Bioprinting of scaled-up meniscal grafts by spatially patterning phenotypically distinct meniscus progenitor cells within melt electrowritten scaffolds

Meniscus injuries are a common problem in orthopedic medicine and are associated with a significantly increased risk of developing osteoarthritis. While developments have been made in the field of meniscus regeneration, the engineering of cell-laden constructs that mimic the complex structure, composition and biomechanics of the native tissue remains a significant challenge. This can be linked to the use of cells that are not phenotypically representative of the different zones of the meniscus, and an inability to direct the spatial organization of engineered meniscal tissues. In this study we investigated the potential of zone-specific meniscus progenitor cells (MPCs) to generate functional meniscal tissue following their deposition into melt electrowritten (MEW) scaffolds. We first confirmed that fibronectin selected MPCs from the inner and outer regions of the meniscus maintain their differentiation capacity with prolonged monolayer expansion, opening their use within advanced biofabrication strategies. By depositing MPCs within MEW scaffolds with elongated pore shapes, which functioned as physical boundaries to direct cell growth and extracellular matrix production, we were able to bioprint anisotropic fibrocartilaginous tissues with preferentially aligned collagen networks. Furthermore, by using MPCs isolated from the inner (iMPCs) and outer (oMPCs) zone of the meniscus, we were able to bioprint phenotypically distinct constructs mimicking aspects of the native tissue. An iterative MEW process was then implemented to print scaffolds with a similar wedged-shaped profile to that of the native meniscus, into which we deposited iMPCs and oMPCs in a spatially controlled manner. This process allowed us to engineer sulfated glycosaminoglycan and collagen rich constructs mimicking the geometry of the meniscus, with MPCs generating a more fibrocartilage-like tissue compared to the mesenchymal stromal/stem cells. Taken together, these results demonstrate how the convergence of emerging biofabrication platforms with tissue-specific progenitor cells can enable the engineering of complex tissues such as the meniscus.


Introduction
The knee joint menisci are wedged-shaped fibrocartilaginous tissues located between the femoral condyle and the tibia plateau [1,2].They play a key role in maintaining stability and in the distribution of loads across the joint, thus protecting the underlying articular cartilage during normal daily activities [3].The meniscus is often classified into two different regions, an inner avascular region, and an outer vascularized region.While the outer region is predominantly composed of circumferentially aligned type I collagen fibers and populated with fusiform-shaped fibroblast-like cells, the inner avascular region is compositionally more similar to articular cartilage and contains a higher proportion of type II collagen [4,5].This unique structure and composition provide the tissue with mechanical properties ideally suited to its function.In spite of this, injuries to the meniscus are common, with an annual incidence rate of 66 per 100 000 people in the United States [6,7].Injuries to the meniscus alter joint homeostasis and significantly increase the risk of developing osteoarthritis (OA) later in life [8,9].For that reason there is increased interest in surgical meniscus repair, with factors such as the type, geometry and location of the injury impacting clinical outcomes [10].Furthermore, in many cases surgical repair is not feasible, motivating the need for innovative strategies to direct meniscal regeneration.
3D bioprinting is emerging as a particularly promising strategy in the field of tissue engineering and regenerative medicine as it allows for the controlled deposition and assembly of biomaterials, bioactive factors and cells, resulting in the development of constructs that mimic key aspects of the structure and composition of native tissues [11][12][13][14][15].Such approaches have been applied to the engineering of meniscal constructs, with the 3D printing of synthetic polymers such as polycaprolactone (PCL) commonly used to fabricate implants with bulk mechanical properties comparable to the native tissue.Furthermore, spatial functionalization of such scaffolds with bioactive factors has been shown to support the formation of meniscus-like tissues both in vitro and in vivo [16][17][18].Bioinks supportive of the different regions of the meniscus have also been developed; for example, meniscus extracellular matrix (ECM) based bioinks have been shown to support the development of meniscus-specific phenotypes [19][20][21][22][23].The limited self-healing capacity of the meniscus also suggests that the incorporation of a cellular component will be required for a successful regenerative implant [21,24,25], especially when targeting larger regions of damaged or diseased tissue.Moreover, the limited engraftment and neotissue formation by invading cells when utilizing clinically available acellular implants further motivates the investigation of cell-based approaches [26][27][28].However, it is unlikely that merely bioprinting cellladen constructs, which mimic the shape of the meniscus, will suffice.Post-printing in vitro maturation, aimed at producing constructs that recapitulate the spatially defined structure, composition, and biomechanics of the native tissue, may also contribute to the development of functional and truly regenerative implants [29].
The identification of a suitable cell population for meniscus regeneration will be integral to satisfactory regenerative outcomes, and it is still a subject of ongoing debate.Primary meniscus cells have been shown to lose their native phenotype upon serial subculture, motivating increased interest in mesenchymal stromal/stem cells (MSCs) as an alternative cell source for meniscus tissue engineering [28,[30][31][32][33].Although MSC-based therapies for meniscus repair have shown some promise, their tendency to undergo hypertrophic differentiation and trigger mineralization via endochondral ossification remains a major concern [34,35].Recently, the identification of a migratory and multipotent cell population within the inner and outer regions of the meniscus has increased interest in their isolation and use for meniscus tissue engineering [36][37][38][39].Using a differential adhesion to fibronectin protocol that has been used to isolate progenitor cells from the articular cartilage [40][41][42], it is possible to isolate cellular subpopulations from within the meniscus which have a high proliferative capacity and are able to maintain their differentiation potential after serial subculture, unlike non-fibronectin selected meniscus cells [43].Phenotypically distinct progenitor cells have been found in both the inner and outer regions of the meniscus, which retain distinct capabilities to differentiate into inner and outer meniscus fibroblasts with limited evidence of hypertrophy [44].Therefore, integrating such meniscus progenitor cells (MPCs) into advanced biofabrication strategies could potentially enable the engineering of functional meniscus constructs that mimic the spatial composition of the native tissue.
In addition to mimicking the spatial composition of the meniscus, functional meniscus tissue engineering will also require recapitulating the spatial structure of the native tissue, which is dominated by the circumferentially aligned collagen fibers that are integral to its biomechanical function [45].Scaffoldfree or self-assembly tissue engineering strategies represent one promising route towards the generation of anisotropic biomimetic meniscus constructs, particularly when combined with appropriate bioactive stimuli [46][47][48][49][50][51].Another approach to control collagen microarchitecture in engineered tissues has been via the use of aligned polymeric scaffolds, which serve as a template to direct the orientation of cells and subsequent ECM deposition [52][53][54][55][56].We have previously used melt electrowritten (MEW) scaffolds within a multiple-tool biofabrication platform to bioprint anisotropic meniscus-like tissues with user-defined collagen architectures [57].By altering the aspect ratio of the pores (or 'microchambers') in these MEW scaffolds, it was possible to preferentially direct collagen deposition parallel to the long axis of these guiding structures.A 30-fold increase in compressive mechanical properties was observed over four weeks of in vitro maturation, resulting in the development of engineered tissues that Experiments used to assess the fibrochondrogenic potential of porcine meniscus-derived progenitor cells (MPCs) in tissue engineering.(A) Initially the fibrochondrogenic potential of both MPCs and non-selected populations derived from both the inner and outer regions was assessed using pellet culture.(B) An investigation of MPCs in the biofabrication of engineered fibrocartilaginous tissues was then undertaken.To achieve this, 3D bioprinting techniques (inkjet and extrusion) were employed to deposit cells into supportive and guiding scaffold structures manufactured using MEW.Inkjet printing was used to deposit cells into small-scale scaffolds, while extrusion based bioprinting was used to deposit cells into the larger, wedged-shaped constructs.
mimicked the compressive properties of the native tissue.Scaling-up such biofabrication platforms may potentially enable the engineering of large, anatomically defined meniscal grafts, especially if combined with cell types capable of generating the spatially distinct regions of meniscus.Furthermore, an inherent advantage of MEW as a fabrication technique is the lower polymer content and the high porosity of the scaffolds due to the small filament diameter that can be obtained, which may improve its in vivo regenerative potential [58,59].
The aim of the study is to combine the biological potential of MPCs, isolated from the inner and outer regions of the meniscus, with a multipletool biofabrication strategy to engineer large (5 mm tall) wedge-shaped meniscus-like tissues with spatially defined structure and composition.Building on previous work reported in the literature [43,44], we first sought to use a differential fibronectin adhesion protocol to isolate MPCs from the inner and outer zone of the meniscus, and then compare their capacity to generate phenotypically distinct meniscus tissue compared to meniscus fibrochondrocytes isolated from the inner and outer zone of the meniscus (figure 1).We then attempted to bioprint fibrocartilage tissue with preferentially aligned collagen fibers by depositing MPCs obtained from either the inner (iMPCs) or outer (oMPCs) region of the meniscus into MEW scaffolds with anisotropic pore networks.Finally, we spatially deposited iMPCs and oMPCs into scaled-up MEW scaffolds with a view to engineering wedge-shaped meniscal tissues with regionally defined composition and structure.

