Efficacy, pharmacokinetic and pharmacodynamic evaluation of apaziquone in the treatment of non-muscle invasive bladder cancer

ABSTRACT

Introduction: Apaziquone (also known as EO9 and Qapzola^TM) is a prodrug that is activated to DNA damaging species by oxidoreductases (particularly NQO1) and has the ability to kill aerobic and/or hypoxic cancer cells.

Areas covered: Whilst its poor pharmacokinetic properties contributed to its failure in phase II clinical trials when administered intravenously, these properties were ideal for loco-regional therapies. Apaziquone demonstrated good anti-cancer activity against non-muscle invasive bladder cancer (NMIBC) when administered intravesically to marker lesions and was well tolerated with no systemic side effects. However, phase III clinical trials did not reach statistical significance for the primary endpoint of 2-year recurrence in apaziquone over placebo although improvements were observed. Post-hoc analysis of the combined study data did indicate a significant benefit for patients treated with apaziquone, especially when the instillation of apaziquone was given 30 min or more after surgery. A further phase III study is ongoing to test the hypotheses generated in the unsuccessful phase III studies conducted to date.

Expert opinion: Because of its specific pharmacological properties, Apaziquone is excellently suited for local therapy such as NMIBC. Future studies should include proper biomarkers.

KEYWORDS:

• Apaziquone
• Qapzola^TM
• bioreductive prodrug
• EO9
• hypoxia activated prodrug
• NAD(P)H:quinone oxidoreductase 1 (NQO1)
• non-muscle invasive bladder cancer

1. Introduction

Apaziquone, originally known as EO9 (3-hydroxy-5-aziridinyl-1-methyl-2-(1H-indole-4,7-dione)prop-β-en-α-ol), was first synthesized at the University of Amsterdam in 1987 [1]. It was developed as one of a series of derivatives of mitomycin C in a program of work sponsored by the Dutch Cancer Society. Its preclinical evaluation was coordinated by the New Drug Development Office of the European Organisation for the Research and Treatment of Cancer (EORTC) in Amsterdam and involved a number of laboratories belonging to the Screening and Pharmacology Group and Pharmacology and Molecular Mechanism groups across Europe [2]. Apaziquone belongs to a class of anticancer drugs known as bioreductive drugs which are prodrugs designed to be activated by enzymes (oxidoreductases) present within the tumor [3]. Depending on the enzymology of individual tumors, it also has the ability to target hypoxic cells and in this context, it also functions as a hypoxia-activated prodrug [3,4]. Apaziquone has had a chequered history in that it failed to show antitumor activity in the clinic following intravenous administration. One key reason for its failure was poor drug delivery to tumors caused by its poor systemic pharmacokinetic properties in conjunction with poor penetration through avascular tissue [5]. Paradoxically, these poor pharmacological properties are beneficial for locoregional therapy and the rationale for testing apaziquone against non-muscle invasive bladder cancer (NMIBC) was generated. The pharmacology of apaziquone has been extensively reviewed elsewhere [6] and the purpose of this article is to (i) to summarize the clinical studies where apaziquone was administered intravenously with the aim of putting its history into an appropriate context and (ii) to focus on the pharmacology, toxicity, and clinical activity of apaziquone against NMIBC.

