Main

GBS, a Gram-positive bacterium, is one of the most common causal agents of neonatal sepsis(1, 2). In the newborn, GBS produced acute pulmonary hypertension, respiratory failure, arterial hypoxemia, decreased cardiac output, and systemic hypotension, resulting in significant morbidity and mortality(2–4).

Endogenous production and release of several vasoactive agents play a determinant role in the pathophysiology of sepsis(5–16). Thus, sepsis-induced pulmonary hypertension has been divided into an early phase related to the arachidonic acid-derived vasoconstrictor TXA2 and a late phase associated with development of pulmonary edema(5–7). Additionally, in an attempt to counteract systemic severe hypotension, circulating catecholamines are increased(8, 9). Moreover, elevated levels of the endothelium-derived vasoconstrictor ET-1 have been found in animal models of sepsis(10) as well as in adult patients with septic shock(11). Unfortunately, to our knowledge, the levels of ET-1 and catecholamines have not been evaluated in GBS sepsis. Together with these vasoconstrictors, endogenous vasodilators are also produced in sepsis(12–16). Currently, there is a great body of evidence that enhanced release of NO via induction of an inducible Ca2+-independent iNOS plays an important role in the loss of systemic(12–14) and pulmonary(15, 16) vascular responsiveness that occurs in Gram-negative sepsis. Gram-positive bacteria have been also recently demonstrated to induce iNOS in several tissues(17, 18). In fact, GBS induced iNOS in macrophages(19, 20) and in pulmonary arteries resulting in a reduced vasoconstrictor response to NA(21).

Therefore, the coexistence of pulmonary hypertension and systemic hypotension may be due to a different sensitivity of the pulmonary and systemic vessels to the vasoactive factors released in sepsis. The aim of the present work was to study the effects of in vitro incubation with heat-killed GBS on the vasoconstrictor responses to NA, the stable TXA2 analog U46619, and ET-1 on isolated pulmonary and mesenteric arteries from piglets.

METHODS

Tissue Preparation and Incubation

Male neonatal piglets (32 animals, 10-17 d of age, 4162 ± 297 g) were used in this study. Piglets were killed by exsanguination, and the lungs and mesenteric vascular beds were rapidly immersed in cold (4 °C) Krebs solution of the following composition (mM): NaCl 118, KCl 4.75, NaHCO3 25, MgSO4 1.2, CaCl2 2.0, KH2PO4 1.2, and glucose 11. The pulmonary (third branch, internal diameter 1-2 mm) and mesenteric arteries (internal diameter 1-2 mm) were carefully dissected free of surrounding tissue and cut into rings of 2-3-mm length(21). The arterial rings were incubated in Krebs solution gassed with 95% O2 and 5% CO2 at 37 °C in the presence of vehicle or heat-inactivated GBS (3 × 107 cfu mL-1) for 20 h. Krebs solution was supplemented with ampicillin (10 μg mL-1) and gentamicin (10 μg mL-1) to avoid bacterial growth during the dissection and incubation procedures. A maximal number of two rings per animal were used in each experimental group. Control and GBS-treated rings from the same animal were always run in parallel. After the incubation, two L-shaped stainless-steel wires were inserted into the arterial lumen, and the rings were introduced in Allhin organ chambers filled with Krebs solution gassed with 95% O2 and 5% CO2 and maintained at 37 °C. One wire was attached to the chamber and the other to an isometric force-displacement transducer (Grass FT07; Grass Instrument Co., Quincy, MA) connected to a polygraph (Grass model 7). The rings were stretched to a resting tension of 0.5 g (pulmonary rings) or 2 g (mesenteric rings) and allowed to equilibrate for 60-90 min. During this period tissues were restretched and washed every 30 min with warm Krebs solution. In some experiments the endothelium was removed by gently rubbing the intimal surface of the rings with a metal rod. The presence or absence of functional endothelium was verified by testing the relaxant effect of acetylcholine (10-6 M) in arteries precontracted with NA 10-6 M.