MEW
MEW scaffolds with a fiber diameter of 15 µm were fabricated on a custom MEW printer built in house [60].Briefly, PCL (50 000 MW, CAPA 6500D, Perstorp) was melted at a temperature of 80 • C around the syringe, increasing to 85 • C around the nozzle (21 G).To extrude the polymer, a pressure of 0.6 Bar with an initial voltage of 6 kV was applied at a distance of 6 mm from the grounded aluminum collector plate.With every layer, the voltage was increased by 0.0125 kV.Fibers were subsequently deposited according to a predesigned architecture, obtaining a scaffold with a pore size/shape of 0.4 × 1.6 mm.For the wedge-shaped scaffold, the starting height was set at 8 mm with an initial voltage of 7 kV.To obtain the wedge-shape profile, at every 1 mm of build, the width in one of the axes was reduced by 2-3 mm (for a total of five layers), to achieve a final scaffold height of 5 mm (figure 6(B)).

Isolation and expansion of MSCs
Porcine bone marrow derived MSCs were obtained from the femur of a pig supplied by a local abattoir (male, approximately 4 months old).The bone marrow was collected under sterile conditions from the femoral shaft, and diluted with expansion medium (XPAN), consisting of hgDMEM supplemented with 10% v/v FBS, 100 U ml −1 penicillin, 100 µg ml −1 streptomycin, and 2.5 µg ml −1 amphotericin B (all Gibco, Biosciences).After obtaining a homogenous suspension using a 16 G needle and filtering through a 40 µm nylon mesh, cells were counted using trypan blue containing 3% acetic acid before plating 23 × 10 6 cells into T175 flasks for expansion.All expansion was performed at 5% pO2, and using XPAN medium supplemented with 5 ng ml −1 FGF2 (Peprotech) seeding at a density of 875.000 cells per T175 flask.Chondrogenic, osteogenic and adipogenic differentiation assays were used to assess the tripotentiality of isolated MSCs.

Meniscus cell isolation and fibronectin selection of MPCs
Medial and lateral knee menisci were obtained from a four month old pig.Under sterile conditions, the menisci were rinsed twice with PBS containing 100 U ml −1 penicillin and 100 µg ml −1 streptomycin (both Gibco, Biosciences).After 30 min of pronase digestion (500 U ml −1 ) at 37 • C, inner and outer regions were plated separately and minced into pieces ∼1-2 mm 3 in size.The tissue samples were then placed in a collagenase (Collagenase type I, Gibco) solution (900 U ml −1 ) for 3 h at 37 • C under constant agitation.When samples had fully digested, the suspension was filtered through a 0.4 µm porous membrane, centrifuged and resuspended in hgDMEM.These cells were then seeded at a density of 850 000 cells per T175 flask (non-selected population), or underwent a process of selection using fibronectin.
For the fibronectin coating, 100 mm petri dishes were coated with a 10 µg ml −1 fibronectin (Promocell) solution in PBS containing 1 mM MgCl 2 and 1 mM CaCl 2 overnight at 4 • C, as previously described [40].Dishes were then blocked using a 1% bovine serum albumin (BSA) for 1 h at RT. Isolated cells from the inner and outer region of the meniscus were resuspended at 4000 cells ml −1 and plated separately onto coated dishes for 20 min at 37 • C in hgDMEM, or into non-coated plates for expansion in XPAN medium with 5 ng ml −1 FGF2.After 20 min, the media in the coated plates was removed together with non-adherent cells, and replaced with XPAN medium with 5 ng ml −1 FGF2.

