TX 72514Circulation Research is published by the American Heart Association. 7272 Greenville Avenue, Dallas,
doi: 10.1161/01.RES.72.4.7681993, 72:768-777Circulation Research
http://circres.ahajournals.org/content/72/4/768located on the World Wide Web at:
The online version of this article, along with updated information and services, is
http://www.lww.com/reprintsReprints: Information about reprints can be found online at [email protected]. E-mail:
Fax:Kluwer Health, 351 West Camden Street, Baltimore, MD 21202-2436. Phone: 410-528-4050. Permissions: Permissions & Rights Desk, Lippincott Williams & Wilkins, a division of Wolters http://circres.ahajournals.org//subscriptions/Subscriptions: Information about subscribing to Circulation Research is online at
by guest on October 18, 2011http://circres.ahajournals.org/Downloaded from
Effect of Dysfunctional Vascular Endotheliumon Myocardial Performance in Isolated
Papillary MusclesKai Li, J.L. Rouleau, L.J. Andries, and D.L. Brutsaert
Vascular endothelium has been shown to modify the contractile characteristics of vascular smoothmuscle, and endocardial endothelium has been shown to modify the contractile characteristics of adjacentmyocardium. In this study, whether vascular endothelium also modifies the contractile characteristics ofadjacent myocardium and whether these effects are additive to those of endocardial endothelium wereinvestigated. Rabbit hearts (n=54) were excised and mounted in a Langendorif preparation. Vascularreactivity was verified by acetylcholine infusion. One group of these hearts had Triton X-100 injected as
a bolus into the coronaries to render the vascular endothelium dysfunctional. The other portion served as
control hearts. Triton X-100 bolus injection resulted in little or no pathological changes on morphologicalexamination; however, the vasodilatory response to acetylcholine in these hearts was abolished, suggestingvascular endothelial dysfunction. Vascular smooth muscle reactivity was verified in Triton X-100-injectedhearts by nitroprusside infusion. In the control Langendorff-perfused hearts, there was little evidence ofvascular endothelial dysfunction, with the coronary perfusion rate increasing from 8.9±0.4 to 11.0±0.3ml/g per minute (p<0.01) in response to acetylcholine. All hearts were then removed, and rightventricular papillary muscles were excised for myocardial mechanical studies. Control Langendorff-perfused hearts had myocardial mechanical characteristics similar to those of muscles from 18 othercontrol hearts without Langendorff perfusion, indicating that the Langendorff perfusion itself had littleeffect on myocardial mechanics. The muscles from the Triton X-100-injected Langendorff hearts hadmarked changes: a shortening of twitch duration (363± 16 versus 449±9 msec, p<0.01) and decreases intotal tension (2.2±0.2 versus 2.9±0.2 g/mm2, p<0.01), dT/dt (9±1 versus 12±1 g/mm2 per second,p<0.05), and maximum velocity of unloaded muscle shortening (Vmax) (0.89±0.06 versus 1.14±0.07 lengthat which maximum developed tension occurred [Lm.]/sec, p<0.05). Endocardial endothelial removal ofthe papillary muscles in the two control groups (with and without Langendorif perfusion) by Triton X-100caused the same changes in twitch characteristics as occurred in muscles from the Langendorff-perfusedhearts injected with Triton X-100 but with intact endocardial endothelium, suggesting that vascularendothelial dysfunction had similar effects on contractile characteristics as endocardial endothelialremoval. Endocardial endothelial removal of the papillary muscles from Langendorif-perfused hearts thathad a bolus injection of Triton X-100 caused further shortening of twitch duration and a further decreasein total tension (1.7±0.2 versus 2.1±0.1 g/mm2, p<O.05), dT/dt, and Vmax (0.69±0.03 versus 0.89±0.01L,../sec, p<O.01), suggesting that the myocardial contractile effects of endocardial and vascularendothelium are additive. Increasing extracellular calcium concentration to 15 mM normalized dT/dt andVm. such that these parameters were similar in all groups of muscles before and after vascular andendocardial endothelial dysfunction. Total tension normalized completely only at higher extracellularcalcium concentrations (20 and 25 mM). The addition of10-` M phenylephrine to muscles with andwithout vascular endothelial dysfunction in a bath with 15 mM calcium normalized tension, alsosuggesting that the changes documented in this study were not due to myocardial damage but toendothelial dysfunction. These results suggest that both endocardial and vascular endothelium modulatethe contractile characteristics of their adjacent myocardium and that their effects are additive.(Circulation Research 1993;72:768-777)KEY WoRDs * vascular endothelium* endocardium * contractility * myocardium
T he endocardial endothelium, the internal lining modifies twitch configuration in a typical manner, i.e.,of the cardiac chamber, has been shown to decreasing time to peak tension and thus decreasingmodulate the contraction of adjacent myocar- tension development. This effect of endocardial endo-
dium.12 Removal of this endocardial endothelial layer thelium on adjacent myocardium has been confirmed byothers in isolated papillary muscles from various species
From the Department of Medicine, University of Sherbrooke(Canada), and the Department of Physiology, University of Ant- Address for correspondence: Jean L. Rouleau, MD, Faculty ofwerp (Belgium). Medicine, University of Sherbrooke, Sherbrooke,Quebec, Canada
Supported by the Medical Research Council of Canada. J.L.R. is a J1H 5N4.senior scholar of les Fonds de la Recherche en Sante du Quebec. Received March 24, 1992; accepted December 12, 1992.
by guest on October 18, 2011http://circres.ahajournals.org/Downloaded from
Li et al Vascular Endothelium and Myocardial Contractility
such as cat, dog, rabbit, and ferret.3-10 In addition, thepresence or absence of this endocardial endotheliallayer has been shown to greatly modify the effects ofplatelets, serotonin, eosinophils, vasopressin, endothe-lin, and other circulating substances on the contractilecharacteristics of isolated papillary muscles.6-10 Vascu-lar endothelium modulates vascular tone by the releaseof varying contractile and vasodilating substances."-'9Several of these vasoactive substances, such as endothe-lium-dependent relaxation factor, angiotensin II, andendothelin, which are released by the vascular endothe-lium of various vessels, have also been shown to havedirect myocardial effects.6,20-30 The exact mediator ormediators that lead to the modulatory effects of en-docardial endothelium on its adjacent myocardium havenot been conclusively identified; however, they appearto enhance myocardial performance by increasing myo-filament calcium responsiveness.5 One recent study sug-gests that contractile proteins are also regulated bysubstances produced by blood vessels31; however, it isunknown to what extent coronary vascular endotheliummodulates the contractile characteristics of the imme-diate adjacent myocardium, particularly in the micro-vascular bed.32 Hence, as vascular endothelium andendocardial endothelium constitute one continuousstretch of tissue, it would be of interest to examinewhether and to what extent vascular endothelium mod-ulates the contractile characteristics of the adjacentmyocardium and to what extent these effects are addi-tive to those of endocardial endothelium.
