Seminars in Pediatric Surgery (2005) 14, 152-158

Ischemia and necrotizing enterocolitis: where, when, and how

Philip T. Nowicki, MD
From the Center for Cell and Vascular Biology, Columbus Childrens Research Institute; and the Division of Neonatology, Department of Pediatrics, College of Medicine and Public Health, The Ohio State University, Columbus, Ohio. INDEX WORDS
Endothelium; Endothelin; Nitric oxide; Intestine; Microcirculation While it is accepted that ischemia contributes to the pathogenesis of necrotizing enterocolitis (NEC), three important questions regarding this role subsist. First, where within the intestinal circulation does the vascular pathophysiology occur? It is most likely that this event begins within the intramural microcirculation, particularly the small arteries that pierce the gut wall and the submucosal arteriolar plexus insofar as these represent the principal sites of resistance regulation in the gut. Mucosal damage might also disrupt the integrity or function of downstream villous arterioles leading to damage thereto; thereafter, noxious stimuli might ascend into the submucosal vessels via downstream venules and lymphatics. Second, when during the course of pathogenesis does ischemia occur? Ischemia is unlikely to the sole initiating factor of NEC; instead, it is more likely that ischemia is triggered by other events, such as inammation at the mucosal surface. In this context, it is likely that ischemia plays a secondary, albeit critical role in disease extension. Third, how does the ischemia occur? Regulation of vascular resistance within newborn intestine is principally determined by a balance between the endothelial production of the vasoconstrictor peptide endothelin-1 (ET-1) and endothelial production of the vasodilator free radical nitric oxide (NO). Under normal conditions, the balance heavily favors NO-induced vasodilation, leading to a low resting resistance and high rate of ow. However, factors that disrupt endothelial cell function, eg, ischemia-reperfusion, sustained low-ow perfusion, or proinammatory mediators, alter the ET-1:NO balance in favor of constriction. The unique ET-1NO interaction thereafter might facilitate rapid extension of this constriction, generating a viscous cascade wherein ischemia rapidly extends into larger portions of the intestine. 2005 Elsevier Inc. All rights reserved.

Traditionally, three factors are believed to conspire in the pathogenesis of NEC: infection, feeding, and ischemia.1 The putative role of ischemia is principally based on the histopathology of the disease; thus, intestine resected for NEC regularly demonstrates coagulation necrosis, the footprint of a preceding ischemic insult.2 While these observations cement a role for ischemia in disease pathogenesis, it remains to address three critical questions: (1)
Address reprint requests and correspondence: Philip T. Nowicki, MD, Columbus Childrens Research Institute, 700 Childrens Drive, Columbus, OH 43205. E-mail: NowickiP@pediatrics.ohio-state.edu. 1055-8586/$ -see front matter 2005 Elsevier Inc. All rights reserved. doi:10.1053/j.sempedsurg.2005.05.003

Where within the gut circulation does the pathophysiology causative of the ischemia occur? (2) When does ischemia occur during the pathogenesis of NEC? (3) How does the ischemia occur, ie, what is the mechanistic basis thereof?

Where does the ischemia occur?

Intestinal vascular anatomy
Arterial anatomy of the intestine can be divided into three segments based on function (Figure 1).3 The rst segment

Nowicki

Ischemia and NEC

153

Figure 1 Schematic representation of the intestinal microcirculation. Small mesenteric arteries pierce the muscularis layers and terminate in the submucosa where they give rise to 1A (1st order) arterioles. 2A (2nd order) arterioles arise from the 1A. Although not shown here, these 2A arterioles connect merge with several 1A arterioles, thus generating an arteriolar plexus, or manifold that serves to pressurize the terminal downstream microvasculature. 3A (3rd order) arterioles arise from the 2A and proceed to the mucosa, giving off a 4A branch just before descent into the mucosa. This 4A vessel travels to the muscularis layers. Each 3A vessel becomes the single arteriole perfusing each villus.

