The significance of serum HMGB1 level in humans with acute paraquat poisoning

High-mobility group box 1 (HMGB1) mediates acute lung injury in a mouse model of paraquat poisoning. However, published reports showing a clinically relevant association between HMGB1 and paraquat exposure are lacking. The objective of the present study was to investigate the potential role of serum HMGB1 level as a prognostic marker of mortality in patients with paraquat poisoning in a clinical setting. This retrospective observational cohort study included a convenience sample of 92 patients with acute paraquat poisoning admitted to the emergency room (ER) of The First Hospital of Jilin University between January 2014 and December 2016. Baseline serum HMGB1 levels and other laboratory parameters were measured on admission. Cumulative incidence of mortality during the first 30 days after admission was 50% (n = 46/92). Serum HMGB1 levels were higher in fatalities than survivors (P = 0.015), 30-day mortality increased with increasing baseline serum HMGB1 level (P < 0.001), and higher serum HMGB1 levels were associated with an increase in 30-day mortality on Kaplan-Meier analysis. Multivariate Cox regression analysis identified baseline serum HMGB1 levels, white blood cell count, and serum lactic acid levels as independent prognostic markers of 30-day mortality. These data suggest that serum HMGB1 levels measured on admission to the ER are an independent predictor of 30-day mortality in patients with acute paraquat poisoning.

www.nature.com/scientificreports www.nature.com/scientificreports/ Methods Study subjects. This study was approved by the ethics committee of The First Hospital of Jilin University and was conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all patients on admission to the emergency room (ER).
This retrospective observational cohort study included a convenience sample of patients presenting to the ER of The First Hospital of Jilin University between January 2014 and December 2016 with acute paraquat poisoning, defined as an oral intake of 2-200 mL paraquat during the prior 24 hours. Exclusion criteria were: (1) diabetes mellitus, hypertension, renal failure, or malignancy; (2) >24 hours since oral intake of paraquat; or (3) incomplete follow-up data. Hospitalized patients were followed up to death. Discharged patients were followed up to death or for 30 days. Twenty healthy subjects were enrolled to investigate baseline HMGBI levels in a group of healthy controls.
Sample collection and biochemical analyses. Each patient provided a peripheral venous blood sample (5 mL) during the first 24 hours after admission to the ER. Sera were isolated, frozen, and stored at −80 °C until use. Serum HMGB1 levels were measured using an enzyme-linked immunosorbent assay kit according to the manufacturer's instructions (Yanhui Biotech, Shanghai, China).
Baseline hematological and laboratory parameters, including white blood cell (WBC), neutrophil, lymphocyte, and platelet counts, hemoglobin, prothrombin time, total bilirubin, albumin, alanine aminotransferase (ALT), aspartate aminotransferase (AST), creatinine, serum amylase, troponin I (cTnI), pH, partial pressure of carbon dioxide (PaCO2), and partial pressure of oxygen (PaO2) were measured using a Hitachi 7600 Clinical Analyzer (Hitachi, Tokyo, Japan), Sysmex CA-7000 System (Sysmex, Kobe, Japan), and Sysmex XE-5000 (Sysmex) 8 , according to standard protocols. Statistical analysis. Statistical analyses were conducted using SPSS, v24 (SPSS, Chicago, IL, USA). Values are expressed as mean ± standard error of the mean (SEM) for normally distributed data or median and interquartile ranges for non-normally distributed data. One-way analysis of variance (ANOVA) and the Kruskal-Wallis H test or Mann-Whitney U test were performed to evaluate between-group differences. Chi-squared test was conducted to evaluate between-group differences stratified by gender and in 30-day mortality. A Kaplan-Meier curve was constructed to describe survival. Univariate and multivariate Cox regression analyses were performed to identify predictors of 30-day mortality; results are summarized as hazard ratios with corresponding 95% confidence intervals (CIs). Variables with a P value < 0.1 on univariate analysis were included in the multivariate analysis. P < 0.05 was considered statistically significant.

Results
Study population. This study enrolled 304 patients with an oral intake of 2-200 ml paraquat. A total of 212 patients were excluded. Of these, 21 patients had diabetes mellitus, 29 patients had hypertension. 19 patients had renal failure, 22 patients had malignancy, 104 patients had inadequate follow up data, and >24 hours had passed since the oral intake of paraquat in 17 patients. The remaining 92 patients were included in the final analysis. These patients had a median age of 33.5 years (range: 13-77 years) and 47 were female and 45 were male.

Comparison of the serum HMGB1 level between fatalities and survivors. Mean serum HMGB1
levels in healthy controls, fatalities, and survivors are shown in Fig. 1. Mean serum HMGB1 level in healthy controls was 0.53 [range, 0.00-1.30] μg/L, and significantly lower than for fatalities or survivors (P < 0.001). Mean baseline serum HMGB1 level was significantly higher in fatalities vs. survivors (11.

