Next Generation Sequencing Spectratyping (NGS-S) Comprehensively Monitors T Cell Receptor Diversity in Children with T Cell Abnormalities

Abstract ²¹⁷³

The adaptive arm of the immune system - the T-cell compartment – may become compromised by inherited or acquired defects resulting in cancer, autoimmunity, or increased susceptibility to microbial infectious agents. A normal polyclonal T cell compartment comprises an estimated number of 2.5 × 10E7 individual T cell clones each expressing a unique antigen recognizing T cell receptor (TCR). The functionality of the T cell compartment is thus – at least partly - reflected by TCR sequence diversity. There is a medical need for rapid and robust diagnostic approaches that accurately monitor TCR diversity in patient samples, e.g. after bone marrow transplantation (BMT).

Previously, complementarity determining region 3 (CDR3) size spectratyping in TCR β-chain subfamilies (Vβ), an immunoscopic technique, was employed for the analysis of T cell diversity. However, spectratyping is limited to the analysis of CDR3 length polymorphisms only. Underlying diversity of TCR Vβ sequences of equal length remain undetected. Furthermore, spectratyping is time consuming and consequently data can only be interpreted with a delay of weeks.

To determine TCR diversity fast and accurately we developed next-generation-sequencing spectratyping (NGS-S), which employs high coverage, massive parallel Roche/454-sequencing of TCR Vβ-chain amplicons.

Three different sample groups were analyzed in parallel by spectratyping and NGS-S: T cells from (1) healthy children (n=6), (2) children at diagnosis of severe aplastic anemia (n=7), and (3) children who underwent haploidentical BMT (n=7). In brief, RNA was extracted from bone marrow derived CD8+ cells and transcribed to cDNA. Amplicon libraries were generated by PCR employing two degenerated wobble primers (VP1, VP2), designed to cover most of the known TCR Vβ gene segments and a universal reverse primer (CP1) located in the conserved TCR region (Figure 1). The mean overall coverage of the CDR3 region achieved was 23.133 per patient.

Figure 1:

View largeDownload slide

Workflow of NGS-S

Figure 1:

View largeDownload slide

Workflow of NGS-S

Close modal

For simultaneous characterization of these individual amplicons we generated the TCR Profiler (Figure 2). This new software tool automatically preprocessed raw sequencing data using a threshold quality value (q=30) to trim the 3' end of the TCR β-chain sequences. Rearranged germline TCR Vβ and Jβ genes were identified by Smith-Waterman local alignment against each human TCR β-chain germline gene of the IMGT/GENE-DB reference directory. Base call reliability was assured by incorporating phred-like quality values provided by the sequencer into the Smith-Waterman local alignment routine. A quality value, q, was computed for each base as the log-transformation of the probability p of a base being incorrectly called, q = − 10 x log₁₀(p). Transformed into a reliability measure, r = 1- (1/10^(q/10)), q-values indicated the probability for each base to be correctly called. CDR3 regions were delimited using specific flanking amino acid sequence motifs. The 5' end motif varies dependently on the rearranged TCR Vβ gene and was identified using the IMGT/GENE-DB reference directory sequence set, whereas the 3' end motif [W/F]GXG (IUPAC code) is conserved in all TCR β-chains. CDR3 length was calculated including all amino acids between the two motifs. Screening for in-frame stop codons was done to exclude non-functional TCR β-chain transcripts.

Figure 2:

View largeDownload slide

TCR-Profiler

Figure 2:

View largeDownload slide

TCR-Profiler

Close modal

The TCR profiler identified on average 16165 of the input sequences as unique CDR3 sequences. Of these a mean of 9840 were predicted to be functionally rearranged.

Whereas spectratyping gave a rough estimate of CDR3 size and allowed T cell deficient samples to be identified, NGS-S determined the exact length and sequence composition of the CDR3, identified the rearranged TCR Vβ and Jβ genes and the specific recombination (Figure 3A).

Figure 3:

View largeDownload slide

Comparison of diversity detection by spectratyping and NGS-S (A). NGS-S, but not spectratyping, allows differentiation between diverse T cell pathologies (B).

Figure 3:

View largeDownload slide

Comparison of diversity detection by spectratyping and NGS-S (A). NGS-S, but not spectratyping, allows differentiation between diverse T cell pathologies (B).

Close modal

Utilization of specific genes, the resulting amino acid composition of the CDR3 region as well as its length and overall diversity were integrated into a novel NGS-S score.

This score reliably differentiated not only between normal and T cell deficient samples, but also between the different T cell deficient groups (SAA and BMT) (Figure 3B). We conclude that NGS-S allows rapid and precise determination of TCR diversity in clinical samples.