Aliza A. Allon, MS, Richard A. Schneider, Ph.D, Jeffrey C. Lotz, Ph.D
Department of Orthopaedic Surgery, University of California, San Francisco, California, USA
Our goal is to optimize stem cell-based tissue engineering strategies in the context of the intervertebral disc environment. We explored the benefits of co-culturing nucleus pulposus cells (NPC) and adult mesenchymal stem cells (MSC) using a novel spherical bilaminar pellet culture system where one cell type is enclosed in a sphere of the other cell type. Our 3D system provides a structure that exploits embryonic processes such as tissue induction and condensation. We observed a unique phenomenon: the budding of co-culture pellets and the formation of satellite pellets that separate from the main pellet.
MSC and NPC co-culture pellets were formed with three different structural organizations. The first had random organization. The other two had bilaminar organization with either MSC inside and NPC outside or NPC inside and MSC outside.
By 14 days, all co-culture pellets exhibited budding and spontaneously generated satellite pellets. The satellite pellets were composed of both cell types and, surprisingly, all had the same bilaminar organization with MSC on the inside and NPC on the outside. This organization was independent of the structure of the main pellet that the satellites stemmed from.
The main pellets generated satellite pellets that spontaneously organized into a bilaminar structure. This implies that structural organization occurs naturally in this cell culture system and may be inherently favorable for cell-based tissue engineering strategies. The occurrence of budding and the organization of satellite pellets may have important implications for the use of co-culture pellets in cell-based therapies for disc regeneration.
From a therapeutic point of view, the generation of satellite pellets may be a beneficial feature that would serve to spread donor cells throughout the host matrix and restore normal matrix composition in a sustainable way, ultimately renewing tissue function.
Spinal intervertebral discs provide flexibility while supporting compressive forces. Each disc is composed of a peripheral, ligament-like annulus fibrosus and a central nucleus pulposus (NP). The NP contains chondrocyte-like cells embedded in a matrix of proteoglycan and type II collagen that is highly hydrophilic, which causes the tissue to swell and resist compression hydrostatically.1 This high-pressure environment is incompatible with blood vessel maintenance, and consequently the disc is the largest avascular tissue in the body as well as a challenging environment for cell function and survival. For these reasons, disc degeneration is common and an underlying cause of various spinal disorders.2
Tissue engineering is a growing and dynamic field with the potential to provide patients with minimally invasive treatments that repair or replace dysfunctional musculoskeletal tissues. A tissue engineering goal for the intervertebral disc is to re-establish pain-free motion by restoring the physical and biochemical properties of the NP matrix. This may be accomplished by stimulating host cells to resume matrix synthesis (particularly aggrecan) and/or by introducing new, more synthetically active cells. Mesenchymal stem cells (MSC) are attractive for this purpose because they can differentiate into a variety of cell types, including chondrocytes, and are a ready source of undifferentiated autologous cells.3 The feasibility of this approach has been demonstrated in several in vivo animal studies.4., 5., 6. However, long-term functional regeneration of adult discs has not been achieved. Ultimately, the degenerative disc is a hostile, often inflammatory environment, which experiences significant mechanical loading.7 When immature MSC are suspended in a carrier and injected into a degenerative disc, they find it difficult to survive, stay in the desired location, and may not receive the environmental cues that enable them to perform optimally to regenerate the tissue.
Pellet culture systems may have benefits in a tissue engineering setting as they can selectively recapitulate embryonic microenvironments for regenerative purposes.8,9 During embryonic development, cartilage and disc formation begins with the aggregation of progenitor cells into a cell condensation. These condensations then progress towards differentiation via the process of tissue induction and begin to secrete matrix.8., 9., 10., 11., 12., 13., 14., 15., 16. Several groups have looked at the concept of co-culture systems of nucleus pulposus cells (NPC) and MSC because signaling between these cell types ultimately occurs in situ during MSC-mediated disc regeneration. Yamamoto and coworkers conducted a 4-day monolayer co-culture study and reported significant increases in proteoglycan synthesis and cell proliferation when non-degenerative NPC and MSC were cultured with direct cell-cell contact.17 They reasoned that MSC were acting as feeder cells, which enhanced the ability of NPC to proliferate and secrete matrix. Richardson and colleagues also employed a similar 2D co-culture system and demonstrated that NPC cause MSC to differentiate into an NP-like phenotype as assessed by gene expression after FACS (fluorescence-activated cell sorting, Becton Dickinson, Franklin Lakes, New Jersey).18 They observed that a 75% NPC / 25% MSC ratio was optimum for MSC differentiation, as indicated by SOX9, collagen 2, and aggrecan gene expression.18 However, using a 3D co-culture system, Le Visage and colleagues noted that a random mixture of MSC and degenerative NPC did not increase GAG production beyond single-cell type controls.19 More recently, Vadala and colleagues have shown that 3D unstructured co-culture of MSC and NPC upregulated key differentiation markers in MSC.20 These papers have looked at the environmental triggers responsible for re-creating the condensation shape and the unique signaling due to co-culturing.
