Acute myeloid leukemia (AML) is a heterogeneous clonal hematopoietic progenitor cell disorder characterized by immature myeloid cell proliferation and bone marrow failure, exhibiting a spectrum of morphological, immunophenotypic, cytogenetic and molecular characteristics (1).

Moreover, AML is an aggressive disease with a poor prognosis (2,3). In addition, >50% of patients with AML are not candidates for intensive chemotherapy therapy due to their age, performance status and/or associated comorbidities (4). Although the long-term overall survival (OS) rates of patients <65 years of age have significantly improved over the past years owing to improved supportive care and an increased use of allogeneic hematopoietic stem cell transplantation (alloHSCT), the prognosis for the elderly AML population is still poor, with a 5-year OS rate of <10% (5).

Furthermore, two-thirds of patients with AML who achieve a complete remission (CR) will relapse within the following 18 months (6), and regrettably, there are no safe and effective curative treatments, apart from alloHSCT, which is a rather aggressive therapeutic modality with high treatment-related morbidity and mortality (5). Therefore, given the significant incidence of relapsed AML and the frequent toxicities associated with standard intensive chemotherapy, an optimal treatment strategy for this population remains unsatisfactory and has yet to be established (4,7). In addition, although several new drugs for the treatment of AML, particularly for elderly patients, have been approved in recent years, such as the FMS-like tyrosine kinase 3 inhibitors, venetoclax, glasdegib or Vyxeos, the medical needs of patients with relapsed or refractory (RR) AML continue to be unmet (1,2,8,9).

Signal transducer and activator of transcription (STAT) is a seven-member family group of latent cytoplasmic transcription factors that are involved in hematopoietic cytokine receptor signaling pathways that mediate several biological processes, such as cell proliferation, differentiation, survival and immune response, by transferring signals from cell-surface cytokines and growth factor receptors to the cell nucleus and subsequently by regulating the transcription of target genes (10). The persistent and aberrant activation of specific STAT factors, particularly STAT3, often results in the growth and survival of tumor cells and, consequently, in the development of a wide range of cancers (11). STAT3 is the STAT family member most strongly associated with tumorigenesis. There are two main mechanisms through which STAT3 promotes tumorigenesis: By acting as a nuclear transcription factor (12) and as a regulator of oxidative phosphorylation (OXPHOS) via interaction with components of the electron transport chain (13).

STAT3 is constitutively activated in leukemic cells from patients with AML. It is now clear that the activation of STAT3 contributes to the development and resistance of AML (10). Furthermore, the assessment of bone marrow samples from patients with AML has demonstrated that the activation of STAT3 is significantly associated with a reduced OS and progression-free survival (PFS) (14).

It has been demonstrated that the blockade of aberrant STAT3 signaling induces tumor cell apoptosis and inhibits tumor growth, confirming its critical role in the molecular pathogenesis of several tumors. It has also been identified as a potential target for the discovery and development of novel anticancer drugs (10,15).

OPB-111077 is a novel orally bioavailable low-molecular-weight compound discovered and developed by Otsuka Pharmaceutical Co., Ltd. as an orally active antitumor agent for the treatment of various types of cancer. In preclinical analyses, it has been shown to be a potent and highly specific STAT3 inhibitor with a good efficacy and safety profile, supporting the initiation of early clinical investigation in humans (16). In fact, a first in-human study of OPB-111077 demonstrated that it could be administered safely, and its pharmacokinetic profile was acceptable for further clinical development (16). Mechanistic analyses have demonstrated that OPB-111077 significantly inhibits the STAT3 activation pathway, with antitumor effects against a wide range of human solid and blood tumor cell lines. Furthermore, OPB-111077 has been shown to exhibit efficacy against several solid and blood cancers both in vitro and in vivo (16,17).

Although in a phase I study, the activity of OPB-111077 in a range of solid tumors was limited, this drug exhibited clinical activity in one subject with diffuse large B-cell lymphoma (16), and hence, it could be more efficient in tumor cells with a high proliferative index, such as AML.

