MicroRNA-320а as a novel biomarker at preclinical stage of necrotizing enterocolitis in term neonates with congenital heart defects

Abstract

BACKGROUND

Necrotizing enterocolitis (NEC) remains a prominent gastrointestinal emergency among infants, particularly term infants with congenital heart defects (CHD) being at high risk. The molecular processes that contribute to NEC have yet to be completely understood. The high mortality rates necessitate an active search for noninvasive biomarkers that can aid in the preclinical diagnosis and prognosis of NEC. MicroRNAs (miRs), which are involved in many biological processes in both health and disease, have been discovered to play an important role in regulating inflammation and immune responses via various signaling pathways.

AIM

To determine the plasma levels of miR-155, miR-221, miR-223, miR-320a, miR-451a as potential NEC biomarkers in term newborns with CHD.

METHODS

This prospective cohort study included twenty-tree term newborns with CHD who underwent cardiac surgery on the median day of life (DOL) = 7. Nine of them developed NEC (Bell’s stage IIA and IIIA) within 1 week of cardiac surgery (NEC newborns). Blood samples were collected before (median DOL = 5) and following (median DOL = 13) cardiac surgery. Levels of plasma miR-155-5p, miR-221-3p, miR-223-3p, miR-320a-3p, and miR-451a were determined using real-time polymerase chain reaction. The functional analysis was executed using the DIANA-miRPath v4.0.

RESULTS

Preoperatively, NEC newborns had significantly lower plasma levels of miR-155 (2.70-fold, P = 0.020), miR-223 (2.42-fold, P = 0.030), and miR-320a (3.62-fold, P = 0.006) than newborns without NEC. Postoperatively, miR-451a levels differed significantly between the newborn groups, showing a 4.70-fold decrease (P = 0.014) in expression when clinical NEC symptoms appeared. According to receiver operating characteristic analysis, miR-320a was found to be the most effective predictive biomarker for NEC [area under the curve (AUC) = 0.835, 63% sensitivity, 100% specificity], while miR-451a was identified as a NEC biomarker (AUC = 0.835, 85.7% sensitivity, 76.9% specificity). Preoperatively, miR-155-5p, miR-223-3p, and miR-320a-3p were differentially expressed and targeted the forkhead box O and Hippo pathways (P < 0.01).

CONCLUSION

Our study demonstrates, for the first time, that plasma miR-320a-3p levels can be used as a preclinical biomarker for NEC in term newborns with CHD.

Key Words: MicroRNA-320a; Term newborns; Necrotizing enterocolitis; Congenital heart defects; Plasma biomarker; Quantitative real-time polymerase chain reaction

Core Tip: The high morbidity and mortality rates associated with necrotizing enterocolitis (NEC) highlight the critical need to identify specific and sensitive biomarkers for early detection and prevention. In this study we evaluated plasma levels of microRNA (miR)-155, miR-221, miR-223, miR-320a, and miR-451a as potential NEC biomarkers in term newborns with congenital heart defects (CHD). For the first time, plasma miR-320a-3p was identified as a predictive biomarker for the risk of NEC in term newborns with CHD at the preclinical stage, and the Hippo signaling pathway may be involved in the early stages of NEC pathogenesis.

INTRODUCTION

Necrotizing enterocolitis (NEC) is a severe multifactorial inflammatory gastrointestinal disease that affects infants in the neonatal period. Although NEC is more commonly associated with prematurity, congenital heart defects (CHD) increase the risk of developing NEC in both preterm and term newborns, with the type of CHD having a massive impact on the causality of NEC[1,2]. Newborns with complex CHD experience oxygenation fluctuations, hypoxia, and hemodynamic instability, necessitating cardiac surgery in the first weeks of life[3].

Several factors, including abnormal intestinal colonization by microorganisms, formula feeding, and ischemia, have been linked to the activation of the mucosal innate immune system in the early stages of NEC development[4]. Ischemia-induced reperfusion causes systemic inflammation, which includes cytokine release, upregulation of adhesion molecules, complement deposition, and immune cell activation and infiltration. Neutrophils, monocytes, and lymphocytes play important roles in inflammation, facilitating tissue damage through the secretion of pro-inflammatory cytokines, chemokines, and reactive oxygen species, which in the case of NEC leads to intestinal epithelial barrier damage[5].

The high morbidity and mortality rates associated with NEC highlight the urgent need to identify specific and sensitive biomarkers to enable early detection and preventive measures. Despite the general understanding of NEC pathophysiology, there are still gaps in the key molecular mechanisms underlying NEC in CHD infants, making it difficult to identify specific laboratory biomarkers.

Some promising key molecules in NEC pathogenesis include small non-coding regulatory RNAs, particularly microRNAs (miRs), which play a crucial role in modulating various biological processes such as inflammation, cellular development, oxidative stress, cell proliferation, differentiation, and apoptosis via post-transcriptional regulation of gene expression[6]. In recent years, miRs have emerged as promising noninvasive biomarkers for various diseases due to their stability in body fluids, varying levels under different conditions, and relatively simple detection[7,8].

In terms of NEC pathogenesis, several studies have investigated the levels of miRs in the intestinal tissues of preterm infants who had NEC surgery[9-11], as well as in plasma and fecal samples from preterm infants[12-14]. Several NEC-associated miRs, including miR-223, miR-451, miR-1290, miR-4725-3p, miR-431, miR-4793-3p, miR-21-3p, miR-132, miR-146b-3p, and miR-410 were identified in tissue samples when compared to controls; however, miR-223, miR-451, miR-431, miR-132, and miR-429 were also found to be associated with spontaneous intestinal perforation[10]. Subsequent investigation revealed that the plasma levels of miR-1290, miR-1246, and miR-375 were specific NEC biomarkers. Notably, miR-1290 demonstrated the ability to efficiently distinguish NEC cases from sepsis in preterm infants with a sensitivity of 0.83 and a specificity of 0.96[13].

