The Impact of Conjugation Mode and Site on Tubulysin Antibody-Drug-Conjugate Efficacy and Stability

Abstract

Antibody-drug conjugates (ADCs) represent a prominent class of biotherapeutics engineered to selectively deliver cytotoxic payloads to tumors, thereby facilitating targeted cell killing. While first-generation ADCs, created by conjugating payloads to surface-accessible lysine or hinge-cysteine residues, have achieved clinical success, several site-specific ADCs with defined drug-to-antibody ratios are currently under clinical investigation. Herein, the efficacy, stability, and pharmacokinetics of ADCs generated by attaching the drug linker to surface-exposed lysine residues, hinge-cysteine residues, and the C'E loop in the CH2 domain (mediated by bacterial transglutaminase) using a tubulysin payload are compared. In N87 xenograft mice, the order of efficacy is C'E loop > hinge-cysteine > lysine-conjugated ADCs. Among the three ADCs evaluated, the site-specific ADC demonstrates superior in vivo stability (minimal payload-linker deconjugation and limited payload metabolism/deacetylation) and favorable pharmacokinetics (longer half-life, low clearance, high exposure). In contrast, the lysine-conjugated ADC exhibits the least stability and poorest pharmacokinetics, which directly correlate with its efficacy. Further investigation into cysteine-engineered site-specific ADCs with payloads conjugated at various sites confirms that both the conjugation chemistry and the site of conjugation significantly influence the in vivo stability and pharmacokinetics of site-specific ADCs.

Keywords: antibody‐drug conjugates; deconjugation; payload metabolism; site‐specific conjugations; tubulysin.

Figure 1

Different conjugation strategies: a,b) 2‐IT and subsequent maleimide chemistry‐based conjugation to introduce tubulysin payload into surface‐exposed lysine residues. Tubulysin payload‐linker (1) or deacetylated tubulysin payload‐linker (2) was attached to aMeso to produce aMeso (lysine) ADC with intact tubulysin (a) and aMeso (lysine) ADC with deacetylated tubulysin (b), respectively. c) aMeso (hinge‐cysteine) ADC was made by reduction of hinge disulfides followed by payload conjugation to free thiols. d) bTGase‐mediated conjugation of tubulysin payload‐linker (3) to Q295 and Q297 residues on N297Q mutated mAb yielded aMeso (N297Q) ADC.

Figure 2

In vivo efficacy of anti‐Meso ADCs made with two different modes of bioconjugation. When tumors reached 100 mm³, mice were treated with a single injection of either 3.5 mg/kg of anti‐Meso or isotype control ADCs. a) ADCs demonstrated a targeting arm‐dependent activity to inhibit tumor growth. b) At low dose level of 1 mg/kg, two ADCs showed a distinct in vivo efficacy.

Figure 3

Comparison of in vivo efficacy of the lysine, hinge‐cysteine, and N297Q ADCs targeting Meso‐dosed i.v. at 0.25 and 0.5 mpk in N87 xenograft model mice.

Figure 4

a) Representative scheme of possible ADC metabolites in vivo, b) concentrations of total Ab (antibody with or without conjugated payload), total ADC (antibody with conjugated active and inactive payload), active ADC (antibody with conjugated active payload) in mice dosed with 0.5 mg/kg of aMeso (lysine) ADC, c) aMeso (hinge cysteine) ADC, and d) aMeso (N297Q) ADC.

Figure 5

Stability of cysteine‐engineered site‐specific ADCs. a) % deconjugation and % succinimide hydrolysis (left) observed for ADCs of interest after incubation in 500 uM glutathione for 5 days or pH9 buffer for 2 days. % deacetylation and % deconjugation observed in vivo (right). b) Position of engineered cysteines.

Figure 6

In vitro and in vivo characterization of PBD‐ADCs constructed with random or site‐specific conjugation chemistry. a) In vitro cell killing activity of PBD ADCs against the N87 cell line expressing the Meso target. Concentration was normalized based on payload molarity. The ADCs showed targeting arm‐dependent activity. b) In vivo efficacy of ADCs dosed via IP at drug equimolar levels (1 mg/kg for lysine, 3 mg/kg for N297A ADC). Site‐specific N297A ADC showed greater activity. c) Body weight loss data after drug administration. No substantial difference between the two ADCs was observed. d) Concentrations of total Ab and total ADC in mice dosed with 0.02 μmol/kg of aMeso (lysine) ADC and aMeso (N297Q) ADC.