- Email: [email protected]
The role of the endothelin-1 pathway as a biomarker for donor lung assessment in clinical ex vivo lung perfusion
http://www.jhltonline.org
The role of the endothelin-1 pathway as a biomarker for donor lung assessment in clinical ex vivo lung perfusion Tiago Noguchi Machuca, MD,a,b Marcelo Cypel, MD, MS,a,b Yidan Zhao, PhD,a Hartmut Grasemann, MD, PhD,a,b Farshad Tavasoli, MD, MS,a Jonathan C. Yeung, MD, PhD,a,b Riccardo Bonato, MD,a,b Manyin Chen, MD,a,b Ricardo Zamel, PhD,a Yi-min Chun,a Zehong Guan, MS,a Marc de Perrot, MD, MS,a,b Thomas K. Waddell, MD, PhD,a,b Mingyao Liu, MS,a and Shaf Keshavjee, MD, MSa,b From the aLatner Thoracic Surgery Research Laboratories, Toronto General Research Institute; and the bToronto Lung Transplant Program, University Health Network, University of Toronto, Toronto, Ontario, Canada.
KEYWORDS: ex vivo lung perfusion; endothelin axis; ET-1; transplant outcomes; primary graft dysfunction
BACKGROUND: Normothermic ex vivo lung perfusion (EVLP) is a preservation technique that allows reassessment of donor lungs before transplantation. We hypothesized that the endothelin-1 (ET-1) axis would be associated with donor lung performance during EVLP and recipient outcomes after transplantation. METHODS: ET-1, Big ET-1, endothelin-converting enzyme (ECE), and nitric oxide (NO) metabolites were quantified in the perfusates of donor lungs enrolled in a clinical EVLP trial. Lungs were divided into 3 groups: (I) Control: bilateral transplantation with good early outcomes defined as absence of primary graft dysfunction (PGD) Grade 3 (PGD3) ; (II) PGD3: bilateral lung transplantation with PGD3 any time within 72 hours; and (III) Declined: lungs rejected after EVLP. RESULTS: There were 25 lungs in Group I, 7 in Group II, and 16 in Group III. At 1 and 4 hours of EVLP, the perfusates of Declined lungs had significantly higher levels of ET-1 (3.1 ⫾ 2.1 vs 1.8⫾2.3 pg/ml, p ¼ 0.01; 2.7 ⫾ 2.2 vs 1.3 ⫾ 1.1 pg/ml, p ¼ 0.007) and Big ET-1 (15.8 ⫾ 14.2 vs 7.0 ⫾ 6.5 pg/ml, p ¼ 0.001; 31.7 ⫾ 17.4 vs 19.4 ⫾ 9.5 pg/ml, p ¼ 0.007) compared with Controls. Nitric oxide metabolite concentrations were significantly higher in Declined and PGD3 lungs than in Controls. For cases of donation after cardiac death, PGD3 and Declined lungs had higher ET-1 and Big ET-1 levels at 4 hours of perfusion compared with Controls. At this time point, Big ET-1 had excellent accuracy to distinguish PGD3 (96%) and Declined (92%) from Control lungs. CONCLUSIONS: In donation after cardiac death lungs, perfusate ET-1 and Big ET-1 are potential predictors of lung function during EVLP and after lung transplantation. They were also associated with non-use of lungs after EVLP and thus could represent useful biomarkers to improve the accuracy of donor lungs selection. J Heart Lung Transplant 2015;34:849–857 r 2015 International Society for Heart and Lung Transplantation. All rights reserved.
Reprint requests: Shaf Keshavjee, MD, MS, Toronto Lung Transplant Program, Toronto General Hospital, 200 Elizabeth St, 9N946, Toronto, ON, M5G 2C4 Canada. Telephone: þ1-416-340-4010. Fax: þ1-416-340-3185. E-mail address: [email protected]
Ex vivo lung perfusion (EVLP) is a contemporary preservation technique that allows for repeated donor lung assessment and treatment in a normothermic metabolically active state. Clinical translation has shown that EVLP
1053-2498/$ - see front matter r 2015 International Society for Heart and Lung Transplantation. All rights reserved. http://dx.doi.org/10.1016/j.healun.2015.01.003
850
The Journal of Heart and Lung Transplantation, Vol 34, No 6, June 2015
perfusion and assessment of high-risk donor lungs provides similar outcomes to lungs that are transplanted using conventional donation criteria, thus representing an important and safe expansion of the donor pool.1,2 Decisions on whether proceed with transplantation during EVLP are currently based primarily on physiologic parameters (gas exchange, airway pressures, pulmonary compliance).3 Future directions in the field point to the development of biomarkers that would (1) improve the diagnostic accuracy by detecting the small, although clinically relevant, percentage of lungs that undergo EVLP and still end up developing severe primary graft dysfunction (PGD); (2) make the decision process more objective and easier for less experienced operators; and (3) increase the current 80% transplant conversion rate after EVLP by using injury-specific targeted therapies. Thus, according to this vision, EVLP will serve as platform for more advanced diagnosis and treatment of donor lungs.4,5 Endothelin-1 (ET-1) is the most abundant and best characterized 21-amino-acid peptide from the endothelin family. It is produced from cleavage of 38-amino-acid big endothelin-1 (Big ET-1) by endothelin-converting enzyme (ECE). Besides being a potent vasoactive peptide, ET-1 also functions as a mediator of acute lung injury.6 Mechanistic studies have shown that ET-1 promotes harmful cross-talk between the endothelial and alveolar compartments by stimulating nitric oxide (NO) production, leading to impairment in alveolar fluid clearance and pulmonary edema.7 More specifically to lung transplantation, ET-1 has been linked to PGD8 and bronchiolitis obliterans.9 Experimental blockade of ET-1 may have beneficial effects on early graft function.10,11 Furthermore, in clinical lung transplantation, ET-1 messenger RNA (mRNA) levels in lung tissue biopsy specimens taken immediately before reperfusion have been shown to correlate with severe PGD.8 In the setting of EVLP, the ET-1 axis would be of particular interest because, unlike most mediators that tend to accumulate in the absence of hepatic or renal clearance mechanisms in the ex vivo circuit, ET-1 is mainly metabolized by the lung.12 We hypothesized that in metabolically active high-risk donor lungs perfused with a clinical intent, the ET-1 axis is active and reflects graft performance during EVLP and after transplantation.
similar to the those described above or an interval between withdrawal of life-sustaining therapies and arrest 460 minutes. Our EVLP technique has been described in detail elsewhere.4 Briefly, donor lungs were placed on the Toronto EVLP system and perfused and ventilated for 4 to 6 hours. Hourly functional assessment included PaO2/FIO2 in the perfusate, peak, mean and plateau airway pressure, dynamic and static compliance, and pulmonary artery pressure. X-ray images of the lung were taken at 1 and 3 hours of EVLP. Lungs declined for transplantation presented a PaO2/FIO2 o400 mm Hg or 415% worsening of any of the above parameters. Additionally, there should be no worsening X-ray infiltrate and no purulent or frothy secretions on the interval bronchoscopy. Donor lungs were divided into three groups: (1) Control—lungs accepted after EVLP, allocated for bilateral transplantation, and the recipient developed no PGD Grade 3 (PGD3) within 72 hours; (2) PGD3—lungs accepted after EVLP, allocated for bilateral transplantation, and the recipient presented PGD3 at any point within 72 hours; (3) Declined—lungs were turned down for transplantation after EVLP evaluation. PGD3 was scored according to the International Society for Heart and Lung Transplantation criteria, with a PaO2/FIO2 o200 and chest X-ray infiltrates within 72 hours.13 The study did not include single-lung transplants, lobar transplants, and recipients bridged to transplant with extracorporeal life support.
Sample collection and protein analysis Perfusate samples were collected at 1 hour and 4 hours of EVLP and stored at –801C for later analysis. Perfusate levels of ET-1 axis proteins were measured using enzyme-linked immunosorbent assay kits (ELISA) for ET-1 (ET-1quantikine ELISA kit; R&D Systems, Minneapolis, MN), Big ET-1 (big ET-1 human ELISA kit; Enzo Life Sciences, Farmingdale, NY), and ECE (ECE ELISA kit; USCN Life Science, Houston, TX).
NO metabolite measurement Total NO metabolite (NOx) levels were measured using a chemiluminescence analyzer (Eco Physics, Switzerland) and a liquid purge vessel system, as previously described.14 All samples were deproteinated using Amicon Ultracel-0.5 10 K centrifugal filters (Millipore catalog UFC501096) by centrifugation at 14,500 g for 15 minutes. Then, sample supernatants (25 μl) were injected into the purge vessel system containing vanadium (III) chloride (0.05 mol/L in 1N hydrochloric acid) as reducing agent.15
Statistical analysis
Methods The cases included in this study were part of the Human Ex vivo Lung Perfusion Trial, conducted by the Toronto Lung Transplant Program to assess high-risk brain-dead donor (BDD) lungs and lungs from donation after cardiac death (DCD).1 Indications for EVLP were (1) last partial pressure of arterial oxygen (PaO2)/ fraction of inspired oxygen (FIO2) of o300 mm Hg; (2) pulmonary edema detected on chest X-ray or during examination of the lungs at procurement; (3) poor lung compliance during examination of the lungs at procurement; and (4) high-risk history (e.g., multiple blood transfusions, questionable history of aspiration). From February 2009 to January 2010, all DCD lungs underwent EVLP. After this period, the EVLP indications for DCD lungs were
EVLP physiologic parameters were compared among the different groups using 2-way analysis of variance. Categoric variables were compared with Fisher’s exact test, and numeric variables were compared with the Mann-Whitney test. For biomarker comparisons, the area under the curve (AUC) was calculated from a receiver operating characteristic (ROC) curve. Statistical analysis was performed using GraphPad Prism 5 (GraphPad Software Inc, La Jolla, CA).
