Introduction

Hyaline cartilage facilitates smooth movement of articular joints and transmission of mechanical forces. The mechanical strength of cartilage tissue is provided by the combination of highly organized collagen arcades and negatively charged proteoglycans that draw water into the tissue1. Persisting damage to this structural organization leads to a change in the distribution of forces and loss in mechanical strength2. Cartilage injury can be post-traumatic, where defects are generally isolated, or it can occur during the progression of osteoarthritis (OA) where defects can emerge simultaneously. Both focal defects and OA impair quality of life leading to pain, reduced mobility, and disability3,4. As the healthy articular cartilage is an avascular tissue, its endogenous healing capacity is limited.

Adult chondrocytes, the cells residing in articular cartilage, are used to treat cartilage defects in autologous chondrocyte implantation5. Due to the low cell density in cartilage, chondrocytes are culture-expanded to obtain a sufficient number of cells for treatment. Expansion of chondrocytes is limited in population doublings6, as they tend to acquire a fibroblastic appearance and lose their chondrogenic phenotype7,8, before becoming senescent. Alternatively, the use of mesenchymal stromal cells (MSCs) for cartilage repair has been evaluated extensively in clinical studies9. Despite their capacity to generate cartilaginous tissue, MSCs have a tendency for differentiation into hypertrophic chondrocytes and subsequent endochondral ossification10. In contrast, MSCs are suggested to have chondro-inductive effects when combined with autologous chondrons for the treatment of focal cartilage defects11.

A distinct population of endogenous progenitor cells that resides in articular cartilage, named articular cartilage-derived progenitor cells (ACPCs), has been described in the last two decades12,13,14,15. The key in vitro characteristics of ACPCs include stem cell-like properties such as clonal expansion, extensive proliferation, and differentiation potential into multiple mesenchymal lineages, including the chondrogenic lineage. ACPCs were first identified in bovine cartilage16, and later also in different species, including equine7,13 and human cartilage17,18. Interestingly, ACPCs were shown not to upregulate type X collagen gene expression in vitro, a marker for hypertrophic differentiation during redifferentiation, contrary to MSCs7,13. The use of an endogenous cartilage progenitor cell population for treatment of cartilage defects and tissue engineering purposes therefore seems favorable over the use of other cell types14,19,20. Yet, isolation protocols and specific characterization for these cells differ greatly amongst researchers. In addition, a wide range of terms is being used to name the cells, like chondrogenic progenitor cells, cartilage stem cells, mesenchymal progenitor cells, or cartilage-derived stem/progenitor cells. For clarity, this review refers to ACPCs to address all endogenous progenitor populations identified in adult hyaline cartilage and characterized in vitro.

The purpose of this review is to systematically evaluate the available literature on ACPCs derived from healthy and diseased adult articular cartilage. We summarize the state-of-the-art research and discuss its potential for clinical use in cartilage repair therapies.

Results

The literature search yielded 1017 studies in EMBASE and 662 studies in PubMed. After duplicate removal, 1064 studies were identified. After title and abstract screening, the full text of 180 studies was screened. A total of 84 studies were then found eligible based on the inclusion and exclusion criteria (Fig. 1).

Fig. 1: Flow diagram of the literature search.
figure 1

Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) workflow showing systematic selection process for studies.

Markers to identify ACPCs in vivo

The presence of ACPCs was first described by Dowthwaite et al. in 200416. Enhanced expression of fibronectin and one of its key receptors, integrin-α5β1, was found in the superficial zone of bovine articular cartilage. Isolation of this fraction resulted in a population with high clonogenicity. As a unique marker or marker set is lacking, MSC or chondrocyte markers are mostly used for identification (Table 1). Classical MSC markers CD10521,22,23, CD16624, CD14625, VCAM26, or combinations including these markers27,28 have been used by others. In essence, this results in the identification of an MSC-like population in articular cartilage. Additional markers have been described to identify ACPCs in the tissue more specifically. Proteins involved in the Notch signaling pathway, like Notch-1, Notch-2, Delta, and Jagged26,29, or integrin-α5β121, proteoglycan 4 (PRG4, or lubricin)30, and laminin31 are used. Alternative approaches to identify ACPCs in cartilage tissue have focussed on visual distinction by an elongated cell morphology of ACPCs in cartilage tissue samples31,32, cell clustering of ACPCs33, proliferation marker Ki-6733,34, and migration of ACPCs upon stimulation of the cartilage35.

Table 1 Identifying an ACPC population in articular cartilage.

Methods for isolation of ACPCs from cartilage

A protocol for selective isolation of ACPC by differential adhesion to fibronectin (DAF) was established16,17,29, taking advantage of the enriched expression of the fibronectin receptor16 and the finding that isolation based on integrins resulted in selection for stem cells rather than transit-amplifying cells36. In two-thirds of the studies using DAF, this protocol is followed by isolation of colonies, that are subsequently formed by the cells that adhere (generally) in 20 min13,14,17,18,22,29,37,38,39,40,41,42,43. Six out of nineteen studies did not perform colony isolation and the complete pool of cells that adhered to fibronectin was isolated34,44,45,46,47,48.

Alternatively, ACPCs are sorted from the total cell population either via immunomagnetic separation or fluorescence-activated cell sorting (FACS). ACPCs were isolated by FACS based on co-expression of CD105 and CD16615, a marker combination that defines a subset of bone marrow-derived MSCs49 and was proposed to select for ACPCs. Another marker set used for cell sorting that resulted in an ACPC population is CD9+/CD90+/CD166+50.

Finally, cells migrating out of cartilage explants, whether or not the cartilage is stimulated in any way, hold progenitor characteristics such as multilineage differentiation potential and colony forming efficiency (CFE)19,28,32,51,52,53. These migratory cells were distinctly different from chondrocytes and osteoblasts54. To stimulate migration of cells, explants were stimulated by nerve growth factor (NGF)52, platelet lysate19, or migrating cells were isolated after partial digestion of the tissue by collagenase28. Cell migration could also be triggered by induction of injury32,53. Cells with progenitor characteristics migrated towards the site of cartilage injury and a role in the repair of adult cartilage upon damage was suggested by the authors32.

Nine studies did not report on any distinct method to isolate a population from the total cell population7,30,55,56,57,58,59,60,61. Five others performed an isolation step after one or two passages in culture24,25,62,63,64. It can therefore be questioned whether these are investigating a population that is different from what is generally referred to as chondrocytes, as most of the studies were also lacking a chondrocyte control group.

In vitro characterization of ACPCs after isolation

Isolated ACPCs are characterized based on their proliferative potential, CFE, differentiation potential, and expression of markers that are also used for their isolation (Table 2). ACPCs could be maintained in culture for up to 30–60 population doublings17,18,37,38,39 and early-passage cells were able to form colonies in culture7,16,18,19,28,32,39,46,52,65,66. Moreover, human ACPCs were found to maintain telomere length and telomerase activity up to at least 20 population doublings18,37. However, ACPCs derived from OA cartilage contained a subpopulation of cells that have reduced proliferative potential and undergo early senescence when cultured in vitro18.

Table 2 Isolation and characterization of ACPCs.

ACPCs could be differentiated into the chondrogenic, osteogenic, and adipogenic lineage, a feature that MSCs also possess67. There is one report of reduced osteogenic differentiation potential of ACPCs42, while 20% of the studies looking into multilineage potential found indications for reduced osteogenesis12,13,19,20,50,68.

Surface marker expression of ACPCs was in general similar to MSCs, with ACPCs being positive for CD90, CD105, CD73, and CD166, while negative for hematopoietic markers, highlighting the challenge to distinguish the two cell types13,19,28,30,40,41,42,51,52,56,63. Of note, about half of the studies mentioned here examine immunophenotype of cells in culture13,19,22,28,30,34,40,42,44,47,51,56,61,66,69, while cells tend to change their phenotype during in vitro expansion70,71. Moreover, investigating marker expression by gene expression or flow cytometry on the bulk populations makes it problematic to define whether these markers are co-expressed or not.

In vitro comparison of ACPCs to other cell types with regard to surface marker expression

Cell surface marker expression and in vitro performance of ACPCs were directly compared to MSCs from various sources, like bone marrow13,14,15,30,32,34,40,51,55,59,64, adipose tissue25,30,40, and infrapatellar fat pad46,59,60. Other cell types compared are chondrocytes17,19,22,25,29,32,34,37,41,42,46,48,53,65,66,72 and other intra-articular cells, like synoviocytes53,59,60, synovial fluid cells53, and periosteal cells59,60,64 (Table 2).

