Dialysis preserves heart function during ex situ heart perfusion

Abstract

Background

Ex situ heart perfusion (ESHP) has been used to optimize donor organs before heart transplantation. However, cardiac function often deteriorates with the development of myocardial edema. The use of dialysis during ESHP could assist in cardiac preservation.

Methods

Male Yorkshire pig hearts were subjected to ESHP for 8 hours with or without dialysis. Hearts were supported during nonworking mode and working mode, and pressure-volume loops and coronary vasomotor function were evaluated. Finally, tissue biopsies were assessed for mitochondrial function, oxidative stress, and inflammation.

Results

Adding dialysis to ESHP significantly enhanced cardiac function, with improved preload recruitable stroke work at 4 hours (64.09 ± 20.13 vs 35.08 ± 13.52, p = 0.010) and 8 hours (64.31 ± 9.08 vs 23.30 ± 19.25, p = 0.0002), maximal elastance at 8 hours (24.67 ± 10.75 vs 10.62 ± 8.471, p = 0.0477), and end diastolic pressure volume relationship at 8 hours (644.7 ± 566.68 vs 86.63 ± 72.05, p = 0.0187). Coronary vasomotor function improved in the dialysis group in endothelium dependent (LogIC50 −7.39 ± 0.25 vs −2.22 ± 0.76, p < 0.0001) and independent (LogIC50 −6.11 ± 0.19 vs −4.79 ± 0.11, p < 0.0001) vasorelaxation. Dialyzed hearts also had reduced sensitivity to endothelin-1 (LogEC50 −7.94 ± 0.5 vs −8.54 ± 0.06, p = 0.0449) and significant changes in endothelin receptor-related protein expression related and oxidative stress.

Conclusions

The combination of dialysis with ESHP improves myocardial and coronary vasomotor preservation and may allow for longer perfusion times.

Background

Patients with heart failure have a poor prognosis with high rates of mortality and hospital admission.1, 2 While improved medical management and utilization of mechanical circulatory support have alleviated some of the mortality associated with the diagnosis, many patients still progress to develop end-stage heart failure. For these patients, heart transplantation remains the gold standard treatment.³ Unfortunately, the demand for donor hearts far exceeds supply,4, 5 with substantial wait times. Therefore, transplant centers have extended their selection criteria to accept older donors, with an increased number of comorbidities, and from greater geographic distances.⁶ Even with extended criteria, only 28% of potential donor hearts are utilized.7, 8 Recently, heart transplant programs have used another donor pool, namely donation after circulatory death (DCD).9, 10 Contrary to previous limitations, current practices in DCD heart donation include preprocurement evaluations, such as echocardiograms, and established protocols for resuscitation and viability assessment of the heart, either in situ or ex situ. This enhances the potential for effective use of DCD donors in heart transplantation.¹¹ Both normothermic regional perfusion and direct procurement with ex situ perfusion are accepted methods for the resuscitation and evaluation of DCD hearts.¹² While normothermic regional perfusion has emerged as a cost-effective and viable method for DCD hearts, ex situ heart perfusion (ESHP) provides the potential to transport hearts across further distances and as a reparative platform.

Ex situ technology has been successfully applied to recover human lungs, livers, and kidneys.¹³ Its implementation limits cold ischemia time, allows for the assessment and optimization of donor organs before transplantation, and has increased the donor pool by 30% with graft function, survival, and quality of life similar to conventional brain death organ utilization.12, 14, 15, 16 Similarly, ESHP is associated with decreased allograft cold ischemia time and the ability to evaluate some predictive elements of myocardial performance.17, 18 However, clinical adoption of ESHP has been limited due to progressive myocardial functional decline12, 19 and worse outcomes post-transplant with prolonged ESHP.²⁰ Without the ability to prolong support of donor hearts by ESHP, this platform will have limited applications.

The functional decline of hearts on ESHP occurs in a time-dependent fashion with evidence of edema, cell death, and metabolic dysregulation.¹⁹ Importantly, the degree of myocardial dysfunction has been out of proportion to the degree of abnormal histological findings.²¹ We need to extend ESHP to allow this technology to be transformative. Herein, we describe the use of dialysis during ESHP to help to reduce myocardial edema and preserve myocardial function. With the ability to improve functional outcomes during ESHP, we may then be able to extend ESHP to become a reparative platform, in addition to its current utility for transplantation.

