Coiled Perimysial Fibers of Papillary Muscle in Rat Heart ...

Coiled perimysial fibers of papillary muscle in rat heart: morphology, distribution, and
changes in configuration.
T F Robinson, M A Geraci, E H Sonnenblick and S M Factor
Circ Res. 1988;63:577-592
doi: 10.1161/01.RES.63.3.577
Circulation Research is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 1988 American Heart Association, Inc. All rights reserved.
Print ISSN: 0009-7330. Online ISSN: 1524-4571
The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://circres.ahajournals.org/content/63/3/577
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published in
Circulation Research can be obtained via RightsLink, a service of the Copyright Clearance Center, not the
Editorial Office. Once the online version of the published article for which permission is being requested is
located, click Request Permissions in the middle column of the Web page under Services. Further information
about this process is available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Circulation Research is online at:
http://circres.ahajournals.org//subscriptions/
Downloaded from
http://circres.ahajournals.org/ by guest on March 23, 2013

Coiled Perimysial Fibers of Papillary Muscle
in Rat Heart: Morphology, Distribution, and
Changes in Configuration
Thomas F. Robinson, Maria A. Geraci, Edmund H. Sonnenblick, and Stephen M. Factor
The morphology, distribution, and configuration of coiled perimysial fibers of rat heart
papillary muscle were studied. Methods included bright-field light microscopy of silver-stained
sections, scanning and transmission electron microscopy, and differentia] interference contrast
light microscopy of unfixed and unstained specimens. Coiled fibers, elliptical in cross section,
are arranged hi a branched network that diverges from the muscle-tendon junction and is
continuous throughout the length of the muscle and into the ventricle wall. Most fibers range hi
diameter from less than 1 fim to 10 fim and are parallel with the long axis of the muscle, although
branching is common and oblique orientations are seen. Several myocytes are associated with
each coiled perimysial fiber. Constituent fibrils (diameter, 40-50 nm) occur in bundles twisted
within the fiber. Small satellite elastic fibers are parallel to the collagen fiber axes. Stereo analysis
of the coiled perimysial fibers reveals helical configurations, as opposed to planar waviness, that
become less convoluted or even straighten as the resting muscle is stretched. Calculations based
on cross-sectional areas of fibers, changes in fiber configurations, and tensile moduli reported for
collagen fibers of other tissues show that the potential tensile strength of the network of coiled
perimysial fibers is sufficient to contribute significantly to the mechanical properties of papillary
muscle. Detailed evaluations of possible roles of the coiled perimysial collagen fiber system as a
function of passive stretch and contraction hi ventricular wall, as well as in papillary muscle,
warrant further study. (Circulation Research 1988;63:577-592)
Connective tissue in skeletal and cardiac muscle
is organized in three levels. The epimysium
is the sheath of connective tissue that
surrounds the entire muscle. The perimysium is
associated with groups of cells, and the endomysium
surrounds and interconnects individual cells. 1
In skeletal muscle, the perimysium has been classified
further to include three constituent fibrous components:
coarse, "crimped" fibers arranged in a criss-
From the Departments of Medicine (T.F.R., M.A.G., E.H.S.,
and S.M.F.); Physiology and Biophysics (T.F.R. and M.A.G.);
Pathology (S.M.F.), The Albert Einstein College of Medicine,
Bronx, New York.
Supported in part by National Institutes of Health Research
Grants HL-24336 (T.F.R.) and Mr. And Mrs. Robert S. Krauser
Grant-in-Aid from the American Heart Association, New York
City Affiliate and Heart Fund (T.F.R.); National Institutes of
Health Grant HL-18824 (E.H.S., S.M.F., and T.F.R.); and
National Institute on Aging Grant PO1-AG 05554 (S.S. and
T.F.R.). The JEM 1200EX was obtained by the Analytical
Ultrastructure Center of AECOM with equipment grants from
NSF and NIH Biotechnology Resources.
Address for correspondence and proofs: Thomas F. Robinson,
PhD, Cardiovascular Research Laboratories, Forchheimer Building,
Room G42, Albert Einstein College of Medicine, 1300
Morris Park Avenue, Bronx, NY 10461.
Received October 13, 1987; accepted April 5, 1988.
577
cross pattern; a loose feltwork of noncrimped fibrils;
and fine, noncrimped bundles of fibrils with no
directional organization. 2
In cardiac trabeculae and papillary muscles, the
epimysium 34 and endomysium 3 - 16 have been structurally
characterized. The perimysium has not been
extensively studied, but it is not so highly organized
as that of skeletal muscle. Groups of cardiomyocytes
are surrounded by weaves of collagen fibers
that are, in turn, interconnected by large collagen
fibers. 8 In the rodent myocardium, other large
perimysial structures imaged by electron microscopy
have been described as "tendonlike" fibers 8
and "sheets." 9 In papillary muscle and Bachmann's
bundle in dog and rabbit hearts, light microscopy of
sections stained with a modified picrosirius red
procedure has revealed "rodlike fibers" and "collagenous
septa." 17
In the present study, a variety of structural techniques
were used to describe and analyze the morphology,
distribution, and changes in configurations
of the branched network of coiled perimysial fibers
(CPFs) that courses from the muscle-chorda tendinea
junction, through the papillary muscle, and into
the wall of rat heart. The same parameters were
Downloaded from
http://circres.ahajournals.org/ by guest on March 23, 2013

