Conference Report: Hot topics in antibody–drug conjugate development

This Conference Report highlights key outcomes from several sessions focused on antibody–drug conjugates (ADCs) at the 2013 American Association of Pharmaceutical Scientists (AAPS) National Biotechnology Conference, held in San Diego, CA, USA from 19–23 May. The National Biotechnology Conference (NBC) is widely attended by representatives from the pharmaceutical industry, biotech, regulatory authorities and academia in the USA and Europe.

The topic of ADCs is no stranger to the NBC – for the past several years ADCs have had a growing presence at this meeting – each year building on the topics of preceding years. The 2012 NBC meeting featured numerous sessions focused on preclinical and translational development of ADCs, including a short course titled, ‘Translational challenges in developing antibody–drug conjugates (ADCs)’, organized by members of the Pharmacokinetics, Pharmacodynamics and Drug Metabolism section of the AAPS. This session described the importance of selection of target antigens, ADME characterization and assay development, and highlighted case examples (including MUC16, trastuzumab emtansine and brentuximab vedotin) to emphasize the importance of early characterization in order to achieve maximal activity in patients.

Recent approvals of Adcetris^® and Kadcyla^®, and ongoing clinical trials of over 20 ADC molecules, have capitulated ADC development into mainstream biotechnology media, generating significant enthusiasm, which was evidenced during the 2013 NBC. The focus on ADCs shifted to clinical development with two short courses, six sessions and numerous contributing papers, with a specific focus on bioanalytical strategies and challenges for clinical immunogenicity and PK assays, drug–drug interactions (DDIs) and in vivo characterization of ADC biotransformation. The complexity of molecular and structural characteristics of ADCs was highlighted throughout the meeting as a driving factor in bioanalytical strategies, thus requiring early interactions between the clinical pharmacologist, clinical team and multi-disciplinary bioanalytical groups to develop novel technologies for PK and immunogenicity and biotransformation assessment.

Our aim is to provide a focused summary of each of the ADC hot topics, including bioanalytical and PK considerations, quantitative evaluation of the impact of immunogenicity on clinical PK, ADME to understand ADC DDIs, risk mitigation and clinical considerations. The views expressed in this commentary are those of the authors from industry and do not necessarily represent those of their affiliations.

Hot topics

▪ Bioanalytical & PK considerations for antibody–drug conjugates

The emergence of contemporary approaches for improved drug delivery led to the emergence of ADCs; presently the US FDA has not issued a Bioanalytical Guidance outlining specific assay requirements for these molecules. The bioanalytical considerations for the measurement of ADCs recently approved by the FDA (brentuximab vedotin and trastuzumab emtansine) were developed based upon the evolving principals for both small and large molecules [1].

Based on PK and bioanalytics included in the Biologics License Applications for the recently approved brentuximab vedotin and trastuzumab emtansine, it is conceivable to deduce that measurement of the ADC, total antibody (Ab), unconjugated small payload components, and perhaps known payload metabolites, are recommended (Table 1). In the case of trastuzumab emtansine, DM1 was available in circulation; however, the soluble fraction of DM1 (concentration ∼5 ng/ml) was measured using sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS PAGE)/ELISA and confirmed to be quite low and unlikely to pose a risk. Thus, ADC thought leaders speaking at the NBC strongly recommend that validated assays should be developed to quantify each component for clinical studies, including: the ADC, total Ab and unconjugated small payload component (Table 1). Representatives from the FDA recommended that sponsors consider requesting Type B or Type C meetings to address questions related to the bioanalytical assay requirements.

A symposium titled ‘Bioanalytical considerations for ADCs’ was led by Boris Gorovits (Pfizer, CT, USA). This session highlighted the importance of fully characterizing PK in early development using rich PK profiles, developing integrated models to support hypothetical catabolism schemes and using exposure–response to identify clinically relevant doses. Limited PK sampling of the most relevant analytes was recommended for late development once fully characterized in early development. A recent publication released by the Antibody Drug Conjugate Working group [2] and speakers in this session recommended analytes for measuring exposure as follows:

• ▪ Conjugated Ab (measuring Ab with minimum of DAR 1 using ligand-binding assay [LBA]);

• ▪ Total Ab (measuring with minimum of DAR 0 and above using LBA);

• ▪ Ab-conjugated drug (measuring total drug conjugated by either LC–MS/MS or LBA);

• ▪ Unconjugated drug (payload not conjugated to Ab measured by LC–MS/MS;

• ▪ Total drug (total unconjugated and conjugated payload measured by LC–MS/MS);

• ▪ Anti-therapeutic Ab (measuring antibodies directed against Ab component of ADC, linker or drug [binding/neutralizing] measured by LBA).

