Danshensu for Myocardial Ischemic Injury: Preclinical Evidence and Novel Methodology of Quality Assessment Tool

Background: Danshensu (DSS) possesses unique bioactivity on the cardiovascular system. However, there is a lack of systematical summary of DSS for acute myocardial ischemia injury and no quality assessment tool for the systematical review of cell experiments. Here, we aimed to assess the preclinical evidences and possible mechanisms of DSS for myocardial ischemia injury, and to develop a quality assessment tool for the systematical review of cell experiments. Methods: Thirty-two studies with 473 animals and 134 cells were identified by searching seven databases. All data analysis was performed using RevMan 5.3. CAMARADES 10-item checklist was used to assess the methodological quality of animal experiments. A new 10-item checklist was first developed to assess the methodological quality of cell studies. Results: The score of study quality ranged from 3 to 7 points in animal studies, while the cell studies scored 3–6 points. Meta-analysis showed that DSS had significant effects on reducing myocardial infarct (MI) size in vivo, and increasing cell viability and reducing apoptosis rate in vitro compared with controls (P < 0.01). The possible mechanisms of DSS for MI are improving circulation, antioxidant, anti-apoptosis, anti-inflammatory, promoting angiogenesis, anti-excessive autophagy, anti-calcium overload, and improving energy metabolism. Conclusions: DSS could exert cardioprotective effect on myocardial ischemia injury, and thus is a probable candidate for further clinical trials andtreatment of AMI. In addition, the newly devloped 10-item checklist for assessing methodological quality of cell study that recommened to use the sysmatic review of cell studies.


INTRODUCTION
Acute myocardial infarction (AMI) is a major cause of death and disability worldwide (Mozaffarian et al., 2016). During AMI, the damage inflicted on the myocardium results in two processes: ischemia and the following reperfusion (I/R) (Ibanez et al., 2015). Thus, timely myocardial reperfusion using either thrombolytic therapy or primary percutaneous coronary intervention (PPCI) is the most effective therapy for limiting myocardial infarct (MI) size, preserving left-ventricular systolic function and reducing the onset of heart failure (Frohlich et al., 2013). However, the mortality and morbidity of patients remain significant with 9% death and 10% heart failure at 1 year after AMI. The fact that a therapeutic intervention administered solely at the time of myocardial I/R injury can reduce MI size by up to half only suggests that myocardial I/R injury may account for up to 50% of the final MI size (Bulluck et al., 2016). Although the process of myocardial reperfusion has been optimized by advances in stent technology, new antiplatelet drugs and novel antithrombotic agents, the studies on drug for necrosis and inflammation of cardiac myocytes during AMI made little evident progress. And there is still no effective therapy for preventing myocardial I/R injury in PPCI patients . Thus, it is urgent to seek new cardioprotective strategies to improve myocardial salvage and cardiac function when AMI happens.
Phenolic and flavonoid compounds with specific structures exhibit strong activity. For example, agrimonolide, curcumin, emodin, resveratrol, baicalin and meletin exerted antiinflammatory, antioxidant, anti-apoptosis effects and regulated metabolism through multiple intracellular signaling pathways (Chen et al., 2016(Chen et al., , 2018aTeng et al., 2016a,b). Previous studies showed that the application of these compounds is beneficial to various diseases (Chen et al., 2018b;Teng and Chen, 2018). Danshen root, Radix Salviae Miltiorrhizae, the dried root of Salvia miltiorrhiza, has been widely used to treat various vascluar diseases including coronary heart disease worldwide . Previous studies showed that one of the main water-soluble phenolic components from Danshen, Danshensu (DSS) [(3-(3,4-dihydroxyphenyl)-lactic acid)], exerted potential cardiovascular protective effects both in vitro and in vivo Cui et al., 2014;Li et al., 2016;Yin et al., 2017;. The specific chemical structures of Danshensu and its derivatives are listed in Table 1. Futhermore, preclinical systematic review can evaluate the efficacy and mechanism of drugs more systematically, establish a test field for further animal experiments, provide reliable information for drug research, and lay a foundation for future clinical research (Sena et al., 2014). what's more, there are no quality assessment tools for a systematical review of cell experiments. Thus, we aimed to provide the preclinical evidence by assessing the efficacy and mechanisms of DSS for expermental AMI, and develop a quality assessment tool for a systematical review of cell experiments.

