Genetic Basis of Congenital Erythrocytosis: Mutation Update and Online Databases

Background

Absolute erythrocytosis is defined by an increased red cell mass as reflected by hemoglobin and hematocrit values above the normal range. It can be either primary (intrinsic to the red cell) or secondary (extrinsic to the red cell) and can be acquired or arise from genetic alterations. Polycythemia Vera (PV) is the most common type of acquired primary erythrocytosis with somatic mutations in the JAK2 gene (MIM #147796) being responsible for almost 98% of the described cases (95% of the mutations involve exon 14 with the p.Val617Phe, whereas only 3% involve exon 12) [Cross, 2011]. Acquired secondary erythrocytosis can develop from various diseases, such as cardiac, pulmonary or renal, or conditions of external hypoxia due to smoking and CO poisoning [reviewed by McMullin, 2008; Patnaik and Tefferi, 2009].

Primary erythrocytosis, also known as Primary Familial Congenital Polycythemia (PFCP) is associated with a subnormal serum erythropoietin (Epo) level. It is caused by a molecular defect in the hematopoietic progenitor cells. Previously diagnosed cases have been found to possess germline gain-of-function mutations in the Epo receptor gene (EPOR; MIM #133171) [Huang et al., 2010 and Table 1]. In contrast, secondary congenital erythrocytosis (CE) is often characterized by inappropriately normal or raised serum Epo [van Maerken et al., 2004]. It can be a consequence of tissue hypoxia caused by hemoglobin variants with increased oxygen affinity due to mutations in the α- or β-globin genes (HBB, MIM #141900; HBA1, MIM #141800; HBA2, MIM #141850) [Percy et al., 2009] or defective bisphosphoglycerate mutase (BPGM; MIM #613896) leading to 2,3-bisphosphoglycerate (2,3-BPG) deficiency [Hoyer et al., 2004]. Secondary CE can also result from defects in components of the oxygen sensing pathway, mutations in the genes that encode the hypoxia-inducible factor 2α (HIF-2α; gene EPAS1; MIM #603349), HIF-prolyl hydroxylase 2 (PHD2, gene EGLN1; MIM #606425) and the von Hippel–Lindau tumor suppressor (pVHL; gene VHL; MIM #608537) have been reported [Lee and Percy, 2011] (Table 1).

Presently, over 160 mutations (including over 100 causing high affinity Hb variants with the remaining in either the EPOR gene or in genes involved in the oxygen sensing pathway) have been described associated with CE. However, in about 70% of CE patients a molecular cause was not identified. Thus, this condition is referred to as idiopathic erythrocytosis (IE) [Finazzi et al., 2006].

Mutation Nomenclature and Accession Numbers

The mutation nomenclature used in this update follows the guidelines indicated by Human Genome Variation Society (HGVS) [den Dunnen and Antonarakis, 2003]. Mutation descriptions have been checked using the Mutalyzer program (https://mutalyzer.nl/). Nucleotide numbering is based on GenBank reference sequences NM_000518.4 for HBB, NM_000558.3 for HBA1, NM_000517.4 for HBA2, NM_199186.2 for BPGM, NM_000121.3 for EPOR, NM_000551.3 for VHL, NM_022051.2 for EGLN1, NM_001430.4 for EPAS1.

The Oxygen-Sensing Pathway

Red blood cell production is regulated by the glycoprotein hormone Epo, which is mainly synthesized in interstitial tubular kidney cells. Epo production is increased under conditions of hypoxia due to anemia or decreased cellular oxygen tension.

