Therapeutic inhibition of HIF-2α reverses polycythemia and pulmonary hypertension in murine models of human diseases

Polycythemia, also known as erythrocytosis, can result from multiple causes, including mutations in proteins involved in oxygen sensing and erythropoiesis, such as in the erythropoietin-erythropoietin receptor (EPO-EPOR), and mutations in hemoglobins (HBB, HBA), 2,3-bisphosphoglycerate mutase (BPGM), VHL, EGLN1 (prolyl hydroxylase domain-2 [PHD-2]), endothelial PAS-domain containing protein-1 (EPAS1, HIF-2α) and IRP1 genes that increase HIF-2α expression levels in renal tissues or endothelia.^1-9  Pulmonary hypertension can develop as a primary disease in which mutations in several disease genes have been identified,¹⁰  including bone morphogenetic protein receptor 2 (BMPR2)¹¹  and VHL in Chuvash polycythemia,^12,13  but in many cases, the causes and pathophysiology are not well understood. Better therapies are needed for polycythemia¹⁴  and pulmonary hypertension.^15,16 

A homozygous germline loss-of-function mutation in the von Hippel-Lindau (VHL) gene at codon 200 (R200W) is associated with a disease commonly known as Chuvash polycythemia, an inherited disease that is endemic in, but not limited to, the Chuvash region of Russia.^6,17,18  In addition to polycythemia, these patients develop pulmonary hypertension, thromboses, and cerebral hemorrhages and die prematurely.¹⁸  A mouse model bearing the same prevalent homozygous Vhl^R200W mutation recapitulates the polycythemia and pulmonary hypertension phenotypes,^13,19  and older Vhl^R200W mice also exhibit pulmonary fibrosis.¹³ 

The VHL protein, which plays a crucial role in cellular oxygen sensing and regulation of α subunits of hypoxia-inducible factors (HIF-α), is a component of an E3 ubiquitin ligase complex that mediates proteasomal degradation of HIF-1α and HIF-2α, also known as EPAS-1, after oxygen-dependent hydroxylation by prolyl-hydroxylases (PHDs), predominantly PHD-2, in normal oxygen conditions.^20-24  HIF-1α and HIF-2α are transcription factors that are involved in mediating the physiological response to changes in oxygen concentrations. At low cellular oxygen (hypoxic) conditions, impaired hydroxylation of HIF-α by PHDs prevents the VHL ubiquitin ligase complex from ligating and degrading HIF-α.²³  Stabilized HIF-1α and HIF-2α translocate into the nucleus where they dimerize with the constitutively expressed aryl hydrocarbon receptor nuclear translocator (ARNT), also known as HIF-1β, to form an active transcription factor complex.²⁵  HIF-α-ARNT complexes bind to hypoxia responsive elements²⁶  and activate transcription of numerous genes, including EPO, endothelin-1, and C-X-C motif chemokine 12 (CXCL-12).^21,27,28  Loss of VHL function leads to accumulation of HIF-α and high expression of the HIF target genes, highlighting the importance of VHL in oxygen sensing and cardiovascular homeostasis. Although both HIF-1α and HIF-2α are crucial regulators for the response to hypoxia, results of previous studies have demonstrated that HIF-2α activation has a predominant role in the pathogenesis of several polycythemia and pulmonary hypertension models.^13,29-31  In patients with Chuvash polycythemia and the VHL^R200W mouse model, impaired degradation of HIF-2α is responsible for both polycythemia and pulmonary hypertension,^12,13,32  as heterozygous deletion of Hif-2α, but not of Hif-1α, prevented the polycythemia and pulmonary hypertension in mice with the Vhl^R200W mutation.¹³  In addition, aged Vhl^R200W mice exhibited pulmonary fibrosis.¹³  Taken together, the results of these studies demonstrate that the PHD2/VHL/HIF-2α axis is important in regulating Chuvash polycythemia, pulmonary vascular homeostasis, and pulmonary fibrosis. Overexpression of HIF-2α led to elevated levels of several of its targets, including EPO, which promotes the production of red blood cells (RBCs); endothelin-1, which is a potent vasoconstrictor in endothelial cells^33,34 ; and Cxcl-12 which promotes fibrosis in lungs of Vhl^R200W mice.¹³ 

