DMBT1 expression and neutrophil-to-lymphocyte ratio during necrotizing enterocolitis are influenced by impaired perfusion due to cardiac anomalies

Necrotizing enterocolitis (NEC) is a very serious disease in preterm and term infants. Incidence rates depend on birth weight and prematurity with a maximum of up to 15% of affected neonates born at less than 30 weeks of gestational age or birth weight of < 1500 g [1]. In contrast, almost 10% of all patients diagnosed with NEC are term neonates, 50% of whom are presenting with persistent ductus arteriosus (PDA) or/and congenital heart disease (CHD) [2]. The increased risk for NEC in these cases is caused by the steal phenomenon leading to reduced perfusion of the intestine and—at least in cyanotic CHD and in preterm infants presenting multiple oxygen desaturations due to PDA—by additional hypoxia. NEC and CHD are shown to be interrelated as main factors for morbidity and mortality in neonates. There is an increased risk of mucosal ischemia due to circulatory alterations, provoking intestinal necrosis [3]. Multiple factors are additionally described as risk factors for NEC (e.g., chorioamnionitis, placental insufficiency, gut ischemia) [4]. The immature immune response of infants and particularly in preterm infants hereby substantially contributes to NEC: intestinal innate and adaptive immunity and inflammatory control are immature and lead to dysfunction of the intestinal barrier [5‐7]. These risk factors seem to alter the intestinal epithelial barrier, as a result of which bacteria are able to translocate from the intestinal lumen into the sub-epithelial tissue and induce inflammation [4]. NEC thus remains a multifactorial induced disease. Research of the last decade therefore focused on the separation of pathology of NEC in patient subgroups [8‐11]. Various publications were able to confirm differences in incidence, mortality, and clinical factors of patients with CHD with concomitant steal phenomenon. In these patients, pathophysiology is primarily based on an impaired intestinal circulation with resulting ischemia and elevated levels of circulating endotoxins and proinflammatory cytokines. A recent publication from Neu proposed to completely separate this intestinal necrosis due to hypoxic mechanisms seen in full-term neonates from the classical NEC of preterm infants [12] as did the most recent study of Klinke et al. [13]. We were recently able to describe that infants with PDA/CHD had more frequently macroscopic intestinal necrosis and positive bacterial culture of intraoperative swabs compared to preterm infants without PDA/CHD [11].

Deleted in malignant brain tumors 1 (DMBT1) is a secreted protein with functions in innate immunity, epithelial cell differentiation, angiogenesis, and tumorigenesis. As member of the secreted scavenger receptor cysteine rich (SRCR) protein family it is involved in various processes of inflammation [14]. This includes binding to various bacteria and interaction with several defense factors (e.g., surfactant protein A/D, secretory IgA) [15, 16]. DMBT1 is expressed in the fetal gastrointestinal tract and additionally in the small and large intestine of preterm infants [17]. The basal and luminal localization in epithelial cells point to functions in epithelial differentiation (basal DMBT1) and local innate immunity (luminal DMBT1). This was supported by the fact that DMBT1 is enormously upregulated in inflammatory processes such as NEC independently from gestational age [17]. The promotor of DMBT1 furthermore contains a binding site for hypoxia-inducible factor 1α, which could be observed with the upregulation of DMBT1 expression in lung tissues and lung epithelial cells after hypoxia [18].

The aim of this study was to compare intestinal DMBT1 expression, intraoperative bacterial detection and the clinical course in infants with NEC according to their cardiac anatomy and to provide new insides into the pathophysiology of different NEC subtypes.

