Ineffective erythropoiesis remains the principal source of disease-related symptoms and morbidity in myelodysplastic syndromes (MDS), resulting in symptomatic anemia, red blood cell transfusion dependence (RBC-TD), iron overload, and progressive physical debility. Potential mechanisms of ineffective erythropoiesis include alterations in both early and late-stage erythropoiesis secondary to clonal stem cell defects (e.g., SF3B1, TET2, and DNMT3A mutations), abnormalities in signaling pathways (e.g., JAK-STAT, SMAD2/3), overproduction of inflammatory cytokines (NLRP3 related ILIβ, IL6 and TNFα), unbalanced iron homeostasis, and free radical mediated damage.^{1, 2} Several strategies have been employed to manage anemia in MDS including erythropoiesis stimulating agents (ESAs), lenalidomide, androgens such as danazol, DNA methyltransferase inhibitors (DNMTi), and combinatorial strategies with these agents.³ Two new drugs that have been added to this arsenal include luspatercept and imetelstat, with luspatercept being US FDA/EMA approved for use in adult patients with RBC-TD beta thalassemia and in patients with low-/intermediate-risk MDS-RS (ring sideroblasts) and MDS/MPN-RS-T (MPN-myeloproliferative neoplasm, RS-T-ring sideroblasts with thrombocytosis).^{4, 5}

Luspatercept is a recombinant activin receptor type IIB fusion protein that is designed to trap TGF-β superfamily ligands (including activin), blocking their engagement with their cognate receptors, which otherwise would have resulted in activation of SMAD2/3 signaling (SMAD2/3 signaling is known to inhibit terminal erythroid differentiation).⁶ While luspatercept can alleviate anemia by enhancing late-stage erythropoiesis, it does not modify disease biology, with progression remaining inevitable. Luspatercept's US FDA/EMA approval was based on the Phase-3, randomized, placebo-controlled MEDALIST trial, where 229 patients with RBC-TD low-/intermediate-risk MDS-RS who failed treatment with ESAs or had an endogenous erythropoietin (EPO) level of >200 U/L, were randomized to receive luspatercept versus placebo (Table 1).⁷ Twenty-eight percent of patients in the luspatercept-treated group versus 8% in the placebo group achieved RBC-TI (transfusion independence) for a consecutive 12-week period during weeks 1 through 24 (p < .001); median duration of RBC-TI was 31 versus 14 weeks, respectively.⁷ This number increased to 33% for the luspatercept arm versus 12% for the placebo arm during weeks 1–48 (p < .001). However, only 14% of patients in the luspatercept arm had their 24-week response for RBC-TI maintained at 1 year. Luspatercept-treated patients experienced higher rates of erythroid-lineage hematologic improvement (HI—53% vs. 12%) and platelet responses (63% vs. 33%), while the rate of leukemic progression was similar between the two treatment arms.⁷ On longer term follow-up, the median duration of RBC-TI ≥ 8 weeks in the heavy RBC-TD patients was 26.9 weeks, whereas the median duration of the longest major erythroid response (mHI-E) was 11.86 weeks (similar to placebo arm), highlighting limitations with durability.⁸ The MEDALIST trial also led to the approval of luspatercept for MDS/MPN-RS-T, an MDS/MPN overlap syndrome with a high frequency of SF3B1 (80%–90%) and JAK2 (50%) mutations, without any clinical trial data supporting safety and efficacy. This was surprising given that 70% of luspatercept-treated patients versus 42% of placebo-treated patients in this study had a ≥30 × 10⁹/L mean increment in their platelet counts, although these counts did remain within upper limits of normal.⁹ It was after approval that a post hoc analysis of the MEDALIST trial found that 23 (10%) patients actually met criteria for MDS/MPN-RS-T, with 14 receiving luspatercept and 9 receiving placebo.¹⁰ RBC-TI ≥ 8 weeks was seen in 64.3% of luspatercept-treated patients versus 22% of placebo-treated patients, with one patient developing a transient ischemic attack on the study drug.¹⁰ Given the higher thrombotic risk associated with MDS/MPN-RS-T, Mayo Clinic is now leading a prospective Phase 2 study to establish safety, efficacy, and an optimal dosing strategy for luspatercept, especially when concomitant cytoreduction with hydroxyurea is warranted (NCT05005182, www.clinicaltrials.gov).¹¹

