Main

Endothelium-derived NO is a potent vasodilator which, in the mature circulation, is a major determinant of intestinal blood flow, contributing to the characteristic low vascular resistance and high level of perfusion of this vascular bed(1). New information suggests that NO also plays a vasodilating role in the immature circulation, perhaps more powerful than that in the adult. NO synthesis is 9-fold greater in the fetus than the mother in sheep pregnancy(2), and NO plays a dominant role in maintaining the low vascular resistance and high perfusion of the umbilical-placental circuit(3,4). NO is also critical for maintaining the caliber of major arteries of the immature rat fetus(5) and thus is likely to be a powerful factor in determining blood flow in vascular beds other than the placenta. In the developing intestinal circulation of the lategestation sheep fetus, recent work has shown that NO plays an important role in contributing to low vascular resistance and high blood flow throughout the GI tract, as well as to the regional distribution of intestinal blood flow between intestinal segments, and across mucosal and muscular layers(6).

As yet no study has addressed the role of NO in the intestine at earlier gestations, though several indirect lines of evidence suggest that NO is likely to be important in the regulation of GI perfusion at mid-gestation and, indeed, that the role of NO at this time may be even more important than later in gestation. Gestational changes in umbilical-placental vessels suggest that the amount of NO released and its relaxant effect in vascular beds of the fetus may be more pronounced earlier in development(7), a possibility that is emphasized by the striking changes in major vessel caliber induced by altering NO synthesis in the very immature circulation of the rat fetus(5). The possibility that the regulatory role of NO in the GI circulation may be enhanced earlier in ontogeny is further supported by the greater NO-mediated intestinal vasodilation in newborn compared with older piglets(8,9). Supporting evidence for the concept that NO may be the dominant vasodilator early in development includes the observation that another important vasodilator, the neuropeptide CGRP, is poorly developed at mid-gestation and does not achieve maturity until the early postnatal period(10,11). Finally, NO appears to play a prominent role in maintaining flow at low perfusion pressures(12), an action which is likely to be of particular importance at the low arterial pressures present at mid-gestation(13).

Resolution of the role of NO in the maintenance of perfusion in the immature GI circulation has implications for understanding the pathogenesis of NEC. NEC has its highest incidence in the immature GI circulation(14) and compromised blood flow and oxygen transport to the intestine are prominent clinical findings in infants suffering NEC(15,16). Endogenous NO may be of critical importance for maintaining high levels of intestinal blood flow, particularly in the small intestine, which is most prone to NEC(6,17) and for preserving intestinal perfusion during hypoxia-ischemic insults that have been postulated to precipitate NEC(9).

The purpose of the present study was to test the hypothesis that NO plays a major role in the regulation of the intestinal circulation at mid-gestation(more powerful than later in development) that is critical to ensuring adequate blood flow and oxygenation of the immature GI tract. The role of NO in regulating GI hemodynamics at mid-gestation was assessed in studies of chronically instrumented fetal lambs before and after inhibition of NO synthase with the stereospecific inhibitor L-NNA. To test the hypothesis that NO has a greater regulatory role earlier in development, results from the mid-gestation fetus were compared with those obtained from late-gestation fetuses(6).

METHODS

All procedures and protocols used in these experiments were in accord with guidelines provided by the National Health and Medical Research Council of Australia and were approved by the Monash University Committee on Ethics in Animal Experimentation.

Surgical preparation. Eight Border-Leicester fetal lambs underwent surgery between 89 and 97 d of gestation (term = 147 d). After 24 h of fasting, general anesthesia was induced in pregnant ewes with 200 mg of i.v. propofol (Diprivan, ICI, Australia). The ewes were intubated and general anesthesia was maintained by positive pressure ventilation using 1-3% halothane, 30% oxygen, and the remainder nitrous oxide.

During surgery, maternal blood gases were monitored at regular intervals via a femoral arterial catheter. Under sterile conditions the fetal lamb's head, left forelimb, and upper thorax were delivered through a uterine incision. A left lateral thoracotomy was performed in the fourth interspace, and the fourth rib was removed to increase exposure of the heart and the pulmonary trunk. After incision of the pericardium, a polyvinyl catheter inserted into the left atrium and a 6- or 8-mm diameter ultrasonic flow transducer (Transonic Systems Inc., Ithaca, NY) was implanted around the main pulmonary trunk. A modified 18-gauge Teflon cannula connected to a polyvinyl catheter was inserted into the pulmonary trunk distal to the flow probe, and the pericardial incision over the left atrium was then closed. Nonocclusive polyvinyl catheters were advanced through the left axillary artery and vein into the brachiocephalic trunk and the superior vena cava, respectively.