Methods

Heart preparation

The experimental protocol was approved by the Toronto General Hospital Research Institute Animal Care Committee (AUP #6417). Animals were treated according to the “Guide for the Care and Use of Laboratory Animals.” Male Yorkshire pigs (42 ± 2 kg) were anesthetized with ketamine (30 mg/kg) and atropine (0.04 mg/kg) given intramuscularly, followed with 2% inhaled isoflurane after endotracheal intubation. A median sternotomy was performed, and the pericardium was opened. Heparin (30,000 U) was given intravenously and blood was collected and passed through a leukocyte filter (Imugard III, Terumo, Tokyo, Japan). The heart was then arrested using histidine-tryptophase-ketoglutarate solution (HTK) cardioplegia, resected, and then cannulated for ESHP as previously described.20, 22, 23, 24 In brief, 1 liter of HTK was administered by gravity through the aortic root cannula. The heart was excised and placed in ice cold HTK for 1 hour of standardized cold storage. During cold storage, the left atrium, aorta, and pulmonary artery were cannulated.

Ex situ heart perfusion

Our ESHP platform and methodology have been described before (Figure 1).²⁴ In brief, the ESHP platform was primed with 2 liter of leukocyte-depleted blood, 1 g of cefazolin, 2 g of magnesium sulfate, 500 mg of methylprednisolone, 20 ml of 8.4% sodium bicarbonate, and 10,000 U of heparin. The fraction of inspired oxygen and oxygen flow through the oxygenator (sweep) were adjusted to maintain a pH of 7.4, PO₂ of 100 to 300 mm Hg, and PCO₂ of 35 to 45 mm Hg.

Figure 1.

A: Experimental design; B: Dialsys ESHP setup diagram. The perfusate is stored in the reservoir and pumped into the oxygenator and heater through the centrifugal pump. Perfusate is split into the dialysis circuit and the heart circuit. In the dialysis circuit, a roller pump pushes perfusate into the dialysis filter and subsequently back into the reservoir. Dialysate flows in a countercurrent fashion against the perfusate and is directed into a waste bag. In the heart circuit, perfusate is directed into the aorta in a retrograde fashion during nonworking mode. Venous return is drained from the pulmonary arteries into the reservoir. In working mode, the perfusate is split between the aorta and left atrial line.

The heart was connected to the ESHP device after 1 hour of standard cold storage in 4°C HTK. The left-sided chambers of the heart were deaired by loading perfusate into the left atrium and manual pumping of the left ventricle by hand to eject the volume out of the left ventricle. The left atrium (LA) line was clamped and retrograde aortic perfusion was commenced at a pressure of 40 mm Hg (nonworking mode). The perfusate was warmed to 37°C over 30 minutes. Dobutamine (10 mcg/min), insulin (5 U/hour), and nitroglycerin (0.5 mcg/min) were continuously infused into the perfusate at the start of reperfusion. Once the heart was rewarmed to 37°C, calcium chloride, sodium bicarbonate, and dextrose were supplemented to adjust perfusate calcium from 1.1 to 1.3 mmol/liter, glucose from 5 to 10 mmol/liter, and bicarbonate from 24 to 30 mmol/liter. Corrections were made every hour for the remainder of the experiment. Hearts were maintained in nonworking mode except for transitions into working mode at 30 minutes, 1.5 hours, 3.5 hours, and 6.5 hours for 30 minutes for functional assessment.

Dialysis

For dialysis, perfusate from the oxygenator was directed to a F3 Fresenius Medical Care dialysis filter (Fresenius Kabi, Bad Homburg, Germany) using a roller pump on the cardiopulmonary bypass machine, while the dialysate was directed in a countercurrent fashion using a 0.25-inch roller pump. After filtration, the perfusate was returned to the reservoir. The dialysate was composed of 139 mM sodium, 4 mM potassium, 0.75 mM magnesium, 109 mM chloride, 3 mM acetate, 1.1 mM dextrose, 32 mM bicarbonate at a pH of 7.4. The dialysis circuit was started 30 minutes before heart reperfusion. The perfusate and dialysate flow rates were adjusted to achieve 100 mm Hg of transmembrane pressure. To preserve dialysate, perfusate and dialysis flow rates were set to a ratio of 2:1.

