Clonal Compositions Involving Epigenetic Regulator Gene Mutations in Clonal Hematopoiesis, Clonal Cytopenias of Undetermined Significance and Chronic Myelomonocytic Leukemia

Introduction: Clonal hematopoiesis of indeterminate potential (CHIP) is defined by the presence and subsequent expansion of leukemia-associated driver mutations (MT). These MT predominantly involve epigenetic regulators such as DNMT3A, ASXL1, and TET2 (DAT, 75%). CHIP is a disease of aging and is associated with increased all-cause mortality. Clonal Hematopoiesis (CH), when present in the context of unexplained cytopenias and MT variant allele fractions (VAF) >20%, is termed as clonal cytopenias of undetermined significance (CCUS). Chronic myelomonocytic leukemia (CMML) is a hematological malignancy characterized by TET2, ASXL1, and SRSF2MT as well as sustained monocytosis. It usually arises in the setting of existing CH and clonal monocytosis. Here, we assessed the clonal compositions involving DAT genes in patients (pts) with CHIP, CCUS, and CMML.

Methods: We included 1008 pts: 54 (5%) with CHIP, 334 (33%) with CCUS, and 620 (62%) with CMML. Demographics, laboratory parameters, and mutational spectrum at diagnosis were collected. We characterized and compared DAT MT in all three cohorts and assessed correlations with clinical parameters. MT locus and type along with protein sequences were analyzed using ProteinPaint. All statistical analyses were performed using R.

Results: Of the 1008 pts, 427 (42%) had DAT MT; 36 had CHIP (67%); 96 had CCUS (29%), and 296 had CMML (48%). The median age was 71 years (IQR 65-76), and 68% were male. CMML pts had higher white blood cell counts, absolute neutrophil counts, absolute monocyte counts, and lower hemoglobin and platelet counts in comparison to CCUS and CHIP pts. The distribution of age and sex were similar among the three groups. After a median follow-up of 11 months (IQR 9.5-12.8), there were no myeloid progressions in the CHIP group and 17 (18%) in the CCUS group (8 of which developed CMML, Table 1).

There were 738 protein-altering or splice-site DATMT in 495 unique MT hotspots. Pts with TET2^MT were older (71 years, compared to DNMT3A^MT: 69 years and ASXL1^MT: 69 years , p<0.022). ASXL1^MT were more common in males (75%, compared to DNMT3A^MT: 52% and TET2^MT: 66%, p=0.005). The mean VAF (mVAF) were 15.5% for DNMT3A^MT, 35% for ASXL1^MT and 39.2% TET2^MT, respectively (p<0.001). We then compared DATMT characteristics in CHIP, CCUS, and CMML groups:

• DNMT3A^MT were most common in CHIP, followed by CCUS, and were very rare in CMML (35% vs. 15% vs. 3%, p<0.001), with mVAF of DNMT3A^MT increasing across this spectrum (CHIP: 6.5% vs. CCUS: 17.1% vs. CMML: 43.4%, p<0.0001). While DNMT3A missense MTs were most common (n=62, 67.4%), the AML-associated R882 hotspot MT was seen in 26 (28%) pts (CHIP: n=3, 5%, CCUS: n=10, 16%, and CMML n=13, 21%, Figure 1a). None of the DNMT3A^MT CHIP and CCUS pts progressed to CMML.

• ASXL1^MT were most frequently seen in CMML (n=178, 28.6%), followed by CHIP (n=4, 7.4%), and CCUS (n=17, 5%, p<0.001). The mVAF of ASXL1^MT was significantly lower in CHIP, followed by CCUS, and CMML (11% vs. 34% vs. 36%, p<0.0001). Truncating ASXL1 MTs were most common (n=205, 53%). Mutational hotspots included G646W (n=81, CCUS: n=6, 3%, CMML: n=75, 37%) and E635 (n=27, CHIP: n=1, 0.5%, CCUS: n=1, 0.5%, CMML: n=27, 13%, Figure 1b). Four ASXL1^MT CCUS pts transformed to CMML (R965, E635, G646W, Q910*, and Q925*); of which 2 co-mutated with TET2^MT, one with SRSF2^MT, and one with an additional ASXL1^MT.

• TET2^MT were frequently encountered in CMML, followed by CCUS and CHIP pts (CMML: n=205, 33% vs. CCUS: n=88, 26.3% vs. CHIP: n=15, 27.8%, p=0.02). The mVAF of TET2^MT progressively increased across the 3 groups (CHIP: 14.1%, CCUS: 31.9%, CMML: 45.5%, p<0.001). TET2^MT occurred across the span of the gene, with no clear hotspots identified, with nonsense MT being most frequent (n=150, 37.3%, Figure 1c). Among the nonsense and frameshift TET2^MT (n=209), 131 (63%) involved the N terminal region, and 78 (37%) involved the catalytic domain. Among 8 CCUS pts that transformed to CMML, 6 had TET2^MT, including 4 with spliceosome co-MT.

Conclusion: While CMML, a disease of aging, arises in the background of CH, the most common CH MT, DNMT3A^MT, plays a minor role in this process, and in fact, DNMT3A^MT tend to occur at younger ages. While DNMT3A ^MT and ASXL1^MT are preferentially observed in hotspots, TET2^MT span the entire gene. Clonal compositions largely involving truncating TET2^MT and frameshift ASXL1^MT, in conjunction with splicing MT, shape the CMML genotype.

Figure 1

View largeDownload slide

Figure 1

View largeDownload slide

Close modal