There is a growing body of evidence suggesting toll-like receptor 4 (TLR4) to play a pivotal role in NEC pathogenesis, initiating a pro-inflammatory cascade resulting in nuclear factor kappa B (NF-κB) activation and the production of pro-inflammatory cytokines[4,15]. Furthermore, during the early reperfusion stage, NF-κB translocation to the nucleus causes cell dysfunction and death by promoting pro-inflammatory and proapoptotic gene expression[16]. MiRs may be involved in NEC pathogenesis through interactions with TLR4 (miR-31, miR-451, miR-203, miR-4793-3p), key transcription factors NFκB/NFκB2 (miR-203), forkhead box protein A1 (miR-21-3p, miR-431, miR-1290)[9,10], nuclear factor I-A (NFIA) (miR-223)[17], nucleotide-binding oligomerization domain-like receptor 3 (NLRP3) inflammasome (miR-146a-5p)[11], and receptor interacting protein kinase 1 (miR-141-3p)[18]. Other miRs, such as miR-155 and miR-221, control inflammation and immune response by modulating the TLR4 signaling pathway[19], whereas miR-320 strengthens intestinal barrier function in vitro, by acting as a negative regulator of pattern recognition receptor nucleotide-binding oligomerization domain-containing protein 2 (NOD2)[20,21], and may also contribute to NEC development.

To date, the role of miRs in NEC pathogenesis in term infants with CHD has remained largely unexplored. In this study, we investigated the plasma levels of miR-155-5p, miR-221-3p, miR-223-3p, miR-320a-3p, and miR-451a in term newborns with CHD both before NEC onset and at the time of NEC diagnosis to assess the prognostic value of these miRs and to gain a better understanding of their role in NEC pathophysiology.

MATERIALS AND METHODS

Study design and sample collection

This study included twenty-tree term newborns with a birth weight greater than 2500 g and a gestational age greater than 37 weeks who had duct-dependent CHD. The spectrum of CHD included: (1) Hypoplastic left heart syndrome; (2) Double-inlet left ventricle; (3) Unbalanced (RV dominant) atrioventricular septal defect; (4) Transposition of the great arteries; (5) Coarctation of the aorta; (6) Pulmonary atresia/stenosis; and (7) Aortic valve stenosis (Table 1). All of the study participants were born at the Almazov National Medical Research Centre in Saint-Petersburg, Russia.

Table 1 Distribution of congenital heart defects in the study cohort, n (%).

Type of congenital heart defect	Total	No-NEC	NEC	Type of cardiac surgery
				No-NEC (n)	NEC (n)
Single-ventricle CHD
Hypoplastic left heart syndrome	2	1 (4)	1 (4)	Damus-Kaye-Stansel procedure (1)	Norwood procedure (1)1
Double-inlet left ventricle + Pulmonary artery/CoAo	2	1 (4)	1 (4)	MBTS placement (1)	Norwood procedure (1)
Unbalanced (right ventricle dominant) atrioventricular septal defect + CoAo	1	1 (4)	0	Modified Amato technique (1)	0
Total	5	3 (13)	2 (9)
Two-ventricle CHD
Transposition of the great arteries	7	5 (22)	2 (9)	ASO (4). ASO + Modified Amato technique (1)	ASO (1). MBTS placement (in left ventricular outflow tract obstruction) (1)2
Coarctation of the aorta	5	3 (13)	2 (9)	Modified Amato technique (2). End-to-end anastomosis (1)	Modified Amato technique (1). End-to-end anastomosis (1)
Pulmonary atresia/stenosis	4	2 (9)	2 (9)	mBTS placement (1). Antegrade palliation (1)	MBTS placement (2)
Aortic valve stenosis	2	1 (4)	1 (4)	Aortic commissurotomy (1)	Aortic commissurotomy (1)
Total	18	11 (48)	7 (30)

¹Mortality related to acute heart failure in congenital heart defects (CHD).

²Mortality related to acute heart failure in CHD and necrotizing enterocolitis.

ASO: Arterial switch operation; CHD: Congenital heart defects; CoAo: Coarctation of the aorta; MBTS: Modified Blalock-Taussig shunt; NEC: Necrotizing enterocolitis

All enrolled neonates had cardiac surgery on a median day of life (DOL) = 7 (interquartile range: 6-8). Nine newborns developed NEC (Bell’s stages IIA and IIIA) within one week after cardiac surgery (hereafter, the NEC newborns) based on both abnormal abdominal radiography and clinical signs. The modified Bell staging criteria were used to determine the severity of the disease[22]. The median infant age at NEC diagnosis was 12 days (interquartile range: 5-16 days). The control group included newborns who did not develop NEC after cardiac surgery (hereafter, the no-NEC newborns).

There were no significant differences in terms of sex, gestational age, birth weight, mode of delivery, Apgar scores at 1 minute and 5 minutes, or type of CHD between NEC and no-NEC newborns during the perioperative period (Table 2).

Table 2 Characteristics of newborns with critical congenital heart defects enrolled in the study, n (%).

