Viable Vitreous Grafts of Whole Porcine Menisci for Transplant in the Knee and Temporomandibular Joints

Abstract

The shortage of suitable donor meniscus grafts from the knee and temporomandibular joint (TMJ) impedes treatments for millions of patients. Vitrification offers a promising solution by transitioning these tissues into a vitreous state at cryogenic temperatures, protecting them from ice crystal damage using high concentrations of cryoprotectant agents (CPAs). However, vitrification’s success is hindered for larger tissues (>3 ml) due to challenges in CPA penetration. Dense avascular meniscus tissues require extended CPA exposure for adequate penetration; however, prolonged exposure becomes cytotoxic. Balancing penetration and reducing cell toxicity is required. To overcome this hurdle, a simulation-based optimization approach was developed by combining computational modeling with microcomputed tomography (μCT) imaging to predict three-dimensional CPA distributions within tissues over time accurately. This approach minimized CPA exposure time, resulting in 85% viability in 4-ml meniscal specimens, 70% in 10-ml whole knee menisci, and 85% in 15-ml whole TMJ menisci (i.e., TMJ disc) post-vitrification, outperforming slow-freezing methods (20%−40%). The extracellular matrix (ECM) structure and biomechanical strength of vitreous tissues remained largely intact. Vitreous meniscus grafts demonstrated clinical-level viability (≥70%), closely resembling the material properties of native tissues, with long-term availability for transplantation. The enhanced vitrification technology opens new possibilities for other avascular grafts.

Graphical Abstract

To address the shortage of large avascular fibrocartilaginous tissues, vitrification holds potential but faces challenges in balancing between cryoprotectant agent penetration and cytotoxicity. A simulation-based optimization approach has produced viable whole meniscus grafts from the knee and temporomandibular joints. These grafts, with clinic-level viability (≥70%), structural integrity, and optimal mechanical function, offer excellent transplantation options due to their sustained availability.

1. Introduction

Musculoskeletal tissue transplantation plays a pivotal role in orthopedic and reconstructive surgeries, especially in preserving the whole tissue structure and function, which significantly enhances the quality of life for patients. However, the limited availability of suitable tissue replacements presents substantial challenges in advancing surgery technology and achieving favorable post-transplantation clinic outcomes. This challenge is particularly acute in the context of interpositional fibrocartilaginous meniscus replacement in the knee joint (knee meniscus) and temporomandibular joint (also known as the TMJ disc).

In the United States, more than one million patients with meniscus injury undergo surgical repair or meniscectomy each year ^[1]. Currently, meniscal allograft transplantation (MAT) involving human cadaver tissue is typically indicated for patients with symptomatic subtotal or total meniscectomy knees. Despite MAT being performed worldwide and yielding favorable long-term outcomes, its widespread clinical adoption has faced significant constraints ^[1–2]. During the period from 2007 to 2011, there were 3,295 MATs performed in the US ^[2b]. This limitation primarily stems from challenges associated with a shortage of available donor meniscal allografts ^[3]. Consequently, this situation significantly increases the incidence of the risk of joint issues later in life by elevating joint contact stress ^[4]. A similar striking issue is found in patients suffering from TMJ disorders (TMD). Over 10 million TMD patients experience TMJ disc degeneration and mechanical dysfunction, leading to surgery in 10% of these cases ^[5]. Although disc discectomy has been used in clinical practice for decades, it often yields unsatisfactory treatment outcomes, such as severe osteoarthritis and joint ankylosis ^[6]. Surprisingly, the potential of using a TMJ disc allograft from a donor has remained largely unexplored due to its limited availability, despite promising results emerging from animal studies ^[7]. This lack of research on donated TMJ disc allografts has forced the use of replacement autografts, such as a temporalis muscle fascia flap and abdominal fat autograft, ^[8] or alloplastic total joint replacement ^[9]. This reliance on autografts results in pain and functional loss, additional surgical sites, prolonged recovery times and elevated risk of surgical site infections ^[10], while alloplastic total joint replacement has shown disadvantages for long-term use, particularly in children, as they may hinder the normal growth of the facial skeleton and require replacement upon reaching facial maturity ^{[10a, 11]}. Therefore, there is an urgent need to address the shortage of donors and provide off-the-shelf availability of grafts that possess natural properties closely resembling those of the native tissues.

To address this challenge, fresh-frozen and cryopreserved allografts are currently used as valuable alternatives for transplantation when fresh donor grafts are unavailable. These grafts provide essential support for patients with immediate needs. However, achieving favorable long-term clinical outcomes still presents challenges due to limitations in preserving viable cells. The formation and growth of ice crystal formation during fresh-freezing and cryopreservation (i.e. slow-freezing) processes result in substantial cell loss^[12]. Within a healthy meniscus, fibrochondrocytes play a vital role in synthesizing and maintaining the extracellular matrix, which is responsible for its distinctive material and structural properties ^[13]. Clinical outcomes have further demonstrated that MAT using fresh meniscus allografts yields superior pain relief, enhanced function and a lower failure rate compared to MAT using frozen grafts ^[14]. In some animal transplantation studies involving conventionally cryopreserved meniscal allografts, while cryopreservation did not appear to affect the morphological appearance or biomechanical properties of menisci, biosynthetic activity was decreased to less than 50% of normal control values, with only 10% of transplanted meniscal cells displaying metabolic activity ^[15]. Additionally, as another vital tissue in the effort to maintain knee joint health and function, alongside the meniscus, donor articular cartilage necessitates the presence of viable chondrocytes, as demonstrated by Cook’s research group^[16]. Their findings indicate that viable donor chondrocytes in osteochondral allografts (with viability exceeding 70% at the time of transplantation) primarily contribute to preserving the health of donor articular cartilage over the long term. Therefore, to maximize the chances of long-term allograft survival, it is essential to manipulate storage conditions to minimize fibrochondrocyte death. Innovative preservation methods for viable allografts are needed to enhance long-term clinical outcomes post transplantation.