Osteogenic differentiation and alizarin red staining
For osteogenic differentiation, 100 000 cells were plated to a six-well plate and maintained for 2-3 d in XPAN medium.After achieving a confluency of 80%, 100 nM of Dexamethasone, 10 mM glycerol phosphate and 0.05 mM ascorbic acid (all from Sigma) were added to regular XPAN.After 21 d, the cell monolayer was washed twice with PBS before the addition of iced ethanol at RT for 10 min.After removing the ethanol, each well was washed first with PBS and then with distilled water. 2 mL of 1% of Alizarin Red (Sigma) was added and incubated for 1 min before rinsing with distilled water and imaging.

Biochemical analysis
The biochemical content of the constructs was analyzed after finishing the in vitro culture.Constructs were washed in deionized water, frozen, and lyophilized to obtain the dry weight.Each construct was digested with papain (3.88 U ml −1 ) in ultrapure water containing 0.1 M sodium acetate, 5 mM L-cysteine-hydrochloride hydrate, 5 mM methylenediaminetetraacetic acid (EDTA) (all from Sigma-Aldrich).The pH was adjusted to 6.5 using 38% HCl.Samples were incubated in papain solution at 60 • C, rotating at 10 rpm for 18 h.Immediately after digestion, the DNA content was quantified using the Hoechst 33 258 dye assay with calf thymus DNA as standard (Merck) reading at 360 nm excitation and 460 nm emission.The amount of sulphated glycosaminoglycan (sGAG) was quantified using the 1,9-dimethylmethylene blue (DMMB) dye-binding assay, with a chondroitin sulfate solution as standard (Blyscan) [61].To exclude any background absorbance from the alginate, the pH of the DMMB was adjusted to 1.5.The 530/590 absorbance ratio was used to generate the standard curve and determine the sGAG concentration in the digested samples.Total collagen content was determined by measuring the hydroxyproline content using the dimethylaminobenzaldehyde and chloramine T assay, and a hydroxyproline/collagen ratio of 1:7.69 [62].Briefly, samples were mixed with 100 µl of 38% HCl, and incubated at 100 • C for 18 h.After cooling, samples were centrifuge at 5000 g for 5 min, and let to dry at 50 • C for 48 h.Dried samples were then dissolved in 200 µl ultra-pure water.2.82% (w/v) chloramine T solution was added and incubated for 20 min at room temperature in the dark, before the addition of a 50% (w/v) 4-(Dimethylamino) benzaldehyde solution (both Sigma) at 60 • C for 20 min.Hydroxyproline levels were estimated from the standard curve at a wavelength of 570 nm.

Histological analysis
Constructs were fixed in 4% (w/v) paraformaldehyde, dehydrated in a graded series of ethanol and xylene baths, embedded in paraffin wax, sectioned at 8 µm using a microtome (Leica Microsystems), and affixed to microscope slides.The sections were stained with hematoxylin and eosin (H&E), alcian blue, picrosirius red and alizarin red.For immunohistochemistry antigen retrieval was carried out by an initial treatment with pronase (3.5 U ml −1 ; Merck) at 37 • C for 25 min, followed by hyaluronidase (4000 units ml −1 ; Sigma-Aldrich) at 37 • C for 25 min for collagen type I and type II.For collagen type X, the antigen retrieval method consisted of an initial treatment with pronase (35 U ml −1 ; Merck) at 37 • C for 5 min, followed by chondroitinase ABC (0.25 U ml −1 ; Sigma-Aldrich) at 37 • C for 45 min.Non-specific sites were blocked using a 10% goat serum and 1% BSA blocking buffer for 1 h at room temperature.Collagen type I (1:400; ab138492; Abcam), type II (1:400; sc52658; Santa Cruz), and type X (1:300; ab49945; Abcam) primary antibodies were incubated overnight at 4 • C, followed by 20 min treatment using a solution of 3% hydrogen peroxide (Sigma-Aldrich).The secondary antibodies for collagen type I (1:250; ab6720; Abcam), type II (1:300; B7151; Sigma-Aldrich) and type X (1:500, ab97228; Abcam) were incubated for 4 h at RT.Samples were then incubated for 45 min with VECTASTAIN Elite ABC before treating them with ImmPACT DAB EqV (both from Vector Labs) at RT. Slides were then imaged using an Aperio ScanScope slide scanner, while sections stained with picrosirius red were imaged using polarized light microscopy.Analysis on the fiber orientation and coherency was carried out using the OrientationJ plugin in ImageJ software [63].

OA preparation
The alginate oxidation was performed as previously described [64].Briefly, 1 g of alginate (MVG, Pronova Biopolymers) was dissolved in 90 ml of deionized water overnight at RT, and mixed with 10 ml of sodium periodate (Honeywell) to achieve a theoretical alginate oxidation of 4% under stirring in the dark at room temperature for 24 h (based on the molecular weight of the repeat unit, MW 198 g mol −1 ).The oxidized alginate was then purified by dialysis against deionized water for 3 d (MWCO 3500 Dalton; Fischer), sterile filtered through 0.22 µm filter, and lyophilized.Further details can be found here [65].
A single droplet was deposited into each well to yield 17 × 10 3 cells/microchamber (valve opening time of 600 µs, which yields droplets of 0.566 µl).Immediately after bioink deposition, a 45 mM CaCl 2 solution in hgDMEM was added to crosslink the alginate for 5 min at RT.The crosslinking solution was then removed and replaced by XPAN medium with 5 ng ml −1 FGF2.After a period of 24 h, all constructs were switch to CDM.Throughout the culture, the constructs were maintained at 5% pO2.

Mechanical testing
To investigate the mechanical properties of the engineered tissues the different groups were tested using a single column Zwick (Zwick, Rowell) with a 10 N load cell.Unconfined compression tests were conducted on the cylindrical plugs within a PBS bath, as described previously [66].The Young's modulus was defined as the slope of the linear phase of the resulting stress-strain curve during the ramp phase of the compression to 10% or 20% strain.The equilibrium modulus was defined as the stress divided by the applied strain at the end of the relaxation phase, while the dynamic modulus was defined as the average stress over the five cycles divided by the dynamic strain amplitude [67].Uniaxial tensile tests were performed on rectangular sections of dimensions 5 × 30 mm as previously described [68].The elastic modulus was calculated from the linear region of the stress-strain curve.