Recently, multiple patients with polycythemia phenotypes similar to patients with Chuvash polycythemia but without VHL^R200W mutations were found to carry mutations in iron regulatory protein 1 (IRP1) in Icelandic pedigrees.³  IRP1 and IRP2 are homologous mammalian cytosolic proteins that maintain cellular iron homeostasis by binding to a RNA stem-loop secondary structure, known as iron responsive elements (IREs) located in the 5′ or 3′ untranslated region (UTR) of messenger RNAs (mRNAs), encoding proteins that are involved in iron import, utilization, storage, and export.^35-37  Binding of the IRPs to the HIF-2α-IRE at 5′ UTR represses translation of HIF-2α.^38,39  We have previously studied mice with targeted deletion of Irp1⁴⁰  and have shown that these mice also develop polycythemia and pulmonary hypertension.⁴  Moreover, cardiac fibrosis and cardiomyocyte degeneration were observed in older Irp1-KO mice (>12 months old) maintained on a low-iron diet.⁴  The polycythemia phenotype of Irp1-KO mice was also reported independently by 2 other research groups.^9,41  Deletion of Irp1 leads to increased translation of Hif-2α, resulting in elevated levels of EPO in serum and endothelin-1 in pulmonary endothelia.⁴  Thus, a major molecular cause of polycythemia and pulmonary hypertension in patients and mouse models with VHL^R200W and IRP1 mutations is excess HIF-2α expression in multiple tissues due to either its decreased degradation or increased synthesis. HIF-2α also regulates expression of the CXCL-12 chemokine,²⁷  and increased expression of Cxcl-12 has been observed in the lungs of Vhl^R200W13  and Irp1-KO mice.⁴  CXCL-12 plays an important role in the progression of pulmonary hypertension⁴²  and in association with C-X-C motif chemokine receptor-4 (CXCR-4) promotes fibrocyte influx into the lungs, leading to fibroblast proliferation and idiopathic pulmonary fibrosis.^43-46 

Structural analysis of HIF-2α has identified a large ligand-binding pocket in its PAS-B domain^47-49  that is absent in HIF-1α.⁵⁰  Peloton Therapeutics Inc (Dallas, TX) and Merck & Co Inc (Kenilworth, NJ) have developed several small-molecule drugs, including PT2385 and MK-6482, also known as PT2977 (molecular structures shown in Figure 1), which can specifically occupy this cavity in the PAS-B domain of HIF-2α^51-53  and can thereby disrupt the binding of HIF-2α to its heterodimerization partner Hif-1β, and inhibit HIF-2α–mediated transcriptional activity. These small molecules inhibit the expression of HIF-2α target genes, including EPO, endothelin-1, and CXCL-12. The second-generation drug MK-6482 (3-[(1S,2S,3R)-2,3-difluoro-1-hydroxy-7-methylsulfonylindan-4-yl]oxy-5-fluorobenzonitrile)^53,54  contains a vicinal difluoro group that replaces the geminal difluoro group in the first-generation HIF-2α inhibitor PT2385, and these chemical modifications have been found to improve the pharmacokinetic profile and increase the potency of the modified drug.⁵³  Since VEGFA is a target of HIF-2α, and it plays a pathogenic role in advanced renal cell carcinoma, MK-6482 has been used in clinical trials to treat renal cell carcinoma, and has shown encouraging antitumor activity, together with a favorable safety profile.^53-56  In this study we treated Vhl^R200W, Irp1-KO, and double-mutant Vhl^R200W;Irp1-KO mice with MK-6482, hypothesizing that by inhibiting the Hif-2α activity, this drug would attenuate polycythemia, pulmonary hypertension, and pulmonary fibrosis in these mice. Our results showed that oral MK-6482 treatment successfully reversed both polycythemia and pulmonary hypertension in all 3 mouse models and normalized the expression of Cxcl-12, a marker for pulmonary fibrosis, in Vhl^R200W mice, indicating promising potential of this HIF-2α inhibitor for treatment of these harmful human diseases.

All procedures involving mice were performed according to protocols approved by the National Institute of Child Health and Human Development (NICHD) Animal Care and Use Committee (protocol 18-038) and met National Institutes of Health (NIH) guidelines for humane care of animals.