Newborns’ characteristic (n)		No-NEC (n = 14)	NEC (n = 9)	P value
					Sex
Female (6)	5 (36)	1 (11)	0.34
Male (17)	9 (64)	8 (89)
Gestational age (week) (23)		39 (39-40)	39 (38-39)	0.26
Birth weight (g) (23)		3345 (2960-3810)	3140 (3020-3580)	0.64
Mode of delivery	Unassisted vaginal delivery (15)	9 (64)	6 (67)	0.41
	Assisted vaginal delivery (1)	0	1 (11)
	Cesarean section (7)	5 (36)	2 (22)
Maternal factors	Chorioamnionitis (3)	3 (21)	0	0.22
	Prelabor rupture of membranes (2)	1 (7)	1 (11)
	Preeclampsia (0)	0	0
	Gestational diabetes (3)	3 (21)	0
Apgar score at 1 minute (23)		7 (7-8)	7 (7-7)	0.78
Apgar score at 5 minutes (23)		8 (8-8)	8 (8-8)	0.51
Intrauterine growth retardation (3)		1 (7)	2 (22)	0.54
Congenital heart defects	Cyanotic (15)	9 (64)	6 (67)	1.0
	Acyanotic (8)	5 (36)	3 (33)
Parameters before cardiac surgery
Feeding	Breast milk (7)	4 (29)	3 (33)	0.91
	Formula (7)	4 (29)	3 (33)
	Breast milk + Formula (9)	6 (43)	3 (33)
Feeding volume (mL/kg/day)		56.3 (28.0-80.8)	42.0 (34.1-72.7)	0.95
Superior mesenteric artery resistance index		0.99 (0.93-1.10)	0.95 (0.94-1.12)	0.79
Antibiotics at sample collection (Ampicillin + Sulbactam) (7)		3 (21)	4 (44)	1.0
Rashkind procedure/balloon valvuloplasty (6)		2 (14)	4 (44)	0.16
Prostaglandin E1 infusion (ng/kg/minute)		10 (10-20)	15 (10-25)	0.61
Inotropic therapy (3)		1 (7)	2 (22)	0.54
Parameters of cardiac surgery
Age at cardiac surgery (days)		7.5 (5.0-8.0)	4.0 (6.0-11.5)	0.72
Palliative surgery (10)		5 (36)	5 (56)	0.42
Duration of cardiac surgery (minute)		235.0 (212.5-347.5)	185.0 (161.0-353.0)	0.33
Aristotle score		7.9 (7.0-11.0)	7.0 (6.3-12.3)	0.50
Cardiopulmonary bypass (17)		10 (71)	7 (78)	1.0
Delayed sternal closure (11)		7 (50)	4 (44)	1.0
Parameters after cardiac surgery
Arterial hypotension on POD1 (13)		8 (57)	5 (56)	1.0
Heart rhythm disturbances on POD1 (9)		5 (36)	4 (44)	1.0
Maximum vasoactive-inotropic score during the first 24 hours after surgery		15.0 (12.6-31.3)	25.0 (7.5-38.8)	0.59
Duration of inotropic therapy (days)		4.0 (2.5-7.0)	10.0 (2.5-20.0)	0.17
Maximum arterial lactate concentration within 24 hours (mmol/L)		3.5 (2.7-5.1)	5.0 (2.7-9.6)	0.48
Arterial lactate concentration at 24 hours (mmol/L)		2.0 (1.8-2.6)	2.3 (1.5-6.5)	0.65
Start of enteral feeding on POD1		11 (79)	5 (56)	0.36
Start of enteral feeding (days)		1.0 (1.0-1.3)	1.0 (1.0-6.0)	0.14
Feeding volume on POD1 (mL/kg/day)		13.1 (6.91-16.0)	3.9 (0.0-10.8)	0.09
Feeding volume on POD3 (mL/kg/day)		33.9 (22.9-56.7)	1.0 (0.0-19.4)	0.004
NEC onset (days)		0	2.0 (1.0-5.0)	-
Intestinal wall thickening on POD1 by ultrasonography (9)		5 (36)	4 (44)	0.38
Pneumatosis intestinalis on POD1 by ultrasonography (4)		0	4 (44)	-
Neonatal intensive care unit duration (days)		6.5 (4.0-16.5)	14.0 (5.5-26.5)	0.17
Antibiotics at sample collection (20)		11 (cefoperazone + sulbactam: 5; cefuroxime: 2; cefotaxime: 1; ampicillin/sulbactam: 2; piperacillin/tazobactam: 1)	9 (cefoperazone + sulbactam: 3; cefuroxime: 1; ampicillin/sulbactam: 2; meropenem: 2; ertapenem: 1)	0.25

Data are n (%) or median (interquartile range). NEC: Necrotizing enterocolitis; POD: Postoperative day.

Blood samples were collected at two time points: (1) Before (median DOL = 5, interquartile range: 5-7); and (2) After (median DOL = 13, interquartile range: 12-15) cardiac surgery. In NEC newborns, the time point of blood sample collection after surgery coincided with the onset of NEC. Blood samples were drawn into EDTA tubes as part of the preoperative and postoperative diagnostic workup. Plasma was obtained from leftover blood after routine laboratory testing, separated within 8 hours of collection, by 2 centrifugation steps at 2000 × g for 15 minutes, and stored at -40 °C until use.

Routine newborn care

All newborns had suspected CHD during the prenatal period, which was later confirmed with postnatal echocardiography and computed tomography. Following birth, infants were admitted to the neonatal intensive care unit and received intravenous prostaglandin E1 infusion to manage duct-dependent circulation before cardiac surgery.

During the preoperative period, newborns were fed according to an internal protocol, that included starting enteral nutrition with a preterm formula (80 kcal/100 mL) and progressing to mixed feeding or feeding with the mother's breast milk (Table 2). After hemodynamic stabilization in the early postoperative period, enteral feeding with a hydrolyzed formula at 2 mL/hour was started if NEC symptoms were not present on ultrasound examination. Newborns with NEC symptoms were kept nil per os, and given total parenteral nutrition, and antibiotic therapy under national management and treatment guidelines[23].

There were no significant differences in potential confounding factors, including maternal factors, feeding practices, antibiotic use, and postnatal hemodynamic changes during the perioperative period (Table 2). The only factor that differed significantly was the feeding volume on postoperative day 3 (P = 0.004), which is expected considering that NEC developed on the median postoperative day 2.