RNA isolation and quantitative real-time PCR
5 weeks after in vitro culture samples were washed in PBS before snap freezing them until further process.RNA was isolated via Trizol method (Sigma-Aldrich) followed by chloroform extraction and isopropanol precipitation as per manufacturer's instructions.RNA was re-suspended in RNase free water, snap frozen in liquid nitrogen and stored at −80 • C. Polymerase chain reaction (PCR) with a high-capacity cDNA reverse transcription kit (Thermofisher) was conducted to transcribe 500 ng of RNA from each sample into cDNA.The levels of gene expression were measured with real-time PCR (Applied Biosystems) using SYBR green master mix (Applied Biosystems) and porcine specific primers (table 1).Relative expression was obtained by 2 -∆∆Ct method and normalized to the arithmetic average of 2 housekeeping genes 18 S and B2M [69].

Extrusion bioprinting
Cell-laden bioinks were prepared by mixing a 5.3% (w/v) oxidized alginate solution and a 21.5% (w/v) gelatin solution (Gelatin type B; Sigma-Aldrich) with a cell suspension (everything in hgDMEM) to obtain a final solution of 3.5% (w/v) of oxidized alginate and 5% (w/v) of gelatin containing 30 × 10 6 cells ml −1 .The bioinks were loaded into a syringe, and then printed using the 3D Discovery bioprinter (RegenHU) in hgDMEM, and then the cell-laden hydrogel were maintained in XPAN medium for 24 h before switching to CDM at 5% pO2.Media exchange was performed twice weekly until the end of the 6-week culture period.

Live/dead imaging
Cell viability was assessed using the live/dead assay 7 d after the bioprinting process.Briefly, constructs were washed in PBS followed by incubation for 1 h in PBS containing 2 µM calcein acetoxymethyl (calcein AM) and 4 µM ethidium homodimer-1 (EthD-1) (both from Bioscience) for 1 h.Samples were then washed in PBS before imaging with a Leica SP8 scanning confocal microscope excited at 494 nm and 528 nm, and read at 517 nm and 617 nm.

Bioreactor culture
All large-scale constructs (wedge-shaped) were cultured in a custom-made bioreactor system previously described after a first week of static culture [70].The constructs were continuously translated at a speed of 0.05 mm s −1 and an amplitude of 3 mm for the entire culture period.Constructs were cultured in static for the first week and switched to dynamic conditions for the remaining five weeks at 5% pO2.

Rheological assessment of the bioinks
The rheological properties of the alginate-based bioink were evaluated using a rheometer (MCR 102, Anton-Paar) equipped with a Peltier element for temperature control.A plate-plate geometry with a diameter of 25 mm (PP25) was used in all the tests.The viscosity as a function of shear rate (0.1-1000 s −1 ) was conducted at a constant temperature of 16 • C.

Scanning electron microscopy (SEM)
For SEM imaging, samples were mounted on SEM pin stubs with carbon adhesive discs and coated with gold/palladium for 60 s at a current of 40 mA using a Cressington 208 h sputter coater.Imaging was carried out in a Zeiss ULTRA plus SEM.

Statistical analysis
Statistical analysis was performed using GraphPad Prism software.Cell experiments were conducted using a pool of cells derived from two different animal donors.Statistical differences were determined by analysis of variance (ANOVA) followed by Tukey's multiple comparison test, or student's t-test where appropriate.Numerical and graphical results are displayed as mean ± standard deviation.Significance was accepted at a level of p < 0.05.Sample size (n) is indicated within the corresponding figure legends.

Fibronectin adherent populations from the inner and outer regions of the meniscus
Porcine menisci were harvested and processed to obtain fibronectin adherent MPCs and non-selected meniscus cells (NS) from the inner and outer regions of the tissue (figure 2(A)).Their chondrogenic capacity was evaluated by pellet culture of 250 000 cells after passage 2 (P2) and 4 (P4).After 4 weeks of chondrogenic culture, pellets derived from inner NS (iNS) and inner MPCs (iMPCs) had an average diameter of 2.2 ± 0.2 mm and 2.1 ± 0.1 mm, respectively, while those derived from outer NS (oNS) and outer MPCs (oMPCs) had an average diameter of 1.5 ± 0.1 mm and 1.4 ± 0.1 respectively (supplementary figure 1).All cell populations displayed a (fibro)chondrogenic phenotype, as evident from the histological analysis where strong staining for sGAG and collagen deposition was observed, with no evidence of pellet mineralization (supplementary figure 2).At earlier passages P4, pellets generated using iNS contained the lowest levels of sGAG, indicating that serial subculture negatively impacted maintenance of its native phenotype; this was also observed in the alcian blue staining (supplementary figure 2).At P4, pellets generated using iNS also contained the lowest levels of collagen (figure 2(E)), further confirming a loss of phenotype in these cells with serial subculture.Such a reduction in collagen synthesis with cell expansion was not observed in iMPCs, which secreted almost twice the amount of collagen as iNS in pellet culture following four passages.Control MSCs isolated from bone marrow were found to secrete comparable levels of sGAG and collagen at P2 and P4 (supplementary figure 3).NS did not demonstrate a capacity to undergo osteogenesis, with only MPCs generating an ECM that stained positive for mineral deposition following culture in osteogenic media (figure 2(F)).It should be noted, however, that the osteogenic potential of MPCs was not comparable to bone marrow derived MSCs, which underwent more robust mineralization in vitro.Overall, these results suggest that fibronectin selection can be used to isolate MPCs from the inner and outer regions of the meniscus; these cells maintain their differentiation capacity with prolonged monolayer expansion, opening their use within advanced biofabrication strategies.