The Vhl^R200W mice were originally generated by Hickey et al.¹⁹  Heterozygous breeding pairs (C57BL/6 background) were kindly provided by Mary Slingo and Peter Robbins⁵⁷  (University of Oxford). The mice were genotyped as described by Hickey et al.¹⁹ Irp1-KO mice were generated as described⁴⁰  and backcrossed with C57BL/6 mice for 10 generations. Vhl^R200W;Irp1-KO mice were made by first crossing Vhl^R200W mice with Irp1-KO mice and breeding the resultant heterozygous mice. All the experiments were conducted in 6- to 11-month-old mice, unless specifically mentioned. The mice were weaned at 3 to 4 weeks of age and were maintained on a normal NIH-07 diet, unless otherwise mentioned.

We received pharmaceutical grade MK-6482 from Peloton Therapeutics Inc and Merck & Co, Inc. MK-6482 was formulated with 10% absolute ethanol, 30% PEG400, 60% water containing 0.5% methylcellulose, and 0.5% Tween 80 (all pharmaceutical grade from Spectrum Chemicals), according to methods used by other investigators.⁵²  The dose was 0.1 mg/g per day. Wild-type (WT), Vhl^R200W, Irp1-KO, and Vhl^R200W;Irp1-KO mice were treated via oral gavage once daily. Control mice received only vehicle. After 5 weeks, we measured complete blood count (CBC) and intraventricular pressure by Millar catheterization, harvested serum, euthanized the mice, and harvested tissues.

We placed the Irp1-KO mice for our main experiment on low-iron diets 2 weeks before the start of drug treatment to maximize their polycythemia phenotype.⁴  The low-iron diet was bought from Harlan Teklad and contained 3.7 ± 0.9 mg iron/kg chow (measured by inductively coupled plasma-mass spectrometry).

The Hep3b human cell line was cultured in Eagle’s minimum essential medium with 10% fetal bovine serum, 100 IU/mL penicillin, 100 μg/mL streptomycin, 2 mM l-glutamine, and 1% nonessential amino acids at 37°C in 5% CO₂. The cells were harvested after growing at either 21% O₂ or 2.5% O₂ in the presence or absence of 20 μM MK-6482 for 16 hours.

We withdrew ∼100 μL of blood from each mouse by using the mandibular bleeding method and funneling into a heparin-coated tube. Hematocrit tubes were ∼80% filled with blood by capillary action and sealed. We centrifuged the sealed tubes in an LW Scientific microhematocrit centrifuge and used a hematocrit reader to measure hematocrit.

Approximately 100 μL of blood was harvested into a heparin-coated tube from each mouse via mandibular bleeds. All the CBC measurements were performed with an IDEXX ProCyte Dx Hematology Analyzer.

Plasma volume was calculated from the known weight of the mice and the centrifuged hematocrit data according to the equation: plasma volume (ml) = total blood volume (ml) × (1 − hematocrit/100). Total blood volume was calculated as 0.079 mL/g × weight of mouse in grams, assuming that a mouse has, on average, 79 mL/kg of body weight blood. In this calculation of plasma volume, we used data obtained only from the mice with comparable body weights to minimize the relative error arising from total blood volume calculation.

EPO, endothelin-1, and Cxcl-12 were measured in serum samples with enzyme-linked immunosorbent assay (ELISA) kits from R&D Systems, according to the manufacturer’s protocol.

Real-time quantitative polymerase chain reaction was performed with primer sequences for EPO, endothelin-1, Cxcl-12, and actin, according to published methods.^4,58  The results were normalized against actin levels.

We invasively measured the intraventricular pressure by using a microtip pressure transducer catheter connected to an electrostatic chart recorder (Millar Instruments). After the mice were anesthetized with 1% to 3% isoflurane, the right external jugular veins were cannulated. The catheter was then advanced into the right ventricle. Correct placement was verified according to the right ventricular pressure curve, and the pressure and volume tracings were then recorded. The body temperature of the mice was maintained at 37°C to 38°C. The data were recorded using LabChart software (AD Instruments). The mice were then euthanized, while under anesthesia, by cervical dislocation or by exsanguination.

Kidney tissue lysates were prepared, and western blot analysis of Hif-2α was performed as previously described⁵⁹  with an Hif-2α antibody (catalog no. NB100-122; Novus Biologicals).

