The novel biomarkers for assessing clinical benefits of continuous renal replacement therapy in pediatric sepsis: a pilot study

Baseline characteristics of patients in the validation set

A total of 27 sepsis patients received CRRT and 19 sepsis patients without CRRT were eligible. To match the severity of patients in the non-CRRT group, 8 patients dead within 7 days after PICU admission were excluded, and 2 patients included in the discovery set were also excluded. Finally, 17 cases were entered into the validation set as the CRRT group. In the non-CRRT group, patients with advanced tumor (n = 3), congenital heart disease (n = 3), or the concentration of biomarkers exceeded the ELISA KIT detection range(n = 3) were excluded. Eventually, 10 sepsis patients without CRRT were included in the validation set as the non-CRRT group. Sera from residual blood after physical examination from 10 healthy children were used as control. In the present study, subjects were divided into discovery set including 3 pediatric patients with severe sepsis received CRRT, and the validation set including 10 healthy controls, 10 patients with severe sepsis without CRRT, and 17 patients with severe sepsis under CRRT, which was schematically summarized in Fig. 1.

Fig. 1figure 1

Flowchart of experimental design for biomarker screening

The median (IQR) age of pediatric patients admitted to the PICU in this cohort was 65 (33.5, 113.5) months in the CRRT group and 44.5 (4.75, 117.25) months in the non-CRRT group. In the validation set, 47% in the CRRT group and 50% in the non-CRRT group were male. The values of APTT were 80.70 (38.6, 155.6) s in the CRRT group and36.1(28.98,38.25) s in the non-CRRT group, respectively (p = 0.015). In addition, there were no significant differences in aspect of weight, age, gender, PRISM III, CRP, WBC, PCT, ALT, TBIL, etc. (Table 1).

Table 1 Baseline characteristics of patients in the validation groupClinical improvement of patients in the validation set

To assess the clinical improvement of patients in CRRT group and non-CRRT group, the pSOFA score were determined at the indicated time. The changes of pSOFA of patients either in CRRT or non-CRRT group have a significantly decrease on 7th day compared with pre-CRRT or on PICU admission, respectively (Additional file 1: Table S1). Moreover, there were 10 CRRT responders and 7 CRRT non-responders according to the definition of validity for CRRT in this study (Additional file 1: Table S2).

TMT-based proteomic analyses

There were 145 differentially expressed proteins identified comparing before and after CRRT for 7 days. Among them, 56 proteins were increased and 89 were decreased expression. Heatmap visualization and a volcano plot of 145 differentially expressed proteins was displayed (Fig. 2A, B). Then, Gene Ontology (GO) analysis revealed that the differentially expressed proteins participated in various biological progress including acute-phase response, acute inflammatory response, leukocyte activation involved in immune response, cell activation involved in immune response, transforming growth factor β(TGFβ) receptor signaling pathway (Fig. 2C, Table 2). Moreover, KEGG pathway analysis indicated that NF-κB signaling pathway, AGE-RAGE signaling pathway, cGMP-PKG signaling pathway, staphylococcus aureus infection, TNF signaling pathway were involved in the process of CRRT-mediated sepsis recovery (Fig. 2D). Furthermore, IPA for the 145 differentially expressed proteins showed that the top 5significantly different canonical pathways were LXR/RXR and FXR/RXR activation, acute phase response signaling, complement system, coagulation system (Fig. 2E). Given the important roles of immune response in sepsis, the differentially expressed protein enriched in cell activation involved in immune response were further analyzed via STRING database. Finally, interactome analysis indicated LYZ, LRG1, FMOD, SERPINA3, SELL, and CD14 could be novel candidates involved in the regulation of complement system (shown as C3 in Fig. 2F), acute-phase response (shown as LPS binding protein [LBP] in Fig. 2F), and TGFβ receptor signaling pathway (shown as TGFβ1 in Fig. 2F).

