The three ADC components

An ADC is only as good as the weakest of its three parts, and bioconjugation is where the three meet. The antibody provides target selectivity and half-life, the linker keeps the payload attached in circulation and releases it (or not) inside the cell, and the payload kills the cell. The conjugation chemistry constrains all three.

Reference products illustrate the combinations. Kadcyla (ado-trastuzumab emtansine) uses lysine conjugation of a non-cleavable SMCC linker to DM1; Enhertu (trastuzumab deruxtecan) uses cysteine conjugation of a cleavable tetrapeptide linker to DXd; Adcetris (brentuximab vedotin) uses cysteine conjugation of a Val-Cit linker to MMAE. The same bioconjugation chemistries serve non-cytotoxic conjugates: PEGylated proteins, antibody-oligonucleotide conjugates, radioconjugates and immune-stimulating antibody conjugates.

Stochastic lysine and cysteine conjugation and DAR distribution

Stochastic bioconjugation is simple chemistry on a complex substrate, and the result is a mixture whose average drug-to-antibody ratio hides a wide distribution.

Lysine conjugation. An IgG carries roughly 80 to 90 lysines, several dozen of them solvent-accessible. An NHS-ester linker acylates lysine amines at random, in two steps or with a pre-formed linker-payload. The product spans DAR 0 to about 8 with a vast number of positional isomers, so the distribution can be described but not resolved. Average DAR is set by the molar excess of linker.

Cysteine conjugation. IgG1 has four interchain disulfides. Partial reduction with TCEP or DTT frees two, four, six or eight thiols, and maleimide-bearing linker-payloads add to them. Because thiols are released in pairs, only even-numbered species form (DAR 0, 2, 4, 6, 8), and the reduction stoichiometry sets the average. Targeting DAR 4 gives a mixture centered on 4; full reduction gives a near-homogeneous DAR 8, the route used for Enhertu. Maleimide-thiol adducts can exchange with albumin in plasma, so stabilized maleimides are now common.

High-DAR species clear faster and aggregate more readily, while DAR 0 contributes no activity and competes for target. Most stochastic processes accept the distribution because preparative HIC at kilogram scale is costly.

Site-specific approaches

Site-specific conjugation moves the heterogeneity problem from purification and analytics into antibody engineering. Each approach creates a unique reactive handle at a defined position, so the product has one DAR and one site per handle. The price is a modified antibody, often an enzyme as a raw material, and licensed technology.

Homogeneous conjugates have shown better pharmacokinetics and wider therapeutic windows in many preclinical comparisons, and one species is easier to characterize than a distribution. Most approved ADCs nevertheless use stochastic chemistry, which shows that heterogeneity is manageable when the linker-payload is well chosen. Site-specific conjugation is a tool for a specific problem, usually a narrow therapeutic index or a hydrophobic payload, not a default.

Measuring DAR: HIC, RP-HPLC and mass spectrometry

Drug-to-antibody ratio is a critical quality attribute, and no single method reports both average and distribution for every bioconjugation chemistry, so a DAR control strategy combines two or three orthogonal methods and validates the one used for release.

DAR is never controlled alone. Unconjugated antibody, free payload and linker-related impurities by RP-HPLC or LC/MS, aggregate by SEC and residual solvents complete the conjugate-specific panel. Under ICH Q5E, a shift in DAR distribution with an unchanged average is still a change in quality.

Manufacturing: HPAPI handling, isolators and scale

ADC bioconjugation manufacturing is solution chemistry at protein scale, performed inside containment designed for compounds hazardous at nanogram exposures. Payloads such as auristatins and maytansinoids have occupational exposure limits in the nanogram-per-cubic-meter range, so the facility, not the chemistry, usually determines which ADC CDMO can make a given product.

A typical campaign follows a fixed sequence: UF/DF of the antibody into conjugation buffer; reduction if the chemistry requires it; addition of the linker-payload in an organic co-solvent under controlled temperature and time; quenching; tangential flow filtration to remove free payload, solvent and reagents; formulation by UF/DF, sterile filtration and bulk fill. Containment comes from isolators or closed systems, pressure cascades, dedicated or single-use product-contact equipment and validated deactivation of payload residues.

Scale is modest by antibody standards: because ADC doses are low, development batches run at tens of grams and commercial batches at hundreds of grams to a few kilograms.

Integrated versus multi-vendor ADC supply chains

An ADC supply chain has at least four nodes, antibody, linker-payload, conjugation and fill, and every hand-off adds shipping, duplicate testing and comparability risk. Choosing between a multi-vendor chain and an integrated ADC CDMO is therefore a risk decision as much as a cost decision.

Whatever the model, compatibility work should start before the antibody process is locked. Free thiol content, lysine accessibility, glycan profile and buffer determine how an antibody behaves in bioconjugation, and linker-payload solubility sets how much can be added before the protein aggregates. An ADC CDMO that produces the antibody as well as the conjugate can run these studies on development material rather than GMP intermediate.