A clear distinction between MSCs and ACPCs based on the expression of markers was only reported once, when equine ACPCs were compared to bone marrow-derived MSCs, an increase in gene expression for CD44 was found13. One-third of the studies directly compare ACPCs to chondrocytes, as these also reside in adult hyaline cartilage, and distinction of these cell types is crucial for isolation and application. The proliferation of ACPCs was faster than chondrocytes in one study19, but slower in a different report42. In addition, ACPCs were found to form more colonies compared to chondrocytes32. A distinction was made between chondrocytes and ACPCs based on high expression of CD9017,25,34, CD4434, CD10546, CD16646, Notch-117, and HLA-ABC25 in ACPCs while culture-expanded chondrocytes showed little to no expression of these markers. Co-expression of CD44 and CD90 was found to distinguish between rat chondrocytes and ACPCs34,45. When ACPCs were sorted from the total pool of chondrocytes by CD49e-expression, a difference was found in the expression of CD29 in chondrocytes (50%) versus ACPCs (100%)66. When ACPCs were treated with platelet lysate, an increased expression of CD166 and decreased expression of CD106 compared to chondrocytes was found19.

Differences between species

Identification of similarities and differences in ACPCs between species is challenging, due to the diversity of isolation procedures and variety of study objectives. Colony formation was identified in several human studies, as well as in the first report on bovine ACPCs. CFE in bovine cartilage cells was reported to be 0.6%16, while all other literature on non-human cells lacked this analysis. In human cells, consistency is found to some extent. CFE in healthy cartilage cells on fibronectin-coated dishes was 1.47%39, while this was almost double (2.8%) in OA cells in the same study. Others reported on CFEs of <0.1%18 and 0.66%66 of OA cells on fibronectin-coated dishes. When OA cells were seeded on uncoated culture plastic, a CFE of <0.01% was found57,58,59. The percentage of colony forming cells increased when cells were culture expanded. Passage one OA cells (isolation method not specified) had 18% CFE30 and the same passage cells that migrated from OA tissue in response to platelet lysate had 7.8% CFE19. Cells that migrated from OA tissue with NGF and were expanded for four passages had increased their CFE to 38.6%52. When CD105+/CD166+-sorted cells were quantified, CFEs of 3.5% (healthy) and 8% (OA) were found in one study15 and 15% (healthy) and 17% (OA) were found in another24. Of note, the latter used cells that were culture expanded for one passage. Overall, when comparing human ACPC studies, it seems that OA tissue contains more colony forming cells than healthy cartilage. Also, CFE increases after culture expansion, possibly as a consequence of culture-related changes in immunophenotype71.

Differences between ACPCs from healthy and osteoarthritic cartilage

ACPCs have been identified in hyaline cartilage from different pathological states. Identification and characterization can contribute to our understanding of their role in homeostasis and disease, as well as their accessibility for clinical use.

In healthy articular cartilage, ACPCs most likely reside in the superficial zone, as Notch-1-expressing cells are found here16 and possess progenitor cell characteristics17,26,29. In addition, enhanced expression of fibronectin and one of its receptors, integrin-α5 and -β1, was found in the superficial zone16. As a direct consequence, most of the cells isolated via DAF originated from the superficial zone. The same group also showed that the CFE of surface zone cells is higher compared to deep zone cells16.

Upon damage of cartilage, ACPCs seem to migrate towards the site of injury73. Cells that migrated into the site of injury were found to possess progenitor-like characteristics32. An increase of CD271-expression was seen in ACPCs from increased OA severity28. Classical MSC markers CD10521,22, VCAM26, or combinations including these markers27,28 were all enhanced in OA cartilage or upon trauma. A shift of expression of PRG430 from the superficial layer to deeper zones was seen in OA, whereas CD271- and CD105-positive cells shifted towards the superficial zone in OA28. In OA cartilage, cell clusters were observed which express ACPC-associated markers like Notch-1, Stro-127, VCAM, FGF-2, and Ki-6726,33. These cells proliferated faster and produced more cartilaginous nodules in vitro compared to cells isolated from macroscopically healthy cartilage33. Contradictory, others found that ACPCs derived from healthy cartilage proliferated faster than OA-derived ACPCs63. Lastly, a high number of CD105/CD166-positive15 and CD146-positive cells25 was found in OA cartilage and these cells had multilineage potential. OA-derived cells also formed more colonies compared to cells from normal human cartilage39 and this increased with OA severity28.

