OX40L’s Dual Role in Fighting Cancer and Autoimmunity

OX40L is a single molecule that sits at the center of two opposing therapeutic goals: revving up the immune system to destroy tumors and dialing it down to stop it from attacking the body’s own tissues. This tension exists because OX40L and its receptor OX40 form a signaling partnership that fundamentally controls how long T cells live, how aggressively they fight, and whether regulatory T cells keep them in check. The same pathway that researchers want to activate in cancer patients is one they need to silence in people with lupus, inflammatory bowel disease, or eczema. How that paradox plays out in the clinic is reshaping how immunologists think about targeted therapy.

The OX40 and OX40L Partnership

OX40 (also called CD134) is a receptor that appears on the surface of T cells after they have been activated by encountering something foreign. It does not sit there permanently. OX40 shows up transiently, mostly on the T cells that have been most recently switched on by antigen.¹ Its binding partner, OX40L (also called TNFSF4), is found on antigen-presenting cells like dendritic cells and macrophages, as well as on some activated T cells, endothelial cells, and mast cells.² When OX40L on an antigen-presenting cell engages OX40 on a T cell, it sends a “keep going” signal: survive longer, divide more, and produce more inflammatory molecules.

This costimulatory signal is not the initial spark that activates a T cell. That job belongs to the T cell receptor recognizing its target. OX40/OX40L is more like a volume knob. Without it, a T cell may activate, do some work, and die. With it, that same T cell expands into a larger army, produces stronger inflammatory responses, and persists as a memory cell that can respond faster the next time.³ The location and timing of the signal matter enormously, because turning up the volume in a tumor is therapeutic, while turning it up in healthy tissue is destructive.

Building Immune Memory Through OX40

One of OX40’s most consequential functions is shaping whether activated T cells become long-lived memory cells or simply flame out after their initial response. Engaging OX40 during the priming phase, when a T cell first encounters its target, leads to more memory T cells being generated, stronger migration of those cells to where they are needed, and enhanced production of inflammatory signaling molecules.⁴

Work on CD8+ T cells (the “killer” T cells that directly destroy infected or cancerous cells) has shown that OX40 signals during priming are not just helpful but necessary for high-quality memory formation. Without OX40, T cells could still mount an initial effector response, but the subset of cells that would normally transition into self-renewing memory cells was sharply reduced. Even the memory cells that did form without OX40 input failed to maintain themselves over time, suggesting the signal affects not just the birth of memory T cells but their long-term survival.⁵ This has obvious implications for cancer, where you want durable immune surveillance, and for autoimmunity, where long-lived self-reactive memory T cells perpetuate disease.

Exploiting OX40 in Cancer Immunotherapy

In the tumor microenvironment, OX40 signaling can be harnessed in at least two complementary ways. The first is straightforward: boosting the expansion and killing power of tumor-reactive T cells. In mouse models, treating tumor-bearing animals with an OX40 agonist antibody roughly doubled the frequency of tumor-reactive CD8+ T cells receiving strong activation signals compared to untreated controls.⁶ Separate work confirmed that tumor-infiltrating CD8+ T cells express OX40 at elevated levels when they encounter tumors, and that activating this receptor enhanced their ability to kill cancer cells and generate tumor-specific memory.⁷

The second mechanism involves regulatory T cells, or Tregs. Inside many tumors, the majority of CD4+ T cells are Tregs, which act as brakes on the immune response and help the tumor evade destruction. When researchers injected an OX40 agonist antibody directly into tumors in mice, it functionally shut down those Tregs and allowed CD8+ killer T cells to reject the tumor. About 80% of mice cleared their tumors after this treatment, an effect that disappeared when CD8+ T cells were removed from the picture.⁸ So OX40 agonism in the tumor setting both empowers the attackers and disarms the suppressors.