Therefore, we examined the cardiac performance ofisolated papillary muscles from Langendorff-perfusedrabbit hearts in which the vascular endothelium hadbeen experimentally made dysfunctional by intracoro-nary Triton X-100 bolus injection, and we compared theperformance of these muscles with muscles from Lan-gendorif-perfused rabbit hearts in which no TritonX-100 bolus injection had been given. The resultsobtained from papillary muscles of the Langendorff-perfused hearts were then compared with those ob-tained from hearts in which no Langendorff perfusionwas done. The effects of damaging the endocardialendothelium by brief immersion of the papillary mus-cles in Triton X-100 were then evaluated in all musclegroups. Our findings suggest that the effects of damageto either the vascular or endocardial endothelium byTriton X-100 result in a similar modulation of thecontractile characteristics of adjacent myocardium: ab-breviation of twitch duration and a decrease in devel-oped tension. Our results also suggest that the effects ofdamage to both of these endothelial layers are additive.
Materials and MethodsFifty-four rabbits of either sex weighing 2.5+±0.3 kg
were used in the study. The rabbits were anesthetizedwith pentobarbital (25-40 mg/kg) and anticoagulatedwith heparin (3,000 units/kg). The chest was openedand the heart was excised and used either for immediateexcision of a right ventricular papillary muscle (group 1,24 hearts) or for the Langendorff preparation (groups 2and 3, 30 hearts).
Langendorff PreparationOnce excised, the hearts were mounted in a Langen-
dorff preparation with a perfusion pressure of 60
LANGENDORFF HEART
Ach Acil
a2Co0W
3
t saline + l
triton
I t/////~~~i///Y/////////
PAPILLARY MUSCLE
E trlyn 1
1 4TIME (hrs)
FIGURE 1. Diagram summarizing the protocol. Protocol A:Effect of acetylcholine (Ach) on coronary flow before thebolus injection of Triton X-100 or saline. Protocol B: Effect ofAch on coronary flow once saline (group 2) or Triton X-100(group 3) had been injected. Protocol C: Mechanical studiesof papillary muscles from rabbit hearts never having hadLangendorffperfusion (group 1) or having had Langendorfperfusion and intracoronary saline (group 2) or Triton X-100(group 3). Protocol D: Papillary muscle mechanical studiesafter endocardial endothelial removal by Triton X-100.
mm Hg. The hearts were perfused with Krebs-Henseleitsolution containing (mM) NaCl 118, KCI 3.5, MgSO42.43, CaCh2 1.25, KH2PO4 1.2, NaHCO3 24.9, and dex-trose 5.0. The solution was kept at a temperature of37°C and bubbled with 95% 02--5% CO2 gas mixture atpH 7.4. The superficial veins and coronary sinus werethen perforated, and the right atria were removed tominimize penetration of the bolus injection of TritonX-100 (Scientillar Mallinkrodt Inc., Paris, Ky.) into theright ventricle via venous return.
Protocol A. The hearts were permitted to stabilize for25 minutes, at which time the coronary perfusion ratewas assessed by accumulating perfusate over a 2-minuteperiod and then dividing by 2 (Figure 1). An acetylcho-line (Sigma Chemical Co., St. Louis, Mo.) infusion wasthen started to assess endothelium-dependent relax-ation of the coronaries. Acetylcholine was delivered bya pump (Harvard Apparatus, South Natick, Mass.) andmixed with the perfusate just above the aortic valve at arate calculated to deliver a final concentration of5 x 10-7 M to the coronaries. This final concentrationwas obtained by adjusting the infusion rate of the 10`M stock solution of acetylcholine to the coronary per-fusion. Coronary perfusion was then calculated by col-lecting coronary effluent. In two hearts, acetylcholinedid not cause an increase in coronary effluent. Thesetwo hearts were discarded because of the risk of endo-thelial dysfunction.
Protocol B. Hearts were then allocated randomly toeither saline (group 2) or Triton X-100 (group 3) bolusinjections (Figure 1). Pilot experiments using TritonX-100 in concentrations of 1:1,000, 1:500, 1:200, and1:100 were studied both in rat and rabbit hearts toestablish the minimal concentration of Triton X-100necessary to abolish the coronary vasodilatory responseto acetylcholine. A 1: 200 concentration injected over a1-second period was found to be the minimal effectiveconcentration. The amount injected was 1% of thecoronary perfusion rate per minute. This was injectedinto the aorta just above the coronary arteries. In thosehearts randomized to saline, an equivalent amount ofsaline was injected in the same manner. Twenty-fiveminutes later, the effects of the Triton X-100 bolus on
769
b
by guest on October 18, 2011http://circres.ahajournals.org/Downloaded from
coronary perfusion rate was evaluated by collecting thecoronary effluent. The effects of acetylcholine on coro-nary perfusion rate were then reevaluated 5 minuteslater in both saline-treated and Triton X-100-treatedhearts to assess the ability of the vascular endotheliumto produce coronary vasodilatation.The ability of coronary vascular smooth muscle to
vasodilate after the Triton X-100 injection was verified bynitroprusside infusion before and after Triton X-100 insix Langendorff-perfused hearts. Nitroprusside was de-livered by Harvard pump and mixed with the perfusatejust above the aortic valve at a rate calculated to delivera final concentration of 3 x 10` M to the coronaries. Thisfinal concentration was obtained by adjusting the infu-sion rate of 1.7 x 10` M stock solution of nitroprusside tothe coronary perfusion. Coronary perfusion was thencalculated by collecting coronary effluent.
