Mesenchymal stem cells (MSCs) have attracted attention as a potential therapy for Acute Respiratory Distress Syndrome (ARDS). At the same time, the use of extracorporeal membrane oxygenation (ECMO) has increased among patients with severe ARDS. To date, early clinical trials of MSCs in ARDS have excluded patients supported by ECMO. Here we provide evidence from an ex-vivo model of ECMO to suggest that the intravascular administration of MSCs during ECMO may adversely impact the function of a membrane oxygenator. The addition of clinical grade MSCs resulted in a reduction of flow through the circuit in comparison to controls (0.6 ±0.35 L min-1vs 4.12 ± 0.03 L min-1, at 240 minutes) and an increase in the transoygenator pressure gradient (101±9 mmHg vs 21±4 mmHg, at 240 minutes). Subsequent immunohistochemistry analysis demonstrated quantities of MSCs highly adherent to membrane oxygenator fibres. This study highlights the potential harm associated with MSC therapy during ECMO and suggests further areas of research required to advance the translation of cell therapy in this population.
Statistics from Altmetric.com
There is growing interest in the potential of mesenchymal stem cells (MSCs) as a therapy for acute respiratory distress syndrome (ARDS). Phase I studies have been reported,1 2 while a larger phase II study has recently completed recruitment (NCT02097641). Several more are under way (NCT02611609, NCT03042143, NCT02804945). However, these trials have excluded patients supported by extracorporeal membrane oxygenation (ECMO), which is increasingly used in patients with severe ARDS.3 While the use of MSCs during ECMO has been described in case reports,4 the compatibility of MSCs and ECMO has not been systematically evaluated. Here, we present data to suggest that the intravascular administration of MSCs during ECMO may have consequences for oxygenator function.
Ex-vivo ECMO was conducted based on our previously described simulated model.5 Briefly, permanent life support circuits (Maquet, Germany) incorporating a Quadrox D Oxygenator, a RotaFlow pump and Bioline tubing were used. Circuits were primed with 500 mL 0.9% sodium chloride (Baxter, Australia), after which this was exchanged for fresh human whole blood (final volume, 420±50 mL). Circulation was commenced at 2000 revolutions per minute (RPM), using a smooth transition clamp to provide a resistance permitting a blood flow of 4 L/min. After baseline sampling, 10 mL of calcium chloride (CSL, Australia), 12 mL of 8.4% sodium bicarbonate (Baxter) and 400 units of sodium heparin (Pfizer, Australia) were added to the circuit to achieve a pH of between 7.3 and 7.5 and an activated clotting time (ACT) ≥180 s. Pressure across the oxygenator was measured using a silicone-based pressure transducer (Omega Engineering, USA). Circulating blood temperature was maintained at 37°C±0.5°C. A solution of saline, adenine, glucose and mannitol (Macopharma, Australia) was infused at a rate of 4 mL/hour to ensure pressure equilibrium within the circuit. Fresh gas flow was set at 2.5 L/min (5% CO2, 21% O2, 74% N2).
Clinical-grade induced pluripotent stem cells (iPSCs) derived human MSCs (Cynata Therapeutics, Australia) were obtained and stored in the vapour phase of liquid nitrogen until use. Cells were suspended in a vehicle composed of Plasmalyte-A (Baxter) (57.5%), Flexbumin 25% (Baxter) (40%) and dimethyl sulfoxide (DMSO 2.5%). iPSC-derived MSCs used in these experiments were between passages 3 and 5. Fourteen ex-vivo ECMO experiments were undertaken, divided as follows: circuits injected with 40×106 MSCs, circulated for 4 hours or until flow reached 0 L/min (n=4); circuits injected with 20×106 MSCs, circulated for 4 hours or until flow decreased by 25% from baseline (n=4); and control circuits, circulated for 4 hours (n=6). When physiological conditions were obtained (PaO2 ≥80 mm Hg, PaCO2 30–50 mm Hg), MSCs in vehicle were thawed to room temperature (>95% viability), agitated to eliminate clumping, and then immediately administered to the circuit, after the oxygenator and before the pump head, by slow injection over 30 s.
Whole blood samples were collected from the circuit at 30 s, 15, 30, 60, 120 and 240 min. Erythrocytes were lysed, and residual cells were washed as described previously.6 Fc receptors were blocked using Human TruStain FcX (BioLegend, USA), and MSCs were stained with mouse antihuman monoclonal antibodies—CD45-PECy7 (BioLegend), CD73-PerCP, CD90-APC and CD105-FITC (Abcam, UK)—according to the manufacturers’ protocol. Precision Count Beads (BioLegend) were then added to determine the fate of circulating MSCs, quantified using a two-laser FACSCanto I flow cytometer (BD Biosciences, USA).
