A single nucleotide polymorphism in the TCL1A promoter is not protective in patients with germ line RUNX1-FPD Open Access

TO THE EDITOR:

Deleterious germ line RUNX1 variants cause familial platelet disorder (FPD) with associated myeloid malignancy, an autosomal dominant disorder.^1,2 Patients with FPD have quantitative and qualitative platelet defects, easy bleeding and bruising, and a 35% to 45% lifetime risk of developing hematologic malignancy (HM).¹ These patients are also prone to the acquisition of somatic mutations in HM-associated genes and the development of clonal hematopoiesis of indeterminate potential (CHIP).³ 

TCL1A is a proto-oncogene that is activated in T-cell prolymphocytic leukemia.⁴ Activation of TCL1A leads to clonal expansion of cells harboring mutations in driver genes for CHIP.⁵ A genome-wide association study revealed that an inherited single nucleotide polymorphism (SNP) in the promoter of TCL1A is associated with slower expansion of CHIP clones associated with TET2, ASXL1, SF3B1, and SRSF2 mutations, but not with DNMT3A.⁵ The reference allele of the variant, rs2887399, is guanine (G), and the alternate, protective allele is thymine (T).⁵ 

To our knowledge, the effect of the TCL1A protective T allele on patients with germ line RUNX1 variants has not been studied before. In this letter, we describe the findings of the effect of the TCL1A SNP on patients with germ line pathogenic (P) or likely pathogenic (LP) RUNX1 variants enrolled in the RUNX1 Natural History Study at the National Institutes of Health (NIH) Clinical Center (ClinicalTrials.gov identifier: NCT03854318).

A total of 159 participants in our natural history study with RUNX1 germ line variants were genotyped for the TCL1A rs2887399 variant. This cohort included 128 P/LP variants, 23 variants of unknown significance (VUS), and 8 benign (B)/likely benign (LB) (supplemental Table 1). Additionally, 50 family controls were genotyped (supplemental Table 1). Of the 128 patients with RUNX1 P/LP variants, 11 had the T/T genotype (8.6%), 37 had G/T (28.9%), and 80 had G/G (62.5%) (Figure 1A). This resulted in a T allele frequency of 23% in the patients with germ line RUNX1 P/LP variants. For the 50 family controls (wild-type RUNX1), 2 were T/T (4%), 17 G/T (34%), and 31 G/G (62%; Figure 1A). This resulted in a T allele frequency of 21% in the family control cohort. The T allele frequency in the RUNX1 variant carriers or the family controls was not statistically different from the one previously reported.⁵ Of the 23 patients with VUS RUNX1, 17 were G/G (74%), 6 were G/T (26%), and zero with T/T (supplemental Figure 1A). For the 8 patients with B/LB RUNX1 variants, 5 were G/G (62.5%), and 3 G/T (37.5%) (supplemental Figure 1B). The remainder of the analysis considered only individuals with germ line P/LP RUNX1 variants as patients. Participants with germ line VUS, B/LB RUNX1 variants were excluded from analysis, as it is unclear if their presentation is consistent with that of FPD.

When examining the prevalence of HM in patients with FPD in relation to their TCL1A genotype, the largest percentage of patients with HM was in the T/T group at 27.27% compared to that of G/T (18.92%) and G/G (8.75%) (Figure 1B). However, there was no statistically significant difference between the 3 groups (P=.097, Fisher exact test); likely due to the relatively small size of the cohort. Additionally, there was a large range of age at diagnosis of HM as expected in RUNX1-FPD (Figure 1C; supplemental Table 2). However, the presence of the TCLI1A protective allele does not seem to delay the onset of HM (P=.61, one-way analysis of variance). Of the patients with HM, 2 had TET2, 3 with ASXL1, 1 with SF3B1, 1 with SRSF2, and 3 with DMT3A somatic mutations (Figure 1C; supplemental Table 2). None of the patients with HM developed a second RUNX1 mutation and their treatment and clinical features at the time of diagnosis are further outlined in supplemental Table 2.

We next analyzed the effect of the TCL1A SNP on clonal expansion. Among the CHIP genes that have been reported to be affected by the TCL1A SNP, we found 12 patients with at least 1 TET2 somatic mutation, 1 with ASXL1, 2 with SF3B1, 2 with SRSF2, and 2 with DNMT3A, at variant allele frequency (VAF) >2%, for a total of 15 patients (supplemental Table 3). For 10 of the 15 patients, we had next generation sequencing (NGS) and/or exome sequencing (ES) data available at 2 or more time points.

