Astragaloside IV Exerts Cardioprotection in Animal Models of Viral Myocarditis: A Preclinical Systematic Review and Meta-Analysis

Astragaloside IV (AS-IV), the essential active component of astragalus, possesses diverse biological activities that have beneficial effects against cardiovascular disease. Here, we conducted a preclinical systematic review of 15 studies including 577 animals to establish the efficacy and potential mechanisms of AS-IV for animal models of viral myocarditis (VM). Six databases were searched from inception to October 2018. Application of the Cochrane Collaboration’s tool 10-item checklist and Rev-Man 5.3 software to analyze risk of bias of studies and data on outcome measures revealed study quality scores ranging from 2 to 5. Compared with the control group, AS-IV induced a marked decrease in mortality (P < 0.05), inflammation of myocardium and pathological score (P< 0.05) and cardiac enzymes expression (P< 0.05), and improved the function of the heart (P< 0.05). The potential mechanisms of AS-IV action were determined as anti-remodeling of myocardium (n = 1), anti-virus (n = 2), antioxidant (n = 2), anti-inflammatory (n = 6), anti-apoptosis (n = 1) and alleviation of myocardial fibrosis (n = 2). The collective results indicate that AS-IV exerts cardioprotective effects in animals with VM via multiple signaling pathways.


INTRODUCTION
Myocarditis is pathologically and clinically defined as an inflammatory disease of the heart muscle based on histological, immunological, and immunohistochemical criteria of cardiomyopathy classification from the World Health Organization (WHO) (Pollack et al., 2015). Myocarditis often arises from infection with common viruses, such as respiratory viruses and enteroviruses, and is most commonly associated with coxsackievirus B (Dennert et al., 2008). Myocarditis infection can be subdivided into three pathological stages (Mason, 2003). In the first stage, viral-mediated lysis directly destroys cardiomyocytes, leading to cardiomyocyte damage and cardiac dilation (McManus et al., 1993). The second stage is imbalance of immune regulation resulting from myocardial cell damage (Lawson, 2000). Finally, the typical manifestation of dilated cardiomyopathy (DCM) occurs due to extensive myocardial injury (Caforio et al., 2002). Viral myocarditis (VM) has the characteristics of myocardial inflammation, often developing into chronic dilated cardiomyopathy, acute myocarditis and even congestive heart failure (Kühl and Schultheiss, 2009). VM accounts for 12% of sudden deaths in patients under 40 years of age and is the leading cause of dilated cardiomyopathy, resulting in 50% of patients deaths within 1-2 years after diagnosis (Cooper et al., 2014). The incidence of myocarditis worldwide was reported a 22 cases per 100,000 individuals in 2013, and about 1.5 million patients were diagnosed with myocarditis according to the ninth revised diagnosis of the International Classification of Diseases (GBD 2015 Disease andInjury Incidence andPrevalence Collaborators, 2016). For VM, antiviral drugs, physical activity restriction and myocardial nutrition are the main routine clinical treatments. Patients should follow the current guidelines for heart failure treatment in cases of left ventricular dysfunction or symptomatic heart failure (Ponikowski et al., 2016). However, establishing the potential benefits of immunomodulators and antiviral therapy is currently at the preliminary research stage (Pollack et al., 2015).
AS-IV, the essential active component of astragalus, has been traditionally used to treat various diseases. This compound is frequently reported to exert protective effects on cardiovascular, nervous and immune systems due to its antioxidant and activitiesantiapoptotic activities as well as regulatory effects on calcium balance (Ren et al., 2013). Systematic studies on animal models play an important role in drug research and provide substantial evidence to support application in the clinic (Sena et al., 2014). The purposes of the present study was to investigate the efficacy and potential mechanisms of action of AS-IV in animal models of VM.

METHODS
The study was carried out according to the Preferred Reporting Items for Systematic Review and MetaAnalyses (Stewart et al., 2015).