Cardiac functional assessment

Left ventricular function was assessed using a conductance catheter (Ventri-Cath-507S, Millar Inc., Houston, TX) inserted transapically and recorded using IOX v1.8.9.13 software (EMKA Technologies Inc., Falls Church, VA). Load-dependent pressure-volume loop parameters were collected under steady-state loading conditions and load-independent pressure-volume loop parameters were collected by occluding the inferior vena cava (IVC) in vivo and the LA line during perfusion. The heart was weighed before perfusion and after 8 hours of perfusion. Percent weight gain was calculated by using the formula:

$$\mathit{Percent}\mathit{weight}\mathit{gain}=[(\mathit{After}\mathit{weight}-\mathit{Before}\mathit{weight})\mathrm{⁄}\mathit{Before}\mathit{weight}]\mathrm{x}100\%$$

Arterial and venous blood and perfusate samples were collected in vivo, and at 30 minutes, 1.5 hours, 3.5 hours, and 6.5 hours of ESHP for blood gas analysis (Siemens Rapidpoint 500 analyzer, Munich, Germany) to obtain measurements of pH, electrolytes, lactate, glucose, PO₂, hemoglobin, hematocrit, and oxygen saturation. Continuous pH and gas measurement in the perfusate were monitored through a Terumo Blood Parameter Monitoring System CDI500 (Terumo, Tokyo, Japan).

Endothelial functional assessment

Endothelium-dependent and endothelium-independent vasomotor functions were assessed in vitro by constructing concentration-response curves with a small-vessel myograft for isometric-tension recording. The left anterior descending coronary artery (LAD) was harvested after each perfusion and stored in 4°C Krebs Henseleit buffer (118 mM NaCl, 4.6 mM KCl, 1.2 mM MgSO₄, 2.5 mM CaCl₂, 1.2 mM KH₂PO₄, 25 mM NaHCO₃, and 11 mM glucose). Segments 5 mm in length were placed in a 25-ml organ chamber where the LAD was suspended between an anchor wire and a wire connected to an isometric force transducer. The tissue was submerged in Krebs Henseleit buffer at 37°C and oxygenated with 95% oxygen and 5% carbon dioxide. The tissue was stabilized for 90 minutes and calibrated to a resting tension of 3 g. Data were acquired with AcqKnowledge software (Biopac Systems Inc., Goleta, CA).

For endothelium dependent and independent vasorelaxation, the LAD was precontracted with a thromboxane A2 analog, U4618 compound (30 nmol/liter) and the dose response to bradykinin (0.25 nmol/liter-1.0 μmol/liter) or to sodium nitroprusside (10 nmol/liter-5 μmol/liter) was determined. Sensitivity to vasospasm was measured by dose response to endothelin-1 (ET1) (0.1 nmol/liter-30 nmol/liter). Three segments of LAD from each heart were used for assessment and data were averaged for analysis. Protein expression of endothelin A receptor and endothelin B receptor were detected by standard western blotting using specific primary monoclonal antibodies (Abcam, Cambridge, United Kingdom) at a concentration of 1:1,000 in a 5% skim milk solution. The western blotting was analyzed with Image Lab 5.0 software. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as a loading control for all blots.

Tissue analysis

Left ventricular biopsies were taken using a 20G core biopsy needle (Monopty Disposable Core Biopsy Instrument, BD, Franklin Lakes, NJ). The mitochondria were extracted using the Mitochondria Isolation kit (MITOISO1, Sigma-Aldrich, St. Louis, MI) and the cytoplasmic proteins were extracted using the NE-PER Nuclear and Cytoplasmic Extraction kit (Thermo Fisher Scientific, Waltham, MA) according to the manufacturer’s protocol. For immunoblotting, total protein extracts were prepared as previously described.²⁵ Protein samples were mixed with equal amounts of sample buffer (Novex Bolt LDS sample buffer, 4×; no. 2107345; Thermo Fisher Scientific) and boiled for 7 minutes. Samples (60-mg total protein per lane) were resolved (Novex Bolt MES SDS-Running Buffer, 20×; no. 2122845; Thermo Fisher Scientific) by sodium dodecyl-sulfate polyacrylamide gel electrophoresis (SDS-PAGE) employing 4% to 12% precast gradient gels (Novex Bolt 4%-12% Bis-Tris Plus; 1.0 mm × 10 wells; no. NW04120BOX; Thermo Fisher Scientific). Samples were electrotransferred using an iBlot system (Thermo Fisher Scientific). Detection was performed with Luminata Crescendo (WBLUR0100, Millipore) on a digital imaging system (Molecular Imager ChemiDoc Imaging System; BioRad). Primary antibodies, including asymmetric dimethyl arginine (AMDA, Cell Signaling #13522), catalase (Cell Signaling #12980), and inducible nitric oxide synthase (iNOS, Novus Biologics #NB300-605), p53 (Enzo Life Sciences #ADI-908-265-010), CC3 (Cell Signaling #14614) were used at a concentration of 1:1,000 in a 5% skim milk solution and detected by standard western blotting.

