The human arthritic hip joint is a source of mesenchymal progenitor cells (MPCs) with extensive multipotent differentiation potential

Background: While multiple in vitro studies examined mesenchymal progenitor cells (MPCs) derived from bone marrow or hyaline cartilage, there is little to no data about the presence of MPCs in the joint capsule or the ligamentum capitis femoris (LCF) of the hip joint. Therefore, this in vitro study examined the presence and compared the differentiation potential of MPCs isolated from the bone marrow, arthritic hyaline cartilage, the LCF and full-thickness samples of the anterior joint capsule of the hip joint. Methods: MPCs were isolated and multiplied in adherent monolayer cell culture. Osteogenesis and adipogenesis was induced in monolayer cell cultures for 21 days using a differentiation medium containing specic growth factors, while chondrogenesis in the presence of TGF-ß1 was performed using pellet-culture for 27 days. Control cultures were maintained for comparison over the same duration of time. The differentiation process was analyzed using histological and immunohistochemical stainings as well as semiquantitative RT-PCR for measuring the mean expression levels of tissue-specic genes. Results: This in vitro research showed that the isolated cells from all four donor tissues grew plastic adherent and showed similar adipogenic and osteogenic differentiation capacity as proven by the histological detection of lipid droplets or deposits of extracellular calcium and collagen type I. After 27 days of chondrogenesis proteoglycans accumulated in the differentiated MPC-pellets from all donor tissues. Immunohistochemical staining revealed vast amounts of collagen type II in all differentiated MPC-pellets, except for those from the LCF. Interestingly all differentiated MPCs still showed a clear increase in mean expression of adipogenic, osteogenic and chondrogenic marker genes. In addition the examination of an exemplary donor sample revealed that cells from all four donor tissues were clearly positive for the surface markers CD44, CD73, CD90 and CD105 by ow cytometric analysis. Conclusions: This study proved the presence of MPCs in all four examined donor tissues of the hip joint. No signicant differences were observed during osteogenic or adipogenic differentiation depending on the source of MPCs used. Further


Background
Tissue loss as well as organ failure due to traumatic, congenital or acquired diseases pose a substantial health threat to our modern society [1][2][3]. Despite medical innovations enabling the transplantation of allogenic tissue and organs, according to the World Health Organization (WHO) there is a substantial lack in potential donors [1,4]. Even more important constant increases in life expectancy have led to a skyrocketing of degenerative diseases which generate both severe expenses for our health care systems and a growing amount of patients that could bene t from progress in the area of regenerative medicine [2,3,5]. These challenges have led to constant new research in the eld of Tissue Engineering (TE) [6,7]. TE aims to regenerate or maintain tissue function through the combined use of living cells as well as biomaterials and speci c growth factors or signalling pathways to guide or maintain cell fate [7].
Disorders of the locomotive system, such as soft tissue defects, diseases of the skelletal systeme and loss of articular cartilage as seen in osteoarthritis, majorly contribute to the rise of chronic, degenerative diseases that are subject to TE [8,9]. Especially the poor capacity for self-regeneration of articular cartilage has been subject to various in vitro and in vivo research [8,10]. While current techniques for cellbased cartilage repair such as autologous chondrocyte transplantation (ACT) have shown bene cial results in the short-term, still there is no reliable technique to restore long-lasting hyaline cartilage in vivo [11][12][13]. In addition, the harvest of mature, autologous-derived cells from healthy tissue causes iatrogen damage at the extraction site and often requires invasive surgery while only offering a small yield of cells [12].
In contrast to mature cells mesenchymal stromal cells (MSCs) offer an extensive multipotent differentiation potential, while vast amounts of MSCs can be isolated from easy accessible tissues minimizing donor site morbidity [14]. Further unlike stem cells they raise less ethical questions and carry less risk of cancer formation after transplantation [15,16]. The minimal requirements cells have to meet to be viewed as MSCs were de ned by the International Society for Cellular Therapy (ISCT) in 2006 [17,18]. These criteria include the expression (≥ 95 % positive) of the surface molecules CD73, CD90 and CD105 as well as the lack of expression (≤ 2 % positive) of haematopoetic cell markers. Furthermore, MSCs have to grow plastic-adherent and show the ability to differentiate into osteoblasts, adipocytes and chondroblasts in vitro which is why they are frequently used for engineering osteochondral or soft tissue [17,18].
When comparing sources of MSC, bone marrow-derived MSCs (BMSCs) are often viewed as the gold standard. Still there is a variety of easy accessible sources for MSCs such as adipose, synovial or perinatal tissue offering a comparable differentiation capacity while causing minimal donor site morbidity [5,19]. Therefore, the goal of this in vitro-study was to further contribute to this subject by proving the presence and comparing the multipotent differentiation potential of mesenchymal progenitor cells (MPCs) in the bone marrow, articular cartilage, full-thickness samples of the joint capsule and the ligamentum capitis femoris (LCF) of the arthritic human hip joint.
Recently, the term MSCs was frequently discussed questionning if it is still up-to-date [20]. Origininally the term MSCs was used as an abbreviation for mesenchymal stem cells which appears inappropriate since MSCs do not function as multipotent stem cells in vivo [20]. Newer studies also queried their description as MSCs that based on the belief that these multipotent cells were found in the stromal section of tissues rising from the mesodermal lineage. The reason for this is that more recent ndings suggested MSCs rise from pericytes located in the perivascular region of various adult tissues [20,21]. To avoid confusion multipotent precursor cells examined in this study will be reffered to as MPCs.

