Primary Familial and Congenital Polycythemia
In contrast to polycythemia vera, PFCP is caused by germline rather than acquired somatic mutations. It is congenital and manifests autosomal dominant inheritance3 and, not infrequently, sporadic occurrence from de novo germline mutations. Like polycythemia vera (Chap. 84), it is primary in that the defect changes intrinsic responses of erythroid progenitors, and erythropoietin levels are low.
To date, 12 mutations of the erythropoietin receptor (EPOR) associated with PFCP have been described (Table 57–1). Nine of the 12 result in truncation of the EPOR cytoplasmic carboxyl terminal, and are the only mutations convincingly linked with PFCP. Such truncations lead to a loss in the negative regulatory domain of the EPOR (Chaps. 32 and 34). Three missense EPOR mutations have also been described, but these have not been linked to PFCP or any other disease phenotype (Table 57–1).
Table 57–1.Summary of Erythropoietin Receptor Gene Mutations ||Download (.pdf) Table 57–1. Summary of Erythropoietin Receptor Gene Mutations
|Type of Mutation ||Mutation ||Structural Defect ||Association with PFCP ||Ref. |
|Deletion (7bp) ||Del5985–5991 ||Frameshift > ter truncation ||Yes ||163,192 |
|Duplication (8 bp) ||5968–5975 ||Frameshift > ter truncation ||Yes ||222 |
|Nonsense ||G6002 ||Trp439 > ter truncation ||Yes ||223 |
|Nonsense ||5986 C>T ||Gln435 > ter truncation ||Yes ||224 |
|Nonsense ||5964C>G ||Tyr426 > ter truncation ||Yes ||162 |
|Nonsense ||5881C>T ||Glu399 > ter truncation ||Yes ||225 |
|Nonsense ||5959G>T ||Glu425 > ter truncation ||Yes ||226 |
|Insertion (G) ||5974insG ||Frameshift > ter truncation ||Yes ||227 |
|Insertion (T) ||5967insT ||Frameshift > ter truncation ||Yes ||228 |
|Substitution ||6148C>T ||Pro 488 > Ser ||No ||192,229 |
|Substitution ||6146A>G ||Asn487 > Ser ||No ||230 |
|Substitution ||2706 A>T ||Unknown ||No ||226 |
Erythropoietin-mediated activation of erythropoiesis involves several steps (also see Chap. 32). First, erythropoietin activates its receptor by inducing conformational changes of its dimers (Chap. 17). These changes lead to initiation of an erythroid-specific cascade of events. The first signal is initiated by conformation change-induced activation of Janus-type tyrosine kinase 2 (JAK2) and its phosphorylation and activation of a transcription factor, signal transducer and activator of transcription 5 (STAT5), which regulates erythroid-specific genes. This “on” signal is negated by dephosphorylation of EPOR by hematopoietic cell phosphatase (HCP), also known as SHP1, that is, the “off” signal. EPOR truncations lead to a loss of the negative regulatory domain of the EPOR, a binding site for HCP, explaining the gain-of-function properties of these EPOR mutations (Fig. 57–1).
Left panel: Erythropoietin binding to a normal erythropoietin receptor (EPOR) results in interaction of a protein kinase (JAK) with the receptor. The interaction leads to phosphorylation of the receptor and initiates a cascade of signaling that ultimately results in erythroid progenitor proliferation and differentiation. This process is self-regulatory. Activated signal transduction molecules, hematopoietic cell phosphatase (HCP) binds to the C-terminal of the EPOR, which is a negative regulatory domain. This interaction dephosphorylates the receptor and turns off the signaling resulting in cessation of erythroid progenitor proliferation. Right panel: Patients with mutated gain-of-function EPOR gene lack the C-terminal portion of the receptor that contains the negative regulatory domain. Erythropoietin binds and the signal transduction pathway is activated by change of configuration of erythropoietin receptor dimer, but because there is there is no structure for HCP to bind on the activated EPOR dimer, the receptor is left in the activated position resulting in unbridled erythroid proliferation and an elevated red cell mass. PO4, phosphate; STAT, signal transducer and activator of transcription.
The morbidity of primary polycythemias, such as polycythemia vera (Chap. 84), is largely the result of increased activated neutrophils and, perhaps, attendant pathologic platelet–endothelial interactions, whereas in secondary polycythemias it is presumably related to an increase in blood viscosity and, in part, to the resulting increased cardiac work.53 In most instances, the etiology of morbidity or mortality, such as associated with congenital disorders of hypoxia sensing, is largely unknown.54 The effect of blood viscosity on oxygen delivery is often oversimplified and the emphasis on the hematocrit alone may lead to ill-advised therapeutic interventions. In the normovolemic state, viscosity increases in a log-linear fashion as hematocrit increases, and the effect is particularly pronounced when the hematocrit rises above 50 percent. The prediction is that oxygen delivery decreases as hematocrit rises significantly above 50, as the greatly increased viscosity reduces blood flow, overshadowing the increased oxygen-carrying capacity of blood with a higher concentration of hemoglobin. However, polycythemia is not a normovolemic state, but is accompanied by an increase in blood volume, which, in turn, enlarges the vascular bed and decreases peripheral resistance (Chap. 34). Thus, hypervolemia can increase oxygen transport, and the optimum for oxygen transport occurs at higher hematocrit values than in normovolemic states. Consequently, despite the attendant increase in viscosity, an increase in hematocrit may generally be of benefit in appropriate secondary polycythemias. However, at some point, the high viscosity causes an increase in the work of the heart and a reduction in blood flow to most tissues and may be responsible for cerebral and cardiovascular impairment.
