HIV's Impact on the Immune Repertoire: New Analytic Methods

Paul David Baum, University of California, San Francisco
Basic-Applied Clinical

The "holes in the repertoire" hypothesis proposes that HIV causes immunodeficiency by diminishing the diversity of the repertoire of T cell receptors (TCRs) that recognize foreign antigens, leading to a failure to recognize and fight off pathogens. If true, this hypothesis would suggest a benefit to starting HAART at an early stage of illness, before much of the TCR repertoire has been lost. It also suggests that TCR diversity measurements could be used to evaluate adjunctive immune therapies (e.g., growth hormone or interleukins 2 or 7) meant to bolster the T cell count. Unfortunately, it has not been possible to verify the presence of "holes in the repertoire" because there has been no practical way to measure TCR diversity in clinical research. I have recently invented an automated method, called AmpliCot, which uses the principle of DNA hybridization kinetics to measure TCR diversity. While AmpliCot is useful as it stands, development of more sophisticated statistical methods could greatly increase the power and range of this novel assay in HIV research applications.

The theoretical development and empirical testing of statistical methods for AmpliCot are proposed in three specific aims. First, the robustness and sensitivity of AmpliCot measurements will be optimized. Tools for this aim will include formal rules defining acceptable or aberrant data, new calculation methods for AmpliCot measurements (using slopes rather than intercepts), and the development of high diversity standards to permit interpolated diversity measurements. Methods will be tested using repeated measurements of replicate cell samples. Improving the assay's ability to detect rare specificities in the repertoire will be crucial in distinguishing between the loss of T cell specificities and the dilution of T cell specificities--a key distinction in HIV immunopathogenesis.

Second, methods will be developed to allow comparison of the diversity of two differently-sized samples. Computational (annealing curve integration) and statistical (regression) approaches to this aim will be devised and then tested using titrated sample inputs of various cell types. In addition to important practical considerations in designing experiments, these methods will also enable diversity comparisons of T cell subpopulations, such as naive CD4+ T cells, between HIV-uninfected subjects who have many of these cells and HIV-infected subjects who often do not.

Third, methods will be devised for calculating the diversity of a whole T cell population in terms of the TCR diversities of its sub-populations. Potential approaches include multinomial or regression calculations. These approaches will be tested with test samples containing deliberate mixtures of different T cell subpopulations. HIV-infected patients often have marked decreases in their CD4+:CD8+ and naive:memory T cell ratios. Statistical methods will be important in determining whether overall diversity changes in HIV-infected patients are due to changes in these ratios, or whether diversity has also been lost in these subpopulations. These methods may allow future analyses of the immune repertoire to proceed without expensive and time-consuming cell sorting.

In conclusion, the statistical tools developed and tested in this project will allow the use of AmpliCot to test the "holes in the repertoire" hypothesis of HIV pathogenesis, and will permit routine measurements of TCR diversity in HIV clinical research.