Autologous Mesenchymal Stem Cell Transplantation in Multiple Sclerosis: A Meta-Analysis

Multiple sclerosis (MS) is considered to be a central nervous system (CNS) chronic inflammatory demyelinating disease, affecting more than 2 million individuals worldwide. In this meta-analysis, we aimed to assess the safety and efficacy of autologous mesenchymal stem cells (aMSCs) in treating MS patients. The PubMed, Embase, Cochrane, Web of Science, and Clinical Trial databases were searched in September 2019. The analysis was conducted for three endpoints: transplant-related mortality (TRM), rate of disease progression, and no evidence of disease activity (NEDA) status. RevMan and the metaprop command of the meta package in R was used in assessing the efficacy and safety of aMSCs. Subgroup analyses were performed for exploration of heterogeneity regarding outcomes. Nine studies comprising 133 patients were included in the meta-analysis. The pooled estimate of TRM was 0% (95% confidence interval (CI) 0%–0.3%). The rate of progression was 16% at 6 months (95% CI 10%–27%) and 35% at 1 year (95% CI 27%–46%). Lower 6-month and 1-year progression rates were significantly associated with intrathecal injection (p = 0.02; p = 0.003). The pooled proportion of NEDA patients at 6 months was 72% (95% CI 58%–89%) and at 1 year was 62% (95% CI 42%–81%). Cell transplantation with aMSCs in MS patients is safe, with the largest benefit profile obtained in patients with aMSCs intrathecal injection.


Introduction
Multiple sclerosis (MS) is a chronic inflammatory demyelinating disease in the central nervous system (CNS), leading to demyelination, neurodegeneration, and gliosis [1]. At present, MS affects approximately 2,500,000 people worldwide, especially young adults, which presents a heavy burden for families and society [1,2]. Although most effective diseasemodifying therapies (DMTs) improve cognitive and physical disability in MS patients, there has been limited efficacy in these treatments and they are often related with side effects [1,3]. Therefore, we need to develop a more effective treatment for MS patients.
Mesenchymal stem cells (MSCs) are derived from bone marrow (BM) and other tissues, including fat, muscle, umbilical cord blood (UCB), dermis, and dental pulp [4,5]. MSCs have been used therapeutically in over 700 clinical trials which include heart diseases, stroke, and autoimmune diseases (multiple sclerosis, graft-versus-host disease, systemic lupus erythematosus, and so on) [5,6]. In addition to their multipotential abilities, MSCs have also displayed immunoregulatory and neuroprotective properties [3,4,7]. Related studies on MSCs using the experimental autoimmune encephalomyelitis (EAE), an animal model for MS, have shown that increased numbers of regulatory T cells (Tregs), decreased demyelination, and inflammation [6]. MSCs also modify microglia cells by increasing the M2 and decreasing the M1 phenotype, exerting a protective role for MS [8]. Currently, the meta-analysis of autologous hematopoietic stem cell transplantation (aHSCT) has been evaluated in treating MS patients [9].
Therefore, the aims of this meta-analysis were to evaluate the efficacy and safety of MSCs by systematically collecting and summarizing all the evidence published about the outcomes including disease activity, Expanded Disability Status Scale (EDSS), and adverse events.

Search Strategy.
We systematically searched all the published studies reporting aMSCs for MS on PubMed, Embase, Web of Science, Cochrane, and Clinical Trials (published up to September 25, 2019). Our search strategies including free words and MeSH terms (multiple sclerosis [MeSH] and transplantation [MeSH] and patient [MeSH]): (a) "mesenchymal stem cells" or "MSCs", and (b) "multiple sclerosis" or "MS", (c) patient. In addition, other included studies were collected manually from references of eligible studies or other articles related to this topic. The abstracts were examined independently by two authors (Y Zhou and X Zhang).

Selection Criteria.
Inclusion criteria for this meta-analysis were as follows: any study (a) on MS; (b) on patients receiving aMSC transplantation; (c) including data on efficacy of aMSC transplantation; (d) reporting mortality and clinical follow-up; (e) including more than 5 patients; and (f) published in English.
The exclusion criteria were as follows: the study (a) did not meet the inclusion criteria; (b) was an editorial, review, case report, or abstract, or was from a clinical conference, comments, or congresses; and (c) involved nonhuman studies.
2.3. Data Extraction. The extracted data from eligible studies were as follows: (a) identity: authors, years, number of included patients; (b) baseline characteristics of patients (age, EDSS, proportion of patients with SPMS, disease duration, and relapses in the previous year); (c) treatments: transplantation methods, cell doses, follow-up period; and (d) outcomes: EDSS, mortality, and disease activity. The data from included articles were independently extracted and processed by two authors (Y Zhou and X Zhang). Any disagreement between the two authors was settled by consultation with a third author.

