Posts Tagged ‘Constructing’

Antibody drug conjugates (ADCs)

Larry H. Bernstein, MD, FCAP, Curator



Processes for Constructing Homogeneous Antibody Drug Conjugates


Processes for Constructing Homogeneous Antibody Drug Conjugates

Igenica Biotherapeutics, 863A Mitten Road, Suite 100B, Burlingame, California 94010, United States
Org. Process Res. Dev., Article ASAP

Abstract Image

Antibody drug conjugates (ADCs) are synthesized by conjugating a cytotoxic drug or “payload” to a monoclonal antibody. The payloads are conjugated using amino or sulfhydryl specific linkers that react with lysines or cysteines on the antibody surface. A typical antibody contains over 60 lysines and up to 12 cysteines as potential conjugation sites. The desired DAR (drugs/antibody ratio) depends on a number of different factors and ranges from two to eight drugs/antibody. The discrepancy between the number of potential conjugation sites and the desired DAR, combined with use of conventional conjugation methods that are not site-specific, results in heterogeneous ADCs that vary in both DAR and conjugation sites. Heterogeneous ADCs contain significant fractions with suboptimal DARs that are known to possess undesired pharmacological properties. As a result, new methods for synthesizing homogeneous ADCs have been developed in order to increase their potential as therapeutic agents. This article will review recently reported processes for preparing ADCs with improved homogeneity. The advantages and potential limitations of each process are discussed, with emphasis on efficiency, quality, and in vivo efficacy relative to similar heterogeneous ADCs.

Antibody drug conjugates (ADCs) are a rapidly growing class of targeted therapeutic agents for treatment of cancer.(1-8) Although the number of ADCs in clinical trials has steadily increased since 2005, many have failed to reach the later stages of clinical development; one has been withdrawn from the market (Mylotarg in 2002), and only two (Adcetris and Kadcyla) are currently approved by the FDA for cancer indications (Figure 1A).(9-11) Thus, far, the approval rate for ADCs has not met early expectations and is lagging behind other antibody-based therapeutics. Based on the number of approved ADCs versus those that have failed to progress into later stage clinical trials, the success rate is reminiscent of that for small molecule drugs. The reasons for the clinical failures of ADCs are often not known or they are still under investigation. More commonly, when the reasons for clinical failure are clear, the information is not made available to the public domain. Emerging preclinical data suggests that heterogeneity, a property shared by most ADCs currently in clinical development (Table 1), may ultimately limit their potential as therapeutic agents.(12, 13)

Table 1. Examples of Heterogeneous ADCs Currently in Clinical Trials for Cancer Indicationsa

a Source: www.clinicaltrials.gov.




Figure 1. (A) Number of ADCs in different stages of clinical development from 2006 to 2014. (B) Structure of a typical IgG antibody showing lysines (red), cysteines (yellow), and glycans (green) as potential conjugation sites.(16)

