|Home | About | Journals | Submit | Contact Us | Français|
Congenital or familial erythrocytosis/polycythemia can have many causes, and an emerging cause is genetic disruption of the oxygen-sensing pathway that regulates the ERYTHROPOIETIN (EPO) gene. More specifically, recent studies have identified erythrocytosis-associated mutations in the HIF2A gene, which encodes for Hypoxia Inducible Factor-2α (HIF-2α), as well as in two genes that encode for proteins that regulate it, Prolyl Hydroxylase Domain protein 2 (PHD2) and the von Hippel Lindau tumor suppressor protein (VHL). We report here the identification of two new heterozygous HIF2A missense mutations, M535T and F540L, both associated with erythrocytosis. Met-535 has previously been identified as a residue mutated in other patients with erythrocytosis, although the mutation of this particular residue to Thr has not been reported. In contrast, Phe-540 has not been reported as a residue mutated in erythrocytosis, and we present evidence here that this mutation impairs interaction of HIF-2α with both VHL and PHD2.
The three principal proteins of the oxygen-sensing pathway regulating the EPO gene are PHD2, HIF-2α, and VHL . PHD2 is a prolyl hydroxylase that site-specifically modifies HIF-2α in an oxygen-dependent manner. The primary site of hydroxylation is Pro-531 of HIF-2α, and this posttranslational modification allows recognition by VHL, a component of an E3 ubiquitin ligase complex [3–5]. VHL recognizes hydroxylated, but not unmodified, HIF. Under normoxic conditions, VHL targets HIF-2α for constitutive degradation. Under hypoxic conditions, this modification is attenuated, allowing stabilization of HIF-2α. HIF-2α then transactivates genes that promote adaptation to hypoxic conditions. A key gene is that encoding for EPO, the central regulator of red cell mass, and the transcriptional upregulation of the EPO results in increased circulating levels of EPO, increased red cell mass, and hence increased oxygen delivery to tissues[6, 7].
Recent studies have identified erythrocytosis-associated mutations in the genes that encode for these three proteins of the oxygen-sensing pathway[1, 8–11]. These include heterozygous mutations in the PHD2 gene, heterozygous mutations in the HIF2A gene, and either homozygous or compound heterozygous mutations in the VHL gene[12–14]. Current evidence indicates that the PHD2 and VHL mutations lead to loss of function of the respective proteins, while the HIF2A mutations lead to a gain of function of HIF-2α . Intriguingly, haplotypes in the HIF2A and PHD2 genes have also been associated with adaptation to high altitudes in Tibetans, highlighting a central role for these genes in hypoxic adaptation[15–17]. All of these issues make the documentation of human mutations in this pathway of considerable interest. In the present report, we identify two new HIF2A mutations associated with erythrocytosis.
Patient A, a 27 year old female, presented with dizzy episodes, and her routine blood picture showed a hemoglobin (Hb) of 17.4 g/dl, a hematocrit (Hct) of 0.51 with a white cell count of 5.8 × 109/l and normal platelet counts. The oxygen dissociation curve and abdominal ultrasound were both normal. She was a smoker. There is no history of thrombosis or pulmonary hypertension and no family history of erythrocytosis. No family members were available for screening. No splenomegaly was detected. No mutations of JAK2 V617F, JAK2 exon 12, VHL, or PHD2 were detected. Repeat Hb level was 17.6 g/dl and at this time her serum EPO was 6.3 mU/ml (reference range 5.0–25.0 mU/ml). She remains asymptomatic with a Hb at this level.
Patient B, an asymptomatic 49-year-old Brazilian male, presented with an increased Hb of 21.0 g/dl, Hct of 0.65, white cell count of 7.3 × 109/l, and platelet count of 236 × 109/l during routine blood tests. There was no history of either thromboembolic events or pulmonary hypertension, nor any family history of erythrocytosis. His father and grandfather both died of acute ischemic cerebral vascular events. He did not smoke, nor did he use any medications. Arterial blood gas analysis showed normal oxygen saturation and p50 values. EPO level was 38.2 mU/mL (reference range 5.0–25.0 mU/ml). Abdominal ultrasound was normal, as were ferritin and C-reactive protein levels. No mutations of JAK2 V617F, JAK2 exon 12, EPOR exon 8, VHL, or PHD2 were detected. The patient has been treated with phlebotomies and acetylsalicylic acid.
Patient C presented at age 35 with a raised Hb of 18.2 g/dl and Hct of 0.52. His white cell count was 3.5 × 109/l and platelet count 200 × 109/l. He is a smoker. There was no splenomegaly and no evidence of renal disease or pulmonary hypertension. His serum EPO level was 7.8 mU/ml (reference range 5.0–25.0 mU/ml). Sequencing both VHL and PHD2 did not detect any mutations. He remains asymptomatic.
Sequencing of exon 12 of HIF2A in these three individuals revealed two novel mutations (Figure 1A). Patient A was heterozygous for a c.1604 T>C mutation (middle panel), which exchanges Met for Thr at amino acid 535 (p.Met535Thr; M535T). Patient B had an identical heterozygous mutation, and this mutation was not present in his only son, who had a normal Hb level (data not shown). Patient C possessed a C to G change at c.1620 (c.1620C>G), resulting in a p.Phe540Leu (F540L) mutation (lower panel). In the case of patient C, a family history of erythrocytosis was confirmed, but no family members were available for screening. The serum EPO levels for patients A and C were within the reference range, while the EPO level for patient B was elevated. It may be noted that in many of the described cases with HIF2A mutations, it is well above the reference range [18–20].
