Xenograft of bio-3D printed scaffold-free cartilage constructs derived from human iPSCs to regenerate articular cartilage in immunodeficient pigs

Abstract

Arthroplasty is currently the only option for reconstruction of large articular cartilage defects, mainly due to osteoarthritis. However, reconstruction with artificial materials has several issues such as deterioration, foreign body reaction, and bacterial infection. This study established a new method for articular cartilage reconstruction that fundamentally solves the problems associated with artificial materials by creating scaffold-free cartilage constructs and implanting them into extensive osteochondral defects without artificial materials. Tubular cartilage constructs were fabricated using a completely scaffold-free Kenzan bio-three-dimensional printing method with chondrogenic spheroids generated from human induced pluripotent stem cell-derived mesenchymal stem/stromal cells (iPSC-MSCs). The constructs were partially cut open to form a patch and implanted into osteochondral defects in the femoral trochlear groove of immunodeficient miniature pigs. The cartilage constructs were elastic and easy to handle, and abundant glycosaminoglycans and collagens were observed in the grafted site after implantation, as well as in the articular cartilage. Cells at this site were positive for human vimentin, indicating that the cartilage constructs were successfully engrafted into the host subchondral bone. Scaffold-free human iPSC-MSC-derived cartilage constructs implanted into osteochondral defects contribute to the regeneration of extensive osteochondral defects in the absence of artificial materials.

1. Introduction

Articular cartilage (AC), also known as hyaline cartilage, consists of chondrocytes and an extracellular matrix (ECM) composed mainly of type II collagen (Col II) and proteoglycans and covers the end of the bone in a joint. Because AC is a non-vascularized tissue with poor self-healing properties, it is repaired by fibrocartilage when damaged. However, fibrocartilage is mechanically inferior to hyaline cartilage; thus, damaged AC eventually deteriorates [1]. Osteoarthritis (OA) is a major joint disease affecting numerous middle-aged and older people worldwide, causing chronic pain and loss of mobility [2]. Eventually, in the late stages of OA, AC loss is extensive. In young individuals, trauma-induced AC damage can be treated with deliberate microfractures, which result in fibrocartilage coverage of the repaired tissue. At the expense of mechanical strength, this approach provides primary pain relief and restoration of joint function [3]. However, professional athletes who received microfracture treatment showed worse performance and poorer post-injury outcomes than pre-injury [4].

As cartilage damage does not heal spontaneously, cartilage regeneration has been studied. However, extensive cartilage defects in OA have not yet been regenerated. Currently, metal-based prostheses are the first choice for end-stage OA; however, various problems such as infection and loosening make their use in younger patients difficult [5]. Experimental attempts have been made to create anatomical cartilage implants using various materials as scaffolds [6]. However, scaffold materials, as well as metals, are prone to infection and elicit an immunological responses [7].

Therefore, we fabricated cartilage constructs using only cells. We developed a bio-three-dimensional (3D) printer based on the Kenzan method, which can fabricate arbitrarily shaped 3D cellular constructs without scaffolds [[8], [9], [10]]. When spheroids consisting of tens of thousands of cells are stacked on microneedles and cultured, adjacent spheroids adhere and fuse to form scaffold-free 3D constructs. This construct avoids the issues associated with the use of artificial materials, as mentioned above. This bio-3D printing technology has been successfully used to fabricate a number of different tissues, including blood vessels [11], nerve conduits [12], trachea [13], myocardium [14], and liver [15], with a wide variety of cell types. We previously reported the fabrication of human induced pluripotent stem cell (iPSC)-derived cartilage constructs in vitro [16]. iPSCs can be generated by adding four reprogramming factors [17,18] and have pluripotent capacity and self-renewal, and unlike embryonic stem cells, which must be harvested from embryos, there are less ethical issues associated with their use. However, while research into regenerative medicine using stem cells, such as iPSCs, is advancing dramatically, the clinical application of human iPSC-derived AC is still a long way off.

For the clinical application of tissue-engineered AC, it is useful to conduct studies in large animal models that are similar in size and loading environment to humans rather than in small animals such as rats and rabbits. Pigs are diurnal and their knees are extended and loaded during the day. However, when they sleep, their knee joints are not loaded. This lifestyle is similar to that of humans; therefore, pigs have been considered a suitable animal model for AC implantation [19,20]. Miniature pig (mini-pig) cartilage is histologically similar to human cartilage, with knee joint size, loading, lack of spontaneous damage healing, and collagen network arrangement similar to those of human joints [[21], [22], [23], [24], [25]]. Furthermore, mini-pig knee joints have been shown to have stress values similar to those of human knee joints [21]. Itoh et al. created an immunodeficient mini-pig model to implant human dermal fibroblast-derived blood vessels for artery-vein bypass using the aforementioned bio-3D printer [26]. This immunodeficient model is desirable for preclinical studies, in which human-derived cells are implanted to evaluate their long-term function.

