Heme oxygenase-1 promoter region (GT)n polymorphism associates with increased neuroimmune activation and risk for encephalitis in HIV infection

Abstract

Background: Heme oxygenase-1 (HO-1) is a critical cytoprotective enzyme that limits oxidative stress, inflammation, and cellular injury within the central nervous system (CNS) and other tissues. We previously demonstrated that HO-1 protein expression is decreased within the brains of HIV+ subjects and that this HO-1 reduction correlates with CNS immune activation and neurocognitive dysfunction. To define a potential CNS protective role for HO-1 against HIV, we analyzed a well-characterized HIV autopsy cohort for two common HO-1 promoter region polymorphisms that are implicated in regulating HO-1 promoter transcriptional activity, a (GT)n dinucleotide repeat polymorphism and a single nucleotide polymorphism (A(-413)T). Shorter HO-1 (GT)n repeats and the 'A' SNP allele associate with higher HO-1 promoter activity.

Methods: Brain dorsolateral prefrontal cortex tissue samples from an autopsy cohort of HIV-, HIV+, and HIV encephalitis (HIVE) subjects (n = 554) were analyzed as follows: HO-1 (GT)n polymorphism allele lengths were determined by PCR and capillary electrophoresis, A(-413)T SNP alleles were determined by PCR with allele specific probes, and RNA expression of selected neuroimmune markers was analyzed by quantitative PCR.

Results: HIV+ subjects with shorter HO-1 (GT)n alleles had a significantly lower risk of HIVE; however, shorter HO-1 (GT)n alleles did not correlate with CNS or peripheral viral loads. In HIV+ subjects without HIVE, shorter HO-1 (GT)n alleles associated significantly with lower expression of brain type I interferon response markers (MX1, ISG15, and IRF1) and T-lymphocyte activation markers (CD38 and GZMB). No significant correlations were found between the HO-1 (GT)n repeat length and brain expression of macrophage markers (CD163, CD68), endothelial markers (PECAM1, VWF), the T-lymphocyte marker CD8A, or the B-lymphocyte maker CD19. Finally, we found no significant associations between the A(-413)T SNP and HIVE diagnosis, HIV viral loads, or any neuroimmune markers.

Conclusion: Our data suggest that an individual's HO-1 promoter region (GT)n polymorphism allele repeat length exerts unique modifying risk effects on HIV-induced CNS neuroinflammation and associated neuropathogenesis. Shorter HO-1 (GT)n alleles increase HO-1 promoter activity, which could provide neuroprotection through decreased neuroimmune activation. Therapeutic strategies that induce HO-1 expression could decrease HIV-associated CNS neuroinflammation and decrease the risk for development of HIV neurological disease.

Keywords: HAND; HIV-associated neurocognitive disorders; HMOX1; HO-1; Heme oxygenase-1; Immune activation; Interferon; Interferon-stimulated genes; Neuroinflammation.

Fig. 1

HO-1 (GT)n repeat length and A(-413)T SNP distributions differ between African Americans and Caucasians. HO-1 (GT)n polymorphism allele lengths were determined by PCR and capillary electrophoresis and the HO-1 A(-413)T SNP was determined by TaqMan SNP Genotyping Assay. a Histogram of the number of alleles of each HO-1 (GT)n repeat length for the entire cohort (HIV− and HIV+). The trimodal distribution of HO-1 (GTN)n repeat lengths was divided into short ‘S’, medium ‘M’, and long ‘L’ alleles as depicted in the histogram. b Histogram of the allelic frequency of each HO-1 (GT)n repeat length in Caucasians and African Americans in the entire cohort. Kolmogorov-Smirnov test revealed a significant (p <  0.001) difference in allelic distribution of HO-1 (GT)n repeat lengths between Caucasians and African Americans. c Allelic frequencies and d genotypic frequencies of the A(-413)T SNP in this cohort, revealing significant allelic and genotypic frequency differences between African Americans and Caucasian subgroups by chi-square (p <  0.001). P value: ***p <  0.001