After luspatercept's approval, important real-world studies failed to replicate the initial enthusiasm for the agent, highlighting nuances with clinical trial selection criteria and the application of response crtieria.^{1, 12, 13} In a Mayo Clinic study, 39 (83% SF3B1 mutant) MDS-RS patients received luspatercept, with 31 patients having RBC-TD. Erythroid responses were documented in 18%, with 16% of RBC-TD patients responding.¹² One reason for the suboptimal response might have been the fact that 95%, 38%, and 26% of these patients had prior exposures to ESA, DNMTi, and lenalidomide, respectively. In a second study assessing responses in 27 patients with chronic myeloid neoplasms without RS (Table 1; not including MPN), only 11% met the criteria for anemia response, including 8% of RBC-TD patients.¹³ Again, in this analysis, 70%, 44%, and 11% of luspatercept-treated patients did have prior exposures to ESA, DNMTi, and lenalidomide, respectively. In this issue of the journal, Lanino et al. present real-world data from an Italian luspatercept compassionate use trial for patients with lower risk MDS-RS (R-IPSS—very low, low, and intermediate risk; FISiM study), who were refractory to, or ineligible for ESA therapy and who were RBC-TD (≥2 units of red cells/8 weeks in the last 16 weeks prior to enrollment).¹⁴ Notably, these patients were not allowed to have received prior DNMTi or lenalidomide and the trial endpoints were like the MEDALIST study. RBC-TI ≥ 8 weeks in the first 24 weeks of therapy was achieved in 62 (30.8%) of 201 enrolled patients, with this number increasing to 39.3% at 48 weeks. Erythroid response (IWG/International Working Group 2006 MDS criteria) was achieved in 35.3% at 24 weeks, while according to the IWG 2018 MDS criteria, 6.9% achieved a major erythroid response and 41.9% achieved a minor erythroid response.^{15, 16} Primary responses were seen at the starting dose level (1 mg/kg) in 33 (41.8%) subjects, while dose increases at 1.3 and 1.75 mg/kg were performed in 24.1% and 34.1% of primary responders, respectively.¹⁴ The median longest duration of primary response was 23.9 weeks (8–70). Low RBC-TD was found to be a predictor for response. Side effects were seen in 35 (17.4%) patients, including 11 (33%) with cardiac events and 2.5% of study patients had progression to AML. Compared with the MEDALIST study, this study had a higher median age for enrolled patients (74 vs. 71 years) and these patients had more comorbidities.^{7, 14}

While the aforementioned MDS studies have focused on the management of MDS-RS patients refractory to, or ineligible for ESA, the question pertaining to the upfront use of this drug in lower risk MDS was recently assessed by the Phase 3, randomized, open-label, COMMANDS study.¹⁷ This study assessed the efficacy of luspatercept versus ESA in lower risk, RBC-TD (2–6 RBC units/8 weeks for ≥8 weeks), MDS patients with or without RS and with <5% BM blasts.¹⁷ Luspatercept was administered to 178 patients while 178 received EPOα, with median treatment durations of 41.6 and 27.0 weeks, respectively. Of note, 73% of assessable patients (n = 355) had RS, while 61% were SF3B1^MT. Of 301 patients included in the efficacy analysis [luspatercept n = 147, EPOα n = 154], 86 patients (58.5%) receiving luspatercept versus 48 patients (31.2%) receiving EPOα achieved the primary endpoint of RBC-TI ≥ 12 weeks with concurrent mean Hb increase ≥1.5 g/dL within the first 24 weeks (p < .0001, odds ratio 3.1). HI-E ≥ 8 weeks was achieved by 109 (74.1%) luspatercept and 79 (51.3%) epoetin alfa patients (p < .0001). Treatment-related AEs were reported by 54 (30.3%) luspatercept and 31 (17.6%) EPOα treated patients, with AML progression being comparable in both arms (2.2% vs. 2.8%).¹⁷ These effects were particularly pronounced in patients with SF3B1^MT-MDS and MDS-RS. In the SF3B1^MT-MDS group, the primary endpoint was achieved in 70% of patients receiving luspatercept versus 31% receiving EPOα, while in the SF3B1^WT group, these response rates were 42% and 32%, respectively. Similarly in MDS-RS patients, the primary endpoint was achieved in 64.8% of patients receiving luspatercept versus 26% receiving EPOα, while in the MDS no RS group these numbers were 41% and 46%, respectively. When stratified by baseline EPO levels, for patients with EPO levels ≤200 U/L, the primary endpoint was achieved in 63% of luspatercept-treated patients versus 36% of EPOα-treated patients, while for patients with EPO levels between 200 and 500 U/L, primary endpoint achievement rates were 41% and 12.1%, respectively. Hence, while the overall response rates in the study were suggestive of luspatercepts superiority to EPOα, some of these data points may be influenced by the higher proportion of SF3B1^MT-MDS-RS patients in this study: a population of patients who do better with luspatercept therapy. An important question raised by this study is the efficacy of luspatercept in comparison to EPOα in lower risk MDS patients without RS. Here the data thus far do not show a clear benefit for the frontline use of this agent. In addition, the incidence of Grade 3/4 thromboembolism was 2.8% in the luspatercept arms versus 0.6% in the EPOα arm, a safety signal that needs to be followed prospectively. In a recent pilot study, investigators assigned 28 lower risk MDS patients (85% with prior ESA exposure) to receive luspatercept in combination with ESA: 18 patients with no prior response to luspatercept (primary resistance), 7 patients with loss of response (secondary resistance), and 3 patients as an exploratory combination arm.¹⁸ Notably, 85.7% of these patients were SF3B1^MT and 43% were RBC-TD. Thirty-six percent of patients achieved a hematological response with the addition of ESA, 71% with secondary resistance, 17% with primary resistance, and 66.6% in the exploratory cohort. There were no responses seen in the SF3B1^WT group (n = 4) and in patients with baseline EPO ≥ 500 U/L.¹⁸ Accurate safety data were not available, but they did not report an increased incidence of thromboembolism.¹⁸