Under normal oxygen tension, the alpha subunits of the hypoxia inducible factor (HIF-1, 2 and 3) are hydroxylated by the dioxygenase PHD (PHD1, 2, and 3) (Fig. 1). Hydroxylated HIF-α is then targeted by the VHL protein (pVHL) for ubiquitin-mediated degradation (the substrate recognition subunit of an E3 ubiquitine ligase complex that, in addition to pVHL, includes Elongin B, C, Rbx1, and Cul2). Under low-oxygen conditions, PHD proteins are unable to modify HIF-α allowing it to escape pVHL recognition and subsequent degradation. HIF-α then forms an active transcriptional complex with nuclear HIF-β (ARNT) and up-regulates expression of more than 200 genes, with one of the target genes being EPO. The major HIF-α isoform involved in the regulation of EPO is HIF-2α, which also regulates genes required for cell survival under low oxygen tension, such as heme synthesis (ALAS2), globin chains production (GATA1) and iron regulation (TRF2, TF) [Lok and Ponka, 1999; Wenger et al., 2005; Haase, 2010; Zhang et al., 2011; 2012]. Once Epo binds its cognate receptor there is initiation of an intra-cellular signaling cascade which inhibits apoptosis and simultaneously promotes the growth and differentiation of erythroid progenitors, thereby adjusting red blood cell mass to oxygen delivery requirements.

Recently, it has been shown that pVHL also form a heterodimeric E3 ligase complex with SOCS1 (suppressor of cytokine signaling 1) to target phosphorylated (p)JAK2 for proteasomal degradation [Russell et al., 2011]. The VHL p.Arg200Trp mutant has altered affinity for SOCS1 and fails to degrade pJAK2 that overactivates the Epo pathway and explains the Epo hypersensitivity of progenitors from Chuvash patients carring this mutation [Ang et al., 2002a, b].

Biological Significance

The identification of CE causal mutations in the HIF pathway genes has established the PHD2:HIF-2α:VHL pathway as the key regulator of adaptation and survival of both cells and the whole organism to hypoxia through Epo regulation [Lee and Percy, 2011].

The first insight into CE came from the studies on Chuvash polycythemia patients where the p.Arg200Trp mutation in the VHL gene led to an autosomal recessive form of erythrocytosis. The VHL p.Arg200Trp loss-of-function mutation results in diminished ubiquitination of the HIF transcription complexes and less proteasomal regulation in normoxia.

Further studies screening individuals with erythrocytosis for defects in HIF-1α, HIF-2α, and the three isofroms of PHD hydroxylases detected mutations in only PHD2 and HIF-2α genes [Percy et al., 2006; 2008a; 2008b]. These results indicated that PHD1 and PHD3 isoforms were unable to compensate for the loss of PHD2 function and there was no redundancy in the oxygen sensing pathway. Furthermore, the different isoforms of HIF and PHD exhibited different specific functions.

The HIF-1 transcription complex was described as the main regulator of Epo from binding studies and for a decade this was believed to be the case. However, the results from erythrocytosis studies caused a paradigm shift resulting in HIF-2α now being recognized as the main isoform that controls Epo. This was borne out by mice and RNA interference studies [Scortegagna et al., 2003; Warnecke et al., 2004]. It is now acknowledged that HIF-1α and HIF-2α regulate different target genes [reviewed by Mole and Ratcliffe, 2008].

Inherited mutations in HIF-2α are all located close to the site of prolyl hydroxylation at Pro531 and this region is crucial for the binding of PHD2 for hydroxylation and VHL for ubiquitination of HIF-2α [Furlow et al., 2009]. Functional analysis of a series of HIF-2α mutations has shown that in most cases the binding of both PHD2 and VHL is decreased, except for the p.Met535Val mutation, whereas VHL binding is retained. Thus, diminishing PHD2 binding alone is sufficient to cause impairment of the oxygen sensing pathway and dysregulation of Epo synthesis.

At the physiological level, VHL mutations may have more profound consequences, not just affecting hemopoiesis but also metabolism and exercise capacity, as both HIF-1α and HIF-2α proteins are stabilized in normoxia [Formenti et al., 2010]. Consequently, a broader range of target genes is upregulated in subjects carrying VHL mutations compared with those with HIF-2α mutations, further highlighting the differing functions of the HIF-α isoforms.

Clinical Significance of Mutation Identification

In CE patients, the elevated number of red blood cells and high hematocrit with a subsequent hyperviscosity, may lead to symptoms and signs ranging from headaches, dizziness, epistaxis, and exertional dyspnea to pruritus after bathing. Moreover, thrombotic and hemorrhagic events leading to premature morbidity and mortality have been reported. Hyperviscosity symptoms are effectively relieved by phlebotomy, but the increased risk of cardiovascular morbidity is not necessarily ameliorated by maintaining a normal hematocrit [McMullin et al., 2005; Finazzi et al., 2006].