ADCs are composed of a cytotoxic drug or “payload” conjugated to a tumor selective monoclonal antibody. The heterogeneity of conventional ADCs arises from the synthetic processes currently used for conjugation.(14) Payloads are typically conjugated to the antibody using amino or thiol specific linkers that react with lysines or cysteines on the antibody surface.(15) A typical antibody contains more than 50 lysines and up to 12 cysteines as potential conjugation sites (Figure 1B).(16) The optimal DAR (drugs/antibody ratio) for most ADCs, however, ranges from 2 to 8 drugs/antibody and is dependent upon a variety of different factors. The discrepancy between the number of potential conjugation sites and the desired DAR, combined with the use of conjugation methods that are not site-specific, result in heterogeneous ADCs that vary in both DAR and conjugation sites. Consequently, conventional heterogeneous ADCs often contain significant amounts of unconjugated antibody in addition to fractions with suboptimal DARs. Unconjugated antibodies can compete for antigen binding and inhibit ADC activity, while fractions with suboptimal DARs are frequently prone to aggregation, poor solubility, and/or instability that ultimately result in a poor therapeutic window.(17, 18)
The relative degree of ADC heterogeneity depends on the methods used for conjugation. For example, Kadcyla, an ADC approved in 2013 for breast cancer, is synthesized using a two-step process in which the linker and payload are conjugated in separate steps (Scheme 1A).(19-21)The linker contains an amino-specific NHS ester that reacts with antibody lysines in the first step and a thiol-specific maleimide group that reacts with a maytansinoid payload in the second step. The process affords a highly heterogeneous mixture of ADC molecules containing from 0 to 10 payloads/antibody with an average DAR of 3.5 drugs/antibody.(22, 23) Additional heterogeneity arises due to distribution of the payloads across dozens of potential conjugation sites. As a result, Kadcyla contains hundreds of different ADC molecules, each with its own unique pharmacological properties.(24)
Scheme 1. (A) General Process for Synthesizing ADCs such as Kadcyla via Lysine Conjugation; (B) General Process for Synthesizing ADCs, such as Adcetris, via Cysteine Conjugation
Conjugation of payloads to antibodies through interchain cysteines reduces ADC heterogeneity relative to lysine conjugation because there are fewer potential conjugation sites. Adcetris, an ADC approved in 2011 for treatment of Hodgkin’s lymphoma, is an example of a cysteine conjugated ADC.(25-27) The process for cysteine conjugation involves partial reduction of four antibody interchain disulfide bonds to generate up to eight reactive thiol groups. The partially reduced antibody is subsequently conjugated to a payload containing a thiol-specific maleimide linker. The payload used for Adcetris is monomethyl auristatin E (MMAE) and contains a protease cleavable maleimide linker (Scheme 1B). Although Adcetris is less heterogeneous than Kadcyla, it is composed of dozens of different ADC molecules containing 0 to 8 payloads with an average DAR of 3.6 drugs/antibody.(28) Like most cysteine conjugated ADCs, Adcetris has a reduced half-life in vivo compared to the parent antibody, cAC10. The diminished half-life has been attributed to rapid clearance of high DAR species (>4 drugs/antibody) and to partial loss of interchain disulfide bonds during the conjugation process.(29, 30)
Although different processes for lysine and cysteine conjugation are used to synthesize Adcetris and Kadcyla, both ADCs contain thio-succinimide bonds between the payload and the antibody, which originate from the use of maleimide linkers in the conjugation processes. Kadcyla contains a thio-succinimide between the linker and the payload (Scheme 1A), while Adcetris contains a thio-succinimide bond between the linker and the antibody (Scheme 1B). Thio-succinimide groups are known to undergo undesired side reactions such as elimination or thiol exchange that can result in premature release of the payloads from the ADC and lead to reduced potency and/or increased systemic toxicity.(31, 32)
Despite the known limitations of conventional heterogeneous ADCs, most ADCs currently in clinical development utilize similar conjugation methods to those described in Scheme 1. As a result, they are likely to possess similar pharmacological properties to Adcetris and Kadcyla, in addition to other less successful ADCs that may have performed poorly in clinical trials. In order to improve the pharmacological properties of current and future ADCs, new site-specific conjugation processes for synthesizing homogeneous ADCs are now being developed.(33-36)
Site-specific conjugation processes for constructing homogeneous ADCs can be divided into three different categories. Two are focused on antibody modification (engineered amino acids and enzyme mediated), while the third category is focused on linker modification. The categories can be subdivided further based on the specific processes that are used (Table 2). Examples from each process were selected based on availability of sufficient preclinical data to enable comparison with similar conventional heterogeneous ADCs. Homogeneous ADCs derived from these processes have only just begun to enter clinical trials. Whether they will outperform their heterogeneous counterparts in clinical trials remains uncertain, but preclinical data suggest that homogeneous ADCs are likely to dominate future clinical trials and will lead to improved clinical results.
Table 2. Summary of Different Processes for Constructing Homogeneous ADCs
All of the processes reviewed here were successfully used to construct ADCs with improved homogeneity over ADCs synthesized using conventional methods. A majority of approaches utilize recombinant antibody engineering to introduce unique functional groups for site-specific conjugation. The unique functional groups were introduced either as point mutations for cysteine and non-natural amino acids or as enzyme recognition tags. These recombinant engineering approaches offer several potential advantages over nonrecombinant approaches. For example, engineered cysteines can be incorporated into dozens of different sites with minimal impact on the functional properties of the antibody. This enables ADCs to be optimized for conjugation efficiency, linker stability, and potency. Engineered non-natural amino acids offer additional advantages due to the diverse array of different functional groups that can be introduced. Furthermore, non-natural amino acids enable a variety of new linker chemistries to be investigated that are not possible with conventional conjugation processes.
The flexibility offered by recombinant processes may also represent their greatest challenge. The importance of the conjugation site for ADC activity is well-established, but additional factors should be considered before selecting a development candidate. Potential effects on antibody expression, conjugation efficiency, linker stability, aggregation, and other factors need to be considered before selecting a specific conjugation site. These factors can ultimately determine the success or failure of an ADC development program. Since antibodies share many of the same properties, it seems likely that optimal conjugation sites will be identified that are broadly effective when used with different antibodies. Other potential challenges for processes involving antibody engineering include increased development time and costs, immunogenicity of engineered sequence tags, scalability, and use of novel linkers and payloads that are not yet clinically validated.
In addition to homogeneity, improvements in other ADC properties such as potency, stability and half-life were observed. In fact, many of the homogeneous ADCs derived from these processes out-performed conventional heterogeneous ADCs in efficacy and safety studies. Much of their success has been attributed to elimination of high DAR species present in conventional ADCs. In general, experimental results are consistent with this conclusion, and many would agree that substantial progress has resulted from these efforts to improve ADC homogeneity. Ironically, the relative contribution of homogeneity to the improved properties of the engineered ADCs could not be determined from most studies because other factors known to effect ADC activity could not be ruled out.
For instance, recombinant approaches for making homogeneous ADCs were designed to introduce conjugation sites in different locations from those used in conventional methods. Since it is now well-established that “location matters”, the observed differences in activity between TDCs (or NDCs) and the conventional ADC controls could result from different conjugation sites, rather than from elimination of high DAR species. Enzyme mediated approaches face similar challenges when comparing homogeneous and heterogeneous ADCs because the conjugation sites are different. Other variables such as linker type (cleavable or noncleavable) and payload (maytansine or PBD) need to be carefully controlled before reaching conclusions about the benefits of homogeneity.
Linker based processes are more suitable for comparing homogeneous ADCs with conventional heterogeneous ADCs because they utilize the same conjugation sites. Once other variables that might impact ADC activity were carefully controlled, the relative benefits of homogeneity were revealed for the first time and the results confirmed that efforts to improve ADC homogeneity have been a worthwhile endeavor.
Most of the processes reviewed here are still in early phases of clinical development. All of the methods have advantages and limitations that will ultimately decide which approach will become the preferred process for manufacturing homogeneous ADCs. It is not yet clear which process will rise above the others as a preferred method, but all of these approaches have contributed valuable information to our knowledge base and resulted in ADCs with improved pharmacological properties over conventional heterogeneous ADCs. Our future challenge will be to apply this knowledge to develop ADCs that will be more effective as therapeutic agents. Our ability to synthesize homogeneous ADCs provides another reason to be optimistic about the future of ADCs.
ACS Editors’ Choice – This is an open access article published under an ACS AuthorChoice License, which permits copying and redistribution of the article or any adaptations for non-commercial purposes.

Read Full Post »

%d bloggers like this: