Current Evidence of Mesenchymal Stem Cells Use in the Treatment of Tendon Disorders: A Systematic Review, Meta-analysis, and Meta-regression of Prospective Clinical Studies

Purpose: Although several studies with animals have reported the effects of mesenchymal stem cells (MSCs) for tendon regeneration, little is known about the efficacy and safety of MSCs in human tendon disorders. We performed this meta-analysis to evaluate the efficacy and safety of MSC therapy in patients with tendon disorders enrolled in prospective clinical studies. Methods: We systematically searched prospective clinical studies investigating the effects of MSCs administration on human tendon disorders with at least a 6-month follow-up period on PubMed-Medline, Embase, and Cochrane Library databases. The primary outcome of interest was the change in pain on motion related to tendon disorders. We performed a pairwise meta-analysis using the fixed-effects model to assess treatment response, which was calculated by the standardized mean difference. Meta-regression analyses were performed to assess the relationship between MSCs dose and pooled effect sizes in each cell dose. Results: Four prospective clinical trials investigating the effect of MSCs on tendon disorders were retrieved. MSCs showed significant pooled effect size (overall Hedge’s g pooled standardized mean difference (SMD) = 1.868; 95% confidence interval [CI], 1.274– 2.462; P < 0.001). The treatment with MSCs improved all the aspects analyzed, i.e. pain, functional scores, radiologic parameters (magnetic resonance image or ultrasonography), and arthroscopic findings. In the meta-regression analysis, there was a significant cell dose-dependent response in pain relief (Q = 9.06, P = 0.029). While three studies reported mild adverse events after MSCs injection, these were not severe and relieved spontaneously. Conclusions: Our meta-analysis revealed that MSC therapy may improve pain, function, radiologic, and arthroscopic parameters in patients with tendon disorders. Due to the

small number of studies in this meta-analysis and considering the increasing MSCs applications, there is a strong need for large-scale randomized controlled trials to confirm the long-term functional improvement as well as the adverse effects of MSC therapies in tendon disorders.

Background
Mesenchymal stem cells (MSCs) treatment is a new regenerative therapy for treating tendon disorder. Preclinical studies have reported that MSC therapy may increase the number of tenocytes and regenerate the injured tendon tissue [1][2][3][4]. While several studies with animals support the treatment of tendon disorders using MSCs, little is known about the efficacy and safety of MSCs to treat these conditions in humans. Although a few clinical reports suggested the therapeutic potentials of MSCs in tendon disorders, they are mostly case reports or case series. Only one randomized controlled trial reported preliminary results (EudraCT Number: 2007-007630- 19), but there are no published results yet [5].
A systematic review of MSC therapy on tendon disorder [6] analyzed three case series [7][8][9] and one matched non-randomized trial [10]. They concluded that MSC treatment is not yet suitable for clinical practice, because the included studies are at high risk of bias.
However, the result should be reconsidered, since three [7,8,10] of the four studies included in this review were not performed with isolated MSCs but bone marrow aspirates or stromal vascular fractions cells. Moreover, this study was not carried out with the metaanalysis methodology, which combines the results from multiple studies. Furthermore, two current clinical studies [5,11], which used isolated MSCs on tendon disorder, were not included in the review.
Although there is an increasing number of published research on stem cell treatments, there are no meta-analyses on this topic to date. Furthermore, concerns regarding possible adverse events of MSC treatments, raised by physicians or scientists reluctant to the therapy [12], should be thoroughly reviewed. Thus, we performed an updated metaanalysis of prospective clinical studies in order to evaluate the efficacy and safety of MSC therapies in patients with tendon disorder.

Search Strategy
The meta-analysis was conducted according to the updated guidelines of the Preferred Reporting Items for Systematic review and Meta-Analysis Protocols (PRISMA-P) [13].

Study Selection Criteria
Identified records were saved to the EndNote software (X7.2; Thomson Reuters). Two independent reviewers (WSC and SYL) screened all the titles and abstracts to identify relevant investigations. The inclusion criteria were as follows: (1) articles reporting a prospective clinical study with at least a 6-month follow-up that (2) described the effect of MSC therapy in patients with any tendon disorder. Although there were no limitations in types of MSCs, i.e. cell origin; autologous or allogeneic, we excluded studies which did not use isolated MSCs, e.g. bone marrow aspirates or stromal vascular fractions cells.
Reviews, basic science articles, comments, letters, and protocols were excluded. When updates of earlier studies were available, we used only the most recent ones.

Outcome Measures and Data Extraction
The primary outcome of interest was defined as pain on motion related to tendon disorder.
All types of pain measurements, e.g. visual analog scale or numeric rating scale, were included. The secondary outcomes analyzed in this study were as follows: 1) functional scores of joint, such as the Constant score, the UCLA score, the modified Mayo elbow performance index, or the Shoulder Pain and Disability Index; 2) radiological parameters to measure tendon defects using magnetic resonance image or ultrasonography; and 3) arthroscopic findings to measure tendon defects with a calibrated arthroscopic probe. For every eligible study, the following data were extracted and entered into a spreadsheet by the two reviewers (WSC and SYL): first author's family name, year of publication, study design, types of tendon disorder, origin of the MSCs, number of patients, MSCs injection methods, cell doses, follow-up duration, safety assessment, and efficacy measurements.

Statistical Analysis
Effect sizes were computed as standardized mean difference (SMD) measures, [18] representing the magnitude of the pretest-posttest difference for each outcome. SMD was calculated separately for all the available control and treatment groups for each study. 6 Heterogeneity between comparable studies was tested with the chi-squared (χ 2 ) and I 2 tests. Values of P > 0.1 and I 2 < 50% were considered statistically significant. Because there was no significant heterogeneity among the four studies (P = 0.658 and I 2 = 0.0%), we used a fixed-effects meta-analysis to quantify the pooled effect size of the studies included. In each analysis by outcome, the following parameters: pain (P = 0.093 and I 2 = 47.0%), functional scores (P = 0.313 and I 2 = 15.3%), radiological parameters (P = 0.406 and I 2 = 0.0%), and arthroscopic findings (P = 0.588 and I 2 = 0.0%) were also analyzed using the fixed-effects model. Additionally, we performed a meta-regression analysis to assess the relationship between MSCs dose and pooled effect sizes in each cell dose. All analyses were performed using the Comprehensive Meta-Analysis Software (version 3.3; Biostat, Englewood, NJ, USA). This study was exempted from Institutional Review Board review as no human subjects were involved.

Description of Included Studies
The primary database search yielded 1,135 records. After duplicates were removed, the titles and abstracts of 897 articles were initially screened and 25 selected for full-text review. The full-text articles were read and 4 were considered relevant by qualitative analysis [5,9,11,19]. The studies selected for final inclusion or exclusion are shown in Fig. 1, and the characteristics of the included studies are summarized in Table 1. In terms of quantitative analysis, these four studies (published from 2015 to 2018) fulfilled our inclusion criteria. Three papers [9,11,19] were open-label prospective studies, while one [5] was an unpublished double-blind randomized controlled trial. The studies identified for meta-analysis included 52 participants. Two studies [9,19] employed adipose tissuederived MSCs and the other two [5,11] administered bone marrow-derived MSCs. The 7 number of cells used in each study ranges from 10 6 to a maximum of 10 8 . Regarding tendon disorder types, most of the studies were performed on rotator cuff tear, but one study [9] was conducted on lateral epicondylitis. The follow-up duration ranged from 6 to 12 months.

Results after Analysis and Publication Bias
The MSC therapies showed a significant pooled effect size (overall Hedge's g pooled SMD 8 = 1.868; 95% confidence interval [CI], 1.274-2.462; P < 0.001) (Fig. 2). The parameters of pain, functional scores, radiological parameters (magnetic resonance image or ultrasonography), and arthroscopic findings all improved with MSC treatment (Fig. 3). In the meta-regression analysis, there was a significant cell dose-dependent responses in pain relief (Q = 9.06, P = 0.029) (Fig. 4). While three studies reported mild adverse events after MSCs injection, these were not severe and relieved spontaneously (  Tendon injuries are a common health problem, which are defined as a painful condition occurring around tendons that limits the function of the affected tendons [20]. Tendons are susceptible to repeated use or degenerative condition. Injuries in those structures are rarely regenerated but repaired by scar tissue and fibrosis. This healed tissue presents inferior tensile strength and is prone to further injuries. Preclinical studies support that MSCs have a regenerative potential as those cells are able to differentiate into proper tendon cell and elicit the secretion of cytokines or growth factors [1]. Therefore, MSCs have been regarded as a possible curative treatment option for tendon degeneration.
Implanted stem cells survive in tendon defects, differentiate into the tenogenic cell lineage and secrete their own extracellular matrix to promote tendon healing [4].

Mazzocca et al. showed that bone marrow-derived stem cells differentiate into tendon-like
cells [21]. Lee et al. also reported that transplanted human adipose tissue-derived stem cells survived for at least 4 weeks in the rat tendon injury model and released humanspecific collagen type I and tenascin-C (TnC) [4]. The expression of TnC is known to increase rapidly during the early period of recovery after tendon injuries, and thus may be used as a marker of tenogenic differentiation [22].
In this meta-analysis, three of the four included studies examined radiological data (magnetic resonance image or ultrasonography) or arthroscopic findings after MSCs injections. These tests could confirm that the injected cells not only relieved pain and improved functions but also regenerated the damaged tissue. Noteworthy, Jo et al conducted the second-look arthroscopic examination at 6 months following MSCs injection as well as MRI follow-up [19]. They reported that the regenerated tendon tissues were 10 identified in all subjects regardless of the location and size of the tear. The defect volumes were decreased in the groups that received mid-dose (5.0 × 10 7 cells) and highdose (1.0 × 10 8 cells). Although this is a macroscopic observation, it may be strong supporting evidence for the regeneration effect of MSCs.
Another important biological mechanism supporting MSC therapy is paracrine effect exerted by these cells [1]. Kinnaird et al. found that growth of endothelial cells and smooth muscle cells may be promoted by the use of medium conditioned with MSCs. This phenomenon might be partly explained by the presence of VEGF and bFGF, which appeared in high levels in the MSCs conditioned medium [23]. The ability of the MSCs to produce a wide range of immunomodulatory and trophic factors has also attracted great attention [24].
There are several concerns regarding the use of MSCs as a treatment option for tendon disorder. Particularly, potential long-term adverse events from the stem cell treatment have been poorly reported in several clinical studies. In the studies included in this metaanalysis, most of the reported adverse events were not related to treatment ( Table 2). The treatment-related side effects were regional swelling following allogeneic stem cell injection [9] or engrafted patch-related chronic synovitis [5]. The joint swelling spontaneously subsided, while the patch-related adverse event needed additional surgery.
Considering the prognosis of the reported adverse events, these side effects might have come from the localized inflammatory response related to the treatment procedure, or to immunologic response against allogeneic cells.
The safety issues related to the MSCs have already been sufficiently assessed in clinical trials in the field of internal medicine, in which MSCs are injected systemically. The POSEIDON trial [25] was designed to investigate the safety and efficacy of autologous and 11 allogeneic MSC therapies for ischemic cardiomyopathy. The study reported that, following trans-endocardial stem cell injection, the treated group showed improvement in structural and functional outcomes, while no serious adverse events including immunologic reactions occurred. Indeed, long-term adverse events from the stem cell treatment and its possible teratogenicity should be thoroughly considered. One animal study reported undesired cartilage formation after the injection of human MSC in eighty-one rat tendon injury models [26]. While there was no histologic evidence of tumor formation in the study, concerns for possible teratogenicity still remain.
Although there are numerous challenges to be overcome and analyzed, it is undisputable that MSC therapy is a potential treatment option to treat tendon disorder. In particular, about 17% of patients with tendon disorder are known to have no effects after undergoing conservative treatment for more than one year [27]. In some patients, the rate of re-tear is fairly high, even following surgical repair for tendon injuries [19].