In this study, we implanted bio-3D printed human iPSC-derived cartilage constructs into extensive osteochondral defects in immunodeficient mini-pigs to evaluate the utility of this new alternative treatment for AC regeneration.

2. Material and methods

2.1. Cell culture and MSC induction

Human iPSC-derived neural crest cells (iNCCs) differentiated from human iPSC line 414C2 were used in this study. Cultures and passages were performed as previously described [16,27]. Briefly, iNCCs were grown in chemically defined medium (CDM) consisting of Iscove's modified Dulbecco's medium (Sigma-Aldrich, St. Louis, MO, USA) and Ham's F-12 (Gibco, Grand Island, NY, USA) at a 1:1 ratio. CDM contained 1 % chemically defined lipid concentrate (Gibco), 15 μg/mL apo-transferrin (Sigma-Aldrich), 450 μM monothioglycerol (Sigma-Aldrich), 5 mg/mL purified bovine serum albumin (BSA; Sigma-Aldrich), 7 μg/mL insulin (Wako Pure Chemical Industries, Osaka, Japan), 10 μM SB431542 inhibitor (Selleck Chemicals, Houston, TX, USA), 20 ng/mL epidermal growth factor (EGF; R&D Systems, Minneapolis, MN, USA), 20 ng/mL basic fibroblast growth factor (bFGF; Wako), and a supplement containing 50 U/ml penicillin and 50 μg/ml streptomycin (Invitrogen, Carlsbad, CA, USA). iNCCs were cultured on fibronectin-coated dishes (Millipore, Billerica, MA, USA) and Accutase (Wako) was used for passaging.

To differentiate iNCCs into induced pluripotent stem cell-derived mesenchymal stem/stromal cells (iPSC-MSCs), CDM was replaced with MSC medium, which contained minimum essential medium Eagle alpha modification (aMEM; Nacalai Tesque Inc., Kyoto, Japan), 10 % fetal bovine serum (FBS; Hyclone, Logan, UT, USA), 5 ng/mL bFGF (Wako), 50 U/ml penicillin, and 50 μg/ml streptomycin. Trypsin-EDTA (0.25 %; Gibco) was used to passage cells. The medium was changed every other day.

2.2. Chondrogenic induction and preparation of spheroids

As previously reported [16], chondrogenic induction was performed using iPSC-MSCs from passage three. Chondrogenic Differentiation Basal Medium supplemented with hMSC chondrogenic single quotes containing dexamethasone, ascorbic acid, ITS + supplement, GA-1000, sodium pyruvate, proline, and l-glutamine (Lonza, Walkersville, MD, USA) was used for chondrogenic differentiation. Additionally, 40 ng/ml platelet-derived growth factor-BB (PDGF-BB; R&D Systems) was used from days 0–10 after the start of differentiation induction, and 10 ng/ml transforming growth factor beta-3 (TGF-β3; Peprotech Inc., Rockey Hill, NJ, USA) and 50 ng/ml bone morphogenetic protein 4 (BMP4; R&D Systems) were added from days 6 and 10, respectively. From the start of induction, 4 × 10⁴ cells were seeded in 96-well non-adherent plates (Sumitomo Bakelite Co., Ltd., Tokyo, Japan), and spheroids reached a diameter of approximately 700 μm after 10 days. The medium was changed twice per week. All the cultures were grown in a humid environment at 37 °C and 5 % CO₂.

2.3. Biofabrication of cartilage constructs

After spheroid-induced chondrogenic differentiation for 10 days in a 96-well plate, the spheroids were placed on a 7-mm-diameter Kenzan with a circular array of 36 microneedles using the Regenova Bio-3D Printer (Cyfuse Biomedical K.K., Tokyo, Japan) to create a 3D tube that had previously been designed using computer-aided design (CAD). Regenova automatically picks up spheroids by suctioning them with a nozzle, and places them on the Kenzan, where a fine-needle array is arranged. Spheroids fuse with each other, and scaffold-free 3D constructs can be created [[8], [9], [10]]. After stacking, the constructs were cultured in a bioreactor with internal reflux function, and the chondrogenic differentiation induction medium was changed twice a week for over a month to allow sufficient maturation of the constructs for implantation.

