Diagnostic workflow for hereditary erythrocytosis and thrombocytosis

A 27-year-old white male was referred to hematology complaining of fatigue and having been found to have an abnormal blood count. He had no previous medical history of note. No family history of a hematological disorder was elicited. He was a nonsmoker and consumed 16 units of alcohol per week.

On examination he was overweight but had no abnormal findings. Blood count showed hemoglobin (Hb) 193 g/L, hematocrit (HCT) 0.59 L/L, mean cell volume 89 fL, mean cell Hb 29.9 pg, mean cell Hb concentration 333 g/L, white blood cells 6.5 × 10⁹/L with normal differential count, and platelet count 167 × 10⁹/L. Further initial investigations were a serum erythropoietin (EPO) 15.9 mIU/mL (normal range, 2.5-10.5), negative for the JAK2 V617F mutation, red cell mass 146% of predicted, and Hb electrophoresis normal.

Since Dameshek introduced the concepts of myeloproliferative disease in the 1950s, definitions of these diseases have been available starting with the definitions of polycythemia vera (PV) and essential thrombocythemia (ET) used in the original Polycythemia Vera Study Group trials. These definitions suggest limits for the Hb and HCT above which there is, by implication, an erythrocytosis. Similarly, limits for platelet counts above the defined limit suggest that there is a thrombocytosis.

An absolute erythrocytosis is present when the red cell mass is greater than 125% of predicted (if this test is available), and it can also be assumed that if the HCT is >0.60 in a man and >0.56 in a woman, then there is an absolute erythrocytosis¹ ; however, anyone with an Hb/HCT above the limits for the definitions of PV may have an erythrocytosis and needs further investigation. Similarly, for platelet counts in the range of ET, further investigation may be indicated.

Some family history of myeloproliferative neoplasm (MPN) is often revealed by patients, and this needs to be assessed in the context of an abnormal blood count. This review will focus on the hereditary causes of myeloproliferative diseases (diseases rather than neoplasms, because these are germline alterations rather than acquired clonal neoplasms), including erythrocytosis and thrombocytosis, and then explore the diagnostic pathways for these disorders.

In sporadic cases of MPNs, with careful inquiry 7% to 8% have other family members with an MPN.²  At the population level, a number of germline predisposition alleles have been identified, such as the JAK2 46/1 haplotype.³  In the Icelandic population and in an Italian cohort, the germline sequence rs2736100 in the TERT gene associates with MPN.^4,5  Multiple germline variants have been described in MPN cohorts in which the variants predispose to MPN,⁶  and, recently, predisposition alleles were shown to be associated with MPNs and age-related clonal hematopoiesis.⁷ 

There have been several extended families described in whom the germline variant leading to the genetic predisposition is identified. In these cases of familial clustering of MPNs, the acquired MPN is indistinguishable from sporadic MPN, with a variety of MPNs and acquired driver mutations in the same family. In 4 families from the same geographical location, a germline copy number variation, leading to germline duplication of ATG2B and GSKIP, predisposes to MPN and progression to myeloid malignancy.⁸  The RBBP6 gene has been shown to be the candidate for genetic predisposition to MPN in 1 extended family.⁹ 

In the 2017 revision, the World Health Organization (WHO) published revised criteria for the definition of PV; an Hb > 165 g/L (men) and >160 g/L (women) or HCT > 0.49 (men) and >0.48 L/L (women) or other evidence of increased red cell mass, JAK2 mutation, and a bone marrow biopsy showing panmyelosis are the major criteria for a diagnosis of PV.¹⁰  These definitions include limits that have changed over time. The current definition uses lower levels of Hb and HCT than previous WHO PV criteria and is based on retrospective studies that suggest that these limits can discriminate between PV and ET.¹¹  It also assumes that it is possible to reliably discriminate between PV and ET on the basis of the morphological appearance of the bone marrow. The British Society for Haematology has reviewed the evidence and suggested that criteria for the diagnosis of PV should be high HCT > 0.52 (men) and >0.48 (women) or an increased red cell mass (>25% above predicted) and a mutation in JAK2, with more detailed criteria for the very rare JAK2 mutation negative cases.¹²  These criteria are likely to be more practical for identifying patients with PV, without flagging too many of those with Hbs within the normal range as being in need of investigation. However, what all of these criteria identify is patients with clonal disease and erythroid proliferation who have the acquired disease, PV. If an erythroid proliferation/erythrocytosis is identified, not an acquired clonal disorder, other causes for the erythrocytosis may need to be investigated.

