Whole-Exome Sequencing of Waldenström Macroglobulinemia Transformation into Aggressive Lymphoma

INTRODUCTION

Waldenström Macroglobulinemia (WM) has an indolent clinical course. However, as in other indolent lymphoproliferative disorders, transformation into aggressive histologies can occur in up to 10% of WM patients.

Recent progress about WM mutational profile has allowed the identification of recurrent mutations in the MYD88, CXCR4 and ARID1A genes.

In turn, mechanisms involved in the transformation to aggressive lymphoma have not been yet described. These findings might be of great interest in the search of new therapeutic strategies for these patients whose prognosis remains extremely poor. Also, the integration of WM transformation genomics with available data regarding transformation in other indolent lymphomas would help to understand the whole process of transformation, and try to define biologic risk criteria, investigate new therapies, and develop preventive strategies in order to improve the outcome of our patients.

Here we describe the major mutations found at the time of diagnosis and transformation, those that were exclusively found at transformation, and the evolutionary pattern observed in four transformed WM by using whole-exome sequencing.

PATIENTS AND METHODS

Patients

Four patients diagnosed with WM and transformed into diffuse large B-cell lymphoma (DLBCL) were included. Tumor DNA samples at diagnosis and transformation, and germline DNA were available in three out the four patients. The diagnostic sample of the fourth patient was unavailable due to bad quality. We added an extra WM relapse sample prior to transformation in one patient. In all cases, sample tumor infiltration by flow cytometry (FCM) was available.

DNA extraction, quantification and quality control

DNA was isolated using DNAzol reagent or Maxwell® 16 System. DNA quality controls prior to enrichment and generation of libraries were performed using fluorometry for DNA quantification, and gel electrophoresis for quality evaluation. Enrichment and generation of libraries was performed using Agilent SureSelect XT2 kit V5. Paired-end sequencingwas carried out using the Illumina HiSeq 2000 platform. The number of reads was set up according to each sample tumor infiltration defined by FCM (depth of coverage range: 150x-250x).

Bioinformatics analysis

DreamGenics® (Oviedo, Spain) supervised the biological procedure and performed bioinformatics analysis. Algorithms and non-commercial pipelines were used to call variants, analyze and compare them. Since our disposal of at least two samples in the four patients, we performed a pair-wise comparative analysis. Since we knew the selected tumor samples' infiltration, we could estimate the percentage of corresponding tumor cells in each sample. MYD88 mutations (known to be constantly mutated in WM) were used to normalize the frequency of the variants.

RESULTS

The most frequently mutated genes at diagnosis were constantly present along disease evolution, as it was MYD88, whose role in WM is well known. Mutations in CD79B were present in two of three patients (67%), and three of four at transformation (75%), a frequency much higher than it should be expected in conventional WM (10%). Mutations in this gene affected the first tyrosine ITAM kinase domain of the receptor, as it has been frequently described in DLBCL.

At transformation, the most frequently acquired mutations affected TP53, CARD11, PIM1 and KMT2D (MLL2). It should be noted that KMT2D was found mutated in all patients at transformation.

A linear evolution was observed in two out of three patients, with cumulative acquisition of new mutations at transformation and maintenance of those detected at diagnosis. However, the patient who had also a relapse sample demonstrated a "branching" model of evolution, sharing variants in all tumor samples (diagnosis, relapse, and transformation), while other variants were only detected at relapse.

CONCLUSIONS

Our study suggests that some genes could be more frequently mutated in this specific subset of patients (i.e. CD79B), which could be considered as potential biomarkers for predicting the risk of transformation in WM patients. In addition, different models of clonal evolution can be found in transformed WM (linear and branching).