Functional analysis of germline ETV6 W380R mutation causing inherited thrombocytopenia and secondary acute lymphoblastic leukemia or essential thrombocythemia.

Germline mutations in ETV6 gene cause inherited thrombocytopenia with leukemia predisposition. Here, we report on functional validation of ETV6 W380R mutation segregating with thrombocytopenia in a family where two family members also suffered from acute lymphoblastic leukemia (ALL) or essential thrombocythemia (ET). In-silico analysis predicted impaired DNA binding due to W380R mutation. Functional analysis showed that this mutation prevents the ETV6 protein from localizing into the cell nucleus and impairs the transcriptional repression activity of ETV6. Based on the germline ETV6 mutation, ET probably started with somatic JAK2 V617F mutation, whereas ALL could be caused by diverse mechanisms: high-hyperdiploidity; somatic deletion of exon 1 IKZF1 gene; or somatic mutations of other genes found by exome sequencing of the ALL sample taken at the diagnosis.

Keywords: Acute lymphoblastic leukemia; myeloproliferative neoplasm; second hit

ETV6-related thrombocytopenia (ETV6-RT) is an autosomal dominant inherited thrombocytopenia (IT) [[1]]. Early leukemic transformation is a major risk of this IT; 2/3 of transformations accounting for B acute lymphoblastic leukemia (ALL) [[8]].

The ETV6 gene encodes an E26 transformation-specific (Ets) family transcriptional repressor. It can bind DNA via a highly conserved Ets DNA-binding consensus site located at the C-terminus. ETV6 plays a critical role in hematopoiesis, embryonic development, megakaryocyte development, and in leukemogenesis. ETV6 mutations prevent the protein from localizing in the nucleus with a significant reduction of the repression activity of the transcriptional factor [[3], [6], [9]]. The mechanisms of ETV6-RT leukemogenesis are still poorly understood.

For methods see supplementary file.

A functional study of novel ETV6 W380R mutation (NM_001987:exon6:c.1138T > A:p.W380R) found in Czech family with highly penetrant autosomal dominant pattern of IT has been performed. This family has been mentioned in Haematologica [[2]] and pedigree is shown in supplemental Figure S1. The proband developed high-hyperdiploid B-ALL at the age 15 and his father developed JAK2 positive (NM_001322195:exon13:c.1849G > T:p.V617F) essential thrombocythemia (ET) at the age 37. The proband achieved complete remission after PBSC transplantation from an unrelated donor at the age 19, since then is in complete remission.

In the X-ray structure of ETV6, W380 creates direct H-bond to DNA backbone (phosphate O2P on thymine 7). This contact was stable during the wt simulation.

In the mutant simulation, the tryptophan 380 was replaced by arginine. At the beginning of the simulation, formation of temporary H-bond between NH2 group of the arginine with O2P on thymine 7 was observed. However, after 430 ns this contact was disrupted and was not restored again (Figure S2a). Similar behavior was observed in the control mutant simulation (contact was disrupted after 418 ns; Figure S2b). This indicates that W380R mutation most probably impairs DNA binding.

Western blot results show presence of ETV6 mutant proteins (including W380R) predominantly in cytoplasm, whereas wt ETV6 localizes into the nucleus.

Indeed, ETV6 protein with W380R variant showed 5 times less transcriptional repressive activity compared to the wt ETV6 protein.

Results of ETV6 functional analyses are summarized in Figure 1.

PHOTO (COLOR): Figure 1. Results of ETV6 functional analyses. (a) ETV6-W380 mutation impairs its localization to the nucleus. HEK293 cells were transfected with indicated constructs and nuclear and cytosolic fractions were extracted 72 h later. Western blotting with FLAG antibody shows a predominant mislocalization of ETV6 mutants W380R, R418G and 385_418del into the cytoplasm. GAPDH and Lamin B were used as the markers of cytoplasmic and nuclear fraction, respectively. (b) ETV6-W380 mutation impairs transcriptional repression activity of ETV6. HEK293 cells were transfected with indicated constructs and the repressive activity upon ETV6-responsive reporter plasmid was determined by a dual luciferase reporter assay. Luciferase activity is normalized to internal control of Renilla luciferase and plotted as the fold repression relative to the empty vector control (set as 1). Representative experiment from 4 biological replicates is shown (mean+s.e.m.). *** P < 0,001; **** P < 0,0001, one-way ANOVA with multiple comparison testing. (c) Aberrant localization of ETV6-W380 mutant into the cytoplasm. HeLa cells transfected with indicated FLAG-tagged constructs were immunostained with FLAG antibody and DAPI 72 h post transfection. Representative images are shown for each construct. (d) Molecular dynamics simulation of ETV6 variant effect. Left: X-ray complex of wt protein – DNA. Contact between W380 and thymine 7 is shown. Middle: MD structure at the beginning of the mutant simulation, contact between R380 and thymine 7 is shown. Right: MD structure at the end of mutant simulation where R380 – thymine 7 contact is disrupted

In patient with ET, JAK2 mutation p.V617F was identified (Figure S3). No other defect suspected to initiate ET has been found.

In proband with ALL, cytogenetics of ALL cells revealed high-hyperdiploid (HeH) karyotype (XY,+X,+4,+5,+6,+10,+11,+12,+14,+17,+18,+21) and partial gain of chromosome 8. MLPA revealed deletion in exon 1 of the IKZF1 gene (Figure S4). Array-CGH results were consistent with the HeH-ALL karyotype and in the IKZF1 showed deletion between g.50 332 244 and g.50 349 149 according to Build 37 (hg19, Feb 2009) reference.

WES of ALL cells collected at the time of diagnosis and disease relapse showed nine somatic gene variants that were not present in the germline DNA (Table I); all of them were confirmed by Sanger sequencing. None of them was detected in ClinVar and HGMD databases. Only DPYSL2 and ARHGAP42 variants were mentioned in Cosmic database in the connection with other malignity types than ALL. Variants of ARHGAP42 and ERBB3 were mentioned in the GnomAD database showing low or unknown frequency in non-Finns European population, respectively.

Table I. Somatic gene variants of unknown significance – potential second hits for ALL onset in the proband. Variants with more than 20 reads found in acute lymphoblastic leukemia blasts at the ALL diagnosis time in more than 10% and not present in the germline DNA in more than 5% were chosen. Kyoto Encyclopedia of Genes and Genomes (KEGG signal pathway) database shows which pathways and associated functions are likely to be encoded in the genome by the gene. Their allele frequency in the NFE (non-Finns European population) was below 0.003% or unknown

A substantial number of patients carrying ETV6 germline mutation develop hematological malignancies. The risk of leukemic transformation is estimated to be up to 25–40%, the age of onset is highly variable (8–82 years). The spectrum of malignancies involves ALL, myeloid malignancies including myeloproliferative disorders or multiple myeloma. Targeted sequencing of a large cohort of childhood ALL patients revealed 31 leukemia-associated ETV6 exonic variants [[5], [10]].

The ETV6 variant p.W380R is present in the HGMD 2020.1 database [[11]], but is not yet listed in the dbSNP [[12]], and GnomAD v2.1.1 [[13]]. Results of our p.W380R in silico and functional analyses showed that this variant reduces DNA binding and affects nuclear localization of the protein causing aberrant cytoplasmic localization and decreased repression of transcription in a dominant-negative effect. It therefore causes a similar defect to the already known pathogenic variants of ETV6 [[7]].

As for the second hit leading to hematological malignancy, somatic mutation of JAK2 seems to be the likely hit in the case of ET. Expression of JAK2 V617F mutation in single hematopoietic stem cell is known to initiate myeloproliferative neoplasms [[14]]. This mutation with the germline ETV6 variant is thus the most probable driver event leading to the ET manifestation.

The situation is more complicated in the case of ALL. Studies regarding germline ETV6 related ALL did not reveal mutations in the remaining wild-type ETV6 allele in most cases. Acquisition of somatic defects in other genes, such as RUNX1, BCOR, and KRAS, is more prominent. However, the exact role of additional mutations in malignant transformation remains to be determined [[14]]. None of 10 somatic gene variants found in ALL blasts of our case has been described previously as a possible second hit for ALL; however, they could also play a role in the leukemogenesis.

In our ALL proband, a HeH karyotype was found. Analyses of clonal stability in samples taken at diagnosis and relapse and the backtracking of the HeH clone to birth indicate that HeH is an early and driving genetic event [[15]]. Chromosomal gains could be thus a driver event of ALL onset in our proband.

Moreover, molecular analysis showed possible IKZF1 inactivation by deletion of the noncoding region including exon 1. IKZF1 deletion predicts poor prognosis in children with B-ALL [[17]]. IKZF1 is also target by somatic alterations in ALL, suggesting that both somatic and inherited variations cooperate in the ALL pathogenesis [[18]].

Identification of genetic variants does not necessarily confirm a diagnosis. The variant may be novel, specific for the family, thus of uncertain pathogenic and clinical significance. In these cases, functional studies aimed at determining the effect of variants on protein function, are useful in the diagnostic process. However, it is a very complex analysis that is not easy to put into routine practice. It is clear that a second hit may not be just one [[19]].

We thank prof. Zhang (Fred Hutchinson Cancer Research Center, Seattle, Washington, USA) for providing ETV6 variant plasmids.

KSK and MD designed and performed experiments, collected, analyzed, interpreted data, wrote and edited the paper; LR, KR, MS, MZK, JB, MP, ZV, FF, SM and SP performed experiments and collected data.

The authors report no potential conflicts of interest.

Supplemental data for this article can be accessed on the https://doi.org/10.1080/09537104.2020.1802416.