We performed transthoracic echocardiography in mice with a high-frequency ultrasound system (Vevo 2100; VisualSonics). We acquired the images with an MS-400 transducer (VisualSonics) with a center operating frequency of 30 MHz, and a broadband frequency of 18 to 38 MHz. The axial resolution and footprint of this MS-400 transducer are 50 µm and 20 × 5 mm, respectively. Two-dimensional images were obtained for multiple views of the heart.

Data are expressed as the mean ± standard deviation. Statistical analyses for multiple comparisons were performed with an ordinary 1-way analysis of variance (ANOVA). Differences were statistically significant at P < .05.

We placed the Irp1-KO mice on a low-iron diet 2 weeks before drug treatment to enhance polycythemia.⁴  We measured hematocrit levels of untreated, vehicle-treated, and MK-6482 (Figure 1)–treated WT, Vhl^R200W, Irp1-KO, and Vhl^R200W;Irp1-KO mice by capillary centrifugation. There was no significant difference in hematocrit levels between the untreated and vehicle-only–treated mice (supplemental Figure 1A, available on the Blood Web site). The hematocrit levels of vehicle-treated Vhl^R200W and Irp1-KO mice were significantly higher than those of vehicle-treated WT mice (Figure 2A-B; supplemental Figure 2A), reconfirming that these mutant mice had polycythemia.^4,19  The hematocrits of vehicle-treated double-mutant Vhl^R200W;Irp1-KO mice were substantially more elevated than those of either the Irp1-KO or the Vhl^R200W mice (Figure 2A-B; supplemental Figure 2A), as previously observed, consistent with their independent mechanisms of elevating HIF-2α, which should be additive.⁵⁸  After drug (MK-6482) treatment, the hematocrits of the Vhl^R200W, Irp1-KO, and Vhl^R200W;Irp1-KO mice decreased dramatically (Figure 2A-B; supplemental Figure 2A). No placebo effect was observed also for the other CBC parameters (supplemental Figure 1A-D). The analyses showed increased hemoglobin and RBC levels in vehicle-treated Vhl^R200W, Irp1-KO, and Vhl^R200W;Irp1-KO mice compared with the WT animals (Figure 2C-D). MK-6482 treatment markedly diminished levels of hemoglobin and RBC counts of the mutant and double-mutant mice, showing complete recovery from polycythemia in all 3 mouse models. We treated 5 Irp1-KO mice with MK-6482 and 5 Irp1-KO mice with vehicle, without maintaining them on a low-iron diet, and found that the effect of the drug on the hematocrits of those mice was comparable with that observed with the Irp1-KO mice receiving the low-iron diet (supplemental Figure 3). Plasma volumes, calculated from the weight of the mice and the centrifuged (spin) hematocrit data, were reduced in the Vhl^R200W, Irp1-KO, and Vhl^R200W;Irp1-KO mice, compared with that in the WT mice (supplemental Figure 2B). During drug treatment, calculated plasma volumes were increased in all the mutant and double-mutant mice, indicating that the observed changes in hemoglobin levels were complemented with the changes in plasma volumes.^18,60  MCV and MCH levels in Irp1-KO mice, but not in Vhl^R200W mice, were lower than those in WT mice (Figure 2E-F), consistent with previously observed systemic iron deficiency in Irp1-KO mice.⁴  Drug treatment did not restore normal MCV and MCH levels in the Irp1-KO mice. The MCHC levels of Vhl^R200W, Irp1-KO, and Vhl^R200W;Irp1-KO mice increased on treatment with the drug (supplemental Figure 2C). Significantly increased reticulocytosis in Irp1-KO mice, as compared with the WT controls (supplemental Figure 2D), indicated that there was stress erythropoiesis in the mice.⁸  However, reticulocytosis was diminished when the Irp1-KO mice were treated with MK-6482, suggesting that the drug mitigated stress erythropoiesis in these mice. We did not find any significant effect of either the mutations or the drug on WBC and platelet counts, although a slight increase in platelet levels was observed with drug treatment in all 3 mouse models (supplemental Figure 2E-F).

EPO is a glycoprotein cytokine secreted by the kidneys in response to hypoxia. It is a specific transcriptional target of HIF-2α, but not of HIF-1α.⁶¹ Because EPO stimulates RBC production, we measured EPO levels in the serum of vehicle- and drug-treated mice by ELISA, and we observed that the elevated EPO levels of Vhl^R200W mice and Irp1-KO mice were significantly diminished during drug treatment (Figure 3A). Notably, Irp1-KO mice had more than fivefold higher EPO levels than Vhl^R200W mice. MK-6482 treatment also significantly decreased the expression of EPO in Hep3b cells when the cells were grown at low (2.5%) oxygen concentrations (Figure 3B), to elicit expression of HIF-2α in response to hypoxia. These results demonstrated that MK-6482 attenuates polycythemia by reducing the EPO level in the mutant mice, and the drug also decreases the expression of EPO, in response to cellular hypoxia. We assessed Hif-2α protein levels in the kidney lysates of the vehicle- or drug-treated Vhl^R200W and Irp1-KO mice, as EPO is a specific target of HIF-2α.⁶¹  Hif-2α protein levels were not affected by drug treatment (supplemental Figure 4), as expected, because MK-6482 inhibits only HIF-2α–mediated transcriptional activity by disturbing HIF-2α-Hif-1β binding and does not contribute to the degradation of HIF-2α protein.⁵³ 

We measured intraventricular pressures by the Millar catheterization method. The data for the untreated and vehicle-treated groups of mice did not show any placebo effect (supplemental Figure 5). The right ventricular pressure (RVP) was markedly elevated in untreated or vehicle-treated Vhl^R200W, Irp1-KO, and double-mutant Vhl^R200W;Irp1-KO mice compared with WT mice (Figure 4A), reconfirming that pulmonary hypertension developed in Vhl^R200W and Irp1-KO mice and demonstrating for the first time that pulmonary hypertension develops in mice with both the Vhl and Irp1 mutations (Vhl^R200W;Irp1-KO).^4,19  However, contrary to the results obtained for hematocrits, the RVPs of Vhl^R200W and Irp1-KO mice^4,13  did not increase further in double-mutant Vhl^R200W;Irp1-KO mice. The drug treatment significantly decreased RVPs of the mutant and double-mutant mice, bringing the values close to WT level (Figure 4A). Untreated Vhl^R200W mice showed systolic flattening of the cardiac interventricular septum (IVS) compared with the normal convex appearance of the IVS in the WT mice throughout systolic contraction (Figure 4B). The D-shaped appearance of the left ventricle in the short-axis views of the untreated VHL mice correlated with elevated RVPs and pulmonary hypertension, as elevated right-side cardiac pressures pushed the IVS toward the left ventricular chamber. Importantly, treatment of the Vhl^R200W mice with the drug eliminated the development of IVS flattening (Figure 4B), further supporting that MK-6482 is an effective drug for attenuating pulmonary hypertension in the mutant mice.

Endothelin-1, a HIF target and potent vasoconstrictor, has been implicated in the pathogenesis of pulmonary hypertension, tissue hypertrophy, and fibrosis.³²  We measured endothelin-1 protein levels in the serum of drug- and vehicle-treated WT, Vhl^R200W, Irp1-KO, and double-mutant Vhl^R200W;Irp1-KO mice and found that the endothelin-1 levels significantly increased in the sera of all the mutant and double-mutant mice compared with the WT animals (Figure 4C). MK-6482 treatment markedly diminished the endothelin-1 level, returning it to near normal levels in all 3 mutant mouse models. In addition, increased endothelin-1 mRNA levels in Vhl^R200W mouse lungs were significantly reduced by drug treatment (supplemental Figure 6). These results suggested that MK-6482 mitigated pulmonary hypertension, in part by decreasing the levels of endothelin-1 released by pulmonary endothelial cells.

Aged Vhl^R200W mice develop pulmonary fibrosis,¹³  and 1-year-old or older Irp1-KO mice fed with a low-iron diet exhibit cardiac fibrosis.⁴  There is no known drug that reverses pulmonary fibrosis.^62,63  Improved therapies are also needed for cardiac fibrosis.⁶⁴  Expression of the chemokine CXCL-12, which, in association with its receptor CXCR-4, mediates the influx of fibrocytes into the lung to promote fibroblast proliferation and idiopathic pulmonary fibrosis,^43-45  is induced by HIF-2α.²⁷  Given that MK-6482 disrupts HIF-2α/ARNT heterodimerization^51-53  and thereby decreases the expression of HIF-2α target genes, we measured the Cxcl-12 expression levels in vehicle- and drug-treated WT, Vhl^R200W, Irp1-KO, and Vhl^R200W;Irp1-KO mice. The mRNA levels of Cxcl-12 in lung and heart tissues and protein expression levels of Cxcl-12 in serum were significantly increased in Vhl^R200W mice compared with WT mice (Figure 5). On MK-6482 treatment, these increased expression levels of Cxcl-12 in Vhl^R200W mice reverted to normal WT levels, indicating that MK-6482 may also be effective in reducing cardiac and pulmonary fibrosis in these mice over time.

