Efficacy and Immunomodulating Properties of Eltrombopag in Aplastic Anemia following Autologous Stem Cell Transplant: Case Report and Review of the Literature

Thrombopoietin receptor agonists (TPO-RA) are currently indicated for the treatment of chronic immune thrombocytopenia and relapsed refractory aplastic anemia. However, the off-label use of these drugs is more and more frequent, including in the setting of aplasia secondary to chemotherapy and hemopoietic stem cell transplant (SCT). Growing evidence suggests that mechanisms of action of TPO-RA go beyond the TPO-receptor stimulation and point at the immunomodulating properties of these drugs. Here, we present a case of prolonged bone marrow aplasia secondary to autologous SCT treated with eltrombopag. We describe the clinical efficacy and the immunomodulating effect of this drug on inflammatory cytokine profile and bone marrow histology. Furthermore, we provide a review of the most recent literature highlighting the efficacy and safety of TPO-RA after SCT and chemotherapy for hematologic conditions.


Introduction
Thrombopoietin receptor agonists (TPO-RA), eltrombopag and romiplostim, are two polyhedric drugs initially licensed for the treatment of immune thrombocytopenia (ITP). Both act by triggering the TPO receptor on megakaryocytes (MK) and hematopoietic stem cells, inducing activation of the JAK2/STAT5 pathway with subsequent MK proliferation and an increase in platelet generation [1,2]. However, the two drugs present some differences both in biological properties and mechanisms of action [3]. In particular, eltrombopag appears to target the earlier precursor of MK, stimulating not only the proliferation pathway but also differentiation [1,4]. Furthermore, growing evidence is pointing at a broader spectrum of action of this small molecule, including immunomodulating properties and iron mobilization [5]. Eltrombopag is administered orally and is currently indicated for the treatment of chronic ITP after failure of first line therapy, relapsed refractory aplastic anemia (AA) and thrombocytopenia secondary to hepatitis C during treatment with interferon (IFN). Romiplostim has a subcutaneous route of administration and is licensed for chronic ITP only. However, the off-label use of these two drugs is more and more frequent, particularly in the setting of aplasia secondary to chemotherapy and hemopoietic stem cell transplant (HSCT), and thrombocytopenia associated with lymphoproliferative disorders and myelodysplastic syndromes (MDS) [6]. In this case report, we describe the clinical efficacy and the immunomodulating effect of eltrombopag on inflammatory cytokine profile in a patient experiencing prolonged aplasia following autologous stem cells transplant (ASCT) for amyloidosis AL. Furthermore, we provide a review of the most recent literature highlighting the efficacy of TPO-RA after HSCT and chemotherapy for hematologic conditions. A 60-year-old woman was referred to our center to receive autologous stem cell transplant (ASCT) due to amyloidosis AL with renal and cardiac disorder. The latter had been diagnosed one year before, and treated with 4 cycles of bortezomib, cyclophosphamide and dexamethasone without significant improvement. Following stem cell mobilization with granulocyte colony-stimulating factor (G-CSF), 6 × 10 6 CD34+ cells/kg were collected. After admission, she was conditioned with Melphalan 200 mg/square meter and two days later, she was reinfused with 4 × 10 6 CD34+ cells/kg. The post-transplant period was complicated by extremely severe acute kidney injury, hemorrhagic shock secondary to hematemesis due to severe mucositis, Pneumococcus pneumonia, sepsis by Bacillus cereus and graft failure requiring 3 further peripheral blood stem cell infusions and prolonged stimulation with G-CSF. After 35 days from ASCT, the patient was transfusion-dependent on red blood cells (RBC) and platelets (PLT), and BM evaluation showed severe aplasia. Eltrombopag was started at 50 mg per day (Figure 1), and subsequently increased to 100 mg per day, but then ceased on day 52 due to apparently scarce efficacy on platelet count. Neutrophils' recovery started from day 45 after ASCT and she was discharged on day 63 with persistent need of RBC and PLT transfusions. On day 91, BM revaluation showed a cellularity of 10% without pathological infiltrate or amyloid deposition. On day 129, due to the persistence of severe thrombocytopenia, a new attempt with eltrombopag at 150 mg per day was performed. The patient progressively became transfusion independent from day 200 for RBC and 249 for PLT and these latter increased above 30 × 10 9 /L from day 262. By day 390, given stable PLT counts > 50 × 10 9 /L, eltrombopag was de-escalated to 100 mg per day and subsequently to 50 mg per day, until discontinuation on day 536. At the last follow up, 30 months after ASCT, amyloidosis is still in remission and the patient displays normal blood counts (Hb 12 g/dL, neutrophils 2070 × 10 6 /L, PLT 151 × 10 9 /L) off therapy. Regarding safety, the patient did not show treatment-emergent adverse events RBC and 249 for PLT and these latter increased above 30 × 10 9 /L from day 262. By day 390, given stable PLT counts >50 × 10 9 /L, eltrombopag was de-escalated to 100 mg per day and subsequently to 50 mg per day, until discontinuation on day 536. At the last follow up, 30 months after ASCT, amyloidosis is still in remission and the patient displays normal blood counts (Hb 12 g/dL, neutrophils 2070 × 10^6/L, PLT 151 × 10 9 /L) off therapy. Regarding safety, the patient did not show treatment-emergent adverse events during eltrombopag therapy, including thrombosis and increase in bone marrow reticulin fibrosis (WHO grade MF-0 at all evaluations).