Search Strategies
Based on the literature retrieval from PubMed, Wanfang database, EMBASE, China National Knowledge Infrastructure (CNKI), Cochrane Library and VIP databases, animal experimental studies of AS-IV for VM were identified. All search strategies were performed from inception until October 2018. We selected the following terms: 1. Astragaloside; 2. Myocarditis; 3. 1 AND 2; 4. Animals NOT humans; 5. 3 AND 4.

Eligibility Criteria
Studies that met the following criteria were included: (1) animal models of VM were established through various ways, (2) treatment groups were administered any dose of AS-IV only, and comparators given non-functional and isasteric liquid (carboxymethyl cellulose or normal saline) or no treatment, and (3) the primary outcome measures were mortality rate, histopathological changes of myocardium, indicators of cardiac ultrasound, cardiac troponin, cardiac enzymes and/or level of ST-segment depression. Cardioprotective mechanisms of AS-IV action against VM were used as the secondary outcome measure. Exclusion criteria were as follows: (1) duplicate publication, (2) not in vivo studies, (3) comparison with other traditional Chinese medicine (TCM), (4) combinations with other medicine, and (5) lack of a control group.

Data Extraction
Two authors independently extracted the following details: (1) name of the first author and publication year, (2) characteristics (species, number, sex, and weight) of the animals used in each study, (3) methods to establish animal models of VM, (4) the therapeutic regimens of treatment and control groups, including method of administration, therapeutic drug dosage and sessions of treatment, (5) primary and secondary outcomes. If the results were obtained at different time-points or following administration of different doses of drug, only the final measured value and data obtained with the highest dose of drug were included. We attempted to contact authors for more information when published data for some records were only shown in a graphical format. In cases where no response was received, data in the graph were measured.

Risk of Bias in Individual Studies
The risk of bias was assessed by two independent authors through applying the ten-item scale (Hooijmans et al., 2014), with minor modifications. Risk of bias of the following domains were assessed: A: sequence generation; B: baseline characteristics; C: allocation concealment; D: random housing and other animal welfare; E: blinding of caregivers and/or investigators; F: random outcome assessment; G: blinding of outcome assessor; H: complete outcome data; I: selective outcome reporting; J: other sources of bias. One point was awarded for each item. Divergent findings between the two authors were settled through resolving by consensus or arbitration by the corresponding author.

Statistical Analysis
RevMan version 5.3 was utilized for statistical analysis. Standardized mean difference (SMD) instead of Mean differences (MDs) was employed as a summary statistic when data were not reported on the same scale. Heterogeneity and choice of effects models were investigated with the aid of standard chi-square test and I² statistic test. Differences were considered significant at P values <0.05.

Study Selection
We identified a total of 540 hints on the basis of pertinent literature retrieval from the databases. After removing 439 reduplicated or irrelevant articles, 101 reports remained. Next, we excluded 53 non-animal studies through screening the titles and abstracts. Overall, 33 articles were excluded after reading the remaining full-text articles owing to: (1) no predetermined outcome index, (2) comparison with other TCM, (3) combination with other medicine, (4) non-viral myocarditis models, and (5) lack of a control group, leading to the final selection of 15 eligible articles Yu et al., 2005;Zhang et al., 2006; The quality of all the included studies was evaluated and scored from 2 to 5. Two studies employed methods of random allocation Liu et al., 2014). Four studies specified the time of model induction before random allocation Chen et al., 2011;Gui et al., 2015). Allocation to different groups was not concealed in any of the studies. Two studies declared compliance with animal welfare regulations Gui et al., 2015). However, no studies reported blinding of caregivers and the methods of random outcome assessment. One study described blinding of outcome assessor . All studies specified complete outcome data and no other sources of bias. Nine studies described the free of selective outcome reporting Yu et al., 2005;Zhang et al., 2006;Liu et al., 2009;Chen et al., 2011;Liu et al., 2014;Gui et al., 2015;Tian et al., 2015;Xiao et al., 2016). The methodological quality is summarized in Table 2.

