White matter protection in congenital heart surgery

Abstract

Background: Neurodevelopmental delays in motor skills and white matter (WM) injury have been documented in congenital heart disease and after pediatric cardiac surgery. The lack of a suitable animal model has hampered our understanding of the cellular mechanisms underlying WM injury in these patients. Our aim is to identify an optimal surgical strategy for WM protection to reduce neurological injury in congenital heart disease patients.

Methods and results: We developed a porcine cardiopulmonary bypass model that displays area-dependent WM maturation. In this model, WM injury was identified after cardiopulmonary bypass-induced ischemia-reperfusion injury. The degree of injury was inversely correlated with the maturation stage, which indicates maturation-dependent vulnerability of WM. Within different oligodendrocyte developmental stages, we show selective vulnerability of O4+ preoligodendrocytes, whereas oligodendrocyte progenitor cells were resistant to insults. This indicates that immature WM is vulnerable to cardiopulmonary bypass-induced injury but has an intrinsic potential for recovery mediated by endogenous oligodendrocyte progenitor cells. Oligodendrocyte progenitor cell number decreased with age, which suggests that earlier repair allows successful WM development. Oligodendrocyte progenitor cell proliferation was observed within a few days after cardiopulmonary bypass-induced ischemia-reperfusion injury; however, by 4 weeks, arrested oligodendrocyte maturation and delayed myelination were detected. Logistic model confirmed that maintenance of higher oxygenation and reduction of inflammation were effective in minimizing the risk of injury at immature stages of WM development.

Conclusions: Primary repair in neonates and young infants potentially provides successful WM development in congenital heart disease patients. Cardiac surgery during this susceptible period should avoid ischemia-reperfusion injury and minimize inflammation to prevent long-term WM-related neurological impairment.

Figure 1

CPB-induced insult and Porcine OL lineage cells. A, Study design of the CPB groups. BC, TOI during surgery and minimum values. DE, Maximum leukocyte numbers and plasma IL-6 concentration. F, Neurological deficit score on POD1–3. G, Time course of deficit after Severe-CPB insult. POD, post-operative day. H, Subdivision of cerebral WM at the level of precentral sulcus. I, OL antibody markers used to immunostain distinct developmental stages of porcine OL lineage cells. J, Olig2⁺Mash1⁺ OPCs. K,L NG2⁺ OPCs display spindle-like (K) or multipolar (L) process-bearing morphology. M, O4⁺Olig2⁺ pre-OLs. N, Olig2⁺CC1⁺ mature OLs. *P<0.05, **P<0.001, vs. Day 1–3, ***P<0.001 vs. Control by ANOVA with Bonferroni comparisons. Scale bar, 50μm. Data are shown as mean±SEM (B–F; n=10). More data are presented in Supplemental Figure 3.

Figure 2

Porcine WM maturation displays area-dependent progression. A–D, CC width at 1, 3, and 7 weeks increases with age. E–G, MBP expression in CC in three age groups. H, Relative changes in MBP expression in three WM areas. I–L, Olig2⁺Mash1⁺ OPCs at 3 and 12 weeks of age in CC and SCWM. M–P, Olig2⁺CC1⁺ mature OLs at 1 and 7 weeks of age in CC and IC. Q, Mash1⁺Olig2⁺ OPC numbers in three WM maturational areas at 1 and 3 weeks of age. R,S, Changes in Olig2⁺ OL lineage cells and in Olig2⁺CC1⁺ mature OLs in three WM areas. T, Porcine WM maturation based on the time course of OL differentiation and MBP expression (Stage1; CC and M-PVWM, Stage 2; SCWM and L-PVWM, Stage 3; IC). * P<0.05, **P<0.01, ***P<0.001 vs. 1week, ^†P<0.01, ^††P<0.001 vs. Stage 1 by ANOVA with Bonferroni comparisons. Scale bar, 50μm. Data are shown as mean±SEM (n=5). Arrows indicate Olig2⁺Mash1⁺ OPCs. More data are presented in Supplemental Table 1.

