Pixantrone: novel mode of action and clinical readouts

ABSTRACT

Introduction: Pixantrone is a first-in-class aza-anthracenedione approved as monotherapy for treatment of relapsed or refractory aggressive diffuse B-cell non-Hodgkin’s lymphoma (NHL), a patient group which is notoriously difficult to treat. It has a unique chemical structure and pharmacologic properties distinguishing it from anthracyclines and anthracenediones.

Areas covered: The chemical structure and mode of action of pixantrone versus doxorubicin and mitoxantrone; preclinical evidence for pixantrone’s therapeutic effect and cardiac tolerability; efficacy and safety of pixantrone in clinical trials; ongoing and completed trials of pixantrone alone or as combination therapy; and the risk of cardiotoxicity of pixantrone versus doxorubicin and mitoxantrone.

Expert commentary: Currently, pixantrone is the only approved therapy for multiply relapsed or refractory NHL, an area with few available effective treatment options. Pixantrone is currently being investigated as combination therapy with other drugs including several targeted therapies, with the ultimate goal of improved survival in heavily pretreated patients. In order for pixantrone to be acknowledged in the treatment of aggressive NHL, the perception of pixantrone as an anthracycline-like agent that has anthracycline-like activity and cardiotoxicity needs to be changed. Further data from ongoing clinical trials will help in confirming pixantrone as an effective and safe option.

KEYWORDS:

• Anthracycline
• cardiotoxicity
• mechanism of action
• non-Hodgkin’s lymphoma
• pixantrone

1. Introduction

Diffuse large B-cell lymphoma (DLBCL) is the most common type of non-Hodgkin’s lymphoma (NHL), accounting for a significant proportion (about 25–30%) of all NHLs [1]. Despite major advances in our understanding of the biology of NHL [1] and current standards of frontline treatment which help patients achieve long-term disease remission or cure, a significant proportion of patients still relapse or become refractory to treatment [2].

Management of relapsed or refractory NHL remains a challenge. Guidelines suggest that patients with relapsed or refractory NHL be given platinum-based salvage regimens, and if they respond well, treatment with high-dose chemotherapy and autologous stem-cell transplantation may be considered as a recommended option in eligible patients [3]. Salvage regimens recommended include R-DHAP (rituximab, cisplatin, cytarabine, and dexamethasone), R-ICE (rituximab, ifosfamide, carboplatin, and etoposide), and R-GDP (rituximab, cisplatin, gemcitabine, and dexamethasone) [3].

However, there is no currently recognized standard of care for patients who have failed first- and second-line treatment and for whom stem-cell transplantation is not an option [4]. Continued use of anthracycline-based chemotherapy is limited by cumulative dose toxicity, and while several salvage regimens may be considered in patients with aggressive multiply relapsed/refractory NHL, re-treatment is rarely curative [5]. Regimens investigated in patients with relapsed/refractory aggressive NHL/DLBCL include R-ICE, R-GEMOX (rituximab, gemcitabine, and oxaliplatin), R-lenalidomide, or R-bendamustine [4].

Pixantrone, an aza-anthracenedione, is one of the treatment options for heavily pretreated NHL patients. It is the first monotherapy approved by the European Medicines Agency as third- or fourth-line treatment for aggressive B-cell NHL [6,7]. The evidence for its efficacy in heavily pretreated patients is acknowledged in current European Society of Medical Oncology guidelines [3], and it is recommended as a third- or fourth-line treatment option in the UK [8]. While pixantrone is commonly perceived to be an anthracycline-like drug, it has a unique chemical structure and distinct pharmacologic properties. This narrative review explores the mode of action of pixantrone, contrasting it with that of the anthracycline doxorubicin and the anthracenedione mitoxantrone, and also provides a brief overview of pixantrone’s efficacy and safety profiles.

2. Chemical structure of pixantrone

The chemical structure of pixantrone is unique compared with other quinone-hydroquinone anthracycline-like drugs. It is different from the prototypic anthracenedione mitoxantrone and the anthracycline doxorubicin [9]. Anthracenediones are a class of synthetic anticancer compounds that differ from doxorubicin and other anthracyclines in that they do not have an amino sugar moiety (daunosamine) and have only three molecular rings compared with four in doxorubicin and other anthracyclines [9]. Pixantrone in turn differs from mitoxantrone by having no hydroquinone, a nitrogen heteroatom inserted in the ring lacking the hydroquinone, and a substitution of (ethylamino)diethylamino for (hydroxyethylamino)-ethylamino side chains [9]. Figure 1 shows the chemical structure of pixantrone compared with doxorubicin and mitoxantrone.

Figure 1. Chemical structures of doxorubicin (DOX), mitoxantrone (MITOX), and pixantrone (PIX).

The arrows indicate that pixantrone differs from mitoxantrone in the lack of the hydroquinone, insertion of a nitrogen heteroatom in the same ring, and substitution of (ethylamino)-diethylamino for (hydroxyethylamino)-ethylamino side chains. Reproduced with permission from Menna et al. [9]. Copyright 2016 American Chemical Society.

3. Mechanism of action

The unique chemical structure of pixantrone results in a different mechanism of action compared with anthracycline doxorubicin and anthracenedione mitoxantrone, as summarized in Figures 2 and 3 (online video).

Figure 2. Simplified representation of proposed mechanism of action of pixantrone (PIX), doxorubicin (DOX) and mitoxantrone (MITOX).

CYP P450, cytochrome P450; Fe, iron; MPO, myeloperoxidase; REDOX, reduction-oxidation; ROS, reactive oxygen species; TOPOII, topoisomerase II.Question marks indicate uncertainty.Data for PIX are from Adnan et al 2010 [24], Evison et al 2007 [23], Evison et al 2008 [22], Evison et al 2009 [21], Beeharry et al 2015 [25] and Ng et al 2017 [20]. Data for DOX are from Patel & Kaufman 2012 [10], Sawyer 2013 [12] and Zhang 2012 [46]. Data for MITOX are from Atwal et al 2017 [47], Faulds et al 1991 [16] and Damiani et al 2016 [48]. © Les Laboratoires Servier, 2018 (published with permission)

3.1. Mechanism of action of doxorubicin and mitoxantrone

Doxorubicin is an anthracycline antibiotic primarily used in the treatment of human cancers [10]. Multiple mechanisms of anthracycline activity (tumor cell killing) have been proposed, but interaction with DNA-topoisomerase II complex is thought to be the primary mechanism for inducing cell death [11]. Tumor growth is disrupted when anthracyclines bind to and block the enzyme topoisomerase II, thus disrupting its activity and activating DNA damage response pathways that lead to cell death [10,12]. Topoisomerase II is an essential enzyme that plays a vital role in DNA replication and transcription as well as chromosome separation and segregation [13]. It catalyzes topological changes in DNA by generating transient double-strand breaks after forming a topoisomerase II-DNA cleavage complex [13]. When anthracyclines bind to topoisomerase II, they intercalate into the DNA molecule and stabilize the DNA-topoisomerase II cleavage complex; consequently, transient double-strand breaks become permanent protein-capped breaks leading to mutations that trigger cell apoptosis [13]. It is also proposed that p53 inhibits topoisomerase II in tumor cells exposed to doxorubicin leading to irreversible DNA damage and cell death [14].

Like doxorubicin, mitoxantrone is a topoisomerase II inhibitor and DNA intercalator [15]. Its mechanism of action is broadly similar to that of doxorubicin [13], with some differences, since it also inhibits RNA synthesis and induces nonprotein-associated strand breaks [16].

3.2. Advances in the understanding of pixantrone’s mechanism of action

Pixantrone differs from doxorubicin and mitoxantrone in its unique chemical structure, and this results in differences in its mechanism of action compared with that of doxorubicin and mitoxantrone. Although pixantrone and anthracycline-like drugs have the potential to cause DNA intercalation, the main cytotoxic mechanism of action of pixantrone is distinct from doxorubicin and mitoxantrone, as described below.