In vitro culture of ACPCs

Three studies made an attempt to optimize growth kinetics examining factors like seeding density, culture systems, and serum concentrations74,75,76. The authors reported on optimal expansion conditions when the medium was supplemented with fetal bovine serum (FBS) and transforming growth factor (TGF)-β1 at 40% and 1 ng/mL, respectively74. However, more recent studies have not used FBS concentrations that were as high as 40%. A passage length of 5 days was optimal for cell yield and the authors reported on reduced costs of expansion by 60%76. Furthermore, a method for expansion on microcarriers eliminated the need for a harvesting step and was thus suggested to prevent dedifferentiation75. A direct comparison of fibronectin versus laminin, another important cell adhesion molecule, for differential adhesion of ACPCs resulted in higher population doubling, increased gene expression of type II collagen, and increased osteogenic and adipogenic differentiation potential of laminin-selected ACPCs69. Likewise, expansion with platelet lysate compared to FBS showed more population doublings and increased expression of chondrogenic genes aggrecan and type II collagen, but at the same time expression of type X collagen was also increased77. Others found increased gene expression of aggrecan, type II collagen, and Sox9, as well as proteoglycan and type II collagen production of ACPCs by application of intermittent hydrostatic pressure46 or mechanical stimulation in a bioreactor system78, and inhibition of CFE by high glucose levels during growth culture60. Moreover, normoxic versus hypoxic conditions revealed greater production of glycosaminoglycans, low alkaline phosphatase expression, and weaker type I collagen staining in both conditions compared to MSCs20. In line, consistently low levels of type X collagen were expressed by ACPCs when normoxia and hypoxia were compared68. A reduction of oxygen tension during culture is also known to delay chondrocyte aging and improve their chondrogenic potential79,80.

In brief, the optimization of culture conditions for ACPCs has been investigated extensively. There are no uniform protocols for expansion and optimal differentiation for cartilage formation. Consensus on these matters would aid in comparing outcomes of studies in the future.

Upon ex vivo injury of bovine cartilage, migratory cells with progenitor characteristics were found32,35. Additional research showed that the phagocytic capacity of these ACPCs was higher compared to chondrocytes and comparable to synoviocytes and macrophages, suggesting a macrophage-like role for ACPCs in cartilage injury81. After treating ACPCs with supernatant from injured explants, proliferation, migration, and expression of immunomodulatory mediators were enhanced, while chondrogenic capacity was impaired27.

Stimulation of chondrogenesis in ACPCs was successful by inhibition of the nuclear factor-κB pathway, the major signaling pathway involved in OA82. Inhibition of this pathway was achieved by an inhibiting peptide34 and magnoflorine45, both resulting in increased chondrogenesis. Interleukin-1Β and tumor necrosis factor-α, inflammatory factors involved in OA, were reported to inhibit migration of ACPCs51. Similarly, β-Catenin and NGF are elevated in OA83,84. Inhibition of the Wnt/β-Catenin pathway promoted proliferation and differentiation62, while NGF failed to stimulate chondrogenesis in ACPCs52. The specific role of these compounds in OA remains to be investigated.

Alternatively, chondrogenesis could be triggered by the direct activation of chondrogenic pathways. Combined mechanical stimulation and shear stress-induced chondrogenesis through an increase in endogenously produced TGF-β1, while overexpression of BMP2 reduced chondrogenesis78. Also, BMP9 was a potent stimulator of chondrogenesis85. Direct treatment of ACPCs with extracellular matrix components Link protein N-terminal peptide44 or nidogen-231 increased expression of chondrogenic genes. The proliferation of ACPCs was promoted by kartogenin86, a small molecule that induces chondrogenic differentiation of MSCs. Finally, sex hormones estrogen and testosterone influenced human ACPC performance87.

To summarize, initial results indicate that ACPCs respond to injury and chondrogenesis can be induced in vitro, which could make the cells interesting as therapeutic targets. These findings could be used to provoke neo-cartilage formation or inhibit inflammation in OA.

Application and translation of progenitors

The potential of ACPCs for tissue engineering, biofabrication, and clinical application has been investigated widely (Table 3). Biofabrication allows for the production of constructs consisting of (bio)materials, bioactive cues, and/or cells, with a detailed predefined architecture88. The extensive proliferative potential of ACPCs combined with their chondrogenic capacity make these cells good candidates to use in tissue engineering and biofabrication approaches to repair or regenerate articular cartilage.