Different studies support the critical role of DMBT1 in epithelial differentiation, fetal, and neonatal innate immunity [16, 17] and especially functions of DMBT1 in the intestine [23]. Accordingly, increased DMBT1 expression with increasing gestational ages has been previously demonstrated in the small intestine of human fetus (gestational age of 14–35 weeks) and in neonates. An upregulation of DMBT1 has been proven in neonatal gastrointestinal diseases to be associated with inflammation (e.g., necrotizing enterocolitis) and infection in preterm patients (e.g., severe sepsis) independent from gestational age [17]. In this study, we focused on NEC in preterm and term neonates with different cardiac anatomy to analyze potential new pathophysiological mechanisms of different NEC subgroups. We postulate that DMBT1 might be differentially expressed regarding the immaturity of the innate immunity alone versus in combination with impaired intestinal perfusion and consecutive cellular hypoxia, since inflammation and hypoxia are known factors which influence DMBT1 expression [17, 18]. Elevated circulating cytokines are described as part of the systemic inflammatory response during NEC, which may also play a direct or indirect role in augmenting the mucosal injury [4].

DMBT1 acts as a binding protein to multiple bacteria, such as group A and B Streptococci, Staphylococcus aureus, and E. coli [16]. The binding results in suppression of bacterial invasion and hindrance of bacterial infections [24]. NEC develops in response to an abnormal interaction of exaggerated bacterial signaling in the premature intestine [4], in which E. coli plays a major role. The results of our study confirm an association of positive intraabdominal swabs and the presence of CHD/PDA, as demonstrated in a recent study [11] and show a predominance of E. coli and Staphylococcus in intraabdominal swabs. DMBT1 addresses a critical factor in the pathogenesis of NEC. Upregulation of DMBT1 in response to bacterial invasion is supported by our observation that the total DMBT1 score was significantly associated with positive bacterial culture of intraoperative swabs. However, an association between different bacterial species and DMBT1 levels could not be confirmed. DMBT1 is able to bind various species of bacteria. Higher levels of DMBT1 can bind a higher load of bacteria [24, 25]. The elevated total DMBT1 expression of the serosa in fulminant NEC also emphasized the positive relation between intense and pronounced inflammation of the intestinal wall and DMBT1 expression.

The total intestinal DMBT1 expression level was slightly but not significantly higher in infants with PDA/CHD compared to infants without cardiac anomalies. This corresponds to the higher inflammation and supports the results of our own data indicating that infants with PDA/CHD had more frequently macroscopic intestinal necrosis and positive bacterial culture of intraoperative swabs compared to preterm infants without PDA/CHD [11]. Cellular hypoxia may also lead to higher DMBT1 expression in intestinal epithelium. This observation has already been made in the respiratory epithelium [18]. Even though the gestational ages between both groups (preterm infants versus infants with PDA/CHD) were slightly different (31.1 versus 36.5 weeks), this cannot explain the differences in DMBT1 expression as shown in our previous study, as upregulation of DMBT1 is independent from gestational age and only the basal level of intestinal DMBT1 expression depends on immaturity [17].

The reduced perfusion and reduced nutrition of the intestine in infants with PDA/CHD may represent an explanation for the negative correlation between DMBT1 levels and CRP. Even cases with the highest DMBT1 expression did not show CRP values > 50 mg/l in every case. This demonstrates that there is a very intense local inflammation with high induction of DMBT1 expression, but the systemic reaction with CRP production is often less intense compared to local DMBT1 expression in infants with impaired intestinal perfusion and impaired systemic circulation due to cardiac anomalies. The initiation of systemic CRP requires intestinal perfusion to induce the systemic reaction.

Intestinal macrophages appear at 11–12 weeks of gestational age in human fetal development followed by a rapid increase of the resident macrophage population during the 12–22-week period [26‐28]. Intestinal macrophages represent an important host defense mechanism. As the first phagocytic cells of the innate immune system, they encounter luminal bacteria that translocate the intestinal epithelium into the lamina propria. The inflammatory responses of the mucosal macrophages contribute to the risk of mucosal injury during NEC [4]. A macrophage-rich infiltrate is involved in the cellular inflammatory response in NEC [29]. It was shown that intestinal macrophages are normally maintained through continuous recruitment of circulating monocytes followed by in situ differentiation in the lamina propria [30, 31]. The higher amount of DMBT1-positive intestinal macrophages in infants with PDA/CHD in comparison to NEC patients with sufficient intestinal perfusion demonstrated a higher inflammatory response of macrophages in the first mentioned NEC subgroup. Even if our results are only able to represent the absolute numbers of DMBT1-positive macrophages and fail at presenting relative numbers of DMBT-1 positive macrophages, this information might contribute to the extended macroscopic intestinal necrosis and positive bacterial culture of intraoperative swabs observed in infants with PDA/CHD [11].

