INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Keywords: interleukin-8; lung; transplantation; cytokines; reperfusion injury

Improvements in immunosuppression, surgical technique, and organ preservation have resulted in high success rates for lung transplantation. In our center, the introduction of low-potassium dextran (LPD) preservation solution allowed safer lung preservation for longer ischemic durations with improved postoperative graft function (1, 2). However, in 10% to 15% of patients, severe reperfusion injury still leads to poor initial graft function. In addition to postoperative mortality, severe reperfusion injury may also contribute to poor long-term outcome by increasing the incidence of acute rejection and chronic graft dysfunction (3, 4).

Cytokines have been shown to play a critical role in modulating inflammatory processes and in enhancing cellular infiltration in injured tissue. Experimental studies have shown that ischemia-reperfusion of solid organs such as the kidney (5), liver (6), heart (7), and lung (8) induces a rapid release of proinflammatory cytokines, including tumor necrosis factor- (TNF-), interferon- (IFN-), interleukin (IL)-1, IL-6, and IL-8. In clinical settings, Boros and coworkers (9) have shown that prolonged elevation of IL-6 and IL-8 in plasma correlated with poor early graft function in liver transplantation, and Le Moine and coworkers (10) have reported that reperfusion of the liver induces a transient systemic release of IL-10. Because IL-10 induces the expression of HLA-G, a histocompatibility class I antigen potentially involved in immunotolerance, its release may contribute to the reduced immunogenic potential and the tolerogenic properties of the transplanted liver (11, 12).

In human lung transplantation, cytokine expression and release have been examined in the setting of infection and acute allograft rejection (13, 14). However, cytokine modulation during ischemia-reperfusion injury and its correlation with donor parameters and graft function has not been systematically explored. Hence, in an attempt to gain insight into the pathophysiological mechanisms of preservation and reperfusion injury, we examined the levels of key pro- and antiinflammatory cytokines at predefined points during the perioperative phase of lung transplantation. We then examined the relationships between the cytokine levels and donor factors, length of ischemic time, and allograft function.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patients

We studied 18 consecutive patients undergoing bilateral lung transplantation. Characteristics of recipients and donors are presented in Tables 1 and 2, respectively.

Anesthesia and Surgery

Retrieval procedures and recipient operation were performed according to our standard protocol (15). When transplantation was performed using cardiopulmonary bypass (n = 6), ventilation and low-pressure perfusion were provided to the newly implanted first allograft during implantation of the second. Total ischemic time (TIT) was divided into cold and warm ischemic periods. The cold ischemic time (CIT) was calculated as the interval from the start of the pulmonary artery flush to the transfer of the lung into the pleural cavity. The warm ischemic time (WIT) corresponded to the period for lung implantation.

Postoperative Management

Data were prospectively collected in order to measure the APACHE score (16) within the first 24 postoperative hours, the arterial oxygen pressure/fraction of inspired oxygen (Pa_O2/FI_O2) ratio on arrival in the intensive care unit (ICU) and after 12, 18, and 24 h, and the mean airway pressure after 12, 18, and 24 h in the ICU. The ICU length of stay was expressed as the ICU-free days in 30 d, which corresponds to the number of days out of the ICU during the first 30 postoperative days.

Fatal Outcome

Four patients died within 30 d of surgery, two from severe reperfusion injury. Both patients had received lungs from marginal donors with diffuse lung infiltrate on chest X-ray, and either an initial Pa_O2 < 300 mm Hg or previous cardiac arrest. A third patient died on postoperative Day 4 from massive pulmonary embolism, and one additional patient died on postoperative Day 12 from a presumed cardiac arrhythmia, despite an uneventful early postoperative course and impending hospital discharge.

Lung Tissue Samples

Lung tissue samples were collected at four time points corresponding to the end of the CIT (sample 1), the end of the WIT (sample 2), and 1 and 2 h of reperfusion (samples 3 and 4, respectively). Sample 1 was taken from the right or left lung at the end of the first CIT, whereas samples 2, 3, and 4 were always taken from the first implanted lung. All samples were immediately snap frozen in liquid nitrogen and stored at 70° C.

Cytokine Measurement

Tissues were homogenized, sonicated, and centrifuged (17, 18). Supernatants were assayed in duplicate using specific enzyme-linked immunosorbent assay (ELISA) kits for human TNF-, IFN-, IL-8, IL-10, IL-12, and IL-18 according to the manufacturer's instructions. The protein content was determined by the Bradford's method (19). The final cytokine profiles were expressed as picogram cytokine per milligram total protein.

Statistical Analysis

Results were compared by one-way analysis of variance (ANOVA), Kruskal-Wallis, and Student's t test when necessary. Differences in cytokine measurement at each time point were analyzed by repeated measures ANOVA (Friedman test). Levels of cytokines were correlated by linear regression to donor and recipient parameters. A p value less than 0.05 is considered significant. Results are expressed as mean ± SD, or as median value and range. The Graphpad software package (Graphpad Software, Inc., San Diego, CA) was used for all statistical analyses.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cytokine Levels

The cytokine profile varied according to the different time points (Table 3). TNF-, IFN-, IL-10, and IL-18 were significantly higher during the ischemic time and decreased after reperfusion. IL-12 significantly decreased during the WIT and remained at lower levels afterward. IL-8 levels, in contrast, tended to increase gradually over time. The two patients who died from ischemia-reperfusion injury had significantly higher levels of IL-8 during the ischemic time and after reperfusion than the remaining 16 patients (Table 4). TNF-, IFN-, IL-10, IL-12, and IL-18 were not significantly different between the two patients who died from ischemia-reperfusion injury and the remaining 16 patients.

Correlation between Donor Parameters and Cytokine Levels

Donor age negatively correlated with IL-10 levels after 1 h (p = 0.0007) and 2 h of reperfusion (p = 0.0006) (Figure 1). Other donor parameters such as sex, etiology of brain death, time on ventilator, smoking history, and Pa_O2 did not correlate with the cytokine levels. Although overall the presence of an abnormal chest X-ray did not correlate with cytokine levels, it should be noted that the two patients who had high levels of IL-8 during the ischemic time and who subsequently died from ischemia-reperfusion injury had diffuse infiltrates on the chest X-ray.

View larger version (10K): [in this window] [in a new window]   Figure 1.   Correlation between donor age and IL-10 levels. Donor age negatively correlated with IL-10 after 1 h (A) and 2 h of reperfusion (B). (A) p = 0.007, r 2 = 0.5; (B) p = 0.007, r 2 = 0.5.

Correlation between Ischemic Times and Cytokine Levels

The length of CIT and TIT positively correlated with the level of IL-18 at the end of the WIT (p = 0.0007). Other cytokines were not influenced by the ischemic time.

Correlation between Graft Function and Cytokine Levels

The level of IL-8 significantly correlated with early graft function (Figure 2). Indeed, the level of IL-8 at 2 h after reperfusion negatively correlated with the Pa_O2/FI_O2 on arrival in the ICU (p = 0.002), and after 12 (p = 0.005), 18 (p = 0.01), and 24 (p = 0.001) h in the ICU. The level of IL-8 at 2 h after reperfusion positively correlated with the APACHE score (p = 0.005) and the mean airway pressure after 12 (p = 0.0002), 18 (p = 0.02), and 24 (p = 0.03) h in the ICU. In addition, the level of IL-8 at 2 h after reperfusion negatively correlated with the ICU-free days in 30 d (p = 0.02). If the APACHE score was broken down into its different components: the Acute Physiologic Score, the age and health status of the recipient (5), we found that IL-8 correlated most significantly with the Acute Physiologic Score (p = 0.006).

View larger version (21K): [in this window] [in a new window]   Figure 2.   Correlation between IL-8 and graft function. IL-8 after 2 h of reperfusion negatively correlated with the PaO2 /FIO2 ratio on arrival in the ICU (A) and after 12 h in the ICU (B). In addition, it positively correlated with the APACHE score measured during the first 24 h in the ICU (C  ) and with the mean airway pressure after 12 h in the ICU (D). The correlations with the PaO2/FIO2 ratio and with the mean airway pressure after 18 h (p = 0.01 and p = 0.02, respectively) and 24 h (p = 0.001 and p = 0.03, respectively) in the ICU are not represented on this figure. (A) p = 0.002, r2 = 0.5; (B) p = 0.005, r2 = 0.5; (C) p = 0.005, r2 = 0.3; (D) p = 0.002, r2 = 0.6.

A similar correlation in graft function and patient outcome was observed even if the two patients that died from reperfusion injury were excluded from the analysis. Other cytokines did not show a consistent correlation with graft function or patient outcome.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The primary finding in this study was the striking relationship between IL-8 and graft function after lung transplantation. IL-8 is rapidly released after reperfusion and levels in lung tissue 2 h after reperfusion correlated with lung function assessed by the Pa_O2 /  FI_O2 ratio, the mean airway pressure, and the APACHE score during the first 24 postoperative hours (Figure 2). In addition, the length of stay in the ICU correlated with IL-8 levels at 2 h after reperfusion.

