Mesenchymal Stromal Cells Laden in Hydrogels for Osteoarthritis Cartilage Regeneration: A Systematic Review from In Vitro Studies to Clinical Applications

This systematic review is focused on the main characteristics of the hydrogels used for embedding the mesenchymal stromal cells (MSCs) in in vitro/ex vivo studies, in vivo OA models and clinical trials for favoring cartilage regeneration in osteoarthritis (OA). PubMED and Embase databases were used to select the papers that were submitted to a public reference manager Rayyan Systematic Review Screening Software. A total of 42 studies were considered eligible: 25 articles concerned in vitro studies, 2 in vitro and ex vivo ones, 5 in vitro and in vivo ones, 8 in vivo ones and 2 clinical trials. Some in vitro studies evidenced a rheological characterization of the hydrogels and description of the crosslinking methods. Only 37.5% of the studies considered at the same time chondrogenic, fibrotic and hypertrophic markers. Ex vivo studies focused on hydrogel adhesion properties and the modification of MSC-laden hydrogels subjected to compression tests. In vivo studies evidenced the effect of cell-laden hydrogels in OA animal models or defined the chondrogenic potentiality of the cells in subcutaneous implantation models. Clinical studies confirmed the positive impact of these treatments on patients with OA. To speed the translation to the clinical use of cell-laden hydrogels, further studies on hydrogel characteristics, injection modalities, chemo-attractant properties and adhesion strength are needed.


Introduction
Osteoarthritis (OA) is a degenerative disease of the whole joint tissues that leads to a progressive loss of articular cartilage (AC), causing chronic pain and disability in the affected patients [1]. Different risk factors, such as age, gender, family history, obesity and traumatic injuries are involved in the pathogenesis of OA [2]. The evolution of OA is characterized by the production of catabolic mediators (interleukin (IL)1β, IL6 and tumor necrosis factor (TNF)α) responsible for the induction of inflammation and production of proteolytic enzymes (aggrecanases and matrix metalloproteinases (MMPs)) that contribute to producing damages in the joint tissues [3,4].
The AC protects bone surfaces within joints by providing a low-friction gliding surface for the articulation, supporting shock-absorption, distributing loads, reducing stresses on the subchondral bone and guaranteeing wear resistance [5]. The AC consists of chondrocytes immersed in an extracellular matrix (ECM) that is composed of 70% water and 30% organic components, such as aggrecan, collagen type 2, minor collagens (type 3-4, 9 and 11), proteoglycans, glycosaminoglycans and glycoproteins [6]. Lesions affecting AC have limited intrinsic capacity for self-regeneration [7]. Surgical, pharmacological or nonpharmacological treatments represent ways to only temporarily relieve the OA symptoms, making the regeneration of AC tissue an unmet clinical problem [8]. Recent progresses have demonstrated the potential of stem-cell-based therapies in the treatment of OA patients that are able to regenerate injured cartilage and at the same time attenuate on-going inflammation within the affected joint [9,10]. Thus, MSCs are stromal cells that can be isolated from different tissues such as bone marrow, adipose tissue, umbilical cord (UC) and blood, and they are capable of differentiating in cartilage, bone and adipose tissue [11,12]. They can produce growth factors, chemokines and cytokines and they possess the ability to migrate in the injured sites [13,14]. In vitro and in vivo studies have proved that MSCs with their secretome promote anti-inflammatory effects and contribute to the formation of cartilage tissue [15][16][17].
One of the most promising strategies in this regard is the use of MSCs combined with biomaterials [18,19]. Biomaterials must possess peculiar characteristics to be used for cartilage regeneration. Among the biomaterials, hydrogels are three-dimensional polymeric matrices that shows interesting features and that have potential for the treatment of cartilage defects [20]. In fact, hydrogels can be injectable or printable and can effectively embed viable cells without hampering their viability [20]. Hydrogel printability enables the creation of well-defined three-dimensional (3D) structures through 3D printing or other biofabrication technologies to mimic cartilage native tissues. Hydrogels are being widely employed as bioinks in 3D printing due to their tunable and injectable features [21,22]. A wide range of natural or synthetic biopolymers are available that can be combined to form hybrid hydrogels [20]. Moreover, gelation of hydrogel matrices can be achieved by physical or chemical crosslinking, creating structures with extraordinary water absorbing ability and a 3D network, such as the ECM [23]. The ECM of AC is rich mainly in polysaccharides and proteins, and hydrogels fabricated from these biopolymers have been widely studied. Recently, the importance of the properties of the hydrogel microenvironment that contribute to regulate stem cell chondrogenesis has been shown [24]. Among them, hydrogel mechanical properties such as stiffness and viscoelastic behavior have a role in guiding the cells to differentiation [25]. Stiffness is the capacity of a hydrogel to resist deformation in response to an applied force [26]. Viscoelasticity is the capacity of a hydrogel to exhibit both viscous and elastic behavior following the application of force [27]. Moreover, microstructural and spatial hydrogel properties such as porosity and anisotropy (hydrogel with well oriented structure) that create the architecture of the hydrogels also significantly influence the chondrogenic differentiation of MSCs [18,28,29]. Finally, it has been demonstrated that functionalized hydrogels with peptide or nanoparticles show positive effects on chondrogenic differentiation, with the ability to regulate cell activity and to show a tunable biodegradation profile [30][31][32][33].
Hydrogels can be excellent hosts for MSCs, and the therapeutic advantage of this strategy is to protect the cells injected into the defect (i.e., from shear forces and needle dimensions) and at the same time to favor their adhesion to the cartilage [20,25,26,34]. In fact, the hydrogels provide mechanical support, elasticity and stiffness and facilitate cell interactions with OA cartilage [5,25,26,35]. Finally, MSCs laden in hydrogel might contribute to ECM remodeling and maintenance of homeostasis [18,23,24]. Different narrative reviews have been published on this topic focusing on specific items [15,35,36]. This systematic review aims to use a defined search strategy that focuses on the main important characteristics of the hydrogels (material type, biofunctionalization, rheological properties, physical property and crosslinking methods), combined with different sources of MSCs and used in in vitro/ex vivo studies, in vivo OA models and clinical trials for favoring cartilage regeneration in OA.

