High-flux hemodialysis with polymethylmethacrylate membranes reduces soluble CD40L, a mediator of cardiovascular disease in uremia

Abstract

Background and hypothesis: Major adverse cardiovascular events (MACE) are the main cause of mortality in hemodialysis (HD). Soluble CD40 ligand (sCD40L) binds to CD40 on endothelial cells (EC) and vascular smooth muscle cells (VSMC), playing a potential role in MACE. HD registries show a reduced mortality for MACE using the polymethylmethacrylate (PMMA) membrane. Study objectives were (i) to confirm the role of sCD40L as independent predictor and mediator of MACE and (ii) to evaluate the effect of PMMA on sCD40L-mediated vascular aging.

Methods: In 201 patients treated by high-flux HD, sCD40L levels were measured and correlated with MACE; 54/201 patients with sCD40L greater than or equal to the median value were randomized for 9 months in two crossover groups alternatively treated with PMMA or polysulfone (PS): sCD40L and dialytic parameters were recorded. In vitro, the role of sCD40L was studied on EC dysfunction and VSMC calcification after incubation with patients' sera: cells engineered to knock down CD40 by siRNA were also used to confirm the role of CD40-CD40L pathway activation.

Results: At study admission, the sCD40L median level of 8.4 ng/mL (interquartile range 2.9-12.7) showed the best statistical performance to identify MACE, which occurred in 51/201 (25.4%) patients. Indoxyl sulfate and p-cresyl sulfate directly correlated with sCD40L levels and induced its release by platelets. In comparison with PS, PMMA treatment significantly reduced sCD40L levels, in accordance with its enhanced mass removal by adsorption. In vitro, sera collected after PMMA treatment reduced EC dysfunction and VSMC osteoblastic differentiation through a mechanism involving the CD40-CD40L pathway.

Conclusion: sCD40L is an independent predictor and mediator of MACE in chronic HD patients. PMMA membrane stably reduced sCD40L under the high-risk cut-off of 8.4 ng/mL. In vitro studies confirmed the role of PMMA in the reduction of EC dysfunction and VSMC calcification in association with sCD40L modulation.

Keywords: hemodialysis; major adverse cardiovascular events (MACE); polymethylmethacrylate (PMMA) membrane; soluble CD40 ligand; vascular aging.

Graphical Abstract

Figure 1:

Clinical study design, flow-chart and randomization for different HD treatment. Upper panel: study design and flow-chart. Lower panel: randomization to Group 1 or Group 2 and HD treatment at different time points.

Figure 2:

Distribution of sCD40L serum levels in enrolled patients. (A) Histograms showing sCD40L level distribution in all enrolled patients (n = 201). The inset depicts sCD40L levels in two subgroups divided according to the median level of 8.4 ng/mL. Boxes represent IQR with the median value shown as a horizontal bar within each box: bars outside each box show minimum and maximum values. (B) ROC curve for MACE observed during the follow-up according to sCD40L levels at study admission. (C) Kaplan–Meier curve showing cardiovascular event free survival in the two subgroups divided according to the median sCD40L level of 8.4 ng/mL.

Figure 3:

Correlation between sCD40L serum levels and PBUT. (A) Linear correlation between sCD40L (ng/mL) and IS (expressed in mg/L): r = 0.5270, r² = 0.2778, P < .00001. (B) Linear correlation between sCD40L (ng/mL) and pCS (expressed in mg/L): r = 0.5774, r² = 0.3334, P < .00001.

Figure 4:

In vitro evaluation of sCD40L release from human platelets incubated with PBUT. Increasing doses (PBUT 10, 50 or 100 µg/mL) of IS and pCS induced a significant release of sCD40L assessed by ELISA after 1 h (gray plots) or 24 h (white plots); (*P < .05 PBUT 10, 50 or 100 µg/mL vs vehicle at both h and 24 h). Data are expressed as average ± 1 SD of five different experiments.

