Potential of periosteal cells in bone and cartilage regeneration: a systematic review

Introduction: The unavailability of adequate human primary cells presents multiple challenges in terms of bone and cartilage regeneration and disease modeling experiments in vitro. Periosteal cells (PCs), which represent promising skeletal stem cell sources, could be a promising strategy in tissue engineering. The present study aimed to summarize the characteristics of PCs to investigate the efficacy of these cells in bone and cartilage regeneration in different models, paying special attention to the comparison of bone marrow stromal cells (BMSCs). Methods: A comprehensive literature search was conducted in Embase, PubMed/MEDLINE, Web of Science, and Scopus for articles published in English until April 2023. Only original researches in which PCs were employed for bone or cartilage regeneration experiments were included. Results: A total of 9140 references were retrieved. After screening the results, 36 publications were considered to be eligible for inclusion in the present literature review. Overall, PCs demonstrated beneficial bone and cartilage regenerative efficacy compared to the bare scaffold since almost all included studies reported positive results. The 9 studies assessing the differences in bone formation capacity between PCs and BMSCs indicated that PCs exhibited stronger in vivo osteogenic differentiation capabilities compared to BMSCs, while the other study demonstrated stronger chondrogenic potential of BMSCs. Discussion: PCs demonstrated beneficial to bone regenerative efficacy compared to the bare scaffold with a low risk of most studies included. However, the cartilage formation capacity of BMSCs still needs to be investigated due to the limited research available and the certain risk of bias. PCs exhibited higher osteogenic capabilities compared to BMSCs in combination with various scaffolds in vivo with good evidence. Further researches are needed to elucidate the comparative benefits of cartilage regeneration. Systematic Review Registration: https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42023411522, CRD42023411522.


Introduction
Over the past decades, therapeutic options capable of repairing and reconstructing bone and cartilage defects have attracted a great deal of scientific and public attention (Grayson et al., 2015;Tamaddon et al., 2018).Normally, small defects can be effectively repaired because of the regenerative potential of bone and cartilage tissues.However, large ones due to multiple diseases remain a great challenge in clinical scenarios (Atala et al., 2012;Su et al., 2018).In addition, the morbidity of musculoskeletal disorders including fractures, osteoporosis, and rheumatic diseases is increasing rapidly due to the increased life expectancy (Roseti et al., 2017).Recently, the conventional approach to cure complex large bone defects includes transplantation of allogenous or autogenous bone grafts harvesting from the radius, fibula, iliac crest, and scapula or the application of substitutes to restore bone integrity (Toros and Ozaksar, 2021).Nevertheless, the inherent shortcomings of this method, such as donor-site morbidity and insufficient autogenous resources, significantly encourage researchers and clinicians to explore alternative treatment options (Dimitriou et al., 2011).Surgical options to manage damaged cartilage include arthroscopic debridement, osteochondral allograft, osteochondral autografts, and, in the presence of osteoarthritis, joint replacement (Goldberg et al., 2017).Bone marrow stimulation techniques, such as micro-fracture, are the most frequently used method in clinical practice for treating small symptomatic lesions of the articular cartilage (Steinwachs et al., 2008).However, the resulting tissue has shown to be a mixed fibrocartilage tissue with varying amounts of type II collagen and inferior to native hyaline cartilage (Goldberg et al., 2017).In this context, tissue engineering based on stem cells and scaffolds has emerged as a potential alternative method for the replacement of defective or malfunctioning tissues.This approach eliminates the inherent limitations of traditional transplantation of bone grafts and provides biological tissue substitutes in various conditions.Through recapitulating critical features of development or tissue repair, stem cell-based tissue engineering can improve tissue formation in vitro or promote tissue regeneration in vivo for the replacement of damaged ones (Charwat et al., 2008;Jukes et al., 2010).
Stem cells are defined as a population of undifferentiated cells with the potential to extensively proliferate from a single cell to different types of cells and tissues (Kolios and Moodley, 2013).Because of the unique ability including self-renew and multidirectional differentiation, tissues that can be engineered using these cells comprise a diverse range from skeletal tissues to epithelial surfaces, which present unprecedented applications.Stem cells are indispensable for the practical use of tissue engineering approaches, and the acquisition of stem cells is important.Among various sources of stem cells used for bone and cartilage regeneration, the bone marrow compartment has been demonstrated to represent a reliable tissue resource to harvest stem cells with convincing evidence of differentiation capacity both in vitro and in vivo (Li et al., 2009;Arthur and Gronthos, 2020).In addition, the periosteum is another essential source of mesenchymal stem cells (MSCs) for cartilage and bone regeneration in addition to the bone marrow compartment, which was originally identified as a reliable resource to harvest MSCs (Bolander et al., 2017;Mendes et al., 2018).
As an essential component covering the outer surface of bone, the periosteum is of great significance in bone physiology during remodeling, development, and growth (Maia Ferreira Alencar et al., 2020).Its structure is heterogeneous, consisting of the following two layers: the outer fibrous layer with fibroblasts, and the inner cambium layer, which contains osteoprogenitor cells, osteoblasts, and pre-osteoblasts that influence bone formation.Activated periosteum produces cartilage and bone, and is colonized by osteoclasts (Hutmacher and Sittinger, 2003).As a primary source of MSCs, PCs have gained a lot of scientific attention for regenerative approaches.The capacity of PCs to develop into bone and cartilage has been demonstrated in several studies (Miyamoto et al., 2004;Chen et al., 2012).In addition, with the help of continuous development of tools and techniques, specific role and regulation of PCs can be investigated more deeply since the challenge of isolating PCs has been overcome.
Previously published systematic reviews have proved the efficacy of BMSCs for bone and cartilage regeneration (Sun et al., 2016;Zhu et al., 2023).However, the role of PCs in tissue engineering remains unclear.Accordingly, it is necessary to summarize the current evidence in terms of the application of PCs in bone and cartilage regeneration.Therefore, the aim of this study is to conduct a systemic review to assess the osteogenic and chondrogenic capacities of PCs.In addition, this review also elucidates the limitations of existing research, paying special attention to the comparison of bone marrow stromal cells (BMSCs).To our knowledge, this is one of the first reviews that summarizes the potential role of PCs in both bone and cartilage regeneration.

