Clonal Dynamics of Refractory Aplastic Anemia in Patients Treated with Eltrombopag

Despite documented success of immunosuppressive therapy (IST) in the treatment of AA, a minority of patients remain refractory, most responses are incomplete, and use of hematopoietic cell transplantation (HCT) is limited in older patients or those with significant comorbidities. While the introduction of the cMpl agonist eltrombopag (EPG) as salvage therapy or in conjunction with IST has revolutionized treatment for refractory AA. It may be effective in improving primary response rates to IST, engaging growth factor receptors with agonistic therapeutics (such as EPG) and also has the potential to promote evolution/expansion of mutant clones, thereby increasing the rate of progression to secondary myelodysplastic syndromes (MDS), a serious complication of AA occurring in 10-20% of patients. Clonogenic somatic mutations typical of MDS in patients with AA and PNH may increase the risk of progression to MDS.

DNA from marrow samples of primary refractory AA patients was subjected to analysis before and after initiation of EPG to evaluate clonal expansion or evolution using a targeted multi-amplicon deep NGS panel of all ORFs of the top 60 most commonly mutated genes in MDS. In addition to the EPG treatment group, a case control cohort matched for age and duration from AA diagnosis to last clinical follow up (who did not receive EPG), was studied.

Among 210 AA patients treated at Cleveland Clinic, we identified 26 who were treated with EPG for IST-refractory AA; median duration of treatment was 56 wks. The overall response rate after 12 weeks of therapy was 58% (15/26), while 31% of patients (8/26) showed stable disease with intermittent transfusions (one of whom underwent HCT). In 3 non-responders, one developed PNH, one had refractory AA/PNH, and one progressed to AML (see below). Expansion of PNH granulocytes after EPG treatment was observed in 23% of patients (6/26). In addition, 15% (4/26) had atypical subclonal chromosomal abnormalities.

Prior to EPG, at least a single somatic event was found in 31% of patients (8/26), with 2 patients harboring 2 mutations. Events included CEBPA, EZH2, BCOR/BCORL1, ASXL1, U2AF1/2, TET2, and DNMT3A mutations. Following EPG therapy, acquisition of new somatic mutations was observed in 23% of cases, including RUNX1, U2AF1, BCOR, RIT1, and CEBPA. In cases with pre-existing clones, 6 clones expanded (e.g., BCOR or ASXL1 from VAF of 8 to 21% and 9 to 29%, respectively) despite clinical hematologic response, while in 2 cases clones disappeared (e.g., U2AF2 and BCORL1). In 54% of cases (14/26), we found detectable levels of a PNH clone at the time of diagnosis. Six of those cases had PNH clonal expansion post-EPG treatment, of which two developed clinically significant PNH clonal burden requiring eculizumab therapy.

In the case-control cohort, 26 AA patients who received IST but were not treated with EPG, were followed for comparable time periods, and no evidence of progression to MDS was recorded. One patient was noted to have trisomy 15 on cytogenetics at diagnosis. "MDS type" molecular mutations were present in 10 patients similar to EPG cohort. Among these patients, 3 had persistent clones of U2AF1, DNMT3A, and STAT3 over one year without acquisition of any new molecular mutations. . PNH granulocytes expanded in 50% of AA cases, decreased in 30% and stayed stable in 20%. Thus, we did not observe any difference in expansion of PNH clones between those treated and untreated with EPG (p=0.73). Unlike for PNH clones, accounting for both new evolution and expansion of preexisting molecular mutations, the frequency of these clonal events was significantly higher in the EPG treated group (p=0.009).

In conclusion, we observed occasional expansion of clones with potentially leukemogenic mutations during treatment with EPG in pts with AA. While higher rates of MDS evolution were not observed in this cohort of EPG treated patients, we found that serial evaluation of somatic mutations can inform clonal evolution and can potentially be used as abiomarker for evaluation of risk for post-AA MDS. Continued use of EPG in such patients should be judicious.