Results From September 2008 to September 2012, we performed 77 EVLPs. Perfusates were not available from 8 donors early in our experience, and 21 donors met the exclusion criteria.
Machuca et al.
ET-1 As Biomarker for EVLP Outcome
851
Figure 1 Study flow diagram shows outcomes after ex vivo lung perfusion (EVLP). The study excluded patients who were bridged with extracorporeal life support (ECLS) and those with unilateral or bilobar transplants. PGD3, primary graft dysfunction Grade 3.
Finally, 48 EVLPs were included in the present study, with 25 Control, 7 PGD3, and 16 Declined (Figure 1). Donor and recipient baseline demographics are summarized in Table 1. Donor age, sex, best PaO2/FIO2, and smoking history were similar among groups. The percentage of DCDs in each group was similar. Recipient variables, such as age, sex, indication for transplant, and percentage of retransplantation, were similar between Control and PGD3 lungs. Detailed donor characteristics for lungs that were transplanted after EVLP are reported in Supplementary Table S1.
EVLP physiology EVLP physiology data are summarized in Table 2. Mean airway pressure, plateau pressure, and PaO2/FIO2 were significantly worse in Declined lungs compared with Controls. Lungs in the PGD3 group presented similar mean airway pressure and plateau pressure compared with Controls. Pulmonary artery pressure was similar in all 3 groups. The reasons to decline each lung after EVLP, along with each donor details, are summarized in Table 3.
ET-1 axis is active during EVLP Although the Big ET-1 concentration in the EVLP perfusate significantly increased over time, from (mean ⫾ standard deviation) 11.2 ⫾ 12.5 pg/ml at 1 hour to 26.8 ⫾ 19.7 pg/ml at 4 hours (p o 0.001), there was no significant change in ET-1 levels, from 2.3 ⫾ 2.6 pg/ml at 1 hour to 1.8 ⫾ 1.4 pg/ml at 4 hours (p ¼ 0.18). Similar to Big ET-1, the levels of ECE increased from 8,405 ⫾ 3,988 pg/ml at 1 hour to 11,330 ⫾ 5,834 pg/ml at 4 hours (p ¼ 0.001).
ET-1 axis at 1 hour is predictive of graft performance on EVLP At 1 hour of EVLP, perfusates from Declined lungs had significantly higher levels of ET-1 (3.1 ⫾ 2.1 vs 1.8 ⫾ 2.3 pg/ml,
p ¼ 0.01) and Big ET-1 (15.8 ⫾ 14.2 vs 7.0 ⫾ 6.5 pg/ml, p ¼ 0.001) compared with Control lungs. On an ROC curve, the AUCs for ET-1 and Big ET-1 to differentiate Declined from Controls were 0.73 and 0.79, respectively. With a cutoff at 9.8 pg/ml, Big ET-1 had a sensitivity of 75% and specificity of 76% to diagnose Declined lungs. The difference between PGD3 and Control lungs with regards to ET-1 and Big ET-1 perfusate levels at 1 hour of EVLP was not significant. The levels of ECE were similar in all 3 groups (Figure 2).
ET-1 axis at 4 hours is predictive of graft performance on EVLP Similar to 1 hour, Declined lungs also had higher levels of ET-1 (2.7 ⫾ 2.2 vs 1.3 ⫾ 1.1 pg/ml, p ¼ 0.007) and Big ET-1 (31.7 ⫾ 17.4 vs 19.4 ⫾ 9.5 pg/ml, p ¼ 0.007) at 4 hours of EVLP compared with Controls. The AUC for a ROC curve of ET-1 in Declined vs Control was 0.75. With a cutoff value of 1.49 pg/ml, there was a diagnostic sensitivity of 80% and specificity of 72%. For Big ET-1, a cutoff value of 24.3 pg/ml had a diagnostic sensitivity of 73.3%, specificity of 75%, and an accuracy of 75% for Declined vs Control lungs. ECE levels were similar among the groups (Figure 3).
ET-1 axis analysis in DCD lungs is predictive of graft performance on EVLP and after transplantation In BDD lungs, the ET-1 axis did not show significant differences among the groups (data not shown). However for DCD cases, PGD3 and Declined lungs had higher ET-1 and Big ET-1 levels at 4 hours of perfusion compared with Controls (PGD3 vs Control: ET-1, p ¼ 0.03; Big ET-1, p ¼ 0.01. Declined vs Control: ET-1, p ¼ 0.007; big ET-1, p ¼ 0.003). There were no differences in ECE levels among groups. The AUCs for Big ET-1 at 4 hours of EVLP in Declined vs Control and PGD3 vs Control were 0.92 and 0.96, respectively (Figure 4). At the same time point, ET-1
852 Table 1
The Journal of Heart and Lung Transplantation, Vol 34, No 6, June 2015 Donor and Recipient Baseline Demographics
Variablesa Donor variables Age, years DCD Male Last PaO2/FIO2, mm Hg Smoker Recipient variables Age, years Male Diagnosis, % ILD Retransplantation Bilateral
Control (n ¼ 25)
PGD3 (n ¼ 7)
p-value
Declined (n ¼ 16)
p-value
45 ⫾ 14 48 48 366 ⫾ 93 52
44 ⫾ 16 43 14 326 ⫾ 104 71
0.981 1 0.195 0.361 0.426
39 ⫾ 17 44 62 335 ⫾ 105 37
0.278 1 0.522 0.574 0.522
48 ⫾ 14 48 30 4 100
49 ⫾ 11 29 71 0 100
1 0.426 0.078 1 1
DCD, donation after cardiac death; ILD, interstitial lung disease; FIO2, fraction of inspired oxygen; PaO2 partial pressure of arterial oxygen; PGD3, primary graft dysfunction Grade 3. a Continous data are presented as mean ⫾ standard deviation and categoric data as percentage.
had an accuracy of 91% to differentiate PGD3 from Controls and of 88% to differentiate Declined from Controls.
compared with Controls at 1 hour of EVLP (p ¼ 0.003 for both comparisons). At 4 hours of EVLP, NOx concentrations were again higher in Declined lungs compared with Controls (p ¼ 0.01; Figure 5).
NOx are significantly higher in lungs declined after EVLP
Discussion
ET-1 signals through ETB receptors in endothelial cells to produce NO, which then acts on epithelial alveolar cells to decrease expression of Naþ/Kþ adenosine 50 -triphosphatase in the basolateral surface.7 This mechanism impairs alveolar fluid clearance and ultimately leads to edema formation. We thus measured NOx in perfusates of high-risk donor lungs that underwent EVLP. Interestingly, NOx concentrations were significantly higher in Declined and PGD3 lungs
We have demonstrated that Big ET-1 and ET-1 measured in perfusate samples during EVLP have prognostic significance. At 1 and 4 hours of perfusion, Big ET-1 and ET-1 levels were higher in lungs that were declined after EVLP compared with EVLP lungs transplanted with good early outcomes (Controls). Analysis of the AUC for Big ET-1 reveals a reasonable accuracy to predict, early (i.e., at 1 hour of perfusion) lungs that would be ultimately declined
Table 2
Physiologic Ex Vivo Lung Perfusion Parameters
Variablea PAP, mm Hg Control PGD3 Declined Airway pressure Mean, cm H2O Control PGD3 Declined Plateau, cm H2O Control PGD3 Declined PaO2/FIO2 mm Hg Control PGD3 Declined
1 hour
2 hours
3 hours
4 hours
p-valueb
9.4 ⫾ 2.1 10.4 ⫾ 2.6 10 ⫾ 2
9.8 ⫾ 2 10.8 ⫾ 2.4 10.3 ⫾ 2
9.8 ⫾ 2.1 11 ⫾ 2.5 10.2 ⫾ 2.3
9.7 ⫾ 2.2 10.7 ⫾ 2.4 10 ⫾ 3
0.269 0.504
7.2 ⫾ 0.7 7.2 ⫾ 0.7 7.8 ⫾ 0.8
7.1 ⫾ 0.9 7.1 ⫾ 0.6 7.8 ⫾ 0.9
7.1 ⫾ 0.9 7.2 ⫾ 0.7 7.8 ⫾ 1.1
7.2 ⫾ 0.9 7.1 ⫾ 0.6 7.9 ⫾ 1.1
0.882 0.015
12.3 ⫾ 1.7 13 ⫾ 2.5 14.7 ⫾ 3.3
12 ⫾ 1.5 12.7 ⫾ 2.2 14.2 ⫾ 3.1
12 ⫾ 1.7 12.5 ⫾ 2.2 13.6 ⫾ 3.1
12.4 ⫾ 1.6 13.2 ⫾ 2.4 14.3 ⫾ 3.1
0.479 0.011
495.9 ⫾ 88.7 552.1 ⫾ 63.2 404.3 ⫾ 103.9
532.5 ⫾ 65.9 567.5 ⫾ 33.5 414.7 ⫾ 99
544.7 ⫾ 58.4 573.6 ⫾ 44.4 448.6 ⫾ 84.3
541.6 ⫾ 58.5 560.3 ⫾ 43.3 446 ⫾ 73.1
0.015 o0.0001
FIO2, fraction of inspired oxygen; PaO2, partial pressure of oxygen; PAP, pulmonary artery pressure. a Data are expressed as mean ⫾ standard deviation. b The p-values correspond to comparison between each respective group with Control using 2-way analysis of variance.
Machuca et al. Table 3
ET-1 As Biomarker for EVLP Outcome
853
Detailed Demographics for Declined Lungs and Reason to Decline During Ex Vivo Lung Perfusion
Donor Age (years)
Type
44 57 53 56
DCD DCD NDD DCD
35 20
Smoker
Last PaO2/FIO2
Abnormal CXR
Bronchoscopy
Main indication for EVLP a
DCD DCDa Low PaO2/FIO2 Edema CXR and at procurement Abnormal CXR Edematous lower lobes
Main reason to decline after EVLP Not meeting EVLP PaO2/FIO2 criteriab Worsening compliance Worsening compliance Not meeting EVLP PaO2/FIO2 criteria,b worsening compliance Worsening compliance Worsening edema CXR and bronchoscopy Frothy secretions, worsening compliance Worsening edema CXR and bronchoscopy Worsening edema CXR, bronchoscopy Worsening edema CXR, bronchoscopy Not meeting EVLP PaO2/FIO2 criteria,b emphysema Not meeting EVLP PaO2/FIO2 criteriab Worsening edema CXR
Yes No No
420 325 297 378
No No Yes Yes
Purulent Clear Clear Bloody
NDD NDD
Yes Yes
404 440
Yes Yes
Clear Mucoid
67
NDD
No
92
Yes
Aspiration?
25
DCD
No
392
Yes
Clear
36
NDD
No
381
Yes
Clear
Low PaO2/FIO2, aspiration suspicion WLST to arrest 70 minutes, fat emboli Edema lower lobes
28
DCD
No
501
Yes
Bloody
Edema lower lobes
49
DCD
Yes
407
No
Clear
20 15
NDD NDD
No No
335 222
Yes Yes
Clear Bloody
16 50
DCD NDD
No No
336 250
Yes Yes
Purulent Clear
60
NDD
Yes
185
Yes
Clear
Heavy smoker, poor deflation, bullae Edema lower lobes Low PaO2/FIO2, edema lower lobes Edema lower lobes Not meeting EVLP PaO2/FIO2 criteriab Worsening edema CXR Low PaO2/FIO2, preserved deflated long CIT (outside procurement team) Poor deflation Not meeting EVLP PaO2/FIO2 criteria,b poor deflation
CIT, cold ischemic time; CXR, chest X-ray; DCD, donation after cardiac death; FIO2, fraction of inspired oxygen; NDD, neurological determination of death; PaO2, partial pressure of arterial oxygen; WLST, withdrawal of life-sustaining therapies. a Represents a case in which the indication for EVLP was a DCD donor. b EVLP PaO2/FIO2 criteria Z 400 mm Hg.
because of subsequent development of poor physiologic performance. This information could be useful to select, early in the course of EVLP, lungs that will require additional therapeutic measures to achieve transplantability criteria. Although this study does not provide the posttransplant function of lungs declined because of worsening physiologic parameters under the protective measures of EVLP, experimental data suggest an association with a high risk of PGD3.3 More importantly, our findings suggest that for DCDs, the ET-1 axis assessment shows a strong correlation between perfusate levels and development of PGD3. Translation of our findings to clinical practice stands to deliver a substantial effect given that EVLP physiologic parameters of lungs that developed PGD3 were similar to those of the Control lungs. Thus, the availability of perfusate biomarkers (possibly Big ET-1 and ET-1 for DCD lungs) in a timely fashion could help improve the precision of donor lung selection, bringing more objectivity to the decision-making process. The stronger significance of the ET-1 axis in DCDs is intriguing and might potentially reflect different mechanisms of donor lung injury. On one hand, DCD lungs may be spared the catecholamine
storm related to brain death, but on the other hand, they are prone to periods of hypotension, low shear stress, aspiration injury, and hypoxia.16 We previously showed different transcriptional signatures between DCD and BDD donor lungs, particularly at the end of cold ischemic time.17 More recently, we showed similar results in a larger cohort of EVLP lungs where lung tissue biopsies at the end of the cold ischemic time demonstrated 285 differentially expressed genes between DCD and BDD.18 ET-1 is mostly known for its pathogenetic involvement in pulmonary arterial hypertension but has also been associated with acute lung injury. Higher circulating levels of ET-1 were found in patients with acute lung injury compared with healthy controls.19 That ET-1 levels fell to levels within normal reference ranges in those patients with acute lung injury/adult respiratory distress syndrome that eventually recovered is further supporting evidence for a potential mediator role of ET-1 in lung injury.20 The evaluation of the ET-1 axis in lung transplantation presents biologic plausibility because its expression is regulated by several mechanisms involved with the organ
854
The Journal of Heart and Lung Transplantation, Vol 34, No 6, June 2015
Figure 2 Endothelin-1 (ET-1) axis protein levels at 1 hour of ex vivo lung perfusion (EVLP). (A) ET-1 levels were significantly higher in Declined lungs compared with Controls (*p ¼ 0.01). (B) Big ET-1 levels were significantly higher in Declined compared with Controls (**p ¼ 0.001). (C) Endothelin-converting enzyme (ECE) levels did not differ among the groups. (D) The area under curve (AUC) of the receiver operating characteristic curve revealed that Big ET-1 at 1 hour of EVLP had a reasonable accuracy (0.79) to differentiate Declined from Control donor lungs. For A-C, bars in the scatter plot represent the mean and standard error. PGE3, primary graft dysfunction Grade 3.
donation process. Hormones (cortisol, adrenaline, insulin), cytokines (interleukin-1β) and physicochemical stimuli (hypoxia, low shear stress) favor increased ET-1 expression, and prostacyclin, heparin, and high shear stress inhibit it.21 In an experimental model of donor lung injury by prolonged cold ischemic time (18–20 hours), grafts transplanted into recipients treated with a non-selective ET-1 receptor antagonist presented significantly better lung function and
were able to sustain life once the contralateral pulmonary artery was clamped.11 In the clinical setting, ET-1 mRNA levels were higher in pre-implantation biopsy specimens from donor lungs that developed PGD3 compared with lungs without PGD3.8 Moreover, the authors also analyzed recipient samples in that study and were able to show that a combination of high donor lung tissue ET-1 mRNA with high recipient pre-transplant serum ET-1 was highly
Figure 3 Endothelin-1 (ET-1) axis protein levels at 4 hours of ex vivo lung perfusion (EVLP). (A) ET-1 levels were significantly higher in Declined lungs compared with Controls (**p ¼ 0.007). (B) Big ET-1 levels were significantly higher in Declined compared with Controls (**p ¼ 0.007). (C) Endothelin-converting enzyme (ECE) levels did not differ among groups. (D) The area under curve (AUC) of the receiver operating characteristic curve revealed that Big ET-1 at 4 hours of EVLP had a reasonable accuracy (0.75) to distinguish Declined from Control donor lungs. For A-C, bars in the scatter plot represent the mean and standard error. PGD3, primary graft dysfunction Grade 3.
Machuca et al.
ET-1 As Biomarker for EVLP Outcome
855
Figure 4 Endothelin-1 (ET-1) axis protein levels in donation after cardiac death (DCD) lungs during ex vivo lung perfusion (EVLP). (A) ET-1 levels at 1 hour of EVLP were significantly higher in Declined lungs compared with Controls (*p ¼ 0.04). (B), Big ET-1 levels at 1 hour of EVLP were significantly higher in Declined compared with Controls (**p ¼ 0.007). (C), ET-1 levels at 4 hours of EVLP were significantly higher in PGD3 (*p ¼ 0.03) and Declined lungs (**p ¼ 0.007) compared with Controls. (D) Big ET-1 levels at 4 hours of EVLP were significantly higher in primary graft dysfunction Grade 3 (PGD3) lungs (*p ¼ 0.01) and Declined lungs (**p ¼ 0.003) compared with Controls. (E) The area under curve (AUC) of the receiver operating characteristic (ROC) curve revealed that Big ET-1 at 4 hours of EVLP had excellent accuracy (0.92) to distinguish Declined from Control donor lungs. (F) The AUC of the ROC curve revealed that Big ET-1 at 4 hours of EVLP had excellent accuracy (0.96) to distinguish PGD3 from Control donor lungs. For A-D, bars in the scatter plot represent the mean and standard error.
