Necrotizing enterocolitis (NEC), is the most common gastrointestinal emergency in the premature infant (Neu, Reference Neu1996). Although NEC is a multifactorial disease, prematurity is a main risk factor. Mucosal injury resulting from ischemia, bacterial colonization and early enteral feeding are also recognized as potentially important contributors to the pathogenesis of NEC (Kliegman et al. Reference Kliegman, Walker and Yolken1993; Neu, Reference Neu2005).
It has been shown that in premature infants at the time of diagnosis of NEC, arginine plasma levels are decreased (Zamora et al. Reference Zamora, Amin, McMillan, Kubes, Fick, Butzner, Parsons and Scott1997; Becker et al. Reference Becker, Wu, Galanko, Chen, Maynor, Bose and Rhoads2000). Arginine is one of the most versatile amino acids in mammalian cells, which is crucial for ammonia detoxification and serves as the precursor of NO (Wu & Morris, Reference Wu and Morris1998; Wu et al. Reference Wu, Jaeger, Bazer and Rhoads2004). NO plays an important role in several pathophysiological aspects of critical illness such as infection, organ injury, but also in inducing gut smooth muscle relaxation, the regulation of mucosal blood flow and maintaining mucosal integrity and the intestinal barrier function (Stark & Szurszewski, Reference Stark and Szurszewski1992; Alican & Kubes, Reference Alican and Kubes1996).
In an experimental piglet model, it was shown that arginine could have therapeutic potential in NEC by attenuating intestinal injury, whereas administration of the NO synthase (NOS) inhibitor L-NAME caused haemorrhagic congestion of the gut wall (Di Lorenzo et al. Reference Di Lorenzo, Bass and Krantis1995). Previously, we showed in a rat model that low arginine plasma levels in combination with a low dose of endotoxin compromised blood flow through the small intestine (Nijveldt et al. Reference Nijveldt, Prins, Siroen, Rauwerda, Teerlink and van Leeuwen2000). Interestingly, Amin et al. (Reference Amin, Zamora, McMillan, Fick, Butzner, Parsons and Scott2002) showed that arginine supplementation in premature infants, reduced the incidence of all stages of NEC.
Recent insights into NO metabolism have shown an important role of endogenously produced inhibitors of NOS, in particular asymmetric dimethylarginine (ADMA; Vallance & Leiper, Reference Vallance and Leiper2004). ADMA as well as symmetric dimethylarginine (SDMA), are synthesized when arginine residues in proteins are methylated by the action of protein arginine methyltransferases. ADMA is an endogenous inhibitor of all isoforms of NOS, while SDMA is not.
Arginine is transported into endothelial cells by the cationic amino acid transporters (CAT) of system y+, where it serves as a substrate for NO synthesis (Closs et al. Reference Closs, Basha, Habermeier and Forstermann1997). Thus, the arginine:ADMA ratio is an important determinant of NO production by NOS (Bode-Boger et al. Reference Bode-Boger, Boger, Kienke, Junker and Frolich1996; Boger et al. Reference Boger, Bode-Boger, Thiele, Creutzig, Alexander and Frolich1998).
Elevated plasma levels of ADMA have been found in diseases related to endothelial dysfunction including peripheral arterial disease, hypertension, hyperlipidemia, diabetes mellitus and hyerhomocysteinemia (Siroen et al. Reference Siroen, Teerlink, Nijveldt, Prins, Richir and van Leeuwen2006). Interestingly, ADMA have been shown to independently predict intima media thickness of the carotid artery in individuals without clinical manifestations of arteriosclerosis (Miyazaki et al. Reference Miyazaki, Matsuoka, Cooke, Usui, Ueda, Okuda and Imaizumi1999; Nanayakkara et al. Reference Nanayakkara, Teerlink, Stehouwer, Allajar, Spijkerman, Schalkwijk, Ter Wee and van Guldener2005). Furthermore, it has been shown that ADMA is an independent risk factor for CHD and in critically ill patients an independent risk factor for intensive-care-unit mortality (Nijveldt et al. Reference Nijveldt, Teerlink, van der Hoven, Siroen, Kuik, Rauwerda and van Leeuwen2003a; Schulze et al. Reference Schulze, Lenzen and Hanefeld2006).