Methods
Isolation and cultivation of MPCs As described earlier, after informed consent and as approved by University of Wuerzburg's institutional review board MPCs were isolated from the femoral marrow reamings, arthritic femoral cartilage, as well as full-thickness samples of the hip joint's anterior capsule and LCF of ve female patients aged 40 to 55 (meag age 49 years), that underwent total hip replacement surgery [22,23]. To summarize the process collected tissues were preparated into small sized pieces that were then digested using collagenase (0.175 U/mL; Serva Electrophoresis, Heidelberg, Germany) in Dulbecco's Modi ed Eagle Medium (DMEM)/Ham's F12 (1:1; Life Technologies GmbH, Thermo Fisher Scienti c, Waltham, Massachusetts). In addition, samples of each tissue were secured to perform Hematoxylin-Eosin (H&E) staining. The cells were then spun, resuspended and disseminated in 175 cm 2 plastic cell culture asks (Greiner Bio-One GmbH, Frickenhausen, Germany). They were covered in standard culture medium consisting of DMEM/Ham's F12, supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin (PS) (all Life Technologies, Thermo Fischer Scienti c, Dreieich, Germany). When the cells reached con uency, they were trypsinated and stored in cryo tubes (Greiner Bio-One Gmbh) at -80°C for later experiments, dissolved in a medium consisting of 50% FBS, 40% DMEM/Ham's F12 and 10% dimethylsulfoxid (DMSO; AppliChem GmbH, Darmstadt, Germany).

FACS analysis of surface antigens
The cells were washed in phosphat buffered saline (PBS), spun and resuspended in standard cell culture medium. After this, cells were counted and seperated into two 2,5 mL eppendorf tubes (Greiner Bio-One GmbH) per tissue sample. Each of the two eppendorf tubes contained a minimum amount of 5 x 10 5 cells per tissue sample. After this 100 μL of an antibody-PBS mixture containing a PBS/1% FBS pre-mix, CD 73 PerCP antibody, CD44 eFLuor antibody, CD105 APC antibody and CD90 FITC antibody (all Thermo Fisher Scienti c GmbH, Dreieich, Germany) was added to one of these two samples for uorescence-activated cell sorting (FACS) analysis. The other sample was only provided with 100 μL of PBS/1% FBS pre-mix and was therefore treated as a negative control. Both samples were then vortexed, stored in the dark (4 °C for 30 min), washed in PBS/1% FBS, spun and resuspended in 2% paraformaldehyd in PBS before being stored in the dark again (4 °C for 15 min). Finally, the cells were washed in PBS, spun, resuspended in PBS/1% FBS and vortexed. The two samples were now ready for FACS analysis which was performed with the BDTM LSR II X device (BD biosciences, Franklin Lakes, NJ, USA) and analyzed using the FlowJo 10.5.3 Software (FlowJo LLC, Ashland, OR, USA).