Animals and creation of the operational immunodeficient pig (OIDP) model.

Nine mini-pigs (Nippon Institute for Biological Science [NIBS]: 4 males, and Göttingen: 3 males and 2 females; Oriental Yeast Industry, Nagano, Japan) were used in this study. Their mean age and body weights were 21.8 [[11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24], [25], [26], [27], [28], [29], [30], [31], [32], [33], [34], [35], [36], [37], [38], [39], [40], [41], [42], [43], [44], [45], [46]] months old and 30.4 (18.2–40.6) kg, respectively. For surgical anesthesia, the mini-pigs were sedated with intramuscular injections of atropine sulfate (0.05 mg/0.1 ml/kg), medetomidine hydrochloride (0.05 mg/0.05 ml/kg), and midazolam (0.5 mg/0.1 ml/kg) followed by a mixture of N₂O:O₂ = 1:1 gas, and general anesthesia was maintained with 2.0 % isoflurane and managed with a ventilator.

In accordance with a previous report [26], the three lobes of the thymus were excised following a midline incision in the ventral region of the neck. The spleen was then removed through a midline abdominal incision. After the gastric corpus was exposed, a gastrostomy tube (GB gastrostomy balloon catheter, tubular type, cannula outer diameter: 24 Fr; Nipro Corporation, Osaka, Japan) was surgically inserted, and the insertion site was ligated with a drawstring suture and fixed to the abdominal cavity wall. One end of the tube was suture-fixed to the skin on the back through a subcutaneous tunnel. A central venous catheter (Medicat L CV-UK kit, single-lumen type, catheter outer diameter: 14 G; Japan Covidien Corporation, Tokyo, Japan) was placed in the anterior vena cava sinus for blood sampling and sutured to the neck. The inside of the vascular catheter was locked with heparinized saline solution (200 units/ml) every 1–3 days.

Tacrolimus (0.25 mg/kg) and mycophenolate mofetil (30 mg/kg) were administered twice daily, and prednisolone (20 mg/body) was administered once daily through a gastrostomy tube for immunosuppression. Blood samples for tacrolimus and mycophenolate mofetil were collected twice a week, and the doses were adjusted to achieve trough blood levels of 5–10 ng/ml and 1–3 μg/ml, respectively.

Ethical review of all procedures in this study was conducted by the Institutional Animal Care and Use Committee and the Human Sample Experiments Ethical Committee of Nihon Bioresearch Inc. (Approval Nos. 380246, 390141, 390193, 409063, 409190, and 410042) and the Kurume University Animal Experiment Committee (Approval No. 2022–175). We confirmed that all methods were carried out in accordance with relevant guidelines and regulations and reported in accordance with the ARRIVE guidelines for the reporting of animal experiments.

2.4. Implantation of cartilage constructs into extensive AC defects

One week after OIDP creation, the same anesthesia was administered, and the knee joints were incised using a medial parapatellar approach to dislocate the patella and expose the trochlear groove. Osteochondral defects (20 × 10 mm, 2 mm deep) were created in the trochlear groove using a bone chisel, and mature cartilage constructs were cut into patches and implanted (Fig. 1a–c, Supplementary Video 1). The constructs were threaded with sutures (7–0 Proline; Ethicon, Somerville, NJ, USA) and fixed onto osteochondral defects with simple interrupted or pull-out sutures (Fig. 1c). The spheroids of the constructs were well-fused to each other and fixed onto the defects using the sutures without tearing. In the control group, only osteochondral defects of the same size as those in the implantation group were created. In eight pigs, we operated on one knee and left the other knee unoperated. In the other pig, we created a defect in one knee and then implanted the construct into the other knee two months later. In both the implanted and control groups, two pins were inserted into the femur and tibia, and an external fixator (IMEX SK Linear External Fixator Small; Kirican Ltd., Tokyo, Japan) was used to limit knee joint motion. Two weeks postoperatively, the external fixator was removed under sedation and the knee joint was allowed to move freely. Ampicillin (1000 mg/5 ml/body) was administered to prevent postoperative infection, and buprenorphine (0.01 mg/0.05 ml/kg) was given for pain relief for up to 4 days after surgery. At 1 or 3 months postoperatively, the mini-pigs were sedated with intramuscular injections of medetomidine hydrochloride (0.05 mg/0.05 ml/kg) and midazolam (0.5 mg/0.1 ml/kg), deeply anesthetized with thiamylal sodium (25 mg/1 ml/kg) and euthanized by exsanguination via the femoral artery. As a result, five implanted knees and five control knees (one month: n = 2, three months: n = 3, per group) were evaluated.

