Plasma Cell Specific cfDNA Methylation Patterns Differentiate MGUS, SMM and MM and Predict Biochemical and Clinical Progression to MM

Background: Predicting progression of Monoclonal Gammopathy of Undetermined Significance (MGUS) and Smoldering Multiple Myeloma (SMM) to Multiple Myeloma (MM) involves bone marrow (BM) biopsies and long term follow up of patients, and the performance of current risk assessment models is limited.

Cell turnover results in the release of cell-free DNA (cfDNA) fragments to blood. The tissue origins of such fragments can be determined using cell type-specific methylation patterns.

We hypothesized that plasma cell (PC)-specific DNA methylation markers and cancer-specific methylation alterations in cfDNA can serve as a minimally-invasive biomarker for diagnosis and surveillance of MGUS, SMM and MM, and may prospectively predict progression dynamics to MM.

Methods: We isolated PC from bone marrow samples of MM patients and healthy individuals, and determined their methylome using deep whole-genome bisulfite sequencing yielded genomic loci that are specifically demethylated in PCs (7 loci), or show disordered patterns of methylation in MM (10 loci). We developed a multiplex PCR-sequencing assay to determine the methylation status of these loci in cfDNA, and analyzed plasma samples from patients with MGUS (n=58), SMM (n=29) and MM (n=37). Results were also correlated with a prospective assessment for biochemical and clinical progression to MM.

Results: cfDNA methylation markers differentiated PC disease states. MM patients had increased PC-cfDNA levels and cancer-specific methylation alterations compared with SMM (AUC=0.81, P=0.0001) MGUS (AUC=0.9, P=0.0001) and healthy controls (AUC=0.94, P=0.0001). PC-cfDNA levels correlated (Pearson-R=0.68-0.86, p=0.0001) with BMPC %, and aberrant PC%. Furthermore, at a median follow up of 16.5 (5-32) months, the assay could significantly predict the progression of premalignant PCD. Patients with MGUS and SMM that had higher levels of PC-cfDNA methylation and cancer-specific methylation alterations had an overall faster biochemical and clinical progression. Utilizing these biomarkers alone had a negative predictive value of 80% for biochemical and 100% for clinical progression within 2 years. The positive predictive value increased upon the replacement of BMPC-20% counts with PC-cfDNA or cancer-specific methylation alterations combined with the 2gr/dL and FLC-Ratio>20 risk stratification model [HR 16-26.5 (CI 3.24-214.6) p<0.0001]. Finally, these results were cross validated with a machine learning approach, indicating that locally disordered methylation patterns in specific loci are the most significant among all laboratory and clinical markers for progression.

Conclusion: cfDNA methylation patterns are a promising biomarker for diagnosis and surveillance of plasma cell disorders, and for identifying patients at risk of progressing to MM. With further development, methylation-based liquid biopsies may offer a non-invasive, practical and essential tool to evaluate individual patient's risk for progression, on their initial MGUS or SMM diagnosis day.