Series: What Early Human Data Tells Us
Early immune-targeting programmes routinely generate evidence of exposure, target binding and biomarker movement before they generate evidence of desired clinical effect. A receptor may be occupied, a proximal signalling marker may change, or an ex vivo assay may confirm that a compound is active. Yet development teams still face the harder question: has the drug altered an integrated human immune response strongly and coherently enough to justify the next study?
That distinction matters. Target engagement shows that a compound has reached the intended biology, but it does not automatically demonstrate meaningful downstream immune modulation. Patient studies may eventually answer the efficacy question, but they are slower, more heterogeneous and often poorly suited to exploring non-therapeutic doses. A controlled immune challenge can provide an intermediate step by activating a defined pathway in healthy volunteers and measuring whether the investigational medicine changes the resulting response.
This is the fourth article in the series What Early Human Data Tells Us. The first three discussed why inflammatory models need benchmarks, how early immune signals should be interpreted, and what in vitro, ex vivo and in vivo challenges can reveal before patient studies begin. This article uses the first-in-human development of the anti-OX40L antibody KY1005, now known as amlitelimab, to examine functional proof-of-pharmacology: evidence that a compound changes an intact adaptive immune response in humans, rather than only a molecular marker.
From Target Engagement To Functional Proof
The keyhole limpet haemocyanin (KLH) challenge is designed to probe adaptive immunity. KLH is a foreign protein to which most participants have not previously been exposed. After intramuscular immunisation, investigators can quantify the developing anti-KLH IgM and IgG response. An intradermal KLH challenge several weeks later then generates a local recall response that can be measured through skin perfusion, erythema and, when needed, tissue or fluid sampling.
This two-stage design connects several levels of biology. It asks whether antigen presentation, T-cell activation, T-cell help to B cells, antibody formation and tissue recall can occur, and whether a drug modifies that sequence. KLH is not a model of atopic dermatitis or any other chronic inflammatory disease. It is a mechanistic tool that isolates a controlled, T-cell-dependent component of adaptive immunity and makes it measurable within intact human physiology.
That made the model relevant to KY1005. The antibody was developed to block OX40 ligand, a co-stimulatory molecule expressed by antigen-presenting cells that interacts with OX40 on activated T cells. A conventional first-in-human study could establish safety, pharmacokinetics and perhaps receptor-level pharmacology. Adding KLH created a more demanding test: did OX40L blockade measurably suppress the generation and recall of an antigen-specific human immune response?
Case Study: KY1005 And The KLH Challenge
The published first-in-human study was conducted at CHDR with Kymab as a randomised, double-blind, placebo-controlled trial in 64 healthy men. Eight cohorts received KY1005 or placebo in a 6:2 ratio. Three cohorts received single doses from 0.006 to 0.05 mg/kg. Five multiple-dose cohorts received loading doses from 0.15 to 12 mg/kg, followed by two maintenance doses four weeks apart.
One week after the third dose, participants in the multiple-dose cohorts were immunised with KLH, a neoantigen. Twenty-one days later, participants received intradermal KLH and a saline control. The investigators measured anti-KLH IgM and IgG, as well as local cutaneous blood perfusion and erythema after the intradermal challenge. These pharmacodynamic endpoints were interpreted alongside safety, pharmacokinetics and exposure-response analyses.
KY1005 was generally well tolerated in the study, with no serious adverse events and only temporary mild or moderate adverse events reported. Its pharmacokinetics were nonlinear, with an estimated terminal half-life of approximately 24 days. More importantly for the development hypothesis, the functional challenge showed a pharmacodynamic pattern. At doses of 0.45 mg/kg and above, KY1005 reduced KLH-induced cutaneous blood perfusion and erythema. Exposure-response modelling also demonstrated statistically significant effects on anti-KLH IgG and the erythema response. That combination is more informative than movement in a single biomarker. The study linked drug exposure to attenuation of both a systemic neoantigen response and a local tissue recall response.
From Healthy Volunteers To Patient Dosing
The KLH study did not show that KY1005 would improve atopic dermatitis. It did something earlier and more specific: it demonstrated that clinically achievable exposure could alter OX40L-dependent immune function in humans and identified a dose range in which that effect became visible.