Why Combining OX40 Agonists with Checkpoint Inhibitors Matters

Checkpoint inhibitors like anti-PD-1 antibodies have transformed cancer treatment, but many tumors, particularly pancreatic cancers, respond poorly to them alone. Adding an OX40 agonist appears to overcome some of that resistance. In one set of experiments using mouse models of pancreatic cancer, animals given anti-PD-1 alone fared no better than untreated controls, with all dying within 50 days. OX40 agonist alone allowed about 43% to survive long-term. But the combination of anti-PD-1 plus OX40 agonist produced near-complete tumor clearance: almost all mice survived more than 225 days with no detectable tumors. In a more aggressive genetic model of pancreatic cancer, the combination roughly doubled median survival compared to either treatment alone.⁹

These are animal results, and the jump from mouse models to human tumors is notoriously unreliable. Still, the logic is compelling: checkpoint inhibitors remove the “don’t attack” signal that tumors hide behind, while OX40 agonists add a “keep fighting” signal to the T cells that respond. Together, they push the immune response from two directions simultaneously.

First-in-Human Safety Data for OX40 Agonists

The first-in-human trial of the OX40 agonist antibody INCAGN01949 in patients with advanced solid tumors provides an early window into what these drugs look like in practice. Immune-related side effects occurred in about 15% of patients and were overwhelmingly mild: fatigue, rash, and itching were the most common. Only one patient experienced a serious immune-related event (colitis), and that patient had previously received another immunotherapy drug, making it difficult to attribute the reaction to INCAGN01949 alone.¹⁰ The low toxicity is encouraging, though it is worth noting that single-agent OX40 agonism in humans has so far shown modest antitumor activity, reinforcing the rationale for combination approaches.

The Autoimmune Side of the Coin

The same signaling power that makes OX40/OX40L attractive for cancer becomes a liability when the immune system is already overactive. Genetic studies have established that the gene encoding OX40L (TNFSF4) is a susceptibility gene for systemic lupus erythematosus, an autoimmune disease in which the immune system attacks joints, skin, kidneys, and other organs. A risk haplotype in the upstream region of the gene correlates with increased OX40L expression on both the cell surface and at the transcript level, suggesting that people who naturally produce more OX40L are more prone to lupus.¹¹ The same gene has also been identified as a susceptibility locus for systemic sclerosis, another autoimmune condition characterized by fibrosis and vascular damage.¹²

In lupus patients specifically, OX40L expressed on myeloid antigen-presenting cells (rather than B cells) appears to drive a particular kind of immune dysfunction. It promotes the generation of T follicular helper cells, which in turn help B cells produce the autoantibodies that cause tissue damage. The frequency of OX40L-expressing antigen-presenting cells in the blood of lupus patients correlates with disease activity, providing a direct link between this molecule and how sick someone is.¹³

Soluble OX40L as a Disease Marker

OX40L does not only exist anchored to cell surfaces. A soluble form circulates in the blood, and its levels can signal autoimmune activity. In early rheumatoid arthritis, patients who tested positive for certain disease-associated autoantibodies had dramatically higher circulating levels of soluble OX40L. Those with a specific antibody marker (IgM rheumatoid factor) had median soluble OX40L levels of about 28 ng/mL, compared to roughly 1.7 ng/mL in antibody-negative patients. A similar gap appeared for anti-citrullinated protein antibodies, another hallmark of aggressive rheumatoid arthritis.¹⁴ Whether soluble OX40L is actively driving disease or is simply a readout of immune activation remains an open question, but the size of the difference hints at diagnostic or monitoring potential.

Blocking OX40L to Calm Autoimmune Disease

If too much OX40/OX40L signaling contributes to autoimmunity, then blocking it should help. Animal studies bear this out. In a mouse model of multiple sclerosis (experimental autoimmune encephalomyelitis), an antibody that neutralized OX40L markedly reduced disease severity and cut the infiltration of immune cells into the spinal cord. The effect worked both preventively and as treatment after disease was already established, and the mechanism appeared to involve blocking the migration of pathogenic T cells rather than preventing their initial development.¹⁵

A parallel story plays out in the gut. In inflammatory bowel disease, the expression of both OX40 and OX40L is elevated at sites of mucosal inflammation.¹⁶ In a mouse model of chronic colitis, blocking OX40L with an antibody markedly reduced both clinical symptoms and tissue damage, decreased the number of CD4+ T cells infiltrating the colon, and suppressed the production of key inflammatory molecules by those cells.¹⁷ Taken together, these findings show that OX40L blockade can pull the immune response down in multiple tissue types and disease contexts.¹⁸

Rocatinlimab and the Atopic Dermatitis Breakthrough

The most advanced clinical proof that blocking OX40 can treat autoimmune-driven disease comes from rocatinlimab, an anti-OX40 antibody developed for moderate-to-severe atopic dermatitis (eczema). In a phase 2b trial, rocatinlimab improved patient-reported itch, sleep disruption, and quality-of-life scores, with benefits appearing quickly and persisting for at least 20 weeks after treatment stopped.¹⁹ That durability is noteworthy because many existing eczema biologics require continuous dosing to maintain their effect.

Two large phase 3 trials, ROCKET-IGNITE and ROCKET-HORIZON, confirmed these results. In IGNITE, about 42% of patients on the higher dose of rocatinlimab achieved at least a 75% improvement in eczema severity at 24 weeks, compared to 13% on placebo. In HORIZON, the drug produced a 75% improvement in about a third of patients versus 14% on placebo. Both trials also met their secondary endpoints for skin clearance.²⁰ These are meaningful response rates for a disease that is notoriously difficult to control in its severe forms, and they validate in humans what animal models predicted: that dampening OX40 signaling can cool down T cell-driven inflammation.

Why OX40 Does Not Always Do What You Expect

A clean narrative of “agonize OX40 for cancer, block it for autoimmunity” is too simple. The reality is that OX40 signaling does not have a single fixed output. Its effect on T cell populations depends heavily on what other signals are present locally, particularly the cytokine environment. When researchers activated OX40 in the presence of TGF-β1, a cytokine that normally pushes T cells toward becoming regulatory T cells, the OX40 signal overrode that push and diverted cells toward inflammatory effector fates instead. But when inflammatory cytokines were blocked while OX40 was stimulated, the result flipped: regulatory T cells actually expanded.²¹

This context-dependence is not just a lab curiosity. It means that an OX40 agonist given to a cancer patient might have different effects in an inflamed tumor bed than in a quiet lymph node, and that the same drug could theoretically promote immune suppression in one tissue compartment while boosting inflammation in another. It also complicates the assumption that OX40 agonism inherently suppresses Tregs. In the right cytokine setting, it may do the opposite. Designing drugs and dosing regimens around this plasticity is one of the major unsolved challenges in the field.

OX40L Beyond the Immune System

The OX40/OX40L axis is not confined to classic immune battles between T cells and pathogens or tumors. It also shows up in cardiovascular disease. OX40L is expressed on activated macrophages and endothelial cells within blood vessel walls, and research has implicated the OX40/OX40L system in the formation of tiny new blood vessels within the walls of larger arteries, a process called vasa vasorum neovascularization that is closely linked to the growth of atherosclerotic plaques.²² This suggests the pathway may be a therapeutic target not just for autoimmunity and cancer but for cardiovascular disease as well, though clinical work in this area lags far behind.

Next-Generation Drug Designs

Recognizing the limitations of single-target approaches, several groups are developing bispecific antibodies that combine OX40 targeting with a second immune target in one molecule. One such construct, ATOR-1015, binds both CTLA-4 and OX40. The idea is that because both targets are often co-expressed on regulatory T cells within tumors, the bispecific antibody concentrates its activity in the tumor microenvironment rather than broadly activating the immune system everywhere. In preclinical work, ATOR-1015 produced immune activation specifically within tumors, suggesting a potential route to stronger efficacy with less systemic toxicity than separately dosing a CTLA-4 inhibitor and an OX40 agonist.²³

Other bispecific designs pair OX40 with CD137 (4-1BB), another costimulatory receptor, aiming to deliver a stronger combined “go” signal to T cells only when both targets are engaged on the same cell.²⁴ The strategy of requiring dual engagement is a safety measure: it narrows the drug’s activity to cells that display both markers, reducing off-target immune activation. Whether this precision translates into better clinical outcomes remains to be seen, but the engineering approach reflects how seriously the field takes the problem of activating OX40 broadly when what you want is local, tumor-specific activation.

On the autoimmune side, novel anti-OX40 antibodies are being designed that not only block the OX40/OX40L interaction but also deplete OX40-expressing T cells, potentially offering a more definitive way to shut down the pathogenic T cells that drive conditions like lupus and inflammatory bowel disease.²⁵ Because OX40 is transiently expressed primarily on the most recently activated T cells, depleting them could selectively remove the cells most actively causing damage while leaving the broader immune repertoire intact. That selectivity, if it holds up in humans, would be a meaningful advantage over existing immunosuppressive drugs that dampen the immune system more broadly.