In this study, we have pioneered the development of vitreous grafts for knee meniscus and TMJ disc replacement. These grafts are created by transitioning donor tissues into an amorphous solid vitreous state during cooling, effectively preventing ice crystal formation at cryogenic temperatures, with the aid of high concentration cryoprotective agents (CPAs). This process, also known as vitrification ^[17], has been demonstrated in small volumes (≤ 3 ml) of vascularized tissues with thin and permeable extracellular matrix (ECM) structure, such as heart valve tissue samples ^[18] and rabbit vein rings ^[19], with the aid of a 55% CPA solution (VS55), licensed from the American Red Cross ^[20]. Scaling up to larger-sized tissues (≥3 mL) poses a formidable challenge in effectively regulating CPA loading periods, as it requires achieving adequate penetration within a limited exposure time, driven by the need to minimize CPA toxicity ^{[18–19, 21]}. This task becomes even more formidable when dealing with avascular fibrocartilaginous tissues like knee menisci and TMJ discs. In contrast to vascularized tissues with thin and permeable tissue structures that facilitate CPA distribution through diffusion and convection, meniscus and TMJ disc tissues are dense and thick with restrictive permeability. These tissues depend exclusively on diffusion to achieve adequate CPA penetration, necessitating prolonged exposure times to enable deep CPA diffusion ^[22]. However, prolonging exposure times carries the risk of increased CPA-induced cytotoxicity ^[21]. Hence, optimizing CPA diffusion by extending the diffusion time within permissible limits, while minimizing toxicity, becomes of paramount importance. While vitrification holds promise in addressing the tissue shortage, its feasibility in maintaining a viable knee meniscus or TMJ disc was yet to be explored. Therefore, in this study, we conducted an extensive investigation into CPA diffusion in avascular fibrocartilaginous tissues to optimize CPA loading time while minimizing cytotoxicity, employing a computational simulation approach. Our hypothesis was that knee menisci and TMJ discs loaded with adequate and toxically tolerable CPAs can achieve clinical-level viability (≥70%), maintain integral ECM structure, and preserve biomechanical functions through vitrification, when compared to fresh tissues. Our objectives were threefold: 1) to prove the concept that enhancing CPA penetration within tissues by extending the CPA loading period is a crucial step for achieving optimal outcomes (illustrated in Figure 1A and B); 2) to develop a computational model combining micro-computed tomography (μCT) imaging to delve into the intricate mechanism of CPA diffusion within meniscal tissues (Figure 1C); and 3) to expand our findings to encompass whole knee menisci and TMJ discs (Figure 1D–G). The leap in cell viability and ECM material properties was achieved for whole knee menisci and TMJ discs grafts, which was not feasible with previous approaches. The vitrification approaches developed in this study can be further applied to design CPA loading protocols for other dense avascular tissue grafts, offering a promising solution to the shortage of donor tissues in transplantation.

Figure 1. Schematic workflow of the study.

(A) Meniscal specimens harvested from the central region of a meniscus. (B) Viability of vitrified samples was improved by extending CPA loading time. (C) A simulation model was established to investigate CPA diffusion kinetics within meniscal tissues. (D) Using the established model, the penetration of CPAs throughout the whole meniscus and TMJ disc was predicted and used for vitrification (E). After convectional warming (F), the whole knee meniscus including anterior horn (A), central (C), and posterior horn (P), and the whole TMJ disc including anterior (A), central (C), posterior (P), medial (M), and lateral (L), were investigated for viability, metabolic activity, histological analysis, and mechanical properties (G).

2. Results

2.1. Viability requires adequate CPA diffusion in vitrified meniscal samples

To investigate the relationship between diffused CPA (VS55) concentration and tissue viability following vitrification, meniscal specimens (5 mm diameter) of two different thicknesses were used: 1 mm and 3.5 mm. Initially, each specimen was loaded with VS55 in six steps, for a total duration of 1.5 hours, and vitrified in a 20-ml glass vial containing 4-ml VS55 supplemented with 0.3M sucrose and 0.3M trehalose (VS55/S/T). Successful vitrification was achieved by cooling at 26.5±0.6°C/min to −100°C, followed by slower cooling to −120°C at 3.7±0.4°C/min (Figure S1, Supporting Information). This process resulted in the formation of a transparent glassy state around the sample (Figure S1A, Supporting Information). During the conventional warming, samples were rapidly warmed to −35°C at 62.1±4.3°C/min in a 37°C water bath. The vitrified 1 mm thick tissues, serving as a smaller volume control group, exhibited preserved living cells comparable to fresh samples, whereas the slow-frozen tissues showed a severe reduction in cell viability (Figure 2A). However, as the thickness increased to 3.5 mm, both vitrification and slow freezing resulted in a decline in cell viability. Nonetheless, when the loading time was extended to 3 hours, the vitrified 3.5 mm thick tissues maintained living cells throughout their full thickness. These findings suggest that viability of vitrified samples depends highly on the concentrations of diffused CPAs. When an adequate concentration of CPAs is diffused throughout the samples, vitrification proves superior to slow freezing in preserving living cells. These results emphasize the critical role of CPA diffusion in achieving optimal tissue viability during vitrification. Quantification analysis of cell viability for the 3.5 mm thick tissues (Figure 2B) revealed no adverse toxic effects from the 3-hour VS55 exposure prior to vitrification. After vitrification, a viability rate of 85.8±5.5% was achieved, whereas slow freezing preserved only 23.9±4.8% of living cells, and 1.5-hour vitrified samples exhibited a viability of 25.1±7.7%. The metabolic activity of the tissues from day 0 to day 4 after warming improved as shown in Figure 2C. Specifically, the 1 mm vitrified samples following 1.5-hour VS55 loading, and the 3.5 mm vitrified samples following 3-hour VS55 loading, exhibited a continuous recovery in metabolic activity over time, ultimately recovering completely by day 4 post-warming. This phenomenon was not observed in slow-frozen samples or 3.5 mm vitrified samples after 1.5-hour VS55 loading.

Figure 2. Meniscal viability and metabolic activity assessed by fluorescence live/dead imaging and alamarBlue assays.

(A) Fresh, slow-frozen (SF), and vitrified meniscal specimens (1 mm and 3.5 mm thick). Each 3.5 mm specimen’s sagittal section is represented through three images: top, middle, and bottom. The groups include fresh, SF, vitrified after the 1.5-hour VS55 loading (1.5 Hr Vit), 3-hour VS55 loaded and unloaded without vitrification (3Hr CPA), and vitrified after the 3-hour VS55 loading (3 Hr Vit). Scale bars represent 200 μm. (B) Cell viability quantified in the middle images of 3.5 mm specimens from fresh (N=n =5) and preserved groups at day 0 post-warming including SF (N=n=4), 1.5 Hr Vit (N=n=6), 3 Hr CPA (N=n=3), and 3 Hr Vit (N=n=4). p-value was determined with a two-sided t-test. (C) Metabolic activity in 1 mm and 3.5 mm thick vitrified tissues compared to SF 3.5 mm thick tissue and the fresh 3.5 mm thick control group, over 0 to 4 days post-warming. Results were normalized to respective specimens (N=3–4; n=6–12) before CPA loading. Results from fresh control samples were normalized to their average values on day 0 (N=3; n=8). A black dashed line at 100% recovery indicates cell metabolic activity of fresh meniscal specimens. N = number of independent menisci; n = number of specimens. p-value determined via two-way ANOVA. Data are mean ± SD.

2.2. μCT-based simulation predicts CPA distribution in menisci and TMJ discs

To explore in-depth studies on CPA diffusion, VS55 concentration and distribution within the 3.5 mm thick meniscal specimens were determined using μCT imaging (Figure 3). The calibration curve of meniscal specimens equilibrated in different VS55 concentrations is shown in Figure 3A. The curve displayed a strong linear correlation (R²>0.99) between Hounsfield Unit (HU) values and the corresponding CPA concentrations, providing a basis for determining CPA concentration within the tissue. Pseudocolor images of CPA-equilibrated meniscal tissues exhibited uniform HU values throughout the meniscal tissues (Figure 3B).

