Tuesday, May 21

Influencing ADC Performance through Linker Design

Webinar Q&A

Webinar Questions and Answers

We have received several questions looking into more details on the content discussed during our webinar on “Influencing ADC performance through Linker Design”. We are very excited to have collected detailed information from our expert speakers on the frequently asked questions and complied the responses for everyone to benefit from. Please send any additional questions to [email protected] and we will send you the responses, as well as include them in our FAQs.

Does Vector laboratories have solutions for ADCs internalization?

ADC internalization is largely a function of the target, the tissue, and the antibody. Conjugation of linkers and payloads to an antibody can influence the various internalization processes to different extents.

Internalization can occur through either non-specific pinocytosis or specific receptor-mediated endocytosis (RME). Receptor-mediated endocytosis can further be divided into either FcgR-mediated or target-mediated drug disposition (TMDD).

Internalization via pinocytosis by endothelial cells is a highly efficient process for IgG elimination that occurs throughout the body, particularly in tissues rich in capillary beds such as skin, muscle, and GI tract. This is a relatively unspecific fluid-phase endocytosis and is likely a highly undesirable pathway for ADCs since conjugates not recycled by the FcRn-mediated salvage pathway will undergo endolysosomal trafficking and release the payload in non-targeted cells, likely resulting in off-target toxicity. This process does not differentiate which proteins in the vicinity of a cell are internalized, and it can be difficult to influence via linker design by itself. However, it has been demonstrated that conjugation of methoxy-terminated PEG4, PEG8, or PEG12 modifiers to different numbers of lysine residues on a ~DAR4 MMAF ADC could reduce the non-specific uptake in an in vitro model of micropinocytosis (Cancer Res 2018;78:2115-2126).

RME and internalization via binding of the Fc domain to FcgRs can occur in many immune cells, such as monocytes, macrophages, and dendritic cells. This would also be a highly undesired pathway for ADCs since internalization and processing would result in payload release in non-targeted cells and off-target toxicity. Conjugation to the hinge region or the Fc domain has been shown to reduce FcgR binding affinities in some cases, and this could be a function of steric congestion and freedom-of-motion afforded by longer PEG spacers, although the effects on internalization were not explored (mAbs 4:3, 362-372, May/June 2012; Understanding the Effect of Discreet PEG Linkers on ADC Structure and Binding, Millipore Sigma, presented at 10th World ADC Conference London, March 3, 2020). It has also been shown that orthogonal PEG modifiers can reduce the non-targeted internalization of ADCs by Kupffer cells, the resident macrophages of the liver (Mol. Pharmaceutics 2020, 17, 3, 802–809). While the contribution of FcgR-mediated internalization of ADCs by Kupffer cells is still unclear, it has been firmly established in tumor-associated macrophages (Mol. Cancer Ther. 2017, 16, 1347–1354). The differences likely reflect the myriad of factors affecting binding to the different FcgR subtypes, including hydrophobicity and aggregation state, and PEG-based linkers may influence non-targeted internalization through these factors.

RME and internalization via TMDD is driven by the antibody binding to the target via the Fab domain, and this is usually the desired pathway for traditional ADC internalization, processing, and payload release. In principle, conjugation to lysine residues or engineered residues near the antigen binding domain has the potential to reduce internalization if binding affinity is negatively affected. However, this is more about the conjugation strategy rather than linker, and there is little to suggest that this process is substantially affected by PEG-based linker design per se.

While there are many literature reports of PEGylation affecting the internalization of lipid nanoparticles, antibody fragments, therapeutic peptides, and small molecules, when used as a linker in a traditional ADC the main role of PEG appears to be preventing undesired internalization by non-targeted cells.

Depending on the payload, ADC design, and type of internalization under consideration, we can recommend the exact Vector product that can be applicable. Please reach out to us at [email protected] for more details.