Experimental Protocol

In pulmonary and mesenteric rings previously incubated as described above, concentration-response curves to NA (10-8 to 10-4 M), U46619(10-10 to 10-6 M), or ET-1 (10-11 to 3 × 10-9 M), were constructed by increasing the organ chamber concentration by cumulative additions after a steady state response was reached after each increment. In some arteries the response to NA, U46619, or ET-1 were performed in the presence of L-NAME, 10-4 M, an inhibitor of NO synthesis), added 30 min before starting the concentration-response curves.

Drugs and Heat-Killed GBS Preparation

The following drugs were used: (-)-noradrenaline bitartrate, acetylcholine chloride, L-NAME, human ET-1, and U46619 (Sigma Chemical Co., London). All drugs were dissolved in distilled deionized water. The concentrations are expressed as a final molar concentration in the tissue chamber.

GBS type III was isolated from the blood of a neonate who developed early-onset sepsis. Bacteria were grown in Todd-Hewitt broth for 18-36 h at 37°C to late log phase and harvested by centrifugation at 5000 rpm for 15 min. Bacteria were resuspended in sterile isotonic saline to a concentration determined by serial viable counts to be 1 × 109 cfu mL-1. Heat-killed bacteria were obtained by heating bacteria to 60 °C for 60 min. GBS killing was confirmed by no growth on blood agar. Endotoxin levels in the heat-killed GBS preparation were undetectable as assayed using a standard Limulus assay kit (Sigma Chemical Co.). Aliquots of heat-killed GBS were stored at -80 °C until the study day.

Statistical Analysis

Results are expressed as means ± SEM of measurements in n arteries. Individual cumulative concentration-response curves were fitted to a logistic equation. The drug concentration exhibiting 50% of the maximal effect(Emax) was calculated from the fitted curve for each ring and expressed as negative log molar (pD2). Statistically significant differences were calculated by means of an unpaired t test. p < 0.05 was considered statistically significant.

RESULTS

Effects of GBS on NA-Induced Contractions

Both pulmonary and mesenteric arterial rings, incubated for 20 h with heat-killed GBS and then transferred to organ baths in the absence of GBS, showed a significant reduction (p < 0.01) in the contractile response to NA (10-8 to 10-4 M) compared with controls(Fig. 1). The reduction of the maximal contraction(Emax) was more marked in mesenteric compared with pulmonary arteries (Table 1). However, a decrease of thepD2 value was observed in pulmonary but not in mesenteric arteries. To determine whether NO production might have influenced the GBS-mediated hyporesponsiveness to NA, the concentration-response curve to NA was performed in the presence of L-NAME (10-4 M, a concentration which inhibits >80% of NO synthesis by the constitutive and inducible isoforms of NOS)(22). Fig. 2 shows that L-NAME induced an upward shift of the concentration-response to NA reversing the hyporesponsiveness to NA induced by GBS in both pulmonary and mesenteric arteries.

Figure 1
figure 1

Concentration-response curves to NA (10-8 to 10-4 M) in pulmonary (A) or mesenteric (B) artery rings preincubated for 20 h in Krebs solution in the presence of vehicle(○) or GBS (3 × 107 cfu mL-1, ▴). The concentration-response curves to NA were carried out in the absence of GBS. Data are expressed as means ± SEM of 8-21 observations. Ordinate, contractile response (mg); abscissa, log NA concentration (M).*p < 0.05 and **p < 0.01 compared with control group.

Full size image
Table 1 Effects of incubation with heat-killed GBS (3× 107 cfu mL-1) for 20 h on the parameters(Emax and pD2) of the concentration-response curves to NA and U46619 in pulmonary and mesenteric arteries (calculated from Figs. 1–4)
Full size table
Figure 2
figure 2

Effects of L-NAME (10-4 M) on GBS-induced hyporeactivity to NA in pulmonary (A) or mesenteric (B) artery rings. Concentration-response curves to NA (10-8 to 10-4 M) were performed in rings previously incubated for 20 h in Krebs solution in the presence of vehicle (○) or GBS (3 × 107 cfu mL-1,▴). L-NAME was added 30 min before the addition of NA. The curves were carried out in the absence of GBS. Data are expressed as means ± SEM of seven to nine observations. Ordinate, contractile response (mg); abscissa, log NA concentration (M).