This session mentioned hydrophobic interaction chromatography (HIC), UV, RP-HPLC, LC–MS/MS and isoelectric focusing in capillary column (ICE) as possible techniques to characterize drug load and suggested several methods for characterizing impurities, including size-exclusion chromatography, SDS-PAGE, isoelectric focusing, ELISA and HPLC (Table 1). A representative from the FDA recommended following bioanalytical method validation, LC–MS/MS and LBA guidance when developing such assays [2–5].

Methods for drug load distribution demonstrate that it can be monitored by HIC, ICE and LC–MS (for characterization). There are a few caveats that were mentioned:

• ▪ HIC – not well resolved;

• ▪ ICE – complicated with charge heterogeneity from monoclonal Ab (mAb) in addition to difference in UV response;

• ▪ LC–MS/MS response decreases with increase of drug load.

▪ Quantitative evaluation of the impact of immunogenicity on PK/PD & efficacy of therapeutic proteins

Distinguishing the incidence of immunogenicity can prove challenging when patients have been previously treated with a biotherapeutic agent and may have residual immunogenicity or greater propensity to develop an immunogenic response. The trastuzumab emtansine (Kadcyla^®) label was presented to illustrate that Kadcyla was measurable in patient serum that tested positive for anti-Kadcyla antibodies at one or more post-dose time points and, thus, may have interfered with the assay’s ability to detect anti-Kadcyla antibodies. The three-tiered approach for trastuzumab emtansine immunogenicity was presented by Surinder Kaur (Genentech, CA, USA), where screening and titration assays were first used to detect anti-therapeutic antibodies and confirm the positive results. Finally, the assays were used to determine which component of the ADC is inducing an immune response – Ab, linker-payload or new epitopes in trastuzumab resulting from conjugation. Implementing a similar tiered approach in clinical studies may be a strategy that programs can implement to understand the specificity of the immune response.

▪ ADME to understand ADC–drug interaction

ADC–drug interaction (Di) assessment should span early to late development. A session dedicated to characterizing the ADME to understand the DDI potential of ADCs featured four speakers including: Karen Thudium (Novartis Pharmaceuticals, NJ, USA); Joe Balthasar (University at Buffalo, NY, USA); Ben Shen (Genentech, CA, USA); and Tae Han (Seattle Genetics, WA, USA). They discussed the importance of characterizing the catabolic, proteolytic and metabolic processes responsible for ADC degradation in preclinical and clinical development. The speakers presented a review of the ongoing clinical DDI studies and a summary of the historical and regulatory precedence (brentuximab vedotin and trastuzumab emtansine) for dedicated DDI studies. In the case of brentuximab vedotin, a dedicated DDI study was completed (using midazolam, rifampicin and ketoconzale as probes). The Kadcyla label was discussed, which cautions the use of trastzumab emtansine with strong CYP3A4 inhibitors due to the potential for an increase in DM1 exposure and toxicity. These suggestions are based solely on preclinical in vitro studies suggesting DM1 is metabolized mainly by CYP3A4, and to a reduced degree by CYP3A5. The metabolic and catabolic pathways driving ADC–Di considerations were highlighted for an early development program (Figure 1). It was recommended to characterize the ADME for the total Ab, ADC and payload/linker components. The deconjugation of the payload/linker component may result in the formation of metabolites – these metabolites should be characterized for cytochrome P450 (CYP) or non-CYP metabolism. Although the risk for PK-based drug interactions between ADC and concomitant medications is low for mAbs as a result of non-overlapping pathways of elimination with small-molecule drugs; the likelihood of an ADC acting as a ‘perpetrator’ and having an effect on the concomitant medication despite sharing the same drug metabolism pathway is relatively low. However, with numerous ongoing combination trials, including ADCs co-administered with small molecules or other large molecules (mAbs) it was recommended to consider concomitant medications in the intended patient population when designing clinical studies. ADC–Di should be evaluated in the context of safety/efficacy on a case-by-case basis. When the ADME/catabolic and metabolic pathways of concomitant medications are known it was recommended that clinical studies include characterization for all analytes with suspected or known interaction potential. However, in the case when these properties are not known, it is recommended that catabolic, metabolic and elimination pathways are defined; if an interaction potential is identified it should be further evaluated in a clinical study (Figure 1). It was recommended that sponsors identify ADC–Di potentials early in clinical development and Type B or Type C are an option to discuss study requirements with the FDA to avoid completing unnecessary dedicated DDI studies or subjecting critically ill patients to certain clinical pharmacology studies.

