Identification and verification of feature biomarkers associated with immune cells in neonatal sepsis

Sepsis is a life-threatening organ dysfunction caused by a dysregulated host response to infection, mainly manifested as an inflammatory response and immunosuppression, and is currently the main cause of death in critically ill patients worldwide [1]. In the US, the present incidence of sepsis is approximately three per thousand, and severe sepsis kills at least 200,000 people annually [2]. Severe sepsis and septic shock account for 30–50% of hospital-reported deaths around the world [3]. Neonatal sepsis (NS) refers to bacteraemia with systemic infection occurring within the first month of life [4]. It is the most common cause of neonatal death, and its associated mortality is currently a major health concern worldwide [5]. NS can be divided into early- and late-onset, with 72 h after birth as the demarcation between the two. Neonatal infections account for an estimated 26% of under-five deaths [6]. In low- and middle-income countries, the reported incidence of NS in 2022 was 17.7% (5425/30577) and the mortality rate was 16.2% (877/5425) [7]. Development of primary and secondary prevention strategies based on different types of infections has become a hot area of NS-related research in recent decades [8].

Immune and inflammatory responses play important roles in the pathogenesis of NS. Currently, the commonly used clinical inflammatory markers are interleukin-6 (IL-6), C-reactive protein (CRP), and procalcitonin (PCT). IL-6 is a cytokine produced by mononuclear phagocytes, endothelial cells, fibroblasts, and decidual, chorionic, amniotic, and trophoblast cells upon stimulation with microbial products [9]. CRP, a protein synthesized in the liver, is currently used as an important biomarker to assess the severity and prognosis of NS [10]. PCT is produced by the parathyroid and neuroendocrine cells and acts as a precursor of calcitonin, which was formally proposed as a diagnostic marker for NS [11‐13] in 2008 and can increase more than 1000-fold during active infection. However, these are not ideal for the diagnosis and prognosis of NS [14]. In the early stages of NS, various immune cells (such as monocytes and macrophages) and released inflammatory mediators and cytokines can induce an excessive inflammatory response, whereas in the late stage, immunosuppression is predominant [15, 16]. Exploring the link between immune responses in the pathogenesis of NS, elucidating the molecular mechanisms involved, and identifying potential therapeutic targets will be of great significance in clinical practice.

Bioinformatic analysis helps to understand the underlying mechanisms of NS by screening gene expression datasets. In the present study, differentially expressed genes (DEGs) between NS and healthy controls were identified through bioinformatic analysis, and the underlying pathology of NS was explored through detection of the immune microenvironment, clustering, and protein–protein network analysis. In addition, we constructed a diagnostic model of six identified DEGs using least absolute shrinkage and selection operator (LASSO) regression analysis. Finally, we confirmed the effectiveness of the diagnostic model of immune-related genes (IRGs) using the GSE25504 dataset. In the present study, we explored the pathogenesis of NS from the perspectives of immunity and inflammation, which can identify potential targets for treating NS.

NS, a life-threatening condition, can lead to microcirculatory disturbances, immune dysfunction, and tissue and organ dysfunction, and is becoming the most common cause of neonatal death worldwide [4]. Hence, NS and its related mortality and complications represent a major global health concern [2‐6].

Impaired inflammatory immune responses during the onset and recovery phases are considered a hallmark of severe NS. Abnormal activation of macrophages and neutrophils occurs in the early stage of NS [17], and the recovery period is mainly characterized by immunosuppression. Sepsis is characterized by upregulation of CD4 + and CD8 + T cells, T helper 17 cells, and regulatory T cells [16], lymphopenia, and loss of immune function. Microarray analysis has indicated abnormalities in the expression of immune-related genes in children with sepsis, including FYN, FBL, ATM, WDR75, FOXO1, and ITK [18]. Alterations in gene expression related to innate immunity have also been reported in NS [19, 20]. The innate immune response in NS is driven by genes involved in innate immunity, such as IL1R2, ILRN, and SOCS3 [21]. The risk of developing NS is also associated with polymorphisms in exon 1 of mannose-binding lectin and Toll-like receptor 4 [22]. Based on the immunomodulatory effects of rhIL-7 in sepsis [23], targeting T cell immunometabolism in early or late sepsis has great therapeutic potential [16]. However, the pathogenesis of NS has not yet been fully established and needs further understanding.

