Jennifer Mitcham, SMARTag business development, Catalent Biologics, offers insight into the new generation of antibody drug conjugates and the increasing importance biologics play in the fight against disease
Jennifer Mitcham, Catalent
A decade ago, the world’s top ten selling medicines were all small molecules. In 2014, just half of the ten biggest sellers were small molecules, with four monoclonal antibodies and a modified form of insulin.
With drugs for the ‘easier’ targets available and many now adequately served with generic medicines, pharma and biotech companies have become more creative. Biologics have become an important part of the fight against disease and the challenges facing the medicines market are how to make safer, more efficacious drugs; exploring innovative combination therapies; and developing more efficient manufacturing processes. Antibody–drug conjugates (ADCs) are one example of how combining a biologic’s targeting mechanism action with the potency of a chemotherapeutic toxin can have significant benefits.
The first generation
The first ADC, Wyeth’s Mylotarg, launched in 2000 and although later withdrawn for lack of efficacy, two others have since reached the market- anti-cancer products Kadcyla (Genentech) and Adcetris (Seattle Genetics). Dozens more are in development and have great potential against solid tumours and haematological malignancies.
ADCs harness the targeting ability of an antibody to direct a highly potent cytotoxic agent to cancer cells, while sparing healthy ones. The two components, antibody and cytotoxin, are joined by a chemical linker. The nature of the linker and the conjugation chemistry critically impacts the characteristics of the resulting product. First-generation ADCs pose challenges in manufacturing and characterisation, mainly because of the variability of the conjugates produced.
These ADCs relied on linkers that reacted with surface-accessible nucleophilic amino acid side chains – either the amine group on lysine, or the thiol moiety on cysteine. Lysine technology, used to make Kadcyla, relies on linkers with active esters including sulfosuccinimide or N-hydroxysuccinimide that react with accessible lysine amines on the antibody. The downside is that the typical antibody has dozens of such reactive lysines, precluding control over the precise location of conjugation, resulting in a heterogeneous mixture of molecules bearing between zero and eight drug molecules, conjugated at a diversity of sites.
Adcetris uses a cysteine conjugation approach that takes advantage of thiol moieties existing as disulfides in an intact antibody. Before conjugation, a reducing agent such as tris(2-carboxyethyl) phosphine or dithiothreitol must be used to reduce the disulfide bonds, releasing free thiol groups that can be reacted with maleimide-containing linkers. A maximum of four of these disulfides can be reduced in an antibody, producing up to eight thiols for conjugation. This number can be varied by altering reducing conditions. Although there is still variability in the number and location of the attached cytotoxic molecules, the overall complexity of cysteine conjugates is lower than that observed when using lysine conjugation.
A new breed of ADCs
The heterogeneity of the first-generation lysine and cysteine conjugation approaches has led to second generation ADC technologies, designed to reduce variability and optimise drug to antibody ratio (DAR). A DAR that is too low results in suboptimal efficacy, while too high makes side-effects more likely. Site-specific conjugation is a new approach targeting tighter control over the DAR, while reducing complexity and variability in terms of conjugation site.
Site-specific conjugation, pioneered by Genentech with its THIOMAB-drug conjugate (TDC), involves engineering reactive cysteine residues at specific locations on the antibody, and using these for conjugation rather than reducing native disulfides. Genentech has reported a comparison between the ADC and TDC equivalents of an anti-MUC16 antibody conjugated to the cytotoxic agent MMAE. The product gave the same cytotoxic effect with half the payload in preclinical studies, and was better tolerated.1 Similar results were seen with TDC versions of Kadcyla (trastuzumab emtansine).2
Further investigations indicated that the precise location of the conjugation site on the antibody also has affects ADC biophysical and functional properties. Conjugates within the antibody’s light chain showed the greatest activity, conjugates in the in the heavy chain moderate activity, and conjugates in the Fc region negligible activity in in vivo efficacy studies. A similar trend was seen for stability, with clearance slowest when conjugation was in the light chain, and fastest when it was in the Fc region.
Early work on TDC encouraged the development of other potential techniques for site-specific bioconjugation. These techniques use protein engineering to move the conjugation site around the antibody so the properties of different potential ADCs can be explored. While all capable of creating more homogeneous ADCs than first generation techniques, not all facilitate creation of the optimal conjugation patterns for any specific combination of antibody and payload.
Allozyne, Ambrx, and Sutro Biopharma have developed ways to introduce non-natural amino acids into the antibody via genetic engineering, then using them to give specific conjugation patterns. Recombinant protein expression systems have engineered transfer RNA synthetase pairs, used to introduce non-natural amino acids bearing bioorthogonal reactive groups anywhere on the antibody backbone, allowing site-specific conjugation.