When considering cleavable vs non-cleavable linkers, you should consider the type of payload you want to conjugate. If your payload is capable of bystander activity, then a cleavable payload would be desired. If you are conjugating a TLR7/8 agonist or a payload where you want to keep the payload in the cell then a non-cleavable linker would be the best choice.

Non-cleavable linkers are an attractive option if high systemic stability of the ADC is a critical feature, or no bystander activity is desired. The lack of a cleavable trigger eliminates the possibility of premature payload release by non-selective extracellular conditions, and the PEG linker residue that remains attached to the payload after lysosomal degradation of the antibody backbone can prevent bystander activity and inhibit membrane permeability, all of which may reduce the potential for off-target toxicity. Without knowing the specific target or indication, there are some general comments that could be made. First, if a truly non-releasable payload is desired it would be important to utilize stable conjugation chemistry without the potential for deconjugation. Second, it would be important to verify that the payload can tolerate functionalization at the position used for linker attachment without substantial negative impacts on activity. Third, it would be important to verify that the steric bulk of the antibody does not impede the interaction of the payload with its target.

It is plausible that a PEG-based linker may prove useful in these cases since the proximity between the payload and the antibody can be increased by increasing the linker length without the concomitant increase in hydrophobicity that would be imparted by longer alkyl-based linkers. Fourth, if orthogonal linkers are used to shield payload hydrophobicity or compensate for payload properties, it should be verified that they do not negatively impact activity. And fifth, if the ADC is internalized it will undergo endolysosomal trafficking and catabolism of the antibody backbone to “release” the payload attached to the linker and the amino acid residue from the antibody. Based on the previous discussion points the linker-payload will likely retain interaction with its target, however the attached linker residue may impede membrane permeability, endosomal escape, and internalization of the free linker-payload, so these factors should also be taken into consideration.

Bioconjugate Techniques and Chemical Linkers in Antibody-Drug Conjugates are very informational. Primary research articles can also be a good source of information.

While the amide bond (and related peptide bond) is indeed enzymatically labile in vivo, and much effort is put into designing therapeutics with amide bioisosteres and peptidomimetics for improved metabolic stability, it does not appear to be a primary concern for ADCs. There are hundreds of proteases ubiquitously distributed throughout the body, however they can be location specific and have different substrate specificities, thus a rigorous answer depends on the exact nature of the ADC, linker, and payload.

Most ADC linkers incorporating amide bonds appear to make it to the lysosome intact. For instance, there are peptide-drug conjugates (J. Med. Chem. 2022, 65, 21, 14337–14347) and antibody-drug conjugates (Pharmaceuticals 2021, 14(3), 247) that utilize polysarcosine linkers, and amide bond instability is not reported as a concern. In a study exploring the stability of non-cleavable ADCs comprised of MMAD payloads conjugated to glutamine residues via a PEG6 linker, no degradation of the amide bonds between the PEG and the glutamine side chain or between the PEG and the payload was observed (PLoS ONE 10(7): e0132282).

In another study exploring the in vivo potency of non-cleavable ADCs with maytansinoid payloads and PEG4 linkers, the main products after lysosomal catabolism of the ADC retained the amide bonds between the PEG and the lysine side chain and between the PEG and the drug moiety. This resulted in a free drug-linker that was not susceptible to efflux and thus more effective against MDR cancer cell lines (J. Med. Chem. 2011, 54, 10, 3606–3623). Thus, while there is always risk, the amide bonds and PEG backbones in ADC linkers appear systemically stable and the more critical factors influencing premature release are deconjugation from the antibody or extracellular activation of cleavable triggers.