RNA isolation

Total RNA was extracted from plasma samples using the TRIzol LS reagent (Thermo Fisher Scientific., Carlsbad, CA, United States), following the manufacturer’s instructions as described previously[24]. Briefly, 100 µL of plasma was mixed with 300 µL of TRIzol LS, which contained 1 µg of Escherichia coli tRNA (Sigma-Aldrich, St. Louis, MO, United States) and 10⁹ molecules of the synthetic oligoribonucleotide synt-cel-miR-39-3p (5′-ucaccggguguaaaucagcuug-3′) (Syntol, Moscow, Russia), which was identical to the mature cel-miR-39-3p of Caenorhabditis elegans (miRBase: MIMAT000001). The RNA precipitation was carried out with isopropanol and GlycoBlue as a coprecipitant (Thermo Fisher Scientific, Waltham, MA, United States). The RNA pellet was dissolved in 12 μL of RNase-free water (Ambion, Austin, TX, United States) and stored at −80 °C.

Reverse transcription and quantitative real-time polymerase chain reaction of miRs

TaqMan MicroRNA Assays were used to detect of the following miRs: (1) Hsa-miR-155 (assay ID: 002623); (2) Hsa-miR-221 (assay ID: 000524); (3) Hsa-miR-223 (assay ID: 002295); (4) Hsa-miR-320 (assay ID: 002277); (5) Hsa-miR-451 (assay ID: 001141); and (6) MiR-39-3p (assay ID: 000200). A reverse transcription reaction was carried out using the TaqMan MicroRNA Reverse Transcription Kit (Applied Biosystems, Waltham, MA, United States) following the manufacturer’s instructions. Reverse transcription was performed in the Veriti Thermal cycler (Applied Biosystems, Waltham, MA, United States), using the following program: (1) 30 minutes at 16 °C; (2) 30 minutes at 42 °C; and (3) 5 minutes at 85 °C. The obtained cDNA was stored at -40 °C until amplification. The qPCR was conducted using TaqMan Universal Master Mix II no UNG (Applied Biosystems, Waltham, MA, United States) following the manufacturer’s protocol on a LightCycler 480 II (Roche, Basel, Switzerland) with the following program: (1) 95 °C for 10 minutes; (2) Followed by 40 cycles of 95 °C for 15 seconds; and (3) 60 °C for 1 minute. The raw data were normalized using the external synthetic control cel-miR-39-3p as described previously[24]. The relative expression of target miRs was calculated using the 2^−ΔΔCt method.

In silico functional analysis of the miRs network

Target-based functional analysis of miR-155-5p, miR-223-3p, miR-320a-3p, and miR-451a was conducted using the DIANA-miRPath v4.0 (TarBase v.8.0) (http://62.217.122.229:3838/app/miRPathv4) database with term annotation sources from the Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Ontology (GO) databases[25].

Statistical analysis

Continuous variables are represented as medians and interquartile ranges, whereas categorical variables are represented as numbers and percentages. For continuous variables, the Mann–Whitney U test was used to calculate differences between the no-NEC and NEC groups, while the Wilcoxon signed-rank test determined differences within groups before and after cardiac surgery. Categorical variables were compared using the two-tailed Fisher’s exact test and the Kruskal-Wallis analysis of variance test. Receiver operating characteristic (ROC) analysis was conducted and the area under the curves (AUC) was calculated to assess the diagnostic accuracy of the miRs for NEC. The Youden index was used to determine the optimal cutoff value. A power analysis was carried out to determine the likelihood of detecting a statistically significant difference between the groups at a 5% significance level, given the sample size and assumed effect size. The nonparametric Spearman rank test was used to determine the correlation between plasma miR levels and routine laboratory parameters. All statistical analyses were conducted using Statistica 10.0 (StatSoft, Tulsa, OK, United States), MedCalc 23.0.2 (MedCalc Software Ltd, Ostend, Belgium), Prism 9 (GraphPad, San Diego, CA, United States), and G*Power 3.1.9.7 (Heinrich Heine University Dusseldorf, Dusseldorf, Germany); P < 0.05 was considered statistically significant.

RESULTS

Clinical laboratory parameters

There were no differences in clinical laboratory parameters between the two groups of neonates prior to and following cardiac surgery. However, no-NEC newborns had significantly lower hemoglobin and lymphocyte levels, as well as higher segmented neutrophil and C-reactive protein (CRP) levels postoperatively (Table 3). In contrast, none of the studied routine laboratory parameters changed statistically after cardiac surgery in NEC newborns.

Table 3 Laboratory parameters of term newborns with critical congenital heart defects before and after cardiac surgery.

Laboratory parameters		No-NEC (n = 14)	P value	NEC (n = 9)	P value	P value between groups
Hemoglobin (g/L)	Before	155.0 (141.0-175.5)	0.003	159.0 (128.5-180.0)	0.050	0.847
	After	129.5 (113.0-146.0)		120.0(109.0-143.0)		0.705
Erythrocytes (× 1012/L)	Before	4.68 (4.24-5.00)	0.110	4.64 (3.70-5.15)	0.208	0.787
	After	4.11 (3.74-4.64)		3.91 (3.55-4.87)		0.900
Platelets (× 109/L)	Before	345.5 (280.0-375.5)	0.583	291.0 (243.5-424.0)	0.674	0.758
	After	372.0 (227.0-501.0)		298.0 (232.0-421.0)		0.244
Leukocytes (× 109/L)	Before	13.65 (11.20-15.65)	0.724	11.45 (8.95-15.35)	0.327	0.375
	After	12.70 (11.70-15.00)		11.70 (10.40-19.40)		0.801
Band neutrophils (%)	Before	1.0 (1.0-1.5)	0.624	1.5 (1.0-2.0)	0.686	0.512
	After	1.0 (0.0-3.0)		1.0 (1.0-2.0)		0.395
Segmented neutrophils (%)	Before	40.0 (35.5-49.5)	0.023	42.5 (40.0-58.0)	0.779	0.217
	After	60.0 (52.0-65.0)		48.0 (44.0-58.0)		0.078
Eosinophils (%)	Before	5.0 (3.0-6.5)	0.103	4.5 (3.6-6.0)	0.447	0.969
	After	3.0 (1.0-5.0)		6.0 (0.0-10.0)		0.314
Lymphocytes (%)	Before	40.0 (36.0-44.5)	0.045	37.0 (25.5-40.5)	0.484	0.247
	After	20.5 (15.0-30.0)		25.0 (16.0-30.0)		0.825
C-reactive protein (mg/L)	Before	7.0 (4.0-10.5)	0.034	1.5 (0.8-5.0)	0.463	0.101
	After	8.2 (5.0-15.0)		6.1 (3.0-18.0)		0.550

Data are median (interquartile range). NEC: Necrotizing enterocolitis.