We hypothesize that in addition to low arginine levels, premature infants with NEC have an increased plasma concentration of ADMA. Increased ADMA levels are associated with reduced NO synthesis and may consequently lead to a decreased mucosal blood flow and barrier function of the gut.
Therefore, the primary aim of this study was to measure the arginine, ADMA, and SDMA concentrations in premature infants with NEC and in control infants without NEC. The secondary aim was to study the relation between arginine, ADMA, and SDMA concentrations and mortality in premature infants with NEC.
We performed a prospective case–control study with a 1:1 ratio of case to control subjects. The hospital ethics committee approved the study and all infants were included after written informed consent of their parents.
All infants with a gestational age < 34 weeks (assessed by last menstrual period of the mother and/or ultrasound before 20 weeks gestation) and a birth weight ≤ 2000 g admitted to the level III neonatal intensive care unit of the VU University Medical Center or the Academic Medical Center, Amsterdam, who developed NEC, were eligible for participation in the study. Only patients with NEC ≥ grade II diagnosed between 08.00 hours and 20.00 hours were included in the study. Exclusion criteria were major congenital or chromosomal anomalies.
NEC was diagnosed and classified according to the criteria of Bell et al. (Reference Bell, Ternberg, Feigin, Keating, Marshall, Barton and Brotherton1978). NEC grade I was characterized by non-specific systemic signs and abdominal signs including increased gastric residuals or abdominal distension. NEC grade II encompasses the signs of grade I, absent bowel sounds or abdominal tenderness. The abdominal radiograph shows intestinal dilation, ileus and pneumatosis intestinalis. NEC grade III is characterized by bowel perforation visualized as pneumoperitoneum on the abdominal radiograph. An independent paediatric radiologist made the diagnosis of pneumatosis intestinalis.
According to the protocol guidelines for parenteral and enteral nutrition, all infants received either breast milk with breast milk fortifier (intake about 57 mg arginine /100 ml; Dupont, Reference Dupont2003), and/or a formula feeding (Nenatal Start®, Nutricia Nederland BV, Zoetermeer, The Netherlands, containing 71 mg arginine /100 ml) and/or total parenteral nutrition (a hospital pharmacy based all-in-one mixture, containing 102 mg arginine/100 ml). After the diagnosis of NEC, a peripheral venous blood sample (0·3 ml) was collected and the infants were treated according to the standard protocol for NEC; nothing by mouth, broad-spectrum antibiotics and parenteral feeding. If this treatment failed or in the case of intestinal perforation, surgery was performed.
For each premature infant with NEC, one premature infant without NEC was selected. The selection was based on the characteristics of the NEC patient including gestational age, birth weight, head circumference, Apgar score at 5 min and postnatal age at the time of the NEC diagnosis. After inclusion, a peripheral venous blood sample (0·3 ml) was collected.
Plasma analysis of arginine, symmetric and asymmetric dimethylarginine
The blood samples taken from the premature infants were immediately placed on ice and centrifuged at 3000 rpm for 10 min at 4°C. Plasma was immediately put in liquid N2, and stored at − 80°C before analysis.
The concentration of ADMA, arginine and SDMA were determined simultaneously by HPLC as described previously (Teerlink et al. Reference Teerlink, Nijveldt, De Jong and van Leeuwen2002). In brief, solid-phase extraction on polymeric cation-exchange columns was performed using monomethylarginine as the internal standard. After derivatization with o-phthaldialdehyde reagent containing 3-mercaptopropionic acid, analytes were separated by isocratic reversed-phase HPLC with fluorescence detection. Intra- and inter-assay CV were better than 1·2 and 3·0 %, respectively. The arginine:ADMA ratio was calculated.
