Novel delivery of cellular therapy to reduce ischemia reperfusion injury in kidney transplantation

Ex vivo normothermic machine perfusion (NMP) of donor kidneys prior to transplantation provides a platform for direct delivery of cellular therapeutics to optimize organ quality prior to transplantation. Multipotent Adult Progenitor Cells (MAPC®) possess potent immunomodulatory properties that could minimize ischemia reperfusion injury. We investigated the potential capability of MAPC cells in kidney NMP. Pairs (5) of human kidneys, from the same donor, were simultaneously perfused for 7 hours. Kidneys were randomly allocated to receive MAPC treatment or control. Serial samples of perfusate, urine, and tissue biopsies were taken for comparison. MAPC‐treated kidneys demonstrated improved urine output (P = .009), decreased expression of injury biomarker NGAL (P = .012), improved microvascular perfusion on contrast‐enhanced ultrasound (cortex P = .019, medulla P = .001), downregulation of interleukin (IL)‐1β (P = .050), and upregulation of IL‐10 (P < .047) and Indolamine‐2, 3‐dioxygenase (P = .050). A chemotaxis model demonstrated decreased neutrophil recruitment when stimulated with perfusate from MAPC‐treated kidneys (P < .001). Immunofluorescence revealed prelabeled MAPC cells in the perivascular space of kidneys during NMP. We report the first successful delivery of cellular therapy to a human kidney during NMP. Kidneys treated with MAPC cells demonstrate improvement in clinically relevant parameters and injury biomarkers. This novel method of cell therapy delivery provides an exciting opportunity to recondition organs prior to transplantation.


| INTRODUC TI ON
The UK kidney transplant waiting list stands at over 5000 patients. To bridge the gap betweensupply and demand, there has been increased use of donation after circulatory death (DCD) and extended criteria donors (ECD). 1 Concerns regarding inferior outcomes from DCD and ECD organs can lead to underutilization of this valuable resource. 2,3 DCD and ECD kidneys are more susceptible to ischemia reperfusion injury (IRI) manifesting as delayed graft function (DGF) and this can diminish long-term graft-survival. 4,5 IRI is the result of hypoxia followed by restoration of blood flow leading to microvascular dysfunction, inflammation, immune activation, and tissue injury. 6 As the transplant community becomes increasingly reliant on marginal donors, new therapeutic approaches to reduce IRI and optimize utilization of kidneys are leading to a greater focus on improving organ preservation.
Normothermic machine perfusion (NMP) is a method of organ preservation that facilitates restoration of cellular metabolism, reviving the organ ex vivo to resume normal physiological functions. 7 Over the last few years, a number of NMP techniques and commercially available devices have been adopted into clinical practice for kidney, heart, liver, and lung transplantation. 8-10 Kidney NMP was first described in 2008. 11 In this system, a pediatric cardiopulmonary bypass machine and membrane oxygenator perfuse an ex vivo kidney with oxygenated red blood cells suspended in crystalloid at 37 o C. 12 A UK multicenter phase III randomized controlled trial (RCT) is currently underway investigating its potential to minimize DGF. 13 NMP provides a unique opportunity to deliver organ-directed, reconditioning therapies in an isolated platform to a metabolically active organ. 14 This could be transformative, facilitating direct delivery of cell therapies and reducing off-target, systemic effects in recipients. A recent study described the use of perfusion technology to preserve a liver for up to 1 week, providing a long potential therapeutic window for optimization and reconditioning. 15 Multipotent adult progenitor cells (MAPC) are adult, bone-marrow derived, mesenchymal origin, stromal cells. 16 MAPC cells represent an attractive "off-the-shelf," nonimmunogenic cell therapy option as they lack major histocompatibility complex (MHC) Class II, or costimulatory molecules (CD80, CD86, and CD40) and have low expression of MHC Class I. 17 MAPC cells release anti-inflammatory, immunomodulatory and pro-tolerogenic cytokines thereby limiting infiltration and proliferation of pathogenic immune cells. 18,19 In the transplant setting, a rat heterotopic heart transplant model demonstrated allogeneic MAPC cells could successfully replace standard pharmacological immunosuppression to maintain long-term graft survival. 20 There have also been a number of successful clinical trials harnessing the immunomodulatory potential of MAPC treatment for graft vs host disease, 21 acute respiratory distress syndrome, 22 myocardial infarction, 23 and an ongoing phase III clinical trial in ischemic stroke. 24 In 2015, the first successful use of allogeneic MAPC therapy in human liver transplantation was also reported. 25 The recipient's leucocyte population displayed diminished immune activation and a pro-tolerogenic profile.
MAPCs and mesenchymal stromal cells (MSC) are inherently very similar. MSCs were also originally isolated from adult bone marrow. MSC culture and expansion characteristics differ from MAPCs resulting in phenotypically different populations; however, they maintain a very similar mechanism of action. 26 There has been a successful RCT investigating autologous MSC therapy vs standard induction immunosuppression (basiliximab) in live donor kidney transplantation. Recipients who received MSC therapy were more likely to recover renal function faster, had decreased rates of rejection, less opportunistic infections, and better renal function at 1 year. 27 However, this has not translated into widespread use of MAPC/MSC therapy in kidney transplantation. One of the main hurdles preventing widespread translation is successful delivery of the cells to the required organ. 28 A number of animal models have explored the possibility of delivering a cell therapy during ex vivo NMP. [29][30][31] To date, there have been no reported studies successfully administering a cell therapy to a human organ during ex vivo perfusion.
Our study aims to investigate the possible benefit of delivering MAPC cell therapy directly to marginal human kidneys in a preclinical NMP model.