LVEDs
Data from meta-analysis of two studies Gui et al., 2015) revealed significant suppressive effects of

Cardiac Enzymes and/or Troponin
One study  reported significant reduction CK-MB and CTnI levels of animals with VM treated with AS-IV (P < 0.05).

Summary of Evidence
To our knowledge, this is the first preclinical systematic review (including 15 in vivo studies on a total of 577 animals) focused on evaluating the efficacy of AS-IV against VM and its potential mechanisms of action. Our findings clearly support cardioprotective effects of AS-IV in animal models of VM, mainly through improving anti-remodeling of the myocardium, anti-inflammatory, anti-apoptosis, anti-virus, antioxidant, and alleviation of myocardial fibrosis.

Methodological Considerations
Assessing of the validity of the included studies is a key component of a systematic review, and subsequently affects its analysis, interpretation and conclusions. Based on the Cochrane Handbook for Systematic Reviews of Interventions (Higgins and  Green, 2011), the validity of an individual study should include two dimensions: external validity and internal validity. The former refers to whether the study is asking the appropriate research questions and the latter is in a manner free from the risk of bias of the included studies. In the present study, an appropriate research question was initially set up to evaluate the efficacy and possible mechanism of AS-IV in the treatment of the animals with VM using preclinical systematic review method. When assessments of methodological quality of included preclinical studies, two main risk of bias tools have been developed, including the Collaborative Approach to Meta-Analysis and Review of Animal Data from Experimental Studies (CAMARADES) 10-item checklist (Van et al., 2007) and the SYstematic Review Centre for Laboratory animal Experimentation (SYRCLE)'s Risk of Bias tool (Hooijmans et al., 2014). The CAMARADES 10-item checklist is specific for stroke and neuroprotective study on animal model while the SYRCLE's Risk of Bias tool is extensively used to assess the methodological quality of animal studies. Here, we selected the SYRCLE's Risk of Bias tool to assess the quality of included study and the result indicated that the quality scores of all studies were generally moderate. Thus, we should treat the present results cautiously when applying the study's findings. Accumulating evidence demonstrates that animal reports on biomedical research are often inadequate in many areas (Kilkenny et al., 2010). An earlier survey evaluated 271 randomly selected articles about animal reports by the National Centre for the Replacement, Refinement and Reduction of Animals in Research (NC3Rs). The results showed that 59% articles overall declared the hypotheses or objectives of the study and the characteristics and number of animals used. Randomization (87%) or blindness (86%) was not used in the majority of the articles to reduce bias of animal selection and outcome assessments. In total, 70% of the articles used statistical methods to fully describe publications and present the results in an accurate manner (Kilkenny et al., 2009). Similarly, the articles selected for our analysis had methodological deficiencies, such as seldom using the blind method and allocation concealment. Therefore, concerns of poor experimental design and lack of transparent reports are raised, which can lead to failure of translation of preclinical animal research into clinical treatments for human diseases (Baker et al., 2014). Reporting guidelines develop clear and standard checklist formats to make animal reports more complete and transparent, and thus increase the value of animal reports in clinical practice (Moher et al., 2015). The Animal Research: Reporting In Vivo Experiments (ARRIVE) guidelines, which provide a checklist of 20 items to promote high-quality, comprehensive reporting, have been accepted by over 300 research journals worldwide (Kilkenny et al., 2010). We recommend adherence to these guidelines for the further design of animal studies. In particular, attention should be paid to calculation of sample size, blinding of outcome assessor and blinding of caregivers (Macleod et al., 2004). As a minimum requirement, harmonized animal research reporting principles (HARRP) has been set up based on a comparative analysis of the ARRIVE guidelines by the Gold Standard Publication Checklist (GSPC) and the Institute of Laboratory Animal Research (ILAR) Guidance. Thus, the HARRP provided the minimal reporting standards for animal-based research (Osborne et al., 2018).

