Mesenchymal stem cells therapy improves ovarian function in premature ovarian failure: a systematic review and meta-analysis based on preclinical studies

Background Studies have revealed that the transplantation of mesenchymal stem cells (MSCs) might be a potential star candidate for premature ovarian failure (POF) in animal experiments. However, individual studies with a small sample size cannot be used to draw a clear conclusion. Therefore, we conducted a systematic review and meta-analysis to explore the potential of using MSCs in the treatment of POF in animals. Methods Seven databases were searched for studies exploring the effect of the transplantation of MSCs on POF in animal models. The PRISMA guideline was followed, and the methodological quality was ensured using SYRCLE’s risk of bias tool. RevMan 5.4 and STATA 12.0 software was performed to meta-analysis. Results In total, 37 studies involving 1,079 animals were included. Significant associations were found for MSCs with the levels of E2 (SMD 2.69 [95% CI 1.97, 3.41]), FSH (-2.02, [-2.74, -1.30]), primary follicles (2.04, [1.17, 2.92]), secondary follicles (1.93, [1.05, 2.81]), and primordial follicles (2.38, [1.19, 3.57]. Other outcomes, such as AMH, LH, INHB, antral follicles, growing follicles, mature follicles, and early antral were also found to be significant. There was no difference in FSH/LH, corpus leteum, follicles, and estruc cycle. Conclusions Our meta-analysis result indicated that the transplantation of MSCs might exert therapeutic effects on animal models of POF, and these effects might be associated with improving the disorder of the sexual cycle, modulating serum hormone expressions to a better state, and restoring ovarian function.


Introduction
Premature ovarian failure (POF) is one of the most common causes of female infertility (1). The main feature is ovarian degeneration before the age of 40, with a possibility of ovarian function fluctuating before complete cessation (2). It usually manifests hypoestrogenism, hypergonadotropism, and amenorrhea (1). POF can arise from a range of conditions, including genetic disorders, iatrogenic factors, chemotherapy injury, autoimmunity, and environmental and infectious causes (3). It has an important social and psychological impact, in addition to physiological consequences such as autoimmune disorders, heart disease, and osteoporosis (4). To date, hormone replacement therapy (HRT) is strongly recommended for women with POF, and it can improve hot flushes and vaginal and urinary symptoms for women with POF (5). In addition to HRT, new methods, such as ovarian transplantation, gene therapy, and stem cell therapy, are also urgently needed to improve the treatment efficacy of POF (6).
Mesenchymal stem cells (MSCs) are characterized by multidirectional differentiation potential. They can be isolated from various tissues, such as the peripheral blood, placenta, umbilical cord (UC), adipose tissue (AD), bone marrow (BM), and menstrual fluid (7). Studies reveal that MSC transplantation is a potential star candidate for POF. Due to their safety and manufacturing, the transplantation of MSCs has mainly focused on animal experiments for POF. Researchers found ovarian function could improve after the transplantation of human umbilical cord MSCs (hUC-MSCs) in a POF rat model (8,9). The transplantation of BM-MSCs can also regain fertility in radiation-associated POF animal models (10). Deng T et al. (11) found that ovarian injury was reduced and ovarian function was improved after UC-MSC transplantation in a chemotherapyinduced POF mice model. In addition, studies have also revealed that Human menstrual blood MSC (HuMenSC) and chorionic plate MSC (CP-MSC) transplantation can be considered suitable treatments for POF (12,13). It is worth noting that these studies were based on small sample sizes and were highly volatile. Because of the small sample size of each study, the treatment may overstate the efficacy of the individual animal experiments.
Studies need to be systematically combined to reach more robust conclusions. Meta-analysis can obtain stable results by increasing the sample size. There is a phenomenon of inconsistency in recent preclinical studies regarding the treatment of POF using MSC transplantation. Therefore, we conducted a systematic review and meta-analysis of preclinical animal data to evaluate the available evidence for MSCs in the treatment of POF.

Methods
We followed the PRISMA guideline for this study (Supplementary Table 1). Supplementary Data Sheet 1 shows the protocol. The study was exempt from research ethics board approval.