High Altitude Polycythemia
Adaptive adjustments of humans living at high altitude involve a series of steps that reduce the steepness of the oxygen gradient between the atmosphere and mitochondria (Fig. 57–2).55 The initial oxygen gradient between atmospheric and alveolar air can be reduced by an increase in respiratory rate and volume. Because dead space and water vapor pressure are constant and acclimatized individuals do not ventilate excessively, the normal sea level gradient of about 60 torr is only reduced to approximately 40 torr at 4540 m (14,900 ft) above sea level.55 Further reduction can be achieved, and at the top of Mount Everest, extreme hyperventilation reduces the gradient to less than 10 torr. A shift in the oxygen dissociation curve to the right, which represents decreased affinity of hemoglobin for oxygen, may be of benefit for short-term high-altitude acclimatization,56 but its usefulness for chronic acclimatization has probably been exaggerated.57 In the unacclimatized subject exposed acutely to high altitude, hyperventilation alkalosis leads initially to a shift of the oxygen dissociation curve to the left, representing an increased affinity of hemoglobin for oxygen, further worsening already present tissue hypoxia. The alkalosis and the hypoxia will, in turn, promote red cell synthesis of 2,3-BPG and cause the oxygen dissociation curve to shift back to a normal or even right-shifted position (Chap. 49). In chronic acclimatization, blood pH is slightly increased, and when this is taken into account, the dissociation curve is shifted approximately to normal.58 It is unlikely that a shift to the right would be to the advantage of high-altitude dwellers, except as a partial compensation for respiratory alkalosis.59 In addition, a right-shifted curve also has a decrease in oxygen loading in the alveolar capillary, minimizing any net gain in off-loading. There is a relationship between higher altitude and hemoglobin concentration response, best studied among Andean highlanders and Europeans in the United States; hemoglobin concentration is almost 10 percent higher in those Andean highlanders living at 5500 m than in those living at 4355 m. Furthermore, native Andean high-altitude dwellers have a gradual increase in their hemoglobin levels with age60 and body weight.61 Although it has been postulated that high hemoglobin–oxygen affinity in the setting of extremely low ambient oxygen may be one such adaptive change,62 increased hemoglobin–oxygen affinity or increased fetal hemoglobin are not adaptive phenotypes of Tibetan or Andean highlanders.63
The oxygen gradient from atmospheric air to the tissues in individuals living at sea level (Lima, Peru) and in Morococha, Peru, at 4540 m (14,900 ft) above sea level.
In a subset of Andean high-altitude native dwellers, namely Quechua and Aymara Indians, polycythemia becomes excessive and often results in chronic mountain sickness with its associated constitutional symptoms and pulmonary hypertension.42,60 This excessive erythrocytosis, called Monge disease or chronic mountain sickness,42,64 is also described in Han Chinese living in Tibet65 and occurs in whites living at high altitudes.66
The polycythemia encountered in high-altitude dwellers is often considered to be a universal, uniform adaptation process to hypoxia that would arise in all normal individuals. In reality, there is marked variability in erythropoietin level and subsequent polycythemic response to chronic hypoxia,60,67 suggesting that some of these factors may be genetically determined; the same degree of hypoxia induces substantial differences in erythropoietin production in response to high altitude.62,68,69 Three distinct adaptations to high altitude appear to have evolved. Andean highlanders have higher oxygen saturation than Tibetans at the same altitudes.62 Tibetan mean resting ventilation and hypoxic ventilatory response are higher than Andean Aymaras, whereas the mean Tibetan hemoglobin concentration is below the Andean mean. High levels of nitric oxide (NO) in the exhaled breath of Tibetans may represent increased physiologic NO. This effect may improve oxygen delivery by inducing vasodilation and increasing blood flow to tissues, thus making the compensatory increase in red cell volume unnecessary.70 Another distinct successful pattern of human adaptation in high-altitude dwellers that contrasts with both the Andean “classic” (arterial hypoxemia with polycythemia) and Tibetan (arterial hypoxemia with normal venous hemoglobin concentration) patterns evolved in Ethiopia. While Ethiopian high-altitude dwellers have hemoglobin concentrations that fall in the normal range (15.9 and 15.0 g/dL for males and females, respectively), they have a surprisingly, as-yet-unexplained high (average of 95.3 percent) oxygen saturation of hemoglobin despite their hypoxic environment (reviewed in Ref. 62). Their cerebral circulation is increased but is insensitive to hypoxia, unlike Peruvian high-altitude dwellers.71 Thus, Ethiopian highlanders maintain venous hemoglobin concentrations and arterial oxygen saturation within the ranges of sea level populations, despite the decrease in ambient oxygen tension at high altitude.72 Tibetans and Ethiopians have lived as mountain dwellers much longer than the Quechua or Aymara Indians,73 suggesting that extreme elevation of red cell mass is a maladaptation that Tibetans avoided by evolving a more efficient, or less detrimental, compensatory mechanism than that which causes Monge disease.