An erythrocytosis can be primary (an intrinsic disorder in the erythroid progenitor cells) or secondary (an external cytokine EPO drives increased red cell production). Primary and secondary causes can be further divided into hereditary (or congenital) and acquired.

The main and predominant cause of primary acquired erythrocytosis is PV; usually there is an acquired JAK2 clone driving red cell production. Primary hereditary causes are very rare, with the main one of note being mutations in the erythropoietin receptor gene (EpoR).

There are many acquired secondary causes of erythrocytosis. These can be hypoxia driven, where hypoxia leads to increased EPO and a drive to red cell production. Such hypoxia can be a central process, such as chronic lung disease or right-to left cardiopulmonary shunts, or a local renal hypoxia, such as renal artery stenosis. Secondary acquired erythrocytosis can also be due to pathological EPO production. A variety of tumors have been described where the tumor is a source of EPO, including cerebellar hemangioblastoma, renal cell carcinoma, hepatocellular carcinoma, and uterine leiomyoma. Exogenous administration of EPO and related substances, such as androgens, can also lead to a secondary acquired erythrocytosis. Many other factors can lead to acquired secondary erythrocytosis, such as living at high altitude, obstructive sleep apnea, smoking, and other behaviors. However, secondary hereditary erythrocytosis can be caused by a number of different defects, including mutations in the genes in the oxygen sensing pathway and high oxygen-affinity Hbs.

There are a number of possible hereditary causes of primary or secondary erythrocytosis (Table 1).

EPO links to a receptor on the cell surface, the erythropoietin receptor (EpoR). The proteins JAK2 and STAT5 then autophosphorylate, and STAT5 dimerizes, translocates to the nucleus, and triggers downstream signaling and production of red cells. Then the process is turned off when another protein, SHP-1, attaches to and downmodulates the receptor. However, mutations occur in the EpoR gene, leading to a truncated protein receptor that has lost the SHP-1 docking site. Thus, when EPO attaches to the mutated receptor, it is switched on but cannot be switched off; therefore, it continues to drive red cell production without further EPO stimulation.¹³  Therefore, with low EPO levels and switched on, mutated receptor increased red cell production results. At least 11 mutations in EpoR have been described with erythrocytosis.¹⁴ 

The SH2B3 gene encodes for the SH2B3 protein, also known as lymphocyte adaptor protein (LNK), which is involved in cell signaling and is a negative regulator of cytokine signaling by attenuating JAK activation. Mutations have been described in SH2B3 in MPNs. These mutations result in a defective LNK protein that does not act as a negative regulator of the JAK/STAT pathway downstream of the cytokines and, thus, lead to increased erythropoiesis, a primary erythrocytosis (with an associated low EPO level). In several cases the mutation was shown to be in the germline; therefore, germline LNK mutations could be a cause of hereditary erythrocytosis.¹⁵ 

The human organism has a sensitive mechanism for sensing oxygen levels and responding to hypoxia. In conditions of normal oxygen levels, the prolyl hydroxylases (PHDs) hydroxylate hypoxia-inducible factor (HIF) and bind the von Hippel–Lindau tumor suppressor protein (VHL). Ubiquitination and degradation of HIF then occurs in the proteasome and, thus, low HIF levels are maintained.

In hypoxia, less hydroxylation occurs, and HIF escapes VHL-mediated degradation. Levels of HIF then rise, and the protein moves to the nucleus and binds to the hypoxia response element in the 3′ region of the target genes. This leads to HIF-regulated transcription and production of a number of proteins, including EPO. Therefore, mutations of the genes in this pathway can lead to failure of HIF breakdown and increased EPO drive, and patients have elevated or normal EPO levels in the presence of an erythrocytosis.

The first defects in the oxygen-sensing pathway were discovered in the VHL gene. A homozygous mutation in the VHL gene C598T was identified in a large cohort of individuals with erythrocytosis in the remote upper Volga region of Russia, Chuvashia.¹⁶  The mutant protein was shown to have reduced activity as a negative regulator of HIF-1–dependent gene transcription and increased expression of HIF-1–regulated genes target genes, including the erythropoietin (EPO) gene. The homozygous mutation in the VHL gene has been identified in patients with congenital erythrocytosis from other areas, and a few compound heterozygotes of the VHL gene with erythrocytosis have been described. Recently, new VHL exon and complex splicing alterations have been described in some cases of hereditary erythrocytosis.¹⁷ 