Fig. 1.

Surgical procedure. a: Cartilage constructs were cut into sheets and threaded through their edges. b: Osteochondral defects in the femoral trochlear groove of the OIDPs were created using a bone chisel. c: Cartilage constructs were fixed onto the defect using pull-out sutures. Scale bar: 5 mm. OIDPs, operational immunodeficient pigs.

2.5. Gross evaluations and histological and immunohistochemical assessments

The two groups at three months postoperatively (each group: n = 3) were compared using macroscopic and histological scoring systems. Gross findings were graded by two investigators using the International Cartilage Repair Society (ICRS) scores (Table 1). The dissected femoral groove was fixed with 10 % formalin (Wako Pure Chemical Industries). Longitudinal sections were cut parallel to the groove and embedded in paraffin, and 5-mm-thick sections were prepared. Tissues were stained with hematoxylin-eosin (HE; Merck, Darmstadt, Germany) and Safranin O and Fast Green (0.1 % Safranin-O; Nacalai Tesque Inc.; 0.05 % Fast Green: Wako) for morphological analyses. Type I collagen (Col I; Southern Biotech, Birmingham, AL, USA) and Col II (Southern Biotech) staining were performed to analyze collagen deposition. Human vimentin (anti-vimentin antibody [SP20] [ab16700]; Abcam, Cambridge, UK) staining was performed to assess whether the repaired tissue was derived from human iPSCs. Histological assessments were scored by two investigators using the modified O'Driscoll histological scoring system [28] (Table 2). Sections of mature cartilage constructs were prepared in the same manner and were histologically evaluated. In addition, the positive areas of glycosaminoglycans (GAGs), Col II, Col I, and human vimentin in regenerated tissue on osteochondral defects were measured using an analytical software program (BZ-H3A0; Keyence, Osaka, Japan).

Table 1.

The International Cartilage Repair Society score for the macroscopic evaluation.

Criteria	Points
Ⅰ. Degree of defect repair
Level with surrounding cartilage	4
75 % repair of defect depth	3
50 % repair of defect depth	2
25 % repair of defect depth	1
0 % repair of defect depth	0
Ⅱ. Integration to border zone
Complete integration with surrounding cartilage	4
Demarcating border <1 mm	3
3/4 of graft integrated, 1/4 with a notable border >1 mm width	2
1/2 of graft integrated with surrounding cartilage, 1/2 with a notable border >1 mm	1
From no contact to 1/4 of graft integrated with surrounding cartilage	0
Ⅲ. Macroscopic appearance
Intact smooth surface	4
Fibrillated surface	3
Small, scattered fissures or cracks	2
Several, small or few but large fissures	1
Total degeneration of grafted area	0
Ⅳ. Overall score
Grade Ⅰ: Normal	12
Grade Ⅱ: Nearly normal	11–8
Grade Ⅲ: Abnormal	7–4
Grade Ⅳ: Severely abnormal	3–1

Table 2.

The Modified O'Driscoll histologic scoring scale for the evaluation of cartilage repair [28].

Scoring scale	Points
Ⅰ. Percentage of hyaline articular cartilage
80–100 %	8
60–80 %	6
40–60 %	4
20–40 %	2
0–20 %	0
Ⅱ. Structural characteristics
A. Surface regularity
Smooth and intact	2
Fissures	1
Severe disruption, fibrillation	0
B. Structural integrity
Normal	2
Slight disruption, including cysts	1
Severe lack of integration	0
C. Thickness
100 % of normal adjacent cartilage	2
50–100 % of normal cartilage, or thicker than normal	1
0–50 % of normal cartilage	0
D. Bonding to adjacent cartilage
Bonded at both ends of graft	2
Bonded at one end/partially at both ends	1
Not bonded	0
Ⅲ. Freedom from cellular changes of degeneration
Normal cellularity, no clusters	2
Slight hypocellularity, <25 % chondrocyte clusters	1
Moderate hypocellularity/hypercellularity, >25 % clusters	0
Ⅳ. Freedom from degenerative changes in adjacent cartilage
Normal cellularity, no clusters, normal staining	3
Normal cellularity, mild clusters, moderate staining	2
Mild or moderate hypocellularity, slight staining	1
Severe hypocellularity, poor or no staining	0
Ⅴ. Reconstitution of subchondral bone
Complete reconstitution	2
Greater than 50 % reconstitution	1
50 % or less reconstitution	0
Ⅵ. Bonding of repair cartilage to de novo subchondral bone
Complete and uninterrupted	2
<100 % but >50 % complete	1
<50 % complete	0
Ⅶ. Safranin O staining
Greater than 80 % homogeneous positive stain	2
40–80 % homogeneous positive stain	1
Less than 40 % homogeneous positive stain	0
Total score	Max27