We worked with mice lacking either Irp1 or functional Vhl, the 2 independent genetic mouse models of polycythemia and pulmonary hypertension that represent models of known human diseases. Phlebotomy therapy has long been used to treat patients with Chuvash polycythemia. However, frequent phlebotomies lead to iron deficiency that can cause 2 conflicting effects on the regulation of HIF-2α. In iron-deficient conditions, increased binding of IRP1 to the HIF-2α IRE at 5′ UTR causes translational repression of HIF-2α.^36,38,65,66  On the contrary, iron deficiency reduces PHD2 activity and subsequently promotes the stabilization of HIF-2α.^23,24  Thus, it is unclear whether phlebotomy represses or aggravates polycythemia.¹⁴  Our results clearly demonstrate that the HIF-2α-inhibitor MK-6482 reduces polycythemia in all 3 mouse models studied. Although Hif-1α deficiency has been shown to cause defective erythropoiesis in yolk sac and embryos,⁶⁷  our results indicate that Hif-2α is the major player in regulating erythropoiesis in our mouse models. The CBC parameters showed that mild anemia developed in all the experimental mice, including the WT mice, after treatment with MK-6482, indicating that the drug worked on target. Low-grade anemia was also found to be the most common adverse effect in patients with MK-6482-treated clear-cell renal cell carcinoma, and this anemia was managed with EPO replacement rather than dose reduction or discontinuation of treatment.^54,68  Pulmonary hypertension is a disease for which there is no cure. However, there are some treatment options available, including oral administration of the drugs ambrisentan, bosentan, or macitentan, all of which are endothelin-1 receptor antagonists.⁶⁹  Indeed, endothelin-1 plays an important role in the development of pulmonary hypertension.⁷⁰  In this study, endothelin-1 levels were increased in the mutant and double-mutant mice, and the drug treatment completely prevented the elevation of endothelin-1 levels in all 3 mutant mouse models, suggesting that this drug has a potential for use in treatment of pulmonary hypertension in patients with Chuvash polycythemia and in patients in whom the disease is caused by the upregulation of endothelin-1 or HIF-2α.

We previously showed that polycythemia develops in Vhl^R200W mice as early as 6 weeks of age,⁵⁸  and pulmonary hypertension develops in Irp1-KO mice as early as 3 months of age.⁴  Because the mice used in this study were 6 to 11 months of age when they received the drug, our results indicate that MK-6482 treatment reversed both polycythemia and pulmonary hypertension phenotypes in the mouse models. The extremely pronounced polycythemia and pulmonary hypertension observed in double-mutant Vhl^R200W;Irp1-KO mice were also reversed almost completely upon treatment with MK-6482. The mitigation of polycythemia and pulmonary hypertension is attributable to the loss of the Hif-2α transcriptional activity, as MK-6482 binds in the ligand binding pocket of the HIF-2α PAS-B domain and disrupts the HIF-2α/ARNT heterodimerization⁵²  necessary to transcriptionally activate target genes, including EPO, endothelin-1, and Cxcl-12 (Figure 6). Thus, daily oral intake of MK-6482 may represent a new approach to the treatment of Chuvash patients and of other patients who are suffering from forms of polycythemia, pulmonary hypertension, and pulmonary fibrosis caused by elevated HIF-2α levels.