Adipogenic and osteogenic differentiation
The MPCs stored at -80°C were defrosted and disseminated in standard cell culture medium in plastic cell culture bottles until reaching con uency. Brie y, the cells were then trypsinated, counted and seeded in six-well plates (Greiner Bio-One GmbH, Frickenhausen, Germany) at a density of 3 x 10 3 cells per cm 2 . Aditionally, MPCs were seeded in two-chamber slides (Greiner Bio-One GmbH, Frickenhausen, Germany) at a density of 5 x 10 3 cells per chamber for immunohistochemical stainings. Cell cultures were stored at 37°C, 5% CO 2 and medium changes were performed every 3 to 4 days (d). After the cells showed con uent growth osteogenesis and adipogenesis were induced as described in previous studies [24,25].
Other wells and chambers were covered with standard culture medium to maintain controls for comparison to the differentiated MPCs. The cell cultures were xed after 21 d.
To detect mineralization of the extracellular matric (ECM) Alizarin Red S (1%; Sigma-Aldrich) stainings were performed as reported previously [25], while immunohistochemical stainings were used to reveal extracellular collagen type I (Col I) formations. The adipogenic differentiation was detected through staining of lipid drops using Oil RedO (all Sigma-Aldrich) as described previousely [26]. In addition, the gene expression levels of the tissue speci c osteogenic marker genes alkaline phosphatase (ALP), Col I, collagen type X (Col X) or osteocalcin (OC), as well as the adipogenic marker genes lipoproteinlipase (LPL) and peroxisome proliferator-activated receptor γ (PPARγ) were examined with RT-PCR.

Chondrogenic differentiation
To further con rm the multilineage differentiation potential of the isolated cells their chondrogenic differentiation potential was tested using pellet culture as described previously [27]. Pellets were covered in chondrogenic differentiation medium as reported earlier [24]. Four pellets per tissue sample were formed for chondrogenic differentiation while four more were maintained in a control medium missing transforming growth factor (TGF)-ß1. The beakers were then centrifuged to promote aggregate formation and stored at 37°C, 5% CO 2 while medium changes were performed every 3 to 4 d. After 27 d of culture the pellets were placed in Tissue-Tek ® Cryomold ® Standard (Sakura Fintek, Torrance, CA, USA) before being shock frozen in liquid nitrogen. The frozen pellets were sectioned at 6 μm thickness and placed on SuperFrost ® cryosection slides (Thermo Fischer Scienti c GmbH). To x the cells the pellets were covered in 3% acetic acid (3 min; Carl Roth GmbH). Alcian blue (1%, Sigma-Aldrich) stainings were performed to verify the formation of proteoglycans as described earlier [28]. Immunohistochemical stainings were used to detect collagen type II (Col II) and Col X. The expression of chondrogenic markers aggrecan (AGG), Col II and sex-determining region Y-box 9 (Sox-9) was tested with RT-PCR.