2.6. Mechanical assessments

The compressive strength was measured using a Digital Force Gauge (ZTA-500 N; IMADA, Aichi, Japan). The post-implanted cartilage (n = 5) 3 months after implantation and native cartilage (n = 5) were trimmed to cylindrical shapes using a biopsy trepan (diameter: 2.0 mm) (Kai Industries, Gifu, Japan). Most of the subchondral bone was excised and measured by using a flat compression attachment. Before implantation, cartilage constructs (n = 6) were formed into appropriate shapes and strengths tested in the same manner.

2.7. Statistical analyses

All data are expressed as mean ± standard deviation. The gross and histological scores, and the positive areas using an analytical software program were compared between the implanted and control groups using the unpaired Student's t-test. Comparisons among the three groups for mechanical assessments were performed using a one-way analysis of variance (ANOVA), followed by Bonferroni's multiple comparison test. Statistical significance was set at P < 0.05.

3. Results

3.1. Fabrication of constructs

Using a bio-3D printer, spheroids were stacked on needle arrays and matured in a reactor with a reflux function for 30–76 days, resulting in the formation of tubular constructs 7 mm in diameter and 2 cm in length (Fig. 2a–c). After removal from the needle array, the tube shape was maintained by fusion of the spheroids; it was elastic and could be easily grasped without breaking (Supplementary Video 2). Histologically, numerous chondrocytes were observed, and the surrounding ECM area produced abundant GAGs, which stained with safranin O and collagen fibers that were positive for both Col I and Col II. Human-specific anti-vimentin antibody staining was strongly positive, in accordance with the cellular localization (Fig. 2d–g).

Fig. 2.

Bio-3D printed scaffold-free cartilage constructs derived from human iPSC-MSCs using a bio-3D printer, Regenova. a: Gross image of cartilage constructs after printing. b, c: Gross image of cartilage constructs after maturation. Histological evaluations of the cartilage constructs after maturation following staining with (d) SOFG, (e) Col I, (f) Col II, and (g) h-vimentin. Solid squares in the lower column were magnified from the upper column. Scale bar: 1 mm in the upper column of d-g, 200 μm in the lower column of d-g. iPSC-MSCs, induced pluripotent stem cell-derived mesenchymal stem/stromal cells; SOFG, safranin O and fast green; Col Ⅰ, type Ⅰ collagen; Col II, type II collagen; h-vimentin, human vimentin. Images are representative of each experiment.

Results of implanting the cartilage constructs on osteochondral defects in immunodeficient pigs.

3.2. Gross assessments

At one month in the implanted group, the osteochondral defect had become grossly filled with regenerated tissue and more than half of the depth had regenerated. White regenerated tissue was observed in the center of the defect; the margins of the defect were repaired with reddish tissue, and most of the borders were integrated (Fig. 3a). Three months after implantation, the entire defect was almost completely filled with white regenerated tissue and the borders were mostly integrated and obscured (Fig. 3b). In the control group at one month postoperatively, the defect was partially covered by white fibrous tissue, but the degree of defect repair was poorer than that in the implanted group, and approximately half of the edge had not been integrated (Fig. 3c). In the control group at three months, the border was generally clear, although it was almost entirely filled with white repaired tissue (Fig. 3d). In all groups, the macroscopic appearance revealed small cracks and fissures on the surface of repaired tissue. There was no significant difference in the ICRS scores between the two groups at three months postoperatively (Supplementary Fig. 1).

Fig. 3.

Gross images after implantation of human iPSC-MSC-derived cartilage constructs into osteochondral defects of OIDPs. a: Implanted group at 1 month. b: Implanted group at 3 months. c: Control group at 1 month. d: Control group at 3 months. Scale bar: 5 mm. Images are representative of each experiment.

3.3. Histological and immunohistochemical assessments

One month postoperatively in the implanted group, abundant chondrocytes were identified by HE staining and further stained with human vimentin, which showed staining findings in accordance with cell localization, indicating that the chondrocytes were derived from human iPSC-MSCs. The ECM surrounding the cells was mostly covered by thick GAG-rich AC-like tissue, and abundant Col I- and Col II-positive collagen fibers were observed (Fig. 4a). In addition, three months after implantation, an extensive layer of cartilage positively stained with human vimentin and safranin O was observed. Collagen fibers were positive for both Col I and Col II (Fig. 4b). In the control group, one month postoperatively, most of the repaired tissue was Col I-positive and covered by a thick fibrous tissue layer without GAGs, and the cells were vimentin-negative (Fig. 4c). At three months postoperatively, the majority of the cartilage was covered with fibrous tissue, although a small internal layer of cartilage was positively stained with safranin O and Col II (Fig. 4d). Chondrocytes in the implanted group were arranged in an orderly manner perpendicular to the articular surface and in layers, similar to the normal AC, and were tightly connected to the host subchondral bone in deeper layers, creating a solid structure (Supplementary Fig. 2). The implanted groups had a significantly greater area that was positively stained with safranin O (p = 0.02) and significantly higher total scores than the control group (p = 0.04) (Fig. 5a). In addition, the positive areas of GAGs staining images were significantly greater in the implant group than in the control group (p = 0.03). The positive areas in Col II and human vimentin staining images tended to be greater in the implant group than in the control group, but the difference was not statistically significant (Fig. 5b).

Fig. 4.

Histopathology of cartilage constructs derived from human iPSC-MSCs after implantation to osteochondral defects of OIDPs. a: Implanted group at 1 month. b: Implanted group at 3 months. c: Control group at 1 month. d: Control group at 3 months. Histological evaluations of cartilage constructs stained with HE, SOFG, Col I, Col II, and h-vimentin. Black triangles indicate the implanted area or the osteochondral defect. Solid squares in the right column are magnified from the left column. Scale bar: 1000 μm in the left column, 500 μm in the right column. HE, hematoxylin and eosin. Images are representative of each experiment.

Fig. 5.

a: The Modified O'Driscoll histologic scoring scale of the implanted and control groups at three months postoperatively (each group: n = 3). The scores for each characteristic of the two groups are the mean of the scores assigned by the two investigators. b: A quantitative analysis of the positive area of Safranin O, Col II, Col I and human vimentin in regenerated tissue at the site of the osteochondral defect of the two groups at three months postoperatively (each group: n = 3). ∗ indicates P < 0.05 by the unpaired Student's t-test.

3.4. Mechanical assessments

The compressive strength was 42.37 ± 14.39 MPa for the postimplanted cartilage and 63.65 ± 25.4 MPa for the native AC. The strength of cartilage constructs before implantation was 0.18 ± 0.05 MPa. The strength of the post-implanted AC was significantly higher than that of the constructs (P < 0.01) (Fig. 6).

Fig. 6.

Mechanical assessments. Compression rupture strength tests were performed on cartilage constructs (n = 6), post-implanted cartilage (n = 5), and native knee joint cartilage of mini-pigs (n = 5) after molding to a diameter of 2 mm via a trepan biopsy. ∗∗ indicates P < 0.01 by a one-way analysis of variance followed by Bonferroni's multiple comparison test.

4. Discussion

In the present study, patch-formed cartilage constructs were implanted into extensive osteochondral defects created in the femoral trochlear groove of mini-pigs. Implanted human iPSC-derived cartilage constructs matured without rejection in immunodeficient mini-pigs, and AC-like tissue was confirmed to be donor-derived by human-specific anti-vimentin antibody staining. Histological examination of the constructs showed many chondrocytes and a GAG-rich ECM, indicating that the cartilage properties were maintained and matured after implantation. The transplant group had higher histological cartilage repair scores and larger positive areas of GAGs than the control group (Fig. 5a and b). The scaffold-free cartilage constructs were strong enough to endure handling during the surgery. Furthermore, the implanted constructs were significantly stronger than before implantation, and their compressive strength increased to a value approximating that of normal cartilage within three months, indicating that maturation also proceeds in vivo (Fig. 6). Although several cartilage regeneration studies are currently underway, the extent of cartilage regeneration remains limited [29,30]. Autologous chondrocyte implantation (ACI), which is indicated for traumatic cartilage injuries with relatively large defects, involves harvesting and culturing chondrocytes from normal cartilage and seeding them onto a scaffold material. ACI requires the sacrifice of cartilage tissue and two surgeries, and dedifferentiation may occur in expanded cultures in addition to marked hypertrophy and ossification of the graft [[31], [32]].

