Longitudinal Contribution of Distinct Hematopoietic Clones during Marrow Recovery in Immune Aplastic Anemia

Introduction: Severe aplastic anemia (SAA) is an immune mediated marrow failure characterized by pancytopenia and depletion of hematopoietic stem and progenitor cells (HSPC) by cytotoxic T cells. Clonal hematopoiesis (CH) has been reported in nearly 70% of AA cases, including somatic mutations mainly in PIGA, BCOR, ASXL1 and DNMT3A. How these clones relate to disease pathophysiology, recovery post immunosuppression (IST), and evolution to myeloid neoplasm is poorly understood. We mapped trajectories of somatic variants in enriched peripheral blood (PB) granulocytes from onset to over 11 years of follow-up in order to infer HSPC contributions throughout the disease course of SAA.

Methods: We screened granulocytes of 194 SAA patients enrolled in clinical trial (NCT 01623167; ATG, cyclosporine, and eltrombopag) for somatic mutations pre- and post-IST (6 months), using targeted error-corrected sequencing (ECS) at minimum variant allele frequency (VAF) of 0.5%. Detected CH was tracked in available longitudinal samples from 86 patients (median follow-up [range] = 5 years [0.5 – 11]) at minimum VAFs of 0.1%.

Results: Patients’ median age was 32 years (3 – 82). At baseline, CH was detected in 64% of patients, most commonly in PIGA (27%), DNMT3A (23%), BCOR (18%), and ASXL1 (14%) at median VAFs <2%, and without hotspots; 25% of cases had multiple mutations in these genes. Only DNMT3A CH was age-related. Post-IST, 71% of cases showed CH, and new mutations were mostly in BCOR and ASXL1 (frequencies slightly increased to 22% and 18%, respectively). Pre-existent BCOR and ASXL1 clones expanded on average 7.5 and 3.5 fold, respectively, unrelated to 6 month hematologic response; VAFs after treatment averaged 8% (up to 37%) and 5% (up to 73%) higher than baseline, respectively, for the BCOR and ASXL1 mutations. In contrast, DNMT3A and PIGA variants were stable. PIGA, BCOR and ASXL1 mutations co-occurred in 28/194 (14%) cases; in the majority, ASXL1 or BCOR dominated clonal expansion after IST.

Longitudinal clonal dynamics were different after 6 months. Most BCOR clones disappeared or were present at very low levels (<1%) in samples 2-5 years post IST, regardless of marked expansion at 6 months. When co-occurring, a decreasing of BCOR clones coincided with late expansion of ASXL1 or PIGA clones (pre-existent or novel). Most PIGA and ASXL1 clones detected pre- or post-IST persisted during follow-up, regardless of IST response. In half of the cases, ASXL1 clones expanded to VAFs >10%, a few co-occurring with novel somatic mutations in MDS-related genes (SETBP1, RUNX1, and U2AF1). As expected, PIGA clonal dynamics were variable. In some cases, expansion of pre-existent small PIGA clones, along with signs and symptoms consistent with clinical PNH, were seen even in complete hematologic responders up to 7.5 years post IST. In three patients who had clinical PNH initially and required anti-complement therapy, there was spontaneous regression of PIGA clones detected by sequencing and GPI-anchor negative granulocytes by flow; all remain in remission off therapy.

Atypical trajectories were observed in 12/86 (14%) cases in which different dominant PIGA, BCOR or ASXL1 clones were seen pre- and post IST, and at last follow-up. In a cohort of SAA patients who were in very long-term remission with complete response (median 20 years, range 14-29 years), clonal landscape was still dominated by PIGA/ASXL1 mutations; VAFs >10% occurred in 7/11 subjects.

Conclusion: Mutation detection from enriched granulocytes supports high rates of clonality in SAA from time of diagnosis. The clonal profile dominated by PIGA, BCOR, and ASXL1 mutations is likely associated with the underlying immune mediated pathophysiology while DNMT3A is age related. HSPCs with CH in BCOR and ASXL1 are major contributors to hematopoiesis during marrow regeneration; long-term ASXL1 trajectories are consistent with linear acquisition of pre-malignant mutations that precede late secondary myeloid neoplasm. Both cell-intrinsic and extrinsic factors may explain the differential recruitment of HSPCs over time and the expansion of distinct clones, CH mutated or wild-type, during disease course. Clonal persistence of PIGA and AXSL1 mutations over 11 years in some cases indicates long HSPC lifespan.