With rapid advances in genomics (Chap. 11), progress has been made in identification of the molecular basis of high-altitude adaptation; most of these advances have been in our understanding of Tibetan adaptation. Several studies reported evidence for positive natural selection in genomic regions in Tibetans; not surprisingly, most are haplotypes comprising genes that are components of hypoxia sensing that are mediated by hypoxia-inducible factors (HIFs) (described in Chap. 32). Two of these selected regions include genes that have undergone the strongest genetic selection, and are thus likely the most beneficial for Tibetan adaptation. These two regions encompass the EPAS1 gene, encoding the α subunit of HIF-2, and the EGLN1 gene, encoding proline hydroxylase 2 (PHD2). PHD2 is the principal negative regulator of both HIF-1 and HIF-2. Both of these haplotypes were shown to be associated with differences in hemoglobin concentration at high altitude by several independent studies.74,75,76 Intriguingly, the EPAS1 haplotype was previously identified as having the strongest Tibetan positive selection and was found to have an unusual haplotype structure that originated by introgression of DNA from Denisovan or Denisovan-related hominins.77 Denisovans, a sister group to the Neanderthals, branched off from the human lineage perhaps 600,000 years ago,78 and available evidence suggests that Denisovan and Neanderthal hominins contributed to the modern homo sapiens admixture before their extinction, likely by interbreeding,79 and that these hominin species provided genetic variations that helped humans adapt to new environments, such as extreme hypoxia associated with high altitude.77 The first Tibetan adaptation gene mutation identified, which changes the encoded PHD2 protein within the selected haplotype, is a missense variant of the EGLN1 gene, c.12C>G,80 that is in near complete linkage disequilibrium with a previously reported missense variant, EGLN1:c.380G>C.80 Both EGLN1:c.[12C>G; 380G>C] (PHD2D4E,C127S) are in cis; that is, the constituting PHD2D4E,C127S locus. Analysis of Tibetans and related populations suggests that 12C>G started on the 380G>C variant that is not Tibetan specific,80,81 that PHD2D4E,C127S originated from a single individual approximately 8000 years ago,81 and that now greater than 80 percent of Tibetans carry this PHD2 variant.81 Functional assessment of homozygous PHD2D4E,C127S recombinant proteins showed that the variant protein has increased hydroxylase activity under hypoxic conditions. Furthermore, native homozygous PHD2D4E,C127S erythroid progenitors have blunted erythropoietic responses to hypoxia by both erythropoietin-specific and erythropoietin-independent mechanisms.81 Although this is the first identified variant that contributes to the molecular and cellular basis of Tibetan adaptation to high altitude, there are other evolutionarily selected genomic regions, and elucidation of their functional impact is, at the time of this writing, unknown.
Understanding the etiology of polycythemia of high altitude is made more complex by a study of inhabitants of the Peruvian mining community of Cerro de Pasco (altitude 4280 m) with excessive erythrocytosis (mean hematocrit: 76 percent; range: 66 to 91 percent). About half of those individuals with a hematocrit greater than 75 percent had toxic serum cobalt levels,82 suggesting that other erythropoiesis promoting factors such as cobalt83 can augment hypoxia induction of erythropoietin, causing extreme polycythemias (Chap. 32). Most high-altitude dwellers do not have measurable levels or a history of exposure to cobalt or other heavy metals.84
Erythrocytosis of Pulmonary Disease
Degrees of arterial hypoxia comparable to those observed in individuals at high altitudes are observed in patients with right-to-left shunting resulting from cardiac or intrapulmonary shunts or to ventilation defects, as in COPD.
Many patients with COPD with severe cyanosis are not polycythemic. This has been attributed to infections and inflammation often present in the lungs, resulting in anemia of chronic inflammation, and to an increase in plasma volume. Why some patients with lung disease and congenital heart disease develop polycythemia, while others do not, is not entirely clear.
