The use of autologous mesenchymal stem cells (MSCs) derived from bone marrow to enhance tendon lesion repair is supported by a growing body of evidence from both experimental and clinical studies [1–5]. In the tendon, influx of a local subpopulation of precursor cells is believed to provide intrinsic post-injury repair [6]. However, these cells appear to be largely derived from peri-tendinous origins such as the paratenon [7] and result in fibrous repair with mechanical, structural and compositional differences from normal tendon. In an attempt to modify the repair towards regeneration, enhancing the small numbers of endogenous MSCs by implantation of large numbers of autologous, culture-expanded MSCs derived from donor tissue such as bone marrow, has been proposed. This hypothesis has been supported by positive results in experimental laboratory animal models of Achilles and patellar tendon injury [4, 5]. However, many laboratory animal models of induced injury have limited relevance to the human disease [8].

Horses, in contrast, suffer naturally-occurring flexor tendon injury with many similarities to human Achilles tendinopathy, making the horse a useful translational model for stem cell therapies. In recent years, a methodology for isolation, expansion and injection of autologous bone marrow-derived MSCs into lesions of the equine superficial digital flexor tendon has been widely accepted into clinical practice for the treatment of this disease [9]. Optimisation of cell-based therapies would ideally require accurate delivery to the target area without significant loss of cellular function or viability [10, 11], although a recent study established that only 24% of injected MSCs were retained at the site of injury after 24 hours [12]. Factors that may influence this poor cell retention include reduced cell viability following transport of the cells to the clinic or damage during the process of intralesional injection.

The current, commercial application of MSCs involves laboratory proliferation to achieve sufficient cells that then are either transported overnight, at 4 to 8°C, in autologous bone marrow aspirate (BMA) at a concentration of 5 × 10⁶ cells/ml [1] or are frozen [13]. The former technique was designed to achieve a fully autologous preparation that can be immediately injected, while the second requires thawing before implantation. These protocols, however, have not been tested in comparison with each other, or with other potential transport media. The first aim of this study was therefore to evaluate cell survival in different transport media.

Once the cells arrive at the clinic, they are injected under ultrasound guidance directly into the lesion within the recipient tendon. The disrupted central area of the tendon provides a cavity into which the cells are slowly injected with minimal injection pressure. Cells respond to external mechanical influences that affect their survival, growth and differentiation. Inappropriate stimulation of signalling pathways can lead to premature apoptosis, a complex cascade of events leading to the final demise of the cell [14]. The injection process could therefore potentially influence post-injection cell survival and metabolism. The most likely cause of cellular damage during injection is sheer stress, caused by the turbulent flow of fluid at the outer edges of the flow channel, where viscous fluid is in contact with the channel walls. Sheer stress is influenced by pressure, cross-sectional area and length of the channel [15], and by sudden changes in geometry, such as the rapid tapering of the vessel diameter from syringe to needle hub [10]. The surface marker phenotype and differentiation ability of rat and human MSCs were reportedly unaffected immediately following injection through catheters and needles at clinically relevant sizes and flow rates [10, 16], although a significant reduction in viability was observed for all catheters 24 hours after infusion. In contrast, another study demonstrated that longer and thinner cannulae were damaging to cells [17]. Other studies investigating cell printing methods to deposit cells in tissue scaffolds demonstrated significant loss of viability with increasing dispensing pressure and reducing nozzle size [18, 19] and suggested that a recovery period was necessary following injection. The technique developed for clinical use in the horse, which has direct relevance to human treatment, utilises a 2 ml syringe and hypodermic needle. A second aim of the present study was therefore to investigate the influence of needle size in this system on both cell viability and the potential delayed effects of injection.

This study demonstrates that transport media but not needle size influenced cell viability, but both had delayed effects on cell metabolism and apoptosis.

This study has demonstrated significant adverse influences on MSCs associated with transport and implantation. It is therefore important to select a transport media and implantation technique that minimises these adverse effects so as to maximise viable cell delivery to the site of injury.

With respect to the selection of transport media, it would appear that any of the wide array of those tested would be satisfactory when implantation occurs within 24 hours, as is customary with the clinical supply of MSCs in the veterinary field. Once cells were suspended for longer than 24 hours, viability of cells stored in BMA was significantly less than if the cells were frozen in cryogenic medium. Although there is, as yet, no guidance on the storage of MSCs, it has been suggested that in vitro storage of haematopoietic stem or progenitor cells should not exceed 2 hours [23]. This suggestion is impractical both for current equine therapeutic use and when considering future, commercial applications of MSC therapy in other species, including humans, because of the necessity to transport the cells from a remote licensed facility to the clinic and our data suggest that up to 24 hours for transportation is possible without significant detrimental effects. For transportation periods suspected to be of greater duration than 24 hours, cells should ideally be frozen, to maximise cell viability. However, the cell suspension media in which MSCs were frozen is not an ideal carrier for intralesional injection, and resuspension in an alternative media would be necessary prior to use. This additional handling step in the clinic could be hazardous for cell survival and iatrogenic infection, and therefore a protocol for transportation periods longer than 24 hours requires further optimisation.

