, we
plotted a scattergram relating the mean myocyte length found in each
reconstructed site to the mean cross-sectional area of the myocyte
profiles in the center section of the reconstructed site (Fig.
6). Thus each point on this plot represents
the mean values for a given site with 22-71 myocytes. The points
for most of the interatrial band myocytes fell on a concave-up curve
for cylinders with a volume of 65,000 µm3. The points for the right
atrial free wall myocytes and the left ventricular myocytes were better
fit by a straight line or perhaps a concave-down curve; the one shown
is that for cylinders whose length is nine times their diameter.
View larger version (23K):
[in this window]
[in a new window]
|
Fig. 6.
Mean myocyte cross-sectional area
(µm2) vs. mean myocyte length
(µm) for several normal (con) and TR sites. Each point represents
mean value for 14-79 myocytes from a single site. Concave-down
curve is that for cylinders whose lengths are 9 times their diameters;
concave-up curve is that for cylinders with a volume of 65,000 µm3.
|
|
Both the smoothed histograms of Fig. 3,
C and
D, and the scattergram of Fig. 6
suggest that the mean lengths of the normal and TR interatrial band
myocytes are at least similar and that myocyte length and breadth are
reciprocally related. Neither the mean lengths (247 vs. 172 µm) nor
the mean widths (282 vs. 338 µm2) of the control vs. TR
interatrial band myocytes were significantly different. The inverse
relationship between myocyte length and cross-sectional area suggests
that myocyte volume is constant but does not reveal whether the
myocytes are of different length-to-breadth ratios at rest or whether
they might have been fixed at different degrees of contraction. To
answer this question, we took the remaining pieces of interatrial band
tissue, embedded them for longitudinal sectioning, and examined the
resulting sections using polarization microscopy. No contraction bands
were observed. We used NIH Image to measure mean sarcomere length in
two sites in each of 10 video micrographs from each block. The
greatest-to-least length ratio for sarcomeres in any site was 1.84 for
control animals and 1.47 for TR animals, whereas the the
greatest-to-least length ratio for reconstructed myocytes in any site
was 2.3 for control animals and 2.5 for TR animals. Thus variations in
contractile state cannot be responsible for the observed differences in
length-to-breadth ratios.
To further explore the relationship between myocyte mean
cross-sectional area and length, we determined for each of 115 normal and 76 TR interatrial band myocytes its mean cross-sectional area (i.e., for each myocyte, the mean value of its cross-sectional area
taken over all of its cross sections). The relationship of the mean
cross-sectional area of each myocyte to its length is shown in Fig.
7. The same curve for cylinders with a
fixed volume of 65,000 µm3 is
drawn with the points. Normal and TR myocytes are clearly well mixed,
again suggesting that TR-induced hypertrophy is not associated with a
change in interatrial band myocyte geometry. It is also clear that the
distribution of points is not well described by the constant-volume
line. However, two qualifications must be made to this statement.
First, it does appear that the upper right-hand margin of
the distribution of points could be fit with a similar curve. Such a
curve might represent the maximal possible myocyte volume, reached by
only a small portion of the myocytes. Second, it must be noted in this
regard that the failure of the points to fall along this line may be
due in part to the division of these myocytes into two or more
populations on the basis of myocyte ploidy (diploid, tetraploid, etc.)
rather than state (normal vs. TR).
View larger version (31K):
[in this window]
[in a new window]
|
Fig. 7.
Mean cross-sectional area
(µm2) vs. myocyte length
(µm) for reconstructed normal (n = 115) and TR (n = 76) interatrial band
myocytes. Each point represents a single myocyte. Curve is that for
cylinders with a volume of 65,000 µm3.
|
|
Variable myocyte ploidy could arise either from variable numbers of
nuclei per myocyte or from variable ploidy per nucleus. To assess these
possibilities, we first isolated myocytes from predominantly the
epimyocardial portion of the interatrial band from two dogs and stained
the myocytes by the Feulgen technique, which is stoichiometric for DNA.
When 300 myocytes from each dog were looked at, 25% were mononuclear
and 75% binuclear in one dog, whereas 59% were mononuclear and 41%
binuclear in the other. DNA content of 200 nuclei was measured for each
dog, revealing that 198 were diploid and 2 tetraploid in one dog and
199 diploid and 1 tetraploid in the other dog. There are clearly two
populations of interatrial band myocytes, diploid and tetraploid, with
varying ploidy almost exclusively due to variable numbers of nuclei per myocyte.