To create cartilage constructs using bio-3D printing, it is necessary to select appropriate cells. In chondrogenic differentiation, iPSC-MSCs have higher expression of SOX9, COL2A1, and Aggrecan and lower expression of hypertrophic markers, such as COL10A1 and RUNX2, than human bone marrow-derived mesenchymal stem/stromal cells (BM-MSCs), indicating that they are a better source for AC regeneration than human somatic stem cells [33]. Cartilage derived from primary chondrocytes and BM-MSCs produces less ECM than iPSC-derived cartilage [16,34]. Somatic stem cells, also called MSCs, have a multipotent differentiation potential; however, several issues prevent their widespread use. For example, they must be harvested from healthy tissues, the differentiated cartilage tends to be hypertrophic, and their proliferative differentiation potential decreases with age [35,36]. In addition, MSCs suffer from chromosomal aberrations and deterioration after repeated passages, making it difficult to obtain a large number of high-quality cells for clinical use in expanded culture [37]. However, human iNCCs can be cryostocked [38] and differentiate into iPSC-MSCs to produce large quantities of cartilage spheroids of uniform quality and size, making them superior for bio-3D printing [16]. There was a previous report in which iPSC-derived cartilage spheroids were produced not through iNCCs; microlevel size control and mass production may be difficult to achieve using the protocols [39].

Although iPSCs have been reported to be associated with a risk of tumorigenesis owing to their pluripotency [[40], [41], [42]], the protocol for inducing chondrogenesis from iPSC-MSCs derived from iNCCs showed excellent chondrogenic differentiation without tumor formation or upregulation of pluripotency markers [27,38]. Cartilage-like tissue using iPSC-MSCs through iNCCs showed no teratoma formation at the transplantation site [43]. In the present study, cartilage constructs fabricated following the protocol showed no malignant cells, and a Col II-positive GAG-rich ECM was observed around abundant chondrocytes. The constructs in the present study were positive for both Col I and Col II. During the remodeling process associated with the maturation of cartilage ECM, Col I gradually disappears from the ECM and is replaced by Col II [[16], [39], [44]]. On the other hand, further long-term evaluations are needed to determine whether there is a trend toward hypertrophy, as with other MSCs. However, there was also no evidence of surface overgrowth or expression of type X collagen in the implanted area of the MSC-based scaffold-free construct [45], suggesting that MSC-based constructs did not lead to hypertrophy.

In previous studies, undifferentiated MSCs were implanted into osteochondral defects and regenerated hyaline cartilage [20,45]. In the present cartilage constructs, it was expected that poorly differentiated constructs would similarly differentiate into hyaline cartilage after implantation; however, this was not the case. Cells that did not differentiate into chondrocytes prior to implantation may not have been undifferentiated iPSC-MSCs, but instead may have differentiated into other cells or reached cellular senescence (Supplementary Fig. 3a–c). In addition, because tissue-engineered constructs seeded with undifferentiated iPSC-MSCs do not differentiate into chondrocytes when implanted [27], it may be necessary to differentiate iPSC-MSCs into cartilage before implantation. After differentiating iPSC-MSCs into cartilage, we fabricated 3D cartilage constructs using a Kenzan bio-3D printer in the present study. An attempt has been made to refine the currently available protocols for producing high-quality chondrogenic spheroids. Thienoindazole derivative (TD-198946)-treated iNCC-MSCs showed upregulation of SOX9, ACAN, and COL2A1 as well as increased GAG production [46]. This new protocol may lead to the fabrication of high-quality cartilage constructs using a bio-3D printer.

We need to improve not only the construct fabrication, but also the surgical protocols, including the suture technique, external fixation, and duration of unloading, to achieve successful implantation of cartilage constructs. In some cases, there are several possible reasons why a part of the defect was repaired with fibrocartilage (Fig. 4b). First, the mechanical load on the construct was considered to be extreme because the defect was created over almost the entire articular surface. Part of the edge could not withstand the friction with the patella and fell off; therefore, such areas had to be repaired with fibrocartilage. However, the position of the suture did not change before or after transplantation, so it may not have been detached due to friction (Supplementary Fig. 4a–e). Second, a hematoma formed between the subchondral bone and the construct due to oozing of the bone, which may have thus inhibited adhesion. Furthermore, bleeding from the bone marrow may have promoted fibrocartilage repair, as in the case of bone marrow stimulation therapy. Finally, we can consider the possibility that the construct itself has a low degree of adhesiveness. However, the transplanted cartilage constructs were tightly connected to the host subchondral bone after only one month (Supplementary Fig. 2). This may be an important indicator of when to start weight bearing after transplantation. Other fixation methods that can withstand large mechanical loads are required to fix constructs to extensive defects. Collagen patches are used in ACI surgery to prevent transplanted tissue from falling off [31]. Other methods, such as sutured anchors and cartilage constructs with pegs (Supplementary Fig. 5), may also be useful for improving the fixation strength. In addition, the edges of the implanted construct, which had an inappropriate thickness, were shaved by friction with the contralateral patella (Supplementary Fig. 4a and b). However, the appropriate cartilage construct regenerated the defects without shaving (Supplementary Fig. 4c–e). The modified protocol is expected to enable the production of high-quality cartilage constructs of controllable size [46].

