Discovering Potent Antibodies, Directly from Blood – a Proof-of-Concept
- Maurits den Boer

- 4 dagen geleden
- 4 minuten om te lezen
Bijgewerkt op: 4 uur geleden
What if the most potent human antibodies are proteins circulating in our blood, invisible to current antibody discovery tools?
This question sits at the heart of our latest work, which is now available as a bioRxiv preprint. In previous articles in this series, we discussed why the most optimized, protective antibodies circulating in our bodies are best discovered at the protein level, directly from serum. This article is the proof of the pudding, where we put our protein-centric antibody discovery platform, AbDirect, to the test.
An exceptional sample that should have worked, but didn’t
First, let me take you back to early 2020, the early days of the COVID-19 pandemic. After release from the ICU at Amsterdam UMC, a severely ill patient exhibited exceptionally strong viral neutralization activity in their blood. At least on paper, this was a great donor for mining antibodies that could help protect and treat new patients.
A prior study applied the current go-to discovery method to this patient and two others (Brouwer et al, Science, 2020). That method is single-cell B cell receptor sequencing (scBCR-seq), and it analyzes patrolling immune cells from large blood draws for potential antibody sequences. This approach has proven very powerful and has produced many important drugs on the market today, but when the sequences from this particular patient were tested, surprisingly, none of them effectively neutralized the virus. This left a conundrum: the patient’s blood clearly contained very potent antibodies, yet the genetic sequencing method didn’t surface them. How could this be?
The answer comes down to where each approach looks. Genetic methods do not look at antibodies, but rather at patrolling immune cells from blood, typically memory B cells. The protective antibodies themselves are produced elsewhere, namely, by plasma cells that reside tucked away in the bone marrow. These antibodies are present in circulation only as proteins, meaning that typical cell-based methods cannot directly sample the protective antibodies that the body actually produces.

Reading the antibodies themselves
To identify valuable protective antibodies, Abvion’s AbDirect approach skips the cells and goes straight to the proteins. It reads antibody repertoires directly from blood – or any other biofluid – capturing the immune system’s active final product.
In the current proof-of-concept study, we did this in three steps:
Capture and screen. From the same patient’s plasma, we pulled down the antibodies that recognize the SARS-CoV-2 spike protein and profiled them. This revealed >1,000 distinct antibodies, gave information on how they bind, and indicated their abundance.
Direct de novo sequencing. Using mass spectrometry alone, we determined complete sequences of the most dominant antibodies straight from the mixture. This is the hard part that the field struggles with, and that is at the core of our platform.
Produce and test. We produced these antibodies in the lab and assessed whether they bound the virus and neutralized it.

What we found
From the same sample analyzed in 2020, we were able to uncover 18 new antibodies of IgG1 and IgA1 isotypes. Their sequences were markedly different from the prior genetic dataset uncovered by scBCR-seq. There was not a single direct match, and germline usage was distinctive. This showed us that, indeed, our direct protein-based approach sampled a unique set of sequences.
All of these circulating antibody sequences could be produced in the lab; 15 of them bound the spike protein across several viral strains, and 8 neutralized the virus. Five did so potently, and three very potently (IC50 ≤1.4 nM, the lowest we tested). Notably, this was all from the exact same sample (same donor, same time point) where the standard genetic method had surfaced no neutralizers.

What it does and doesn’t mean
While very enticing as a proof-of-concept, the genetic methods remain faster and have proven enormously successful. This is especially true when you have good access to the right cells. Today, we therefore see AbDirect as a complementary tool for high-value antibodies that are out of their reach, either because the relevant cells are inaccessible or because the results come up short.
What this work does show is that discovering antibodies directly at the protein level can surface high-potency candidates that were overlooked by an established method. AbDirect can do this from minimal amounts of material, any biofluid holding antibodies – even decades-old archived samples – without using any cells in the process. That opens the door to disease areas where other approaches struggle, including in infectious diseases, autoimmune conditions, and cancer.
Looking ahead
This proof-of-concept represents an important milestone for Abvion, solidifying the foundation of what we are building. Our next steps are to broaden this approach beyond a single donor and into disease areas where accessing real circulating human antibodies could matter the most.
This preprint is the work of our team at Abvion together with our collaborators at Amsterdam UMC and Utrecht Univeristy, with the full method, data, and author list in the preprint. You can read it here: Direct serological antibody discovery by integrative proteomics (bioRxiv)
If you struggle to discover antibodies using conventional cell-based methods, we'd be glad to talk. Send me an email at contact@abvion.bio and follow us on LinkedIn to stay informed about our latest updates.