Comparison of the serum HMGB1 level in females and males. Mean serum HMGB1 levels in
females and males are shown in Fig. 2. There was no significant difference in mean serum HMGB1 level between females and males (t = 1.60, P = 0.11) (Fig. 2).
Association of serum HGB1 level with mortality rate. All patients were treated according to standard procedures in the ER. Cumulative incidence of mortality during the first 30 days after admission was 50% (n = 46/92). The ratio of survivors to fatalities in each HMGB1 tertile is shown in Table 2. The 30-day mortality rate was significantly increased across tertiles from lowest (tertile 1, 26.7%; tertile 2, 43.3%) to highest (tertile 3, 78.1%) serum HMGB1 levels. A Kaplan-Meier curve was constructed to describe survival. The log rank test of a difference in survival between tertiles was significant (P < 0.001) (Fig. 3).

Risk factor analysis for 30-day mortality. Multivariate Cox proportional hazards regression analyses
identified baseline serum HMGB1 level, WBC count, and serum lactic acid levels as independent prognostic markers of 30-day mortality in patients with acute paraquat poisoning (Table 3).

Discussion
The results of the current study investigating acute paraquat poisoning in humans showed that 30-day mortality was increased in patients with higher baseline serum HMGB1 levels, and baseline serum HMGB1 level was an independent prognostic marker of 30-day mortality in this patient population.
Paraquat is a poisonous herbicide that has toxic effects due to its redox activity. Paraquat induces oxidative damage and cell death by generating ROS, including hydrogen peroxide, superoxide anion, and hydroxyl radicals. HMGB1 in involved in infection, organ dysfunction and immune responses in acute injury and chronic disease, including myocardial infarction, stroke, acute lung injury, acute liver injury 7,9 , ischemia-reperfusion injury, sepsis, fibrotic disorders, and cancer 10 . HMGB1 is an alarmin, with distinct intracellular and extracellular effects. It is present in the cell nucleus and cytoplasm and can be translocated to the extracellular space 11 . HMGB1can be actively released from cells, including macrophages, monocytes, natural killer cells, dendritic cells, endothelial   Table 1. Baseline characteristics of patients stratified by serum HMGB1 tertile (n = 92). MAP, mean arterial pressure; PaCO 2 , partial pressure of carbon dioxide; PaO 2 , partial pressure of oxygen; WBC, white blood cell; Hg, hemoglobin; CK, creatine kinase; BUN, blood urea nitrogen; AST, aspartate transaminase; ALT, alanine transaminase. www.nature.com/scientificreports www.nature.com/scientificreports/ cells and platelets, or passively produced by necrotic and damaged cells 12,13 . In the extracellular space, HMGB1 mediates inflammation and chemotaxis 12 and has far reaching effects causing remote organ dysfunction in sterile injury 14 .
Death in paraquat poisoning is usually caused by ALI 15,16 . The development and severity of ALI is controlled by pulmonary neutrophil infiltration [17][18][19] . Findings from our previous study revealed that the HMGB1-TLR4-I-L23-IL-17A axis regulated neutrophil infiltration and ALI in a mouse model of paraquat     www.nature.com/scientificreports www.nature.com/scientificreports/ poisoning 7 . Conversely, administration of HMGB1-specific neutralizing antibodies or the HMGB1-binding antagonist thrombomodulin can rescue mice and rats 20-22 from lethal sepsis. Taken together, data from our studies and previous animal experiments suggest that neutralization of HMGB1 represents a potential therapeutic option for the treatment of paraquat poisoning. However, the results should be interpreted with caution, as the specific mechanism of action of HMGB1 in humans remains to be elucidated.
The previous reports and findings from the current study suggest an association between HMGB1 and the mechanism of paraquat toxicity, in which HMGB1 may function as a regulator of cytokine, chemokine, and growth factor activity, coordinating the inflammatory and immune response 7,23 . Further investigations regarding this association are warranted to expand understanding of paraquat toxicity and inform the development of targeted therapeutics.
Some studies have determined that initial leucocyte level and arterial lactate are predictors of mortality in patients with paraquat poisoning 17,[24][25][26] . This is consistent with our findings. Reliable predictors of prognosis are important to inform clinical treatment decision making in this patient population.
This study was associated with several limitations. First, this was a cross-sectional study and cause and effect of HMGB1 in the etiology of paraquat poisoning could not be elucidated. Second, oral paraquat poisoning was based on patient self-report. Evidence suggests that plasma paraquat levels have prognostic value in patients with acute paraquat poisoning 27 . However, plasma paraquat levels were not measured in the current study as access to testing is limited by time and cost, and most ER physicians do not rely on plasma paraquat measurements for clinical decision making.
In summary, the results of the current study suggest that serum HMGB1 level measured on admission to the ER is a feasible index for assessing the outcome of patients with acute paraquat poisoning, and strategies that decrease serum HMGB1 levels may have role in the management of these patients. Further studies elucidating the role of HMGB1 in paraquat poisoning are required to guide the development of effective clinical interventions.

Data Availability
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.