Figure 3. CPA concentration profiles within meniscal specimens using a μCT imaging-based simulation method.

(A) Scanned transverse section (green area in the insert figure) at half thickness with the calibration curve for meniscal specimens equilibrated in different VS55 solutions (N=3, n=12 for samples equilibrated in 0% VS55; N=n=3 for other conditions). (B) Pseudocolor μCT images of meniscal tissues equilibrated in different concentrations of VS55 solutions. (C) A pseudocolor μCT image (left) alongside a computational image (right) of a 3.496 mm thick VS55-loaded sample. (D) Concentration profiles of the same VS55-loaded sample using both computational modeling and the μCT curve, with a best-fit diffusivity value of 4.6×10⁻¹⁰ m²/s (R²=0.9696). (E) Diffusivity values of VS55 at 0°C (N=3; n=9) and 22°C (N=3; n=6) with a two-sided t-test for significance. Data are mean±SD. Panels (F) and (G) depict VS55 distribution after loading samples of varying thicknesses and loading times at 0°C. Solid lines represent experimental conditions for viability assessment; dashed lines represent simulated conditions. (H) Representative validation of VS55 distribution in a 3.443 thick meniscal specimen after a 1-hour stepwise loading period (10 min each step) at 0°C using modeling (D=3.82×10⁻¹⁰ m²/s). N = number of independent menisci; n = number of specimens.

Based on the linear relationship between HU values and CPA concentrations, the CPA concentration in meniscal tissues after a 1.5-hour loading period can be determined from μCT data. The pseudocolor and the computational modeling images of a VS55-loaded sample (Figure 3C) showed higher CPA concentration at the edges and lower concentration in the center area of the sample. The concentration profiles obtained from the computational model using the best-fit diffusivity to the μCT data (Figure 3D) were consistent with the pseudocolor μCT images. Through this approach, the diffusivities of VS55 in meniscal specimens were determined at two different temperatures: 0°C (CPA loading temperature) and room temperature 22°C (as a control group). The results indicated that CPA molecules diffused significantly faster at 22°C than at 0°C (Figure 3E). The CPA concentration profiles of tissues of different thicknesses (1–3.5 mm) within a total exposure time of 1.5 hours, as well as for 3.5 mm thick tissues with various total exposure times (1.5–7 hours), were simulated using the mean VS55 diffusivity value at 0°C (3.82 × 10⁻¹⁰ m²/s) (Figure 3F and G). The simulation curves show that a 1.5-hour CPA loading period resulted in diffusion of over 95% CPA in 1 mm thick tissues at 0°C. However, in the case of 3.5 mm thick tissues, only 50% of CPAs were able to penetrate the center area within the same loading time. To achieve a >80% CPA concentration of 3.5 mm thick tissues, a longer exposure time of 3 hours was required, suggesting that higher concentrations of CPAs improved viability of vitrified tissues.

To validate our computational model, we compared the CPA distribution in 3.5 mm thick meniscal specimens after a 1-hour VS55 stepwise loading protocol with the model-predicted values. The experimental data aligned well with the modeled curve, as demonstrated by the example shown in Figure 3H, with an associated R² value of 0.9061.

Leveraging this model, we effectively incorporated diffusion principles into the larger-sized geometries of the whole meniscus and TMJ disc. This was achieved through simulation of CPA distribution across the entire structure, as shown in Figure 4. To predict CPA concentration in the whole meniscus after a specified loading period, we extracted a transverse section at half thickness and three sagittal sections from distinct regions. Both the modeling and μCT data-derived pseudocolor images of the transverse section (Figure 4A) exhibited a similar CPA concentration distribution, affirming the agreement between them. Concentration profiles from the horn regions (black lines on the transverse sections in Figure 4A) further confirmed the strong correspondence between the modeling and experimental data, with an R²=0.9562 for the anterior and 0.8161 for the posterior region (Figure 4B). This validation reinforces the accuracy of our CPA diffusion model for the whole meniscus. Additionally, validation of the CPA concentration profiles was also observed in the sagittal sections from different regions (Figure 4C). It was observed that a loading period of 2 hours did not result in a penetration of VS55 beyond 50%, whereas the lowest VS55 concentration across the three regions was found to range between 50–75% after a 3-hour loading. Based on the findings of our study, it was observed that a 3-hour CPA loading period did not have adverse effects on tissue viability prior to vitrification (Figure 2). Therefore, in subsequent experiments, the 2-hour and 3-hour loading protocols were employed for vitrification of whole menisci.

Figure 4. Simulation and μCT-validation of CPA distribution in the whole meniscus and TMJ disc.

(A) Simulation for CPA concentration through a 3D whole meniscus model after 3-hour VS55 loading (left); its transverse sections at half thickness from the model (middle) and experimental μCT data (right). Concentration profiles (black lines) on both transverse and sagittal sections of the anterior (A), central (C), and posterior (P) horns are shown in (B) and (C), respectively. Simulated profiles after 120-min (red), 180-min (blue), and 240-min (green) loading periods using modeling (D=3.82×10⁻¹⁰ m²/s). Experimental μCT data (black circle with line) after a 180-min CPA loading align well with simulation data. The x-axis represents the distance normalized by its actual length. (D) Simulation for CPA concentration through a 3D whole TMJ disc model after 3-hour VS55 loading (left). Red dashed area in the inferior view (left top) represents the TMJ disc: anterior (A), central (C), posterior (P), medial (M), and lateral (L). Sagittal sections of three bands (A-M-P, A-C-P, and A-L-P) represent CPA concentrations with average values of 82.22%, 82.20% and 82.29% respectively (right). (E) CPA concentrations from μCT data across TMJ discs (N=3, n=9 for each region; N=3, n=45 for whole disc average) after 3-hour loading. N = number of independent discs; n = number of measurements. No significance in different regions observed using a one-way ANOVA. Data are mean ± SD.

An identical simulation approach was employed for the TMJ discs. The distribution of VS55 concentration across the TMJ disc following the 3-hour loading period from both the computational model and the experimental μCT data is illustrated in Figure 4D and E. Linear calibration curves for TMJ discs equilibrated in different concentrations of VS55 solutions are shown in Fig S2. Given the similar structure between a knee meniscus and TMJ disc, we utilized the same mean VS55 diffusivity value (3.82×10⁻¹⁰ m²/s) to predict its concentration distribution within TMJ discs. Our model demonstrates a close alignment with experimental μCT data for the distribution of approximately 82±15% VS55 throughout the sagittal sections of anterior-posterior bands in the TMJ disc (Figure 4D). The experimental μCT data reveals an average concentration ranging from 70.6% to 78.6% across five regions, as shown in Figure 4E. No significant difference in VS55 concentrations was observed from μCT data across five distinct regions after the 3-hour loading.

2.3. Loading 3 hours enhances viability in vitrified menisci and TMJ discs

For upscaling to 10-ml whole meniscus vitrification, 2-hour and 3-hour VS55 loading protocols were utilized. Following CPA loading, each meniscus placed in a 50-ml centrifuge tube containing 10-ml VS55/S/T was successfully vitrified by cooling at 12.3±0.9 °C/min until −100°C, followed by a slower cooling to −120°C at 4.2±0.5 °C/min (Figure S1, Supporting Information). This resulted in the formation of a transparent, glass-like state encompassing the sample (Figure S1A, Supporting Information). For the conventional warming phase, samples were initially warmed slowly to −100°C at 18.6±1.8 °C/min, followed by a faster warming to −35°C at 53.9±4.4 °C/min in a 37°C water bath (Figure S1, Supporting Information). Viability assessment of vitrified whole menisci was conducted using fluorescence live/dead staining and metabolic activity analysis, with fresh and slow-frozen samples included for comparison. To comprehensively evaluate the distribution of living and dead cells, transverse sections at half thickness of the whole menisci were cut and imaged, enabling observation of all three regions with the lowest CPA concentration (Figure 5A). As expected, the 3-hour vitrification method exhibited superior maintenance of living cells across all regions compared to slow freezing. The 2-hour vitrified menisci retained only a slim layer of viable cells in the inner region.

Figure 5. Viability and metabolic activity of vitrified vs fresh and slow-frozen menisci.

(A) Transverse sections at half thickness show fresh (a, e, i, m), slow-frozen (b, f, j, n), 2-hour vitrified (c, g, k, o), and 3-hour vitrified (d, h, l, p) menisci. Images (a-d) are 2× objective magnification and stitched together. Images of regions (e-p) within dashed boxes (red = anterior (A), blue = central (C), and yellow = posterior (P)) presented at 10×. Scale bars: 2 mm (a-d), 200 μm (e-p). (B) Cell viability from fresh (N=4), slow-frozen (N=3), 2-hour vitrified (N=3), and 3-hour vitrified (N=3) menisci. p-value determined with two-way ANOVA with Bonferroni post-hoc test. (C) Cell viability quantified in outer (grey), middle (red), and inner (blue) layers within sagittal sections. No significant differences within layers of each group using two-way ANOVA with Bonferroni post-hoc test. (D) Metabolic activity between vitrified menisci for fresh (N=3; n=11), 2-hour (N=3; n≥10) and 3-hour (N=4; n≥11) durations, along with SF menisci (N=3; n≥11), over 0 to 4-day post warming. Metabolic activity normalized to corresponding average fresh meniscus values on day 0 within the same batch. N = number of independent menisci; n = number of sagittal sections. p-value determined via two-way ANOVA. Data are mean ± SD.

Quantitative analysis of viability was conducted in sagittal sections from the three regions (Figure 5B), with each region further divided into three layers: inner, middle, and outer (Figure 5C). The 3-hour vitrification showed a significantly higher viability in whole menisci, achieving an average of >70%, a significant improvement over slow freezing (~21.6%) and 2-hour vitrification (~40.3%) groups. Additionally, region-dependent viability was not observed for each group. When comparing different layers on the sagittal sections, although no significant difference was observed, there was an increasing trend in viability from the outer layer to the inner layer in the two vitrified groups. This trend can be attributed to the increase in CPA concentration as the thickness decreases from the outer, middle, to the inner layer on the sagittal section.

Metabolic activity of whole menisci from different groups was assessed and compared in different regions, as shown in Figure 5D. Among all three regions, 3-hour vitrification exhibited significantly better preservation of metabolic activity than slow-freezing and 2-hour vitrification immediately after warming, with the posterior horn showing the most notable improvement, reaching around 70% viability on day 0. In accordance with the metabolic activity pattern observed in fresh tissues, the 3-hour vitrified menisci consistently exhibited an increase, reaching up to and above 70% of metabolic activity in all regions during the 24-hour period. In contrast, the slow-frozen and 2-hour vitrified groups maintained much lower metabolic activity levels over the course of 4 days.

To continuously upscale the volume to 15 ml, whole TMJ discs were gradually loaded with VS55 for a duration of 3 hours predicted by our computational model. The transparent glassy state around the sample was observed by cooling at 9.2±0.3 °C/min to −100°C, followed by a slower cooling to −120°C at 2.4±0.4 °C/min (Figure S1, Supporting Information). During warming, the samples were first slowly warmed to −100°C at 13.9±1.1 °C/min, and then warmed faster to −35°C at 30.1±4.4 °C/min, in a 37°C water bath (Figure S1, Supporting Information). The vitrification of whole TMJ discs resulted in significantly improved cell viability compared to slow-freezing and comparable to fresh discs, as shown in Fig 6A using fluorescent live/dead imaging. The quantitative analysis revealed viability rates across all regions ranging from 82.4% to 87.5% after vitrification, a significant improvement over the slow-frozen samples which exhibited less than 40% viability (Figure 6B). A similar observation was obtained regarding the metabolic activity of the TMJ discs from day 0 to day 4 post-warming, as shown in Figure 6C. The vitrified TMJ discs displayed a continuous recovery in metabolic activity over time, fully recovering completely by day 2 post-warming, aligning with the observed increase in metabolic activity seen in fresh tissues. In contrast, this phenomenon was not observed in slow-frozen samples.

Figure 6. Viability and metabolic activity of 3-hour vitrified vs fresh and slow-frozen whole TMJ discs.

(A) Sagittal section of anterior (A), central (C), posterior (P), medial (M), and lateral (L) for fresh, slow-frozen (SF), and 3-hour vitrified (3 Hr Vit) TMJ discs. The whole anterior-posterior band images captured via 2× magnification and manually stitched together. Scale bars represent 3 mm. Five specific region views were captured via 10× magnification. Scale bars represent 200 μm. The region views for anterior (red), central (blue), and posterior (yellow) align with locations shown within the anterior-posterior bands through dashed boxes. (B) Quantitative viability analysis and comparison of five distinct regions obtained from three groups (N=4). p-value determined via two-way ANOVA with Bonferroni post-hoc test. (C) Metabolic activity comparison over a 4-day post-warming period in fresh, slow-frozen and 3-hour vitrified whole TMJ discs compared to their respective fresh values (N=4). p-value determined using two-way ANOVA. Data are mean±SD. N = number of independent TMJ discs.

2.4. ECM and biomechanics of 3-hour vitrified menisci closely resemble fresh samples

To evaluate the effect of vitrification on the ECM structure and compositions of whole menisci, we conducted histological imaging with hematoxylin & eosin (H&E), Safranin O, and Sirius red stains. The results are depicted in Figure 7A–D, captured using both brightfield and polarized light microscopes. The use of polarized light microscopy was particularly employed to focus on the examination of collagen fiber structure within the Sirius red-stained samples. In addition, we conducted a comparison of the polarized light image with its corresponding brightfield counterpart at the same location. To validate the collagen fiber structure observed in polarized light imaging, we conducted SEM imaging on transverse sections of meniscal samples extracted from the anterior horn regions. For whole menisci, images of the whole transverse sections through all three stains revealed that no significant difference in ECM structure was observed among all four groups. When observing each region in detail, it becomes evident that slow-freezing and two vitrification methods effectively maintained ECM compositions in the anterior horn and central regions. Only a slight decrease in glycosaminoglycans (GAGs) in the posterior horn was observed in Safranin O staining images post slow-freezing and vitrification compared to fresh tissues. Polarized light microscopy of Sirius red stained samples did not reveal notable variations in collagen fiber shape and content among the four groups, consistent with SEM observations (Figure S6, Supporting Information). Both polarized light images and SEM images of anterior horns of meniscal samples showed only slight differences in collagen fiber structure patterns between slow frozen and 2-hour vitrified samples compared to the fresh sample, whereas the 3-hour vitrified meniscal sample exhibited a pattern closely resembling the fresh sample. No apparent differences in ECM structures were observed in sagittal sections across all three regions of the whole menisci, as shown in Figure S3–S5 (Supporting Information).