Full size image

Effects of GBS on U46619-Induced Contractions

As shown in Fig. 3A, the maximal contractile responses to U46619 were similar in control and in GBS-treated pulmonary arteries and no change in the pD2 values was observed (Table 1). In the absence of endothelium (Table 1) or in the presence of 10-4 M L-NAME (Table 1, Fig. 4A), GBS was again unable to modify the contractile responses to U46619.

Figure 3
figure 3

Concentration-response curves to U46619 (10-10 to 10-6 M) in pulmonary (A) or mesenteric (B) artery rings preincubated for 20 h in Krebs solution in the presence of vehicle (○) or GBS (3 × 107 cfu mL-1, ▴). The concentration-response curves to U46619 were carried out in the absence of GBS. Data are expressed as means ± SEM of 11-12 observations. Ordinate, contractile response (mg); abscissa, log U46619 concentration (M).*p < 0.05 compared with control group.

Full size image
Figure 4
figure 4

Effects of L-NAME (10-4 M) on the concentration-response curves to U46619 in untreated and GBS-treated pulmonary(A) or mesenteric (B) artery rings. Concentration-response curves to U46619 (10-10 to 10-6 M) were performed in rings previously incubated for 20 h in Krebs solution in the presence of vehicle(control, ○) or GBS (3 × 107 cfu mL-1, ▴). L-NAME (10-4 M) was added 30 min before the addition of U46619. The curves were carried out in the absence of GBS. Data are expressed as means± SEM of 6-18 observations. Ordinate, contractile response (mg); abscissa, log U46619 concentration (M).

Full size image

Fig. 3B shows that, in contrast to pulmonary arteries, GBS-treated mesenteric arteries showed a significant reduction (p< 0.05) of the maximal contractions induced by U46619 compared with controls without affecting the pD2 value. Pretreatment with L-NAME induced an upward shift of the concentration-response to U46619 in the GBS-treated but not in untreated mesenteric arteries (Fig. 4B). Therefore, L-NAME reversed the hyporesponsiveness to U46619 induced by GBS in mesenteric arteries (Table 1).

Effects of GBS on ET-1-Induced Contractions

As shown in Fig. 5 the incubation for 20 h with GBS produced a marked decrease in the contractile response to ET-1 in both pulmonary and mesenteric arteries. Because the maximal concentration of ET-1 used (3 × 10-9 M) did not reach maximal contractile effect, neither the Emax nor the pD2 values could be calculated in these experiments. The GBS-induced reduction of the maximal contraction to ET-1 was more marked in mesenteric compared with pulmonary arteries (40 ± 5% versus 60 ± 4, p < 0.01). Fig. 6 shows that GBS-induced hyporesponsiveness was completely reversed in the presence of L-NAME.

Figure 5
figure 5

Concentration-response curves to ET-1 (10-11 to 3 × 10-9 M) in pulmonary (A) or mesenteric(B) artery rings preincubated for 20 h in Krebs solution in the presence of vehicle (○) or GBS (3 × 107 cfu mL-1,▴). The concentration-response curves to U46619 were carried out in the absence of GBS. Data are expressed as means ± SEM of seven to nine observations. Ordinate, contractile response (mg); abscissa, log ET-1 concentration (M). *p < 0.05 and **p < 0.01 compared with control group.

Full size image
Figure 6
figure 6

Effects of L-NAME (10-4 M) on the concentration-response curves to ET-1 (10-11 to 3 × 10-9 M) in untreated and GBS-treated pulmonary (A) or mesenteric(B) artery rings. Concentration-response curves to ET-1 were performed in rings previously incubated for 20 h in Krebs solution in the presence of vehicle (○) or GBS (3 × 107 cfu mL-1,▴). L-NAME (10-4 M) was added 30 min before the addition of ET-1. The curves were carried out in the absence of GBS. Data are expressed as means± SEM of 7-10 observations. Ordinate, contractile response (mg); abscissa, log ET-1 concentration (M).