Erythrocytosis of Eisenmenger Syndrome
Patients with right-to-left shunting (Eisenmenger syndrome) develop a degree of erythrocytosis comparable to that observed with similar degrees of desaturation at high altitudes.85 The hematologic changes associated with this syndrome include hyperviscosity caused by erythrocytosis. Erythrocytosis is present in most patients, but excessive phlebotomy may cause microcytosis and some have attributed this effect to the exacerbation of the symptoms of hyperviscosity.28 In view of recent understanding of physiology of HIF regulation, it may not be the microcytosis, per se, that is detrimental, but the induced iron deficiency that inhibits PHD2 and increases HIF, which then directly causes pulmonary vasoconstriction and enhanced pulmonary vascular pressure (Chaps. 32 and 34).
Obstructive Sleep Apnea-Induced Syndrome
In the colorfully named pickwickian syndrome,86 polycythemia is characterized by its association with extreme obesity and somnolence. Today, the more widely studied OSA may not always be associated with obesity87 but can, if severe, cause arterial hypoxemia, hypercapnia, somnolence, and secondary polycythemia.88
Heavy smoking will result in the formation of carboxyhemoglobin, which does not transport oxygen (Chap. 50), and causes an increase in oxygen affinity of the remaining normal hemoglobin. Carboxyhemoglobin increases in relationship to the number of cigarettes or cigars smoked each day (Table 57–2). This leads to tissue hypoxia, erythropoietin production, and stimulation of red cell production.89 Smoking may also cause a reduction in plasma volume.90 Either augmentation of red cell mass or shrinkage of plasma volume could easily explain the rise in the hematocrit.
Table 57–2.Blood Oxygen Capacity in Smokers with Polycythemia ||Download (.pdf) Table 57–2. Blood Oxygen Capacity in Smokers with Polycythemia
|Subject ||Hemoglobin (Hgb) (g/dL) ||Carboxyhemoglobin (COHb) (g/dL) ||Hgb-COHb (g/dL) ||Affinity Correction (g/dL) ||Adjusted Hgb (g/dL) |
|Healthy male nonsmokers ||16 (14–18) ||0.16 (0.08–0.25) ||15.8 (14–18) ||0 ||16 (14–18) |
|Male smokers with increased Hgb concentration ||20 (17–23) ||2 (1–3) ||1816–21 ||1.5 (0.5–2.0) ||16.5 (15–19) |
Polycythemia Secondary to High-Affinity Hemoglobins
Hemoglobins with certain amino acid substitutions manifest an increased affinity for oxygen, producing tissue hypoxia and compensatory erythrocytosis (Chap. 49). Mutations affecting amino acids of the α1β2-globin chain interface affect normal rotation within molecules and impair the rate of deoxygenation. Changes in the carboxyl terminal and penultimate amino acids also impair intramolecular motion and tend to keep molecules in a high-affinity state. Alterations in the amino acids lining the central cavity of hemoglobin destabilize the binding of 2,3-BPG in this cavity and lead to increased oxygen affinity (Chaps. 47 and 49). Finally, some heme pocket mutations interfere with deoxygenation. Most hemoglobins with a mutation involving amino acids in the heme pocket are unstable and are associated with hemolytic anemia and cyanosis. Inheritance of these disorders is autosomal dominant. High-affinity hemoglobins result from mutations in any of three globin genes; those from α-globin gene mutations are congenital and life-long. β-Globin gene mutations are not apparent at birth but manifest after fetal to adult hemoglobin switching at approximately 6 months of life, while γ-globin gene mutations cause transient increase of hemoglobin concentration at birth lasting only about 6 months.
Polycythemia Secondary to Red Cell Enzyme Deficiencies
Deficiencies of red cell enzymes in early steps of glycolysis sometimes cause a marked decrease in the levels of 2,3-BPG (Chap. 47). Occasionally, mild polycythemia occurs in patients with methemoglobinemia as a result of cytochrome b5 reductase (methemoglobin reductase) deficiency (Chap. 50).
The same pathophysiology as that seen in high-affinity hemoglobins is also exhibited in mutations of the 2,3-BPG mutase gene, resulting in low 2,3-BPG. Because these mutations are very rare, with only a single family comprehensively studied,91 it is not clear if the mode of inheritance is recessive or dominant.
This condition, as well as other high-affinity hemoglobins, can only be conclusively confirmed by direct measurement of a hemoglobin dissociation curve, conveniently expressed as the partial pressure of oxygen required to saturate 50 percent of hemoglobin (p50O2); when equipment for this is not available, p50 can be estimated from pH, pO2 and hemoglobin oxygen saturation of venous blood.92,93
Chemically Induced Tissue Hypoxia
A number of chemicals have been suspected of causing histotoxic anoxia and secondary polycythemia, but the only chemical with a predictable capacity to cause erythrocytosis is cobalt.83 Cobalt administration increases erythropoietin production by increasing HIFs (see Chap. 32).94
Congenital Disorders of Hypoxia Sensing
Chuvash polycythemia is the only known endemic congenital polycythemia. The condition is caused by an abnormality in the oxygen-sensing pathway and causes thrombotic and hemorrhagic vascular complications that lead to early mortality; survival beyond age 65 years is uncommon.48,95 Inheritance is autosomal recessive, and affected patients tend to have normal blood gases, normal calculated p50, normal to increased erythropoietin levels, absence of genetic linkage to erythropoietin and EPOR loci, and no evidence of abnormal hemoglobin.95 In a study of five multiplex Chuvash families with Chuvash polycythemia, a homozygous mutation of the von Hippel-Lindau (VHL) gene (598C>T); that is, VHLR200W, was found in all affected individuals. This mutation impairs the interaction of VHL protein (pVHL) with both HIF-1α and HIF-2α, thus reducing the rate of ubiquitin-mediated destruction of HIF-1α and HIF-2α (Chap. 32). As a result, the level of HIF-1 and HIF-2 heterodimers increases and leads to the increased expression of target genes, including the erythropoietin (EPO), vascular endothelial growth factor (VEGF), and plasminogen activator inhibitor genes (PAI-1), among others.4,5 Figure 57–3 depicts the effect of this mutation on hypoxic sensing. The role of circulating erythropoietin in the Chuvash polycythemia phenotype is indisputable. However, there must be other factors associated with the Chuvash polycythemia VHL mutation that contribute to the polycythemic phenotype, as the erythroid progenitors of Chuvash polycythemia patients are hypersensitive in vitro to extrinsic erythropoietin; the mechanism of this observation is not fully explained.4,5 Some, but not all, patients with other VHL mutations have erythropoietin hypersensitive erythroid colonies96,97,98; in these patients there is also increased expression of the RUNX1 and NFE1 genes, which can stimulate erythropoiesis.6
Elongins B, C, and proteins Rbx1, Cul2 E2, and NEDD 8 are interacting proteins that facilitate von Hippel-Lindau (VHL) function. Interaction of mutated VHL protein with HIF-1α. The Chuvash VHL mutation leads to the impaired interaction with HIF-1α, which results in impaired degradation in 26S proteasome and augmented hypoxia sensing. CP, Chuvash polycythemia.
Despite increased expression of HIF-1α, HIF-2α, and VEGF in normoxia, Chuvash polycythemia patients do not display a predisposition to tumor formation. Imaging studies of 33 Chuvash polycythemia patients revealed unsuspected cerebral ischemic lesions in 45 percent, but no tumors characteristic of VHL syndrome.99 There also, is a high prevalence of this disorder on the Italian island of Ischia.52 The Chuvash VHLR200W mutation has also been described in whites in the United States and Europe, and in people of Punjabi/Bangladeshi Asian ancestry.100 Some patients with congenital polycythemia have proved to be compound heterozygotes for the VHLR200W mutation and other VHL mutations. Additionally, two distantly related Croatians with polycythemia are homozygous for VHLH191D, the first example of a homozygous VHL germline mutation other than VHLR200W causing polycythemia.51,54,101,102,103,104,105
A small number of cases of congenital polycythemia that appear to have a mutation of only one VHL allele confound an obvious pathophysiologic explanation. In a Ukrainian family, two children with polycythemia were heterozygotes for VHL 376G>T (D126Y), but the father with the same mutation was not polycythemic.104 An English polycythemic patient was a heterozygote for VHL 598C>T106; but the inheritance of the deletion of a VHL allele, or null VHL allele, in a trans position was not excluded. Subsequently, two polycythemic VHL heterozygous patients were described in whom a null VHL allele was more rigorously excluded101,102; the molecular mechanism of their polycythemic phenotype remains to be elucidated.
To address the question of whether the VHL 598C>T substitution occurred in a single founder or resulted from recurrent mutational events, haplotype analysis of eight highly informative single nucleotide polymorphic markers covering 340 kb spanning the VHL gene was performed on 101 subjects bearing the VHL 598C>T mutation and 447 normal unrelated individuals from Chuvash, Southeast Asian, white, Hispanic, and African American ethnic groups.49 Polymorphism of the VHL locus in normal controls (having a wild VHL 598C allele) and subjects bearing Chuvash polycythemia VHL 598T were in strong linkage disequilibrium. These studies indicated that, in most individuals, the VHL 598C>T mutation arose in a single ancestor between 51,000 and 12,000 years ago. However, this is not the case for a Turkish polycythemic family with a VHL 598C>T mutation wherein the VHL 598C>T mutation occurred independently.102
Chuvash polycythemia homozygotes have decreased survival because of thrombotic and hemorrhagic complications, mostly in the venous circulation,99 and thus are under negative selection pressure. The high frequency of the mutation in some areas may be the result of random factors (“drift”), but it is also possible that propagation of the VHL 598C>T mutation is the result of a survival advantage for heterozygotes. Such an advantage might be related to a subtle improvement of iron metabolism, erythropoiesis, embryonic development, energy metabolism,106 or some other as yet unknown effect. Indeed, heterozygotes were shown to be less likely to be anemic compared to control subjects.107 Another potential protective role of a mildly augmented hypoxic response is improved protection against bacterial infections, as the hypoxia-mediated response has been reported to be essential for the bactericidal action of neutrophils.108
Classic von Hippel-Lindau Syndrome
VHL syndrome is an autosomal dominant genetic abnormality affecting the posttranslational control of HIF-1α.109,110,111 The syndrome is characterized by a propensity for developing renal cell carcinomas, retinal hemangioblastomas, cerebellar and spinal hemangioblastomas, pancreatic cysts, and pheochromocytomas. The tumors result from a somatic mutation in addition to the germline mutation, that is, loss-of-heterozygosity. Polycythemia is not part of VHL syndrome but hemangioblastomas of the central nervous system, and, less commonly, pheochromocytoma and renal cancer, have been associated with polycythemia mediated by paraneoplastic erythropoietin production.111 Other patients with VHL syndrome also develop acquired polycythemia.99,111 The VHL gene codes for 213 amino acids, and more than 130 germline mutations associated with classic VHL syndrome have been identified, virtually all of them 5′ to the codon 200 position that is mutated in Chuvash polycythemia.112 Figure 57–4 depicts the schematic effect of the Chuvash polycythemia mutation in the context of other previously found VHL mutations.