3.2.1. Weak topoisomerase II inhibition

Pixantrone is only a weak inhibitor of topoisomerase II [7], unlike doxorubicin and mitoxantrone. Indeed, early evidence suggested that pixantrone was able to intercalate into DNA but caused low levels of double-strand breaks in human tumor cells [17]. Other studies showed that DNA intercalation and binding of topoisomerase II leading to a stabilized DNA-topoisomerase complex and then DNA double-strand breaks did not correlate with pixantrone’s potency as a cytotoxic agent [18]. Moreover, p53 was not essential for pixantrone to induce cell death [19]. Topoisomerase II inhibition is less efficacious with pixantrone than with doxorubicin, as indicated by less decatenated DNA product after treatment in NHL cells [20].

3.2.2. DNA alkylation and covalent adduct formation

Pixantrone intercalates the major and minor grooves of DNA resulting in the formation of DNA adducts [21–24]. Removal of the 5,8-substituents (hydroquinone moiety) and introduction of a nitrogen heteroatom into the pixantrone chromophore might lead to improved hydrogen bonding and thus potentially greater affinity for DNA [23]. Pixantrone is reported to display a 10- to 100-fold greater propensity to form DNA adducts compared with mitoxantrone [23]. Formaldehyde, a by-product of lipid oxidation that is found in cancer cells, is an important facilitator of pixantrone covalent binding to DNA, i.e. DNA alkylation [9]. The pixantrone-DNA adducts are more stable than the mitoxantrone-DNA adducts after formaldehyde activation [23]. Pixantrone alkylates DNA through the amino structures in each side chain of the drug forming covalent pixantrone-DNA adducts. These adducts are formed at the N2 amino group of guanine in the DNA dinucleotide and are mediated by a single methylene linkage provided by formaldehyde [22]. Pixantrone also forms two- to five-fold more covalent adducts at the 5ʹ-CpG dinucleotide sites when cytosine (C5-position) is methylated [21]. Moreover, cancer cells deficient of CpG methylation are 12-fold less sensitive to pixantrone relative to the wild-type cells suggesting specific pixantrone activity for DNA hypermethylation site [21].

3.2.3. Cell death after successive rounds of aberrant mitosis

In vitro experiments show that pixantrone has little or no effect on cell cycle at therapeutic concentrations (i.e. equal to or lower than plasma C_max) [25]. Tumor cells exposed to pixantrone undergo mitosis, but with aberrations [25]. Defective kinetochore attachments lead to mis-segregation of chromosomes and generation of micronuclei in pixantrone-treated cells, and cell death occurs only after successive rounds of aberrant mitosis [25]. At the American Association for Cancer Research (AACR) 2017 Annual Meeting, Ng and colleagues presented a study conducted in NHL cell lines further showing that pixantrone (1–100 nM) impairs chromosomal segregation (as indicated by micronuclei formation) without triggering mitotic checkpoint activation [20]. Pixantrone induces less DNA double-strand breaks compared with doxorubicin, as quantified by the number of γH2AX positive foci, although the long-term antitumor effect, measured by long-term clonogenic assays, is similar for the two drugs [20].

4. Pixantrone: preclinical evidence

Preclinical studies with pixantrone demonstrated anticancer activity, particularly in hematologic malignancies, where pixantrone generally showed equal or superior activity versus other drugs [26,27]. Preclinical studies of pixantrone as monotherapy suggested a therapeutic effect similar to doxorubicin and mitoxantrone; however, less cardiotoxicity was observed (Table 1) [27–31]. This was attributed to the removal of the 5,8 substituents (hydroquinone moieties) from pixantrone’s chemical structure, which makes pixantrone unable to bind iron and in turn promote iron-catalyzed oxidative damage. Figure 4 illustrates the comparative cardiotoxicity of repeated treatment cycles with doxorubicin or mitoxantrone resulting in marked degenerative cardiomyopathy compared with minimal changes associated with pixantrone treatment [31].

Table 1. Preclinical profile of pixantrone monotherapy.

Efficacy [27]
In vitro murine and human tumor cell lines	Pixantrone cytotoxic potency < MITOX, DOX
	Pixantrone cytotoxic potency in leukemia and lymphoma cell lines > solid tumor cell lines
In vivo murine models of leukemia and lymphoma	Pixantrone curative
In vivo murine models of solid tumors	Pixantrone antitumor activity = MITOX, DOX
Cardiac effects
In vitro rat embryo myocardial cell line [28]	Pixantrone is less cardiotoxic than DOX (ID50 > 50 µg/mL vs. 1 µg/mL) No significant ROS production with pixantrone compared with DOX
In vitro primary rat cardiac myocytes [29]	No significant reduction in mitochondrial membrane potential (a potential mechanism for cardiotoxicity) in pixantrone-treated cells Less myocyte toxicity after pixantrone exposure versus DOX and MITOX
In vivo murine models: DOX-naïve	Pixantrone cardiotoxicity negligible [27,30] DOX cardiotoxic [30]
In vivo murine models: anthracycline pretreated [31]	Preexisting cardiomyopathy unchanged with pixantrone but worsened with MITOX and DOX

= similar to; > greater than; DOX: doxorubicin; ID₅₀: inhibitory dose (50%); MITOX: mitoxantrone; ROS: reactive oxygen species

Figure 3. Mode of action video. Available online at Supplemental data. © Les Laboratoires Servier, 2018 (published with permission).

Figure 4. Morphologic evaluation of cardiac lesions in mice following repeated treatment cycles of either doxorubicin (Dx), mitoxantrone (Mx) or pixantrone (Pix).

The mean total score (MTS) represents the product of the severity and degree of extension of myocardial lesions evaluated on a median section of whole heart and scored according to literature methods.*p < 0.05; ***p < 0.001.Reproduced with permission from Cavalletti et al. [31]. Copyright 2007 Springer International Publishing AG.

5. Pixantrone for relapsed or refractory aggressive NHL

5.1. Efficacy

Promising results from a phase II trial in 33 patients mainly with relapsed aggressive NHL (mostly DLBCL) [32] prompted investigation of pixantrone in a pivotal phase III trial. This phase III trial of pixantrone in adult patients with relapsed or refractory aggressive NHL demonstrated its efficacy and manageable toxicity in this difficult-to-treat patient population [33] and confirmed a place for pixantrone as a useful treatment option in these patients [34]. The multicenter open-label phase III trial randomly assigned patients with relapsed or refractory aggressive NHL (DLBCL; transformed indolent lymphoma; peripheral T-cell lymphoma, not otherwise specified; primary anaplastic large-cell lymphoma, null cell type; follicular lymphoma, grade 3) to intravenous pixantrone 50 mg/m² (85 mg/m² pixantrone dimaleate) on days 1, 8, and 15 of a 28-day cycle (n = 70) or comparator drugs (n = 70; vinorelbine, oxaliplatin, ifosfamide, etoposide, mitoxantrone, gemcitabine, or rituximab) at prespecified standard doses and schedules [33]. Patient demographic and clinical characteristics were mostly balanced at baseline including previous anthracycline cumulative dose (median doxorubicin dose equivalent of 293 versus 316 mg/m² for pixantrone and comparator groups, respectively). However, three pixantrone recipients had a history of congestive heart failure and two had uncontrolled cardiomyopathy [33].

Significantly more patients treated with pixantrone experienced a complete response/unconfirmed complete response (CR/CRu) at the end of treatment (primary efficacy end point; Figure 5) compared with comparator drugs [33]. CR and overall response rates (ORR) were higher with pixantrone than comparator regimens (see Figure 5). Pixantrone was associated with longer progression-free survival (p = 0.005) and overall survival (p = NS, not significant) compared with comparator groups (5.3 vs. 2.6 months; hazard ratio [HR] 0.60 [95%CI 0.42–0.86], and 10.2 vs. 7.6 months; HR 0.79 [95%CI 0.53–1.18], respectively) [33].

Figure 5. Efficacy of pixantrone monotherapy in patients with refractory or relapsed aggressive NHL in a phase III trial.

*p < 0.05, **p < 0.01 vs. comparator agent group [33].CR, complete response; CRu unconfirmed CR; ORR, overall response rate.

