ETV6-NTRK3 was a rare but recurrent molecular event in hematological malignancies.(19–22) Limited number of hematopoietic tumors have so far been reported, including AML(5, 8, 17, 23, 24), CEL(25), B-ALL(12, 17, 21, 22, 26–28) and MDS transformed AML.(29, 30) The very aggressive clinical characteristics of this patient described here was similar to the firstly reported ETV6-NTRK3 positive patient diagnosed as AML-M2.(31) Both of the two patients (Case 1 and 15) were presented with splenomegaly and accompanied by severe fibrosis of the BM samples at the time of diagnosis, and died in a short period with no remissions achieved. The manifestation was also observed in the patient (Case 7) with a diagnosis of CEL and previous history of breast and pancreatic carcinomas. We speculate that this distinct manifestation might be associated with ETV6-NTRK3 fusion gene. The AML-M3 cell line AP-1060 was obtained from the patient (Case 5), who had relapsed for the fourth time resistant to all-trans retinoic acid and arsenic trioxide.(24, 32) The patient (Case 3) initially diagnosed as MDS aggressively progressed to AML and was died for persistent resistant to chemotherapy soon after the transformation.(29) In addition, four patients introduced in Case 1, 3, 4, 15, respectively, were refractory to each treatment cycle with a short survival time, indicating primary drug resistance and poor outcome of the ETV6-NTRK3 fusion gene in AML. Furthermore, eosinophilia was described in Case 7 and 15, suggesting that ETV6-NTRK3 fusion gene may be associated with myeloid neoplasms with eosinophilia. ETV6-NTRK3 positive B-ALL is associated with relapse and/or refractory diseases (Case 11 and 12). B-ALL harboring an ETV6-NTRK3 fusion gene accounts for approximately 1% of the Philadelphia-like cases(21), and is characterized by rapid proliferation and infiltration of central nervous system in preclinical models.(27) However, ETV6-NTRK3 was relatively frequently detected in some rare non-hematologic malignancies/solid tumors, like congenital fibrosarcoma (CFS), congenital mesoblastic nephroma (CMN) and secretory breast carcinoma (SBC).(9) The presence of the ETV6-NTRK3 fusion gene indicate an excellent prognosis and highly sensitive to typical chemotherapy in some solid tumors as CFS and cellular CMN.(33)
There was a correlation between the expression of the ETV6-NTRK3 fusion gene and trisomy 11 in CMN as well as CFS, among which the polysomies of chromosomes 8 and 17 were also well known.(6, 34) We found a striking correlation between trisomy 8 and ETV6-NTRK3 fusion gene expression in hematological malignancies with available karyotypes (6/10 in total, 4/6 in AML/CEL and 2/4 in B-ALL, Table 1). Trisomy 8 is one of the most common numerical aberrations in AML, with an incidence between 10% and 15%.(20, 35) Trisomy 8 is rarely observed in ALL. Notably, Valentina Nardi et al demonstrated that the patient (Case 12) relapsed due to an expansion of an ETV6-NTRK3 rearranged subclone that also carried trisomy 8, confirming the association described above. Furthermore, FISH analysis showed that 93.8% of the ETV6-NTRK3 positive SKK-1 cells, derived from the patient 3, carried trisomy 8, which was considered to gradually expand and have a growth advantage.(29) AML with trisomy 8 is classified as intermediate risk according to the European Leukemia Net (ELN) classification.(36) In the 11q23/MLL-rearranged AML, which implied an adverse prognosis, the addition of chromosome 8 seem to independently predicted a more favorable outcome.(37) However, CBFB/MYH11 rearrangement AML with trisomy 8 had an inferior OS.(38) Here, the co-occurrence of trisomy 8 does not change the adverse prognosis of the ETV6-NTRK3 fusion gene.
However, trisomy 11 and 17 was absent in hematological malignancies, which was obviously different from non-hematological malignancies. We also found that ETV6-NTRK3 fusion gene was related to complex karyotype with a cryptic t(12;15)(p13;q25) translocation, which has never been detected in hematological malignancies using a conventional technique. Nevertheless, the chromosome translocation could be detected in several cases of solid tumors.(39) Monosomy 7 was noted in myeloid leukemias harboring non-complex karyotype (Case 4 and 7, 2/2). Trisomy 21 and deletion 17p (containing the TP53 gene) were recurrent in ETV6-NTRK3 positive B-ALL (Case 9, 12 and 13, 3/4 and 2/4 proportion respectively).
Various breakpoints of ETV6-NTRK3 have been identified. It was previously considered that exon 1 to exon 5 of the ETV6 were retained in the ETV6-NTRK3 fusion product of solid tumor, but leukemic fusion variant contained exon 1 to exon 4, excluding exon 5 of the ETV6.(8, 23–25, 40, 41) We here found that the breakpoints of ETV6 were not specific to distinguish solid tumors from leukemias. The two variants of fusion transcripts identified in leukemias may function differently in cell transformation due to the entire central domain encoded by partly of exon 4 and the entire exon 5, enabling further protein-protein interactions. There are no differences in clinical outcome between the two fusion variants. The reciprocal NTRK3-ETV6 fusion gene was not detected in all reported leukemias but could be detected in solid tumors as CFS.(42)
All the patients treated with traditional chemotherapy had disease progression while on treatment or persistent chemo-resistant, clearly highlighting that leukemia with ETV6-NTRK3 fusion is not chemo-sensitive. These observations are in contrast with the excellent clinical response on the larotrectinib, a highly selective TRK inhibitor. It has shown robust and durable anti-tumor activity in solid tumors and also in two single cases of B-ALL related to relapse/refractory.(4, 26, 28, 43–46) It has been demonstrated that larotrectinib was effective on AML cell lines carrying TRK rearrangements (IMS-M2, M0-91) as well as primary BM cells of the patient reported here. (27, 47)
In conclusion, we identified the ETV6-NTRK3 fusion gene in a patient with AML remaining refractory and surviving for a short term. Moreover, we present a study of this rare entity from a comprehensive literature search as a specific and important subgroup of hematological malignancies with poor outcome and usually an additional copy of chromosome 8, which were different from solid tumors with good prognosis and always trisomy 11. Most importantly, the translocation t(12;15)(p13;q25) is highly cryptic and easily overlooked in conventional cytogenetics. Newly diagnosed AML presented with myeloproliferative neoplasm clinal features such as splenomegaly, myelofibrosis, eosinophilia or trisomy 8 cytogenetic abnormality, may harbor ETV6-NTRK3 fusion transcript. Our study highlights the significance of combining multiple molecular techniques to early identify the cryptic t(12;15)(p13;q25)/ETV6-NTRK3 in hematological malignancies and also emphasizes the value of TRK inhibitors on patients with ETV6-NTRK3 positive hematological malignancies.