INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Investigators have shown that inflammatory cytokines such as interleukin 1 (IL-1) and tumor necrosis factor- (TNF-) decrease contractility of isolated cardiac muscles via stimulation of nitric oxide (NO) production (1, 2). Subsequent studies demonstrated that the inducible form of nitric oxide synthase (iNOS) is found in the vascular and endocardial endothelia as well as cardiac smooth muscle. This enzyme, whose synthesis is induced after the administration of lipopolysaccharide (LPS) or cytokines, produces large amounts of NO over long periods of time. Therefore, given that NO is produced in large amounts in the myocardium after LPS administration, and given the negative inotropic effects of NO on the heart, it has been proposed that NO plays a major role in sepsis-induced myocardial dysfunction (3). To that effect, some studies have shown that nitric oxide synthase (NOS) inhibition improves myocardial function in sepsis (4). In contrast, other data argue against the role of NO in LPS or in TNF-/IL-1 pathophysiologic depression of myocardial contractility (7). Moreover, some investigators have suggested that the administration of NO to endotoxemic animals preserves rather than deteriorates ventricular function (10, 11). Thus, the role of NO in the pathophysiology of myocardial depression in sepsis is not well defined.

The reasons for these conflicting findings are not fully known, but may relate to several factors; first of which is the type of NOS inhibitor used. Because nonselective NOS inhibitors primarily target the constitutive form of NOS (cNOS), the suppression of constitutively produced NO may lead to loss of normal NO-mediated vasoregulatory and immunomodulatory functions. A second variable relates to the type of models used, as many of the studies were performed in vitro where many conditions are difficult to duplicate in vivo. Third, while investigators have demonstrated increased myocardial NO production in vivo after LPS administration (12, 13), none of these studies demonstrated that increased iNOS expression corresponds to diminishing myocardial contractility. This raises the possibility that increased myocardial NO may be a response to infection or inflammation, but may not depress contractility.

We undertook this study to define the relationship between myocardial NO content and myocardial contractility in a well established in vivo model of endotoxemia. This model offered the following advantage: myocardial contractility decreases at 16 h after LPS administration, and recovers at 48 h (14). This allowed us to study myocardial contractility and NO levels ex vivo during sickness and recovery. Furthermore, we investigated the effects of both nonselective and selective NOS inhibition on contractility and NO concentrations at these time periods. Additionally, we studied the hemodynamics in this model to better define the systemic actions of LPS and NOS inhibition.

We hypothesized that LPS administration would result in elevated NO levels in the myocardium, and that the time course of myocardial NO would correspond to a decrease in myocardial contractility. That is, myocardial NO content would be increased when contractility is decreased, and myocardial NO would be low when contractility has recovered. We further hypothesized that pretreatment of endotoxic animals with a selective iNOS inhibitor, by blocking mostly excess NO formation, results in improved contractility, whereas pretreatment with a nonselective inhibitor, by suppressing all NO synthesis, may have adverse effects on myocardial contractility.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This study was approved by the Institutional Animal Care and Use Committee. Male Sprague-Dawley rats weighing 250 to 350 g were used in this study. All chemicals were purchased from Sigma (St. Louis, MO). The rats were allowed to acclimatize to 12 h light/dark cycles for at least 5 d in the animal housing facility, and were given free access to food and water. Animals were randomized into two main batches, one that received endotoxin (LPS groups), and one that was injected with a similar volume of normal saline (control groups). All agents were administered intraperitoneally. In each batch, three groups of animals (n = 8 each group) were studied at baseline, 4, 16, and 48 h after the injection of either LPS or saline.

LPS Groups

Animals were randomized into three groups (n = 8 each group). Two groups were pretreated, 30 min before the LPS injection, with either the selective iNOS inhibitor S-methylisothiourea sulfate (SMT/LPS group), or the nonselective NOS inhibitor N^G-nitro-L-arginine methyl ester (L-NAME/LPS group). These agents were administered intraperitoneally at an equimolar dose of 50 µM/kg. The doses chosen are similar to those used by other investigators (15). An identical volume of normal saline was administered to the LPS only group. Thirty minutes later, Escherichia coli 05:B55 LPS (Difco Laboratories, Detroit, MI) was administered at a dose of 10 mg/kg. We had determined that this dose caused no mortality, and that contractility decreased at 16 h, and recovered by 48 h after LPS injection. Animals were studied at baseline (0 h), and then at 4, 16, and 48 h after LPS administration. Because SMT may possess antioxidant effects in addition to iNOS inhibition, we studied the effects of the iNOS inhibitor aminoguanidine (AG) which has not been reported to have antioxidant effects.

