Analytical Validity of Deep Sequencing to Characterize Tissue-Infiltrating Lymphocytes

Lymphocytes infiltrate most human tissues, comprising a critical component of the adaptive immune response. Infiltrating lymphocyte populations have previously been associated with outcomes for a variety of clinical issues, including cancer, alloimmunity, and autoimmunity. Quantitative detection of lymphocyte infiltration has potential as a clinical biomarker of disease activity. Furthermore, locally infiltrating disease-associated clones can be monitored longitudinally in peripheral blood, enabling less invasive follow-up. Deep sequencing of lymphocyte variable regions is an ideal candidate for clinical characterization of tissue-infiltrating lymphocytes. Tissue biopsies produce low genomic DNA yields generally and also contain only a small percentage of cells that have undergone variable region recombination. In this study, we first performed two analytical validity experiments to determine the sensitivity of T cell receptor (TCR) repertoire sequencing, and then made measurements of infiltrating lymphocytes in non-small cell lung carcinomas (NSCLCs), gastrointestinal (GI) biopsies, and cerebrospinal fluid (CSF).

The first analytical validity experiment was designed to determine whether we could accurately characterize rare T cells in a background of cells with unrearranged TCRs. We therefore spiked genomic DNA from three unique T cell clones into background unrearranged human genomic DNA derived from immortal lung cancer carcinoma cells. The dilution series was designed such that the high end of the range surveyed the equivalent of 40,000 T cells, and the low end of the range surveyed the equivalent of 1,200 T cells. We then performed commercial TCR repertoire sequencing (GigaMune Rep-Seq™). The second analytical validity experiment was designed to determine reproducibility and accuracy across a dilution series. Here, we spiked two TCR clones into blood genomic DNA from a normal human individual. These DNA mixes were created at clonal cell to normal blood genomic DNA ratios of 1:1000, 1:100, 4:100, and 1:10. Then, we subjected these DNA mixtures to commercial TCR repertoire sequencing in triplicate. Finally, as proof-of-principle, we used deep TCR sequencing to characterize tissue-infiltrating lymphocytes and peripheral blood from patients with NSCLC, multiple sclerosis (MS), and graft-versus-host disease (GVHD).

In the two analytical validity experiments, we detected all expected clones, at quantities as low as 50 copies per reaction. In the first experiment, as a metric of detection noise, we calculated a normalized count of primer-dimer reads (“background”) across the dilution series. The background for the lowest spike-in mixture was 96% lower than a no-template negative control. In the second analytical validity experiment, the average variation coefficient across replicate measurements was 10%; regression analysis of observed versus expected TCR counts had an r²of 0.93. In our proof-of-principle experiments, we were able to detect oligoclonal TCR in CSF biopsies from two MS subjects. Each CSF contained a single clone at >39% of the total TCR repertoire. Matched peripheral blood from one of these subjects contained the CSF-infiltrating clone at 8% of the total TCR repertoire. We were also able to reproducibly detect TCR clones in GI biopsies from two GVHD subjects. For each of the GI biopsies, top GI-infiltrating clones represented >10% of the total peripheral blood TCR repertoire +30 days after GVHD diagnosis. Finally, we were able to detect tumor-infiltrating TCR clones in NSCLCs and matched peripheral blood from two subjects. In one of the NSCLC subjects, three of the top five tumor-infiltrating clones were among the four most common TCR clones in matched peripheral blood.

TCR repertoire sequencing can be used to quantitatively, sensitively, and accurately detect tissue-infiltrating T cell clones. We have shown that tissue-infiltrating T cell clones proliferate significantly in the peripheral blood, across disease conditions as divergent as MS, GVHD, and NSCLC. In the future, we will conduct analytical and clinical validity studies that will eventually be used as evidence to justify use of deep TCR sequencing as a clinical diagnostic in autoimmunity, alloimmunity, and oncology.

Johnson:GigaGen Inc.: Employment, Equity Ownership. Löhr:GigaGen Inc.: Employment, Equity Ownership. Howell:GigaGen Inc.: Employment, Equity Ownership. Hsu:GigaGen Inc.: Employment, Equity Ownership. Meyer:GigaGen Inc.: Consultancy, Equity Ownership.