An EGLN1 (PHD2) gene was discovered in a family with erythrocytosis, a change on one allele, C950G, leading to a protein alteration of proline to arginine at codon 317. Affected siblings had erythrocytosis with normal or increased EPO levels.¹⁸  In vitro studies showed that the mutation had abnormal activity that would cause erythrocytosis, and a mouse model provided further evidence of the mutation as a cause of erythrocytosis.¹⁹  A number of additional mutations in PHD2 have been documented in individuals with erythrocytosis.¹⁴  In 1 case, an individual was found to have a paraganglioma 13 years after presentation.²⁰  The mutation was also found in the tumor tissue, as well as absence of the wild-type EGLN1 allele, thus resulting in loss of heterozygosity. This suggests, in this case, that EGLN1 was acting as a tumor-suppressor gene.

An EPAS1 (HIF2A) gain-of-function mutation was identified in 3 generations of a family associated with erythrocytosis: a G1609T change leading to a change at codon 537 from glycine to tryptophan. In vitro studies showed that the altered protein bound PHD2 and VHL differently than did the wild-type protein, degraded it more slowly, and induced downstream genes.²¹  Other mutations in EPAS1 associated with congenital erythrocytosis have been identified in other kindreds.

Thus, a number of genes in the oxygen-sensing pathways have been shown to be mutated and cause hereditary erythrocytosis.¹⁴  The clinical presentation of these cases is not clear. Some features have been described, such as early development of varicose veins in Chuvash polycythemia. Early severe thrombotic events and emerging pulmonary hypertension are noted with other defects. Such clinical signs may alert to the possibility of an oxygen-sensing pathway mutation, but there is no clear phenotype. Recently, a gain-of-function mutation in the EPO gene itself, leading to a familial disorder with high EPO levels, has been described.²² 

Oxygen is transported to the tissues bound to Hb. The oxygenation and deoxygenation of Hb occurs at the heme iron binding site, and the affinity for oxygen depends on the nature of the Hb. The Hb oxygen dissociation curve describes this relationship. A high oxygen-affinity Hb has a left-shifted oxygen-dissociation curve because oxygen is tightly bound. At the tissue level, this results in relative hypoxia, EPO production, and secondary erythrocytosis. Approximately 100 high oxygen-affinity variants have been described, with both α and β globin gene mutations resulting in stable and unstable Hbs. These have an autosomal-dominant inheritance.²³ 

One percent of Hb is normally in the methemoglobin form, which impairs oxygen binding and transport. The presence of a large amount of methemoglobin leads to cyanosis, and a compensatory erythrocytosis develops. Congenital methemoglobinemia can arise because of a deficiency in cytochrome b₅ reductase or an abnormal M Hb. NADH–cytochrome b₅ reductase catalyzes electron transfer from NADH to cytochrome b₅ and is encoded by the CYB5R3 gene. More than 40 mutations of this gene have been described, and inheritance is autosomal recessive. Type I mutations lead to a defect in the erythrocytes only, whereas type II mutations have accompanying neurological defects.²⁴ 

Binding of 2,3 bisphosphoglycerate (2,3-BPG) to Hb converts the Hb molecule to a low oxygen-affinity state, shifting the oxygen-affinity curve to the right. Therefore, deficiency of 2,3-BPG moves the oxygen dissociation curve to the left, less oxygen is delivered to tissues, and a compensatory erythrocytosis results. In the glycolytic pathway, the production of 2,3-BPG involves the conversion of 1,3-BPG to 2,3-BPG catalyzed by bisphosphoglycerate mutase (BPGM). Mutations in the BPGM gene lead to an abnormal functioning of BPGM and deficiency of 2,3-BPG.²⁵ 

Families have been identified with mutations in the SLC30A10 gene who have the syndrome of hepatic cirrhosis dystonia, erythrocytosis, and hypermanganesemia. Manganese induces EPO gene expression, and increased EPO levels are seen in these patients, causing the erythrocytosis.²⁶  Erythrocytosis has also been reported in families who have been described with increased adenosine triphosphate levels associated with low 2,3-BPG levels with autosomal-dominant inheritance.²⁷ 

The WHO has recently revised the criteria for ET. The major criteria are platelet count ≥ 450 × 10⁹/L, a proliferative bone marrow appearance with megakaryocyte predominance, not meeting the criteria for other myeloid neoplasms, and presence of a JAK2, CALR, or MPL mutation; the minor criteria are presence of a clonal marker or no evidence of a reactive thrombocytosis. All 4 major or 3 major and 1 minor criteria are required to make the diagnosis, showing evidence of an acquired clonal neoplasm. The lower limit of the platelet count is unchanged from previous WHO criteria, and in patients with an increased platelet count > 450 × 10⁹/L who do not fulfill the criteria may have other than an acquired clonal neoplasm.¹⁰ 

A platelet count above the normal range can be caused by a primary disorder arising in the bone marrow compartment. It can also be secondary or reactive, and the result can be spurious if other cellular factors, such as microspherocytes, schistocytes, or bacteria, are being mistakenly counted as platelets. Numerous reactive secondary causes can lead to an elevated platelet count, including infection and inflammation, postoperatively or following other causes of tissue damage, and hyposplenism. The platelet count is often elevated, with hemorrhage, iron deficiency, malignancy, and hemolysis, and it can rebound following myelosuppressive therapy; the presence of any such condition needs to be considered in patients presenting with an elevated platelet count.

Primary acquired thrombocytosis is seen with the acquired clonal disorder ET, but it is also seen in other myeloid malignancies, including PV, primary myelofibrosis and prefibrotic myelofibrosis, myelodysplasia with isolated del5q and myelodysplastic/MPN with ring sideroblasts and thrombocytosis, chronic myeloid leukemia, and myelodysplastic/MPN unclassified. However, there are rare individuals with a primary thrombocytosis who may have a hereditary or congenital germline cause for the disease.

Rare families have been described with a clearly inherited thrombocytosis (Table 2). The thrombopoietin (THPO) gene was isolated in 1994; following this, a number of alterations in this gene were discovered in families with autosomal-dominant hereditary thrombocytosis. These gene alterations result in translational inhibition of THPO messenger RNA and elevated thrombopoietin (TPO) levels in serum, causing thrombocytosis. Affected family members in these kindreds had elevated TPO levels.²⁸  Associated distal limb defects are described, with some of these mutations indicating a role for TPO in vasculogenesis.²⁹ 

The myeloproliferative leukemia virus oncogene (MPL) codes for the TPO receptor MPL. A polymorphism in MPL, MPL Baltimore, was found to be associated with mild thrombocytosis in African American women who were heterozygous for the polymorphism, with extreme thrombocytosis in homozygotes. This polymorphism is restricted to individuals of African American descent.³⁰  Families have been described with a point mutation in the transmembrane domain of the MPL gene with an autosomal-dominant inheritance. This mutation activates intracellular signaling and cell survival. This has been found in a number of Italian children with hereditary thrombocytosis; follow-up of this cohort demonstrated a large number of major thrombotic events and, with aging, development of splenomegaly and bone marrow fibrosis.³¹  In Arab families, a point mutation in MPL (P106L) is causative for hereditary thrombocytosis with high TPO and platelets levels in homozygotes and mild thrombocytosis in heterozygotes.³²  Another nearby point mutation (R102P) is also associated with thrombocytosis in heterozygotes (homozygotes have congenital amegakaryocytic thrombocytopenia). It is thought that, in the heterozygotes, subnormal cell surface expression of wild-type MPL in platelets induces defective TPO clearance.³³ 

The JAK2 acquired point mutation V617F is associated with PV and ET, and other acquired mutations in JAK2 exon 12 are also seen in PV. However, in recent years, other germline mutations in JAK2 have been associated with hereditary thrombocytosis (Table 2) with autosomal-dominant inheritance. JAK2V617I was found in a family with high-penetrance hereditary thrombocytosis. The mutant is shown to induce cytokine hyperresponsiveness that would produce the phenotype.³⁴  Residues other than 617 mutations have been described at JAK2: R564Q in a family³⁵  and H608N.³⁶  In other families, additional germline JAK2 mutations have been found and demonstrated to be active in altering signaling R867Q and in 1 family 2 JAK2 mutations were located in cis in the pseudokinase and kinase domains. These mutants have altered constitutive signaling, leading to growth factor independence and hypersensitivity to TPO.³⁷ 

Of note, a patient with 2 germline mutations (E846D and R1063H), 1 inherited from each parent, is described with erythrocytosis and atypical megakaryocytes but a normal platelet count, suggesting that various germline JAK2 mutations have roles in MPNs.³⁸ 

The patient referred with an elevated blood count may have an obvious diagnosis of sporadic MPN with an identified driver mutation. However, familial MPN, reactive and secondary causes, and then hereditary reasons need to be considered. In the referred patient, it is first necessary to take a careful history from the patient, with particular care taken to inquire about family history of likely MPN and vascular disease. Other factors that may suggest reasons for the elevated blood count, such as drug use, including recreational use, and lifestyle factors, should be considered. A confirmatory blood count after a suitable interval from the original one is required to confirm a sustained elevated count.

Initial investigations would then include a screen for mutations in JAK2 and proceeding to CALR and MPL, if appropriate. Serum EPO will exhibit a low, an increased, or an inappropriately normal level, suggesting either a primary or secondary erythrocytosis. A bone marrow examination with cytogenetic examination should be considered to determine whether any MPN is present and to confirm any such diagnosis. This should be undertaken in any patient for whom the explanation for the abnormal blood count is not obvious. Other initial investigations could include iron studies to consider reactive causes, such as iron deficiency, and C-reactive protein as a marker for inflammation (Table 3; Stage 1 investigations)

Having carried out stage 1 investigations and considered the possibility of a hereditary erythrocytosis or thrombocytosis, further investigation is directed at identifying the hereditary cause and ruling out any alternative explanation, if appropriate. With an erythrocytosis, a red cell mass is a useful way of confirming its presence, although this is becoming an increasingly difficult test to obtain. Imaging, including renal ultrasound, computed tomography scan of the abdomen, and neuroimaging, may be indicated to search for lesions. A venous P₅₀ measurement (the partial pressure of oxygen required to achieve 50% Hb saturation) will show whether the oxygen-dissociation curve is shifted, and Hb electrophoresis may detect if an abnormal Hb is present. Specific sequencing of genes of the oxygen-sensing pathway can be carried out for known mutations. However, whole-genome sequencing,³⁹  if available, will find mutations, and next-generation sequencing (NGS) panels for testing genes that can cause hereditary erythrocytosis, including the Hb genes, are becoming available, and it is likely that such NGS panels will replace some of the other tests.

With a possible hereditary thrombocytosis, after eliminating possible reactive causes, specific tests could include a TPO level, because this is elevated with some of the hereditary causes and, if found, is useful confirmation. Full sequencing of THPO, MPL, and JAK2 is required to search for the described defects to confirm a hereditary cause (Table 3; Stage 2 investigations).

Careful history taking in individuals with sporadic MPN suggests that a number will have a family history of MPN as suggested by ≥2 cases of MPN in a family tree. It would not be routine practice to investigate the genome to identify the genetic predisposition; currently, this would only be done as a research investigation. However, identification of such family cases may be useful, because it is suggested that a blood count be performed in healthy relatives to identify MPNs at an early stage.

At the end of investigations, there remain patients who have an erythrocytosis for which no cause can be identified; these patients are categorized as having idiopathic erythrocytosis. The number of such patients has diminished with time as new reasons for an erythrocytosis are discovered and new investigations, such as NGS, become available. Any new discoveries should be explored in patients with idiopathic erythrocytosis. These patients need to be followed up long term so that they can be investigated further, and long-term outcomes can be established.

Initial investigation of the patient demonstrated a clear erythrocytosis with a raised EPO level, suggesting a secondary erythrocytosis. Further testing did not reveal any mutations in the oxygen-sensing genes, which were investigated as mutations in individual genes were described. A P₅₀ of 23.91 mm Hg (normal, 27-33) was found, suggesting a left-shifted oxygen-dissociation curve. However, as part of a whole-genome sequencing research project, DNA sequence analysis detected a heterozygous G>A substitution at nucleotide c.269 in the patient’s BPGM gene. This change was also detected in his mother, but not in his father. An assay for 2,3-BPG showed lower levels in the patient and his mother compared with controls, confirming the functional effect of the mutation. The mother did not have an elevated Hb; however, it was 155 g/L, which is at the upper end of the normal range for a female.⁴⁰  This patient with erythrocytosis was found to have a causative inherited mutation in the BPGM gene.

A young patient and/or a possible family history of MPN might suggest a hereditary cause for an increased blood count. This should be investigated by specific genetic investigations after other initial tests to rule out the more common acquired disorders. Upon completion of the diagnostic pathway, there remain individuals in whom a hereditary disorder is likely but in whom no abnormality can be identified. These individuals are subjects for future research.

Mary Frances McMullin, Centre for Medical Education, Whitla Medical Building Queen’s University Belfast, Lisburn Rd, Belfast BT9 7AB, United Kingdom; e-mail m.mcmullin@qub.ac.uk.