Vitamin D Receptor Is a Sepsis-Susceptibility Gene in Chinese Children

Background

Sepsis is commonly seen among critically ill children worldwide, with a prevalence rate of 8.2% and an in-hospital mortality rate as high as 25% [1]. The Resolution on Sepsis by the United Nations World Health Assembly in 2017 recognized sepsis as a global threat in children and a priority to address during the next decade [2]. Sepsis is a polygenic and multifactor symptom, with ambiguous etiology [3]. There is evidence from studies of twins with late-onset sepsis [4,5] showing that genetic variability may influence the susceptibility to sepsis through the innate immune system. These genetic variants explain different outcomes of pediatric patients under standardized treatments, and also provide important clues for new mechanisms of sepsis.

The candidate gene approach assumes that the gene with a known biological function is the host one regulating the investigated traits [6,7]. Using this approach, our previous study revealed that the gene encoding vitamin D receptor (VDR) may be a candidate for sepsis risk, as some variants showed a cumulative effect on neonatal sepsis cases [8]. VDR is a kind of nuclear receptor that plays a central role in 1α, 25-dihydroxyvitamin D3’s biological actions, and regulates mass gene expression, cellular proliferation and differentiation, and immune response, largely in a ligand-dependent manner [9,10].

Animal experiments have shown that the activation of VDR can protect or attenuate organ injury through inhibiting cell apoptosis, and has even been shown in a mouse model to reverse sepsis-induced immunosuppression through enhancing autophagy [11–13]. Therefore, we developed the hypothesis that the VDR gene is a candidate gene of sepsis, and designed a case-control study among 267 children with sepsis and 283 healthy children by genotyping 9 variants in the VDR gene to see whether they can predict the risk of sepsis among children in China.

Material and Methods

STUDY CHILDREN:

This study was a hospital-based, case-control, genetic association study. Recruitment was carried out at the Emergency Department and Intensive Care Unit (ICU) of the Capital Institute of Pediatrics, Beijing, China. A total of 267 children with sepsis who met the criteria for treatment during the period from October 2017 to April 2020 and 283 healthy controls were included. The conduct of this study was approved by the Ethics Review Committee of the Capital Institute of Pediatrics in Beijing, China (approval ID SHERLL 2013075). All participants read and signed the informed consent form. If children were unable to sign the consent form, the guardians signed on their behalf. This study complied with the Declaration of Helsinki.

INCLUSION AND EXCLUSION CRITERIA:

The inclusion criteria were: a) patients under 12 years old admitted to the Pediatric Intensive Care Unit (PICU) diagnosed with sepsis; b) patients were included if they fulfilled criteria of the International Pediatric Sepsis Consensus Conference: Definitions for sepsis and organ dysfunction in pediatrics [14]. The exclusion criteria were as follows: a) patients with known autoimmune disease and cancer; b) using immunosuppressant or immunomodulator; c) congenital organ dysfunction and d) diagnosis of sepsis or shock over 72 h.

DNA EXTRACTION AND QUALITY CONTROL:

Venous blood samples were taken in 5-mL Vacutainer tubes with K3-EDTA from each participant. Plasma was separated by centrifugation at 4°C, then frozen in a freezer at −80°C. The RelaxGene Blood DNA System (Tiangen Biotech, Beijing, China) was used to extract genomic DNA from white blood cells according to the manufacturers’ guidelines. The specific process of DNA extraction is provided in Supplementary File 1. Then, we used a spectrophotometer to determine concentration (at A260 nm) and purity (at A260/A280 ratio) of DNA.

VARIANT SELECTION:

Nine variants in the VDR gene were selected: rs9729, rs2107301, rs2189480, rs2239185, rs3782905, rs4516035, rs7139166, rs11168266, and rs11168293. The selection of these variants was based on published papers [15–19] and the NCBI-Gene website analysis (https://www.ncbi.nlm.nih.gov/gene/).

GENOTYPING:

The 9 variants in the VDR gene were amplified by polymerase chain reaction (PCR) and sequenced by 3730 sequencing analysis. The primer sequences were designed according to the genomic sequence deposited in the NCBI database; 10% of the samples were randomly selected and re-genotyped to ensure the consistency of results. Sequencing results’ alignment and multiple comparisons were analyzed by Chromas Lite version 2.01 (http://www.technelysium.com.au). PCR was performed using the following parameters: denaturation for 5 min at 95°C, 30 cycles of 95°C for 15 s, Tm°C for 15 s (annealing temperatures are provided in Supplementary File 2), and extension for 1 min at 72°C, with a final extension for 7 min at 72°C. The DNA extraction procedure and primer sequences are provided in Supplementary File 2.

STATISTICAL ANALYSIS:

For database management and statistical analysis, we used STATA software Release 14.1 (Stata Corp, TX). Continuous variables are expressed as mean (standard deviation), and compared using the t test or Mann-Whitney U test between sepsis and healthy control groups according to its distribution. Categorical variables are described as number (percentage) and we performed the χ² test to assess differences between groups. Genotypes and allele differences were compared using the χ² test or Fisher’s exact test (if 1 observation’s frequency was lower than 5). Sepsis risk conferred by different genotypes was calculated by logistic regression analysis after adjusting for age and sex. Effect size was described as odds ratio (OR) and 95% confidence interval (95% CI). Two-sided P<0.05 was considered statistically significant. We used the additive model and dominant model to calculate risk prediction of the 9 studied variants for sepsis risk.

