Clinical Characteristics and Metabolomics Investigation for Patients with Dengue Virus 2 Infection

Dengue is an endemic viral disease affecting tropical and subtropical regions around the world. The aim of this study was to retrospectively study the clinical characteristics of dengue fever (DF) and evaluate the metabolomic changes. The medical records of 50 patients presented with dengue virus 2 (DENV-2) RNA positive were reviewed. Serum was collected from 20 patients diagnosed with DF and 20 healthy volunteers. The changes in serum metabolites were explored by UPLC-Q-TOF MS spectrometer. Fever, nausea and vomiting, asthenia, skin hyperemia were the most common clinical symptoms. The results of the laboratory examinations showed that leukopenia (76%, 38/50), thrombocytopenia (92%, 46/50), increase of aspartate aminotransferase (96%, 48/50), increase of free fatty acids (FFA) (68%, 34/50), increase of creatine kinase (CK) (44%, 22/50) were more common. Based on orthogonal projection to latent structures-discriminant analysis (OPLS-DA), 2 up-regulated and 15 down-regulated metabolites were identied, contributing to DF progress to some extent. Among them, LysoPC (18:2(9Z,12Z)) were highly positive correlated with PE (21:0/20:5(5Z,8Z,11Z,14Z,17Z)) and 13'-Hydroxy-alpha-tocopherol. The identied biomarkers were mainly involved in glycerophospholipids metabolism pathway. The


Introduction
Dengue is one of the important mosquito-borne viral diseases, posing a global health threat to public health. The accurate data on prevalence rate is di cult to obtain due to a vast majority of asymptomatic and self-managed patients. It is estimated that there are 390 million dengue virus (DENV) infections (60 million symptomatic infections) per year in 130 countries [1][2]. Most notably, the mortality is increased from 960 to more than 4032 during 2000-2015 [3]. In addition, economic burden of dengue is substantial. A systematic analysis showed the annual global cost of dengue was $8.9 billion (95% UI, 3.7 billion- 19.7 billion) in 2013 [4]. In China, high risk areas of dengue outbreaks were Guangzhou and Yunnan Provinces [5]. However, three outbreaks, especially large epidemic in 2017, were reported in Zhejiang Province with a typical subtropical climate [6].
Dengue virus (DENV) with 4 serotypes (DENV-1, DENV-2, DENV-3, DENV-4) could cause a wide spectrum of disease, including dengue fever (DF), dengue hemorrhagic fever (DHF) and dengue shock syndrome (DSS) [7]. Delayed diagnosis of dengue could increase the risk of severe dengue. Recently, virological tests and serological tests are used for diagnosis of DENV infections in clinic. However, these methods vary in their sensitivity and speci city at different stages [8]. Therefore, it is urgent to develop a rapid and accurate tool for detection of DENV infections in the early stage. The aim of present study was to analyze the clinical characteristics of DF retrospectively and identify metabolites of serum in order to provide rapid diagnosis of DF.

Medical Records
The medical records of 50 hospitalized patients presented with DENV-2 RNA positive were reviewed during the outbreak attacked Hangzhou, China in 2017.
The data of demographics, clinical characteristics, laboratory results, treatment history, hospitalization and clinical outcomes were collected.

Sample preparation and UPLC-Q-TOF MS spectrometer test
Serum samples were collected from 20 hospitalized DF patients (DF group) and 20 healthy volunteers (HC group) at The First A liated Hospital, College of Medicine, Zhejiang University from August to November in 2017. Serum samples were pretreated with triple volumes of acetonitrile and centrifuged at 15,000 rpm for 15 min to extract supernatant. The samples were tested by UPLC-Q-TOF MS spectrometer (Waters, Manchester, UK) using ACQUITY UPLC HSS T3 C18 column (100 mm × 2.1 mm, 1.8 μm) (Waters, Manchester, UK).

Data processing and multivariate analysis
Nuclear Magnetic Resonance (NMR) data were recorded on AV-400 spectrometer and proton coupling patterns were subjected to multivariate analysis. Metabolites quanti cation were identi ed using Chroma TOF (V 4.3x, LECO, United States) for deconvolution and peak matching. Further sorting and annotation were interpreted with Human Metabolome Database (HMDB) (http://www.hmdb.ca/) and LIPID MAPS (http://www.lipidmaps.org/). Differential metabolites responsible for distinguishing between DF patients and healthy volunteers were analysis by principal component analysis (PCA), partial least squares discrimination analysis (PLS-DA), and orthogonal projection to latent structures-discriminant analysis (OPLS-DA) using metaX (http://metax.genomics.cn). R2Y was used to estimate the goodness of t, and Q2Y was used to assess the quality of prediction [9]. In addition, the value of variable importance in the projection (VIP) greater than 1 and log2FC greater than 2 or lower than 0.5 was considered as signi cant contribution to classi cation. Furthermore, the potential biomarkers involved in metabolic pathways was performed based on Metabolomics Pathway Analysis (MetPA) (https://www.metaboanalyst.ca/). A p value < 0.05 was considered statistically signi cant.