begins with the two mesenteric artery trunks, the superior and inferior mesenteric arteries, that give rise to progressively smaller mesenteric arteries in a consecutive, tree-like branching pattern. This segment ends at the terminal branches of the mesenteric arterial arcade and functions as a conduit for delivery of blood to the gut wall. The second segment begins with the terminal branches of the mesenteric arterial arcade, very small arteries (200 m) that pierce the two muscularis layers and descend directly into the submucosa. There, each artery gives rise to several 1A arterioles, which in turn generate 2A arterioles. These vessels remain entirely within the submucosa; as well, 2A arterioles interconnect with multiple 1A vessels forming arteriolararteriolar shunts. The second segment functions as a bridge between the macro- and microcirculations; as well, the 1A2A plexus serves as a manifold to pressurize the downstream microcirculation. The third segment begins with 3A arterioles which arise from the 1A2A plexus and descend toward the mucosa. Each 3A arteriole provides perfusion of a single villus; however, just before its entry into the mucosa, each 3A gives rise to a 4A arteriole that turns back into the muscularis. In this context, the microcirculations of the muscularis and mucosal layers of the gut wall run in parallel from the site of the 3A 4A branch point. The third segment gives rise to the capillaries and thus

functions as the terminal portion of the intestinal arterial circulation, ie, the site responsible for water and solute exchange. Two salient features of venular anatomy bear mention. First, collecting venules from each villus drain into what is essentially a venous sinus within the mucosa; thus, venous drainage from individual villi is admixed with that of surrounding villi, permitting ascent of absorbed substances in a relatively generalized manner. Second, venules do not generally run in immediate proximity to their arterial counterparts within the gut wall, until the level of the submucosa. Therein, 1V venules run in tight proximity to their 1A counterparts, being 30 m apart. This circumstance facilitates venular-to-arteriolar diffusion. The major sites of resistance, and hence ow regulation, are the 1A and 2A arterioles within the submucosa and the small arteries immediately upstream there from, ie, the second segment of the intestinal circulation.4,5 This physiology is important as it suggests that noxious luminal stimuli, eg, infectious agents and their toxins, are not in direct contact with sites of resistance regulation; indeed, the length between the villus base and submucosa is 500 m, a distance that functionally precludes passive interstitial diffusion in the absence of extensive mucosal damage. Therefore, if luminal factors (eg, inammatory mediators) are to

154 alter perfusion, then some mechanism must exist to permit their ascent from the mucosa to submucosa.

Seminars in Pediatric Surgery, Vol 14, No 3, August 2005 villous integrity, an early histopathological marker of NEC.2 It is also important to note that physical disruption or vasoconstriction of the third segment of the gut circulation beginning with the 3A arterioles has the potential of altering upstream gut hemodynamics. Mucosal damage initially generated by a nonvascular mechanism (eg, inammation) could exert a direct effect on the physical integrity or contractile state of the villous 3A arteriole and reduce ow through the arteriole. Loss of villous blood ow, and hence O2 delivery, will further compromise villus integrity and lead to localized villous sloughing. However, if the extent of this mucosal damage is substantial, then ow through multiple (ie, 106) 3A arterioles would become compromised. This circumstance has the potential to attenuate the ow rate in the upstream 1A2A plexus within the submucosa, in the same way that a clogged drain lls a sink. Although run-off through the plexus would occur, a generalized compromise of ow rate within a circumscribed area of the submucosal plexus could occur. As will be discussed below, the mechanostimulus of reduced ow rate can act as a trigger to alter production of endothelium-derived vasoconstrictor and vasodilator agents, leading to progressive vasoconstriction and further ischemia. This event would have the effect of spreading vascular malfeasance within the submucosal plexus, an act that could compromise downstream (ie, 3A) perfusion in immediately contiguous areas. An important clinical hallmark of NEC is its capacity for exceptionally rapid spread; vascular pathophysiology ascendant into the submucosa, followed by spread via the plexus-like nature of these vessels could participate in this rapid spread.

Anatomic considerations in NEC pathogenesis

The importance of the terminal mesenteric arteries and the 1A2A submucosal plexus in regulating intestinal vascular resistance suggests that this site would be pivotal in generating, in whole or in part, the ischemia relevant to NEC. Unfortunately, this presumption, while based on a sound physiological footing, has not yet achieved experimental substantiation. Animal studies of the newborn intestinal circulation have limited their focus on total intestinal blood ow; thus, specic information regarding the intestinal microcirculation is limited to a single study of terminal mesenteric arteries in piglets.6 This study conrmed the importance of this portion of the gut arterial vasculature in resistance regulation and demonstrated that reactivity of newborn arteries was signicantly less efcient than that of their weanling counterparts. Although specic information regarding the putative site of vasculopathy in NEC is yet forthcoming, two possibilities bear discussion based on the evolving concept that the initial trigger of NEC pathogenesis is mucosal inammation. First, it is entirely possible that inammatory mediators ascend into the submucosal 1A2A arteriolar plexus via venules and lympthatics, particularly the latter. Venular arteriolar and lymphaticarteriolar proximity exists within the submucosa.7 This anatomy facilitates venule-to-arteriole and lymphatic-to-arteriole diffusion of vasoactive agents. The diffusive process plays a pivotal role in generating the requisite vasodilation and increased gut perfusion during nutrient absorption; thus, nitric oxide generated within venules8 and lymphatics9 draining the mucosa induces vasodilation of 1A and 2A arterioles, and the immediately upstream terminal mesenteric arteries (ie, the second portion of the gut circulation) that regulates resistance, generating postprandial hyperemia. It is entirely feasible that the same ascending pathway could be taken by other, perhaps unwanted travelers, such as proinammatory mediators generated within the downstream villi and mucosa. These agents could include platelet activating factor or cytokines. Both have been identied in human infants with NEC and are capable of inducing signicant vascular dysfunction, including the direct and indirect generation of vasoconstriction and hence ischemia.10,11 Ascent thereof would permit their rapid diffusion within the critical 1A2A submucosal plexus, leading to a widespread dissemination of inammation-induced vasculopathy. Second, it is possible that the 3A arteriole present within each villus becomes exposed to noxious agents in the immediately adjacent gut lumen; indeed, the 3A arteriole is separated from the luminal gut environment by a single line of epithelial cells and 10 m of interstitial space. Thus, elements of an inammatory response generated by toxins or sclerotic agents have ready access to these arterioles. Induction of 3A vasoconstriction would accelerate loss of

When does ischemia occur?

Clearly, ischemia can cause intestinal damage; alternatively, however, it is possible that intestinal damage, caused by factors other than ischemia, reduces the need for perfusion (or eliminates the vascular space requisite for perfusion to occur), and so reduces intestinal blood ow. It is most likely that neither extreme is correct: ischemia is certainly not the sole basis for NEC-related tissue damage, while it is very likely that ischemia occurs at some time before complete tissue destruction. It is more realistic to query the temporal positioning of ischemia in the cascade of events that leads to NEC. In this context, if ischemia were the instigating perturbation leading to NEC, then it should occur early in disease pathogenesis; alternatively, a more secondary role for ischemia might place it in the midst of the cascade of events leading to NEC. Unfortunately, existing data from human and animal studies fail to provide signicant insight into this issue. Very early descriptions of NEC suggested a primary or early role for ischemia in disease pathogenesis. Thus, Touloukian and coworkers12 described 25 infants with NEC, 18 of which were 2500 g who had sustained evidence of signicant birth asphyxia. Based on the physiology of div-

Nowicki

Ischemia and NEC

155 however, articial feeding of the pups is requisite for the development of NEC, suggesting that ischemia per se is not sufcient to generate tissue damage in this model.26 Moreover, the technical challenge of quantifying gut perfusion in rat pups has thus far precluded direct measurement of intestinal perfusion therein. A piglet model of NEC recently developed in our laboratory utilizes direct reduction of intestinal perfusion via cerclage of the mesenteric artery, but requires coinfusion of lipopolysaccharide (LPS) into the gut lumen to generate signicant tissue damage. The sequence of the insults, ie, LPS then ischemia or vice versa, does not affect the nal degree of tissue damage; however, both are necessary. Therefore, while it is clear that reduction of ow through the mesenteric artery by 60% for 6 hours is requisite for tissue damage in this model, the relative role of this ischemia with respect to luminal LPS remains unclear. The concept of ischemia as a primary cause of tissue damage in NEC generally assumes that this damage is mediated by a profound compromise of parenchymal O2 delivery; hence, blood ow and tissue O2 delivery fall below the critical level necessary to maintain mitochondrial pO2, leading to cell demise. However, there are at least two other means by which modest levels of ischemia might cause tissue damage by indirect means. First, modest ischemic episodes followed by reperfusion, the ischemiareperfusion cycle (I/R), can generate tissue damage of both vascular and parenchymal elements via generation of reactive oxygen species at reperfusion.27 I/R has been proposed as an important mechanism in NEC pathogenesis.28 Second, sustained ow reduction to 50% of baseline signicantly disrupts the intrinsic vascular effecter systems which regulate vascular resistance in newborn intestine, leading to progressively worsening vasoconstriction29,30; stated otherwise, the degree of ow reduction (50%) is not sufcient to cause tissue damage, but is of sufcient magnitude to alter the downstream vascular regulatory mechanisms so that the ischemia is sustained and possibly worsened.