Types of Outcome
Measures. The following endpoints were used to assess the safety and efficacy of aMSC on transplantation MS patients. The transplant-related mortality (TRM) was defined as death within 100 days of aMSCs transplantation, and the data of overall mortality (OM) was derived from the entire follow-up of all included studies. Progression events were defined as increasing 1 point (baseline EDSS ≤ 5:5) or 0.5 points (baseline EDSS > 5:5), and the EDSS score was, respectively, assessed at 6 or 12 months. No evidence of disease activity (NEDA) in MS patients was defined as without any disability progression, clinical relapse, or new MRI lesion (T2 or gadoliniumenhancing) over a limited period.

Statistical
Analysis. The meta-analysis was completed using RevMan5.3 (The Cochrane Community, London, United Kingdom) and the metaprop command of the meta package in R to assess the safety (the transplant-related mortality) and efficacy (progression rate and the proportion of no evidence of disease activity). We used odds ratio (OR) and related 95% CI to analyze the data and log transformation to calculate pooled proportions under the fixed and random effects model. The chi-squared value test and inconsistency index statistic (I 2 ) were used to evaluate statistical heterogeneity across each study. A value of p < 0:1 or I 2 > 50% was indicated significant heterogeneity, and then the related data was analyzed with the random effects model. Otherwise, we used the fixed effects model. Subgroup analyses used the chi-squared test to analyze the subgroup data, with p < 0:05 indicating statistical significance.
2.6. Standard Protocol Approvals, Registrations, and Patient Consents. Written informed consent was acquired from all patients, and original data was approved by the local ethics committees. Any identifiable patient data were not found in this meta-analysis.

Study Screen.
The search strategy retrieved 16, 22, 69, 46, and 9 studies from the PubMed, Embase, Web of Science, Cochrane, and Clinical Trials databases, respectively. A total of 82 duplicated studies were excluded, and the titles and abstracts of the remaining articles were screened by two reviewers. Subsequently, 25 studies with full text were evaluated. Among these studies, 3 were excluded due to less than 5 cases, 3 due to lack of full texts, 5 due to being review and conference articles, and 5 due to relevant data not being extractable. Finally, in this meta-analysis, 9 studies [10][11][12][13][14][15][16][17][18] on aMSCs for the treatment of MS were analyzed which included 133 transplanted patients ( Figure 1).

Study
Characteristic. The included 9 studies, along with the basic characteristics and sample sizes, are summarized in Table 1. Nine studies were all reported in English. Eight out of 9 studies were open-label uncontrolled studies and 1 was a randomized study. The MSCs in the eight studies were derived from the bone marrow, whereas one study was from adipose. Five studies used intrathecal injection for cell transplantation, and a mean number for every patient was less than 50 × 10 6 cells. Three studies adopted intravenous injection with cell dosages of 1-4 × 10 6 MSCs/kg, while only one study included both methods. The follow-up period ranged from 6 to 96 months. For reporting the outcomes, all included studies reported the change of EDSS and magnetic resonance imaging (MRI), and they also reported some mild adverse effects.

Subgroup Analysis.
We performed subgroup analysis based on age, EDSS, MS duration, and transplantation method. The subgroup analysis of disease progression rates at 6 months indicates intrathecal injection was more beneficial than intravenous injection on the disability progression rates (p = 0:02, I 2 = 82:7%). However, there was no significant difference between the age, EDSS, and MS duration of different groups (Figures 5(a)-5(d)). The subgroup analysis of disease progression rates at 1 year also indicates intrathecal injection was more beneficial than intravenous injection on the disability progression rates (p = 0:003, I 2 = 88:9%). There were no significant differences between the age, EDSS, and MS duration of the different groups (Figures 6(a)-6(d)).
We performed subgroup analysis on the percentage of patients with NEDA at 6 months and 1 year, and the results showed that there were no significant differences between the age, EDSS, MS duration, and transplantation method of the different groups (data not shown).

Publication
Bias. The assessment of publication bias cannot be conducted due to the number of included studies being less than 10.

Discussion
In recent years, the increase in reported MS patients treated with aMSC transplantation has led to debates about this therapeutic approach as a progressive MS treatment [19].
In this meta-analysis, we hope to provide useful information based on published articles.  there are some limitations: (1) the studies used different MSC types and sources which included adipose tissue and bone marrow, and the MSC transplantation procedure was not uniform in each study; (2) the characteristics of MS patients in each study were different: the majority of patients in the study were SPMS patients and the degree of disability was relatively advanced. Some studies even only included SPMS patients, and this type is highly likely to continue to progress after MSC transplantation; (3) the doses and transplantation methods of MSCs were different in each study; (4) this study included only 9 studies and 133 patients and was unable to perform some important subgroup analyses, such as different disease subtypes and cell sources; (5) the EDSS in assessing disease progression was significantly subjective, had poor reproducibility and low consistency, and it is difficult to compare and combine disease progression in different states [20,21].
In this meta-analysis, we chose to use the rate of disease progression and percentage of NEDA as the main efficacy endpoints. The rate of disease progression was 16% at 6 months, while the rate reached 35% at 1 year. However, the 2-year disease progression rate was only 17.1% in the metaanalysis of autologous hematopoietic stem cell transplanta-tion (aHSCT) for MS, and the 5-year disease progression rate was only 23.3% [9]. Considering that the above results may be related to the proportion of SPMS patients included in the study, since the study of aMSC transplantation in the treatment of MS mainly included SPMS patients, the proportion was mostly higher than 50%, and several studies even reached 90%-100%. Moreover, in the meta-analysis of aHSCT-treated MS patients, it was found that the maximal benefit profile was obtained in MS patients with relapsingremitting stage [9]. It is not easy to compare the treatment outcomes of aMSCs with those reported drugs for MS in clinical trials, since the majority of patients enrolled in receiving aMSC transplantation generally have much higher EDSS scores and more aggressive stages of MS than the ones accepting disease-modifying drugs (DMDs). Among approved DMDs for MS, especially for SPMS, there has yet to be identified the most effective drugs to prevent disease progression [22]. The rate of progression after 6 months and 1 year was less than 10% and 20%, respectively, in SPMS patients treated with siponimod [23]. At the same time, in the clinical trial on the effect of natalizumab on disease progression in SPMS patients, the rate of progression after 2 years was 16% [24]. It       for the two studies was 3.0-6.5, while the median EDSS is higher than 6 in this meta-analysis [23,24]. Despite this, the results of this meta-analysis are robust, and the efficacy of aMSC transplantation for MS should be interpreted cautiously due to the limited studies and lack of control data. In terms of safety, no transplant-related deaths were observed in all included studies. However, in the metaanalysis of aHSCT for MS, the pooled estimate of transplant-related mortality (TRM) was 2.1% [9]. There were 2 deaths in the overall studies during the follow-up, one due to severe spastic quadriplegia and the other due to choking on food [16]. In addition, there were some adverse effects reported in the enrolled studies which displayed as temporary and light side effects, including fever, slight headache, backache, nausea/vomiting, iatrogenic meningitis, and urinary/respiratory infection. All in all, it is safe to treat MS with aMSC transplantation.
In fact, subgroup analysis shows that the rate of disability progression of aMSC transplantation with intrathecal injection was less than that with intravenous injection both at 6   Hetorogeneity: tau 2 = 2.06; chi 2 = 11.84, df = 2 (p = 0.03); I 2 = 83% Test for overall effect: Z = 1.46 (p = 0.14) 53.5% 0.26 [0.12, 0.49] Subtotal (95% CI) Hetorogeneity: tau 2 = 0.00; chi 2 = 0.96, df = 3 (p = 0.81); I 2 = 0% Test for overall effect: Z = 3.94 (p < 0.0001) Hetorogeneity: tau 2 = 0.81; chi 2 = 18.87, df = 6 (p = 0.004); I 2 = 68% Test for overall effect: Z = 3.06 (p = 0.002) Test for subgroup differences: chi 2 = 0.00, df = 1 (p = 0.96), I 2 = 0%    months and 1 year, and there was statistical difference (p < 0:05). In this analysis, the 6-month disease progression rate for EDSS ≤ 6:5 is less than EDSS > 6:5 and the 1-year disease progression rate for age ≤ 44 y is also less than age > 44 y, unfortunately there were no statistical differences (p > 0:05). Therefore, the largest benefit profile of aMSC transplantation for MS can be obtained in patients with aMSC intrathecal injection. Some data also demonstrates that intraperitoneal injection of adipose MSCs in EAE resulted in higher Treg cells expression and IL-4 production compared with intravenous route [25]. Animal studies on MS have shown that MSCs suppressed pathogenic effector CD4 + T cells, increased numbers of Tregs, and modulated effector CD8 + T cell subsets [6,26]. In addition, MSCs exert a protective role in EAE by modifying microglia cells, including increasing the M2 and decreasing the M1 phenotype [8]. Furthermore, it should be noted that MSCs can be obtained from many human tissues,  especially adipose tissue, which includes a majority of stem cells. Therefore, although this meta-analysis on MSCs in the treatment of MS has a relatively high rate of disability progression compared to other treatment regimens, we still have reason to expect the efficacy of MSC. In this meta-analysis, we demonstrate the safety of aMSCs for the treatment of MS patients. However, the efficacy of aMSC transplantation in MS should be interpreted cautiously compared with those reported treatment regimes, and more random clinical trials are needed to clarify the efficacy of MSCs for treating MS. This meta-analysis also shows the significant association of aMSC intrathecal injection with lower disability progression rate. All in all, comprehensive consideration of results indicate that aMSC transplantation is safe, and to better evaluate the efficacy more studies need to be investigated in the future.

Conclusions
Cell transplantation with aMSCs for treatment of MS is safe, with a largest benefit profile being obtained in patients with aMSC intrathecal injection.