2. Overview of NMIBC and available therapies

Bladder cancer is the ninth most common cancer worldwide with 429,800 new cases diagnosed in 2012 [7,8]. In the USA, it is the fourth most common cancer in males with an estimated 76,960 new cases of bladder cancer (both sexes) diagnosed in 2016 and 16,390 deaths [9]. Similarly, it is the fourth most common cancer in Europe with an estimated 151,252 cases in 2012 (both sexes) and 52,411 deaths [10]. The majority of these cases (70–75%) present as NMIBC that are defined as papillary tumors (low and high grade) confined to the mucosa and invading the lamina propria (stage Ta and T1 respectively) and includes high grade, flat tumors confined to the mucosa (carcinoma in situ [CIS] or Tis) [11]. Treatment of NMIBC (including high-risk G3 and CIS) typically involves surgical removal by transurethral resection of bladder tumors (TURBT), but NMIBC has a tendency to recur after TURBT alone. Adjuvant intravesical instillation of chemotherapy or immunotherapy has been shown to reduce tumor recurrence rates and prevent disease progression [12]. Commonly used chemotherapeutic drugs for G1/G2 and Ta/T1 tumors include mitomycin C, valrubicin, epirubicin, cisplatin, and doxorubicin and European Association of Urology (EAU) and American Urological Association guidelines recommend single, immediate post-TURBT intravesical instillation for low- and intermediate-risk NMIBC [13,14]. Meta-analysis data have shown that this treatment reduces recurrence rates by 11.7–13.0% compared to TURBT alone [15–17]. In the USA, four agents have been approved by the FDA for use in NMIBC; ThioTEPA in 1959, Bacillus Calmette-Guérin (BCG) Tice in 1989, BCG Connaught in1990, and valrubicin in 1998 [18]. ThioTEPA is no longer widely used because of toxicity issues and mitomycin C is commonly used off-label as an intravesical treatment for NMIBC. The use of mitomycin C is also associated with rare but severe toxicities [19–22]. Intravesical immunotherapy with BCG has been widely used since the 1970s and is frontline treatment for patients presenting with high-risk lesions (G3 and CIS) and is not administered immediately after surgery [23,24]. Apaziquone has been evaluated against low-risk (G1/G2 and Ta/T1) tumors.

Despite adjuvant intravesical therapy, recurrence rates remain high (30–70%) and EAU and AUA guidelines recommend frequent cystoscopic and cytologic surveillance that continues for the lifetime of the patient. Recurrence of the disease together with the cost of intensive surveillance strategies contributes to the high cost of treating this disease [25]. Some authors have stated that bladder cancer has the highest lifetime treatment costs per patient of all cancers due to the high recurrence rate and sustained invasive monitoring requirements [26]. There is therefore a strong pharmacoeconomic need for new therapies that can delay recurrence rates in NMIBC. In addition, the relative lack of approved treatment options for NMIBC and the fact that no new therapies have been approved for NMIBC in the twenty-first century graphically illustrates the need to develop new drugs for the management of NMIBC [18]. Finally, NMIBC is a nonlethal disease and as most patients are old with other comorbid conditions such as cardiovascular and pulmonary diseases [27,28], the safety of drugs used after TURBT is also of paramount importance. The toxicity profile of novel therapeutics is also therefore an important consideration in the management of these patients.

3. Introduction to apaziquone

Apaziquone was originally developed at the University of Amsterdam in a research project sponsored by the Dutch Cancer Society (Box 1). The original patent was published in 1987 (WO87/06227) and maintained by the University of Amsterdam until Kyowa Hakko licensed the compound in 1994 [29]. Kyowa Hakko released the license in 1997/98 and the patent was maintained by the NDDO Oncology BV and INC Research. Spectrum Pharmaceuticals (Irvine, California) in-licensed exclusive worldwide rights to EO9 (currently branded as Qapzola) from INC Research in 2001. Formulations designed specifically for the intravesical administration of apaziquone were developed [30,31] and patented in 2007 (WO2007092964). These formulations led to greater stability of apaziquone in human urine [32]. In 2008, Spectrum Pharmaceuticals and Allergan signed an exclusive collaboration for the development and commercialization of EO9 (apaziquone) but Spectrum Pharmaceuticals subsequently regained exclusive rights to apaziquone in 2013.

3.1. Chemistry

The original synthesis of apaziquone (EO9) was reported in 1987 [1] and since then various methodologies have been described [33–35] including large-scale synthesis methods [36,37]. Whilst apaziquone is structurally related to mitomycin C, its mechanism of action is very different particularly with respect to NQO1 activation (see below in Section 3.2). The chemical structure of EO9 is presented in Figure 1. It is an indolequinone that has three active centers (i) an aziridinyl group at C5, (ii) a vinylic group at C2, and (iii) the hydroxymethyl group at C3. The mechanism(s) of action are presented in Figures 1 and 2 and these include the generation of reactive oxygen species via redox cycling [38] (Figure 1) and alkylation reactions via (i) enzymatic reduction and (ii) proton assisted aziridine ring opening (Figure 2).

