Research populations
A total of 1,620 major trauma patients were enrolled in this study, all of whom were Han Chinese people from Chongqing in western China (n=920), Zhejiang Province in eastern China (n=350) and Guizhou Province in southern China (n=350). Patients were admitted to Daping Hospital of Army Medical University, Chongqing Emergency Medical Center, the Second Affiliated Hospital of Zhejiang University and the Affiliated Hospital of Guizhou Medical University between 2005 and 2019. The inclusion criteria were as follows: 1) age between 18 and 65, 2) injury severity score (ISS) greater than 16, and 3) survival for more than two days after injury. The exclusion criteria were as follows: 1) combined with penetrating injuries and 2) severe brain injury or preexisting cardiovascular, respiratory, renal, hepatic, hematological or immunological diseases. Ethics approval for this study was obtained from the Ethics and Protocol Review Committees of Army Medical University, Chongqing Emergency Medical Center, Zhejiang University and Guizhou Medical University (Trial registration: ClinicalTrials.gov, NCT01713205. Registered on 18th October 2012, retrospectively registered). Before enrollment, written informed consent was obtained from the patients or their next of kin, which covered the collection of relevant clinical data and explicit DNA analysis. Patient confidentiality was preserved according to the guidelines of the Declaration of Helsinki.
Sepsis and MODS evaluation
Sepsis is defined as an acute change in total sequential (sepsis-related) organ failure assessment (SOFA) score greater than or equal to 2 points consequent to an infection (33). MOD scores are evaluated on each hospitalization day. Briefly, pulmonary, renal, hepatic, neurological, cardiac and hematological parameters were scored from 0 to 4 every day. The MOD scores ranged from 0 to 4, and the total score ranged from 0 to 24 (six organs). Failure of organ function was considered to be 3 or more points for more than two consecutive days (34). The presence of sepsis and multiple organ dysfunction scores were determined by individuals who did not know the patients’ genotypes.
Genotyping
Genomic DNA was isolated from peripheral whole blood by QuickGene-610L (Fujifilm, Tokyo, Japan). The concentration and purity of DNA samples were checked by a Thermo Scientific Nanodrop ND-1000 spectrophotometer (Isogen Life Science, De Meern, Netherlands). DNA samples were stored at −80°C. SNP genotyping was performed by an improved multiplex ligation detection reaction (iMLDR) technique, as described in our previous report (35). Approximately 10% of the samples were genotyped in duplicate to assess the accuracy of iMLDR. Genotyping was performed by researchers who did not know the patients’ clinical data.
Expression plasmid construction
A plasmid containing human CXCL16-I123A181 cDNA with a C terminal flag and His-tag was purchased from Vigene Biosciences (Jinan, Shandong Province, China). Plasmids encoding CXCL16-I123V181, T123A181 and T123V181 cDNA mutations were generated by site-directed mutagenesis (Stratagene, La Jolla, CA). The primers for 123T were 5’-TGAGGCCTGAGAAgTTGGGGGGCTGGTAGGAA-3’ (forward) and 5’-CAACTTCTCAGGCCTCAGAGGGGGCA-3’ (reverse). The primers for 181V were 5’- CCCAaCTGCCAGACTGTGGCCCGCA-3’ (forward) and 5’-ACAGTCTGGCAGtTGGGCCTGAGGCTGGGGA -3’ (reverse). The sequences of the four plasmids were checked by direct sequencing (BGI Genomics, Beijing, China). To obtain recombinant human CXCL16 proteins, CHO cells were cultured in a 6-well plate (F12 medium with 10% FCS, penicillin-streptomycin, 3 mM glutamine). After washing was performed, CHO cells were cultured in serum-free medium and transfected with 4 μg of CXCL16-I123A181, I123V181, T123A181 or T123V181 plasmids. Transfected cells were cultured in serum-free medium for 48 hours. According to the manufacturer's instructions, the supernatants were subsequently purified by a His-bind purification kit (Merck-Millipore, Massachusetts, the United States) to obtain purified CXCL16-I123A181, I123V181, T123A181 and T123V181 proteins.
Cell chemotaxis and adhesion assay
The polycarbonate membrane in the middle of the Transwell insert (Corning, New York, the United States) divides the chamber into two parts: the upper chamber and the lower chamber. THP-1 cells (human monocyte cell strain) were cultured in the upper chamber of the Transwell insert. The purified human CXCL16-I123A181, I123V181, T123A181 and T123V181 proteins were dissolved in the medium of the lower chamber (100 ng/ml). Chemotactic cells that passed through the membrane pores were dyed and subsequently counted to calculate the chemotaxis ability of four kinds of human CXCL16 protein (Fig. 2B).
For the cell adhesion test, THP-1 cells were cultured in serum-free medium and transfected with 4 μg of CXCL16-GFP-I123A181, I123V181, T123A181 and T123V181 expression plasmids. Raw264.7 cells that express CXCL16 receptor- CXC chemokine receptor 6 (CXCR6) were seeded at a density of 1x104 in a 6-well plate (36). Raw264.7 cells were cultured for 48 hours until they reached 70%-80% confluency and were subsequently cocultured with THP-1 cells expressing CXCL16-GFP-I123A181, I123V181, T123A181 and T123V181. The plate was incubated for 60 minutes at 37°C to enable cell binding. Nonadherent THP1 cells were washed away. The plate was read at 488 nm in a fluorescent plate reader. All assays were performed in triplicate. The number of THP-1 cells that adhered to Raw264.7 cells is presented as the mean fluorescence intensity (MFI).
Molecular modeling of CXCL16 polymorphisms
Using MAFFT (v7.38) (https://mafft.cbrc.jp/alignment/software/algorithms/algorithms.html), the CXCL16 proteins of 11 species were sequenced, and the evolutionary tree was calculated. The structurally conserved regions and loop regions of CXCL16 were analyzed by SMART (https://www.megasoftware.net/). The optimized structures were compared by PyMOL (version 0.97). Polymorphism Phenotyping v2 (Polyphen2, http://genetics.bwh.harvard.edu/pph2/) was employed to predict the effect of the two mutations. Using Modeler, Phyre2 and Raprotx, the full-length CXCL16 sequence was modeled. Using the wild-type CXCL16 structure as the template, the structure models of I123T and A181V are established. Using GROMACS 5.12, the structural models of I123T and A181V were modeled. The distances of amino acids and the area and volume of the hydrophobic pocket were analyzed using the distance geometry method.
Statistical Analysis
The relevance of CXCL16 tag SNPs for clinical outcomes was analyzed by three genetic models, that is, the dominant, recessive, and allele-dose models. To assess sepsis risk, adjusting for age, sex, and injury severity for confounding effects, odds ratios with 95% confidence intervals were calculated by multivariable logistic regression models. Adjusted by age, sex, and injury severity for confounding effects, the association of tag SNPs with MOD scores was determined by linear regression analysis. A P-value < 0.05 after Bonferroni correction for multiple testing was employed to indicate significance. SPSS 13.0 software (SPSS Inc., the United States) was utilized for the statistical analysis. Chemotaxis and adhesion data were compared using One-way analysis of variance (One-way ANOVA).