Fig. 2

HIV-infected individuals with “short” HO-1 (GT)n repeat lengths have significantly decreased risk for HIV encephalitis. a Percentage of HIV+ subjects with HIVE in those with a short ‘S’ HO-1 (GT)n allele (SS, SM, SL) and those without a short ‘S’ allele (MM, ML, LL). HIV+ subjects with a short ‘S’ allele have a significantly reduced risk (odds ratio = 0.62) of having a post-mortem diagnosis of HIVE by chi-square analysis (p = 0.04). b Percentage of HIV+ subjects with HIVE with each HO-1 (GT)n allele genotype (‘S’ short, ‘M’ medium, ‘L’ long). c Percentage of HIV+ subjects with HIVE with each HO-1 A(-413)T SNP genotype groups by chi-sqaure test (p = 0.36). *p <  0.05

Fig. 3

HO-1 (GT)n repeat length does not correlate with plasma or CNS viral load or blood CD4 T-cell count in HIV-infected individuals. Correlations between an HIV+ individual’s shortest (GT)n allele length and a plasma viral load (HIV copies/ml), b CSF viral load (HIV copies/ml), c brain parenchyma viral load (HIV copies/g), and d peripheral blood CD4 T-lymphocyte count (cells/mm³). Solid lines represent best-fit linear regression models with dashed lines showing 95% confidence bands. Statistical analysis was performed by Spearman rank correlation. e Plasma viral load, f CSF viral load, g brain parenchyma viral load, and h peripheral blood CD4 T-lymphocyte count in HIV+ individuals with a short ‘S’ allele (SS, SM, SL) and those without a short ‘S’ allele (MM, ML, LL). Lines and error bars indicate median ± 95% confidence interval. All viral load data is log₁₀ transformed. Differences between groups were analyzed by Mann-Whitney test. n.s., not significant

Fig. 4

HIV-infected individuals with short ‘S’ HO-1 (GT)n alleles have significantly lower CNS expression of markers of type I interferon responses and T-lymphocyte activation. Prefrotanal cortex RNA expression of neuroimmune markers was compared between HIV-infected subjects (without HIVE) with a short ‘S’ HO-1 (GT)n allele (SS, SM, SL) and those without a short ‘S’ allele (MM, ML, LL). Neuroimmune markers analyzed were a MX1, b ISG15, c IRF1, d CD38, e GZMB, f CD8A, g CD68, h CD163, i CD19, j PECAM1, and k VWF. Lines and error bars indicate median ± 95% confidence interval of log₁₀ transformed RNA expression data. Red median lines and errors bars denote significant differences. Differences between groups were analyzed by Mann-Whitney test. n.s., not significant

Fig. 5

Length of the shortest HO-1 (GT)n allele correlates positively with CNS expression of markers of type I interferon responses and T-lymphocyte activation in HIV-infected individuals. Correlations were determined between the length of an HIV-infected (without HIVE) individuals shortest HO-1 (GT)n allele and the prefrontal cortex expression of the RNA markers a MX1, b ISG15, c IRF1, d CD38, e GZMB, f CD8A, g CD68, h CD163, i CD19, j PECAM1, and k VWF. Significant associations were determined by Spearman rank correlation. Solid lines represent best-fit linear regression models with dashed lines showing 95% confidence bands. Red regression lines denote significant differences. r, Spearman rank coefficient; n.s., not significant

Fig. 6

Prefrontal cortex type I interferon responses and T-lymphocyte activation marker expression trends lower in HIV-infected Caucasian and African American individuals with short ‘S’ HO-1 (GT)n alleles. Prefrotanal cortex RNA expression of neuroimmune markers was compared between HIV-infected subjects (without HIVE) with a short ‘S’ HO-1 (GT)n allele (SS, SM, SL) and those without a short ‘S’ allele (MM, ML, LL) in self-identifying Caucasian and African American subgroups. Neuroimmune markers analyzed were a MX1, b ISG15, c IRF1, d CD38, and e GZMB, those that significantly associated with the HO-1 (GT)n polymorphism in the full HIV+ cohort. Lines and error bars indicate median ± 95% confidence interval of log₁₀ transformed RNA expression data. Red median lines and errors bars denote significant differences. Differences in expression between subjects with short ‘S’ alleles and those without ‘S’ alleles in Caucasian and African American subgroups were analyzed by Mann-Whitney test