Congenital sideroblastic anemias (CSAs) are nonclonal disorders that result in ineffective erythropoiesis with the accumulation of bone marrow RS, due to hereditary defects in heme biosynthesis.³ Luspatercept was recently successfully used in a 51-year-old man with nonsyndromic X-linked sideroblastic anemia (ALAS2 c.1231C→T; p.Arg411Cys) who was RBC-TD and had prior treatment with pyridoxine and iron chelation, potentially opening the doors for a formal evaluation for luspatercept in these hereditary diseases.¹⁹

Here, we carefully scrutinize data, leading to the approval of luspatercept in ESA refractory or ineligible patients with MDS-RS, discuss existing real-world data and limitations of the agent in patients with prior exposures to MDS/anemia-directed therapies and MDS without RS, comment on the lack of disease-modifying activity and the limited durability (especially in heavily RBC-TD patients) of responses, explore the possibility of its use in CSA, and finally highlight an important randomized study assessing the upfront management of lower risk MDS-related anemia. While the COMMANDS study is an overall positive study, questions still remain on the role of luspatercept in the frontline management of patients with MDS without RS, MDS with isolated deletion(5q), and in MDS patients with prior exposures to MDS-directed therapies. Quality of life studies and cost comparisons will also be an important facet moving forward, especially in patients with heavy RBC-transfusion dependence, where very few patients actually achieve sustained HB levels ≥10 g/dL. Finally, while the drug is currently approved for MDS/MPN-RS-T, we encourage administration in the context of a clinical trial to ensure safety and to document efficacy.

In summary, the elucidation of the TGFβ/SMAD signaling pathway in MDS, apart from leading to regulatory drug approval for luspatercept, has also opened the doors for discovery of newer agents, including small-molecule inhibitors of the TGFβ pathway (Vactosertib and Galunisertib) and ligand traps (AVID200 and Bintrafuspα).²⁰ While these advancements bring hope and respite to affected patients and the MDS community at large, we underscore the dire need to continue to find agents with durable responses and with disease-modifying activity. Considering the above elaborated limitations and relatively short follow-up data on luspatercept, we plan to continue using ESAs as the frontline treatment for low-risk MDS, with or without RS. Our second-line options depend on other factors, including presence of deletion 5q (lenalidomide) or other nonerythroid features that might warrant the use of DNMTi.

CONFLICT OF INTEREST STATEMENT

Mrinal M. Patnaik has received research funding from Kura Oncology, Stem Line pharmaceuticals, Epigenetix, and CITI pharmaceuticals.