Figure 7. Representative histological images and mechanical properties of whole menisci.

(A) Representative transverse sections with anterior horn (A), central (C), and posterior horn (P) stained with H&E, Safranin O, and Sirius red for fresh, slow-frozen (SF), 2-hour vitrified (2 Hr Vit), and 3-hour vitrified (3 Hr Vit). Magnification: 2×. Scale bars represent 5 mm. Corresponding brightfield images are at 20× with 200 μm scale bars (B-D). The polarized light images, along with their corresponding brightfield counterparts, were captured at 20× with 200 μm scale bars. Equilibrium contact modulus (E) and permeability (F) of the superficial layer in three regions for fresh (N=6; n≥14), SF (N=3; n≥6), 2-hour vitrified (N=3; n≥7), and 3-hour vitrified (N=5; n≥12) menisci, obtained through microindentation tests. N = number of independent menisci; n = number of measurements of all samples in each region. p-value determined with a two-sided t-test. Data presented as the mean ± SD.

To investigate region-specific mechanical properties of both whole menisci and TMJ discs among various groups, microindentation testing was conducted in this study. Equilibrium contact modulus and permeability were assessed to explore the effect of vitrification on tissue biomechanics in both superficial and intermediate layers of whole menisci. The setup of microindentation testing for menisci is illustrated in Figure S7 (Supporting Information). The results indicated that the equilibrium contact modulus of the 3-hour vitrified samples remained unchanged across all three regions in both layers after warming (Figure 7E and Figure S8A). Conversely, the slow-frozen samples showed an elevated equilibrium contact modulus in the posterior horn’s superficial layer, and 2-hour vitrified samples exhibited an increased equilibrium contact modulus in the central region’s superficial layer, in comparison to the fresh samples. Nonetheless, both 2-hour and 3-hour vitrification led to a reduction in permeability within the superficial layer of the central region (Figure 7F). Interestingly, all three preservation methods exhibited a decrease in permeability within the intermediate layer of the posterior horn (Figure S8B, Supporting Information).

ECM and biomechanics of 3-hour vitrified TMJ discs closely resemble fresh samples

In the evaluation of the ECM structure and composition of TMJ discs, three primary regions, comprising the anterior, central, and posterior, were subjected to histological imaging within three different groups. No significant alteration in ECM structure and composition was observed between fresh and 3-hour vitrified tissues for all three stains, whereas the slow-frozen tissues showed big pores and more severe damage to the collagen network when compared to both fresh and vitrified tissues (Figure 8A–D). A consistent pattern was observed in both SEM and polarized light images of central regions of TMJ discs. When compared to slow-frozen samples, the 3-hour vitrified TMJ disc samples exhibited collagen fiber structures more closely resembling those of the fresh samples (Figure S6, Supporting Information). For mechanical properties, the equilibrium contact modulus and permeability values for the whole TMJ discs did not show significant differences among the three groups (i.e. fresh, slow-frozen, and vitrified). Furthermore, a more in-depth analysis was conducted by comparing them on a region-specific basis. There were no significant differences in the equilibrium contact modulus across all regions when comparing fresh and 3-hour vitrified samples. However, a decrease in permeability of the posterior region was observed in the 3-hour vitrified compared to fresh samples. In the case of slow-frozen tissues, both the equilibrium contact modulus and permeability exhibit significant alterations in the posterior regions when compared to fresh tissues (Figure 8E and F). Notably, in the anterior regions of slow-frozen tissues, we observed a decrease in equilibrium modulus and a significant increase in permeability. These findings align with the alterations seen in trypsin-treated tissues (Figure S7), and can be attributed to damage to the ECM structure, which is evident from the presence of large pores observed in histological images (Figure 8B), resulting from ice crystal formation.

Figure 8. Representative histological images and mechanical properties of whole TMJ discs.

(A) Representative sagittal sections incorporating anterior (A), central (C), and posterior regions (P) stained using H&E, Safranin O, and Sirius red, are displayed for fresh, slow-frozen (SF), and 3-hour vitrified (3 Hr Vit). Magnification: 2×. Scale bars represent 2 mm. The corresponding magnified views of the anterior (B), central (C), and posterior regions (D) are displayed at 20× magnification with 200 μm scale bars. The polarized light images, along with their corresponding brightfield counterparts, were captured at 20× with 200 μm scale bars. Location-dependent equilibrium contact modulus (E) and permeability (F) of the superficial layer for fresh (N≥3; n=59 total), SF (N≥3; n=23 total), and 3-hour vitrified TMJ discs (N≥3; n=26 total), were obtained through microindentation tests. p-value determined with a two-sided t-test. N = number of independent TMJ discs; n = number of measurements. Data are mean ± SD.

3. Discussion

Vitrification shows tremendous promise in addressing the shortage of tissues and organs by allowing them to be stored at cryogenic temperatures for extended periods without the damaging effects of ice crystal formation. It has succeeded in maintaining viability of samples with VS55 in small volumes (≤3 ml) and has potential for larger vascularized tissues and small animal organs using higher CPA concentrations ^{[18–19, 23]} or nanowarming ^[24]. However, applying vitrification to avascular cartilaginous tissues in VS55 has been limited to small volumes (≤3 ml) ^[20]. This limitation arises from the denser structure, less permeable ECM, increased thickness, and lack of blood vessels in these tissues, making successful vitrification for larger-sized tissues (>3 ml) exceptionally challenging. Even with the incorporation of VS83 and nanowarming, improvements are still constrained ^[25]. Recognizing the urgent and extensive demand for suitable meniscus replacements in the knee joint and TMJ, this study specifically addressed the challenge of vitrifying larger-sized avascular fibrocartilaginous tissues by achieving CPA penetration within a limited timeframe while minimizing CPA-associated cytotoxicity.