Potential weaknesses of this phase III trial have been discussed previously [8,35]. The trial recruited fewer patients than was originally planned and only 54% of those in the pixantrone and 56% in the comparator group had received prior therapy with rituximab, which became a standard of care while the trial was already ongoing. Moreover, the majority of patients presented with relapsed rather than refractory disease. In a retrospective, real-world analysis of high-risk patients with relapsed or refractory DLBCL, pixantrone has shown only limited activity and, therefore, the authors have called for a further evaluation [36]. To address these concerns, post hoc subgroup analyses of the phase III trial were conducted, which confirmed the superior efficacy of pixantrone as a salvage therapy. Among 97 patients with aggressive B-cell lymphoma (DLBCL, transformed indolent lymphoma and follicular lymphoma, grade 3 confirmed by blinded centralized review) who were receiving study drug as their third or fourth line of therapy had an ORR of 43.6% versus 12.8% (p = 0.005) and a CR/CRu of 23.1% vs. 5.1% (p = 0.047) for pixantrone versus the comparator agent group [37]. Within this subgroup, analysis of those who had previously received rituximab indicated that treatment with pixantrone was associated with a better response than comparator drug (ORR 45% vs. 11.1%; p = 0.033) [37].

5.2. Safety

Adverse events associated with pixantrone were manageable. These were consistent with expectations of cytotoxic therapy in heavily pretreated patients [33]. In the phase III trial described, grade 3 or 4 adverse events were reported in 76% of the pixantrone group and 52% of the comparator agent group, of which neutropenia and leukopenia were the most common adverse events [33]. Pixantrone-related neutropenia was usually transient, and the severity of neutropenia did not increase over time. Based on preclinical evidence, myocardial effects were expected to be less frequent with pixantrone than with doxorubicin or mitoxantrone. However, cardiac adverse events were more common with pixantrone than with comparator drugs (35.3% vs. 20.9%). The authors noted that baseline cardiac morbidity profiles were different between the study groups and may have contributed to these cross-group differences. In addition, as previously mentioned, more patients in the pixantrone arm presented preexisting heart failure or uncontrolled cardiomyopathy. Moreover, the frequency of these events did not increase with increased pixantrone exposure and they were mostly asymptomatic reductions in left ventricular ejection fraction of grade 1 or 2 severity [33]. This is a remarkable finding. When pixantrone is converted to doxorubicin equivalents, it transpires that the majority of patients were exposed to a lifetime cumulative anthracycline dose (prior doxorubicin plus pixantrone) of >400 mg/m², which is usually associated with a 5% risk of heart failure [8]. Patients treated with six cycles of pixantrone were exposed to a lifetime cumulative anthracycline dose of approximately 600 mg/m² [38]. These facts denote the safety of pixantrone in patients with prior exposure to doxorubicin and form a rationale for evaluating pixantrone in special populations such as, for example, the vulnerable elderly.

5.3. Ongoing clinical trials

Pixantrone is currently being investigated as combination therapy with rituximab in the ongoing PIX306 randomized multicenter study in adult patients with relapsed or refractory DLBCL or follicular grade 3 lymphoma who are ineligible for high-dose (myeloablative) chemotherapy and stem-cell transplant (NCT01321541) [39]. Rituximab, which is the backbone of treatment in patients with CD20-positive NHL, has shown efficacy in patients with relapsed or refractory aggressive lymphomas, and is generally well tolerated [40]. The comparator regimen is gemcitabine plus rituximab; gemcitabine has shown promising efficacy in patients with relapsed or refractory lymphoma and has a favorable adverse event profile, making it amenable for use in combination therapy [41].

Additional ongoing trials of pixantrone in combination therapy regimens include the phase II GOAL study, which is investigating pixantrone in combination with obinutuzumab in patients with relapsed aggressive B-cell lymphoma (NCT02499003) [42], and a phase I study of pixantrone, bendamustine, and rituximab in patients with relapsed/refractory B-cell NHL (NCT01491841) [43]. See Section 7 for further discussion on the future of pixantrone in combination regimens.

6. Risk of cardiotoxicity and its mechanism

6.1. Doxorubicin and mitoxantrone

The development of doxorubicin-associated cardiomyopathy is closely linked to the administered dose and subsequent accumulation of doxorubicin in the heart [14].