von Hippel-Lindau (VHL) gene structure and mutation. Three exons of VHL genes are depicted encoding for UTR (untranslated portion of mRNA), and coding sequences (CDs). VHL domains β, α, β are shown. The relative number of reported VHL gene mutations are depicted in vertical lines. The location of the Chuvash polycythemia mutation is depicted by the diamond.
It is not clear why mutations of a single gene lead to these two diverse phenotypes. It has been suggested that quantitative differences in loss of activity could explain the variable phenotypes among VHL mutations,113 but the VHL gene may also have other functions, possibly as a result of interactions with other modifying factors, that can contribute to the onset of disease and that await future clarification. Another plausible explanation of polycythemia versus cancer predisposition syndrome is that almost all polycythemic subjects have germline mutations of both VHL alleles, whereas those with VHL cancer predisposition syndrome have only a single germline mutation and then acquire a somatic mutation that is essential for tumor genesis.
EGLN1 Gene Mutations, Proline Hydroxylase Deficiency
Another principal negative regulator of HIFs is PHD2 (encoded by the EGLN1 gene), which targets the α subunit of HIF for degradation. The first loss-of-function mutation of PHD2 (PHD2P317R) was identified in a family in which heterozygotes had mild or borderline polycythemia.114 Since then, 25 additional patients with unexplained polycythemia who are heterozygote carriers of different PHD2 mutations have been reported.115 Almost all patients with PHD2-associated polycythemia have normal erythropoietin levels. Whether the cause of polycythemia in this case is haploinsufficiency or a dominant negative effect remains to be determined.
EPAS1 (HIF-2α) Gain-of-Function Mutations
Affected patients have heterozygous missense mutations in the coding sequence of the EPAS1 gene that encodes HIF-2α, and typically have elevated erythropoietin levels.115,116 There is heterogeneity in these gain-of-function HIF-2α mutations, but their existence supports the critical role of HIF-2α in controlling the expression of erythropoietin. Some patients with EPAS1 mutations, similar to Chuvash polycythemia, have erythropoietin hypersensitive colonies, thus sharing features of both primary and secondary polycythemias.97
An explanation for the hypersensitivity of erythroid colonies bearing mutations that augment HIF stabilization remains to be discovered. It has been proposed that mutated VHLR200W protein hinders suppression of cytokine signaling SOCS1-mediated JAK2 degradation, via binding of a negative regulator of erythropoiesis, SOCS1,117 to the extreme 3′ coding region of the VHL gene. Other observations are not consistent with this proposed mechanism: Another closely positioned VHL polycythemia mutation, VHLH191D, is not associated with erythropoietin hypersensitivity,96 while other, more upstream, mutations such as VHLP138L are.80 Furthermore, the hypersensitivity of erythroid colonies is also seen in some HIF-2α mutations.86 Interestingly, in some, but not all, of these families, upregulation of NFE2, which enhances erythropoiesis, has been found.6,118
Unexplained Congenital Polycythemias with Elevated or Inappropriately Normal Levels of Erythropoietin
The majority of patients with congenital polycythemias with inappropriately normal or elevated erythropoietin levels do not have VHL mutations, EGLN1 or EPAS1 mutations, hemoglobinopathies, or 2,3-BPG deficiency, and the molecular basis of polycythemia in these cases remains to be elucidated. Some such families show dominant inheritance,119 while in others inheritance is recessive, and in some it is sporadic. Lesions in genes linked to hypoxia independent regulation of HIF, as well as oxygen-dependent gene regulation pathways, are leading candidates for mutation screening in polycythemic patients with normal or elevated erythropoietin without VHL, EGLN1 (proline hydroxylase), or EPAS1 (HIF-2α) mutations.