We initially used small meniscal specimens with two different thicknesses: 1 mm and 3.5 mm. Through analysis involving fluorescence live/dead staining images and alamarBlue tests, we found, following a 1.5-hour VS55 loading (a procedure widely used for vitrification of blood vessels ^[19], rat hearts ^[24a] and kidneys ^[26]), that 1 mm thick vitrified samples had preserved cells comparable to fresh samples. However, when the thickness was increased to 3.5 mm, both vitrification and slow freezing reduced cell viability. When the VS55 loading time was extended to 3 hours, fortunately without causing any CPA toxicity, the vitrified 3.5 mm tissues maintained living cells throughout their full thickness, marking a promising development. These results are encouraging and lead to the following observations: 1) The VS55 loading time commonly used for vascularized tissues and animal organs falls short in ensuring adequate CPA penetration within dense avascular tissues; 2) The viability of vitrified tissues highly depends on diffused CPA concentration; 3) Vitrification enables sufficiently CPA-loaded meniscal tissues to maintain cell viability and thus has great potential to replace conventional cryopreservation methods; 4) The CPA loading time for adequate CPA penetration needs to be optimized for each different type of tissue.

The use of a simulation-based optimization approach was critical to explore the intricate process of CPAs diffusing across tissues of different sizes and shapes, and to improve more rapidly the testing and mechanistic understanding of the vitrification process. μCT imaging-based computational modeling was used to quantify diffusivity by virtue of its rapidity, non-invasiveness, and capacity to capture 3D data. Its efficacy has been demonstrated in previous applications, including the derivation of diffusion coefficients for low Me₂SO concentrations within tissue-engineered collagen scaffolds ^[27], as well as for high concentrations of vitrification solutions within porcine arteries ^{[21, 28]}. In the present study, μCT imaging was adapted to determine VS55 concentrations within meniscal tissues and was synergistically combined with a 3D computational model, which facilitated the derivation of VS55 diffusivity through the fitting of μCT-derived data. Additionally, the effect of temperature on VS55 diffusion was explored.

To the best of our knowledge, this is the first work to report CPA diffusivity in meniscal tissues, and the diffusivities measured are in good agreement with those measured in other collagenous tissues. A comparison between our data and CPA diffusion data in prior literature is shown in Table S3. The simulation curves of VS55 concentration profiles within meniscal specimens yield a compelling rationale for the observed viability outcomes. Notably, the predictions elucidate the relationship between VS55 diffusion and viability. For the 1 mm thick tissues, a 1.5-hour VS55 loading period led to >95% center loading, resulting in comparable viability to fresh tissues. However, for 3.5 mm thick tissues, the simulation predicted a mere 50% VS55 diffusion in center areas after 1.5-hour loading, leading to reduced viability (23%). Extending the loading to 3 hours yielded 85% viability due to >80% VS55 center loading. This achievement highlights the crucial role of time in ensuring sufficient CPA diffusion for successful vitrification.

VS55 diffusivity was subsequently employed to predict concentration profiles for varying exposure times throughout the whole menisci and TMJ discs, scaled up to total volumes of 10 ml and 15 ml, regardless of their specific geometries. This predictive capacity provides invaluable insight into the optimization of CPA loading times for larger-sized tissue vitrification. Following the simulation of CPA distribution throughout the whole 3D tissue geometry models, our fitting analysis on three primary regions of each meniscus—the anterior horn, the central region, and the posterior horn—in both transverse and sagittal sections, while five regions of TMJ discs, comprising anterior, posterior, medial, lateral, and central, was analyzed. The 3D simulation revealed a location-dependent CPA concentration. The μCT validation reinforces the accuracy of our CPA diffusion models for both meniscus and TMJ disc. Notably, the lowest CPA concentration in each region was 50–75% VS55 through the whole meniscus following 3-hour loading, in contrast to less than 50% VS55 diffused within the tissue after 2 hours. For CPA distribution throughout the whole TMJ disc, both the computational model and experimental μCT data confirmed more than 75% VS55 penetration after the 3-hour loading period. As a result, the 3-hour VS55 loading protocol was used for the vitrification of whole TMJ discs.

The efficacy of the optimized 3-hour VS55 loading period was then assessed on whole knee menisci and TMJ discs, examining both cellular and tissue level impacts. As the control group, whole fresh-frozen menisci exhibited less than 10% metabolic activity upon thawing, indicating a scarcity of viable cells (Figure S9, Supporting Information). Although the addition of 10% DMSO combined controlled-rate slow-freezing increased overall metabolic activity to approximately 40%, this level may not be sufficient to sustain long-term clinic outcomes post-transplantation, as it remains significantly lower than clinical standards of 70% viability. The 2-hour vitrified menisci displayed a depth-dependent viability pattern. Only a thin layer of living cells was preserved on the surface of the inner layer, resembling our prior observations involving larger-sized articular cartilage ^[25]. However, the 3-hour vitrified knee menisci yielded significant improvements in both cell membrane integrity and metabolic function. These improvements led to a viability rate of approximately 70% on day 0 post-rewarming, indicating the optimal timeframe for transplantation. Upon examining the ECM structure of the whole knee menisci, no significant alterations were detected in the overall tissue structure and collagen networks within the 3-hour vitrified groups as compared to the fresh tissues. Nevertheless, a decrease in GAG content was observed in 3-hour vitrified menisci when compared to fresh tissues. This reduction could be attributed to leaching resulting from prolonged exposure to CPA solutions. A statistically significant decrease in permeability observed during biomechanical tests may be associated with tissue dehydration during cooling and warming processes. The loss of interstitial fluid can lead to increased density of the ECM, potentially enhancing the strength modulus and reducing the permeability of the ECM structure. While this decline in permeability was noted in the central region of 3-hour vitrified menisci, it did not exert a substantial impact on the biomechanical strength (equilibrium contact modulus) of the tissue. The TMJ discs, vitrified with this new approach, had nearly 85% viability and a full recovery of metabolic activity after day 2 post warming, even when scaling the volume up to 15 ml. Moreover, the collagen network of vitrified TMJ discs closely resembled that of fresh tissues. In contrast, the ECM structure of slow-frozen samples exhibited severe damage such as pores and cracks. No significant changes in biomechanical strength (equilibrium contact modulus) of vitreous TMJ discs across all regions were observed when compared to fresh tissues. These findings provide a solid foundation for assessing the long-term outcomes post-transplantation in our forthcoming in vivo animal studies.

Notably, the 3-hour duration of VS55 loading did not give rise to CPA-associated toxicity concerns for knee menisci and TMJ discs. Conventional convective warming proved a successful combination with this CPA loading for 3 hours, which can be primarily attributed to the addition of sucrose and trehalose. Our differential scanning calorimetry (DSC) analysis revealed that these additives significantly improved the glass forming ability of VS55, raising its glass transition temperature from −121.2 ± 0.4°C to −115.9 ± 0.2 °C (Figure S6, Supporting Information). As such, the critical cooling rates (CCR) decreased from 2.5°C/min to below 2°C/min and the critical warming rates (CWR) CCR dropped from 50°C/min to below 10°C/min. The addition of these sugars ensured that our systems met their CCRs and CWRs, provided that the tissues could be fully penetrated by the CPA solutions. However, our optimized 3-hour loading protocols for both knee meniscus and TMJ disc were unable to achieve 100% CPA penetration throughout the entire tissue, which could potentially lead to the risk of ice crystal formation at tissue locations with lower CPA concentrations. This might be one of the reasons why the improved viability reached 70% instead of a perfect 100%. In fact, extending exposure time until achieving 100% CPA penetration could induce CPA toxicity to cells, causing damage to the cells before vitrification. Achieving a dedicated balance between CPA penetration and its toxicity is crucial. Additionally, as CCR and CWR decreased, the slow cooling and warming procedures used in our systems could help mitigate thermal stress, especially within larger-sized volumes of the system. According to our findings, there was no evidence of thermal stress-induced damage during vitrification and warming, such as cracking or structure fractures.