The chest incision was closed and a wide bore catheter was sutured to the anterior chest wall for measurement of amniotic fluid pressure. Through a second uterine incision, both fetal groins were exposed, and catheters were inserted into a femoral artery bilaterally and a femoral vein, and advanced into the abdominal aorta and inferior vena cava, respectively.

The fetus was returned to the uterus, amniotic fluid losses were replaced with warm isotonic saline, antibiotics administered into the amniotic cavity(1 g of procaine penicillin, 1 g of dihydrostreptomycin sulfate), and the uterine incisions then closed. The catheters and flow transducer cable were exteriorized through a s.c. tunnel and protected in a pouch located above the ewe's right flank. Each day thereafter antibiotics were administered to the amniotic cavity, catheters were flushed, and the dead space filled with heparinized saline (1000 U/mL).

Experimental protocol. After 72-h recovery from surgery, the ewe was brought to the laboratory and placed in a study cage with free access to food and water. Pulmonary trunk flow was measured continuously by an ultrasonic volume flowmeter (model T208, Transonic Systems). Abdominal aortic blood pressure and amniotic fluid pressure were measured by pressure transducers (model CDX-111, COBE, Lakewood, CO). Hemodynamic variables were displayed on a Neotrace 800ZF chart recorder (Neomedix Systems, Sydney, Australia), recorded on computer at a sampling rate of 200 Hz with an analog to digital converter, and analyzed off-line with interactive software.

After recording of baseline hemodynamics and blood gas variables, ventricular outputs and regional blood flows were measured with radioactive microspheres. NOS was then inhibited with the specific NOS inhibitor L-NNA(Sigma Chemical Co., St. Louis, MO). L-NNA was dissolved in saline to a concentration of 5 mg/mL and infused into the inferior vena caval catheter at a rate of 0.68 mL/min to achieve cumulative doses of approximately 10 and 25 mg/kg of estimated fetal body weight. After a period of stabilization, hemodynamic, blood gas, and blood flow measurements were repeated at each L-NNA dose. On completing the experiment, the ewe and fetus were killed with an overdose of pentobarbitone sodium (150 mg/kg), and an autopsy was performed on the fetus to verify catheter placement.

Blood flow and resistance measurements. Blood flows were measured with radioactive microspheres (141Ce, 113Sn,85 Sr, 95Nb, and 46Sc, New England Nuclear, Boston, MA) using the reference sample technique(18). After ultrasonication, approximately one million 15-µm diameter microspheres were injected per microsphere label. At baseline and at an L-NNA dose of 25 mg/kg, two different microsphere labels were injected simultaneously, one into the left atrium to measure LV output and the other into the superior vena cava to measure RV output(19), whereas reference samples were drawn simultaneously from the pulmonary trunk, axillary artery, and descending aorta. At an L-NNA dose of 10 mg/kg, LV output was obtained with a single microsphere label injected into the left atrium, whereas reference samples were withdrawn from the axillary artery and descending aorta. Reference samples were withdrawn at a rate of 2.06 ml·min-1 using a mechanical pump (Harvard Apparatus, South Natick, MA). Reference sample withdrawal was commenced 10 s before microsphere injection and terminated 1.5 min after its completion. Fetal blood withdrawn during reference sampling was simultaneously replaced with maternal blood.

At the end of the experiment, the GI tract was totally removed, fixed in 10% formalin for a period of 1 wk, and then dissected into its major anatomical segments, namely the stomach, duodenum, jejunum, ileum, cecum, colon, and rectum. Each GI segment was blotted dry, weighed, and packed into plastic counting vials to a height of ≤ 2 cm. Radioactivity in tissue and reference blood samples was counted using a well type scintillation counter(model 1282 Compugamma, LKB-Wallac, Turku, Finland), with windows set at the appropriate energy levels.