Table 3 Application and translation of ACPCs.

Under the influence of intermittent hydrostatic pressure, the performance of rabbit ACPCs embedded in alginate was enhanced significantly. These cultures were pretreated for 1 week with a TGF-β3-containing medium but did not receive any exogenous growth factors thereafter. After two and four weeks, glycosaminoglycan, collagen, and DNA content were significantly higher than groups not treated with intermittent hydrostatic pressure46. Two studies investigating equine ACPC performance in hydrogels both reported good outcomes. When the cells were embedded in gelatin methacrolyl (gelMA) hydrogel cultured in a chondrogenic medium, mainly a difference was found in the expression of zonal markers compared to bone marrow-derived MSCs. Expression of PRG4 was increased in ACPC-loaded gels, while type X collagen expression was decreased compared to MSCs13. Furthermore, when equine ACPCs were embedded in gelMA/gellan and gelMA/gellan/HAMA hydrogels and cultured in a chondrogenic medium, these produced more glycosaminoglycans and type II collagen than chondrocytes, whereas the performance of MSCs in the same gels was comparable to ACPCs89. Similar to hydrogels, printed scaffolds have also been successfully seeded with ACPCs. Human ACPCs seeded on fibrin-polyurethane composite scaffolds were responsive to mechanical stimulation. The cells produced more glycosaminoglycans and aggrecan gene expression was increased without the addition of exogenous growth factors78. Furthermore, human ACPCs could also be seeded onto polycaprolactone/polylactic electrospun nanofibrous scaffolds where the cells attached and spread over the fibers. Further research has to shed light on the chondrogenic performance of the cells in this specific setting40.

Besides tissue engineering, ACPCs were successfully used in several biofabrication techniques. It was shown that equine ACPCs have the potential to be bioprinted and while exact mechanisms remain to be elucidated, an interplay between MSCs, ACPCs, and chondrocytes was found to be important for neo-cartilage synthesis13. The same cells were also successfully used for encapsulation in various hydrogels90,91,92,93 in combination with biofabrication techniques like extrusion-based bioprinting13,89, digital light processing94, and volumetric bioprinting95, while maintaining cell viability. While these are only the first indications to use ACPCs with various techniques, additional research is necessary to assess chondrogenic performance of the cells in these settings. Nevertheless, initial results are promising to move forward with this cell population.

Several attempts were made to take the next steps in the application of ACPCs for in vivo cartilage formation and repair. These are important to translate in vitro findings and define the potential of ACPCs for the clinic.

When DAF-selected ACPCs were applied in a caprine model for cartilage defect filling using a cell-seeded type I/III collagen membrane (Chondro-Gide®), ACPC-seeded scaffolds showed good lateral integration with the surrounding tissue and type II collagen-positive repair tissue. However, no difference was found between chondrocyte- or ACPC-treated defects17. In the same study, engraftment into the growth plate of developing chick hind limbs of isolated and culture-expanded ACPCs was shown. Contradictory, DAF-selected bovine ACPCs that were injected intramuscularly in immune-deficient mice failed to produce cartilage matrix38. In an equine model, DAF-selected ACPCs were applied in a layered biofabricated osteochondral plug and showed good integration with the native cartilage, but the repair tissue contained mainly type I collagen91. When autologous and allogeneic ACPCs were directly compared in an equine cartilage defect model, an advantage of autologous over allogeneic cells was seen in histological outcomes96.

When human ACPCs were used in immune-deficient mice, the cells were successful in the production of cartilage matrix, whereas MSCs produced mainly bone55. The cells in this study were not isolated using any distinct method for ACPC isolation but were 2D expanded in low density with low glucose. Furthermore, migratory human ACPCs expanded using platelet lysate outperformed chondrocytes in an in vivo ectopic chondrogenesis assay in athymic mice19.

Finally, an attempt was made to proceed to human application, by using ACPCs to replace chondrocytes for matrix-assisted autologous chondrocyte transplantation, similar to the caprine study mentioned earlier. The pilot study with 15 patients12 reported on repair tissue rich in type II collagen and proteoglycans and without types I and X collagen. Furthermore, IKDC and Lysholm questionnaire scores improved significantly. However, there was no direct comparison between ACPCs and expanded chondrocytes in this study.

While the discussed studies provide initial evidence of in vivo chondrogenic potential of these cells, further investigation is essential to ascertain promise cartilage repair and clinical translatability.