Kymab subsequently evaluated KY1005 in a randomised, placebo-controlled phase 2a study in 89 adults with moderate-to-severe atopic dermatitis. Participants received a 200 mg loading dose followed by 100 mg every four weeks, a 500 mg loading dose followed by 250 mg every four weeks, or placebo. At week 16, mean percentage reductions in Eczema Area and Severity Index were 69.9% and 80.1% in the low- and high-dose groups, respectively, compared with 49.4% with placebo; the high-dose comparison reached statistical significance.
The healthy-volunteer and patient regimens cannot be converted one-to-one: body weight, loading and maintenance schedules, pharmacokinetics and disease biology all matter. Even so, the phase 2a doses sat within the broader multiple-dose window in which functional KLH effects had been observed. The sequence is therefore a useful example of translation across stages. The KLH model first identified human functional activity and an exposure-response relationship; the patient study then tested whether dosing in that pharmacologically active region produced clinical benefit.
Commercial Impact — And The Limits Of Early Proof
Early human pharmacology can also influence how an asset is valued. It would be inaccurate to attribute a transaction to one study, but the KY1005 programme had more than a plausible target and encouraging preclinical data. By the time Sanofi agreed to acquire Kymab in 2021, the asset had demonstrated functional proof-of-pharmacology in healthy volunteers and positive phase 2a results in atopic dermatitis. Sanofi announced an upfront payment of approximately $1.1 billion, with up to $350 million in potential milestones, and highlighted KY1005 as a central part of the acquisition rationale.
The KLH data contributed to that evidence package by reducing a specific uncertainty: whether blocking OX40L could meaningfully change human adaptive immune function at tolerated exposures. That is commercially relevant because a coherent human pharmacology story can support dose rationale, strengthen confidence in mechanism and make subsequent patient data easier to interpret. It does not determine asset value on its own, but it can change the quality of the questions that remain.
The later history is equally instructive. In July 2026, Sanofi announced that it would discontinue development of amlitelimab in atopic dermatitis and would not seek global regulatory review, concluding that the totality of efficacy and safety evidence did not support further development in that indication and that the programme would not meaningfully improve the standard of care. Development in coeliac disease was still ongoing at the time of the announcement.
That outcome does not invalidate the early KLH result. It clarifies its scope. The challenge study showed that the antibody engaged its intended human biology and functionally modulated an adaptive immune response. It could not establish the magnitude of benefit required in a competitive, heterogeneous chronic disease, predict every safety consideration across a large programme, or guarantee commercial differentiation. Proof-of-pharmacology is a de-risking tool, not a promise of approval.
What This Data Tells Us
The first lesson is that meaningful early pharmacology is more than target engagement. A functional challenge asks whether molecular activity propagates through an integrated immune system to alter antibody formation, cellular coordination or tissue physiology. When several mechanism-relevant endpoints move coherently and in relation to exposure, the result is more persuasive than a proximal biomarker alone.
The second lesson is that model selection should follow the mechanism. KLH was useful for KY1005 because OX40L is involved in antigen presentation and T-cell co-stimulation. An acute innate challenge would have asked a different question. The system, stimulus, sampling schedule and endpoints must be chosen around the biological decision the programme needs to make.
The third lesson is that early challenge data can improve dose and study design without pretending to predict efficacy. In the KY1005 programme, the KLH response identified a pharmacologically active dose region before the asset entered a patient trial. That allowed the next study to start with stronger human evidence and fewer assumptions. The eventual atopic dermatitis outcome also shows why the boundary must remain clear: functional immune modulation is necessary for many mechanisms, but it may not be sufficient for clinically or commercially meaningful benefit.
For development teams, that is the practical value of controlled immune challenges. They turn an otherwise quiet pathway into a measurable human experiment and test whether a compound can change the response it was designed to change. Used alongside pharmacokinetics, safety, biomarkers and later patient data, they help distinguish a molecule that merely reaches its target from one that produces functional human pharmacology — while preserving a realistic view of what early evidence can and cannot prove.