METHODS
The prensent study was conducted according to the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guidelines (Liberati et al., 2009).

Search Strategies
The following databases were searched from the inception to February, 2018. They are PubMed, CENTRAL, Web of Science, Chinese National Knowledge Infrastructure, Wanfang Data Information Site, VIP information database and Chinese Biomedical Literature Database. The keywords used are as follows: (1) "Danshensu OR Salvianic acid A OR tanshinol OR 3,4-dihydroxyphenyl" AND "myocardial infarction OR myocardial ischemia OR myocardial I/R" limited on animals; (2) "Danshensu OR Salvianic acid A OR tanshinol OR 3,4dihydroxyphenyl" AND "hypoxia OR injury OR damage OR impair" limited on cells. The literature search in this article was last updated in February, 2018.

Eligibility Criteria
Inclusion criteria were prespecified as follows: (1) experimental animal models of myocardial ischemia or myocardial cell injury models; (2) analyzed treatments received the DSS [to enhance stability and activities of danshensu, Danshensu derivatives Cui et al., 2014; have been widely developed and thus was also included]; (3) comparator interventions received vehicle or non-functional liquid with equal volume or conventional medicine or no treatment; (4) the primary outcome measures were MI size in animal model, and the viability and/or apoptosis rate in cell model. The second outcome measures were biochemical markers of MI, left ventricular ejection fraction (LVEF), left ventricular shortening fraction (LVFS), and/or level of STsegment depression, and mechanism indices. Exclusion criteria were prespecified as follows: (1) experimental models unrelated to myocardial ischemic injury; (2) the use of DSS in combination with other drugs in treatment group; (3) no control group or comparing DSS with other herbs in control group; (4) duplicate publication; (5) case report, clinical trial and review article.

Data Extraction
The detailed information of included studies were collected by two indenpendent authors. For animal studies, standard extraxted form on the basis of our previous publication was adopted (Zhang K. J. et al., 2017). For cell studies, the information extracted from included studies was as follows: (1) the first author's name and year of publication; (2) the features of cells, such as name, source (organism, age, tissue) and primary/subcultured; (3) model of cell injury; (4) the information of treatment/control group, including therapeutic drug, dosage, method of administration, duration of treatment; (5) mean value and standard deviation of outcomes.
If the outcomes were displayed through multiple time points, only the peak time point was included. Only the data of highest dose was selected, if the experimental group included gradient doses of drug therapy. Owing to some data were only displayed by graphics, the correspondence authors were be contacted for further information. If there were no responding, the data in the graph were measured by the digital ruler software.  Wang et al., 2017Tang et al., 2018Cui Q. B. et al., 2013Xu et al., 2013 Flexible nano-liposom of DSS (The specific chemical structure is unknown) Chen et al., 2014

Risk of Bias in Individual Studies
The risk of bias (ROB) in animal studies was refered to the ten-item scale using our previously publications (Zhang K. J. et al., 2017). Owing to absence of the ROB assessment scale in cell experiments, we developed a novel 10-item checklist as follows: (A) peer reviewed publication; (B) use appropriate cells to study; (C) cell lines with reliable source or validated by appropriate methods; (D) assess toxicity of treatment on cells; (E) culture environment (culture media/sera, pH/CO 2 and temperature); (F) random allocation to treatment or control; (G) blinded induction of model; (H) blinded assessment of outcome; (I) calculation of the sample size necessary to achieve sufficient power; and (J) statement of potential conflict of interests. Each item was awarded one point. All studies were evaluated by two independent authors. Possible divergences in the evaluation process were properly addressed through contacting with correspondence authors.

