Acquired immunological imbalance after surgery with cardiopulmonary bypass due to epigenetic over-activation of PU.1/M-CSF

Abstract

Background: It has been shown that severe insult to the immune system may trigger prolonged macrophage characteristics associated with excessive release of monocyte colony stimulating factor (M-CSF). However, it is unclear how persistent is the macrophage-like characteristics in circulating monocytes (MO). In this study, 20 patients who underwent non-emergent cardiopulmonary bypass had their monocytes characterized before surgery and 3 months after surgery.

Methods: We assessed the macrophage characteristics of MO using cytokine production, surface marker expression, an ability to stimulate T cells, and methylation of the promoter region of the gene encoding PU.1, a critical component to M-CSF production. MO function as well as activation and differentiation potential were longitudinally assessed.

Results: At 3 months after cardiopulmonary bypass, monocytes exhibited increased expression of MRP8, transforming growth factor-β/latency-associated peptide, suppressor of cytokine signaling 3 while phagocytic properties were increased. Concomitantly, we observed a decreased expression of CD86, a decreased ability to form regulatory dendritic cells, and a diminished ability to stimulate T cells. These characteristics were accompanied by a persistent increase in the secretion of M-CSF, over-activation of PU.1, and decreased methylation of the PU.1 promoter region. Serum levels of C-reactive protein and anti-cytomegalovirus IgG antibody titers were also elevated in some patients at 3 months after surgery.

Conclusions: We concluded that at 3 months after cardiopulmonary bypass, monocytes continued to express a new macrophage-like milieu that was associated with the persistent activation of the PU.1/M-CSF pathway.

Keywords: Cardiopulmonary bypass; Epigenetics; Granulocyte colony stimulating factor; Monocytes; PU.1.

Fig. 1

Change in the phenotype and kinase makeup of the post-CPB. Fresh MO were stained with flow cytometry antibodies. There was a significant difference in post hoc analysis baseline in frequency  %CD14^(high)CD16^(high) MO (Me = 13.54; IQR[6.15, 24.39] ) vs t + 24 h and vs t + 7d, t_24h: Me = 22.63; IQR[13.32, 45.06] ΔRank = − 15.00; CI_95% = [− 27.81, − 2.19]; p = 0.034; CLES = 0.86; t_7d: Me = 20.58; IQR[18.11, 33.09]; ΔRank = − 15.00; CI_95% = [− 27.81, − 2.19]; p = 0.034; CLES = 0.71 (a). There was no significant difference in %CD14^(high)CD16^(low) (b). A significant increase in predominantly M2-promoting kinases were seen in peripheral blood MO population at 3 months. The population SOCS3^(high) cells was significantly increase at t_3m: Me = 88.46; IQR[30.56, 97.60] as compared to pre-CPB level: Me = 24.77; IQR[24.17, 25.83] ΔpMe = − 39,41; CI_95%: ΔpMe[− 72.47, − 3.16]; p = 0.047; CLES = 0.78 (c). This shift is very clear if data from the same patient are shown (d). Asterisk indicate data points which were significantly different when compared vs baseline. CI_95%: 95% confidence interval for difference in sum of ranks. CLES common language effect size

Fig. 2

Changes in monocyte function after surgery. Functional properties of the circulating MO were tested using, phagocytosis assays (a), ability to stimulate allogeneic T cells (b), functional plasticity by incubate cells in vitro with IL-4 and GM-CSF and measuring an emergence of DC specific marker (c) or ability to induce proliferative T cells response. a There was a significant increase (t₀: Mean = 10.7 ± 0.05 vs t_3m: Mean = 1.64 ± 0.49 CI_95%: Δx = [0.23, 0.90]; p = 0.003; d = 1.36) in the phagocytic capabilities of peripheral blood monocytes taken from patients at 3 months after the surgery, as shown by uptake of zymosan. b The ability of peripheral blood monocytes to stimulate allogeneic T cells was significantly (t₀: Mean = 0.65 ± 0.13 vs t_3m: Mean = 0.46 ± 0.31; CI_95% Δx [− 0.33, − 0.04]; p = 0.016; d = − 0.68) diminished at 3 months. c There was a significant decline in the emergence of CD1a⁽⁺⁾ (immature DC marker) (t₀: Me = 1964.58; IQR[745.82, 7586.04] vs t_3m: Me = 519.19; IQR[428.24, 785.52]; CI_95%: ΔpMe[761.49, 5459.51]; p = < 0.001; CLES = 1.0) and CD83 (mature DC marker) (t₀: Me = 64.72; IQR[43.22, 82.55] vs t_3m: Me = 445.24; IQR[240.07, 747.04]; CI_95%: ΔpMe [101.62, 964.67]; p = 0.047; CLES = 0.75). d Stimulation with IL-4 and GM-CSF did not recover MO ability to stimulate T cells to pre-CPB levels (t₀: Me = 39.50; IQR[30.75, 48.50] vs t_3m: Me = 26.40; IQR[18.25, 34.00]; ΔRank = 21.50; CI_95% = [6.83, 36.17]; p = 0.015; CLES = 0.83). Asterisk indicate data points which were significantly different when compared vs baseline. CI_95%: Δx-95% confidence interval for difference in means. CI_95%: 95% confidence interval for difference in sum of ranks. CI_95%: ΔpMe-95% confidence interval for difference in pseudomedians