Full size image

DISCUSSION

In the present report we have studied the effects of prolonged incubation with heat-killed GBS on the vascular contractile responses induced by NA, the TXA2 mimetic U46619, and ET-1 on isolated pulmonary and systemic(mesenteric) arteries from piglets. The results demonstrate that GBS reduced the vascular responsiveness to NA, U46619, and ET-1 in mesenteric arteries and this effect was reversed by the presence of the NOS inhibitor L-NAME. In pulmonary arteries, GBS also reduced the response to NA and ET-1, and L-NAME reversed this hyporesponsiveness. In contrast, GBS did not affect the contractions induced by U46619 in pulmonary arteries both in presence or absence of L-NAME.

In a recent study we have reported that prolonged (20 h) incubation with GBS or Escherichia coli LPS, reduced the contractile responses to NA in piglet pulmonary arteries(21). This hyporesponsiveness was potentiated by the NO precursor L-arginine and reversed by L-NAME. Dexamethasone, which inhibits the induction of iNOS, and cycloheximide, an inhibitor of protein synthesis, when coincubated with GBS or LPS, completely reversed the reduced response to NA. Moreover, GBS and LPS induced a marked increase in iNOS activity, indicating that the vascular hyporesponsiveness was related to overproduction of NO as a consequence of the induction of iNOS. In the present study, incubation of mesenteric arteries with GBS also decreased the contractile responses induced by NA, U46619, and ET-1. The reversal by L-NAME of these reduced responses suggested that an enhanced NO production may be responsible for GBS-induced vascular hyporesponsiveness. It has been found that LPS leads to the induction of iNOS resulting in an overproduction of NO, which contributes to the severe hypotension seen in Gram-negative sepsis(12–14). More recently, iNOS induction in several tissues has also been demonstrated with some Gram-positive bacteria(17, 18) including GBS(19–21). Moreover, lipoteichoic acid, a component of the peptidoglycan layer of the cell wall in most Gram-positive bacteria, has been described as being responsible for this induction(23, 24).

In spite of the pulmonary hypertension that appears in the sepsis syndrome, induction of iNOS and decreased responses to pressor agents have been reported after in vitro incubation of pulmonary arteries from piglets(21) or adult rats(16, 25) with LPS or GBS. Moreover, Curzen et al.(26) have described hyporesponsiveness to ET-1 in pulmonary arteries from LPS-treated rats. Unfortunately, they did not evaluate the possible role of NO in this hyporesponsiveness. In the present study we have demonstrated that prolonged in vitro incubation with GBS reduces the pulmonary vascular response to NA and ET-1, but not to the TXA2 mimetic U46619. Interestingly, in mesenteric arteries, the GBS-induced hyporesponsiveness to U46619 was less marked than that to NA or ET-1. The reason for this lack of hyporesponsiveness to U46619 in GBS-treated pulmonary arteries is unknown, but a reduced sensitivity to NO-mediated vasodilatation can be involved. In fact, we have previously reported that pulmonary arteries precontracted with U46619 were relatively insensitive to the relaxant action of the NO donor sodium nitroprusside or 8-bromo-cGMP, the stable analog of cGMP(27). In contrast, inhaled NO was able to reduce the pulmonary hypertension produced by U46619 infusions(28) or GBS-sepsis(29, 30). Interestingly, the response to U46619 after iNOS induction varies among species and vascular beds. Thus, isolated piglet pulmonary arteries incubated with endotoxin(27) or isolated mesenteric beds from endotoxin-treated rats were not hyporesponsive to U46619 despite significant iNOS induction(22), whereas in isolated rat lung(31), isolated piglet mesenteric arteries(27), and isolated perfused rat(32), or rabbit hearts(33), the vasoconstrictor responses to U46619 were attenuated by endotoxin. In addition, increased responses to U46619 have been reported after treatment of isolated perfused guinea pig lungs with tumor necrosis factor-α(34) which is one of the most important cytokines released in septic shock or in experimental models of sepsis induced by Gram-negative and Gram-positive bacteria including GBS(20, 35, 36). Thus, a cytokine-mediated increase in pulmonary arterial sensitivity to TXA2 has been also proposed as a mechanism for the persistence of pulmonary hypertension in sepsis(34). Because TXA2 is an important mediator in GBS sepsis-related pulmonary hypertension(6, 7), one could speculate that the lack of GBS-induced hyporesponsiveness to U46619 may explain the pulmonary hypertension despite iNOS induction in the lung. However, the relationship between these events is unclear, because TXA2-mediated pulmonary hypertension appears in the early phase of experimental sepsis(5–7), whereas iNOS induction seems to be a delayed process. In fact, we have observed iNOS induction in piglet pulmonary arteries after incubation for 20 h with GBS but not after 1 or 5 h(21). The uptake of bacteria by pulmonary intravascular macrophages and the subsequent release of inflammatory mediators are central to the pathophysiology of GBS-induced pulmonary hypertension(37). The absence of these cells in our model is another limitation of this study which may explain, at least in part,in vivo, in vitro, and organ- and species-differences. Moreover, due to its cytotoxic effects, NO may play a role in the edema and vascular injury that accompanies the late phases of sepsis-mediated pulmonary hypertension.