Statistical Analysis
All data analysis was implemented by RevMan 5.3. We calculated the standard mean difference (SMD) with 95% Confidence Intervals (CIs). Heterogeneity was assessed using the Cochrane Q-statistic test and the I 2 -statistic test. When I 2 > 50%, indicating substantial heterogeneity and a random-effects model was adopted. Conversely, a fixed-effects model was used. Sources of heterogeneity were searched as far as possible and subgroup analysis was carried out when necessary. Funnel plots were used to intuitively reflect publication bias when more than ten studies were identified. Sensitivity analysis was carried out to prove the robustness of our main results. P < 0.05 was considered statistically significant.
For cell studies, 257 potentially relevant studies were identified through database searching, of which 39 were duplicated. After preliminary screening of the titles and abstracts, we excluded 14 studies because of case report, clinical observation, and review article. Then secondary screening was conducted by reading the full text of the remaining 25 studies, nine studies were excluded because of: (1) failed to obtain full text; (2) not DSS intervention; (3) inappropriate cell model; (4) combined use of other drugs in treatment group; (5) compared with herbs or active herbal compounds; (6) no control group; (7) master dissertation or doctoral dissertation; (8) no available data. Ultimately, 16 eligible studies Guo et al., 2006;Dong et al., 2009;Zhao et al., 2012;Cui G. et al., 2013;Cui Q. B. et al., 2013;Xu et al., 2013;Yin et al., 2013;Fan et al., 2016;Hu et al., 2016;Meng et al., 2016;Wang et al., , 2017Gao et al., 2017;Tang et al., 2018) were selected ( Figure 1B).

The level of ST-segment elevation and/or LVEFand/or LVFS
One study  showed DSS significantly inhibit the ST-segment elevation compared with control (P < 0.001).
Anothe study  showed positive effects of DSS on improving LVEF and LVFS compared with control (P < 0.05).

Summary of Evidence
This is the first preclinical systematic review to estimate the efficacy and mechanisms of DSS for Myocardial ischemic injury through both in vivo and in vitro. Thirty-two studies with 473 animals and 134 cells were selected. The quality of the included studies was generally moderate. The present study indicated that DSS exerted cardioprotective function in myocardial ischemic injury, mainly through improving circulation, antioxidant, antiapoptosis, anti-inflammatory, promoting angiogenesis, antiexcessive autophagy, anti-calcium overload and improving energy metabolism.

Limitations
First, only English and Chinese database were searched, which may lead to selective bias (Guyatt et al., 2011); Second, negative studies were defficult to publish, which may overstimate the true effecacy of DSS (Franco et al., 2014); Third, no study had used an animal with comobidity, such as diabetes, hypertension, or hyperlipidemia (Heusch, 2017); Finally, primary adult cardiomyocytes were considered to be the suitable subjects to validate cardioprotection in vitro (Lecour et al., 2014), but only one study  used it.

Implications
The damage inflicted on the myocardium during AMI is a complex process, which involves multiple factors. The main mechanisms of injury are energy metabolism disorders, excessive oxy radical, calcium overload, inflammation and mitochondrial dysfunction (Turer and Hill, 2010). Thus, cardiovascular protection drugs generally work through one or combined aspects of the above targets (Heusch, 2015;Ibanez et al., 2015;Qi and Young, 2015). The present study reveals that DSS can reduce MI size, increase cell viability, and reduce apoptosis in myocardial ischemic injury through multiple targets. Hence, our findings provide a preclinical evidence-based approach to develop DSS for acute MI. The possible mechanisms of DSS for myocardial ischemic injury are summarized as follows: (1) improving circulation by expanding the coronary artery and antiplatelet action; (2) antioxidant by promoting the Nrf2 accumulation in nucleus through activation of PI3K/AKTsignaling pathway; (3) anti-apoptosis via upregulation of Bcl2/Bax and reducing the expression of caspase 3; (4) anti-inflammatory; (5) promoting angiogenesis by motivating the homing of stem cells and increasing the expression of VEGF; (6) anti-excessive autophagy via activation of m-TOR signaling pathway; (7) anti-calcium overload by reducing the expression of TRPC6 via inhibiting JNK/NF-KB signaling pathway; (8) improving energy metabolism.
Preclinical systematic review is a novel means that provides robust and informative evidence for animal researches. It contributed to decisions on the utility of further preclinical experiments. Additionally, important information about the translational probability of bench to bed was presented according to the preclinical evidence rating (Sena et al., 2014). In this study, the average CAMARADES score of the included studies is generally medium and still needs further improvement. For animal studies, the main defects are lacking of sample size estimation, blinded induction of model, and blinded assessment of outcome (Zhang K. J. et al., 2017). In addition, all the animal experiments are performed in young and healthy small animals which lack the comorbidities, such as aged, diabetes, hypertension and hyperlipidemia (Van Hout et al., 2016;Heusch, 2017). Cardioprotective effects from a longer-term perspective, such as chronic inflammation, remodeling, heart failure or mortality are not considered (Heusch, 2017). These are main failure factors to translate cardioprotective strategies from the animal study to the clinical practice (Lecour et al., 2014;Heusch, 2017). Reporting guidelines set a detailed checklist of predetermined criteria to make reports of biomedical research more complete and transparent, and thus increasing their value in Scientific exploration, clinical practice and inform policy (Moher et al., 2015). In 2010, the Animal Research: Reporting of in vivo Experiments (ARRIVE) guidelines were published on the initiative of the UK National Centre for the Replacement, Refinement and Reduction of Animals in Research (Kilkenny et al., 2010). The ARRIVE guidelines are organized into twenty sections, covering the key points on reporting study design, experimental procedures, and experimental animals to improve comprehensive and transparent on reporting (Karp et al., 2015). Items 6 and 10 of ARRIVE guideline emphasizes the importance of sample size calculation and blinding. Therefore, we suggest the further animal studies should refer to the ARRIVE and sample size estimation, blinded induction of model, and blinded assessment of outcome should be focused on. In addition, the following factors need to be considered: (1) using large mammals that are closer to humans in anatomy and physiology; (2) experimental animals have comorbidities with aged, diabetes, hypertension, hyperlipidemia or other risk factors; (3) use of primary outcome measures that are closer to clinical practice.
For cells, there are lacking of randomization and blinding (Zhang K. J. et al., 2017). In addition, primary adult cardiomyocytes were considered to be the suitable subjects to validate cardioprotection in vitro (Lecour et al., 2014), however, only one study  used it. Thirdly, most cell experiments mimicked the acute I/R injury model using compounds, such as t-BHP, which cannot reflect the complex conditions in vivo (Lecour et al., 2014). Thus, we suggest the further studies focused on randomization and blinding in cell experiments, and it is best to use primary adult cardiomyocytes in ischemic buffer with oxygen-deficient environment, then mimics perfusion using normal buffer with normoxic environment.
