Page 193 - Haematologica May 2022

Letters to the Editor
Erythrocytosis associated with EPAS1(HIF2A), EGLN1(PHD2), VHL, EPOR or BPGM mutations: the Mayo Clinic experience
Germline mutations in the oxygen-sensing pathway (VHL-HIF2A-PHD2) or erythropoietin (EPO) signaling (EPOR) are relatively rare but may result in erythrocytosis with normal p50 measurement (oxygen tension at which hemoglobin is 50% saturated) accompanied by either an elevated or inappropriately normal EPO (VHL-HIF2A- PHD2) or subnormal EPO (EPOR).1 On the other hand, a left shift of the oxygen dissociation curve, with venous p50 <24 mmHg may result from high-oxygen affinity (HOA) hemoglobin variants, defective 2,3-bisphospho- glycerate mutase (BPGM) causing 2,3-BPG deficiency or methemoglobinemia.1 The incidence, clinical course and management of hereditary erythrocytosis has not been well-characterized due to its rare occurrence. In that regard, we recently reported on 41 patients with HOA variant associated erythrocytosis; over half of the patients manifested one or more symptoms thought to be related to increased hematocrit while thrombosis was documented in a quarter of the patients.2 Neither hemat- ocrit level nor active phlebotomy showed significant cor- relation with either thrombotic or non-thrombotic symp- toms, which might have resulted from the limited sample size.2 In a recent study which included 270 patients with idiopathic erythrocytosis, 1.1% harbored EPOR muta- tions, while pathogenic variants involving genes in the hypoxia pathway were identified in 23% of patients.3 Accordingly, we share the Mayo Clinic clinical and labo- ratory experience with hereditary erythrocytosis result- ing from genetic alterations in the oxygen-sensing path- way (VHL-HIF2A-PHD2), EPOR or BPGM.
All patients that underwent hereditary erythrocytosis evaluation at the Mayo Clinic over the last 10 years (2012-2021), were retrospectively recruited after obtain- ing Institutional Review Board approval. Polycythemia vera was excluded with JAK2 exon 12-15 sequencing. Hereditary erythrocytosis testing was pursued at the Mayo Clinic laboratory utilizing an algorithmic approach which included p50 measurement, serum EPO level (Epo), and DNA sequencing by polymerase chain reac- tion (PCR) of EPOR (exon 8), hypoxia-inducible factor 2 a (HIF2A) encoded by endothelial PASS domain protein 1 (EPAS1) (exons 9 and 12), prolyl hydroxylase 2(PHD2) encoded by EGL-9 family hypoxia inducible factor 1(EGLN1)(exons 1-5), von Hippel Lindau (VHL) (three coding exons and intron/exon boundaries) and BPGM (exons 1-4) as detailed in our prior work.4
Of 592 patients tested at the Mayo Clinic for HIF2A/PHD2/EPOR alterations, 14 pathogenic variants were identified in HIF2A (n=6, 1%), PHD2 (n=3, 0.5%), EPOR (n=2, 0.3%), while two of 421 (0.5%) and one of 446 (0.2%) patients harbored BPGM and VHL variants, respectively. In addition, 22 variants of uncertain signifi- cance (VUS) were reported; EPOR (n=1), HIF2A (n=3), PHD2 (n=10), BPGM (n=2), VHL (n=6), resulting in com- bined (pathogenic + VUS) Mayo Clinic incidence rates of 0.5%, 1.5%, 2.2%, 1% and 1.6% for EPOR, HIF2A, PHD2, BPGM, and VHL aberrations, respectively.
Table 1 summarizes oxygen-sensing pathway (PHD2/HIF2A/VHL) pathogenic variants including clini- cal course of ten patients with median follow-up of 2 years, (range, 0.2-10 years). HIF2A pathogenic variants were noted in six patients; four harbored the heterozy- gous HIF2A c.1121T>A, p.(Phe374Tyr) alteration in exon 9, previously reported in association with neuroen-
docrine tumors with or without erythrocytosis.5 A 57- year-old male with heterozygous HIF2A c.1121T>A mutation presented with a hemoglobin (Hb)/hematocrit (Hct)/Epo of 17.9 g/dL/54.4%/93.4 mIU/mL, diabetes mellitus and prior cerebrovascular accident (CVA)/ left ventricular thrombus, was started on phlebotomy, con- tinued aspirin with anticoagulation and did not experi- ence additional thromboses. The second case was a 56- year-old female with heterozygous HIF2A c.1121T>A mutation. Hb/Hct/Epo at presentation; 19.1 g/dL/57%/40.8 mIU/mL, with hypertension and hyper- lipidemia, developed multiple thromboses; myocardial infarction, followed by CVA, inferior vena caval throm- bus post-diagnosis, the latter occurred despite ongoing phlebotomy and aspirin/clopidogrel. The remainder two patients with heterozygous HIF2A c.1121T>A mutations were 68- and 71-year-old males with hypertension and hyperlipidemia respectively, Hb/Hct/Epo at diagnosis were 19.1/57.2/20.7 and 17.2/52/7.7, both did not expe- rience thrombosis with the former receiving phlebotomy and the latter low dose aspirin.
Additionally, a 61-year-old female harbored a het- erozygous missense alteration in HIF2A c.1620C>A, resulting in amino acid substitution p.Phe540Leu (F540L) previously reported by our group.4 She had a history of hypertension, presented with Hb/Hct/Epo of 16.1/47.8/7.3 and did not experience thrombosis while on low-dose aspirin. On the other hand, a 69-year-old hypertensive male with heterozygous HIF2A c.1609G>A, mutation with Hb/Hct/Epo of 23/58.7/175 at diagnosis, developed a CVA with ongoing phlebotomy. An elevated Epo level (range, 20.7-175, reference range; 2.6-18.5 mIU/mL) was noted in four of six patients with HIF2A pathogenic variants, which in all instances was accompa- nied by phlebotomy. All patients had one or more cardio- vascular risk factors, with three patients (50%) experienc- ing thrombosis, two of which occurred with ongoing phlebotomy, suggesting the lack of benefit of phleboto- my.
Of three patients with PHD2 pathogenic variants; a 35- year-old female with family history of erythrocytosis, current smoker, without history of prior thrombosis, and Hb/Hct/Epo 17.2/52.6/11.2, demonstrated a PHD2 c.1111C>T, p.(Arg371Cys) missense variant. This variant has been reported in the human gene mutation database,6,7 and involves a highly conserved amino acid in the Fe(2+) 2-oxoglutarate dioxygenase domain, critical for hydroxy- lation of HIF; functional studies have not been performed but studies involving (Arg371His) have shown decreased ability of PHD2 to bind and hydroxylate HIF. On the other hand, two patients harbored previously reported PHD2 c.461C>A, p.(S154*) and c.1030C>T, p.(Arg344*) nonsense variants predicted to result in a premature stop codon in exon 1 and 3, respectively, and expected to be loss of function mutations.4 This included a 67-year-old male with PHD2 c.461C>A and a 60-year-old female with PHD2 c.1030C>T mutation, Hb/Hct/Epo at diagno- sis were 17.8/50.7/10.3 and 17/not available/30, both did not experience thrombosis, former had known coronary artery disease and was on low-dose aspirin while the lat- ter was hypertensive and receiving phlebotomy along with aspirin.
A pathogenic variant in VHL was detected in a 19-year- old male, compound heterozygous (L188V and R200W) for the previously described VHL mutations,8 who pre- sented with erythrocytosis (Hb/Hct 19/57) and a marked- ly elevated EPO level at 1465 mIU/mL. He was managed with phlebotomy every 4 weeks, in addition to aspirin and did not experience thrombosis.
haematologica | 2022; 107(5)
1201