Concerning the clinical presentation, the Chuvash cohort, homozygous for the VHL p.Arg200Trp mutation, has been most extensively studied. Homozygous patients were compared with a spouse control group and with age and sex matched community controls including VHL heterozygotes. Chuvash patients had reduced survival rates as compared with the control groups and presented a higher prevalence of arterial and venous thromboses and of hemorrhagic events [Gordeuk et al., 2006]. They also had a higher risk to develop venous varicosity. Patients with Chuvash polycythemia had lower blood pressures than heterozygote carriers whereas those had lower blood pressures than the controls [Gordeuk et al., 2004]. Cancer incidence in the Chuvash polycythemia cohort was not increased. There have been no reports on tumor development in patients with CE type 2 (VHL related), except for two cases of isolated hemangioblastoma [Woodward et al., 2007].

Cardiopulmonary physiology has been investigated in Chuvash polycythemia patients and compared with two control groups. Participants were studied at baseline and then subjected to hypoxia. Mild hypoxia induced a greater increase in ventilation in the Chuvash patients compared with the controls and they did not tolerate moderate hypoxia. They had abnormally high pulmonary artery pressures and hypoxia provoked a further abnormal rise. Physiological studies showed that Chuvash patients appeared to be in a situation characteristic of acclimatization to the hypoxia resulting from high altitude [Bushuev et al., 2006; Smith et al., 2006]. These patients should be regularly monitored for cardiopulmonary function.

Retrospective studies in the original Chuvash population failed to show significant effects of phlebotomy treatment and aspirin on the occurrence and severity of thromboembolic (and hemorrhagic) events [Gordeuk et al., 2004; Gordeuk et al., 2006]. Further studies are needed, as particularly in patients with pulmonary hypertension (PHT) discernable risks of phlebotomy treatment have to be calculated very carefully with regard to a possible negative influence of iron deficiency on PHT [Sable et al., 2012].

EGLN1 and EPAS1 mutations are mostly described in single case reports and only sparse clinical information is available. This includes a few thromboembolic events occurring even at young ages [Percy et al., 2008a]. PHT has also been described in individuals with the EPAS1 p.Gly537Arg mutation [Gale et al., 2008]. The underlying physiological changes are similar to those observed in patients with Chuvash polycythemia [Formenti et al., 2011]

Several cases with EGLN1 and EPAS1 mutations developed paraganglioma. This will be discussed in the section entitled “Risk of tumor development in patients with CE.”

Whereas the majority of adult patients with EPOR mutations had only mild symptoms some cases were reported to present with severe and even fatal clinical complications such as arterial hypertension, intracerebral hemorrhage, deep vein thrombosis, coronary disease, and myocardial infarction [Prchal et al, 1995; Sokol et al, 1995; Kralovics et al, 1997a; 1998;]. In the majority of patients, the disease could be controlled by antihypertensive treatment and by phlebotomies, either regularly performed aimed to maintain the hematocrit at an almost normal level or initi­ated only in the presence of hyperviscosity symptoms. Although a 40-year-old male patient had been phlebotomized regularly, it was reported that he died from myocardial infarction [Prchal and Sokol, 1996].

CE patients with high oxygen affinity hemoglobins usually are asymptomatic but hyperviscosity symptoms and thromboembolic episodes have been reported and related to the high hematocrit [Fairbanks et al, 1971; Weatherall et al., 1977]. In contrast to other CE types, phlebotomy cannot be recommended as general treatment of choice since erythrocytosis in these patients is primarily a requirement due to general tissue hypoxia. Therefore, phlebotomy treatment will be limited to single symptomatic events. In severe symptomatic cases, regular exchange transfusion has to be considered. Cases have been described showing unusual incidences of spontaneous abortion in female carriers caused by the lower oxygenation of the fetus resulting from either the alteration in the physiological oxygen gradient affinity between fetal and maternal blood or by placental infarction caused by the high viscosity of the mother's blood [Koller et al., 1980, Bento et al., 2000].

