Increased generation of cyclopentenone prostaglandins after brain ischemia and their role in aggregation of ubiquitinated proteins in neurons

Abstract

The cyclopentenone prostaglandin (CyPG) J₂ series, including prostaglandin J₂ (PGJ₂), Δ¹²-PGJ₂, and 15-deoxy-∆¹²,¹⁴-prostaglandin J₂ (15d-PGJ₂), are active metabolites of PGD₂, exerting multiple effects on neuronal function. However, the physiologic relevance of these effects remains uncertain as brain concentrations of CyPGs have not been precisely determined. In this study, we found that free PGD₂ and the J₂ series CyPGs (PGJ₂, Δ¹²-PGJ₂, and 15d-PGJ₂) were increased in post-ischemic rat brain as detected by UPLC-MS/MS with 15d-PGJ₂ being the most abundant CyPG. These increases were attenuated by pre-treating with the cyclooxygenase (COX) inhibitor piroxicam. Next, effects of chronic exposure to 15d-PGJ₂ were examined by treating primary neurons with 15d-PGJ₂, CAY10410 (a 15d-PGJ₂ analog lacking the cyclopentenone ring structure), or vehicle for 24 to 96 h. Because we found that the concentration of free 15d-PGJ₂ decreased rapidly in cell culture medium, freshly prepared medium containing 15d-PGJ₂, CAY10410, or vehicle was changed twice daily to maintain steady extracellular concentrations. Incubation with 2.5 μM 15d-PGJ₂, but not CAY10410, increased the neuronal cell death without the induction of caspase-3 or PARP cleavage, consistent with a primarily necrotic mechanism for 15d-PGJ₂-induced cell death which was further supported by TUNEL assay results. Ubiquitinated protein accumulation and aggregation was observed after 96 h 15d-PGJ₂ incubation, accompanied by compromised 20S proteasome activity. Unlike another proteasome inhibitor, MG132, 15d-PGJ₂ treatment did not activate autophagy or induce aggresome formation. Therefore, the cumulative cytotoxic effects of increased generation of CyPGs after stroke may contribute to delayed post-ischemic neuronal injury.

Fig. 1

Brain concentrations of PGD₂ and its metabolite cyclopentenone prostaglandins are increased after middle cerebral artery occlusion (MCAO) in rat in a COX activity-related manner. Rats were treated via oral gavage with vehicle (2% methyl cellulose) or COX inhibitor piroxicam (30 mg/kg) 1 hour prior to MCAO (n=6 per group; naïve, n=4). Rats were sacrificed 24 h after reperfusion and ipsilateral brains dissected into penumbra and core samples. Free PGD₂ and its metabolites in each infarct area were detected and quantified by UPLC-MS/MS. Pir: piroxicam; Veh: vehicle; NS: not significant. * P < .05; &: not detected. Data are means +/− SD

Fig. 2

The cyclopentenone prostaglandin 15d-PGJ₂ induces cell death through a cyclopentenone ring moiety-dependent mechanism. Primary neuronal cells were treated with vehicle (Veh), CAY10410 (CAY, 2.5 μM) or 15d-PGJ₂ for 96 h. Cell viability was measured using the MTT assay. n = 54–72 wells per group (4 plates run on 2 experimental days). * P < 0.001. Data are means +/− SE

Fig. 3

15d-PGJ₂ induces primary neuronal cell death via a caspase-3 independent mechanism. Cells were treated with vehicle (Veh, methyl acetate), CAY10410 (CAY), 15d-PGJ₂ for 24 h – 96 h. Pro-apoptosis reagents, MG132 and staurosporin, were incubated with primary neurons at 10 μM for 16 h as positive controls. (a). Caspase-3 and PARP cleavage: Cells were harvested and immunoblotted using anti-cleaved caspase-3 and anti-PARP antibodies. β-actin was used as a loading control. Full: full length protein; Clvd: cleaved. (b). Representative immunocytochemical photos of rat primary neurons after treatment with Veh or 2.5 μM 15d-PGJ₂. Upper panel: Cleaved caspase-3 was detected using anti-cleaved caspase-3 antibody (green) and anti-NeuN (red) antibodies (48 h after treatment). Middle panel: Activated caspase 3/7 was detected using a fluorogenic substrate for activated caspases-3 and 7 (green, 72 h after treatment). Lower panel: Double-stranded DNA breaks were detected using TUNEL (green, 72 h after treatment). All photos: 60X using a confocal microscope; scale bar = 40 μm. Blue is DAPI nuclear stain

Fig. 4

Ubiquitinated proteins accumulate and aggregate in primary neurons after chronic treatment with 15d-PGJ₂. A and B: Cells were incubated with Vehicle (Veh, methyl acetate), CAY10410 (CAY), or 15d-PGJ₂ for 96 h (n = 3 per group) then harvested with RIPA buffer. Both RIPA -soluble and -insoluble fractions were collected. Poly-ubiquitinated proteins were detected by Western blot using an anti- poly-ubiquitinated conjugates antibody. Poly-ubiquitinated proteins were quantified and normalized to vehicle-treated groups. (a) RIPA-soluble fraction. (b): RIPA-insoluble fraction. Data are means +/− SE. * P < 0.05. (c) Detection of ubiquitinated protein by immunocytochemical staining using anti-ubiquitinated proteins (red) and anti-MAP-2 (green) antibodies. Primary neurons were treated with vehicle (Veh), 2.5 μM CAY10410, or 2.5 μM 15d-PGJ₂ for 72 h. MG132 (10μM, 16 h incubation), a protease inhibitor, was used as a positive control. Blue is DAPI nuclear stain. Photos are at 240X. (d) Primary neurons were treated with vehicle (Veh) or 2.5 μM 15d-PGJ₂ for 24 h or 72 h prior to fixation and immunocytochemical staining with anti-ubiquitinated proteins antibody (green). Photos are at 60X and 240X. All photos: scale bar = 25μm

Fig. 5

Long-term treatment of primary neuronal culture with 15d-PGJ₂ impairs proteasome activity without autophagy activation. (a) Proteasome activity 72 h after treatment with vehicle (Veh), CAY10410 (CAY) or 15d-PGJ₂ normalized to the vehicle-treated group. Data are means +/− SE. n = 8 per group. * P < 0.01 vs Veh. (b) Cells were treated with vehicle or 15d-PGJ₂ for 72 h or 96 h, or 10 μM MG132 (MG, positive control, 16 h treatment). Lysates were then immunoblotted with anti-LC3B antibody. β-actin was used as a loading control. (c) Primary neurons were treated with Veh, 2.5 μM 15d-PGJ₂ for 96 h or MG for 16 h then fixed with formalin and immunostained using anti-LC3B antibody (green). Aggresomes were detected using a ProteoStat aggresome detection kit (red). Blue is DAPI nuclear stain. Photos are 180X taken with an Olympus confocal microscope; scale bar = 20 μm

Fig. 6

Continuous incubation with 15d-PGJ₂ modifies endogenous neuronal proteins and increases protein aggregates. (a–b) Rat primary neurons were treated with biotinylated (b-) 15d-PGJ₂ or vehicle (Veh). (a) b-15d-PGJ₂ adducts to endogenous proteins in primary neurons after treatment for 24 h – 72 h. Cell lysates underwent SDS-PAGE and biotin-incorporated proteins (b-proteins) were detected using HRP-conjugated Streptavidin (upper). β-actin was used as a loading control (lower). (b) Immunocytochemical detection (60X, inset 240X) of ubiquitin and b-15d-PGJ₂ using anti-ubiquitin conjugates (Ub-conjugates, green) and anti-biotin (red) antibodies. Blue is DAPI nuclear stain. Scale bar = 60 μm; inset = 20 μm. (c) UCH-L1 aggregates in primary neurons after 15d-PGJ₂ treatment. Primary neurons were harvested with RIPA buffer 96 h after treatment with 15d-PGJ₂ or vehicle. RIPA -soluble (upper) and -insoluble (lower) fractions were collected, and UCH-L1 was detected by Western blot with anti-UCHL-1 antibody and quantified. Graphs are means +/− SE normalized to vehicle. n=3 per group. * P<0.05 vs vehicle. (d) Immunocytochemical staining (180X) of rat primary neurons 96 h after incubation with vehicle, 2.5 μM CAY 10410 or 2.5 μM 15d-PGJ₂. UCHL-1 and ubiquitinated proteins were visualized with anti-UCHL-1 (red) and anti-ubiquitinated conjugates (Ub-conjugates, green) antibodies, respectively. Blue is DAPI nuclear stain. Scale bar = 20 μm. Photos (b,d) were taken with an Olympus confocal microscope