Abstract

Extracts or active components from Acorus gramineus Aiton (EAAGA) have been clinically used for cognition impairment more than hundreds of years and are still used in modern times in China and elsewhere worldwide. Previous studies reported that EAAGA improves cognition impairment in animal models. Here, we conducted a preclinical systematic review to assess the current evidence of EAAGA for cognition impairment. We searched 7 databases up until June 2019. Methodological quality for each included studies was accessed according to the CAMARADES 10-item checklist. The primary outcome measures were neurobehavioral function scores evaluated by the Morris water maze test, electrical Y-maze test, step-down test, radial eight-arm maze test, and step-through test. The secondary outcome measures were mechanisms of EAAGA for cognition function. Finally, 34 studies involving 1431 animals were identified. The quality score of studies range from 1 to 6, and the median was 3.32. Compared with controls, the results of the meta-analysis indicated EAAGA exerted a significant effect in decreasing the escape latency and error times and in increasing the length of time spent in the platform quadrant and the number of platform crossings representing learning ability and memory function (all ). The possible mechanisms of EAAGA are largely through anti-inflammatory, antioxidant, antiapoptosis activities, inhibition of neurotoxicity, regulating synaptic plasticity, protecting cerebrovascular, stimulating cholinergic system, and suppressing astrocyte activation. In conclusion, EAAGA exert potential neuroprotective effects in experimental cognition impairment, and EAAGA could be a candidate for cognition impairment treatment and further clinical trials.

1. Introduction

With the average life expectancy increasing, there is concern about the proportion of cognitive impairment in the global population, which results from degeneration of the brain and very high prevalence in elderly individuals [1]. The World Health Organization estimates that the number of people over the age of 60 will be around 2 billion in 2050, while the number of cognitive impairment patients is expected to rise rapidly along with the aging population worldwide [2, 3]. However, so far, clinical trials have not identified efficacious neuroprotective therapies for cognitive impairment patients [4]. Thus, given the huge translational gap between the animal studies and clinical trials, seeking or developing innovative neuroprotectants is urgently needed.

For more than a millennium, traditional Chinese medicine (TCM), a main form of complementary and alternative medicine, has been used in Asian countries, especially in China, Japan, and Korea, to alleviate various symptoms of cognitive deficits and to facilitate learning and memory [5]. Acorus gramineus Aiton (AGA) (record 2322 (http://www.theplantlist.org.)), the dry rhizomes of Acorus gramineus Solander (Shi Changpu), is listed officially in the Chinese Pharmacopoeia and used in oriental medicines for more than hundreds of years to treat neurological disorders. AGA possessed various pharmacological effects on the central nervous system, including neuroprotective effects [6, 7], central inhibitory effects [8], inhibitory effects on excitotoxic neuronal death [9], and stroke [10], and amelioration in learning and memory [5]. AGA may be effective for the improvement of amnesia [9]. AGA contains different extract fractions: volatile oil, composing mainly of β-asarone (63.2–81.2%), and α-asarone (8.8–13.7%) [11], as well as water extract, ethyl ether extract, ethyl acetate extract, N-butanol extract, and the defatted decoction fractions. AGA is often used as a component in some Chinese herbal formulas. Among 75 of the most famous Chinese herbal formulas characterized as improving intelligence both in ancient and modern time in China, more than half contain AGA, such as Kai-Xin-San [12] and Chong-MyungTang [13].

Systematic reviews are believed to be preferred; only data that from systematic reviews will be considered as the highest level of medical evidence basis for the levels of evidence from the Centre of Evidence-Based Medicine in Oxford [14]. Preclinical systematic reviews are a powerful approach to analyze and synthesize the results of an intervention from animal data into a useful document that can help to shape further basic research, optimize the experimental studies, and enhance the success rate of future clinical trials [15]. Thus, we conducted a preclinical systematic review to assess the current evidence of extracts or active components from Acorus gramineus Aiton (EAAGA) and active component for animal models of cognitive impairment.