| Human kidney normothermic machine perfusion
On arrival at our center kidneys were surgically prepared on ice.
The renal artery was cannulated to facilitate connection to the NMP circuit. The ureter was also cannulated so urine output could be measured and sampled. Kidney pairs were perfused simultaneous for 7 hours with an oxygenated red-cell-based perfusate at a mean temperature of 36.5°C and mean arterial pressure of 75 mm Hg, according to published protocols. 32 The volume of perfusate in the circuit was kept constant by matching the urine output with a crystalloid solution via continuous infusion. All physiological parameters accessible during NMP were recorded and analyzed including: perfusate blood gas analysis, biochemical analysis, urine production, and flow rate and scored according to the validated "quality assessment tool". 33 The kidney histopathology pre-and postperfusion was assessed by a consultant histopathologist ( Figure S2D).

| MAPC treatment
The MAPC cells used in this study were obtained in collaboration with Athersys Inc (Cleveland, OH). The MAPC cells were researchgrade MultiStem® cultures isolated from human bone marrow with consent from a single healthy donor as previously described. 34 The MAPC cells were confirmed to express phenotypic markers, 35 be of >95% purity, 36 and were fluorescently labeled with cytoplasmic dye CellTracker Red CMPTX (ThermoFisher,Waltham, MA).
For MAPC treatment of kidneys during NMP, following 60 minutes of perfusion, the right or left kidney was randomly allocated to a prescribed MAPC dose (50 × 10 6 cells). Immediately prior to infusion the MAPC cryovial aliquots were gently thawed in a water bath at 37°C and for delivery cells were resuspended in 10 mL of perfusate.
This was infused via a 3-way tap on the arterial cannula ( Figure S3).
Control kidneys simply received a 10 mL bolus of crystalloid. The cell dose was extrapolated from phase II/III clinical trials of systemic MAPC therapy taking into account the average weight of a human kidney and volume of circulating perfusion fluid. 29

MAPC cells viability was evaluated using a AnnexinV/Propidium
Iodide flow cytometry assay and revealed 90.9% viability immediately after thawing prior to infusion into the NMP circuit. To gauge potency, a T cell proliferation assay demonstrated MAPCs had immunomodulatory capability immediately upon thawing at various MAPC:peripheral blood mononuclear cells ratio ( Figure S2E&F).
This technique utilizes microbubbles of inert gas (sulfur hexafluoride in a phospholipid shell) to increase ultrasound signal return.
Each bubble is approximately 2-3 μm in diameter, allowing it to pass through the capillary bed but not the interstitium. CEUS has made it possible to assess the distribution of perfusion at a microcirculation level. 37 We have previously described this technique on porcine kidneys on a cold perfusion circuit. 38 To investigate tissue perfusion CEUS on kidneys pre-and post-MAPC infusion. CEUS was performed using a Philips EPIQ7 Ultrasound machine. During NMP CEUS recordings were taken at 60 minutes (before MAPC infusion) and 4 hours later. Microflow imaging designed to detect blood flow within the small vessels at high resolution with minimal artefact was also performed using the Philips EPIQ7 machine. Detailed methodology of this technique is included in the Data S1.