ing mammals,13 this group proposed that asphyxia induced a massive sympathetic discharge that generated intestinal ischemia leading to NEC. Subsequent work by this group,14 as well as by Alward and coworkers,15 conrmed that massive postnatal asphyxia induced mucosal destruction in newborn piglets, evidence that appeared to support the asphyxiaischemia hypothesis. However, the hypothesis did not bear up under further scrutiny. Epidemiological studies of NEC failed to nd correlation between birth asphyxia, or between postnatal hypoxia (another potential source of sympathetic discharge) and NEC.16-18 It was also demonstrated that sustained stimulation of extrinsic sympathetic nerves in newborn intestine (ie, the mechanism of gut ischemia proposed by Touloukian) caused only a transient reduction in ow rate or tissue O2 uptake.19 More recent human studies have described superior mesenteric artery (SMA) blood ow velocity during late fetal life and after birth. Rhee and colleagues20 evaluated antenatal SMA ow growth-retarded fetuses, a demographic group that historically demonstrates a high incidence of NEC. To this end, Doppler ultrasonography was used to measure SMA blood ow velocity and to calculate a pulsatility index, from which an indirect estimate of vascular resistance was calculated. Forty percent of these fetuses demonstrated compromised SMA ow velocity and manifest a pulsatility index suggestive of increased intestinal vascular resistance; however, the subsequent incidence of NEC in these infants was not different from the remaining 60% of infants whose fetal SMA hemodynamic prole was normal. In a study of preterm infants, Kempley and coworkers21 reported an increased, rather than a decreased, SMA ow velocity in infants with Bells stage II NEC when compared with age-matched, unfed controls. No correlation was noted between SMA ow velocity before the onset of NEC-like symptoms and the incidence of the disease in these infants. It is important to stress that the Doppler technique measures blood ow velocity (mm/sec), a term that is only indirectly related to the volume of blood ow (ml/min); it is the latter term that is relevant to O2 transport physiology.22 Furthermore, measurement of SMA hemodynamics provides little insight into downstream intramural microvascular events. Indeed, substantial compromise of downstream ow must occur before SMA ow velocity is affected. These caveats acknowledged, the data of Rhee and Kempley indicate that a global reduction of intestinal blood ow does not precede the development of NEC; stated in the most simplistic sense, it does not appear that ischemia, at least in a global sense, is the initial insult. Animal models of NEC suggest both indirect and direct participation of ischemia in tissue damage, but do not specically address the question of timing. The rat pup model of NEC, perhaps the best accepted model to date generates disease by repetitive exposure of formula fed pups to hypoxia and hypothermia.23,24 Based on data from larger animals, it is likely that both hypoxia and hypothermia compromise intestinal perfusion in the rat pup model25;

How does the ischemia occur?

Regulation of the newborn intestinal circulation
The most salient and distinguishing feature of the newborn intestinal circulation is its very low resting vascular resistance and hence high rate of blood ow when compared with older subjects. While myriad factors contribute to regulation of intestinal vascular resistance, two stimuli, one constrictor and one dilator, dominate control of vascular resistance within the newborn intestine.31 The principal constrictor stimulus in the newborn intestinal circulation is the peptide endothelin-1, or ET-1, as evidenced by the signicant vasodilation that occurs when following attenuation of the ET-1 system.32 ET-1 is produced in endothelial cells in a constitutive fashion and exerts its constrictor effect by binding to ETA receptors on