Histology and Immunohistochemistry
Native tissue samples were xed in 4% paraformaldehyde, dehydrated and embedded in para n. They were then sectioned and stained with H&E (Sigma-Aldrich) as described in earlier studies [22]. Monolayers were xed in ice-cold methanol, while pellet cultures were placed in Tissue-Tek ® Cryomold ® Standard (Sakura Fintek, Torrance, CA, USA) before being shock frozen in liquid nitrogen. The frozen pellets were sectioned, placed on SuperFrost ® cryosection slides (Thermo Fischer Scienti c GmbH) and xed with 3% ice-cold acetone (Carl Roth GmbH). Alizarin Red S, Oil redO and alcian blue stainings were performed as outlined previously [25,26,28]. In addition, immunohistochemical stainings were performed on monolayers and pellets using the following antibodies: Col I -monoclonal anti Col Iα1 (5 μg/mL; Abcam pls, Cambridge, Great Britain); Col II -polyclonal Col IIα1 antibodies (5 μg/mL; Acris Antibodies GmbH, Herford, Germany); Col X -polyclonal Col X antibodies (5 μg/mL; Abcam pls). The immunostainings were visualised with the Avidin-Biotion complex method using the protocols, biotinylated antibodies, blocking serum and peroxidase from the VECTASTAIN ® Universal Elite ® ABC Kit (Vector Laboratories, Burlingame, CA, USA) and the VECTOR ® NovaRED™ peroxidase substrate kit (Vector Laboratories). The slides and wells were counterstained with hemalaun. For control stainings the primary antibodies were replaced with non-immune IgG antibodies (Sigma-Aldrich).
RNA isolation and semiquantitative RT-PCR analysis RNA was isolated from control and differentiated samples of adipogenic, osteogenic (21 d) and chondrogenic (27 d) MPCs using Trizol reagent (Invitrogen), as well as further puri cation steps, including DNase treatment, as described in the user's guide (NucleoSpin ® RNA II kit, Macherey-Nagel GmbH & Co. KG, Düren, Germany). To create cDNA 1 μg of puri ed RNA was combined with random hexamer primers (Thermo Fischer Scienti c) and Promega ® M-MLV reverse transcriptase (Promega GmbH, Mannheim, Germany). 1 μL of cDNA served as a pattern for ampli cation in a 30 μL reaction volume containing GoTaq ® DNA polymerase (Promega GmbH) with forward and reverse gene-speci c primers (5 pmol each). The primers and their speci c sequences, annealing temperatures and cycle numbers are listed in Table 1. Elongation factor 1α (EEF 1α) was used as the housekeeping gene. The nal products of RT-PCR were seperated through electrophoresis on 2% agarose (Biozym Scienti c GmbH, Hessisch Oldendorf, Germany) gels containing 5 μL per 100 mL GelRed ® (Biotium, Fremont, CA, USA). The band densities for each primer pair were then examined to measure the mean ratio and standard deviation for all of the tested genes in comparison to the expression of the EEF 1α housekeeping gene.

Statistical analysis
Semiquantitative RT-PCR experiments were performed on four different tissues all taken from ve different donors (n = 5) and expressed as mean values standard deviation. Statistical signi cance was de ned using the Student's t test with p < 0,05 being considered as signi cant.

Results
H&E stainings of native tissue samples H&E staining was succesfully performed on all collected native tissue samples from bone marrow, hyaline cartilage, LCF and the joint capsule, coloring acidophile structures red and basophile structures blue ( Figure 1). Stained bone marrow tissue samples ( Figure 1, bone marrow), derived from the femoral reaming, showed parts of bone trabeculae that were composed of collagen bres and osteocytes. In addition, osteoblasts could be assumed at the periphery of bone trabeculae. Samples from degenerated, arthritic hyaline cartilage (Figure 1, cartilage) revealed few chondrones that consisted of condrocytes in lacunaes. These chondrones were surrounded by an irregularly, basophilic stained ECM. The LCF ( Figure   1, LCF) mainly consisted of connective tissue formed by collagen bres, brous and adipose tissue. This connective tissue contained blood vessels of varying size and was coated by synovium. Similar to the LCF the joint capsule ( Figure 1, capsule) consisted of a brous and synovial stratum. The brous stratum was formed by connective, ligamentous tissue built by collagen bres and adipose tissue and lled with numerous small blood vessels. The thin synovial stratum coated the brous stratum and formed folds, the so-called plicae synoviales. In both the LCF (Figure 1, LCF) and the joint capsule ( Figure 1, capsule) broblasts appeared as spindle shaped cells in the connective tissue. Mononuclear cell in ltration was observed around small blood vessels of the brous stratum in the tissue samples derived from the joint capsule ( Figure 1, capsule).

