Combination Treatment with an Antibody–Drug Conjugate (A1mcMMAF) Targeting the Oncofetal Glycoprotein 5T4 and Carboplatin Improves Survival in a Xenograft Model of Ovarian Cancer

Epithelial ovarian cancer (EOC) is one of the most common gynaecologic malignancies, with 50% of all cases occurring in women aged older than 65 years. It is the eighth most frequent cause of cancer death in women [1]. The most common subtype of EOC, high-grade serous cancer (70%), is characterised by high levels of genomic instability resulting in frequent genomic aberrations and morphological heterogeneity. This makes the identification of key driver mutations and the development of therapeutics against such mutations challenging [2, 3]. Whilst this genomic instability initially sensitises high-grade serous EOCs to DNA-damaging agents and sensitisers, such as carboplatin and poly-ADP ribose polymerase inhibitors [4‐6], ultimately, the inherent molecular heterogeneity and plasticity of EOC cells mean that 80–90% of women originally treated with these agents will develop resistance [7].

Antibody–drug conjugates (ADCs) are an emerging class of anticancer drug that combine the selectivity of a monoclonal antibody with the cytotoxic potency of established chemotherapeutics. Selectively targeting tumour cells thought to initiate relapse could prolong progression-free survival whilst limiting potential off-target side effects [8]. The 5T4 oncofetal glycoprotein exhibits several properties that make it a highly attractive target antigen in this regard. 5T4 is rarely expressed by normal adult tissues [9, 10], but is detected in a high proportion of primary and metastatic cancers, including EOC [11‐13]. Mechanistically, 5T4 expression has been shown to associate with several processes that are important in the spread of cancer cells and relapse [14]. These include epithelial-mesenchymal transition [15, 16], the directional movement of cells through potentiation of CXCL12/CXCR4 chemotaxis [17, 18] and inhibition of canonical Wnt/beta catenin and concurrent potentiation of non-canonical pathway signalling [19]. Consistent with this role in tumour development and spread, 5T4 expression in colorectal, gastric and all subtypes of ovarian cancer has been correlated with a poorer clinical outcome [11, 20‐24].

These characteristics have underpinned the clinical development of various immunotherapies [12, 14] including a number of ADCs [25‐27]. A1mcMMAF is a 5T4 humanised monoclonal antibody (A1) chemically linked by a non-cleavable maleimidocaproyl linker to deliver a microtubule-disrupting agent, monomethyl auristatin F (MMAF). A1mcMMAF has been shown to be rapidly internalised on binding of the antibody portion and to have potent in vivo activity in a variety of tumour models [26] leading to the initiation of a phase I dose-escalation study [25]. Recent work has identified that combining a 5T4 ADC with various other chemotherapeutics enhances its effects on survival [28, 29]. These data support further investigation of A1mcMMAF in combination with chemotherapy in other cancers.

The effectiveness of ADCs depends on numerous factors including tumour expression levels of the target antigen, as well as the ability of tumour cells to internalise and cleave the conjugated drug and undergo apoptosis in response. More recently, it has been demonstrated that this activity can be further enhanced by concurrent administration of other chemotherapeutics [28, 29]. Our study confirms that 5T4 is expressed by all histotypes of EOC, but particularly in women with high-grade serous carcinoma and supports the findings demonstrated in other published series [11‐13]. 5T4 cell membrane expression makes this an attractive target for antibody-delivered therapies, such as ADCs, that require binding to a cell surface marker and subsequent internalisation to deliver their cytotoxic payload intracellularly. In this cohort, 5T4 expression appears to be an independent poor prognostic factor for survival, even after accounting for tumour stage, histotype and adjuvant therapy. This is consistent with the hypothesis that 5T4 expression may have a functional role in relapse and/or metastasis. Selective elimination of tumour-initiating cells expressing 5T4 using ADCs was shown to be effective at prolonging survival in mouse models of both non small cell lung carcinoma (NSCLC) and pre-B acute lymphoblastic leukemia [26‐28]. Demonstrating similar effects in EOC, a condition in which most women will relapse following their initial surgery and chemotherapy, could offer additional options for disease control. Conventional therapy tends to reduce tumour burden and improves symptoms but may fail to eradicate tumour-initiating cells, leading to eventual recurrence and drug resistance. Therapies targeting tumour-initiating cells may, therefore, work best alongside conventional therapies that debulk the tumour mass. Our intraperitoneally seeded xenograft mouse model sought to emulate the effect of adjuvant treatment regimes following complete surgical debulking.