To the best of our knowledge this is the first report of NO-mediated vascular hyporesponsiveness induced by GBS in systemic arteries. Moreover, the percentage of reduction in the contractile response induced by GBS was greater in mesenteric compared with pulmonary arteries for any of the stimuli studied. These results may be relevant due to the association of GBS sepsis and hypotension. In fact, in neonates with early-onset GBS sepsis, the development of hypotension is one of the most sensitive predictors of mortality(4, 38), and in newborns with Gram-positive or Gram-negative sepsis an association between plasma nitrite plus nitrate(metabolites of NO) and shock has been reported(39). Whether iNOS induction is beneficial in sepsis-induced shock or is an undesirable collateral effect remains controversial. Experiments using NO inhibitors have strongly implicated NO as a cytotoxic agent which plays a role in antimicrobial and inflammatory responses(40), whereas mortality related to experimental endotoxemia was abolished in mutant mice lacking the iNOS gene(41).

The coexistence of pulmonary hypertension and systemic hypotension in septic shock states important therapeutical problems. Systemic administration of NOS inhibitors to patients with sepsis reversed systemic hypotension but significantly enhanced mean pulmonary arterial pressure and pulmonary vascular resistances(42). Thus, inhaled NO (as a treatment for pulmonary hypertension) in combination with NOS inhibitors (as a treatment for systemic hypotension) has been studied in a porcine model of sepsis and proposed as a new therapeutic regimen for septic shock(43). However, in experimental models of GBS sepsis where systemic hypotension is not observed, probably due to the short duration of the GBS infusion, treatment with NOS inhibitors worsens the hemodynamic situation, increasing both pulmonary and systemic vascular resistances and decreasing cardiac output(44, 45). These data partially contradicts our findings, because L-NAME did not enhance the U46619-induced contractions in GBS-incubated pulmonary arteries. Further studies, including the vascular responses of pulmonary and systemic vessels after prolonged in vivo GBS exposure, would be necessary to elucidate the pathophysiologic and therapeutical implications of sepsis-induced iNOS induction and the possible involvement of the changes in vascular tone which accompanies this process.

In conclusion, prolonged incubation of piglet mesenteric arteries with heat-killed GBS produced a marked hyporesponsiveness to NA, U46619, and ET-1 due to an enhanced NO release, suggesting a role for iNOS induction in the systemic hypotension associated with GBS sepsis. GBS-treated pulmonary arteries exhibited a NO-mediated hyporesponsiveness to NA and ET-1 but not to the TXA2 mimetic U46619. This absence of pulmonary hyporesponsiveness to U46619 may contribute to the persistence of pulmonary hypertension in GBS sepsis despite iNOS induction in the lung.