In the present study, bioinformatic analysis and GSEA of DEGs in the merged dataset showed significant enrichment of immune and inflammatory responses, including acute inflammatory response, bacterial detection (including coagulase-negative Staphylococcus, Enterococcus species, et al. [19, 24]), and regulation of macrophage activation, which play important roles in the pathogenesis of NS. Most infiltrating immune cells were significantly different in patients with NS compared to the control group; activated CD8 + T and B cells, CD56 natural killer cells, naïve dendritic cells, and T helper cells were significantly enriched in the sepsis group, whereas activated dendritic cells, memory CD8 + T cells, macrophages, plasmacytoid dendritic cells, and neutrophils were significantly enriched in the control group. GSEA of IRGs showed that their functions were significantly enriched in the activation of immune response, macrophage, and the regulation of T cells. The diagnostic model of DEG containing five genes (PROS1, TDRD9, RETN, LOC728401, and METTL7B) and that of IRG with one gene (NSUN7) were constructed using LASSO algorithm, and their diagnostic performance verified by correlation and logical analyses showed good area under the curve (AUC) scores. Additionally, the DEG and IRG models were verified in the GPL6947 and GPL13667 sub-datasets, respectively. The IRG model performed better than the DEG model. The IRG model contained only NSUN7 suggesting that this gene may be important for the diagnosis and treatment of NS. Finally, statistical analysis of the validation datasets suggested that the risk scores may be related to gestational age and birth weight, regardless of sex.

Current knowledge of human B and T cells in sepsis is sparse, discordant, and at variance with findings reported from animal models. Our research find the activated B cell and activated CD8 T cells showed lesser expression in sepsis cases compared to control. These data are in agreement with those published in previous studies. Hotchkiss et al. [25] demonstrated that patients with sepsis show a severe B-cell deficiency. Monserrat et al. [17] pointed that B-cell lymphopenia affects the B-cell subsets heterogeneously, with marked reduction of CD19 + CD23 + B cells (activated regulatory B cells) and CD19 + CD5 + B cells (natural responder B-1a cells), but with normal numbers of CD19 + CD69 + early activated B cells. Similar findings were reported by other groups [26]. Meanwhile it is established that septic shock is associated with a severe exhaustion and depletion of T lymphocytes [27]. So the present results establish an association between decreased lymphocytes and sepsis but do not establish causality between lymphocyte apoptosis and outcome in patients with sepsis, which required further investigation.

Sun RNA methyltransferase 7 (NSUN7) belonging to the methyltransferase superfamily is located on chromosome 4p14 and consists of 12 exons and 718 amino acids. It reduces protein activity and motility of sperms and is associated with male infertility [28]. High expression of NSUN7 is associated with shortened survival in low-grade gliomas [29]. The overall survival in Ewing sarcoma is significantly associated with NSUN7 immunoreactivity, an independent favourable prognostic marker [30]. NSUN7 may also serve as a pivotal biomarker for predicting biochemical recurrence in patients with prostate cancer [31]. An increase in the mean precursor strength of plasma protein polypeptides, such as NSUN7, is associated with sepsis [32]. NSUN7 may also be associated with psychiatric disorders, including schizophrenia, bipolar disorder [33], and major depressive disorders. In eukaryotes, the NSUN family is the major RNA m5C modifying enzyme and includes seven family members (NSUN1–7). The biological function and significance of RNA m5C modification in maintaining mRNA stability is essential during early embryonic development and in the post-embryonic immune system. NSUN7 has been systematically studied in male sperm motility, but its mechanism of action in tumours and sepsis has not been elucidated. In the present study, NSUN7 expression was up-regulated in the NS group. Combined with bioinformatic analyses, NSUN7 may be used as a biomarker for the pathogenesis of NS.