Search Strategy
A systematic review was conducted in Pubmed and Embase databases from January 2011 to July 2022 considering the following keywords: adipose stem cells; mesenchymal stem cells; stromal cells; osteoarthritis; knee osteoarthritis; cartilage; chondrogenesis; carti-Cells 2022, 11, 3969 3 of 20 lage regeneration and differentiation. The query box used for each study was "((mesenchymal stem cells OR stromal cells OR adipose stem cells) AND (hydrogel)) AND (cartilage OR chondrogenesis OR differentiation OR cartilage regeneration) AND (osteoarthritis OR knee osteoarthritis))", filters "all fields". Two independent researchers (Y.S. and E.G.) performed the screening process. Moreover, to overcome problems related to risk of bias assessment, we did not use a validated tool of assessment, but we scored the bias risk only if we found items that were not reported or unclearly reported. Finally, any disagreements were resolved by consensus with a third reviewer (C.M.).

Selection Process
The selection of studies to be included was carried out following the PRISMA guidelines for systematic reviews. Starting from the results of Embase and Pubmed databases, a screening of the title and abstract and subsequently of the entire text of the article was performed using the free tool Rayyan online Software (https://www.rayyan.qcri.org, Qatar) (accessed on 30 July 2022). Articles written in other languages, abstracts, reviews, full texts not available, editorials or conference proceedings were excluded. The entire selection process is represented in the flow chart shown in Figure 1.