An ADC could have chemical stability or physical stability. Chemical stability of an ADC generally refers to the bonds between the antibody, the linker, and the payload, and an orthogonal linker can be used to improve the stability of the bond between a cleavable linker and a payload by restricting enzyme access to the enzymatically labile cleavable trigger. The efficacy of this approach depends on the cleavable trigger, the enzyme, and the linker (Bioconjugate Chem. 2021, 32, 10, 2257–2267). In this context, stability is synonymous with payload release, thus improved stability means a reduced rate of payload release, and its overall effect on in vivo efficacy depends on the balance between systemic stability and intracellular payload release. It should be noted that a reduced rate of intracellular payload release may not affect the desired end point (Bioconjug Chem. 2019 Nov 20;30(11):2982-2988. doi: 10.1021/acs.bioconjchem.9b00713), and this could provide a strategy to find the optimal payload release rate. If the payloads are on a branched-type linker where they are all exposed at the ends of each of the branching arms there may be no improvement in chemical stability.

Physical stability of an ADC generally refers to aggregation. In this case the ability of a linear PEG spacer or an orthogonal PEG modifier to improve aggregation is highly dependent on the reactive group for conjugation, the cleavable trigger, and payload. This was detailed in the Webinar, but in brief, if the cleavable trigger and payload can be chemically modified to improve properties without adversely affecting activity, a linear PEG spacer can be sufficient to reduce the propensity for aggregation (or it may not be needed at all). If, on the other hand, the payload cannot be altered without unacceptable reductions in potency or if it just an inherently hydrophobic moiety, an orthogonal PEG modifier may be able to shield the payload from undesired associations. If these types of payloads are positioned at the terminus of linear PEG spacers or branching PEG arms, aggregation issues may be exacerbated.

The answer to the second question again depends on the type of stability in question. In general, it appears that more payloads conjugated to fewer sites on the antibody can improve an ADCs efficacy, reduce an ADC’s toxicity, improve an ADCs PK, and improve the physical stability, relative to the same number of payloads conjugated to more sites on the antibody. However, as always, this depends on the conjugation site, the conjugation chemistry, the payload, and the linker design. Mersana’s various publications and posters on the Synthemer scaffold are worth looking at for more details on this topic.

The antibody subclass would likely affect the choice of conjugation chemistry more than the PEG-based linker design. Stochastic conjugation strategies targeting lysine residues should always be evaluated on a case-by-case basis, but this approach always results in a heterogenous population of conjugates regardless of the antibody subclass. The four subclasses, IgG1, IgG2, IgG3, and IgG4 are highly conserved but have significant differences in the constant region, particularly in the hinge regions and CH2 domains, thus conjugation strategies targeting these regions would need to be evaluated. For instance, the four subtypes differ in the number of amino acids and inter-heavy chain disulfide bonds in the hinge region, which affects rigidity and stability. These differences may affect the possible DAR values and conjugation efficiencies when conjugating maleimide-based reagents to the reduced cysteine residues. The differences in the hinge region may also affect conjugation with the various reactive groups for disulfide re-bridging, which have been shown to result in different conjugation efficiencies and disulfide scrambling within the single IgG1 subclass (Bioconjug Chem. 2021 Sep 15;32(9):1947-1959; Org. Biomol. Chem., 2018,16, 1359-1366; ACS Omega. 2020 Jan 28; 5(3): 1557–1565).

While the subclasses do differ in the constant region, they all conserve Q295 thus, in principle, enzymatic ligations using mTG are applicable to all the subclasses. Similarly, almost all IgG-type antibodies contain a conserved glycan at position N297, and presumably oxidation or enzymatic glycan remodeling are unaffected by subclass, so glycan-mediated conjugation strategies can also likely be applied to all subtypes. The different subclasses have different effector functions due to the differences in the Fc regions that affect binding to FcgRs and C1q, and this may also affect the binding of affinity peptides that are employed for proximity-directed conjugations. Some affinity peptide reagents, such as the AJICAP technology, have demonstrated applicability to multiple subclasses (Angew. Chem. Int. Ed. 10.1002/anie.201814215), however, others should be evaluated on a case-by-case basis if not documented in the literature. Most conjugation technologies are developed using the IgG1 subclass, and while in principle many should be applicable to the other subclasses they would need to be evaluated carefully.

Explore BioDesign