MYD88L265P Detection in IgM Monoclonal Gammopathies: Methodological Considerations for Routine Implementation

In IgM monoclonal gammopathies MYD88L265P is a prognostic and predictive biomarker of therapy response. MYD88L265P detection is mainly performed by allele-specific quantitative PCR (ASqPCR), however recently, droplet digital PCR (ddPCR) has been proved to be suitable for MYD88L265P screening and minimal residual disease monitoring (MRD). This study compared ASqPCR and ddPCR to define the most sensitive method for MYD88L265P detection in bone marrow (BM), peripheral blood (PB) sorted or unsorted CD19+ cells, and in plasma cell-free DNA (cfDNA). Overall, the analysis showed a good concordance rate (74%) between the two methods, especially in BM samples, while discordances (26%) were mostly in favor of ddPCR (ddPCR+ vs. ASqPCR-) and were particularly evident in samples with low mutational burden, such as PB and cfDNA. This study highlights ddPCR as a feasible approach for MYD88L265P detection across different specimen types (including cfDNA). Interestingly, its high sensitivity makes CD19+ selection dispensable. On the other hand, our results showed that MYD88L265P detection on PB samples, especially with ASqPCR, is suboptimal for screening and MRD analysis. Finally, significantly different MYD88L265P mutational levels observed between Waldenström Macroglobulinemia and IgM monoclonal gammopathy of undetermined significance patients suggest the need for further studies in order to identify possible correlations between mutational levels and risk of progression to Waldenström.


Introduction
IgM monoclonal gammopathies consist of a broad spectrum of diseases, ranging from apparently benign to malignant conditions, in which variable levels of serum IgM monoclonal proteins are detected [1,2]. The IgM monoclonal gammopathy of undetermined significance (IgM-MGUS) plays a relevant role as a pre-malignant condition, most notably for Waldenström Macroglobulinemia (WM), with a risk of progression of 10% within 5 years from diagnosis [3,4]. IgM-MGUS is defined by the absence of bone marrow (BM) infiltration by lymphoma cells, the presence of a serum monoclonal Immunoglobulin M (IgM) paraprotein lower than 3 g/dL and the absence of end-organ damage [4]. WM is characterized by the heterogeneous infiltration of monoclonal lymphocytes, lymphoplasmacytic cells, and plasma cells in BM, which are responsible for the production of IgM paraprotein in the serum [5,6].
Based on current criteria, the diagnosis of WM requires histological confirmation of monoclonal lymphoplasmacytic cells infiltration by BM aspirate and trephine biopsy: both flow cytometry (MFC) and immunohistochemistry (IC) are advisable in achieving a correct diagnosis [5][6][7][8].
Important progresses in understanding the biology of WM have been made by whole genome sequencing (WGS), that identified MYD88 L265P as a recurrent somatic mutation in patients with WM [9,10]. Gradually, several studies using different techniques, such as Sanger sequencing, polymerase chain reaction (PCR), and allele-specific PCR (ASqPCR), confirmed that MYD88 L265P is highly frequent in WM (up to 95% of patients) and IgM-MGUS (at least 50%) (Table S1), whereas it is absent in patients with multiple myeloma (MM) (including the IgM isotype) and rarely found in other indolent lymphoid disorders [11][12][13][14][15][16].
It has been shown that mutation status influences clinical presentation and outcome of WM. Indeed, MYD88 wild type (MYD88 WT ) patients are characterized by an inferior overall survival, as well as by higher levels of circulating B cells, positivity for CD23 surface marker and a spectrum of specific additional mutations [17,18]. Moreover, MYD88 WT is associated to high rates of progression to symptomatic WM, transformation to aggressive lymphoma, and inferior responses to Bruton's tyrosine kinase (BTK) inhibitors [17][18][19][20][21][22].
In spite of its diagnostic, prognostic, and predictive role, recognized by the European Society of Medical Oncology (ESMO) clinical guidelines [23], the detection of MYD88 L265P has not been standardized yet. A number of distinct assays and methods are available, characterized by different levels of application and sensitivity (Table S1); however, there is lack of consensus about the optimal specimens and techniques, both in terms of operating procedures, test sensitivity, and results interpretation [24][25][26]. Despite ASqPCR currently being the most widely used method, it suffers from some technical restraints. Due to its variable sensitivity (up to 1 × 10 −3 ), ASqPCR obtained superior results in highly infiltrated specimens, as BM aspirates (mainly enriched in tumor content by CD19+ cells selection, not routinely practiced in many diagnostic laboratories) [27], while did not perform concordantly well in unsorted BM or peripheral blood (PB) from WM and especially in IgM-MGUS samples (Table S1). Moreover, its suitability for assessing MRD analysis in low infiltrated samples or cell-free tumor DNA (cfDNA) from plasma or other biological compartments, such as cerebrospinal fluid or pleural effusions, need further investigation [28][29][30].
Recently, droplet digital PCR (ddPCR) emerged as a reliable and sensitive diagnostic tool that might overcome some of the ASqPCR limitations, with a higher sensitivity (up to 5 × 10 −5 ) [31][32][33]. Moreover, the emerging concept of "liquid biopsy" has been applied also to WM, as a non-invasive, patients friendly, potential alternative to BM aspiration; in fact, plasma has been shown superior to PB, that is hampered by high false negative rates, in particular in patients previously treated with B-cell depleting agents [33,34].
Due to the relevance of MYD88 L265P both in daily management of WM patients as well as in prospective clinical trial investigating the efficacy of novel agents, it is crucial to find an agreement among different diagnostic laboratories about the most sensitive, appliable, and standardizable molecular technique for mutation detection.
Thus, the aim of this study was to directly compare the performance of ASqPCR and ddPCR in different tissues and at distinct time points in order to identify the most suitable technique and the most useful specimen for MYD88 L265P detection, in order to establish recommendations on laboratory practice for WM and IgM-MGUS marker screening at diagnosis and MRD analysis.