For example, Ambrx incorporated p-acetyl phenylalanine into antibodies using mammalian expression systems. The arylketone side chain can be used to conjugate cytotoxic drugs using oxime bonds. Ambrx and Agensys show that ADCs made in this way are, chiefly, just as efficacious as engineered or conventional cysteine-conjugated ADCs, but have better pharmacokinetic and safety properties. This is attributed to the greater stability of the oxime bond by researchers at both companies.3,4
Sutro Biopharma can introduce a non-natural amino acid bearing an azide-containing side chain into antibodies made in a cell-free expression system. Resulting antibodies are conjugated to cytotoxic drugs via copper-catalysed or copper-free click chemistry.
Enzymatic techniques
Pfizer employs an alternative strategy. It uses the enzyme transglutaminase to mediate conjugation reactions between glutamine residues and primary amines, with the site directed using a glutamine tag, engineered into the primary sequence of the antibody. Pfizer and scientists at ETH University, Zurich, have shown that site-specific ADCs can be made in this way, as natural glutamine residues in the antibody backbone are not affected.
Pfizer has also shown that a small glutamine-containing tag can be used as a transglutaminase substrate when introduced in surface-accessible antibody sites, before being conjugated to a cytotoxic payload using an amine-functionalised linker. Cleavable and non-cleavable linkers can be used, along with a variety of payloads. ADCs with comparable efficacy but better safety and tolerability than conventional products can be made in this way. It is scalable and compatible with multiple different payloads, and amide bonds in the conjugates are generally stable, with no need to incorporate non-natural amino acids within the antibody.
Redwood Bioscience, part of Catalent Biologics, introduced aldehyde-tagged antibodies created via naturally occurring enzyme, formylglycine-generating enzyme (FGE), to produce site-specific conjugation. SMARTag technology uses FGE to introduce formylglycine residues into protein backbones via insertion of a short FGE consensus sequence in the conserved regions of the antibody’s heavy or light chains. The ‘tagged’ antibody is then produced in cells that overexpress FGE, converting a cysteine within the consensus sequence to a formylglycine residue bearing a bioorthongonally-reactive aldehyde group. With FGE-mediated chemoenzymatic modification occurring cotranslationally during antibody production, no additional enzymes or protein-modifying agents are required during conjugation. Expression levels of 3 g/L and high yields (>95%) of formylglycine conversion have been demonstrated up to 100 L scale. After standard purification methods, proprietary aldehyde-specific conjugation chemistries form stable carbon–carbon bonds between payload and antibody at the tag locations.
This technique can create site-specific ADCs with maytansine payloads conjugated at different locations on an anti-HER2 antibody. Site-specific conjugation achieved via a carbon–carbon bond results in a highly stable product that is independent of the local protein microenvironment, demonstrated by serum stability experiments. However, the nature of the different conjugation sites influences pharmacokinetic behaviour and efficacy in tumour xenograft models, indicating that the conjugation site matters when building a better ADC. Although the aldehyde tag approach entails the introduction of a few exogenous amino acids into the antibody backbone, in silico and ex vivo tests assessing the immunogenicity potential of various aldehyde tag sites and conjugation products showed that the risk was low, and was similar to the native, untagged antibody. The safety of SMARTag ADCs has been demonstrated in exploratory toxicology studies conducted in rat and cynomolgus monkeys, with doses up to 60 mg/kg being well tolerated in single and repeat dose studies.
What’s next?
Creating new ADC molecules is complex, with many different parameters needing to be orchestrated to generate the optimal conjugate. However, the ability of site-specific conjugation to produce increased homogeneous populations of ADCs should be a significant advantage in the development process. Manufacture, characterisation, and CMC processes should all benefit from a degree of simplification. The ability to optimise both DAR and payload placement may also improve the therapeutic index for ADCs.
The field is still relatively young, with the majority of new ADCs still created using the conventional cysteine or lysine conjugation processes that generate heterogeneous ADCs. However, site-specific ADCs are now starting to reach early phase clinical development. With various different techniques available for making site-specific ADCs, and preclinical studies showing the potential to offer significant benefits in terms of efficacy, pharmacokinetics, and safety over first-generation iterations, the future of ADCs is bright.
References
1 Junutula JR, et al. Site-Specific Conjugation of a Cytotoxic Drug to an Antibody Improves the Therapeutic Index. Nat. Biotechnol. 26(8) 2008: 928–932.
2 Junutula JR, et al. Engineered Thio-Trastuzumab-DM1 Conjugate with an Improved Therapeutic Index to Target Human Epidermal Growth Factor Receptor 2–Positive Breast Cancer. Clin. Cancer Res. 16, 2010: 4769–4778
3 Tian F, et al. A General Approach to Site-Specific Antibody Drug Conjugates. Proc. Natl. Acad. Sci. 111(5) 2014: 1766–1771.
4 Jackson D, et al. In Vitro and In Vivo Evaluation of Cysteine and Site Specific Conjugated Herceptin Antibody-Drug Conjugates. PLOS One 14 January 2014: 1–14