Blood flows were calculated from the general relation Qtissue = [Qreference ×Rtissue]/Rreference, where Q is flow (mL/min) and R is radioactivity (counts/min). Ventricular outputs were calculated using an adaptation of this general relation for the fetal circulation as previously described(19). Thus, fetal LV output (QLV) was calculated as [Qreferennce ×RLA]/RLA≫CA, where RLA is the radioactivity of the label injected into the left atrial cavity and RLA≫CA is the radioactivity of the same label in the carotid arterial reference sample. Fetal RV output(QRV) was equal to [Qreference ×RRV]/RV≫PT, where RRV is the radioactivity of venous label passing into the right ventricle, calculated as the injected radioactivity of this label minus that portion crossing the foramen ovale to appear in the LV output, and RV≫PT is the radioactivity of the venous label in the pulmonary reference sample(19). RV output at an L-NNA dose of 10 mg/kg was obtained by interpolation from the measured pulmonary trunk flow, using baseline and 25 mg/kg L-NNA measurements of pulmonary trunk obtained with the ultrasonic flow probe and RV output obtained with radioactive microspheres. Global and regional GI vascular resistance(RGI, mm Hg·mL-1·min-1·100 g-1) was calculated as MAP/QGI, where MAP is mean abdominal aortic blood pressure and QGI is the appropriate GI blood flow.

Blood gas measurements. Blood gas tensions and pH corrected to 40°C (Radiometer ABL 500 blood gas analyzer, Radiometer, Copenhagen, Denmark), as well as Hb concentration ([Hb]) and Hb oxygen saturation(Sao2) (Radiometer OSM-2 hemoximeter) were measured in abdominal aortic blood. Abdominal aortic oxygen content (CAo2, mL/dL) was calculated as (0.0136 × [Hb] × SaO2) + (0.003 × PO2) and GI oxygen delivery (DO2, mL O2·min-1·100 g-1) was calculated as QGI × CAO2.

Statistics. All data are presented as mean ± SE. Data were compared with one-way and two-way repeated measures ANOVA and a post hoc Newmann-Keuls test as appropriate. All analyses were performed using the computer statistical package Statistica (StatSoft), and p < 0.05 was considered significant.

RESULTS

Systemic hemodynamics and blood gas changes. Baseline hemodynamics (Table 1) and blood gases(Table 2) in the mid-gestation fetal lambs were comparable to those reported in previous studies(13,20,21). After administration of 10 mg/kg L-NNA, arterial pressure rose by 13 ± 3 mm Hg (p < 0.01), the combined LV and RV output (CVO) fell by 145 ± 25 mL·min-1·kg-1 (p < 0.001), whereas heart rate was unchanged. At 25 mg/kg L-NNA, the CVO fell by a further 105 ± 39 mL·min-1·kg-1(p < 0.01), but MAP and heart rate were not significantly different from 10 mg/kg L-NNA (Table 1). Apart from an increase in Hb concentration (p < 0.001), blood gas variables were unchanged between baseline and 10 mg/kg L-NNA. However, the dose of 25 mg/kg L-NNA was associated with decreases in pH (p < 0.05) and SaO2 (p < 0.05), a rise in PCO2 (p< 0.05) and a further increase in Hb concentration (p < 0.001), but no change in arterial oxygen content (Table 2).

Table 1 Systemic hemodynamic responses to L-NNA
Full size table
Table 2 Arterial blood gas values before and after L-NNA
Full size table

GI blood flow responses. Under baseline conditions, average GI blood flow was 93 ± 7 mL·min-1·100 g-1. Moreover, a substantive differential pattern of regional blood flow was evident along the GI tract with stomach and small intestinal and large intestinal blood flows being significantly different from one another(p < 0.001) and the highest flow occurring in the small intestine (Table 3).

Table 3 GI blood flow (mL·min-1·100 g-1) responses to L-NNA
Full size table

After the infusion of 10 mg/kg L-NNA, average GI blood flow decreased to 41 ± 8 mL·min-1·100 g-1 (p< 0.001). Significant flow reductions occurred in all regions of the GI tract (p < 0.01 to 0.001), and blood flow to the stomach was still lower than to either the small or large intestine (Table 3). However, the extent of the flow reduction in the small intestine (59± 12 mL·min-1·100 g-1) was more pronounced than in the large intestine (27 ± 7 mL·min-1·100 g-1, p < 0.01) and, as a result, small and large intestinal flows were not significantly different from one another at 10 mg/kg L-NNA. Neither average GI blood flow nor regional blood flows to the stomach, small intestine or large intestine changed significantly between 10 and 25 mg/kg L-NNA (Table 3).

As illustrated in Fig. 1, similar degrees of blood flow reduction occurred at 10 mg/kg L-NNA in the duodenal (46 ± 16 mL·min-1·100 g-1), jejunal (61 ± 17 mL·min-1·100 g-1), and ileal (58 ± 10 mL·min-1·100 g-1) portions of the small intestine. Likewise, blood flow decrements in the cecum (28 ± 8 mL·min-1·100 g-1), colon (28 ± 9 mL·min-1·100 g-1), and rectum (20 ± 8 mL·min-1·100 g-1) did not differ significantly from one another. No further change in blood flow to regions of the small and large intestine occurred at 25 mg/kg L-NNA, apart from a further reduction in blood flow to the rectum (p< 0.05;Fig 1).

Figure 1
figure 1

Regional intestinal blood flow under baseline conditions (open bars) and after L-NNA infusion at 10 mg/kg i.v. (solid bars) and 25 mg/kg i.v. (hatched bars). Values are means ± SE; n = 8 fetal lambs.**p < 0.001; *p < 0.05 vs baseline.†p < 0.05 vs L-NNA (10 mg/kg).

Full size image

GI vascular resistance. Under baseline conditions, average GI vascular resistance was 0.41 ± 0.06 mm Hg·mL-1·min-1·100 g-1. In similar manner to blood flow, a substantial differential pattern of regional vascular resistance was evident along the GI tract, with small and large intestinal vascular resistances being significantly different from the stomach (p < 0.001), but not from one another(Fig. 2).

Figure 2
figure 2

Vascular resistance (R) of stomach, small intestine, and large intestine (A) and regional vascular resistance (B) under baseline conditions (open bars) and after L-NNA infusion at 10 mg/kg i.v. (solid bars) and 25 mg/kg i.v. (hatched bars). Values are means ± SE; n = 8 fetal lambs. *p < 0.01 vs baseline; †p < 0.05 vs L-NNA (10 mg/kg).

Full size image

In conjunction with the blood flow changes, vascular resistance in all major regions of the GI tract increased significantly after 10 mg/kg L-NNA(p < 0.01), with increments of 181 ± 28% in the stomach, 194 ± 37% in the small intestine and 120 ± 25% in the large intestine (Fig. 2A). No further change in vascular resistance occurred at 25 mg/kg L-NNA.

Regional vascular resistance increased dramatically in all segments of the small intestine at 10 mg/kg L-NNA, with increments ranging from 174 ± 44% for the duodenum to 199 ± 44% for the jejunum (Fig. 2B). Increases in regional vascular resistance were less marked for the large intestine, ranging from 92 ± 24% for the rectum to 166 ± 49% for the colon. No significant change in regional vascular resistances occurred between 10 and 25 mg/kg L-NNA, apart from a further small rise within the rectum (p < 0.05;Fig. 2B).

GI oxygen delivery. Under baseline conditions, average GI oxygen delivery was 5.0 ± 0.4 mL O2·min-1·100 g-1. A substantial differential pattern of regional oxygen delivery was evident along the GI tract with stomach and small and large intestinal oxygen delivery being significantly different from one another (p < 0.001) and the highest oxygen delivery occurring in the small intestine (Fig. 3A). After 10 mg/kg L-NNA, oxygen delivery decreased by 44 ± 3% to the stomach (p < 0.001), 44 ± 4% to the small intestine (p< 0.001), and 26 ± 10% to the large intestine (p < 0.05). Oxygen delivery to the small and large intestine did not change significantly between 10 and 25 mg/kg L-NNA, but fell further by 22% in the stomach(p < 0.05; Fig. 3A).

Figure 3
figure 3

Oxygen delivery of stomach, small intestine, and large intestine (A) and regional oxygen delivery(B) under baseline conditions (open bars) and after L-NNA infusion at 10 mg/kg i.v. (solid bars) and 25 mg/kg i.v.(hatched bars). Values are means ± SE; n = 8 fetal lambs. Note that oxygen delivery to the small intestine significantly exceeded that to the large intestine in the baseline conditions, but not after L-NNA infusion. **p < 0.001; *p < 0.05vs baseline; †p < 0.05 vs L-NNA (10 mg/kg).

Full size image

Regional oxygen delivery decreased significantly in all segments of the small intestine after 10 mg/kg L-NNA, amounting to reductions of 37 ± 6% in the duodenum (p < 0.001), 43 ± 5% in the jejunum by (p < 0.001) and 45 ± 3% in the ileum (p< 0.001). Regional oxygen delivery also fell in the cecum (33 ± 8%; p < 0.01) and colon (41 ± 4%, p < 0.05), but did not change significantly in the rectum (Fig. 3B). No further change in regional oxygen delivery to the small and large intestine occurred between 10 and 25 mg/kg L-NNA, apart from the rectum which was 38 ± 12% lower than baseline (p < 0.05).

