Tibetan Gain-of-Function Variant of Prolyl Hydroxylase 2 (EGLN1) and Selected SNPs of HIF-2-Alpha (EPAS1) Are Associated with Lower Hemoglobin Values in Tibetans

Hypoxia is the principal driver of erythropoiesis. Tibetan highlanders are protected from polycythemia despite living at high altitude. Whole genome sequencing studies narrowed evolutionary selected Tibetan haplotypes to two loci, EGLN1 and EPAS1 (Huerta-Sanchez et al, Nature, 2014, 512:194). Our previous work identified a functional variant of the principal negative regulator of hypoxia inducible transcription factors (HIFs), prolyl hydroxylase 2 (PHD2^D4E/C127S, encoded by EGLN1^{12C>G, 380G>C}), that has a lower K_m for oxygen, thereby explaining the lack of polycythemia, and is present in ~88% of Tibetans (Lorenzo FR et al, Nat Genet. 2014, 46:951). Additionally, a highly selected Tibetan EPAS1 (encoding HIF-2a) haplotype has introgressed from prehistoric Denisovans (Huerta-Sanchez et al, Nature, 2014, 512:194). This Tibetan EPAS1 haplotype has no exonic variants but a unique epigenetic profile when analyzed by ENCODE software.

We analyzed genomes of 260 healthy adult Tibetans residing in different altitudes in China, India and the US with appropriate IRB approvals and informed consents. Since EGLN1^{12C>G, 380G>C} is located in a GC rich region that is not easily detectable by NGS and Sanger sequencing, we developed a high resolution melting assay for its detection. For the EPAS1 haplotype we analyzed 10 SNPs (5 Denisovan and 5 non-Denisovan, each with unique linkage disequilibrium (LD) value) in its evolutionary selected Tibetan region and correlated them with hemoglobin levels.

We found that EGLN1^12C>G (D4E) is always in cis (i.e complete LD) with the EGLN1^380G>C (C127S) missense variant, constituting a unique Tibetan-specific positively selected haplotype. The prevalence of PHD2^D4E/C127S variants increased with altitude, seen in 52 out of 55 individuals (94.5%) at altitudes above 4000 m, consistent with positive selection of PHD2^D4E/C127S at high altitude. Across all altitudes, after adjusting for age, gender and MCHC, the mean hemoglobin level was lower in PHD2^D4E/C127S heterozygotes (14.4 g/dl; p=0.009), and tended to be lower in homozygotes (14.7 g/dl; p=0.10) than in PHD2 wild-type subjects (15.2 g/dl). Combining the homozygotes and heterozygotes, the hemoglobin was lower (14.6 g/dl; p=0.015) compared to PHD2 wild-type subjects. This lower hemoglobin level in those with PHD2^D4E/C127S variants also extended to the subset of individuals residing at low altitudes (200 m). Thus, blunting of the HIF pathway by PHD2^D4E/C127S persists even at the ambient oxygen of lower altitudes, although none these low-altitude residents were found to be anemic. We then evaluated contributions of EPAS1 SNPs to hemoglobin as the response variable using linear regression controlled for age, sex, altitude and EGLN1 genotype. Five of the 10 selected EPAS1 SNPs remained significant after multiple testing corrections (FDR <0.05). At low altitude these SNPs are associated with lower hemoglobin concentration, but the, number of Tibetans with wild-type EGLN1 and EPAS1 haplotypes at higher altitude (>3,500 m; ~11,500 feet) was not sufficient for analysis.

In conclusion, we show that the gain-of-function PHD2^D4E/C127S is a Tibetan specific variant that forms a part of the genetic basis of high altitude adaption by lowering hemoglobin at all altitudes. Tibetans have a unique EPAS1 haplotype introgressed from prehistoric Denisovans, but no coding mutations in this EPAS1 haplotype were found. However, selected SNPs in this Tibetan EPAS1 haplotypes also contribute to their decreased hemoglobin. At the time of writing of this report we have obtained samples from additional 110 Tibetan subjects from Yushu area of Tibetan Plateau residing at altitude 3500 meters and higher (average 3700m). This will permit more rigorous analyses of combined effect of these two EGLN1 and EPAS1 positively selected Tibetan variants on hemoglobin levels at various altitudes. Our investigation of epigenetic alterations of the DNase hypersensitive regions and methylated regions of the Tibetan selected EPAS1 haplotype and the effect of this EPAS1 haplotype on transcriptome and determination of its functional relevance is in progress.