A prospective study was carried out in the emergency intensive care unit (ICU) of Zhongshan Hospital, Fudan University, Shanghai, China. From December 2017 to September 2019, 101 septic patients who met the clinical criteria for sepsis-3 were enrolled (1). There were 104 non-septic patients in the ICU (poly-trauma, cerebral trauma, intracranial hemorrhage, cerebrovascular accidents, and hypertensive emergencies) recruited as non-septic ICU controls. Furthermore, septic patients were divided into survival group and non-survival group according to the survival of 90 days.

Exclusion criteria included age below 18 years, pregnancy, severe chronic respiratory disease, severe chronic liver disease (defined as Child-Pugh score > 10), malignancy, immune disease, using high-dose immunosuppressive therapy, and AIDS patients. In addition, 46 age- and gender-matched healthy volunteers with no medical problems were obtained from the medical examination center of Zhongshan Hospital, Fudan University, China. The flowchart of this study is shown in Supplementary Figure 1 .

Patients’ characteristics (age, gender, and previous health status), as well as clinical data including Sequential Organ Failure Assessment (SOFA) and Acute Physiology and Chronic Health Evaluation II (APACHE II) scores, source of primary infection, and ICU mortality, were obtained after ICU admission. The characteristics of patients, mechanical ventilation, and vasopressor treatment are shown in Table 1 . This study was approved by the Ethics Committee Study Board of Zhongshan Hospital, Fudan University, Shanghai, China (number: B2014-082). Written informed consent was obtained from patients, the next of kin, or guardians on the behalf of the participants before enrollment according to the Declaration of Helsinki.

Gpr174 knock-out (Gpr174-KO) mice generated by a homologous recombination method were provided by Shanghai Southern Model Biotechnology Co. Ltd. (Shanghai, China). Age- and gender-matched C57BL/6 mice were obtained from Fudan University, Shanghai, China. Mice were housed under a specific pathogen-free condition with a 12 h-light/12 h-dark cycle, 22 ± 2°C. Animal experiments were approved by the Ethics Committee of Laboratory Animal Science, Fudan University, China (number 201804001Z).

The CLP-induced septic mouse model was established as described previously (19). In short, mice were anesthetized with 1% pentobarbital (i.p. 10 ml/kg of body weight). After that, the cecum was ligated in half, and a 21-gauge needle was used to puncture the stump once to squeeze out a small number of feces. The cecum was placed back into a normal intraabdominal position, and the abdominal incision was closed by applying sample running sutures. Then, pre-warmed normal saline (50 ml/kg of body weight) was injected subcutaneously. Sham-operated control mice underwent the same surgical procedures without puncture or ligation. The survival rate was monitored daily for 1 week.

The expressions of GPR174 mRNA in peripheral blood of all samples were tested by quantitative PCR on Day 1 (D1) and Day 7 (D7) after ICU admission. Peripheral blood mononuclear cells (PBMC) were isolated from fresh anticoagulant blood by Ficoll lymphocyte separation solution (Lymphoprep, Axis-Shield, UK). The total RNAs from all enrolled subjects and mice were extracted using TRIzol reagent (Life Technologies, Carlsbad, CA) and then 10 μg of RNA samples were reverse-transcribed into cDNA with Prime Script™ reagent kit (Takara, Dalian, China) following the manufacturer’s instructions. Quantitative PCR was performed with SYBR^® Premix Ex Taq™ II (Takara, Dalian, China) on a 7500 Real-time PCR system (Applied Biosystems, Carlsbad, CA), according to the manufacturer’s instructions.

Primers used were as follows:

human-GPR174 sense 5’-ATCATCTGCCTTGCCTGTGTACTC, antisense 5’-CGCCAATGGTCATCATAACAACGG; human-GAPDH sense 5’-AAGGTCGGAGTCAACGGATT, antisense 5’- CTCCTGGAAGATGGTGATGG; mouse-Gpr174 sense 5’- TTGGTCTGCATCAGTGTGCGAAG, antisense 5’-CAGGCAGGCAAGGCAGATGATC; mouse-Gapdh sense 5’-GGAGAGTGTTTCCTCGTCCC, antisense 5’-ACTGTGCCGTTGAATTTGCC; mouse-Cfh sense; 5’- GTATCAAAACGGATTGTGACGT, antisense 5’- TAACACATGTCACAGTGTCTGA; mouse-Oas1g sense 5’- TAAGAAACAGCTGTACGAGGTT, antisense 5’- CCAGATGAGGATGGTGTAGATT; mouse-Spp1 sense 5’- AAACACACAGACTTGAGCATTC, antisense 5’- TTAGGGTCTAGGACTAGCTTGT.

Blood samples of septic and ICU non-septic patients and mice were centrifuged (3500 rpm, 15 min, room temperature) after collection. The serum was immediately frozen in liquid nitrogen (LN2) and stored at -80°C for further use. The levels of IL-1β, IL-2R, IL-6, IL-8, and IL-10 were measured by ELISA kit (R&D Systems, Inc., Minneapolis, MN) according to the manufacturer’s protocol. Serum concentrations of CRP and PCT were measured by IMMAGE800 analyzer (Beckman Coulter, Inc. CA) and VIDAS B.R.A.H.M.S PCT analyzer (Biomerieux, Lyon, France). Routine blood tests and blood gas analysis were conducted in the clinical laboratory of Zhongshan Hospital, Fudan University, Shanghai, China.