| Mesoscale Discovery™ Multiplex
To facilitate the measurement of multiple proteins of interest in 1 assay, a custom electrochemiluminescent Mesoscale Discovery™

| High performance liquid chromatography
Protein was precipitated with 1:10 perchloric acid 60%; samples were centrifuged, filtered and, then, analyzed by high performance liquid chromatography (HPLC) to quantify kynurenine and tryptophan as described. 39 Culture supernatants and perfusate were acidified with sodium acetate, pH = 4, to give a final concentration of 15 mM.

| In vitro endothelial cell line perfusate stimulation model
An immortalized endothelial cell line (human microvascular endothelial cell line [HMEC]-1 cells, ATCC) was used to investigate the impact of the perfusate secretome on endothelial function. The perfusate from the NMP series was centrifuged at 1500 G for 10 minutes at 4°C and the acellular portion reserved. HMEC-1 cells were cultured to confluence and were treated with the acellular perfusate diluted in modified MCDB131 media. In chamber slides, cells were stimulated with 25 µL of acellular perfusate to evaluate protein expression using immunohistochemistry. In 6-well plates, HMEC-1 cells were stimulated with 250 µL of acellular perfusate to evaluate gene expression using reverse transcription polymerase chain reaction (RT-qPCR). In both situations, cells were stimulated for 4 hours. RNA extraction was performed on-column using Qiagen RNeasy Plus Mini-kit as per manufacturer's instructions. cDNA synthesis was carried out using the Tetro cDNA synthesis kit. RNA sequence quantification was carried out using TaqMan RT-qPCR on a StepOnePlus™ Real-Time PCR System. The following primers were used; ICAM1 TaqMan Hs00164932_m1, S1PR1TaqMan Hs00173499_m1, and GAPDH TaqMan Hs02786624_g1 as a housekeeping gene.

| MAPC tracking in kidney tissue
MAPC cells were prelabeled with CellTracker CMPTX Red fluorescent dye (CT24552, ThermoFisher) facilitating cell tracking to evaluate partitioning and engraftment within the kidney. Core biopsies were taken at time 0, 1 hour, 4 hours after cell infusion and wedge biopsies at the end of perfusion. Samples were fixed in formalin and stored as paraffin embedded blocks. These were subsequently cut into 4 µM sections for fluorescence microscopy imaging using a Zeiss Axioimager.
Coverslips were mounted with Vectashield Antifade Mounting medium with DAPI (Vectorlabs). Fluorescent imaging was performed using a Zeiss Axioimager or for high resolution Leica SP8 UV AOBS within 2 weeks of staining. Images were processed and analyzed using Zen, LAS X, or Fiji software.

| Statistical analysis
Continuous variables are reported as mean ± standard error of mean where appropriate. Data were assessed for normality, t tests, or analysis of variance (ANOVA) were then applied as appropriate.

| Kidneys included in preclinical series
Five pairs of kidneys (n = 10) were included in the study ( Table 1).
The donors included represent a heterogeneous cohort with ages ranging from 52 to 77 years, but all were from either DCD donors or had characteristics consistent with ECD status. Cold ischemic times were significantly extended due to the delays inherent in a kidney being offered for research only purposes. All kidneys pairs were deemed untransplantable due to a suspicion regarding nonrenal malignancy identified at retrieval.

| Renal physiology
Serial measurements of physiological parameters were recorded during NMP. Perfusate samples were analyzed in real time to assess adequate oxygenation, metabolic requirements, and biochemical parameters and compared between the kidney pairs; this demonstrated equivalent organ physiology associated with MAPC cell infusion. For physiological parameters such as, potassium, lactate, renal blood flow, and renal resistance, the kidneys were well matched throughout the 7-hour perfusion timeline ( Figure 1A-D).
During NMP the ureter of the kidney was cannulated. In the kidneys treated with MAPC cells, there was significantly higher urine output compared to control kidneys, P = .009 ( Figure 1E).

