Tumor cell-free DNA detection in CSF for primary CNS lymphoma diagnosis


 
 
 Primary CNS lymphoma (PCNSL) is a rare disease accounting for around 3% of the primary brain tumors. In the vast majority of cases, PCNSL are classified as diffuse large B-cell lymphoma according to the WHO classification. The diagnosis is based on cranial MRI, in combination with a brain biopsy. In case of classical MRI findings, the identification of lymphoma cells in the cerebrospinal (CSF) or vitreous fluid by cytology and flow cytometry might obviate brain biopsy. The presence of the somatic mutation p.Leu265Pro (L265P) in MYD88 is detectable in 50 to 80% of PCNSL, and might also be helpful to confirm the diagnosis. The aim of our study was to evaluate the contribution of a highly sensitive digital droplet PCR, targeting the mutation L265P MYD88, for the detection of tumoral circulating DNA from CSF supernatant.
 
 
 
 We identified 9 PCNSL expressing the L265P mutation at diagnosis or relapse. The mutation was found by an allele specific PCR technique either on biopsy or in CSF cells. Circulating DNA was isolated from CSF supernatant with the « QiaAmp® circulating nucleic acid» kit. The quantity of DNA collected was estimated by quantitative PCR for a reference gene (albumine) with 7900HT (Life technologies™) device. Subsequently, the L265P MYD88 mutation was quantified by digital droplet PCR Biorad™: The droplets generated were amplified by PCR, detected with the QX200 Reader, and analyzed with the QuantaSoft™ software.
 
 
 
 The circulating DNA concentration was low, varying between 0 and 2.2 ng/mL of CSF. However, the mutation was detected in the circulating DNA from CSF supernatant in 6 out of 9 cases (66%). The fractional abundance varied from 2.6 to 85%. In 3 cases, the mutation was detected even though cytology and flow cytometry did not reveal leptomeningeal disease. For 3 other cases, the mutation was not detected: The genome copy number was below 1 copy/µL, indicating a low analytical sensitivity for theses samples.
 
 
 
 This study shows that circulating DNA is present in low concentration in CSF and can be amplified by a sensitive digital PCR for the L265P MYD88 mutation. The detection of circulating PCNSL DNA in CSF is possible and might be used to improve the non-invasive diagnosis of PCNSL. It might also help to select patients for targeted therapies.


