The Use of Peripheral Blood-Derived Stem Cells for Cartilage Repair and Regeneration In Vivo: A Review

Background Peripheral blood (PB) is a potential source of chondrogenic progenitor cells that can be used for cartilage repair and regeneration. However, the cell types, isolation and implantation methods, seeding dosage, ultimate therapeutic effect, and in vivo safety remain unclear. Methods PubMed, Embase, and the Web of Science databases were systematically searched for relevant reports published from January 1990 to December 2019. Original articles that used PB as a source of stem cells to repair cartilage in vivo were selected for analysis. Results A total of 18 studies were included. Eight human studies used autologous nonculture-expanded PB-derived stem cells (PBSCs) as seed cells with the blood cell separation isolation method, and 10 animal studies used autologous, allogenic or xenogeneic culture-expanded PB-derived mesenchymal stem cells (PB-MSCs), or nonculture-expanded PBSCs as seed cells. Four human and three animal studies surgically implanted cells, while the remaining studies implanted cells by single or repeated intra-articular injections. 121 of 130 patients (in 8 human clinical studies), and 230 of 278 animals (in 6 veterinary clinical studies) using PBSCs for cartilage repair achieved significant clinical improvement. All reviewed articles indicated that using PB as a source of seed cells enhances cartilage repair in vivo without serious adverse events. Conclusion Autologous nonculture-expanded PBSCs are currently the most commonly used cells among all stem cell types derived from PB. Allogeneic, autologous, and xenogeneic PB-MSCs are more widely used in animal studies and are potential seed cell types for future applications. Improving the mobilization and purification technology, and shortening the culture cycle of culture-expanded PB-MSCs will obviously promote the researchers' interest. The use of PBSCs for cartilage repair and regeneration in vivo are safe. PBSCs considerably warrant further investigations due to their superiority and safety in clinical settings and positive effects despite limited evidence in humans.


INTRODUCTION
Articular cartilage covering the surface of joints plays a very important role in bearing loads, absorbing mechanical shocks, and enabling synovial joints to articulate with low friction (Chen et al., 2017). Acute trauma, repetitive joint use, and degenerative joint disease may lead to cartilage and osteochondral injuries (Saw et al., 2011;Fu et al., 2014a). Articular cartilage has a very limited regenerative and self-healing potential due to its avascular, aneural, and alymphatic characteristics and a low number of progenitor cells (Redondo et al., 2018). Many attempts have been made to identify the ideal treatment for cartilage lesions, including bone marrow stimulation (BMS) techniques (Jin et al., 2011), osteochondral autografts and allografts (Makris et al., 2015), and cell-based cartilage repair procedures, including autologous chondrocyte implantation (ACI) (Riboh et al., 2017), mesenchymal stem cell (MSC)based therapy (Fu et al., 2014a;Li et al., 2016) and tissueengineered cartilaginous grafts (Zhao et al., 2018;Ding et al., 2019;Wang et al., 2019;Zhang et al., 2019). Since BMS techniques, osteochondral transplantation, and ACI have limitations and shortcomings, such as fibrocartilage regeneration, donor site complications, graft failure, dedifferentiation of seed cells, and two-stage invasive surgical procedures (Fortier et al., 2010;Andriolo et al., 2017;Riboh et al., 2017), MSCs, which are multipotent progenitor cells with an intrinsic potential for multilineage differentiation, self-renewal, low immunogenicity, anti-inflammatory activity, and immunomodulatory effects by suppressing the graft-versushost reaction, may be obtained from multiple tissues of individual patients, and these cells are easily cultured, amplified, and purified (Goldberg et al., 2017;Guadix et al., 2017). MSCs are widely used in cartilage repair and regeneration as seed cells without concerns regarding increasing the risk of cancer (Hernigou et al., 2013;Liu et al., 2018;Han et al., 2019). An increasing number of studies have suggested that peripheral blood (PB) is a potential alternative source of MSCs, which have shown similar chondrogenic differentiation potential with bone marrow-derived MSCs (BM-MSCs) in both in vitro and in vivo studies (Fu et al., 2014a;Wang et al., 2016a). PB-derived stem cells (PBSCs) can be obtained by a minimally invasive procedure with fewer complications than bone marrow (BM) harvesting, which has been reportedly associated with haemorrhage, chronic pain, neurovascular injury, and even death (Bain, 2003). Moreover, PBSCs also have the ability to be used in autologous transplantation, which greatly benefits patients in clinical applications and facilitates the development of a one-stage surgical solution and other cell-based therapies (Spaas et al., 2012;Hopper et al., 2015a;Saw et al., 2015).
