Designing next generation antibody drug conjugates

Abstract

The new class of anti-cancer therapies known as antibody drug conjugates (ADC’s) are demonstrating significant improvements in the overall survival of patients with advanced malignancies, often without the penalty of excessive drug toxicity.

A novel property, that of ‘acid switching’ whereby an ADC can dissociate from the target antigen at endosomal pH 6.0 whilst maintaining affinity at pH 7.4, has been shown to confer enhanced accumulation of intracellular ADC and greater release of conjugated cytotoxic when compared to the parent ADC in vivo. This property has resulted in increased efficacy, notably in mouse models expressing intermediate and low levels of target antigen.

Given the potential for acid switching, we assessed if this property can be engineered into further ADCs and identified Enfortumab Vedotin (EV) as a potential target. EV is an anti-nectin-4 ADC with FDA approval for locally advanced or metastatic urothelial carcinoma. It is of further interest as human nectin-4 is expressed by tumours where there is currently great unmet clinical need, specifically oesophageal, triple negative breast and ovarian cancer, as well as the urothelial malignancies.

Initial in silico modelling identified amino acid residues for histidine replacement, however this strategy failed to identify an acid switched variant of Enfortumab. Nevertheless, analysis of these results led to a collaboration that solved the crystal structure of the Enfortumab fab fragment and human nectin4 in complex. This new data highlighted the presence of histidine 83 and 89 on human nectin-4 positioned within the binding epitope of the Enfortumab antibody, data that provided an insight into the intrinsic pH dependency of the wild-type (WT) interaction and a potential rationale for our early unexpected results.

Structural analysis subsequently identified Y49 in the second complementarity determining region (CDR) of the Enfortumab light chain as a residue that uniquely interacted with H89 of human nectin4. We show that a Y49H mutation confers a KD change from 5.8 nM at pH 7.4 to 14.9 nM at pH 6.0 and further demonstrate that this mutant has increased in vitro intracellular accumulation vs WT in multiple cancer cell lines. When conjugated to the cytotoxic monomethyl auristatin E (MMAE) as an ADC, this mutant antibody has increased in vitro cytotoxicity in cancer cell lines with low nectin-4 expression when compared to the WT ADC.

Our mutagenesis data also guided the identification of amino acid residues within the different CDRs for randomisation, with subsequent selection via phage display. This study identified two novel mutational combinations that demonstrate pH dependence in surface plasmon resonance experiments. The lead candidate; a light chain, CDR3, N92A, S93Y, F94I mutant antibody (labelled ‘AAYI’) demonstrated pH dependent binding and maintained greater affinity for nectin-4 vs the Y49H mutant at pH 7.4 (KD 1.19 nM) and pH 6.0 (KD 4.13 nM). This mutant demonstrated greater intracellular accumulation and efficacy in vitro, vs WT and Y49H mutant antibodies. In vivo experiments are planned in the near future. Overall, this work contributes towards strengthening the foundations upon which the next generation of ADCs will be built. We demonstrate that an in depth understanding of the molecular interaction between antibody and target is highly informative. This data, in combination with an understanding of the antigens physiological role, its internalisation and intracellular sorting, can lead to production of unique antibodies with the potential to deliver greater efficacy in models that could ultimately translate into improved survival benefit in the clinic.

More information

Identifiers

ORCID for Sally Ward: orcid.org/0000-0003-3232-7238

ORCID for Peter Johnson: orcid.org/0000-0003-2306-4974

Catalogue record

Export record