Acupuncture for neurogenesis in experimental ischemic stroke: a systematic review and meta-analysis

Acupuncture has been used for patients with stroke and post-stroke rehabilitation for thousands of years. Previous studies reported that acupuncture enhanced stroke recovery through neurogenesis. Hence, we conducted a systematic review and meta-analysis for preclinical studies to assess the current evidence for acupuncture effect on neurogenesis in treating ischaemic stroke. Studies were obtained from six databases, including PubMed, EMBASE, Cochrane Library, Chinese National Knowledge Infrastructure, VIP information database, and Chinese Biomedical Literature Database, Ultimately, 34 studies containing 1617 animals were identified. Neurogenesis markers of Brdu, Nestin, PSA-NCAM, NeuN and GFAP were selected as major outcomes. The pooled results of 15 studies marked with Brdu showed significant effects of acupuncture for improving proliferation when compared with control groups (P < 0.01); 13 studies marked with Nestin showed significant effects of acupuncture for increasing proliferation when compared with control groups (P < 0.01); 4 studies marked with PSA-NCAM showed significant effects of acupuncture for enhancing migration when compared with control groups (P < 0.01); 4 studies marked with NeuN showed significant effects of acupuncture for stimulating differentiation when compared with control groups (P < 0.01). The findings suggest that acupuncture is a prospective therapy targeting neurogenesis for ischemic stroke.

Ischemic stroke is one of the leading causes of death and long-term disability worldwide 1 . The only Food and Drug Administration-approved thrombolytic for treating ischemic stroke is recombinant tissue plasminogen activator (rtPA). Unfortunately, rtPA must be administered within 4.5 hours of stroke onset to be effective, and it often results in intracranial hemorrhage 2 . These two factors largely restrict the clinical use of rtPA. Thus, given its widespread occurrence and devastating impact of sufferers and their caregivers, better treatment of ischemic stroke is urgently needed.
Regenerative strategies, particularly with regard to neurogeneisis, offer long-term hope for many patients who have suffered a stroke. Neurogenesis naturally occurs throughout adulthood in the subventricular zone (SVZ) and the subgranular zone (SGZ) of the dentate gyrus (DG) in the hippocampus of the brain. Furthermore, SVZ-derived neural progenitor cells (NPCs) can migrate. When ischemia occurs, genetic fate-mapping studies show that these cells can migrate ectopically into the ischaemic penumbra of the striatum and cerebral cortex 3 , and later, new neurons are generated to replace those lost. This neurogenic response, however, is generally considered too weak to yield enough neurons for compensation and recovery of lost neurocytes and their functions 4 . Instead, there has been some success in supplying neural stem/progenitor cells (NSCs/NPCs) to replace the injured neural cells after acute stroke 5 . Therefore, promoting endogenous neurogenesis appears to be a promising therapeutic strategy for treating stroke, and any treatment that can increase NCSs will test this hypothesis 6 .
Acupuncture is a kind of classical traditional Chinese medicine (TCM) that has been used for patients with stroke and post-stroke rehabilitation for thousands of years 7 . One systematic review of preclinical study indicates that acupuncture decreases infarct volume, ameliorates neurological impairment and plays a neuroprotective role in animal models with acute ischemia 8 . However, the scientific mechanisms of acupuncture for stroke have so far not been fully clarified. Recent reports suggest that acupuncture can enhance stroke recovery through neurogenesis 9 . Therefore, we undertake a preclinical systematic review and meta-analysis to assess the current evidence for acupuncture effect on neurogenesis in treating ischaemic stroke. This review is valuable in several aspects. It can inform the planning and improve the likelihood of success of future clinical trials, provide empirical evidence to improve the rigor of the conducting and reporting of preclinical research, and contribute to both 'reduction' and 'refinement' in animal experimentation 10 .