We have demonstrated that the 5T4-targeting ADC, A1mcMMAF, has effect both in vivo and in vitro. Using paired wild-type and 5T4 knockout SKOV3 cells, we demonstrated that the specificity of A1mcMMAF is conferred by the antibody portion, whilst the anti-proliferative effect is conferred by the drug conjugate. At higher dose ranges, reductions in cell viability are seen in vitro and reductions in tumour loads are seen in vivo with the control ADC, suggesting that there is the potential for non-specific killing of non-antigen-bearing cells. This has been previously noted in the early preclinical development phase, where numbers of NSCLC-derived spheroids decreased when treated with the control ADC [26]. It is thought that this may result from phagocytosis of cleaved ADC produced by secreted tumour-associated proteases [26]. Reassuringly, toxicology studies in cynomolgus monkeys, which demonstrate comparable normal tissue staining patterns to humans, demonstrated that, neither the non-specific effect, nor the 5T4-specific effect of the ADC lead to any significant adverse events [26, 37‐39]. Furthermore, phase I trials of the 5T4 ADC, PF-06263507, demonstrated a favourable safety profile for the drug [25].

Treatment with A1mcMMAF attenuates tumour growth and leads to dose-dependent reductions in tumour load. Previous work by Sapra et al. using various primary culture and cell line models suggest that the response to 5T4 ADCs is governed more by the presence of 5T4 expression, rather than antigen density, histological type or sensitivity to auristatin compounds [26]. Whilst we recognise that the demonstration of effect in a more high-grade serous cancer cell line such as OVSAHO or OVCAR 8 would add to the case for clinical trials of this treatment in patients with ovarian cancer, the evidence of effect in SKOV3 cells, which express high levels of 5T4, but are thought to represent a clear cell/endometrioid phenotype, support using 5T4 expression as an inclusion criterion in future trials, over histotype alone.

In our study, tumour loads increased rapidly following cessation of A1mcMMAF monotherapy (5 mg/kg). We have demonstrated that combination therapy with carboplatin, the first-line chemotherapeutic used in ovarian cancer, leads to a prolonged delay in tumour growth and an increase in overall survival. A potential explanation for the enhanced activity of MMAF in conjunction with carboplatin may lie in their distinct antineoplastic effects. MMAF inhibits cell division by inhibiting tubulin polymerisation, leading to cell-cycle arrest in mitotically active cells. Alkylating agents, such as carboplatin, damage DNA by the induction of DNA adducts and are effective in both actively dividing cells, as well as, the quiescent cells that are unaffected by A1mcMMAF. Synergy between platinum-based drugs and anti-mitotic agents has been proposed to occur because the cell-cycle arrest induced by antimitotic drugs hinders DNA repair and leads to the accumulation of toxic platinum–DNA adducts [40]. Synergism of 5T4 ADCs has been shown with other classes of drugs. Shor et al. demonstrate a superadditive effect when using A1mcMMAF with both taxanes and PI3 K/mTOR inhibitors preclinically [29]. They propose that the mechanism for the synergy between A1mcMMAF and taxanes is that, despite both being microtubule agents, the distinct sites of action of MMAF and taxol result in distinct effects on tubulin, which cooperate to modulate cell-cycle progression and microtubule dynamics. We have previously demonstrated that the combination of dexamethasone and A1mcMMAF is effective in B-ALL [28].

5T4 scores highly when assessed against predefined criteria for priority ranking target antigens for translational research [12, 41]. Its potential association with tumour progression, high expression levels in a wide range of different cancers, expression on tumour-initiating cells, cell surface location and immunogenicity have made it an equally attractive target for other immunotherapeutic modalities. Following promising results from phase I and II clinical trials of modified vaccinia Ankara-5T4 (Trovax^®) [42‐52], a viral vector designed to induce a 5T4-specific immune response, a phase II clinical trial of this drug was undertaken in relapsed ovarian cancer (TRIOC; NCT01556841). Results are awaited but a phase III trial of Trovax^® in metastatic renal cell carcinoma demonstrated that survival was improved in patients that mounted an antibody response [42]. Unfortunately, when all patients were considered no overall difference in survival was seen. Antibody–drug conjugates have the advantage that they do not rely on a patient’s ability to mount an immune response, making their activity more predictable in the short term. Like vaccine approaches, ADC efficacy may be limited by preformed antibodies. In the phase I trial of the 5T4 ADC, PF-06263507, 9% of patients exhibited anti-drug antibodies prior to treatment, rising to 17% of patients post-treatment [25]. Chimeric antigen receptor T cells provide a further alternative that avoids the requirement for patients to mount a specific immune response. Preclinical studies have demonstrated the feasibility of engineering 5T4-specific chimeric antigen receptor T cells that are active in ovarian cancer mouse models [13]. However, the chimeric antigen receptor T cell pipeline is extremely labour intensive and requires additional infrastructure, increasing potential costs for healthcare providers and patients when compared with ADCs. Antibody–drug conjugates, therefore, have the most potential for use in clinical practice.

Antibody–drug conjugates are one of the most rapidly growing fields of immunotherapeutics with over 100 in different stages of development [53]. The potential of 5T4 ADCs to improve the effectiveness of platinum-based chemotherapy regimes in ovarian cancer offers an exciting opportunity to prolong progression-free survival and overall survival in women in both the primary treatment and relapse settings. The completion of the recent phase I dose-escalation study of A1mcMMAF in solid cancers [25] lays the foundation for further clinical trials to determine its efficacy in women with EOC.