Thiamine deficiency activates hypoxia inducible factor-1α to facilitate pro-apoptotic responses in mouse primary astrocytes

Abstract

Thiamine is an essential enzyme cofactor required for proper metabolic function and maintenance of metabolism and energy production in the brain. In developed countries, thiamine deficiency (TD) is most often manifested following chronic alcohol consumption leading to impaired mitochondrial function, oxidative stress, inflammation and excitotoxicity. These biochemical lesions result in apoptotic cell death in both neurons and astrocytes. Comparable histological injuries in patients with hypoxia/ischemia and TD have been described in the thalamus and mammillary bodies, suggesting a congruency between the cellular responses to these stresses. Consistent with hypoxia/ischemia, TD stabilizes and activates Hypoxia Inducible Factor-1α (HIF-1α) under physiological oxygen levels. However, the role of TD-induced HIF-1α in neurological injury is currently unknown. Using Western blot analysis and RT-PCR, we have demonstrated that TD induces HIF-1α expression and activity in primary mouse astrocytes. We observed a time-dependent increase in mRNA and protein expression of the pro-apoptotic and pro-inflammatory HIF-1α target genes MCP1, BNIP3, Nix and Noxa during TD. We also observed apoptotic cell death in TD as demonstrated by PI/Annexin V staining, TUNEL assay, and Cell Death ELISA. Pharmacological inhibition of HIF-1α activity using YC1 and thiamine repletion both reduced expression of pro-apoptotic HIF-1α target genes and apoptotic cell death in TD. These results demonstrate that induction of HIF-1α mediated transcriptional up-regulation of pro-apoptotic/inflammatory signaling contributes to astrocyte cell death during thiamine deficiency.

Fig 1. HIF-1α activation in mouse primary astrocytes.

Cells were treated with 10μM pyrithiamine (PT) up to 14d to induce thiamine deficiency relative to 3μM thiamine control (CTL). Representative Western blots are shown for expression of HIF-1α in WCL (A) and nuclear lysates (C). Actin was used as a loading control for WCL while p84 was used for nuclear samples. Densitometry of mean protein expression +/- SD includes n = 3 independent experiments for (B) WCL and (D) nuclear lysates. E) Real time-PCR analysis of mRNA expression +/- SD of the established HIF-1α target genes LDHA, GLUT1 and VEGF. Data are normalized to Actin as a loading control and the control sample using the 2-^ΔΔCt method. (★) Represents a statistically significant difference of p<0.05 compared to CTL based on the results of a one-way ANOVA with Tukey’s post-hoc test.

Fig 2. Effects of thiamine deficiency on the p53 pro-apoptotic pathway.

Cells were treated with 10μM pyrithiamine up to 14d to induce thiamine deficiency compared to 3μM thiamine control (CTL). A) Representative Western blots are shown for expression of p53 and pro-apoptotic target genes Bax and Bid in WCL. Actin was used as a loading control. B) Representative Western blot of p53 nuclear localization in thiamine deficiency. P84 is used as a loading control. Densitometry of mean protein expression +/- SD includes n = 3 independent replicates for WCL (C) and nuclear lysates (D). (★) Represents a statistically significant difference of p<0.05 compared to CTL based on the results of a one-way ANOVA with Tukey’s post-hoc test.

Fig 3. Effect of thiamine deficiency on expression of pro-apoptotic HIF-1α target genes.

Primary astrocytes were treated with 10μM pyrithiamine up to 14d to induce thiamine deficiency compared to 3μM thiamine control (CTL). A) Representative Western blots are shown for expression of MCP1, BNIP3, Nix and Noxa in WCL. Actin was used as a loading control. B) Densitometry of mean protein expression +/- SD includes n = 3 independent replicates. C) Real time-PCR analysis of mRNA levels +/- SD of the HIF-1α target genes MCP1, BNIP3, Nix and Noxa. Data are normalized to Actin as a loading control and the control sample using the 2^-ΔΔCt method. (★) Represents a statistically significant difference of p<0.05 compared to CTL among n = 3 independent replicates based on the results of a one-way ANOVA with Tukey’s post-hoc test.