578 Circulation Research Vol 63, No 3, September 1988
FIGURE 1. Darkfield stereomicroscope photograph of fixed, isolated, cylindrical right ventricular papillary muscle. Bar,
500 fim; Ch, chorda tendinea; J, muscle-tendon junction; m, central region of muscle; LC, lateral collagen cord. Original
magnification, x44.
also investigated for connective tissue fibers in the
chorda tendinea. The numerous CPFs of papillary
muscle are large relative to the size of myocytes.
The functional implications of the structural characteristics
of this impressive, branched, continuous
array of collagen fibers are discussed for both in
vivo and in vitro function in view of the extensive
use of the isolated papillary muscle as a model of
myocardial contraction. 18
Materials and Methods
Tissue Preparation
Twenty-nine rat hearts were used specifically for
these experiments (photographs from previously
studied hearts were also reexamined). Sixteen young
adult (15 male and one female) Wistar rats (4-6
months old, 250-300 g) were used, and 10 male
Fisher rats (strain 344, barrier reared at the National
Institute on Aging; either 3.5 or 23 months old) were
used. Three Sprague-Dawley rats (heart weights,
250-300 g) were used.
The rats were anesthetized with ether, and their
hearts were quickly removed and placed in oxygenated
Tyrode's solution (270-275 mosm) with the
following composition (rnM): Na + 151.3, Ca 2+ 2.4,
K + 4.0, Mg 2+ 0.5, Cl" 147.3, H2PO4~ 12.0, and
glucose 5.5. After the hearts in oxygenated Tyrode's
solution were clear of blood, they were placed in
high potassium (30 rnM) Tyrode's solution until
beating ceased. This solution was maintained at
30° C and bubbled with 95% O2-5% CO2.
Both left and right ventricular papillary muscles
were used. Papillary muscles examined with stereomicrophotography
were from the right ventricle and
were exposed by cutting the anterior border of free
right wall and folding it back. Papillary muscles
ranged in configuration from fan shaped to cylindrical,
and the muscle-tendonjunctions varied in shape,
as did the occurrence of collagen cords (third-order
chordae tendineae) attached to the sides of muscles.
The appearance of the CPF network depended on
the overall appearance of the muscles. The present
study describes only CPF networks in cylindrical
muscles (Figure 1).
Some isolated papillary muscles were fixed at
slack length, whereas others were stretched before
fixation. The muscles were continuously superfused
with Tyrode's solution (pH, 7.2; temperature, 30° C)
and bubbled with 95% Or5% CO2. The nontendinous
ends of the papillary muscles were inserted
Downloaded from
http://circres.ahajournals.org/ by guest on March 23, 2013

REST
Robinson et al Coiled Perimysial Fibers of Papillary Muscle
MODEL CONSIDERATIONS FOR A SIMPLE COILED FIBER (h«lix)
fOYTTX /
TTD
Eitannon
Example. Fiber Length itays constant, only coiled configuration changes
Convolution
Indei:
Change tn P
From REST:
Pouible
Sarcomere
Length (M*0 :
D 16 -u% trr
1.4
1 2 •IU
FIGURE 2. Diagrams that show method for calculation of convolution index (CI) and that show changes in pitch (P),
configuration, and CI for a uniformly coiled thin fiber of constant length as it is stretched or compressed from rest. P,
distance between two corresponding points along the coil; FL, fiber length; D, diameter of coil. CI=FLJP;
FL=[(3.14xD)'+P 2 ] m . (These diagrams demonstrate the geometric changes for helical fibers that are very thin relative
to D. Most coiled perimysial fibers, however, are actually thick relative to D. The range and analysis of configurational
changes ofCPFs is thus similar but more complex than that shown.)
into a collet, which was attached to a micrometer
that was used to make external changes in length,
and the tendinous ends were tied with 5-0 braided
silk (Ethicon, Somerville, New Jersey). Muscles
were fixed at slack length or at approximately 15%
stretch beyond slack length, as described
previously. 13
For studies of continuity of the CPF branched
network with the CPF network of the ventricle wall,
three hearts were mounted on a perfusion apparatus
linked to a manometer and were infused with bariumgelatin
at 30 mm Hg. Their aortas were ligated, and
the hearts were fixed for silver staining.
Wedges of living myocardium were cut with a
razor blade from living, unfixed, unstained papillary
muscle and ventricular wall and were immersed in
oxygenated relaxing solution [high potassium (30
mM) Tyrode's solution]. The moist, thin ends of the
wedges were mounted on slides in drops of Tyrode's
solution and quickly viewed in the light microscope
by means of differential interference contrast in a
manner analogous to that used by Broom 19 to view
wet porcine valve leaflets.
Fixation
Muscles fixed at either slack or stretched lengths
were fixed by substitution of Tyrode's buffer with
Tyrode's buffer that contained either 3.7% formaldehyde
(for subsequent silver staining) or 6% glutaraldehyde
(for subsequent ultrastructural studies).
Hearts filled with barium-gelatin were sectioned
in a cryostat microtome at 100-120-j/.m thickness.
2 0
>2 2
579
Fixation for silver staining was performed by
immersion of papillary muscles in 3.7% buffered
formaldehyde for a minimum of 2 weeks.
Fixation for ultrastructural studies was performed
by immersion of papillary muscles in 6%
glutaraldehyde in Tyrode's for 3 hours at room
temperature or overnight at 4° C, rinsing several
times in buffer, postfixation in 1% osmium tetroxide
for 1 hour, and rinsing in buffer. Some muscles were
fixed with 0.5% tannic acid in the primary
fixative. 311 For muscles that were analyzed by
stereomicrophotography, fixation was performed in
situ with minimum delay after photography.
Silver Staining
The formaldehyde-fixed muscles were prepared
as described previously 3 ; longitudinal cryosections
were cut at a thickness of 80 /xm, and the floating
sections were stained with a modification of the
silver-staining technique of Hasagawa and Ravens, 20
which in turn was a modification of the silver
carbonate method of Del Rio Hortega. 21
Stereomicrophotography
Stereomicrophotography was performed with a
Zeiss SV-8 stereomicroscopc (Zeiss, Thornwood,
New York) fitted with an MC-63 camera and Kodak
Tri-X, Panatomic X, or Ektachrome 400 film.
Light Microscopy
Light microscopy of silver-stained sections was
performed on a Zeiss Photomicroscope III, and
Downloaded from
http://circres.ahajournals.org/ by guest on March 23, 2013

