thebench.bio #1 Ecosystem for the Bio-Tech Industry
All Antibody Engineering questions

Question 19 of 20 Advanced Practical

What would a cross-reactivity testing strategy for an antibody candidate include?

Short answer

Test binding against the closest family members and across a panel of normal human tissues, plus a general stickiness assessment. For a therapeutic, tissue cross-reactivity by immunohistochemistry across a range of human tissues is expected, and the same work identifies which animal species is relevant for safety testing.

Why

Cross-reactivity has two faces: binding to related proteins, and binding to anything else. The first is tested against purified family members, ideally with the same format and conditions, and against cells expressing each. The second needs a broader net.

ICH S6(R1) states that for monoclonal antibodies the immunological properties should be described in detail, including antigenic specificity and any unintentional reactivity or cytotoxicity toward human tissues distinct from the intended target, and that such cross-reactivity studies should be carried out by appropriate immunohistochemical procedures using a range of human tissues. The same guideline ties this to species selection: a relevant species is one in which the molecule is pharmacologically active because the epitope is expressed, and relevant species for antibodies are those that express the epitope and show a similar tissue cross-reactivity profile to human tissues. Developability work adds a third axis, nonspecific stickiness, measured with polyspecificity reagents and baculovirus particle binding, which predicts poor pharmacokinetics even without a specific off-target.

Tissue work is read by a pathologist and generates findings that need interpretation rather than a pass or fail.

How to work it through

  1. List the closest homologs by sequence and by structural similarity of the epitope region.
  2. Test binding to each purified protein and to cells expressing them, at a concentration range.
  3. Run a polyspecificity or nonspecific binding panel to detect general stickiness.
  4. Commission tissue cross-reactivity by immunohistochemistry across a range of normal human tissues with pathologist review.
  5. Repeat the tissue panel in candidate animal species to identify a relevant species.
  6. Decide for each finding whether it is a real target, an artifact or a risk to monitor, and record the reasoning.

An example

A candidate against one member of a four-member receptor family binds the intended target with a KD of 0.8 nM and shows no measurable binding to the other three up to 1 micromolar. A polyspecificity screen puts it in the low-risk group. The human tissue panel shows expected staining in the target tissue and unexpected cytoplasmic staining in one organ, which the pathologist attributes to a known artifact pattern and which does not appear with a non-binding control antibody.

The cynomolgus monkey panel matches the human one, supporting that species for safety work.

Where the answer stops holding

Tissue cross-reactivity uses fixed or frozen tissue and detects binding, not consequence, so findings need interpretation. A homolog panel cannot cover unrelated proteins that share a surface. Negative results bound only the conditions and formats tested.

A strong answer covers

  • Test close homologs as purified protein and on cells
  • ICH S6(R1) expects tissue cross-reactivity across a range of human tissues
  • Species relevance depends on epitope expression and a similar tissue profile
  • Nonspecific stickiness is a separate measurable risk
  • Findings need pathologist interpretation, not a simple pass or fail

Common mistakes

  • Testing only the closest homolog and calling the molecule specific
  • Choosing a toxicology species before confirming the epitope is expressed
  • Treating every tissue panel finding as a safety signal without interpretation