Tissues (lung, liver) were excised at 0 h, 12 h, 24 h, 72 h, and 1 and 2 weeks after CLP, washed with DPBS, fixed with 4% formalin buffer, and paraffin embedded. There were 4-6 μm sections cut and placed on glass slides, deparaffinized in xylene, and rehydrated in a series of alcohol solutions. Sections were then washed in distilled water, and stained with H&E for histopathological examination.

Total RNA was digested by DNase I (Qiagene) and separated by Dynabeads^® mRNA DIRECT™ Kit (Life Technologies). The isolated mRNAs were used for mRNA-seq libraries with a KAPA Stranded mRNA-Seq Kit according to the manufacturer’s instructions. Libraries were sequenced on the Hiseq Xten system (Illumina) with a reading length of 150 base pairs (bps). Differentially expressed genes (DEGs) between Gpr174-KO and Wild type (WT) mice were identified using the Bioconductor package RUVSeq (version 1.0.0, http://www.bioconductor.org). Hierarchical cluster analysis of the DEGs was carried out by the “hclust” function of the “stats” package in R software (https://www.r-project.org/). The heatmap for this cluster analysis of the DE genes was drawn with heatmap.2 function in the “gplots” package. Gene ontology analysis (GO analysis) performed by Blast2GO software (version 4) was used to provide a further understanding of these results.

Normal distribution data were expressed as means and standard deviations (S.D.) with Student’s t-test or ANOVA test. Non-normal distribution continuous data were expressed as medians with the 25th and 75th quartiles applying Mann-Whitney U test or Kruskal-Wallis test followed by Dunn’s multiple comparisons post-test. Categorical data were expressed as frequency and percentage. Non-parametric Spearman’s rank correlation coefficient was performed to test correlations between two parameters. For analyzing the independent predictors of 90-day mortality, binary logistic regression was used to determine the discriminative power of GPR174 mRNA for 90-day mortality. Receiver-operating characteristic (ROC) curves were constructed and the area under the curve (AUC) was determined with a 95% confidence interval (CI). The bootstrap and Venkatraman’s test were used for comparing the AUC using the “pROC” package in R (20). Kaplan-Meier curves of disease-free survival were plotted and compared by Cox regression analysis in the groups layering by ROC cut-off point. For murine survival studies, Kaplan-Meier analyses followed by log-rank tests were performed. All statistical analyses were done using SPSS 16.0, R 4.0.2, and GraphPad Prism version 5.01 (GraphPad Software, San Diego, CA). P values less than 0.05 were considered statistically significant.

The demographic and clinical characteristics of septic patients and non-septic ICU controls are presented in Table 1 . No age effects were observed between septic patients and non-septic ICU controls (61.5 vs. 61.6, P=0.937). The 90-day mortality was 25.7% in septic patients, while 2% in non-septic ICU controls. After ICU admission (Day 1), septic patients had significantly higher APACHE II and SOFA scores than non-septic ICU controls. Non-survival septic patients also had significantly higher APACHE II and SOFA scores, and the utilization rate of mechanical ventilation and vasoactive agents than survivors ( Table 2 ). The levels of serum PCT, IL-2R, IL-6, and the counts of WBC, neutrophil, and lymphocyte were also significantly higher in septic patients than that in non-septic ICU controls ( Table 1 ). However, there was no difference between non-survivors and survivors in sepsis patients at Day 1 (D1) for these inflammatory markers ( Table 2 ).

The relative expressions of GPR174 mRNA in septic patients at D1 were significantly lower than that in non-septic ICU controls and healthy volunteers ( Figure 1A ). The levels of GPR174 mRNA had no significant difference between survivors and non-survivors in patients with sepsis at D1 ( Figure 1B ). However, lower expressions of serum GPR174 mRNA were observed in non-survivors compared to survivors at Day 7 after admission (D7) ( Figure 1B ). Interestingly, with the recovery of sepsis, the concentration of GPR174 mRNA in septic patients returned to the level of healthy subjects ( Figure 1C ).

Our results showed that GPR174 mRNA levels were significant negative correlations with APACHE II and SOFA scores ( Figure 2 ). As for other biomarkers, a significant negative correlation was observed between GPR174 mRNA and CRP, while there is a positive correlation for the count of Lym at D7 ( Figure 3 ). However, no significant correlations were found among the levels of GPR174 mRNA and PCT, IL-2R, IL-6, IL8, IL-10, and the count of Neu, WBC at D7.

Then, we detected the individual change of those markers from D1 to D7 both in survivors and non-survivors in sepsis. Interestingly, GPR174 mRNA showed remarkable differences both in the non-survivor and survivor groups (ascending in the survivors and descending in non-survivors, respectively), while other biomarkers did not show such results, even for APACHE II and SOFA scores ( Supplementary Figure 2 ).

