Design of limiting dilution analysis experiments for helper T lymphocyte precursor frequency determination in the context of allogeneic bone marrow transplantation.

Helper, interleukin 2 (IL-2) producing, T lymphocyte precursor (HTLp) frequency determination by limiting dilution analysis (LDA) is of value for quantifying alloreactivity in allogeneic bone marrow transplantation (BMT). LDA assays are labour-intensive and time-consuming to perform and the numbers of donor and recipient cells available are limited. It is therefore important that the design of the experiment yields reliable frequencies with a minimum of effort and a realistic cell requirement. We have critically evaluated the methods proposed for LDA design by Strijbosch et al. [Strijbosch, L.W., Buurman, W.A., Does, R.J., Zinken, P.H., Groenewegen, G., 1987. Limiting dilution assays. Experimental design and statistical analysis. J. Immunol. Methods 97, 133] and by Blackett and Gordon [Blackett, N.M., Gordon, M.Y., 1996. Optimizing limiting dilution assays: frequency and 'ability' measurements of haemopoietic progenitor cells. Br. J. Haematol. 92, 507 (see comments)] and found them inadequate for this application. The estimation of the HTLp frequency is traditionally based on the single-hit Poisson model and the adequacy of this model was compared with that of a double-hit model. The results were in favour of the single-hit model. Ten different LDA experimental designs were explored by Monte Carlo simulations. The optimal design exploits the maximal numbers of cells that can be obtained for analysis to estimate HTLp frequencies in the range 1:1,000,000-1:20,000 with a coefficient of variation of 10-20% and with a minimum of manual labour.