Hypoxic Response-Dependent Genetic Regulation Revealed By Allele-Specific Expression in Reticulocytes of Chuvash Polycythemia

Homozygosity for the VHL^R200W mutation in Chuvash polycythemia (CP) leads to decreased degradation of the α subunits of hypoxia inducible factor (HIF)-1 and HIF-2 by the hypomorphic variant of VHL, the principal negative regulator of HIFs. The constitutively activated HIFs directly regulate the transcription of a suite of hypoxic responsible genes, including the principal regulators of erythropoiesis, vessel development, and glycolytic metabolism, which further trigger a downstream cascade of gene expression. Besides these transcriptional factors, cis acting elements play an important role in the hypoxic gene regulatory network. To assess the extent of cis regulatory variation in hypoxic gene expression, we compared allele-specific expression (ASE) in purified reticulocytes between VHL^R200W homozygote individuals and age- and gender-matched wild type control individuals living at the same altitude of ~200 meters from the Chuvash population.

Cell fractions of reticulocytes were purified from 17 VHL^R200W homozygotes and 13 wild type individuals. Total RNA was extracted, depleted of ribosomal RNA and hemoglobin transcripts, and reverse transcribed. Strand-specific libraries were constructed for 125 bp paired-end sequencing to 30-45 million read pairs per sample using Illumina HiSeq 2500 platform. The samples were collected and processed in three batches across two years, with VHL genotype randomized in each batch. The sequencing data were mapped to human reference genome and analyzed for differential expression and differential ASE between VHL^R200W homozygotes and wild type individuals.

At 5% false discovery rate (FDR, i.e., <5 false positives in 100 detected genes), 1,267 genes were differentially expressed with more than 1.2-fold change in CP patients, 703 elevated and 564 decreased. Genes up-regulated in CP were enriched (fold enrichment >5, FDR <0.05) in REACTOME pathways of epigenetic remodeling (Packaging of telomere ends, DNA methylation, HDACs deacetylate histones, PRC2 methylates histones and DNA, Deposition of new CENPA-containing nucleosomes at the centromere, HATs acetylate histones) and oxidative stress induced senescence (DNA damage/telomere stress induced senescence, Senescence-associated secretory phenotype, Oxidative stress induced senescence). Genes decreased in CP were enriched in REACTOME pathways of cell cycle (E2F-enabled inhibition of pre-replication complex formation, Nuclear pore complex disassembly, SUMOylation of DNA replication proteins) and DNA damage repair (Activation of ATR in response to replication stress, SUMOylation of DNA damage response and repair proteins).

ASE was analyzed between CP and wild type individuals to assess hypoxic response-dependent genetic effects on gene expression. For the 1,267 genes differentially expressed in the CP, we selected genes containing exonic SNPs with heterozygous alleles for ASE analysis. With a null hypothesis of no cis acting regulation on the gene expression, both alleles are expected to be expressed at the same level, whereas allelic imbalance indicates linked cis regulation. At a given bi-allelic SNP, individuals with ≥2 read counts covering each of the reference and alternative alleles and with ≥20 total counts were included in the analysis. Exonic SNPs with at least one individual in each of the CP and wild type group were further selected to test for differential ASE between the CP and wild type groups, using a generalized linear model. A total of 147 genes passed the filtering and were analyzed, among which 32 were detected to have significant CP-dependent ASE at 5% FDR. Some of these genes may have important roles in hypoxic responses in CP reticulocytes, for example NEIL3, encoding a DNA glycosylase that initiates the first step in base excision repair by cleaving bases damaged by reactive oxygen species, and STOM, encoding an integral membrane protein that localizes to the cell membrane of red blood cells, loss of which is associated with hereditary stomatocytosis.

Our study reveals plethora of gene expression changes in CP reticulocytes compared to wild type controls, among which 22% could be regulated by hypoxic response-specific cis genetic variations. These observations indicate the prominence of cis elements in hypoxic response, for which substantial inter-individual differences exist even among a relatively isolated population.