Search Strategy
A systematic review was conducted in Pubmed and Embase databases from January 2011 to July 2022 considering the following keywords: adipose stem cells; mesenchymal stem cells; stromal cells; osteoarthritis; knee osteoarthritis; cartilage; chondrogenesis; cartilage regeneration and differentiation. The query box used for each study was "((mesenchymal stem cells OR stromal cells OR adipose stem cells) AND (hydrogel)) AND (cartilage OR chondrogenesis OR differentiation OR cartilage regeneration) AND (osteoarthritis OR knee osteoarthritis))", filters "all fields". Two independent researchers (Y.S. and E.G.) performed the screening process. Moreover, to overcome problems related to risk of bias assessment, we did not use a validated tool of assessment, but we scored the bias risk only if we found items that were not reported or unclearly reported. Finally, any disagreements were resolved by consensus with a third reviewer (C.M.).

Selection Process
The selection of studies to be included was carried out following the PRISMA guidelines for systematic reviews. Starting from the results of Embase and Pubmed databases, a screening of the title and abstract and subsequently of the entire text of the article was performed using the free tool Rayyan online Software (https://www.rayyan.qcri.org, Qatar)(accessed on 30 July 2022). Articles written in other languages, abstracts, reviews, full texts not available, editorials or conference proceedings were excluded. The entire selection process is represented in the flow chart shown in Figure 1.   As shown in Figure 2, all the information extracted from selected papers were grouped and the review was organized focusing on the following main items: hydrogel features (material type, biofunctionalization, rheological properties, physical property and crosslinking methods), MSC types and cell loading, experimental design (in vitro or ex vivo), in vivo OA models and clinical trials. Figure 2, all the information extracted from selected papers were grouped and the review was organized focusing on the following main items: hydrogel features (material type, biofunctionalization, rheological properties, physical property and crosslinking methods), MSC types and cell loading, experimental design (in vitro or ex vivo), in vivo OA models and clinical trials.

Literature Search Strategy Results
The initial literature search retrieved 95 articles using PubMed and 110 articles using Embase by using the mentioned keywords. The selected references were submitted to a public reference manager Rayyan Systematic Review Screening Software (Qatar) to eliminate duplicate articles (n = 101). All the remaining abstracts (n = 104) were screened for excluding conference presentations (n = 9), reviews (n = 22), full text not available (n = 1) and medical hypothesis (n = 1) not matching with the inclusion criteria. A total of 71 articles were considered eligible. By evaluating the full text of all of the articles, a total of 29 articles were excluded because they focused on the following: regeneration of meniscus (n = 5), osteochondral lesions (n = 10), bone (n = 1), exosome treatment (n = 1), non-OA pathologies (n = 2), adipose or bone differentiation (n = 5) and hydrogels without MSCs (n = 5). As reported in Figure 1, a total of 42 studies were finally included in this review: 25 articles concerned in vitro studies, 2 in vitro and ex vivo results, 5 in vitro and in vivo results, 8 in vivo studies and 2 clinical trials.

Literature Search Strategy Results
The initial literature search retrieved 95 articles using PubMed and 110 articles using Embase by using the mentioned keywords. The selected references were submitted to a public reference manager Rayyan Systematic Review Screening Software (Qatar) to eliminate duplicate articles (n = 101). All the remaining abstracts (n = 104) were screened for excluding conference presentations (n = 9), reviews (n = 22), full text not available (n = 1) and medical hypothesis (n = 1) not matching with the inclusion criteria. A total of 71 articles were considered eligible. By evaluating the full text of all of the articles, a total of 29 articles were excluded because they focused on the following: regeneration of meniscus (n = 5), osteochondral lesions (n = 10), bone (n = 1), exosome treatment (n = 1), non-OA pathologies (n = 2), adipose or bone differentiation (n = 5) and hydrogels without MSCs (n = 5). As reported in Figure 1, a total of 42 studies were finally included in this review: 25 articles concerned in vitro studies, 2 in vitro and ex vivo results, 5 in vitro and in vivo results, 8 in vivo studies and 2 clinical trials.