Isolated Papillary Muscle StudiesOnce the 17 hearts not used for the Langendorff
studies (group 1) were excised from the chest and the 28hearts used for the Langendorff and papillary musclestudies (groups 2 and 3) were removed from the Lan-gendorif preparation, the right ventricle was opened,and a right ventricular papillary muscle was excised.The papillary muscle was then mounted in a bath withKrebs-Henseleit solution similar to the one used in theLangendorff preparation, but in this experimental situ-ation the bath was kept at 29°C. The base of the musclewas held by a stainless-steel clamp, and the other endwas tied to a lever identical to that described in detail byBrutsaert et al.33 The muscles were stimulated at 10%above threshold at 12 stimuli per minute with a modelS-88 stimulator (Grass Instrument Co., Quincy, Mass.)through platinum field electrodes. The preload wasadjusted so that the muscle length was at the length atwhich maximum developed tension occurred (1,nx).
Protocol C. The muscles were stabilized at Lm for 2hours, and then isometric and isotonic contractionswere recorded at a speed of 100 mm/sec on a model2400 s recorder (Gould Inc., Cleveland, Ohio) (Figure1). After the elastic damping of the force-length-velocity lever feedback system was adjusted to compen-sate for electromechanical transients, the maximumvelocity of unloaded muscle shortening (Vm.,) was ob-tained by abruptly decreasing the load on the muscle atthe time of activation (zero load clamp).33 The output ofthe amplifier of the Gould 2400 s recorder was con-nected to an electronic differentiator, and both of thesewere fed into an analog-to-digital converter (model DT2821-F- 801, Data Translation Inc., Marlborough,Mass.), which in turn was connected to a microcom-puter (model 286, COMPAQ Computer Corp., Hous-ton, Tex.). Analysis of force-length characteristics wasperformed by custom-made software running underMS-DOS. Once basal values at 1.25 mM calcium (physi-ological) were recorded, extracellular calcium was in-creased to 15 mM by adding CaCI2 from a stock solution(high calcium), and repeat isometric, isotonic, andunloaded contractions were recorded.
Protocol D. Once the basal values for groups 1-3muscles were obtained, the endothelial layer of theendocardium was removed by immersing the papillarymuscle in 1% Triton X-100 dissolved in Krebs-Hense-
abundantly (Figure 1). This technique has been shownto destroy the endothelial layer of the endocardiumwithout damaging myocardial cells.2 Muscles were thenpermitted to restabilize for 2 hours, and repeat isomet-ric, isotonic, and unloaded contractions were recordedat 1.25 mM calcium concentration from the papillarymuscles of all three groups. This was done to assess
whether removing the layer of endothelial cells coveringthe endocardium altered twitch characteristics similarlyin papillary muscles in which vascular endothelium hadbeen rendered dysfunctional or not. To assess whetherthe changes in contractile characteristics caused byaltering endothelial function were reversible by increas-ing extracellular calcium concentration, repeat isomet-ric, isotonic, and unloaded contractions were recordedat 15 mM extracellular calcium concentrations in allthree groups of muscles (groups 1-3). As the totaltension of muscles with vascular or endocardial endo-thelial dysfunction did not completely normalize atextracellular calcium concentrations of 15 mM, twofurther sets of experiments were conducted. In the firstset of experiments, 10-5 M phenylephrine in the pres-ence of 10-5 M propranolol (Sigma) was added to a bathwith 15 mM extracellular calcium for eight muscles withvascular endothelial dysfunction (group 3) and eightcontrol muscles (group 1) with intact and functionalendocardial and vascular endothelium, and isometric,isotonic, and unloaded contractions were then re-
corded. In a second set of experiments, a dose-responsecurve from 1.25 to 25 mM extracellular calcium was
performed before and after endocardial endothelialremoval in six muscles to assess whether maximalcalcium activated force at saturating bath calcium was
changed by the removal of endocardium. Muscles hadexcessive ectopic contractions or went into contractureat extracellular calcium concentrations higher than 25mM.
The muscle cross section was measured by assuming a
cylindrical shape of the muscles and dividing muscleweight by length. The average cross section of the rabbitpapillary muscles was 0.51±0.03 mm2.
Morphological EvaluationScanning electron microscopy, light microscopy, and
confocal laser-scanning light microscopy were used todetect morphological changes in Langendorff-perfusedhearts. The viability of the endothelial cells was verifiedby confocal laser-scanning light microscopy using pro-pidium iodide before fixation to stain the nuclei of deadcells. After fixation, F-actin in endocardial endothelialand myocardial cells was stained by Bodipy-phallacidin.For scanning electron microscopy, three control and
six Triton X-100-treated rabbit hearts were perfusion-fixed with 2% glutaraldehyde in 0.1 M Millonig buffer atthe end of a Langendorff experiment. Endocardial andmyocardial tissue blocks were further processed as
previously described.34,35 Specimens were examinedwith a Cambridge Stereoscan scanning electron micro-scope operating at 20 kV.
For confocal laser-scanning light microscopy, one
control and four Triton X-100-treated rabbit heartswere perfused for 10 minutes with a Krebs-Henseleitsolution containing 0.1 mM propidium iodide beforefixation. The hearts were then perfusion-fixed with 4%
leit solution at 29°C for 1 second and then washing formaldehyde in 0. 1 M Millonig buffer for 5 minutes.
by guest on October 18, 2011http://circres.ahajournals.org/Downloaded from
Li et al Vascular Endothelium and Myocardial Contractility 771
50
[10t(saine) ( ) triton infusion
30
ry~I
0 10
0__
GROUP 2 GROUP 3 GROUP 3
10 (saline) (triton) (triton)
FIGURE 2. Bar graph showing the effect of acetylcholine(Ach) on coronary flow before and after saline injection ingroup 2 Langendorff-perfused hearts and the effects ofAch or
nitroprusside (NP) on coronary flow before and after TritonX-100 injection in group 3 Langendorff-perfused hearts. Ingroup 2 hearts, the vasodilatory effect ofAch was maintainedthroughout the study, whereas in group 3 hearts, Triton X-100abolished the vasodilatory effect ofAch but did not modify theeffect ofNP, suggesting that Triton X-100 produced endothe-lial but not vascular smooth muscle dysfunction.
After dissection, the tissue was further fixed for at least18 hours. The fixed tissue was then rinsed with phos-phate-buffered saline, treated with 0.1% glycine inphosphate-buffered saline, and stained with 0.165 ,uMBodipy-phallacidin in phosphate-buffered saline. Spec-imens were mounted in Cytofluor and observed with an
MRC- 600 laser-scanning confocal microscope (Bio-Rad, United Kingdom). The argon-ion laser beam was
used for double fluorescence emission with the Al andA2 filter set. Images at both wavelengths were electron-ically merged and printed with a Sony videoprinter.For light microscopy, the hearts were fixed as for
confocal laser-scanning light microscopy, dehydrated inethanol, and embedded in Historesin or JB-4. Semithinsections, 1-2 ,um thick, were cut on glass knives andstained with azure II. Plastic sections were observedwith a Polyvar II light microscope.
StatisticsResults are expressed as mean+SEM. For multiple
comparisons, a one-way analysis of variance followed bya Student-Newman-Keuls test was used. A paired t testwas used to evaluate the effects of interventions on thesame muscles. An unpaired t test was used for compar-ison of different groups of muscles.
ResultsLangendorffPerfusion and Coronary Vasodilatation
Protocol A. Before the injection of Triton X-100, theinfusion of acetylcholine caused a similar increase incoronary perfusion rate in both groups of Langendorff-perfused hearts (Figure 2, Table 1).
Protocol B. The injection of Triton X-100 (group 3)led to a significant decrease in coronary perfusion rate(Table 1). The bolus injection of Triton X-100 alsoabolished the vasodilatory effects of acetylcholine butdid not modify the vasodilatory effects of nitroprusside(Sigma) (Figure 2, Table 1). The vasodilatory response
to acetylcholine was not modified by saline injection(group 2).
Isolated Right Ventricular Papillary Muscle StudiesProtocol C. Before endocardial endothelial removal,
the contractile characteristics of papillary muscles ob-tained from hearts that did not undergo Langendorffperfusion (group 1) and right ventricular papillarymuscles from hearts that underwent Langendorff per-
fusion and had a bolus injection of saline (group 2) weresimilar (Table 2). However, the contractile characteris-tics of papillary muscles from hearts that underwentLangendorff perfusion and had a bolus injection ofTriton X-100 (group 3) had considerably different con-
tractile characteristics (Figure 3, Table 2). Peak twitchtension, maximum rate of tension development (dT/dt),time to peak tension, time to half tension decline, andVm,, decreased significantly, indicating an abbreviationof twitch duration and a slight decrease in contractility.
Protocol D. The removal of endocardial endotheliumin right ventricular papillary muscles from hearts with-out vascular endothelial dysfunction (groups 1 and 2)showed characteristics similar to those described previ-ously for this intervention. Twitch duration decreased,and there was a mild decrease in peak twitch tensionand dT/dt and a smaller decrease in Vma,x (Figure 3,Table 2). These changes were very similar to thosecaused by creating vascular endothelial dysfunctionalone by bolus injection of Triton X-100 into the coro-
nary arteries (group 3) (Figure 3, Table 2). Endocardialendothelial removal of the right ventricular papillarymuscle from the Langendorff-perfused hearts in which a
bolus injection of Triton X-100 was given and vascularendothelial dysfunction was created (group 3) resultedin further shortening of twitch duration and a furtherdecrease in peak twitch tension, dT/dt, and Vm,, (Figure3, Table 2).
TABLE 1. Bolus Triton X-100 Injection and the Effects of Acetylcholine on Coronary Flow in Langendorff-Perfused Rabbit Hearts
Coronary flow (ml/g per minute)
n Baseline 1 Ach 1 Baseline 2 After Triton Ach 2
Group 2 8 8.6±0.3 12.1+0.5 8.9+0.4 ... 11.0+0.3
Group 3 13 9.0+0.3 12.4+0.6 9.7+0.4 7.1+0.5* 7.0+0.5t
Baseline 1, values before acetylcholine (Ach) injection; Ach 1, response after Ach injection; baseline 2, response before injection of TritonX-100; after Triton, response to bolus injection of Triton X-100; Ach 2, reevaluation of effects of Ach after Triton X-100 injection; group2, saline-injected hearts with Langendorff perfusion; group 3, Triton X-100-injected hearts with Langendorff perfusion. Values are
mean+SEM.*p<0.05 vs. corresponding baseline value.tp<0.01 vs. corresponding value for group 2.
by guest on October 18, 2011http://circres.ahajournals.org/Downloaded from
TT, total twitch tension; dT/dt, maximum rate of tension development; TTPT, time to peak tension; RT!, time from the onset of the twitchto half relaxation; Vma, maximum shortening velocity of an unloaded (zero load) contraction; group 1, control muscles fromnon-Langendorif-perfused hearts; -EE, endocardial endothelial removal; group 2, control muscles from saline-injected Langendorff-perfused hearts; group 3, muscles from Triton X-100-injected Langendorff-perfused hearts; -VE, vascular endothelial dysfunction; PHE,phenylephrine.
*p<0.01 and tp<0.05 vs. baseline.*p<O.05 and §p<0.01 vs. -VE.
Mechanical studies done at 15 mM calcium showedthat the depression in peak twitch tension, dT/dt, andVmax observed at physiological concentrations of calcium(1.25 mM) were nearly completely reversible, with dT/dt
3- 8 10Ca++ 1.25 mM Co++ 15 mM Ca++ 15 n
E 8 XE conrolcontrolE. -EE 6.
1.1~~~~~~~~2 4.
z E
H 21
200 m
and Vm,, being similar in all groups (groups 1-3) beforeand after endocardial endothelial removal (Table 2).Nevertheless, the use of 15 mM calcium concentrationsdid not totally abolish the changes in twitch configura-
FIGURE 3. Representative curves of twitches frompapillary muscles with functional endocardial andvascular endothelium (control), with dysfunctionalvascular endothelium (- VE), with the endocardialendothelium removed (-EE), and with dysfunc-
PHEN tional vascular endothelium after endocardial endo-thelium removal (-VE-EE). At physiological cal-cium concentration (1.25 mM), endothelial removal
itrol or dysfunction caused a decrease in contractility-VE and abbreviation of twitch duration. At high cal-
cium concentrations, twitch duration remained\ shorter and total tension was lower, but other in-
dexes ofcontractility had normalized. When 10-5Mphenylephrine (PHEN) in the presence of 10-5 Mpropranolol was added to high extracellular calciumconcentrations, total tension normalized in papil-lary muscles with dysfunctional vascular endothe-lium (-VE), but contraction duration remainedabbreviated.
by guest on October 18, 2011http://circres.ahajournals.org/Downloaded from
Li et al Vascular Endothelium and Myocardial Contractility 773
10
cmEE0e
z0
U)
z
a
4
2
O control*-EE
*10+v***
0/a'
1.2
E 0.9
E
%-00.6z .0
zW 0.3
0.0
10 15 20 25
Y=1.037-0.032 X
r=O.941
20 25 30 35
CALCIUM CONCENTRATION (mM)
FIGURE 4. Graphs showing the effect of calcium concentra-tion on tension. With increasing calcium concentrations, thedifference in total tension between papillary muscles with(control) and without (-EE) endocardial endothelium de-creases and becomes insignificant (top panel). The differencein total tension (A tension) between papillary muscles beforeand after endocardial endothelial removal with Triton X-100is shown in the bottom panel. *p<0.05 and **p<0.01 for-EE vs. control.
tion documented at lower calcium concentrations.Twitch duration remained shorter in all groups in whichendocardial or vascular endothelium was removed (Ta-ble 2). Also, a mild decrease in total tension persisted,a decrease that can be explained by the shortened timeto peak tension in those groups (Table 1). However, theaddition of 10` M phenylephrine to 15 mM calciumconcentration resulted in complete normalization of thetwitch tension, time to peak tension, dT/dt, and Vma,, ofpapillary muscles with intact endocardium but dysfunc-tional vascular endothelium (Figure 3, Table 2). Never-theless, twitch duration remained shorter. Finally, in-creasing extracellular calcium concentrations tosaturating bath calcium (25 mM) eliminated all differ-ences in total tension between papillary muscles withand without endocardial endothelium (Figure 4).
MorphologyScanning electron microscopy of the endothelial layer
of the large coronary vessels and their smaller brancheswas normal in both saline-injected (group 2) and TritonX-100-injected (group 3) Langendorff-perfused hearts(Figures SA and 5B). The viability of these endothelialcells was documented by lack of propidium iodidestaining of the nuclei and no change in the pattern ofactin filament staining in both groups (Figure 6A). Inarterioles and capillaries, there was occasional damageof endothelial cells in several vessel segments of group 3hearts; however, large areas were observed withoutpropidium iodide staining of nuclei, and the smoothmuscle cells of these vessel segments and their sur-
rounding myocytes were intact (Figure 6B). Thechanges in the microvascular bed were more markedthan those in the macrovascular bed (Figure 6C), withonly occasional damage of endothelial and myocardialcells being found in group 3 hearts, with no associationoccurring between damage of the two cellular groups. Ingroup 2 hearts, no changes in the vascular bed werefound; however, as occurred in group 3 hearts, occa-sional damage to myocardial cells was found. Interstitialedema was detected in both group 2 and group 3 hearts.
Scanning electron microscopy of the endocardial en-dothelium of the left and right ventricles revealed manytranscellular holes along the periphery of endocardialendothelial cells in both groups of hearts, indicating thatthis was not due to Triton X-100. However, the endocar-dial endothelium of the papillary muscles from the rightventricle usually showed less or no transcellular holesthan elsewhere in the heart, with most cells havingwell-developed intercellular borders (Figure 5C). Withconfocal laser-scanning light microscopy, very few en-docardial endothelial cells were found with propidiumiodide-stained nuclei, indicating cellular viability (Fig-ure 6D). However, at the periphery of the endocardialendothelial cells, circular structures outlined with F-ac-tin staining distorting the organization of peripheralactin bands were found (Figure 6D). Endocardial en-dothelial damage was similar in both saline-injected andTriton X-100-injected hearts.
DiscussionBy creating vascular endothelial dysfunction with a
bolus injection of Triton X-100 in Langendorfr-perfusedhearts, the myocardial performance of subsequentlyisolated papillary muscles from the same hearts wasmodified in a manner typical of that created by endocar-dial endothelial dysfunction, i.e., a decrease in twitchduration with a concomitant decrease in peak twitchtension with little or no change in VmaX.2,6 Endocardialendothelial removal had the same effects on papillarymuscle characteristics whether vascular endotheliumwas dysfunctional or not. These findings suggest thatcoronary vascular endothelium directly modulates thecontractile characteristics of adjacent myocardium in amanner similar to that of endocardial endothelium andthat these effects appear to be additive.
In this study, vascular endothelial dysfunction wascreated by a bolus injection of Triton X-100 into theaorta, just above the coronary arteries. This resulted inlittle or no change in the morphological characteristicsof both macrovascular and microvascular endotheliallayers or of the adjacent myocardium. Lack of celldamage was verified by staining with propidium iodide.In view of this apparently largely intact vascular endo-thelium, the functional integrity of the vascular endo-thelium was verified before isolation of the papillarymuscles by examining coronary perfusion rate respon-siveness to acetylcholine in the Langendorff heart prep-aration.11,16 Before the bolus injection of Triton X-100,acetylcholine caused relaxation of adjacent vascularsmooth muscles, suggesting the presence of intact func-tioning vascular endothelium. After the bolus injectionof Triton X-100, the lack of change in coronary perfu-sion rate in response to acetylcholine suggests that theTriton X-100 injection resulted in endothelial dysfunc-tion.1112,16 That coronary perfusion increased similarly
by guest on October 18, 2011http://circres.ahajournals.org/Downloaded from
FIGURE 5. Scanning electron microscopy ofperfused rabbit hearts treated with Triton X-100. PanelA illustrates the intact natureof the endothelial layer of the large coronary vessels. Endothelial cells have an elongated cell shape, an intact surface structurewith short microvilli4 and some bleblike protrusions. Magnification, x 3,000. Panel B illustrates a fractured coronary capillaryshowing a smooth luminal surface (arrow). Fine collagen fibers (arrowheads) surround the basal lamina covering the capillary andmyocytes. Magnification, x10, 700. Panel C illustrates that the endocardial endothelium near the tendon of a right ventricularpapillary muscle has an elongated cell shape, with well-developed intercellular borders (arrows) and a few holes through theendothelial layer (arrowheads). Magnification, x2,100.