Figure 1. Redox reactions and generation of reactive oxygen species following the reduction of apaziquone by one electron oxidoreductases. The quinone nucleus of apaziquone is reduced by NAD(P)H dependent one electron oxidoreductases such as cytochrome P450 reductase to a semiquinone radical which redox cycles back to the parent quinone in the presence of oxygen generating superoxide anions. Hydrogen peroxide is generated via superoxide dismutase (SOP) mediated reactions and this in the presence of trace metals can lead to the formation of hydroxyl radicals and subsequent damage to cellular macromolecules.

Figure 2. Mechanisms for DNA alkylation by apaziquone: (i) Pathways involving one- and two-electron reduction. Proton-coupled single-electron reduction of apaziquione generates stablised radical A, which can be further reduced (via B) to key intermediate C; C can be alkylated via cation D, formed by acid-catalysed dehydration, to give DNA monoadduct E. can undergo a second dehydration, giving dienyl cation F, which can be trapped in a second DNA alkyation event, giving interstrand cross-linked species G. (ii) Proton-accelerated mechanism. The conjugation of the aziridinyl lone-pair (acting as a vinylogous amide) into the quinone ring allows for an addition proton-accelerated DNA alkylation mechanism. Thus, protonation leads to cation A, which is in resonance with B; ring-opening of the reactive aziridinium moiety of B by DNA leads to aza-quinone species C, which can then undergo reduction to D. Acid-catalysed dehydration of D gives methylene indolium cation E, which can function as an active alkylating agent.

3.2. Mechanism of action and pharmacodynamics

The mechanism of action of apaziquone is complex and has been extensively reviewed elsewhere [6]. Briefly, the pharmacodynamics properties of apaziquone can be segregated into (i) factors that influence the bioreductive activation process and (ii) how cells respond to the type and extent of DNA damage induced following bioreductive activation. Several factors influence the bioreductive activation process and cell kill is determined by a complex interaction between the expression of one and two electron oxidoreductases in cells, the presence or absence of oxygen, and extracellular pH (pHe). Several oxidoreductases are capable of reducing apaziquone but the major ones are the two-electron reductase NAD(P)H:quinone oxidoreductase 1 (NQO1, also known as DT-diaphorase) and a series of one-electron reductases including cytochrome P450 reductase [6]. Reduction of apaziquone by purified NQO1 and cytochrome P450 reductase leads to free radical generation and DNA-damaging species in cell-free assays [38–40] (Figures 1 and 2). The types of DNA damage observed in cell-free assays include interstrand cross-links, mono-adducts, and strand breaks [39,41–43]. With regards to the induction of DNA cross-links, there are some discrepancies in the literature with some studies reporting cross-linking [41,42], whereas others report no cross-link induction following reduction of apaziquone by NQO1 in cell-free assays [43]. Species-dependent differences (rat vs. human NQO1) may account for these differences but similar conflicting results were reported following the analysis of DNA interstrand cross-link induction in human cancer cells using alkaline elution or comet assay techniques [41,44]. Whilst the extent of DNA interstrand cross-link induction maybe uncertain, it is clear that apaziquone is reduced by one- and/or two-electron reductases to species that are capable of damaging DNA by free radical generation and alkylation.