156 adjacent vascular smooth muscle, thus acting in a paracrine fashion. The constriction generated by ET-1 is sustained and profound, and a signicant increase in ET-1 production, generated in the absence of counterbalancing vasodilator stimuli, leads to tissue hypoxia.33 Although constitutively produced, ET-1 production can be increased by a wide range of stimuli, including reduced ow rate,34 hypoxia,35 and inammatory cytokines.36,37 The principal dilator stimulus in newborn intestine is nitric oxide, or NO.38 The NO relevant to vascular regulation is generated during the reduction of L-arginine to Lcitrulline by the endothelial isoform of nitric oxide synthase (eNOS).39 eNOS is a constitutive enzyme but its activity can be increased by chemical agonists and mechanical stimuli, particularly ow rate.40 The amount of NO production is also determined by eNOS expression. The promoter region for the eNOS gene contains cis-acting responsive to ow rate, hypoxia, and the nuclear transcription factors AP-1, SP-1, and NF-B.41 Perhaps of greatest importance to regulation of the newborn intestinal circulation is that developmental regulation of eNOS occurs during perinatal life.42 Thus, eNOS expression within mesenteric artery is low during fetal life, increases after birth in association with feeding, then decreases with subsequent maturation. The low vascular resistance characteristic of newborn intestine reects substantial generation of eNOS-derived NO when compared with ET-1,31 a circumstance that mirrors the increased expression of eNOS within the newborn intestinal circulation.42 This balance favoring dilation facilitates an increased basal rate of blood ow, and hence O2 delivery to the newborn intestine, and is designed to meet the substantial oxidative demand of newborn intestine, a demand likely consequent to the tremendous rate of tissue growth present in early postnatal intestine.

Seminars in Pediatric Surgery, Vol 14, No 3, August 2005 promised endothelial cell production of NO, once again in an age-specic manner; thus, accumulation of cGMP, a marker of NO production, was decreased under both basal conditions and after stimulation of eNOS with substance P in artery segments from newborn subjects exposed to I/R.44 The conclusion from these data that is most germane to NEC is the age-specicity of the response. The effects of I/R were signicantly greater in the newborn, most likely because developmental regulation of eNOS within the newborn intestinal circulation places a far greater eNOS expression during early perinatal life. Sustained low-ow perfusion The most physiologically relevant stimulus for eNOS expression and activity is the mechanostimulus of ow rate or, more accurately, the wall shear stress generated by the ow of blood against the static endothelial surface.45 Shear stress activates eNOS and induces its upregulation, and the resulting increase in NO production generates a drop in vascular resistance and the consequent increase in ow rate in process termed ow-induced dilation. Flow-induced dilation occurs in newborn intestine,46 and direct correlations exist among ow rate, NO production, and vascular resistance within mesenteric vessels from 3-day-old, but not weanling piglets.47 If an increased ow rate increases shear stress and hence eNOS activity/expression, then it might be predicted that a sustained reduction of ow rate might have the opposite effect. This predication was conrmed in newborn piglet intestine.29 Reduction of ow rate to 50% of baseline for 6 hours through an isolated segment of small intestine in 3-day-old piglets caused a progressive rise in vascular resistance. A similar experimental manipulation in older subjects did not have a signicant effect on vascular resistance. The change noted in newborn intestine was signicantly attenuated by prior blockade of eNOS activity, indicating that the change in resistance was secondary to a loss of NO production. In a subsequent study, it was demonstrated that the contractile response to the vasoconstrictor agents angiotensin II, norepinephrine, and ET-1 was significantly enhanced in newborn intestine subjected to sustained low-ow perfusion.30 Thus, the dose-response curves for these agents were signicantly left-shifted: the magnitude of vasoconstriction to physiologically relevant concentrations of these agents was greater. Inammatory mediators Inammation plays an important, if not seminal, role in disease pathogenesis. While it is clear that inammation alters microvascular permeability, other effects also occur, including upregulation of ET-1 and the ETA receptor in response to cytokines.36,37 We have begun to investigate the effects of proinammatory mediators on the newborn intestinal circulation, approaching the problem both in situ and in vitro. Lipopolysaccharide (LPS) infused through the lumen of an isolated gut loop in a 1-day-old piglet caused a transient rise in vascular resistance that could be attenuated

