Work

Concurrent STAT3, DNMT3A, and TET2 mutations in T‐LGL leukemia with molecularly distinct clonal hematopoiesis of indeterminate potential

Year: 2016

Type: article

Abstract: T cell large granular lymphocytic leukemia (T-LGLL) is a rare neoplasm with persistently increased LGL counts, typically with a CD3/CD8/TCR α/β positive immunophenotype. The clinical presentation is frequently of patients in their 5th through 7th decade of life with neutropenia, lymphocytosis, anemia, and splenomegaly. Recent investigations into the molecular profile of T-LGLL have demonstrated that STAT3 mutations are present in ∼40% of cases 1. STAT3 mutations have also been reported in hepatosplenic T cell lymphoma 2, anaplastic large cell lymphoma 3, extranodal NK/T cell lymphoma 4, and enteropathy-associated T cell lymphoma 5. TET2 and DNMT3A mutations have been described in peripheral T cell lymphomas 6 but not in T-LGLL. Myelodysplastic syndromes can have similar clinical presentations as T-LGLL (neutropenia, anemia) and frequently demonstrate recurrent mutations in genes including TET2, DNMT3A, ASXL1, and others 7, 8. The mere presence of mutations in these genes, however, does not indicate a malignant hematopoietic neoplasm. Clonal hematopoiesis of indeterminate potential (CHIP) is a recently proposed pre-neoplastic entity in which clonal hematopoiesis is demonstrated by the presence of a somatic mutation at >2% mutant allele frequency (MAF) and absence of definitive morphologic and clinical evidence of a hematologic malignancy 9. A 73-year-old male presented with progressive fatigue over the past year. The patient's past medical history was significant for celiac disease, hypogonadism, and type 1 diabetes mellitus. During a recent hospitalization for fevers, diarrhea, and vomiting (ascribed to infectious gastroenteritis), anemia and lymphocytosis were noted and further hematologic evaluation was undertaken. Complete blood count demonstrated WBC 14.4 K/cu mm (normal, 4.4–11.0), Hgb 9.3 g/dL (normal, 13.5–17.5), and platelets 274 k/cu mm (normal, 150–400). A manual WBC differential demonstrated 75% large granular lymphocytes [10.8 k/cu mm], 4% small lymphocytes, 16% neutrophils, 3% monocytes, and 2% eosinophils (Fig. 1a). No abnormal nuclear lobation, cytoplasmic hypogranularity, or other dysplastic morphologic features were noted in the myeloid series. Bone marrow evaluation demonstrated a normocellular marrow with trilineage hematopoiesis, unremarkable M:E ratio, and no increase in blasts. Dysplastic features were not identified in any lineage. CD3 (data not shown) and CD8 immunohistochemical stains highlighted linear arrays and aggregates of CD8+ T cells in a sinusoidal distribution (Fig. 1b). Flow cytometric immunophenotyping demonstrated a population of NK/T cells with dim CD5 and CD7 expression, increased CD57 expression, and absent expression of tested KIR isoforms (Fig. 1c). Cytogenetic studies revealed a normal male karyotype and fluorescence in situ hybridization (FISH) studies for myelodysplasia (MDS)-related abnormalities were negative. Morphologic, immunophenotypic, and molecular characterization of T-LGL leukemia. A: Peripheral blood smear demonstrating lymphocytosis with frequent large granular lymphocytes (inset). B: CD8 immunohistochemistry of bone marrow biopsy demonstrating linear arrays of CD8+ T-LGLs. C: Flow cytometric immunophenotyping demonstrates the neoplastic CD3+/CD8+ T-LGLL population (purple) with reduced expression of CD2, CD5, and CD7; subset CD16 negativity and increased CD57 expression, and lack of KIR isoform expression. Normal lymphocyte (green), granulocyte (blue), and monocyte (orange) populations included as internal controls. D: Mutant allele frequencies (MAF) in whole blood, sorted CD3+ T cells, and sorted CD33+ myeloid cells are expressed as a percentage of DNA sequence reads containing the mutant. Values marked with an asterisk indicate analysis via Sanger sequencing with a limit of detection of ∼20% (STAT3 p.Y640F and TET2 p.E1144K mutations); DNMT3A p.R882S and p.E907del mutations have been evaluated with high throughput sequencing with a limit of detection of 0.5%. [Color figure can be viewed at wileyonlinelibrary.com] Targeted high throughput sequencing of whole peripheral blood demonstrated the presence