| Kidney injury biomarkers
NGAL and KIM-1 represent promising biomarkers in acute kidney injury or transplantation research. 40 During NMP we can analyze these biomarkers to evaluate reconditioning. We assessed urinary KIM-1 and NGAL concentration over time and compared treatment groups. This demonstrated that MAPC-treated kidneys showed a significantly lower concentration of NGAL P = .012 ( Figure 1F).
Concentrations of urinary KIM-1 were matched between both treatment groups throughout perfusion ( Figure S1A).
We also investigated if MAPC therapy had an effect on flavin mononucleotide (FMN) concentration. FMN has emerged as a promising biomarker of organ viability during liver ex vivo perfusion.
Increased FMN measured the perfusate is a marker of mitochondrial complex 1 injury. 41 There was a nonsignificant reduction in FMN production in MAPC-treated kidneys ( Figure S1B).

| Ex vivo ultrasound
To determine if the administration of a MAPC cell bolus was associ-   (Figure 2A). There was also a significant increase in anti-inflammatory cytokine IL-10, P = .047 ( Figure 2B). Alongside this was a nonsignificant reduction in other pro-inflammatory cytokines; IL-6, IL-1α, IL-17, and IL-8 but his was not the case for all cytokines; TNFα, MIP-1β, IL-2, and IFNγ.
(Figure S1 C).  This revealed that MAPC-treated kidneys had significantly higher IDO activity (elevated Kyn:Trp ratio) following 7 hours of perfusion when compared to IDO activity in paired control kidneys, P = .050

| MAPC perfusate secretome effect on HMEC-1 endothelial cell line model
An in-vitro endothelial cell line model was established to better understand the impact of MAPC perfusate secretome on the endothelium and microvascular integrity following reperfusion.
Samples of perfusate taken after 7 hours of NMP were added to HMEC-1 cells in culture. ICAM-1 (activation status) and S1PR1 (microvascular barrier integrity) expression was analyzed in the HMEC-1 cells in response to perfusate from pairs of kidneys (control perfusate vs MAPC perfusate) ( Figure 2D-G). ICAM-1 protein expression was significantly increased in the control group, P < .001. However, in the MAPC-treated group this upregulation was not as marked, P = .022. In contrast, S1PR1 gene and protein expression in HMEC-1 cells was downregulated by control perfusate; however, MAPC perfusate maintained S1PR1 gene expression at unstimulated levels, P = .032; and increased protein expression when compared with control perfusate, P < .001.
This preservation of S1PR1 protein was also seen on immunofluorescence in the tissue of NMP perfused kidneys, P = .030 ( Figure 2H-J).   Figure S2B,C).

