Eltrombopag for patients with moderate aplastic anemia or uni-lineage cytopenias

Abstract

There is no standard or widely effective treatment of patients with moderate aplastic anemia (MAA) or hypo-productive uni-lineage cytopenias (UC). Eltrombopag (EPAG), a small molecule thrombopoietin mimetic, has previously been shown to result in durable multi-lineage hematologic responses with low toxicity in patients with refractory severe aplastic anemia (SAA). Its safety and efficacy in MAA are unknown. This prospective phase 2 study enrolled previously untreated and treated MAA and UC patients with clinically relevant cytopenias. EPAG was administered at doses escalating from 50 to 300 mg/d. Hematologic responses were assessed at 16 to 20 weeks. Responding patients were continued on EPAG until reaching defined robust or stable blood counts. EPAG was reinstituted for relapse. Thirty-four patients were enrolled between 2012 and 2017, including 31 with MAA and 3 with UC. Seventeen patients responded in at least 1 eligible lineage by the primary end point. A striking improvement in anemia was observed in a patient with Diamond-Blackfan anemia. EPAG was well tolerated, and it was discontinued for robust or stable blood counts in 12 of 17 patients after a median of 8 months. A majority required re-initiation of EPAG for declining counts, and all regained response. Two of 34 patients developed non-chromosome 7 bone marrow cytogenetic abnormalities while taking EPAG, without dysplasia or increased blasts. Somatic mutation allele frequencies in cancer genes did not increase overall on EPAG. EPAG is a well-tolerated oral treatment of cytopenias in patients with MAA/UC. This trial was registered at www.clinicaltrials.gov as #NCT01328587.

Conflict of interest statement

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

Figures

Graphical abstract

Figure 1.

Summary of EPAG responses. (A) Bar charts show the absolute numbers and percentage of responders (green shaded) and nonresponders (light blue shaded) for all patients, patients with and without previous IST, and patients with and without germline mutations. (B) Bar charts show the absolute numbers and percentage of responders (color shaded) for patients eligible for an erythroid response, patients eligible for a myeloid response, and patients eligible for bi-lineage responses, both at the primary end point and at best response during the extension study.

Figure 2.

Longitudinal mapping of responses, relapses, and cytogenetic progression. All enrolled patients are shown, grouped into responders and nonresponders at the primary end point of 16 to 20 weeks. From baseline until the primary end point or early discontinuation due to intolerance (UPN27), all patients were taking EPAG, designated by gray bars. Seventeen patients met hematologic response criteria at the primary end point and were continued on EPAG on the extension arm of the study, designated by solid green bars. Twelve of 17 patients reached robust or stable response criteria (see the Methods section) and were taken off EPAG (hatched green bars). Seven of these patients relapsed after EPAG was discontinued; EPAG was restarted, and all of them again responded (solid yellow bars). Four patients again reached a robust response and came off EPAG a second time (hatched yellow bars). The 2 patients with cytogenetic progression are shown with the specific chromosomal change and the number of abnormal metaphases. The 3 patients with germline mutations are indicated (TERT, RPS19, and MPL).

Figure 3.

Acquired somatic mutations in relation to EPAG treatment. (A) Targeted deep sequencing for a panel of genes previously reported as associated with myeloid cancers (MC) or aplastic anemia (AA) was performed on bone marrow cells from all patients at baseline and at the primary end point. Each row indicates a specific MC/AA gene found mutated in at least 1 patient, and each column represents each patient’s samples at baseline and at the primary end point. The panel of MC/AA genes is given in the supplemental Methods, and the mutations detected are listed in supplemental Table 4. Patients are grouped according to response (green at primary end point designates response) and the presence of abnormal cytogenetics (pink box). (B) The percentage of variant alleles is shown on the y-axis at baseline and the primary end point for the 9 individual mutations detected in 6 patients, with a P value (paired Student t test) comparing VAF at baseline and primary end point. (C) The percentage of glycosylphosphatidylinositol-negative (GPIneg) PNH granulocytes in 11 patients with detectable clones at baseline and/or the primary end point are plotted, with a P value (paired Student t test) comparing baseline and primary end point. The 3 patients with germline mutations (UPN10, UPN30, and UPN34) had no somatic mutations detected.

Figure 4.

Longitudinal follow-up of mutated clone size on and off EPAG. Candidate gene mutations were analyzed at baseline, primary end point, and the latest time point available on and off EPAG. Each panel shows the results in a single patient, with mutated clone size off EPAG (at baseline and long-term follow up) designated by open symbols and clone size on EPAG (at primary end point or during long-term follow-up) designated by closed symbols. Timeline bars designating time on and off EPAG are shown for each patient, with the same color key as used in Figure 2. Specific somatic mutations are listed in supplemental Table 7.

Figure 4.

Longitudinal follow-up of mutated clone size on and off EPAG. Candidate gene mutations were analyzed at baseline, primary end point, and the latest time point available on and off EPAG. Each panel shows the results in a single patient, with mutated clone size off EPAG (at baseline and long-term follow up) designated by open symbols and clone size on EPAG (at primary end point or during long-term follow-up) designated by closed symbols. Timeline bars designating time on and off EPAG are shown for each patient, with the same color key as used in Figure 2. Specific somatic mutations are listed in supplemental Table 7.

Figure 5.

Impact of EPAG on HSPC frequency. The frequencies of CD34+/CD38– primitive HSPCs, CD34+/CD38+ maturing HSPCs, and total CD34+ HSPCs within bone marrow mononuclear cells is shown for 12 responding (A) and 10 nonresponding (B) patients. P values for paired Student t tests comparing baseline (Pre) and primary end point (Post) are given. *P < .05.