Biofabrication of structurally organized and phenotypically distinct fibrocartilaginous tissues
After having demonstrated the potential of fibronectin selected MPCs, we next explored if our previously developed MEW PCL scaffolds could be used to engineer spatial organized fibrocartilaginous tissue using this cell source, similar to that observed using bone marrow derived MSCs [57].By using a temporal bioink, based on oxidized alginate, it is possible to provide temporal support for the cells after the jetting process.After crosslinking, the rapidly degrading hydrogel allows the cells to self-organize following the microchamber architecture [57].To this end, MPCs and bone marrow derived MSCs were inkjet bioprinted into MEW PCL scaffolds with a pore (or microchamber) geometry of 0.4 × 1.6 mm, with a total height of 1 mm (figure 3(A)).For these and subsequent experiments, we used cells at passage 4 unless otherwise stated.Building on previous investigations, a one-time cABC treatment was applied at day 14 of chondrogenic culture to enhance collagen network maturation [71].Following 5 weeks of in vitro culture, robust chondrogenesis of MPCs was observed within the MEW scaffolds, as evident by alcian blue staining for sGAG deposition (figure 3(B)).Small nodes of calcification were observed in the tissues generated using MSCs, which was not observed for the MPCs.MSCs, iMPCs and oMSCs all secreted abundant levels of sGAG, with MSCs secreting the highest levels of sGAGs after the 5 weeks in culture (figure 3(D)).All three cell populations deposited similar amounts of collagen (figure 3(E)).
Next, the compressive and tensile properties of the engineered constructs were assessed [67].No significant difference was observed in either the compressive ramp or dynamic modulus of the different constructs (figures 3(F) and (G)), although all constructs were over 25 times stiffer than that of empty MEW scaffolds, with the ramp modulus values ranging between 75.24 and 99.20 kPa at 10% strain, and 98.40 and 181.5 kPa at 20% strain.Similar to native meniscus tissue, which displays dramatic tension-compression nonlinearity, all bioprinted tissues were over an order of magnitude stiffer in tension than in compression.This can be at least partially attributed to the high aspect ratio of the MEW fibers, which preferentially increases the tensile properties of the resulting construct.After 5 weeks of culture, the tensile stiffness of tissues bioprinted using iMPCs was higher than those generated using oMPCs (figure 3(H)).
We next sought to investigate if preferentially aligned collagen was laid down by MPCs within the MEW scaffolds.Picrosirius red staining confirmed robust collagen deposition in all groups (figure 4(A)).We then used polarized light microscopy (PLM) to analyze the degree of orientation of the collagen fibers in the engineered tissues.From the color map, most of the fibers appeared to be oriented parallel to the long axis of the scaffold pores or microchambers, with a small portion of the fibers also oriented parallel to the short axis.Further quantification indicated that most of the fibers were oriented preferentially at 0.1 ± 0.1 • , following the physical boundaries imposed by the MEW fibers.Furthermore, coherency values, which are a measure of the collagen fibers dispersion, confirmed a preferential organization of the secreted collagen network (figure 4

(C)).
To characterize tissue phenotype, immunohistochemistry was undertaken to identify the collagen types present within the engineered tissues (figure 5(A)).Tissues generated using oMPCs stained more intensely for collagen type I, but less intensely for collagen type II.Light staining for collagen type X was observed in all groups.qPCR confirmed that oMPCs express the lowest levels of type II collagen, adopting a phenotype more representative of the outer region of the meniscus, with a higher ratio of collagen type I to type II expression (figures 5(B)-(D)).Tenomodulin expression, which is a known tendon marker expressed in the outer region of the meniscus, was also 8-fold higher in constructs generated using oMPCs compared to the other cell populations (figure 5(G)).Conversely, the expression of type I collagen relative to type II collagen was lower in iMPC compared to oMPCs, indicating a more inner zone or hyaline-like cartilage phenotype (figure 5(D)).Collagen type X expression was dramatically lower in the MPCs compared to MSCs.Aggrecan expression was also ∼five-fold higher in MSCs compared to  the MPCs (figure 5(F)).Taken together, these results indicate that the MPCs display a unique phenotype that resembles that of their native niche.

Biofabrication of scaled-up, wedged-shaped fibrocartilaginous tissues
The native meniscus has a unique wedge-shaped profile which is important for the stability of the knee joint.In order to recapitulate this anatomical feature, we 3D scanned a porcine meniscus to reconstruct its profile (figure 6(A)).An iterative MEW process was implemented to print scaffolds with a similar 'stepped' profile (figure 6(B)).This process allowed us to print constructs of up to 5 mm height without any detrimental effect on scaffold porosity and interconnectivity (figure 6(B) 1-4).To spatially deposit MPC or MSC laden bioinks into these anatomically defined MEW scaffolds, we implemented an extrusion-based bioprinting process.Specifically, vertical extrusion (in the z-axis) allowed the controlled deposition of shear-thinning oxidized alginate-gelatin bioinks (supplementary figure 4) into the pores of the MEW scaffold.To promote nutrient diffusion within such scaled-up constructs, sacrificial channels of gelatin (that melt at 37 • C) were also 3D printed in the construct (supplementary figure 5).Live/dead analysis 7 d after bioprinting confirmed excellent cell viability, indicating that the biofabrication approach could successfully be used for the engineering of large constructs (figure 6(C)).
We next sought to engineer tissues with regionally distinct phenotypes resembling the native meniscus by spatially depositing iMPCs and oMPCs into the inner and outer regions of the scaled-up MEW scaffold (supplementary figure 6).To that end we bioprinted two different constructs: (1) using only MSCs, and (2) using iMPCs and oMPCs into distinct regions of the MEW scaffold to engineer constructs more mimetic of the native meniscus (figure 7(A)).One week after bioprinting, the different constructs were transferred to a custom bioreactor for another 5 weeks of culture under dynamic conditions.After a total of 6 weeks in culture, robust tissue formation was observed macroscopically throughout the scaffold (figure 7(B)).Histological staining for total collagen and sGAG also demonstrated ECM deposition throughout the entire scaffold for both MSCs and MPC groups, with little evidence of calcification (figure 7(C)).Immunochemistry revealed a collagen type II rich hyaline cartilage-like matrix within the MSC constructs, whereas a more type I collagen rich fibrocartilage-like matrix was generated by the MPCs more similar to the native meniscus (figure 7(C), supplementary figure 7).PLM confirmed that the collagen orientation followed the architecture of the MEW scaffolds pores/microchambers (supplementary figures 8(A) and (B)), confirming that this physical guidance is not lost in our scaled-up tissues.Quantitative evaluation through biochemical assays demonstrated that there were no major differences in total levels of ECM production across the different groups (supplementary figure 8(C)).