Surface markers on isolated MPCs
FACS analysis was performed ( Figure 2) using cells from all four tissue samples of a single donor to exemplary de ne the percentage of isolated cells that expressed the surface antigens CD44, CD73, CD90 and CD105 ( Table 2). All isolated cells from bone marrow, cartilage, LCF and the joint capsule were strongly positive (≥ 95 %) for CD44, CD90 and CD105 (Table 2, CD44 + , CD90 + , CD105 + ). The majority of isolated cells was also strongly positive (< 95 %) for the surface marker CD73 ( Table 2, CD73 + ). While > 70% of cells isolated from bone marrow, cartilage and the joint capsule expressed CD73, only about half of the cells from the LCF were positve for this surface marker ( Table 2, CD73 + ). Matching results were shown within the coexpression of CD-antigens ( Figure 2; Table 2).

Histological analysis of adipogenic differentiation
In comparison to control cultures ( Figure 3 Osteogenic marker gene expression after 21 d RT-PCR revealed an upregulation of the mean mRNA expression levels of osteogenic marker genes collagen type X (Col X), alkaline phosphatase (ALP) and osteocalcin (OC) in all of the osteogenic differentiated cultures compared to their controls (Figure 4, b). The expression of the osteogenic marker gene OC in MPCs derived from bone marrow was signi cantly higher (p < 0,05) than from MPCs isolated from cartilage (Figure 4, b, OC). Interestingly, the mean expression levels of Col I mRNA in comparison to negative controls only showed a signi cant increase in MPCs derived from bone marrow (Figure 4, b, Col I). Other than that, the expression of osteogenic marker genes did not differ signi cantly within MPCs derived from different native tissues. These ndings are contrary to the clearly positive immunohistochemical stainings for Col I in all osteogenic differentiated MPCs ( Figure 5, c, differentiation d 21).
Histological and immunohistochemical analysis of chondrogenesis After 27 d of pellet-culture in presence of TGF-ß1, positive alcian blue stainings proved the presence of proteoglycans in all chondrogenic differentiated cell-pellets derived from bone marrow, cartilage, the joint capsule and LCF (Figure 6, a, differentiation d 27). In addition, chondrogenic differentiated pellets containing cells from bone marrow and cartilage showed a chondrogenic phenotype especially in outer regions of the pellets (Figure 6, a, bone marrow, cartilage). Despite clearly postive alcian blue stainings, hyaline cartilage-speci c attributes, such as chondrones in areas with a low density of cells, were less visible in pellets from the joint capsule ( Figure 6, a, capsule) and almost non-existent in pellets from LCF ( Figure 6, a, LCF). Corresponding to alcian blue stainings immunohistochemical stainings revealed vast amounts of collagen type II (Col II) in outer regions of the differentiated pellets from bone marrow, cartilage and the joint capsule after 27 d (Figure 6, b, differentiation d 27). In comparison to the strong immunohisochemical stainings of Col II in differentiated pellets from bone marrow and cartilage ( Figure  6, b, bone marrow, cartilage) the staining appeared less intense in differentiated pellets from the joint capsule ( Figure 6, b, capsule). In contrast, chondrogenic differentiated pellets from LCF showed no visible stainings of Col II after 27 d (Figure 6, b, LCF). All controls remained negative for alcian blue (Figure 6, a, control d 27) and Col II (Figure 6 Immunohistochemical stainings for Col X were performed to examine possible chondrocyte hypertrophy during chondrogenic differentiation (Figure 6, c). While controls remained mostly negative (Figure 6, c, control d 27), differentiated pellets showed slightly stronger stainings especially in inner, cell-rich regions of the sections (Figure 6, c, differentiation d 27) that showed a less chondrogenic phenotype in earlier alcian blue and Col II stainings. The strongest matrix staining for Col X after 27 d was seen in the differentiated pellets containing cells from cartilage (Figure 6, c, cartilage).

Expression of chondrogenic marker genes after 27 d of pellet-culture
RT-PCR showed an upregulation of the mean mRNA levels of the examined chondrogenic marker genes AGG, Col II and Sox-9 in all chondrogenic differentiated pellets after 27 d (Figure 4, c). The mean mRNA expression levels of AGG and Col II were highest in chondrogenic differentiated pellets from the joint capsule although observed differences were not from statistical signi cance (Figure 4, c, AGG, Col II). Interestingly, the mean expression of Col II mRNA was upregulated in differentiated pellets from LCF (Figure 4, c, Col II) although the immunohistochemical stainings remained negative for Col II in corresponding pellets from LCF (Figure 6, b, LCF). Signi cant differences in the mean mRNA expression levels between the pellets from varying native tissues were not observed.