Materials and methods
The present systematic review was registered at PROSPERO under number CRD42023411522 and performed in accordance with the Preferred Reporting Items for Systematic Reviews and Metaanalyses (PRISMA) guidelines (Moher et al., 2015).As this study did not involve human or animal subjects, ethics approval was not required.
The guiding question for this systematic review was formulated according to the PICO format; (P) indicates the participants, (I) means the intervention, (C) represents the comparison, and (O) is for the outcome (Schardt et al., 2007).Does the use of PCs (I) improve the rate of bone and cartilage regeneration (O), compared to formation ability using other types of cells (C) in various animal models (P)?

Search strategy
A systematic literature search was performed in Embase, PubMed/MEDLINE, Web of Science, and Scopus as sources for literature published up to April 2023 to identify suitable publications.These four databases were selected since they are the largest pharmaceutical and biomedical databases, which would be unlikely to lessen the number of articles.Defense Technical Information Center was used to search gray literature.The search strategy was shown in Table 1.Three components were included: bone regeneration and/or cartilage regeneration, PCs and tissue engineering.In addition, the electronic search of the databases was complemented by a manual search in reference lists of chosen articles to improve completeness.

Eligibility criteria
Publications that fulfilled the following inclusion criteria were selected: 1) all preclinical controlled animal model studies with PCbased approaches for bone and/or cartilage regeneration; 2) data were measured as new bone and/or cartilage formation (%) with the utilization of PCs-based strategies.
The exclusion criteria included: 1) review articles, abstracts, letters, editorials, correspondences, and case reports; 2) PCs that were genetically modified or not isolated from the periosteum.

Study selection and data collection process
The information retrieved from the database was compiled, and any duplicate entries were removed.The title and abstract were evaluated based on eligibility criteria by the two authors separately.Studies considered ineligible by the two authors were excluded immediately, while studies considered ineligible by one author but eligible by the second author were retained for reading the full text.Researches not excluded were read in full text by two reviewers, who then chose studies that met the eligibility criteria and conducted data extraction.Any disagreements were then resolved through discussion and consensus with all the reviewers.Data from selected studies were retrieved and gathered in detail in one document.Reports of the following variables were extracted from each study: author(s), year of publication, species, age, sex, animal model, tissue origin, types of tissue regeneration, source of MSCs, defect type, implant site, scaffold, density, scaffold size, treatment duration, measurement, and main findings.

