A novel PGAM5 inhibitor LFHP-1c protects blood-brain barrier integrity in ischemic stroke

Abstract

Blood-brain barrier (BBB) damage after ischemia significantly influences stroke outcome. Compound LFHP-1c was previously discovered with neuroprotective role in stroke model, but its mechanism of action on protection of BBB disruption after stroke remains unknown. Here, we show that LFHP-1c, as a direct PGAM5 inhibitor, prevented BBB disruption after transient middle cerebral artery occlusion (tMCAO) in rats. Mechanistically, LFHP-1c binding with endothelial PGAM5 not only inhibited the PGAM5 phosphatase activity, but also reduced the interaction of PGAM5 with NRF2, which facilitated nuclear translocation of NRF2 to prevent BBB disruption from ischemia. Furthermore, LFHP-1c administration by targeting PGAM5 shows a trend toward reduced infarct volume, brain edema and neurological deficits in nonhuman primate Macaca fascicularis model with tMCAO. Thus, our study identifies compound LFHP-1c as a firstly direct PGAM5 inhibitor showing amelioration of ischemia-induced BBB disruption in vitro and in vivo, and provides a potentially therapeutics for brain ischemic stroke.

Keywords: Blood–brain barrier; Brain microvascular endothelial cells; Ischemic stroke; LFHP-1c; NRF2; PGAM5; Surface plasmon resonance; Target identification.

Graphical abstract

Figure 1

LFHP-1c attenuates brain edema, protects BBB integrity and attenuates brain microvascular endothelial inflammation in rats at 72 h after ischemia onset. (A) When administered at four and 24 h after tMCAO, LFHP-1c (5 mg/kg) attenuated tMCAO-induced brain edema at 72 h post-ischemia, n = 8 per group. (B) and (C) When administered at 4 h after tMCAO, LFHP-1c (5 mg/kg) attenuated tMCAO-induced Evans blue leakage in the brain of rats, n = 5–7 per group. (D)–(F) Post-ischemic treatment with LFHP-1c (5 mg/kg) reduced the content and area of endogenous IgG extravasation to brain tissue, n = 3–4 per group. The scale bar represents 1 mm. (G) and (H) LFHP-1c treatment prevented tMCAO-induced degradation of tight junction proteins ZO-1, occludin and claudin-5 in isolated rat brain microvessels at 72 h post-ischemia, n = 4 per group. (I) Post-ischemic treatment with LFHP-1c attenuated mRNA expression of Vcam-1, Icam-1, Mcp-1, Tnf-α, Mmp 2 and Mmp 9 in isolated rat brain microvessels at 72 h post-ischemia, n = 4 per group. (J) LFHP-1c alleviated the protein expression of VCAM-1 in isolated rat brain microvessels at 72 h post-ischemia, n = 4 per group. Results are expressed as mean ± SEM. ^∗∗P < 0.01, ^∗∗∗P < 0.001 versus Sham group; ^#P < 0.05, ^##P < 0.01, ^###P < 0.001 versus Vehicle group.

Figure 2

LFHP-1c prevents brain microvascular endothelial cells injury under OGD/R condition. (A) Schematic illustration of in vitro BBB model. The in vitro BBB model was established with co-culture of primary rat brain microvascular endothelial cells and primary rat astrocytes. After exposure to OGD/R injury, the in vitro BBB disruption was assessed by FITC-dextran permeability and TEER. (B) The in vitro BBB model was pretreated with LFHP-1c (1, two or 5 μmol/L) for 9 h, followed by exposed to 4 h OGD and 3 h reoxygenation, and TEER was measured at the time points of OGD onset, reoxygenation onset and the end of reoxygenation, n = 3 per group. (C) FITC-dextran (40 kDa) permeability across the in vitro BBB model was measured at 1, 2, three and 6 h after the end of reoxygenation period, n = 3 per group. (D) FITC-dextran (40 kDa) permeability across the rBMECs monolayer was measured at 1 h after the end of reoxygenation, n = 3 per group. (E) rBMECs were treated with LFHP-1c (1, two or 5 μmol/L) for 9 h, then exposed to OGD injury for 4 h, and the content of LDH release to culture medium was measured, n = 4 per group. (F) and (G) TUNEL staining shows that pretreated with LFHP-1c reduced apoptosis in rBMECs after OGD/R injury, n = 3 per group. The scale bar represents 50 μm. (H) and (I) LFHP-1c prevented the degradation of tight junction proteins ZO-1, occludin and claudin-5 induced by OGD/R, n = 3 per group. (J) and (K) Fluorescence staining of ZO-1 shows that LFHP-1c improved ZO-1 expression after OGD/R injury, n = 3 per group. The scale bar represents 50 μm. (L) and (M) LFHP-1c reduced the expression of Vcam-1 mRNA (L) and VCAM-1 protein (M) after exposure to OGD/R, n = 3 per group. Results are expressed as mean ± SEM. For Fig. 3B–D and Fig. 3F to M: ^∗∗P < 0.01, ^∗∗∗P < 0.001 versus Control (Ctrl) group; ^#P < 0.05, ^##P < 0.01, ^###P < 0.001 versus OGD/R group. For Fig. 3E: ^∗∗P < 0.01 versus Control (Ctrl) group; ^##P < 0.01, ^###P < 0.001 versus OGD group.