580 Circulation Research Vol 63, No 3, September 1988
FIGURE 3. Light micrograph of silver-stained, 100-pjn-thick, longitudinal section of left ventricle at low magnification.
Branched network of coiled perimysial fibers is continuous through the papillary muscle (P)from muscle-tendon junction
(J; see Figure 4) to ventricle wall (W). Individual branched coiled perimysial fibers can be traced over distances
equivalent to many myocyte lengths (arrowheads). (Average myocyte length, ~100 \un). Field represented in micrograph
is 5,600x4,000 fim.
micrographs were taken with Kodachrome 25 film.
Differential interference contrast light microscopy
was performed with a Zeiss WL microscope, and
micrographs were taken with Kodak Technical Pan
film. Calibration micrographs were taken with a
slide graticule (Graticules Ltd, Tonbridge, UK) or
hemocytometer. Optical sectioning was performed
to trace the branched network of CPFs through the
papillary muscle and into the ventricle wall (on
cryosections of fixed, barium-gelatin-infused hearts).
Electron Microscopy
Fixed and rinsed muscles were dehydrated through
a graded ethanol series.
For scanning electron microscopy, dehydrated
samples were critical point-dried in carbon dioxide
in a Samdri Model 790 dryer (Tousimis, Rockville,
Maryland), equilibrated in liquid nitrogen, fractured
with a precooled razor blade, and dried by evaporation
of the nitrogen under vacuum. Samples were
then mounted on stubs, coated with gold-palladium,
and viewed in a JSM 25.
For transmission electron microscopy, dehydrated
samples were infiltrated with rotation in
Spurr's resin 22 or Embed 812 (EMS, Fort Washington,
Pennsylvania), embedded, and cured in an
oven at 70° C. Sections were cut with a diamond
knife (Diatome USA, Fort Washington, Pennsylvania)
on an LKB Ultratome III (Rockville, Maryland)
or a Reichert Ultracut E (Warner Lambert,
Edison, New Jersey), collected on copper grids,
stained with 1% uranyl acetate in 40% ethanol-
60% water and in Reynold's aqueous lead citrate, 23
and viewed in a JEOL 100CX or JEOL 1200EX
electron microscope. Calibration micrographs were
taken with a carbon grating replica (54,864 lines/in.;
Ladd Research, Burlington, Vermont).
Measurements
Diameters and cross-sectional areas of CPFs and
myocytes were estimated from electron micrographs
of sections cut from blocks oriented for
cross sections of the papillary muscles; perpendicularity
of the sections was affirmed by the appear-
Downloaded from
http://circres.ahajournals.org/ by guest on March 23, 2013

Robinson et al Coiled Perimysial Fibers of Papillary Musde 581
100H m
FIGURE 4. Light micrograph of silver-stained section of muscle-tendon junction of left ventricular papillary muscle.
Convoluted collagen fibers of perimysium and epimysium are stained with silver. Coiled perimysial fibers (CPFs) are
large relative to myocyte size, and the CPF network extends from muscle-tendon junction through entire muscle. CPFs
are highly coiled at slack length of muscle, and are interconnected by struts to other CPFs. Branches of CPFs
(arrowheads) vary from parallel to oblique to muscle axis and have been followed over many myocyte lengths by optical
sectioning. EC, epimysial collagen fibers; S, struts; bar, 100 fim. Original magnification, x/90.
ance of the myofilaments. The apparent diameters
of fibers in muscle cross sections are larger than
their intrinsic diameters because CPFs are helical
and the coiled fiber is always cut oblique to its own
axis. To minimize overestimation of fiber diameters
because of their coiled configurations, many of the
measurements were made from stretched muscles,
whose CPFs have convolution indexes lower than
those at slack length. The straightened fibers are,
however, still twisted like a stretched, coiled telephone
cord. In addition, measurements were made
of fibers in which most fibrils were perpendicular to
the plane of section.
Sarcomere lengths were averaged from total
lengths of eight to 15 sarcomeres. Measurements
were made from calibrated prints of Kodachrome
micrographs of midmuscle in regions immediately
adjacent to CPFs. No correction factor was applied
for possible shrinkage during sample preparation.
Convolution indexes, diameters, and pitches of
helices. The actual CPF thickness (d) is as much as
approximately one third to one half of the diameter
of the coil (D). The precision of a convolution index
(CI) estimate is thus more difficult to define than in
the simple case of a very thin fiber wound around a
relatively large hollow core, as shown in Figure 2.
Therefore, CIs were estimated from one half the
diameters of coils (D/2), where D is measured as the
peak-to-peak amplitude of the coil and the pitch (P)
is the coil length divided by the turn, according to
the Pythagorean theorem. This can be visualized as
the opening and flattening of a cylinder constructed
around a regular coil to become a rectangle, which
is divided into two right triangles by their shared
hypotenuse. 24 Therefore, in accordance with the
definitions in Figure 2,
and
= [(3.14D) 2 + P 2 ]
CI = FL/P
Results
A branched, continuous network of perimysial
and epimysial fibers extends from the muscle-
Downloaded from
http://circres.ahajournals.org/ by guest on March 23, 2013
1/2

582 Circulation Research Vol 63, No 3, September 1988
tendon junction, through the length of papillary
muscle, and into the ventricle wall, as seen most
eflFectively by light microscopy of silver-impregnated
sections (Figure 3). The most striking features of
these perimysial fibers are their highly coiled configurations
at slack length, their large size relative to
myocytes, and the extent to which their array
appears to form an orderly framework, described
here for cylindrical papillary muscles (Figure 1). In
single, low-magnification micrographs (Figure 3),
the continuity of branched CPFs can be followed
over distances equal to many myocyte lengths. At
higher magnifications, the CPF network is seen to
originate in the chorda tendinea, from which CPFs
diverge and branch (Figure 4). The branching and
continuity of the CPFs have been followed over
distances of many cell lengths by means of optical
sectioning within the 100-/xm-thick sections.
The overall appearance of CPFs is intrinsic and
not the result of fixation or staining artifact. Figure
5 is a differential interference contrast light micrograph
of a wedge of wet, unfixed, unstained myocardium.
CPFs can be studied with differential
interference contrast microscopy or with polariza-
CPF
tion microscopy because of the birefringent nature
of collagen fibrils. The images reveal the same
highly convoluted configurations seen with light
microscopy of silver-stained tissue as well as with
ultrastructural techniques.
CPFs range in diameter from less than 1 /im to
several microns and are distributed across the muscle
in a ratio of one CPF: several myocytes (Figure
6). The cross-sectional profiles of the CPFs are
varied and range from slightly elliptical to very
elliptical; in some cases they are even ribbonlike.
CPFs are oriented in a primarily axial manner. The
ribbonlike CPFs are distinguished from other ribbonlike,
collagenous fibers that are perpendicular to
the long axis of the muscle and are not helical but
are wavy and often follow the contours of underlying
cells. 16 The degree to which the apparent eccentricity
of elliptical cross-sectional profiles is a function
of the coiled nature of the CPF rather than the
fiber's intrinsic shape depends on the thickness of
the fibers and the pitch of the coil. Higher power
transmission electron micrographs of cross sections
reveal constituent fibrils on the order of 40-50 nm
in diameter, often grouped in bundles of two or
40pm
FIGURE 5. Differential interference contrast light micrograph of wet, unfixed, unstained, birefringent coiled perimysial
fiber (CPF) in situ in thin end of cut wedge of left ventricular papillary muscle. Convoluted path of fiber is apparent from
alternate light and dark regions (arrow). Pitch of coiled perimysial fiber is same order of magnitude as that observed in
fixed and stained samples. Bar, 40 \im. Original magnification, y.533.
Downloaded from
http://circres.ahajournals.org/ by guest on March 23, 2013