Both mutations share two key features with previously described HIF2A mutations. First, they are heterozygous. Second, they affect residues that are C-terminal and in close proximity to the primary site of prolyl hydroxylation in HIF-2α, Pro-531(Figure 1B). In fact, the M535T mutation is now the third mutation that has been reported to affect Met-535, the first two being M535I and M535V[18, 19]. The M535V mutation has previously been shown to impair interaction of HIF-2α with PHD2 . Furthermore, HIF-1α Met-568, which corresponds to HIF-2α Met-535, is a contact residue for PHD2 in the cocrystal structure of HIF-1α (556–574):PHD2, therefore making it likely that Met-535 of HIF-2α is one as well.
The F540L affects a residue that has, thus far, not been a mutational target in the context of erythrocytosis. To examine the functional consequences of the F540L mutation, we proceeded as follows. To assess its effect on the interaction between HIF-2α and VHL, we performed a competition assay in which we incubated immobilized (biotinylated) and hydroxylated (Hyp-564) HIF-1α (556–574) peptide with VHL in the absence or presence of wild type or F540L Hyp-531 HIF-2α (527–542) peptide. In the absence of HIF-2α peptide, VHL binds to immobilized Hyp HIF-1α, as expected (Figure 2A, lane 3). In the presence of Hyp HIF-2α peptide, this binding is diminished, and we find that the mutant HIF-2α is a substantially weaker inhibitor than wild type (compare lanes 5 and 4), implying that this mutation weakens the interaction with VHL.
To assess the effect of the F540L mutation on the interaction of HIF-2α and PHD2, we performed a direct binding assay in which we examined the binding of PHD2 to immobilized glutathione S-transferase fused to either wild type or F540L HIF-2α (516–549). We find substantially less binding to the latter than the former (compare lanes 4 and 3), indicating that this mutation also weakens the interaction with PHD2. We therefore conclude that the F540L mutation impairs the interaction of HIF-2α with both PHD2 and VHL.
Taken together, these observations support the assignment of the M535T and F540L HIF2A mutations as new causes of erythrocytosis. The F540L mutation, in addition, is noteworthy for several reasons. First, the F540L mutation affects a residue that is the most distant from Pro-531 of any of the erythrocytosis-associated mutations that have been reported thus far. Second, like many, though not all, HIF-2α mutations, it impairs interaction with both its upstream modifying enzyme, PHD2, and its downstream ubiquitin ligase, VHL. Third, it is a seemingly conservative substitution of a bulky hydrophobic amino acid with another bulky hydrophobic residue, yet it produces significant functional defects. It may be noted in this regard that Phe-540 is conserved in HIF-2α proteins from mammalian species, human HIF-1α and HIF-3α, as well as HIF-2α proteins from chicken, frog, and zebrafish (Figure 1B). The importance of this residue is further highlighted by x-ray crystallographic studies, which show that the corresponding residue in HIF-1α, Phe-572, is a contact residue for both PHD2 and for VHL [22–24].
PCR-direct sequencing was performed for exon 12 of HIF2A as described previously.
Polyhistidine and Flag-tagged VHL and PHD2 were purified from baculovirus-infected insect cells using Ni-NTA-agarose (Qiagen) as described . Bacterial expression vectors for glutathione S-transferase (GST)-HIF-2α (516–549) and GST-HIF-2α (516–549) F540L were prepared using pGEX-5X-1 and oligonucleotides encoding the indicated sequences by standard recombinant DNA methods. Authenticity of the constructs was verified by DNA sequencing. GST, GST-HIF-2α (516–549), and GST-HIF-2α (516–549), were purified from E. coli transformed with the appropriate vectors using glutathione (GSH)-agarose (GE Healthcare).
Ten ng of recombinant VHL was incubated without or with 1 μg of biotinylated hydroxyproline (Hyp)-564 HIF-1α (556–574) prebound to 10 μl of streptavidin-agarose (Sigma) in the absence or presence of 5 nM wild type (WT) or F540L Hyp-531 HIF-2α (527–542) peptide. Peptides were obtained from Genscript. The resins were washed, eluted, and the eluates subjected to SDS-PAGE and western blotting using anti-Flag antibodies (Sigma), essentially as described [13, 21]. Quantification was performed using a Chemi Doc-It system (UVP Inc).
Ten ng of recombinant PHD2 was incubated with 1 μg of E. coli-expressed GST, GST-HIF-2α (516–549), or GST-HIF-2α (516–549) F540L prebound to 10 μl GSH-agarose. The resins were washed, eluted, and the eluates subjected to SDS-PAGE and western blotting using anti-Flag antibodies, as described [12, 21]. Quantification was performed using a Chemi Doc-It system.
Student’s t test was employed to analyze binding data.
We thank the patients who participated in the study. This work was supported in part by NIH grant R01-CA153347 (FSL), the University of Pennsylvania Scholars program (YJC), and a fellowship from FAPESP, Proc. 2010-17465-8, Brazil (PCJLS). The technical assistance of the Laboratory of Genetics and Molecular Cardiology group, Heart Institute group (University of Sao Paulo Medical School) is gratefully acknowledged.
† Conflict of interest: All authors declare there are no conflicts of interests.