Premium
B220 analysis with the local lymph node assay: proposal for a more flexible prediction model
Author(s) -
Betts Catherine J.,
Dearman Rebecca J.,
Kimber Ian,
Ryan Cindy A.,
Gerberick G. Frank,
Lalko Jon,
Api Anne Marie
Publication year - 2007
Publication title -
journal of applied toxicology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.784
H-Index - 87
eISSN - 1099-1263
pISSN - 0260-437X
DOI - 10.1002/jat.1269
Subject(s) - local lymph node assay , false positive paradox , lymph node , lymph , immunology , medicine , pathology , computer science , skin sensitization , artificial intelligence , sensitization
The mouse local lymph node assay (LLNA) has been developed and validated for the identification of chemicals that have the potential to induce skin sensitisation. In common with other predictive test methods the accuracy of the LLNA is not absolute and experience has revealed that a few chemicals, including for instance a minority of skin irritants, may elicit false‐positive reactions in the assay. To improve further the performance of the LLNA, and to eliminate or reduce false‐positives, there has been interest in an adjunct method in which the ability of chemicals to cause increases in the frequency of B220 + lymphocytes in skin‐draining lymph nodes is measured. Previous studies suggest that the use of B220 analyses aligned with the standard LLNA may serve to distinguish further between contact allergens and skin irritants. In the original predictive model, chemicals were regarded as being skin sensitisers if they were able to induce a 1.25‐fold or greater increase in the percentage of B220 + cells within lymph nodes compared with concurrent vehicle controls. Although this first prediction model has proven useful, in the light of more recent experience, and specifically as a consequence of some variability observed in the frequency of B220 + lymphocytes in nodes taken from vehicle control‐treated animals, it is timely now to reconsider and refine the model. As a result a new prediction model is proposed in which reliance on the use of absolute thresholds is reduced, and in which small changes in control values can be better accommodated. Copyright © 2007 John Wiley & Sons, Ltd.