Quality Control (QC) analysis and metabolites annotation
There was no statistical difference in age and gender between DF group and HC group. The QC samples were used to evaluate the repeatability and stability

Discussion
The global burden of dengue has increased 400% during 13 years, presenting a great challenge to prevention and control [10]. The current diagnostic procedures for DENV infections had different their own limitations [8]. The results of our retrospective study indicated that the clinical symptoms were mild-tomoderate for most hospitalized patients during the outbreak in Hangzhou. Some degree of leucopenia and thrombocytopenia, increase of transaminase, FFA and CK were common during the febrile phase. In addition, the changes in serum metabolites were investigated by UPLC-Q-TOF MS spectrometer. There were 17 potential biomarkers were identi ed, mainly involved in glycerophospholipid metabolism pathway.
In the present study, all enrolled patients were infected by DENV-2. Previous study also demonstrated DENV-2 was the dominant serotype in Zhejiang Province, however, other three serotypes were identi ed during the same period [11]. Most people infected by DENV remain asymptomatic or self-limited [10]. We found fever, nausea and vomiting, asthenia, skin hyperemia were the most common clinical symptoms in hospitalized patients. These symptoms were relieved after supportive treatments. However, critical phase of dengue could be developed from febrile, especially for infants, old age, pregnant women with hypertension or vascular disorders. [10]. Although there were 26% patients with hypertension in our study, none of hospitalized patients progressed to severe dengue.
Signi cantly lower white blood cell and platelet were observed in present and previous studies [12]. In addition, laboratory investigations showed transaminase increased, especially for aspartate aminotransferase, indicating the combination of liver and musculoskeletal involvement [13].
Previous metabolomic studies detected serum metabolome and lipidome changes in patients infected with DENV [14]. The changes in host lipid homeostasis could not only satisfy the replication of offspring virus, but also support the energy of lipolysis supply [15][16]. In present study, 2 up-regulated and 15 downregulated serum metabolites were contributed to DF. The down-regulated serum metabolites were phosphatidylcholine (PC), phosphatidylethanolamine (PE), sphingomyelin (SM), lysophospholipids (LPL), which might be attributable to the alternation of cell membrane composition caused by DENV, consistent with previous results in human serum [14]. In addition, glycerophospholipid metabolism pathway was the main pathway of metabolic disorder in DF. Furthermore, the potential glycerophospholipids metabolites have been identi ed as useful biomarkers for detection of DHF from DF and HC [17]. The glycerophospholipids were linked to thrombocytopenia, vascular permeability and viral spherule membrane formation during DENV-2 infection [18]. In addition, the decrease of glutathione and total antioxidant capacity in patients with dengue fever leads to mild disorder of antioxidant system [19]. 13'-Hydroxy-alpha-tocopherol regarding as an antioxidant was reported to be bene cial to thrombocytopenia due to the increase of platelets [20]. The reason for this phenomenon is associated with the antioxidant effect of vitamin E on the in ammatory response of viral infection. The increase of FFA and the abnormality of glycerin phospholipid in clinic further con rmed the important relationship between lipid metabolism and virus reproduction after dengue virus infection.
Although the differences of metabolites were identi ed for DF from HC, there were several limitations in our study. Firstly, all patients were infected by DENV-2 due to the outbreaks in 2017. Second, the clinic data were collected and analyzed retrospectively. Third, the sample size was small and no patients progressed to severe dengue. Therefore, the results of this study need higher quality studies for con rmation in the future.

Conclusions
In conclusion, the major clinical symptoms of DF were fever, nausea and vomiting, asthenia, skin hyperemia, leukopenia, thrombocytopenia, and increase of transaminase. In addition, DENV-2 induced alterations in human glycerophospholipids, especially for the decrease of PC, PE, SM, LPL. These results should be supported in further animal infection models and even prospective randomized controlled clinical trials.

Declarations
Ethical Approval. This study was approved by the recommendations of the Ethics Committee of The First A liated Hospital, College of Medicine, Zhejiang University (No. 2020622). To protect personal privacy, identifying information of each patient in the electronic database was encrypted.
Consent for publication. All authors have seen and approved the content and ful l the journal's requirements for authorship.
Availability of data and materials. All data in the manuscript.

Competing Interests. None
Funding. This work was supported by the National Science Foundation for Young Scientists of China (No. 81803589).
Author Contributions. WY and YQQ developed the concept and designed the experiments. JPT and ZW collected clinical data from the patient records. JPT, JJZ and LZ did UPLC-Q-TOF MS spectrometer test. WY and ZW analyzed the metabonomic data. JPT and WY wrote the paper. All authors discussed the results and implications and commented on the manuscript at all stages.