Using binary logistic regression analysis adjusted by age and gender, both GPR174 mRNA (OR = 0.004, P = 0.003) and CRP (OR = 1.010, P = 0.031) at D7 were found to be independent predictors of 90-day mortality in patients with sepsis ( Table 3 ).

We further investigated the prognostic value of GPR174 mRNA in sepsis with the ROC analysis. The area under the ROC curve (AUC) of GPR174 mRNA at D7 was higher than that at D1 (0.83 vs. 0.60; P < 0.001). Furthermore, via comparing the AUC of GPR174 mRNA with APACHE II score, SOFA score, CRP, PCT at D7, we found that the AUC of GPR174 mRNA was higher than that of PCT (0.83 vs. 0.66; P = 0.05). However, there was no difference when compared with CRP (AUC = 0.72, P = 0.21), APACHE II score (AUC = 0.88; P = 0.704), and SOFA score (AUC = 0.92; P = 0.26) ( Figure 4 ).

The samples were divided into higher and lower GPR174 mRNA groups according to ROC cut-off point and the K-M survival curve was obtained by Cox regression analysis. At D1, the cut-off level for GPR174 mRNA is 0.071, and the sensitivity and specificity were 0.880 and 0.338, respectively. At D7, the cut-off level for GPR174 mRNA is 0.101, and the sensitivity and specificity were 0.890 and 0.720, respectively. The Kaplan-Meier curves showed that no significant difference was found between the two groups at D1 [HR = 0.756, 95% CI = (0.291, 1.966), P = 0.566], while the mortality of the lower GPR174 mRNA group at D7 was significantly higher than that of the higher GPR174 mRNA group [HR = 0.224, 95% CI = (0.090, 0.561), P = 0.001] ( Figure 5 ).

To further study the expression of Gpr174 mRNA in sepsis, a CLP-induced septic mice model (n = 20 per group) was used to test the expression of Gpr174 mRNA in PBMC and spleen. Compared to control mice with sham surgery, the levels of Gpr174 mRNA in PBMC significantly decreased in septic mice, then met an ascending inflection point at 72 h ( Supplementary Figure 3A ), while it continuously decreased in the spleen ( Supplementary Figure 3B ).

The pathological injury of the liver and lung after CLP was confirmed by H&E staining (n = 5 per group). The results showed that clear polygonal hepatic lobules with obvious hepatic cords and sinuses were seen in the liver in the sham surgery group. However, the boundaries of some hepatic lobules were not clear after CLP, and hepatic cord disorder could be seen in hepatic lobules. The damage was most severe at 72 h, which was characterized by cell edema, infiltration of inflammatory cells, and hepatic sinus congestion ( Supplementary Figure 3C ). As for lung tissues, the alveolar cavity was clearly visible in the sham surgery group, while most of the alveolar cavity showed exudation and edema. Severe inflammatory cells infiltration was observed at 72 h after CLP ( Supplementary Figure 3D ).

Compared with sham mice, the mortality rate of Gpr174-KO septic mice was dramatically decreased from 55% to 25% (n = 20 per group) ( Supplementary Figure 4A ). In addition, Gpr174-deficiency significantly alleviated the pathology scores of lung and liver, respectively. H&E staining showed that inflammatory cell infiltration and edema of lung tissue were less severe in Gpr174-KO + CLP mice compared to WT + CLP mice ( Supplementary Figure 4B ). Moreover, the hepatic cells edema was less in Gpr174-KO + CLP mice than that in WT + CLP mice ( Supplementary Figure 4C ).

To evaluate the effect of GPR174 on the dysregulated systemic inflammation, the serum levels of IL-1β, TNF-α, and IL-10 were examined at 24 h after the CLP challenge (n = 5 per group). The results showed that the levels of IL-1β and TNF-α were significantly decreased in Gpr174-KO+CLP mice than in the WT+CLP mice, and the levels of IL-10 were significantly increased ( Supplementary Figure 5 ).

Then, RNA-seq was used to explore the mechanism of GPR174 in the pathogenesis of sepsis. A total of 10,389 genes were identified compared between Gpr174-KO mice and WT mice at 24 h after the CLP challenge (n = 3 per group). Gpr174-KO+CLP mice and WT+CLP mice were well separated by cluster PCA analysis ( Figure 6A ). We defined the identified genes as differentially expressed genes (DEPs) if there was a log2FC in excess of 1.5 or less than -1.5, FDR < 0.05. A total of 360 genes changed significantly as shown in the heatmap ( Figure 6B ). GO analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis were performed to provide a further understanding of these results. A similar differential effect was observed in the molecular and immunologic pathways that were impacted between Gpr174-KO+CLP mice and WT+CLP mice. Further validation of the findings was carried out by qPCR. Intriguingly, we found that representative potential mediators of the immune response, including dendritic cell-specific transmembrane protein and macrophage mannose receptor 1 were highly upregulated, while interleukin-12 subunit alpha and Sialoadhesin were downregulated ( Figure 6C ).