To approach the question of the effect of ploidy on the myocyte
breadth-length relationship, we repeated the reconstruction study with
propidium iodide staining to show the locations and numbers of the
nuclei in the reconstructed myocytes. In two cases, we were successful
in removing the coverslips from the slides and additionally staining
with propidium iodide; in these cases the task of determination of
myocyte mean cross-sectional area did not have to be repeated. In the
other cases, however, new sets of serial sections had to be cut, new
reconstructions were made, and both lengths and myocyte mean
cross-sectional areas were determined along with nuclear counts.
In all, 62 mononuclear and 168 binuclear myocytes were reconstructed.
Lengths (149 ± 53 vs. 206 ± 80 µm;
P < 0.001), cross-sectional areas
(290 ± 86 vs. 304 ± 110 µm2;
P = 0.17), and volumes (41,633 ± 15,470 vs. 57,583 ± 21,535 µm3;
P < 0.001) of the mononuclear
myocytes were lower than those of the binuclear myocytes. Scattergrams
of the two populations are shown in Fig. 8, with the
mononuclear and thus generally diploid myocytes in Fig.
8A and the binuclear and thus
generally tetraploid myocytes in Fig.
8B. The curve in Fig.
8A again represents cylinders with a
volume of 65,000 µm3, whereas
Fig. 8B also includes the curve for
cylinders with twice that volume. All of the mononuclear myocytes fell
close to or below the 65,000 µm3
curve, whereas quite a few of the binuclear myocytes exceeded that
curve, falling close to or below the 130,000 µm3 curve. On the other hand,
many mononuclear and binuclear myocytes fell well beneath the 65,000 µm3 curve.
View larger version (20K):
[in this window]
[in a new window]
|
Fig. 8.
Mean cross-sectional area
(Axs;
µm2) vs. myocyte length (µm)
for mononuclear and binuclear normal interatrial band myocytes.
A: mononuclear myocytes. Curve is for
cylinders with volume of 65,000 µm3.
B: binuclear myocytes. Curves in
B are for cylinders of 65,000 (left) and 130,000 µm3
(right).
|
|
Do interatrial band myocytes get longer in TR-induced right atrial
hypertrophy?
Because it was not possible to randomly select myocytes for
reconstruction inasmuch as relatively few myocytes were studied, and
because there was so much heterogeneity of myocyte length, it was
unclear whether interatrial band myocyte length changed in TR-induced
right atrial hypertrophy. Another approach to this question was made
possible by the measurements of mean cross-sectional area and length of
individual myocytes. The myocytes were grouped into classes 100 µm2 wide, and the mean lengths
of the myocytes in each of those classes were compared. We already know
from the previous morphometric study that interatrial band myocyte
cross-sectional area does not change in TR-induced right atrial
hypertrophy (9). Accordingly, if the normal and TR myocytes within a
given breadth class are of comparable lengths, there could have been no
increase in length. As shown in Fig. 9, this is in fact
the case. For this plot, we pooled the data from all of the
reconstructions for which myocyte mean cross-sectional area was
determined (274 normal and 76 TR myocytes). As before, there is a trend
for narrower myocytes to be long and broader myocytes to be short. The
normal and right atrial hypertrophy length values in each class are
very similar except for the narrowest myocytes, for which the numbers
of myocytes are very small. Overall, these data suggest that there is
no length hypertrophy in right interatrial band myocytes in tricuspid
regurgitation animals.
View larger version (30K):
[in this window]
[in a new window]
|
Fig. 9.
Mean length (µm) vs. cross-sectional area
(µm2) of normal vs. TR right
interatrial band myocytes. Lengths of normal and TR myocytes in 6 mean
Axs classes are
plotted side-by-side. Bars indicate SEs.
|
|
| |
DISCUSSION |
In our previous study, we found that myocytes in the right atrial
portion of the interatrial band showed no change in cross-sectional area with right atrial hypertrophy (9). Here we have used
reconstructions to demonstrate that these interatrial band myocytes
also show no change in length with right atrial hypertrophy. Right
atrial appendage myocytes, on the other hand, show roughly
proportionate changes in cross-sectional area and length. Some of the
interatrial band myocytes may fail to hypertrophy because they are
already as large as is possible, perhaps constrained by their ploidy. For others, an alternative explanation must be sought.
The apparent failure of interatrial band myocytes to increase in length
or breadth is consistent with our previous demonstration of their
failure to increase in girth (9). Although our sample size was
sufficient for a demonstration of increased length and breadth of right
atrial free wall myocytes, it could be considered too small to conclude that there was
definitely no hypertrophy of interatrial band myocytes. Nonetheless,
Fig. 7 gives very little reason to believe that a larger sample size
would reveal interatrial band myocyte hypertrophy in TR dogs. Indeed,
Fig. 7 suggests only the possibility of a modest hypertrophy in
breadth, contradicted by our earlier and larger study.
Determination of myocyte lengths by serial section reconstruction.
For our purposes there were two problems with the use of isolated
myocytes for the determination of interatrial band myocyte size. First,
because there are no macroscopic landmarks marking the border between
the larger epimyocardial myocytes and the smaller endomyocardial
myocytes, samples taken for myocyte isolation would likely include a
variable percentage of the smaller myocytes. If the failure of
epimyocardial myocytes to hypertrophy is indeed due to their having
reached a maximum possible myocyte size, then the inclusion of these
smaller myocytes would constitute a potentially serious problem.
Second, those interatrial band myocytes that are very long or very
broad may not be as easily isolated without damage as other myocytes.
Accordingly, we believed it was necessary to use an in situ method for
these studies. The reconstruction method, although laborious, enables
the determination of the dimensions and volumes of individual myocytes
and obviates potential problems of selective myocyte loss during
myocyte isolation. The method works well, as shown by the close
correspondence between values obtained for left ventricular myocyte
lengths by the reconstruction and myocyte isolation methods. Several
difficulties remain, especially that of nonrandom selection imposed by
the necessity of being able to recognize the site to be photographed in
section after section. We have since found that this problem is much
reduced by the use of video microscopy, with which the last section
collected can be compared with the next to be collected to ensure that
precisely the same area is captured.
Why do interatrial band myocytes fail to hypertrophy?
Interatrial band myocytes do not increase in cross-sectional area (9)
nor in length (this study) in TR-induced right atrial hypertrophy,
whereas right atrial appendage myocyte cross-sectional area and length
increased markedly and approximately proportionately. There are two
broad plausible reasons for the failure of interatrial band myocytes to
hypertrophy: that they are not subjected to elevated stresses in this
model or that the myocytes are incapable of responding to hypertrophic
stimulation. Regarding the first reason, there are no
physiological data to evaluate the proposition that the right atrial
portion of the interatrial band is not subject to elevated stresses in
TR-induced right atrial hypertrophy, but it seems unlikely given the
position of this thickened muscle bundle. A failure to respond to
hypertrophic stimulation could have several sources, most relating to
signal transduction and intracellular signaling. Hypertrophic stimuli
produce a cascade of effects, beginning with some form of signal
transduction at the plasmalemma, then involving complex intracellular
signaling, transcriptional activation, translation, and finally myocyte
growth. A defect at any step along this cascade could limit myocyte
size. Plausible defects include interatrial band-specific deficits in some signaling pathway(s), and size- and/or ploidy-related
diminution of signaling intensity.
Are interatrial band myocytes in any way fundamentally different from
other atrial myocytes and thus plausibly different in their signaling
pathways? The interatrial band, or Bachmann's bundle, has been
suggested to be part of an atrial conduction system (30). However, the
evidence for this proposition is not convincing. First, although
interatrial impulse conduction along the band is indeed rapid, this is
likely explained by the greater breadth of its constituent myocytes
(11) and its origination near the sinoatrial node. Second, although
early morphological studies emphasized the Purkinje-like appearance of
the myocytes, by which was meant that the myocytes were poor in
organized myofilaments, subsequent studies showed that with adequate
fixation the myocytes are in fact rich in organized myofilaments (11,
33). Third, the defining characteristic feature of all myocytes in the
ventricular conduction system is the absence of a transverse tubule
system (33). However, interatrial band myocytes are richer in
transverse tubules than any other atrial myocytes (9), albeit not as
richly endowed as working ventricular myocytes. Further, the absence of
transverse tubules from some myocyte profiles does not indicate the
absence of transverse tubules from the entire myocyte (see Fig. 5).
Even if interatrial band myocytes were part of an atrial conduction
system, the question as to why the myocytes fail to respond to
hypertrophic stimulation would remain, because myocytes in the
ventricular conduction system are in fact capable of hypertrophy (20,
25). Thus, if there is a deficit in signaling in interatrial band
myocytes, it is unlikely to be due to their being specialized atrial
conduction system myocytes.
