Serum GDF15 in β-Thalassemia: A Quantitative Marker of Ineffective Erythropoiesis?

Introduction: Ineffective erythropoiesis (IE) is a crucial hallmark of β-Thalassemia (β-Thal) and sets the goal for treatment of both Transfusion-Dependent (TDT) and Non Transfusion-Dependent Thalassemia (NTDT) patients. The Growth Differentiation Factor (GDF) group has a relevant role in the molecular regulation of IE. Specifically, GDF11 contributes to the inhibition of RBC maturation and it is targeted by activin traps, such as luspatercept. However, its role is still debated; i.e., in a mouse model of β-Thal, the absence of GDF11 alone is not sufficient to mitigate IE (Guerra A et al., Blood, 2019). GDF15 increases from early until late phases of erythroid differentiation and negatively regulates erythroid cell development in-vitro, modulating maturation and apoptosis. (Ranjbaran R et a.l, Exp Cell Res, 2020). A few clinical studies found elevated serum GDF15 levels in β-Thal (Tanno T et al., Nat Med, 2007; Huang Y et al., Int J Med Sci, 2019), but data are sparse and a clear correlation with the severity of the pathology is still missing.

Methods: We run an observational study at our institution. At routine checks, patients were asked to consent to a specific blood sample for GDF15, whereas hemoglobin (Hb), serum erythropoietin (Epo), ferritin (Ftn), iron, and transferrin saturation (TSat) were measured as part of clinical practice. In a small subset of patients, a consent for serial sampling was added. Serum GDF15 was measured by ELISA (DuoSet DY957, R&D Systems). Demographics were collected from clinical records. Statistical analysis was performed using Statistica 10 (Statsoft).

Results: GDF15 levels were measured in 458 individuals: 267 TDT, 77 NTDT, 45 β-Thal trait carriers (BTC), and 69 healthy (H) subjects. Median (IQR) levels of GDF15 were significantly different among diagnoses (P<0.0001), and specifically measured 0.22 (0.16-0.34) in H, 0.48 (0.28-0.96) in BTC, 1.35 (0.40-5.46) in NTDT and 5.95 (3.19-10.52) ng/mL in TDT. (FIG 1A). In TDT patients, a mild but highly consistent negative correlation was observed between Hb and GDF15 levels (R=-0.31, P=0.002). GDF15 levels correlated positively with Epo (R=0.60, P<0.0001), TSat (R=0.32, P=0.0003) and serum iron (R=0.27, P=0.003). In addition, they correlated with length of transfusion interval in splenectomized patients (R=0.65, P<0.001). In a small subset of longitudinal data in TDT, the transfusion cycle had a strong and uniform effect on GDF15 levels. FIG 1B shows a single individual (female, 38 years old, splenectomized): mean (±SD) GDF15 levels were 5.27 (± 1.99) and 2.10 (±0.97) ng/mL pre-transfusion and 8 days post-transfusion, respectively (P<0.01). After 21 days post-transfusion, they were 3.74 (±0.76), showing a trend of variation opposite to Hb. In another individual with a prenatal diagnosis of severe TDT (homozygous for IVS I:110), GDF15 levels were high at 5 months and showed a progressive decline after the start of regular transfusion therapy. After reaching the 10 mg/dL threshold for pre-transfusion Hb, GDF15 levels were approximately halved compared to 5 months of age (6.2 vs 2.7 ng/mL) (FIG 1C). In NTDT patients, GDF15 correlated positively with TSat (R=0.40, P<0.0001), serum iron (R=0.39, P<0.0001) and Ftn (R=0.25, P=0.01). Among TDT and NTDT patients, 45 and 19 were paediatrics, respectively. No significant differences were observed at different ages. In BTC, GDF15 correlated negatively with Hb levels (R=-0.43, P=0.002) and positively with Epo (R=0.78, P=0.02), TSat (R=0.77, P<0.0001), serum iron (R=0.62, P<0.0001) and Ftn (R=0.59, P<0.0001). No significant correlations were observed for H subjects.

Discussion: GDF15 levels correlated with the severity of β-Thal phenotype, showing a 26-fold (TDT), a 6-fold (NTDT) and a 2-fold (BTC) increase compared to controls. In TDT patients, higher GDF15 levels correlated with lower Hb and higher Epo, which are typically observed as a result of IE in thalassemia. In addition, GDF15 correlated with markers of altered iron metabolism, such as TSat and serum iron. In individual patients, GDF15 showed strong and consistent variation with treatment. GDF15 was also associated with quantitative markers of disease in NTDT and BTC patients. This is to our knowledge the larger sample of patients carrying β-thal mutations in which GDF15 levels were measured and correlated with the severity of the disease. These results show that GDF15 may be a suitable and useful quantitative marker of IE.

Figure 1

View largeDownload slide

Figure 1

View largeDownload slide

Close modal