Due to non-normal distribution of the amino acids, non-parametric tests were used, and data are presented as median and range. The Mann-Whitney U test was used to investigate differences in plasma amino acid concentrations between the premature infants with NEC and the control group, and between surviving and non-surviving NEC patients. A value of P < 0·05 was considered statistically significant. Statistical analysis was performed using SPSS 11·0 for Windows (SPSS Inc., Chicago, IL, USA).
Between September 2001 and October 2003, a total of thirty-three premature infants with NEC were eligible to enter the study. Of these thirty-three infants, twenty-three were excluded from the study (thirteen infants with NEC grade I and ten infants who were diagnosed and treated at night or during the weekend). The ten premature infants with NEC ≥ grade II were included in the study. Six premature infants with NEC were treated by surgery, while four infants were treated with antibiotics and parenteral nutrition only. The control group consisted of ten case-controlled premature infants without NEC. Patient and nutrition characteristics are presented in Table 1. The patient characteristics were not different between the NEC and control group. Breast milk feeding, parenteral nutrition energy intake and the intake of arginine, carbohydrate, protein and fat were not different between the two groups. Plasma concentrations of arginine, ADMA, SDMA and the arginine:ADMA ratio are shown in Table 2. The arginine concentration in premature infants with NEC was lower (P = 0·001) compared to the infants without NEC. The ADMA concentration in infants with NEC was also lower (P = 0·009) compared to the control group. Plasma SDMA was not different between the two groups. The arginine:ADMA ratio was lower (P = 0·023) in the NEC group.
Of the ten NEC patients, four patients died because of circulatory and respiratory insufficiency. In the non-surviving NEC infants, arginine and arginine:ADMA concentrations were lower (P = 0·01, P = 0·038 respectively) than in the surviving infants (Table 3).
The plasma concentrations of the premature infants without NEC and the surviving NEC patients are shown in Table 4. The plasma concentration of arginine, ADMA and SDMA were lower in the surviving NEC patients (P = 0·022, P = 0·011, P = 0·022 respectively). The arginine:ADMA ratio was not different between the two groups.
Our study shows that premature infants with NEC have significantly lower plasma concentrations of arginine than premature infants without NEC. The lower concentration of arginine in NEC patients is in line with the results of other studies (Zamora et al. Reference Zamora, Amin, McMillan, Kubes, Fick, Butzner, Parsons and Scott1997; Becker et al. Reference Becker, Wu, Galanko, Chen, Maynor, Bose and Rhoads2000). In addition, ADMA and the arginine:ADMA ratio are also significantly lower in premature infants with NEC. Furthermore, we found in non-surviving NEC patients a significantly lower arginine concentration and arginine:ADMA ratio than in surviving NEC patients. However, compared to the control patients, the surviving NEC patients had still significantly lower arginine and ADMA concentrations, but the arginine:ADMA ratio was not significantly different. These results indicate that NEC patients have less substrate available for NO synthesis, which is important to sustain blood flow in the gut of these premature infants. Moreover, these results suggest that a decreased availability of substrate for NO synthesis is associated with an increased risk of mortality.
In adult mammals, approximately 60 % of net arginine synthesis occurs in the kidney, whereas in neonates, arginine is synthesized in the enterocytes of the small intestine (Wu & Morris, Reference Wu and Morris1998). Wu & Knabe (Reference Wu and Knabe1995) studied the arginine synthesis in the enterocytes of fetal piglets and concluded that there was little synthesis of arginine in the enterocytes of preterm piglets. This near absence of intestinal arginine synthesis in preterm piglets is in sharp contrast to the high rate of intestinal arginine synthesis in term piglets (Wu, Reference Wu1997). In the enterocytes of preterm piglets, there is low activity of the Δ1–pyrroline-5-carboxylate synthase enzyme, which may limit the conversion of glutamine into citrulline (Wu & Knabe, Reference Wu and Knabe1995). Furthermore, there is negligible expression of argininosuccinate synthase and argininosuccinate lyase. Both enzymes are essential for the conversion of citrulline into arginine. Notwithstanding that these studies are performed in animal models, decreased expression of argininosuccinate synthase and argininosuccinate lyase and the incapability of the fetal kidney to synthesize arginine may account for the decreased level of arginine in premature infants.