Size alone could account for decreased signaling intensity. Signals
arising at the sarcolemma could be diluted in enlarged myocytes,
particularly if the myocyte surface-to-volume ratio was not maintained.
Although surface-to-volume ratio appears to be maintained in the
hypertrophied ventricular myocyte by proliferation of transverse
tubules (26), transverse tubules do not proliferate during atrial
myocyte hypertrophy (9), and because of this the myocyte
surface-to-volume ratio falls. This could produce a
diminution of membrane-associated elements of the signal transduction pathways (e.g., phospholipase C) relative to myocyte volume and consequent dilution of the response to hypertrophic stimulation in
otherwise normal myocytes. However, this probably does not have a great
impact on the potential for myocyte hypertrophy, given that right
atrial free wall myocytes grow to nearly the size of interatrial band
myocytes in right atrial hypertrophy despite a relative paucity of
transverse tubules (9) and thus a lower surface-to-volume ratio.
Furthermore, increasing myocyte volume and decreasing surface-to-volume
ratio cannot be the sole determinants of a limit to myocyte
hypertrophy, given that binuclear myocytes can attain greater volumes
than mononuclear myocytes (Fig. 8). However, they could yield an
apparent decreased sensitivity to hypertrophic stimulation. Such a
decreased sensitivity has been reported for rat ventricular myocytes
that are enlarged because of pressure overload (29) or aging (16).
It is likely that the conjoint effects of size and ploidy are involved
in setting the upper limits on myocyte size. Cardiomyocyte size and
ploidy are closely related in human cardiac hypertrophy. Therefore,
markedly enlarged myocytes with correspondingly markedly elevated
ploidy levels are found in hypertrophy in the atria (27), ventricles
(1, 35), and ventricular conduction system (27). The presence of larger
myocytes with higher ploidy might indicate either that hypertrophic
stimulation causes increased ploidy or that preexisting higher ploidy
permits greater hypertrophy in response to hypertrophic stimulation.
Recent work favors the proposition that myocyte size depends on the
ploidy level, which is determined near puberty (5-7). The
different degrees of hypertrophy present in myocytes of different
ploidy thus in part reflect the different degrees of hypertrophy
possible. Similarly, it has been noted that rat ventricular myocytes,
which like dog atrial myocytes are very rarely more than tetraploid,
never hypertrophy beyond ~85% in volume; human ventricular myocytes
with their often much higher ploidy may hypertrophy beyond 300% (34).
This effect of ploidy is probably tied to gene dosage, i.e., the number
of copies of the gene available for transcription, which is an
important factor in determining the rate of mRNA production. Although
contractile proteins are needed in large amounts in hypertrophying myocytes, there are single copies of the genes for any of their isoforms within the diploid genome (12). Gene dosage can be increased
by increasing ploidy, and polyploid cells do show a higher level of
transcription and increased size (8, 21, 24). If there were appreciable
barriers to mRNA diffusion or if mRNA stability were low, the
concentration of mRNA within the cytoplasm could fall off rapidly, as
has been reported for cardiac myocytes (28). The shape of the mRNA
concentration gradient could then regulate myocyte size by determining
where the rate of protein synthesis could not outstrip that of protein
degradation (18, 19).
In conclusion, the failure of interatrial band myocytes to enlarge in
TR-induced canine right atrial hypertrophy may be a consequence of
their already large size in normal dogs, which alone may reduce
signaling intensity after hypertrophic stimulation. Furthermore, as the
ratio of myocyte volume to myocyte ploidy drops with increasing myocyte
volume, the myocyte may reach a maximal volume because the rate of
protein synthesis at the myocyte periphery cannot support further
myocyte growth. However, the volumes of some interatrial band myocytes
are far below the maximum volume shown to be possible at their ploidy
level. This indicates that other factors must be at work in preventing
the hypertrophy of interatrial band myocytes.
| |
ACKNOWLEDGEMENTS |
We are very grateful to Larry D. Moy and Craig Silberberg for
exceptional technical work and to Marie B. Anderson for bibliographic research. We are grateful also to Dr. George Cooper IV and Mary Barnes
of the Charleston, SC, Veterans Affairs Medical Center, who generously
taught us their technique for cardiomyocyte isolation.
| |
FOOTNOTES |
This work was supported in part by National Heart, Lung, and Blood
Institute Grant HL-18468, by the Duke University Medical Center Small
Grants Program, and by the Research and Development Service of the
Department of Veterans Affairs Medical Center (Durham, NC).