Fig. 2figure 2

Serum proteomics analysis in 3 pediatric patients with severesepsis as a discovery set. A Heatmap visualization of 145 differentially expressed proteins. B Volcano plot of differentially expressed proteins. C Distribution of altered proteins for biological processes. Only GO terms that were significantly overrepresented (p < 0.05) are shown. D Significantly enriched pathways identified by KEGG pathway analysis (p < 0.05). E The canonical pathways obtained by IPA analysis. The Y-axis represents the ratio of the number of genes from the dataset that map to the pathway and the number of all known genes ascribed to the pathway. F Interconnections between differentially expressed proteins in the process of cell activation involved in immune response

Table 2 The biological processes enriched by differential expressed proteinsConfirmed analysis of serum candidates including LYZ, LRG1, FMOD, SERPINA3, SELL, andCD14 in the validation set

As shown in Fig. 3A, B, serum level of LYZ significantly increased on the 7th day compared with the 1stday in the CRRT group (p < 0.05) but not CRRT group (p > 0.05), and serum LYZ levels were significantly higher on the 7th day in the CRRT group than the non-CRRT group (p < 0.001). In addition, serum levels of LRG1 (Fig. 3C, D) significantly increased on the 7th day compared with the 1st day in the CRRT group (p < 0.01) but not CRRT group (p > 0.05), and serum LRG1 levels were significantly higher on the 7th day in the CRRT group than the non-CRRT group (p < 0.05).However, serum levels of FMOD, SERPINA3, and SELL were no significant changes on the 7th day compared with the 1st day either in the CRRT group or the non-CRRT group (Fig. 3E–J). Importantly, serum CD14 levels decreased on the 7th day compared with the 1st day in the non-CRRT group, but not in the CRRT group (Fig. 3K, L).

Fig. 3figure 3

Confirmed analysis of serum candidates including LYZ, LRG1, FMOD, SERPINA3, SELL, and CD14 using ELISA in the validation set. A, B serum LYZ levels, C, D serum LRG1 levels, E, F serum FMOD levels, G, H serum SEPRINA3 levels, I, J serum SELL levels, K, L serum CD14 levels. * indicates P < 0.05. ** indicates P < 0.01. ***indicates P < 0.001

Comparison of the levels of serum novel biomarkers in CRRT responders and non-CRRT responders

Serum LRG1 levels were appeared significantly elevated on the CRRT 1st day compared with CRRT initiation (413.257 ng/mL vs. 277.824 ng/mL, P = 0.047), and serum LYZ levels changed with the similar tendency (55.233 μg/mL vs. 39.05 μg/mL, P = 0.05) (Fig. 4A, B). However, there were no significantly changes in serum levels of CD14, SELL, FMOD, and SERPINA3 (Fig. 4C–F). On CRRT 7th day, serum LRG1, LYZ, and FMOD levels were significantly increased compared with pre-CRRT (Fig. 4A, B, E), but serum CD14 levels were significantly decreased (Fig. 4C). However, there were no obvious changes during CRRT in non-responders (Fig. 4).

Fig. 4figure 4

The levels of serum novel biomarkers in both CRRT responders (n = 10) and non-responders (n = 7). A LRG1, B LYZ, (C) CD14, D SELL, E FMOD, F SERPINA3. * indicates P < 0.05

Correlation analysis of serum biomarkers involved in sepsis and CRRT treatment to clinical indexes in pediatric patients with sepsis

To further explore the potential roles of confirmed serum LYZ and LRG1 in sepsis and for assessing the clinical benefits of CRRT, Pearson’s correlation analysis was performed between either LYZ or LRG1 and the clinical indexes on the 1stday.In the CRRT group, serum LYZ level was positively correlated to platelet count (Fig. 5A) and level of albumin (Fig. 5D), but negatively correlated to ALT and AST (Fig. 5B, C). Similarly, serum LRG level was positively correlated to total platelet count (Fig. 5E) and negatively correlated to total bilirubin level (Fig. 5F). However, there was no correlation between these two candidates and clinical parameters in the non-CRRT group (Table 3).

Fig. 5figure 5

Correlation analysis of serum biomarkers and clinical indexes in pediatric patients with sepsis in the CRRT group. AD Correlation analysis between serum LYZ level and platelet count, AST, ALT, or ALB in the CRRT group. EF Correlation analysis between serum LRG1 level and total platelet count or TBIL

Table 3 The correlation analysis between candidate indicators and clinical parameters in the CRRT group and the non-CRRT groupPPI network analysis of confirmed biomarkers of both LYZ and LRG1

Focusing on the potential interaction network of LYZ and LRG1, leucocytes adhesion to vascular endothelial cell, acute inflammatory response, defense response to G+ bacterium, regulation of Toll-like receptor signaling pathway, as well as complement activation were the most important, which influenced each other; the remaining other biomarkers including CD14, FMOD, SERPINA3, and SELL had also involved into these process with potential interaction with TGFβ, CRP, LBP, etc. (Fig. 6).

Fig. 6figure 6

The protein–protein analysis using STRING software online

留言 (0)

沒有登入
gif