Search strategy
We searched three English (PubMed, Embase, and Web of Science) and four Chinese (VIP, Wanfang, CBM, and CNKI) databases from inception to August 2022. Searches included terms relating to mesenchymal stem cells, premature ovarian failure, and animal models (Supplementary Data Sheet 2). Study language was not a restriction. With the help of two medical librarians (LL and LS), two authors (CG and TS) independently searched and screened the databases. secondary outcomes were other sex hormone outcomes, such as luteinizing hormone (LH), anti-Mullerian hormone (AMH), FSH/ LH, inhibin B (INHB), and other follicle outcomes, such as primordial follicles, growing follicles, antral follicles, mature follicles, atretic follicles, corpus leteum, follicles, early antral, and estruc cycle. The type of study included was randomized controlled studies. We excluded case studies, observational studies, reviews, and human studies. If they were duplicates, studies with the most complete information were included in the analysis.

Data extraction
Using pretested forms, two authors (CG and TS) independently screened articles for eligibility, and extracted data. Study author, publication year, author country, sample size, type and dose of MSCs, route of transplantation, type of animals, and assessment of outcomes were extracted. The closest to 4-week outcomes were extracted when results for more than one time point appeared. Conflicts were resolved through consensus with a third author (DF).

Quality assessment
The methodological quality of each included study was assessed with the use of SYRCLE's risk of bias tool (14). The contents of the evaluation covered 10 entries in six aspects (Supplementary  Table 2). Each entry was categorized as "Yes", "No", or "Unclear" regarding bias. Two reviewers (CG and TS) independently assessed the quality of each included study and were mediated by the third reviewer (DF).

Data analysis
Data analysis and forest plots were combined for meta-analysis using Review Manager 5.4.1 and Stata version 12. Standardized mean difference (SMD) with a 95% confidence interval (95% CI) was approved for each of the included studies. I 2 statistics (> 50% indicating substantial heterogeneity) assessed the variation studyto-study. If substantial heterogeneity existed, a random effects model was used. Otherwise, a fixed effects model was applied. A Funnel plot and Begg's and Egger's tests were provided to assess the potential publication biases if more than 10 studies for one outcome were included. We performed subgroup analyses for animal species (rats, mice, rabbits, or mice and rats), induced (chemotherapy, autoimmune, or radiation), source of MSCs (human or animal), types of MSCs (UCMSCs, BMMSCs, AFMSCs, CPMSCs, MenSCs, ADMSCs, or PMSCs), the transplantation dose (10 5 , 10 6 , or 10 7 ), and route of transplantation (intravenously injected or intra ovary injected). We considered p-values of less than 0.05 to be statistically significant.

Assessment of study quality and risk of bias
The quality and risk of bias for studies included in the metaanalysis are presented in Supplementary Table 2. Only six of the 37 studies mentioned sequence generation. The baseline characteristics were comparability. Allocation concealment, attrition, reporting, and other bias were not found. Performance bias and detection bias were unclear for all the included studies.

E2
In total, 30 studies (n = 568 animals) provided data for this comparison (286 vs. 282 animals). E2 was noted to be higher in the MSC group (SMD 2.69, 95% CI 1.97 to 3.41) in the random-effect model (I 2 = 88) ( Figure 2). A funnel plot showed asymmetry (Supplementary Figure 1), and Begg's (z = 4.46, p = 0.001) and Egger's (t = 6.50, p = 0.001) tests also indicated that there was a statistical difference.  Flow diagram of the study selection process for this systematic review and meta-analysis.

Secondary follicles
The effect size for secondary follicles was pooled from 12 studies (n = 238 animals). A significant association was found between MSC therapy and an increase in the secondary follicle counts (SMD 1.93, (95% CI 1.05 to 2.81) in the random-effect mode (I 2 = 83%) (   Forest plot summarizing the effects of MSCs on E2 in preclinical models of POF.

Subgroup analysis
Most of the subgroup analyses did not change substantially, except for a few variables that resulted from fewer data ( Table 2). Forest plot summarizing the effects of MSCs on FSH in preclinical models of POF. Forest plot summarizing the effects of MSCs on primary follicles in preclinical models of POF.

Discussion
This systematic review and meta-analysis examining 37 studies (1,079 animals) found that mesenchymal stem cell transplantation might effectively improve ovarian function in premature ovarian failure in animal model studies.
According to the source of the isolated tissue, MSCs can be classified into different types. Multiple previous studies indicated that MSC transplantation was effective in repairing tissue damage, including neurodegenerative diseases, skin injury, cardiovascular disease, bone damage, and diabetes (46)(47)(48)(49). For POF, MSC therapy had promising potential for clinical use but some safety and manufacturing issues have kept it mostly in the animal model stage.
In preclinical studies, MSCs appeared to be able to improve certain functional measures of damaged ovaries, including sex hormones and the number of follicles at different stages, in POF animal models. However, MSCs from different tissues produced different effects. One study has shown (11) that primary and secondary follicles increased, atretic follicles decreased, and ovarian volume also can increase after UC-MSC treatment in POF mice. Another study found that BM-MSCs could regulate and activate ovarian function and regain fertility in a radiationinduced POF rat model (10). Meanwhile, serum hormone levels, including E2 and FSH, were restored after the transplantation of CP-MSC in a chemotherapy-induced POF mouse model (13). Overall, our results are generally consistent with previous research and showed that different MSCs from different sources are different in repairing ovarian function in this study. In addition to the properties of specialized cells, the sample size and included studies are also important and affect treatment outcomes, and small samples often produced false negative results.
It is important to note that, from the analysis of the 37 studies, it is suggested that MSC transplantation enhanced the primordial follicle formation and follicle growth up to the secondary follicle stage, but not to the antral follicle stage, thus contributing to hormonal status as far as estrogen levels are concerned while not contributing to re-establishing the estrus cycle. The reason for this may be the limitations of the original articles concerning their methodological quality or their small sample size. Meanwhile, the significant changes found in the analyses regarding different parameters should be evaluated and commented on with respect to the duration of the effect and observation period.
Although they have been shown to be safe in animal models, the risk cannot be ignored concerning MSCs being transplanted into patients. Several small-sample clinical trials indicated that ovarian function and fertility could improve in women with POF after MSC therapy. In their phase I study (1), the authors found MSCs were safe to use for more than a year, and menstruation also can resume for POF patients. Other clinical results (3,50,51) also showed that transplantation of MSCs could improve ovarian features, including increasing the number of follicles and improving ovarian hormone production, in POF patients. In these human studies, it seems that the improvement is somewhat transient or temporary. Thus, it is important to have a similar conclusion in animal studies and provide some evidence and guidance to the readers for the next step. Forest plot summarizing the effects of MSCs on secondary follicles in preclinical models of POF. Forest plot summarizing the effects of MSCs on primordial follicles in preclinical models of POF.        In this comprehensive, thorough, and complete systematic review and meta-analysis study, there is further evidence, obtained by increasing the sample size in POF animal models, that MSC treatment could rescue ovarian function. The results of this study will provide more powerful preclinical medical evidence support for the application of MSCs in high-quality clinical trials. Our study further demonstrates the ability of MSCs to treat POF. However, it is still a tough task for the transformation from animal to human. More detailed mechanisms of MSC treatment for POF need to be discovered in the future. In addition, continuous follow-ups are needed to evaluate the long-term effects of this new intervention.

Route of transplantation
There are still several limitations of this study that need consideration. First, we only searched the commonly used English and Chinese databases, which would lead to some language bias. Second, the included studies were generally of low methodological quality, and the sample size of each study was small. Third, some outcomes could not be subgroup analyzed because of small sample sizes. In addition, the various sources of MSCs and animal species could limit the value of MSCs in the treatment of premature ovarian failure.

Conclusions
In summary, this systematic review and meta-analysis indicated that MSCs might exert therapeutic effects on animal models of POF, and these effects might be associated with improving the disorder of the sexual cycle, modulating the serum hormone expression to a better state, and restoring ovarian function.

Data availability statement
The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding author.

SUPPLEMENTARY TABLE 2
Study quality and risk of bias.