Inflammation

Perfusate was collected at 0.5, 4, and 8 hours of perfusion for measurement of cytokines by utilizing a commercial Millipore MILLIPLEX Porcine cytokine/chemokine array analyzed with a BioPlex 200 (Eve Technologies, Calgary, AB, Canada) to measure granulocyte macrophage colony stimulating factor (GM-CSF), interferon gamma (IFN-γ), IL-1α, interleukin 1 receptor alpha (IL-1ra), IL-1β, IL-2, IL-4, IL-6, IL-8, IL-10, IL-12, IL-18, and tumor necrosis factor alpha (TNF-α).

Statistical analysis

Statistical analysis was performed in Graphpad Prism version 9.4 (Graphpad Holdings LLC, San Diego, CA). Continuous data were expressed as mean ± SD. Normal distribution was tested before conducting a parametric test. The unpaired t-test was used to compare single timepoints and repeated-measures one-way analysis of variance (ANOVA) was used to compare multiple timepoints within a group. Data that did not meet the normality test were analyzed using Kruskal-Wallis test. The best fit-values of the LAD dose-response curves were compared using the extra sum-of-squares F test. Significance was set at α = 0.05 for all statistical tests.

Results

No significant differences in left ventricle (LV) function were observed between the dialysis and control groups during the pre- and post-ESHP for the first 90 minutes (Table 1). Hearts treated with ESHP + dialysis, however, did have significantly higher preload recruitable stroke work (PRSW) at 4 (64.09 ± 20.13 vs 35.08 ± 13.52, p = 0.010) and 8 hours (64.31 ± 9.08 vs 23.30 ± 19.25, p = 0.0002). In addition, hearts with ESHP + dialysis had significantly higher maximal elastance (Emax) (24.67 ± 10.75 vs 10.62 ± 8.471, p = 0.0477) and end diastolic pressure volume relationship (EDPVR) at 8 hours of ESHP (644.7 ± 566.68 vs 86.63 ± 72.05, p = 0.0187). However, there was a significant decline in maximum rate of pressure change (dP/dT max) over 8 hours with ESHP alone and ESHP + dialysis (Table 2). In a mixed-effects analysis, both treatment and time were statistically significant. While PRSW showed that the dialysis treatment affected both function and time, end systolic pressure volume relationship (ESPVR), EDPVR, dP/dT max, Tau, Emax, stroke work, and cardiac index showed that only time significantly affected function (p < 0.05). This was accompanied by the significant reduction in weight gain in the ESHP + dialysis group as compared to the control (22.25% ± 6.185% vs 72.60% ± 13.92%, p = 0.0025; Figure 2).

Table 1.

Dialysis and Control Group PV Loop Data Over Time

Variable	Group	1.5 hours	4 hours	8 hours	p-value
PRSW	Dialysis	44.49 ± 12.34	64.09 ± 20.13	64.31 ± 9.08	0.0348
	Control	48.1 ± 10.71	35.08 ± 13.52	23.3 ± 19.25	0.0164
ESPVR	Dialysis	12.42 ± 9.48	12.07 ± 7.39	13.72 ± 4.94	0.8618
	Control	9.08 ± 9.63	6.77 ± 3.45	7.51 ± 7.04	0.7204
EDPVR	Dialysis	0.64 ± 0.61	1.36 ± 1.02	3.05 ± 1.59	0.0218
	Control	0.60 ± 0.91	1.12 ± 1.46	0.96 ± 0.86	0.5456
Tau	Dialysis	9.60 ± 1.91	10.43 ± 0.43	20.5 ± 13.87	0.1831
	Control	10.59 ± 1.47	11.58 ± 1.70	20.83 ± 12.23	0.2150
dP/dT max	Dialysis	1175 ± 289.5	1136 ± 328.2	883.3 ± 174.7	0.0187
	Control	2070 ± 1162	1797 ± 872.1	1441 ± 918.2	0.0432
dP/dT min	Dialysis	−1098 ± 306.4	−1236 ± 467.2	−890.7 ± 307.2	0.0934
	Control	−1451 ± 751.7	−1314 ± 719.3	−1496 ± 788.9	0.2286
Emax	Dialysis	19.38 ± 12.98	21.78 ± 7.56	24.67 ± 10.75	0.4839
	Control	21.34 ± 8.30	14.28 ± 4.00	10.62 ± 8.47	0.0818

Abbreviations: dP/dT max, maximum rate of pressure change; dP/dT min, minimum rate of pressure change; EDPVR, end diastolic pressure volume relationship; ESPVR, end systolic pressure volume relationship; Emax, maximal elastance, PRSW, preload recruitable stroke work; PV, pressure volume; Tau, exponential decay of ventricular pressure during isovolumetric relaxation.