In addition to the presence of oxidoreductases, the oxygenation state of cells has a significant bearing on the pharmacodynamic response. In the presence of oxygen, apaziquone is reduced predominantly by the two-electron reductase NQO1 and several studies have demonstrated that IC₅₀ values are typically inversely proportional to NQO1 activity [45–48]. In the absence of oxygen however, significant potentiation of apaziquone activity is only seen in cells that have low or no NQO1 activity [47,49], particularly those that harbor the C609T polymorphic variant of NQO1 [49,50]. In cells with low or no NQO1, one-electron reductases assume a prominent role under these conditions, reduction of apaziquone is oxygen sensitive with the semiquinone rapidly redox cycling back to the parent compound resulting in the production of reactive oxygen species (Figure 1). When oxygen is low or absent, the half-life of the semiquinone increases providing more time to either alkylate DNA directly or be reduced further to produce the hydroquinone [51]. The experimental evidence in vitro clearly indicates that the activity of apaziquone is strongly dependent upon NQO1 activity and cause cell death in an oxygen independent manner but in cells that are devoid of NQO1 activity, apaziquone is able to selectively target hypoxic cells. The response of tumors in vivo could however not be predicted based on NQO1 activity [52], but it is likely that poor systemic pharmacokinetics and penetration through avascular tissue compromised drug delivery resulting in suboptimal levels of apaziquone in tumors [5]. This suboptimal accumulation of apaziquone in tumors is possibly sufficient to exert a radiosensitizing effect. The ability to target hypoxic cells suggests that combination therapy with radiotherapy could be effective and this was indeed demonstrated in three syngeneic rat lung cancer models and a rhabdomyosarcoma in vivo (Table 1) [53], but this strategy was never considered as a clinical strategy.

Table 1. In vivo antitumor activity of apaziquone (EO9) in combination with radiation in experimental rat tumors. The tumors were two squamous cell lung carcinomas (L17 and L42), a lung adenocarcinoma (L27), and a rhabdomyosarcoma (BA1112). All tumors were grown subcutaneously and EO9 was administered intraperitoneally (daily × 5 at 0.4 mg/kg). When EO9 was combined with 5 daily doses of 4Gy, additive or synergistic effects were observed. This data were originally reported by Kal et al [53].

Specific growth delay
Regimen	L17	L42	L27	BA1112
EO9	0.1	1.9	1.5 (3.3)	0
EO9 + Enflurane	0.1	2.2	2.6	Not done
5 × 4Gy	1.5	4.4	19.7^a (7.4)	5.6
EO9 + 4Gy	2.2	4.6	18^a (11.0)^b	Not done
4Gy + EO9	1.8	3.9	31^a (14.5)	7.2

^aCure rates: 2/6, 2/3, and 4/5, respectively; ^bcure rate 1/5; between brackets results of 5 × 2Gy. Enflurane was used as an inhalational anesthesia for irradiation of the rats.

The activity of apaziquone is strongly influenced by pHe. Under mild acidic extracellular conditions (pH 6.0), the cytotoxic potency of indolequinones containing aziridine ring functionality (including apaziquone) is significantly enhanced in vitro [44,54]. The activity of apaziquone remains inversely proportional to NQO1 activity under acidic pHe conditions, but potency is significantly enhanced [55]. In contrast to mitomycin C where reducing the pH increases substrate specificity and reduction by NQO1 [56], the substrate specificity of apaziquone for NQO1 is not affected by pH [55]. Whilst the cytotoxic potency of apaziquone is enhanced under acidic conditions, its chemical stability is decreased resulting in enhanced formation of EO5A in the absence of cells [54]. These apparently paradoxical properties can be explained by proton-assisted aziridine ring opening and enhanced nucleophilic attack (Figure 2), the outcome of which would vary depending on the type of nucleophile (e.g. water or DNA) present in the system.

In summary, the pharmacodynamic properties of apaziquone are determined by a complex ‘cocktail’ of parameters including (i) the presence or absence of NQO1, (ii) the presence of one-electron reductases such as cytochrome P450 reductase, (iii) the presence or absence of oxygen, (iv) acidic pHe, and (v) repair of DNA damage. The preclinical data suggests that apaziquone should (i) target the aerobic and hypoxic fraction of NQO1 rich tumors equally efficiently, (ii) target the hypoxic fraction of tumors that have low NQO1 activity, particularly those that harbor the inactivating NQO1 single nucleotide polymorphic variant, and (iii) target cells that reside in an environment where pHe is mildly acidic. Most of these parameters are related to the bioreductive activation process but as stated above, the second important parameter that determines response is how the cell responds to the damage induced in terms of detoxification and repair. As apaziquone can induce strand breaks via the induction of ROS, antioxidant defense mechanisms could modulate activity and antioxidants such as catalase and Tempol have been shown to reduce DNA damage and apaziquone cytotoxicity [57,58]. Despite the fact that apaziquone has been studied for 30 years, comparatively little is known about the role DNA repair plays in determining response and further studies are required to address this issue.