Arginine is a physiological precursor for the synthesis of NO which plays an important role in regulating vascular tone, the development of gastrointestinal circulation, and in preserving the integrity of the gastrointestinal mucosal barrier (Stark & Szurszewski, Reference Stark and Szurszewski1992; Alican & Kubes, Reference Alican and Kubes1996; Wu & Morris, Reference Wu and Morris1998; Wu et al. Reference Wu, Jaeger, Bazer and Rhoads2004). Conversely, the NOS inhibitor L-NNA preferentially reduces the gastrointestinal blood flow as shown in fetal sheep (Fan et al. Reference Fan, Smolich, Wild, Yu and Walker1996). In a rat model, the NOS inhibitor L-NMMA enhanced intestinal damage after administration of endotoxin. However, this damage could be reversed by administration of arginine (Hutcheson et al. Reference Hutcheson, Whittle and Boughton-Smith1990). The mucosal permeability of the feline ileum increased after infusion with the NO synthesis inhibitor L-NAME, while infusion with the NO donor sodium nitroprusside or arginine reversed this effect (Kubes, Reference Kubes1992). These animal studies indicate that a low plasma concentration of arginine, and inhibition of the arginine–NO pathway, could contribute to decreased synthesis of NO, thereby resulting in decreased gastrointestinal blood flow and impairment of the intestinal mucosal barrier.
Interestingly, in contrast to our hypothesis, a significantly lower ADMA concentration was found in premature infants with NEC. Because ADMA is an endogenous inhibitor of all isoforms of NOS, a decrease of ADMA in premature infants with NEC could be associated with attempted regulatory changes aimed at a preservation of NO production in the presence of low arginine concentrations.
Although we cannot designate a single mediator for the observed effect, theoretically an increased metabolic turnover of ADMA may be responsible for the lower levels in NEC patients. ADMA is eliminated from the body by renal excretion as well as being metabolized by the enzyme dimethylarginine dimethylaminohydrolase (DDAH) which is widely distributed, in particular in the pancreas, liver and kidney (Kimoto et al. Reference Kimoto, Tsuji, Ogawa and Sasaoka1993; Kimoto et al. Reference Kimoto, Whitley, Tsuji and Ogawa1995). Induction of this enzyme or an increased renal excretion in NEC patients could be responsible for the decreased ADMA concentration (Yudkoff et al. Reference Yudkoff, Nissim, Pereira and Segal1984; Kimoto et al. Reference Kimoto, Whitley, Tsuji and Ogawa1995). Nijveldt and co-workers (Nijveldt et al. Reference Nijveldt, Teerlink, van Guldener, Prins, van Lambalgen, Stehouwer, Rauwerda and van Leeuwen2003b; Nijveldt et al. Reference Nijveldt, Siroen, Teerlink, van Lambalgen, Rauwerda and van Leeuwen2004) showed a reduced systemic ADMA concentration in rats during endotoxemia, indicating increased DDAH activity. Also, acute Escherichia coli endotoxaemia in human subjects decreased the plasma arginine:ADMA ratio (Mittermayer et al. Reference Mittermayer, Namiranian, Pleiner, Schaller and Wolzt2004). Since arginine and other cationic amino acids such as ornithine and lysine and also ADMA and SDMA are transported into endothelial cells by CAT, changes in the expression of CAT mRNA by endotoxemia may influence ADMA transport (Closs et al. Reference Closs, Basha, Habermeier and Forstermann1997). Hattori et al. (Reference Hattori, Kasai and Gross1999), showed that expression of CAT 1 and 2 mRNA in the lung, heart, and kidney was increased by injection of lipopolysaccharide–interferon γ in rats whereas CAT 2A mRNA was abundantly expressed in the liver independent of lipopolysaccharide–interferon γ treatment. The abundant expression of CAT 2 mRNA in the liver indicates a potentially high uptake of dimethylarginines in this organ. The important function of endotoxemia inducing DDAH, as well as the expression of the CAT, provides a possible explanation for lower ADMA concentration in NEC patients.