Value represented as mean ± SD.

Table 2.

Dialysis and Control Group PV Loop Parameter Percent Change From 1.5-Hour Assessment at 4 Hours and 8 Hours

Variable	Group	4 hours	8 hours
PRSW	Dialysis	146.89 ± 45.54	148.93 ± 23.13
	Control	72.2 ± 18.77	46.88 ± 31.75
	p-value	0.0035	0.0002
ESPVR	Dialysis	110.54 ± 34.02	148.01 ± 89.73
	Control	94.95 ± 51.99	102.01 ± 106.23
	p-value	>0.9999	0.7075
EDPVR	Dialysis	221.53 ± 141.45	644.7 ± 566.68
	Control	117.25 ± 107.07	86.63 ± 72.05
	p-value	0.8311	0.0187
Tau	Dialysis	112.54 ± 24.82	238.19 ± 206.86
	Control	109.8 ± 10.71	199.24 ± 124.48
	p-value	>0.9999	>0.9999
dP/dT max	Dialysis	97.06 ± 14.58	76.15 ± 8.06
	Control	92.68 ± 35.53	66.45 ± 11.88
	p-value	>0.9999	0.9303
dP/dT min	Dialysis	114.88 ± 34.79	84.7 ± 35.22
	Control	89.71 ± 15.1	104.2 ± 29.83
	p-value	0.3625	0.5342
Emax	Dialysis	133.94 ± 49.51	147.54 ± 49.11
	Control	73.81 ± 35.4	54.74 ± 59.2
	p-value	0.1409	0.0173

Abbreviations: dP/dT max, maximum rate of pressure change; dP/dT min, minimum rate of pressure change; EDPVR, end diastolic pressure volume relationship; ESPVR, end systolic pressure volume relationship; Emax, maximal elastance; PRSW, preload recruitable stroke work; PV, pressure volume; Tau, exponential decay of ventricular pressure during isovolumetric relaxation.

Value represented as mean ± SD.

Figure 2.

Changes in weight over duration of ex situ heart perfusion (ESHP control blue) and with dialysis (red). Data are shown as mean ± SD.

To explore the benefits of dialysis, we examined the concentrations of electrolytes in the perfusate and found a significant increase in potassium and sodium concentration over time in the control group, whereas the dialysis group maintained steady levels of both electrolytes (Figure 3). There were no significant differences between the 2 groups for hematocrit and lactate production over time.

Figure 3.

Changes in perfusate over duration of ex situ heart perfusion (ESHP) in control (blue) and with dialysis (red). (A): Potassium concentration; (B) Sodium concentration; (C) Hematocrit; (D) Lactate concentration. Data is shown as mean ± SD.

Figure 4 illustrates coronary vasomotor function from LADs²⁶ from both dialysis and control hearts. The dialysis group demonstrated significantly improved endothelin-dependent (Edep) (LogIC50 dialysis −7.39 ± 0.25, control −2.22 ± 0.76, p < 0.0001) and endothelin-independent (Eind) vasorelaxation (LogIC50 dialysis −6.11 ± 0.19, control −4.79 ± 0.11, p < 0.0001). The dialysis group demonstrated reduced sensitivity to ET1 (LogEC50 dialysis −7.94 ± 0.5, control −8.54 ± 0.06, p = 0.0449). In addition, the coronary artery from the control heart had significantly higher protein expression of endothelin A and endothelin A/B receptor ratios than the dialysis heart (Figure 5; p = 0.0485; p = 0.0051, respectively). The LV from the dialysis group had significantly less asymmetric dimethyl arginine protein than control LV samples (p = 0.0027). In contrast, dialysis hearts also had significantly higher LV expression of catalase (p = 0.0107), while having significantly lower expression of p53 (p = 0.0493) compared to control hearts. There were no significant differences in endothelin B receptor, inducible nitric oxide synthase, or cleaved caspase-3 expression between the 2 groups (Figure 6).

Figure 4.

Coronary vasomotor function with ex situ heart perfusion in control (blue) and dialysis (red). (A) endothelial-dependent vasomotor function; (B) Endothelialindependent vasomotor function; (C) Vasospasm sensitization. Data is shown as mean + SD.

Figure 5.