Although increasing evidence has shown that PBSCs are a potential alternative source of chondrogenic progenitor cells for cartilage repair, reviews describing the application of PBSCs for cartilage repair and regeneration in vivo are lacked. The purpose of this review was to evaluate the treatment efficacy and safety of using PBSCs for cartilage regeneration in vivo and attempt to clarify treatment details about cell types, isolation methods, optimal dosages, and implantation methods.

METHOD
This review was conducted in accordance with Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines and a PRISMA checklist using PubMed, EMBASE, and Web of Science to search for relevant studies published from 1 January 1990 to 31 December 2019 (Charlesworth et al., 2019). The search terms used in the selection were "(peripheral OR blood OR circulating OR circulation) AND (mesenchymal OR stem cell OR stromal cell OR progenitor cell OR mononuclear cell OR primitive cell) AND (cartilage OR chondrogenesis OR chondral OR osteochondral OR osteoarthritis) AND (vivo OR human OR patient OR animal OR mouse OR rat OR rabbit OR dog OR sheep OR pig OR horse OR ovine)". YRC, XY, and FZY independently screened study titles and abstracts from the beginning. Only original research studies published in full English that used PB as the source of chondrogenic progenitor cells for cartilage repair were included in the analysis. Both print journals and e-published journals were eligible for inclusion and screening. However, all non-English language studies, review articles, letters, editorials, conference, patents, and meeting abstracts and studies not involving cartilage regeneration were excluded. Duplicates were excluded. In addition, studies of primary cells that were not derived from the PB and studies that were not related to in vivo animal or human experiments or only used non-PB sources were excluded. Disagreements between the authors were resolved by discussion and consensus.
To avoid the omission of relevant studies, we investigated all reference lists of the eligible studies for studies that were likely not identified by the initial retrieval criteria. Unpublished studies were not included in this review. A flowchart of the literature search is shown in Figure 1. We reviewed human studies first, and then reviewed the animal studies according to the order of the publication date. Preoperative characteristics of patients and animals, treatment details, and the treatment efficacy and safety of PBSCs were assessed.

Description of the Included Studies
Overall, 1,898 publications were retrieved from the initial search. A total of 1,685 potential studies were retained for further identification after 84 non-English language studies and 129 review articles, letters, editorials, conference, and meeting abstracts were excluded. Furthermore, 1,619 duplicates and studies of unrelated topics based on the title and abstract, and 49 basic studies that did not conduct experiments in vivo were excluded. We identified 17 in vivo studies consisting of 7 human trials and 10 animal studies published between 1990 and 2019 using this retrieval strategy. All reference lists of the 17 included studies were investigated, and an additional human trial (Jancewicz et al., 2004) was identified and included in this review. Finally, data from the 18 studies [8 human studies (Jancewicz et al., 2004;Saw et al., 2011;Skowroński et al., 2012;Saw et al., 2013;Skowroński and Rutka, 2013;Turajane et al., 2013;Fu et al., 2014a;Saw et al., 2015) and 10 animal studies (Spaas et al., 2012;Broeckx et al., 2014a;Broeckx et al., 2014b;Fu et al., 2014a;Deng et al., 2015;Hopper et al., 2015b;Zhao et al., 2018;Daems et al., 2019;Broeckx et al., 2019a;Broeckx et al., 2019b) published by investigators from seven countries or regions] were analyzed.

Efficacy and Safety of Treatment
We assessed the adverse events and the clinical, radiographic, and histologic results to determine the treatment efficacy and safety ( Table 3).
Radiological examination, which is a non-invasive examination method, was widely used to evaluate the efficacy of cartilage repair and regeneration. Seven human studies used MRI (Jancewicz et al., 2004;Skowroński et al., 2012;Saw et al., 2013;Skowroński and Rutka, 2013;Fu et al., 2014a), X-ray (Saw et al., 2011;Saw et al., 2015) or CT (Fu et al., 2014a) to evaluate the repair effect and reported improved MRI morphologic scores, regenerative tissue with the same signal as normal cartilage, subchondral bone recovery, or reappearance of the medial compartment. However, radiological examination was rarely used in the animal studies. Three animal studies (Spaas et al., 2012;Daems et al., 2019;Broeckx et al., 2019a) reported no significant radiographic changes.
Four human studies (Saw et al., 2011;Saw et al., 2013;Fu et al., 2014a;Saw et al., 2015) evaluated cartilage repair with the method of second-look arthroscopy and suggested that cartilage regeneration was comparable to BM-MSCs with improved ICRS macroscopic scores, cartilage regeneration similar to normal cartilage tissue, or significantly fewer wear lines and synovial hyperaemia.

DISCUSSION
Researchers have conducted investigations of PBSCs in cartilage repair and regeneration because of the advantages of PBSCs and limitations of chondrogenic progenitor cells from other sources, such as bone marrow (Bain, 2003), synovial membranes (Murata et al., 2018), and adipose tissue (Kuroda et al., 2015). Increasing evidence has shown that PB-MSCs have a similar potential for proliferation and trilineage differentiation as BM-MSCs and might be a promising source of seed cells for cartilage repair (Wang et al., 2016b) since Fernández et al. (Fernandez et al., 1997) reported the presence of stromal cells in hG-CSFmobilized PB from patients with breast cancer for the first time in 1997. However, PBSCs were not used to treat chondral defects and promote cartilage regeneration in vivo until 2004, as reported by investigators in Poland (Jancewicz et al., 2004).
For the first time, this review comprehensively evaluated the feasibility, efficacy, and safety of using PBSCs for cartilage repair and regeneration in vivo by analyzing the preoperative characteristics, therapeutic details, outcomes, and adverse events reported in currently published literature. This review might provide new insights and strategies for further foundational research and clinical applications of PBSCs.
Autologous nonculture-expanded PBSCs are easy to harvest and manipulate from G-CSF-activated PB without the concerns of disease transmission, immune rejection, and ethical issues (Fu et al., 2014a;Saw et al., 2015). PBSCs are currently the most commonly used cell type for cartilage repair in all stem cell types derived from PB. It has been demonstrated that noncultureexpanded PBSCs comprise haematopoietic stem cells (HSCs), fibrocytes, a population of MSCs/mesenchymal progenitor cells (MSCs/MPCs), white blood cells, platelets, growth factors, and a small percentage of red blood cells (Stroncek et al., 1997;Cesselli , 2009). When PBSCs are being cultured, other impure cell types except MSCs/MPCs are not present anymore. To a certain degree, the cell composition of nonculture-expanded PBSCs is similar to that of the bone marrow-derived buffy coat (BMBC), which is separated from bone marrow using a Ficoll gradient centrifugation system. The bone marrow-derived buffy coat has been widely used as a source of MSCs for cartilage repair and regeneration and has achieved good to excellent results (Fortier et al., 2010;Jin et al., 2011). Several possible mechanisms of action of PBSCs might contribute to cartilage repair. Hopper et al. (Hopper et al., 2015a;Hopper et al., 2015c) found that PBSCs stimulate the upregulation of eight genes associated with chondrogenic differentiation of knee infrapatellar fat padderived MSCs, increase the total number of MSCs, increase native chondrocyte migration, and accelerate the rate of cell movement. Exogenous MSCs, HSCs, and growth factors in PBSCs initiate cartilage regeneration and augment endogenous MSC recruitment from bone marrow to subchondral drilling sites (Khaldoyanidi, 2008;Onuora, 2015;Saw et al., 2015). Deng et al. (2015) suggested that PBSCs prevent the progression of papain-induced knee OA in a rat model by reducing articular surface fibrillation, irregularity, and erosion, and by inhibiting chondrocyte necrosis and loss of chondrogenic proteins. HSCs and non-HSCs, such as MSCs, endothelial progenitor cells, and very small embryonic-like (VSEL) cells, contained in PBSCs might play an important role through a paracrine mechanism (Kucia et al., 2007;Onuora, 2015). Although the term "PBSCs" had different expressions in different studies, such as PB progenitor cells (PBPCs) (Saw et al., 2011) and PB mononuclear cells (PBMCs) (Hopper et al., 2015b), we found that the cell acquisition method and cell composition were basically the same. For the convenience of expression, "PBSCs" was used uniformly in this paper. The transplantation of autologous culture-expanded PB-MSCs requires two procedures for obtaining patient cells and transplanting the cells after cultivation, which prolongs hospital stays, increases costs and risks contamination related to in vitro culture, possibly limiting the clinical application of autologous PB-MSCs (Saw et al., 2013;Fu et al., 2014b). Moreover, an agerelated decline in MSC numbers, proliferation, and clonogenicity, which lead to more difficult culture in vitro and a longer culture cycle than MSCs from other tissue sources, might be another significant cause for the lack of clinical applications of autologous culture-expanded PB-MSCs (Kassis et al., 2006;Bourzac et al., 2010;Chong et al., 2012;Spaas et al., 2013;Wang et al., 2016b). For example, MSCs derived from bone marrow, synovium or adipose tissue reached 80-90% confluence within 7 to 14 days (Zhang et al., 2014;Jin et al., 2016;Shimomura et al., 2016). However, MSCs derived from PB did not achieve the same confluence until about 21 days after primary culture . It takes longer to obtain the culture-expanded PB-MSCs than other tissue-derived MSCs. The presence of MSCs in human PB is debatable and their identification may be hampered, among others, by: (i) their low frequency in PB of healthy individuals, and (ii) the large biological variations related to donor age, pathology, disease status, and corresponding treatment regimens (Fox et al., 2007;Moll et al., 2019). Most investigators agree that their frequency in blood is low in healthy individuals, but that the amounts of circulating MSCs may increase under special mobilization conditions, thus supporting the notion that MSCs can be transiently found circulating in blood (Moll et al., 2020). Jain et al. provide evidence that MSCs can be found in PB and apheresis product of patients treated with a typical G-CSF-based HSCs mobilization regimen by using flow cytometry (Jain et al., 2020). However, a systematic review strongly indicated the existence of MSCs in the PB of animals (Wang et al., 2016b), this might be because researchers could improve the success rate of PB-MSCs in animal studies by optimizing mobilization and culture procedures, prolonging the culture time, and increasing the number of animals and the frequency of blood drawn (Pitchford et al., 2009;To et al., 2011;Spaas et al., 2013). To et al. noted in a baboon model that MSC mobilization and colony-forming unit fibroblast (CFUF) in PB in response to G-CSF did only occur when adding stem cell factor . Pitchford et al. found, that MSCs/CFU-F were not found in mice PB post-mobilization with G-CSF, but when adding vascular endothelial growth factor and CXCR4-antagonist (Pitchford et al., 2009). Spaas et al. systematically studied the isolation and culture methods, cell characteristics, and clinical safety of equine PB-MSCs, and applied them to many veterinary clinical studies, such as promoting cartilage repair, cutaneous wound healing, and healing of tendon and ligament lesions (Spaas et al., 2013;Beerts et al., 2017;Martinello et al., 2018;Broeckx et al., 2019a). Allogenic or xenogeneic MSCs banks, improving the mobilization and purification techniques, and shortening the culture cycle might effectively account for deficiencies in autologous MSCs, reduce the burden on both patients and treatment providers, and promote the development of singlestage procedures (Moroni and Fornasari, 2013;Pescador et al., 2017).
MSCs inhibit immune responses and are not restricted by the HLA system through immune evasion and immune privilege mechanisms (Paterson et al., 2014;Vega et al., 2015). Moreover, the strong immunomodulatory and immunosuppressive properties of MSCs may play an important role in modifying graft-versus-host reactions during allogenic transplantations (Le Blanc and Ringden, 2007). Two animal studies used allogenic native and chondrogenic-induced PB-MSCs as a treatment for degenerative joint disease in horses and significantly improved the short-and long-term effects without serious adverse events (Broeckx et al., 2014a;Broeckx et al., 2014b). Vega et al. (2015) performed an RCT to assess the feasibility and safety of treating osteoarthritis with allogeneic MSCs in humans, and they concluded that allogeneic MSCs might be a convenient and effective alternative to autologous MSCs for the treatment of OA in the knee without serious transplantation-related adverse events. A number of published papers have indicated that transplanted MSCs influence the local microenvironment of cartilage by paracrine actions, such as the secretion of various growth factors, cytokines, and chemokines, to exert antiinflammatory, anti-apoptotic, and anti-fibrotic effects on chondrocytes (Kuroda et al., 2015;Mancuso et al., 2019). Another possible mechanism of action of MSCs in cartilage repair and regeneration is that transplanted progenitor cells migrate to damaged cartilage areas and differentiate into chondrocytes and osteocytes (Cesselli et al., 2009). The fate of MSCs injected into the articular cavity can be monitored by labelling with green fluorescent protein (GFP) or carboxyfluorescein diacetate succinimidyl ester (CFDA-SE) (Guest et al., 2008;Sato et al., 2012). Murphy et al. found that the implantation of MSCs into the knee joints of goats with OA showed a strong and sustained effect in promoting cartilage repair. However, further tracing the labelled MSCs showed that the cell retention rate was very low, usually about 3%, and most cells disappeared within a few days (Murphy et al., 2003). This suggests that MSCs may not directly differentiate into chondrocytes to participate in tissue repair in vivo, but promote cartilage regeneration through other mechanisms. In recent years, more and more researchers believed that exosomes secreted by MSCs played an important role in cartilage repair and regeneration (Marote et al., 2016;Yan and Wu, 2019;Jin et al., 2020;Liu et al., 2020). Exosomes are generally considered as communication vectors between cells, and carry a large number of complex nucleic acids (mRNA and miRNA lncRNA), proteins and lipids that can regulate and restore extracellular matrix (ECM) homeostasis (Colombo et al., 2014). For example, MSCs exosomes with overexpressing of miR-140-5p blocked other Wnt signals in vitro by inhibiting vral simian leukemia viral oncogene homolog A (RalA) and activating sex determining region Y-box 9 (SOX9), and regulate the expression of Col II and aggrecan (ACAN) in vivo to promote cartilage regeneration (Tao et al., 2017). It may also be an important mechanism for PBSCs to promote cartilage repair.
Blood cell separation is the most commonly used method for collecting PBSCs. It is a developed and simple technique that has been widely used in the treatment of systemic blood diseases. In a monocyte suspension isolated by blood cell separation, CD105 + cells have been shown to be more abundance than CD34 + cells, and the proportion of CD105 + cells increased after cryopreservation (Saw et al., 2011). However, there is no study on the subsequent isolation and culture of PB-MSCs from PBSCs collected by blood cell separation. The current standard methods of PB-MSC isolation are DGC (such as Ficoll, Lymphoprep, and Percoll) and PA (Bourzac et al., 2010).
As one of the most fundamental parameters that might influence the outcome of cartilage repair (Gupta et al., 2016), the optimal density or dosage of PBSCs used for cartilage regeneration in different methods and species has not been fully investigated. Skowroński et al. (Skowroński and Rutka, 2013) reported a slightly poorer outcome of cartilage repair in a group treated with a bone marrow concentrate than a group treated with fresh condensed PBSCs, and they attributed this result to the lower cell count in the suspension obtained from bone marrow. The main concern of using nonculture-expanded PBSCs to promote tissue regeneration is the low content of MSCs within harvests. The number of HSCs (with a CD34 + surface marker) and MSCs (with a CD105 + surface marker) were quantified by flow cytometry in a study carried out by Saw et al. (Saw et al., 2011). The flow cytometry result showed that the proportion of CD105 + cells in fresh PBSC suspension was 7.24% (2.32×10 6 cells/ml). Interestingly, the proportion of CD105 + cells reached 8.39% (2.69×10 6 cells/ml) after cryopreservation. However, the CD105 + cell counts vary between different studies. Turajane et al. (Turajane et al., 2013) reported that a proportion of CD105 + cells ranging from 0.75 to 0.88%. The difference of the proportion of CD105 + cells in the two studies was probably due to the younger patients in the previous study and the older patients in the latter.
To increase the yield of MSCs from autologous PB, repeated intra-articular injections were implemented in some studies (Saw et al., 2011;Saw et al., 2013;Turajane et al., 2013;Saw et al., 2015), and they speculated that this method is more efficacious than a single injection for the enhancement of cartilage repair on the basis of a suggestion from an animal study (Saw et al., 2009). However, repeated IA injections of culture-expanded allogeneic MSCs is not recommended due to a significant adverse response that might be initiated by immune recognition of allogeneic MSCs after a second exposure (Joswig et al., 2017).
Currently, the optimal seeding density of MSCs also remains unknown. A systematic review showed that the dose of MSCs for cartilage repair varies from 2×10 6 -7.7×10 7 cells in human clinical studies (Goldberg et al., 2017). Gupta et al. (Gupta et al., 2016) found that an MSC dose of 2.5×10 7 with the IA injection method for treating OA showed the best improvement for relieving pain and the lowest adverse events compared with other higher dose groups. They hypothesized that a higher cell dosage causes cell aggregation and subsequent cell death due to limited space in the knee joint. A prospective RCT demonstrated that an intra-articular injection of cultured MSCs with a mean dose of 1.46×10 7 cells for treating OA is effective in improving clinical and magnetic resonance observation of cartilage repair tissue (MOCART) scores after a 2-year follow-up (Wong et al., 2013). Given the limited evidence of clinical application of PB-MSCs in cartilage repair and regeneration, the optimal therapeutic dose of PB-MSCs remains to be further studied.
Moreover, a number of studies have reported concomitant procedures, such as abrasion arthroplasty (Beckmann et al., 2015), autologous bone grafting to restore bone mass (Sadlik et al., 2017), treatment of co-existing pathologies (Wong et al., 2013), and BMS (Jin et al., 2011), PRP (Broeckx et al., 2019a) and HA (Charlesworth et al., 2019) to repair cartilage defects. Thus, the abovementioned methods are recommended to supplement PBSCs for cartilage repair and regeneration. A rigorous postoperative rehabilitation programme is required to protect grafts and avoid the effusion of PBSC suspensions (Skowroński and Rutka, 2013;Fu et al., 2014a).
Compared with other tissue-derived MSCs, the culture of PB-MSCs was relatively difficult, which resulted in less reports of its application in vivo, but it does not affect its application prospects. On the contrary, it is ethically more suitable for clinical application due to its unique advantages, such as minimally invasive sample acquisition procedure, repeatable sampling, and high recognition of patients (Fu et al., 2014a;Fu et al., 2014b;Wang et al., 2016a;Chen et al., 2019). In this review, we have summarized all the currently published researches on the use of PBSCs for cartilage repair and regeneration in vivo. Although only 5 human and veterinary clinical studies (Saw et al., 2013;Skowroński and Rutka, 2013;Daems et al., 2019;Broeckx et al., 2019a;Broeckx et al., 2019b) had a control group, the results were still very useful for readers, and can reflect the progress and problems in this field to a certain extent.

CONCLUSION
This review evaluated the use of PBSCs in cartilage repair and regeneration in vivo for the first time. Autologous PBSCs are easy to obtain and are free of transmittable diseases, infection risks, and medical ethical restrictions. They are currently the most commonly used cell type for cartilage repair among all stem cell types derived from PB. Blood cell separation technology is developed, simple, and convenient, making it the most commonly used method to obtain PBSC suspensions. Allogeneic culture-expanded PB-MSCs are more widely used in animal research and are potential seed cell types for cartilage repair and regeneration in the future. DGC and PA are the most commonly used methods for PB-MSC isolation. Improving the purification technology and shortening the culture cycle of culture-expanded PB-MSCs will obviously promote the researchers' interest. PBSCs are safe in cartilage repair and regeneration. Although all reviewed articles indicated that using PB as a cell source enhances cartilage repair and regeneration in vivo by the IA injection and surgery implantation methods, we should maintain a prudent attitude towards the positive therapeutic effect of PBSCs considering the deficiency of studies with a high level of evidence, incomplete assessment system of outcomes, and combined use of multiple other treatments. In summary, the use of PBSCs in cartilage repair and regeneration warrants considerable efforts for further investigations due to its superiorities and safety in clinical settings and positive effects despite limited evidence in human.