Although the reduction in cell viability was small, in vivo delivery of an increased number of dead cells potentially has two adverse consequences: first, a reduced efficacy; and second, the presence of dead cell debris, which may induce inflammation. While the first consequence can be compensated for by higher cell numbers, this compounds the problems of the second.

Human MSCs stored in saline at either 4°C or room temperature showed rapidly decreasing colony-forming unit frequency as storage times exceeded 1 hour, despite a population viability of over 90%, indicating that simple measurement of viability may not fully reflect either the self-renewal or proliferation capacity of MSCs [24]. Additionally, the differentiation potential (assessed by trilineage differentiation) gradually decreased over time, regardless of storage temperature, when compared with that of freshly harvested MSCs from culture [24]. This study therefore utilised more than one assessment of cellular health to evaluate effects other than immediate cell death. Other studies have reported that viability and osteogenic potential of human MSCs were not significantly altered by storage for 24 hours in phosphate-buffered saline [25] and that storage of up to 6 hours in an electrolyte solution (plasmalyte A) had no significant adverse effect on human MSCs, as measured by viability, differentiation capability and expression of cell surface markers [26]. These studies suggest that a loss in differentiation function is not an insurmountable, inevitable consequence of storage or shipment. In the current study, cells suspended in physiologic saline maintained surprisingly good viability and high rates of proliferation following 24 hours of storage, which is in line with previously reported, short-term viability in saline [24].

The reason for a more rapid rate of cell death in cells suspended in BMA, in comparison with saline or plasma, is not clear. The acellular fraction of bone marrow aspirate contains various growth factors and cytokines, and its composition may gradually alter during storage, with degradation of some beneficial growth factors and a gradual deterioration in quality, with a correspondingly negative influence on cell viability.

The current protocol for treating clinical tendon injuries involves a combination of MSCs and BMA [1, 9] although the relative contributions of each to tendon healing have not been clarified. Maintenance of the current protocol, which has been demonstrated to be both safe and efficacious [9] with increased emphasis on the importance of implantation within 24 hours of dispatch, should ensure sufficient cells remain viable.

This study utilised allogeneic BMA, serum and plasma, due to the practical and ethical difficulties in securing autologous fluid from each horse for which MSCs were available for experimental use. Future clinical use of allogeneic serum-containing transport media may necessitate heat inactivation of complement to further optimise cell viability. Pilot studies performed in our laboratory suggest that heat treatment of allogeneic plasma, bone marrow aspirate and platelet-rich plasma (56°C for 1 hour) results in a small, but measurable, improvement in cell viability (J. Dudhia, unpublished observations). The particular role of complement in cell death observed in this study is unknown and further work should be carried out to investigate this effect.

We have demonstrated that although the needle size did not have an effect on immediate cell viability, or on differentiation potential, there were significant effects on cellular processes. A cell damaged, but not killed, by the process of injection could enter one of three states: viable with a desired phenotype (that is, capable of differentiation and proliferation), surviving but dysfunctional, or apoptotic [18]. Dysfunctional cells could be further subdivided into those that sustain mechanical stress but remain able to restore their original capabilities, those with permanent phenotypic change, and those that later become apoptotic. The decrease in cell metabolic activity observed immediately post injection indicates that the injection process does alter cell health, although the subsequent increase in cell metabolic activity by 2 hours post injection suggests that the cells have recovered their original capabilities, as shown in other studies [18]. Although it is likely that the mechanism of action of MSCs in vivo involves processes more complex and wider reaching than simple in situ differentiation, we have demonstrated that this facility remains intact following injection.

The proportion of early apoptotic cells present immediately post injection decreased over the subsequent 24 hours. This correlates with the early stage in apoptosis at which externalisation of cell membrane phosphatidylserine residues occurs and maximum Annexin V binding would be expected. Indeed, the time from initiation of apoptosis to completion can occur in as little as 2 to 3 hours [14], which may explain the rapid decrease in the proportion of early apoptotic cells from 4 hours post injection. Although we did not measure caspase 3 levels or activation in the present study, analysis of this protein in future studies may enable the apoptotic pathways to be consolidated in better detail.

The peak in early apoptotic cells for the 19G group occurred at the 2-hour time point, whereas the peak for the 21G and 23G groups was immediately post injection. This observation may represent a difference in the intensity of mechanical sheer stress between the needle gauges, and supports previous observations that an increase in needle diameter decreases cell damage in human MSCs [19]. The smaller 21G and 23G needles may exert sufficient cell stress to induce apoptosis immediately, whereas the 19G needle may induce a lesser and more limited stress responsible for fewer cells undergoing delayed apoptosis. Indeed, variations in the fluid flow shear stress caused by needle walls have been shown to induce subtle phenotypic alterations and alter cell fate [27].

We have demonstrated that, even following optimisation of these two necessary processes, a degree of cell death and altered cell function, albeit temporary, is to be expected. These results have implications for the efficacy of intralesional injection of MSCs for tendon repair. Every effort should be made to ensure that cells are reimplanted within 24 hours of resuspension in the laboratory and a needle with a bore larger than 21G should be used for injection.