Control Groups

Three other groups of animals were assigned to the control batch that did not receive LPS, but received an equivalent volume of normal saline. One group (time control) received only normal saline intraperitoneally; the other two groups received SMT or L-NAME 30 min before the saline injection. Both NOS inhibitors were given at the equimolar dose of 50 µM/kg. Animals were studied at baseline (0 h), and at 4, 16, and 48 h after saline injection.

Papillary Muscle Preparation

For each time period, rats were anesthetized with the intraperitoneal injection of pentobarbitone (35 mg/kg). Through a median sternotomy, the heart was rapidly excised and placed in oxygenated (95% O₂ and 5% CO₂) Krebs-Henseleit solution (KHS) in a water-jacketed tissue bath at 32° C. The left ventricular chamber was opened and the free wall papillary muscle resected. The parameters recorded were peak tension (PT), time to peak tension (tPT), and the average rate of tension (dT/dt) calculated as PT divided by tPT as described previously (14). The mean of 12 contractions at each recording time was taken as the data point for that time interval. Preparations were usually stable for 6 h. Solutions of KHS were prepared fresh on the day of use; the pH was between 7.30 and 7.40.

Myocardial NO Content

NO content in left ventricular (LV) tissue was measured as the NO oxidation products nitrite/nitrate (NO_x) by the chemiluminescence technique (16). NO oxidation products were expressed as micrograms per gram of tissue wet weight (µg/gtww). As the papillary muscle was being removed for contractility studies, a segment (about 2 to 3 mm) of adjoining LV free wall was excised, avoiding any major blood vessel.

To ensure that differences in regional myocardial NO production would not influence results, in a separate set of experiments we assessed NO content in papillary muscle and adjacent LV in LPS animals at baseline, 4, 16, and 48 h.

Hemodynamic Measurements

Separate groups (n = 8 each group) of animals were anesthetized with 15 to 20 mg/kg of pentobarbitone, and placed in supine position with the limbs secured. Animals were breathing spontaneously and body temperature was maintained at 37° C with a temperature-regulated heating pad. One percent lidocaine was used for local anesthesia, and the internal jugular artery was exposed and cannulated with polyethylene (PE-20) tubing. The tubing was connected to a Gould 2400 chart recorder (Gould Instruments, Valley View, OH) for measurements of mean arterial blood pressure () and heart rate (HR). For cardiac output () measurements, a thermistor probe was inserted into the internal carotid artery to the level of the aortic arch. The probe was connected to a thermodilution cardiac output computer, Cardiotherm 500 (Columbus Instruments, Columbus, OH). Animals were allowed to stabilize for 30 min before recording any measurements. The depth of the anesthesia was checked every 20 to 30 min by stimulating whiskers and footpads; additional anesthesia was administered if needed. Five measurements were recorded from each animal, and the mean calculated.

Data Analysis

We used the software Sigmastat (Jandel, San Rafael, CA) to analyze the findings. Data were compiled and expressed as mean ± SD. Data were analyzed by two-factor analysis of variance (ANOVA), one factor being time, the other drug. When differences reached statistical significance, separate one-way ANOVA for repeated measures were done. If significance was found, a post hoc analysis (Newman-Keuls) was used to determine the source of significance. The null hypothesis was rejected at the 5% level.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animals that received endotoxin displayed visible signs of systemic illness including lethargy, piloerection, lack of grooming, periocular drainage, epistaxis, and diarrhea. These symptoms were noted as early as 4 h after LPS in some animals, were present in all animals by 8 to 10 h, and continued over the next 10 to 14 h. By 48 h after LPS injection, all animals had visibly recovered and there was no mortality. Unlike the LPS animals, the rats that received normal saline (control groups) did not show any symptoms.

There was a 33% mortality in the animals that received LPS and were pretreated with L-NAME (L-NAME/LPS group). These animals appeared more lethargic than the LPS only and the SMT/LPS groups. Mortality in the L-NAME/LPS group occurred between 10 and 14 h after LPS injection. Because of the increased mortality in this group, it was necessary to include more animals to achieve the number of eight animals in each group. In the L-NAME/LPS group, the animals that survived appeared normal by 48 h. There was no mortality in the SMT/LPS group, and there were no noticeable differences in this group's illness or recovery when compared with the LPS only group.

Papillary Muscle Contractility

Because there were no changes in tPT in any of the groups, the average rate of tension generation (dT/dt) reflected only changes in PT. Thus for ease of illustration, we show only the changes in PT. Two-way ANOVA revealed significant differences between treatment groups and time periods.

LPS groups. Figure 1A shows the changes in PT for the LPS groups. LPS caused a decrease in PT at 16 h (p < 0.0001 compared with baseline) with recovery to baseline levels occurring by 48 h. There was no change in contractility at the 4-h time period after administration of LPS.

View larger version (45K): [in this window] [in a new window]   Figure 1.   Time versus peak tension in grams in LPS (A) and control (B) groups. Differences between LPS, L-NAME, SMT groups and between time periods by two-way ANOVA. *Significance to baseline value. Significance to baseline and to the LPS or saline groups at same time period.

In the L-NAME/LPS animals that survived, an even more pronounced drop in contractility at 16 h (p < 0.001 versus LPS group) was observed. Furthermore, recovery of contractility did not occur by 48 h in the L-NAME/LPS group as it did in the LPS only group.

Pretreatment with an equimolar dose of SMT (SMT/LPS group) prevented the LPS-induced reduction in contractility at 16 h. No change in contractility was noted at the 4-h period.

Control groups (Figure 1B). The administration of L-NAME caused a significant drop in PT at 16 h (p < 0.0001 versus baseline). Recovery to baseline levels did occur by 48 h in the L-NAME group. There were no changes in contractility in either the saline only, or the SMT groups.

Myocardial NO Content in the LPS Groups (Figure 2A)

LPS only group. Two-way ANOVA revealed significant differences in both time and treatment. LPS caused a significant increase in tissue NO content noted at 4 h (p < 0.007 versus baseline). This increase was more pronounced by 16 h and was sustained at 48 h.

L-NAME/LPS group. L-NAME did not alter baseline myocardial NO release assessed 30 min later; however, the pretreatment with this agent reduced LV NO levels at all other time intervals (p < 0.03, 0.0001, and 0.01 when compared with LPS alone at 4, 16, and 48 h). L-NAME prevented the increase in myocardial NO concentrations at 4 h as the NO concentration was no different from the baseline value. LV NO content was elevated when compared with baseline at 16 and 48 h in the L-NAME/LPS groups, but these levels were lower when compared with the LPS only group and the SMT/LPS group (p < 0.0001).

SMT/LPS group. When SMT was administered 30 min before LPS, it reduced NO concentrations at the 4-h time period (p < 0.01 versus LPS group), but this level was still higher than baseline, and higher than the 4-h value in the L-NAME/LPS group. NO content at the 16-h time period was also reduced compared with the LPS only group (p < 0.007). However, LV NO level at 16 h in the SMT/LPS group was higher than the L-NAME/LPS group.

Myocardial NO Levels in the Control Groups

LV NO concentrations did not change with either saline, L-NAME, or SMT (Figure 2B).

View larger version (25K): [in this window] [in a new window]   Figure 2.   LV nitrite/nitrate (NOx) content expressed as µg per g of LV mass in LPS groups (A) and normals (B). Treatment and time different by two-way ANOVA. *Significance to baseline; Significance to baseline and LPS or saline group; Significance to LPS and SMT groups at same time period.

Hemodynamic Measurements in the LPS Groups (Figure 3)

LPS only group. The administration of LPS did not cause a change in at any time period. LPS did cause an increase in at 16 h (p < 0.0001), which persisted until 48 h. As a result systemic vascular resistance (Rsv) decreased at 16 and 48 h. LPS did not affect HR.

L-NAME/LPS group. Pretreatment with L-NAME caused an increase in at 4 h that was sustained until 16 h. decreased by 4 h after L-NAME (p < 0.0001 versus baseline); however, by 16 h, had rebounded above baseline value (p < 0.001) while blood pressure remained elevated. By 48 h and had returned to baseline value. L-NAME also caused an increase in Rsv at 4 and 16 h (p < 0.001 for both). L-NAME did not affect HR.

SMT/LPS group. SMT pretreatment prevented the increase in at 16 and 48 h observed with LPS alone. SMT affected neither nor HR.

Hemodynamic Measurement in the Control Groups (Figure 3)

There were no changes in hemodynamics in the saline only group. L-NAME caused an increase in at 4 and 16 h (p < 0.0006 and 0.01 versus baseline), and returned to baseline values by 48 h. L-NAME also caused a decrease in at 4 and 16 h (p < 0.005 and 0.001 versus baseline). returned to baseline value by 48 h. As a result of changes in and , Rsv increased at 4 and 16 h after L-NAME (p < 0.0001). The administration of SMT to the control rats did not affect any hemodynamic parameters. HR was not affected by any agent.

The results of AG administration are shown in Figure 4. Similar to SMT, the administration of this iNOS inhibitor restored LV contractility at 16 h (p < 0.001 when compared with LPS and L-NAME/LPS groups).

View larger version (19K): [in this window] [in a new window]   Figure 4.   Effects of AG administration on peak tension in endotoxic animals. AG had similar effect as SMT, and restored peak tension at 16 h when compared with LPS and L-NAME/LPS animals. Significance to LPS and L-NAME/LPS groups.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The main findings in this in vivo model of endotoxemia are that LPS causes a depression in myocardial contractility at 16 h, and this is associated with increased myocardial NO content at that time. However, elevated LV NO levels cannot be entirely related to decreased contractility, because at 48 h, when contractility had recovered in the LPS group, myocardial NO content was the same as the 16-h time period. Furthermore, SMT, a selective iNOS inhibitor, decreased myocardial NO at the 4- and 16-h periods when compared with the LPS groups at the same time periods, but these levels were higher than the L-NAME/LPS groups. Additionally, SMT restored myocardial contractility in endotoxemic rats at 16 h, and did not affect . However, for equimolar doses, the nonselective NOS inhibitor L-NAME increased , decreased , prevented the early (4-h) increase in myocardial NO in endotoxemic rats, depressed NO levels further at 16 and 48 h, exacerbated the drop in contractility, and increased mortality. Furthermore, L-NAME increased , decreased , and depressed contractility in control animals that did not receive LPS.

To our knowledge, this is the first in vivo study that relates LV NO content to contractility in endotoxemia during both illness and recovery. We chose to assess contractility in a manner independent of imposed loads from venous return, right and left ventricles, and sympathetic activity. Furthermore, we chose a model whereby contractility decreases at a later time period (16 h) so as to curtail any controversy of the early versus late effects of NO on sepsis-induced myocardial contractility (17). We evaluated LV NO content rather than serum levels because the latter may not reflect tissue levels. Moreover, measurements of tissue NO_x accurately reflect NOS activity in vivo (18).

Multiple in vitro studies have linked NO production to depressed myocardial function. Investigators also detected release of NO by the myocytes in response to the activated medium. Furthermore, the use of nonselective NOS inhibition completely restored the reduced response to baseline levels (1). Other investigators have demonstrated increased myocardial NO production in vivo. Liu and coworkers (12) demonstrated increased iNOS messenger RNA (mRNA) in the rat heart as early as 20 min after LPS administration. Inducible NOS mRNA increased markedly in the heart and remained elevated until 24 h. Similar results were obtained by Bateson and coworkers (13). Therefore, the weight of the evidence suggests that iNOS is activated early on in the septic rat myocardium presumably leading to the production of large amounts of myocardial NO. Yet, none of these studies evaluated myocardial function simultaneously. Thus, it is unclear from these experiments whether the increased NO production in the rat myocardium in vivo and in endotoxemia is neutral, beneficial, or detrimental. Moreover, although these studies confirmed increased NO production in the heart after LPS administration, none extended their observation beyond 24 h.

Because NO has been shown to decrease myocardial contractility in vitro in endotoxemia, NOS inhibition would have been expected to improve contractility in vivo. This has not been easy to demonstrate. In two studies of animal model endotoxemia, Herbertson and coworkers (5) and Kaszaki and coworkers (6), found that while the use of a nonselective NOS inhibitor improves LV contractility, its overall effects on were detrimental owing to the effects of NO inhibition on pulmonary arterial pressure. These studies illustrate the difficulty of performing experiments in whole animals in sepsis in which multiple changes occur simultaneously in several systems.

Other studies have been unable to attribute myocardial depression to NO in animal models (8, 9). On the other hand, there is evidence that early (minutes to few hours) contractile dysfunction in sepsis may be NO independent (19), and these studies did not assess contractility past the 6-h time period.

Paradoxically, there is evidence that NO has beneficial effects on the myocardium in sepsis. Ishihara an coworkers (10) evaluated the effects of NO inhalation on LV contractility. NO inhalation maintained LV contractility and prevented its impairment in a rat model of endotoxemia. Additionally, Zhang and coworkers (11) studied the effects of the NO donor 3-morpholinosydnonimine (SIN-1) in a dog model of endotoxemia. This drug increased cardiac index and regional blood flow without a detrimental impact on . Moreover, NO may have other beneficial effects in sepsis. It inhibits platelet aggregation and may regulate neutrophil recruitment by controlling microvascular permeability (20).

These results indicate that the effects of NO on myocardial function in endotoxemia defy a simple model. We attempted to clarify this dilemma by relating LV NO content to contractility. We postulated that some constitutive NO is necessary for normal cardiac function, and hence nonselective inhibition would result in further deterioration of contractility. We assumed that excess NO is harmful, and that selective inhibition would improve contractility by inhibiting mostly excess NO. Our data corroborate some of these assumptions: inhibition of excess NO synthesis by SMT prevented myocardial dysfunction, and a greater inhibition of NO synthesis by L-NAME exacerbated myocardial dysfunction. Our findings also indicate that some level of increased NO production in the myocardium in response to LPS is vital. Furthermore, recovery of myocardial function (48 h) occurred in spite of continued elevation of myocardial NO content. Thus, our results point to a greater complexity in the myocardial effects of NO in endotoxemia. The simple hypothesis that elevated NO results in myocardial dysfunction and that NOS inhibition improves this dysfunction is no longer adequate.

It may be that increased LV NO content does not affect contractility. If this were the case, then L-NAME and SMT would have other effects on LV function unrelated to NOS inhibition. For example, there is evidence to indicate that L-NAME may increase oxidative stress in sepsis (21). Additionally, because SMT contains a thiol group, it may posses antioxidant effects, although Szabo and coworkers failed to attribute such actions to SMT in vitro (22).

It is equally possible that high NO concentrations at certain times are more detrimental than other times in endotoxemia. Thus, it is not only the tissue content of NO that is important; the associated cellular milieu at that time may be equally pivotal. Accordingly, the biological chemistry of NO must be considered. Because NO is a free radical, it readily interacts with thiol and metal-containing enzymes via redox additive chemistry (23). Because such enzymes generally serve as major signaling proteins, NO can thus affect cellular function. However, whether the result is beneficial or detrimental depends on which interaction takes place. The implication is that NO may have different effects on the myocardium depending not only on its concentration, but also the timing and associated cellular conditions such as oxidative stress or thiol availability. For example, N-nitration of DNA and covalent modifications of tyrosine residues by NO, both potential mechanisms of cell damage and toxicity, are more likely to occur under conditions of oxidative stress (24).

Interestingly, L-NAME also decreased contractility and increased in normal animals and appeared to do so without a reduction in myocardial NO content. Although the reasons for this adverse effect of L-NAME in vivo are unknown, they have been demonstrated in other studies (25, 26). Some investigators have postulated that L-NAME impairs perfusion to the heart. However, in previous studies in our laboratory we showed that L-NAME does not impair myocardial perfusion (27, 28). As was the case in this study, others have shown that baseline NO production in the heart is quite low (29); thus, L-NAME may have depressed LV NO content in the normal animals, but this decrease may have been beyond the detection limit of our assay.

Limitations of this study. We did not assess LV contractility and LV NO content from identical locations, and this may be a source of error because regional differences in NO production may exist. However, in additional experiments, we found no difference in NO content between papillary muscle and adjacent LV fragment (data not shown).

The methodology used in this project does not allow for determination of NOS enzyme activities, nor for identification of cells responsible for NO production. However, this was not the purpose of this study. One of our stated aims was to correlate LV NO content with an index of LV contractility, and this purpose was accomplished.

While SMT is a powerful inhibitor of iNOS, and has been shown to be superior than other treatment strategies in rat models of sepsis (15, 30), SMT inhibits cNOS as well (22). Rosselet and coworkers (30) demonstrated in normal rats that high-dose SMT (1 mg/kg/h) increased , whereas low-dose SMT (0.1 mg/kg/h) did not. In the present study the equimolar doses of the two NOS inhibitors had different hemodynamic effects in endotoxemic and normal rats. Hence, we postulate that the predominate action of SMT was on iNOS inhibition.

As noted previously, SMT may possess antioxidant properties in addition to its effects on iNOS. It is then possible that antioxidant effects caused improved LV contractility. Therefore, in separate groups of animals (n = 3), we administered the selective iNOS inhibitor AG (31), at a dose of 60 mg/kg intraperitoneally, 30 min before LPS administration. Similar to our findings with SMT, AG restored papillary muscle contractility at 16 h (Figure 4).

In an in vivo study such as this one, it is difficult to isolate the factors responsible for LV dysfunction; while in vitro, where the environment is more controlled, conclusions can be more readily made. However, we noted from the preceding discussion that in vitro results have not always been reproducible in whole animals. Ultimately, the effects of NO in sepsis must be determined in vivo. For these reasons, we tested the effects of LPS and NOS inhibitors in vivo, as opposed to adding agents directly to the tissue bath. We should note that in preliminary experiments, we did examine the effects of L-NAME on contractility by adding it at increasing concentrations to the muscle bath. After adjustment of the pH of the medium to 7.4 using Hepes, papillary muscle contraction decreased with high doses of L-NAME. The physiological significance of L-NAME decreasing contractility at high doses, in vitro, and in altered KHS is not readily apparent. One such in vitro study added NOS inhibitors to an isolated muscle bath after contractile tension was found to be decreased 16 h after LPS administration. In that study, neither selective nor nonselective NOS inhibition reversed the depression in contractile function (32).

Finally, it has been suggested that NO interacts with superoxide to form the potent cytotoxic oxidant peroxynitrite. It has been further hypothesized that it is peroxynitrite rather than NO that is responsible for harmful effects (33). Indeed, this hypothesis has provided a biochemical rationale to account for the perplexing effects following the administration of NOS inhibitors. We did not assess peroxynitrite tissue concentrations in this study. It is conceivable that the detrimental effects on the myocardium at 16 h were caused by peroxynitrite. However, other investigators have suggested that peroxynitrite formation may represent a major detoxification and anti-inflammatory pathway for the removal of O₂ without the concomitant formation of H₂O₂. These investigators believe that the physiological actions of peroxynitrite may depend on the oxidative and nitrosative stress of the tissue at that time (24). Thus, the determination of peroxynitrite levels alone will not be meaningful without the establishment of the corresponding oxidative and nitrosative stress of the myocardium at the same time periods. This is certainly beyond the stated aims of this study. However, because we showed that in vivo myocardial NO concentrations do not necessarily correlate with decreased contractility, and postulated that the cellular milieu may play a large part, the issue of oxidative and nitrosative stress needs to be investigated.

We conclude that selective iNOS inhibition is beneficial to myocardial contractility, whereas nonselective is clearly detrimental and increases mortality. Furthermore, our results indicate that the effects of NO in the myocardium in sepsis cannot be simply explained by elevated NO content, but may be related to the timing and the associated tissue milieu as well.