Patients and Samples Collection
BM and PB samples were collected at baseline and during follow-up (FU) from 227 patients affected by WM and IgM-MGUS. Three patient series, from hematological Italian centers routinely involved in management of IgM monoclonal gammopathies (Torino, Pavia and Varese), were tested for MYD88 L265P mutation by both ASqPCR and ddPCR. All patients provided written informed consent for sample collection and analysis (Table S2).
Overall A group of 60 samples, including 40 healthy subjects and 20 MM patients were used as negative control.
For plasma recovery, PB samples, processed within 4 h from the drawing, were centrifuged at 1300× g for 13 min at room temperature. Plasma was transferred in separate tubes and centrifuged at 1800× g for 10 min at room temperature, before being stored at−80 • C in 1ml aliquots.
BM (previously filtered through a 1ml syringe), PB, and PB leftover (after plasma separation) were treated with erythrocytes lysis buffer (NH4Cl) (at 1:5, 1:2, and 1:2 dilution, respectively), left 15 min at room temperature (lying flat at dark), and centrifuged at 450× g for 10 min, at room temperature. The supernatant was discarded and the cell pellet was resuspended in 10-15 mL NH4Cl and centrifuged at 450× g for 10 min, at room temperature. The supernatant was discarded and the cell pellet was resuspended in 0.9% NaCl (q.s.), cells were counted, dispensed in 5-10 × 10 6 stocks, and stored as dry pellets at −80 • C, for further DNA extraction.
MNC B-cells were collected from BM and PB by lymphoprep standard density gradient centrifugation, while CD19+ cells were further isolated from MNCs by immunomagnetic adsorption on MiniMACS separation columns using an anti-CD19 antibody (Miltenyi Biotec GmbH), in accordance with the manufacturer recommendations. The purity of CD19+ separated cells was assessed by flow cytometry using anti-CD19 monoclonal antibodies (Becton Dickinson).

Nucleic Acid Extraction
Genomic DNA (gDNA) was extracted from WBC, MNC, and CD19+ cells by MaxWell RSC system with blood RSC kit (Promega) (Torino samples), by Puregene Blood DNA isolation kit (Qiagen) (Pavia samples), by Maxwell ® 16 LEV Blood DNA Kit (Promega, Madison, WI, USA) (Varese samples), in accordance with the manufacturer recommendations. cfDNA was extracted by Maxwell RSC with LV ccfDNA kit (Promega), in accordance with the manufacturer instructions.

ddPCR Assays for MYD88 L265P Detection
ddPCR mutation detection assay was performed, as previously described by Drandi et al., 2018 [33], using the QX200 Droplet Digital PCR System (Bio-Rad Laboratories, Hercules, CA, USA). Briefly, a single set of primers was combined with two competitive probes, in one assay for MYD88 L265P (FAM labeled) and one for MYD88 WT (HEX labeled) (CSTM DDPCR HEX/FAM ASSAY BIO-RAD). The cut-off for mutation was settled based on the highest MYD88 L265P level detected within the control group. Each experiment included a known highly mutated positive control sample (MUT/WT ratio 6.8 × 10 −1 ), a negative control (healthy donor or MM gDNA), and a no template control (NTC). Gate setting was performed based on the positive control results.
All relevant experimental details are reported following the updated digital MIQE2020 (Table S3) [35].