Biomarker-based treatment selection is a popular topic in oncology. However, few successful biomarkers have been discovered thus far, with the majority of these being molecular, such as tyrosine kinase inhibitors in chronic myeloid leukemia (18).

Previous studies have analyzed the association between ex vivo drug testing and clinical outcomes in adult patients with AML. Functional ex vivo assays that predict a patient's clinical response to anticancer drugs for guiding cancer treatment have long been a goal, but few have yet proven to be reliable (19,20).

The present study conducted a phase Ib dose-escalation and biomarker-driven study to assess the safety and efficacy profiles of OPB-111077 in patients with RR AML. In order to identify and select the subpopulation most sensitive to the study drug and optimize disease management, a precision medicine, personalized, ex vivo test was first performed that evaluated the pharmacological activity of OPB-111077 directly in individual patient bone marrow samples.

To analyze the mechanisms of action of the compound, in a preliminary preclinical phase, a total of 19 patients with AML were analyzed at the Vivia Biotech laboratories in a proliferation assay. This was the starting point to further expand the number of samples reaching statistical significance and converging in population models in order to achieve a better characterization of OPB-111077. As shown in Fig. 1, OPB-111077 exerted anti-proliferative rather than cytotoxic activity, as it exerted a more prominent effect on proliferating cells than on the population of non-proliferative cells. A comparison between a reference anti-proliferative approved drug in AML, decitabine, was performed. Population dose response curves of the proliferating cells were generated using both the novel OPB-111077 compound and decitabine. The pharmacological profiles revealed a high interpatient variability in the patient samples incubated with OPB-111077 and in those incubated with decitabine (Figs. S2 and S3), suggesting the need for a precision medicine (PM) test to select the best patient candidates. The overlapping population curves of the proliferating cells showed similar activity of OPB-111077 vs. decitabine.

Once the pharmacodynamic model of OPB-111077 in the AML patient samples was established, a phase Ib investigator-sponsored trial using this assay as a selection criterion was launched. A total of 47 patients with RR AML were screened, and 12 were ultimately enrolled in the study between September 7, 2017 and March 31, 2020 at three Spanish sites (Fig. S1): Hospital La Fe (Valencia), Hospital 12 de Octubre (Madrid), and Hospital San Pedro de Alcántara (Cáceres). In total, 26 patients were excluded using the personalized medicine sensitivity test, as their results were below the primary acceptance criteria of the 70th percentile.

Dose response curves of OPB-111077 in bone marrow samples from the screened subjects, highlighting those sensitive and resistant treated patients, are displayed in Fig. 2A. A stratification based on the percentile AUC and represented in a heatmap was performed to aid in the selection of patients to be included in this phase Ib clinical trial (Fig. 2B). Few samples crossed the sensitive (green) vs. resistant (red) threshold. These samples near the threshold may have slightly shifted their activity from 72 h shown to 96 h or 120 h (data not shown), which were also measured and could serve to decide on patient inclusion.

The patient demographics and baseline clinical characteristics are summarized in Table I. The median age was 76 years, and 91.7% were male. No differences were observed in the patient screening failure, except in the frequency of the NPM1 mutation (Table SII). A total of 5 (42%) patients with AML were refractory; the median (range) of relapse was 2 (1-6) (Table I). In addition, 5 (42%) patients were treated with the first level dose (DL1) of OPB-111077 (200 mg), while 7 patients (58%) were escalated the second dose level (DL2) of 250 mg. The median total doses administered were 17,000 mg and 8,250 mg for DL1 and DL2, respectively. The study treatment dose was only reduced in 1 patient treated with DL2.

The present phase I dose-escalation trial was performed to assess the safety, tolerability and efficacy of OPB-111077 in patients with RR AML treated with doses ranging from 200-250 mg/day for 4 weeks.

In the current trial, no DLTs were observed, and therefore, the MTD (primary study endpoint) was not reached, confirming the good safety profile of OPB-111077. This good safety and tolerability profile has also been reported in previously published studies with OPB-111077 (16,17). Likewise, the most frequently reported TEAEs were nausea, vomiting and fatigue.