Monitoring of Clonal Evolution during Eltrombopag Therapy
To investigate efficacy and clonal evolution under eltrombopag therapy, BM was performed at 3, 6, and 12 months of eltrombopag treatment. It showed a progressive improvement of cellularity along with an increasing polyclonal lymphoid infiltrate of double-negative T-CD3+ cells ( Figure 2). No signs of dysplasia or blasts increase were reported. Cytogenetics was persistently normal. Flow-citometry for paroxysmal nocturnal hemoglobinuria clones was performed at 3 months from eltrombopag start and was negative. By NGS analysis on peripheral blood after 6 months of eltrombopag, a mutation of the RB1 gene, with a variant allele frequency of 52.8%, was detected. A re-evaluation at 12 months did not confirm this finding and showed an absence of mutations.

Monitoring of Clonal Evolution during Eltrombopag Therapy
To investigate efficacy and clonal evolution under eltrombopag therapy, BM was performed at 3, 6, and 12 months of eltrombopag treatment. It showed a progressive improvement of cellularity along with an increasing polyclonal lymphoid infiltrate of double-negative T-CD3+ cells ( Figure 2). No signs of dysplasia or blasts increase were reported. Cytogenetics was persistently normal. Flow-citometry for paroxysmal nocturnal hemoglobinuria clones was performed at 3 months from eltrombopag start and was negative. By NGS analysis on peripheral blood after 6 months of eltrombopag, a mutation of the RB1 gene, with a variant allele frequency of 52.8%, was detected. A re-evaluation at 12 months did not confirm this finding and showed an absence of mutations.  Figure 3 shows the cytokine levels during eltrombopag treatment. The T helper1 (Th1) cytokines IFN-γ and TNF-α were at lower limit of normality of controls and remained stable along eltrombopag treatment. The T helper 2 (Th2) cytokines IL-6 and IL-10 levels and the immunoregulatory cytokine IL-17 were higher than in controls, and persisted elevated during therapy, with a peak of IL-6 levels at 12 months. Finally, the inhibitory cytokine TGF-β was progressively increased along eltrombopag treatment, although within the normal range.

Literature Review
Several studies have been published about the use of TPO-RA for aplasia secondary to HSCT and chemotherapy both in adult and pediatric settings (Tables 1 and 2). A total of 1218 patients received TPO-RA post-HSCT, of whom, 581 (48%) eltrombopag, 122 (10%) romiplostim, and 515 (42%) both agents. The great majority were retrospective studies, including small series of around 30 patients, a larger study of 86 cases, a meta-analysis of 378 patients, and 6 prospective trials (Table 1).  Overall, they highlight a highly variable efficacy in post HSCT aplasia, and a recent meta-analysis on 378 patients documented a pooled response rate of 73% (95% CI: 68-78%) for both eltrombopag and romiplostim [34]. However, the majority of the studies analyzed the efficacy of these drugs in an allogenic HSCT setting and only few data are available for post ASCT aplasia. To the best of our knowledge, only 8 articles, for a total of 28 cases, reported the use of TPO-RA after ASCT, and pooled efficacy may be calculated from 3 retrospective studies only, including 8 patients with 100% response rate [8][9][10]13,19,27,32,39]. Notably, responding patients within the various reports could discontinue TPO-RA after persistent recovery of cytopenia and bone marrow function. However, such small series carry the bias of reporting only successful cases and a prospective evaluation would be necessary.
Far fewer studies are available about the use of TPO-RA after chemotherapy in hematologic malignancies (2 retrospective, 2 prospective trials, and 1 case report), for a total of 233 patients, the great majority receiving eltrombopag. Different efficacy endpoints were considered, often expressed as median time to PLT count recovery [43][44][45][46][47]. The latter was reported to be about 22.5 (16-43) days in acute myeloid leukemia patients and 7.4 (4.9-9.9) days in lymphoma patients [45,46]. Interestingly, even in these cases, TPO-RA discontinuation was possible after cytopenia recovery. Finally, novel anti-cancer drugs such as immune checkpoint inhibitors (nivolumab, ipilimumab, pembrolizumab, etc.) reactivate the immune system to recognize tumor specific antigens but may in turn cause autoimmune hyperinflammatory reactions, including autoimmune cytopenias. In this context, TPO-RA have been used in 6 cases with a good efficacy and safety profile [48][49][50][51][52][53]. Similarly, CAR-T cell therapy has been related to late-onset cytopenias, resembling primary autoimmune ones, often responding to immunosuppressive drugs [54]. A case report about the use of romiplostim for thrombocytopenia following CAR-T cells in a patient affected by relapsed/refractory diffuse large B-cell lymphoma was recently published [55] (Table 3). Some ITP patients have been reported to benefit from switching from one TPO-RA to another. However, no data are available in the setting of aplastic anemia post-chemotherapy or after HSCT, except for a case switched from eltrombopag to romiplostim after umbilical stem cell transplant [56].
Regarding safety data, TPO-RA appear to have a good profile also in these off-label settings. From the meta-analysis of Yao et al., the pooled rate of adverse events (AE) was 3% (95%CI: 1-7%), with no severe AEs reported [34]. Similarly, the phase II study conducted by Ahmed S et al. highlighted no differences in grade 2 or higher AE rates between the eltrombopag and placebo arm, particularly regarding bone marrow fibrosis, cataract, vision changes, or thrombosis [39]. Finally, we identified only one case of pulmonary embolism secondary to eltrombopag therapy, while the most frequent reported AE was a transient elevation of liver function tests.
Research on ClinicalTrials.gov (accessed on 22 January 2022) returned 29 registered trials regarding the use of TPO-RA in aplasia post-chemotherapy for hematologic disease and post-HSCT (Table 4). The results will be important to better define efficacy, dosage, timing, and side effects of these drugs in these off-label settings.

Immunomodulatory Effects of Eltrombopag
Several mechanisms of action have been proposed to explain the immunomodulating properties of eltrombopag. The most agreed hypothesis is the restoration of an antigen-specific tolerance through induction of T cell anergy by increased exposure to PLT antigens [57]. More interesting, eltrombopag appeared to be able to recover immune tolerance by improving T-regulatory cells (T-regs) function and TGF-α levels [58]. The latter are key immunomodulatory factors inducing the suppression of self-reactive clones, but also contributing to bone marrow microenvironment homeostasis. In fact, the TGF-beta superfamily is implied in the regulation of late-stage erythropoiesis that is affected in patients with aplastic anemia and bone marrow failure. These observations are also supported by a recent study by Lozano et al., suggesting a possible role of platelet-derived ectosomes (a type of microparticles) in inducing T-regs differentiation in TPO-RA treated patients, and the involvement of additional cellular pathways (i.e., FOXP3, PTPN1, and PPARγ) that may modulate immune response and induce immune tolerance [57]. Moreover, eltrombopag was shown to restore Fcγ receptor (FcγR) balance through the inhibition of FcγRIIb on monocytes, mitigating phagocytic capacity, often hyperactivated in ITP patients [59]. In this context, the drug may also promote the switch of macrophage phenotype from the M1 proinflammatory type to the M2 anti-inflammatory type, dampening autoinflammation and disease chronicization [60]. Finally, eltrombopag may exert immunomodulating properties by promoting B regulatory cells' function [61]. An additional mechanism related to autoinflammation and bone marrow failure is the interference of eltrombopag with iron metabolism [62][63][64][65][66]. In fact, the drug exhibits iron-chelating properties that may alleviate bone marrow stress and improve hematopoiesis by avoiding iron-induced damage mechanisms, i.e., reactive oxygen species (ROS) generation. Concurrently, iron chelation could contribute to restoring immune functions, often impaired by disturbance of iron homeostasis [5,67].
The new insights on immunomodulatory effects of eltrombopag are extremely interesting even though most of the current evidence is limited to ITP. Thus, further studies would be important to evaluate these mechanisms in other settings such as primary bone marrow failures (AA and MDS) and aplasia secondary to chemotherapy and HSCT, where the degree of immune activation and the composition of bone marrow microenvironment are different.