Resistin (RETN), located on chromosome 19p13.2, encodes an anti-retro-transcriptional protein and belongs to the resistance protein-like gene family. Its encoded protein, a 114 amino acid polypeptide (12.5 kDa) hormone, is secreted by adipocytes and is a member of the cysteine-rich small secreted protein gene family [34, 35]. RETN activates monocytes and macrophages and induces the release of proinflammatory cytokines including lipopolysaccharides, IL-1, IL-6, and tumour necrosis factor (TNF)-α [36‐38]. RETN promotes endothelial cell activation and smooth muscle cell proliferation [39]. Elevated RETN levels have been reported in sepsis samples [40‐43]. Clinical observations have indicated that plasma RETN levels are highly correlated with the levels of inflammatory markers, such as CRP and IL-6 [44]. Additionally, RETN increases endothelial cell permeability, thereby promoting the adhesion and infiltration of endothelial cells and monocytes. RETN also mediates immunosuppression, directly suppresses neutrophil function, and is associated with poor outcomes in sepsis [45]. These findings suggest a link between RETN, immunity, and inflammation. In the present study, RETN expression was up-regulated in the NS group, indicating that RETN may be involved in the occurrence and development of NS.

Protein S1 (PROS1), located on chromosome 3q11.1, is a vitamin K-dependent plasma protein that activates coagulation factors V and VIII by activating protein C while promoting the clearance of early apoptotic cells [46]. Tyrosine kinase receptor (TAM receptor) regulates the basic mediator of inflammatory response; PROS1 acts as a ligand of TAM receptor; and the expression of proinflammatory factors, such as TNF-α and CCL3, is increased during PROS1 deficiency [45]. PROS1 expression is positively correlated with neutrophil count and activity and oxidative burst, and is a potential therapeutic target for decompensated cirrhosis and sepsis [46]. PROS1 can be used as a targeted drug for the treatment of inflammatory diseases, such as spinal cord injury and ankylosing spondylitis [47]. In the present study, PROS1 expression was up-regulated in the NS group. The role of PROS1 in the coagulation mechanism has been systematically studied; however, its role in NS has not been elucidated.

Methyltransferase 7B (METTL7B) belongs to the methyltransferase-like protein family, and is located on chromosome 12. To date, the function of METTL7B is unclear, although several studies have linked it to specific disease states, subcellular localization, and cellular processes [48, 49]. A recent study found that METTL7B has methylase activity, which can methylate intracellular alkanethiol molecules and reduce associated cellular toxicity [49, 50]. METTL7B expression is associated with immune cells, such as B cells, CD4 + T cells, CD8 + T cells, monocytes, neutrophils, macrophages, and activated mast cells. Clinical studies have shown that METTL7B responds to inflammatory signals via Janus Kinase 1 [51]. In the present study, METTL7B expression was up-regulated in the NS group, indicating that METTL7B may be involved in the occurrence and development of NS.

Tudor domain-containing protein 9 (TDRD9) is a DEXH-box RNA helicase, which is involved in PIWI-interacting RNA formation [52]. TDRD9 is a DNA damage and repair-associated gene and is mainly expressed in sperms [53]. It can be used to predict disease-free survival in cancers, such as clear cell renal cell carcinoma and thyroid cancer [54, 55]. In addition to the male reproductive system, it is mainly expressed in the blood cells, including monocytes and dendritic cells, which play important roles in the innate immune response [56].

The novelty of our study is as follows. First, we used bioinformatic analysis to investigate the molecular mechanisms of NS from the perspectives of immunity and inflammation. Second, we found that NSUN7, PROS1, TDRD9, RETN, LOC728401, and METTL7B may be potential diagnostic biomarkers for NS, particularly NSUN7. However, this study has some limitations. First, we could not determine whether a causal relationship exists between the differences in gene expression and pathophysiological mechanisms of NS or if it is simply a compensatory change. Second, the study was a retrospective data analysis; therefore, we lacked detailed clinical and prognostic data, which limited further exploration of the genes for their clinical characteristics and outcomes. Finally, our study was based on bioinformatic analysis of transcriptome data from public datasets, which may be inconsistent with the actual situation. Further clinical trials are needed to validate our findings.

Through bioinformatic analysis of published transcriptional data, NSUN7, PROS1, TDRD9, RETN, LOC728401, and METTL7B were identified as potential biomarkers of NS from the perspective of immune cell infiltration combined with logistic regression. More importantly, the developed diagnostic models provide a new perspective for future research on the pathogenesis of NS.