What Is Claudin 6 and Its Role in Cancer?

Claudin-6 is a protein that helps seal the junctions between cells during embryonic development and then virtually disappears from normal adult tissue. When it resurfaces on tumor cells, it becomes both a driver of cancer progression and a bullseye for some of the most promising targeted therapies now in clinical testing. That combination of absence from healthy tissue and presence on tumors makes claudin-6 unusually attractive as a drug target, and a wave of antibody-drug conjugates, CAR-T cells, and bispecific antibodies are already being built around it.

A Fetal Protein That Should Not Be There

Claudin-6 belongs to the claudin family, a group of proteins that form the backbone of tight junctions, the structures that glue neighboring cells together and control what passes between them. Tight junctions maintain cell polarity, regulate permeability, and participate in signaling that governs cell growth and differentiation.¹ Claudin-6 is one of the earliest molecules turned on when embryonic stem cells commit to becoming epithelial tissue, and under normal circumstances its expression is almost entirely restricted to embryonic and fetal life.² By the time a person is born, healthy tissues have essentially stopped making it.

That developmental shutdown is what makes claudin-6 so interesting for cancer biology. Many tumors reactivate genes that were only meant to run during fetal development, and claudin-6 is one of the clearest examples. Its expression in tumor cells with little or no detectable presence in normal adult tissues creates a natural therapeutic window: a drug that homes in on claudin-6 could, in theory, hit cancer while leaving the rest of the body alone.³

How Claudin-6 Gets Switched Back On

In many cancers, the claudin-6 gene is kept silent by chemical tags called methyl groups that sit on its promoter region, essentially locking the gene shut. When those methyl marks are removed or disrupted, the gene turns back on and claudin-6 floods the cell surface. Research in breast cancer cells has shown that this methylation recruits a specific protein that strips acetyl groups from nearby proteins involved in packaging DNA, tightening the chromatin structure and keeping the gene inaccessible. When that silencing cascade fails, claudin-6 re-emerges, and the cell takes on a more invasive character.⁴

The flip side of that same coin has been demonstrated experimentally. Treating breast cancer cells with drugs that strip away methyl marks and inhibit the enzymes that remove acetyl groups can push claudin-6 expression back up. In one study, this combination was enough to sensitize cancer cells to programmed cell death and wipe out their ability to form colonies.⁵ That finding cuts both ways: claudin-6 seems to act as a tumor suppressor in some settings, and yet in other settings its presence on the tumor surface drives aggressive behavior. The context matters enormously, and this duality is part of what makes claudin-6 biology complicated.

Which Cancers Express Claudin-6

Germ cell tumors are the most reliable expressors. A study examining a broad panel of tumors found claudin-6 in every single germ cell tumor tested, including seminomas, embryonal carcinomas, yolk sac tumors, and the trophoblastic cells of choriocarcinomas. Among more than 800 non-germ-cell tumors, only about 8% were positive.⁶ That near-universal expression in germ cell tumors, combined with near-total absence in other cancers, led researchers to propose claudin-6 as a diagnostic marker for primitive germ cell malignancies.

Within testicular germ cell tumors specifically, more recent work has added nuance. Claudin-6 was found in all precancerous lesions and all primary non-teratoma components, but expression dropped sharply in teratomas and dropped even further in post-chemotherapy teratoma tissue. The average fraction of positive cells went from roughly 88% in primary non-teratoma components to under 5% in post-chemotherapy teratoma.⁷ That has practical implications: patients whose tumors evolve into teratoma after chemotherapy may lose the target that claudin-6 therapies depend on.

Outside germ cell tumors, endometrial cancers of the aggressive non-endometrioid type also show substantial expression. Among serous endometrial tumors, about half showed high claudin-6 levels, and roughly a fifth of carcinosarcomas were positive. Clear cell tumors were rarely positive, and undifferentiated tumors showed no high expression at all.⁸ Ovarian cancer is another focus, though expression there is variable. In one small study, five of seven patients were claudin-6-positive, but the amount of protein varied widely from patient to patient, and even within a single tumor there were distinct patches of positive and negative cells.⁹ That patchiness, called intratumoral heterogeneity, is a challenge for any therapy that relies on the target being uniformly present.

Claudin-6 as a Driver of Invasion and Drug Resistance

When claudin-6 shows up on tumor cells, it does more than just sit on the surface. Research has linked it to cell invasiveness and the breakdown of surrounding tissue, mainly through signaling pathways that activate enzymes capable of chewing through the matrix around a tumor.¹⁰ That process, epithelial-to-mesenchymal transition, is a key step in how a cancer cell goes from a stationary mass to a migratory, metastasis-prone threat.