In conclusion, the raised hematocrit and increased viscosity associated with CE may lead to a number of clinical complications including increased thromboembolic events at young ages, but in the absence of good clinical data and follow up, it is difficult to obtain a clear picture of the clinical situation. At present, it is not possible to make unambiguous general treatment recommendations. However, the identification of the underlying genetic defect aids avoidance of possible pitfalls in the treatment (e.g., cytoreductive treatment for CE, or phlebotomy in hemoglobinopathies). A specified diagnosis also aids organization of adequate monitoring (e.g., for PHT in VHL and EPAS1 cases) and to counsel the patient.

It is advisable that patients document their family histories concerning all known cardiovascular events as well as causes of death and life spans of relatives. Bleeding tendencies should also be noted and further investigated when indicated. In the absence of clinical data, phlebotomy should be considered in CE mostly in special situations such as relief of possible hyperviscosity symptoms and as secondary prophylaxis if a thromboembolic event has occurred [Finazzi et al., 2006].

Our database will attempt to gather information on treatments and outcomes in pregnancies, with special risks for mother and fetus in CE, to aid future counseling. Genetic counseling is important and contraceptive methods with low thromboembolic risk should be chosen when indicated.

Carriers of VHL Mutations with CE

Although CE type 2 is considered a recessive disease the occurrence of individuals heterozygous for VHL mutations with erythrocytosis has been described. Eleven independent cases of erythrocytosis patients heterozygous for VHL mutations were reported in the literature. In contrast, other carriers of the same VHL mutation exhibited normal hematological parameters. In the case of heterozygous carriers with erythrocytosis the presence of other VHL mutations or a VHL null allele or deletion that could affect the apparently wild-type VHL allele had been ruled out and no mutations in the other genes associated with CE were found [Bento et al., 2005; Cario et al., 2005; Pastore et al., 2003a, b; Percy et al., 2003; Percy et al., 2007; Perrotta et al., 2006; Randi et al., 2005]. In addition 15 patients with erythrocytosis but only heterozygous VHL mutation are registered in the erythrocytosis database (Table 2). Interestingly, in two of these cases, the wild type VHL allele showed an unexplained low RNA expression (data not shown).

Diagnostic Strategies

When diagnosing a patient with erythrocytosis it is important to exclude acquired secondary (pulmonary, renal, and cardiac) or acquired primary (PV due to JAK2 mutations) causes. The family history and the determination of serum Epo levels are very useful in the decision regarding which molecular tests should be performed first. If available, determination of p50 (percentage at which Hb is half saturated with oxygen) can be helpful in establishing the presence of a hemoglobin variant with high oxygen affinity. Sequencing of the candidate genes is mandatory for a definitive diagnosis. Information on the laboratories performing molecular studies on CE genes is available at www.erythrocytosis.org. Based upon the serum Epo level and familial data, it is possible to establish an algorithm to decide which genes should be sequenced in each case of erythrocytosis (Fig. 2). A comparable algorithm with a specific focus on diagnostics in affected children and adolescents has been published recently [Cario et al., 2013].

Future Prospects

Significant advances have been made during the past decade in the CE field with the identification of causal mutations in the EPOR gene and the elucidation of the genes directly implicated in the hypoxia sensing mechanism (VHL, EGLN1, and EPAS1). Presently, over 160 mutations have been associated with CE but despite this about 70% of the CE patients, and 12–35% of PFCP cases, still remain unexplained at the molecular level. The absence of erythrocytosis in a child heterozygous for a deleterious nonsense EPOR mutation [Kralovics et al., 1998] and the observation of individuals heterozygous for VHL mutations with erythrocytosis confirm that other genes or epistatic factors must be implicated in the clinical manifestation of CE. The incoming use of next-generation sequencing is expected to further expand the number of genes involved in CE. With the implementation of the internet-based erythrocytosis database, it is hoped that it will allow the establishment of clinical and genotype–phenotype correlations in larger groups of individuals.