correlated with PGD3. Interestingly, the inverse situation, low donor lung tissue ET-1 mRNA and low recipient plasma ET-1 also correlated with the absence of PGD. This highlights the point that although we are focusing on donor lung–related predictive factors, ultimate models of prediction of PGD after lung transplantation should include consideration of recipient factors that undoubtedly do contribute to PGD.22 Several mechanisms have the potential to impair the alveolar–capillary barrier through ET-1, namely: (1) the increase in capillary hydrostatic pressure; (2) the increase in capillary permeability mediated by upregulation of vascular
endothelial growth factor; (3) the recruitment of inflammatory cells; and (4) the decrease in alveolar fluid clearance.6 Our observation of increased NOx concomitant with increased levels of ET-1 axis proteins in Declined lungs points toward ET-1–mediated impairment of alveolar fluid clearance as a plausible mechanism leading to physiologic deterioration of damaged human lungs that cannot be rescued with the current EVLP technique. Evidence of NO as a mediator of impaired alveolar fluid clearance is found in the work of Kaestle et al,23 who performed isolated perfusion with rat lungs. They demonstrated in a model of hydrostatic pulmonary edema that
856
The Journal of Heart and Lung Transplantation, Vol 34, No 6, June 2015
Figure 5 (A) Perfusate concentrations of nitric oxide metabolites (NOx) at 1 hour of ex vivo lung perfusion (EVLP) were significantly higher in primary graft dysfunction Grade 3 (PGD3) and Declined lungs compared with Controls (**p ¼ 0.003 for both). (B) Perfusate NOx levels at 4 hours of EVLP were significantly higher in Declined lungs compared with Controls (*p ¼ 0.01). The bars in the scatter plot represent the mean and standard error.
increased fluid filtration is responsible for only 30% of the net fluid shift, with the other 70% attributable to NO-induced impairment in alveolar fluid clearance. Whereas the NO synthase inhibitor L-NG-nitro-L-arginine methyl ester prevented edema formation, the addition of NO (S-nitrosoglutathione and 1-propamine 3-[2-hydroxy-2-nitroso-1-propylhydrazine] NONOate) to the perfusate blocked fluid reabsorption. More recently, Comellas et al7 performed mechanistic studies providing evidence pointing to ETB receptors as the main responsible factor for impairment in alveolar fluid clearance. Combining rat lung perfusion models with cell culture experiments, they demonstrated the important crosstalk between endothelial and epithelial alveolar cells through ET-1-ETB receptor signaling via NO to decrease the expression of Naþ/Kþ adenosine 50 -triphosphatase in the basolateral membrane of alveolar epithelial cells, ultimately leading to pulmonary edema. Interestingly, the addition of ET1 to the perfusate, but not its instillation into the airway, led to a decrease in alveolar fluid clearance by 40% to 50%. These findings in experimental isolated lung perfusion corroborate our findings in clinical EVLP and also reinforce the importance of measuring ET-1 levels in the vascular compartment. Differently from chemokines and cytokines, which will tend to accumulate in the EVLP perfusate due to the lack of renal and/or hepatic clearance mechanisms, ET-1 is partially produced and cleared in the pulmonary circulation. In models of lung perfusion, 31% to 38% of radiolabeled ET-1 was cleared by a single pass.12 The clearance of ET-1 is closely related to ETB receptors because its selective blockade significantly increases circulating ET-1 levels.24 Interestingly in our study, levels of Big ET-1 and ECE significantly increased over time, whereas the levels of ET-1 remained relatively stable from 1 to 4 hours. These findings are suggestive of active ET-1 clearance mechanisms in donor lungs perfused ex vivo. Because the ET-1 axis can be manipulated pharmacologically, our findings have therapeutic implications. Although selective blockade of ET-1 has been shown to elicit pulmonary edema, its non-selective blockade may be beneficial in improving alveolar fluid clearance and
avoiding the formation of pulmonary edema. In this scenario, EVLP would provide the ideal platform for donor lung treatment and reassessment ex vivo before moving forward to a successful transplantation.5 The main limitation of our study is the sample size. Nevertheless, this represents the largest clinical EVLP experience to date using marginal donor lungs. Further validation with an independent set of patients will be required before bringing this biomarker to clinical use. In conclusion, we have shown temporal changes and the prognostic relevance of the ET-1 axis in EVLP for clinical lung transplantation. In DCD lungs, higher levels of big ET-1 and ET-1 were observed at 4 hours of EVLP and demonstrated a strong correlation with PGD after transplantation. Further studies will be instrumental to validate our results and determine whether ET-1 will be an important biomarker for translation to clinical decision making in donor lung selection.
Disclosure statement S.K. and M.C. were principal investigators for the Toronto ExVivo Lung Perfusion Trial, sponsored by XVIVO Perfusion (Gothenburg, Sweden). S.K., M.C., and T.K.W. are founding members of Perfusix Inc, a company that provides ex vivo organ perfusion services. None of the other authors has a financial relationship with a commercial entity that has an interest in the subject of the presented manuscript or other conflicts of interest to disclose. T.N.M. is supported by a Research Fellowship from the American Society of Transplantation.
Supplementary data Supplementary data associated with this article can be found in the online version at www.jhltonline.org.
References 1. Cypel M, Yeung JC, Liu M, et al. Normothermic ex vivo lung perfusion in clinical lung transplantation. N Engl J Med 2011;364:1431-40.
Machuca et al.
ET-1 As Biomarker for EVLP Outcome
2. Cypel M, Yeung JC, Machuca T, et al. Experience with the first 50 ex vivo lung perfusions in clinical transplantation. J Thorac Cardiovasc Surg 2012;144:1200-6. 3. Yeung JC, Cypel M, Machuca TN, et al. Physiologic assessment of the ex vivo donor lung for transplantation. J Heart Lung Transplant 2012; 31:1120-6. 4. Machuca TN, Cypel M, Keshavjee S. Advances in lung preservation. Surg Clin North Am 2013;93:1373-94. 5. Machuca TN, Hsin MK, Ott HC, et al. Injury-specific ex vivo treatment of the donor lung: pulmonary thrombolysis followed by successful lung transplantation. Am J Respir Crit Care Med 2013;188:878-80. 6. Comellas AP, Briva A. Role of endothelin-1 in acute lung injury. Transl Res 2009;153:263-71. 7. Comellas AP, Briva A, Dada LA, et al. Endothelin-1 impairs alveolar epithelial function via endothelial ETB receptor. Am J Respir Crit Care Med 2009;179:113-22. 8. Salama M, Andrukhova O, Hoda MA, et al. Concomitant endothelin-1 overexpression in lung transplant donors and recipients predicts primary graft dysfunction. Am J Transplant 2010;10:628-36. 9. Tikkanen JM, Koskinen PK, Lemstrom KB. Role of endogenous endothelin-1 in transplant obliterative airway disease in the rat. Am J Transplant 2004;4:713-20. 10. Khimenko PL, Moore TM, Taylor AE. Blocked ETA receptors prevent ischemia and reperfusion injury in rat lungs. J Appl Physiol (1985) 1996;80:203-7. 11. Shennib H, Lee AG, Kuang JQ, et al. Efficacy of administering an endothelin-receptor antagonist (SB209670) in ameliorating ischemiareperfusion injury in lung allografts. Am J Respir Crit Care Med 1998; 157:1975-81. 12. Dupuis J, Goresky CA, Stewart DJ. Pulmonary removal and production of endothelin in the anesthetized dog. J Appl Physiol (1985) 1994;76:694-700. 13. Christie JD, Carby M, Bag R, et al. Report of the ISHLT Working Group on Primary Lung Graft Dysfunction part II: definition.
857
14.
15.
16.
17.
18.
19. 20.
21. 22.
23.
24.
A consensus statement of the International Society for Heart and Lung Transplantation. J Heart Lung Transplant 2005;24:1454-9. Grasemann C, Ratjen F, Schnabel D, et al. Effect of growth hormone therapy on nitric oxide formation in cystic fibrosis patients. Eur Respir J 2008;31:815-21. Braman RS, Hendrix SA. Nanogram nitrite and nitrate determination in environmental and biological materials by vanadium (III) reduction with chemiluminescence detection. Anal Chem 1989;61:2715-8. Tremblay LN, Yamashiro T, DeCampos KN, et al. Effect of hypotension preceding death on the function of lungs from donors with nonbeating hearts. J Heart Lung Transplant 1996;15:260-8. Kang CH, Anraku M, Cypel M, et al. Transcriptional signatures in donor lungs from donation after cardiac death vs after brain death: a functional pathway analysis. J Heart Lung Transplant 2011;30:289-98. Zamel R, Bai XH, Yeung JC, et al. Differential gene regulation between brain and cardiac death donor lungs during ex vivo lung perfusion. J Heart and Lung Transplant 2012;31:S146-7. Mitaka C, Hirata Y, Nagura T, et al. Circulating endothelin-1 concentrations in acute respiratory failure. Chest 1993;104:476-80. Langleben D, DeMarchie M, Laporta D, et al. Endothelin-1 in acute lung injury and the adult respiratory distress syndrome. Am Rev Respir Dis 1993;148:1646-50. Shao D, Park JE, Wort SJ. The role of endothelin-1 in the pathogenesis of pulmonary arterial hypertension. Pharmacol Res 2011;63:504-11. Diamond JM, Lee JC, Kawut SM, et al. Clinical risk factors for primary graft dysfunction after lung transplantation. Am J Respir Crit Care Med 2013;187:527-34. Kaestle SM, Reich CA, Yin N, et al. Nitric oxide-dependent inhibition of alveolar fluid clearance in hydrostatic lung edema. Am J Physiol Lung Cell Mol Physiol 2007;293:L859-69. Dupuis J, Goresky CA, Fournier A. Pulmonary clearance of circulating endothelin-1 in dogs in vivo: exclusive role of ETB receptors. J Appl Physiol (1985) 1996;81:1510-5.