Ex Vivo Studies
Two in vitro studies previously described [49,60] also showed ex vivo results that are reported in Table 2. They focused on the integration and adhesive capacity [60] of the MSCladen hydrogels to human OA cartilage and on their resistance to different strains delivered by a traumatic impact system [49]. Moreira-Teixeira et al. [60] analyzed the interaction and adhesion of a Dex-TA-based hydrogel to human OA cartilage with and without platelet lysate, evidencing by electron microscopy a close interaction with the cartilage specimen. He et al. [49] studied an engineered cartilage construct (GelMA hydrogel-BMSCs chondrogenically differentiated for 28 days) subjected to a traumatic impact system, evaluating the cell viability, cartilage gene modifications and the elastic modulus. N.I., not indicated.
In vivo subcutaneous dorsal implantation studies were performed in nude mice in four studies [47,57,59,72] and one in rats [74]. All MSCs used were of human origin and one of them overexpressed the long intergenic non-coding RNA regulator of reprogramming (Linc-ROR) [72]. All these studies evidenced an increase in chondrogenic markers from 2 to 8 weeks. Interestingly, Feng Q. et al. [47] and Feng L. et al. [72] reported a decrease in hypertrophic markers COL10 and MMP13. Only 4 out of 13 reported the dimension of the needles used for in vivo injection [69,70,75,76].

Clinical Trials
As reported in Table 4, two clinical trials were performed in OA patients with knee lesions [77,78]. Both studies treated the patients with multiple drill holes that were filled with non-crosslinked natural HA hydrogel combined with hUCB-MSCs [77,78]. In both studies the hydrogel characteristics in terms of rheological and physical properties were not reported. Different clinical (IKDC, WOMAC and VAS, KSS for pain, and arthroscopy) radiological (MRI) and histological parameters were analyzed to define the safety and/or the efficacy of the treatments. All studies performed a knee joint injection of embedded cells one, two or three times. In two out of nine rat OA model studies, human umbilical cord blood (hUCB)-MSCs or mixed human embryonic stem cells with MSCs were used. All studies evidenced from 4 to 9 weeks an increase in chondrogenic markers (COL2, or ACAN or SOX9 or proteoglycan or GAG) [47,57,59,61,66,[69][70][71][72][73][74][75][76] associated in some studies with a decrease in hypertrophic factors or inflammation or reduction in bone osteophytes or apoptosis [47,66,[69][70][71][72]75,76].
In vivo subcutaneous dorsal implantation studies were performed in nude mice in four studies [47,57,59,72] and one in rats [74]. All MSCs used were of human origin and one of them overexpressed the long intergenic non-coding RNA regulator of reprogramming (Linc-ROR) [72]. All these studies evidenced an increase in chondrogenic markers from 2 to 8 weeks. Interestingly, Feng Q. et al. [47] and Feng L. et al. [72] reported a decrease in hypertrophic markers COL10 and MMP13. Only 4 out of 13 reported the dimension of the needles used for in vivo injection [69,70,75,76].

Clinical Trials
As reported in Table 4, two clinical trials were performed in OA patients with knee lesions [77,78]. Both studies treated the patients with multiple drill holes that were filled with non-crosslinked natural HA hydrogel combined with hUCB-MSCs [77,78]. In both studies the hydrogel characteristics in terms of rheological and physical properties were not reported. Different clinical (IKDC, WOMAC and VAS, KSS for pain, and arthroscopy), radiological (MRI) and histological parameters were analyzed to define the safety and/or the efficacy of the treatments.