We hypothesize that by mimicking cell condensations through the use of a pellet system, and by allowing for tissue induction via a bilaminar organization, disc regeneration can be enhanced. As a first step toward testing this hypothesis, we generated bilaminar pellets of MSC and NPC and analyzed their behavior in culture. We observed a unique phenomenon: the budding of co-culture pellets and the formation of satellite pellets that separate from the main pellet. Interestingly, the satellite pellets were composed of both cell types and had a specific organization in which the MSC were on the inside and NPC on the outside. The occurrence of budding and the organization of satellite pellets may have important implications for the use of co-culture pellets in cell-based therapies for disc regeneration.
Bovine NPC were isolated from caudal discs of healthy adult cows within 48 hours of sacrifice. The NP tissue was carefully separated by gross dissection and digested in 0.5% collagenase/dispase and 2% antibiotic/antimycotic in low glucose Dulbecco's Modified Eagle Medium (DMEM) at 37 °C for 4 to 6 hours with constant stirring. The cells were then plated in tissue culture flasks and expanded to the fourth passage in NPC Media (DMEM with 1% antibiotic/antimycotic, 1.5% 400 m osmolarity, and 5% fetal bovine serum (FBS)) at 37 °C with 5% CO2. Culture media was changed twice a week.
Commercially available human MSC were purchased (Lonza, Basel, Switzerland) and expanded to the sixth passage in monolayer culture using growth media (DMEM low glucose with 1% antibiotic/antimycotic and 10% FBS) at 37 °C with 5% CO2. Culture media was changed twice a week.
Human nucleus pulposus samples were obtained from a consenting 55-year-old female patient undergoing surgery for scoliosis. The tissue was digested and the cells expanded. In addition, bovine MSC were isolated from femur tissue and the cells were expanded.
Human MSC and bovine NPC were used to make co-culture pellets. The cross-species human-MSC / bovine-NPC pellets enabled us to trace the location of the cells via their lineage. Three different types of pellets were formed, each consisting of 500,000 cells total: pellets of 100% one cell type, pellets of MSC and NPC with randomized organization, and pellets of MSC and NPC organized into a bilaminar structure. The pellets containing both MSC and NPC were formed with 3 different cell number ratios of 25/75, 50/50, and 75/25 respectively (Figure 1). To produce the 100% one-cell type pellets, 500,000 cells were pipetted into a 15 mL polypropylene tube and centrifuged at low speeds (300 g) for 5 minutes. To create the randomized pellets, both cell types were added to the same tube, pipetted to ensure thorough mixing, and centrifuged at low speed for 5 minutes. In order to form the bilaminar organized pellets, the cell type that would form the inner sphere of the pellet was added to a 15 mL polypropylene tube and centrifuged at low speed for 5 minutes. Subsequently, the second cell type that would form the outer shell was gently added to the same tube. The cells were then centrifuged again at low speed for 5 minutes. Organized pellets were formed for all 3 ratios with MSC on the inside and NPC on the outside and vice versa. All pellets were cultured in 2 mL of growth media for 3 days with caps loosened to allow for gas exchange. After 3 days, the pellets became spherical and were transferred to Costar 24-well ultra low attachment microplates (Corning Inc., Lowell, Massachusetts) for the remainder of their culture time. Media was changed 3 times a week.
We also made same-species pellets as controls for species interactions. We made pellets with human MSC and human NPC and pellets with bovine MSC and bovine NPC.
Before being pelleted, cells were labeled with fluorescent cell membrane markers (either DiO or Dil, Invitrogen, Carlsbad, California). After the pellets had reached their desired culture time, they were embedded in OCT Tissue-Tek® (Sakura Finetek USA, Torrance, California) and frozen sectioned at 7 nm using a cryostat. Sections were then counterstained with the Hoechst dye, a fluorescent nuclear marker (Sigma, St. Louis, Missouri). Images were taken using epi-fluorescent microscopy.
At the end of the culture time, the pellets were fixed in 10% buffered formalin overnight and processed for paraffin sectioning at 7 nm thickness. At this stage, the sections were immunostained using the human specific antibodies Lamp1 and Lamp 2 (Lysosomal-associated membrane proteins 1 and 2, Abcam plc, Cambridge, England). The sections were then counterstained with hematoxylin.
