Author + information
- Received August 12, 1996
- Revision received July 14, 1997
- Accepted August 5, 1997
- Published online November 15, 1997.
- Jacob Gurevitch, MDA,*,
- Inna Frolkis, MD, PhDA,
- Yael Yuhas, PhDB,
- Beatriz Lifschitz-Mercer, MDA,
- Esther Berger, PhDA,
- Yosef Paz, MDA,
- Menachem Matsa, MDA,
- Amir Kramer, MD, PhDA and
- Rephael Mohr, MDA
- ↵*Dr. Jacob Gurevitch, Department of Thoracic and Cardiovascular Surgery, Ichilov Hospital, Elias Sourasky–Tel-Aviv Medical Center, 6 Weizman Street, Tel-Aviv 64239, Israel.
Objectives. This study sought to assess the importance of locally released or paracrine myocardial tumor necrosis factor-alpha (TNF-alpha) in the evolution of postischemic myocardial dysfunction and to use immunohistochemical studies to localize TNF-alpha within the myocardium.
Background. TNF-alpha is implicated as a systemic mediator in the development of myocardial ischemia–reperfusion injury by promoting leukocyte myocardial infiltration, and it has been shown to originate from noncardiac peripheral mononuclear cells. We have recently documented in a blood-free environment the release of TNF-alpha from the ischemic-reperfused myocardium.
Methods. Isolated rat hearts undergoing 1 h of global cardioplegia-induced ischemia and 30 min of reperfusion were investigated with use of the modified Langendorff model. Hearts were randomly divided into three subgroups: group A, control group; and groups B and C, isolated hearts receiving cardioplegic solution containing monoclonal hamster antimurine TNF-alpha antibodies (group B) or hamster IgG (group C).
Results. Significant amounts of TNF-alpha were detected in group A and group C effluent on 1 min of reperfusion (752 ± 212 and 958 ± 409 pmol/ml, respectively). However, in group B, TNF-alpha was below detectable levels. In this group, postischemic left ventricular peak systolic pressures, first derivative of the rise in left ventricular pressure (dP/dtmax), pressure-time integral, coronary flow and O2consumption improved (analysis of variance [ANOVA] p < 0.0001 for all variables) compared with values in groups A and C; creatine kinase levels decreased (p < 0.005); and myocardial structure was preserved. Immunohistochemical staining localized TNF-alpha to cardiac myocytes and to endothelial cells.
Conclusions. Anti–TNF-alpha neutralizes local TNF-alpha release from cardiac myocytes after ischemia and improves myocardial recovery during reperfusion, indicating that postischemic paracrine TNF-alpha release plays an active role in myocardial dysfunction.
Tumor necrosis factor-alpha (TNF-alpha), or cachectin, is a polypeptide hormone considered a common systemic mediator of stress-induced situations such as septic shock during endotoxemia , wasting in cancer and a variety of cardiac disease states [3–7]. It was first introduced in 1975 as an endotoxin-induced serum factor . Experimental evidence has recently shown that TNF-alpha is also involved in the pathogenesis of myocardial ischemia-reperfusion injury.
The current hypothesis is that TNF-alpha is primarily produced by noncardiac-activated macrophages in response to ischemia and reperfusion [9–11]. Support for this hypothesis is the finding of increased levels of TNF-alpha in the peritoneal macrophages collected from rats subjected to myocardial ischemia–reperfusion injury . The circulating TNF-alpha stimulates intracellular adhesion molecule-1 (ICAM-1) expression on cardiac myocytes , which in turn promotes adhesive interaction between transmigrated neutrophils and cardiac myocytes of the reperfused myocardium [13, 14]. The end result is the release of harmful substances, such as oxygen free radicals, leukotrienes and cytokines from these neutrophils [15–17].
In contrast to the preceding, we recently observed that rat myocardium releases significant amounts of TNF-alpha after prolonged ischemia. This latter observation raises the possibility that myocardial dysfunction that occurs after ischemia–reperfusion may develop directly in response to the compartmentalized production of TNF-alpha in the heart, as opposed to the systemic production of TNF-alpha by circulating mononuclear cells.
To examine this new pathway we tested whether treatment with anti–TNF-alpha monoclonal antibodies given before ischemia can attenuate ischemia-reperfusion myocardial injury in a blood-freeenvironment. The second goal of this study was to identify cellular origins of TNF-alpha production within the myocardium by using immunohistochemical strategies.
Male Wistar rats were anesthetized by intraperitoneal injection of pentobarbital sodium (30 mg/kg body weight) and heparinized. Their hearts were rapidly excised, immersed in cold saline solution (4°C) with heparin, and mounted on the stainless-steel cannula of a modified Langendorff perfusion apparatus.
Retrograde aortic perfusion was initiated at a perfusion pressure of 85 mm Hg with an oxygenated modified Krebs-Henseleit (KH) buffer solution of the following composition (mmol/liter): NaCl 118, KCl 4.7, CaCl 2.0, MgSO47H2O 1.2, KH2PO41.2, glucose 11.1 and NaHCO325. The perfusate was bubbled continuously with 95% O2and 5% CO2, maintaining a pH of 7.4 to 7.5. Values of Po2and Pco2in the perfusion solution were 450 to 550 mm Hg and 25 to 30 mm Hg, respectively.
The heart temperature was monitored by a thermistor implanted in the right ventricular wall and carefully maintained at 37°C or 31°C (at ischemia) by water jacketing the perfusate reservoir and the isolated heart. The right atrium was removed, and the heart was paced to 300 beats/min at 4 V with use of an external pacemaker (Devices Limited, Implants Division, type E4162), ensuring identical heart rates for all hearts. A water-filled latex balloon was placed in the left ventricular cavity through a small incision in the left atrium and was connected to a Mennen Medical PI 32284 pressure transducer. The balloon was tied and inflated to a volume that produced 0 mm Hg diastolic pressure. Zero calibration of the pressure transducer was examined throughout the experiment.
Left ventricular peak systolic pressure, time to peak systolic pressure, relaxation time, the first derivative of the rise and fall in left ventricular pressure (dP/dtmax, dP/dtmin), the area calculated under the left ventricular developed pressure curve (pressure-time integral) and coronary flow were measured. These variables were continuously recorded, and measurements were taken at 10-min intervals.
All animals received humane care as described in “Principles of Laboratory Animal Care” formulated by the National Society for Medical Research and the “Guide for the Care and Use of Laboratory Animals” prepared by the National Academy of Sciences and published by the National Institutes of Health (NIH publication no. 80-23, revised 1985).
Twenty-seven rats were randomly divided into three subgroups of nine animals each. Group A (the control group), undergoing ischemia and reperfusion; group B, isolated hearts receiving hamster antimurine TNF-alpha monoclonal antibodies (anti-TNF mAb) in the cardioplegic solution (50 mg/ml, total dose given 1.5 mg); and group C, isolated hearts receiving hamster IgG in the cardioplegic solution (50 mg/ml, total dose given 1.5 mg), serving as additional control to group B.
Control measurements were recorded after a 15-min period of stabilization, and each heart was perfused thereafter for 30 min. Warm cardioplegic solution was administered for 2 min (37°C, perfusion pressure 73 mm Hg, KCl = 16 mEq/liter in KH solution), and a 60-min period of global ischemia at 31°C was applied to the arrested heart.
Creatine kinase (CK) activity was measured spectophotometrically in the effluent at 1 min of reperfusion after ischemia. Measurements of left ventricular function were taken every 10 min during the 30 min of the reperfusion period. Finally the hearts were dried at 90°C for 24 h to achieve a constant dry weight. Throughout the study, experiments were alternated between the control and the experimental limbs to avoid bias or differences in results.
1.3 TNF-Alpha Determination
Efferent perfusate fluid samples for TNF-alpha measurements were withdrawn at baseline measurements (15 min after stabilization), after 30 min of perfusion, immediately after ischemia (1st ml), at 10 minutes of reperfusion and at termination of reperfusion (30 min); the samples were then immediately stored at −70°C until assay. TNF-alpha levels were determined on the basis of cytotoxic activity assay on mouse L929 cells according to the methods described by Wallach and Yuhas et al. . Each assay included a standard curve of recombinant human TNF-alpha (specific activity 2.5 × 107IU/mg protein) kindly provided by Dr. Armat from Reprogen Ltd, Rehovot, Israel. The limit of detection was 1 U/ml.
Anti-TNF mAb was purchased from Genzyme; ∼50 to 100 ng of this antibody will completely neutralize 1 U of murine TNF. Anti-TNF mAb was previously recognized to completely neutralize rat TNF [21, 22]. For group C control rats we used hamster IgG purchased from Jackson.
1.5 Oxygen Consumption
Perfusate afferent and efferent gases were measured after 15 min of stabilization, 1 min before administration of cardioplegic solution, at 10 min of reperfusion and at the end of reperfusion. Samples were withdrawn from the Langendorff perfusion apparatus and from the right ventricle by using a tiny polyethylene catheter inserted through a pulmonary artery incision. O2consumption was calculated by using the following formula :
1.6 Immunohistochemical Localization of Myocardial TNF-Alpha
To visualize the presence and anatomic localization of TNF-alpha within the myocardium, immunohistochemical studies were performed in three groups of isolated hearts: group 1, four hearts undergoing 1 h of cardioplegic ischemia; group 2, four hearts undergoing similar ischemia receiving anti-TNF mAb in the cardioplegic solution (50 mg/ml, total given dose of 1.5 mg; hearts from these two groups were freshly excised immediately after ischemia); and group 3, four isolated hearts excised after 120 min of perfusion (without ischemia) that served as a control group.
Excised hearts were fixed in 10% neutral buffered formalin (Z-fix; Anatech, Ltd.) at room temperature for 18 to 24 h. The tissue was then dehydrated, embedded in paraffin and sectioned (5-μm sections) in a standard manner. Immunohistochemical studies were performed by using a streptavidin-biotin immunoperoxidase method according to the manufacturer’s instructions (Zymed Lab. Inc.). All reagents were prepared with a dilluent consisting of Tris buffer pH 7.6 supplemented with 0.2% monolaurate (Tween 20) and 1% bovine serum albumin. To minimize background staining, all sections were first blocked with normal goat serum for 10 min at room temperature. The slides were incubated with two concentrations (1:100 or 1:250) of primary antibody directed against human TNF-alpha (Genzyme). Sections were allowed to incubate in a humidified chamber for 1 h at room temperature. The slides were then rinsed in Tris buffer and incubated for 10 min at room temperature with a biotin-conjugated goat anti-rabbit secondary antibody. To rule out nonspecific antibody staining, slides for negative control were stained with 1) normal nonimmune rabbit serum (1:100 dilution) instead of the primary antibody (n = 4); 2) secondary antibody alone (1:100 dilution, n = 4). After blocking the endogenous peroxidase activity with methanol and hydrogen peroxide, the slides were incubated with streptavidin-peroxidase complex. Diaminobenzidine was used as a chromogen to visualize the presence and distribution of TNF-alpha. Sections were then rinsed in Tris buffer counterstained in hematoxylin, dehydrated, cleared and mounted with use of a synthetic mounting medium.
To compare the intensity of TNF-alpha expression in a variety of cell types of the myocardium, relative intensity of staining was assayed by two independent observers. In this assay TNF-alpha staining ranged from the lowest intensity of positive staining (scored as 1) to very intense, dark staining (scored as 3). No staining was scored as 0. Mean value ± SE was calculated. Three or four consecutive tissue sections of each heart were assayed. The tissue sections of the same hearts were used for histologic studies using hematoxylin-eosin staining.
Results are presented as mean value ± SE. To avoid differences in baseline values, the value of the 1st 15 min of contraction of each heart was used as the heart’s individual control. All control values for left ventricular function, coronary effluent and O2consumption were considered as 100%. Between-group cytotoxic activity, CK levels and immunostaining intensity measurements were analyzed by paired Student ttest. Two-way analysis of variance (ANOVA) with repeated measurements for drug and time effect was calculated for all variables before and after ischemia. Significance was established at the level of p < 0.05.
All baseline variables taken after 15 min of stabilization were statistically similar for the three groups of hearts (Table 1). Moreover, the three groups maintained comparable variables during 30 min of the preischemic perfusion period (Fig. 1).
2.1 TNF-Alpha Release
Significant amounts of TNF-alpha were detected in group A and C effluent on the 1st min of reperfusion (752 ± 212 and 958 ± 409 pmol/ml, respectively). In all other samples taken during baseline measurements (after 15 min of stabilization), before ischemia (after 30 min of perfusion) and at 10 and 30 minutes of reperfusion, TNF-alpha was below detectable levels. All effluent cytotoxic activity was neutralized by preincubation with anti-TNF-alpha mAb: 62.5 ng of antibodies completely neutralized 1 U of effluent TNF (n = 6), whereas 31.25 ng of antibodies provided 80 ± 3% of complete neutralization (n = 6, p < 0.002), thus confirming that the cytotoxic activity was due to TNF-alpha. In all samples withdrawn from group B effluent (hearts treated with anti-TNF mAb), TNF-alpha was below detectable levels.
For additional evidence that primary TNF-alpha is indeed synthesized de novo by the isolated heart, we constructed another study subgroup. Isolated rat hearts (n = 4) received a nonspecific protein synthesis blocking agent (1.5 mg of cycloheximide) in the cardioplegic solution just before ischemia. After 1 h of ischemia, effluent TNF-alpha was below detectable levels.
2.2 Hemodynamic Changes
After ischemia, a significant deterioration in preischemic values was observed in all group A and C myocardial performance variables (Fig. 1). Hearts treated with anti-TNF mAb (group B) had a significant improvement over control group values in postischemic left ventricular peak systolic pressures (Fig. 1A), dP/dtmax(Fig. 1B) and dP/dtmin, the area calculated under the left ventricular developed pressure curve (pressure-time integral, Fig. 1C), coronary flow (Fig. 1D) and O2consumption (Fig. 1E) (ANOVA with repeated measurements, p < 0.0001 for all of these variables). The significant improvement in myocardial recovery of group B hearts was already observed at 10 min of reperfusion, and was maintained thereafter.
Group B postischemic effluent CK levels were significantly lower than those in groups A and C (42.7 ± 2.2 vs. 73.2 ± 5.6 and 67.0 ± 4.9 U/liter, respectively, p < 0.005). Adding IgG to the cardioplegic solution (group C, serving as an additional biologic control) did not protect against the ischemia-reperfusion myocardial injury, and no hemodynamic and enzymatic variables differed significantly from those in group A (control group).
2.3 Intramyocardial Immunochemical Localization of TNF-Alpha
TNF-alpha immunostaining was not detected in the hearts after 2 h of KH nonischemic perfusion (Fig. 2A). There was obvious TNF-alpha immunostaining in hearts after 1 h of ischemia (Fig. 2B). TNF-alpha staining was primarily localized in cardiac myocytes (immunostaining intensity 2.79 ± 0.06) and, to a lesser degree, in the endothelium of myocardial vessels (immunostaining intensity 1.35 ± 0.03, p < 0.001). Smooth muscles of the vessels were not immunostained. TNF-alpha staining was not detected in hearts pretreated with anti TNF mAb in the cardioplegic solution (group B), indicating compete neutralization of TNF-alpha synthesized during ischemia (Fig. 2C). There was no nonspecific myocardial immunostaining for TNF-alpha when ischemic hearts were incubated with either normal nonimmune rabbit serum instead of the primary antibody, or with secondary antibody alone (two groups of negative controls; data not shown).
2.4 Histologic Studies
The tissue sections of the same hearts were used for histologic studies using hematoxylin-eosin staining. After 2 h of nonischemic KH perfusion, myocardial structure was preserved (Fig. 2D), whereas after 1 h of cardioplegia-induced ischemia, myocardial damage was observed (Fig. 2E). Hearts pretreated with anti-TNF mAb in the cardioplegic solution (group B) demonstrated almost complete restoration of normal histologic features (Fig. 2F).
In this study significant amounts of TNF-alpha were detected in the effluent of the isolated rat heart undergoing cardioplegia-induced arrest and prolonged ischemia. Immunohistochemical studies localized TNF-alpha to cardiac myocytes and, to a lesser extent, to vascular endothelial cells. Primary TNF-alpha was produced during ischemia and was detectable only in the 1st min of reperfusion. We have proved here that this paracrine TNF-alpha release plays a dominant role in the ischemia-reperfusion injury, as neutralization with specific monoclonal antibodies against this cytokine (given before ischemia, in the cardioplegic solution) completely eliminated TNF-alpha from the effluent and concurrently attenuated the postischemic myocardial injury. Anti-TNF mAb improved the postischemic recovery and myocardial mechanical performance, decreased the amount of cellular necrosis and preserved myocardial structure.
3.1 TNF-Alpha: A Myocardial Depressant Factor
TNF-alpha is a proinflammatory cytokine with potent negative inotropic properties. Elevated levels of serumTNF-alpha were associated with depressed myocardial function . These TNF-alpha concentration- and time-dependent negative inotropic effects were reversible , and only recently appeared to be initiated by the activation of TNFR1 cell surface receptors. Treatment with anti-TNF mAb was shown to prevent myocardial dysfunction during experimental burn shock and in several cardiac pathologic conditions such as acute viral myocarditis and acute allograft rejection . The involvement of serum TNF-alpha in myocardial ischemia–reperfusion injury in the rat was indicated, as passive systemicimmunization against this TNF-alpha increased survival rates .
3.2 Primary Intramyocardial Synthesis of TNF-Alpha
In previous studies TNF-alpha was assumed to originate not from the ischemic heart but from systemic activated macrophages. We have demonstrated that primary TNF-alpha is synthesized in the ischemic isolated heart. As no TNF-alpha immunoreactivity was shown in the myocardium of 2 h of perfused control hearts, TNF-alpha cannot be derived in the cytoplasmic pool. In our study we deliberately used a blood-free perfusion model to exclude the possibility of involvement of systemic blood-borne factors in TNF-alpha production. However, we noted additional evidence that TNF-alpha is indeed synthesized de novo by the isolated heart in a relatively short ischemic time: Hearts treated with cycloheximide did not release TNF-alpha after 1 h of global ischemia.
By performing in vivo experimental regional ischemia, Herskowiz et al. found TNF-alpha messenger ribonucleic acid (mRNA) expression after only 30 min in the postischemic reperfused myocardium. Kapadia et al. demonstrated that when isolated hearts were stimulated with endotoxin in vitro, primary TNF messenger ribonucleic acid (mRNA) expression occurred within 30 min.
3.3 Sources of Intramyocardial TNF-Alpha
Immunohistochemical studies localized TNF-alpha to cardiac myocytes and to a lesser extent to endothelial cells and not to smooth muscle cells. TNF-alpha immunostaining was not apparent either in nonischemic normally perfused hearts or in hearts pretreated with anti-TNF mAb. The former findings emphasize the importance of ischemia to primary TNF-alpha induction or synthesis, whereas the latter strengthen the notion that TNF-alpha was neutralized not only in the effluent but probably in the tissue as well, and therefore its postischemic paracrine deleterious effect was eliminated. It is not unlikely that TNF-alpha release from the heart is derived quantitatively from different sources in various situations. Kapadia et al. found TNF-alpha in cardiac myocytes, endothelial and smooth muscle cells in endotoxin-stimulated isolated hearts; however, Herskowitz et al. found immunostaining of TNF-alpha in microvessels.
3.4 Possible Mechanisms for Anti-TNF Myocardial Protection
TNF-alpha is probably a linking milestone in the evolution of the ischemia-reperfusion injury. The negative inotropic effect of TNF-alpha is associated with decreased levels of intracellular calcium during the systolic contraction sequence and with decreased contractility. TNF-alpha is involved in the release of free radicals from the myocardium , a self-amplifying process, as free radical production has been shown to further increase TNF-alpha. TNF-alpha also acts by way of the NO pathway and causes myocardial dysfunction after ischemia .
Monoclonal antibodies against TNF-alpha (given in cardioplegic solution) saturate the extracellular environment of the ischemic heart, neutralize the TNF-alpha (released by the ischemic myocardium) and probably prevent the binding of TNF-alpha to TNFR1 receptors. Thus, a reasonable explanation for the cardioprotective effect of anti-TNF mAb might be that by eliminating TNF-alpha from the ischemic cascade, the entire snowball effect is significantly decreased, and therefore ischemic damage is also decreased. By a similar mechanism, soluble TNF-alpha binding proteins were shown to modulate the negative inotropic properties of TNF-alpha. However, we do not believe that TNF-alpha is the only factor responsible for the evolution of ischemic damage. By eliminating other factors (e.g., free radical scavenging) postischemic myocardial recovery is also improved, as all factors are probably linked.
TNF-alpha and TNF-alpha receptors are dynamically regulated in different pathologic situations. In failing human hearts TNF-alpha overexpression was one of several maladaptive mechanisms responsible for the progressive cardiac decompensation . Therefore, TNF-alpha might trigger its self-induction. In other words, the production of TNF-alpha by the ischemic heart might be a self-amplifying process. Anti-TNF mAb might intervene and stop the process. Anti-TNF mAb binds TNF-alpha and prevents its binding to cell membrane receptors that might be responsible for the further increase in TNF-alpha synthesis after ischemia. Anti-TNF mAb might prevent the induction of TNF-alpha in the ischemic myocardium in a yet unexplored way. This hypothesis might explain why there was no immunostaining for TNF-alpha in the anti-TNF mAb treatment group.
3.5 Limitations of the Study
The current study was designed to examine the role of paracrine TNF-alpha in the ischemic heart. Caution must be taken in withdrawing direct clinical conclusions from the use of anti-TNF mAb in a clinical setting, as the model used was an isolated rat heart model and the experiments were performed in a blood-free environment. Future studies using blood and neutrophil reperfusion need to be carried out.
3.6 New Insights
Our findings of local TNF-alpha synthesis and release from the isolatedischemic reperfused myocardium and the protective role of anti-TNF mAb on ischemia-reperfusion injury in a blood- and leukocyte-free environment can shed new light on previous hypotheses. In other words, TNF-alpha does not act only as an important substance in the recruitment of circulating leukocytes to sites of inflammatory lesions [9–15, 36]; it is probably synthesized during ischemia and released on reperfusion from the myocardium itself. TNF-alpha possesses a local insult or deleterious effect (myocardial depression) without the essential need for systemic involvement in this process. By diminishing this deleterious effect with specific treatment—monoclonal antibodies—against TNF-alpha, we have presented here unequivocal proof of the active role of local or paracrine myocardial TNF-alpha in the evolution of ischemia-reperfusion injury.
All represented data were statistically analyzed by Yael Villa, MSc, School of Mathematics, Tel-Aviv University. We thank Lynda Hemi for help in preparing the manuscript.
☆ This study was supported by the Research Fund of the Department of Thoracic and Cardiovascular Surgery, Elias Sourasky–Tel-Aviv Medical Center.
- analysis of variance
- anti-TNF mAb
- hamster antimurine tumor necrosis factor-alpha monoclonal antibodies
- creatine kinase
- first derivative of the rise in left ventricular pressure
- first derivative of the fall in left ventricular pressure
- intracellular adhesion molecule-1
- tumor necrosis factor
- tumor necrosis factor receptor-1
- Received August 12, 1996.
- Revision received July 14, 1997.
- Accepted August 5, 1997.
- The American College of Cardiology
- Tracey KJ,
- Beutler B,
- Lowrey SF,
- et al.
- Yamada T,
- Matsumori A,
- Sasayama S
- Levine B,
- Kalman J,
- Mayer L,
- Fillit HM,
- Packer M
- Carswell EA,
- Old LJ,
- Kassel RL,
- Green S,
- Fiore N,
- Williamson B
- Dinerman JL,
- Mehta JL
- Youker KA,
- Hawkins HK,
- Kukielka GL,
- et al.
- Gurevitch J,
- Frolkis I,
- Yuhas Y,
- et al.
- Wallach D
- Sheehan KCF,
- Ruddle NH,
- Schreiber RD
- Rabinovici R,
- Bugelski PJ,
- Esser KM,
- Killegas LM,
- Vernick J,
- Feuerstein G
- Neely JR,
- Liebermeister H,
- Battersby EJ,
- Morgan HE
- Torre-Amione G,
- Kapadia S,
- Lee J,
- Bies RD,
- Lebovitz R,
- Mann DL
- Giroir BP,
- Horton JW,
- White J,
- McIntyre KL,
- Lin CQ
- Murphy HS,
- Shayman JA,
- Till GO,
- et al.
- Pogrebniak HW,
- Matthews WA,
- Pass HI
- Finkel MS,
- Oddis CV,
- Jacob TD,
- Watkins SC,
- Hattler BG,
- Simmons RL
- Kapadia S,
- Torre-Amione G,
- Yokoyama T,
- Mann DL
- Torre-Amione G,
- Kapadia S,
- Lee J,
- et al.