F I G U R E 1
Establishing the feasibility of MAPC therapy during kidney NMP. Panels A-D depict real-time physiological parameters measured during the NMP timeline. Panel (A) electrolyte balance (potassium), (B) cellular metabolism (lactate), (C) renal blood flow and (D) intrarenal resistance recorded throughout the perfusion timeline. Panel E depicts the impact of MAPC treatment on urine output; n = 5pairs, paired t-test, **P < .01 (MAPC-treated kidneys mean urine output 1436mL ± 281 vs control 960mL ± 189). Panel F demonstrates MAPC treatment effect on a biomarker of kidney injury NGAL concentration measured at serial time points during perfusion, repeatedmeasures 2-way ANOVA with Sidak's multiple comparison's post-hoc test, *P < .05 **P < .01 (MAPC-treated mean NGAL concentration 7196pg/ml ± 379 vs control 12400pg/ml ± 1014 at 7 hours). Panel G demonstrates still images taken from one pair of kidneys recorded during Doppler Ultrasound MicroFlow Imaging. Blue colouring represents overlaid Doppler signal from blood flow within microvasculature. Cytokine profiling revealed that MAPC therapy during kidney NMP was associated with an anti-inflammatory, pro-tolerogenic cytokine profile. This included decreased expression of IL-1β, a pro-inflammatory cytokine associated with endothelial activation.
Decreased IL-1β expression during ex vivo lung perfusion correlates with improved survival following lung transplantation. 46 There was also upregulation of pro-tolerogenic, anti-inflammatory IL-10 and this was accompanied by increased IDO activity. MAPC cells modulating differential cytokine expression in perfusate may be key in their potential role of minimizing IRI.
Tracking MAPC cells after NMP delivery revealed cells could be found throughout the kidney. The distribution was similar to that  [56][57][58] However, there are a number of limitations in this study. NMP as an experimental system is limited in its ability to measure the posttransplantation effectiveness of a therapy. These issues . Panel D demonstrates ICAM-1 gene expression, paired t test, n = 5 pairs, p = ns (MAPC treated mean fold change on qPCR 9.6 ± 3.1 vs 6.5 ± 0.9). Panel E S1PR1 gene expression, paired t test, *P < .05 (MAPC treated mean fold change on qPCR −4.9 ± 1.7 vs control 0.6 ± 0.6). Panels F and G depict perfusate stimulated HMEC-1 cells immunohistochemistry staining for ICAM-1 and SIPR1 protein expression (control vs MAPC treated). Graph F demonstrates comparison of ICAM-1 protein expression between treatment groups (MAPC treated 23.9 ± 3.3 vs control perfusate 37.3 ± 9.2). Graph G demonstrates comparison of S1PR1 protein expression between treatment groups (MAPC treated 76.5 ± 4.7 vs control perfusate 57.6 ± 4.9). Analyzed with 1-way ANOVA and Tukey's test posthoc analysis, *P < .05, **P < .01, ***P < .001. Panels H and I depict S1PR1 staining visualized using fluorescence microscopy on kidney biopsies taken during NMP, blue stain is DAPI, red is S1PR1. Panel J demonstrates a significant upregulation of S1PR1 protein expression as quantified by mean fluorescence intensity, paired t test, n = 5 pairs, *P < .

| Conclusion
This is the first reported series of cell therapy successfully delivered directly to human donor kidneys in an isolated ex vivo perfusion platform. Kidneys treated with MAPC cells during NMP demonstrate improvement in clinically relevant parameters and a reduction in injury F I G U R E 3 Determining MAPC cell fate and physical distribution following delivery during kidney NMP. MAPC cells were prelabeled with a red cytoplasmic fluorescent dye (CellTracker Red CMPTX, ThermoFisher) to facilitate tracking of the cells through the kidney following 7 hours of perfusion. The red labeled cells are the MAPC cells as noted by white arrows. Counter nuclear staining is in blue with DAPI. The green staining is endothelial marker CD31 and highlights the kidneys vasculature. The pink staining is Aquaporin 1 a marker for proximal tubular cells. Panels A and B are sections taken from the kidney cortex after 6 hours of MAPC treatment during NMP. These represent typical confocal microscopy images from kidneys treated with MAPC cells. The white letter (G) identifies the glomerulus. Panel C is a typical section taken from the kidney's medulla; vessels identified by the CD31 green stain are identified by the green (V). Panel D is a high magnification image taken of three MAPC cells that are resident in the interstitium next to a peritubular capillary. Panel E depicts the confocal images of a blood vessel from the renal cortex with a MAPC cell in the lumen and cells that have mobilized out of the vessel into the nearby tissue. MAPC, Multipotent Adult Progenitor Cells; NMP, normothermic machine perfusion

A B
C D E and pro-inflammatory biomarkers. This effect may be mediated by changes to circulating cytokines or through secreted soluble anti-inflammatory mediators. NMP represents a novel cell therapy delivery system. This represents a paradigm shift, providing an exciting opportunity to directly treat organs prior to transplantation to minimize IRI.
A future clinical trial evaluating this modality of delivery could result in the transplantation of otherwise discarded organs, thereby reducing the transplant waiting list and offering hope to patients with renal failure.

ACK N OWLED G M ENTS
We thank all the member of the perfusion and theatre team at the Freeman Hospital for technical support. We are very grateful to staff at NHSBT Barrack Road Histocompatibility and Immunogenetics for help with tissue typing.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data related to this study are present in the paper or the Data S1 found at the end of this article. The data that support the findings of this study are available from the corresponding au-