Although the pathophysiology of NEC is multifactorial, there is some evidence suggesting that the development of NEC is associated with intestinal gram-negative bacterial growth (Neu, Reference Neu1996; Neu, Reference Neu2005; Westerbeek et al. Reference Westerbeek, van den, Lafeber, Knol, Fetter and van Elburg2006). Gram-negative bacteria produce the endotoxin lipopolysaccharide, inducing endotoxemia and the endogenous production of platelet activating factor and TNF (Feuerstein & Hallenbeck, Reference Feuerstein and Hallenbeck1987; Beutler & Cerami, Reference Beutler and Cerami1988; Bone, Reference Bone1991; Zanetti et al. Reference Zanetti, Heumann, Gerain, Kohler, Abbet, Barras, Lucas, Glauser and Baumgartner1992). The combination of lipopolysaccharide, platelet activating factor and TNF increases intestinal epithelial permeability and often causes necrosis which could contribute to the development of NEC (Hutcheson et al. Reference Hutcheson, Whittle and Boughton-Smith1990; Hsueh et al. Reference Hsueh, Caplan, Qu, Tan, De Plaen and Gonzalez-Crussi2003). Therefore, endotoxemia due to bacterial growth and release of lipopolysaccharide in premature infants can not only contribute to the development of NEC but can also induce DDAH and the expression of CAT, thereby increasing elimination of ADMA.
The lower concentration of ADMA could be a compensatory mechanism to counteract the disadvantageous effect of low arginine concentrations in NEC patients. However, since the decrease of arginine is larger than the decrease of ADMA, the arginine:ADMA ratio is lower in NEC patients, suggesting a reduced capacity for NO synthesis.
In this study we did not measure the oxidation products nitrate and nitrite as an indirect determination of endogenous formation of NO. Because the nitrate and nitrite concentration is influenced by endogenous NO synthesis, dietary intake and also by the excretion in urine, faeces and expired air, nitrate and nitrite concentration cannot be used as a truly meaningful indicator of NO production in premature infants (Baylis & Vallance, Reference Baylis and Vallance1998).
In conclusion, the present study confirms the existence of low arginine concentrations in premature infants with NEC. In contrast to our hypothesis, the ADMA concentration is lower in these infants. In spite of the low ADMA concentration, the arginine:ADMA ratio is lower in NEC patients than in controls. Furthermore, a low arginine plasma concentration and a low arginine:ADMA ratio are associated with an increased risk of mortality in NEC patients. These findings suggest that despite the increased need for arginine, premature infants with NEC have lower availability of this substrate to synthesize NO, which may be involved in the pathophysiology of NEC. The results of our study suggest the association between arginine, ADMA and NEC. However it remains to be determined whether low levels of arginine contribute to the pathophysiology of NEC, or if low arginine levels are a consequence of NEC. As suggested in the few other studies to date, we agree that the supplementation of arginine may prevent the development of NEC in premature infants. Due to the low incidence of NEC (on average 6 out of 300 patients per year), and a high local nosocomial infection rate in the VU Medical Center (van der Zwet et al. Reference van der Zwet, Kaiser, van Elburg, Berkhof, Fetter, Parlevliet and Vandenbroucke-Grauls2005) a multicentre randomized controlled study of arginine supplementation in premature infants is needed to determine whether arginine supplementation reduces the development of NEC.
The authors thank Ms Sigrid de Jong for measurement of the amino acids. M. C. Richir and M. P. C. Siroen contributed equally to this work.