For most of the history of antibody engineering, the assumption was that you needed both a heavy chain and a light chain to bind an antigen well. The Fab, the smallest functional binding unit of a conventional antibody, contains both. Then in 1993, a team at the Vrije Universiteit Brussel discovered that camels make antibodies that throw out the light chain entirely — and the heavy-chain-only variable domain still binds antigens with high affinity and specificity.
That single observation kicked off the field of nanobodies — the smallest single-domain antibody fragments — and, three decades later, has produced approved drugs, dozens of clinical candidates, and a fundamentally different way of thinking about protein binders. This guide unpacks what nanobodies are, why they work, and how to design with them. You can explore the approved nanobodies in the Antibody Design Lab.
The Hamers-Casterman discovery
In the early 1990s, an undergraduate course at the Vrije Universiteit Brussel needed a source of antibodies for a practical exercise. The group had access to dromedary serum. When they ran it on an SDS-PAGE gel under reducing conditions, they saw something strange: in addition to the expected heavy and light chain bands, there was an extra band with no corresponding light chain. Raymond Hamers, Cécile Casterman, and colleagues followed up and published the result in Nature in 1993.
It turned out that camelids — camels, llamas, alpacas, and related species — produce a unique class of antibodies that consist of two heavy chains and no light chains at all. The variable region of these heavy chains, called VHH, is fully functional on its own. A single VHH domain can bind antigens with affinities in the nanomolar range, often comparable to full conventional antibodies. The same architecture was later also identified in some shark antibodies (called VNAR or IgNAR).
What makes a VHH special
At the level of primary sequence, a VHH looks similar to a human VH3 framework. About 80 percent of the residues are conserved across the two. But camelid evolution made several small adaptations that let the VHH function on its own:
- Hydrophilic residues at the former VL interface.The face of a normal VH that would pair with a VL is hydrophobic. In a VHH, this surface is replaced by polar and charged residues that make the standalone VHH soluble. This is the main reason VHHs don't aggregate the way isolated human VH domains do.
- Extended CDR-H3. The third CDR loop of a VHH is often longer than in human antibodies (15 to 25 residues, sometimes more). This long loop can reach into clefts and enzyme active sites that conventional antibodies cannot access. Some VHH-target structures show the CDR-H3 inserted deep into a substrate-binding pocket.
- Disulfide bond between CDRs. Many VHHs have an extra non-canonical disulfide bond that connects CDR-H1 or CDR-H2 to CDR-H3, locking the loops into a defined geometry.
The combination of these features makes VHHs unusually robust: they refold after thermal denaturation, they tolerate high urea concentrations, and many can be expressed in functional form from E. coli inclusion bodies after refolding.
Why size matters
A standard IgG is about 150 kDa and roughly 14 nm across at its widest. A nanobody is about 12-15 kDa and 4 nm long. That difference has practical consequences across the entire drug pipeline:
Tissue penetration
Small molecules diffuse into solid tumors and tissues much faster than large proteins. Nanobodies penetrate dense tumor stroma and reach cells deep inside the tissue better than full IgG, which is one reason imaging agents based on radio-labeled nanobodies have become a major area of development.
Production cost
Conventional antibodies require mammalian cell culture (CHO, HEK293) for proper folding and glycosylation, which is expensive and slow. Nanobodies fold correctly in microbial hosts like E. coli and Saccharomyces cerevisiae and can be produced at gram-per-liter scale. That cuts cost dramatically.
Stability
Many VHHs survive temperatures above 70 °C, refold after denaturation, and remain functional after lyophilization. That opens routes to oral or inhaled formulations that would be impossible for conventional antibodies, which are typically destroyed by stomach acid or aerosolization stress.
Rapid clearance
The flip side of small size is fast clearance. A bare nanobody is below the renal filtration cutoff (~70 kDa) and gets excreted in the urine within hours. For most therapeutic applications, you need to extend the half-life — by fusion to an albumin-binding nanobody, by PEGylation, or by linking multiple nanobodies into a multivalent construct.
Approved nanobody drugs
Caplacizumab (Cablivi)
Caplacizumab was the first nanobody to receive FDA approval, in 2019, for the treatment of acquired thrombotic thrombocytopenic purpura (aTTP). aTTP is a rare disorder in which patients develop autoantibodies against ADAMTS13, the protease that cleaves von Willebrand factor (vWF). Without ADAMTS13, vWF multimers grow huge, snare platelets, and cause widespread microvascular thrombosis.
Caplacizumab is a humanized bivalent anti-vWF nanobody. It binds the A1 domain of vWF and blocks its interaction with the platelet GPIbα receptor, preventing the unchecked platelet aggregation. Bivalent format gives it extra avidity, and the molecule is small enough to be administered as a daily subcutaneous injection during the acute phase of aTTP.
Ozoralizumab (Nanozora)
Ozoralizumab is a trivalent humanized nanobody approved in Japan in 2022 for rheumatoid arthritis. It contains two anti-TNFα binding domains and a third anti-human serum albumin domain that extends the half-life by hitchhiking on circulating albumin. The two anti-TNFα domains give it strong avidity for TNFα trimers, which is the active signaling form of the cytokine.
Compared to conventional anti-TNFα antibodies like Adalimumab and Infliximab, Ozoralizumab is much smaller and is positioned as a once-monthly subcutaneous injection. Whether it ultimately wins meaningful market share against entrenched conventional anti-TNFα products will depend on real-world differentiation.
Nanobody design today
Immunization-derived libraries
The original route is to immunize a camelid (commonly a llama or alpaca) with the antigen, collect peripheral blood lymphocytes, amplify the VHH repertoire by PCR, clone it into a phage display vector, and pan against the antigen. This gives you genuine immune-matured VHHs that have been through somatic hypermutation in the animal, so affinities tend to be high. The downside is that immunization is slow, requires animal facilities, and isn't feasible for every target.
Synthetic VHH libraries
More recent approaches build large synthetic VHH libraries on top of a single optimized framework. The library is constructed by randomizing the three CDR loops with degenerate primers or chip-synthesized oligos. You then screen the library by phage display, ribosome display, or yeast display against any target you like, no animal required.
Synthetic libraries miss out on natural affinity maturation, but they make up for it by being huge (10^10 or more variants), fast to screen, and completely platform-portable. Hits from initial selections can be matured in vitro by introducing additional diversity into the binding loops.
Computational design
AI-assisted approaches to nanobody design are growing fast. Tools like NanoBodyBuilder2 (and ImmuneBuilder more broadly) can predict VHH structure from sequence in seconds. Sequence design tools like AntiFold can propose CDR variants conditioned on a target backbone. The combination — predict structure, design CDR variants, score them, predict the consequences — is starting to look like a real in-silico-first nanobody pipeline.
What you can do in the Antibody Design Lab
The Antibody Design Lab includes nanobody-aware workspaces:
- Caplacizumab — first FDA-approved nanobody, for aTTP
- Ozoralizumab — trivalent anti-TNFα nanobody for RA
- ImmuneBuilder structure prediction for any VHH sequence
- AntiFold-based CDR variant design starting from approved nanobody scaffolds
Bottom line
Nanobodies are the result of an evolutionary accident in camelids that turned into a major engineering opportunity. Their small size, ease of production, intrinsic stability, and unique geometry make them well-suited to applications where conventional antibodies struggle: imaging, intracellular targeting, oral and inhaled delivery, and rapid synthetic library design. They're not a universal replacement for IgG, but they have carved out a clear and growing niche in the protein engineering toolkit, with two approved drugs already and many more on the way.