Robinson et al Coiled Perimysial Fibers of Papillary Muscle 583
FIGURE 6. Low-magnification transmission electron micrograph of cross-section of central region of right ventricular
papillary muscle. Coiled perimysial fibers (CPF) are several microns in diameter and are interspersed among and closely
opposed to myocytes. M, myocyte; *, CPF; bar, 10 \tm. Original magnification, y.1,900.
three subfibers (Figure 7). The orientation of fibrils
varies across the fiber: some fibrils are in perfect
cross-sectional alignment, whereas neighboring
fibrils are oblique or even longitudinal.
Elastic fibers are associated with each CPF (Figure
7). They run parallel with the axis of the CPF and
are composed of amorphous and microfibrillar components.
Fibroblast processes are visible in most of
the random and serial photographs of CPFs and are
often rich in rough endoplasmic reticulum, Golgi
apparatus, and secretory vesicles. The cell processes
wrap around the CPF, as might be expected
from recent studies on synthesis of collagen types I
and III by myocardial fibroblasts. 25 The extracellular
space enclosed by the processes is less electrondense
than that outside the processes. This observation
was made in muscles that had undergone
primary fixation in solutions containing only glutaraldehyde
as well as in solutions that contained
tannic acid or the cationic dyes ruthenium red,
safranin-O, or Alcian blue. 3
The CPFs are highly convoluted, but the issue of
whether the fibers are true coils (helices) as opposed
to wavy in only two dimensions (as in a sine wave)
was approached in two ways. The first and more
direct approach was to view scanning electron
microscopic images of the CPFs in stereo. In fortuitous
fractures of dehydrated or critical point-dried
samples, some of the CPFs remain intact and isolated
from surrounding tissue. Fibers, like that in
Figure 8, have a pitch (convolution repeat distance)
within the range observed in embedded samples and
in unfixed, unstained wedge samples and are thus
suitable for stereo analysis. The convoluted configuration
of the CPF is seen in stereo imaging to be
truly coiled (Figure 8B).
The second approach to the issue of helicity
depends on the observation that most convolutions
are minimal in stretched muscles; that is, most
fibers appear straight or much less convoluted at
15% muscle stretch (Figure 9) than at slack lengths,
where most fibers appear highly coiled. We can
predict that a true coil, when stretched, will have
crossover points. An analogy can be made with a
coiled telephone cord that is highly stretched; it
becomes almost straight but the crossover points
between inside and outside surfaces of the coil
leave thickened, regularly spaced bulges. Bulges,
spaced either in regular or irregular patterns, are
seen along the lengths of stretched CPFs (Figure 9).
Downloaded from
http://circres.ahajournals.org/ by guest on March 23, 2013

584 Circulation Research Vol 63, No 3, September 1988
i-ER
FIGURE 7. Transmission electron micrograph of cross-section of central region of right ventricular papillary muscle.
Coiled perimysial fiber has intrinsically elliptical cross-section, but eccentricity is exaggerated because plane of section
cut normal to muscle is oblique to coiled fiber. Differential alignment ofapposed constituent fibrils, in true cross-section
(arrow) and in oblique section (double arrowheads), are due to both the plane of section relative to the coil of the fiber
and the substructural twists of fibrils within the fiber. Fibrils are arranged in bundles within the fiber (arrowheads). Small
elastic fibers, associated with the coiled perimysial fiber, have amorphous and microfibrillar components. Coiled
perimysial fiber is enveloped by fibroblast process; the extracellular space between the fiber and fibroblast is less
electron-dense than that between the fibroblast and myocyte. cf, collagen fibrils; E, elastic fiber; Fb, fibroblast; rER,
rough endoplasmic reticulum; x!4,700 bar, 1 fim. Original magnification, x.14,700.
Furthermore, although considerable variation in
regularity of coiling exists among fibers or regions
of fibers, most of the CPFs are coiled with a pitch
that corresponds to several sarcomere lengths. In a
sample calculation, for example, a convolution index
of about 1.4 (see "Materials and Methods" and
Figure 2) at slack length would yield a convolution
index of 1.2 as a consequence of a 15% stretch. The
convolution indexes of many of the CPFs in muscles
stretched by 15% are in the range of 1.1-1.2,
although some regions actually seem to have a
convolution index of 1.0 (Figure 9).
Some of the variations in pitch of a given CPF or
between different CPFs within a single muscle in a
given state are so large that the CPFs must be
visualized as a population with distributions of
diameters, pitches, regularities of pitch, and starting
points of pitch at any given muscle length.
CPF convolutions were studied at two selected,
extreme mechanical states of papillary muscles,
slack length, and 15% stretch beyond slack length.
The concept of continuous change in convolution of
each CPF throughout the entire range of muscle
lengths is not, however, only deduced from different
CPFs in different muscles fixed at predetermined
lengths. Wedges of living myocardium were
stretched and compressed, in a qualitative manner.
The CPFs stretched and compressed reversibly,
like coiled springs.
The CPFs are laterally connected to myocytes
(Figure 10) and to each other (Figure 4) by struts
that are on the order of 0.1 Aim in diameter. The
struts have much smaller diameters than those of
the CPFs. They also are quite numerous and can be
seen draped along the sides of CPFs isolated for
scanning electron microscopy.
Downloaded from
http://circres.ahajournals.org/ by guest on March 23, 2013

Robinson et al Coiled Perimysial Fibers of Papillary Muscle 585
Downloaded from
http://circres.ahajournals.org/ by guest on March 23, 2013
FIGURE 8. Scanning electron
micrograph of right ventricular
papillary muscle fractured while
frozen. Muscle was fixed and processed
at slack length. Panel a:
Muscle axis (double-ended arrow)
and coiled perimysial fiber (CPF).
Sample calculation for one
selected turn of coil (see Figure 2
and "Materials and Methods"):
coil diameter (D)I2, 22 fim; pitch
(P), 73 yjn; convolution index (CI),

586 Circulation Research Vol 63, No 3, September 1988
15H m
FIGURE 9. Light micrograph of silver-stained longitudinal section of left ventricular papillary muscle fixed at 15%
stretch beyond slack length shows a single coiled perimysial fiber with differential configuration along its length.
Adjacent average sarcomere lengths are 2.3 fim along straight region (between arrows) and 2.1 \xm along nearly straight
region (between arrowheads). Compare with convoluted fibers at slack length (Figures 3 and 4). s, sarcomeres; bar, 15
lim. Original magnification, xl,480.
The configurations of collagen fibrils in chordae
tendineae differ in muscle-chordae tendineae preparations
that have been fixed at slack length (Figure
11 A) from muscle-chordae tendineae preparations
that have been stretched before fixation (Figure
1 IB). Quantitation of the relative changes in segment
length of the muscle itself versus that of the
chorda tendinea as a function of total stretch of the
muscle-chordae tendineae preparations remains to
be studied, as do the accompanying relative changes
in CPF convolutions and waviness of collagen fibrils
in chordae. Cross sections reveal that the abundant
elastic fibers in chordae tendineae contain a large
proportion of microfibrils and that the collagen
fibrils vary noticeably in diameter (Figure 11C).
Discussion
Configurations and Orientation of Coiled
Perimysial Fibers
CPFs comprise a branched network that is continuous
throughout the length of the papillary muscle
of rat heart (Figure 12). They are longitudinal or
oblique to the long axis of the muscle. Although
CPFs do not surround groups of myocytes in transverse
section, they are classified as perimysial
because each fiber is associated with more than two
myocytes in longitudinal directions and often in
transverse directions.
Several types of imaging have revealed the configurations
of the CPFs. Stereo viewing in the scanning
electron microscope was performed on fibers from
muscles that were processed in the critical point
dryer, frozen, and then fractured to reveal fibers free
of surrounding tissue. The truly helical fibers whose
surfaces are thus analyzed have the same projected
two-dimensional conformation as fibers studied by
transillumination in the light microscope, in both
silver-stained sections and unfixed, unstained wedges
of myocardium viewed with differential interference
contrast optics. The convoluted, embedded fibers
studied in the light microscope were not subject to
forces of fracture, and the wet wedges of tissue were
not subject to fixation, dehydration, or staining protocols
but were imaged by virtue of the birefringent
nature of the constituent collagen fibers. CPFs studied
in serial sections by transmission electron micros-
Downloaded from
http://circres.ahajournals.org/ by guest on March 23, 2013

Robinson et al Coiled Perimysial Fibers of Papillary Muscle 587
FIGURE 10. Scanning electron micrograph of lateral connecting struts (arrowheads) between coiled perimysial fiber
(CPF) and surface of myocyte (M) in left ventricular papillary muscle. Bar, 3 pm. Original magnification, x7,600.
copy had tortuous paths relative to neighboring
myocytes and different alignment of constituent fibrils
across the fiber and along its length. These observations
are consistent with the concept that CPFs are
helically configured as opposed to convoluted in two
dimensions.
Many of the bundles of collagen fibrils" and
rodlike collagen fibers 17 referred to in other studies
of papillary muscles are probably part of the axial,
branched network of CPFs. The CPF network is
distinct, however, from other components of the
myocardial perimysium in papillary muscle in the
weaves and septa of collagen fibers that surround or
separate myocytes or groups of myocytes, 38917 the
transverse collagenous ribbons that follow cell
contours, 16 the axial tendonlike fibers, 8 and the
collagenous strands that interconnect the weaves. 26
The network of CPFs is also different from the
highly regular, three-dimensional crisscross array
of perimysial fibers in bovine and sheep skeletal
muscle. 2 The less-regular array of CPFs in rat
papillary muscle could reflect the differences in
sizes, configurations, and orientations of myocytes
in cardiac and skeletal muscle, as well as difference's
in degree of elaboration of perimysium as a
function of muscle size (e.g., the perimysium is
more ordered and elaborate in skeletal muscles of
large mammals than in those of small mammals. 2
The structural and configurational differences
between perimysial arrays in skeletal and cardiac
muscles presumably reflect a difference in their
functions. The dependence of the pitch of CPFs on
the degree of stretch of the muscle supports this
idea. The convolution indexes of CPFs decrease as
the papillary muscle is stretched (Figures 2 and 12);
this is expected from the continuity of the branched
network of CPFs between anchor points in the
muscle-tendon junction and the ventricle wall. In
contrast, the perimysial "neutral connecting fibers"
of skeletal muscle are transverse to the direction of
muscle shortening, remain slack as myocytes change
length, and apparently serve only as scaffolds for
attachments of myocytes. 2728
Estimates of Tensile Strength
of Coiled Perimysial Fibers
In the absence of any direct measurements of the
mechanical properties of CPFs, tensile strengths
can be estimated as the product of cross-sectional
areas of the CPFs and the modulus of tensile
strength of collagen fibers from similar tissue. For
example, if the modulus of skeletal muscle fascia is
Downloaded from
http://circres.ahajournals.org/ by guest on March 23, 2013