The role of the endothelin-1 pathway as a biomarker for donor lung assessment in clinical ex vivo lung perfusion Tiago Noguchi Machuca, MD,a,b Marcelo Cypel, MD, MS,a,b Yidan Zhao, PhD,a Hartmut Grasemann, MD, PhD,a,b Farshad Tavasoli, MD, MS,a Jonathan C. Yeung, MD, PhD,a,b Riccardo Bonato, MD,a,b Manyin Chen, MD,a,b Ricardo Zamel, PhD,a Yi-min Chun,a Zehong Guan, MS,a Marc de Perrot, MD, MS,a,b Thomas K. Waddell, MD, PhD,a,b Mingyao Liu, MS,a and Shaf Keshavjee, MD, MSa,b From the aLatner Thoracic Surgery Research Laboratories, Toronto General Research Institute; and the bToronto Lung Transplant Program, University Health Network, University of Toronto, Toronto, Ontario, Canada.
KEYWORDS: ex vivo lung perfusion; endothelin axis; ET-1; transplant outcomes; primary graft dysfunction
BACKGROUND: Normothermic ex vivo lung perfusion (EVLP) is a preservation technique that allows reassessment of donor lungs before transplantation. We hypothesized that the endothelin-1 (ET-1) axis would be associated with donor lung performance during EVLP and recipient outcomes after transplantation. METHODS: ET-1, Big ET-1, endothelin-converting enzyme (ECE), and nitric oxide (NO) metabolites were quantified in the perfusates of donor lungs enrolled in a clinical EVLP trial. Lungs were divided into 3 groups: (I) Control: bilateral transplantation with good early outcomes defined as absence of primary graft dysfunction (PGD) Grade 3 (PGD3) ; (II) PGD3: bilateral lung transplantation with PGD3 any time within 72 hours; and (III) Declined: lungs rejected after EVLP. RESULTS: There were 25 lungs in Group I, 7 in Group II, and 16 in Group III. At 1 and 4 hours of EVLP, the perfusates of Declined lungs had significantly higher levels of ET-1 (3.1 ⫾ 2.1 vs 1.8⫾2.3 pg/ml, p ¼ 0.01; 2.7 ⫾ 2.2 vs 1.3 ⫾ 1.1 pg/ml, p ¼ 0.007) and Big ET-1 (15.8 ⫾ 14.2 vs 7.0 ⫾ 6.5 pg/ml, p ¼ 0.001; 31.7 ⫾ 17.4 vs 19.4 ⫾ 9.5 pg/ml, p ¼ 0.007) compared with Controls. Nitric oxide metabolite concentrations were significantly higher in Declined and PGD3 lungs than in Controls. For cases of donation after cardiac death, PGD3 and Declined lungs had higher ET-1 and Big ET-1 levels at 4 hours of perfusion compared with Controls. At this time point, Big ET-1 had excellent accuracy to distinguish PGD3 (96%) and Declined (92%) from Control lungs. CONCLUSIONS: In donation after cardiac death lungs, perfusate ET-1 and Big ET-1 are potential predictors of lung function during EVLP and after lung transplantation. They were also associated with non-use of lungs after EVLP and thus could represent useful biomarkers to improve the accuracy of donor lungs selection. J Heart Lung Transplant 2015;34:849–857 r 2015 International Society for Heart and Lung Transplantation. All rights reserved.
Reprint requests: Shaf Keshavjee, MD, MS, Toronto Lung Transplant Program, Toronto General Hospital, 200 Elizabeth St, 9N946, Toronto, ON, M5G 2C4 Canada. Telephone: þ1-416-340-4010. Fax: þ1-416-340-3185. E-mail address: [email protected]
Ex vivo lung perfusion (EVLP) is a contemporary preservation technique that allows for repeated donor lung assessment and treatment in a normothermic metabolically active state. Clinical translation has shown that EVLP
1053-2498/$ - see front matter r 2015 International Society for Heart and Lung Transplantation. All rights reserved. http://dx.doi.org/10.1016/j.healun.2015.01.003
850
The Journal of Heart and Lung Transplantation, Vol 34, No 6, June 2015
perfusion and assessment of high-risk donor lungs provides similar outcomes to lungs that are transplanted using conventional donation criteria, thus representing an important and safe expansion of the donor pool.1,2 Decisions on whether proceed with transplantation during EVLP are currently based primarily on physiologic parameters (gas exchange, airway pressures, pulmonary compliance).3 Future directions in the field point to the development of biomarkers that would (1) improve the diagnostic accuracy by detecting the small, although clinically relevant, percentage of lungs that undergo EVLP and still end up developing severe primary graft dysfunction (PGD); (2) make the decision process more objective and easier for less experienced operators; and (3) increase the current 80% transplant conversion rate after EVLP by using injury-specific targeted therapies. Thus, according to this vision, EVLP will serve as platform for more advanced diagnosis and treatment of donor lungs.4,5 Endothelin-1 (ET-1) is the most abundant and best characterized 21-amino-acid peptide from the endothelin family. It is produced from cleavage of 38-amino-acid big endothelin-1 (Big ET-1) by endothelin-converting enzyme (ECE). Besides being a potent vasoactive peptide, ET-1 also functions as a mediator of acute lung injury.6 Mechanistic studies have shown that ET-1 promotes harmful cross-talk between the endothelial and alveolar compartments by stimulating nitric oxide (NO) production, leading to impairment in alveolar fluid clearance and pulmonary edema.7 More specifically to lung transplantation, ET-1 has been linked to PGD8 and bronchiolitis obliterans.9 Experimental blockade of ET-1 may have beneficial effects on early graft function.10,11 Furthermore, in clinical lung transplantation, ET-1 messenger RNA (mRNA) levels in lung tissue biopsy specimens taken immediately before reperfusion have been shown to correlate with severe PGD.8 In the setting of EVLP, the ET-1 axis would be of particular interest because, unlike most mediators that tend to accumulate in the absence of hepatic or renal clearance mechanisms in the ex vivo circuit, ET-1 is mainly metabolized by the lung.12 We hypothesized that in metabolically active high-risk donor lungs perfused with a clinical intent, the ET-1 axis is active and reflects graft performance during EVLP and after transplantation.
similar to the those described above or an interval between withdrawal of life-sustaining therapies and arrest 460 minutes. Our EVLP technique has been described in detail elsewhere.4 Briefly, donor lungs were placed on the Toronto EVLP system and perfused and ventilated for 4 to 6 hours. Hourly functional assessment included PaO2/FIO2 in the perfusate, peak, mean and plateau airway pressure, dynamic and static compliance, and pulmonary artery pressure. X-ray images of the lung were taken at 1 and 3 hours of EVLP. Lungs declined for transplantation presented a PaO2/FIO2 o400 mm Hg or 415% worsening of any of the above parameters. Additionally, there should be no worsening X-ray infiltrate and no purulent or frothy secretions on the interval bronchoscopy. Donor lungs were divided into three groups: (1) Control—lungs accepted after EVLP, allocated for bilateral transplantation, and the recipient developed no PGD Grade 3 (PGD3) within 72 hours; (2) PGD3—lungs accepted after EVLP, allocated for bilateral transplantation, and the recipient presented PGD3 at any point within 72 hours; (3) Declined—lungs were turned down for transplantation after EVLP evaluation. PGD3 was scored according to the International Society for Heart and Lung Transplantation criteria, with a PaO2/FIO2 o200 and chest X-ray infiltrates within 72 hours.13 The study did not include single-lung transplants, lobar transplants, and recipients bridged to transplant with extracorporeal life support.
Sample collection and protein analysis Perfusate samples were collected at 1 hour and 4 hours of EVLP and stored at –801C for later analysis. Perfusate levels of ET-1 axis proteins were measured using enzyme-linked immunosorbent assay kits (ELISA) for ET-1 (ET-1quantikine ELISA kit; R&D Systems, Minneapolis, MN), Big ET-1 (big ET-1 human ELISA kit; Enzo Life Sciences, Farmingdale, NY), and ECE (ECE ELISA kit; USCN Life Science, Houston, TX).
NO metabolite measurement Total NO metabolite (NOx) levels were measured using a chemiluminescence analyzer (Eco Physics, Switzerland) and a liquid purge vessel system, as previously described.14 All samples were deproteinated using Amicon Ultracel-0.5 10 K centrifugal filters (Millipore catalog UFC501096) by centrifugation at 14,500 g for 15 minutes. Then, sample supernatants (25 μl) were injected into the purge vessel system containing vanadium (III) chloride (0.05 mol/L in 1N hydrochloric acid) as reducing agent.15
Statistical analysis
Methods The cases included in this study were part of the Human Ex vivo Lung Perfusion Trial, conducted by the Toronto Lung Transplant Program to assess high-risk brain-dead donor (BDD) lungs and lungs from donation after cardiac death (DCD).1 Indications for EVLP were (1) last partial pressure of arterial oxygen (PaO2)/ fraction of inspired oxygen (FIO2) of o300 mm Hg; (2) pulmonary edema detected on chest X-ray or during examination of the lungs at procurement; (3) poor lung compliance during examination of the lungs at procurement; and (4) high-risk history (e.g., multiple blood transfusions, questionable history of aspiration). From February 2009 to January 2010, all DCD lungs underwent EVLP. After this period, the EVLP indications for DCD lungs were
EVLP physiologic parameters were compared among the different groups using 2-way analysis of variance. Categoric variables were compared with Fisher’s exact test, and numeric variables were compared with the Mann-Whitney test. For biomarker comparisons, the area under the curve (AUC) was calculated from a receiver operating characteristic (ROC) curve. Statistical analysis was performed using GraphPad Prism 5 (GraphPad Software Inc, La Jolla, CA).