Figure 3

Developing porcine WM displays maturation-dependent vulnerability to CPB-induced insults. A,B, Caspase-3⁺ cells in CC in different CPB groups. C, Caspase-3⁺ cell number after CPB-induced insult. D, Relationship between Caspase-3⁺ cell number after Severe-CPB insult and WM maturation. E–H, Co-immunostaining of Caspase-3⁺ with different OL lineage antibody markers, including Mash1, PDGFRα, O4, and CC1. I, Percentage of Caspase-3⁺ cells immunostained with OL antibodies. J–L, Image of M-PVWM stained with Caspase-3, O4, and CC1 on day 3 after Severe-CPB insult, and numbers of different Caspase-3⁺ cell in CC and M-PVWM. M–O, Image of CC stained with TUNEL, O4, and CC1 on day 3 after Severe-CPB insult and numbers of different TUNEL⁺ cell in CC and M-PVWM. P, Percentage of apoptotic O4⁺CC1-negative pre-OLs per total apoptotic cells in CC and M-PVWM after Severe-CPB insult, as determined by Caspase-3 and TUNEL assays. Q,R, Images of NG2⁺ OPCs in M-PVWM after Mild- and Severe-CPB insults. *P<0.05 vs. O4⁺ and CC1⁺, respectively, **P<0.001 vs. Mash1⁺, PDGFR ⁺, and CC1⁺, respectively, ***P<0.001 vs. Control, ^†P<0.05, ^††P<0.001 vs. other three groups by ANOVA with Bonferoni comparisons. Data are shown as mean±SEM (n=5). More data are presented in Supplemental Table 3 and Figure 5.

Figure 4

CPB induces proliferation of OL progenitors in WM on post-operative day 3. AB, Olig2⁺Ki67⁺ cells in M-PVWM after Mild- and Severe-CPB insults. C, Number of PVWM and SCWM Olig2⁺ cells in different CPB groups. D, Number of Olig2⁺Ki67⁺ cells in PVWM and SCWM. EF, PVWM and SCWM NG2⁺Ki67⁺ cells after Severe-CPB insult. G, PVWM and SCWM NG2⁺ cell number in different CPB groups. H, Number of PVWM and SCWM NG2⁺Ki67⁺ cells in different CPB groups. I–K, M-PVWM Olig2⁺Mash1⁺ cells in different CPB groups. L, Percentage of PVWM Olig2⁺Mash1⁺ cells in different CPB groups. *P<0.05, **P<0.01, ***P<0.001 vs. Control by ANOVA with Bonferroni comparisons. Scale bar, 50μm. Data are shown as mean±SEM (n=5). More data are presented in Supplemental Table 3.

Figure 5

Arrested OL maturation and delayed myelination occur at 4 weeks postoperatively in Severe-CPB insult. A–D, MBP expression in CC of different CPB groups, and relative changes within these groups. E–H, CC Olig2⁺CC1⁺ cells in different CPB groups, and density of CC CC1⁺ mature OLs within these groups. I, Olig2⁺CC1-negative immature OL percentage in CC. J,K, Change of Olig2⁺ cell density in PVWM and SCWM (J) and in IC (K) after CPB. L, IC CC1⁺ and Olig2⁺CC1-negative cell density in different CPB groups. *P<0.05, **P<0.01, ***P<0.001 vs. Control by ANOVA with Bonferroni comparisons. Scale bar, 50μm. Data are shown as mean±SEM (n=5). More data are presented in Supplemental Table 3.

Figure 6

Severe-CPB insult results in acute cortical neuronal injury and long-term WM axonal injury. A–C, SMI-312⁺ neurofilaments in cortex layer V/VI on post-operative day 3 in different CPB groups. D–F, NeuN⁺Caspase-3⁺ cells in cortex layer II/III on post-operative day 3 in the CPB groups. G–I, SMI-32⁺ neurofilaments in SCWM at week 4 after surgery in the CPB groups. J,K, NeuN⁺Caspase-3⁺ cell number in cortex layer II/III (G) and V/VI (H). L, Density of SMI-32⁺ neurofilaments in SCWM in the CPB groups. *P<0.01, **P<0.001 vs. Control by ANOVA with Bonferroni comparisons. Data are shown as mean±SEM (n=5). More data are presented in Supplemental Table 3 and Figure 8.

Figure 7

Maintaining high TOI and reducing inflammation minimize the risk of CPB-induced injury at immature stages of WM development. AB, Effects of minimum TOI and maximum plasma IL-6 concentration on total cell death (Caspase-3⁺ cells) at three different stages of WM maturation. CD, Effects of minimum TOI and maximum plasma IL-6 level on cell death in cerebral cortex. EF, Damage probability curves as a function of TOI and IL-6 level at three different stages of WM maturation. These stages are defined as shown in Figure 2T. Stage 1 corresponds to CC and M-PVWM; Stage 2, SCWM and L-PVWM; Stage 3, IC. For example, a bypass corresponding to a TOI of 40% is associated with a probability of damaging nearly 80% for stage 1, 45% for stage 2, and only 15% for stage3 (E). IL-6 level of 400pg/ml is associated with a probability of 68% for stage 1 and only 25% for stage 3 (F). *P<0.05, **P<0.001 TOI more than 45% vs. less than 45%, ***P<0.05 IL-6 less than 250 pg/ml vs. more than 250 pg/ml by ANOVA with Bonferroni comparisons. Data are shown as mean±SEM (n=10).