588 Circulation Research Vol 63, No 3, September 1988
Downloaded from
http://circres.ahajournals.org/ by guest on March 23, 2013
"**.- --=«i~

Robinson et al Coiled Perimysial Fibers of Papillary Muscle 589
3Hm
FIGURE 11. Transmission electron micrographs of chordae tendineae of papillary muscles. Panel a: Longitudinal
section, slack length. Collagen fibrils (cf) are wavy. Bar, 3 tun. Original magnification, x.6,700. Panel b: Longitudinal
section, muscle—chorda tendinea preparation stretched 15% beyond slack length. Collagen fibrils are relatively straight.
Bar, 3 /xm. Original magnification, x.6,200. Panel c: Transverse section. Collagen fibrils and microfibril-rich elastic fibers
are interspersed. Fibroblasts are more abundant in periphery than in central regions of chorda tendinea. Bar, 3 fim.
Original magnification, x.6,800.
used 29 [15 /xN/(/im)2], the ultimate tensile strengths
of CPFs with range in cross-sectional area of 0.19-
68 (/Am) 2 would be 2.9-1,020 FIN. These calculated
values correspond to elliptical CPFs whose crosssectional
areas can be approximated as ellipses with
minor and major axes of 0.2 x 0.3 fim and 3x7 /tm.
If the calculation is made for only the smaller axes
(0.2 and 3 /im) the ultimate tensile strengths are
calculated as 1.9-424 /AN. The best estimates are
probably intermediate between these two sets of
values because at least some of the eccentricity of
the elliptical cross-sectional profile arises from the
transverse sectioning of a coil, which will not yield
an intrinsically circular cross section. In addition,
to relate stress to strain in an axially loaded helical
spring, corrections must be made for small values of
D/d (see "Materials and Methods") and for large
pitch angles (a>10°). For the CPF in Figure 8, D/
d—2.7 and a=*47°; in this case, the corrections
balance and yield a correction factor ~1 according
to the formula of Rover. 30
These calculations are intended to serve only as
preliminary estimates in the absence of direct mea-
surements of isolated CPFs. The values may represent
an overestimate of the tensile strengths actually
invoked during the cardiac cycle because not all
the fibers fully straighten, even in vitro at 15%
stretch beyond slack length, and because the moduli
of the rat cardiac CPFs relative to skeletal
muscle fascia depend on an incompletely understood
set of variables that includes the amounts and
ratio of type I and type III collagen, degree of cross
linking, diameters of constituent fibrils, and glycosaminoglycan
and glycoprotein contents. Differences
in these parameters can result in altered
mechanical function; for example, Viidik 31 has shown
that human chordae tendineae are more extensible
than skeletal muscle tendon and that chordae from
the left ventricle can be stretched 21% compared
with 14% in the lower-pressure right ventricle.
A preliminary comparison of estimates of tensile
strengths can be made between individual CPFs and
published values of measurements made from individual
myocytes freshly isolated from rat heart. 32 - 33
This type of analysis is complex because stiffness
depends on myocyte length, and myocytes display
Downloaded from
http://circres.ahajournals.org/ by guest on March 23, 2013

590 Circulation Research Vol 63, No 3, September 1988
Slack
Branched
Endothelium Fibers
FIGURE 12. Diagrams that summarize multiple observations
of configurations of the branched network of coiled
perimysial fibers (CPFs) and epimysial fibers at slack
length (top panel) and 15% stretch beyond slack length
(bottom panel) in cylindrical papillary muscle. The
branched CPF network is continuous from muscle-tendon
junction to ventricle wall. CPFs diverge from the musclechorda
tendinea junction and are highly convoluted at
slack length and less convoluted or even straight at 15%
stretch. Obliquely oriented CPFs are more aligned with
the muscle axis at 15% stretch. The biaxially arranged
epimysial collagen fibers are also less convoluted and
more highly aligned at 15% stretch ("cargo net
model" 313 ). Collagen fibrils of chordae tendineae are
crimped at slack length and straight when stretched.
These various changes in configuration are correlated
with a striking increase in resistance to passive stretch in
papillary muscles at 15% stretch (near "L^") and with
a sarcomere length at the upper limit of maximal crossbridge
overlap. 3 ' 13 - 43
large stress relaxation. 33 However, if we consider a
cell stiffness (measured force of an isolated cell
relative to length-normalized length perturbation) of
10 ^iN, 33 it seems reasonable to suggest that the
stiffness of the CPF array, with its higher modulus
but smaller total cross-sectional area, is significant
relative to the total stiffness of the muscle at long
lengths.
The questions then become what fraction of the
population of CPFs exert their full tensile strengths
when muscles are at 15% stretch and what increases
in myocyte stiffness occur at corresponding sarcomere
lengths? The effective values of CPF tension
are the magnitudes of the axial vectors of tension
for branches of a CPF network that are not colinear
with the muscle axis. Although a quantitative estimate
of vector components would be premature and
is not attempted, the oblique angles of some of the
CPFs would suggest that the tensile strengths actually
manifest are somewhat less than the calculated
ultimate tensile strengths. If this were true, then the
actual tensile strengths of the CPF and myocyte
systems could be fairly closely matched. Caulfield
and Borg 8 point out that the presence of more
tendonlike collagen fibers in papillary muscle than
in other muscle of the ventricle could account for its
higher stiffness constant.
Possibility of Energy Storage and Energy
Dissipation in Elastic and Collagenous
Components of Coiled Perimysial Fibers
The presence of satellite elastic fibers parallel to
the CPF axes warrants further investigation. Elastic
fibers adjoin essentially all of the CPFs, although
their cross-sectional area is far smaller than that of
the bundles of collagen fibrils. The elastic fibers
could serve as elastic forces of resistance to stretch.
A possibility (not a conclusion) is that stored energy
from both sources could then help promote subsequent
shortening or could preload the muscle for
the subsequent contraction. This type of model has
been invoked to explain biphasic stress-strain curves
in elastic fiber-collagen systems in bovine neck
ligament. 34 Additional or alternative functions have
been ascribed to tendons subject to rapid loading.
Energy dissipation might act as shock absorption to
prevent structural damage to the muscle during
sudden stretch. The elastin might act against the
creep phenomenon to restore collagen configuration.
35 Another possibility is that the elastic fibers
might serve as part of a scaffold in the domain of
remodeling close to the fibroblast. 25 A model for
extracellular compartments of fibroblasts of cornea
and tendon has been proposed, although a role for
elastin was not proposed in those studies. 3637
A role of energy storage by the collagen of CPFs
should not be ruled out without direct experimental
tests. Tendons in jumping animals such as Alsatian
dogs and kangaroos have been shown to store large
amounts of elastic energy (on the order of 5 J/g) 38 - 39
over defined ranges of stretch. CPFs look like
springs, but whether the coiled configurations of the
CPFs contribute significant elastic storage of energy
with gradual onset over the physiological range of
lengths remains to be determined. The possibility of
springlike behavior of the CPFs is consistent with
but not proven by the observation that CPFs are
coiled, even when isolated from the muscle, and
change convolution continuously with stretch and
compression (authors' published observations). This
CPF configuration in the absence of external forces
implies a prestressed system with reversibly distensible
constraints built into the coil itself. An energy
storage system comprised of CPFs acting in concert
with the myocytes themselves could provide the
Downloaded from
http://circres.ahajournals.org/ by guest on March 23, 2013