We further analyzed the NLR and MLR in dependence of cardiac anatomy. Both NLR and MLR are discussed as outcome parameter in adult intensive care unit patients [32]. A rise in blood neutrophils is part of an appropriate inflammatory response [4]. NLR is a simple parameter for easy assessment of a patient’s inflammatory status [33]. Yang et al. reported that the NLR may be a useful marker for early NEC diagnosis and could allow to distinguish the severity of NEC [34]. During NEC, neutrophils emigrate into the intestines and peritoneum, and increased margination of neutrophils in the microvasculature is known [4], which was also seen in our study population (Fig. 2i/k).

Intestinal macrophage populations are recruited from circulating monocytes followed by in situ differentiation in the lamina propria [30, 31]. Accordingly, monocytes guarantee that macrophage-rich infiltrates are present in NEC [29]. We observed higher MLR in infants with PDA/CHD compared to infants with normal cardiac anatomy, supporting the role of monocytes/macrophages in NEC of infants with PDA/CHD.

There are different aspects due to inflammation in NEC. First, the limited mucosal immune system increases the risk of inflammatory injury and NEC in preterm infants. Second, elevated circulating cytokine levels due to intestinal inflammation and increased inflammatory status increase intestinal injury [4]. There is an underlying elevation of circulating endotoxins and proinflammatory cytokines in term and predominantly preterm neonates, which predisposes neonates to intestinal necrosis [6]. The stimulation of nuclear factor κB (NF-κB) is induced by pro-inflammatory stimuli, which subsequently induces the transcription of pro-inflammatory mediators such as interleukin 6 and 8 (IL-6/IL-8), tumor necrosis factor alpha (TNFα), interferon γ, and platelet-activating factor (PAF), ultimately leading to intestinal necrosis. The role of IL-6 in neonatal sepsis or intestinal necrosis has additionally been well established, presenting itself as a promising new biomarker [35]. A correlation of DMBT1-/- mice and colitis could be confirmed, in which elevated levels for IL-6 and TNF during inflammation could be seen [23].

It remains unclear if the increased DMBT1 levels in NEC patients may augment the intestinal injury or may help to bind bacteria [25] and to suppress inflammation as already observed in gut and lung epithelial cells [23, 36]. It may also be possible that DMBT1 has different functions in different time windows.

The study has some limitations: First, both subgroups include preterm infants. To exactly define the exclusive role of gestational age and cardiac malformation on DMBT1 expression as well on NEC manifestation, it may be useful to examine greater study populations and create subgroups enabling examination of the distinct role of prematurity and cardiac malformation. Second, improvement of the pathologic handling of resected intestinal tissue (e.g., longitudinal versus transverse cutting) may enable accurate comparability of tissue samples in terms of assessment of DMBT1-positive structures and in terms of the ratio of DMBT1-positive macrophages.

In conclusion, we confirm that DMBT1 expression may be influenced by various states of inflammation or cellular hypoxia in the neonatal intestine. The hypothesis of different pathologic conditions in various NEC subtypes is supported by the observation that DMBT1 expression depends on NEC subgroup and that NEC in infants with PDA/CHD is characterized by a significantly increased number of DMBT1-positive macrophages. Furthermore, the neutrophil-to-lymphocyte ratio is significantly higher in infants with an impaired intestinal perfusion due to cardiac anomalies.