Claudin-6 also appears to blunt the effectiveness of chemotherapy. In triple-negative breast cancer cells, introducing claudin-6 raised the dose of adriamycin needed to kill them and boosted the cells’ ability to survive and form colonies despite treatment. The mechanism involves activation of a signaling cascade that enhances stem-cell-like properties in the tumor cells, making them inherently harder to kill.¹¹ Separately, claudin-6 has been shown to trigger a self-protective form of autophagy, essentially a recycling program that cancer cells use to survive under the stress of chemotherapy drugs like adriamycin and paclitaxel.¹² If claudin-6 helps tumors resist standard treatment, that strengthens the case for targeting it directly.

The Immune Landscape Around Claudin-6-Positive Tumors

Claudin-6 does not exist in an immune vacuum. Analysis in bladder cancer found that its expression level tracks closely with the types of immune cells that infiltrate the tumor. Higher claudin-6 expression correlated with more macrophages, natural killer cells, and certain T-cell subtypes entering the tumor. But the relationship is not straightforwardly good news: claudin-6 was positively associated with M2 macrophage polarization, a state in which macrophages tend to suppress rather than activate the immune response. That association may mean that claudin-6-high tumors create an immune environment that shields them from attack, potentially making them less responsive to conventional immunotherapy like checkpoint inhibitors.¹³ Understanding the immune microenvironment is relevant for predicting which patients might benefit from claudin-6-targeted treatments versus standard immunotherapy.

Antibody-Drug Conjugates Aimed at Claudin-6

Antibody-drug conjugates, or ADCs, are essentially guided missiles: an antibody that recognizes the tumor target is chemically linked to a potent toxin, so when the antibody lands on a claudin-6-positive cell, the toxin gets dragged inside and kills it from within. Several claudin-6 ADCs are in development. The most extensively published preclinical candidate, CLDN6-23-ADC, uses a toxin called MMAE. It selectively binds claudin-6 without cross-reacting with other family members, gets rapidly internalized by cells that carry the target, and produced lasting tumor regressions in mouse models of bladder, endometrial, and ovarian cancer. Tumors expressing claudin-6 shrank and stayed small for months, while tumors lacking the protein were unaffected.¹⁴

A separate ADC approach has been tested for liver cancer. A different antibody conjugated to a toxin called DM1 showed strong antitumor activity in both liver cancer cell lines and primary tumors, and worked even better in combination with sorafenib, an existing standard drug for that disease.¹⁵ These remain preclinical results, not human trials, but the consistency across multiple tumor types and drug payloads has generated considerable momentum.

CAR-T Cells and the RNA Vaccine Twist

Chimeric antigen receptor T cells, or CAR-T cells, are immune cells engineered to recognize a specific target on tumor cells. Claudin-6 CAR-T cells have entered human testing in one of the more innovative trial designs in recent oncology. The BNT211-01 trial, run by BioNTech, tested claudin-6-directed CAR-T cells with and without a novel RNA vaccine designed to boost the engineered cells once they are inside the body. In preclinical work, the RNA vaccine concept had already shown that delivering the CAR target antigen to immune compartments throughout the body could stimulate CAR-T cells to expand even at doses that would otherwise be too low to work.¹⁶

In the phase 1 trial itself, 22 patients with relapsed or treatment-resistant claudin-6-positive solid tumors received the CAR-T cells. Roughly half experienced cytokine release syndrome, a common side effect of CAR-T therapy, though only one case was severe. The treatment produced an overall response rate of about a third, including one complete response. The disease control rate reached two-thirds, meaning the majority of patients saw at least stabilization. Patients with germ cell tumors treated at the higher dose level responded most frequently, with a response rate of roughly 57%.¹⁷ Those numbers are striking for solid tumors, where CAR-T cells have historically struggled far more than in blood cancers.

The RNA vaccine component appeared to help. Patients who received the vaccine alongside CAR-T cells at the lower dose level showed a trend toward greater CAR-T cell expansion than those who received CAR-T cells alone.¹⁷ If confirmed in larger trials, this booster-vaccine strategy could address one of the biggest weaknesses of CAR-T therapy in solid tumors: the engineered cells tend to peter out before the job is done.

Beyond adult cancers, claudin-6 CAR-T cells have also been tested in preclinical models of pediatric brain tumors. In mice carrying orthotopic tumors mimicking atypical teratoid/rhabdoid tumors, the CAR-T cells demonstrated potent and specific antitumor activity, suggesting the approach could eventually extend to some of the most difficult childhood cancers.¹⁸

Bispecific Antibodies and T-Cell Engagers

A third class of claudin-6 therapies sidesteps the complexity of engineering a patient’s own cells. Bispecific antibodies are synthetic molecules with two arms: one grabs claudin-6 on the tumor, and the other grabs a receptor on nearby T cells, physically dragging immune cells into close contact with cancer cells and triggering killing. Several of these are racing through development. AMG 794, from Amgen, redirected T cells to kill claudin-6-positive lung and ovarian cancer cells at extremely low concentrations and shrank established tumors in mice.¹⁹ XmAb541 uses a 2+1 format designed for avid tumor binding and is entering a first-in-human dose-escalation trial in patients with germ cell tumors and other advanced solid tumors.²⁰ A third candidate, ARC101, was specifically engineered to avoid any cross-reactivity with claudin-9 or other human membrane proteins.²¹

Why Specificity Is So Hard

The biggest engineering challenge with claudin-6 therapies is not getting them to work but getting them to be specific. Claudin-6 is nearly identical to claudin-9, a family member found in normal tissues including the inner ear and certain epithelial surfaces. The two proteins differ by only three amino acids in their extracellular domains, the portion of the protein that antibodies must latch onto.²¹ A drug that cannot tell them apart could attack healthy tissues.