source of the elastic recoil and suction exhibited
during the cardiac cycle. 40 - 42
Functional Implications of Configurational
Changes in Coiled Perimysial Fibers and
Epimysial Fibers
To describe changes in muscle length, CPF configuration,
and sarcomere lengths on a quantitative
basis, a starting point needs to be selected. Slack
length has been used as this arbitrary point where
resting tension approaches zero and the muscle
appears straight (Figure 12). However, the accuracy
of estimates of the percent changes in convolution
indexes and sarcomere lengths does not permit
an assessment of whether the CPFs or the
myocytes pick up the load first during stretch or if
the two systems act in concert. Uncertainty attributable
to normal variations in size and pitch of
CPFs, extent of differential preset stretch, possible
differences in myocyte stimulation and shrinkage
during fixation and preparation, and uncertainty in
measurements of coil pitch limit the assessment of a
differential onset of stretch in CPFs and myocytes
in intact muscle.
CPFs and myocytes are interconnected (Figure
10). The cardiac endomysial connective tissue is
mechanically linked to myocytes and their contractile
lattices and cytoskeletons, 15 which have significant
tensile properties, 32 but the slack and compliance
in their interconnections is not yet known. A
current approach to further investigate this question
involves our use of living muscles monitored for
CPFs, sarcomere lengths, and muscle force simultaneously
as a function of muscle length.
CPFs are abundant in the ventricular walls, and
their changes in configuration with ventricular distension
in isolated, arrested rat hearts (authors' unpublished
observations) suggest a possible role in passive
compliance that currently is under investigation.
Acknowledgments
We thank Dr. J. Capasso for helpful discussion and
for providing samples of fixed, mechanically characterized
papillary muscles for structural studies; Drs.
Mengjia Zhao and Mark Flomenbaum for assistance
with the whole heart preparations; Drs. S. Seifter and
O. Blumenfeld for helpful discussions and comments
on the manuscript; Dr. N. Broom for advice and
encouragement with differential interference contrast
of collagen in wet tissue; Ms. L. Cohen-Gould for
excellent assistance with many aspects of the project;
Mrs. R. Dominitz for excellent technical assistance
with the silver impregnation; Ms. J. Fant and Mr. F.
Macaluso for use of facilities and help at the Analytical
Ultrastructure Center; Mrs. K. Cohen for
translation of articles from German; Ms. C. Donner
for the illustrations in Figures 2 and 12; and the
reviewers for helpful comments.
Robinson et al Coiled Perimysial Fibers of Papillary Muscle 591
References
1. Walls EW: The microanatomy of muscle, in Boume GH
(ed): Structure and Function of Muscle. New York, Academic
Press, 1960, vol 6, pp 21-61
2. Rowe RWD: Morphology of perimysial and endomysial
connective tissue in skeletal muscle. Tissue Cell 1981;
13:681-690
3. Robinson TF, Cohen-Gould L, Factor SM: Skeletal framework
of mammalian heart muscle: Arrangement of inter- and
peri-cellular connective tissue structures. Lab Invest 1983;
49:482-498
4. Klein W, Bock P: Elastica-positive material in the atria]
endocardium: Light and electron microscopic identification.
ActaAnat 1983;116:106-113
5. Holmgren E: Uber die Trophospongien der quergestreiften
Muskelfasem, nebst Bemerkungen uber den allgemeinen
Bau dieser Fasern. Arch Mikrosk Anat 1907;71:165-247
6. Nagel A: Die mechanischen Eigenschaften von Perimysium
intemum and Sarkolemm bie der quergestreiften Muskelfaser.
Z Zellforsch Mikrosk Anat 1935;22:694-706
7. Winegrad S, Robinson TF: Force generation among cells in
the relaxing heart. Eur J Cardiol 1978;7(suppl):63-70
8. Caulfield JB, Borg TK: The collagen network of the heart.
Lab Invest 1979;40:364-372
9. Robinson TF: Lateral connections between heart muscle
cells as revealed by conventional and high voltage transmission
electron microscopy. Cell Tiss Res 1980;211:353-359
10. BorgTK, Caulfield JB: The collagen matrix of the heart. Fed
Proc 1981;4O:2O37-2O41
11. Robinson TF, Winegrad S: A variety of intercellular connections
in heart muscle. / Mol Cell Cardiol 1981 ;13:185-195
12. Robinson TF: The physiological relationship between connective
tissue and contractile filaments in heart muscle.
Einstein Quarterly 1983 ;1:121-127
13. Robinson TF, Cohen-Gould L, Remily RM, Capasso JM,
Factor SM: Extracellular structures in heart muscle, in
Harris P, Poole-Wilson PA (eds): Advances in Myocardiology,
Vol. 5. New York, Plenum Publishing Company, 1985,
pp 243-255
14. Frank J, Beydler S: Intercellular connections in rabbit heart
as revealed by quick-frozen, deep-etched, and rotaryreplicated
papillary muscle./ UltrastrucRes 1985^0:183-193
15. Robinson TF, Factor SM, Capasso JM, Wittenberg BA,
Blumenfeld OO, Seifter S: Morphology, composition and
function of structs between cardiac myocytes of rat and
hamster. Cell Tiss Res 1987 ;249:247-255
16. Robinson TF, Cohen-Gould L, Factor SM, Eghbali M,
Blumenfeld OO: Structure and function of connective tissue
in cardiac muscle: Collagen type III in endomysial struts and
pericellular fibers. Scanning Microsc 1987;2:1005-1015
17. Spach MS, Dolber PC: The relation between discontinuous
propagation in anisotropic cardiac muscle and the "vulnerable
period" of reentry, in Zipes DP, Jalife J (eds):
Cardiac Electrophysiology and Arrhythmias. Orlando, Fla,
Grune and Stratton, 1985, pp 241-252
18. Schlant RC, Sonnenblick EH: Normal physiology of the
cardiovascular system, in Hurst JW (ed): The Heart, Arteries
and Veins. New York, McGraw-Hill, 1986, ed. 6, 37-72
19. Broom ND: Simultaneous morphological and stress-strain
studies of the fibrous components in wet heart valve leaflet
tissue. Conn Tiss Res 1978;6:37-50
20. Hasagawa T, Ravens JR: A metallic impregnation method
for the demonstration of cerebral vascular patterns. Ada
Neuropathol 1968;10:183-188
21. Del Rio Hortega P: El metodo del carbonata argentico:
Revision general de sus tecnicas y aplicaciones en histologia
normal y patoligica. Arch Histol Norm Pathol (Buenos
Aires) !943;2:231-243
22. Spurr AR: A low viscosity epoxy resin embedding medium
for electron microscopy. J Ultrastruct Res 1969;26:31-43
23. Reynolds ES: The use of lead citrate at high pH as an
electron opaque stain in electron microscopy. J Cell Biol
1963,17:208-212
Downloaded from
http://circres.ahajournals.org/ by guest on March 23, 2013