Subgroup analyses and sensitivity analyses.
To explore other factors which potentially affected the outcome measures, we stratified the included studies according to variables as shown in Table 3. During the course of neurogenesis, the differentiation process chronologically depended on mature proliferation and migration. Based on the above results of overall effects for acupuncture on neurogenesis, we conducted the subgroup analyses and sensitivity analyses of Brdu, Nestin, and PSA-NCAM markers based on the time point of 7 days after ischemic injury, and of NeuN marker at 14 days after ischemia. However, no subgroup analysis and sensitivity analysis of GFAP marker were conducted at any time point as the pooled result of the overall effect did not have significant difference. (1) For Brdu + cells, the stratified analysis showed that significant differences in effect sizes were observed relative to the brain site of the target detection (P < 0.001), animal species (P = 0.01), the type of MCAO model (P < 0.001), anesthetic used (P < 0.001), duration of treatment (P < 0.001), and the baseline median of positive cells (P = 0.01). No significant differences in effect sizes was observed relative to animal weight (P = 0.79). (2) For Nestin + cells, the stratified analysis showed that significant differences in effect sizes were observed relative to the brain site of the target detection (P < 0.001), animal species (P = 0.006), anesthetic used (P < 0.001), animal weight (P = 0.001), duration of treatment (P < 0.001), and the baseline median of positive cells (P = 0.02). No significant differences in effect sizes was observed relative to the type of MCAO model (P = 0.14). (3) For PSA-NCAM + cells, the stratified analysis showed that significant differences in effect sizes were observed relative to animal species (P < 0.001), the type of MCAO model (P < 0.001), and the baseline median of positive cells (P = 0.01). No significant differences in effect sizes were observed relative to the brain site of the target detection (P = 0.15), and animal weight (P = 0.11). (4) For NeuN + cells, the stratified analysis showed that significant differences in effect sizes were observed relative to the brain site of the target detection (P < 0.001), the type of MCAO model (P < 0.001), animal weight (P < 0.001), and duration of treatment (P < 0.001). No significant differences in effect sizes was observed relative to animal species (P = 0.22).
Sensitivity analyses showed that the results did not substantially alter after removing any one trial. Furthermore, respectively excluding the one in the three studies 11,13,27 that contributed least to the overall estimate of each marker also did not materially alter the results (Brdu: MD

Meta-regression.
To further explore heterogeneity among the studies, meta-regression was conducted to investigate the effect of characteristics on the positive markers (Table 3). (1) For Brdu + cells, we found that animal species (Adjusted R 2 = 14.33%), the type of MCAO model (Adjusted R 2 = 23.17%), duration of treatment (Adjusted R 2 = 31.16%), and the baseline median of positive cells (Adjusted R 2 = 24.43%) accounted for a significant proportion of the between-study heterogeneity in studies. (2) For Nestin + cells, anesthetic used (Adjusted R 2 = 6.22%), duration of treatment (Adjusted R 2 = 7.70%), and the baseline median of positive cells (Adjusted R 2 = 48.78%) were significant sources of heterogeneity. Meta-regression was not conducted in PSA-NCAM, NeuN and GFAP markers as less than 10 studies were included in each outcome measure.

Discussion
Efficacy of acupuncture. This systematic review found that acupuncture improved neurological deficits and reduced brain edema in experimental ischemia, and that the mechanisms mostly involved with enhancing endogenous neurogenesis including proliferation, migration and differentiation of NSCs.
Methodological considerations. There were some limitations to consider in interpreting our study.
Firstly, we were unable to meta-analyze all trials due to insufficient data in several primary studies included [41][42][43][44] . Although they all claimed positive effects of acupuncture on neurogenesis after ischemia. Secondly, publication bias existed by the asymmetry of the funnel plot and statistical analysis with Egger's test. Some non-positive studies have been missed inevitability, as negative findings are less likely to be published 45 . Publication bias inflated estimates by a flawed methodological design in smaller studies, and/or a lack of publication of small trials with opposite results. Thus, the overall positive findings of treatment with acupuncture should made with caution. Thirdly, the qualitative score ranging from 3 to 6 points indicated poor methodological quality in the included studies. Randomization, blinding, sample-size estimation and the handling of all data should be reported clearly, as these are the core standards of rigorous study design 46 . Although 32 studies described random allocation to a treatment group, none of them reported details of how the animals were randomized. Only one study described masked induction of stroke model. None of the included studies reported the sample-size calculation or whether investigators were blind to the outcome. Low quality of the included studies suggested that the results should be interpreted with caution.
Heterogeneity interpretation. The heterogeneity to some extent existed in the meta-analyses of neurogenesis that represented with five specific biological markers, i.e. Brdu, Nestin, PSA-NCAM, NeuN and GFAP. According to the sub-analyses and meta-regression, however, a significant linear relation was demonstrated between the brain site of neurogenesis and acupuncture therapy, animal species, type of MCAO model, anesthetic, duration of treatment and the baseline median of positive cells. In addition, duration of treatment and the baseline median of positive cells were the most significant proportion of the between-study heterogeneity in studies. Higher level of median baseline accounted for greater beneficial effect, which would contributed to different trials adopted various counting methods to calculate the positive cells, as some studies counted cells under microscope with different folds, visual fields and slices of sample, or even with unequal sizes of interested areas. Also, either optical density or grayscale was used in image analysis.
Implication for further studies. In the present study, acupuncture was found to be effective in neurogenesis, particularly from day 7 to day 14 after ischemia; neurocyte proliferation peaked at day 7 while differentiation peaked at day 14. These results suggested an optimum time window in stroke for acupuncture therapy. Meanwhile, we found that the most used acupuncture points were Baihui (GV 20), Zusanli (ST 36), Dazhui (GV 14) and Quchi (LI 11). They were used individually or in combination. Based on TCM theory, Baihui (GV 20) is located on the highest place of the head where all the yang meridians meet 47     clear the mind, lift the spirits, tonify yang, strengthen the ascending function of the spleen, eliminate interior wind, and promote resuscitation 48 . Thus, Baihui (GV 20) is a principle acupoint which is often selected for stroke patients. After all, several included studies speculated on how acupuncture enhanced neurogenesis, and proposed the following possible biological mechanisms: up-regulating the expression of the signaling molecules of the PKA-CREB transduction system in the hippocampus of the ipsilateral hemispheres 11 , enhancing the brain-derived neurotrophic factor (BDNF) and vascular endothelial growth factor (VEGF) signaling pathways 44 , improving the levels of neurotrophic factors BDNF, NGF, bFGF, NSE and FGF-2 13,31,32,34,36,40 , activating the ERK signal transduction pathway 15 , and up-regulating the expression of GAP-43, TrkB and down-regulating the expression of Rho Kinase 31 . Thus, we suggest that the endogenous neurogenesis after ischemia should be further performed with their physiological function by electrophysiology and other methodologies. Meanwhile, future well-designed studies are warranted to fully clarify the mechanisms of acupuncture inducing endogenous neurogenesis.

Conclusion
Our findings indicate that acupuncture ameliorates neurological deficits and reduces brain edema in experimental ischemic stroke, and the mechanisms positively correlate with endogenous neurogenesis, in which acupuncture therapy can promote the proliferation, migration and differentiation of NSCs.

Methods
Search strategy. Studies of acupuncture on neurogenesis treatment after experimental ischemic stroke models were derived from PubMed, EMBASE, Cochrane Library, Chinese National Knowledge Infrastructure (CNKI), VIP information database, and Chinese Biomedical Literature Database. We also manually searched abstracts of academic seminars and reference lists of identified publications, and we contacted researchers to retrieve unpublished and/or unreported materials that might be relevant. All searches were performed from the establishment of each database up to August 2015. Database searches were conducted using the following terms:   (acupuncture OR electroacupuncture (EA) OR moxibustion) AND (stroke OR cerebral ischemia) AND (neurogenesis OR neural regeneration OR neurotization). Chinese Databases were also searched using the above search terms in Chinese. All searches were limited to studies on animals.

Eligibility.
We included all identified studies of acupuncture in neurogenesis of experimental ischemic stroke where the outcome was measured as bromodeoxyuridine (Brdu) and/or Nestin and/or polysialylated form of the neural cell adhesion molecule (PSA-NCAM) and/or neuronal nuclear antigen (NeuN) and/or glial fibrillary acidic protein (GFAP). Neurological function score (NFS) or/ and infarct volume (IV) or/ and brain water content (BWC) were also extracted as outcome indicators, as appropriate. Brdu, being used for monitoring cell  proliferation, is a synthetic thymidine analog that incorporates DNA of dividing cells during the S-phase of the cell cycle. Nestin is defined as a class VI intermediate filament protein that is known as a NSC/ NPC marker during development of the central nervous system (CNS). Polysialic acid (PSA) is a linear homopolymer of alpha2-8-N acetylneuraminic acid whose primary vector in vertebrates is the neural cell adhesion molecule (NCAM). PSA-NCAM participates in neural plasticity and neurogenesis and it is particularly considered towards cell migration. NeuN, a DNA-binding protein that is abundantly and exclusively expressed in neurons of the CNS, is considered to be a hallmark of neuronal differentiation during development. Not only do neurons immunolabeled NeuN differentiate from gliocyte and endothelial cells, but also NeuN identifies morphologically distinct classes of neurons in brain structures. GFAP is an intermediate-filament protein expressed uniquely in astrocytes and vulnerable to reactive gliosis that follows injuries to the CNS. It is described as one of the labels of astrocytic differentiation in vertebrates 53 , 54 . NFS, IV and BWC indicators are also known to contribute to the efficacy of acupuncture in neurogenesis of ischemic stroke 8 .  Inclusion criteria were as follows: (1) studies of experimental ischemic stroke induced by temporary middle cerebral artery occlusion (MCAO) or permanent MCAO; (2) randomized studies of manual acupuncture (MA) or electroacupuncture (EA); and (3) studies using at least one of the following neurogenesis indexes Brdu, Nestin, PSA-NCAM, NeuN and GFAP as outcome measures; (4) studies results comparing experimental groups with control groups receiving placebo/ sham acupuncture or no treatment. Exclusion criteria were that the studies used a non-focal cerebral ischemia model (such as global, traumatic models, or hypoxic-ischemic models), that they used young animals, that they had no control group, or that the same study was published in duplicate. Data extraction. The following details were extracted from each study: (1) publication year and the first author's name, type of ischemic stroke (transient, or permanent), and ischemic time; (2) animal model details, namely species, sex, weight and number of animals per group; (3) treatment information, including timing for initial treatment, types and methods of treatment, duration of treatment, and comparable treatment of control group; (4) outcome measure information, including the phase of neurogenesis, outcome indexes, timing for outcomes assessments and the corresponding p values. (5) For each comparison, we extracted data of mean value and standard deviation from each treatment and control group of each study. If the data for meta-analysis were missing or only expressed graphically, we tried to contact the authors for the missing, or more specific information. Otherwise we only performed qualitative analysis.
Quality assessment. Two authors independently assessed the methodological quality of the included articles according to the CAMARADES 10-item checklist 55 : (1) peer-reviewed publication; (2) statements of temperature control; (3) randomization to treatment or control group; (4) blinded induction of model; (5) blinded assessment of outcome; (6) use of anesthetic without significant intrinsic neuroprotective activity; (7) appropriate animal model; (8) sample size calculation; (9) compliance with animal welfare regulations; and (10) declaration of potential conflict of interests. Each study was given an aggregate quality score out of a possible total of 10 points, and the group median was calculated. Discrepancies were resolved after discussion between the two authors or were referred to an arbitrator.

Statistical analysis.
Meta-analyses and sub-analyses were performed using RevMan V.5.1 software, and analyses of public bias and meta-regression were performed using STATA/SE 12.0 software. Outcome indicators of Brdu, Nestin, PSA-NCAM, NeuN, GFAP, NFS, IV, and BWC were all considered as continuous data, and these indicators conducted an estimate of the combined effect sizes utilizing mean difference (MD) with the random effects model, respectively. Publication bias was assessed with a funnel plot and Egger's test 56 . The I 2 statistic was used for assessment of heterogeneity. To clarify the impact of factors potentially modifying the outcome measures, we also conducted sensitivity analyses and subgroup analyses according to the following variables: encephalic region of neurogenesis, animal species, anaesthetic used, MCAO model induction, acupuncture method used, and baseline median of positive cells. Difference between n groups was assessed by partitioning heterogeneity and using the χ 2 distribution with n-1 degrees of freedom (df), where n meant the number of groups.