Fig 4. Effects of thiamine repletion on the expression of HIF-1α regulated pro-apoptotic proteins.

Primary astrocytes were treated with 10μM pyrithiamine treatment for 4d (PT). Subsequently, 3μM thiamine was repleted into the culturing media for up to 5d (5R). A) Representative Western blot of HIF-1α and the established target gene LDHA in WCL. Expression of the pro-apoptotic target genes MCP1, BNIP3, Nix and Noxa in WCL are also shown. B) Densitometry of mean protein expression +/- SD is included with Actin as a loading control. (★) Represents a statistically significant difference of p<0.05 among n = 3 independent replicates based on the results of a one-way ANOVA with Tukey’s post-hoc test.

Fig 5. Effect of HIF-1α inhibition on expression of pro-apoptotic proteins.

To achieve pharmacological inhibition of HIF-1α, 10μM YC1 was supplemented into PT containing media after a loading dose of 20μM for 24h. YC1 +/- pyrithiamine treatments lasted a total of 4d. A) WCL was assessed for expression of HIF-1α, LDHA, MCP1, BNIP3, Nix and Noxa. B) Densitometry of mean protein expression +/- SD is included with Actin as a loading control. (★) Represents a statistically significant difference of p<0.05 among n = 3 independent replicates based on the results of a one-way ANOVA with Tukey’s post-hoc test.

Fig 6. TD associated pro-apoptotic protein expression is reduced following inhibition of HIF-1α.

Representative Western blots of cleaved Caspase-3 and cleaved Parp in WCL after A) treatment with 10μM pyrithiamine (PT) up to 14d, C) treatment with 10μM pyrithiamine for 4d with YC1 or E) treatment with 10μM pyrithiamine for 4d followed by 3μM thiamine repletion up to 5d (5R). Densitometry of mean protein expression +/- SD of each treatment set is shown with Actin as a loading control (B, D, F). (★) Represents a statistically significant difference of p<0.05 compared to CTL among n = 3 independent replicates based on the results of a one-way ANOVA with Tukey’s post-hoc test.

Fig 7. TD associated apoptosis is reduced following HIF-1α inhibition.

Primary astrocytes were treated with 10μM pyrithiamine (10μM PT) for 4d, 10μM pyrithiamine for 4d with YC1 (10μM PT+YC1) or treatment with 10μM pyrithiamine for 4d followed by 3μM thiamine repletion for 2d (10μM PT+2R). Representative plots of TUNEL assay analyzed by flow cytometry (A) with a quantitative summary of n = 3 independent replicates +/- SD (B). C) Representative microscopy images of TUNEL assay performed on fixed cells are shown after treatment with PT for 4d, PT + YC1 for 4d or PT for 4d with 2d of repletion. D) N = 3 independent replicates of the Cell death ELISA +/- SD. E) Representative plots of PI/ Annexin V staining analyzed by flow cytometry with a summary of n = 3 independent replicates +/- SD (F). (★) Represents a statistically significant difference of p<0.05 compared to CTL among n = 3 independent replicates based on the results of a one-way ANOVA with Tukey’s post-hoc test.

Fig 8. Schematic representation of the hypothesized role of HIF-1α in alcohol-induced neurological damage.

The poor diet of chronic alcohol consumers and subsequent loss of intestinal thiamine transport contributes to TD in these patients. We have demonstrated that TD induces HIF-1α signaling and pro-apoptotic/inflammatory HIF-1α target genes such as MCP1, BNIP3, Nix and Noxa in astrocytes. Independent of TD, ethanol metabolism by CYP2E1 leads to an increase in oxygen consumption resulting in the development of a hypoxic microenvironment and an increase in ROS. In astrocytes, this may also lead to stabilization of HIF-1α and subsequent cellular death. Overall, this would suggest that apoptosis in either uncomplicated alcoholism or in conjunction with TD is mediated by a HIF-1α response to induce pro-apoptotic/inflammatory signaling, as observed in ischemic disease.