Thrombopoietin measurement as a key component in the evaluation of pediatric thrombocytosis

Abstract

JAK2, MPL, and CALR mutations, which underlie essential thrombocythemia (ET) in most adults, are infrequent in children. Consequently, additional tests are needed to confirm pediatric ET diagnoses. We report a child with suspected ET and p JAK2, MPL, and CALR analyses. Serum thrombopoietin was markedly elevated, leading to analysis of the thrombopoietin gene, TPHO, which contains an upstream open reading frame (uORF) known to repress THPO translation. Sequencing revealed a de novo, germline stopgain mutation in the uORF, explaining the elevated thrombopoietin and thrombocytosis. This suggests that screening thrombopoietin levels and, if elevated, THPO 5’ UTR sequencing could be diagnostic.

Introduction

Essential thrombocythemia (ET) is a myeloproliferative disorder characterized by persistent thrombocytosis and increased megakaryocytes in the absence of a reactive or secondary cause or other myeloid neoplasia. Approximately 85 to 90% of adults with ET have clonal disease and an acquired JAK2, MPL, or CALR mutation.[1] Current algorithms for the diagnosis of ET include molecular genetic studies for a pathogenic variant in one of these genes.[2-4] ET is a rare disorder in childhood and, in striking contrast to adults, only 25 to 30% of children diagnosed with ET will have an acquired JAK2, MPL, or CALR mutation.[5,6] Thus, 70-75% of children are classified as having ‘triple negative’ disease. X chromosome inactivation studies in girls with triple negative disease argued against clonal disease.[5] However, targeted deep sequencing performed in two cohorts of children with ET revealed that 32 to 40% of those lacking JAK2, MPL, or CALR mutations had somatic changes in one or more other genes previously implicated in myeloproliferative neoplasms.[6,7] In these triple negative cases, any given other gene (e.g., ASXL1 and IRF8) was mutated in just a few patients. Thus, the diagnosis of ET in children remains challenging and the possibility remains that other genes may underlie the thrombocytosis.

Results

Our patient is a male who was referred at 16 months of age for persistent thrombocytosis, with platelet counts averaging between 1.2 and 1.6 × 10⁶/μL and otherwise p complete blood counts. Past medical history was remarkable for recurrent otitis and mild eczema. There was no family history of thrombocytosis and both parents had p platelet counts. On physical exam, he had mild hepatosplenomegaly (liver edge and spleen tip palpable 1 and 2 cms below the right and left costal margins, respectively), but no other findings. Laboratory studies for secondary causes were p. Bone marrow examination demonstrated a mildly hypercellular marrow (95%) with megakaryocytic hyperplasia (up to 22 megakaryocytes per high powered field), no dysplasia, and no fibrosis (Figure 1). Chromosomal analysis revealed a p male karyotype and a myelodysplastic syndrome fluorescence in situ hybridization panel was p, showing no evidence of −5/5q-, −7/7q-,trisomy 8 or 20q-. Targeted JAK2 V617 pyrosequencing, CALR amplification and sequencing, and MPL gene bidirectional sequencing did not detect a variant. Variants of unknown significance in genes associated with hematologic disease (LYST, FANCI, PRF1, PNPLA1, and CD36) were noted in the initial interpretation of clinical whole exome sequencing (Baylor Medical Genetics Laboratory), however, none were considered to be responsible for the patient’s thrombocytosis nor clinically significant.

Figure 1.

Bone marrow histopathologic evaluation showing mild bone marrow hypercellularity with markedly increased megakaryocytes and p reticulin staining. A) 10X and B) 20X magnification of bone marrow biopsy. C) Reticulin stain of bone marrow biopsy under 20X magnification. D) 20X magnification of bone marrow aspirate.

With a presumptive diagnosis of ET, the patient was managed over 2 years, during which time his platelet count ranged between 0.8 − 2.0 × 10⁶/μL off medication (mean and median 1.2 × 10⁶/μL) . The peripheral blood counts remained otherwise within p range.

To further explore the pathophysiology underlying this triple negative patient’s disease, a serum TPO measurement (Quest Diagnostics) was obtained to determine if thrombopoiesis was being driven by elevated TPO levels. The TPO concentration was extremely elevated at > 4000 pg/mL (reference range 7-99 pg/mL). This value was significantly higher than observed in ET or reactive thrombocytosis, in which the TPO concentrations are infrequently > 1,000 pg/mL,[8-13] and suggested the presence of a pathogenic variant driving TPO production.

Inherited pathogenic variants in the 5’ untranslated region (UTR) of THPO, which encodes TPO, have been reported in kinships with thrombocytosis in multiple family members, including children, with autosomal dominant transmission and elevated TPO levels.[14-17] The 5’ UTR of the various THPO mRNA isoforms contains 5-7 upstream open reading frames (uORFs), which regulate translation.[18] uORF₇, present in all of the THPO mRNAs, exerts the strongest translational inhibition, with its translation extending past the physiologic AUG (Figure 2A). Given the highly elevated TPO level in our patient, we hypothesized that he may have a pathogenic variant in THPO uORF₇. Polymerase chain reaction amplification of the THPO 5’ UTR was performed on peripheral blood DNA (forward primer 5’AATGTGAGAGAATTCAGGGCTT3’ and reverse primer 5’ TTTGGGAGAATGGGTTCCCC3’) followed by bidirectional sequencing and comparison to the THPO coding sequence (NM_001290028.1). This revealed a heterozygous pathogenic variant, THPO:c.−34C>T, resulting in a stopgain in uORF₇ (Figure 2B). Sequencing of buccal DNA indicated the variant was germline. The variant was absent in both parents, indicating that the mutation was de novo (Figure 2A). The variant was considered novel, based on its absence in dbSNP and in the Exome Aggregation Consortium database, which reports the variants identified in approximately 60,706 unrelated individuals subjected to exome sequencing as part of various disease-specific and population genetic studies, with efforts made to specifically exclude individuals with severe pediatric disease (http://exac.broadinstitute.org/about).[19]

Figure 2.

A de novo, germline, stopgain mutation within uORF₇ of THPO. A) Schematic of uORF₇ and translation of wild type THPO. Sequence chromatograms of the proband’s parents are shown below. B) Schematic of the truncated uORF₇ and effects on THPO translation. The sequencing chromatogram of the proband blood and buccal cell DNA is shown below. Asterisks indicate two previously identified mutations, c.−47delG and c.−31G>T.[14,15]

Re-evaluation of the prior clinical whole exome sequencing also revealed the variant, which had been initially excluded from interpretation given it was neither a previously reported pathogenic variant nor, given its noncoding position, predicted to affect splicing. Similar to previously reported inherited pathogenic variants in uORF₇,[14,15] the c.−34C>T variant increases the intercistronic distance between uORF₇ and the main ORF, and eliminates extension of uORF₇ translation over the physiologic AUG, allowing increased ribosomal initiation at the THPO start site and markedly enhanced TPO expression (Figure 2).

Discussion

Proper diagnosis and identification of the underlying cause of ET is critical for prognosis, treatment, and genetic counseling. This case is the first description of suspected ET in childhood caused by a de novo germline pathogenic variant in the 5’ UTR of THPO. It illustrates that THPO uORF₇ variants may underlie not just familial thrombocytosis, but also sporadic, chronic thrombocytosis masquerading as ET. Additionally, while similar mutations have been reported previously, the potential impact of novel mutations may be unappreciated during the analysis of whole exome sequencing data due to their location outside of the coding region of THPO. This report brings to light the fact that base substitutions in the 5’ UTR may be clinically significant even in the absence of family history, and that those who analyze clinical whole exome sequencing must consider them.

While studies have previously reported elevated TPO levels in patients with ET,[8,11-13] those levels were substantially lower than those seen in our patient and in previously reported patients with inherited THPO uORF₇ pathogenic variants.[14,15] Determination of TPO level is not included in current algorithms for the evaluation of persistent thrombocytosis or suspected ET. However, the discovery of a markedly elevated TPO level would direct further investigations, as in our case. Given the low proportion of JAK2, MPL, and CALR variants and clonality in children with suspected ET,[5] we propose that screening for an elevated TPO level be included in their initial evaluation. If found to be markedly elevated (e.g., >1000 pg/mL), evaluation for variants in the 5' UTR of THPO may prove diagnostic.

Abbreviations key

ET

essential thrombocythemia

TPO

thrombopoietin

uORF

upstream open reading frame

UTR

untranslated region