Two of the patients with TCL1A T/T genotype had multiple somatic mutations in TET2 and other CHIP genes at high VAF, and both had developed HM. For FPD_90.1, 2 TET2 mutation clones were detected at high VAF at 6 time points over the course of 26 months, and a new TET2 mutation and a SF3B1 mutation were detected at the sixth time point (Figure 2A,B; supplemental Table 3). Despite the diagnosis of chronic myelomonocytic leukemia (CMML), FPD_90.1 had not undergone treatment. FPD_133.1 carried 2 TET2 somatic mutations, 1 ASXL1, and 1 DNMT3A (supplemental Table 1C) and had also been diagnosed with CMML. This patient had 2 sequencing time points, 5 months apart, which showed significant VAF increases for both TET2 mutations, and the ASXL1 and DNMT3A mutations (Figure 2A,C,E; supplemental Table 3). It should be noted that FPD_133.1 started chemotherapy for CMML between these 2 time points and eventually underwent bone marrow transplantation.

Five other patients had 1 TET2 mutation each and sequencing results at 2 or more time points (FPD_9.1, FPD_9.3, FPD_126.1, FPD_37.2, and FPD_105.8) (Figure 2A; supplemental Table 3). The TET2 VAF for FPD_9.1 (TCL1A G/T) stayed relatively high through 5 time points (Figure 2A; supplemental Table 3). This patient had also been diagnosed with monoclonal gammopathy of undetermined significance. FPD_9.3, FPD_126.1, and FPD_105.8 (TCL1A G/G) had relatively low and stable TET2 VAF. Finally, FPD_37.2 (TCL1A G/G) had 1 TET2 mutation at the initial time point, which disappeared at later time points (Figure 2A; supplemental Table 3), even though the patient had a SRSF2 mutation with an increasing VAF over time (Figure 2D; supplemental Table 3).

FPD_42.4 (TCL1A G/T) had a large SF3B1 mutation clone with a stable VAF over 4 time points. However, this patient was diagnosed with myelodysplastic syndrome with ring sideroblasts prior to these genetic screenings (Figure 2B; supplemental Table 3). FPD_125.1 (TCL1A G/G) had a SRSF2 mutation at high VAF, which was relatively stable (Figure 2D; supplemental Table 3). Finally, FPD_4.2 (TCL1A T/T) had a DNMT3A mutation at a relatively low and stable VAF over 5 time points (Figure 2E; supplemental Table 3).

Previous reports indicated that the T allele could reduce the expression of TCL1A in blood, especially in B cells.⁵ Therefore, we performed intracellular flow cytometry to assess the expression of TCL1A protein in the peripheral blood cells from patients with FPD. Flow cytometry was completed on 18 patients of varying TCL1A genotypes (supplemental Table 4). The data confirmed expression of TCL1A in the B cells from patients with FPD (supplemental Figure 2A). However, there was no significant difference in the percentage of TCL1A-positive cells within the B cells or the mean fluorescence intensity between the TCL1A genotype in either the fresh or frozen peripheral blood cells (one-way analysis of variance /Student t; α = 0.05; supplemental Figure 2B). This could be due to the small sample size.

Overall, although the TCL1A T allele was found to have protective qualities in clonal hematopoiesis in the general population, this does not seem to hold true for patients with RUNX1-FPD. Furthermore, the T allele does not seem to protect against HM development. We hypothesize that the germ line RUNX1 pathogenic variants in these patients are masking any protective abilities of the T allele SNP in TCL1A.

Acknowledgment: The authors are grateful to the patients and their family members for their participation in the RUNX1-Famililal Platelet Disorder clinical study. The authors thank Yevgeniya Abramzon of the National Human Genome Research Institute (NHGRI) Genomics Core for helping develop the assay for genotyping patients.

This work was supported by the Intramural Research Programs of the NHGRI and the National Heart, Lung, and Blood Institute of the National Institutes of Health.

Graphs were created with GraphPad Prism 10.

Contribution: A.K. conducted the experiments, analyzed the genotyping, genetic, and patient data, and wrote the manuscript; N.T.D. provided genetic counseling on the genetic variants on the patients and collected genetic testing data from patients; A.C.M. analyzed flow cytometry data; M.K. assisted in setting up the flow cytometry experiment and provided guidance on analysis; U.H. and R.S. assisted with genotyping the patients; S.C., K.C., and J.D. collected patient and family history and coordinated patient visits and sample collection; D.J.Y. served as the clinical director for the National Institutes of Health study protocol; E.B. helped plan the project and supervised experiments; P.P.L. served as the protocol principal investigator; and all authors discussed and revised the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Erica Bresciani, Oncogenesis and Development Section, Translational and Functional Genomics Branch, Division of Intramural Research, National Human Genome Research Institute, National Institutes of Health, 50 South Drive, Building 5, Room 5154, Bethesda, MD 20892; email: erica.bresciani@nih.gov; and Paul P. Liu, Oncogenesis and Development Section, Translational and Functional Genomics Branch, Division of Intramural Research, National Human Genome Research Institute, National Institutes of Health, 50 South Drive, Building 5, Room 5154, Bethesda, MD 20892; email: pliu@nih.gov.