Cochrane ROB tool that designed for evaluating the quality of clinical trials was used worldwide (O'Connor et al., 2011). Similarly, some tools have been developed to evaluate the quality of preclinical animal experiments in vivo (Hooijmans et al., 2014;Zhang K. J. et al., 2017). However, there are no quality assessment tools for cell experiments. Thus, we developed an evaluation tool for quality of cell experiments. Detailed description of the assessment tool is as follows: (1) peer review is an inseparable part of science. Non-peer reviewed articles are considered to be of inferior quality and negligible value (Costello et al., 2013). Therefore, it is necessary to use peer review as a part of quality assessment (Table 5A: peer reviewed publication); (2) The use of cell models to simulate clinical disease is important in modern research. However, they do not accurately represent organisms and cannot reproduce the complex interactions in vivo. Thus, it is necessary to select cells that can maximally simulate clinical disease. For example, primary adult cardiomyocytes are most closely related to the characteristics of the adult heart in cell experiments (Lecour et al., 2014). Hence, in the present study, they are the appropriate cells for simulating myocardial ischemic injury. Here we call on researchers to choose the appropriate cells based on actual situation (Table 5B: use appropriate cells to study); (3) quality control of cell lines is essential to ensure the reproducibility of biomedical research. Although cell line certification has been widely recommended for many years, misidentification and cross-contamination are still a serious problem. Science and nature, the two top journals, also have repeatedly called attention to this issue (Freedman et al., 2015;Neimark, 2015;. Thus, we recommend using cell lines with reliable source or validated by appropriate methods (Table 5C: cell lines with reliable source or validated by appropriate methods); (4) efforts to improve the safety of drugs have always been the key to pharmacological research (Kalgutkar and Dalvie, 2015;Nussinov and Tsai, 2015;Insel et al., 2018). Therefore, special attention should be paid to the toxicity of treatment in preclinical studies ( Table 5D: assess toxicity of treatment on cells). (5) The culture environment is crucial for cell growth and subsequent experiments. It is necessary to select a suitable culture environment according to the cell species [ Table 5E: culture environment (culture media/sera, pH/CO 2 and temperature)]; (6) the items of random allocation and blinding are essential to RoB (O'Connor et al., 2011;Zhang K. J. et al., 2017) (Table 5F: random allocation to treatment or control; 5G: blinded induction of model; 5H: blinded assessment of outcome); (7) experiments should have enough sample size to allow for proper statistical analysis and to ensure that the results are robust and reliable. But to avoid waste of resources and prevent exposure of research participants to risk associated with the interventions, the calculation of the sample size necessary to achieve sufficient power is needed (Stokes, 2014) (Table 5I: calculation of the sample size necessary to achieve sufficient power); (8) conflicts of interest are one source of bias. They exist when professional judgment or actions regarding a primary interest may be unduly influenced by a secondary interest (Bero, 2017) (Table 5J: statement of potential conflict of interests).
Improving the transparency and quality of reporting is the imperative of modern scientific research. Numerous publications have emphasized the lack of transparency in reporting, but researches in life sciences, still often lack reporting on the design, conduct and analysis of the experiments (Landis et al., 2012). This resulted in the Consolidated Standards of Reporting Trials (CONSORT) Statement, which was first published in 1996 (Begg et al., 1996), and had been proven to improve the transparency of clinical researches in subsequent implementation (Plint et al., 2006). In 2004, CAMARADES 10-item checklist was first developed to assess the quality of animal experiments (Macleod et al., 2004). Subsequently in 2012, CAMARADES called for transparent reporting to optimize the predictive value of preclinical research (Landis et al., 2012). They recommend that, at minimum, a rigorous study design consist of randomization, blinding, sample-size estimation and the handling of all data should be reported. In the present study, we emphasized the importance of the above recommendations, the items F, G, H, and I in the cell study checklist are consistent with the CAMARADES claim. In addition, misidentification, crosscontamination, cytotoxicity, and culture environment are core issues widely existed in cell studies. Therefore, we recommend that authors should report on exact source of the cells, the method of verifying the cell type, toxic effect of the drug on the cells, and the detailed data of the cell culture environment. We hope and expect that this tool will fill in the blank areas for preclinical systematic review and meta-analysis of cell studies, subsequently improving the transparency and quality of the in vitro experiments. Furthermore, this tool should be tested for its applicability and validity in practice. We look forward to the feedback from users on this new field of evidence-based cell experiments to make this tool more rigorous, scientific, and valuable.

CONCLUSION
The findings of present study demonstrated that DSS exerted cardioprotective function in AMI, largely through improving circulation, antioxidant, anti-apoptosis, anti-inflammatory, promoting angiogenesis, anti-excessive autophagy, anticalcium overload and improving energy metabolism. Thus, DSS is a probable candidate for further AMI treatment and clinical trials. In addition, the newly devloped 10-item checklist for assessing methodological quality of cell study that recommened to use the sysmatic review and meta-analysis of cell studies.

AUTHOR CONTRIBUTIONS
XB and QZ contributed equally to this work. XB, QZ, QT, PZ, ZZ, GZ, and YW designed the study. QT, PZ, and ZZ collected the data. XB, QZ, and QT performed all analyses. XB and QZ wrote the manuscript.