Brain-targeted hypoxia-inducible factor stabilization reduces neonatal hypoxic-ischemic brain injury

Abstract

Hypoxia-inducible factor-1α (HIF1α) is a major regulator of cellular adaptation to hypoxia and oxidative stress, and recent advances of prolyl-4-hydroxylase (P4H) inhibitors have produced powerful tools to stabilize HIF1α for clinical applications. However, whether HIF1α provokes or resists neonatal hypoxic-ischemic (HI) brain injury has not been established in previous studies. We hypothesize that systemic and brain-targeted HIF1α stabilization may have divergent effects. To test this notion, herein we compared the effects of GSK360A, a potent P4H inhibitor, in in-vitro oxygen-glucose deprivation (OGD) and in in-vivo neonatal HI via intracerebroventricular (ICV), intraperitoneal (IP), and intranasal (IN) drug-application routes. We found that GSK360A increased the erythropoietin (EPO), heme oxygenase-1 (HO1) and glucose transporter 1 (Glut1) transcripts, all HIF1α target-genes, and promoted the survival of neurons and oligodendrocytes after OGD. Neonatal HI insult stabilized HIF1α in the ipsilateral hemisphere for up to 24 h, and either ICV or IN delivery of GSK360A after HI increased the HIF1α target-gene transcripts and decreased brain damage. In contrast, IP-injection of GSK360A failed to reduce HI brain damage, but elevated the risk of mortality at high doses, which may relate to an increase of the kidney and plasma EPO, leukocytosis, and abundant vascular endothelial growth factor (VEGF) mRNAs in the brain. These results suggest that brain-targeted HIF1α-stabilization is a potential treatment of neonatal HI brain injury, while systemic P4H-inhibition may provoke unwanted adverse effects.

Keywords: Birth asphyxia; Erythropoietin; Hypoxic-ischemic encephalopathy; Intranasal; Prolyl-hydroxylase.

Fig. 1.

Expression of $\text{HIF}1\text{α}$ in neonatal brains. (A) Schematic outline of the regulation of $\text{HIF}1\text{α}$ stability under hypoxia (light gray area) versus degradation in normoxia. See the text for details. Selective $\text{HIF}1\text{α}$ target-genes and the chemicals that modulate $\text{HIF}1\text{α}$ stability are labeled. (B) Immunoblotting showed a rapid decline of $\text{HIF}1\text{α}$ from postnatal day 2 to P10 (24% of the P2 level) to P90 (18% of the P2 level) in mouse brains, as opposed to a stable level of $\text{HIF}1\text{β}$/ARNT over the period (n=4 for each time-point; p-value were determined by one-way ANOVA). (C) Immunostaining showed preferential anti-$\text{HIF}1\text{α}$ reactivity in the cerebral cortex (Ctx), but not in the white matter (WM) of P2 mouse brains. Double-labeling confirmed the presence of $\text{HIF}1\text{α}$ in the nucleus of NeuN⁺ neurons and absence in Olig2⁺ OPC/OL. Shown are typical results in three animals. (D) Iron staining by the Perls’s blue method preferentially labeled the axons in WM in P2 mice (n=3 examined). St: striatum. N: neuron. OPC/OL: oligo-progenitor cell/oligodendrocyte. Scale bar: $50\text{μm}$.

Fig. 2.

Post-HI induction of $\text{HIF}1\text{α}$ in neonatal brains. (A) Photoacoustic imaging with co-registered anatomical image (B-mode ultrasound) and the cortical oxygen saturation (sO₂) map of P7 rats after ligation of the right common carotid artery (RCCAo) from under 100% oxygen (left) to 20.9% oxygen (middle) and to 10% oxygen (right). The sO₂ is indicated in a red-white-blue scale from high to low saturation. Labeled are the mean sO₂ value and SD for the RCCAo- and contralateral/left (L)-hemisphere from three animals. (B) The ipsilateral post-HI cortex showed a poor recovery of sO₂, even under 100% oxygen, at both 4 h (87%) and 24 h (83%) recovery in P7 rats (n>5, p-value determined by one-way ANOVA). (C, D) Histological analysis of post-HI brains showed more $\text{HIF}1\text{α}$ positive cells in white-mater (C; MBP staining) and most of neurons in cortex (D; NeuN staining, n=4 at 6 h after HI). Scale bar: $50\text{μm}$. (E) Similarly, P10 mice showed an increase of $\text{HIF}1\text{α}$ protein at 4 h (4.37±0.97, p<0.05) and 24 h (7.15±1.46, p<0.001) after HI compared with unchallenged (UN) mice. The $\text{HIF}1\text{β}$ protein remained stable at both 4 h and 24 h after HI insult. N=3 for each group. p-value determined by one-way ANOVA. NS: not significant.