Figure 3

Target verification of LFHP-1c in endothelial cells. (A) Schematic illustration of target protein capture of LFHP-1c in endothelial cells based on SPR. LFHP-1c was printed on the 3D photo-crosslinking chip via the chip array printer through C–H covalent bond connection, and then the lysates of endothelial cells flowed through the surface of the chip. Finally, the proteins were dissociated from the chip and conducted LC–MS/MS analysis and compared with UniProt database. (B) LFHP-1c binds with ΔN21-PGAM5 in kinetic level determined by SPR, and the K_D value was about 0.961 μmol/L. (C) LFHP-1c concentration-dependently inhibited the dephosphorylation activity of ΔN21-PGAM5 at molecular level, n = 6 per group. (D) LFHP-1c inhibited the dephosphorylation activity of PGAM5 in isolated mitochondria from mouse brain-derived Endothelial cells.3 (bEnd.3) in a concentration-dependent manner, n = 3 per group. (E) Schematic illustration of target identification in rBMECs lysates. Photoaffinity probe HP-62 binds with proteome in rBMECs through photoaffinity labeling, and then clicks with biotin-PEG3-N3 based on copper-catalyzed azide–alkyne cycloaddition (CuAAC), subsequently enriched by Streptavidin Mag Sepharose™ beads, and finally separated by SDS-PAGE followed by immunoblotting. (F) Evaluation of HP-62 effect on the dephosphorylation activity of PGAM5 compared to the parent compound LFHP-1c at molecular level, and the results reveal that HP-62 retained the dephosphorylation activity of PGAM5, n = 6 per group. (G) Pull-down/Western blotting for target validation of PGAM5 with the photoaffinity probe HP-62, and a representative blot shown here. Results are expressed as mean ± SEM. ^∗P < 0.05, ^∗∗∗P < 0.001 versus Control (Ctrl) group.

Figure 4

LFHP-1c targets PGAM5 to facilitate nuclear translocation of NRF2 for endothelial protection in stroke. (A) The expression of PGAM5, p-DRP1 (Ser 637), NRF2 and HO-1 were detected at indicated times after LFHP-1c treatment in rBMECs by immunoblotting, n = 3 per group. (B) and (C) IP/Western blotting results reveal that LFHP-1c treatment significantly reduced the interaction of PGAM5 with NRF2, n = 3 per group. (D) LFHP-1c facilitates nuclear translocation of NRF2 in rBMECs, n = 3 per group. (E) Silencing efficiency of siRNA against PGAM5 in rBMECs was verified by immunoblotting, n = 3 per group. (F) and (G) LFHP-1c treatment or knockdown of Pgam5 facilitates nuclear translocation of NRF2 in rBMECs, and LFHP-1c treatment after si-PGAM5 did not affect the nuclear translocation of NRF2, n = 3 per group. (H) Fluorescence staining of NRF2 shows that LFHP-1c treatment or knockdown of Pgam5 promoted NRF2 nuclear translocation, n = 3 per group. The scale bar represents 20 μm. (I) and (J) The protein expressions of PGAM5, p-DRP1 (Ser637), NRF2 and HO-1 in rat brain microvessels were detected at 72 h after tMCAO onsetand representative blots are shown here, n = 4 per group. (K) The mRNA expression of Ho-1 and Nqo1 was measured by RT-PCR, n = 4–5 per group. Results are expressed as mean ± SEM. For Fig. 4C, ^∗∗∗P < 0.001 versus IgG plus DMSO group, ^##P < 0.01 versus PGAM5 plus LFHP-1c (2 μmol/L) group. For Fig. 4D, ^∗P < 0.05, ^∗∗∗P < 0.001 versus ‘0’ time point group. For Fig. 4E, ^∗∗∗P < 0.001 versus scramble group. For Fig. 4G, ^∗∗∗P < 0.001 versus scramble plus DMSO group. For Fig. 4J and K, ^∗P < 0.05, ^∗∗P < 0.01, ^∗∗∗P < 0.001 versus Sham group; ^#P < 0.05, ^##P < 0.01, ^###P < 0.001 versus Vehicle group.

Figure 5

LFHP-1c ameliorates brain ischemic injury in tMCAO model of Macaca fascicularis through PGAM5–NRF2 pathway. (A) Experimental design of tMCAO to study the effects of LFHP-1c against ischemic stroke in Macaca fascicularis. (B)–(G) The brain infarct volume (B, D, and E) and edema volume (C, F, and G) were measured at three and 7 days after tMCAO onset by MRI, n = 2–7 per group. (H) and (I) Neurological deficit induced by tMCAO-injury was evaluated at 1, 3, 5, 7 and 14 days post-ischemia. (J) and (K) The protein expressions of VCAM-1, PGAM5, p-DRP1 (Ser 637), NRF2 and HO-1 in cerebral cortex of the brain of Macaca fascicularis were detected at 14 days after tMCAO onset, n = 2–4 per group. (L) The mRNA expression of VCAM-1, ICAM-1 and occludin was measured by RT-PCR, n = 2–4 per group. Results are expressed as mean ± SEM. ^∗P < 0.05, ^∗∗P < 0.01 versus Sham group; ^#P < 0.05, ^##P < 0.01, ^###P < 0.001 versus Vehicle group.

Figure 6

Schematic diagram of LFHP-1c protection on BBB disruption and ischemic brain injury. LFHP-1c is here reported as a novel PGAM5 inhibitor with beneficial effects on protection of brain microvascular endothelial cells (BMECs) and blood–brain barrier (BBB) integrity after stroke. By direct binding to PGAM5, LFHP-1c not only inhibits the PGAM5 enzyme activity, but also impairs the interaction between PGAM5 and NRF2, subsequently facilitates nuclear translocation of NRF2, thereby promotes the transcription activity of NRF2 to prevent endothelial injury, and finally protects against BBB disruption and ischemic brain injury.