Prolonged dialysis during ex vivo lung perfusion promotes inflammatory responses

Abstract

Ex-vivo lung perfusion (EVLP) has extended the number of transplantable lungs by reconditioning marginal organs. However, EVLP is performed at 37°C without homeostatic regulation leading to metabolic wastes' accumulation in the perfusate and, as a corrective measure, the costly perfusate is repeatedly replaced during the standard of care procedure. As an interesting alternative, a hemodialyzer could be placed on the EVLP circuit, which was previously shown to rebalance the perfusate composition and to maintain lung function and viability without appearing to impact the global gene expression in the lung. Here, we assessed the biological effects of a hemodialyzer during EVLP by performing biochemical and refined functional genomic analyses over a 12h procedure in a pig model. We found that dialysis stabilized electrolytic and metabolic parameters of the perfusate but enhanced the gene expression and protein accumulation of several inflammatory cytokines and promoted a genomic profile predicting higher endothelial activation already at 6h and higher immune cytokine signaling at 12h. Therefore, epuration of EVLP with a dialyzer, while correcting features of the perfusate composition and maintaining the respiratory function, promotes inflammatory responses in the tissue. This finding suggests that modifying the metabolite composition of the perfusate by dialysis during EVLP can have detrimental effects on the tissue response and that this strategy should not be transferred as such to the clinic.

Keywords: dialysis; ex vivo; ex vivo lung perfusion; ischemia-reperfusion; lung; pig; transplantation.

Figure 1

Monitoring of the physiological parameters (A), electrolyte (B), and metabolite (C) concentrations in the perfusates in the Gold Standard (five pigs) and Pediatric Dialysis groups (four pigs). (A) From left to right: lung compliance, pulmonary artery pressure, pO2 on FiO2 30%. (B) From left to right: Na²⁺, Ca²⁺, Cl⁻ concentrations. (C) From left to right: glucose, lactate production, lactate deshydrogenase (LDH). To compare the data between the two groups, an unpaired t-test was used when the data passed the Shapiro normality test. For cases that did not pass the Shapiro normality test, a non-parametric Mann–Whitney test was performed. The p-value classes are reported as *p < 0.05 or **p < 0.01 at specific time points. The highest p-value class common to several time points is reported above a line. The error bars represent standard deviations.

Figure 2

Cytokine protein and transcript analyses in the Gold Standard and Pediatric Dialysis groups. (A) Cytokines were detected in the perfusates using a porcine cytokine magnetic bead panel kit. At specific time points, when the data followed a normal distribution after Log₁₀ transformation, a t-test was used to compare the data between the two groups; alternatively, a non-parametric Mann–Whitney test was used, *p < 0.05, **p < 0.01 or exact value in case of tendency. (B) The cytokine gene expression ratios were established from the RNA-seq gene counts at 6h and 12h divided by the gene counts at 0h. The same statistical analysis was done as in A at specific time points.

Figure 3

Functional enrichments and predictions of modulated pathways/functions induced by EVLP in the Gold Standard and Pediatric Dialysis groups. Differentially expressed gene lists (DEGs) were established from the RNA-seq results of lung tissue undergoing EVLP at 6h versus 0h and at 12h versus 0h, in the Gold Standard and Pediatric Dialysis groups. (A) The DEGs were subjected to a functional enrichment analysis using the Hallmark gene sets. The functional enrichments were selected based on a −Log₁₀ adjusted p-value superior to 3 and over three contributing genes in at least one of the four conditions. (B, C) The DEGs were loaded and processed through the IPA core analysis for identification of predicted activated of inhibited canonical pathways (B) and functions (C) that were selected based on their biological relevance, absolute value of z-score superior to 2 with a −Log₁₀ p-value superior to 1.3 and over three contributing genes, in at least one of the four cases. The 15 top functions/pathways are shown. The contributing genes to the functional enrichments are reported in Additional file 4 .

Figure 4

Functional enrichments of the gene expression comparisons between the Gold Standard and Pediatric Dialysis groups. (A) A Wilcoxson test was performed to compare the expression fold change (FC) at 6h and 12h between the two groups. An adjusted p-value of 0.05 was used to filter the genes of the expression ratio comparisons and the resulting gene lists were subjected to an enrichment analysis using the IPA and Reactome datasets. The enriched functions with −Log10 p-values > 3 and > 3 contributing genes were selected. The enriched functions in the comparison of FC that is higher in the Pediatric Dialysis group are in red; they are blue when the comparison of FC is higher in the Gold Standard group. (B) The mean gene fold changes of the contributing genes of some of the enriched functions shown in A are illustrated as an heat map, in blue when down-modulated and in red when upregulated at the reported timing versus 0h.

Figure 5

C5a levels in the Gold Standard and Pediatric Dialysis groups. The porcine C5a was detected by ELISA in the perfusates. The raw data followed a normal distribution with equal variance, and were compared between the two groups with an unpaired t-test. No statistically significant difference was found.