To consider scaling up to larger fibrocartilaginous avascular tissues such as human intervertebral discs, an extended CPA loading period can be optimized using the simulation approach established in this study. In the case where the CPA loading period surpasses the maximum acceptable exposure time (due to CPA-induced cytotoxicity) before achieving the required CPA penetration, our simulation approach can provide invaluable insights into CPA concentration and distribution throughout the tissues. To prevent the formation of ice crystals during the warming process, faster warming methods such as nanowarming or removing surrounding vitrification solution ^[29] could be considered in conjunction with the optimized CPA penetration developed in this study. As a result, due to the heightened thermal gradient resulting from faster warming methods, it will be crucial to exercise precise control over thermal stress within upscaling systems.

While our research predominantly focused on vitrification of fibrocartilaginous avascular tissues, the applicability of our optimization approach extends to diverse musculoskeletal tissues. By gaining a deeper understanding of how CPAs permeate each distinct tissue type, such as articular cartilage, bone, and the cartilage-bone interface, our approach holds the potential to enhance future optimization of CPA loading procedures for large-sized OCAs. The diffusion of VS55 was assumed to be homogeneous and isotropic in this study. The simulation approach developed in this study was validated against experimental data obtained from μCT, indicating a reasonable alignment between the simulation results and the empirical data. This alignment serves as an indicator of the accuracy and reliability of the simulation under the assumption of homogeneous and isotropic diffusion. Exploration of the inhomogeneous and anisotropic diffusion properties of CPAs was not included in this research, despite location-specific and direction-dependent alignment of meniscal matrix structures ^[30]. This could be a valuable avenue for future investigations, especially when exploring different types of tissues. Furthermore, in our study, the pig model, a large animal model, was chosen due to its established comparability to humans in terms of morphological, mechanical, and biochemical properties ^[31]. However, it is essential to acknowledge the differences between porcine and human menisci when applying these findings to human tissues. For instance, human menisci possess a less dense ECM structure than porcine menisci. Consequently, human menisci are expected to present fewer challenges in terms of CPA penetration and hold greater potential for successful vitrification. Nevertheless, assessments regarding cellular tolerance to CPA toxicity and microstructure consideration remain to be explored when translating this research to vitrification of human menisci.

4. Conclusion

In conclusion, the present study makes four significant contributions towards addressing the shortage of suitable knee meniscus and TMJ disc replacements. First, we have pinpointed a critical obstacle impeding the application of vitrification to larger-sized avascular fibrocartilaginous tissues, which revolves around achieving adequate CPA penetration. Second, we have effectively tackled this challenge by developing a μCT imaging-based computational modeling approach to optimize the CPA loading period, thereby ensuring adequate CPA penetration. Third, we effectively facilitated the upscaling of vitrified tissue volumes to 15 ml, while leading to substantial improvements in viability (≥70%) that hold great promise for future clinical applications. Fourth, by focusing on the whole knee meniscus and TMJ disc, this research lays the groundwork for future development of long-term preservation methods of viable human meniscal and TMJ disc grafts. It introduces a valuable repository of high-quality vitreous tissue, serving both transplantation needs and diverse research pursuits. Notably, this pioneering work represents the first successful application of vitrification strategies to preserve whole viable grafts in both tissue types which opens new avenues for the successful preservation of various avascular tissues and holds the promise of facilitating widespread adoption of viable graft transplantation and research endeavors.

5. Experimental section

Study design

The objectives were to (1) identify the main barrier, i.e., inadequate CPA penetration, hindering vitrification for avascular musculoskeletal fibrocartilaginous tissues, –and (2) address this challenge by optimizing CPA loading periods. We used 4-ml small meniscal specimens for proof of concept. Our results demonstrated that viability of vitrified samples strongly depended on diffused CPA concentrations. To gain deeper insights into the CPA diffusion mechanism in meniscal tissues, we developed a 3D computational model using finite element analysis, determining best-fit CPA diffusivity values from experimental μCT scans. With this approach, we accurately predicted CPA distribution throughout the whole meniscus and TMJ disc, considering various geometries and loading periods. Using these predicted protocols, we evaluated the viability of vitreous samples through fluorescent live-dead imaging and alamarBlue assays, comparing findings with fresh and slow-frozen samples. Histological analysis and microindentation assessed ECM structure, compositions, and mechanical properties. Experiments were replicated at least three times (N≥3), with blinded tissue analysis. Pre-established criteria were used for exclusions, considering contamination, inflammation, and defects. Outliers from time-course metabolic activity assessments were excluded starting from the day when sample contamination occurs (the RFU values were typically ten times higher than those of fresh samples). Outliers from the biomechanical data were excluded due to incomplete creep curves caused by improper experimental setups. Menisci and TMJ discs underwent randomization from a pool of N pigs. Each individual meniscus or TMJ disc within every experimental group originated from distinct pigs, where n represents the count of specimens or measurements, accounting for cases where multiple specimens or measurements were taken from an individual meniscus or TMJ disc.

Tissue dissection

For investigations involving 4-ml small meniscal specimens and 15-ml whole TMJ discs, menisci and TMJ discs were procured from male and female Yorkshire pigs (8–9 months old, approximately 150 kg) within 6 hours of death, acquired from a local slaughterhouse. For the 4-ml system, cylindrical meniscal specimens (5 mm diameter) of two thicknesses (1 mm or 3.5 mm) were prepared in the superior-inferior direction in the center region of each meniscus using a micro dissecting trephine and a custom cutting tool with parallel razor blades. For the 10-ml whole menisci studies, intact medial and lateral menisci were procured from male domestic Yorkshire pigs (4–6 weeks, around 10 kg) from Valley Brook Research Inc. (Madison, GA). Leg removal occurred within 30 minutes of the animal’s demise following approved research protocols by the Institutional Animal Care and Use Committee (IACUC) at the Medical University of South Carolina (MUSC). All tissues were surplus sources, eliminating the need for specific euthanization for this research. All procured tissues were immediately placed in Dulbecco’s Modification of Eagle’s Medium (DMEM) with glucose (1 g/L), L-glutamine, and sodium pyruvate (Corning, VA) supplemented with penicillin (100 IU/ml) and streptomycin (100 μg/ml) (Millipore Sigma, MO) (DMEM/P/S) at 4°C and randomized for subsequent experiments.

