How an antibody “rebuilds” its target: why some anti-CD20s call on complement, while others kill directly

Scientists have visualized what exactly happens to the CD20 receptor on B cells when therapeutic antibodies (rituximab, obinutazumab, etc.) attach to it. Using a new version of RESI super-resolution microscopy, they saw this in whole living cells at the level of individual proteins and linked the pattern of nanoclusters with different mechanisms of drug action. The result: “type I” antibodies (e.g., rituximab, ofatumumab) assemble CD20 into long chains and superstructures — this “plants” the complement better. “Type II” antibodies (e.g., obinutazumab) are limited to small oligomers (up to tetramers) and provide stronger direct cytotoxicity and killing through effector cells. The work was published in Nature Communications.

Background of the study

• Why CD20? Anti-CD20 antibodies are the workhorse in the treatment of B-cell lymphomas/leukemias and some autoimmune diseases. There are several drugs on the market, but they behave differently in the cell and produce different clinical profiles.
• Two mechanistic camps. Conventionally, there are antibodies of type I (rituximab, ofatumumab) and type II (obinutazumab, etc.). The former are more likely to include complement (CDC), the latter more often provide direct cell death and killing through effector cells (ADCC/ADCP). This has long been known from biochemistry and functional tests - but why exactly this is so at the nanometer level was unclear.
• What was missing from previous methods.
  • Classical immunofluorescence and even many super-resolving approaches do not see “single molecule” in a living membrane when the targets are tightly packed and dynamic.
  • Cryo-EM provides amazing details, but usually outside the context of a whole living cell.
    As a result, the "geometry" of CD20 under the antibody (which clusters, chains, sizes) had to be guessed from indirect data.
• Why Geometry Matters. Complement is “turned on” when C1q simultaneously captures correctly positioned Fc domains—it’s literally a matter of distances and angles. Likewise, the efficiency of ADCC/ADCP depends on how the antibody exposes its Fc to effector cell receptors. So, nano-architecture of CD20+antibody = key to function.
• What was the authors' goal? To show in whole living cells (in situ) what exactly different anti-CD20s do with CD20: what oligomers and superstructures arise, how this is related to complement incorporation and killing, and whether it is possible to control the mechanics through antibody design (binding angles, hinges, valence, bispecifics).
• Why is this necessary in practice?
  • Next-generation design: learning to “tweak the handles” of a structure to obtain the desired mechanism of action for a specific clinical task or tumor context.
  • Meaningful combinations: understand where a “complementary” drug is more appropriate, and where a “direct killer” is more appropriate.
  • Quality control/biosimilars: have a physical “fingerprint” of correct clustering as a biomarker of equivalence.

In short: therapeutic antibodies work not only "according to the recipe of the mechanism", but also according to the geometry that targets impose on the membrane. Before this work, we did not have a tool to see this geometry in a living cell with the accuracy of individual molecules - this is the hole that the authors are closing.

Why was this necessary?

Anti-CD20 antibodies are the basis of therapy for B-cell lymphomas and leukemias, and a means of “switching off” B-cells in some autoimmune diseases. We knew that “type I” and “type II” act differently (complement versus direct killing), but what this difference looks like at the nanometer level in the cell membrane was unclear. Classical methods (cryo-EM, STORM, PALM) in living cells did not reach the resolution of “one protein” precisely for dense, dynamic complexes. RESI does this.

What did they do?

• We used multi-target 3D-RESI (Resolution Enhancement by Sequential Imaging) and DNA-PAINT labeling to simultaneously highlight CD20 and its associated antibodies in the membrane of whole cells. Resolution is the level of individual molecules in an in situ context.
• We compared type I (rituximab, ofatumumab, etc.) and type II (obinutazumab; as well as clone H299) and quantitatively analyzed what CD20 oligomers they form - dimers, trimers, tetramers, and higher.
• We tested the relationship between the “pattern” and the function: we measured complement binding, direct cytotoxicity, and killing via effector cells. We also played with the geometry of the antibodies (example: flipping the Fab arms in CD20×CD3 T-cell engager) to understand how the flexibility/orientation of the hinge shifts the function between type I and II.

The main findings in simple words

• Type I makes chains and “platforms” of CD20 — at least hexamers and longer; this is a geometry that is convenient for C1q, so complement is better included. Example: rituximab, ofatumumab.
• Type II is limited to small assemblies (usually up to tetramers), but it has higher direct cytotoxicity and more powerful killing through effector cells. Example: obinutazumab.
• Geometry matters. Change the flexibility/orientation of the Fab arms of the CD20xCD3 bispecific antibody and its behavior shifts from “type II” to “type I”: CD20 clustering ↑ and direct cytotoxicity ↓ – a clear structure-function relationship.

Why is this important for therapy?

• Next-generation design: It is now possible to design antibodies specifically for a desired mechanism (more complement or more direct killing) by tailoring binding angles, hinges, and valence to achieve the desired CD20 nano-architecture.
• Personalization and combinations. If the “complement” route works better in a particular tumor, it is worth reaching for “type I” (or antibodies/bi-specifics that build long CD20 chains). If direct death is more important, choose “type II” and enhance it with effector pathways.
• Quality control and biosimilars. RESI effectively provides a geometry test: a model can be trained to recognize the “signature” of the correct CD20 oligomers and used as a biophysical control in the development of biosimilars.

A bit of mechanics (for those interested)

• According to cryo-EM and new images, type I (eg, rituximab) binds to CD20 at a shallow angle, bridges CD20 dimers, giving chains with platforms for C1q; ofatumumab does a similar thing, but with a smaller step in the chain and "plants" complement even more stably. Type II (obinutazumab) has a steeper angle and a different stoichiometry (1 Fab to 2 CD20), so it remains in the trimer-tetramer zone.

Limitations and what's next

• These are cell models with carefully controlled conditions. The next step is to confirm key CD20 cluster patterns in primary tumor samples and correlate them with clinical response.
• RESI is a complex technique, but the team emphasizes its versatility: it can map any membrane target and its antibodies—from EGFR/HER2 to PD-L1—and also link nano-architecture to function.

Conclusion

Antibodies work not only "according to the recipe of the mechanism", but also according to the geometry that they impose on the receptor in the membrane. It has become possible to see this geometry - and this opens the way to a more precise design of immunopreparations, where the desired clinical effect is set at the level of nanometers.

Research source: Pachmayr I. et al. Resolving the structural basis of therapeutic antibody function in cancer immunotherapy with RESI. Nature Communications, July 23, 2025. doi.org/10.1038/s41467-025-61893-w