Fig. 3

Changes in post-CPB M-CSF production. Production of M-CSF in response to LPS (a) and PMA and ION (b) was assessed in vitro together with flow cytometric expression of M-CSFR and TLR on circulating MO (c). a There was a significant difference in all data points in M-CSF production with LPS stimulation t₀: Me = 121.00; IQR[48.50, 189.00] t_24h: Me = 4567.00; IQR[1298.50, 5755.00]; ΔRank = − 31.00; CI_95% = [− 44.28, − 17.72]; p = 0.003; CLES = 0.89; t_7d: Me = 2053.00; IQR[375.50, 7036.50]; ΔRank = − 30.00; CI_95% = [− 43.28, − 16.72]; p < 0.001; CLES = 0.89; t_3m: Me = 1346.00; IQR[192.00,5627.00] ΔRank = − 21.00; CI_95% = [− 34.28, − 7.72]; p = 0.003; CLES = 0.79. b Production of M-CSF after PMA and ION stimulation was increased at all data point when compared with baseline t₀: Me = 93.00; IQR[38.00, 192.50]; t_24h: Me = 3321.00; IQR[530.00, 6532.00]; ΔRank = − 22.00; CI_95% = [− 34.73, − 9.27]; p = 0.002; CLES = 0.94; t_7d: Me = 2341.00; IQR[265.00, 8356.50]; ΔRank = − 21.00; CI_95% = [− 33.73, − 8.27]; p = 0.002; CLES = 0.69; t_3m: Me = 341.00; IQR[106.00, 4249.00]; ΔRank = − 23.00; CI_95% = [− 35.73, − 10.27]; p = 0.002; CLES = 0.88. c Receptor expression of TLR4 was unchanged in terms of both positive cells and receptor density (MFI). Asterisk indicate data points which were significantly different when compared vs baseline. CI_95%: 95% confidence interval for difference in sum of ranks. CI_95%: ΔpMe-95% confidence interval for difference in pseudomedians. CLES common language effect size

Fig. 4

Expression of M-CSF receptor significantly increases and persists after CPB. a The expression of M-CSF receptors in terms of positive cells was not significantly different in a group analysis on the surface of circulating MO. b The CD115/M-CSFR receptor density was significantly increased at all times vs baseline after the surgery t₀: Me = 2314.64; IQR[1032.07, 4978.19] t_24h: Me = 10,425.65; IQR[8041.68, 16,667.99]; ΔRank = − 22.50; CI_95% = [− 35.59, − 9.41]; p = 0.002; CLES = 0.82; t_7d: Me = 12,824.02; IQR[4165.53, 16,443.61]; ΔRank = − 18.5; CI_95% = [− 31.59, − 5.41]; p = 0.007; CLES = 0.82; t_3m: Me = 9858.70; IQR[7541.30, 16,547.60] ΔRank = − 25.00; CI_95% = [− 38.09, − 11.91]; p = 0.001; CLES = 0.82. Asterisk indicate data points which were significantly different when compared vs baseline. CI_95%: 95% confidence interval for difference in sum of ranks. CLES: common language effect size

Fig. 5

Increase in PU.1 activity coupled with epigenetic regulation of its gene promoter. Intracellular expression of PU.1 and methylation of SPI were measured in the circulating MO. a There was a significant increase in the percentage of expression of PU.1 t₀: Me = 70.76; IQR[53.16, 79.72] t_24h: Me = 90.14; IQR[83.88, 92.94]; ΔRank = − 4.00; CI_95% = [− 7.15, − 0.85]; p = 0.018; CLES = 0.83; t_3m: Me = 94.99; IQR[94.47, 95.68]; ΔRank = − 11.00; CI_95% = [− 14.15, − 7.85]; p < 0.001; CLES = 1.00. b There was also a significant increase in antigen density (MFI) at all data points t₀: Me = 545.30; IQR[500.20, 638.62]; t_24h: Me = 3839.20; IQR[3729.39, 3944.17]; ΔRank = − 7.00; CI_95% = [− 10.15, − 3.85]; p = 0.001; CLES = 1.00; t_3m: Me = 4129.11; IQR[3993.60, 4222.01]; ΔRank = − 11.00; CI_95% = [− 14.15, − 7.85]; p < 0.001; CLES = 1.00. c Again, this pattern was demonstrated on the MO obtained from the same patient (grayed graph is isotype control and auto fluorescence, green is baseline while red represent patient MO PU.1 reactivity at 3 months after CPB). d There was a significant decrease in methylation of SP11 gene at 3 months t₀: Me = 49.65; IQR[39.50, 70.71]; t_3m: Me = 4.61; IQR[4.27, 5.56]; ΔRank = 12.00; CI_95% = [5.71, 18.29]; p = 0.004; CLES = 1.00. Asterisk indicate data points which were significantly different when compared vs baseline. CI_95%: 95% confidence interval for difference in sum of ranks. CI_95%: ΔpMe-95% confidence interval for difference in pseudomedians. CLES: common language effect size