Antibody-Drug Conjugates: Following the Molecule From Bloodstream to Tumor Cell

An antibody-drug conjugate carries a potent toxin to a tumor using an antibody as the guide and a chemical linker as the release switch, so the cell-killing molecule stays tethered until it reaches the target. That much is a clean idea. What the mechanism cannot tell you is how much any particular conjugate actually helps a patient, because the approved agents rest on very different amounts of proof, ranging from an early approval that was later pulled to phase 3 trials with clear survival gains. The most useful way to hold both facts at once is to follow a single molecule from the bloodstream to the tumor cell, then step back and ask what the trials showed.

Key points#

• An ADC has three working parts: an antibody that finds a tumor marker, a linker that holds the toxin during transit, and a payload that does the killing.
• The linker, though small, decides where and whether the drug is released, which shapes both effectiveness and side effects.
• The drug-to-antibody ratio and the bystander effect are design levers, not guarantees of benefit.
• "It is an antibody-drug conjugate" describes a mechanism, not the strength of the evidence behind a specific drug.

Why not just give the chemotherapy?#

Ordinary chemotherapy travels everywhere the blood goes and damages fast-dividing healthy tissue, such as bone marrow and the gut lining, alongside the cancer. That collateral harm is what caps the dose. An antibody-drug conjugate tries to change the arithmetic: instead of flooding the body with a tolerable amount of a moderately toxic drug, it delivers a tiny amount of an extremely toxic drug to a narrower address. A 2023 review of approved conjugates in the journal Cancers is worth reading with a skeptical eye, because it treats each of the three parts as an engineering compromise rather than a solved problem. That framing is a healthy corrective to any marketing that offers you "targeted" as a synonym for "precise."

Following the molecule: three parts, three jobs#

The antibody: the guide, not the weapon#

The antibody accounts for most of the conjugate's mass and provides the navigation. It latches onto a protein that shows up more often on tumor cells than on healthy ones, for example HER2 on certain breast cancers, CD33 on myeloid leukemia cells, Nectin-4 on urothelial cancer, or Trop-2 across several solid tumors. Two things are easy to overlook. First, "more often" is not "only," so the target protein almost always appears on some normal cells as well, and that overlap is one root of side effects. Second, the antibody is essentially a courier here. It does not kill anything on its own, so choosing a good target matters more than how tightly the antibody happens to bind.

The linker: a seal that should only break in the right room#

The linker is the chemical bridge between the antibody and its cargo, and it carries far more responsibility than its size implies. Think of it as a tamper-proof seal that is supposed to open only after delivery. Cleavable linkers are designed to snap apart under conditions found inside or around tumor cells, such as an acidic pocket, particular enzymes, or the chemically reducing interior of the cell. Non-cleavable linkers hold together until the entire antibody is broken down inside the cell's lysosome, releasing the payload only then. Non-cleavable designs tend to survive longer in the bloodstream, which limits release in the wrong place, while cleavable designs permit a behavior that turns out to matter in the clinic: the bystander effect, described further below.

The payload: potency counted in single molecules#

The payload is a cytotoxic drug so poisonous that it could never be given by itself. The Cancers review sorts the common payloads into two broad families: microtubule inhibitors such as the auristatins (MMAE) and maytansinoids (DM1, DM4), and DNA-damaging agents such as calicheamicins and topoisomerase I inhibitors including the deruxtecan and SN-38 payloads. Because so little payload reaches the tumor, the toxin has to be brutally effective in small quantities, which is exactly why it would be unsafe as a free-floating drug.

Two design levers that get oversold#

Drug-to-antibody ratio#

The drug-to-antibody ratio (DAR) is the average number of payload molecules carried by each antibody. It is tempting to assume more is better, but the review describes a sweet spot of roughly two to four. Load too little and the conjugate is underpowered. Load too much and the antibody can turn unstable, get cleared from the body faster, and lose the very selectivity that made the whole approach attractive in the first place.

The bystander effect#

Tumors are patchy. Within one mass, some cells carry a great deal of the target protein and their neighbors carry little or none. A membrane-permeable payload freed by a cleavable linker can slip out of the cell it first entered and kill nearby cells that the antibody could never have reached. That is the bystander effect, and it helps explain why some conjugates work against tumors with uneven target expression. It cuts both ways, though: a drug that can drift to neighboring tumor cells can also drift to neighboring healthy ones, which is part of why bystander-capable agents carry distinctive toxicities. It is a real mechanism, not a free safety margin.

Mechanism is not the same as proof#

At this point the story leaves chemistry and enters the trial record, and the approved agents are not on equal footing.

Gemtuzumab ozogamicin, a CD33-directed conjugate carrying a calicheamicin payload, was the first ADC approved, back in 2000, then voluntarily withdrawn in 2010 after a confirmatory trial failed to show benefit and raised safety concerns. It came back in 2017 on a fractionated, lower-dose schedule that the FDA approval says improved the benefit-risk balance in CD33-positive acute myeloid leukemia. The whole arc is a reminder that a sensible mechanism can still fail its first honest test, and that changing how a drug is dosed can change the verdict.

Trastuzumab deruxtecan sits on firmer ground. It joins a HER2 antibody to a topoisomerase I inhibitor with a high DAR and a strong bystander effect. In the randomized phase 3 DESTINY-Breast04 trial it improved both progression-free and overall survival against physician's-choice chemotherapy in previously treated HER2-low metastatic breast cancer; long-term follow-up in Nature Medicine reported median overall survival near 22.9 months versus 16.8 months for chemotherapy in the HER2-low group. Enfortumab vedotin, a Nectin-4 conjugate with an MMAE payload, cleared a similarly high bar in the phase 3 EV-301 trial, with median overall survival around 12.9 months versus about 9.0 months on chemotherapy after prior platinum and immunotherapy.

The lesson is that "it is an antibody-drug conjugate" names a mechanism, not a level of proof, and you have to look up the second thing yourself. Some approved agents stand on randomized overall-survival gains against an active comparator; others reached the market on earlier signals and single-arm response rates that a later trial may or may not confirm. When you read about a new conjugate, the questions worth keeping in mind are which trial supported it, what it was compared against, whether overall survival or only a surrogate endpoint improved, and in which specific population.