3.3. Pharmacokinetics and metabolism

Full details of preclinical and clinical pharmacokinetics following intravenous administration have been reviewed in detail elsewhere [6]. The suboptimal pharmacokinetic profile of apaziquone is illustrated by the results of phase I studies where apaziquone is rapidly cleared from the systemic circulation with half-lives ranging from 0.8 to 19 min [59]. Furthermore, the predominant metabolite found in blood and urine following intravenous administration was the inactive, aziridine ring open metabolite EO5A [54,59]. Following intravesical administration, two clinical studies have investigated the pharmacokinetic and safety of apaziquone [60,61]. In the first of these studies [61], apaziquone was administered 2 weeks after TURBT and for the first six patients entered into the trial, each received an escalating dose of apaziquone with one dose per week (for a total of 6 weeks). At all doses administered, no detectable levels of apaziquone or EO5A were found in the blood. At the end of the 1-h intravesical instillation, over 60% of the administered dose of apaziquone (but not EO5A) was recovered from urine, the concentration being proportional to the dose administered [61]. Overall exposure parameters within the bladder considerably exceeded plasma AUC values following intravenous administration indicating that therapeutically relevant concentrations of drug were delivered to tumors in the bladder [6,61]. The second study was conducted in a total of 20 patients with NMIBC and a single 4 mg/40 ml dose of apaziquone was administered intravesically within 6 h of TURBT [60]. No apaziquone or EO5A were detected in the blood of patients during and after the instillation. In acidified urine, a degradation product known as EO-9-Cl has been reported [31], but the presence of this product was not observed in clinical studies. Pharmacokinetic studies following the intravesical administration of apaziquone to NMIBC patients are therefore limited but in both studies, high levels of apaziquone (4 mg/40 ml) were delivered into the bladder but no detectable levels of drug or known metabolites reached the systemic circulation.

3.4. Clinical efficacy

A number of clinical trials were conducted following the intravenous administration of apaziquone and details of these are reviewed elsewhere [6]. This review will focus only on the clinical trials conducted to date following the intravesical administration of apaziquone to NMIBC patients. In a phase I/II pilot study, 12 patients with low-risk NMIBC were administered apaziquone intravesically (once a week for 6 weeks; 0.5–16 mg/40 ml) and antitumor effects were determined against marker lesions left in the bladder at TURBT [61]. A total of eight complete responses as defined by both visual loss of the marker lesion (Figure 3) and pathological confirmation of no tumor at the marker lesion site was reported. Biomarker studies were conducted in the phase I/II study using immunohistochemical analysis of NQO1 and Glut-1, the latter being an endogenous marker of hypoxia [61]. Whilst a broad range of NQO1 and Glut-1 protein expression was detected in tumors, there was no correlation between the response of the marker lesion and biomarker expression. These studies were only conducted on 12 patients and no further biomarker studies were performed in subsequent clinical trials. Because of the low number of patients, it is advised to include extensive biomarker studies (focusing on NQO1 and other hypoxia biomarkers) in future studies. Specific PET imaging probes might also help to determine rate of hypoxia in tumors.

Figure 3. The presence of a marker lesion prior to (A) and after (B) a 6 week course of apaziquone. The results were first presented in the review article by Phillips et al, reproduced with permission [6].

Using an identical marker lesion study design, 31 out of a total of 46 patients with multiple pTa or pT1 tumors achieved a complete response in phase II trials at a dose of 4 mg/40 ml [62]. Marker lesion studies have been conducted for many other cytotoxic and immune response modifiers, but the complete response rate achieved by apaziquone (67%) was the highest reported complete response rate [63]. Two-year follow-up studies demonstrated that long-term results were good in comparison to other ablative studies with a recurrence-free rate of 49.5% at 2 years [64]. Similar results were reported using patients enrolled on the phase I/II clinical pilot study [65]. Phase II studies of intravesically administered apaziquone to high-risk NMIBC have also reported encouraging results [66]. These results were promising, but it is however acknowledged that the number of phase II studies is limited and randomized phase III studies are required [67].