Levels and prognostic values of the studied plasma miRs in the perioperative period

In the preoperative period, NEC newborns had significantly lower levels of circulating miR-155, miR-223, and miR-320a than no-NEC infants. Specifically, the median level of plasma miR-155 was reduced 2.70-fold (P = 0.020), miR-223 was reduced 2.42-fold (P = 0.030), and miR-320a levels decreased 3.62-fold (P = 0.006) when compared to no-NEC infants (Figure 1A). In the postoperative period, the miR-451a level was the only one that differed significantly between the newborn groups, with a 4.70-fold decrease (P = 0.014) observed in NEC newborns. The levels of miR-221 remained consistent between the two newborn groups throughout the entire perioperative period.

Figure 1 Diagnostic potential of the studied microRNAs in the plasma samples of term newborns with critical congenital heart defects in distinguishing the risk of necrotizing enterocolitis development. A: The relative expression levels of target microRNAs (miRs); B: Receiver operator characteristic (ROC) curve for each target miR in the preoperative period; C: ROC curve of each target miRs in the postoperative period; D: ROC curve of the combined three miRs (miR-155 + miR-223 + miR-320a) vs miR-320a alone for predicting necrotizing enterocolitis (NEC) development before clinical symptoms; E: ROC analysis statistics for the studied miRs during the perioperative period. No-NEC_BS and No-NEC_AS represent no-NEC newborns before and after cardiac surgery, NEC_BS and NEC_AS represent NEC newborns before and after cardiac surgery, respectively. AUC: Area under the curve; MiRs: MicroRNAs; NEC: Necrotizing enterocolitis.

Additionally, the miR-223 level in the plasma samples of no-NEC newborns before surgery was significantly higher than that of NEC newborns following surgery. Besides, the miR-320a level in plasma samples from no-NEC newborns after surgery was significantly higher than that of NEC newborns before surgery (Figure 1A). Interestingly, the levels of any studied miRs did not differ statistically in no-NEC newborns during the perioperative period, but, at the same time, the plasma level of miR-451a in NEC newborns after surgery was significantly lower than before surgery, as well as compared to no-NEC newborns.

The ROC analysis revealed that miR-155, miR-223, and miR-320a detected newborns at risk of developing NEC with AUC values of 0.808, 0.788, and 0.856, respectively (Figure 1B). The optimal cut-off values for NEC prediction were found to be 0.172, 0.103, and 0.081 for miR-155, miR-223, and miR-320a, respectively, with sensitivities of 87.5%, 75.0%, and 62.5% and specificities of 76.9%, 92.3%, and 100.0%. Surprisingly, at the onset of NEC, only miR-451a showed promise as a valuable biomarker of disease, with an AUC of 0.835 and sensitivity and specificity values of 85.7 and 76.9, respectively (Figure 1C).

Although miR-155, miR-223, and miR-320a had promising AUC values, indicating that they could be used as prognostic biomarkers for NEC before clinical symptoms appeared, a logistic regression-based combined biomarker panel performed no better than individual miRs. The combined panel had an AUC of 0.856, which was identical to miR-320a alone (Figure 1D). Figure 1E shows the ROC analysis statistics for the investigated miRs during the perioperative period.

The power analysis revealed that the sample size of 14 no-NEC and 9 NEC newborns had 80%-90% power to identify differences in miR expression levels between the groups using pre-defined cut-off values at a 5% significance level.

Correlation analysis

Correlation analysis revealed that the two newborn groups had distinct miR interaction profiles. Both groups showed only positive correlations between different miRs. No-NEC newborns showed moderate-strength correlations between miRs, which expanded and became stronger in the postoperative period (Figure 2). Conversely, NEC newborns had fewer miR-miR interactions, but these interactions were stronger and decreased in number postoperatively. Particularly, the correlation network of miR-223 and miR-320a in the preoperative period differed between the two newborn groups. The miR-221 showed no correlation with other miRs in NEC newborns during the perioperative period. Surprisingly, miR-451a lost its correlation with other miRs in NEC newborns postoperatively, despite being the only miR that could be used as a biomarker of NEC based on AUC value postoperatively.

Figure 2 Correlation heatmap between the plasma microRNA levels and clinical laboratory parameters of term newborns with critical congenital heart defects across the perioperative period. The color scale bars represent nonparametric Spearman correlations. The red color indicates a strong positive correlation, while purple color indicates a strong negative correlation. Only significant correlations are shown (P < 0.05). No necrotizing enterocolitis (NEC)_BS and No-NEC_AS represent newborns without NEC before and after cardiac surgery; NEC_BS and NEC_AS represent newborns, who developed postoperative NEC, before and after cardiac surgery, respectively. CRP: C-reactive protein; MiRs: MicroRNAs; NEC: Necrotizing enterocolitis.

The correlation analysis of miRs with routine laboratory parameters varied between infant groups (Figure 2). Preoperatively, no-NEC newborns had positive moderate-strength correlations between miR-155 and miR-223 with band neutrophil counts, and miR-221 with leukocyte counts. In contrast, NEC newborns showed a correlation pattern for miR-451a: (1) A strong negative correlation with segmented neutrophil counts; and (2) A strong positive correlation with lymphocyte counts (Figure 2). Postoperatively, no significant correlations were found in no-NEC newborns, whereas miR-155 and miR-451a had strong negative correlations with platelet counts and with leukocyte counts in NEC newborns. There was no correlation between miRs and inflammatory markers such as CRP concentration.