Discussion
Due to the tripotent differentiation capacity of MPCs they possess huge potential for the regenerative, cell-based treatment of musculoskeletal or soft tissue defects as well as traumatic or degenerative cartilage disorders [13,29]. Since MPCs were rst isolated from bone marrow by Friedenstein et al. they have been found to residue in a variety of adult or perinatal-derived tissues [19,30,31]. Interestingly, MPCs isolated from easy accesible, abundant sources, such as adipose or synovial tissue, have shown to possess a potentially superior differentiation capacity compared to bone marrow-derived MPCs (BMPCs) and thus sparked great interest among clinicians and researchers [19,32]. These MPC-subpopulations, which can be harvested causing minimal donor site morbidity while offering a high yield of cells from native tissues, could lead to substantial progress in TE when combined with smart biomaterials and suitable growth factors [3,5]. Therefore, the goal of this study was to further contribute to the detection and examination of innovative and valuable sources of multipotent precursor cells. In order to do so the LCF and full-thickness samples of the joint capsule in arthritic human hip joints were examined regarding the presence of MPCs. The characteristics of isolated cells were then compared to better-known sources of MPCs simultaneously derived from bone marrow and arthritic hyaline cartilage of the same hip joints as well as similar MPC-subpopulations analyzed in earlier studies. In addition, the characterization of cells isolated from all four donor tissues was set in relation to the ISCT's minimal criteria for the de nition of MSCs.
First and formost, our study showed that multipotent, plastic-adherent growing cells could be isolated from bone marrow, arthritic hyaline cartilage, the LCF and full-thickness samples of the joint capusles of the examined arthritic hip joints. Further cells cells from an exemplary tested donor sample were clearly positive for the surface antigens CD44, CD73, CD90 and CD105 independent of utilised native tissue. The presence of BMPCs [26,33] and MPCs found in the super cial zone of arthritic or healthy hyaline cartilage [34,35] has been frequently con rmed through earlier in vitro-studies, matching our ndings. While MPCs have previously also been obtained from the joint capsule of the human hip, interestingly, the tissue utilized for the isolation of cells was exclusively derived from the synovial stratum [36]. In addition, there is no data exclusively describing the characterization of MPCs isolated from the LCF, although Steinert et al. and Cheng et al. described the isolation of MPCs from the anterior cruciate ligament (ACL) of the knee in earlier studies [24,37]. Even though the two ligaments are not identical the ACL and the LCF share various histological and anatomical similarities and both form intraarticular ligaments in synovial joints coated by a synovial membrane [38]. Regardless, both of these ndings support the complementary and novel results in our study in which MPCs could be isolated from the LCF and fullthickness tissue samples of the arthritic hip's joint capsule, including the ligamentous tissue of the brous stratum. Moreover, all four tissues examined in our study rise from the LPM and are well vascularized, making them home to pericytes. Considering the current knowledge about the embryonic developement of MPCs this supports our studies results [20]. Hyaline cartilage represents an exception since it is avascular. This led to previous debates about if multipotent precursors, which have been shown to be located close to the surface in arthritic and healthy hyaline cartilage, may be distinct from MPCs or whether precursors other than PSCs contribute to the formation of MPCs [20,39].
Secondly, we found that ≥ 95% of the examined cells derived from bone marrow, hyaline cartilage, the joint capsule and the LCF of one representative donor showed a clearly positive expression of CD44, CD90 and CD105 as shown through FACS analysis. Although a large number of isolated cells were also positive for CD73, the percentage did not exceed 95% independent of the type of donor tissue. This means that all isolated cells of the exemplary chosen donor carried a compilation of surface antigens which is viewed as highly characteristic for multipotent precursor cells described as MSCs [18]. The presence of these surface markers on BMPCs has previously been proved in various studies [33,[40][41][42]. Alsalameh et al. and Pretzel et al. also showed that MPCs derived from healthy and arthtitic cartilage, coexpress the MSC-like surface markers CD105 and CD166 [34,35]. Multiple