Immunoblotting of coronary and left ventricular tissue from ex situ hearts perfused for 8 hours. (A, B) Quantification of endothelin A (ETA in A) and endothelin B (ETB in B) normalized to GAPDH from immunoblot shown in (E). Left anterior descending artery was utilized for this analysis. (C) Ratio of normalized ETA/ETB. (D) Quantification of AMDA normalized to GAPDH in immunoblot shown in (E). *p < 0.05, **p<0.01. Data are shown as mean ± SD. AMDA, asymmetric dimethyl arginine; ETA, endothelin A; ETB, endothelin B. LV left ventricle. Coro coronary.

Figure 6.

Quantification (A-D) and immunoblotting (E) for markers of oxidative stress and apoptosis signalling in LV after 8 hours of perfuion in control (blue) and with dialysis (red). Quantifiation of catalase (A); p53 (B); CC3 (C); iNOS (D). Representative immunoblot of LV lysate using the indicated antibodies. *p < 0.05. CC3, cleaved caspase 3; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; iNOS, inducible nitric oxide synthase.

Extravasation and leukocyte infiltration due to inflammation has been thought to contribute to dysfunction and weight gain. However, the cytokine panel showed no significant differences between the dialysis and control perfusate samples at the start, 1.5 hours, 4 hours, and 8 hours of perfusion (Table 3). There were also no differences found among the LV biopsies between the 2 groups (Table 4). The postfiltration dialysate contained IFNγ, IL-1β, IL-6, IL-10, and IL-18.

Table 3.

Perfusate and Dialysate Cytokines Over Time

Cytokine	Time point	Perfusate control	Perfusate dialysis	Dialysate	p-value
GM-CSF	MB	190.81 ± 241.9	23.49 ± 25.83		0.58
	1.5 hours	0.79 ± 1.3	10.96 ± 20.41		0.80
	4 hours	13.93 ± 17.64	24.75 ± 7.08		0.69
	8 hours	18.4 ± 27.16	14.74 ± 19.77		1.00
IFNy	MB	259.95 ± 225.84	76.91 ± 70.7		0.47
	1.5 hours	16.11 ± 36.01	129.86 ± 206.4	14.26 ± 11.52	0.74
	4 hours	44.23 ± 46.01	7.93 ± 17.74	22.51 ± 12.27	0.50
	8 hours	77.12 ± 107.79	207.32 ± 266.37	20.1 ± 9.18	0.83
IL-1a	MB	0	0		1.00
	1.5 hours	8.58 ± 5.36	7.75 ± 4.24		1.00
	4 hours	86.94 ± 93.06	70.88 ± 38.84		1.00
	8 hours	460.04 ± 426.77	401.88 ± 230.13		1.00
IL-1b	MB	0 ± 0	0.9 ± 2.02		0.85
	1.5 hours	203.41 ± 202.11	146.4 ± 113.2	2.75 ± 2.35	0.97
	4 hours	2,303.29 ± 2,176.72	1,941.04 ± 1,036.7	1.66 ± 1.41	1.00
	8 hours	15,067.21 ± 9,392.12	11,619.23 ± 5,282.57	1 ± 0.72	0.94
IL-1ra	MB	0.3 ± 0.66	3.8 ± 8.5		0.88
	1.5 hours	68.78 ± 68.95	73.94 ± 78.37		>0.99
	4 hours	802.07 ± 471	1,556.63 ± 1,360.46		0.75
	8 hours	2,588.38 ± 1,690.04	3,796.31 ± 3,175.51		0.93
IL-6	MB	23.34 ± 6.32	18.3 ± 5.26		0.61
	1.5 hours	900.94 ± 393.48	687.88 ± 428.22	25 ± 4.08	0.90
	4 hours	3,771.58 ± 930.37	4,388.5 ± 713.44	27.27 ± 4.01	0.72
	8 hours	8,084.35 ± 1,455.04	9,759.36 ± 1,580.36	28.03 ± 4.17	0.40
IL-8	MB	2.81 ± 6.28	1.94 ± 2.94		1.00
	1.5 hours	280.11 ± 266.12	331.7 ± 394.7		1.00
	4 hours	21,446.38 ± 10,766.24	24,399.24 ± 14,086.2		0.99
	8 hours	63,511.07 ± 44,661.05	66,439.78 ± 50,938.85		>0.99
IL-10	MB	0 ± 0	0.34 ± 0.77		0.85
	1.5 hours	0 ± 0	1.22 ± 2.73	7.46 ± 3.59	0.85
	4 hours	1.95 ± 4.36	10.82 ± 7.92	7 ± 2.45	0.25
	8 hours	19.61 ± 10.68	67.55 ± 51.15	6.47 ± 2.05	0.36
IL-12	MB	68.01 ± 30.51	43.95 ± 21.73		0.58
	1.5 hours	102.75 ± 48.12	74.26 ± 39.7		0.81
	4 hours	126.76 ± 42.26	96.31 ± 33.41		0.67
	8 hours	202.56 ± 73.93	189.66 ± 83.32		1.00
IL-18	MB	118.68 ± 55	114.77 ± 52.81		>0.99
	1.5 hours	231.83 ± 44.47	193.46 ± 41.6	96.26 ± 19.3	0.58
	4 hours	301.9 ± 50.77	274.72 ± 32.98	99.99 ± 16.83	0.82
	8 hours	387.22 ± 119.83	348.41 ± 65.69	88 ± 14.78	0.96
TNF-α	MB	0 ± 0	10.09 ± 22.56		0.85
	1.5 hours	533.59 ± 704.6	338.26 ± 449.48		0.98
	4 hours	1,488.4 ± 990.84	645.87 ± 554.24		0.47
	8 hours	1,044.94 ± 422.81	536.57 ± 353.53		0.27