in response to nitroprusside before and after the TritonX-100 injection supports an endothelium-mediated de-fect in vasodilatation after Triton X-100 injection andsuggests that the vasodilatory capacity of coronaryvascular smooth muscle was not altered.
In addition to abolishing the vasodilator response toacetylcholine, the bolus injection of Triton X-100 led toa 27% decrease in basal coronary flow. This decreasecould have been due to suppressed basal release ofcoronary vasodilating substances such as endothelium-derived relaxing factor,'6 much as has been shown tooccur with the use of another detergent (CHAPS) ornitro-L-arginine in the guinea pig isolated perfusedheart.36 An alternative possibility that cannot fully beexcluded at present is a compressive efTect of thecoronary vascular bed caused by perivascular edema.However, because some perivascular edema was pre-sent in both Triton X-100-injected and saline-injectedhearts, this second possibility is less probable.The Langendorff perfusion itself did not appear to
significantly alter myocardial performance of subse-quently isolated papillary muscles, nor did it modulateendothelium-dependent contractile characteristics. Thebolus injection of Triton X-100 in the coronary vascu-lature resulted in a change in contractile characteristicssimilar to those found after endocardial endothelial
removal.2 This is compatible with a direct effect of thevascular endothelium on the contractile characteristicsof adjacent myocardium, much as has been shown forthe endocardial endothelium and its adjacent myocar-dium.2 Endocardial endothelial removal in this studyhad a similar effect on the contractile characteristics ofall groups of muscles, whether intracoronary TritonX-100 bolus injection had been given or not, suggestingthat functionally intact endocardial endothelium waspresent in all preparations. It would thus appear, fromthese observations, that vascular endothelium modu-lates myocardial performance, regardless of the func-tional status of the endocardial endothelium, and thatthis occurs independently and to a similar degree inboth endothelial layers. It would also seem that themodulatory effects of both systems on myocardial con-traction are complementary. Finally, because the effectsof endocardial endothelium on myocardial contractilecharacteristics occurred despite some breaks in theendocardial endothelial layer and the effects of vascularendothelium could be inhibited despite a largely intactlayer of cells, it would appear that the effects ofendothelium on myocardial contractile characteristicsare not dependent on an intact cellular barrier. Never-theless, it is also possible that complete removal of the
by guest on October 18, 2011http://circres.ahajournals.org/Downloaded from
Li et al Vascular Endothelium and Myocardial Contractility
FIGURE 6. Confocal scanning light microscopy ofperfused rabbit heart treated with Triton X-100. PanelA: En face view of theendothelial layer of a coronary artery. Vascular endothelial cells were viable, as indicated by the lack ofstaining of the nuclei withpropidium iodide and the normal (yellow color) actin filament system. Bar, 25 gim. Panel B: Optical section through myocardiumofthe left ventricle from a perfused heart that was fixed with glutaraldehyde. Part of the microvascular tree showed highly elongatedand aligned nuclei of damaged endothelial cells in arterioles (arrows) and more rounded capillary nuclei (arrowheads) stainedwith propidium iodide. A large part of the tissue did not contain propidium iodide-stained nuclei. No damaged myocytes wereobserved in this area. Bar, 100 gm. Panel C: Optical section through myocardium showing myocytes and an arteriole. The striatedpattern of actin filaments (yellow color) and the absence ofpropidium iodide-stained nuclei indicate that the myocytes were viableat the time ofstaining. The tangentially sectioned arteriole, in the middle of the figure, showedfilaments ofconcentrically arrangedsmooth muscle cells (arrowhead). The elongated blue-colored nuclei in the arteriole are nuclei ofdamaged endothelial cells, sincethey are orientedparallel to the blood flow. Again, the absence ofpropidium iodide-stained nuclei indicates the viability of vascularsmooth muscle cells. Bar, 50 gm. Panel D: Confocal laser scanning lhght microscopy of an en face preparation of endocardialendothelium from right ventricular papillary muscles of hearts receiving intracoronary Triton X-100 (group 3) and stained withpropidium iodide. Only two dead cells with a propidium iodide-stained nucleus (blue color) were found, attesting to the viabilityof the endocardial endothelial cells. Peripheral actin bands (arrowheads) outlined the intercellular borders of endocardialendothelial cells. Bar, 10 gtm.
vascular endothelium would have resulted in an evenmore marked alteration of contractile characteristics.