Tumor cell-free DNA detection in CSF for primary CNS lymphoma diagnosis To the editor, Primary central nervous system lymphoma (PCNSL) is a rare disease accounting for around 3% of primary CNS tumors. Its diagnosis is usually based on cranial MRI and brain biopsy (including immunophenotyping for faster diagnostic confirmation [4]). The identification of lymphoma cells in the cerebrospinal fluid (CSF) or vitreous fluid by cytology (generally associated with flow cytometry) in association with typical neuroimaging allow faster and less invasive diagnosis. PCNSL characterization, frequently leads to diagnosis of diffuse large B-cell lymphoma (DLBCL), belonging to the ABC subgroup [3]. Moreover, MYD88 mutations are detectable in 58 to 76% of PCNSL cases, in about 30% of ABC DLBCL patients, and in the majority of lymphoplasmacytic lymphoma cases [6,7,11,13]. Since this mutation is not described in glioblastoma or in other solid metastatic tumors, its detection in the cerebrospinal fluid (CSF) [10] could be helpful for PCNSL diagnosis without invasive surgical biopsies, such as IL10 concentration [12] and micro-RNA profiling [1]. The MYD88 L265P mutation detection in cell DNA from vitreous aspirates [2] and CSF [10] was reported to improve the PCNSL diagnosis. The aim of our study was to evaluate the contribution of cell-free (cf) DNA from the CSF with a valuable molecular tool detecting the tumor-specific mutation MYD88 L265P , using ddPCR in known MYD88 L265P PCNSL.
This retrospective study was conducted between August 2016 and June 2018 on a series of 11 MYD88 L265P PCNSL patients without ocular infiltration. The MYD88 mutation status was established either on brain biopsy (n = 7) or in cell DNA from CSF (n = 4) with an allele specific (AS) PCR technique. CSF samples at initial diagnosis (n = 9) or relapse (n = 5) were processed within 4 h after lumbar puncture. After CSF centrifugation ( Fig. 1), the cell pellet and the previously discarded supernatant (1.5-5 mL) were collected for cfDNA isolation and ddPCR for the detection of the NM_002468:exon5:c.T778C(p.Pro265Leu) MYD88 variant. Sensitivity thresholds were established by a dilution study with the lower limit of quantification and detection found to be 0.9 copy/μL and 0.2 copy/μL, respectively. The specificity was evinced by the absence of L265P-positive droplets in 10 CSF samples from nonlymphomatous lesions.
The presence of cfDNA was detected in PCNSL CSF with a median value of 3.1 cfDNA copies/μL ddPCR mix (Fig. 1). Substantial variations of the amount of cfDNA were observed and four cases exhibited less than 1 copy/ μL ddPCR mix, even though special care was given to the parameters affecting the quantity and quality of cfDNA, such as pretreatment delay, sufficient CSF volume, DNA isolation process and storage. The MYD88 L265P mutation was detectable in 10 out of 14 cell-free CSF samples, and not in the four cases with less than 1 copy wild-type MYD88/μL. In these samples, MYD88 L265P was detected in the CSF cell DNA using ddPCR only (#10R) or AS PCR (#09D). The MYD88 L265P detection rate in CSF combining both CSF fractions achieved 86% (12/14 cases). Two cases, at relapse, remained negative for MYD88 L265P detection in CSF, most probably due to a low cfDNA input or possible clonal evolution. The median fractional abundance (FA) was 7%, varying from 2.6 to 92.9%. FA was higher than in previous studies using plasma [5,9], probably because CSF directly bathes the brain tumor, without background hematopoietic DNA retained by the blood-brain barrier. Furthermore, mutated cfDNA FA was higher than in the cell pellet DNA in six out of seven available samples. Moreover, in three cases (#01D; #03D; #07R), the L265P variant could only be detected in the cell-free fraction. Finally, cfMYD88 L265P was present in the absence of lymphoma cells using cytology and flow cytometry (FCM) in three cases at diagnosis (#01D; #02D; #03D), as it was described in recurrent/refractory CNS * Correspondence: rimelenv@ghrmsa.fr † V. Rimelen, G. Ahle and J. Voirin contributed equally to this work. 1 Laboratoire d'Hématologie-Groupe Hospitalier de la Région de Mulhouse Sud-Alsace, Hôpital E. Muller, 20 avenue de Dr Laennec, 68100 Mulhouse, France Full list of author information is available at the end of the article lymphoma [8]. Even if hot spot mutation is predominant in PCNSL, our cost-effective, highly sensitive ddPCR approach is limited to a restricted number of mutations and will miss PCNSL bearing other mutations. This is the first report comparing cell and cell-free tumor load in CSF from PCNSL, showing the contribution of cell-free tumor detection in CSF for diagnosis. This study shows that detection of tumor cell and cell-free DNA is feasible using a workflow combining FCM and molecular biology. Moreover, ddPCR could be used for the tumor characterization of actionable mutations and longitudinal monitoring of the disease. We anticipate that this technique might also be applicable to other brain tumors with known hotspot mutations.

Funding
Not applicable.

Availability of data and materials
Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.
Authors' contributions BD, GA, AD and VR designed the study. GA, FL and JV procured clinical specimens. RH, IA provided CSF samples. VR and NZ performed the ddPCR experiments. IH, AD, LZ performed cytology and FCM analysis. VR analyzed the data and prepared the original manuscript. GA, BD, EP, AD provided analytical advice, data interpretation, and assisted with manuscript preparation. All authors read and approved the final manuscript. VR, AG and BD contributed equally to this work.
Ethics approval and consent to participate All patients had given written informed consent. An institutional review board had approved the study, which was conducted in accordance with the declaration of Helsinki.

Consent for publication
Not applicable.

Competing interests
The authors declare that they have no competing interests.

Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.