Biological processes regulated by the studied miRs

KEGG-pathway enrichment analysis using the pathways union method in DIANA-miRPath v4.0 revealed no shared biological processes involving all four differentially expressed miRs. Three potential NEC prognostic biomarkers, miR-155, miR-223, and miR-320a, were discovered to be simultaneously involved in two biological processes: (1) The forkhead box O (FoxO) signaling pathway (with 6 genes regulated by miR-223, 18 by miR-155, and 20 by miR-320a); and (2) Prostate cancer (Table 4). Remarkably, the combination of miR-155 and miR-320a, which had higher AUC values in the preoperative period, targeted genes in eight biological processes, including the Hippo signaling pathway, various cancer pathways, and protein processing in the endoplasmic reticulum, whereas the combination of miR-223 and miR-320a targeted genes in two processes (Table 4). Interestingly, no KEGG pathways were enriched by genes targeted by miR-451a alone or in combination with other studied miRs, after false discovery rate correction.

Table 4 Kyoto encyclopedia of genes and genomes and gene ontology pathways regulated by the differentially expressed microRNAs.

MiRs	Term name	Term genes (n)	P value	False discovery rate	Target genes
Kyoto Encyclopedia of Genes and Genomes (pathways union)
MiR-155-5p miR-223-3p, miR-320a-3p	Forkhead box O signaling pathway	139	7.31e-10	1.66e-8	18 by miR-155-5p, 6 by miR-223-3p, 20 by miR-320a-3p
	Prostate cancer	101	5.77e-8	7.84e-7	15 by miR-155-5p, 4 by miR-223-3p, 15 by miR-320a-3p
MiR-155-5p, miR-320a-3p	Hippo signaling pathway	164	1.20e-10	4.07e-9	21 by miR-155-5p, 31 by miR-320a-3p
	Viral carcinogenesis	265	1.07e-10	4.07e-9	29 by miR-155-5p, 42 by miR-320a-3p
	Pathways in cancer	555	8.83e-9	1.50e-7	53 by miR-155-5p, 60 by miR-320a-3p
	Transcriptional misregulation in cancer	206	1.13e-7	1.28e-6	21 by miR-155-5p, 32 by miR-320a-3p
	Colorectal cancer	88	3.43e-7	3.33e-6	15 by miR-155-5p, 15 by miR-320a-3p
	Chronic myeloid leukemia	79	5.08e-6	2.88e-5	12 by miR-155-5p, 14 by miR-320a-3p
	Hepatitis B	177	8.83e-6	4.62e-5	22 by miR-155-5p, 21 by miR-320a-3p
	Protein processing in endoplasmic reticulum	194	2.04e-4	4.91e-4	20 by miR-155-5p, 22 by miR-320a-3p
MiR-223-3p, miR-320a-3p	Signaling pathways regulating pluripotency of stem cells	156	1.67e-6	1.26e-5	5 by miR-223-3p, 22 by miR-320a-3p
	Longevity regulating pathway	105	5.56e-5	1.89e-4	4 by miR-223-3p, 14 by miR-320a-3p
Gene Ontology (genes intersection)
MiR-155-5p, miR-223-3p, miR-320a-3p	Cytokine-mediated signaling pathway	319	2.93e-3	2.63e-2	2 by miR-155-5p, 2 by miR-223-3p, 2 by miR-320a-3p

MiRs: MicroRNAs.

GO enrichment analysis using the genes intersection method revealed a common pathway, the cytokine-mediated signaling pathway, which includes all three potential preclinical NEC biomarkers: (1) MiR-155; (2) MiR-223; and (3) MiR-320a (Table 4). Two genes in this pathway, interleukin (IL)-6 cytokine family signal transducer (IL6ST) and POU Class 2 Homeobox 1 (POU2F1) were discovered to be common target genes for these miRs.

DISCUSSION

There is currently limited information available on the role of miRs in NEC pathogenesis in term newborns with CHD. To the best of our knowledge, this is the first study to report plasma miR expression profiles as potential early biomarkers for NEC in these patients. We investigated miRs that may target key receptors (TLR4), transcription factors [NF-κB and signal transducer and activator of transcription 3 (STAT3)], and downstream effector genes within functional networks related to inflammation, angiogenesis, apoptosis, cell proliferation/migration, cell adhesion/chemotaxis, barrier function, and hypoxia/oxidative stress. Specifically, we investigated the diagnostic potential of miR-155-5p, miR-221-3p, miR-223-3p, miR-320a-3p, and miR-451a in term newborns with CHD, both before the onset of clinical NEC symptoms appeared and after NEC was diagnosed. Our findings revealed that plasma expression of miR-155-5p, miR-223-3p, and miR-320a-3p was lower in NEC newborns than in no-NEC newborns prior to NEC onset; however, only miR-451a showed altered expression when clinical NEC symptoms appeared. ROC analysis identified miR-320a as the most effective predictive biomarker for NEC (AUC = 0.835, 63 % sensitivity, 100 % specificity at the cut-off value ≤ 0.081), while miR-451a emerged as a potential biomarker during NEC diagnosis.

Altogether, the downregulation of miR-155-5p, miR-223-3p, and miR-320a-3p may contribute to an increase in pro-inflammatory signals mediated by TLR4 and the key transcription factor NF-κB. This can exacerbate the inflammatory response by producing pro-inflammatory cytokines and chemokines, ultimately leading to extensive mucosal injury and NEC.