Discussion

With this review, we aimed to systemically evaluate the available literature on adult ACPCs and their use for cartilage tissue engineering and repair therapies. We are the first to provide a thorough overview of research from the last two decades that demonstrates the presence of a progenitor cell population residing in adult hyaline cartilage (Fig. 2). Although a great effort was made to study the identity and applications of ACPCs, many uncertainties remain. As a result of differences in isolation protocols, characterization, and culture expansion, most cell populations discussed in the literature are likely to be heterogeneous populations and difficult to compare between laboratories. This stresses the need for this systematic review to expose certain inconsistencies and arrive at a shared definition of ACPCs.

Fig. 2: Isolation, characterization, and application of articular cartilage-derived progenitor cells.
figure 2

Schematic overview of the identification of articular cartilage-derived progenitor cells (ACPCs) in cartilage, isolation methods, and applications of ACPCs. Created with BioRender.com.

The reviewed literature employs a wide variety of procedures for the isolation and characterization of ACPCs. Broadly speaking, three main methods for ACPC isolation are described. The method using DAF, used in 42% of the investigated studies, is based on the enriched expression of integrin-α5β1, as first described by Dowthwaite et al.16. The other two main methods are based on the expression of (a combination of) cell surface markers (19%)15,50 or migratory capacity (6%)51. Most populations isolated through these methods employed multilineage potential, responded to acute injury or mobilized during OA, and were able to produce hyaline cartilage extracellular matrix in vitro or in vivo.

The heterogeneity in isolation and characterization creates discrepancies between donors and laboratories. Direct comparisons of ACPC populations isolated through different procedures are lacking and would aid to improve our understanding of the populations. The identification of a unique cell marker would facilitate extensive and coordinated research into the cell type. This could pave the way towards clinical application or cell targeting to promote cartilage regeneration in OA. Recently, Gdf5-expressing cells in developing joints were identified to contribute to joint cell lineages97. Co-expression of Lgr5 and Col22a1 was identified as an important lineage marker towards juvenile articular chondrocytes in the developing mouse joint98. Additionally, single-cell RNA sequencing has revealed several novel markers that are potentially specific for ACPCs in human OA cartilage99.

The available literature suggests that ACPCs resemble MSCs67 in vitro based on surface marker expression and multilineage potential. The comparison to MSCs is often made due to the fact that MSCs (derived from various tissues) are a useful cell type for clinical application and are currently applied9. As the general view on the origin and role of MSCs is changing100, characterization of ACPCs based on MSC features might not be the way to go and other routes should be investigated. More recent work has shed light on the cellular basis of bone and cartilage formation by identifying skeletal stem cells in mice and humans101,102, a cell type that might be closer related to (the origin of) ACPCs in adult hyaline cartilage. Although the comparison to clinically used chondrocytes is relevant, research into similarities between ACPCs and skeletal stem cells or more downstream progenitor cells is lacking and finding resembling features would contribute to knowledge about the origin and identity of ACPCs.

Establishing the role of ACPCs in cartilage development and homeostasis, as well as their response upon injury or in OA would provide additional insights into their physiological function in mature cartilage. Regeneration in the early stages of OA could be stimulated or progression of the disease halted. Several studies discussed here suggested that ACPCs have a possible role in immunomodulation, based on their capacity to migrate upon injury32,35, excretion of inflammatory mediators27 and phagocytic capacities81. Others found a higher prevalence of these cells upon cartilage damage and OA21,22,26,27,28,30,32,33,34,35. Of note, these are all in vitro indications for which in vivo validation is essential. During OA, cell density and clustering in cartilage increases103, for which ACPCs might partly be responsible. On the other hand, there is contradicting evidence that Prg4-expressing cells from the synovium migrate into sites of acute cartilage injury and contribute to cartilage repair97. In order to expand the application of ACPCs to OA besides repair of chondral defects alone, immunomodulatory properties should be demonstrated in vivo as is known for MSCs104.

The described ACPC populations generally surpassed other cell types in proliferative potential and producing cartilage extracellular matrix in vitro19,20,25,42,46. In addition, most studies implanting animal-derived ACPCs in vivo confirmed their chondrogenic potential17,38,91, and even two studies using human cells reported on successful neo-cartilage formation19,55. As isolation methods do not seem to be associated with in vivo outcomes of cell performance and tissue formation, the challenge remains to compare findings between studies. Furthermore, differences between species and pathological states could influence cell performance. Donor age might play an important role, although none of the studies investigated this specifically. Nevertheless, the cells’ potency of prolonged in vitro expansion17,19,37 combined with limited tendency towards hypertrophic differentiation13,14,19,68 and their ability to form neo-cartilage can make ACPCs an appropriate cell type for repair of focal chondral defects.