In this study, intensive morphological investigationallowed us to verify the cellular integrity of the whole ofthe coronary vascular tree. The endothelial layer wasessentially intact with only occasional evidence of dam-aged endothelial cells. Experiments with acetylcholinepermitted us to also verify the functional integrity of themacrovascular system.1'12 To what extent the acetylcho-line experiments provided information about the func-tional integrity at the microvascular level remains opento question. However, because damage to vascularendothelial cells in Triton X-100-injected hearts ap-peared to be greater in the microvascular tree, there is
reason to believe that vascular endothelium in themicrovascular tree was also dysfunctional and that thiscontributed to the contractile changes documented inthe papillary muscles of the Triton X-100-injectedhearts. That vascular endothelium in the microvasculartree is in direct apposition to myocardium, much asendocardial endothelium is, would also appear to favorthe view that changes in contractile characteristicscaused by vascular endothelial dysfunction are at least,in part, the result of microvascular endothelial dysfunc-tion. Finally, although the presence of a well-developedglycocalyx on vascular endothelial cells may diminishthe penetration and the effects of detergents on mem-branes, detergents such as Triton X-100 (the nonionic
775
by guest on October 18, 2011http://circres.ahajournals.org/Downloaded from
detergent polyoxyethylene ether) seem to bind to cellmembranes even at very low concentrations and mayindeed, without lysis, change the functional propertiesof these membranes.37The results of studies done at extracellular calcium
concentrations of 15 mM suggest that the changes intotal twitch tension, dT/dt, and Vmax caused by endothe-lial removal were largely reversible and not due tomyocyte damage. At 15 mM calcium, the dT/dt and Vma,values of all groups were similar, indicating completereversibility of these abnormalities. As previously de-scribed,2 the twitch configuration remained shorter, afinding that can explain the slightly lower total tensiononce endocardial or vascular endothelial dysfunctionoccurred. Also, morphological studies with propidiumiodide found a similar but rare incidence of damagedmyocardial cells in both saline-injected and TritonX-100-injected hearts, suggesting that any myocardialdamage was the result of the Langendorff perfusionrather than the result of the Triton X-100 injection.Taken together, these findings suggest that the changesin contractile characteristics that resulted from thevarious interventions done in this study were the resultsof endothelial rather than myocardial dysfunction.
Nevertheless, as opposed to previous studies in myo-cardium from other species,25 increasing extracellularcalcium concentration to 15 mM was insufficient tonormalize either total twitch tension or time to peaktension. Only at greater extracellular calcium concen-trations of 20 and 25 mM was the difference in totaltwitch tension abolished between normal muscles andmuscles with the endocardial endothelium removed.This apparent discrepancy with the literature is in factcompatible with species differences in contractile char-acteristics38 and with recent concepts regarding themechanism by which endothelium modulates myocar-dial contractile characteristics. As compared with otherspecies in which the myocardial contractile effects ofendocardial endothelium have been evaluated, at phys-iological calcium concentrations, rabbit myocardium hasa relatively poor performance.38 It responds rather wellto increases in extracellular calcium concentrations, butas opposed to other species such as the cat, even at 15mM extracellular calcium concentration, rabbit myocar-dium is far from reaching maximal performance.3839 Afurther increase in extracellular calcium concentrationto at least 25 mM is necessary to reach maximalperformance. Alternatively, a further intervention suchas the addition of phenylephrine or endothelin638 isrequired to reach its full potential. It would thus appearthat, in rabbit papillary muscles, increasing the extra-cellular calcium concentration to at least 25 mM orincreasing the myofibrillar affinity for calcium at 15 mMcalcium concentration22,31,39-41 is necessary for the myo-cardium to reach maximal performance, with additionalinterventions not necessary in the cat or ferret.9,38
Recently, Wang and Morgan5 and McClellan et a131have demonstrated that both endocardial and vascularendothelium appear to modify the contractile charac-teristics of adjacent myocardium by increasing myofila-ment calcium responsiveness or affinity. In this studydone in rabbit myocardium, the addition of phenyleph-rine to 15 mM extracellular calcium concentrations wasnecessary to completely normalize total twitch tensionand time to peak total tension in papillary muscles with
dysfunctional vascular endothelium. Our findings arethus in full agreement with current knowledge56'31,38and reinforce the findings linking endothelial effects onmyocardium to changes in myocardial myofibrillar cal-cium affinity. Nevertheless, as occurred with endocar-dial endothelium,2 when vascular endothelium is re-moved, normalizing twitch tension is not accompaniedby a normalizing of twitch duration, underscoring theprobable multifactorial nature of the relation betweenendothelium and its adjacent myocardium.That modulation of the performance of adjacent
myocardium by endocardial and by vascular endothe-lium is complementary is not surprising. In early em-bryonic cardiogenesis, there is no coronary circulation,and exchange of substances between the luminal bloodand adjacent myocardium- and perhaps modulation ofits performance-is accomplished exclusively throughthe endocardial endothelium.1 The function of theendocardial endothelium is subsequently comple-mented and partially replaced by the endothelium of thecoronary circulation, the development of which appearsto accompany the formation of compact myocardium inthe outer portion of the ventricular wall. A reminiscenceof this developmental feature is present in fish andreptiles, in which capillaries were observed in theepicardial and outer compact regions of the ventricularwall only. In higher animals including, although notalways unequivocally, humans, the intertrabecularspaces regress as the coronary circulation develops. Thespaces are either reduced to strands of endotheliumwithout lumen or may give rise to capillaries in thecentral and inner portions of the myocardium withpersistence of ventricular communication through thethebesian and arteriosinusoidal vessels.
In conclusion, it would appear from this study thatcoronary vascular endothelium modulates the contrac-tile characteristics of adjacent myocardium in a mannersimilar to that of endocardial endothelium and thatthese effects are additive. Thus, it would appear that the"cross talk" between different cell types in the heart ismore widespread than previously thought.
AcknowledgmentsWe would like to acknowledge the expert technical assis-
tance of Hugues Gosselin and Lise C6teDougherty and theexpert secretarial and editorial help of Marguerite Cloutier.