Indeed, miR-155 plays a major role in regulating inflammation and immune response by modulating the TLR4 signaling pathway, acting as either a pro- or anti-inflammatory factor based on the cellular environment[19]. MiR-155 blocker reduces inflammation and improves intestinal barrier function in septic mice by inhibiting NF-κB signaling[26]. MiR-155 serves as a pro-inflammatory mediator by elevating the expression of cytokines such as IL-6 and tumor necrosis factor alpha (TNF-α) via suppressor of cytokine signaling 1 suppression[27]. Inhibiting miR-155 reduced inflammation in a lipopolysaccharide (LPS)-treated epithelial cell line model and NEC rat intestinal tissues by suppressing IκBα/NF-κB p65 signaling and lowering TNF-α and IL-6[28]. Elevated miR-155 levels stimulate pro-inflammatory activity in patients with inflammatory bowel disease (IBD) and worsen experimental colitis in mice[29,30]. Despite the established role of miR-155 in inflammatory diseases, our research found no significant difference in postoperative miR-155 levels between no-NEC and NEC newborns exhibiting clinical NEC symptoms. However, miR-155 levels were substantially elevated in no-NEC newborns vs NEC newborns before NEC onset during the preoperative period (P = 0.020), indicating a potential anti-inflammatory role in the early stages of NEC pathogenesis, possibly through the suppression of the myeloid differentiation factor 88 and/or transforming growth factor-β-activated kinase 1-binding protein 2 adapters within the LPS-activated TLR4 signaling pathway[27].

MiR-223 has been shown to regulate immune responses by modulating macrophage function, dendritic cell maturation, neutrophil activation and infiltration, inflammasome activation, and maintaining innate immunity balance to prevent excessive inflammation[31]. Downregulation of miR-223 in TLR-stimulated macrophages increases pro-inflammatory cytokines IL-6 and IL-1β by interacting with STAT3; while upregulation of miR-223 inhibits NLRP3, leading to decreased inflammation and reduced macrophage activation[31]. Overexpression of miR-223 has been observed in the intestinal tissues of preterm infants with NEC, implying a role in inflammation and tissue damage via the miR-223/NFIA axis and establishing its importance in inflammatory conditions[17]. In contrast, in our study, plasma miR-223 levels were higher in no-NEC newborns than in NEC newborns during the perioperative period, indicating its role in innate immunity control and subsequent inflammation mitigation in NEC pathogenesis in term infants with CHD. This is consistent with miR-223's anti-inflammatory functions, such as promoting M2 macrophage polarization, inhibiting neutrophil activation and neutrophil trap formation[31].

In our study, miR-320a was identified as the most promising early biomarker for NEC development. Several investigations have demonstrated the potential of miR-320a as a non-invasive biomarker for the diagnosis and monitoring of IBD[20,21], and plasma elevated miR-320 levels may serve as a surrogate marker of mucosal inflammation, including subclinical intestinal inflammation, in asymptomatic patients[20,21]. Elevated miR-320a levels in the quiescent colonic mucosa of Crohn’s disease patients and ulcerative colitis suggested a role in the mucosal sensitization to environmental factors, allowing for differentiation between active and remission stages[32]. Furthermore, elevated miR-320a-3p levels have been linked to markers of hypoperfusion, such as blood lactate concentration, as well as an increased risk of mortality in cardiogenic shock[33]. It is known that miR-320a controls multiple target genes across different biological pathways, including the canonical Wnt/β-catenin signaling pathway[34], which controls intestinal stem cells, promotes Paneth cell formation, and is downregulated by TLR4 activation in experimental NEC[35]; the phosphoinositide 3-kinase/protein kinase B/mammalian target of rapamycin signaling pathway[36], involved in NEC pathogenesis through the control of oxidative stress and inflammation[37]; and the vascular endothelial growth factor signaling pathway[38]. In vitro and in vivo studies have shown that miR-320a strengthens intestinal barrier integrity by stabilizing the tight junction complex in intestinal cells[20,21], and negatively regulates NOD2, a key cytosolic receptor in inflammatory responses, thereby facilitating immune tolerance to bacteria in the intestine[39]. Furthermore, it has been found that under hypoxic conditions, the overexpression of miR-320a, induced by hypoxia-inducible factor (HIF-1α), promotes barrier formation in intestinal epithelial cells[40]. Therefore, in the context of CHD, circulatory hypoxia activates HIF-1α in intestinal epithelial cells, leading to increased miR-320a expression and improved barrier integrity. Consistent with this, our study discovered that low preoperative plasma miR-320a levels in term newborns with CHD were linked to subsequent postoperative NEC development. These findings suggest that the miR-320a cut-off level determined in our study may represent a threshold for intestinal barrier function to become critically affected.

