Human meniscus allograft augmentation by allogeneic mesenchymal stromal/stem cell injections

Abstract Meniscus allograft transplantations (MATs) represent established surgical procedures with proven outcomes. Yet, storage as frozen specimens and limited cellular repopulation may impair graft viability. This proof‐of‐concept study tests the feasibility of injecting allogeneic mesenchymal stromal/stem cells (MSCs) in meniscus allograft tissue. We investigated the injectable cell quantity, survival rate, migration, and proliferation ability of MSCs up to 28 days of incubation. In this controlled laboratory study, seven fresh‐frozen human allografts were injected with human allogeneic MSCs. Cells were labeled and histological characteristics were microscopically imaged up to 28 days. Mock‐injected menisci were included as negative controls in each experiment. Toluidine blue staining demonstrated that a 100‐µl volume can be injected while retracting and rotating the inserted needle. Immediately after injection, labeled MSCs were distributed throughout the injection channel and eventually migrated into the surrounding tissues. Histological assessment revealed that MSCs cluster in disc‐like shapes, parallel to the intrinsic lamination of the meniscus and around the vascular network. Quantification showed that more than 60% of cells were present in horizontally injected grafts and more than 30% were observed in vertically injected samples. On Day 14, cells adopted a spindle‐shaped morphology and exhibited proliferative and migratory behaviors. On Day 28, live/dead ratio assessment revealed an approximately 80% cell survival. The study demonstrated the feasibility of injecting doses of MSCs (>0.1 million) in meniscus allograft tissue with active cell proliferation, migration, and robust cell survival.

and proliferation ability of MSCs up to 28 days of incubation. In this controlled laboratory study, seven fresh-frozen human allografts were injected with human allogeneic MSCs. Cells were labeled and histological characteristics were microscopically imaged up to 28 days. Mock-injected menisci were included as negative controls in each experiment. Toluidine blue staining demonstrated that a 100-µl volume can be injected while retracting and rotating the inserted needle. Immediately after injection, labeled MSCs were distributed throughout the injection channel and eventually migrated into the surrounding tissues. Histological assessment revealed that MSCs cluster in disc-like shapes, parallel to the intrinsic lamination of the meniscus and around the vascular network. Quantification showed that more than 60% of cells were present in horizontally injected grafts and more than 30% were observed in vertically injected samples. On Day 14, cells adopted a spindle-shaped morphology and exhibited proliferative and migratory behaviors. On

| INTRODUCTION
The knee menisci play a vital role in joint function and primarily contribute to load transmission, shock absorption, stability, and nutrient diffusion in the knee. 1,2 The C-shaped meniscus is composed of water (72%) and a densely packed collagen network of mainly type 1 (90%) and type 2 collagen fibers. 1 Compared with hyaline cartilage, proteoglycan content is relatively low but nevertheless important in determining viscoelastic properties of the meniscus. 1,2 Meniscal tears are the most common knee injury and frequently cause functional impairment and altered joint homeostasis, which ultimately results in a higher likelihood for early osteoarthritic changes. 3 Due to the poor vascularization in the white zone, 4 the natural healing capacity of avascular meniscal tears is relatively limited, thus justifying surgical intervention. 5 Arthroscopic partial meniscectomy remains the most commonly performed knee surgery when a meniscal repair seems unfeasible. 6,7 Nevertheless, the resected meniscus area should be limited to the absolute minimum, because resection is directly proportional to increased peak contact forces and early degeneration of the articular cartilage. 8,9 After meniscectomy, a reasonable number of patients remain symptomatic or develop joint dysfunction and/or pain. Young patients without arthritis are suitable candidates for meniscus allograft transplantation (MAT). Best outcomes after MAT are reported in relatively younger patients (<50 years) with a stable and well-aligned knee joint and minor articular cartilage degeneration. 10 Overall graft survival rates are currently estimated as 73.5% at 10 years and 60.3% at 15 years, with a mean time to failure of approximately 7.8 years. 11,12 The ultimate goal is to restore knee biomechanics and improve functional outcome in patients, while providing sufficient chondroprotection to delay cartilage degeneration. There is no consensus on the chondroprotective value of MAT, but it is unlikely to be equivalent to a native intact meniscus. 13,14 Solicitation of expert opinions through a survey from the International Meniscus Reconstruction Experts Forum (IMREF) showed that the majority of surgeons (68%) prefer fresh-frozen (−80°C) over cryopreserved meniscus allografts, mainly for logistical reasons and the relatively low associated cost. 15 Current donor cell survival rates in cryopreserved grafts range from 4% to 54%, 11 whereas meniscus cells do not survive fresh-frozen storage at −80°C. Therefore, insertion of a "dead" acellular piece of donor tissue provides a plausible explanation for tissue degradation with concomitant graft shrinkage (65% minimal, 20% mild, and 16% moderate) in the first 3 months posttransplantation as well as graft tearing and extrusion. [16][17][18][19] These complications pose a serious concern, because preoperative graft sizing is already challenging and it could further deteriorate the biomechanical environment of the knee joint and reduce the chondroprotective effect. [16][17][18][19] Rapid cellular repopulation after implantation may mitigate the initial lack of viable cells in the graft, but this notion does not conform to empirical observation. Due to the densely packed collagen fiber network, cell migration into the graft is time-consuming and largely incomplete. Rodeo et al. 20 have presented histology findings demonstrating that only a thin layer of the grafts' contact surface contains cells from the synovial membrane (fibroblasts) and signs of remodeling at 6 months after transplantation.
Although the importance of initial meniscus allograft viability to outcome and durability remains unclear, 10,13 early remodeling and biological incorporation of meniscus grafts may facilitate favorable outcomes, patient rehabilitation and return to activities of daily living, work, and sports activities. The development of improved integration strategies using cell-based procedures may repopulate and remodel meniscal tissue to prevent early degradation and shrinkage.
Although the meniscus is a challenging but valuable structure, cellmediated treatment may support the preservation and/or rejuvenation of meniscus architecture and function, while augmenting long-term outcomes of interventions for meniscus repair. 21,22 Accruing evidence suggests that human multipotent progenitor cells have therapeutic potential in orthopedics and that adipose-derived mesenchymal stromal cells (aMSCs) are a viable option for cartilage regeneration, (fibro-)chondrocyte differentiation, and meniscus engineering. 23 Analogous to a cell injection technique developed for repopulating collagen meniscal implants (CMI), 24 this proof-ofconcept study examined the feasibility of injecting mesenchymal stromal/stem cells (MSCs) in meniscus allograft tissue, as well as established the injectable cell quantity and volume, survival, migration, and proliferation up to 28 days of incubation. using the same protocol as described by Crespo-Diaz and others. 25 These MSCs have been extensively characterized by RNA-seq and cell surface marker expression. 26,27 The mesenchymal cells we used to represent the immature perivascular stromal fibroblasts that have the potential to differentiate into multiple mesenchymal lineages in cell culture. Certainly, these cells are not related to embryonic pluripotent stem cells because they do not robustly express the pluripotency markers Sox2, Oct4/POU5F1, or Nanog. 26

| Injection with MSCs
The cell-based injection experiments were subdivided in short-term magnification. If cells were present, images were taken at ×20 magnification to permit manual cell counting and evaluation with the ImageJ application. Total cell counts for each slide were calculated and averaged for two consecutive slides.

| Histology
For histological analysis, samples were either paraffin-embedded or visualized by cryosections. Tissues were first fixed with 10% buffered formalin for 5-7 days, after which stepwise dehydration with increasing ethanol concentrations was performed. Samples were cleared in xylene and immersed with paraffin for a minimum of 4 h in a 60°C oven. After mounting, tissues were entirely sectioned on a microtome at 10-µm thickness and stained with Mayer's H&E Y (Sigma), 25 Safranin-O (Sigma-Aldrich), 28

| RESULTS
A total of seven adult meniscus allografts (5 medial and 2 lateral) were used for the experiments. None of the grafts exhibited any structural defects upon macroscopic screening.

| Stain injection
The needle was horizontally introduced at the synovial attachment. Fresh live-dead staining immediately after injection showed a significant amount of viable (green) round cells at the injection site in a circular trend. After 2 days of incubation, a considerable number of living cells were observed but at a lower density. Dying (red) cells in microscopic images could either be injected MSCs or native donor meniscus cells, as indicated by the control sample ( Figure 6A).
Although the red fluorescent signals cannot discriminate between native and exogenous cells, we conservatively assume that at least some of the observed signals detected here represent injected cells that perished. Yet, surviving MSCs with a fibroblast-related spindlelike cell morphology were clearly detected on Day 2 ( Figure 6A).   introduction to the dense microenvironment of the meniscus. In addition, while the tissue is decellularized ("dead"), the ECM is expected to have residual bioactive proteins (e.g., bound to glycosaminoglycans associated with ECM proteins) that may instruct how MSCs behave upon injection. Moreover, based on histological analysis at 14 days after injection, MSCs appear to expand in a horizontally aligned fashion as they were initially injected ( Figure 3D).