Fig. 3.

Protective effects of GSK360A in in-vitro oxygen-glucose deprivation (OGD). (A, B) OGD (1% O₂ for 2 h) was performed in MAP⁺ cortical neurons at 10 day-in-vitro (DIV) and 3 or $30\text{μM}$ GSK360A or the vehicle (V) were added after hypoxia to compare the survival and expression of key $\text{HIF}1\text{α}$ target-gene mRNAs at 24 h. GSK360A provided dose-dependent increase of neuronal survival and the $\text{HIF}1\text{α}$ target-gene expression. (C, D) Similarly, GSK360A conferred dose-dependent increase of survival and $\text{HIF}1\text{α}$ target-gene mRNAs in RIP⁺ OLs at 5 DIV. Data were drawn from four biological replicates from >3 independent cultures in each group for viability assay, and three biological replicates from >3 independent cultures in each group for mRNA detections. *p<0.05, **p<0.01, and ***p<0.001 compared to OGD+V-treated cells by one-way ANOVA. Scale bar: $25\text{μm}$.

Fig. 4.

Protective effects of GSK360A via ICV injection against neonatal HI brain injury. (A) RT-PCR analysis showed a dose-dependent increase of $\text{HIF}1\text{α}$ target-genes Ho1, Epo, and Glut1 mRNA, but not Vegf mRNAs, and an inverse reduction of the neuroinflammatory markers (Tspo, Opn, and $\mathit{\text{Tnf}}\text{α}$ mRNAs) at 24 h post-HI in the ipsilateral hemisphere of P7 rats receiving ICV injection of GSK360A (1.2 mg/kg) *p<0.05, **p<0.01 and ***p<0.001 compared with the HI/Vehicle group by one-way ANOVA (n>6 for each). (B, C) HI-injured rats showed TUNEL-positive cell deaths in the ipsilateral cerebral in cortex, striatum and hippocampus at 24 h recovery. Scale bar: $50\text{μm}$. (B) ICV-injection of GSK360A reduced the numbers of TUNEL-positive cell death (left) to 40.8% ± 15.46% (n=19 in 5 animals) compared with vehicle-treated rats (p<0.001 by unpaired t test, n=15 in 4 animals). (C) Among the apoptotic cells, over 50% reduction of TUNEL-NeuN double-positive neurons and TUNEL-O4 double-positive OLs (p<0.001, n>28 in 5 animals) compared with the vehicle-treated group (p<0.001, n>19 in 4 animals). (D) Representative photographs of brain tissue loss (indicated by arrows) in P7 rats after HI insult and ICV-injection of vehicle or GSK360A. ICV injection of GSK360A reduced brain tissue loss from 48.25± 13.73% in the vehicle-treated to 37.12% ± 4.54% in GSK360A-treated group at 7 d post-HI (p=0.026, n>10 for each). p-values were determined by unpaired t-test

Fig. 5.

Comparison of IN-versus-IP application of GSK360A against neonatal HI brain injury. (A) IN-application of GSK360A (6 mg/kg) induced the ODD-Luc activity (7.3±3.2-fold, p=0.02) and $\text{HIF}1\text{α}$ protein levels (3.8±0.37-fold, p<0.05, n=4) at 8 h in P15 neonatal brains (n=4). (B, C) Comparison of the 7-d brain tissue loss (B) and mortality (C) in P7 rats following Rice-Vannucci HI and IP-injection of GSK360A (10, 20, 30, 50 mg/kg or vehicle) or IN-delivery of GSK360A (3 mg/kg or vehicle). IP-injection of GSK360A not only failed to decrease brain tissue loss, but elevated the mortality rate at a high dose (n=15 for each dose of IP-treatment). In contrast, IN-delivery of GSK360A reduced brain atrophy without mortality (n=14 each for V and GSK360A group). p=0.03 by unpaired t test. (D, E) RT-PCR analysis of Ho1,Vegf, and Glut1 mRNAs in the brain (D) and Epo mRNA in brain-versus-kidney (E) at 24 h after IP injection of 30 mg/kg GSK360A or IN injection of 3 mg/kg GSK360A in P7 rats. Both treatments elevated the brain Ho1, Vegf, Glut1, and Epo mRNAs, but only IP-GSK360A treatment caused a marked increase of Epo mRNAs in the kidney (n>3). (F, G) IP-injection of GSK360A (30 mg/kg) significantly elevated the blood EPO protein level (>2600 pg/ml, p<0.0001, n>5 for each) and leukocyte count (192%± 89%, p<0.001, n>9 for each) in P7 rats. *p<0.05, **p<0.01, and ***p<0.001 by one-way ANOVA.