Other Inappropriate Secondary Erythrocytoses
Renal Polycythemia and Post–Renal Transplant Erythrocytosis
Absolute erythrocytosis has been observed in a considerable number of patients with solitary renal cysts, polycystic renal disease, or hydronephrosis. In most of these cases, erythropoietin assays on cyst fluid, serum, or urine have disclosed the presence of erythropoietin.120 Patients with polycystic disease have a hematocrit value slightly higher than normal and definitely higher than would have been expected of patients with uremia. In some patients on prolonged dialysis treatment, cystic transformation occurs in the native kidneys. This acquired cystic disease is occasionally associated with marked erythrocytosis.121 In patients with pheochromocytoma or paraganglioma and erythrocytosis, erythropoietin assays of serum and urine have disclosed higher-than-normal levels, and the erythrocytosis is most likely caused by excessive erythropoietin secretion by the tumor. This assumption has been supported by the presence of erythropoietin mRNA in tumor cells.122 Wilms tumors123 and paraganglioma124 are also occasionally associated with an erythrocytosis. Many of these cases may have a somatic VHL gene mutation that, in combination with a germline mutation of another allele, may constitute an unrecognized VHL syndrome. A patient with congenital erythrocytosis and recurrent paraganglioma with a PHD2 mutation was described. Tumor tissue exhibited a loss of heterozygosity of PHD2 in the tumor, suggesting that PHD2 could be a tumor-suppressor gene.34
Partial obstruction of the renal artery would be expected to cause renal tissue hypoxia and a physiologic stimulation of erythropoietin production. Nevertheless, it has proved quite difficult to induce erythrocytosis in laboratory animals by placing a Goldblatt clamp on the renal arteries.125 Only a few of the many patients who have arteriosclerotic narrowing of the renal arteries have been reported to be polycythemic.126
Post–Renal Transplantation Erythrocytosis
Although the full molecular basis of post–renal transplant erythrocytosis remains unknown, angiotensin II (Chaps. 32 and 34) plays an important role in its pathogenesis.127 Increased activity of the angiotensin II–angiotensin receptor 1 pathway makes the erythroid progenitors hypersensitive to angiotensin II.128,129 Furthermore, angiotensin II can modulate release of erythropoiesis stimulatory factors (Chap. 32) including erythropoietin and insulin-like growth factor (IGF)-1.130,131 Studies of venous effluents have determined that the native rather than the transplanted kidneys are the source of the inappropriate production of erythropoietin,132 and in some patients, removal of the native kidneys has led to rapid restoration of normal hematocrit values.133 The condition is rarely seen in patients with nonrenal solid-organ allografts. The role of angiotensin II in augmenting erythropoiesis was confirmed by anemia in angiotensin-converting enzyme–knockout mice.134 Prior to the late 1990s when the use of angiotensin-converting enzyme inhibitors increased as a means to reduce proteinuria, the incidence of erythrocytosis in renal transplant patients was approximately 8 to 10 percent within the first 2 years after engraftment.
Polycythemia with Connective Tissue Tumors
Occasionally, there is an association of erythrocytosis with large uterine myomas.35 Usually, the tumor has been huge and extirpation has routinely been followed by a hematologic “cure.” The suggestion that the tumor interferes with pulmonary ventilation has not been supported by the normal arterial blood gas findings in the few patients so studied. Another possible mechanism is that the large abdominal mass causes mechanical interference with the blood supply to the kidneys, resulting in renal hypoxia and erythropoietin production. Inappropriate erythropoietin secretion by smooth muscle cells has been demonstrated both in uterine myomas and in one case of cutaneous leiomyoma.35,135 Rare cases of polycythemia attributed to a myxoma of the atrium,36 hamartoma of the liver,37 and focal hyperplasia of the liver38 have been documented.
In adequately studied patients with erythrocytosis and cerebellar hemangiomas, arterial blood gas tensions have been normal. That the tumors are directly responsible for the polycythemia can be surmised from the identification of erythropoietin in cyst fluid and stromal cells and from a case in which erythropoietin mRNA was present in the tumor.136 Although in these cases a mutation of the VHL gene was not sought, it is likely that these tumors were a manifestation of an underlying VHL syndrome as cerebellar hemangiomas are an integral feature of VHL syndrome.
In 1958, McFadzean and coworkers reported that almost 10 percent of patients in Hong Kong with hepatocellular carcinoma developed erythrocytosis.137 Since then, this association has been recognized as an important clinical clue in the diagnostic consideration of patients with liver disease.138 The cause of erythrocytosis is probably inappropriate production of erythropoietin by the neoplastic cells.139 Normal hepatocytes, and to a lesser degree nonparenchymal liver cells, produce small amounts of erythropoietin, both constitutively and in response to hypoxia.
Congenital Polycythemia and Pheochromocytoma
Pheochromocytomas have been described in association with congenital erythrocytosis.140 In a growing number of reports, several individuals with congenital polycythemia have developed recurrent pheochromocytomas, paragangliomas, and sometimes somatistatinomas.141,142,143 The tumors in these patients are heterozygous for gain-of-function mutations of the EPAS1 gene (encoding HIF-2α), and an erythropoietin transcript is present in tumor tissues (Chaps. 32 and 57). Even though these tumors may be recurrent, they bear the same heterozygous mutations of the EPAS1 gene. However, these mutations are generally not found in nontumor tissues, so the etiology of the association of these tumors with polycythemia is not certain; it is also possible that they may be associated with postgonadal genetic mosaicism, wherein the EPAS1 mutation predisposes to tumor development.141,142,143 However, in one family the EPAS1 mutation was inherited and also associated with the development of recurrent pheochromocytomas/paragangliomas.116
Chapter 38 has additional discussion.
Aldosterone-producing adenomas,144 Bartter syndrome,145 and dermoid cyst of the ovary146 have been described in association with erythrocytosis. Erythropoietin levels were found to be elevated in the serum and returned to normal after extirpation of the tumors. A number of pathogenetic mechanisms have been suggested (Chaps. 32 and 38), including decreased plasma volume; mechanical interference with renal blood supply; hypertensive damage to renal parenchyma; functional interaction between aldosterone, renin, and erythropoietin; and inappropriate secretion of erythropoietin by the tumors. Mild polycythemia may be present in patients with Cushing syndrome, but its pathophysiologic basis is not entirely clear (Chap. 38).
The erythropoietic effect of androgens is of considerable practical importance.147 For many years, it was assumed that the higher red cell count in males was caused by androgens because the hemoglobin levels of boys and girls were identical up until the time of puberty. It was not until pharmacologic doses of testosterone were administered to women with carcinoma of the breast that the full erythropoietic potency of androgens was appreciated.148 Since then, various androgen preparations have been used in the treatment of refractory anemia, occasionally causing dramatic erythropoiesis, with hemoglobin values climbing into the polycythemic range (Fig. 57–5).
Erythropoietic response to testosterone derivatives in a patient with myelofibrosis. Hgb, hemoglobin; Hct, hematocrit.
The mechanism of androgen action on erythropoiesis appears to be complex, related both to their capacity to stimulate erythropoietin production149 and their capacity to induce differentiation of marrow stem cells directly.147 These two effects have specific structural requirements. Androgens with the 5α-H configuration stimulate renal and extrarenal erythropoietin production, whereas androgens with the 5β-H configuration enhance the differentiation of stem cells.149 Testosterone administration is associated with an increase in erythropoietin levels and a decrease in hepcidin levels.150 Although erythropoietin levels declined with continued testosterone administration, they remained inappropriately high despite improved hemoglobin levels, suggesting a new set point.150
Polycythemia at birth is a normal physiologic response to intrauterine hypoxia and to the high oxygen affinity of red cells containing very high proportions of hemoglobin F (Chap. 7). It may become excessive and even symptomatic, especially in infants of diabetic mothers, or if the clamping of the cord is delayed, permitting placental blood to boost the blood volume of the infant.151 Because it is difficult to recognize symptoms of hyperviscosity in the neonate, many pediatricians perform a partial plasma exchange transfusion if the venous hematocrit is above 65 percent at birth.152
One study of 25,000 neonates in Utah153 showed that the average hematocrit at birth would be considered “polycythemic” in adults, while 2 weeks later it has fallen to “anemic” levels. This dramatic decrease of red blood cells in neonates during their first days of life likely contributes to neonatal jaundice (Chap. 33).154
APPARENT (RELATIVE) POLYCYTHEMIA
Some believe that apparent polycythemia is merely a mild absolute polycythemia accentuated by a compensatory reduction in plasma volume. Others suggest that it is caused by a primary reduction in plasma volume and have associated it with hypertension, obesity, and stress. When the red cell mass is documented to be normal, spurious polycythemia is also an appropriate term. Its clinical significance has also been disputed. The high hematocrit with its associated high viscosity is believed by some to be a risk factor heralding cerebral and cardiac complications, while others believe it is merely a well-tolerated anomaly. Because the designation apparent polycythemia155 is noncommittal, it is used here.
The main clinical associations with apparent polycythemia are obesity, hypertension, and smoking. In obese patients, the finding of a normal red cell volume may be spurious because if the volume is expressed in terms of lean body weight, some of these patients would have a significant increase in red cell mass. In hypertensive patients, there is no adequate explanation for the apparent increase in red cell production or decrease in plasma volume. Sleep apnea (common in patients with congestive failure), excessive production of atrial natriuretic factor, increased adrenal activation, decreased aldosterone secretion, and hypoxic vasoconstriction are all factors that have been invoked,156,157,158 but with uncertainty. Chronic administration of diuretics to treat hypertension may be a more likely cause.158