Two phase III trials (SPI-611 and 612) were completed and the results were presented at the American Urological Association annual meeting in 2016 [68,69]. Both studies had similar designs and a single dose of apaziquone (4 mg/40 ml) was administered intravesically to patients with low-risk NMIBC (TaT1 and G1/G2 lesions) within 6 h of TURBT and retained for 1 h. The control arm consisted of the administration of the placebo in both studies. The primary endpoint was the 2-year recurrence rate and the secondary endpoint was time to recurrence. Taking both studies together, 1614 patients were enrolled across 152 centers in the USA, Poland, and Canada (study 612 only). The studies were analyzed individually and both did not meet their primary endpoint of a statistically significant difference in the 2-year recurrence rate between treatment and placebo arms. When the results of the two trials were pooled, however, statistically significant differences were obtained in both the 2-year recurrence rate (apaziquone 38.8% vs. placebo 45.5%; p = 0.0218) and time to recurrence (apaziquone 18.2 months vs. placebo 16.8 months; P = 0.0096; Hazard ratio 0.79) [68]. Furthermore, post hoc analysis of data from both studies demonstrated that a significant difference in time to recurrence (Hazard ratio 0.48; p = 0.0096) was observed in the 117 patients who received apaziquone >30 min post-TURBT compared to placebo (100 patients). Patients in the apaziquone 31–90 min subgroup also showed a significant absolute decrease in the 2-year recurrence rate (apaziquone 28.2% vs. placebo 50%; p = 0.001) resulting in a 43.6% relative reduction in 2-year recurrence rates [69], but no improvements were found in the other 2 subgroups (<30 min, later than 90 min). The result in the <30 min could be explained by the fact that blood released during surgery could inactivate apaziquone [70]. Loadman et al. [70] indeed demonstrated that EO9 is rapidly degraded by blood in vitro. However, the FDA’s Oncologic Drugs Advisory Committee recently determined that apaziquone did not show sufficient statistical evidence of treatment effect of apaziquone over placebo based on the primary endpoint results of individual studies and that the post hoc analysis conducted by Spectrum Pharmaceuticals was hypothesis generating. The FDA have granted an SPA agreement for a new phase III study to address the issues arising from SPI-611 and 612, particularly with regards to the timing of apaziquone administration post-TURBT. Spectrum plans to initiate enrolment in this study.

Preclinical toxicology studies in mice were reported in 1993 and the most obvious finding was a lack of myelosuppression, a result that was considered unique for what was essentially an alkylating agent [2]. No myelosuppression was also reported in the phase I and II studies when EO9 was administered intravenously [59,71]. Dose-limiting toxicity in these studies was reversible proteinuria that was attributed to high levels of NQO1 found within the kidney. In the phase I/II study following intravesical administration to patients with NMIBC, 6 patients received escalating doses of apaziquone (0.5–16 mg/40 ml) weekly for 6 weeks. A further 6 patients received weekly apaziquone for 6 weeks at the highest nontoxic dose (4 mg/40 ml). Grade 2 and 3 dysuria and hematuria were observed at doses above 8 mg/40 ml, but 4 mg/40 ml was well tolerated [61]. Apaziquone was also well tolerated in phase II studies when administered either as 6 weekly infusions at 4 mg/40 ml [62] and as a single dose administered within 6 h of TURBT [60]. In phase III studies (single dose of 4 mg/40 ml administered within 6 h of TURBT), apaziquone was well tolerated with a safety profile that was indistinguishable from placebo, with a similar incidence of adverse events (80.0% vs. 78.5%, respectively) with dysuria as the most common treatment-related adverse event in both the apaziquone and placebo group 4.6 versus 4.1%, respectively) [68]. Hence, the safety profile of apaziquone is very good and at 4 mg/40 ml, apaziquone was well tolerated following both single and multiple instillations. This compares favorably with other agents such as the structurally related (but mechanistically very different) anticancer drug mitomycin C used to treat NMIBC where rare but severe side effects have been reported [19–22]. Of course, the occurrence of rare adverse events caused by apaziquone cannot be ruled out until larger numbers of patients have been treated, but the early signs are that apaziquone has a favorable toxicity profile. In the context of managing elderly NMIBC patients with comorbidities, the excellent safety profile of apaziquone is very encouraging.

