In the early stages of ADC drug development, linker design was often underestimated due to the initial belief that simply adding a cytotoxin to an antibody was sufficient for ADC drug development.
However, as iterations of ADCs have been developed, and as successes and failures have been experienced in the clinic, it has now been clearly demonstrated in at least two ways that the linker is a key element of the overall ADC design:
1.The linker can be architecturally modified to optimize the therapeutic index (TI);
2.The linker must ensure that the correct amount of cytotoxin is delivered to the correct cell (Figure 1).
To develop a successful ADC, it is essential to find the appropriate combination of antibody, linker, and payload. Different cancer treatment targets require distinct ADC linker structural designs. Various structural elements of the linker, including the attachment site, the release mechanism, the solubilizing portion, and the proper design of these different units, can regulate the ADC’s functionality, safety, and manufacturability, establishing a suitable Target Product Profile (TPP) (Table 1). This article provides a brief overview of how linker design impacts ADC performance, aiming to serve as a reference for our future ADC development.
Linker Design and ADC Potency
Payload potency is a major driver of ADC potency because reduced levels of antigen expression limit the delivery of ADCs and payloads to target cells. Different classes of payloads tend to have different cytotoxicity, influenced by their specific MOA; for example, DNA cross-linkers such as PBD dimers, cyclopropylbenzindole dimers (CBIs), or indolylbenzodiazepine dimers (IGNs) are among the most potent payloads, with IC50s typically in the PM or sub-PM ranges in in vitro cell-killing assays.
In xenograft mouse models, ADCs with these payloads can induce cure after a single injection as low as 100ug/kg. Anti-tumor agents that interfere with microtubule protein polymerization, such as the auristatins derivatives auristatins and the medenosin derivatives maytansinoids, typically exhibit sub-nM IC50 values and can induce cure in xenograft mouse models with a single injection within mpk.
Topoisomerase inhibitors, such as camptothecin derivatives, lie at the lowest end of the ADC payload with IC50 values in the nM range. In heterozygous mouse models, a single dose in the 10 mpk range is required to achieve cure.
However, for a given payload, ADC activity can be affected by modifying the linker. One of the key parameters affecting the potency of ADCs is the ability of their released payload or active metabolite to have a cytotoxic effect on neighboring cells. The effect of the linker on this so-called bystander activity can be well demonstrated in in vitro assays in which antigen-positive and antigen-negative cells are co-cultured.
For example, in an in vitro co-culture containing antigen-negative Namalwa cells and antigen-positive Colo 205 cells, the bystander activity of ADCs directed against the tumor antigen CanAg, which connects different maytansinoids payloads via disulfide-bonded cleavable linkers or non-cleavable linkers, was investigated. Cells were treated with different ADCs at fixed concentrations in order to be able to kill all antigen-positive cells but not antigen-negative cells.
Know more about cleavable linkers: Cleavable linkers play a pivotal role in the success of antibody-drug conjugates (ADCs)
Need to buy cleavable linker products, you can go to: https://axispharm.com/product-category/cleavable-linkers/
For instance, in an in vitro co-culture of antigen-negative Namalwa cells and antigen-positive Colo 205 cells, the bystander activity of an ADC targeting the tumor antigen CanAg was studied, where the ADC was conjugated to different maytansinoid payloads through either a disulfide-cleavable or a non-cleavable linker. Cells were treated with different ADCs at fixed concentrations sufficient to kill all antigen-positive cells without affecting antigen-negative cells.
For the ADC with a disulfide-cleavable linker, as the proportion of Colo 205 cells in the co-culture increased, the killing effect on Namalwa cells also increased, indicating that the metabolites of the cleavable linker ADC could diffuse and kill neighboring cells. Using a non-cleavable MCC-DM1 ADC, although it effectively killed Colo 205 cells, Namalwa cells were unaffected regardless of the Colo 205 cell level in the co-culture, indicating that the charged Lys-MCC-DM1 metabolite lacked efficacy against neighboring cells.
For these reducible disulfide-based linkers, the cleavage efficiency can also be adjusted by varying the steric hindrance around the disulfide bond. Using the CanAg-targeting huC242 antibody, a series of maytansinoid-based ADCs with varying levels of steric hindrance on the disulfide bond were produced. All these ADCs were prepared by conjugating the maytansinoid payload to an SPP linker via lysine. As the number of methyl substituents on the carbon adjacent to the disulfide bond increased (0 for DM1, 1 for DM3, 2 for DM4), various maytansinoid payloads were linked to the SPP linker.
When evaluated in a Colo 205 xenograft mouse model, despite unfavorable PK leading to much lower tumor exposure levels, the least hindered ADC, huC242-SPP-DM1, showed better efficacy than other cleavable linker forms and resulted in complete responses in some animals. In this specific model, the cleavage ability of the disulfide bond was shown to be a determining factor for ADC efficacy.
As part of the preclinical development of the FR-α-targeting ADC mirvetuximab soravtansine, researchers compared various maytansinoid-based ADC linker forms, including three disulfide-cleavable linkers (SPP-DM1, SPDB-DM4, and Sulfo-SPDB-DM4) and one non-cleavable linker (MCC-DM1).
It was found that in high-expressing KB cells, the IC50 was approximately 0.1 nM. In low-expressing Igrov-1 and JEG-3 cell lines, the activity of the non-cleavable linker decreased by 1-2 log10, while the efficacy of the cleavable linker ADCs remained nearly unchanged. When evaluated in vivo, ADCs with disulfide-cleavable linkers also maintained higher efficiency.
Bystander activity not only helps improve efficacy against tumors with varying levels of antigen expression but may also be more beneficial for the treatment of solid tumors.
Another illustrative example of adjusting ADC efficiency by modulating linker stability comes from Immunomedics’ development of ADCs using the topoisomerase I inhibitor SN-38.
Compared to ADCs with more stable, enzyme-cleavable linkers, the in vivo efficacy of SN-38 ADCs with hydrolysable linkers is superior because the release level of the payload with bystander effect is higher in the TME. The FDA-approved ADC Trodelvy utilizes this hydrolysable linker form.
Typically, in vitro ADC potency correlates with DAR values. However, other factors can influence ADC performance in vivo, favoring lower DARs. Therefore, DAR optimization is an integral part of ADC development once the appropriate combination of antibody, payload, and release mechanism has been established.
For most regulatory-approved ADCs, optimal DAR studies were conducted during preclinical development. For example, based on comprehensive results from in vivo efficacy and tolerability studies in mouse xenografts, Brentuximab ADC with the MC-Val-Cit-PABC-MMAE linker showed the optimal TI with DAR 4 compared to DAR 2 and DAR 8 ADCs.
Preclinical studies also compared deruxtecan DAR 3.4 and DAR 8 candidates. Given that DAR 8 ADC showed higher efficacy in low HER2-expressing cell lines compared to DAR 3.4 ADC, DAR 8 ADC DS-8201a was selected as a clinical candidate.
Conversely, another ADC targeting Trop-2, DS-1062, used DAR 4 in clinical trials, as did DS-7300a targeting B7-H3, indicating that the optimal drug load may also depend on antigen expression levels and profiles.
The two FDA-approved calicheamicin ADCs, Mylotarg and Besponsa, although targeting antigens CD33 and CD22 in hematologic malignancies, have different DARs. Besponsa has DAR 5-7, while Mylotarg has DAR 2-3, further emphasizing the importance of drug load as a critical design parameter and the need for DAR optimization in ADC candidate selection.
Some discrepancies between the in vitro potency and in vivo efficacy of high-DAR ADCs relate to the hydrophobic nature of the payload and some linker elements, such as the commonly used Val-Cit-PABC.
Tumor exposure is a crucial determinant of ADC efficacy. High DAR ADCs with increased hydrophobicity from Val-Cit-PABC linkers can be cleared more quickly, counterbalancing the higher potency of high DAR ADCs with their faster clearance.
One method to reduce linker-payload hydrophobicity is to introduce hydrophilic groups, which has been shown to improve the PK profile of some ADCs. For example, introducing sulfo groups into the linker of disulfide-based maytansinoid ADCs increased their hydrophilicity.
Read more: The High Hydrophilicity of ADC Linker Has Become the Main Trend in Modification
The sulfo-SPDB-DM4 linker-payload was more effective in vivo against low FR-α-expressing cancer cell lines compared to SPDB-DM4.
Introducing simple polar segments into the linker can also significantly impact ADC efficacy. This is demonstrated in Synaffix’s HydraSpace technology-based ADCs. By incorporating polar sulfonamide spacer units into various linker payloads and then site-specifically conjugating them to remodeled glycan residues on the antibody.
Compared to ADCs using the same elements but lacking polar sulfonamide spacers, incorporating polar sulfonamide spacers into ADCs reduces aggregation tendencies, improves PK, and significantly enhances in vivo efficacy.
Carbohydrates, as highly hydrophilic moieties, have also been used to improve ADC solubility and PK. Specifically, these moieties can serve as part of the ADC linker release unit, cleaved by enzymes specific to the lysosome, such as β-glucuronidase and β-galactosidase.
In the Karpas-299 xenograft model, a single dose of 0.5 mpk of Brentuximab MMAE DAR 8 ADC with a β-glucuronidase release unit achieved complete remission in all animals. In contrast, in another study, the non-β-glucuronidase release unit MC-Val-Cit-PABC-MMAE DAR8 ADC at a single 1 mpk dose resulted in only partial remission in most animals.
Introducing hydrophilic and solubilizing moieties into the linker is not the sole method to reduce the overall hydrophobicity of ADCs. Shielding linker-payload hydrophobicity is another effective approach to improving ADC PK. The antibody itself can interact with the linker-payload to attenuate hydrophobicity.
PK Study related resources:
Bioanalytical Service – PK Studies
Intrinsic relationship of PK/PD study
For instance, the reason why ADCs with short linkers lacking particularly hydrophilic characteristics (such as MC or MCC linkers) still exhibit acceptable PK might be attributed to the shielding effect provided by the antibody. Another way to shield hydrophobic linker-payload is through amphiphilic polymers such as polyethylene glycol (PEG), PSAR, or hydrophilic dendrimers. These amphiphilic polymers serve as interfaces between the linker-payload and the solvent to shield the hydrophobicity. This approach has been demonstrated to improve ADC PK, leading to better in vivo efficacy.
Studies have found that positioning these solubilizing elements on the side chains of the linker-payload, rather than linear spacers, maximizes the protective effect. Longer linear spacers, on the contrary, increase the exposure of the linker-payload to the solvent, adversely affecting the hydrophobicity of the ADC.
To maximize the effectiveness of these linker designs, optimization of the length of the amphiphilic polymer side chains is also required since short chains may not provide sufficient shielding, while longer chains may affect payload release. The optimal side chain length may vary depending on the nature of the polymer, payload, and release unit.
Since the P-glycoprotein (P-gp) efflux mechanism plays a crucial role in multidrug resistance (MDR) phenotypes, exhibiting broad specificity for hydrophobic substrates, cytotoxic payloads often exhibit hydrophobicity, making them prone to P-gp uptake.
Therefore, adding hydrophilic linkers to these cytotoxic payloads has been shown to be a viable strategy to reduce P-gp uptake and restore cytotoxic activity in MDR cell lines.
In one study, ImmunoGen demonstrated the benefits of hydrophilic non-cleavable linkers in MDR cancer cell lines. In this study, DM1 was conjugated to EPCAM antibodies with three different linkers, one prepared with a hydrophobic MCC linker and the other two with hydrophilic linkers of higher solubility, one containing a charged sulfonate group and the other containing a PEG4 unit.
At similar drug loads, comparable in vitro efficacy of these three ADCs was observed in Colo 205 colon cancer cell lines. However, in naturally high P-gp-expressing HCT-15 cell lines and engineered high P-gp transporter-expressing Colo 205 cell lines, the potency of ADCs with PEG4 and 3-sulfonate non-cleavable linkers was four times that of ADCs with MCC linkers. These results were also confirmed in the Colo 205 MDR mouse xenograft model.
In this model, PEG4 ADC inhibited tumor growth after a single injection of 10 mpk for 20 days. Meanwhile, MCC ADC only slowed tumor growth. These results were attributed to the lower affinity of the DM1 metabolites produced by the hydrophilic non-cleavable linker compared to Lys-MCC-DM1 substrates for P-gp efflux pumps.
MD1 related products:
DM1-MCC-PEG3-Biotin | CAS:2183472-94-6
Stability of ADC Linkers to Metabolic Degradation
ADC linker design can affect the stability of payloads to metabolic degradation, as demonstrated in a series of PNU-159682 anthracycline-based THIOMAB ADCs. The cytotoxin PNU-159682 features a glucosamine crucial for its activity, and its deglycosylation was used to explore the effect of ADCs generated through coupling on the stability of engineered cysteine residues.
It was found that spatial hindrance near the disulfide cleavage unit did not play a significant role in distinguishing deglycosylation; however, the linker length had a significantly greater impact on payload metabolism. Compared to ADCs with 6-8 Å linkers, the metabolism of payloads in ADCs with 16 Å linkers increased approximately threefold in crab-eating macaques and doubled in mouse whole blood.
The application of maytansinoid compounds as cytotoxic payloads in ADCs provides a particularly useful case study to describe the driving factors of degradation metabolism for MTD evaluation. Using sulfur-containing maytansinoids functionalized with cleavable thiol or non-cleavable linkers, a detailed comparison of the spatial shielding effects of methyls around disulfides and degradation metabolism driving factors of non-cleavable linkers was conducted.
It was found that the MCC-DM1 linker could resist reverse Michael reaction, which might be due to the higher pKa value of the thiol donor compared to cysteine residues on the antibody.
In contrast, SPP-DM1 and SPDB-DM4 linkers relied on intracellular thiols for the release of DM1 or DM4, followed by methylation through cellular methyltransferases. In the liver, this methylated maytansinoid is further metabolized into methyl sulfoxide and methyl sulfone through oxidation.
DM4 Linkers:
sulfo-SPDB-DM4 | CAS:1626359-59-8
The glycoprotein CanAg is a novel glycoform of mucin 1 (MUC1), highly glycosylated, rich in fucose and sialic acid, and expressed on the surface of human colon cancer, pancreatic cancer cells, and gastric cancer cells. The plasma stability of a series of CanAg-maytansinoid ADCs was evaluated, indicating a half-life of 2 days for the huC242-SPP-DM1 ADC without spatial hindrance and a half-life of 4.6 days for the huC242-SPDB-DM4 ADC with larger spatial hindrance.
Compared to the non-cleavable MCC-DM1 linker, the plasma stability of these two cleavable linker ADCs was reduced. In clinical trials, DLTs of unstable ADCs exhibited characteristics of free active payloads, whereas DLTs of stable linker ADCs may be attributed to the distribution and metabolic degradation of the antibody (ADC). However, the first 10 maytansinoid ADCs in clinical trials showed a negative correlation between tolerance and the chemical stability of the linker.
This was evidenced by the MTD of ADCs using the SPP-DM1 linker ranging from 6 to 8 mpk, the SPDB-DM4 linker ranging from 3.5 to 7 mpk, and the MTD of ADCs using the MCC-DM1 linker at 3.6 mpk. This result emphasizes that a more stable linker does not necessarily result in a higher MTD clinically, and the chemical nature of released metabolites is as important as the plasma stability of the linker.
ADC Linker Design Considerations for PK, Hydrophobicity, and Safety
The PK profile of ADCs is quite complex in terms of its absorption, distribution, metabolism, and excretion ( ADME ) properties. Several key features of ADCs, including linker stability, binding site, DAR, cytotoxin, and overall hydrophobicity, can influence its PK profile(Table 2).
Early work on developing anti-CD30 ADCs using the MC-Val-Cit-PABC-MMAE linker demonstrated that increasing the drug-to-antibody ratio (DAR) from 2 to 4 and 8 resulted in decreased plasma PK. This pioneering work has had a significant impact in the ADC field, setting a strong precedent that the optimal drug load for ADCs does not exceed 4. It wasn’t until the recent successes of DAR8 ADCs Enhertu and Trodelvy that this boundary was broken. These ADCs employed an improved drug-to-antibody ratio strategy, which focused on combining payloads with lower potency and higher DAR. Enhertu and Trodelvy were developed utilizing the maytansinoid DXd and SN-38, both of which are cytotoxic in the nanomolar range. Trodelvy achieved this by introducing a polyethylene glycol-type polymer, while Enhertu reduced overall ADC hydrophobicity by incorporating a less hydrophobic enzyme-cleavable unit, GGFG.
As mentioned earlier, reducing the hydrophobicity of ADCs can be achieved by introducing polar units into the linker, such as sulfonates, sulfonamides, lactose, and phosphates, which can also prevent intracellular cleavage in diphosphate solubilizing and release units. However, the development of quaternary ammonium salt linkers represents another advancement in reducing hydrophobicity. These linkers are tetra-substituted (quaternized) at the N-terminus of dimethylvaline, linked to PAB amino esters and MMAE, resulting in a DAR of 8 for the ADCs. Compared to the tertiary amine of MMAE in vedotin’s amino ester, this ADC exhibits improved PK properties.
Additionally, incorporating a glucuronic acid hydrolase-cleavable release unit into these quaternary ammonium salt linkers during their development can confer benefits and impart additional hydrophilicity to the ADCs. Further development of this glucuronic acid linker has demonstrated improvements in PK and therapeutic index (TI) for MMAE-based ADCs.
Seagen has undertaken pioneering work around linker optimization based on the glucuronic acid lipid release mechanism, which has been evaluated in combination with polyethylene glycol components of different lengths (n=2-24). These DAR8 MMAE ADCs were subsequently evaluated in PK and tolerability studies, revealing a significant correlation between polyethylene glycol content and tolerability; however, the PEG number reached a plateau from 8 to 24 units. This indicates that 8 units are sufficient to shield the hydrophobic MMAE, reducing non-specific cellular uptake, suggesting an optimal value for PEG length.
Hepatotoxicity is a key driver of ADC safety, and the relationship between ADC hydrophobicity and liver-mediated clearance is driven by increased hydrophobicity of high payload ADC species, which can be rapidly cleared by Kupffer cells in the liver. Therefore, appropriate linker design can also modulate safety risks associated with ADC hydrophobicity.
Amino esters and MMAE Linkers: