Unbiased NGS Approach for IGHV Mutational Status, Clonal Space and Immune Repertoire Assessment in Chronic Lymphocytic Leukemia

Introduction: Chronic lymphocytic leukemia (CLL) is the most common leukemia in western countries. The mutational load of HV immunoglobulin genes (IGHV) is one of the most relevant prognostic markers, a sequence homology of IGHV to its germline ≥98% (unmutated UM-CLL) is indicative of a worse prognosis, in turn cases with mutated IGHV (M-CLL) show a better outcome. The current gold standard method for evaluating the mutational status is based on the BIOMED-2 protocol followed by Sanger sequencing. This strategy, however, does not provide insight into subclonal architecture and intraclonal diversity, it might require laborious cloning techniques in cases with biallelic rearrangements, and can be subject to methodological biases inherent in the use of multiplex PCR. ARTISAN-PCR followed by long-read sequencing allows the unbiased amplification of the full IGHV repertoire present in a sample. The combination of long-read single-molecule sequencing capability, low upfront investment, and reduced maintenance, makes Oxford Nanopore sequencing technology (ONT) an ideal candidate for facilitating testing accessibility.

Methods: Subjects included 51 CLL patients and 12 healthy donors to obtain controls B-cell repertoires. BIOMED-2 PCR was performed according to the recommendation of the ERIC group. ARTISAN PCR was performed as earlier described by our group (Koning et al.). BIOMED-2 amplicons were sequenced by Sanger and ONT. ARTISAN PCR amplicons were sequenced by ONT. Internal NGS validation was performed by comparing ONT vs PacBio sequencing. A bioinformatic pipeline was developed for mutational status, clonal space and non-clonal B-cell repertoire characterization.

Results: We performed a pre-validation step in 10 samples by comparing BIOMED-2 amplicons subject to Sanger sequencing vs ARTISAN amplicons measured by ONT sequencing. There was a 100% agreement in VDJ identification and mutational status classification. The VH genes coverage was significantly higher with Nanopore vs Sanger sequencing (mean coverage= 99.3% vs 88.4%) with an error rate of 2.5 variants/kb.

We then compared the ARTISAN-ONT performance with BIOMED-2 Sanger in an expansion cohort. ARTISAN-ONT performed with an equivalent mutational status classification in 96% of the cases (49/51). The correlation of IGHV homology was r = 0.85, p <0.0001. The sensitivity of ARTISAN-Nanopore was 91% and the specificity was 100%. In the discordant cases both methods lead to the identification of distinct clonal IGHV rearrangements. By ONT sequencing these cases had a low read count that could account for this discrepancy.

Since ARTISAN-Nanopore allows the characterization of the full B-cell repertoire, we further analyzed its characteristics contrasting with the repertoire of healthy donors.

ARTISAN-Nanopore allowed the unequivocal detection of overexpressed clonal reads in all cases, healthy donors showed no evidence of clonal expansion. The Clonal Space (CS, i.e., proportion of clonal reads/non-clonal reads) was 0.66 for M-CLL and 0.71 for UM-CLL. One sample showed bi-clonality with two overexpressed clonal rearrangements.

Healthy donors showed an average repertoire richness of 1178 vs 2806 in M-CLL and 2817 in UM-CLL as assessed by ⁰D Hill Index. CLL showed a significant lower number of common rearrangements: 63 in UM-CLL, 31 in M-CLL as compared with 1076 in healthy donors. On the other hand, CLL cases showed higher mean aminoacidic charge and larger CDR3 mean length on non-clonal B-cells (charge 0.75 in M-CLL, 0.58 in UM-CLL vs 0.15 in healthy donors; CDR3 length 18.2 aa in M-CLL, 19.55 in UM-CLL and 15.8 in healthy donors). This data indicates disturbances in the diversity and composition of the remaining non-clonal B-cell repertoire in CLL.

Conclusion: Here we present an amplicon NGS-based strategy to determine the IGHV mutational status and the size of the occupied clonal space by the tumor clone, which also enables the characterization of the non-clonal B-cell repertoire in CLL patients. The strategy it fast, reliable, shows high specificity and sensitivity. This method relies on ONT sequencing, thus implies a relatively low investment being suitable for small laboratories around the world. This new methodology represents a cost-effective approach that should facilitate molecular testing availability in CLL.