4. Conclusion

There is a clear need for new therapeutic approaches to treating NMIBC [18]. From both a pharmacoeconomic and patient perspective, the development of new therapies that can delay the recurrence of NMIBC and are well tolerated would have a significant financial impact on health-care providers. Apaziquone demonstrated excellent activity against NMIBC in marker lesion studies, but the clinical improvement in phase III studies required for drug approval is yet to be demonstrated. It is important to state that the design of the phase III studies was significantly different from the phase I/II studies. With hindsight, the presence of blood in the bladder shortly after TURBT should have been considered as inactivating metabolism of apaziquone by blood [70] and this could have had a significant bearing on the outcome of these trials. A further phase III study has been designed and the results of this trial will be definitive.

5. Expert opinion

Apaziquone was initially developed as a bioreductive drug designed to target NQO1 rich and/or hypoxic tumors. Initial clinical studies following intravenous administration failed to show ant-tumor activity at doses that caused toxicity and it was not developed further. These early clinical trials did not include potential biomarkers (e.g. hypoxia, NQO1) or test the drug in combination with other modalities and it is feasible that the lack of target expression in the patient’s tumors could account for its poor efficacy in clinical trials [72]. Moreover, the drug showed a very unfavorable pharmacokinetic profile leading to poor drug delivery to tumors and development was halted, until it was recognized that such a profile would be ideal for local administration. Any leakage to the systemic circulation would immediately be taken care of by a rapid degradation and elimination. Early clinical studies by bladder instillation showed very promising data with limited local and no systemic toxicity. Significant activity against marker lesions in phase I and II studies clearly demonstrates that apaziquone has anticancer activity, but this failed to translate into positive results in phase III trials. Whilst the two, randomized phase III studies failed to reach statistical significance, there was some evidence from post hoc analysis of pooled results of significant activity. The FDAs Oncologic Drugs Advisory Committee were not convinced of the validity of the statistical analysis and a further phase III trial is required to test the hypotheses generated from the completed studies. Despite these disappointing results, it is our opinion that the development of apaziquone should continue and future studies should take several aspects into account such as the instability at acidic conditions and protein binding, as well as timing of the instillation (between 31 and 90 min) after surgery. The stability issue of apaziquone is taken care off by using a buffered formulation (pH 9.2) for the drug, which prolonged stability of apaziquone during instillation, even when mixed with acid urine [31]. Other future applications of apaziquone would be regional treatment for other malignancies, even in the prevention setting. Moreover, apaziquone is an excellent radiosensitizer, which offers possibilities for combinations with radiotherapy, including bladder cancer, but also other types of localized tumors would be eligible. From a wider perspective, the developmental history of this drug shows that compounds with poor pharmacological profile are not necessarily precluded from further development as loco-regional therapies.

Box 1. Drug summary box: apaziquone

Drug names:

Apaziquone: Apaziquonum; NSC-382459; NOR-701; EOquin; EO 9 (pharmaceutical)

Drug description:

Apaziquone is an indolequinone bioreductive prodrug which is

activated by oxidoreductases (NQO1) to DNA damaging

species which can kill cancer cells^6

Formula: C15H16N2O4

Chemical structure:

3-hydroxymethyl-5-aziridinyl-1-methyl-2-(1H-indole-4,7-dione) prop-beta-en-alpha-ol

IUPAC name:

5-(aziridin-1-yl)-3-(hydroxymethyl)-2-[(E)-3-hydroxyprop-1-enyl]-1-methylindole-4,7-dione

Molecular weight: 288.303 g/mol

PubChem CID: 5813717

Declaration of interest

GJ Peters has received research funding from Spectrum pharmaceuticals, but not for apaziquone, HR Hendriks is the owner of Hendriks consulting. G Reddy is an employee and shareholder of Spectrum pharmaceuticals. 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

Acknowledgement

This paper was written in the framework of the EORTC Pharmacology and Molecular Mechanism (PAMM) Group.

Additional information

Funding

This paper was not funded.