Discussion and Conclusions
We presented a case of prolonged bone marrow aplasia secondary to ASCT treated with eltrombopag at the same dosage used for AA. During treatment, our patient did not show neither hepatologic toxicity nor thrombotic events and bone marrow evaluations did not display an increase in fibrosis. Our observations are consistent with the several evidence about the efficacy and safety of TPO-RA in post-HSCT and aplasia post chemotherapy, where the reported pooled response is about 70%, similarly to primary ITP.
One of the major concerns about the use of TPO-RA, particularly in the oncologic setting, is clonal evolution. In our case, sequential BM evaluation did not show an increase in blasts percentage, nor the emergence of cytogenetic aberrations; furthermore, the detection of RB1 gene mutation by NGS was not confirmed in the subsequent analysis and PNH clone testing was negative. In the literature, the risk of clonal escape is controversially reported for AA and MDS, which, however, carry an intrinsic risk of leukemic evolution [68][69][70]. In this context, a recent trial analyzing the efficacy and safety of eltrombopag added to immunosuppressive therapy in severe AA, also collected sequential mutational data by NGS, and did not detect clonal evolution after a median follow-up of 24 months [71].
Another concern with eltrombopag is the increase in reticulin bone marrow fibrosis. Although specific data in the setting of aplastic anemia following chemotherapy, HSCT, or novel anti-cancer drugs are not available, several studies in ITP and primary aplastic anemia demonstrated that it is mainly low-grade fibrosis, remitting after TPO-RA discontinuation [3,72,73]. Notably, our patient and many of those described in the literature were able to discontinue TPO-RA after robust recovery of cytopenias and bone marrow cellularity. This is also reported for ITP patients in about 1/3 of cases and is likely linked to an immunomodulatory effect of the drug.
Accordingly, in our study, we demonstrated an immunomodulatory effect of eltrombopag on different cytokines. Particularly, we observed that the patient's IL-6 and IL-10 levels were increased compared to controls, in agreement with the several reports of increased levels in plasma cell dyscrasias [74,75]; additionally, IL-10 is a negative regulator of Th1 cytokines and might have accounted for the reduced levels of TNF-α and IFN-γ observed during eltrombopag therapy. Consistently, the reduction of the latter two cytokines may have resulted in decreased inhibition of Th 2 response with an amplification of the autoimmune process. Regarding immunoregulatory cytokines, IL-17 was higher than controls, further amplifying the proinflammatory and autoimmune response. Interestingly, the inhibitory cytokine TGF-β, which was initially at the lower limit of normal, increased over time, in an attempt to downregulate the pro-inflammatory/autoimmune response. This might be related with a restoration of T-regs/T-helper 17 cells ratio already described in patients treated with eltrombopag [58].
Another interesting finding is the histologic demonstration of a polyclonal double negative T-cells infiltrate in patient bone marrow, increasing over time in parallel with bone marrow cellularity. These cells belong to a neglected class of T cells called CD3+CD4-CD8and have been described to play a homeostatic role in suppressing exuberant immune responses [76]. Along with the immune reconstitution after ASCT, this observation may further suggest an immunomodulatory effect of eltrombopag in this setting.
Finally, the new insights on the immunomodulatory effects of TPO-RA are intriguingly expanding the spectrum of use of these drugs. In the near future, treatment of "peripheral" (ITP) and "central" (AA, MDS) autoimmune diseases will take advantage of TPO-RA to either stimulate the stem cells compartment and exert a slower, but likely less toxic, immunomodulatory effect.
In conclusion, the case presented and the literature review highlight the efficacy and safety of eltrombopag in aplastic anemia following ASCT, HSCT, and chemotherapy for hematologic malignancies. Furthermore, cytokine studies and bone marrow findings support the emerging immunomodulatory effect of the drug. Altogether, growing evidence suggests eltrombopag use in these off-label settings, particularly in the case of prolonged cytopenias following poor graft function or chemotherapy.

Institutional Review Board Statement:
The study was conducted in accordance with the Declaration of Helsinki, and approved by the Ethics Committee of Comitato Etico Milano Area 2 (protocol code 2093 and date of approval 9 March 2021).