Cryopreservation methods by vitrification and slow freezing

VS55 composed of dimethyl sulfoxide (Me₂SO, 3.1M), formamide (3.1M) and propylene glycol (PG, 2.4M) was employed for all CPA loading procedures. VS55 supplemented with sucrose (0.3 M) and trehalose (0.3 M) (VS55/S/T) was used for tissue vitrification. Sucrose and trehalose were added to facilitate glass formation and prevent ice crystal formation during cooling and warming. All samples were gradually infiltrated with VS55 (8.4 M) in 1×Euro Collins (EC) solution at 0°C. The precooled diluted vitrification solutions were added in six sequential 15-min (total 1.5 hours), 20-min (total 2 hours) or 30-min (total 3 hours) steps at increasing concentrations (0%, 18.5%, 25%, 50%, 75%, and 100%) on ice ^{[19–20, 32]}. Each sample was then transferred to a 20-ml glass scintillation vial containing 4 ml VS55/S/T (for small meniscal specimen), or a 50-ml centrifuge tube containing 10-ml (for whole meniscus) or 15-ml (for whole TMJ disc) VS55/S/T. For the 4-ml system, the top of the vitrification solution was then covered with 1 ml of 2-methylbutane (Sigma-Aldrich, MO) to prevent direct air contact ^[20]. Subsequently, all samples were successfully vitrified and stored in a −135°C mechanical storage freezer (Revco^™, ThermoFisher Scientific, USA) for over 48 hours. The cooling and warming profiles for all tissue samples are shown in Figure S1, Supporting Information, and the corresponding cooling and warming rates reported in Table S1, Supporting Information.

For conventional cryopreservation by slow freezing, all samples were exposed to Me₂SO (10% v/v) in DMEM with 1% P/S for 30 min at 4°C and then transferred respectively to the same corresponding containers with the same corresponding CPA volumes of Me₂SO (10% v/v) with those used for vitrification. Conventional cryopreservation was performed by cooling slowly to −80°C at −1°C/min using a control-rate freezer (Kryo 560, Planer, Middlesex, UK) ^{[12, 20]}. The slow-frozen samples were then stored at −160°C in vapor phase nitrogen for a minimum of 48 hours.

Warming of all vitrified and slow-frozen tissue samples was accomplished by convectively warming in a 37°C water bath, whereupon the vitrification solution was removed in seven sequential steps at 0°C into DMEM/P/S culture medium as previously described ^{[19–20, 32]} and 10% Me₂SO was removed in DMEM/P/S twice at 4°C for 30 min in total. All fresh and preserved samples underwent viability assessment using fluorescence live/dead staining and metabolic activity using an alamarBlue assay (BIORAD, California). Comprehensive protocols for viability and metabolic activity assessment are available in Supporting Information.

VS55 diffusivity in meniscal tissues determined using a μCT-based simulation approach

μCT imaging was used to determine CPA concentration in meniscal tissues after the 90-min stepwise VS55 loading. A thin transverse section at half thickness of each sample was scanned using a Scanco μCT 40 device (Scanco Medical, Southeastern, PA) (Figure 3A). Comprehensive scanning and reconstruction protocols are available in Supporting Information. The greyscale values (μ) of the pixels on the diameters of the μCT-scanned transverse sections were extracted using a Radial Profile Plot plugin in ImageJ software (NIH) and converted to HU values (equation (1) in Supporting Information). Referring to the correlations between HU and CPA concentration found in this study, the corresponding CPA concentration values were determined. COMSOL Multiphysics version 5.5 (COMSOL, Burlington, MA) was used to build a 3D model for the diffusion of CPAs into the meniscal specimen with a diameter of 5 mm and its measured thickness. A time-dependent CPA concentration boundary condition was used to simulate the six CPA loading steps. The diffusion was assumed to be homogeneous and isotropic, and the diffusivity was quantified by fitting the modeling curves to the experimental data of samples through a least squares method as previously described ^{[21, 27, 33]}. To simulate the effect of exposure time on CPA concentration, increasing or decreasing total exposure time was achieved by equally increasing or decreasing the time at each step; for example, adding 5 min per step increases the total exposure time by 30 min. A validation of the computational model was conducted through a comparison between the modeling curves and the experimental data of samples loaded with VS55 at 0°C over a total exposure time of 60 min (10 min each step).

Simulation and validation of CPA concentration throughout the whole meniscus and TMJ disc

The whole 3D geometry models of meniscus and TMJ disc were captured using a Skyscan 1176 μCT scanner (Bruker, Belgium), reconstructed, and then created with a second-order tetrahedral mesh for CPA concentration simulation in the COMSOL Multiphysics software. The concentration profiles of the meniscus after different CPA loading periods were simulated using the 3D model and compared to those from μCT data for 3 hours, using the VS55 diffusivity determined at 0°C. This analysis encompassed both transverse and sagittal sections. The ensuing evaluation involved the calculation of the R-square value. Leveraging the analogous structure shared between knee meniscus and TMJ disc, we used the determined mean value of VS55 diffusivity at 0°C derived from meniscal tissues to predict CPA concentrations across the whole 3D TMJ disc model after the 3-hour loading period and compared it with the experimental μCT data. The protocols for 3D model capture and μCT validation are available in Supporting Information.

Histological analysis of whole menisci and anterior-posterior bands of TMJ discs

Whole menisci collected from the four experimental groups (fresh, slow-frozen, 2-hour vitrified, and 3-hour vitrified) and TMJ discs collected from the three experimental groups (fresh, slow-frozen, and 3-hour vitrified) underwent fixation in a 10% formalin solution for a duration of 48 hours. Subsequently, these samples were dehydrated and embedded in paraffin. Microtome transverse or sagittal sections 5 μm in thickness were prepared and stained with Hematoxylin & Eosin (H&E), Safranin O, and Sirius red. Bright-field images were captured using ×2 and ×20 objectives on a BZ-X810 microscope (Keyence, Itasca, IL). The polarized light images, along with their corresponding brightfield counterparts, were captured using ×20 objective on an Olympus BX53 microscope (Olympus, Center Valley, PA).

Mechanical microindentation testing of whole menisci and TMJ discs

Whole menisci were collected from the four experimental groups: fresh, slow-frozen, 2-hour vitrified, and 3-hour vitrified. Whole TMJ discs were collected from the three groups: fresh, slow-frozen, and 3-hour vitrified. All preserved samples were incubated for one hour at 37°C following warming and CPA removal steps prior to microindentation testing. Each meniscus was separated into three sections consisting of the anterior horn, central region, and posterior horn. Each section was initially tested along the superficial layer of the meniscus. Samples were then shaved and tested along the intermediate layer of the meniscus (Figure S7, Supporting Information). Only the superficial layer of each whole TMJ disc was tested. Thickness measurements were taken using a current sensing micrometer with the values averaged to determine the meniscus section thickness. Sample attachment to the sample holder used a small amount of cyanoacrylate. PBS solution was used to immerse the sample, and the holder was aligned to ensure that the indention tip was perpendicular to the surface tested. The microindentation testing protocols and data analysis methods are available in Supporting Information.

Statistical analysis

The raw data is available in the Supporting Information. Statistical analyses were performed using OriginPro software (OriginLab, Northampton, MA). Data are shown as mean ± SD (SD=error bars). Details regarding statistical tests, the number of replicates, and p values can be found in the corresponding figure and figure legends. Differences were considered significant when p<0.05.