DISCUSSION

NO has been shown to exert a powerful vasodilatory influence in GI circulation of mature animals(22–24). However, it was not until recently that endogenous NO was shown to be a significant contributor to the modulation of vascular tone in the GI circulation of late-gestation fetal lambs(6). In this study, we have extended our investigation to the mid-gestation fetus, and shown that NOS substantially affects GI vascular tone at this age. Moreover, our data indicate that a characteristic in utero pattern of basal intestinal blood flow favoring the small intestine is dramatically altered by the administration of L-NNA. As we elaborate below, our results suggest that NO plays a substantially greater role in the regulation of intestinal vascular tone in the mid-gestation fetus than it does later in gestation.

The baseline values we obtained for MAP, blood gas values, and CVO for the mid-gestation fetal sheep are consistent with values published in other studies(13,20). In accord with previously published gestational data(21), average blood flow was less at mid-gestation (93 mL·min-1·100 g-1) than at late-gestation (136 mL·min-1·100 g-1)(6). Values of blood flow in different regions of the GI tract ranged between 39 and 134 mL·min-1·100 g-1, levels that are comparable to those previously reported for fetal lambs at this stage of gestation(13,21). In a manner similar to the late-gestation fetal sheep(6), the mid-gestation fetus also displayed a differential pattern of blood flow between the small and large intestine, with much higher blood flows directed to the small intestine (Table 3).

An important feature of our experimental design was to evaluate the response to incremental doses of L-NNA, as it is known that specific effects of NOS inhibition in the mature circulation are dose related(24). In a manner similar to the late-gestation fetus, some of the systemic variables in the mid-gestation fetus, specifically arterial blood gases, pH, and Hb showed dose-related changes. Although changes in SaO2 and PCO2 which occurred after L-NNA were statistically significant, they were small in magnitude, so that values remained within normal ranges reported in other studies(13,20). pH at 10 mg/kg L-NNA was within the normal range but a mild acidosis was evident at 25 mg/kg L-NNA. This was most likely a consequence of significantly reduced placental blood flow causing reduced oxygen transport to the fetus, or of increased systemic vascular resistance leading to inadequate oxygen delivery. Hb rose progressively with increasing L-NNA concentration, probably reflecting increases in vascular permeability and promotion of plasma leakage that accompanies NOS inhibition(25,26).

Although basal arterial blood pressure at mid-gestation was lower than at late-gestation(6), the increase in blood pressure at both the 10 and 25 mg/kg doses of L-NNA were similar to those observed in late-gestation fetuses(6). That no accompanying bradycardia occurred after NOS inhibition in mid-gestation animals is probably related to the immaturity of vagal effects upon heart rate at this early stage of fetal development(27). One explanation for the progressive decrease in CVO after L-NNA is that the substantial increase in blood pressure associated with NOS inhibition may have exceeded the compensatory capacity of the immature fetal heart to maintain output in the face of the increased afterload. An alternative factor which may have contributed to the decreased CVO was a reduction in circulating blood volume accompanying the hemoconcentration observed after NOS inhibition(6,26).

The major finding of our study was that NOS inhibition in the mid-gestation fetus was associated with a substantial decrease in GI blood flow. As in the late-gestation fetal lambs(6), responses to NOS inhibition achieved steady state in the mid-gestational GI circulation between the 10 and 25 mg/kg doses of L-NNA (Table 3). In the regional GI vascular beds, the substantial blood flow reduction evoked by L-NNA largely abolished the basal differential pattern of higher blood flow in the small intestine compared with that in the large intestine. Particularly remarkable was the blood flow reduction evident in segments of the small intestine, where an average blood flow fell by 73 mL·min-1·100 g-1 overall. This flow reduction was twice that of the large intestine overall(Table 3), and in specific segments was much greater in magnitude than across the segments of the large intestine(Fig. 1). By contrast, in the late-gestation fetus, the differential pattern of blood flow between the small and large intestine was attenuated, but not entirely abolished by NOS inhibition(6).

We ascribe the powerful vasoconstriction in the immature GI circulation directly to the loss of the vasodilating effect of NO. No baroreflex-induced vasoconstriction would be expected under the conditions of our study as MAP rose, and as this would promote vasodilation rather than vasoconstriction. In any case, reflex responses would be expected to produce less vasoconstriction in the immature circulation(5,28), not greater vasoconstriction as we observed. It is also unlikely that the GI blood flow changes accompanying NOS inhibition in this study were related to the induction of a migratory motor complex(29), a pattern of electrical and contractile activity that is associated with cyclical increases in intestinal blood flow in the mature animal(30), as the migratory motor complex is absent or quiescent at mid-gestation(31).

Comparison of blood flow decreases with previously published data from late-gestation fetuses(6) indicates that percentage decreases in blood flow and increase in vascular resistance were significantly greater at mid-gestation (Fig. 4). These observations suggest that endogenous NO exerts a much greater influence on the regulation of GI vascular tone at this early stage of gestation than it does later in fetal life. That NO should have a prominent role in the mid-gestation GI circulation of the lamb is supported by recent evidence that NOS is critical in maintaining the caliber of the ductus arteriosus and great vessels of the immature fetal rat(5). Interestingly, our findings complement those in the postnatal intestinal circulation(8,32), where the role of NO in promoting vasodilation is enhanced at the younger age of 3 d compared with 35 d.

Figure 4
figure 4

Illustrations of the substantially greater effects of L-NNA infusion upon the GI circulation of the mid-gestation fetal lamb compared with the late-gestation fetal lamb. (A) Decrease in blood flow and (B) increase in vascular resistance after L-NNA infusion (25 mg/kg i.v.) in stomach, small intestine and large intestine of mid-gestation (open bars, n = 8) and late-gestation (solid bars, n = 11) fetal lambs. Values are means ± SE; **p < 0.001; *p < 0.05vs mid-gestation.

Full size image

A possible explanation for the dominant vasodilator role of NO in the mid-gestation GI circulation may lie in a different pattern of interaction with other vasoactive influences(33–36) determining the balance of vascular tone in the immature circulation. Specifically, the vasodilatory effects of NO may operate in the presence of limited opposing vasoconstricting effects of the sympathetic nervous system which is poorly developed at this age(20,27,28,37). Importantly, when NO production is inhibited in the more mature intestinal circulation, there is a compensatory vasodilation effected by release of the neuropeptide CGRP(11,38–41). Although present by mid-gestation(10) CGRP does not attain peak activity until the early neonatal period(11,39). This suggests that NO may be the dominant vasodilator at mid-gestation, with CGRP assuming a greater role as the neonatal period is approached. Whether compensatory vasodilation due to other factors is also less pronounced in the GI circulation at mid-gestation remains to be determined; prostaglandins, which support NO-induced vasodilation in mature mesenteric vessels(42), may also contribute to vasodilation in the immature circulation, though perhaps to a lesser extent than NO(4).

In the present study, resting oxygen delivery to the GI circulation (5.0± 0.4 mL O2·min-1·100 g-1) was approximately half that of the near term fetus (8.6 ± 1.0 mL O2·min-1·100 g-1), presumably reflecting lesser requirements for growth and metabolic demand at mid-gestation. After NOS inhibition in the mid-gestation fetus, oxygen delivery was substantially decreased, particularly in the small intestine (Fig. 3), where the basal pattern of higher oxygen delivery compared with the large intestine was completely abolished. Compared with mid-gestation, changes in the regional pattern of GI O2 delivery in the late-gestation were less pronounced and confined to the small intestine, with O2 delivery to some segments of the large intestine being unaffected by L-NNA infusion(6). The late-gestation fetal GI circulation is more sensitive to hypoxia-ischemia than the neonate(43). In this respect, the mid-gestation GI circulation may be even more susceptible than the late-gestation GI circulation to hypoxia-ischemic insult because of the greater role of NO-mediated vasodilation in maintaining basal flow, and the sensitivity of NO production to hypoxia-ischemic insult(9).

The substantial dependence upon NO production of the immature ruminant GI circulation has important implications for the pathogenesis of NEC, which has a high incidence in premature infants. Although the human GI tract differs anatomically and physiologically from that of the ruminant, both species have high blood flow levels in the small intestine(44,45) that may be dependent upon NO production. As our study shows, endogenous NO is in large measure responsible for maintaining high levels of blood flow in the immature GI tract, particularly in the small intestine, a site at which NEC occurs at high frequency(17). Endothelium dysfunction resulting from damage by factors such as hypoxia-ischemia, may act as trigger events leading first to disruption of the key vasodilating role of NO and eventually to significant intestinal necrosis(9).

In summary, the inhibition of endogenous NO in the midgestation fetal sheep, results in substantial blood flow reductions across all segments of the GI tract. These effects are substantially greater at mid-gestation than later in fetal development and after birth. We speculate that if the synthesis and activity of NO were to be impaired, the immature GI circulation would be particularly vulnerable to ischemic insult.