Generally, a haplotype is a combination of multiple alleles on one chromosome. We did some haplotype-based statistical analysis to explore the interactions of these variants. The Haplo.em program was used to compute the haplotype frequencies for the variants in different groups. Haplo.glm and haplo.cc were used to calculate effect sizes for each variant and haplo.score was used to estimate an individual’s phenotype as a function of each inferred haplotype. Simulate P (p_sim) was the statistical value after 1000 replicates. All the statistical analyses based on haplotype were implemented in the program Haplo.stats software (version 1.4.0) developed using R language (http://www.r-project.org).

Interaction analysis was implemented using the open-source multifactor dimensionality reduction (MDR) software package Release 3.0.2 available from http://www.multifactordimensionalityreduction.org/. Interaction circle graphs were used to visualize the nature of the dependencies.

Results

BASELINE CHARACTERISTICS:

Table 1 shows the baseline characteristics of sepsis cases and healthy controls. Controls were significantly older than cases (45 months vs 29 months, p<0.001), and males were overrepresented among cases.

SINGLE VARIANT ANALYSIS:

Table 2 shows the genotype distributions and allele frequencies of 9 variants in the VDR gene. The genotype distributions differed very significantly for rs2107301 and rs2189480 between cases and controls (p: 0.01 and 0.004, separately). The mutant allele frequencies of rs2107301 (C) and rs2189480 (C) were significantly higher in healthy controls than in sepsis groups (p: 0.003 and 0.001, respectively). There was no hint of significant differences in the other 7 variations, either in genotype distributions or allele frequencies, between the 2 groups. The effect size of each variations’ genotype was calculated using their wild homozygous genotype as a reference (Table 2). The mutant homozygous genotypes of rs2107301 (CC) and rs2189480 (CC) were associated with reduced risk compared to the wild homozygous genotype (OR: 0.44 and 0.43, 95% CI: 0.21–0.92 and 0.23–0.81, p: 0.03 and 0.009, respectively). Among the genotypes of rs2189480, the CA genotype also showed a lower risk of sepsis than the AA type (OR: 0.62, 95% CI: 0.43–0.90, p: 0.01). In the genotypes of the remaining variations, the differences in effect size were not significant. All p values were calculated after adjusting for age and sex in logistic regression analysis.

The unadjusted and adjusted risk predictions of the 9 studied variants in the VDR gene for sepsis mortality risk were calculated in both additive and dominant models to correct for some mutant homozygotes that were not sufficiently numerous, which has an impact on the predictive value (Table 3). In general, 2 variants, rs2107301 and rs2189480, showed significant protective effects under the additive and dominant model. The effect sizes were: rs2107301 in the additive model (OR: 0.68, 95% CI: 0.51–0.91, p: 0.008), rs2107301 in the dominant model (OR: 0.65, 95% CI: 0.46–0.93, p: 0.02), rs2189480 in the additive model (OR: 0.64, 95% CI: 0.49–0.84, p: 0.001), and rs2189480 in the dominant model (OR: 0.58, 95% CI: 0.41–0.83, p: 0.003). All the results were meaningful regardless of whether factors were adjusted or not.

HAPLOTYPE ANALYSIS AND HAPLOTYPE–PHENOTYPE ASSOCIATION:

Table 4 presents the derived haplotype frequencies and risk estimates for pediatric sepsis. Haplotype C-T-A-C-C-T-C-G-G (alleles arranged by order of rs9729, rs2107301, rs2189480, rs2239185, rs3782905, rs4516035, rs7139166, rs11168266, and rs11168293, with the same hereafter) was the most common (frequency less than 1% is not displayed). Frequency of haplotype C-C-C-C-C-T-C-A-G was significantly higher in controls than in cases (4% vs 1%, p: 0.02) and was significantly associated with a 0.59-fold decreased risk of pediatric sepsis (95% CI: 0.12–0.76, p: 0.02). Other haplotypes had no significant association with sepsis risk.

We explored a haplotype–phenotype association by taking all haplotypes of the 9 studied variants as a whole, and tested the comprehensive correlation of haplotypes with all collected baseline characteristics (Table 5). A significant association was noted for high-density lipoprotein (HDL) in sepsis patients after simulation correction (psim<0.05).

INTERACTION ANALYSIS:

The interaction of the 9 variants under study in predisposition to sepsis in children is displayed in Figure 1. Blue lines and green lines represent antagonism, and the intensity of blue was stronger than that of green; red lines or orange lines represent synergistic effect, and the intensity of red was stronger than that of orange (this study did not produce red or orange lines, indicating that antagonism was dominant among various sites). Overall, there was no evidence of synergistic interaction between variants.

Discussion

LIMITATIONS:

Several limitations should be acknowledged for this study. Firstly, there was a significant difference between groups in sex and age. Although we carried out statistical adjustments, the bias cannot be eliminated. Secondly, as an observational case-control association study, our results cannot prove the cause-effect relationship between VDR gene and pediatric sepsis risk. Thirdly, a small-sample bias may exist in this study, so the results should be considered as preliminary. Fourthly, the genetic variation coverage in the VDR gene was limited.

Conclusions

Taken together, our findings indicate that the VDR gene may be a sepsis-susceptibility gene in children from China. In particular, 2 variants in the VDR gene, rs2107301 and rs2189480, played a leading role in predicting pediatric sepsis risk. We agree that future large-scale, well-designed studies are warranted to further confirm or refute the findings of this study.