Abbreviation: GM-CSF, granulocyte macrophage colony stimulating factor; MB, ESHP baseline.

Data are expressed as mean ± SD.

Table 4.

Cytokine Concentration in the Left Ventricle and LAD

Cytokine	Location	Dialysis	Control	p-value
GM-CSF	LV	42.42 ± 12.48	39.34 ± 7.82	>0.99
	LAD	0	0	NA
IFNy	LV	232.1 ± 44.52	217.6 ± 28.34	0.80
	LAD	57.1 ± 118.3	52.35 ± 117.1	NA
IL-1a	LV	131.4 ± 43.61	153.4 ± 51.06	0.69
	LAD	73.9 ± 46.39	122.4 ± 185.9	0.69
IL-1b	LV	3121 ± 818.9	4201 ± 1198	0.54
	LAD	4154 ± 2463	5087 ± 6252	0.69
IL-1ra	LV	35.89 ± 10.04	31.61 ± 6.85	0.69
	LAD	20.81 ± 14.34	28.19 ± 36.19	0.55
IL-2	LV	40.55 ± 13.69	35.75 ± 10.73	0.73
	LAD	1.44 ± 1.99	2.09 ± 3.87	NA
IL-4	LV	22.02 ± 13.59	15.76 ± 19.37	0.31
	LAD	0	0	NA
IL-6	LV	1156 ± 201	981.7 ± 233.3	0.31
	LAD	1214 ± 242	795.7 ± 298.4	0.06
IL-8	LV	2141 ± 857.2	1781 ± 634.6	0.69
	LAD	1183 ± 464.4	980.6 ± 883.9	0.54
IL-10	LV	15.2 ± 5.91	14.32 ± 5.14	0.75
	LAD	8.58 ± 2.65	7.82 ± 2.3	0.84
IL-12	LV	1.66 ± 2.79	1.23 ± 2.75	0.72
	LAD	0	0	NA
IL-18	LV	131.8 ± 20.83	140.7 ± 18.11	0.42
	LAD	120.1 ± 13.57	121.9 ± 18.51	>0.99
TNF-α	LV	2.42 ± 5.4	1.3 ± 2.9	>0.99
	LAD	0	0	NA

Abbreviation: GM-CSF, granulocyte macrophage colony stimulating factor; LAD, left anterior descending coronary artery; NA, not applicable.

Data are expressed as mean ± SD.

Discussion

ESHP has been shown to be noninferior to traditional cold storage and is showing great promise in DCD.12, 27 Large animal and human experimental studies have shown that as perfusion time is extended, there is a significant decline in cardiac function.19, 28 While this decline is multifaceted, the removal of built-up waste products and replenishment of metabolites may help extend heart perfusion to more than 24 hours as has been demonstrated by continuous plasma exchange in neonatal pig hearts.²⁹ Therefore, we investigated the effects of dialysis with the addition of a small dialysis circuit to our current ESHP setup.²⁴