of DNMT3A p.R882S (4% MAF) and p.E907del (38% MAF, in-frame deletion of one amino acid) mutations, a STAT3 p.Y640F mutation (40% MAF), and a TET2 p.E1144K (40% MAF) mutation (Fig. 1d, method details in Supporting Information). The DNMT3A p.R882S mutation commonly occurs in CHIP and is also the most frequently reported DNMT3A mutation in myeloid malignancies. The DNMT3A p.E907del mutation has not been previously reported in the COSMIC database but may have deleterious functional consequences. The TET2 p.E1144K missense mutation has been reported twice in AML in the COSMIC database. Flow cytometric sorting of CD3+ and CD33+ leukocyte subsets with subsequent targeted high throughput and Sanger sequencing was performed to investigate whether the myeloid-associated mutations were present on the CD3+ T-LGLL clone or background CD33+ myeloid cells. The analysis demonstrated the T-LGLL clone to harbor the DNMT3A p.E907del, STAT3 p.Y640F, and TET2 p.E1144K mutations, whereas the DNMT3A p.R882S mutation was only detected in the CD33+ myeloid cells. Oral methotrexate plus folic acid therapy and growth factor support with weekly epoetin alfa administration were initiated with symptomatic improvement of fatigue after 2 months of therapy. Follow-up analysis of CBC parameters demonstrated some improvement of the patient's anemia (Hbg 10.2). Therapy is ongoing. The clinical differential diagnosis for T-LGLL includes bone marrow failure syndromes and myelodysplastic syndromes. While the presence of STAT3 mutated cytotoxic T-lymphocyte populations has been observed in MDS and aplastic anemia 10, in this case, the absence of dysplastic morphologic features, the low MAF of the DMNMT3A p.R882S mutation, and the restriction of other myeloid-associated mutations to the CD3+ T-LGLL clone support classification as CHIP rather than a concurrent MDS. In addition to expanding the mutational spectrum of T-LGLL to include DNMT3A and TET2 mutations, this case highlights the importance of morphologic assessment in differentiating CHIP from MDS. Pre-malignant mutations in the myeloid lineage (i.e., CHIP) are well-characterized. Given the limit of detection of Sanger sequencing, the presence of low level STAT3 and TET2 mutations at in CD33+ myeloid cells cannot be entirely excluded. If present, this could suggest the possibility that these mutations are present in an early stem cell progenitor. The concept of stem cell progenitors acquiring mutations has been previously reported in lymphoid malignancies such as angioimmunoblastic T cell lymphoma 11, chronic lymphocytic leukemia 12, and hairy cell leukemia 13. In summary, this case is the first description of concurrent DNMT3A, TET2, and STAT3 mutations in T-LGLL. Flow cytometric sorting and high throughput sequencing analysis also demonstrated background trilineage hematopoiesis with concurrent and molecularly distinct CHIP. Identification of pathogenic DNMT3A and TET2 mutations within the T-LGLL clone expands the mutational spectrum of this entity. Additionally, these mutations are potentially actionable; TET2 and DNMT3A mutations may predict better response to hypomethylating agents in myeloid malignancies when the MAF is >10% 14, 15. Although epigenetic alterations have been characterized in T cell lymphomas and agents that inhibit histone deacetylases have shown significant clinical activity, the role of hypomethylating agents in T-cell neoplasms has yet to be defined. As sequencing platforms become more widely available, it will be important to better characterize the clinical relevance of lineage specific mutations while further exploring the therapeutic applicability of epigenetic modifiers and targeted agents including in rare hematologic malignancies such as T-LGLL. PWR and SES wrote the paper. MJC analyzed pathologic samples. GF performed flow cytometric sorting. RP and PWR analyzed molecular data. SES, DB, and BJD participated in the clinical care of the patient. Philipp W. Raess,1 Michael J. Cascio,1 Guang Fan,1 Richard Press,1,2Brian J. Druker,2 Diana Brewer,2 and Stephen E. Spurgeon2* 1Department of Pathology, Oregon Health and Science University, Portland, Oregon; 2Knight Cancer Institute, Oregon Health and Science University, Portland, Oregon

Source: American Journal of Hematology

DOI: 10.1002/ajh.24586