Results From September 2008 to September 2012, we performed 77 EVLPs. Perfusates were not available from 8 donors early in our experience, and 21 donors met the exclusion criteria.
Machuca et al.
ET-1 As Biomarker for EVLP Outcome
851
Figure 1 Study flow diagram shows outcomes after ex vivo lung perfusion (EVLP). The study excluded patients who were bridged with extracorporeal life support (ECLS) and those with unilateral or bilobar transplants. PGD3, primary graft dysfunction Grade 3.
Finally, 48 EVLPs were included in the present study, with 25 Control, 7 PGD3, and 16 Declined (Figure 1). Donor and recipient baseline demographics are summarized in Table 1. Donor age, sex, best PaO2/FIO2, and smoking history were similar among groups. The percentage of DCDs in each group was similar. Recipient variables, such as age, sex, indication for transplant, and percentage of retransplantation, were similar between Control and PGD3 lungs. Detailed donor characteristics for lungs that were transplanted after EVLP are reported in Supplementary Table S1.
EVLP physiology EVLP physiology data are summarized in Table 2. Mean airway pressure, plateau pressure, and PaO2/FIO2 were significantly worse in Declined lungs compared with Controls. Lungs in the PGD3 group presented similar mean airway pressure and plateau pressure compared with Controls. Pulmonary artery pressure was similar in all 3 groups. The reasons to decline each lung after EVLP, along with each donor details, are summarized in Table 3.
ET-1 axis is active during EVLP Although the Big ET-1 concentration in the EVLP perfusate significantly increased over time, from (mean ⫾ standard deviation) 11.2 ⫾ 12.5 pg/ml at 1 hour to 26.8 ⫾ 19.7 pg/ml at 4 hours (p o 0.001), there was no significant change in ET-1 levels, from 2.3 ⫾ 2.6 pg/ml at 1 hour to 1.8 ⫾ 1.4 pg/ml at 4 hours (p ¼ 0.18). Similar to Big ET-1, the levels of ECE increased from 8,405 ⫾ 3,988 pg/ml at 1 hour to 11,330 ⫾ 5,834 pg/ml at 4 hours (p ¼ 0.001).
ET-1 axis at 1 hour is predictive of graft performance on EVLP At 1 hour of EVLP, perfusates from Declined lungs had significantly higher levels of ET-1 (3.1 ⫾ 2.1 vs 1.8 ⫾ 2.3 pg/ml,
p ¼ 0.01) and Big ET-1 (15.8 ⫾ 14.2 vs 7.0 ⫾ 6.5 pg/ml, p ¼ 0.001) compared with Control lungs. On an ROC curve, the AUCs for ET-1 and Big ET-1 to differentiate Declined from Controls were 0.73 and 0.79, respectively. With a cutoff at 9.8 pg/ml, Big ET-1 had a sensitivity of 75% and specificity of 76% to diagnose Declined lungs. The difference between PGD3 and Control lungs with regards to ET-1 and Big ET-1 perfusate levels at 1 hour of EVLP was not significant. The levels of ECE were similar in all 3 groups (Figure 2).
ET-1 axis at 4 hours is predictive of graft performance on EVLP Similar to 1 hour, Declined lungs also had higher levels of ET-1 (2.7 ⫾ 2.2 vs 1.3 ⫾ 1.1 pg/ml, p ¼ 0.007) and Big ET-1 (31.7 ⫾ 17.4 vs 19.4 ⫾ 9.5 pg/ml, p ¼ 0.007) at 4 hours of EVLP compared with Controls. The AUC for a ROC curve of ET-1 in Declined vs Control was 0.75. With a cutoff value of 1.49 pg/ml, there was a diagnostic sensitivity of 80% and specificity of 72%. For Big ET-1, a cutoff value of 24.3 pg/ml had a diagnostic sensitivity of 73.3%, specificity of 75%, and an accuracy of 75% for Declined vs Control lungs. ECE levels were similar among the groups (Figure 3).
ET-1 axis analysis in DCD lungs is predictive of graft performance on EVLP and after transplantation In BDD lungs, the ET-1 axis did not show significant differences among the groups (data not shown). However for DCD cases, PGD3 and Declined lungs had higher ET-1 and Big ET-1 levels at 4 hours of perfusion compared with Controls (PGD3 vs Control: ET-1, p ¼ 0.03; Big ET-1, p ¼ 0.01. Declined vs Control: ET-1, p ¼ 0.007; big ET-1, p ¼ 0.003). There were no differences in ECE levels among groups. The AUCs for Big ET-1 at 4 hours of EVLP in Declined vs Control and PGD3 vs Control were 0.92 and 0.96, respectively (Figure 4). At the same time point, ET-1
852 Table 1
The Journal of Heart and Lung Transplantation, Vol 34, No 6, June 2015 Donor and Recipient Baseline Demographics
Variablesa Donor variables Age, years DCD Male Last PaO2/FIO2, mm Hg Smoker Recipient variables Age, years Male Diagnosis, % ILD Retransplantation Bilateral
Control (n ¼ 25)
PGD3 (n ¼ 7)
p-value
Declined (n ¼ 16)
p-value
45 ⫾ 14 48 48 366 ⫾ 93 52
44 ⫾ 16 43 14 326 ⫾ 104 71
0.981 1 0.195 0.361 0.426
39 ⫾ 17 44 62 335 ⫾ 105 37
0.278 1 0.522 0.574 0.522
48 ⫾ 14 48 30 4 100
49 ⫾ 11 29 71 0 100
1 0.426 0.078 1 1
DCD, donation after cardiac death; ILD, interstitial lung disease; FIO2, fraction of inspired oxygen; PaO2 partial pressure of arterial oxygen; PGD3, primary graft dysfunction Grade 3. a Continous data are presented as mean ⫾ standard deviation and categoric data as percentage.
had an accuracy of 91% to differentiate PGD3 from Controls and of 88% to differentiate Declined from Controls.
compared with Controls at 1 hour of EVLP (p ¼ 0.003 for both comparisons). At 4 hours of EVLP, NOx concentrations were again higher in Declined lungs compared with Controls (p ¼ 0.01; Figure 5).
NOx are significantly higher in lungs declined after EVLP
Discussion
ET-1 signals through ETB receptors in endothelial cells to produce NO, which then acts on epithelial alveolar cells to decrease expression of Naþ/Kþ adenosine 50 -triphosphatase in the basolateral surface.7 This mechanism impairs alveolar fluid clearance and ultimately leads to edema formation. We thus measured NOx in perfusates of high-risk donor lungs that underwent EVLP. Interestingly, NOx concentrations were significantly higher in Declined and PGD3 lungs
We have demonstrated that Big ET-1 and ET-1 measured in perfusate samples during EVLP have prognostic significance. At 1 and 4 hours of perfusion, Big ET-1 and ET-1 levels were higher in lungs that were declined after EVLP compared with EVLP lungs transplanted with good early outcomes (Controls). Analysis of the AUC for Big ET-1 reveals a reasonable accuracy to predict, early (i.e., at 1 hour of perfusion) lungs that would be ultimately declined
Table 2
Physiologic Ex Vivo Lung Perfusion Parameters
Variablea PAP, mm Hg Control PGD3 Declined Airway pressure Mean, cm H2O Control PGD3 Declined Plateau, cm H2O Control PGD3 Declined PaO2/FIO2 mm Hg Control PGD3 Declined
1 hour
2 hours
3 hours
4 hours
p-valueb
9.4 ⫾ 2.1 10.4 ⫾ 2.6 10 ⫾ 2
9.8 ⫾ 2 10.8 ⫾ 2.4 10.3 ⫾ 2
9.8 ⫾ 2.1 11 ⫾ 2.5 10.2 ⫾ 2.3
9.7 ⫾ 2.2 10.7 ⫾ 2.4 10 ⫾ 3
0.269 0.504
7.2 ⫾ 0.7 7.2 ⫾ 0.7 7.8 ⫾ 0.8
7.1 ⫾ 0.9 7.1 ⫾ 0.6 7.8 ⫾ 0.9
7.1 ⫾ 0.9 7.2 ⫾ 0.7 7.8 ⫾ 1.1
7.2 ⫾ 0.9 7.1 ⫾ 0.6 7.9 ⫾ 1.1
0.882 0.015
12.3 ⫾ 1.7 13 ⫾ 2.5 14.7 ⫾ 3.3
12 ⫾ 1.5 12.7 ⫾ 2.2 14.2 ⫾ 3.1
12 ⫾ 1.7 12.5 ⫾ 2.2 13.6 ⫾ 3.1
12.4 ⫾ 1.6 13.2 ⫾ 2.4 14.3 ⫾ 3.1
0.479 0.011
495.9 ⫾ 88.7 552.1 ⫾ 63.2 404.3 ⫾ 103.9
532.5 ⫾ 65.9 567.5 ⫾ 33.5 414.7 ⫾ 99
544.7 ⫾ 58.4 573.6 ⫾ 44.4 448.6 ⫾ 84.3
541.6 ⫾ 58.5 560.3 ⫾ 43.3 446 ⫾ 73.1
0.015 o0.0001
FIO2, fraction of inspired oxygen; PaO2, partial pressure of oxygen; PAP, pulmonary artery pressure. a Data are expressed as mean ⫾ standard deviation. b The p-values correspond to comparison between each respective group with Control using 2-way analysis of variance.