The addition of dialysis to ESHP significantly reduced weight gain of the hearts with an associated improved PRSW over time as compared to the control group. In addition, we found that dialysis significantly increased EDPVRs over time. Interestingly, EDPVR is a load-independent parameter and is commonly thought to have an inverse relationship with compliance. But it has also been demonstrated to be an important measure of contractility.30, 31 In vivo, increased EDPVR can be a surrogate for ventricular hypertrophy. However, it is unlikely that ventricular hypertrophy develops within 8 hours of ESHP and dialysis, as these hearts also have significantly lower percent weight gain than control hearts. Rather, the increased EDPVR may reflect increased contractility32, 33 with the associated preservation of end-systolic Emax, a load-independent measurement of contractility.³⁴ We observed that hearts supported on ESHP with dialysis exhibited significantly improved preservation of Emax at 8 hours compared to the control group. Interestingly, both dialysis and control hearts exhibited a significant decline in dP/dT max, a load-dependent parameter, over time. This suggests that while dialysis may help to prevent decline in contractility, both control and dialysis hearts exhibited impaired response to LV loading over time. Therefore, the use of dialysis in hearts supported with ESHP will only partially prevent a decline in cardiac function. Further research is needed to look at other approaches to better preserve heart function during ESHP.

Coronary vasomotor function, an important predictor of the future development of cardiac allograft vasculopathy, is affected by ischemia reperfusion injury and oxidative stress during ESHP.20, 35, 36 In this study, coronary arteries from dialyzed hearts were more responsive to both endothelial dependent and independent vasorelaxation agents and were less sensitized to ET1 compared to control hearts. This may be due to the improved oxidative stress milieu conferred in dialyzed hearts as is reflected in the observed catalase upregulated in dialysis hearts, which helps reduce reactive oxygen species.³⁷ Furthermore, asymmetric dimethyl arginine, a competitive inhibitor of nitric oxide synthase that causes local vasoconstriction, was significantly downregulated in dialyzed hearts.³⁸ Albeit iNOS was not significantly different between the 2 groups, the significantly upregulated ADMA in control hearts can impair iNOS function, thus, resulting in less nitric oxide production in control hearts.³⁹ We were also able to show that p53 protein expression was significantly upregulated in the LV of control hearts that may have further exacerbated the oxidative damage during ischemia reperfusion.⁴⁰ However, there was no significant difference between the 2 groups in markers of apoptosis, such as cleaved caspase-3 expression. This is consistent with others groups showing that there is no enhanced apoptosis during ESHP.¹⁹ p53 may increase complex I to V synthesis and assembly.⁴¹ While this may be beneficial for LV function, it may also contribute to further reactive oxygen species (ROS) production (Figure 6).⁴²

Considering the role of inflammation, we were only able to detect 5 different cytokines in the filtered dialysate and there was no significant difference in these cytokines between the 2 groups. Thus, the degree of inflammation did not necessarily contribute to the development of edema. Interestingly, the use of filtration of the blood for perfusion has been shown to increase edema formation during ex vivo lung perfusion.43, 44

There are several limitations of this study. Our experimental heart procurement protocol does not mimic standard DCD protocols. Rather, we procure these hearts after exsanguination and the immediate administration of 4°C cardioplegia; warm ischemia and the catecholamine storm are thus avoided. As such, our experimental design may provide less injury prior to perfusion than is sustained clinically. We also observed significant differences in in vivo LV function between our animals. This difference was addressed in both studies by completing repeated measure analysis and comparing percent change in heart function instead of raw pressure volume (PV) loop values between animals. The clinical applicability of our protocol would be limited by the high volume of dialysate used adding to the impracticability of this approach. In addition, the 1-hour cold ischemia time included in our study could represent a significant limitation and potentially reduce the clinical relevance of this study.

In conclusion, this study demonstrates that addition of dialysis to ESHP mitigates the development of myocardial edema and functional decline resulting in improved function and preserved coronary vasomotor function. These findings contribute the field of ESHP technology development to help realize its potential as a potential reparative platform in the future.

CRediT authorship contribution statement

Frank Yu participated in study design, data analysis, and reading of the manuscript. Roberto Ribeiro participated in study design, acquisition of data, and reading of the manuscript. Roizar Rosales participated in acquisition of data and reading of manuscript. Ludger Hauck participated in study design, acquisition of data, and reading of manuscript. Daniela Grothe participated in study design, acquisition of data, and reading of manuscript. Juglans Alvarez participated in study design, acquisition of data, and reading of manuscript. Mitchell Adamson participated in acquisition of data and reading of manuscript. Vivek Rao participated in study design, data analysis, and reading of manuscript. Mitesh Badiwala participated in study design, data analysis, writing of manuscript, and acquisition of funding. Filio Billia participated in study design, data analysis, and writing of manuscript.

Disclosure statement

F.B. has research funding from Abbott Laboratories.

Acknowledgments: None.