Humans living at high altitudes, including Tibetans, develop genetic adaptations and show improved survival, despite living with comparatively low oxygen availability.^71,72  A high-frequency missense mutation in the EGLN1 gene encoding PHD2 is responsible for such adaptive response in Tibetans.⁷³  However, exposure to high-altitude hypoxia causes polycythemia characterized by augmented erythropoiesis and elevated hematocrit levels, and pulmonary hypertension^10,74  associated with elevated RVPs and increased endothelin-1 levels⁷⁵  in most nonadapted people. Our results suggest that treatment with the specific Hif-2α inhibitor MK-6482 can also mitigate hypoxic polycythemia and hypoxic pulmonary hypertension, indicating that this drug may be used to treat high-altitude polycythemia and pulmonary hypertension in nonadapted individuals. However, because this class of drugs may also cause defects in respiratory control, some caution should be taken, and individuals should be monitored carefully when MK-6482 is used as a therapy for patients who are dependent on hypoxic ventilatory drive.⁴⁹ 

An interesting observation we had here regarding the regulation of the Hif-2α targets EPO and endothelin-1 is that the translational inhibition of Hif-2α through binding of Irp1 to Hif-2α-IRE at the 5′ UTR plays a more pronounced role in regulating EPO than the Vhl protein–mediated proteasomal degradation of Hif-2α, whereas the reverse is true of the regulation of endothelin-1. A major difference between these 2 pathways for downregulation of Hif-2α is that the Irp1/Hif-2α IRE binding pathway specifically decreases Hif-2α expression, but Hif-1α expression remains unaffected, because Hif-1α has no IRE, whereas the Vhl protein–mediated pathway degrades both Hif-1α and -2α. In our previous work,⁵⁸  we showed that Tempol, a small, stable nitroxide molecule⁷⁶  corrects polycythemia in Vhl^R200W mice via translational repression of Hif-2α expression, but it did not attenuate pulmonary hypertension in those mice (M.C.G., unpublished results, 15 December 2015). As we observed in our previous study,⁵⁸  the serum EPO levels increased significantly, but only slightly in untreated Vhl^R200W mice, compared with WT animals. These results are consistent with a previously observed hypersensitivity to EPO and involvement of the JAK/STAT5 pathway.^77-79  Notably, inhibition of JAK-1 and JAK-2 by the inhibitor ruxolitinib was shown to reduce the severity of polycythemia in 3 patients.⁸⁰  Our results, however, imply that increased Hif-2α expression is the predominant factor that causes polycythemia in Vhl^R200W mice and clearly in Irp1-KO mice.

Suppression of Hif-2α signaling by a small-molecule Hif-2α inhibitor has recently been found to diminish pulmonary hypertension in rodents exposed to hypoxia for 4 to 5 weeks.³³  In another study, pharmacological inhibition of Hif-2α by a Hif-2α translational inhibitor C76 has been observed to reduce pulmonary hypertension in prolyl hydroxylase-2–deficient Egln1^Tie2Cre mice and Sugen 5416/hypoxia PAH rats.³⁰  Our study with 3 independent mouse models that had mutations comparable to those in 2 defined groups of patients suggested that polycythemia, pulmonary hypertension, and pulmonary and cardiac fibrosis can be mitigated by treatment with a small-molecule inhibitor of Hif-2α, MK-6482, a drug that has been shown to have a favorable safety profile in renal cancer trials.^53,54  Thus, multiple patients with pulmonary hypertension and/or polycythemia caused by mutations of critical genes known to be involved in the HIF-2α pathway, and others as yet undefined, may benefit from the potent and selective Hif-2α inhibitor MK-6482, which showed promising efficacy and tolerability in humans.

The authors thank Peloton Therapeutics Inc (Dallas, TX) and Merck & Co Inc (Kenilworth, NJ) for their generous gift of the drug MK-6482 (PT2977); Peter Robbins and Mary Slingo of Oxford University for providing the heterozygous Vhl^R200W mice; Laura Schmidt of Frederick National Laboratory for Cancer Research for helping to obtain the drug; Eric A. Meade of Merck for critically reading the manuscript; and members of the T.A.R. laboratory for contributing to constructive discussions.

This work was supported by the National Institutes of Health, Eunice Kennedy Shriver National Institute of Child Health and Human Development Intramural Research Program (grant ZIAHD001602).

Contribution: M.C.G. designed the project, performed most of the experiments, analyzed the data, and wrote the paper; D.-L.Z. performed experiments, analyzed the data, and reviewed the paper; W.H.O. performed experiments and reviewed the paper; A.N. and D.A.S. performed experiments, analyzed the data, and reviewed the paper; W.M.L. provided guidance on drug use and formulation and reviewed the paper; and T.A.R. designed and supervised the project and wrote the paper.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Tracey A. Rouault, National Institute of Child Health and Human Development, Bldg 35A, Rm 2D824, Bethesda, MD 20892-3755; e-mail rouault@mail.nih.gov.