Bio-3D printed scaffold-free cartilage constructs derived from human iPSC-MSCs show cartilage regeneration in extensive osteochondral defects. Therefore, cartilage constructs may be suitable for the extensive regeneration of AC. However, the present study has several limitations that should be mentioned, including the disregard for the effects of lineage and sex. One of the limitations is that we were only able to follow up for three months after transplantation. Most studies on cartilage regeneration have followed up for more than one year after transplantation and reported that hyperplasia of the transplanted tissue and degradation of the regenerated cartilage occurred after one year [31]. Since xenotransplantation causes severe rejection immediately after transplantation, either genetically engineered immunodeficient models or drug immunosuppression are essential. Immunodeficient pigs have been already developed, but mature immunodeficient pigs are difficult to obtain. Conventional drug-induced immunodeficient pigs have a limit at which human cells can be confirmed as viable after transplantation. In fact, human iPSC-derived cartilage has only been followed for only one month by using the conventional immunodeficient pigs [39]. In contrast, by using the OIDP model created by removing the thymus and spleen and administering immunosuppressive drugs, human dermal fibroblast-derived blood vessels have been shown to be stable for three months [26]. However, the OIDP weakens after three months, making it difficult to obtain data beyond six months. It is therefore remarkable that in this study we were able to confirm that the large-scale xenograft of the human iPSC-derived cartilage constructs engrafted and survived for three months in the porcine model. In the future, histological and gene expression analyses of cells after implantation as well as functional evaluations by imaging, such as fluorography, should be conducted over a longer period. However, we will have to wait for the development of new immunodeficient pigs to obtain longer-term data.

It was not easy to supply the medium evenly throughout the large constructs, so it was beneficial for the maturation of the construct that the tubular shape allowed the medium to reflux into the lumen of the cartilage construct with the shape. The tube-shaped construct was cut along its longitudinal axis to create a cartilage tissue with an expanded surface area. This process results in a naturally curved shape that ideally fits the typical concave shape of the groove, which is designed to accommodate the patella. The contact pressure in the patellofemoral (PF) joint increases from 0° to 90° of knee flexion, and the force during walking can be 0.8 times the body weight [47]. We confirmed that cartilage constructs can be grafted to the osteochondral defect in the PF joint even under knee flexion during gait.

The Kenzan bio-3D printing method can be used to fabricate constructs with complex shapes using CAD data of articular surface shapes [16]. In addition, it can be applied to create a construct as large as the femorotibial joint surface by docking multiple Kenzans. If the fabrication protocol and bioreactor system for media penetration could be further improved to produce high-strength cartilage constructs with articular surface shapes, a surface-replacing cellular joint implant could be realized (Supplementary Fig. 5, Supplementary Video 3).

5. Conclusion

We fabricated cartilage constructs from human iPSC-derived cells without scaffolds using a bio-3D printer and implanted them into extensive articular cartilage defects in immunodeficient pigs. The transplanted cartilage constructs were successfully engrafted into the host and regenerated the articular cartilage defects.

Data availability

The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request.

Author contributions

T.N. performed the experiments, analyzed the data, and drafted the manuscript.

A.N. performed the experiments and analyzed the data.

D.M. implanted the cartilage constructs, interpreted the data, and edited the manuscript.

H.Y. implanted the constructs and performed gross and histological assessments.

S.K. implanted the constructs.

Y.N. provided advice on the manuscript.

D.Z., C.Z., Y.I.,M. Ikeya. and J.T. provided iNCCs derived from human iPSC line 414C2.

M. Itoh. developed the operational immunodeficient pig model.

M.M. provided advice on the research design and interpreted the data.

K.N. developed the bio-3D printer system, designed and supervised the experiments, interpreted the data, and edited the manuscript.

Declaration of competing interest

K.N. is a co-founder and shareholder of Cyfuse Biomedical KK, chief technical officer of Arktus Therapeutics, and an inventor/developer designated on the patent for the bio-3D printer. Patent title: Method for Production of Three-Dimensional Structure of Cells; patent number: JP4517125. Patent title: Cell structure production device; patent number: JP5896104. The other authors declare no competing interests regarding the publication of this article.

Appendix A. Supplementary data

The following are the Supplementary data to this article.

figs1.

figs2.

figs3.

figs4.

figs5.