T DXd for HER2-Positive Cancer: Mechanism and Pharmacokinetics

Trastuzumab deruxtecan, marketed as Enhertu and commonly abbreviated T-DXd, is an antibody-drug conjugate that delivers a potent cell-killing chemical directly to tumors expressing the HER2 protein. What sets it apart from older HER2-targeted therapies is a combination of engineering choices: a high drug-to-antibody ratio of roughly eight payload molecules per antibody, a linker designed to release the drug inside tumor cells, and a payload that can slip through cell membranes and damage neighboring cancer cells that don’t even express HER2. Those design features shape both how the drug behaves in the body and why it has produced striking results in clinical trials across breast cancer, gastric cancer, and other solid tumors.

How T-DXd Is Built

An antibody-drug conjugate is, conceptually, a guided missile: a targeting antibody ferries a toxic payload to tumor cells. T-DXd uses trastuzumab, the same anti-HER2 antibody found in Herceptin, as its homing component. Attached to that antibody via a cleavable tetrapeptide linker are molecules of DXd, a derivative of the chemotherapy drug exatecan and a potent inhibitor of the enzyme topoisomerase I.

One of the more consequential design decisions is the drug-to-antibody ratio, or DAR. Older conjugates like trastuzumab emtansine (T-DM1) carry an average of about 3.5 drug molecules per antibody. T-DXd carries roughly eight, meaning each antibody that reaches a tumor cell delivers more than double the cytotoxic cargo.¹ That higher loading is possible because the conjugation chemistry attaches DXd to interchain cysteine residues in a relatively homogeneous fashion, confirmed by mass spectrometry analyses identifying conjugation at specific positions on both the heavy and light chains of the antibody.²

The linker connecting DXd to trastuzumab is a glycine-glycine-phenylalanine-glycine tetrapeptide designed for cleavage by cathepsin proteases inside tumor cells.¹ That cleavability is intentional: once the conjugate is internalized and reaches the acidic environment of a cell’s endosomes and lysosomes, cathepsin enzymes cut the linker and free the DXd payload.³ Despite being cleavable inside cells, the linker is stable in the bloodstream. In mouse models carrying HER2-positive tumors, the pharmacokinetic profiles of intact T-DXd and total antibody (conjugated plus unconjugated) tracked almost identically, confirming that DXd doesn’t fall off the antibody prematurely during circulation.⁴

How DXd Kills Cancer Cells

Once freed from the antibody inside a tumor cell, DXd goes to work on topoisomerase I, an enzyme that manages the twisting and untwisting of DNA during replication. Normally, topoisomerase I creates a temporary single-strand nick in the DNA helix, allows it to unwind, and then reseals the break. DXd locks the enzyme onto the DNA at the cleavage site, preventing that resealing step. When the cell’s replication machinery collides with the trapped complex, single-strand nicks are converted into double-strand breaks, a far more lethal form of DNA damage.⁵

Laboratory work has confirmed this chain of events by measuring molecular markers of DNA damage after T-DXd treatment. In HER2-positive breast and gastric cancer cell lines, T-DXd treatment increased levels of γH2AX, a well-established marker of double-strand breaks, as well as other DNA damage-response proteins. Comet assays, which visualize DNA fragmentation in individual cells, showed significantly increased tail moments in treated cells, directly confirming that DNA was being shattered.⁶ The accumulation of irreparable double-strand breaks ultimately triggers programmed cell death.

The Bystander Effect

Tumors are not uniform. Even within a single tumor mass, some cells display high levels of HER2 on their surface while others display low levels or none at all. An antibody-drug conjugate that can only kill the cells it binds to directly will miss those HER2-negative neighbors, leaving behind a resistant population that can regrow the tumor. T-DXd was designed to address this problem through what researchers call the bystander effect.

The key property enabling bystander killing is the membrane permeability of free DXd. Once released inside a HER2-positive cell, DXd can diffuse out through the cell membrane and enter adjacent cells regardless of their HER2 status. Researchers demonstrated this in a straightforward experiment: when HER2-positive cells were cultured alongside HER2-negative cells and treated with T-DXd, both populations died. The same experiment with T-DM1 killed only the HER2-positive cells, because T-DM1’s payload, a maytansine derivative called DM1, has low membrane permeability and stays trapped in the cell that internalized it.⁷

Imaging studies using phosphor-integrated dot technology have visualized this process in living tumors. After T-DXd treatment, the drug was initially concentrated in HER2-rich areas of the tumor, consistent with antibody-guided delivery. Over time, the released DXd spread into adjacent HER2-negative zones.⁸ In co-culture studies with labeled cell populations, bystander apoptosis in HER2-negative cells reached about 23% by day six with T-DXd, while T-DM1-treated cultures showed no difference from untreated controls.¹

How T-DXd Compares with T-DM1

Because T-DM1 (Kadcyla) was the standard HER2-targeted ADC before T-DXd arrived, the comparison between them is instructive. The two drugs differ in three fundamental ways: payload type, linker design, and drug loading.

Payload: T-DXd carries DXd, a topoisomerase I inhibitor. T-DM1 carries DM1, a microtubule inhibitor that prevents cell division by blocking the assembly of the internal scaffolding cells need to pull chromosomes apart.
Linker: T-DXd uses a cleavable peptide linker cut by cathepsin enzymes in tumor cell endosomes. T-DM1 uses a non-cleavable thioether linker, meaning its payload is only freed when the entire antibody is broken down inside the lysosome.
Drug loading: T-DXd carries a DAR of roughly eight. T-DM1 carries about 3.5.

These differences compound. The cleavable linker and high membrane permeability of DXd together produce a bystander effect that T-DM1 largely lacks. The higher DAR means more payload delivered per binding event. And the topoisomerase I mechanism of DXd acts through a different pathway than T-DM1’s microtubule inhibition, which matters for patients whose tumors have already progressed on T-DM1.¹ In the phase 3 DESTINY-Breast03 trial, T-DXd demonstrated a median progression-free survival of 28.8 months compared with 6.8 months for T-DM1 in HER2-positive metastatic breast cancer.⁹

However, the more active bystander effect also contributes to a different safety profile. Pharmacovigilance data show that T-DXd carries a significantly higher proportion of respiratory adverse events, including interstitial lung disease, and more hematologic fatal outcomes compared with T-DM1.¹⁰

How the Body Handles T-DXd

Systemically, T-DXd behaves the way you’d expect for a large antibody-based molecule. Population pharmacokinetic modeling describes the intact conjugate using a two-compartment model with linear elimination, meaning the drug distributes between blood and tissues and is cleared at a rate proportional to its concentration.¹¹ The released DXd payload, by contrast, follows a simpler one-compartment model. In preclinical studies, the linker’s stability meant that circulating DXd levels remained low, with most of the drug staying attached to the antibody until it reached target tissue. After intravenous dosing in cynomolgus monkeys, intact T-DXd stayed predominantly in the blood without accumulating in any specific non-tumor tissue.¹²

Once DXd is released, the body eliminates it primarily through bile. Physiologically based pharmacokinetic modeling estimates that biliary clearance accounts for about 62% of total DXd elimination, with renal excretion handling roughly 33%. Metabolism by the liver enzyme CYP3A4 contributes less than 5%, which is pharmacologically significant because it means drugs that inhibit or induce CYP3A4 are unlikely to dramatically alter DXd levels.¹³

Dosing, Patient Factors, and Tumor Type

T-DXd is dosed by body weight, typically at 5.4 mg/kg every three weeks for breast cancer. Researchers have tested whether individual patient characteristics meaningfully change drug exposure. In population pharmacokinetic analyses pooling data from multiple trials, covariates like country, tumor size, sex, age, and liver function markers were all statistically significant but clinically modest: most shifted steady-state drug exposure by less than 20%. The exceptions were patients with very high body weight (around the 95th percentile, about 86 kg) and those with low albumin levels (around the 5th percentile, 31 g/L), where changes in exposure were somewhat larger.¹¹

Analyses stratified by renal function, hepatic function, and race found no clinically meaningful differences in exposure, and current evidence does not support dose adjustment for these subgroups.¹¹ Tumor type, however, does matter. In gastric cancer patients, T-DXd clearance is faster than in breast cancer patients. At the same 6.4 mg/kg dose, steady-state exposures in gastric cancer were lower than in breast cancer, more closely resembling what breast cancer patients see at 5.4 mg/kg.¹⁴ This is why the approved dose for HER2-positive gastric cancer is 6.4 mg/kg rather than 5.4 mg/kg.

Clinical Results in HER2-Low Breast Cancer

Perhaps the most unexpected aspect of T-DXd’s clinical story is its activity in HER2-low breast cancer, a category of tumors with minimal HER2 expression that had never been considered candidates for HER2-targeted therapy. The DESTINY-Breast04 trial enrolled patients with metastatic breast cancer classified as HER2-low and compared T-DXd against standard chemotherapy. Among all enrolled patients, T-DXd produced a median progression-free survival of about 10 months versus 5 months for chemotherapy, and median overall survival of roughly 23 months versus about 17 months.¹⁵ Extended follow-up at a median of 32 months confirmed the durability of this benefit, with overall survival of 22.9 months for T-DXd versus 16.8 months for chemotherapy in the overall population.¹⁶

A systematic review and meta-analysis covering both HER2-positive and HER2-low populations confirmed that T-DXd significantly outperformed comparators in progression-free survival, overall survival, and objective response rates across both groups.¹⁷ The bystander effect described earlier is thought to be a major reason T-DXd works even in tumors with very low HER2 expression: the small number of HER2-positive cells serve as entry points for the drug, and the membrane-permeable DXd payload spreads to kill surrounding HER2-negative cells.

Safety Concerns Worth Knowing

The most closely watched toxicity with T-DXd is interstitial lung disease, or ILD, a potentially serious inflammatory condition of the lungs. In pharmacovigilance data, the safety signal for ILD and pneumonitis with T-DXd is among the strongest of any reported adverse event. Researchers have proposed four possible mechanisms: direct uptake of T-DXd by lung cells that express some HER2, nonspecific uptake by normal lung tissue, bystander killing from DXd released by nearby tumor cells, and circulating free DXd from deconjugation in the bloodstream.¹⁸ Exposure-safety analyses have found a statistically significant relationship between higher T-DXd exposure and increasing risk of ILD across dose levels ranging from 0.8 to 8 mg/kg.⁹

Beyond lung toxicity, the most frequent side effects resemble standard chemotherapy reactions. A meta-analysis and pharmacovigilance study identified gastrointestinal disorders and blood-related problems as the most common categories of adverse events. Neutropenia, a dangerous drop in infection-fighting white blood cells, was the most frequent severe event, occurring at grade 3 or higher in about 21% of patients.¹⁹ In the DESTINY-Breast03 trial specifically, the median time to first neutropenia event was about 64 days, with anemia appearing at a median of 70 days and low platelet counts at around 132 days. Roughly a third of patients developed anemia of any severity, typically in the early cycles. These hematologic effects tended to decrease over time.²⁰

How Tumors Develop Resistance

As with any targeted therapy, tumors eventually find ways around T-DXd. Early research into resistance mechanisms has identified two broad categories. The first involves changes to the drug’s target. Mutations in the ERBB2 gene (which encodes HER2) can reduce or alter the receptor on the cell surface, making it harder for trastuzumab to bind and deliver its payload.

The second category involves the cancer cell learning to tolerate DXd itself. Researchers have found that loss of the protein SLFN11, which normally sensitizes cells to DNA-damaging agents, can make tumor cells resistant to T-DXd’s payload. Copy number gains in ABCC1, a gene encoding a drug efflux pump that physically ejects small molecules from cells, have also been observed in resistant tumors.²¹ Genomic profiling of tumors that progressed on T-DXd has found enrichment of mutations in NFE2L2, a gene that activates production of the same ABCC1 pump, and KEAP1, which normally keeps NFE2L2 in check. When KEAP1 is lost, the NFE2L2 pathway runs unchecked and drug efflux ramps up.²²

Understanding these resistance pathways matters practically. If a tumor loses HER2 expression, switching to a different HER2-targeted drug won’t help, but a treatment targeting a different pathway might. If the tumor retains HER2 but has acquired efflux-pump overexpression, future strategies could include combining T-DXd with efflux-pump inhibitors or switching to a different payload class entirely.

Combining T-DXd with Immunotherapy

One emerging research direction is pairing T-DXd with immune checkpoint inhibitors. The rationale is straightforward: by killing cancer cells and scattering their contents into the surrounding tissue, T-DXd may help alert the immune system to the tumor’s presence. Preclinical models have tested whether adding dual checkpoint blockade, targeting both the PD-1/PD-L1 pathway and a second immune checkpoint like CTLA-4 or TIGIT, could enhance this effect. The results have provided scientific support for moving such combinations into clinical testing.²³ Multiple clinical trials are now exploring T-DXd combined with various immunotherapy agents, though mature efficacy and safety data from these combinations are still forthcoming.

The broader platform of DXd-based conjugates extends beyond HER2. The same linker-payload technology has been adapted to target other surface proteins, including TROP2 (in datopotamab deruxtecan, approved for non-small cell lung cancer) and HER3 (in patritumab deruxtecan, still in development).²⁴ The shared DXd backbone means that lessons learned about T-DXd’s pharmacokinetics, bystander activity, resistance patterns, and safety signals are likely to apply, at least in part, to these newer agents as well.