Iron binding and reductive bioactivation, both resulting in the generation of reactive oxygen species (ROS) and ultimately leading to cellular dysfunction and cell death was the predominantly held theory for the cardiomyotoxicity associated with anthracyclines/anthracycline-like drugs [9,12]. ROS are formed after one-electron reduction of the quinone moiety to a semiquinone, which in turn regenerates its parent quinone by reducing oxygen to the ROS superoxide anion (O₂^–) and hydrogen peroxide (H₂O₂) [14]. It is proposed that ROS overproduction causes oxidative stress in cardiomyocytes by the formation of hydroxyl radicals (˙OH) from ROS via reactions requiring iron [9]. Anthracyclines and mitoxantrone bind Fe(II), forming drug-iron complexes [28]. While these complexes are unstable, they may react with oxygen to form stable drug-Fe(III) complexes, generating ROS in the process [9]. With doxorubicin, but not mitoxantrone, self-reduction of Fe(III) complexes may occur that regenerates Fe(II) and ROS [44].

In addition, the secondary alcohol metabolite of doxorubicin, doxorubicinol, is cardiotoxic and shows no cardiac clearance, and thus its accumulation in the heart increases the risk of congestive heart failure [45]. Anthracycline secondary alcohol metabolites are formed by two-electron reduction of the side-chain carbonyl moiety [14]. Since anthracenedione side chains lack carbonyl groups [45], secondary alcohol metabolite formation does not play a role in the cardiotoxicity of mitoxantrone and pixantrone compared with doxorubicin.

It is important to note that there is evidence that does not support the role of iron binding and ROS formation as the mechanism of anthracycline and anthracycline-like cardiotoxicity, as reviewed elsewhere [9]. For example, the administration of high-dose antioxidants to patients receiving doxorubicin did not prevent or reduce anthracycline-associated cardiomyopathy [14]. Moreover, preclinical cardiac protection by iron chelators did not correlate with their potency in chelating free iron or in removing it from anthracycline-iron complexes [9].

Anthracycline drug efficacy (killing of tumor cells) is mediated predominantly via topoisomerase IIα, which is expressed at high levels in tumor cells but not at all in cardiomyocytes [12]. With recent evidence for the role of topoisomerase IIβ being constitutively expressed in cardiomyocytes and other quiescent cells [12,46], an alternative mechanism of cardiotoxicity has been proposed: using a mouse model with cardiomyocyte-specific deletion of the Top2b gene (encoding topoisomerase IIβ), Zhang et al. demonstrated that these cardiomyocytes were protected from doxorubicin-induced cardiotoxicity [46]. They showed that doxorubicin, in the presence of topoisomerase IIβ, activates DNA damage and suppresses the transcription of genes involved in mitochondrial biogenesis and function, leading to mitochondrial damage. ROS generation would result from these effects, rather than from redox cycling of doxorubicin [46].

Some investigators have suggested different mechanisms of cardiotoxicity between doxorubicin and mitoxantrone. Mitoxantrone is not activated via reduction (unlike doxorubicin), but by oxidative metabolism (e.g. by myeloperoxidase [47] or cytochrome P450 [48] to free radical intermediates (quinone, quinonediimine) [48]. Mitoxantrone appears to have a weaker capacity to enter redox cycling generating ROS and it has been proposed that it induces energy imbalance [48], although further research is required to confirm this.

6.2. Pixantrone

Pixantrone appears to be less redox active (less potential for iron binding and ROS production) [28,29] and shows greater selectivity for topoisomerase IIα than topoisomerase IIβ in stabilizing enzyme-DNA covalent complexes [29], albeit at higher than clinically relevant drug concentrations [9]. In vitro assays have shown that pixantrone is more selective for topoisomerase IIα than topoisomerase IIβ at all tested concentrations (1–10 µM), which exceed plasma levels in patients; in contrast, mitoxantrone loses its selectivity for topoisomerase IIα at higher concentrations, and is more active at 10 μM (not achieved clinically) against topoisomerase IIβ [29]. Thus, pixantrone avoids the topoisomerase IIβ inhibition seen with anthracyclines or mitoxantrone that could cause cardiotoxicity.