Address for reprint requests: P. C. Dolber, Div. of Cardiology, Dept.
of Medicine, Box 3475, Duke Univ. Medical Center, Durham, NC 27710.
Received 16 May 1997; accepted in final form 16 June 1998.
| |
REFERENCES |
1.
Adler, C. P. Histochemically determinable changes
in cardiac insufficiency and their functional significance.
Basic Res. Cardiol. 81, Suppl. 1: 179-192, 1986.
2.
Bachman, G.
The inter-auricular time interval.
Am. J. Physiol.
41:
309-320,
1916.
3.
Bauman, R. P.,
J. C. Rembert,
and
J. C. Greenfield, Jr.
Vascular responses of right atria to chronic tricuspid regurgitation (Abstract).
Clin. Res.
40:
399A,
1992.
4.
Bishop, S. P.,
and
J. L. Drummond.
Surface morphology and cell size measurement of isolated rat cardiac myocytes.
J. Mol. Cell. Cardiol.
11:
423-433,
1979[Medline].
5.
Brodsky, V. Y.,
A. L. Chernyaev,
and
I. A. Vasilyeva.
Variability of the cardiomyocyte ploidy in normal human hearts.
Virchows Arch. B Cell Pathol.
61:
289-294,
1991[Medline].
6.
Brodsky, V. Y.,
D. S. Sarkisov,
A. M. Arefyeva,
and
N. W. Panova.
DNA and protein relations in cardiomyocytes. Growth reserve in cardiac muscle.
Eur. J. Histochem.
37:
199-206,
1993[Medline].
7.
Brodsky, V. Y.,
D. S. Sarkisov,
A. M. Arefyeva,
and
N. W. Panova.
Polyploidy in cardiac myocytes of normal and hypertrophic human hearts: range of values.
Virchows Arch.
424:
429-435,
1994[Medline].
8.
Brodsky, V. Y.,
and
I. V. Uryvaeva.
Genome Multiplication in Growth and Development. Cambridge, UK: Cambridge University Press, 1985.
9.
Dolber, P. C.,
R. P. Bauman,
J. C. Rembert,
and
J. C. Greenfield, Jr.
Regional changes in myocyte structure in a model of canine right atrial hypertrophy.
Am. J. Physiol.
267 (Heart Circ. Physiol. 36):
H1279-H1287,
1994[Abstract/Free Full Text].
10.
Dolber, P. C.,
E. C. Beyer,
J. L. Junker,
and
M. S. Spach.
Distribution of gap junctions in dog and rat ventricle studied with a double-label technique.
J. Mol. Cell. Cardiol.
24:
1443-1457,
1992[Medline].
11.
Dolber, P. C.,
and
M. S. Spach.
Structure of canine Bachmann's bundle related to propagation of excitation.
Am. J. Physiol.
257 (Heart Circ. Physiol. 26):
H1446-H1457,
1989[Abstract/Free Full Text].
12.
Emerson, C. P., Jr.,
and
S. I. Bernstein.
Molecular genetics of myosin.
Annu. Rev. Biochem.
56:
695-726,
1987[Medline].
13.
Gerdes, A. M.,
and
F. H. Kasten.
Morphometric study of endomyocardium and epimyocardium of the left ventricle of dogs.
Am. J. Anat.
159:
389-394,
1980[Medline].
14.
Hewett, K.,
M. J. Legato,
P. Danilo, Jr.,
and
R. B. Robinson.
Isolated myocytes from adult canine left ventricle: Ca2+ tolerance, electrophysiology, and ultrastructure.
Am. J. Physiol.
245 (Heart Circ. Physiol. 14):
H830-H839,
1983.
15.
Huebner, G.
Electron microscopic investigation of cardioplegia. Electron microscopy of various forms of cardiac arrest in correlation with myocardial function.
Methods Achiev. Exp. Pathol.
5:
518-539,
1971[Medline].
16.
Isoyama, S.
Age-related changes before and after imposition of hemodynamic stress in the mammalian heart.
Life Sci.
58:
1601-1614,
1996[Medline].
17.
Jacobson, S. L.
Culture of spontaneously contracting myocardial cells from adult rats.
Cell Struct. Funct.
2:
1-9,
1977.
18.
Johnson, H. A.
Adaptive growth without mitosis.
In: Cardiac Hypertrophy, edited by N. R. Alpert. New York: Academic, 1971, p. 21-23.
19.
Kohn, R. R.
Some thoughts on adaptive growth of heart muscle.