The NPC started as small rounded cells. They often required several days to attach to the tissue culture flask. Many of the cells never attached at all and were discarded. Though the cells were initially seeded in very small flasks (12.5 cm2, Falcon, Becton Dickinson, Franklin Lakes, New Jersey), it took up to 2 weeks for the cells to reach confluence. Once the cells were confluent and passaged, their growth rate significantly increased, and they were easily expanded to the fourth passage.
Immediately after centrifugation the pellets appeared flattened. Spheres formed within 48 hours and reached a maximum size of approximately 2 mm in diameter. After one week of culture, co-culture pellets began to exhibit budding (Figure 2). During the course of the second week of culture, these buds separated from the main pellet entirely to form numerous independent satellite pellets of various sizes. As the satellite pellets budded off of the main pellet, the main pellet did not noticeably decrease in size. At the 3-week time point, several of the larger satellite pellets also began to exhibit budding.
The 100% MSC pellets did not exhibit budding nor did any satellite pellets form at any point. The 100% NPC pellets exhibited budding and satellite pellet formation within the same time frame as the co-culture pellets (obviously in this case the satellite pellets were composed entirely of NPC). The same-species pellets (eg, human-MSC and human-NPC) exhibited the same behavior as the cross-species pellets (human-MSC and bovine-NPC). There was no difference in the budding or satellite formation rate between the bilaminar co-culture pellet and the random co-culture pellets.
Both the frozen (Figure 3) and paraffin histological sections confirmed that the main pellets maintained their structure throughout the culture time even as satellite pellets budded off them (Figure 4). The histology of the satellite pellets confirmed that they were composed of both cell types. Most surprisingly, the satellite pellets all appeared to have the same structure with MSC on the inside and NPC on the outside (Figure 5). This structural organization was independent of the structure and ratio of the main pellet that they stemmed from.
How MSC and NPC influence each other's behavior in 3D culture has very important implications for cell-based, disc regeneration therapies. Presumably, a cell-based therapy will always elicit an interaction between the newly introduced cells and the host cells since NPC would not be removed during a minimally invasive injection. Previous studies have shown that cell-cell contact between MSC and NPC was necessary to induce a stimulatory effect.18,19
Several groups have shown that unstructured co-culture offers some advantage over using MSC or NPC alone in the search of a cell-based strategy to regenerate the nucleus.17., 18., 19., 20. It has been shown that these cell types have the ability to communicate with one another to increase MSC differentiation, cell proliferation, and matrix production.17., 18., 19., 20. In our system, we have observed a novel phenomenon whereby co-culture pellets exhibit budding and the formation of satellite pellets. This occurrence may have important implications as to the behavior of constructs when administered in vivo during disc regeneration. What remains unclear as of yet is which underlying biological mechanisms are at work. However, we believe that such mechanisms will be rooted in the molecular and cellular processes underlying cartilage formation where similar behaviors exist.10., 11., 12., 13., 14., 15., 16.
In order to ensure that our results were not due to species incompatibility between human and bovine cells, we wanted to test the extent to which co-culture pellets composed entirely of human cells and entirely of bovine cells would exhibit the same dynamic behaviors. Our results indicate that budding and satellite pellet formation are generalizable phenomena that arise in response to certain cell configurations rather than being due to species incompatibility.
Overall, our experiments led to the novel and yet unexplained discovery that MSC and NPC, when co-cultured in any arrangement and in any ratio, will form bilaminar satellite pellets with MSC on the inside and NPC on the outside. This discovery implies that structural organization is intrinsic to these cell populations, can occur spontaneously in this cell culture system, and may be inherently favorable for cell-based tissue engineering strategies. We believe that bilaminar organization may accelerate MSC differentiation and the cells’ ability to resist environmental factors. At the center of a pellet, cells experience higher pressure and lower oxygen tension than on the outside. This microenvironment is therefore more conducive to MSC differentiation. In addition, the NPC readily form an epithelial-like layer on the outside of the tissue. This layer may ultimately protect the pellet from negative external signals, such as inflammatory cytokines.
From a therapeutic point of view, the phenomenon of budding and the generation of satellite pellets may be a beneficial feature that would serve to distribute donor cells throughout the host matrix. Over the long term, we envision using such a bilaminar pellet system to form the basis for new cell-based therapies for disc degeneration. Several pellets could be injected using a minimally invasive procedure into the nucleus pulposus of patients with back disc disease.2,21 We anticipate that these satellite pellets will enhance the ability of this technique to repopulate and regenerate the disc nucleus. These pellets would then serve to restore normal matrix composition in a sustainable way and ultimately renew tissue function.
The authors have no financial disclosures. Address correspondence to: Jeffrey Lotz, 533 Parnassus Ave, Box 0514, San Francisco, CA 94143.