Somatic Hypermutation and V-J Gene Usage for CLL Prognosis: Evaluating Data from Miseq NGS Vs PCR-Sanger Sequencing Approaches

Introduction: Somatic hypermutation (SHM), an important process that evolved to increase the affinity of immunoglobulin molecules, occurs in the germinal center during B-cell maturation. The presence of SHM, defined as ≥ 2% difference from the germline sequence of the immunoglobulin heavy chain variable (IGHV) region genes, provides important prognostic information for patients with chronic lymphocytic leukemia (CLL), as leukemic clones that fail to undergo SHM are more aggressive tumors. In addition, presence of the V3-21 variable segment portends a worse prognosis independent of mutation status or other prognostic indicators. The gold standard method used to determine SHM status and presence of V3-21 rearrangements requires two steps: a PCR/capillary electrophoresis (CE) based method to detect clonality, followed by Sanger sequencing. This multistep approach is labor intensive and time consuming.

In order to perform a clinical evaluation of the relative merits of NGS and Sanger sequencing approaches, DNA was tested from anonymized, blinded clinical samples obtained from the University of Bologna (n=62).

Methods: We examined a next-generation sequencing (NGS) approach for determining SHM status for IGHV region and compared these results to results from PCR-based clonality testing using the BIOMED-2 IGH FR1 master mixes, followed by gel extraction and Sanger sequencing. The MiSeq SHM Assay employs two master mixes. One amplifies genomic DNA between the leader (VHL) region and the joining (J) region. The other amplifies from the framework1 (FR1) to J region. The NGS assay design of V and J consensus primers allowed PCR products to be sequenced directly on the MiSeq platform. PCR was followed by amplicon purification using the AMPureXP PCR system. Purified equimolar amounts of amplicons were pooled to form a library. A portion of the library was sequenced using the MiSeq v3 Reagent kit (600 cycles). Sequencing data were analyzed using LymphoTrack™ bioinformatics software.

Results: Prior to testing clinical samples the analytical performance of the SHM assay was evaluated using diluted contrived samples, which exhibit known V-J rearrangement and SHM status. For limit of detection (LOD), limit of blank (LOB), linearity, and precision and reproducibility (P/R) the assay demonstrated excellent linearity (R²>0.99), sensitivity (1%), repeatability (<2% variance) and reproducibility (<2% variance). In addition, good concordance (R²>0.99) was demonstrated between the NGS assay and the traditional method of using gel extraction and Sanger sequencing in evaluating the SHM status of cell line DNAs.

The clinical performance of the NGS SHM assay was evaluated by testing over 62 CLL samples that had also been tested using the traditional CE assay for clonality. The overall diagnostic accuracy was greater than 96%, sensitivity (ST) was 94%, specificity (SP) was 73%, positive predictive value (PPV) was 94%, and negative predictive value (NPV) was 72%. Forty-six samples that were tested for SHM status by both the MiSeq SHM assay and traditional Sanger sequencing had a concordance of 96%. Twenty-five samples were tested for SHM status by both MiSeq VHL/J and FR1/J assays with a concordance of 99%. Furthermore, the VHL/J mix detected clonality in SHM samples where mutation rate appeared to interfere with IGH FR1 priming. It was also observed that the CE assay could give a false impression of clonality if multiple clones were of very similar size – an occurrence quickly resolved by NGS sequencing showing the different DNA sequences of the same size from multiple rearranged clones thus indicating a polyclonal situation.

Conclusion: Both the NGS SHM MiSeq assay and PCR capillary electrophoresis-based clonality assays identified a majority of clonal IGH V-J rearrangements in clinical CLL samples. However, the BIOMED-2 IGH FR1 assay missed certain of the clonal rearrangements and determination of SHM status was a labor intensive, multistep process, whereas data output from the NGS assay included the DNA sequence, V-J usage and SHM status of CLL patient samples. The NGS assay also provided frequency distributions of V region and J region segments important in detecting IGH V3-21, a determinant of poor prognosis in CLL. The MiSeq SHM assay and traditional PCR-based clonality-Sanger sequencing assays demonstrated excellent analytical performance and good clinical concordance.