Figure 5:

Modulation of sCD40L and Hepcidin serum levels. (A) ELISA showing sCD40L serum levels at different time points (T0, T3, T6, T9) in Group 1 (left panel) and Group 2 (right panel). Each 3-month treatment with PMMA induced a significant decrease of sCD40L (Group 1: *P < .001 T3 vs T0 and T9 vs T6; Group 2: *P < .001 T6 vs T3). By contrast, each 3-month treatment with PS after PMMA was associated with a significant increase of sCD40L (Group 1: *P < .001 T6 vs T3; Group 2: *P < .001 T9 vs T6). No significant difference was observed in Group 2 between T0 and T3 during PS treatment. In both groups a significant difference of sCD40L between T0 and T9 was observed (Group 1 and 2: *P < .001 T9 vs T0): in addition, in both groups, sCD40L at T9 was lower than the cut-off risk of 8.4 ng/mL (bold dotted line). (B) ELISA showing Hepcidin serum levels at different time points (T0, T3, T6, T9) in Group 1 (left panel) and Group 2 (right panel). Each 3-month treatment with PMMA induced a significant decrease of Hepcidin (Group 1: *P < .001 T3 vs T0, ^#P < .01 T9 vs T6; Group 2: *P < .001 T6 vs T3). By contrast, each 3-month treatment with PS after PMMA was associated with a significant increase of Hepcidin (Group 1: ^#P < .01 T6 vs T3; Group 2: ^#P < .01 T9 vs T6). No significant difference was observed in Group 2 between T0 and T3 during PS treatment. In Group 1 but not in Group 2, a significant difference of Hepcidin levels between T0 and T9 was observed (*P < .001 T9 vs T0). Wilcoxon signed ranks test was used for statistical analysis.

Figure 6:

Evaluation of sCD40L mass removal with different membranes. ELISA showing sCD40L serum levels in 10 patients enrolled in Group 2. Data are expressed as average ± 1 SD in two different dialysis sessions of the same week (A, Monday; B, Friday). In comparison with PS, PMMA was associated with a significant higher sCD40L mass removal at all time points considered in both HD sessions (*P < .05 PMMA vs PS at 1 h, 2 h and 4 h).

Figure 7:

In vitro assessment of EC viability, ROS production, Nrf2 mRNA expression, monocyte adhesion and EndMT. All in vitro studies on EC were performed using sera collected at different time points from Group 1 (n = 24). (A) Wild-type, siRNA control and siRNA CD40 EC viability evaluated by MTT assay after 96 h incubation with sera collected at different time points (results are expressed as quantitative fluorescence determination in y-axis). PMMA treatment significantly reduced uremic sera-induced EC cytotoxicity (*P < .05 T3 vs T0 and T9 vs T6). PS treatment after PMMA induced a significant decrease of EC viability (*P < .05 T6 vs T3). A significant increase of EC viability was also observed between T0 and T9 (*P < .05 T9 vs T0). A significant cell viability augmentation was observed in siRNA CD40 EC but not in siRNA control EC challenged with T0 sera (*P < .05 siRNA CD40 EC vs siRNA control). Despite the decrease of cytotoxicity observed by down-regulating the CD40 receptor, siRNA CD40 EC viability was significantly lower than that observed with control (*P < .05 siRNA CD40 EC vs control). (B) ROS production by wild-type, siRNA control and siRNA CD40 EC incubated with sera collected at different time points (results are expressed as quantitative fluorescence determination in y-axis). PMMA treatment significantly reduced ROS production (*P < .05 T3 vs T0 and T9 vs T6). PS treatment after PMMA induced a significant increase of ROS production (*P < .05 T6 vs T3). A significant decrease of ROS production was also observed between T0 and T9 (*P < .05 T9 vs T0). By contrast, a significant ROS reduction was observed in siRNA CD40 EC but not in siRNA control EC challenged with T0 sera (*P < .05 siRNA CD40 EC vs siRNA control). However, ROS generation by siRNA CD40 EC was significantly higher than that observed with control (*P < .05 siRNA CD40 EC vs control). (C) Nrf2 mRNA expression in wild-type, siRNA control and siRNA CD40 EC incubated with sera collected at different time points (results are expressed as fold variation with respect to control). PMMA treatment significantly enhanced Nrf2 expression (*P < .05 T3 vs T0 and T9 vs T6). PS treatment after PMMA induced a significant reduction of Nrf2 expression (*P < .05 T6 vs T3). A significant increase of Nrf2 expression was also observed between T0 and T9 (*P < .05 T9 vs T0). A significant Nrf2 mRNA augmentation was also detected in siRNA CD40 EC vs siRNA control EC (*P < .05 siRNA CD40 EC vs siRNA control EC). (D) Monocyte adhesion to wild-type, siRNA control and siRNA CD40 EC after incubation with sera collected at different time points (results are expressed as number of adherent cells/microscopic field). PMMA treatment significantly reduced monocyte adhesion (*P < .05 T3 vs T0 and T9 vs T6). PS treatment after PMMA significantly increased monocyte adhesion (*P < .05 T6 vs T3). A significant reduction of monocyte adhesion was also observed between T0 and T9 (*P < .05 T9 vs T0). A significant decrease of monocyte adhesion was found in siRNA CD40 EC but not in siRNA control EC challenged with T0 sera (*P < .05 siRNA CD40 EC vs siRNA control). Despite the decrease of inflammatory cell binding observed by down-regulating the CD40 receptor, monocyte adhesion to siRNA CD40 EC was significantly higher than that observed with control (*P < .05 siRNA CD40 EC vs control). (E, F) EndMT characterization by FACS analysis for the endothelial antigen CD31 (E) and for the fibroblast molecule type 1 collagen (F) in wild-type, siRNA control and siRNA CD40 EC incubated with sera collected at different time points (results are expressed as quantitative fluorescence determination in y-axis). PMMA treatment significantly reduced EndMT [*P < .05 T3 vs T0 and T9 vs T6 for CD31 increase in (E) and for type 1 collagen decrease in (F)]. PS treatment after PMMA significantly increased EndMT [*P < .05 T6 vs T3 for CD31 decrease in (E) and for type 1 collagen increase in (F)]. A significant decrease of EndMT was also observed between T0 and T9 [*P < .05 T9 vs T0 for CD31 increase in (E) and for type 1 collagen decrease in (F)]. A significant EndMT reduction was also observed in siRNA CD40 EC but not in siRNA control EC challenged with T0 sera [*P < .05 siRNA CD40 EC vs siRNA control for CD31 increase in (E) and for type 1 collagen decrease in (F)]. EndMT in siRNA CD40 EC was significantly higher than that observed with control [*P < .05 siRNA CD40 EC vs control for CD31 decrease in (E) and for type 1 collagen increase in (F)].

Figure 8:

Evaluation of in vitro calcification and Runx2 mRNA expression of VSMC. All in vitro studies on VSMC were performed using sera collected at different time points from Group 1 (n = 24). (A) Red alizarin staining of wild-type, siRNA control and siRNA CD40 VSMC incubated with sera collected at different time points (results are expressed as number of positive cells/microscopic field). PMMA treatment significantly reduced VSMC calcification (*P < .05 T3 vs T0 and T9 vs T6). PS treatment after PMMA induced a significant increase of red alizarin staining (*P < .05 T6 vs T3). A significant decrease of VSMC calcification was also observed between T0 and T9 (*P < .05 T9 vs T0). By contrast, a significant red alizarin staining reduction was observed in siRNA CD40 VSMC but not in siRNA control VSMC challenged with T0 sera (*P < .05 siRNA CD40 vs siRNA control). Red alizarin staining of siRNA CD40 VSMC was significantly higher than that observed with control (*P < .05 siRNA CD40 VSMC vs control). (B) Runx2 expression in wild-type, siRNA control and siRNA CD40 VSMC incubated with sera collected at different time points (results are expressed as fold variation with respect to control). PMMA treatment significantly reduced Runx2 expression (*P < .05 T3 vs T0 and T9 vs T6). PS treatment after PMMA induced a significant increase of Runx2 expression (*P < .05 T6 vs T3). A significant decrease of Runx2 expression was also observed between T0 and T9 (*P < .05 T9 vs T0). A significant Runx2 mRNA reduction was detected in siRNA CD40 VSMC vs siRNA control VSMC incubated with T0 sera (*P < .05 siRNA CD40 vs siRNA control).