Discussion
The regeneration of cartilage in OA disease remains an unmet problem that still requires the development of new approaches [8]. Hydrogels represent a promising tool, since they can easily embed viable cells such as MSCs or chondrocytes and can be easily injected in the defect area [20]. It has been shown that the hydrogels create a microenvironment that influences the cells' characteristics, mainly due to their specific properties that have a positive or negative impact on the regulation of stem cell behavior [60]. Different papers have considered rheological and physical properties of the hydrogels, evidencing their direct role on cell chondrogenic differentiation. It has been shown that the material properties of hydrogels such as porosity, stiffness and viscoelasticity could modulate the cell characteristics; however, we have found that only a few papers have considered these important parameters. In fact, only in 7 [12,20,45,46,54,56,60] out of 42 papers the hydrogels' porosity was analyzed by scanning electron microscope (SEM) analysis. Moreira Teixeira et al. [60] demonstrated that neither the culture medium nor the platelet lysate affected the pore size of a dextran-based hydrogel. It has been shown that pores ranging from 50-300 µm are suitable for modulating the cell shape, but also for cell adhesion, migration and diffusion of the nutrients and for stimulating the cell differentiation. The stiffness is an important hydrogel parameter, showing its capacity to resist deformation. This characteristic is fundamental for knee cartilage regeneration that is constantly under loading. However, we found that 17 papers [12,20,42,45,46,[48][49][50][51]53,54,56,[59][60][61][62]65,66,69,70,74] considered in their studies the stiffness. Interestingly, only Favi et al. [12] considered porosity (focusing on pore dimension, interconnectivity and fiber orientation) of the bacterial cellulose-based hydrogel. Viscoelasticity is the capacity of a hydrogel to exhibit both viscous and elastic behavior, and it has been shown that the increase in hydrogel stress relaxation promotes chondrogenesis. Only Yu et al. [70] considered the porosity, stiffness and viscoelasticity for developing an ECM-mimicking hydrogel. They demonstrated that a ThHA hydrogel functionalized with collagen type 1 (ThHA-Col) displayed the rheological properties that protect the cell survival and growth, having a stiffness close to the native microenvironment [70]. Moreover, they also evidenced that ThHA-Col exhibited shear-thinning properties that protect the cells during the injection.
It is well known that MSCs represent a promising cell tool for chondrogenic differentiation, and as reported in Tables 1-4 human MSCs derived from bone marrow or adipose tissue or umbilical cord blood alone or combined with chondrocytes are the cell types mainly analyzed. However, MSCs from different animal sources (rabbit, canine, equine and rat) were also used to define their chondrogenic potentiality in hydrogels. MSCs have the potentiality to differentiate but at the same time are an important source of bioactive molecules that exert specific effects on chondrocyte proliferation and migration, as well as on immunomodulation. The embedding of MSCs in hydrogels represents an interesting approach for treating cartilage defects by injection. Hydrogel injection is a fundamental feature for the translation to the clinic, and only some in vivo studies [69,70,75,76] have reported the needle gauge (25 and 29 gauge), considering that it could be a parameter that could affect the cell viability by creating a shear stress that should be lower than 5 kPa, as previously reported [79]. Moreover, it is also important that injected hydrogel adhere well and remain stable in the defect area. Interestingly, Moreira Teixeira et al. [60] considered dextran-tyramine hydrogel adhesion to human OA cartilage, evidencing that the presence of tyramine residues contributed to the fixation of the hydrogel to collagen fibers or other matrix proteins of the cartilage. The authors also evidenced that the use of platelet lysate did not improve the cartilage adhesion but contributed to cell migration. The capacity of hydrogels to function as chemo-attractants for the cells is another important point for the clinical translation, since it contributes to assure a better integration of the hydrogel with the surrounding cartilage tissue. Finally, the mechanical properties of the engineered cartilage construct are another important point discussed by He et al. [49], using a loading system to mimic cartilage pressure.
Subcutaneous in vivo implantation studies [47,52,57,59,72,74] of MSC-laden hydrogels in nude mouse or rat models contributed only to defining the chondrogenic potentiality of the cells in a closer in vivo microenvironment but did not help for understanding OA disease effects.
In vivo studies based on the use of OA animal models [61,[69][70][71]73,75,76] are fundamental for defining not only the chondrogenic potentiality of embedded MSCs in hydrogels but also for defining their effects on counteracting the inflammation that is a known feature in OA. Interestingly, Bhattacharjee et al. [69], Yu et al. [70] and Kim et al. [75], using collagenase-induced or ACL transection and medial meniscectomy or ACL and medial collateral transection in rats for inducing OA, demonstrated a significant reduction in inflammation and an increase in chondrogenic markers, such as collagen type 2 or GAG in treated animals. It is well known that the progression of OA disease is also characterized by the presence of osteophytes, and Yu et al. [70], Kim et al. [75] and Xing et al. [71] evidenced that MSCs embedded in hydrogel and injected in the knee were also effective in reducing osteophyte formation and restoring bone density.
The application of HA-based hydrogels embedded with umbilical-cord-blood-derived MSCs (Caristem ® ) is an approach already used in two clinical studies [77,78]. The authors applied hydrogel-laden MSCs to patients with OA knee lesions that were pre-treated with multiple drill holes. In one study [77], the follow-up was evaluated at 24 weeks and 7 years to define the efficacy and safety profile of the treatment. They evaluated different clinical scores (VAS, IKDC and MRI) and histological samples did not evidence undesired effects, but the number of included patients was limited to only three for each group. Regarding the other study [78], patients who underwent HTO for medial unicompartimental OA were preliminary treated with multiple drill holes and then divided into two groups, one treated with hydrogel-laden MSCs (32 patients) and one with bone marrow concentrate (42 patients). At 1 year follow-up, clinical and radiological outcomes were considered, and no differences were evidenced in term of WOMAC and KSS pain between the two groups. In a second look arthroscopy the ICRS grade was better in MSC-laden hydrogel treated patients, confirming that this treatment was the most efficient. The main limitations are the number of treated patients [78] and the use of MSC-laden hydrogel combined with multiple drill holes and compared with bone marrow concentrate, and not with MSCs alone as in the other study [77]. Finally, only one study considered the patients' malalignment by performing HTO before the MSC-laden hydrogel treatment [78].