Despite the great deal of research that has been done on ACPCs, certain actions need to take place in order to close the gaps and reach consensus between researchers and laboratories. As noted before, isolation based on a unique marker is crucial to ascertain similarity in cell population between laboratories. Comparison of culture media and additives for ACPCs in a recent systematic review105 highlights the importance of consistency to align research. As ACPCs currently have no discrete set of cell surface markers that can be used to isolate the cells from tissue, the question remains whether ACPCs are a distinct cell type or it refer to a heterogeneous mix of many cell types. The establishment of a cell marker and consistency in isolation and culture protocols ascertain comparability between populations. Another limitation might be the availability of tissue for cell isolation. Cartilage from OA patients is generally more accessible than healthy cartilage, as it is redundant after knee replacement surgery. A direct comparison of ACPC populations of healthy and OA cartilage would shed light on differences in performance. In the same view, investigation of allogeneic use of ACPCs is valuable, as this would greatly improve the potential of application by availability, reduction of costs, and preselection of chondrogenic cells.

The current systematic review is limited by the restriction to cell populations that are isolated from adult hyaline cartilage. The comparison and relation to cell types in the developing joint are lacking and would contribute to our further understanding of the origin of the populations discussed here and their role in joint development and homeostasis. However, the current review and discussed literature are predominantly directed at clinical translation as opposed to etiology or the role of a cartilage progenitor cell in development.

Arriving at a shared definition of a homogenous cell population that can be isolated and characterized in a comparable manner is crucial. This work could be used as a basis for research groups and clinicians to harmonize study protocols and characterization. First, studies should report on the origin of the cell in terms of species, anatomical location of the hyaline cartilage, and disease state. Second, the method of isolation should be described in detail and preferably identical to one of the established protocols. Finally, the phenotype of the isolated populations should be examined directly following isolation and culture media (and additives) as well as expansion time and/or passage number should be reported and synchronized.

To conclude, the available literature indicates that a cell population with progenitor-like characteristics resides in adult hyaline cartilage, which has extensive chondrogenic and proliferative potential. These features highlight the suitability of ACPCs as a cell source for focal chondral repair. In addition, it is crucial to investigate the role of ACPCs in development and disease, in order to determine their potential to slow down or reverse OA. If the current challenges can be overcome and consensus can be reached on this population, ACPCs hold great potential as a cell type for tissue engineering and for the repair of cartilage damage in both focal cartilage injury and OA.

Methods

Literature search

A systematic search of the literature was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines on adult endogenous ACPCs. The review protocol was prospectively registered with PROSPERO (registration number CRD42020184775). The electronic databases of EMBASE and PubMed were searched using the following search terms: (cartilage AND (articular OR hyaline OR knee OR hip OR ankle)) AND (progenitor OR progenitor cell OR multipotent cell OR chondroprogenitor OR multipotent cell OR cartilage-derived OR articular cartilage-derived OR (stem cell OR MSC OR mesenchymal stem cell OR mesenchymal stromal cell AND (cartilage-derived OR cartilage resident))). A final search was performed on 17 February 2021. Two authors (M.R. and J.V.K.) independently screened all selected studies for eligibility, first by title and abstract followed by full-text screening. After duplicate removal, inconsistencies between the researchers were discussed in a consensus meeting.

Inclusion and exclusion criteria

Inclusion criteria that were used during the title, abstract, and full-text screening for eligible studies included: adult endogenous cartilage stem/progenitor cells; knee, hip, or ankle cartilage; in vitro and/or in vivo and/or in man studies; English language; reviews, case reports, conference papers, studies of which the full texts were not retrievable, studies investigating cell line of chondroprogenitors, cells other than endogenous cartilage-derived progenitors, and lineage-tracing studies were excluded. Extracted data from the selected studies included species, anatomical location of cartilage, isolation procedure, cell characterization, and application of the cells. The quality of a study was considered inferior if methods or results are poorly reported. Study limitations/inconsistencies are discussed at the end of a paragraph in the results.