Machuca et al. Table 3
ET-1 As Biomarker for EVLP Outcome
853
Detailed Demographics for Declined Lungs and Reason to Decline During Ex Vivo Lung Perfusion
Donor Age (years)
Type
44 57 53 56
DCD DCD NDD DCD
35 20
Smoker
Last PaO2/FIO2
Abnormal CXR
Bronchoscopy
Main indication for EVLP a
DCD DCDa Low PaO2/FIO2 Edema CXR and at procurement Abnormal CXR Edematous lower lobes
Main reason to decline after EVLP Not meeting EVLP PaO2/FIO2 criteriab Worsening compliance Worsening compliance Not meeting EVLP PaO2/FIO2 criteria,b worsening compliance Worsening compliance Worsening edema CXR and bronchoscopy Frothy secretions, worsening compliance Worsening edema CXR and bronchoscopy Worsening edema CXR, bronchoscopy Worsening edema CXR, bronchoscopy Not meeting EVLP PaO2/FIO2 criteria,b emphysema Not meeting EVLP PaO2/FIO2 criteriab Worsening edema CXR
Yes No No
420 325 297 378
No No Yes Yes
Purulent Clear Clear Bloody
NDD NDD
Yes Yes
404 440
Yes Yes
Clear Mucoid
67
NDD
No
92
Yes
Aspiration?
25
DCD
No
392
Yes
Clear
36
NDD
No
381
Yes
Clear
Low PaO2/FIO2, aspiration suspicion WLST to arrest 70 minutes, fat emboli Edema lower lobes
28
DCD
No
501
Yes
Bloody
Edema lower lobes
49
DCD
Yes
407
No
Clear
20 15
NDD NDD
No No
335 222
Yes Yes
Clear Bloody
16 50
DCD NDD
No No
336 250
Yes Yes
Purulent Clear
60
NDD
Yes
185
Yes
Clear
Heavy smoker, poor deflation, bullae Edema lower lobes Low PaO2/FIO2, edema lower lobes Edema lower lobes Not meeting EVLP PaO2/FIO2 criteriab Worsening edema CXR Low PaO2/FIO2, preserved deflated long CIT (outside procurement team) Poor deflation Not meeting EVLP PaO2/FIO2 criteria,b poor deflation
CIT, cold ischemic time; CXR, chest X-ray; DCD, donation after cardiac death; FIO2, fraction of inspired oxygen; NDD, neurological determination of death; PaO2, partial pressure of arterial oxygen; WLST, withdrawal of life-sustaining therapies. a Represents a case in which the indication for EVLP was a DCD donor. b EVLP PaO2/FIO2 criteria Z 400 mm Hg.
because of subsequent development of poor physiologic performance. This information could be useful to select, early in the course of EVLP, lungs that will require additional therapeutic measures to achieve transplantability criteria. Although this study does not provide the posttransplant function of lungs declined because of worsening physiologic parameters under the protective measures of EVLP, experimental data suggest an association with a high risk of PGD3.3 More importantly, our findings suggest that for DCDs, the ET-1 axis assessment shows a strong correlation between perfusate levels and development of PGD3. Translation of our findings to clinical practice stands to deliver a substantial effect given that EVLP physiologic parameters of lungs that developed PGD3 were similar to those of the Control lungs. Thus, the availability of perfusate biomarkers (possibly Big ET-1 and ET-1 for DCD lungs) in a timely fashion could help improve the precision of donor lung selection, bringing more objectivity to the decision-making process. The stronger significance of the ET-1 axis in DCDs is intriguing and might potentially reflect different mechanisms of donor lung injury. On one hand, DCD lungs may be spared the catecholamine
storm related to brain death, but on the other hand, they are prone to periods of hypotension, low shear stress, aspiration injury, and hypoxia.16 We previously showed different transcriptional signatures between DCD and BDD donor lungs, particularly at the end of cold ischemic time.17 More recently, we showed similar results in a larger cohort of EVLP lungs where lung tissue biopsies at the end of the cold ischemic time demonstrated 285 differentially expressed genes between DCD and BDD.18 ET-1 is mostly known for its pathogenetic involvement in pulmonary arterial hypertension but has also been associated with acute lung injury. Higher circulating levels of ET-1 were found in patients with acute lung injury compared with healthy controls.19 That ET-1 levels fell to levels within normal reference ranges in those patients with acute lung injury/adult respiratory distress syndrome that eventually recovered is further supporting evidence for a potential mediator role of ET-1 in lung injury.20 The evaluation of the ET-1 axis in lung transplantation presents biologic plausibility because its expression is regulated by several mechanisms involved with the organ
854
The Journal of Heart and Lung Transplantation, Vol 34, No 6, June 2015
Figure 2 Endothelin-1 (ET-1) axis protein levels at 1 hour of ex vivo lung perfusion (EVLP). (A) ET-1 levels were significantly higher in Declined lungs compared with Controls (*p ¼ 0.01). (B) Big ET-1 levels were significantly higher in Declined compared with Controls (**p ¼ 0.001). (C) Endothelin-converting enzyme (ECE) levels did not differ among the groups. (D) The area under curve (AUC) of the receiver operating characteristic curve revealed that Big ET-1 at 1 hour of EVLP had a reasonable accuracy (0.79) to differentiate Declined from Control donor lungs. For A-C, bars in the scatter plot represent the mean and standard error. PGE3, primary graft dysfunction Grade 3.
donation process. Hormones (cortisol, adrenaline, insulin), cytokines (interleukin-1β) and physicochemical stimuli (hypoxia, low shear stress) favor increased ET-1 expression, and prostacyclin, heparin, and high shear stress inhibit it.21 In an experimental model of donor lung injury by prolonged cold ischemic time (18–20 hours), grafts transplanted into recipients treated with a non-selective ET-1 receptor antagonist presented significantly better lung function and
were able to sustain life once the contralateral pulmonary artery was clamped.11 In the clinical setting, ET-1 mRNA levels were higher in pre-implantation biopsy specimens from donor lungs that developed PGD3 compared with lungs without PGD3.8 Moreover, the authors also analyzed recipient samples in that study and were able to show that a combination of high donor lung tissue ET-1 mRNA with high recipient pre-transplant serum ET-1 was highly
Figure 3 Endothelin-1 (ET-1) axis protein levels at 4 hours of ex vivo lung perfusion (EVLP). (A) ET-1 levels were significantly higher in Declined lungs compared with Controls (**p ¼ 0.007). (B) Big ET-1 levels were significantly higher in Declined compared with Controls (**p ¼ 0.007). (C) Endothelin-converting enzyme (ECE) levels did not differ among groups. (D) The area under curve (AUC) of the receiver operating characteristic curve revealed that Big ET-1 at 4 hours of EVLP had a reasonable accuracy (0.75) to distinguish Declined from Control donor lungs. For A-C, bars in the scatter plot represent the mean and standard error. PGD3, primary graft dysfunction Grade 3.
Machuca et al.
ET-1 As Biomarker for EVLP Outcome
855
Figure 4 Endothelin-1 (ET-1) axis protein levels in donation after cardiac death (DCD) lungs during ex vivo lung perfusion (EVLP). (A) ET-1 levels at 1 hour of EVLP were significantly higher in Declined lungs compared with Controls (*p ¼ 0.04). (B), Big ET-1 levels at 1 hour of EVLP were significantly higher in Declined compared with Controls (**p ¼ 0.007). (C), ET-1 levels at 4 hours of EVLP were significantly higher in PGD3 (*p ¼ 0.03) and Declined lungs (**p ¼ 0.007) compared with Controls. (D) Big ET-1 levels at 4 hours of EVLP were significantly higher in primary graft dysfunction Grade 3 (PGD3) lungs (*p ¼ 0.01) and Declined lungs (**p ¼ 0.003) compared with Controls. (E) The area under curve (AUC) of the receiver operating characteristic (ROC) curve revealed that Big ET-1 at 4 hours of EVLP had excellent accuracy (0.92) to distinguish Declined from Control donor lungs. (F) The AUC of the ROC curve revealed that Big ET-1 at 4 hours of EVLP had excellent accuracy (0.96) to distinguish PGD3 from Control donor lungs. For A-D, bars in the scatter plot represent the mean and standard error.
