Activation of HIF-2a-EPO Axis in Kidney or Liver Is Sufficient to Drive Erythrocytosis in a Novel Inducible HIF-2a Transgenic Mouse Model

Erythropoietin (Epo) plays a critical role in the regulation of red cell mass (RCM) according to tissue oxygen levels. Hypoxia inducible factor-2a (HIF-2a) is known to be a master regulator for Epo expression. The liver is the major source of Epo in mice during embryonic and early postnatal development. In adult erythropoiesis, the kidneys become the major source of Epo and only a small portion of systemic Epo is produced by non-renal tissue, such as hepatocytes. The hepatocytes retain the capacity for Epo production in adult hematopoiesis and can increase Epo production in certain situations, such as severe hypoxia, post-nephrectomy, or certain liver cancer conditions. Erythrocytosis is defined as an absolute increase in RCM, and can be described as either primary or secondary, depending on whether the defect is intrinsic or extrinsic to the erythrocyte progenitors. The main causes of the secondary erythrocytosis include the defects in the oxygen-sensing pathway (VHL-PHD2-HIF2a signaling pathway) resident in the kidneys and liver, that transduces changes in oxygen delivery to changes in Epo gene transcription. Human mutations which cause erythrocytosis are usually genetic or somatic mutations; thus, both intrinsic and extrinsic effects to the erythrocyte progenitors could contribute to erythrocytosis. It is unclear whether kidney or liver specific upregulation of HIF-2a and downstream target Epo production will be sufficient to develop secondary erythrocytosis. It is also unclear whether any other cell type(s) in/out of the bone marrow (BM) also contribute to erythrocytosis phenotypes.

To model the gain-of-function (GOF) mutations of HIF-2a and excess Epo production in erythrocytosis, we generated an inducible HIF-2a transgenic allele with human HIF-2a mutant (P531A and N847A, double point mutant; DPM) under control of tetracycline response elements and CMV mini promoter. Using kidney and liver specific Cre (Pax3-Cre and Alb-Cre), Rosa26-loxP-Stop-loxP (LSL)-rtTA-IRES-GFP driver, and administering doxycycline, we could induce a stable HIF-2a protein (DPM) in the kidneys and liver. HIF-2a induced cells can also be detected by their GFP expression. We used Pax3-Cre/LSL-rtTA-GFP and Alb-Cre/LSL-rtTA-GFP mice without DPM transgenic allele as our control. After doxycycline administration, both Pax3-Cre/LSL-rtTA-GFP/DPM (Pax3-Cre/DPM) and Alb-Cre/LSL-rtTA-GFP/DPM (Alb-Cre/DPM) mice quickly developed typical erythrocytosis, while control mice did not show any abnormal phenotypes. We subsequently measured the peripheral blood (PB) Epo level with ELISA. The expression levels of Epo in the PB of Pax3-Cre/DPM and Alb-Cre/DPM mice were significantly higher than that from the control mice (control, 107 pg/ml; Alb-Cre/DPM, 2976 pg/ml; Pax3-Cre/DPM, 1810 pg/ml). The values of hematocrit (Ht, %: control, 39.9±1.2; Alb-Cre/DPM, 51.3±10.0; Pax3-Cre/DPM, 48.6±1.0) and hemoglobin (Hb, g/dl: control, 15.2±0.67; Alb-Cre/DPM, 20.8 ±2.84; Pax3-Cre/DPM, 17.3±1.04) in Alb-Cre/DPM and Pax3-Cre/DPM mice were significantly elevated compared to those in the control mice (control, n=10; Alb-Cre/DPM, n=11; Pax3-Cre/DPM , n=3, p<0.05 for both). White blood cell counts and platelet counts were comparable among the groups. Interestingly, PB monocytes from Alb-Cre/DPM mice were significantly increased compared to those from Pax3-Cre/DPM mice and control mice. The impact of these elevated monocytes on erythrocytosis phenotype is unclear. Prolyl hydroxylase domain protein (PHD) is critical for the degrading process of the HIF protein. Previously, it has been reported that PHD knockout in hematopoietic cells with Vav1-Cre caused erythrocytosis. However, when we induced HIF-2a DPM in hematopoietic cells using Vav1-Cre/LSL-rtTA-GFP/DPM, these mice did not show obvious erythrocytosis phenotype despite DPM and GFP expression in all the blood lineages.

In conclusion, we have generated a novel inducible HIF-2a transgenic allele and established both kidney- and liver-driven erythrocytosis mouse models. We found that kidney or liver specific Epo overexpression resulting from stable HIF-2a DPM mutant is sufficient to develop erythrocytosis, while HIF-2a DPM mutant induced in hematopoietic cells cannot develop erythrocytosis. Our study suggests that targeting VHL-PHD2-HIF2a-EPO pathway in kidney/liver will benefit patients with erythrocytosis.