Disease Monitoring in Multiple Myeloma Patients Using Liquid Biopsy and Next Generation Sequencing of IGH Gene Rearrangements

Background:

Current criteria used to assess response lag behind the extraordinary evolution in the treatment of multiple myeloma (MM) patients (pts), and more sensitive techniques are being explored to detect true minimal residual disease (MRD) for new complete remission (CR) definitions. In recent years, next generation sequencing (NGS) technologies have emerged. NGS of immunoglobulin (IgH) gene rearrangements are very sensitive and also allow the identification of small subclonal population that can be monitored over time during treatment, something not possible with flow cytometry or PCR. However, the patchy pattern of bone marrow infiltration observed in MM leads to some degree of uncertainty regarding MRD-negative results, irrespectively of the technique adopted. This highlights the value of applying "liquid biopsy" as a non-invasive strategy for monitoring MRD through the analysis of circulating cell-free tumor DNA (ctDNA). The objective of the current study was to measure residual tumor burden in sequential plasma samples of a cohort of MM pts by NGS of the IgH gene rearrangements.

Methods:

We retrospectively analyzed 14 MM pts homogeneously treated between 2011 and 2015 with all clinical data available. We obtained serial tumor and plasma samples at diagnosis and at specified time points during treatment cycles and up to 24 months of follow-up. Genomic DNA (gDNA) was extracted from immunomagnetically selected CD138+ plasma cells at diagnosis (n=14). ctDNA was extracted from 500uL of plasma (Qiagen) at diagnosis (n=14) and at follow-up time points (n=58). IgH gene rearrangements were amplified, quality assessed (Agilent hsDNA kit) and sequenced on Ion Torrent PGM as previously described (Gimondi et al., ASH 2015). Raw reads were filtered for quality, length (>250bp) and presence of both forward and reverse primers. Reads were subsequently aligned using IgBlast against IMGT germline database and aggregated into clonotypes based on identity of CDR3, V and J segments (MigMap). Post-processing analyses were performed using VDJtools and customized R scripts.

Results:

PCR products quality assessment from ctDNA amplification of the entire IgH-VDJ region revealed the presence of both short (150-250bp) and long amplicons (310-360bp). Raw reads were subjected to filtering using our custom bioinformatic workflow to retain only complete IgH-VDJ gene rearrangements and discard low-quality reads. Three pts could not be evaluated due to low quality sequencing reads in all samples. At least 3 follow-up time points were available for all the remaining 11 pts whereas 6 pts had 4 time points. At diagnosis, both plasma and tumor samples revealed a high level of heterogeneity (range 1980-7753 clonotypes) with only a small fraction of shared clonotypes (346±262, mean±SD). Among the shared ones, the clonotype with the highest frequency in plasma corresponded to the tumor-associated one identified in CD138+ cells. Interestingly, in the plasma of 3 pts, additional clonotypes were detected at relatively high frequencies (range 1-16%) suggesting the presence of subclones.

IgH-NGS at follow-up time points revealed that the clonotype identified at diagnosis (range 4-31% of total reads) could be easily tracked over time in plasma samples, at frequencies as low as 0.00001%. Frequencies of the tumor-associated IgH gene rearrangement in plasma showed a patient-specific modulation and reflected the tumor burden assessed according to the International Myeloma Working Group-Uniform Response Criteria. At the time of CR, the tumor-associated clonotype was undetectable in the plasma of pts who would not subsequently relapse. In patients that would lately experience progressive disease, the tumor specific clonotype was still detectable at low frequencies (range 0.00001-0.03%) in all plasma samples suggesting that liquid biopsy can be used for MRD monitoring.

Conclusions:

Despite the limited number of pts and follow-up samples analyzed, we demonstrate that NGS of IgH gene rearrangements from ctDNA can be used for MM disease monitoring, thus representing a non-invasive alternative strategy for clinical management. The analysis of retrospectively collected plasma samples revealed that ctDNA quality is essential for a NGS characterization of IgH gene rearrangements. Plasma samples collection and processing represent critical steps that need to be considered designing prospective liquid biopsy studies.