Implications
This study confirmed that AS-IV induces a significant reduction in mortality, improves myocardial pathology, and suppresses expression of cardiac enzymes, supporting its utility as a therapeutic option for patients with VM. However, translation of preclinical experiments remains challenging in predicting the effectiveness of therapeutic strategies in clinical trials (Hackam, 2007). The differences in drug doses and time of drug administration between humans and animals are considered the main reasons underlying failure of translation of research from bench to patient bedside (Baker et al., 2014). It is important to note that the 15 articles included in this study are not consistent in terms of drug dose and time of drug administration. Thus, we recommend establishing drug concentration gradients and grouping administration times to establish the optimal treatment strategy. In view of the differences between clinical trials and animal research, further high-quality randomized controlled trials of AS-IV for VM are required.
Several methods are commonly used for establishing animal models of myocarditis. (1) CVB3 or encephalomyocarditis virus (EMCV) injection: the advantages of this model is use of virus, which is close to clinical practice and suitable for studying CVB3 replication. However, this method is associated with high mortality rates and requires high biosafety standards. (2) Reovirus or murine adenovirus type 1 (MAV-1) injection is a unique model of pediatric myocarditis but is not clinically relevant.
(3) Trypanosoma cruzi (T. cruzi) infection model, which could recapitulate the course of Chagas disease, but requires a long period for study. (4) Immunization with α-myosin heavy chain (α-MyHC) or troponin I peptide and complete Freund's adjuvant (CFA). The limitation of this model, which is suitable for studying transition from myocarditis to DCM, is non-physiological disease induction. (5) T cell receptor specific to α-MyHC (TCR-M) transgenic mice. The advantages of this model are biosafety and suitability for studying the pathophysiology of heart-specific T cells. However, the lack of non-specific T cells presents a drawback. (6) Programmed cell death protein-1 (PD-1)/PD-1 ligand (PD-L1) deficiency. This model is suitable for investigating the side-effects of anti-PD-1/ PD-1L therapy but involves multiple organs. (7) A human leukocyte antigen (HLA-DQ8) transgenic mouse model is beneficial for studying cardiac antigen presentation but a humanmouse chimeric system is lacking at present (Błyszczuk, 2019). All the studies included in this review utilized a myocardial model induced by CVB3 injection. Since CVB3 is closely related to the pathogenesis of myocarditis in human, this method of modeling is rational. We therefore recommend application of CVB3 to establish models of VM.

Mechanisms
Based on the findings of the included studies, the possible mechanisms of AS-IV mediated cardiovascular protection are as follows: (1) antioxidant through increasing T-SOD, GSH-PX and CAT levels Luo et al., 2008) to inhibit release of ROS and MPO , (2) anti-inflammatory activity through suppression the TLR4/P38MAPK/MCP-1 Zhou et al., 2012) and IL23/IL17 pathways  and inhibition of NF-κB  as well as MIF  and A20 , (3) alleviation of myocardial fibrosis through inhibiting TGF-β1, Smad2/3 and Smad4 and enhancing Smad7  and increasing TIMP-1to reduce MMP-1 release , (4) prevention of viral spread through increasing IFN-γ  and decreasing CAR expression , (5) inhibition of apoptosis through enhancing the Bcl-2/ Bax ratio , (6) anti-remodeling of myocardium through suppressing the expression of IGF-1, IGF-1R and IGFBP3 .

CONCLUSION
The collective research clearly demonstrates that AS-IV decreases mortality, inflammation of the myocardium, pathological score and cardiac enzyme activity and improves heart function.
Moreover, AS-IV exerts potential cardioprotective function of in VM primarily via improving anti-remodeling of myocardium, anti-inflammatory, anti-apoptosis, anti-virus, antioxidant and alleviation of myocardial fibrosis. Thus, AS-IV is a cardioprotective recruit for further clinical trials on VM.