Conclusions
The need for new approaches to restore cartilage in OA disease is growing, and the use of MSC-laden hydrogel is a regeneration method that has been underlined in this systematic review, evidencing overall positive results that we summarized in Figure 5. The positive in vitro results using different hydrogels and cell types were confirmed in in vivo OA animal models that well represent the progression of OA disease. Promising clinical trials confirmed the positive effect of these treatment on patients with OA. However, some aspects remain to be elucidated, mainly those focused on the material characteristics of the hydrogels used and on the cell type. Moreover, other aspects such as 3D bioprinting and crosslinking should be investigated in depth to provide better biocompatibility, as well as personalized and customized cartilage regeneration strategies. Additional hydrogel injection modalities, strength of adhesion to OA cartilage and the chemoattractant role of the hydrogel need further studies that are fundamental to speeding the translation to the clinical use of this approach.  The positive in vitro results using different hydrogels and cell types were confirmed in in vivo OA animal models that well represent the progression of OA disease. Promising clinical trials confirmed the positive effect of these treatment on patients with OA. However, some aspects remain to be elucidated, mainly those focused on the material characteristics of the hydrogels used and on the cell type. Moreover, other aspects such as 3D bioprinting and crosslinking should be investigated in depth to provide better biocompatibility, as well as personalized and customized cartilage regeneration strategies. Additional hydrogel injection modalities, strength of adhesion to OA cartilage and the chemo-attractant role of the hydrogel need further studies that are fundamental to speeding the translation to the clinical use of this approach. Acknowledgments: Rayan is available for free at http://Rayyan.Qcri.org and is fully founded by the Qatar Foundation, a non-profit organization in the State of Qatar.