In: Cardiac Hypertrophy, edited by N. R. Alpert. New York: Academic, 1971, p. 19-20.
20.
Kuribayashi, T.,
K. Furukawa,
H. Katsume,
H. Ijichi,
and
Y. Ibata.
Regional differences of myocyte hypertrophy and three-dimensional deformation of the heart.
Am. J. Physiol.
250 (Heart Circ. Physiol. 19):
H378-H388,
1986.
21.
Laurent, G. J.,
M. P. Sparrow,
and
D. J. Millward.
Turnover of muscle protein in the fowl: changes in rates of protein synthesis and breakdown during hypertrophy of the anterior and posterior latissimus dorsi muscles.
Biochem. J.
176:
407-417,
1978[Medline].
22.
Lewin, B.
Genes V. Oxford, UK: Oxford University Press, 1994.
23.
Lexell, J.,
and
C. Taylor.
Smoothed histograms: a visual aid for the analysis of distributions of muscle fiber areas.
Muscle Nerve
14:
826-828,
1991[Medline].
24.
Nagl, W.
Polyploidy in differentiation and evolution.
Int. J. Cell Cloning
8:
216-223,
1990[Abstract].
25.
Okada, R.,
and
K. Fukuda.
A histopathological study on the hypertrophy of the atrioventricular conduction system.
Jpn. Circ. J.
45:
441-445,
1981[Medline].
26.
Page, E.,
and
L. P. McCallister.
Quantitative electron microscopic description of heart muscle cells. Application to normal, hypertrophied and thyroxin-stimulated hearts.
Am. J. Cardiol.
31:
172-182,
1973[Medline].
27.
Rumyantsev, P. P.,
I. L. Erokhina,
E. M. Antipanova,
and
M. G. Martynova.
DNA and sex chromatin content in nuclei of conductive system and working myocytes of normal and hypertrophied heart.
Acta Histochem. Suppl.
39:
225-237,
1990[Medline].
28.
Russell, B.,
and
D. J. Dix.
Mechanisms for intracellular distribution of mRNA: in situ hybridization studies in muscle.
Am. J. Physiol.
262 (Cell Physiol. 31):
C1-C8,
1992[Abstract/Free Full Text].
29.
Schunkert, H.,
L. Jahn,
S. Izumo,
C. S. Apstein,
and
B. H. Lorell.
Localization and regulation of c-fos and c-jun protooncogene induction by systolic wall stress in normal and hypertrophied rat hearts.
Proc. Natl. Acad. Sci. USA
88:
11480-11484,
1991[Abstract/Free Full Text].
30.
Sherf, L.,
and
T. N. James.
Fine structure of cells and their histologic organization within internodal pathways of the heart: clinical and electrocardiographic implications.
Am. J. Cardiol.
44:
345-369,
1979[Medline].
31.
Silverman, B. W.
Density Estimation for Statistics and Data Analysis. New York: Chapman and Hall, 1986.
32.
Sommer, J. R.
To cationize glass (Abstract).
J. Cell Biol.
75:
245A,
1977.
33.
Sommer, J. R.,
and
P. C. Dolber.
Cardiac muscle: ultrastructure of its cells and bundles.
In: Normal and Abnormal Conduction in the Heart, edited by A. Paes de Carvalho,
B. F. Hoffman,
and M. Lieberman. Mount Kisco, NY: Futura, 1982, p. 1-28.
34.
Vliegen, H. W.,
A. V. Bruschke,
and
A. Van der Laarse.
Different response of cellular DNA content to cardiac hypertrophy in human and rat heart myocytes.
Comp. Biochem. Physiol. A Physiol.
95:
109-114,
1990.
35.
Vliegen, H. W.,
A. Van der Laarse,
C. J. Cornelisse,
and
F. Eulderink.
Myocardial changes in pressure overload-induced left ventricular hypertrophy. A study on tissue composition, polyploidization and multinucleation.
Eur. Heart J.
12:
488-494,
1991[Abstract/Free Full Text].
36.
Zellner, J. L.,
F. G. Spinale,
D. M. Eble,
K. W. Hewett,
and
F. A. Crawford, Jr.
Alterations in myocyte shape and basement membrane attachment with tachycardia-induced heart failure.
Circ. Res.
69:
590-600,
1991[Abstract/Free Full Text].
Am J Physiol Heart Circ Physiol 275(4):H1225-H1235
0002-9513/98 $5.00
Copyright © 1998 the American Physiological Society
Top
Abstract
Introduction