Discussion
This study demonstrates the capacity of MPCs to synthesize neo-fibrocartilage following prolonged passaging and to be used in the bioprinting of anisotropic and phenotypically defined meniscal constructs.These cells, isolated based on binding to fibronectin, maintained their capacity to generate a fibrocartilage-like tissue upon serial expansion.By jetting zone-specific MPCs into MEW scaffolds with elongated pores, it was possible to engineer structurally organized meniscal tissue with significant tension-compression non-linearity.These meniscal-like tissues recapitulated key features of the inner and outer zone of the meniscus, including differential expression of collagen type I, collagen type II, and zone-specific markers such as tenomodulin (TNMD).Finally, we biofabricated a 5 mm high MEW scaffold with a wedged-shaped profile similar to that of the native tissue, into which we spatially deposited iMPCs and oMPCs, thereby enabling the biofabrication of scaled-up meniscal constructs.Taken together, this work motivates (1) the use of MPCs for meniscus tissue engineering, and (2) the application of MEW in combination with bioprinting technologies for the biofabrication of meniscus-like tissues.
Based on differential adhesion to fibronectin [39,43,44,72], we were able to isolate progenitor cell populations from both the inner and outer zone of the meniscus.MPCs maintained their capacity to generate meniscal-like tissue upon serial passage, making them a particularly attractive cell source for meniscus TE strategies where large numbers of cells are required to fabricate human scale tissues.Unlike MSCs, which have an inherent tendency to become hypertrophic and progress along an endochondral pathway [34,73], MPCs expressed type X collagen (a marker of hypertrophy) at very low levels, and the resultant tissues did not calcify in culture.Furthermore, while MPCs generate a calcified tissue in osteogenic culture conditions, they do so to a lesser extent than bone marrow derived MSCs.Similar results have been observed with articular cartilage progenitor cells [39][40][41]74], which have also been used for meniscus repair applications [42].Zone specific MPCs also appear superior to the use of inner or outer zone meniscal fibrochondrocytes (non-selected cells) for meniscus tissue engineering.Unlike non-selected fibrochondrocytes, which displayed a diminished capacity for ECM production with serial passaging, MPCs secreted comparable levels of sGAG and collagen at passage 2 and 4. It should be noted that the MPCs were isolated form animals that had not reached full skeletal maturity.Furthermore, changes in phenotype might be expected with skeletal maturation [75,76], which might lead to the generation of MPCs with even greater zonal definition.A limitation of this study was that we did not undertake a detailed characterization of the MPCs (e.g.flow cytometry) following the fibronectin isolation protocol.A recent investigation using fibronectin adhesion for the isolation of progenitor cells revealed that commonly used MSC-associated surface markers (e.g.CD90 and CD105) were not able to distinguish between selected and non-selected populations, indicating the need for more specific markers [43,77].Integrin marker expression levels has been suggested as potential target to identify progenitor cells in the meniscus tissue [72].Future work is required to better characterize the progenitor cell populations isolated from the different regions of the meniscus, and to identify culture conditions that best maintain their phenotype during monolayer expansion.
MPCs were also able to generate structurally organized grafts in response to the physical constraints provided by the MEW scaffolds.We have previously shown that MSCs can generate such anisotropic tissues [57], and here we demonstrate that tissue-specific progenitor cells can also respond to such architectural cues and generate a collagenous matrix that aligns parallel to the long axis of the MEW scaffold pores.Such structured organization has been observed with different scaffolding systems [54][55][56], and has been previously hypothesized to be directed by cell generated forces [78].This results in collagen alignment that is dependent on the aspect ratio and width of the pores of the scaffold.The tensile and compressive properties of the tissues generated by the MPCs and MSCs were comparable, dramatically increasing over the five-week culture period.Importantly, the resultant tissues displayed significant tension-compression nonlinearity, mimicking the mechanical behavior of the native tissue.Despite this, the tensile stiffness of the engineered tissues was still significantly lower than that of the native tissue, where modulus values range between 100 and 200 MPa in the circumferential direction [79].Surprisingly, the iMPCs generated a stiffer tissue than the oMPCs.This could be linked to differences in the relative production of type I and II collagens, and/or due to altered levels of sGAG accumulation which can negatively impact collagen network maturation [80][81][82].Further work is required to improve the tensile mechanical properties of these grafts, which might involve the implementation of biochemical and biomechanical stimulation regimes which have shown to enhance the tensile properties of engineered tissues [46,[83][84][85].In particular, employing strategies to develop native-sized hierarchical collagen fibers will likely be important to the engineering of functional meniscal grafts [86].
The tissues generated following the jetting of iMPCs and oMPCs into the MEW scaffolds were phenotypically distinct.Immunohistochemistry and gene expression analysis revealed oMPCs exhibit a higher Col I to Col II ratio compared to iMPCs and MSCs, with the latest displaying the highest individual values of gene expression, overall mimicking the native collagen distribution [5,87].Furthermore, oMPCs expressed higher levels of TNMD, a marker more specific in the outer region of the meniscus, indicating meniscus derived cells retain regionally distinctive phenotypes [68].The bioprinted tissues generated from both MPCs populations were also phenotypically distinct from that derived from MSCs.Significantly higher levels of GAGs (and higher type II collagen expression) were observed in chondrogenically primed MSCs, suggesting these cells are generating a more GAG-rich, hyaline cartilage-like tissue that may be more suited to engineering the inner zone of the meniscus.It should be noted, however, that even the inner zone of the meniscus is a mix of type I (40%) and type II (60%) collagen [5].It is also important to highlight that the meniscus contains dramatically lower levels of proteoglycans than articular cartilage [88].Therefore, the sGAG and type II collagen rich tissue generated by MSCs is not truly mimetic of the inner zone of the meniscus.Furthermore, previous studies have shown that tissues engineered using meniscus-derived cell populations and MSCs not only differ in composition, but also in terms of collagen matrix organization, emphasizing the need to adequately select the cell source for meniscus tissue engineering [82].Together, these results point to the potential of using iMPCs and oMPCs to biofabricate regionally defined constructs mimicking the zonal phenotypes observed in the native meniscus.
Having demonstrated the capacity to engineer anisotropic and phenotypically distinct fibrocartilaginous tissues using MPCs derived from the inner and outer regions of the meniscus, we next sought to biofabricate wedge-shaped meniscal tissues with regionally distinct compositions.Despite the significant advancements achieved by MEW in scaffold fabrication for tissue engineering, there remain limitations driven by electrostatic buildup and electrostatic repulsion [60,89].In this study, a maximum height of 5 mm was reached while maintaining control over polymer deposition.Furthermore, the limited congruency of the scaffold described in this work could be overcome by employing non-planar printing [90][91][92].The application of z-printing enabled the rapid and controlled deposition of zone-specific MPCs within specific regions of the MEW scaffold [67,93,94].Tissue engineering such scaled-up constructs requires satisfying the nutrient demands of large numbers of cells, which if not appropriately considered can lead to cell death and core regions devoid of matrix [70].In order to address this challenge, we implemented the use of dynamic bioreactor culture and introduced microchannels into the bioprinted construct as it has previously been shown to enhance collagen and sGAG synthesis in large engineered tissues [93,95,96].For the creation of microchannels, a fugitive gelatinbased bioink was deposited using z-bioprinting leaving behind a network of microchannels [97,98].To generate regionally defined tissues, we deposited iMPCs within the inner region of the construct, and oMPCs within the outer region of a wedged-shaped scaffold.The MPCs were again found to generate a more fibrocartilaginous meniscal graft compared to MSCs, which instead supported the development of a more hyaline cartilage-like tissue rich in type II collagen.In this proof-of-concept study, not all spatial features of the meniscus, such as the presence of a transition zone between the avascular and vascular regions of the tissue, were considered.The integration of zone-specific biochemical cues (e.g.localized loading of growth factors) and biophysical cues (e.g.providing anisotropic stimulation regimes) might further enhance the development of more regionally defined meniscal grafts in long-term culture [18,22,83,99,100].Achieving true zone-specific functionality requires the accurate recapitulation of many key features, including the localized confinement of the microvasculature and the establishment of the transition zone.Imbalances in this organization have been correlated with the onset and progression of joint diseases, such as OA.Potential strategies to achieve this goal might include the bioprinting of tissues with spatially defined vasculature [101], or the inclusion of biomaterials that can enhance or inhibit the process of vascularization [102,103] and/or promote region-specific meniscal cell phenotypes [22].
While our findings point to the potential of MPCs as a cell type for tissue engineering and 3D bioprinting, further investigation is necessary to determine their true regenerative potential and clinical feasibility.Exploring the allogeneic use of MPCs could greatly enhance their clinical applicability, as it has the potential to increase availability, reduce costs, and enable cell preselection [104].Moreover, conducting comparative studies between MPCs and non-selected meniscus-derived populations for transplantation is crucial for advancing our understanding in this field.Limitations such as patient burden and costly cellexpansion phases have driven meniscus repair and regeneration towards single-stage procedures and off-the-shelf cellular or biomaterial-based products.However, the promising results obtained by the 'Cell Bandage' treatment for avascular tears, where autologous bone marrow MSCs are seeded onto a collagen matrix harvested from the iliac crest two weeks before meniscus surgery, point to the potential of more complex cell-based approaches [105].In conclusion, considerations regarding clinical implementation and market acceptance are important factors that need addressing for successful translation into the clinic.

