Real-world comparative effects of curative and disease-modifying therapies on ineffective erythropoiesis in beta-thalassemia

Abstract Introduction In beta-thalassemia (BT), reduced or absent production of beta globin leads to the accumulation of excess alpha globin, which drives ineffective erythropoiesis (IE). In turn, IE leads to anemia and, indirectly, iron overload (IO). Parameters associated with IE and IO are elevated erythropoietin (EPO; renal hormone that stimulates erythropoiesis), high growth differentiation factor 15 (GDF-15; marker of cellular stress), low hepcidin (HAMP; hepatic hormone controlling iron absorption), and high erythroferrone (ERFE; erythroid hormone that suppresses hepcidin). Treatment options include regular transfusions with iron chelation; luspatercept, an erythroid maturation agent; and curative therapy with allogeneic hematopoietic stem cell transplant (allo-HSCT) or novel gene-modified autologous HSCT (auto-HSCT). The aim of this study was to explore the real-world impact of various treatment options in reducing IE in BT and provide parameters to compare current and future treatments for BT. Methods Patients with BT or HbE BT were enrolled from the Children's Hospital of Philadelphia and the Hospital of the University of Pennsylvania Thalassemia Centers. Patients were categorized into four treatment groups: allo-HSCT, luspatercept (all but one also received regular transfusions), auto-HSCT (all received beti-cel), and regular transfusions. Blood samples were obtained at a median of 36.3 (range 5.8-72) months (mo) after receiving allo-HSCT, 17 (6-109) mo after auto-HSCT, and 47 (5.7-56.5) mo after initiation of luspatercept. HAMP, ERFE, EPO and GDF-15 were measured by ELISA and hemoglobin (Hb) was obtained from routine clinical lab results obtained within 3 days of the research sample. Statistical comparisons among treatment groups were carried out with the Kruskal–Wallis test, followed by Dunn's post hoc correction for multiple comparisons. Results The study included 58 subjects with a mean age of 20.2 ± 14.1 (range 1-65) years, 39.7% male. Patients who underwent allo-HSCT (N=5) had Hb 11.6 ± 1.7 g/dL, with lower EPO (11.6 ± 5.9 mIU/mL), ERFE (1.3 ± 0.8 ng/mL), and GDF-15 (668.2 ± 632.3 pg/mL), and higher HAMP (86.7 ± 48.6 ng/mL) consistent with normalization of erythropoiesis and iron metabolism. Compared to allo-HSCT, patients who underwent auto-HSCT (N=17) achieved similar Hb (11.1 ± 1.0 g/dL) with slightly elevated but not statistically significant differences in EPO (27.5 ± 12.4) and GDF-15 (1574.8 ± 1094.2) and similar ERFE (1.6 ± 1.7) and HAMP (80.7 ± 28.8) levels. Taken together, these data suggest that control of IE following gene-modified auto-HSCT approaches that with allo-HSCT. The transfusion group (N=30) had lower Hb (10.0 ± 1.0 g/dL) with evidence of IE, including intermediate EPO (72.4 ± 77.8), ERFE (4.3 ± 3.7), GDF-15 (4594.0 ± 2717.0), and HAMP (68.0 ± 43.2) levels. GDF-15 was higher compared with both allo- (p<0.0049) and auto-HSCT (p=0.0071). Within the transfusion group, patients were stratified by Hb into 3 groups: high (>10.5 g/dL, N = 10), intermediate (9.5–10.5 g/dL, N = 8), and low (<9.5 g/dL, N = 12). Patients with lower Hb exhibited progressively higher EPO, ERFE, and GDF-15 levels and lower HAMP (all p<0.001), consistent with increased IE. The patients receiving luspatercept (N=6, Hb 10.1 ± 1.3 g/dL) had higher EPO (187.0 ± 140.0, p<0.001), ERFE (16.3 ± 7.5, p<0.001), and GDF-15 (7721.8 ± 1853.7, p<0.001) compared with both allo and auto-HSCT, with significantly lower HAMP (19.4 ± 17.5, p<0.01) compared with auto-HSCT. Compared to patients receiving transfusions, patients receiving luspatercept had comparable Hb but higher ERFE (p<0.05) and lower HAMP (p=0.093), consistent with significantly elevated markers of IE. Conclusions Our analysis of markers of IE and IO demonstrates that erythroid and iron homeostasis were fully restored in the group of BT patients who underwent allo-HSCT. Some markers of IE (EPO and GDF-15) remain mildly elevated in those who received gene-modified auto-HSCT; further study with larger numbers of patients is needed to determine if differences in these markers are clinically meaningful. These data also provide evidence to support a higher pre-transfusion Hb goal among regularly transfused individuals to better control IE. Finally, although luspatercept may reduce transfusion requirements, our data show that its use is associated with elevated markers of IE, suggesting that patients treated with luspatercept may benefit from close monitoring for sequelae of active IE.