592 Circulation Research Vol 63, No 3, September 1988
24. Woodcock CLF, Frado L-Ly, Rattner JB: The higher order
structure of chromatin: Evidence for a helical ribbon arrangement.
J Cell Biol 1984;99:42-52
25. Eghbali M, Eghbali M, Robinson TF, Seifter S, Blumenfeld
OO: Role of fibroblasts in accumulation of collagens in
growing and adult rat hearts (abstract). J Molec Cell Cardiol
1987;19(supplIV):160
26. Weber KT, Janicki JS, Pick R, Abrahams C, Shroff SG,
Bashey RI, Chen RM: Collagen in the hypertrophied,
pressure-overloaded myocardium. Circulation 1987;
75(suppl I):I-40-l-47
27. Schmalbruch H: Skeletal Muscle. New York, Springer-
Verfag, 1985, pp 14-22
28. Feneis H: Ueber die Anordnung und die Bedeutung des
Bindegewebes fuer die Mechanik der Skelettmuskulatur.
Gegenbaurs Morphol Jahrbuch 1935;76:161-202
29. Viidik A: Properties of tendons and ligaments, in Skalak R,
Chien S (eds): Handbook of Bioengineering. New York,
McGraw-Hill, 1987, p 63
30. Berry WR: Spring Design. London, Emmott and Company,
1961, pp 20-21
31. Viidik A: Chordae tendineae—morphology and biomechanics
of an unusual collagenous tissue, in Spring E, Jauhiaincn
T, Honkavaara T (eds): Proceedings of the First Nordic
Meeting on Medical and Biological Engineering. Helsinki,
Finnish Society for Medical and Biological Engineering,
1970, pp 200-202
32. Brady AJ: Passive stiffness of rat cardiac myocytes. J
BiomechEng 1984; 106:25-30
33. Brady AJ, Famsworth SP: Cardiac myocyte stiffness following
extraction with detergent and high salt solutions. Am J
Physiol 1986;250:H932-H943
34. Hoeve CAJ, Flory PJ: The elastic properties of elastin. J Am
Chem Soc 1958;80:6523-6526
35. Harkness RD: Mechanical properties of connective tissues
in relation to function, in Parry DAD, Creamer LK (eds):
Fibrous Proteins: Scientific, Industrial and Medical Aspects.
London, Academic Press, 1979, pp 207-230
36. Birk DE, Trelstad RL: Extracellular compartments in matrix
morphogenesis: Collagen fibril, bundle, and lamellar formation
by cornea! fibroblasts. J Cell Biol 1984^9:2024-2033
37. Birk DE, Trelstad RL: Extracellular compartments in tendon
morphogenesis: Collagen fibril, bundle, and macroaggregate
formation. J Cell Biol 1986;103:231-240
38. Bennett-Clark HC: Energy storage in jumping animals, in
Perspectives in Experimental Biology. Elmsford, New York,
Pergamon Press, 1976, vol. 1, pp 467-479
39. Alexander R McN: The mechanics of jumping by a dog
(Canis familiaris). J Zool Lond 1974;173:549-573
40. Robinson TF, Factor SM, Sonnenblick EH: The heart as a
suction pump. Sci Amer 1986;254:84-91
41. Hori M, Yellin EL, Sonnenblick EH: Left ventricular diastolic
suction as a mechanism of ventricular filling. Jap Circ J
1982;46:124-129
42. Yellin EL, Hori M, Yoran C, Sonnenblick EH, Gabbay S,
Frater RWM: Left ventricular relaxation in the filling and
nonfilling intact canine heart. Am J Physiol 1986;
250:H620-H629
43. Robinson TF, Winegrad S: The measurement and dynamic
implications of thin filament lengths in heart muscle. J
Physiol 1979;286:607-619
KEY WORDS • collagen fibers, fibrils • connective tissue •
muscle, cardiac • elastic fibers • perimysium
Downloaded from
http://circres.ahajournals.org/ by guest on March 23, 2013