Conclusion
In these series of experiments, we established a novel biofabrication framework that enabled the engineering of zone-specific meniscal tissue grafts.We first demonstrated the benefit of utilizing MPCs over MSCs or meniscal fibro-chondrocytes as a cell source for meniscus tissue engineering.We then demonstrated that it is possible to engineer structurally organized meniscal grafts, that displayed significant tension-compression nonlinearity, by inkjetting either iMPCs or oMPCs into structurally defined MEW scaffolds.Zone specific MPCs were able to generate tissues with distinct regional phenotypes following printing, including altered ratios of collagen type I to collagen type II expression, and differential expression of regional markers such as TNMD.Finally, we developed a bioprinting strategy to fabricate scaled-up wedge-shaped scaffolds using MEW.These findings encourage further investigation on the application of MPCs and MEW for the biofabrication of meniscal grafts for regenerative medicine applications.

Figure 1 .
Figure 1.Experiments used to assess the fibrochondrogenic potential of porcine meniscus-derived progenitor cells (MPCs) in tissue engineering.(A) Initially the fibrochondrogenic potential of both MPCs and non-selected populations derived from both the inner and outer regions was assessed using pellet culture.(B) An investigation of MPCs in the biofabrication of engineered fibrocartilaginous tissues was then undertaken.To achieve this, 3D bioprinting techniques (inkjet and extrusion) were employed to deposit cells into supportive and guiding scaffold structures manufactured using MEW.Inkjet printing was used to deposit cells into small-scale scaffolds, while extrusion based bioprinting was used to deposit cells into the larger, wedged-shaped constructs.

Figure 2 .
Figure 2. Chondrogenic and osteogenic potential of the different meniscus resident cell populations.(A) Schematic of the cell isolation process for non-selected (NS) and meniscus progenitor cells (MPCs).(B) Immunohistochemistry staining for collagen type I, type II and type X after 28 d of culture (scale bar is equal to 1 mm).Biochemical analysis at day 28 for the different cell populations, (C) DNA content, (D) sGAG content and (E) collagen content.All error bars denote standard deviation, significance was considered * p ⩽ 0.05, n = 4. $ = significance between P2 and P4.(F) Alizarin red staining following 3 weeks of culture with osteogenic media.Scale bar is equal to 1 mm.

Figure 3 .
Figure 3. Histological analysis and biochemical content of the engineered tissues.(A) Schematic of the bioprinting process into the MEW scaffolds.(B) Histological staining for sGAG (Alcian blue) and calcium deposits (Alizarin red).Scale bar is equal to 400 µm.Biochemical analysis at day 35 for the different cell populations, (C) DNA content, (D) sGAG content and (E) collagen content.Biomechanical analysis at day 35 for the different cell populations (F) compression stress-strain curves of constructs from each group, (G) compression modulus of the constructs derived from the linear region of the compressive stress-strain curves, (H) equilibrium and (I) dynamic modulus.(J) Tensile stress-strain curves of constructs from each group and (K) tensile modulus of the constructs derived from the linear region of the tensile stress-strain curves.All error bars denote standard deviation, significance was considered * p ⩽ 0.05, n = 4.