Pixantrone is reported to not induce significant ROS production in rat embryo myocardial cell experiments in vitro and is far less toxic to these cells than doxorubicin [28]. While cell-free models have suggested high concentrations of pixantrone capable of producing semiquinone free radicals and thus having the potential to generate ROS [29], a study using an ex vivo human myocardial strip model exposed to clinically relevant concentrations of pixantrone found redox inactivity (no detectable H₂O₂). Notably, in the setting of prior doxorubicin exposure, mitoxantrone acts synergistically with the accumulated doxorubicin to form H₂O₂ whereas pixantrone did not increase myocardial levels of H₂O₂ [45]. Moreover, unlike mitoxantrone, pixantrone is reported to inhibit doxorubicinol formation from residual doxorubicin [45].

Taken together, these data show that pixantrone’s unique chemical structure allows it to behave differently than anthracyclines and mitoxantrone. Much preclinical evidence supports an inherently reduced cardiotoxicity of pixantrone, including in conditions of prior doxorubicin/anthracycline exposure (see Section 4). An open-label, multicenter, comparative phase II study that investigated the efficacy and safety of pixantrone substituted for doxorubicin as first-line treatment of NHL also reported promising evidence of cardiac safety with pixantrone [49].

7. Combination therapy: future prospects

While pixantrone has been investigated in phase I/II trials in several different combination regimens [49–52], none appear to have advanced to phase III trials or widely adopted in real-world practice [53]. This section focuses on potentially effective combinations under investigation or worthy of further study.

7.1. Pixantrone plus etoposide and bendamustine, with or without rituximab

Rational combination therapy seeks to maximize synergistic therapeutic effect and minimize clonal resistance while not significantly increasing toxicity. Ideal candidate drugs for use in combination with pixantrone would be non-anthracycline topoisomerase IIα inhibitors (since pixantrone only weakly inhibits topoisomerase IIα at clinically relevant concentrations) such as etoposide, or downregulators of mitotic checkpoints such as the antimetabolite alkylator bendamustine [35]. The PREBEN phase I/II trial is currently recruiting participants and aims at investigating the combination of pixantrone, etoposide, bendamustine, and, in CD20 positive patients, rituximab in patients with relapsed aggressive B- or T-cell NHL (NCT02678299) [54]. There is already evidence of potential efficacy for this combination from an observational study in 30 heavily pretreated patients with aggressive NHL reported at the American Society of Hematology conference in 2016 [55]. Substantial and durable responses were seen in some patients (ORR 50%), particularly early in the course of therapy, suggesting that it could be used as a bridging therapy to non-myeloablative allogeneic transplant in younger patients.

7.2. Preclinical evidence for novel combination therapies

A preclinical study reported that combination of pixantrone with the Bruton’s tyrosine kinase inhibitor ibrutinib or with the phosphoinositide 3-kinase-delta inhibitor idelalisib was particularly promising compared with combinations with other targeted agents (e.g. rituximab) or other chemotherapy agents (e.g. bendamustine, etoposide) [56]. Pixantrone plus ibrutinib was effective in all five ABC-DLBCL cell lines, demonstrating synergism in four cell lines and an additive effect in one cell line. For pixantrone plus idelalisib, activity was demonstrated against 10 cell lines out of 11 tested; of these, the effect of the combination was synergistic in nine cell lines, and additive in one cell line [56].

8. Conclusions

Pixantrone’s mode of action is specific to its unique chemical structure (see Figures 2 and 3, and mode of action video online). As a first-in-class drug, its therapeutic potential may have been underestimated in clinical practice [9,57]. Emerging data from pharmacological studies have shed light on its novel mode of action, which underscores its therapeutic effect as a monotherapy observed in the pivotal phase III trial in patients with relapsed or refractory aggressive NHL, and lack of severe cardiotoxicity that is associated with typical anthracyclines like doxorubicin. Pixantrone could be useful in combination therapies. With further data on the safety and efficacy of its combination with rituximab from the ongoing phase III clinical trial, it is hoped that pixantrone’s place as a useful therapeutic option in the treatment of NHL can be further expanded.