2. Materials and Methods

2.1. Search Strategies

Experimental studies of EAAGA for cognitive impairment were identified in the databases, including PubMed, Embase, Web of Science, Wanfang database, China National Knowledge Infrastructure (CNKI), CBM, and VIP information database. All searches were performed from inception to June 2019. Studies about assessing the effectiveness of AAGA for improving cognitive function impairment in animals were identified. The search terms were as follows: (Acorus tatarinowii Schott OR Rhizoma acori graminei OR Acorus calamus OR Acorus gramineus Soland OR acorus gramineus aiton OR Acori graminei rhizoma OR Acori tatarinowii rhizoma OR grassleaf sweetfalg Rhizome) AND (cognitive function impairment OR amnesia OR dementia OR Alzheimer’s disease).

2.2. Inclusion Criteria

Experimental studies on EAAGA for cognitive impairment models were included, regardless of publication status or animal species, gender, age, and methods of model establishment. The primary outcome measurements were Morris water maze test (MWM test), electric Y-maze test (EY-M test), radial eight-arm maze test (RAM test), Step down test (SD test), and/or Step through test (ST test). The secondary outcome measures were mechanisms of EAAGA for learning and/or memory function.

2.3. Exclusion Criteria

Exclusion criteria were prespecified as follows: (1) the article was a review, case report, comment, clinical trial, abstract, or editorial; (2) the article was a clinical or in vitro study; (3) the article was not a research about cognitive impairment model; (4) EAAGA was used as combination; (5) there was no control group; and (6) the article was a duplicate publication.

2.4. Data Extraction

The information of each included study was extracted: (1) author and publication year, animal model species, method of anesthesia, and random method; (2) characteristics of animals, including species, sex, animal number, and weight; (3) treatment information from treatment and control groups, including drug, dose, method of treatment, timing for initial treatment, frequency, and duration of treatment; and (4) outcome measures, sample size, and corresponding data including mean value, standard deviation, and intergroup differences. If outcomes were presented at different time points, we extracted data from the last time point. If studies utilized dose gradient of the drug, we extracted data from the highest dose of EAAGA and active component since the dose-response relationship. If the data were incomplete or presented in graphs, we tried to contact the authors for data needed or calculated using relevant software. Information of the mechanism studies of EAAGA and active component for cognitive impairment models among the included articles was extracted.

2.5. Quality Assessment

The methodological quality of included studies was evaluated by two independent reviewers using Collaborative Approach to Meta-Analysis and Review of Animal Data from Experimental Studies (CAMARADES) 10-item checklist [16]. For calculating an aggregate quality score, each item of this scale was attributed one point.

2.6. Statistical Analysis

Meta-analysis was conducted via RevMan version 5.3. To estimate the effect of EAAGA on cognitive impairment, the random effects model and standard mean difference (SMD) with 95% confidence intervals (CIs) were calculated. Heterogeneity was assessed via statistics test. If probability value was less than 0.05, the difference was considered statistically significant. In addition, to explore potential sources of high heterogeneity, subgroup analyses were performed according to animal species and models. Difference between groups was determined by partitioning heterogeneity and utilizing the distribution with degrees of freedom (df).

3. Results

3.1. Study Selection

We identified 2368 potentially relevant papers after systematical search from six databases. After removing duplicates, 1887 studies remained. By reading titles and abstracts, 1602 articles were excluded that were reviews, case reports, comments, abstracts, clinical trials, letters, or editorials. After reading the remaining 285 full-text articles, 228 studies were excluded for at least one of following reasons: (1) not an animal study; (2) the article was not a research about cognitive impairment; (3) the study did not access the effects of AGA or active component on the animal model of cognitive impairment; (4) EAAGA was not used as a monotherapy; and (5) lack of control group. Ultimately, 34 eligible articles [5, 6, 10, 11, 1746] were selected (Figure 1).

3.2. Characteristics of Included Studies

Sixteen studies [5, 6, 10, 11, 1727, 37] were published in English, and 18 studies were in Chinese between 1999 and 2019. In total, 34 studies with 1431 animals were included. Ten species were referred, including Sprague-Dawley (SD) rat (, 16.49%), Wistar rats (, 9.08%), Kunming mice (, 37.04%), ICR mice (, 16.49%), NIH mice (, 11.74%), AβPP/PS1 double-transgenic mice (, 1.82%); APPswe/PS1dE9 double transgenic mice (, 1.54%), C57BL/6 mice (, 1.68%), senescence-accelerated prone-8 (SAMP8) mice (, 1.82%), and FMR1gene knock mice (, 2.31%). The weight of SD rats ranged from 200 g to 650 g, the weight of Wistar rats used ranged from 30 g to 250 g, and the weight of mice ranged from 17 g to 50 g. Twenty-two studies used male rodents, 1 study used female rodents, 5 study used both female and male rodents, and the remaining 6 studies did not provide gender details. Sodium pentobarbital was used to induce anesthesia in 8 studies, and chloral hydrate was used in 2 studies [20, 21], 1 study [41] used phenytoin sodium, 1 study [17] used CO2, and 1 study [10] used isoflurane, while the remaining 21 studies did not report the type of anesthetics. Cognitive impairment models were induced by lead [17], noise stress [18], LPS [19], amyloid beta 1-42 [11, 21, 26, 28, 29, 37, 41, 46], D-gal plus AlCl3 [22], scopolamine [5, 24, 30, 3436, 42, 45], ethanol [5, 32, 3436], sodium nitrite [5, 32], corticosterone [23], Ibotenic acid [25], chronic restraint stress [31], pentobarbital sodium [32], D-galactose [33, 38], AlCl3 [40], streptozotocin (STZ) [43], pent ylenetet razol (PTZ) [44], and NaNO2 [3436]. As an intervention, fourteen studies [6, 17, 20, 22, 23, 26, 27, 32, 35, 37, 39, 41, 42, 46] used β-asarone, eight studies [18, 19, 21, 24, 33, 38, 40, 44] used α-asarone, three studies [10, 25, 44] utilized AGA, twelve studies [5, 11, 22, 2832, 35, 36, 43, 45] used essential oil, seven studies [11, 28, 29, 3336] researched water extract, four studies [11, 28, 29, 32] used defatted decoction, and one study [18] researched ethyl acetate extract. Normal distilled water control was used in 2 studies [17, 33]; Tween 80 control was used in 6 studies [5, 6, 18, 20, 27, 32]; normal saline control was used in 24 studies; 0.5% methylcellulose solution containing 1% Tween 80 control was used in 1 study [24], and 2% propylene glycol containing 2% polyethylene glycol stearate control was used in 1 study [43]. Neurobehavioral function indices as primary outcome measures were carried out by the Morris water maze test (MWM test) (), step-down test (SD test) (), electrical Y-maze test (EY-M test) (), step-through test (ST test) (), and radial eight-arm maze test (RAM test) (). The characteristics of the 34 studies are shown in Table 1.

3.3. Study Quality

The quality scores of the 34 included studies varied from 1/10 to 6/10 with the average of 3.32. One study [40] got 1 point; 11 studies [29, 3136, 38, 39, 42, 44] got 2 points; 9 studies [5, 6, 18, 24, 25, 27, 30, 43, 45] got 3 points; 4 studies got 4 points; 7 studies got 5 points; and 2 studies [20, 37] got 6 points. Thirty-four studies were published. Sixteen studies described control of temperature [6, 10, 1726, 30, 37, 41, 45]. Random allocation was declared in 28 studies [5, 6, 11, 17, 1923, 2628, 3039, 4146]; 1 study [42] used random block allocation method, and 2 studies used the method of random digit table [34, 41]. Two studies [23, 37] described the use of blinded assessment of outcome. Thirteen studies did not use anesthetics with significant intrinsic neuroprotective activity, and the remaining 21 studies did not report the type of anesthetics [5, 6, 18, 19, 24, 27, 3040, 4245]. Sixteen studies reported compliance with animal welfare regulations [5, 10, 11, 1722, 24, 27, 28, 37, 41, 43, 45]. Four studies mentioned statement of potential conflict of interests [11, 20, 28, 37]. None of the included studies reported allocation concealment, sample size calculation, and the utilization of animal or model with relevant comorbidities. The quality scores for the included studies are shown in Table 2.

3.4. Effectiveness

The Morris water maze test, including the probe test and the spatial test, was conducted in 28 studies [6, 10, 11, 17, 19, 20, 22, 23, 2531, 3339, 4146]. Twenty-seven studies reported the spatial test using the escape latency as an outcome measure. Meta-analysis of 20 studies with 27 comparisons showed EAAGA significantly decreased the escape latency compared with the control (, , 95% CI [−1.37 to −0.82], ; heterogeneity: , ; ; Figure 2(a)). In the probe test, meta-analysis of 16 studies [17, 19, 20, 22, 26, 27, 2931, 33, 34, 37, 38, 41, 44, 45] with 19 comparisons showed EAAGA were significant for increasing number of platform crossings (, , 95% CI [1.25 to 1.94], ; heterogeneity: , ; ; Figure 2(b)) compared with controls. Meta-analysis of 6 studies [17, 20, 22, 29, 34, 44] with 7 comparisons showed a significant effect of EAAGA in increasing the length of time spent in platform quadrant compared with control (, , 95% CI [0.90 to 2.67], ; heterogeneity: , ; ). As the values of were greater than 50%, we sequentially omitting each study; two studies [20, 22] were removed and markedly reduced the heterogeneity (, , 95% CI [1.55 to 3.12], ; heterogeneity: , ; ; Figure 2(c)). Two studies [20, 22] used relatively large doses of β-asarone that might have potential toxic effects [47]. Meta-analysis of 3 studies [20, 23, 25] for increasing percentage of time in the platform quadrant (, , 95% CI [2.86 to 5.15], ; heterogeneity: , ; ; Figure 2(d)). Three studies [17, 22, 23] showed there were not a significant difference in improving the swimming velocity compared with controls.

The step-down test, including the training test which represents learning ability and retention test which represents memory ability, was conducted in 6 studies [5, 32, 3436, 40]. Meta-analysis of 5 studies with 19 comparisons showed EAAGA were significant for increasing right reaction latency in the retention test (, , 95% CI [0.87 to 1.43], ; heterogeneity: , ; ; Figure 3(a)) and 1 study [5] for increasing right reaction latency () in the training test. Meta-analysis of 3 studies [32, 35, 36] with 16 comparisons showed EAAGA were significant for decreasing the error times (, , 95% CI [−1.30 to −0.83], ; heterogeneity: , ; ; Figure 3(b)) in the retention test and 1 study [5] for decreasing the error times () in the training test.

The electrical Y-maze test was conducted in 3 studies [5, 32, 36]. Meta-analysis of 3 studies showed EAAGA were significant for decreasing error reaction times (, , 95% CI [−1.56 to −0.88], ; heterogeneity: , ; ; Figure 4).

The step-through test was conducted in 4 studies [24, 3436]. Meta-analysis of 4 studies with 7 comparisons showed EAAGA were significant for decreasing latency in the retention test (n = 134, SMD = 1.26, 95% CI [0.81 to 1.71], P < 0.00001; heterogeneity: χ2 = 8.09, df =6 (P = 0.23); ; Figure 5(a)) and 2 studies [35, 36] with 5 comparisons showed EAAGA significantly decreased the number of errors in the retention test (, , 95% CI [−1.45 to −0.60], ; heterogeneity: , ; ; Figure 5(b)) compared with controls.

The eight-arm maze test was conducted in 3 studies [10, 18, 21]. Meta-analysis of 2 studies [10, 21] showed EAAGA were significant for increasing number of correct choices (, , 95% CI [0.00 to 2.29], ; heterogeneity: , ; ; Figure 6(a)) and 2 studies [10, 18] with 3comparisons showed EAAGA significantly decreased the number of errors in the training test (, , 95% CI [−3.36 to −1.37], ; heterogeneity: , ; ; Figure 6(b)) compared with controls.

3.5. Neuroprotective Mechanisms

The mechanisms of neuroprotection of EAAGA on cognitive impairment were studied in 34 included articles [5, 6, 10, 11, 1746] as follows: (1) reduction of oxidative reactions by increasing the activity of SOD [30, 35, 39, 41, 43] activity, while decreasing the activity of SOD and AChE [18, 24], decreasing the levels of MDA [24, 30, 33] and nitric oxide [21], decreasing the mRNA levels of hsp 70, increasing the levels of VC, VE, and GSH, and increasing the activity of CAT and G6PD [18]; (2) inhibition of apoptosis by increasing the mRNA levels of Bcl-2, BDNF, CREB [6, 23, 42], Bcl-w and Bcl-2 [26], and c-jun [35], decreasing the mRNA levels of Bax [23], increasing the expression of BDNF, CREB [23], Bcl-w, and Bcl-2 [26], decreasing the expression of caspase-3, p-JNK [26], and BACE1 [19], and preventing cell loss [10], Aβ, and Tau protein [38]; (3) repression of inflammatory reactions by decreasing the expression of TNF-α and IL-1β mRNA levels [19]; (4) repression of autophagy by decreasing LC3, ROCK, and beclin1 expression and increasing p62, GAP43, MAP2, and SYN expression [27]; (5) protection of cerebrovascular by increasing rCBF and the Na-K-ATP activity, decreasing pyruvic acid contents, and decreasing the mRNA levels of ET-1, eNOS, and APP [22]; (6) promotion of cognitive function by increasing the levels of 5-HT, NE, DA, and NE [5] and suppression of astrocyte activation [37]; (7) stimulation of cholinergic system by increasing AchE and ChAT neurons [25]; (8) improvement of memory impairments through regulation of synaptogenesis, which is mediated via Arc/Arg3.1 and Wnt pathway [17]; (9) neuroprotection through damage of Akt pathway [40]; (10) inhibition of neurotoxicity by decreasing the expression of DCx and nestin, decreasing nestin positive cells [11], decreasing Aβ plaques depositions, and decreasing NOS activity [29]; (11) regulation of synaptic plasticity by increasing the expression of SYP and GluR1 [20, 46] and decreasing the expression of GAP-43 and PSD-95 [46]; and (12) inhibition of chronic stress by decreasing plasma cortisol levels [41]. Characteristics of mechanism studies of EAAGA on experimental ischemic stroke are shown in Table 3 and Figure 7.

4. Discussion

As far as we know, it is the first preclinical systematic review that determined the efficacy of EAAGA for learning and memory function. In the present study, 34 studies with 1431 animals showed that EAAGA significantly improve learning and memory function, suggesting the potential neuroprotective functions of EAAGA in cognitive function impairment. However, given methodological weaknesses, the overall available evidence from the present study should be interpreted cautiously.

Some limitations should be considered while interpreting this study. First, we only searched databases in Chinese and English. The absence of studies published in other languages may cause certain degree selective bias [48]. Second, the methodological quality of included studies showed some inherent drawback. Most of the research had methodological flaws in aspects of blinding, randomization, allocation concealment, sample size calculation, and lacking statement of potential conflict of interests [49, 50]. The studies without adequate sample sizes, allocation concealment, or randomization may result in inflated estimates of treatment efficacy [51, 52]. Lower quality trials could attribute to statistically significant 30–50% exaggeration of treatment efficacy [53]. Third, no study adopted animals with comorbidities, which would have created more relevant models for human pathology [49]. Thereby, the results should be interpreted cautiously.

The poor design of animal research hindered the translation of animal research into effective preclinical drug treatments for human disease [54, 55]. Thus, it is necessary to take a rigor experimental design to overcome methodology quality issues for further research. The Animal Research: Reporting of In Vivo Experiments (ARRIVE) [56, 57] is a reporting guideline consisting of a 20-item checklist that provides recommendations on Introduction, Methods, Results, and Discussion which were recommended to be utilized as guidelines when designing and reporting animal research on EAAGA for improving the cognitive function impediment. Meanwhile, many drugs that exerted significant effects in animal researches failed to translate into effective clinical drug treatments [58, 59]. One of the possible reasons is the application of drug doses and the timing of drug administration in animal models that are inapplicable for human disease [55]. In the present study, doses of EAAGA and timing for initial administration in animal models were inconsistent among the 34 included studies. Thus, we suggest further studies to determinate the optimal gradient doses and timing of administration in animal models of cognition impairment.

The present study showed that EAAGA had cognitive enhancing effects through different mechanisms as follows: (1) reduction of oxidative reactions by increasing the activity of SOD [30, 35, 39, 41, 43] activity, while decreasing the activity of SOD and AChE [18, 24], decreasing the levels of MDA [24, 30, 33] and nitric oxide [21], decreasing the mRNA levels of hsp 70, increasing the levels of VC, VE and GSH, and increasing the activity of CAT and G6PD [18]; (2) inhibition of apoptosis by increasing the mRNA levels of Bcl-2, BDNF, CREB [6, 23, 42], Bcl-w and Bcl-2 [26], and c-jun [35], decreasing the mRNA levels of Bax [23], increasing the expression of BDNF, CREB [23], Bcl-w, and Bcl-2 [26], decreasing the expression of caspase-3, p-JNK [26], and BACE1 [19], and preventing cell loss [10], Aβ, and Tau protein [38]; (3) repression of inflammatory reactions by decreasing the expression of TNF-α and IL-1β mRNA levels [19]; (4) repression of autophagy by decreasing LC3, ROCK, and beclin1 expression and increasing p62, GAP43, MAP2, and SYN expression [27]; (5) protection of cerebrovascular by increasing rCBF and the Na-K-ATP activity, decreasing pyruvic acid contents, and decreasing the mRNA levels of ET-1, eNOS, and APP [22]; (6) promotion of cognitive function by increasing the levels of 5-HT, NE, DA, and NE [5] and suppression of astrocyte activation [37]; (7) stimulation of cholinergic system by increasing AchE and ChAT neurons [25]; (8) improvement of memory impairments through regulation of synaptogenesis, which is mediated via Arc/Arg3.1 and Wnt pathway [17]; (9) neuroprotection through damage of Akt pathway [40]; (10) inhibition of neurotoxicity by decreasing the expression of DCx and nestin, decreasing nestin positive cells [11], and decreasing Aβ plaques depositions, decreased NOS activity [29]; (11) regulation of synaptic plasticity by increasing the expression of SYP and GluR1 [20, 46] and decreasing the expression of GAP-43 and PSD-95 [46]; and (12) inhibition of chronic stress by decreasing plasma cortisol levels [41]. However, cellular and molecular alteration mechanisms of EAAGA and active components for cognition impairment have not been clearly explored yet, which presented an exciting investigative direction of further studies. All 5 measuring methods for learning and memory ability were used in the 34 included studies, which showed that the measuring methods for cognition impairment were inconsistent. The diverse measuring methods for learning and memory ability need further study.

5. Conclusions

Although some factors such as study quality may undermine the validity, EAAGA exert potential neuroprotective effects in cognition impairment. In addition, AGA and active components may be a promising candidate for clinical trials.

Conflicts of Interest

The authors declare no conflicts of interest.

Acknowledgments

This project was supported by the Young and MiddleAged University Discipline Leaders of Zhejiang Province, China (2013277) and Zhejiang Provincial Program for the Cultivation of High-level Innovative Health Talents (2015). We would like to thank LetPub (http://www.letpub.com) for providing linguistic assistance during the preparation of this manuscript. This work was supported by the grant from the National Natural Science Foundation of China (81573750/81473491/81173395/H2902).