Although the clinical activity (i.e., an ORR of 25%) may be considered modest (25,26), it was much higher than the response observed in the aforementioned published phase I studies with OPB-111077 (i.e., an ORR of 1/145) (16,17). It is also an even higher response rate compared to other new drugs with different mechanisms of action, such as MDM2 antagonist RO6839921. Uy et al (27) reported a response rate of 7.7% in their phase 1 study. The same occurred in the phase 1 study on CWP232291, in which Lee et al (28) described a low number of responses. However, it should be noted that the patients included in the present clinical trial had a very poor prognosis; they were elderly (many of them >70 years of age), a difficult-to-treat population (29), and the majority were refractory to standard therapy (30). Tolcher et al (16) reported clinical activity (durable PR) in only one subject, with diffuse large B-cell lymphoma, from a population of 18 patients with unselected and mostly solid tumors, while in the study conducted by Yoo et al (17), no patients with hepatocellular carcinoma achieved complete or partial responses with OPB-111077. A plausible explanation for this finding is that, unlike other phase I trials, in the present study, the population was selected based on a biomarker that enabled the upfront identification and enrollment of those AML patients with the highest sensitivity to the study drug, discarding those hypothetically resistant ones and thus minimizing the likelihood of treatment failures. This is supported by the differences in both the AUC and EC50 values that were found between patients achieving PR or TF as the optimal responses. However, those differences did not meet the statistical significance criteria, probably due to the small sample size. Other research groups have also implemented this ex vivo personalized medicine sensitivity test in the AML population to improve prognostic risk stratification, tailor treatments, and minimize drug resistance. As in the current analysis based on the expression of a biomarker, other researchers have found strong correlations between the ex vivo sensitivity test and the clinical response to chemotherapy in AML patients in their respective studies (31,32).

One of the mechanisms through which STAT3 promotes oncogenesis is through the activation of OXPHOS (16). Of note, OXPHOS has been reported to be involved as a mechanism of resistance to chemotherapy in AML (33). Therefore, the use of drugs targeting OXPHOS may be an appropriate therapeutic approach for the treatment of refractory and relapsed AML (34,35). Other drugs have been proposed to function through the OXPHOS of leukemic cells, such as IACS-010759(36) and ME-344(37) (4). However, in contrast to OPB-111077, phase 1 studies of the use of these drugs in relapsed/refractory acute myeloid leukemia have not yet been conducted.

As demonstrated in the present study, drugs such as decitabine, similar to OPB-111077, exert an anti-proliferative effect on tumor cells. Therefore, the combination of both can increase anti-tumor activity. In this regard, the authors of an ongoing trial evaluating the combination of OPB-111077 with decitabine and venetoclax for the treatment of AML have suggested that the combination of OPB-111077 and venetoclax reduces tumor cell proliferation and increases apoptosis rates to a greater extent than exposure to any single study drug (38). Notably, the effects obtained with the combination were even more pronounced in AML cells that were genetically engineered to increase OXPHOS (38). Pollyea et al (39) also demonstrate that the combination of venetoclax and a hypomethylating agent such as azacitidine can eradicate leukemic cells by disrupting energy metabolism through suppression of OXPHOS. This is in line with the similar activity and weak toxicity found in the preliminary preclinical study we performed, which may suggest a similar clinical profile; thus, their use in combination could increase the chance of achieving an overall response.

Certain limitations of the present study are the small number of patients included, although this is due to of the nature of a phase I clinical trial and the strategy used for patient selection. The employment of an ex vivo test for selection could hinder patient treatment in this aggressive disease.

In conclusion, OPB-111077 as a monotherapy has exhibited a good safety and tolerability profile in patients with RR AML. Additionally, some clinical response was found compared to previous studies performed with the same study drug (16,17). The innovative biomarker-driven design used in the present study to select the patient population upfront based on their sensitivity to the study drug may partly explain these improved results over previous studies. This innovative phase IB biomarker selection design may help to lower the high attrition rate of new drugs.