A similar role for IL-8 has been observed in human liver transplantation (9, 20). Mueller and coworkers (20) have shown that IL-8 peaked in the blood taken from the systemic circulation 2 h after reperfusion and rapidly decreased thereafter. In addition, the levels of IL-8 in the effluent hepatic veins immediately after reperfusion and in the systemic circulation 2 and 24 h after reperfusion correlated with the severity of reperfusion injury (9, 20). In contrast to the findings in liver transplantation, we did not find any correlation between the levels of other proinflammatory cytokines, including TNF-, IFN-, IL-12, and IL-18 and early lung function (9, 20).

IL-8 is a potent neutrophil chemotactic factor, promoting neutrophil adhesion, migration, and degranulation (21). Neutrophils have been implicated in the pathogenesis of ischemia-reperfusion injury after lung transplantation and are primarily involved in the delayed inflammatory response that occurs within a few hours after reperfusion (24, 25). Hence, the release of IL-8 early after reperfusion could initiate neutrophil recruitment and secondary lung injury.

The potential importance of IL-8 has also been demonstrated in acute inflammatory lung disease (26). The level of IL-8 in the bronchoalveolar lavage (BAL) was shown to be significantly higher in patients with acute respiratory distress syndrome (ARDS) than in patients at risk of developing ARDS and in patients with chronic obstructive lung disease (COPD) (27, 28). In addition, the level of IL-8 was even higher among patients who died than among those who survived ARDS (28).

Recently, Fisher and coworkers (29, 30) have shown that lung transplant patients who died from early lung dysfunction had significantly higher levels of IL-8 in the donor BAL before retrieval. This is consistent with our findings: the two patients who died from severe reperfusion injury in our series had significantly higher levels of IL-8 in lung tissue during the ischemic time and after reperfusion than the remaining 16 patients (Table 4). Such high levels of IL-8 before and after reperfusion may be a marker of previous aspiration, ongoing infection, or neurogenic pulmonary edema that should preclude the use of these donor lungs (31). In contrast to Fisher and coworkers (29), we have further observed that IL-8 is an important cytokine even in patients who do not develop life-threatening reperfusion injury. Indeed, in these patients, IL-8 increases significantly 2 h after reperfusion and its level correlates significantly with lung function. This finding suggests that the administration of a neutralizing monoclonal antibody against IL-8 at the time of reperfusion might prevent neutrophil infiltration and reduce tissue injury. Similar findings have been observed experimentally by Sekido and coworkers (34). They have shown that IL-8 increases significantly in the lung tissue of rabbits 3 h after reperfusion and that the intravenous administration of anti-IL-8 antibody at the beginning of the reperfusion period markedly reduced infiltration of neutrophils, fibrinous exudation in the alveolar lumina, and destruction of alveolar architecture (34).

In contrast to liver transplantation, we did not find a significant release of the antiinflammatory cytokine IL-10 after reperfusion in lung transplantation (9, 20). However, we did observe a significant decline in the release of IL-10 in lung tissue after reperfusion in older donors (Figure 1). Interestingly, the release of IL-10 has also been shown to be decreased in older male mice subjected to the stressful event of trauma-hemorrhage (35, 36). This finding may, thus, explain why lungs from older donors are more susceptible to ischemic injury and are associated with a higher mortality rate than lungs from younger donors (37).

IL-12 and IL-18 are two cytokines involved in the immune response to bacterial infections (38). Lately, Lentsch and coworkers (39) and Daemen and coworkers (40) have shown in a murine model of warm ischemia that these cytokines may also play a role in ischemia-reperfusion injury of the liver and kidney by inducing the release of TNF- and IFN-, and by enhancing major histocompatibility complex (MHC) class I and II expression. We observed that both IL-12 and IL-18 were significantly higher during the ischemic time than after reperfusion. In addition, IL-18 was the only cytokine that correlated with the length of ischemic time in our study. Because longer ischemic times have been shown to induce the expression of MHC class II (4), our finding suggests that long ischemic times may influence acute rejection and subsequent chronic allograft dysfunction through the release of IL-18.

In conclusion, this study demonstrates the importance of IL-8 in predicting early graft function. It also shows that IL-18 correlates with the length of ischemic time and confirms that IL-12 and IL-18 may play a role in human lung reperfusion injury. Although donor factors did not influence the release of proinflammatory cytokines, we observed that older donors had lower levels of IL-10 after reperfusion. Further studies are required to delineate the role of the pro- and antiinflammatory cytokines in ischemia-reperfusion injury. This understanding will hopefully lead to strategies designed to modify the expression and release of these cytokines. For instance, neutralization of IL-8 at the time of reperfusion or administration of IL-10 in selected donors before lung retrieval may be beneficial in reducing early graft dysfunction.