9. Expert commentary

The major gap in the available treatment options for NHL is that there is a lack of a standardized treatment regimen for patients with multiply relapsed/refractory aggressive NHL; currently, pixantrone is the only approved treatment that is indicated for third- or fourth-line therapy in NHL. Treatments truly tailored to individual patients remain elusive in this patient population.

NHL includes a diverse spectrum of immune-system cancers but next-generation gene-sequencing and gene-expression profiling is helping greatly to understand the pathogenesis of this disease, and thus reveal treatment opportunities [1]. The ultimate goal of treatment in the difficult-to-treat patients is to improve the dismal survival rates seen in this population. To this end, pixantrone is being explored in combination therapies, including newer targeted agents. One of the biggest challenges is overcoming any perceptions of pixantrone that may relate to its often reported anthracycline-like activity and cardiotoxicity. Further efficacy and safety data from ongoing clinical trials, especially in combination therapies, will help in this regard. Additional challenges include integration of the large volume of genomic data into clinical practice, for example defining which treatment is best for genetic subgroups of DLBCL [1].

10. Five-year view

If the new treatments under development (such as CAR-T cells, checkpoint inhibitors, and bispecifics, among others) show real, impressive results, the treatment of NHL, as well as other malignancies, will become very different compared with how it is today. Changing treatment practices will mean that patient selection will become a very important issue, and the sequence of treatments will be a relevant topic of discussion, so that we can ensure we are achieving the best outcomes for our patients. Meanwhile, patients with relapsed/refractory NHL will still require salvage therapy that is both active and safe and does not overlap with cardiac toxicity from the current R-CHOP therapy. Pixantrone has the potential to meet this medical need and enter avenues for combination with new generation treatment modalities.

Key issues

• Management of relapsed or refractory NHL is a significant challenge. While these patients can be treated with platinum-based salvage regimens, and subsequent high-dose chemotherapy and autologous stem-cell transplantation in eligible patients who respond well, there is no currently recognized standard of care for patients who have failed first- and second-line treatment and for whom stem-cell transplantation is not an option.

• An important treatment option for heavily pretreated patients is the aza-anthracenedione pixantrone, which is approved as monotherapy for the third- or fourth-line treatment for aggressive B-cell NHL. While pixantrone is commonly perceived to be an anthracycline-like drug, it has a unique chemical structure and distinct pharmacologic properties that mean it behaves differently to anthracyclines and mitoxantrone. Much preclinical evidence supports a lower incidence of cardiotoxicity with pixantrone, including in conditions of prior doxorubicin/anthracycline exposure.

• When used to treat relapsed or refractory aggressive NHL in clinical trials, patients treated with pixantrone had a greater rate of complete response and greater overall response rates versus comparators, as well as a longer progression-free survival. When used as third- or fourth-line therapy pixantrone was associated with a better response than comparator drugs. Adverse events with pixantrone are manageable and consistent with expectations of cytotoxic therapy in heavily pre-treated patients.

• Additional combination therapy regimens being investigated include pixantrone and rituximab in patients with relapsed or refractory DLBCL or follicular grade 3 lymphoma, pixantrone and obinutuzumab in patients with relapsed aggressive B-cell lymphoma, and pixantrone, bendamustine, and rituximab in patients with relapsed/refractory B-cell NHL.

Declaration of interest

G Minotti has received research funds from Cell Therapeutics Inc., has received institutional research funds from CTI Life Sciences, and consultant fees from Servier, Italy. V Cattan and A Egorov are employees of Servier. H Han received institutional research funds from CTI Life Sciences. F Bertoni received institutional research funds from CTI Life Sciences. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed. One peer reviewer received research funding from and has participated in advisory boards with Servier but peer reviewers have no other relevant financial relationships to disclose.

Supplemental data for this article can be accessed here.

Acknowledgments

The authors would like to thank Tracy Harrison, who wrote the outline draft of the manuscript on behalf of Springer Healthcare Communications, and Sheridan Henness, PhD, of Springer Healthcare Communications for writing the first draft. This assistance was funded by Servier, France.

Additional information

Funding

This manuscript was funded by Servier, France.