researchers also isolated MPCs expressing the surface antigens CD73, CD90 and CD105 from the ACL [24,37] or the synovial membrane and synovial uid of the hip joint's capsule [19,43,44]. This strongly supports our ndings in which cells expressing the surface markers CD44, CD73, CD90 and CD105 can also be found in fullthickness samples of the joint capsule and LCF of the human hip joint. Nonetheless, independant of the utilised native tissue the representative cells examined in our study did not fully meet the ISCT's minimal criteria for the de nition of MSCs in vitro since the percentage of cells positive for CD73 did not exceed 95% and cells were not tested negative (≤ 2%) for haematopoetic surface markers [17]. Still cells isolated from all four tissues of the exemplary chosen donor clearly expressed MSC-like surface antigens and would most certainly ful ll the ISCT's formal criteria if further steps of puri cation had been performed.
Thirdly, mulitpotent differentiation potential varied but was proven for all cells isolated from bone marrow, cartilage, the joint capsule and LCF of the human hip joint as requested by the ISCT's minimal criteria for MSCs [17]. Adipogenic and osteogenic differentiation capacity, as proven by histological, immunohistochemical and molecularbiological analysis, showed hardly any signi cant differences between the different donor tissues of cells. Lastly, chondrogenic differentiation was induced using a TGF-ß1 based differentiation medium and pellet culture system as described earlier by Johnstone et al. [45]. Three dimensional culture systems add mechanical stimuli, similar to the enchondral ossi cation in the human growth plate, that are necessary for chondrogenic differentiation in vitro [46][47][48]. Although MPCs derived from all four tissues showed successfull chondrogenesis in histological and molecularbiological analysis, immunohistochemical stainings of Col II (Figure 6, b)  Still the chondrogenic phenotype as well as the intensity of immunohistochemical stainings of Col II in differentiated pellet-sections containing BMPCs were superior to those containing cells from the joint capsule.
As mentioned above there is no scienti c data about the isolation and multipotent differentiation capacity of MPCs from the LCF. Similar to the LCF in the hip, the ACL forms an intraarticular, vascularized ligament in the knee. Although of course this does not allow a fully equal treatment of MPCs derived from both tissues we compared our ndings for MPCs derived from the LCF to those from studies that isolated MPCs from the ACL [24,37]. Matching the results of the trilineage differentiation of MPCs derived from the LCF in our study Steinert et al. found that the chondrogenic differentiation capacity of BMPCs was higher compared to cells isolated from the ACL, while their osteogenic and adipogenic differentiation potential did not differ signi cantly [24]. In contrast to our ndings Steinert et al. found positive immunohistochemical stainings of Col II in chondrogenic differentiated pellet-sections containing cells derived from the ACL in the presence of TGF-ß1 [24]. Although Cheng at al. found similar results regarding the adipogenic differentiation of both sources of MPCs, they reported a higher osteogenic and similar chondrogenic differentiation capacity of BMPCs compared to the cells derived from the ACL [37].
Finally, until today there are many uncertainties surrounding the group of multipotent precursor cells commonly described as MPCs or MSCs. These include the de ning criteria for MSCs in vitro released by the ISCT as well as the functionalities, characterization and localization of potential MSCs in vivo [20]. The functional capacities of cells described as MSCs in vitro including multipotent differentiation and colony formation suggested that MSCs may represent a group of stem cells [20]. However unlike stem cells, MSCs showed limited self-renewal after multiple passages of cell culture in vitro [49]. In addition, the implantation of MSCs in vivo lead to spontaneous formation of bone tissue but not to formation of cartilage or fat, showing that the multipotent differentiation capacity of so called MSCs in vitro does not represent their functionality in vivo [20,49]. Due to these insights, this group of cells was no longer reffered to as mesenchymal stem cells but rather as MSCs, since they were believed to reside in the stroma of various tissues rising from the mesodermal lineage [20]. This was not wrong to assume since MSCs are closely linked to the somatic layer of the lateral plate mesoderm (LPM). Cells from the somatic layer of the LPM later form bone, cartilage, tendons and