correlated with PGD3. Interestingly, the inverse situation, low donor lung tissue ET-1 mRNA and low recipient plasma ET-1 also correlated with the absence of PGD. This highlights the point that although we are focusing on donor lung–related predictive factors, ultimate models of prediction of PGD after lung transplantation should include consideration of recipient factors that undoubtedly do contribute to PGD.22 Several mechanisms have the potential to impair the alveolar–capillary barrier through ET-1, namely: (1) the increase in capillary hydrostatic pressure; (2) the increase in capillary permeability mediated by upregulation of vascular
endothelial growth factor; (3) the recruitment of inflammatory cells; and (4) the decrease in alveolar fluid clearance.6 Our observation of increased NOx concomitant with increased levels of ET-1 axis proteins in Declined lungs points toward ET-1–mediated impairment of alveolar fluid clearance as a plausible mechanism leading to physiologic deterioration of damaged human lungs that cannot be rescued with the current EVLP technique. Evidence of NO as a mediator of impaired alveolar fluid clearance is found in the work of Kaestle et al,23 who performed isolated perfusion with rat lungs. They demonstrated in a model of hydrostatic pulmonary edema that
856
The Journal of Heart and Lung Transplantation, Vol 34, No 6, June 2015
Figure 5 (A) Perfusate concentrations of nitric oxide metabolites (NOx) at 1 hour of ex vivo lung perfusion (EVLP) were significantly higher in primary graft dysfunction Grade 3 (PGD3) and Declined lungs compared with Controls (**p ¼ 0.003 for both). (B) Perfusate NOx levels at 4 hours of EVLP were significantly higher in Declined lungs compared with Controls (*p ¼ 0.01). The bars in the scatter plot represent the mean and standard error.
increased fluid filtration is responsible for only 30% of the net fluid shift, with the other 70% attributable to NO-induced impairment in alveolar fluid clearance. Whereas the NO synthase inhibitor L-NG-nitro-L-arginine methyl ester prevented edema formation, the addition of NO (S-nitrosoglutathione and 1-propamine 3-[2-hydroxy-2-nitroso-1-propylhydrazine] NONOate) to the perfusate blocked fluid reabsorption. More recently, Comellas et al7 performed mechanistic studies providing evidence pointing to ETB receptors as the main responsible factor for impairment in alveolar fluid clearance. Combining rat lung perfusion models with cell culture experiments, they demonstrated the important crosstalk between endothelial and epithelial alveolar cells through ET-1-ETB receptor signaling via NO to decrease the expression of Naþ/Kþ adenosine 50 -triphosphatase in the basolateral membrane of alveolar epithelial cells, ultimately leading to pulmonary edema. Interestingly, the addition of ET1 to the perfusate, but not its instillation into the airway, led to a decrease in alveolar fluid clearance by 40% to 50%. These findings in experimental isolated lung perfusion corroborate our findings in clinical EVLP and also reinforce the importance of measuring ET-1 levels in the vascular compartment. Differently from chemokines and cytokines, which will tend to accumulate in the EVLP perfusate due to the lack of renal and/or hepatic clearance mechanisms, ET-1 is partially produced and cleared in the pulmonary circulation. In models of lung perfusion, 31% to 38% of radiolabeled ET-1 was cleared by a single pass.12 The clearance of ET-1 is closely related to ETB receptors because its selective blockade significantly increases circulating ET-1 levels.24 Interestingly in our study, levels of Big ET-1 and ECE significantly increased over time, whereas the levels of ET-1 remained relatively stable from 1 to 4 hours. These findings are suggestive of active ET-1 clearance mechanisms in donor lungs perfused ex vivo. Because the ET-1 axis can be manipulated pharmacologically, our findings have therapeutic implications. Although selective blockade of ET-1 has been shown to elicit pulmonary edema, its non-selective blockade may be beneficial in improving alveolar fluid clearance and
avoiding the formation of pulmonary edema. In this scenario, EVLP would provide the ideal platform for donor lung treatment and reassessment ex vivo before moving forward to a successful transplantation.5 The main limitation of our study is the sample size. Nevertheless, this represents the largest clinical EVLP experience to date using marginal donor lungs. Further validation with an independent set of patients will be required before bringing this biomarker to clinical use. In conclusion, we have shown temporal changes and the prognostic relevance of the ET-1 axis in EVLP for clinical lung transplantation. In DCD lungs, higher levels of big ET-1 and ET-1 were observed at 4 hours of EVLP and demonstrated a strong correlation with PGD after transplantation. Further studies will be instrumental to validate our results and determine whether ET-1 will be an important biomarker for translation to clinical decision making in donor lung selection.
Disclosure statement S.K. and M.C. were principal investigators for the Toronto ExVivo Lung Perfusion Trial, sponsored by XVIVO Perfusion (Gothenburg, Sweden). S.K., M.C., and T.K.W. are founding members of Perfusix Inc, a company that provides ex vivo organ perfusion services. None of the other authors has a financial relationship with a commercial entity that has an interest in the subject of the presented manuscript or other conflicts of interest to disclose. T.N.M. is supported by a Research Fellowship from the American Society of Transplantation.
Supplementary data Supplementary data associated with this article can be found in the online version at www.jhltonline.org.
References 1. Cypel M, Yeung JC, Liu M, et al. Normothermic ex vivo lung perfusion in clinical lung transplantation. N Engl J Med 2011;364:1431-40.
Machuca et al.
ET-1 As Biomarker for EVLP Outcome
2. Cypel M, Yeung JC, Machuca T, et al. Experience with the first 50 ex vivo lung perfusions in clinical transplantation. J Thorac Cardiovasc Surg 2012;144:1200-6. 3. Yeung JC, Cypel M, Machuca TN, et al. Physiologic assessment of the ex vivo donor lung for transplantation. J Heart Lung Transplant 2012; 31:1120-6. 4. Machuca TN, Cypel M, Keshavjee S. Advances in lung preservation. Surg Clin North Am 2013;93:1373-94. 5. Machuca TN, Hsin MK, Ott HC, et al. Injury-specific ex vivo treatment of the donor lung: pulmonary thrombolysis followed by successful lung transplantation. Am J Respir Crit Care Med 2013;188:878-80. 6. Comellas AP, Briva A. Role of endothelin-1 in acute lung injury. Transl Res 2009;153:263-71. 7. Comellas AP, Briva A, Dada LA, et al. Endothelin-1 impairs alveolar epithelial function via endothelial ETB receptor. Am J Respir Crit Care Med 2009;179:113-22. 8. Salama M, Andrukhova O, Hoda MA, et al. Concomitant endothelin-1 overexpression in lung transplant donors and recipients predicts primary graft dysfunction. Am J Transplant 2010;10:628-36. 9. Tikkanen JM, Koskinen PK, Lemstrom KB. Role of endogenous endothelin-1 in transplant obliterative airway disease in the rat. Am J Transplant 2004;4:713-20. 10. Khimenko PL, Moore TM, Taylor AE. Blocked ETA receptors prevent ischemia and reperfusion injury in rat lungs. J Appl Physiol (1985) 1996;80:203-7. 11. Shennib H, Lee AG, Kuang JQ, et al. Efficacy of administering an endothelin-receptor antagonist (SB209670) in ameliorating ischemiareperfusion injury in lung allografts. Am J Respir Crit Care Med 1998; 157:1975-81. 12. Dupuis J, Goresky CA, Stewart DJ. Pulmonary removal and production of endothelin in the anesthetized dog. J Appl Physiol (1985) 1994;76:694-700. 13. Christie JD, Carby M, Bag R, et al. Report of the ISHLT Working Group on Primary Lung Graft Dysfunction part II: definition.
857
14.
15.
16.
17.
18.
19. 20.
21. 22.
23.
24.
A consensus statement of the International Society for Heart and Lung Transplantation. J Heart Lung Transplant 2005;24:1454-9. Grasemann C, Ratjen F, Schnabel D, et al. Effect of growth hormone therapy on nitric oxide formation in cystic fibrosis patients. Eur Respir J 2008;31:815-21. Braman RS, Hendrix SA. Nanogram nitrite and nitrate determination in environmental and biological materials by vanadium (III) reduction with chemiluminescence detection. Anal Chem 1989;61:2715-8. Tremblay LN, Yamashiro T, DeCampos KN, et al. Effect of hypotension preceding death on the function of lungs from donors with nonbeating hearts. J Heart Lung Transplant 1996;15:260-8. Kang CH, Anraku M, Cypel M, et al. Transcriptional signatures in donor lungs from donation after cardiac death vs after brain death: a functional pathway analysis. J Heart Lung Transplant 2011;30:289-98. Zamel R, Bai XH, Yeung JC, et al. Differential gene regulation between brain and cardiac death donor lungs during ex vivo lung perfusion. J Heart and Lung Transplant 2012;31:S146-7. Mitaka C, Hirata Y, Nagura T, et al. Circulating endothelin-1 concentrations in acute respiratory failure. Chest 1993;104:476-80. Langleben D, DeMarchie M, Laporta D, et al. Endothelin-1 in acute lung injury and the adult respiratory distress syndrome. Am Rev Respir Dis 1993;148:1646-50. Shao D, Park JE, Wort SJ. The role of endothelin-1 in the pathogenesis of pulmonary arterial hypertension. Pharmacol Res 2011;63:504-11. Diamond JM, Lee JC, Kawut SM, et al. Clinical risk factors for primary graft dysfunction after lung transplantation. Am J Respir Crit Care Med 2013;187:527-34. Kaestle SM, Reich CA, Yin N, et al. Nitric oxide-dependent inhibition of alveolar fluid clearance in hydrostatic lung edema. Am J Physiol Lung Cell Mol Physiol 2007;293:L859-69. Dupuis J, Goresky CA, Fournier A. Pulmonary clearance of circulating endothelin-1 in dogs in vivo: exclusive role of ETB receptors. J Appl Physiol (1985) 1996;81:1510-5.