▪ Risk mitigation in ADC development

A session devoted to: ‘Risk mitigation in ADC development’ featured Clifford Dilea (Bristol-Myers Squibb, NJ, USA) and Don Mager (University at Buffalo, NY, USA). The speakers presented different approaches to aid in risk mitigation strategies in development of ADCs and drug development using PK/PD strategies.

Don Mager reviewed the semi-mechanistic population PK model that was recently published to describe the DAR PK of Kadcyla and its uses in risk mitigation of both very early and late development of ADCs [6]. Clifford Dilea reviewed the tumor kinetic model and its uses in predicting tumor effect from mouse xenograft models to man, as well as using the model with current available data to suggest the most optimized doses and dosing regimens for later development.

▪ ADCs in clinical development: initiating Phase 1 & beyond

Short courses with a limited number of seats are held annually at the NBC; the aim of these short courses is to provide a forum to present new and emerging data on selected topics in biotechnology. This year’s agenda boasted two sessions on ADCs. The clinical development of ADCs were discussed on Sunday 20 May 2013 in a course titled, ‘ADCs in clinical development: initiating Phase 1 and beyond’, which featured six speakers: Sanela Bilic (Novartis Pharmaceuticals, NJ, USA), Don Mager and Sandhya Girish (Genentech, CA, USA), Joseph Boni (Pfizer, PA, USA), Sarah Scrieber (FDA, MD, USA), Steve Olsen (Genentech, CA, USA), and Nancy Whiting (Seattle Genetics, WA, USA). This course examined both theoretical and practical considerations for ADC development based on the two recent approvals to increase awareness in the scientific community. Nancy Whiting and Steve Olsen discussed the late-stage clinical development of Adcetris and Kadcyla, respectively. This session highlighted the importance of understanding which PK metrics drive response. Don Mager highlighted the features of a semi-mechanistic population PK model that was recently published to describe the multivalent trastuzumab emtansine PK [6]. This model demonstrated the deconjugation process (from higher to lower drug to Ab ratios) and highlighted the shorter half-life of TDM1 relative to trastuzumab naked Ab. Joseph Boni covered the strategies employed in the clinical development of inotuzumab ozogamicin while identifying key learnings, misconceptions and emerging truths of the calicheamicin conjugates. Sandhya Girish described the Kadcyla submission package, including the discussion on immunogenicity presented in the submission, the clinical pharmacology strategies and various models used during the development of Kadcyla. Sarah Scrieber provided perspective on the relevant assays required for submissions and highlighted the importance of partnering with the FDA to discuss submission requirements. The high enrollment in this course and active attendee participation suggests a strong interest in defining a clear path forward for clinical development.

Summary & next steps in ADC development

ADCs became a central theme of the 2013 AAPS NBC; these molecules boast a new generation of targeted drug delivery for highly potent cytotoxic payloads. The sessions in 2013 built on previous years (preclinical and translational strategies) – the focus shifted to cover key questions that clinical teams face given limited historical precedence. Characterization of ADCs disposition can be one of the most important factors in understanding their efficacy and toxicity profiles. As sponsors initiate clinical develop programs it is important to consider highly integrated interactions with bioanalytical colleagues as assay specificity played a key role the submissions of brentuximab vedotin and trastuzumab emtansine. It was recommended throughout the meeting to merge small- and large-molecule drug-development paradigms in an effort to build successful clinical development plans for ADCs. Based on the precedent set by brentuximab vedotin, and more recently by trastuzumab emtansine, the current regulatory expectation is that PK of each analyte is characterized, metabolism and DDI potentials are evaluated, and immunogenicity assays are developed in accordance with FDA recommendations for mAbs, which sufficiently distinguish the different components contributing to an immunogenic potential.