After termination of each experiment, oxygenators were flushed with 1000 mL 0.9% NaCl, perfused with 500 mL paraformaldehyde (Merck, Germany) and rinsed again with 1000 mL 0.9% NaCl. Samples of the heat and gas exchange fibres were then retrieved.
For immunohistochemistry analysis, fibre sections were blocked for 2 hours in phosphate buffered saline (PBS) containing 2% heat inactivated sheep seerum (HISS) (Sigma, Australia) and 0.5% Triton X-100 (Sigma). Samples were incubated overnight at 4°C with primary antibodies: mouse Ab to CD105-FITC (Abcam, 1:10), or mouse Ab to CD90-PE (Abcam, 1:50), and rabbit antibody to β1 tubulin-488 conjugated (Abcam, 1:50). The fibres were washed with PBS followed by 1-hour incubation with secondary antibodies to boost the fluorescent signal followed by additional PBS washes. Images were acquired with a wide-field Nikon deconvolution (TiE) microscope, and confocal z-stacks were acquired with Zeiss LSM710 Airyscan.
In all circuits to which MSCs were administered, blood flow through the ECMO oxygenator decreased by at least 25% within 4 hours, with a corresponding increase in the transoxygenator pressure gradient (table 1). When 40×106 MSCs were administered, flow was reduced to <1.5 L/min in all circuits by 4 hours, and in one case within 30 min. ACT was maintained ≥180 s throughout all experiments. Microscopy of deconstructed oxygenators demonstrated the widespread adherence of MSCs to plastic fibres (figure 1).
This is the first study to directly address the feasibility of MSC therapy during ECMO. Our data suggest that intravascular administration of MSCs during ECMO may have important consequences for oxygenator function, as well as for their efficacy as a therapy for severe ARDS in this setting. This may have occurred due to the characteristic plastic adhesiveness of MSCs. A limitation of our study is that we did not assess gas exchange across the oxygenator.
MSC use during ECMO has been described previously but has either been administered before the commencement of ECMO,7 by intratracheal administration8 or during a pause in flow.4 These methods of administration may not always be possible in severely ill patients with ARDS who are reliant on continuous high-flow ECMO for oxygenation.
At a circuit concentration of 48–95×103 MSCs/mL, our study may have underestimated the effect of MSCs on oxygenator performance. Previous studies in ARDS have used up to 10×106 MSCs/kg,1 which assuming an average blood volume of 70 mL/kg9 and equal distribution would result in a higher circulating cell concentration (approximately 143×103 MSCs/mL). This must be weighed against the possibility that indirect intravascular administration, peripherally or after the membrane oxygenator, may reduce the number of MSCs reaching the oxygenator surface, most likely as a result of entrapment in the pulmonary circulation. These data also support further investigation of MSC therapy during extracorporeal carbon dioxide removal or cardiopulmonary bypass, both of which share functionally similar membrane gas exchange devices with ECMO.10
Alternative routes of administration, such as intrabronchial, may have advantages in the setting of ECMO, when prolonged bronchoscopy is possible. This should be evaluated in any future study. In light of these results, further investigations using MSCs in ECMO should explore safety considerations in an intact animal model before progression to clinical trials.
Contributors JEM, DFM and JFF conceived and designed the study. VvB and MVM designed and modified the ex-vivo ECMO model. JEM, VvB and MVM conducted the ex-vivo experiments. KKK, MAR, NB and JYS analysed the generated samples. All authors have participated in the drafting and critical revision of the manuscript.
Funding This study was supported by the Intensive Care Society (UK) New Investigator Award and by the National Health and Medical Research Council (NHMRC), Australia (APP1079421). VvB is funded by Gålöstiftelsen Foundation, Olle Enqvist Byggmästare Foundation, Karolinska Institutet Fernström Foundation, and the Royal Swedish Academy of Sciences, Claes Adelsköld Foundation.
Competing interests Cynata Therapeutics provided mesenchymal stem cells in kind to this study. The company was not involved in the commission, design or analysis of the study.
Ethics approval The study was approved by the Metro North Ethics Committee (HREC/16/QPCH/221).
Provenance and peer review Not commissioned; externally peer reviewed.
Correction notice This article has been corrected since it was published Online First. Errors with in Table 1 was corrected.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.