| Long-term experiments
Both on H&E and Safranin-O stain, cells were oriented in a similar way as the circumferential meniscus collagen fibers were aligned.
Interestingly, cells were observed distal to the initial injection channel by horizontal dispersion; few or none appeared to migrate across the vertical plane.
To optimize injection procedures relative to the lamellar organization of the meniscus, we examined both horizontally and vertically injected allografts for the presence of injected cells. We for aggrecan/ACAN and type X collagen/COL10A1. 28,29 Yet, to date, only one published in vitro study attempted to improve cellular repopulation of meniscus allografts. This study reported outcomes of MSC seeding on the surface of a decellularized needle-punched meniscus allografts (28G and 1 mm spacing). 27 After 28 days of tissue culture, seeded cells were infiltrating through the created channels, whereas nonpunctured seeded grafts appeared to remain empty. These results emphasize the formidable barrier that high-density collagen fibers may pose for cell migration within meniscus grafts and the importance of increased graft porosity for spontaneous host cell influx. Moreover, the creation of space channels in the graft is somewhat similar to the meniscus "trephination" technique that has been described as an approach to induce blood supply for tear healing in the avascular meniscal zone. 30 A review by Michiewicz et al. 33 on preservation and sterilization methods of meniscus allografts discussed that fresh-frozen meniscus allografts are the most frequently used type of allograft, even though there are potential concerns with this tissue. The rationale of this study is to restore fresh-frozen meniscus allografts with living cells to rejuvenate the graft before implantation into the patient. However, the use of freshly stored allografts is increasing. Although such tissues already contain live resident cells, injection of stromal cells in fresh, viable meniscus allograft tissue may still be considered. Injection of MSCs in fresh grafts could potentially re-enforce the viability of the fresh tissue, whereas injection into frozen meniscus tissue would restore the viability of the decellularized graft. STRUIJK ET AL.

| 723
However, MSCs injected into fresh meniscus allografts may interact with the native meniscus cells and both cell types may mutually alter their biological behaviors and/or cell survival.
The current study identified several technical and cellular limitations that are not insurmountable but will require resolution in the future.  34 showed that the MSC viability, proliferation, and metabolism function are not affected by needle passage as small as 30G. As fluid pressure is exponentially related to the radius of the needle, we are confident that needle sizes 18, 20, and 21G selected for the majority of the experiments were not modifying the cells by mechanical sheering of fluid pressure. 34 In addition, MSCs injected into meniscus tissue encounter backpressure from injecting fluid into a dense meniscus, and it is still uncertain whether this could have altered the effective dosing of MSCs upon injection. We have tested fluid pressure upon injection with a manometer but observed that the resulting data were highly variable, thus precluding the definition of a maximal limit on the observed fluid pressure during meniscus injection.
Finally, samples were not biomechanically tested after injection.
Mechanical characteristics and material properties of the graft could potentially be compromised depending on the bore size of the needle and the number of injections. This feasibility study was solely targeting the biological cell performance in the graft; however, biomechanical analysis of cell enhanced meniscus grafts will be essential for any path towards clinical application.
The current study results are opening careful perspectives for clinically augmenting the initial viability of meniscus allografts, aiming to minimize the risk of graft shrinkage, optimize biological integration, and prolong MAT durability. Strikingly, MSCs were able to migrate following the alignment of collagen bundles in the meniscus, which provided a very encouraging avenue for consideration of MSCs in cellular repopulation of meniscus allografts in facilitating knee repair. Whether injected MSCs promote tissue recovery and synthesize relevant matrix proteins is a critical next step of this study. Although this study yielded encouraging study results, many factors remain to be clarified before this technique can be translated to in vivo models.

| CONCLUSION
The study demonstrated that injection of more than 0.1 million MSCs is clinically feasible and may suffice in improving the biological properties of meniscus allografts. Cells maintain proliferation and migration properties for at least 14 days postinjection and survival rates of (>80%) after 28 days are observed. These findings attest to the feasibility of repopulating meniscus allograft tissues with viable cells that may accommodate biological integration after joint surgery.

ACKNOWLEDGMENTS
We thank the members of the Dietz, Saris, and van Wijnen laboratories, including Peggy A. Bulur, Alexander M. Bornschlegl, and Leila Bagheri, for stimulating discussions, technical advice, and/or assistance with equipment and reagents. This study was supported with intramural funds from the Center of Regenerative Medicine, Mayo Clinic (to DBFS). Additional support was provided by NIH grant AR049069 (to AJvW) and the Department of Lab Medicine and Pathology. We appreciate the support from JRF Ortho® (Centennial, CO) who provided us with meniscus allografts. We wish to thank the ON Foundation for supporting this study project with a kick-starter grant. We also appreciate the support and generosity of William and Karen Eby. Other authors declare to have no conflict of interests.