Perturbations of endothelial function in newborn intestine

This environment, wherein an endothelium-derived vasodilator stimulus (NO) dominates vascular control to generate a high basal ow rate might be dramatically disrupted by a loss of endothelial integrity; specically, events that compromise eNOS activity and/or expression and simultaneously intensify ET-1 expression could lead to rapid decompensation of vascular homeostasis and tissue oxygenation. Three perturbations known to compromise endothelial integrity or function have been used to evaluate this possibility in newborn intestine. Ischemia-reperfusion I/R (60 minutes of total ischemia followed by 2 hours of reperfusion) increased intestinal vascular resistance in 3-day-old, but not weanling piglets and compromised the capacity of the 3-day-old to demonstrate intestinal vasodilation in response to a variety of hemodynamic stimuli.43 Subsequent studies revealed that this magnitude of I/R com-

Nowicki

Ischemia and NEC

157 (ie, ow interruption followed by restoration in 1A, 2A, and 3A arterioles). The net effect of these circumstances would be a generalized loss of the balance between NO and ET-1 within a critical portion of the intestinal circulation; thus, what began as a relatively isolated episode of mucosal inammation is rapidly extended by means of its effects on vascular function. Here, the instigating event is inammation with vascular pathophysiology, ie, ischemia playing a secondary role, with the principal activity occurring within the intestinal microcirculation. This scenario, while at present hypothetical, has several important clinical correlates. First, it offers an explanation for the oftentimes rapid spread of NEC, ie, via the progressive spread of vascular dysfunction within the submucosal arteriolar plexus, the principal site of resistance regulation within the intestine. Second, it provides a linkage between inammatory events and vascular dysfunction. Third, it proposes the specic sites of vascular dysfunction as well as the putative timing of ischemic events, ie, as occurring secondary to an initial inammatory response.

by blockade of endothelin ETA receptors. Subsequent immunohistochemistry demonstrated a loss of eNOS expression in villus 3A arterioles and submucosal vessels, as well as an increase in endothelin ETA receptors in terminal mesenteric arteries. In a second study, terminal mesenteric arteries from 1-day-old piglets were exposed to the proinammatory cytokine IL-1. This treatment signicantly compromised ow-induced dilation. A doubling of the ow rate increased vessel diameter by 26 5% in control vessels, but only 4 5% in vessels treated with IL-1. This IL-1 effect was duplicated by infusion of an L-arginine analog to reduce eNOS-derived NO, while the IL-1 effect was attenuated by blockade of endothelin ETA receptors. These observations suggest that the IL-1 compromise of ow-induced dilation was mediated by disruption of the NO:ET-1 balance. Furthermore, application of IL-1 to newborn endothelial cells in 1o culture reduced eNOS expression, while a similar action increased ETA expression in a 1o culture of newborn vascular smooth muscle cells, conrming that IL-1 can alter the NO and ET-1 vascular effector systems in elements from newborn arterioles.

Hypothetical relevance to NEC pathogenesis

Collectively, these studies indicate that disruption of endothelial cell function has the potential to disrupt the normal balance between NO and ET-1 within the newborn intestinal circulation, and that such an event can generate significant ischemia. In this context, it is important to note that NO and ET-1 each regulate the expression and activity of the other. An increased [NO] within the microvascular environment reduces ET-1 expression and compromises ligand binding to the ETA receptor (thus decreasing its contractile efcacy), while ET-1 compromises eNOS expression.48 Thus, factors that upset the balance between NO and ET-1 will have an immediate and direct effect on vascular tone, but also exert an additional indirect effect by extenuating the disruption of balance between these two factors. It is not difcult to construct a hypothesis that links the perturbations of I/R and sustained low-ow perfusion with an initial inammatory insult. Initiation of an inammatory process at the mucosalluminal interface could have a direct impact on villus and mucosal 3A arterioles, damaging arteriolar integrity and disrupting villus hemodynamics. Ascent of proinammatory mediators to the submucosal 1A2A arteriolar plexus could occur via draining venules and lymphatics, generating damage to vascular effector systems therein; these mediators might include cytokines and platelet activating factor, as these elements have been recovered from human infants with NEC.10,11 This event, coupled with a generalized loss of 3A ow throughout a large portion of the mucosal surface, could compromise ow rate within the submucosal arteriolar plexus. This occurrence could have two possible effects: generation of changes in eNOS and ET-1 expression secondary to sustained ow reduction, or endothelial damage in response to localized episodes of I/R

Acknowledgment
This work is supported by the National Institute for Digestive Diseases, Diabetes, and Kidney (NIDDK), grant number DK065306 awarded to P. Nowicki, MD.