The activity of individual miRs can be modulated by the broader miRs network, resulting in diverse effects, including increased risk of NEC development. Functional analysis of the preoperatively differentially expressed miRs in no-NEC vs NEC newborns showed potential regulatory patterns of the Hippo signaling pathway (via miR-155 and miR-320a) and the FoxO signaling pathway (via miR-155, miR-223, and miR-320a) involved in NEC pathogenesis. Additionally, miR-155, miR-223, and miR-320a have two common gene targets: (1) IL6ST; and (2) POU2F1. IL6ST, also known as gp130, is a transmembrane protein that initiates STAT3 signaling when phosphorylated resulting in the activation of pro-inflammatory gene expression[41] and can activate the Hippo signaling pathway[42]. POU2F1 is a transcription factor that controls the expression of many genes involved in development, differentiation, and stress response[43]. The IL-6/IL-6R/STAT3 signaling pathway plays a crucial role in modulating neutrophil activity, apoptosis, and protease activation, making it a key regulator of inflammation[44]. The Hippo signaling pathway, a negative regulator of both cell proliferation and apoptosis, plays an essential role in regulating the innate immune response via positive or negative modulation of NF-κB signaling, given the molecules involved[45], as well as in maintaining intestinal homeostasis and promoting regeneration after injury[46]. Yes-associated protein (YAP) and transcriptional co-activator with PDZ-binding motif (TAZ), both downstream effectors of the Hippo signaling pathway, have been linked to epithelial cell polarity[47], alveolar epithelial repair, and fibrotic remodeling prevention[48], adherens junction dynamics[49], and epithelial integrity[50]. In silico functional analysis revealed that miR-320a targets 31 genes within the Hippo signaling pathway, including three regulators, the large tumor suppressors large tumor suppressor (LATS) 1 and LATS2, which encode LATS1/2, and WWTR1, which encode TAZ. Within the signaling cascade, LATS1/2 kinases phosphorylate YAP and TAZ, causing nuclear exclusion and degradation in the proteasome[51]. Recent research has shown that suppressing YAP/TAZ activity in alveolar epithelial cells during lung injury causes pathological alveolar remodeling, loss of cell differentiation, and increased neutrophilic inflammation[48]. Furthermore, Hippo signaling modulates the FoxO signaling pathway by interacting with YAP, a nuclear co-factor, and the forkhead transcription factor FoxO1, which has a negative impact on the antioxidant response and cardiomyocyte survival during ischemia/reperfusion[52]. Components of the FoxO signaling pathway, such as FoxO3a, were found to be involved in IBD pathogenesis via negative regulation by miR-223[53], and FoxO3a overexpression reduced inflammation by inhibiting NLRP3 in intestinal epithelial cells in a mouse model[54]. Therefore, given that low preoperative levels of miR-320a were associated with NEC development and high levels of miR-320a enhanced barrier formation in intestinal epithelial cells, we hypothesized that miR-320a inhibits LATS1/2 kinases within the Hippo signaling pathway, resulting in increased YAP activity that promotes epithelial repair and integrity and reduces local inflammation, possibly through decreased neutrophil recruitment. Moreover, YAP interacts with and stabilizes HIF-1α in the nucleus[55], enhancing transcriptional activity and increasing miR-320a expression. This creates a positive feedback loop that may protect against NEC development. Thus, the Hippo signaling pathway appears to play a role in the early stages of NEC pathogenesis, but further investigation is needed.

Our findings revealed that, of the five miRs studied, only plasma miR-451a levels showed a significant decrease in expression during the clinical stages of NEC. In contrast to our findings, miR-451 and miR-223 were upregulated in the small bowel tissues of preterm newborns with NEC and newborns with spontaneous intestinal perforation[10], as well as in fecal samples of preterm newborns at the onset of NEC[14]. Current understanding suggests that the underlying processes involved in NEC development may differ between term and preterm newborns, implying that associated biomarkers could also vary. MiR-451 is a critical regulator of numerous genes and biological signaling pathways, influencing processes such as cell proliferation, angiogenesis, apoptosis, epithelial-mesenchymal transition, invasion, and migration, as well as embryonic development, hematopoietic system differentiation, epithelial cell differentiation and polarity[56]. MiR-451a has been linked to erythropoiesis and the modulation of cytokine production by dendritic cells[57,58]. Findings revealed that miR-451 upregulation inhibited neutrophil chemotaxis by downregulating p38 mitogen activated protein kinase phosphorylation, and systemic administration of miR-451 significantly reduced neutrophil migration in a rheumatoid arthritis mouse model[56,59]. The anti-inflammatory effects of miR-451 were demonstrated by inhibiting NLRP3-induced pro-inflammatory cascades[60] and preventing LPS-induced TLR4 signaling[61] in microglia-mediated neuroinflammation. Furthermore, it has been established that NLRP3 plays a critical role in maintaining the epithelial barrier, and reducing gut inflammation[60]. Although miR-451a does not directly interact with inflammasome signaling, it targets macrophage migration inhibitory factor[62], preventing NLRP3 inflammasome activation[63] and decreasing TLR4 expression[64], which may limit intestinal inflammation during NEC pathogenesis in term newborns with CHD, but further functional investigations are needed to elucidate its specific involvement.

MiR-221, a highly conserved miRNA, plays a role in cancer and inflammatory diseases like rheumatoid arthritis and atherosclerosis by regulating cytokine production, cell proliferation, cell cycle, and migration[65]. Similar to miR-155-5p, miR-221-3p is regulated by TLRs and plays an important role in TLR4-mediated immune responses to LPS[66,67]. However, our study discovered no evidence of miR-221 involvement in NEC pathophysiology in term newborns with CHD.

The current study has a few limitations. The small sample size, particularly the limited number of NEC cases (n = 9), may have an impact on the generalizability of the findings. Additional large-scale, multicenter studies are required to confirm the diagnostic and prognostic potential of the identified miRs. Furthermore, in vitro and/or in vivo functional analyses are required to determine the involvement of miR-320a-3p in the regulation of the FoxO and Hippo signaling pathways, as well as its role in NEC pathogenesis. Furthermore, it is essential to investigate whether miR-451a, which was identified as a diagnostic biomarker for NEC in our study, is specific to NEC or common to other inflammatory disease processes in the early neonatal period.

CONCLUSION

This is the first study to report plasma miR expression profiles as early biomarkers of NEC risk in term newborns with CHD. MiR-320a could serve as a promising predictive biomarker for NEC in term newborns with CHD at the preclinical stage (AUC = 0.835, 63% sensitivity, 100% specificity at the cut-off value ≤ 0.081), while miR-451a emerged as a potential biomarker during NEC diagnosis (AUC = 0.835, 85.7% sensitivity, 76.9% specificity at the cut-off value ≤ 0.108). In silico functional analysis of preoperatively expressed miRs (miR-155-5p, miR-223-3p, and miR-320a-3p) provided new insights into early NEC pathogenesis, implying the involvement of the FoxO and Hippo signaling pathways and emphasized the need for additional research to provide deeper insights into their roles in NEC pathogenesis.

ACKNOWLEDGEMENTS

We would like to thank doctors of the Perinatal Centre of Almazov National Medical Research Centre, who treated and cared for newborns, and neonatal nurses for help in the investigation.