ligaments in all limbs, subcutaneous abdominal and gluteofemoral fat, as well as multipotent progenitor cells for the homeostatic preservance of these tissues [50]. While MSC-like cells have since been isolated from close to all of these tissues rising from the somatic layer of the LPM in vitro, their exact location in vivo remains unclear. Interestingly, perivascular located pericytes and adventitial progenitors cells, found in the tunica adventitia of large blood vessels, remain some of the only cells examined in situ to meet all MSC-de ning criteria [51]. Hence these perivascular stem cells (PSCs), even when isolated from tissues that do not rise directly from the somatic layer of the LPM, showed adipogenic, osteogenic and chondrogenic differentiation capacity in in vitro cultures [21]. This led to the belief that contrary to what their name would suggest MSCs may not reside in the stromal tissue section but rather rise from PSCs which are located in the perivascular region of almost all tissues of the human body [21]. This challenged our earlier understandings of MSCs as tissue resident, stem cell-like cells isolated from tissues of mesodermal origin and opens up new possibilities for the use and isolation of cells commonly de end as MSCs.
These contrasts between the characteristics of multipotent progenitor cells presented in vitro and in vivo not only led to disagreement regarding their naming but also questionned the bene ts of the ISCT's minimal criteria for the de nition of MSCs that solely builds on in vitro ndings [18]. By introducing these minimal criteria the ISCT tried to standardise the scienti c use of multipotent precursor cells and to thereby facilitate the comparison of similar cells described as MSCs [18]. However research has shown that cells meeting the criteria for MSCs derived from different origins not only present distinct compilations of surface markers but that the expression of surface antigens also changes throughout their cultivation in vitro [39,49]. In addition, MSCs derived from varying sources require a speci c composition of growth factors for trilineage differentiation and hence vary regarding their differentiation capacity [19,20,49]. This raises further doubt whether the ISCT's criteria or any standardizing classi cation at all is adequate to de ne cells reffered to as MSCs and rather indicates that MSC-like cells isolated from different tissues represent subpopulations with unique tissue-speci c characteristics that offer both advantages and disadvantages for their clinical use [39,52].
Finally all cells isolated from bone marrow, articular cartilage, the LCF and the joint capsule of arthritic hip joints grew plastic-adherent and were capable of differentiation into adipocytes, osteocytes and chondrocytes. In addition, all cells from the examined representative donor samples expressed the surface antigens CD44, CD73, CD90 and CD105 which are viewed as characteristic for MSCs.
Nonetheless, since cells from the exemplary tested donor sample did not fully meet the ISCT's minimal criteria regarding the expression of the surface antigen CD73, along with the mentioned confusion surrounding the correct naming of multipotent progenitor cells, we reffered to the isolated cells as MPCs.
The ndings in this study represent information gathered in vitro and may not fully correspond to the in vivo functionalities of examined cell subpopulations. Therefore, additional in vitro-studies are necessary to fully determine suitable growth factors and tting biomaterials to optimize the chondrogenic differentiation of MPCs derived from full-thickness samples of the joint capsule and the LCF. When considering the most suitable source of MPCs for clinical use one has to bear in mind the donor site morbidity caused by the isolation of these cells as well as the varying yield obtained from different tissues [19,39]. Ultimately the target for all inidcations of tissue engineering is to identify the tting source of MPCs, offering the best desired differentiation capacity, while causing the least donor site morbidity during withdrawal. This study adds further evidence and information to possible sources of MPCs and their characteristics in vitro. Availability of data and materials

Conclusions
The datasets used and analysed during the current study are available from the corresponding author on reasonable request.

Ethics approval and consent to participate
The study design and experiments were approved by University of Wuerzburg's institutional review board.
In addition, participating patients agreed to the use of collected and examined surgical waste after undergoing total hip replacement surgery.