References
1. Kliegman RM, Fanaroff AA. Necrotizing enterocolitis. N Engl J Med 1984;310:1093-103. 2. Balance WA, Dahms BB, Shenker N, et al. Pathology of neonatal necrotizing enterocolitis: a ten-year experience. J Pediatr 1990;117: S6-13. 3. Bohlen HG. Intestinal microvascular adaptation during maturation of spontaneously hypertensive rats. Hypertension 1983;5:739-45. 4. Bohlen HG. Determinants of resting and passive intestinal vascular pressures in rat and rabbit. Am J Physiol 1987;253:G587-95. 5. Christensen KL, Mulvany MJ. Mesenteric arcade arteries contribute substantially to vascular resistance in conscious rats. J Vasc Res 1993;30:73-9. 6. Reber KM, Nowicki PT. Pressure and ow characteristics of terminal mesenteric arteries in postnatal intestine. Am J Physiol 1998;274: G290-8. 7. Unthank JL, Bohlen HG. Lymphatic pathways and role of valves in lymph propulsion from small intestine. Am J Physiol 1988;254:G389-98. 8. Bohlen HG. Mechanism of increased vessel wall nitric oxide concentrations during intestinal absorption. Am J Physiol 1998;275:H542-50. 9. Bohlen HG, Lash JM. Intestinal lymphatic vessels release endothelialdependent vasodilators. Am J Physiol 1992;262:H813-8. 10. Caplan MS, Sun X, Hsue W, et al. Role of platelet activating factor and tumor necrosis factor-alpha in neonatal necrotizing enterocolitis. J Pediatr 1990;116:960-4. 11. Ford H, Watkins S, Reblock K, et al. The role of inammatory cytokines and nitric oxide in the pathogenesis of necrotizing enterocolotis. J Pediatr Surg 1997;32:275-82. 12. Touloukian RJ, Berdon WE, Amoury RA, et al. Surgical experience with necrotizing enterocolitis in the infant. J Pediatr Surg 1967;2:289-401.

158
13. Scholander PF. The master switch of life. Sci Am 1963;209:92-106. 14. Touloukian RJ, Posch JN, Spencer R. The pathogenesis of ischemic gastroenterocolitis of the neonate: selective gut mucosal ischemia in asphyxiated neonatal piglets. J Pediatr Surg 1972;7:194-205. 15. Alward CT, Hook JB, Helmrath TA, et al. Effects of asphyxia on cardiac output and organ blood ow in the newborn piglet. Pediatr Res 1978;12:824-7. 16. Yu VY, Joseph R, Bajuk B, et al. Perinatal risk factors for necrotizing enterocolitis. Arch Dis Child 1984;59:430-4. 17. Uauy RD, Fanaroff AA, Korones SB, et al. Necrotizing enterocolitis in very low birth weight infants: biodemographic and clinical correlates. National Institute of Child Health and human Development Neonatal Research Network. J Pediatr 1991;119:630-8. 18. Stoll BJ. Epidemiology of necrotizing enterocolitis. Clin Perinatol 1994;21:205-18. 19. Nowicki PT, Miller CE, Hayes JR. Effect of sustained mesenteric nerve stimulation on intestinal oxygenation in developing swine. Am J Physiol 1991;260:G333-9. 20. Rhee E, Detti L, Mari G. Superior mesenteric artery ow velocity waveforms in small for gestational age fetuses. J Matern Fetal Med 1998;7:120-3. 21. Kempley ST, Gamsu HR. Superior mesenteric artery blood ow velocity in necrotizing enterocolitis. Arch Dis Child 1992;67:793-6. 22. Granger HJ, Nyhof RA. Dynamics of intestinal oxygenation: interactions between oxygen supply and uptake. Am J Physiol 1982;243: G91-6. 23. Barlow B, Santulli TV. Importance of multiple episodes of hypoxia or cold stress on the development of enterocolitis in an animal model. Surgery 1975;77:687-90. 24. Caplan MS, Hedlund E, Adler L, et al. Role of asphyxia and feeding in a neonatal rat model of necrotizing enterocolitis. Pediatr Pathol 1994;14:1017-28. 25. Nowicki PT, Miller CE, Edwards RC. Effects of hypoxia and ischemia on autoregulation in postnatal intestine. Am J Physiol 1991;261: G162-7. 26. Barlow B, Santulli TV, Heird WC, et al. An experimental study of acute neonatal entercolotis: the importance of breast milk. J Pediatr Surg 1974;9:587-95. 27. Boros M, Takaichi S, Hatanaka K. Ischemic time-dependent microvascular changes and reperfusion injury in the rat small intestine. J Surg Res 1995;59:311-20. 28. Chan KL, Hui CW, Chan KW, et al. Revisiting ischemia and reperfusion injury as a possible cause of necrotizing enterocolitis: role of nitric oxide and superoxide dismutase. J Pediatr Surg 2002;37:828-34. 29. Nowicki PT. Effects of sustained ow reduction on postnatal intestinal circulation. Am J Physiol 1998;275:G758-68. 30. Nowicki PT. Effects of sustained low-ow perfusion on the response to vasoconstrictor agents in postnatal intestine. Am J Physiol 1999; 276:G1408-16.