Statistical Analysis
For methods comparison, ddPCR results were expressed as MUT/WT ratio, while ASqPCR as allele frequency/100. To calculate the correlation and agreement between the methods, we evaluated the test-retest reliability for continuous variables by a singlemeasurement, consistent, 2-way mixed-effects model, Inter Class Correlation (ICC) analysis, with a 95% confident interval (CI) [36]. Correlation analyses and their representation plots were performed using IBM SPSS Statistics (version 25.0. Armonk, NY, USA: IBM Corp.) or GraphPad5 Software (GraphPad Software Inc., San Diego, CA, USA).
Since no difference in terms of mutation level was observed between WBC and MNC ( Figure S2c

Discussion
In this study ASqPCR and ddPCR were compared in order to define the abilities and aptitudes of each method for MYD88 L265P mutation detection and quantification in different tissue samples, such as BM, PB, and plasma-cfDNA, in sorted or unsorted CD19+ cells, collected from WM and IgM-MGUS patients routinely followed in clinical practice in three Italian hematological centers.
Overall, the analysis showed a good concordance rate (74%) between the two methods, especially in BM samples, while discordances (26%) were mostly in favor of ddPCR (ddPCR+ vs. ASqPCR-). Notably, ddPCR showed a higher sensitivity in detecting MYD88 L265P in samples with low mutational burden, such as PB, where median mutational levels were lower than in BM ( Figure S1).
Moreover, by selecting CD19+ cells no differences in frequency of patients carrying the MYD88 L265P were observed in comparison to WBC or MNC, even if the mutation level was of course different. As expected, both BM and PB CD19+ cells showed a higher mutational burden compared to unselected MNC/WBC samples. Thus, despite enriching the sample in tumor cells content, cell selection is dispensable for mutational screening by ddPCR, due to its superior mutation detection ability in unsorted samples, compared to ASqPCR.
The concept of the "high sensitivity" of ddPCR is relevant in the context of treatment tailoring based on patients' genotype. Recent studies evaluated the efficacy and safety of novel BTK inhibitors (i.e., acalabrutinib, zanubrutinib) on MYD88 WT WM patients, so defined by ASqPCR in unsorted BM aspirates, with a declared limit of detection of 0.5% (5 × 10 −3 ). Scientists must be aware that the misclassification of patients with a mutation level below the sensitivity of the employed PCR method may lead to an overestimation of the efficacy of novel treatments. As a consequence, the results of the direct comparison between ASqPCR and ddPCR in unselected samples underline how ddPCR could benefit patients stratification for treatment assessment [37,38].
This study also highlights a significantly different mutation level between WM and IgM-MGUS. Interestingly, IgM-MGUS showed a median mutation level almost one logarithm lower than WM. This difference is statistically significant and draws attention to the need for further studies on larger patients' series to identify possible correlations between mutational levels and risk of progression to WM.
Finally, the analysis of 64 plasma-cfDNA confirmed the role of ddPCR for mutational screening and MRD monitoring on less-invasive tissue sources, representing an attractive alternative to BM biopsy particularly in asymptomatic patients, such as those affected by IgM-MGUS. Actually, this study describes the sensitivity of ddPCR for MYD88 L265P mutation detection in a group of samples collected in clinical routine. Of note, a multicenter clinical trial for non-invasive diagnostics and monitoring of MRD in WM and in IgM-MGUS patients is currently ongoing, in the context of the Fondazione Italiana Linfomi (BIO-WM trial: NCT03521516), with the primary endpoint of demonstrating that the MYD88 L265P mutation rate detected in plasma by ddPCR is superimposable to the rate detected in BM.

Conclusions
ddPCR is a feasible approach for MYD88 L265P detection, more sensitive than ASqPCR across different specimen types (including plasma-cfDNA) and in distinct diseases. MYD88 L265P detection on PB samples, especially with ASqPCR, is suboptimal for MYD88 L265P screening and MRD analysis. CD19+ selection, despite enriching the sample tumor content, is dispensable for mutational screening, suggesting that the implementation of ddPCR assay in routine diagnostic laboratories might avoid the need of CD19+ selection. There is a need for further studies on larger patients' series, in order to identify possible correlations between mutational levels and risk of progression to WM. Based on the above conclusions we finally propose an algorithm for the best use of the most convenient PCR methods for MYD88 L265P detection, based on the available type of specimens (Table 1) in order to establish a uniform testing approach for determining the MYD88 mutational status. These recommendations, if followed, will improve routine clinical practice within different laboratories with the intention to standardize protocols and procedures for the management of both clinical routine and multicenter clinical trials.

Conflicts of Interest:
The authors declare no conflict of interest.