Limit of Detection – Serum Antibody Analysis

Aug 2, 2024 | Science, Services

Project Summary

Antibodies are a major constituent of serum, accounting for 10% of protein in serum. Per mL of serum, there’s typically at least 6mg of IgG – the most abundant antibody isotype. At Abterra Biosciences, our Alicanto antibody discovery platform mines the serum antibody repertoire to identify functional, high affinity, and diverse monoclonal antibodies. Serum antibody titer is a phenotype often used for screening patients, tracking seroconversion in immunized animals, or monitoring disease status. Alicanto connects individual antibodies to the desired phenotype directly.

In traditional antibody discovery campaigns, there are practical limits on how rare an antibody clone can be. For hybridoma and single B cell screening, the natural limit is the number of hybridomas or B cells screened. For phage display, the limit is the effective library size. However, for serum antibodies, the limit of detection has not been established. In this blog post, we answer this question by looking at monoclonal antibodies (mAbs) spiked into a polyclonal (pAb) background, as would typically be analyzed by Alicanto.

Challenges of serum antibody analysis:

Mass spectrometry (MS)-based proteomics is used by Alicanto to analyze serum antibodies. MS, particularly bottom-up MS, whereby proteins are digested into 7-20 amino acid peptides for high throughput identification poses several challenges that must be addressed by Alicanto.

Challenge	Alicanto Solution
Protein sequence composition can result in peptides that are not amenable to analysis by bottom-up MS (e.g. too short, or too small, or not ionizable)	Multiple enzymatic digests increase the probability that a clone will produce peptides amenable to mass spec. For simplicity in this blog post, we only focus on tryptic peptides
Sequence homology can make data interpretation challenging, since it’s not clear from which clone a particular peptide may have been generated	Shared peptides are discounted compared to unique peptides, and region of the antibody matters (e.g framework covering peptides are less important for antibody sequencing compared to CDR covering peptides). We’ve developed rigorous false positive controls to prevent incorrect sequencing, and peptide-weighting scheme

Experimental Design

For this experiment, we chose the two therapeutic monoclonal antibodies novilumab and eculizumab, due to their reliability in displaying at least one unique CDR3-covering tryptic peptide in previous mass spectrometry runs. A protein G-purified serum sample from a human donor provided a pAb background into which we spiked either novilumab or eculizamab, at a range of dilutions spanning 20% to 0.1% by mass.

The spiked samples were then digested with trypsin and processed using our in-house sample preparation protocols. Each dilution was prepared in triplicate, as well as 100% (all spiked-in mAb) and 0% (all pAb background) controls. The resulting tryptic peptides were analyzed using our standard LC-MS methods and analyzed with Alicanto, specifically searching for the unique target CDR3 peptides for nivolumab and eculizamab.

Experimental set up: two monoclonal antibodies (novilumab and eculizumab) were individually spiked into a human polyclonal IgG background at 6 different concentrations in addition to controls that were 100% monoclonal and 100% polyclonal (15 samples total)

In this blog post, we report mAbs as detectable at two confidence levels. A higher confidence level, required in our standard Alicanto results, and a lower confident level used here only to evaluate limits of detection . For standard Alicanto, corroborating spectra with high individual confidence scores are required to identify a clone. For detection, we permitted a laxer definition that relied on the identification of at least 1 confident spectrum identification and peptides covering 100% of the CDR3.

Note: The 100% mAb spike in is equivalent to a monoclonal antibody sequences that are routinely sequenced with Valens.

Results

Standard Alicanto exhibits a consistent limit of detection of 5%, but can identify clones as low as 0.1%

Identification rate by Alicanto for limit of detection of each antibody

Proportion of replicates in which the mAb was detected through standard Alicanto analysis.

Standard Alicanto analysis resulted in the detection of both mAbs at 5% of the sample volume by mass in all 3 replicates. Eculizumab was detected in at least 1 of the three replicates when spiked in as low as 0.1% (6.6 fmole).

Proteomic detection is possible as low as 1% for both mAbs.

Spectrum identified at each dilution for limit of detection

Number of spectra identified for each mAb at each dilution matching the CDR3. While more spectra enable confident identification of clones, a small number of high confident peptide-spectrum matches can still result in identification of mAbs below 1% spike in.

By relaxing the Alicanto selection criteria to accept at least 1 spectrum supporting the CDR3 and 100% CDR3 coverage by peptides, we find that both mAbs are detectable as low as 1% of the sample in at least 1 replicate.

Conclusions

Replicate analysis enables the identification of clones at lower than 5% of the sample. For the two mAbs analyzed, both could be detected in all replicates down to 5% of the sample. For rarer clones, more than 1 replicate could be needed to enable confident identification.

Sequence composition influences clone detectability. In this analysis, we only used 1 common enzymatic digest. In standard Alicanto, multiple enzymes are used to overcome sequence composition challenges.

Antigen purification enriches relevant serum antibodies that are extremely rare in total IgG. In a typical Alicanto antibody discovery campaign, we focus only on antigen-specific antibodies. We use affinity chromatography to enrich for antigen-specific antibodies which often make up < 5% of the total IgG. This means that a clone that comprises 5% in the antigen purified samples (consistently above the limit of detect for Alicanto), will be only 0.025% of total IgG.

VHH Discovery with Alicanto versus Phage Display

Project Summary Camelids such as llamas and alpacas produce conventional antibodies as well as heavy-chain only antibodies (HCAbs) from which the small, stable, antigen-binding domain (VHH) can be extracted. Target-specific VHHs can be identified from immunized...