GDF15 and Erythroferrone Mark Erythropoietic Response to ACE-011 (Sotatercept) in Thalassemia

Background

The hemoglobin (Hb) response to the activin receptor type IIA ligand trap ACE-011 (Sotatercept) in non-transfusion-dependent thalassemia (NTDT), and the transfusion requirement response in TDT are described, but the mechanisms of action, clinical predictors and markers of response, are unclear. In principle, ACE-011 may act on early and/or late erythroblasts to decrease ineffective erythropoiesis (IE), both in healthy subjects and in thalassemia. As erythropoiesis intimately links to iron metabolism, changes in markers of iron metabolism relative to those of erythropoiesis may inform the mechanisms and developmental stage at which ACE-011 acts.

Methods

Markers of IE and iron metabolism in 46 thalassemia patients (30 NTDT, 16 TDT) were taken before and during a median follow up of 722.5 days, (IQR 830.5, range 152-1427) of escalating doses of ACE-011 (3-weekly 0.1 to 1.0 mg/kg s.c. injections), depending on the protocol, as part of the approved study (Cappellini, et al. 2018 under review). Markers of erythropoiesis included Hb, Reticulocytes (Ret), soluble transferrin receptor 1 (sTfR), growth differentiation factor 11 and 15 (GDF11, GDF15), erythroferrone (ERFE). Markers of iron metabolism included plasma hepcidin, serum ferritin (SF), transferrin saturation (TSAT), and non-transferrin-bound iron (NTBI).

Data were analyzed using longitudinal multilevel model for change (LMMC) on STATA (Version 14) as generalized linear mixed model with random and fixed effects to account for repeated measures and highly complex temporal data structure. Final LMMCs were built to explain the change in Hb from baseline and the behavior of key biomarkers. Independent variables were entered into the models based on the study design (predictors from design) and theoretical background (predictors of interest and control variables). P value <0.05 was considered statistically significant.

Results

Predictors of response.

In NTDT, Hb response associated positively with dose and duration of exposure (both at p<0.0001), and negatively with baseline EPO (p=0.002), GDF15 (p<0.0001) or TSAT. In TDT, accounting for transfusion effect, baseline GDF15 and EPO positively predicted Hb response while ERFE was a negative predictor (all at p<0.0001).

Biomarker changes with time

In NTDT, significant changes with time on study were increases in Hb, sTfR, ERFE and Ret and decreases in hepcidin, bilirubin, and NTBI. GDF15 showed no change. In TDT, GDF15, sTfR and ERFE increased, whereas hepcidin decreased. The Hb change was insignificant in TDT (as expected per protocol) and was dose-independent, however the mean 41% reduction in transfusion iron load rate (ILR) was dose dependent, implying an equivalent net production of Hb in TDT (Table 1).

Other significant relationships

Absolute hepcidin level in NTDT and TDT was negatively predicted by ERFE (p<0.0001): the first longitudinal demonstration of this association in thalassemia patients. GDF11, the target for ACE-011 that was shown previously to negatively regulate erythroid differentiation (Dussiot et al, Nat Med 2014), fell significantly on study (preliminary data). Significant reduction in indirect bilirubin in NTDT, implying reduced hemolysis, suggests improved quality of produced erythrocytes.

Interpretation and conclusions

NTDT patients allow a cleaner interpretation of biomarker changes as these are confounded in TDT by increased bone marrow stress from less transfusion. In NTDT, sTfR and ERFE (total erythropoiesis) increase on study while GDF15 (IE) and EPO do not. Thus Hb gain may result from increased effectiveness of late stage erythropoiesis (sTfR+ and ERFE+) possibly from decreased apoptosis and more rapid maturation (Carrancio et al, BJH 2014) in this compartment. We speculate that increased survival of those erythroid progenitor cells expressing TfR (and hence sTfR) and ERFE (hence increased ERFE) also explains the observed hepcidin reduction. Despite exposure to lower hepcidin, there is no iron loading: NTBI falls against stable SF in NTDT, consistent with shunting of iron into the sink of effective erythropoiesis. The reduction in GDF11 in vivo, not previously reported for ACE-011, is also consistent with increased apoptosis of the early progenitor pool and with improved erythroblast differentiation relative to proliferation as suggested in murine β-thalassemia treated with ACE-011 (Dussiot, et al. Nat Med 2014).

View largeDownload slide

View largeDownload slide

Close modal