Seminars in Pediatric Surgery, Vol 14, No 3, August 2005

31. Nankervis CA, Dunaway DJ, Nowicki PT. Determinates of terminal mesenteric artery resistance during the rst postnatal month. Am J Physiol 2001;280:G678-86. 32. Nankervis CA, Nowicki PT. Role of entothelin-1 in regulation of the postnatal intestinal circulation. Am J Physiol 2000;278:G367-75. 33. Yanagisawa M, Kurihara H, Kimura S, et al. A novel potent vasoconstrictor peptide produced by vascular endothelial cells. Nature 1998; 332:411-5. 34. Kuchan MJ, Frangos JA. Shear stress regulates endothelin-1 release via protein kinase C and cGMP in cultured endothelial cells. Am J Physiol 1993;264:H150-6. 35. Kourembanas S, Marsden PA, McQuillan LP, et al. Hypoxia induces endothelin gene expression and secretion in cultured human endothelium. J Clin Invest 1991;88:1054-7. 36. Woods M, Mitchell JA, Wood EG, et al. Endothelin-1 is induced by cytokines in human vascular smooth muscle cells: evidence for intracellular endothelin converting enzyme. Mol Pharmacol 1999;55:902-9. 37. Yoshizumi M, Kurihara H, Morita T, et al. Interleukin 1 increases the production of endothelin-1 by cultured endothelial cells. Biochem Biophys Res Commun 1990;166:324-9. 38. Nankervis CA, Nowicki PT. Role of nitric oxide in regulation of vascular resistance in postnatal intestine. Am J Physiol 1995;268: G949-58. 39. Marletta MA. Nitric oxide synthase structure and mechanism. J Biol Chem 1993;268:12231-4. 40. Davis ME, Cai H, Drummond GR, et al. Shear stress regulates endothelial nitric oxide synthase expression througth c-Src by divergent signaling pathways. Circ Res 2001;89:1073-80. 41. Kuboki K, Jiang ZY, Takahara N. Regulation of endothelial constitutive nitric oxide synthase gene expression in endothelial cells and in vivo: a specic vascular action of insulin. Circulation 2000;101:676-81. 42. Reber KM, Su BY, Clark KR, et al. Developmental expression of eNOS in postnatal swine mesenteric artery. Am J Physiol 2002;283: G1328-35. 43. Nowicki PT, Nankervis CA, Miller CE. Effects of ischemia and reperfusion on intrinsic vascular regulation in the postnatal intestinal circulation. Pediatr Res 1993;33:400-4. 44. Nowicki PT. Postnatal changes in gut hemodynamics: a possible role for substance P. Am J Physiol 1998;274:G1142-50. 45. Koller A, Kaley G. Endothelial regulation of wall shear stress and blood ow in skeletal muscle microcirculation. Am J Physiol 1991; 260:H862-8. 46. Nowicki PT, Miller CE. Flow-induced dilation in newborn intestine. Pediatr Res 1995;38:783-91. 47. Reber KM, Mager GM, Miller CE, et al. Relationship between ow rate and NO production in postnatal mesenteric arteries. Am J Physiol 2001;280:G43-50. 48. Boulanger C, Luscher TF. Release of endothelin from the porcine aorta. Inhibition by endothelium-derived nitric oxide. J Clin Invest 1990;85:587-90.