References1. Brutsaert DL: The endocardium. Annu Rev Physiol 1989;51:
263-2732. Brutsaert DL, Meulemans AL, Sipido KR, Sys SU: Effects of
damaging the endocardial surface on the mechanical performanceof isolated cardiac muscle. Circ Res 1988;62:358-366
3. Smith JA, Shah AM, Lewis MJ: Factors released from the endocar-dium of the ferret and pig. J Physiol (Lond) 1991;439:1-14
4. Li K, Calderone A, Rouleau JL: Reduced myocardial alpha-adrenergic contractile responsiveness in pacing-overdrive model ofheart failure in the dog: Potential role of the endocardial endo-thelium. (abstract) Eur Heart J 1990;11(suppl):412
5. Wang J, Morgan JP: Endocardial endothelium modulates myofil-ament Ca2' responsiveness in aequorin-loaded ferret myocardium.Circ Res 1992;70:754-760
6. Li K, Stewart DJ, Rouleau J-L: Myocardial contractile actions ofendothelin-1 in rat and rabbit papillary muscles: Role of endocar-dial endothelium. Circ Res 1991;69:301-312
7. Shah AM, Meulemans AL, Brutsaert DL: Myocardial inotropicresponses to aggregating platelets and modulation by the endocar-dium. Circulation 1989;79:1315-1323
by guest on October 18, 2011http://circres.ahajournals.org/Downloaded from
epinephrine from adrenergic nerve endings of blood vessels ismodulated by endothelium-derived relaxing factor.Am JHypertens1990;3:211-218
15. Martin W, Furchgott RF, Villani GM, Jothianandan D: Depres-sion of contractile responses in rat aorta by spontaneously releasedendothelium-derived relaxing factor (EDRF). J Pharmacol ExpTher 1986;237:529-538
16. Stewart DJ, Baffour R: Functional state of the endothelium deter-mines the response of endothelin in the coronary circulation. Car-diovasc Res 1990;24:7-12
17. Stewart DJ, Holtz J, Pohl U, Bassenge E: Balance between endo-thelium-mediated dilating and direct constricting actions of sero-
tonin on resistance vessels in isolated rabbit heart. Eur Heart J1987;143:132-134
18. Warner TD, Mitchell JA, De Nucci G, Vane JR: Endothelin-1 andendothelin-3 release EDRF from isolated perfused arterial vesselsof the rat and rabbit. J Cardiovasc Pharmacol 1989;13(suppl 5):S85-S88
19. Yanagisawa M, Kurihara H, Kimura S, Tosnobe Y, Kobayashi M,Mitsui Y, Yazaki Y, Goto K, Masaki T: A novel potent vasocon-
strictor peptide produced by vascular endothelial cells. Nature1988;332:411-415
20. Ishikawa T, Yanagisawa M, Kimura S, Goto K, Masaki T: Positiveinotropic action of novel vasoconstrictor peptide endothelin on
guinea pig atria. Am J Physiol 1988;255(Heart Circ Physiol 23):H970-H973
21. Hu JR, Harsdorf RV, Lang RE: Endothelin has potent inotropiceffects in rat atria. Eur J Pharmacol 1988;158:275-278
22. Wang J, Paik G, Morgan JP: Endothelin 1 enhances myofilamentCa2l responsiveness in aequorin-loaded ferret myocardium. CircRes 1991;69:582-589
23. Watanabe T, Kusumoto K, Kitayoshi T, Shimamoto N: Positiveinotropic and vasoconstrictive effects of endothelin-1 in in vivo and
in vitro experiments: Characteristics and the role of L-type calciumchannels. J Cardiovasc Pharmacol 1989;13(suppl 5):S108-S111
24. Moravec CS, Reynolds EE, Stewart RW, Bond M: Endothelin is apositive inotropic agent in human and rat heart in vitro. BiochemBiophys Res Commun 1989;159:14-18
25. Vigne P, Lazdunski M, Frelin C: The inotropic effect of endothe-lin-1 on rat atria involves hydrolysis of phosphatidylinositol. FEBSLett 1989;249:143-146
26. Tohse N, Sperelakis N: cGMP inhibits the activity of single calciumchannels in embryonic chick heart cells. Circ Res 1991;69:325-331
27. George WJ, Wilkerson RD, Kadowitz PJ: Influence of acetylcho-line on contractile force and cyclic nucleotide levels in the isolatedperfused rat heart. J Pharmacol Exp Ther 1973;184:228-235
28. Dzau VT: Implications of local angiotensin production in cardio-vascular physiology and pharmacology. (abstract) Am J Cardiol1987;59:59A
29. Re RN: The myocardial intracellular renin-angiotensin system.AmJ Cardiol 1987;59:56A-58A
30. Lindpaintner K, Ganten D: The cardiac renin-angiotensin system:An appraisal of present experimental and clinical evidence. CircRes 1991;68:905-921
31. McClellan G, Weisberg A, Kato NS, Ramaciotti C, Sharkey A,Winegrad S: Contractile proteins in myocardial cells are regulatedby factor(s) released by blood vessels. Circ Res 1992;70:787-803
32. Kramer BK, Nishida M, Kelly RA, Smith TW: Endothelins: Myo-cardial actions of a new class of cytokines. Circulation 1992;85:350-356
33. Brutsaert DL, Claes VA, Goethals MA: Effect of calcium onforce-velocity-length relations of heart muscle of the cat. Circ Res1973;32:385-392
34. Andries LI, Meulemans AL, Brutsaert DL: Ultrasound as a novelmethod for selective in vitro and in vivo damage of the endocardialendothelium: A functional and morphological study. Am J Physiol1991;261:H1636-H1642
35. Andries LJ, Brutsaert DL: Junctional organization and permeabil-ity of endocardial endothelium. Cell Tissue Res (in press)
36. McLeod JD, Piper PJ: Effect of removing the endothelium on thevascular responses induced by leukotrienes C4 and D4 in guinea-pigisolated heart. Eur J Pharmacol 1992;212:67-72
37. Helenius A, Simons K: Solubilization of membranes by detergents.Biochem Biophys Acta 1975;415:29-79
38. Brutsaert DL, Sys SU: Ventricular function: Is the total more thanthe sum of the parts? Circulation 1991;83:1444-1449
39. Endoh M, Blinks JR: Actions of sympathomimetic amines on theCa2+ transients and contractions of rabbit myocardium: Reciprocalchanges in myofibrillar responsiveness to Ca2-mediated througha- and ,3-adrenoceptors. Circ Res 1988;62:247-265
40. Thieleczek R, Heilmeyer LMG Jr: Inositol 1,4,5-trisphosphateenhances Ca2+-sensitivity of the contractile mechanism of chemi-cally skinned rabbit skeletal muscle fibers. Biochem Biophys ResCommun 1986;135:662-669
41. Kelly RA, Eid H, Kramer BK, O'Neill M, Liang BT, Reers M,Smith TW: Endothelin enhances the contractile responsiveness ofadult rat ventricular myocytes to calcium by a pertussis toxin-sensitive pathway. J Clin Invest 1990;86:1164-1171
by guest on October 18, 2011http://circres.ahajournals.org/Downloaded from