Figure 4 .
Figure 4. Collagen fiber orientations for the different cell populations following 5 weeks days of in vitro culture.(A) Picrosirius red, polarized light (PLM) and color map imaging of the collagen fiber distributions.Scale bar is equal to 400 µm.Here cyan/blue colors denote fibers oriented at 0 degrees while purple/red denote fibers oriented at 90 degrees.(B) Quantification of the fiber orientation.(C) Quantification of the fiber coherency, where values approaching to 1 indicate that fibers are aligned towards a preferential direction.All error bars denote standard deviation.

Figure 5 .
Figure 5. Phenotypical characterization.(A) Immunohistochemical staining for collagen type I (Col I), collagen type II (Col II) and collagen type X (Col X) deposition.Scale bar is equal to 400 µm.qPCR analysis of the engineered tissues using the different cell types, showing relative gene expression of (B) collagen type I, (C) collagen type II, (D) collagen type I/type II ratio, (E) collagen type X, (F) aggrecan and (G) TNMD, normalized against the housekeeping genes B2M and 18 S. Statistically significant difference are marked with an * (p ⩽ 0.05).

Figure 6 .
Figure 6.MEW fabrication of the wedged-shaped scaffold and bioprinting of cells.(A) Outline of the biofabrication process of wedge-shaped constructs.(B) Top-macroscopic images of the melt electrowritten scaffold during and post printing.Scale bar is equal to 4 mm.Bottom-scanning electron microscopy images of different regions of the scaffold.Scale bar is equal to 1 mm.(C) Cell viability (MSCs) in cross section and bottom plane after 7 d in culture.Scale bars are equal to 2 mm and 400 µm respectively.

Figure 7 .
Figure 7.In vitro chondrogenesis within the wedge-shaped scaffold bioprinted with MSCs or MPCs and cultured in a dynamic culture system for 6 weeks.(A) Outline of the groups and the dynamic culture conditions.(B) Macroscopic images of the constructs following the 6 weeks of culture.Scale bar is equal to 2 mm.(C) Histological and immunohistochemical staining for total collagens (PR), sulphated glycosaminoglycan (AB), calcium deposition (AR), collagen type I (Col I), collagen type II (Col II) and collagen type X (Col X) deposition.Scale bar is equal to 2 mm.

Table 1 .
Primer sequences for real-time PCR.
at 16 • C, using a pressure of 0.07 MPa, a translational speed of 4 mm s −1 and a 27 G needle (Adhesive Dispensing Ltd).After printing, the constructs were incubated for 30 min in a bath of 45 mM CaCl 2 tissue bioRxiv https://doi.org/10.1101/2022.09.07.506986 (posted online 09 September 2022) (accessed on 01 April 2023) [72] Wang J, Roberts S, Li W and Wright K 2022 Phenotypic characterization of regional human meniscus progenitor cells Front.Bioeng.Biotechnol. 10 1003966 [73] Pelttari K, Winter A, Steck E, Goetzke K, Hennig T, Ochs B G, Aigner T and Richter W 2006 Premature induction of hypertrophy during in vitro chondrogenesis of human mesenchymal stem cells correlates with calcification and vascular invasion after ectopic transplantation in SCID mice Arthritis Rheum.54 3254-66 [74] McCarthy H E, Bara J J, Brakspear K, Singhrao S K and Archer C W 2012 The comparison of equine articular cartilage progenitor cells and bone marrow-derived stromal cells as potential cell sources for cartilage repair in the horse Veterinary J. 192 345-51 [75] Gannon A R, Nagel T, Bell A P, Avery N C and Kelly D J 2015 Postnatal changes to the mechanical properties of articular cartilage are driven by the evolution of its collagen network Eur.Cell Mater.29 105-21 [76] Tsinman T K, Jiang X, Han L, Koyama E, Mauck R L and Dyment N A 2021 Intrinsic and growth-mediated cell and matrix specialization during murine meniscus tissue assembly FASEB J. 35 e21779 [77] Sun H et al 2020 Single-cell RNA-seq analysis identifies meniscus progenitors and reveals the progression of meniscus degeneration Ann.Rheum.Dis.79 408 [78] Knychala J, Bouropoulos N, Catt C, Katsamenis O, Please C and Sengers B 2013 Pore geometry regulates early stage human bone marrow cell tissue formation and organisation Ann.Biomed.Eng.41 917-30 [79] Tissakht M and Ahmed A M 1995 Tensile stress-strain characteristics of the human meniscal material J. Biomech.28 411-22 [80] Douglas T, Heinemann S, Bierbaum S, Scharnweber D and Worch H 2006 Fibrillogenesis of collagen types I, II, and III with small leucine-rich proteoglycans decorin and biglycan Biomacromolecules 7 2388-93 [81] Makris E A, MacBarb R F, Paschos N K, Hu J C and Athanasiou K A 2014 Combined use of chondroitinase-ABC, TGF-β1, and collagen crosslinking agent lysyl oxidase to engineer functional neotissues for fibrocartilage repair Biomaterials 35 6787-96 [82] McCorry M C and Bonassar L J 2017 Fiber development and matrix production in tissue-engineered menisci using bovine mesenchymal stem cells and fibrochondrocytes Connect.Tissue Res.58 329-41 [83] Zhang Z-Z et al 2019 Orchestrated biomechanical, structural, and biochemical stimuli for engineering anisotropic meniscus Sci.Transl.Med.11 eaao0750 [84] Huey D J and Athanasiou K A 2011 Tension-compression loading with chemical stimulation results in additive increases to functional properties of anatomic meniscal constructs PLoS One 6 e27857 [85] Lee J K, Huwe L W, Paschos N, Aryaei A, Gegg C A, Hu J C and Athanasiou K A 2017 Tension stimulation drives tissue formation in scaffold-free systems Nat.Mater.16 864-73 [86] Puetzer J L, Ma T, Sallent I, Gelmi A and Stevens M M 2021 Driving hierarchical collagen fiber formation for functional tendon, ligament, and meniscus replacement Biomaterials 269 120527 [87] Sanchez-Adams J, Wilusz R E and Guilak F 2013 Atomic force microscopy reveals regional variations in the micromechanical properties of the pericellular and extracellular matrices of the meniscus J. Orthopaedic Res. 31 1218-25 [88] Chen S, Fu P, Wu H and Pei M 2017 Meniscus, articular cartilage and nucleus pulposus: a comparative review of cartilage-like tissues in anatomy, development and function Cell Tissue Res.370 53-70