Quality assessment
The quality assessment in selected studies was evaluated independently by 2 authors based on the risk of bias (RoB) tool of Systematic Review Centre for Laboratory Animal Experimentation (SYRCLE) (Hooijmans et al., 2014).The tool contains 8 criteria designed to determine the appraisal of validity, which was assessed as low, high, or unclear.The following 8 questions were included: 1) Was the allocation sequence adequately generated and applied?2) Were the groups similar at baseline or adjusted for confounders?3) Was the allocation adequately concealed?4) Were the animals randomly housed during the experiment?5) Were the caregivers and/or investigators adequately blinded? 6) Were animals selected at random during outcome assessment?7) Was the outcome assessment adequately blinded?8) Were incomplete outcome data adequately addressed?Furthermore, the other two questions were applied to avoid excessive items being rated as unclear because of poor reporting details of included studies: 1) Was it stated that the experiment was randomized at any level?2) Was it stated that the experiment was blinded at any level?"(Hutmacher and Sittinger, 2003;Chen et al., 2012).When evaluating the quality of included studies, the quality of question 4 was scored as low if all experimental interventions were present in one animal.In addition, the risk of bias for the sixth question was always considered low if the outcome of control and intervention groups of included studies was assessed at the same time.Any disagreements were then resolved through discussion and consensus with all the reviewers.

Study selection
A total of 9140 papers were initially retrieved from electronic search, including 1852 articles from PubMed/MEDLINE, 2570 articles from Embase, 2010 articles from Scopus and 2708 from Web of Science.A manual search of the included references yielded a further 5.After removing the duplicates, 4634 publications remained.None of the 55 articles retrieved from the gray literature was considered eligible.Of these, 4492 were excluded after the assessment of abstracts and titles.After the full-text reading, 106 publications were excluded since they did not meet the eligibility criteria.Finally, 36 studies were included in this systematic review (Figure 1).Among them, 30 studies evaluated the potential of PCs in bone formation, 4 studies assessed the PCs in cartilage regeneration and 2 included both in one study.
Regarding the cartilage capacity, all 6 studies evaluated the chondrogenic ability of PCs and demonstrated promising results (Li et al., 2011;Matsushima et al., 2011;Chang et al., 2012;Leijten et al., 2016;Iuchi et al., 2020;Yoo et al., 2021).4 studies reported significant differences in new cartilage regeneration (Li et al., 2011;Chang et al., 2012;Leijten et al., 2016;Yoo et al., 2021).One study (Li et al., 2011) compared the chondrogenic ability of BMSCs and PCs and showed the stronger chondrogenic potential of BMSCs.The chondrogenic ability of synovial membrane MSCs, adipose-derived MSCs and muscle membrane MSCs were also evaluated and demonstrated (Li et al., 2011).Furthermore, another study compared the capacity of bone and cartilage formation in periosteum from different sources (Matsushima et al., 2011).The results showed that cranial and mandibular periosteal tissues increased the bone and cartilage formation capacity most and least prominently, respectively.

Quality assessment
Figure 2 and Table 4 summarize the risk of bias in the included studies.Regarding selection bias, 18 studies included the randomization of the experimental process, while the sequence generation of the remaining 18 studies was considered an unclear risk of bias since they did not mention the randomization.Among the 36 included studies, 26 indicated that the baseline characteristics such as age, gender and weight were similar between the experimental and control groups.19 of the studies were considered a low risk of bias since they mentioned the allocation concealment.However, 3 studies presented a high risk of bias in allocation concealment because the experimenters were aware of which group the samples came from.Furthermore, for performance bias, 21 researches were assessed as low risk in terms of "random housing," while other the 15 studies had an unclear risk because the authors could not determine if the animals were randomly housed in the experiments.Unclear bias risks in terms of blinding were identified in 23 studies.Regarding detection bias, 34 studies were assessed as low bias risk in "random outcome assessment," while 2 studies had a high risk of bias because animals were not randomly selected.In the seventh item, 10 of the included studies were considered a low risk of bias because of the use of blinding for outcome assessment.For attrition bias, 30 studies were assessed as low risk, while 3 studies presented a high risk of bias because of the non-use or exclusion of incomplete data.For the two additional questions, 24 studies stated that the experiment was randomized at any level, while only 11 researches indicated that the experiment was blinded at any level.
Overall, 25 of the included studies presented a low risk of bias, and 11 researches were regarded as an unclear risk of bias, none of the included studies were scored as a high risk in the quality assessment.19 of the 27 studies which reported positive results in new bone formation present a low risk of bias although the other 8 publications showed an unclear risk of bias.However, half of the studies (3/6) that evaluate the cartilage formation capacity of PCs showed an unclear risk of bias.In addition, 6 studies indicated the greater bone formation capacity of PCs compared to BMSCs with a low risk of bias, and one showed the stronger chondrogenic potential of BMSCs also presents a low risk of bias.

Discussion
The objective of this study was to summarize the potential of PCs in terms of bone and cartilage regeneration.Despite an exhaustive search, only 36 articles informed the conclusions of our study, most of which focused on the osteogenic capacity of PCs.To our knowledge, this is the first review that focuses on the characteristics and efficacy of these cells in bone and cartilage regeneration in different models.
The potential of PCs for bone regeneration was first proposed in the 19th century (Nakahara et al., 1991).PCs as the source of MSCs in humans for bone tissue generation have also been proved in current studies.After conducting a comprehensive systematic review, the authors found most publications reported positive results in new bone formation with a combination of PCs and multiple scaffolds, including β-tricalcium phosphate (β-TCP), 3D collagen, BioOss, Collagraft, and Polydioxanone/pluronic F127 (Jaquiéry et al., 2005;Eyckmans and Luyten, 2006;Sakata et al., 2006;Agata et al., 2007;Chen et al., 2011;Lee et al., 2011;Ryu et al., 2011;Chang et al., 2012;Chen et al., 2012;van Gastel et al., 2012;Annibali et al., 2013;Lee et al., 2013;Chen et al., 2015;Katagiri et al., 2019).Scaffolds with PCs present significantly higher bone regeneration efficacy than bare scaffolds (Perka et al., 2000;Miyamoto et al., 2004;Moreira-Gonzalez et al., 2005;Sakata et al., 2006;Ribeiro et al., 2010b;Chen et al., 2011;Knothe Tate et al., 2011;Li et al., 2011;Chang et al., 2012;Chen et al., 2012;Stockmann et al., 2012;Zhang et al., 2012;Lee et al., 2013;Chen et al., 2015;Leijten et al., 2016;Yin et al., 2018;Lammens et al., 2020).However, certain biocompatible scaffold materials may not be suitable for in vivo implantation (Annibali et al., 2013).Moreira-Gonzalez et al. (2005) found that when repairing rabbit cranial bone defects, the sole implantation of 45S5 bioactive glass was unfavorable for defect repair, possibly due to the release of soluble silica from 45S5 bioactive glass into the environment, which influenced cell metabolism.In addition, in vivo experiments using β-TCP scaffolds indicated that scaffolds loaded with human PCs exhibited more neoangiogenesis and mature bone formation compared to those loaded with BMSCs (Chen et al., 2011).Studies have revealed that the characteristics of scaffolds may influence the behavior of implanted cells and ultimately impact the regenerative outcomes of bone tissue engineering (Ryu et al., 2011).To achieve cellular bone reconstruction and remodeling on a scaffold material, two key aspects need to be considered.The first one is that the provided cells should possess strong osteogenic ability, be non-immunogenic, and be easily obtained and manipulated.In addition, the scaffold material should exhibit good biocompatibility, strong osteoconductive properties, excellent absorbability, support MSCs attachment, and promote rapid vascularization (Perka et al., 2000).
Another interesting area regarding the osteogenic differentiation capability of PCs of the included studies is the influence of donor's age and sources.Regarding the potential influence of donor cell age on osteogenic differentiation capability, researchers concluded that as donor age increases, the thickness and cellular structure of the periosteum decrease (Jaquiéry et al., 2005).The osteogenic potential of PCs from different donor sources can vary among different tissues.For example, one of the included studies compared the capacity of bone and cartilage formation in periosteum from different sources (Matsushima et al., 2011).After 20 weeks of the implantation of PCs, the calvarial periosteum exhibited significantly higher expression of the runx2 and BSP, indicating strong osteogenic potential.On the other hand, the mandibular periosteum constructs showed slower development, and overall gene expression levels analyzed were not high.Accordingly, the osteogenic differentiation abilities of PCs to bone defect may be influenced by factors such as the age of donor cells and the donor sources.
In addition to the osteogenic potential of PCs, recently, researchers have focused on studying the potential of PCs to differentiate into cartilage and exploring their ability to repair bone defects.The inner layer of the periosteum contains osteoprogenitor cells, chondrocytes, and other osteogenic precursor cells, which can serve as the main source for chondrocyte production.In vitro experiments have shown that different types of induction culture media can promote the differentiation of PCs into osteoblasts, chondrocytes, and adipocytes, indicating the characteristics of mesenchymal stem cells (van Gastel et al., 2012).Chang et al. (2012) prepared functional PCs sheets from the periosteum of the rabbit tibia and transplanted them into the tibial tendon tunnel.Morphological and histological staining after 8 weeks demonstrated enhanced fibrocartilage formation at the tendon-bone interface, increased collagen fibers, and glycosaminoglycan deposition.In the present study, all 6 studies assessing the chondrogenic ability of PCs demonstrated promising results, and 4 of them reported significant differences in new cartilage regeneration.Accordingly, the potential of PCs in cartilage regeneration could be a promising strategy in tissue engineering.
Inducing MSCs to differentiate into cartilage can be achieved through various methods, such as modifying cell-loaded biomaterials with biomimetic elements like proteins or peptides, and performing in vitro pretreatment of the implant.Essentially, these approaches aim to create a microenvironment conducive to cartilage formation.Scholars abroad have found that when PCs micro-aggregates are integrated into biomaterials without exogenous growth factors, compared to single-cell-loaded biomaterials, the former exhibits upregulation of cartilage formation genes and improved formation of cartilage tissue in vivo (Leijten et al., 2016).Different sources of periosteal tissue may have an impact on cartilage formation.Iuchi et al. isolated PCs from the skull, mandible, radius, and ilium, and combined them with three-dimensional hydroxyapatite-poly(l-lactic acid-co-εcaprolactone) (HA-P[LA/CL]) scaffolds, which were then implanted into nude mice.PCs from the tibia of the lower leg showed better bone formation and maturation of chondrocytes in the engineered phalanges (Iuchi et al., 2020).
One of the main objectives of the present study was to pay attention to the comparison between the PCs and BMSCs.The results of those studies evaluating the differences between PCs and BMSCs indicated that PCs exhibited stronger in vivo osteogenic differentiation capabilities.For example, Chen et al. cultured human PCs and BMSCs and compared their osteogenic differentiation capabilities in vitro and in vivo (Chen et al., 2011).The results showed that human PCs demonstrated greater mineralization ability than BMSCs, with higher expression levels of osteopontin, BMP-2, and osteocalcin genes.Studies have shown that the periosteum contains more MSCs compared to bone marrow stroma, and PCs express more osteoprogenitor and chondroprogenitor cells than BMSCs (Zhu et al., 2006;van Gastel et al., 2012).However, no significant differences were found in the histomorphometric analysis of new bone formation among the different sources of MSCs in another study (Stockmann et al., 2012).Ribeiro et al. implanted carriers containing autologous PCs and BMSCs into extraction sockets of adult Beagle dogs.Although the PCs group showed a trend towards higher new bone area values, there were no significant differences in the formation of mineralized nodules and expression of bone markers between the two groups (Ribeiro et al., 2010a).Other sources of MSCs, such as dental pulp stem cells, adiposederived MSCs; periodontal ligament cells and muscle membrane MSCs have been also investigated in certain studies.However, conclusive conclusions cannot be drawn due to the experimental variabilities that existed and the limited available research.In addition, only one study evaluated the chondrogenic ability of BMSCs and PCs (Li et al., 2011).Although the results showed the stronger chondrogenic potential of BMSCs, the limited available research restricts our ability to draw conclusions.Accordingly, further research is needed to elucidate the differences between PCs and BMSCs and determine which MSCs from different tissue sources have the advantages in terms of chondrogenic potential.
The present study has some limitations.First of all, despite an extensive study, only 36 studies were selected in this systematic review, and only 6 included articles evaluated the cartilage regeneration capacity of PCs, which restricted us from drawing conclusions.In addition, because of the dissimilarity in settings, such as animal models and scaffold types, and most importantly, outcome characterization, a metaanalysis was not feasible.Therefore, a systematic narrative synthesis approach was adopted in accordance with the research questions proposed to thematically explore the results.Further clinical trials and experimental studies are required to confirm the results of this study.

Conclusion
After conducting a comprehensive literature review, the potential role of PCs in bone and cartilage regeneration has been demonstrated in the current literature.PCs demonstrated beneficial to bone regenerative efficacy compared to the bare scaffold with a low risk of most (19/27) studies reported.However, the cartilage formation capacity of BMSCs still needs to be investigated due to the limited researches available and the certain risk of bias.Moreover, PCs exhibited higher osteogenic capabilities compared to BMSCs in combination with various scaffolds in vivo with good evidence.However, the comparative benefits between the PCs and other sources of MSCs in cartilage regeneration remain uncertain.Further researches are required to confirm these results and determine the advantages of MSCs from different tissue origins in terms of chondrogenic and osteogenic potential.

FIGURE 1
FIGURE 1Flow chart of the literature search and results.

TABLE 1
Electronic databases used and search strategies.

TABLE 2
Characteristics and main results of the included studies.
(Continued on following page)Frontiers in Bioengineering and Biotechnology frontiersin.orget al.

TABLE 2 (
Continued) Characteristics and main results of the included studies.

TABLE 3
Information related to scaffolds used in tissue engineering.

TABLE 3 (
Continued) Information related to scaffolds used in tissue engineering.