Structural studies have pinpointed the molecular basis for distinguishing the two. One key investigation mapped the antibody-binding surface of claudin-6 at atomic resolution and found that specificity comes down to a single amino acid position: residue 156. In claudin-6, that position holds a glutamine; in claudin-9, it is a leucine. The difference in side-chain bulk at the gamma carbon is enough to let a carefully designed antibody fit snugly on claudin-6 but clash sterically with claudin-9. In experiments where that residue was mutated to dozens of alternatives, any substitution with a bulky branch at the beta or gamma carbon destroyed antibody binding, confirming that the specificity depends on the precise size and shape at that single point.²² The preclinical bispecific candidate CTIM-76, for instance, demonstrated at least a 500-fold selectivity for claudin-6 over claudin-9 in cell-based assays.²³ That kind of razor-thin molecular discrimination is what makes or breaks a safe drug.

Claudin-6 as a Diagnostic Marker

Given that claudin-6 lights up so reliably in germ cell tumors, pathologists have naturally asked whether it could help diagnose difficult cases. For germ cell tumors arising outside the brain, the evidence is strong: the near-100% positivity rate across seminomas, embryonal carcinomas, and yolk sac tumors makes it a useful addition to the diagnostic panel.⁶

But claudin-6 has not panned out as a reliable marker for all tumor types. In the central nervous system, where a rare and aggressive childhood cancer called atypical teratoid/rhabdoid tumor (AT/RT) can be difficult to distinguish from other embryonal tumors, claudin-6 staining turned out to be disappointingly nonspecific. Only about 29% of AT/RTs were positive, and claudin-6 also turned up in 60% of medulloblastomas and a fifth of malignant gliomas, undermining its diagnostic value in that setting.²⁴ The lesson is that claudin-6 expression is a strong signal in certain tumor families and a much weaker one in others.

Expression Heterogeneity and What It Means for Treatment

One of the trickiest realities of claudin-6-targeted therapy is that not every cell in a positive tumor actually carries the protein. The ovarian cancer data showing distinct positive and negative patches within the same tumor illustrate a broader problem.⁹ If a drug kills only claudin-6-positive cells, the negative cells survive and can regrow the tumor. In testicular germ cell tumors, the sharp drop-off in claudin-6 expression after chemotherapy raises a related concern: the treatment itself may select for claudin-6-negative cells, leaving a resistant population that a claudin-6 drug cannot touch.⁷

These patterns suggest that claudin-6 therapies will likely work best in combination, either with conventional chemotherapy, with drugs targeting other tumor proteins, or with treatments that can address the negative cells independently. Expression testing before treatment, and possibly re-testing during treatment, will probably be essential for selecting the right patients and monitoring for escape. The field is still working out the best biopsy strategies and scoring systems to guide those decisions, and no standardized diagnostic test for claudin-6 has been widely adopted yet.

Claudin-6 in Breast Cancer and the Tumor Suppressor Paradox

Breast cancer presents a particularly paradoxical picture. In some breast cancer cell lines, claudin-6 is undetectable, and restoring it experimentally increases tight junction strength and reduces the malignant phenotype.²⁵ That is classic tumor-suppressor behavior: when the protein is present, it acts as a brake on cancer. Yet in the very same cell lines, claudin-6 has also been shown to promote resistance to chemotherapy and enhance stem-cell-like traits. The abnormal expression of claudin-6 can disrupt tight junction integrity through various mechanisms, contributing to tumor development.¹

The resolution to this apparent contradiction likely lies in context: expression level, timing, location within the cell, and which signaling partners are available all determine whether claudin-6 behaves as a brake or an accelerator. In early-stage disease or when re-introduced into cells that have lost it, it may reinforce normal cell architecture. In advanced or treatment-resistant disease, it may co-opt survival pathways that help the tumor persist. This dual personality is not unique to claudin-6; several other tumor-associated proteins toggle between suppressor and promoter roles depending on the cellular context. But it does complicate the therapeutic narrative: simply eliminating claudin-6 is not an obvious universal strategy, because in some settings you might be removing a brake rather than a fuel line. The current therapeutic approaches that use claudin-6 as a homing signal rather than trying to modulate its function directly sidestep this problem neatly. They do not care whether claudin-6 is helping or hurting the tumor; they just use it to deliver a lethal payload.