Analysis of Driver Mutational Hot Spots in Blood-Derived Cell-Free DNA of Patients with Primary Central Nervous System Lymphoma Obtained before Intracerebral Biopsy
2020; Elsevier BV; Volume: 22; Issue: 10 Linguagem: Inglês
10.1016/j.jmoldx.2020.07.002
ISSN1943-7811
AutoresManuel Montesinos‐Rongen, Anna Brunn, Armin Tuchscherer, Peter Borchmann, Elisabeth Schorb, Benjamin Kasenda, Janine Altmüller, Gerald Illerhaus, Maximilian I. Ruge, Mohammad Maarouf, Reinhard Büttner, Martin‐Leo Hansmann, Michael Hallek, Marco Prinz, Reiner Siebert, Martina Deckert,
Tópico(s)Acute Myeloid Leukemia Research
ResumoIn newly diagnosed systemic diffuse large B-cell lymphoma, next-generation sequencing of plasma-derived cell-free DNA (cfDNA) detects somatic mutations as accurate as genotyping of the tumor biopsy. A distinct diffuse large B-cell lymphoma entity confined to the central nervous system is primary central nervous system lymphoma (PCNSL), which requires intracerebral biopsy and neuropathologic analysis to establish the diagnosis. So far, a biomarker for diagnosis and follow-up of PCNSL that can be investigated in blood has not been identified. This article addresses the question whether somatic mutations of the CD79B and MYD88 driver genes of PCNSL can be detected in cfDNA at disease diagnosis. Stereotactic biopsies and cfDNA of 27 PCNSL patients were analyzed for CD79B and MYD88 mutations. As control, cfDNA derived from six healthy volunteers was used. CD79B and MYD88 hot spot mutations were identified in 16 of 27 (59%) and 23 of 27 (85%) PCNSL biopsies, respectively, but only in 0 of 27 (0%) and 1 of 27 (4%) corresponding cfDNA samples, respectively. In cfDNA of one of four patients with Waldenstrom disease, as a further control, the MYD88 L265P mutation was readily detected, despite complete clinical remission. These data suggest that in PCNSL even if they carry such mutations, alterations of CD79B and MYD88 cannot be reliably detected in blood-derived cfDNA obtained before intracerebral biopsy. In newly diagnosed systemic diffuse large B-cell lymphoma, next-generation sequencing of plasma-derived cell-free DNA (cfDNA) detects somatic mutations as accurate as genotyping of the tumor biopsy. A distinct diffuse large B-cell lymphoma entity confined to the central nervous system is primary central nervous system lymphoma (PCNSL), which requires intracerebral biopsy and neuropathologic analysis to establish the diagnosis. So far, a biomarker for diagnosis and follow-up of PCNSL that can be investigated in blood has not been identified. This article addresses the question whether somatic mutations of the CD79B and MYD88 driver genes of PCNSL can be detected in cfDNA at disease diagnosis. Stereotactic biopsies and cfDNA of 27 PCNSL patients were analyzed for CD79B and MYD88 mutations. As control, cfDNA derived from six healthy volunteers was used. CD79B and MYD88 hot spot mutations were identified in 16 of 27 (59%) and 23 of 27 (85%) PCNSL biopsies, respectively, but only in 0 of 27 (0%) and 1 of 27 (4%) corresponding cfDNA samples, respectively. In cfDNA of one of four patients with Waldenstrom disease, as a further control, the MYD88 L265P mutation was readily detected, despite complete clinical remission. These data suggest that in PCNSL even if they carry such mutations, alterations of CD79B and MYD88 cannot be reliably detected in blood-derived cfDNA obtained before intracerebral biopsy. The term liquid biopsy has recently been coined to describe minimally invasive tools for the detection of tumor-derived material in blood or other body fluids. Such derivates include intact circulating tumor cells, or cell-free components. Liquid biopsy on cell-free DNA (cfDNA) has recently shown to be applicable for early detection and monitoring of various tumor entities.1Corcoran R.B. Chabner B.A. Application of cell-free DNA analysis to cancer treatment.N Engl J Med. 2018; 379: 1754-1765Crossref PubMed Scopus (413) Google Scholar In newly diagnosed diffuse large B-cell lymphoma (DLBCL), plasma cfDNA genotyping using next-generation sequencing (NGS) detects somatic mutations as accurate as genotyping of the tumor biopsy.2Rossi D. Diop F. Spaccarotella E. Monti S. Zanni M. Rasi S. Deambrogi C. Spina V. Bruscaggin A. Favini C. Serra R. Ramponi A. Boldorini R. Foa R. Gaidano G. Diffuse large B-cell lymphoma genotyping on the liquid biopsy.Blood. 2017; 129: 1947-1957Crossref PubMed Scopus (144) Google Scholar This putatively opens the possibility also of plasma cfDNA sequencing in diagnosing and/or monitoring primary central nervous system (CNS) lymphoma (PCNSL), a distinct DLBCL entity confined to the CNS with unique molecular and clinical features.3Deckert M. Paulus W. Kluin P. Ferry J. Lymphomas.in: Louis D.N. Ohgaki H. Wiestler O.D. Cavenee W.K. WHO Classification of Tumours of the Central Nervous System. Revised 4th ed. IARC, Lyon, France2016: 272-277Google Scholar,4Kluin P. Deckert M. Ferry J.A. Primary diffuse large B-cell lymphoma of the CNS.in: Swerdlow S.H. Campo E. Harris N.L. Jaffe E.S. Pileri S.A. Stein H. Thiele J. WHO Classification of Tumours of Haematopoietic and Lymphoid Tissues. Revised 4th ed. IARC, Lyon, France2017: 300-302Google Scholar PCNSLs are histogenetically derived from germinal center B cells; however, they differ from systemic DLBCL by their less favorable prognosis.5Deckert M. Engert A. Brück W. Ferreri A.J. Finke J. Illerhaus G. Klapper W. Korfel A. Küppers R. Maarouf M. Montesinos-Rongen M. Paulus W. Schlegel U. Lassmann H. Wiestler O.D. Siebert R. DeAngelis L.M. Modern concepts in the biology, diagnosis, differential diagnosis and treatment of primary central nervous system lymphoma.Leukemia. 2011; 25: 1797-1807Crossref PubMed Scopus (142) Google Scholar, 6Montesinos-Rongen M. Brunn A. Bentink S. Basso K. Lim W.K. Klapper W. Schaller C. Reifenberger G. Rubenstein J. Wiestler O.D. Spang R. Dalla-Favera R. Siebert R. Deckert M. Gene expression profiling suggests primary central nervous system lymphomas to be derived from a late germinal center B cell.Leukemia. 2008; 22: 400-405Crossref PubMed Scopus (136) Google Scholar, 7Montesinos-Rongen M. Küppers R. Schlüter D. Spieker T. Van Roost D. Schaller C. Reifenberger G. Wiestler O.D. Deckert-Schlüter M. Primary central nervous system lymphomas are derived from germinal-center B cells and show a preferential usage of the V4-34 gene segment.Am J Pathol. 1999; 155: 2077-2086Abstract Full Text Full Text PDF PubMed Scopus (152) Google Scholar, 8Montesinos-Rongen M. Siebert R. Deckert M. Primary lymphoma of the central nervous system: just DLBCL or not?.Blood. 2009; 113: 7-10Crossref PubMed Scopus (59) Google Scholar In contrast to the vast majority of other sites in the body where DLBCL may manifest, the blood-brain barrier (BBB) restricts exit of cells and soluble factors from the CNS. Thus, it is an intriguing question whether NGS of cfDNA is also a suitable approach in PCNSL, particularly if applied before disruption of the BBB by surgical measures. To study the clinical value of cfDNA sequencing in PCNSL, the current study took advantage of the high frequency of CD79B (>20%) and MYD88 (>50%) mutations in PCNSL.9Montesinos-Rongen M. Godlewska E. Brunn A. Wiestler O.D. Siebert R. Deckert M. Activating L265P mutations of the MYD88 gene are common in primary central nervous system lymphoma.Acta Neuropathol. 2011; 122: 791-792Crossref PubMed Scopus (115) Google Scholar,10Montesinos-Rongen M. Schäfer E. Siebert R. Deckert M. Genes regulating the B cell receptor pathway are recurrently mutated in primary central nervous system lymphoma.Acta Neuropathol. 2012; 124: 905-906Crossref PubMed Scopus (46) Google Scholar This research demonstrates that cfDNA samples from healthy controls and PCNSL patients investigated before diagnostic biopsy do not differ significantly in their mutation frequency of both CD79B and MYD88, irrespective of the presence of CD79B and MYD88 mutations in the PCNSL frozen tissue obtained by subsequent biopsy. Moreover, the extensive control analyses suggest that low-level CD79B and MYD88 mutations might exist in cfDNA in healthy individuals. Stereotactic biopsies of 27 HIV-negative patients with PCNSL were included in this study. The diagnosis of PCNSL was based on a combination of standard neuropathology and immunohistochemistry, according to the World Health Organization classification of tumors of the CNS.4Kluin P. Deckert M. Ferry J.A. Primary diffuse large B-cell lymphoma of the CNS.in: Swerdlow S.H. Campo E. Harris N.L. Jaffe E.S. Pileri S.A. Stein H. Thiele J. WHO Classification of Tumours of Haematopoietic and Lymphoid Tissues. Revised 4th ed. IARC, Lyon, France2017: 300-302Google Scholar These included staining for hematoxylin and eosin, CD20, CD10, BCL6, MUM1, BCL2, MYC, and MIB1, as described.11Brunn A. Nagel I. Montesinos-Rongen M. Klapper W. Vater I. Paulus W. Hans V. Blümcke I. Weis J. Siebert R. Deckert M. Frequent triple-hit expression of MYC, BCL2, and BCL6 in primary lymphoma of the central nervous system and absence of a favorable MYC(low)BCL2 (low) subgroup may underlie the inferior prognosis as compared to systemic diffuse large B cell lymphomas.Acta Neuropathol. 2013; 126: 603-605Crossref PubMed Scopus (49) Google Scholar Samples were snap frozen immediately after neurosurgical removal and stored at −80°C until use for molecular biological analysis. Blood samples were collected immediately preceding surgery to obtain a tumor sample to establish PCNSL diagnosis. In addition to PCNSL patients, blood samples from six healthy controls and from four patients in clinically complete remission after treatment for Waldenstrom disease were included. For patients' details, see Table 1. The study was approved by the local Ethics Commissions (06-187 and 07-109) and performed in accordance with the Declaration of Helsinki.Table 1Patient DetailsGroupNFMMean∗In years.Median∗In years.Minimum∗In years.Maximum∗In years.PCNSL2791971.075.043.086.0HC64239.241.521.047.0WD43165.863.559.077.0F, female; M, male; HC, healthy control; PCNSL, primary central nervous system lymphoma; WD, Waldenstrom disease.∗ In years. Open table in a new tab F, female; M, male; HC, healthy control; PCNSL, primary central nervous system lymphoma; WD, Waldenstrom disease. High-molecular-weight DNA was extracted from frozen tissue using the Gentra Puregene DNA Purification Kit (Qiagen, Hilden, Germany), according to the manufacturer's instructions. cfDNA was extracted from fresh EDTA-blood samples (<2 hours after blood collection) with the QIAamp Circulating Nucleic Acid Kit (Qiagen). Significant contamination of the cfDNA samples with cell-derived DNA was excluded by TapeStation analysis (Supplemental Figure S1). DNA derived from the PCNSL biopsies, and from corresponding plasma-derived cfDNA of the 27 PCNSL patients, six plasma-derived cfDNAs from healthy controls, and four plasma-derived cfDNAs from patients with Waldenstrom disease in complete clinical remission were subjected to PCR for hot spot mutations in CD79B [CD79B_F: 5′-CTCTGATCTCCATCCCTCTCC-3′; CD79B_R: 5′-TGGGCCTGCCCCTCTCCTTA-3′; 92 bp from 63929413 to 63929504, according to NC_000017.11 (GRCh38.p12 under https://www.ncbi.nlm.nih.gov/nuccore); 40× cycles] and MYD88 [MYD88_F: 5′-CTTGCAGGTGCCCATCAG-3′; MYD88_R: 5′-CAGGATGCTGGGGAACTCT-3′; 75 bp from 38141125 to 38141199, according to NC_000003.12 (GRCh38.p12); 40× cycles]. PCR products were analyzed by QIAxcelAdvanced system (Qiagen) and subjected to NGS. Library preparation, NGS, and data processing were performed at the Cologne Center for Genomics. Purified PCR amplicons (CD79B and MYD88; see PCR Analysis) were end repaired and A-tailed and adapter ligated using the Illumina TruSeq nano kit (Illumina, San Diego, CA) and protocol without further PCR amplification. After validation (2200 TapeStation; Agilent Technologies, Santa Clara, CA) and quantitation (Qubit System; Invitrogen, Waltham, MA), amplicon libraries were quantified using the KAPA Library Quantification kit (Peqlab, Erlangen, Germany) and the 7900HT Sequence Detection System (Applied Biosystems, Waltham, MA) and subsequently sequenced on an Illumina MiSeq sequencing instrument using a paired-end 2 × 150 bp protocol. FASTQ output data were aligned to the GRCh37 human genome sequence using bwa-aln version 0.6.212Li H. Durbin R. Fast and accurate short read alignment with Burrows-Wheeler transform.Bioinformatics. 2009; 25: 1754-1760Crossref PubMed Scopus (26648) Google Scholar to generate bam and bai files for visualization in Integrative Genomics Viewer.13Robinson J.T. Thorvaldsdottir H. Winckler W. Guttman M. Lander E.S. Getz G. Mesirov J.P. Integrative genomics viewer.Nat Biotechnol. 2011; 29: 24-26Crossref PubMed Scopus (7591) Google Scholar To test for a biological background of mutations at driver gene hot spot (ie, the presence of hot spot driver mutations detectable at low level in healthy individuals), MYD88 L265P and L281P (triplet code L = CTG and P = CCG) mutations were investigated by NGS. The position MYD88 L281P is located within the primer MYD88_R (positions 1 to 3) applied for PCR in the NGS approach; thus, any kind of possible biological background will be overwritten. Results were reported as variant allelic frequency (VAF; ie, the highest value, excluding the germ line relative to total reads). VAF for MYD88 L281P of healthy controls represents the technical background. MYD88 L265P and L281P mutations were further investigated by droplet digital PCR (ddPCR). Assays were designed by the manufacturer (Bio-Rad, Feldkirchen, Germany) for T to C mutation. In contrast to the NGS approach, any kind of biological background at position L281P will not be overwritten. All ddPCRs were performed in duplicate; only samples that had generated a minimum of 8000 droplets were considered for analysis. Sufficient material of cfDNA to allow analysis was available in 18 of 27 PCNSL cases, 2 of 6 healthy controls, and 3 of 4 Waldenstrom disease cases (Supplemental Table S1). Amplifications were performed in a reaction volume of 20 μL on the QX200 Droplet Digital PCR System (Bio-Rad). Each PCR contained 10 μL Bio-Rad PCR mix for probes, 1 μL of each (mutated and unmutated) amplification primer/probe mix, and 8 μL of cfDNA. PCR cycling was performed on a C1000 thermo cycler (Bio-Rad), according to manufacturer's instructions. Results were analyzed with Quantasoft software version 1.7.4 (Bio-Rad) and reported as VAF. Statistical significance between mean ratios of different groups was assessed by two-sided t-test; P < 0.05 was considered significant. Plasma cfDNA isolated from peripheral blood collected immediately before stereotactic biopsy and frozen PCNSL tissue obtained thereafter was analyzed for CD79B and MYD88 mutations in 27 patients with newly diagnosed PCNSL3Deckert M. Paulus W. Kluin P. Ferry J. Lymphomas.in: Louis D.N. Ohgaki H. Wiestler O.D. Cavenee W.K. WHO Classification of Tumours of the Central Nervous System. Revised 4th ed. IARC, Lyon, France2016: 272-277Google Scholar (Table 1). Controls included plasma cfDNA from six healthy controls and four patients with Waldenstrom disease in complete clinical remission (Table 1). The mean value of the isolated cfDNA concentration for all samples was 2.4 ng/μL (median, 2.5 ng/μL; minimum, 1.1 ng/μL; maximum, 9.1 ng/μL). There was no significant difference in cfDNA between mean DNA concentration of PCNSL and healthy controls (P = 0.699). Mean coverage of PCNSL biopsy-derived DNA for CD79B was 80,253 (median, 71,980; minimum, 51,661; maximum, 137,781; coverage defined as total reads); and for MYD88, it was 26,108 (median, 26,029; minimum, 13,083; maximum, 37,139). Mean coverage of cfDNA from PCNSL patients for CD79B was 95,422 (median, 96,814; minimum, 73,728; maximum, 130,780); and for MYD88, it was 31,187 (median, 32,175; minimum, 12,554; maximum, 44,421). For all cfDNA samples of healthy controls (Table 1), the mean cfDNA VAF of CD79B Y196X was 0.20% (median, 0.20%; minimum, 0.19%; maximum, 0.25%); and for MYD88 L265P, it was 0.27% (median, 0.24%; minimum, 0.22%; maximum, 0.45%). Using the VAF values of the MYD88 L265P healthy controls, the cutoff VAF value (mean VAF in healthy controls + 3 × SD) was set to 0.53%. Using this cutoff, CD79B and MYD88 mutations were identified in 16 of 27 (59%) and 23 of 27 (85%) PCNSL biopsies (Table 2), respectively. This frequency is in line with previous reports.9Montesinos-Rongen M. Godlewska E. Brunn A. Wiestler O.D. Siebert R. Deckert M. Activating L265P mutations of the MYD88 gene are common in primary central nervous system lymphoma.Acta Neuropathol. 2011; 122: 791-792Crossref PubMed Scopus (115) Google Scholar,10Montesinos-Rongen M. Schäfer E. Siebert R. Deckert M. Genes regulating the B cell receptor pathway are recurrently mutated in primary central nervous system lymphoma.Acta Neuropathol. 2012; 124: 905-906Crossref PubMed Scopus (46) Google Scholar Among PCNSL, the mean VAF was 27.66% and 43.68% for CD79B and MYD88, respectively (Table 3). Furthermore, regarding CD79B, the mutated PCNSL cases have a mean VAF of 46.55%, and the unmutated PCNSL cases have a mean VAF of 0.19%. With respect to MYD88, the mutated PCNSL cases have a mean VAF of 51.25%, and the unmutated PCNSL cases have a mean VAF of 0.19% (Table 3).Table 2Mutation Status of the CD79B and MYD88 Genes Detected by NGSCaseDiagnosisSexAge, yearsTumor biopsy DNACell-free DNACD79B∗CD79B Y196.%ReadsCoverageMYD88†MYD88 L265P.%ReadsCoverageCD79B∗CD79B Y196.%ReadsCoverageMYD88†MYD88 L265P.%ReadsCoverage1PCNSLM43Mut75.4753,36770,713Mut80.5222,49527,937WT0.1714384,793WT0.216228,9112PCNSLM65WT0.2412351,661Mut48.1011,17523,233WT0.18200114,226WT0.249138,0363PCNSLM63Mut69.0747,18368,313Mut39.5510,81527,347WT0.24254106,492WT0.3010836,4424PCNSLM66Mut44.7632,21871,980Mut45.8613,73829,954WT0.24249103,159WT0.165434,4205PCNSLM78Mut86.3553,92562,451Mut47.7212,42126,029WT0.1615799,852WT0.247933,1386PCNSLF71WT0.1812067,460Mut37.51975726,013WT0.1611774,573WT0.297927,1807PCNSLF82Mut44.9230,21667,269WT0.306120,350WT0.14166118,493WT0.2810638,0188PCNSLF79WT0.179254,946Mut65.2716,30124,976WT0.1413996,814WT0.175733,5949PCNSLM75Mut31.2224,78779,394Mut47.8417,42936,433WT0.20201101,425WT0.248535,46410PCNSLF77WT0.2413355,869Mut82.4128,05934,048WT0.2221697,454WT0.258031,48811PCNSLM67Mut34.8319,57056,185Mut45.8916,18635,268WT0.28289102,504WT0.2810837,94912PCNSLM51Mut32.1020,96565,307Mut47.4910,83422,813WT0.2524398,531WT0.299933,58513PCNSLF76WT0.2216373,723Mut39.0512,09830,977WT0.2216174,692WT0.226931,50914PCNSLF72WT0.2818466,017WT0.217837,139WT0.2217076,680WT0.2310144,42115PCNSLM72Mut40.9724,32359,364Mut50.04982019,625WT0.2420486,241WT0.319430,38916PCNSLM65Mut3.27232871,179Mut81.8326,59232,496WT0.2521787,299WT0.215426,13617PCNSLM57WT0.2621784,621Mut46.2712,84027,750WT0.2216473,728WT0.3510630,64518PCNSLM80Mut48.3339,42181,572Mut51.9216,44231,668WT0.1310078,782WT0.216932,55019PCNSLM86Mut41.9030,67373,200Mut46.5613,08228,100WT0.2926189,249WT0.166540,11920PCNSLF85Mut10.23881886,201Mut29.82760125,491WT0.2217578,674WT0.2610238,49321PCNSLF80WT0.2421088,361Mut4.81161833,642WT0.2016681,300WT0.286723,71522PCNSLM78Mut43.6252,790121,019Mut84.8914,06616,569WT0.19238127,468WT0.368924,45923PCNSLF57WT0.09109124,597Mut0.6010417,460WT0.1311286,848WT0.206332,17524PCNSLM75Mut92.61127,600137,781Mut64.8814,38722,175WT0.12123105,556WT0.185430,30325PCNSLM82Mut45.1549,373109,350Mut89.8916,48818,343WT0.13146109,028Mut0.8811112,55426PCNSLM82WT0.10121119,359WT0.091515,984WT0.12157130,780WT0.204221,09527PCNSLM52WT0.099398,938WT0.141813,083WT0.1311791,745WT0.132015,25628HCM46––––––––WT0.1917493,755WT0.226629,61529HCF47––––––––WT0.1915278,084WT0.256325,40830HCM42––––––––WT0.2013468,284WT0.2612448,29231HCF38––––––––WT0.1911661,922WT0.2411749,15532HCF41––––––––WT0.2017689,282WT0.226931,34133HCF21––––––––WT0.2524197,987WT0.4520044,60934WDM77––––––––WT0.12137109,931WT0.173621,24535WDF59––––––––WT0.14189133,032Mut2.8471325,14036WDF60––––––––WT0.11130113,198WT0.8119924,59537WDF67––––––––WT0.14163119,278WT0.161710,888F, female; M, male; –, not applicable; HC, healthy control; Mut, mutation; NGS, next-generation sequencing; PCNSL, primary central nervous system lymphoma; WD, Waldenstrom disease; WT, no mutation.∗ CD79B Y196.† MYD88 L265P. Open table in a new tab Table 3Details of VAF of PCNSL BiopsiesGroupVAF, %∗The highest value, excluding the germ line relative to total reads.MeanMedianMinimumMaximumPCNSL CD79B27.6631.220.0992.61PCNSL CD79B mutated46.5544.193.2792.61PCNSL CD79B unmutated0.190.220.090.28PCNSL MYD8843.6846.560.0989.89PCNSL MYD88 mutated51.2547.720.6089.89PCNSL MYD88 unmutated0.190.170.090.30PCNSL, primary central nervous system lymphoma; VAF, variant allelic frequency.∗ The highest value, excluding the germ line relative to total reads. Open table in a new tab F, female; M, male; –, not applicable; HC, healthy control; Mut, mutation; NGS, next-generation sequencing; PCNSL, primary central nervous system lymphoma; WD, Waldenstrom disease; WT, no mutation. PCNSL, primary central nervous system lymphoma; VAF, variant allelic frequency. Remarkably, with the notable exception of the MYD88 L265P mutation in PCNSL Patient 25 (VAF = 0.88%) (Table 1), NGS of cfDNA did not yield a VAF for both genes above the experimental cutoff, neither in any of the other 26 PCNSL patients nor in any of the six healthy controls. In plasma-derived cfDNA of one of the four patients with Waldenstrom disease, the MYD88 L265P mutation was readily detected (VAF = 2.84%; Case 35) (Table 2) despite complete clinical remission. In six healthy controls and a representative subset of 16 patients covering all four mutational combinations in PCNSL (CD79BmutMYD88mut, CD79BmutMYD88unmut, CD79BunmutMYD88mut, CD79BunmutMYD88unmut), ultradeep sequencing was performed. Mean coverage was 667,781 (median, 659,270; minimum, 569,676; maximum, 792,437) and 187,042 (median, 179,007; minimum, 111,180; maximum, 310,463) for CD79B and MYD88, respectively. In this ultradeep sequencing, among the 16 PCNSL patients, the mean VAF of cfDNA was 0.51% (median, 0.15%; minimum, 0.09%; maximum, 3.31%) for CD79B and 0.16% (median, 0.16%; minimum, 0.12%; maximum, 0.24%) for MYD88. For healthy controls, the VAF of cfDNA was 0.14% (median, 0.12%; minimum, 0.09%; maximum, 0.24%) and 0.16% (median, 0.16%; minimum, 0.13%; maximum, 0.20%) for CD79B and MYD88, respectively. There was no significant difference in cfDNA between mean VAF of PCNSL and healthy controls, neither for CD79B Y196X (P = 0.324) nor for MYD88 L265P (P = 0.859). Ultradeep sequencing of the generated PCR products indicated that mutations at the hot spots might also occur in cells of healthy individuals at low level. To follow this hypothesis that activating mutations of MYD88 may also occur in healthy controls, the current study took advantage of the fact that after 16 amino acids downstream of the mutational hot spot L265P (CTG > CCG) there is an identical triplet. However, a similar mutation at this position (ie, an L281P mutation) has not yet been detected in PCNSL. For all cfDNA samples of PCNSL (Table 1), the mean cfDNA VAF of L265P was 0.26% (median, 0.24%; minimum, 0.13%; maximum, 0.88%), which is significantly higher (P < 0.001) than the mean cfDNA VAF of L281P of 0.03% (median, 0.03%; minimum, 0.01%; maximum, 0.07%). There was no significant difference (P = 0.391) between the corresponding cfDNA mean VAF of mutated (0.27%) and unmutated (0.21%) PCNSL for MYD88 L265P. There was also no significant difference (P = 0.169) between the corresponding cfDNA mean VAF of mutated (0.04%) and unmutated (0.03%) PCNSL for MYD88 L281P. Similarly, for all cfDNA samples of healthy controls (Table 1), the mean cfDNA VAF of L265P was 0.27% (median, 0.24%; minimum, 0.22%; maximum, 0.45%), which is significantly higher (P < 0.001) than the mean cfDNA VAF of L281P of 0.04% (median, 0.03%; minimum, 0.04%; maximum, 0.08%) in healthy controls. Furthermore, there was no significant difference between the corresponding cfDNA mean VAF of mutated PCNSL and healthy controls, neither for MYD88 L265P (P = 0.978) nor for MYD88 L281P (P = 0.918). The above finding led us to further analyze MYD88 L265P and L281P by ddPCR. Because of limited material, this analysis could not be performed for all samples (Supplemental Table S1). The mean ratio of positive and negative droplets in cfDNA for MYD88 L265P was 0.19% for PCNSL patients (median, 0.20%; minimum, 0.08%; maximum, 0.31%), which, again, is significantly different (P = 0.012) from the mean cfDNA VAF of L281P of 0.09% (median, 0.10%; minimum, 0.00%; maximum, 0.18%). For healthy controls, the mean ratio of positive and negative droplets in cfDNA for MYD88 L265P was 0.22% (median, 0.22%; minimum, 0.00%; maximum, 0.43%); and it was 0.96% for Waldenstrom disease patients (median, 0.33%; minimum, 0.23%; maximum, 2.31%). Excluding the patient (Case 35) with an MYD88 L265P mutation in cfDNA, the mean ratio of positive and negative droplets was 0.28% (median, 0.28%; minimum, 0.23%; maximum, 0.33%) (Supplemental Table S1). In summary, there was no significant difference in cfDNA mean VAF values of PCNSL and healthy controls for MYD88 L265P, but there was a significant difference in cfDNA over all samples regarding the mean VAF values between MYD88 L265P and MYD88 L281P. The current study established a cutoff level for the detection of mutations for the NGS approach, which was set to a VAF of 0.53%. With this criterion, CD79B and MYD88 mutations could be detected in 16 of 27 (59%) and 23 of 27 (85%) PCNSL biopsies, respectively. With the exception of the MYD88 L265P mutation in PCNSL Patient 25 cfDNA, NGS of cfDNA did not yield a VAF for both genes above the cutoff, neither in any PCNSL patient nor in any healthy control. Furthermore, there was a significant difference in cfDNA regarding the mean VAF values between MYD88 L265P and MYD88 L281P in PCNSL as well as controls (Table 2). These data are in line with a study of 14 PCNSL patients in whom targeted deep sequencing did not identify MYD88 mutations in plasma cfDNA; however, when the same samples were subjected to ddPCR, the MYD88 L265P mutation was identified in 8 of 14 (57%) cfDNAs with VAF between 0.09% and 0.69%.14Hattori K. Sakata-Yanagimoto M. Suehara Y. Yokoyama Y. Kato T. Kurita N. Nishikii H. Obara N. Takano S. Ishikawa E. Matsumura A. Hasegawa Y. Chiba S. Clinical significance of disease-specific MYD88 mutations in circulating DNA in primary central nervous system lymphoma.Cancer Sci. 2018; 109: 225-230Crossref PubMed Scopus (44) Google Scholar As nontumor control samples were not included in the latter study,14Hattori K. Sakata-Yanagimoto M. Suehara Y. Yokoyama Y. Kato T. Kurita N. Nishikii H. Obara N. Takano S. Ishikawa E. Matsumura A. Hasegawa Y. Chiba S. Clinical significance of disease-specific MYD88 mutations in circulating DNA in primary central nervous system lymphoma.Cancer Sci. 2018; 109: 225-230Crossref PubMed Scopus (44) Google Scholar the question of whether detection of these mutations by ultradeep sequencing is tumor related could not been answered. In another series, NGS detected somatic mutations in cfDNA of 8 of 25 (32%) PCNSL patients, with six patients harboring the MYD88 c.T778C (ie, L265P) mutation.15Fontanilles M. Marguet F. Bohers E. Viailly P.J. Dubois S. Bertrand P. Camus V. Mareschal S. Ruminy P. Maingonnat C. Lepretre S. Veresezan E.L. Derrey S. Tilly H. Picquenot J.M. Laquerriere A. Jardin F. Non-invasive detection of somatic mutations using next-generation sequencing in primary central nervous system lymphoma.Oncotarget. 2017; 8: 48157-48168Crossref PubMed Scopus (57) Google Scholar This apparent discrepancy with the data may be explained by differences in the time point of blood sample collection and in preceding medication. The optimal time point for collection of tumor and plasma samples critically determines detection of mutations.1Corcoran R.B. Chabner B.A. Application of cell-free DNA analysis to cancer treatment.N Engl J Med. 2018; 379: 1754-1765Crossref PubMed Scopus (413) Google Scholar In the present study, blood was collected immediately preceding surgery, whereas Fontanilles et al15Fontanilles M. Marguet F. Bohers E. Viailly P.J. Dubois S. Bertrand P. Camus V. Mareschal S. Ruminy P. Maingonnat C. Lepretre S. Veresezan E.L. Derrey S. Tilly H. Picquenot J.M. Laquerriere A. Jardin F. Non-invasive detection of somatic mutations using next-generation sequencing in primary central nervous system lymphoma.Oncotarget. 2017; 8: 48157-48168Crossref PubMed Scopus (57) Google Scholar obtained blood after neurosurgery. Thus, surgery-induced BBB disturbance may facilitate tumor DNA leakage into the circulation. Although 20% (5/25) of patients in the Fontanilles et al15Fontanilles M. Marguet F. Bohers E. Viailly P.J. Dubois S. Bertrand P. Camus V. Mareschal S. Ruminy P. Maingonnat C. Lepretre S. Veresezan E.L. Derrey S. Tilly H. Picquenot J.M. Laquerriere A. Jardin F. Non-invasive detection of somatic mutations using next-generation sequencing in primary central nervous system lymphoma.Oncotarget. 2017; 8: 48157-48168Crossref PubMed Scopus (57) Google Scholar study underwent open neurosurgical resection, unequivocally disrupting the BBB, all patients in the current study were stereotactically biopsied, which minimizes trauma and BBB alteration. Furthermore, extravasation of intact and/or disintegrating tumor cells from the CNS may be enhanced by corticosteroids, which induce rapid tumor cell apoptosis followed by DNA fragment deliberation. All of the patients in the Fontanilles et al15Fontanilles M. Marguet F. Bohers E. Viailly P.J. Dubois S. Bertrand P. Camus V. Mareschal S. Ruminy P. Maingonnat C. Lepretre S. Veresezan E.L. Derrey S. Tilly H. Picquenot J.M. Laquerriere A. Jardin F. Non-invasive detection of somatic mutations using next-generation sequencing in primary central nervous system lymphoma.Oncotarget. 2017; 8: 48157-48168Crossref PubMed Scopus (57) Google Scholar study, but none of the patients in the current study, had received corticosteroids before blood collection and neurosurgical intervention. Alternatively, biological factors may account for discrepancies in the various studies. One might hypothesize that healthy individuals carry CD79B and MYD88 mutations at low frequency, as mutations in these driver genes are early events in lymphomagenesis.16Vater I. Montesinos-Rongen M. Schlesner M. Haake A. Purschke F. Sprute R. Mettenmeyer N. Nazzal I. Nagel I. Gutwein J. Richter J. Buchhalter I. Russell R.B. Wiestler O.D. Eils R. Deckert M. Siebert R. The mutational pattern of primary lymphoma of the central nervous system determined by whole-exome sequencing.Leukemia. 2015; 29: 677-685Crossref PubMed Scopus (103) Google Scholar When tumor precursor cells with CD79B and MYD88 mutations receive further oncogenic hits, they may develop into PCNSL. Thus, in this regard, CD79B and MYD88 mutations may resemble the BCL2 translocation, which also occurs in healthy individuals and per se does not yield the diagnosis of follicular lymphoma.17Limpens J. Stad R. Vos C. de Vlaam C. de Jong D. van Ommen G.J. Schuuring E. Kluin P.M. Lymphoma-associated translocation t(14;18) in blood B cells of normal individuals.Blood. 1995; 85: 2528-2536Crossref PubMed Google Scholar Depending on assay sensitivity, almost all healthy individuals carry one or multiple cell clones of the t(14;18) with a frequency of 1 to 100 rearranged cells in 106 cells.18Schüler F. Hirt C. Dölken G. Chromosomal translocation t(14;18) in healthy individuals.Semin Cancer Biol. 2003; 13: 203-209Crossref PubMed Scopus (74) Google Scholar Interestingly, the frequency of clones with the t(14;18) translocation increases with age.19Liu Y. Hernandez A.M. Shibata D. Cortopassi G.A. BCL2 translocation frequency rises with age in humans.Proc Natl Acad Sci U S A. 1994; 91: 8910-8914Crossref PubMed Scopus (329) Google Scholar As the risk for PCNSL also increases with age,5Deckert M. Engert A. Brück W. Ferreri A.J. Finke J. Illerhaus G. Klapper W. Korfel A. Küppers R. Maarouf M. Montesinos-Rongen M. Paulus W. Schlegel U. Lassmann H. Wiestler O.D. Siebert R. DeAngelis L.M. Modern concepts in the biology, diagnosis, differential diagnosis and treatment of primary central nervous system lymphoma.Leukemia. 2011; 25: 1797-1807Crossref PubMed Scopus (142) Google Scholar one might hypothesize that the frequency of mutations in CD79B and MYD88 also increases with age. This hypothesis is also supported by the fact that the vast majority of MYD88 mutations have been assigned to a signature related to aging.20Chapuy B. Stewart C. Dunford A.J. Kim J. Kamburov A. Redd R.A. et al.Molecular subtypes of diffuse large B cell lymphoma are associated with distinct pathogenic mechanisms and outcomes.Nat Med. 2018; 24: 679-690Crossref PubMed Scopus (817) Google Scholar Thus, under conditions of high sensitivity, which is the case for ddPCR and ultradeep sequencing, genetic alterations may become detectable, but, per se, may be not tumor specific, as also concluded by Schüler et al.18Schüler F. Hirt C. Dölken G. Chromosomal translocation t(14;18) in healthy individuals.Semin Cancer Biol. 2003; 13: 203-209Crossref PubMed Scopus (74) Google Scholar Interestingly, Singh et al21Singh M. Jackson K.J.L. Wang J.J. Schofield P. Field M.A. Koppstein D. Peters T.J. Burnett D.L. Rizzetto S. Nevoltris D. Masle-Farquhar E. Faulks M.L. Russell A. Gokal D. Hanioka A. Horikawa K. Colella A.D. Chataway T.K. Blackburn J. Mercer T.R. Langley D.B. Goodall D.M. Jefferis R. Gangadharan Komala M. Kelleher A.D. Suan D. Rischmueller M. Christ D. Brink R. Luciani F. Gordon T.P. Goodnow C.C. Reed J.H. Lymphoma driver mutations in the pathogenic evolution of an iconic human autoantibody.Cell. 2020; 180: 878-894.e819Abstract Full Text Full Text PDF PubMed Scopus (51) Google Scholar demonstrated that mutations in genes involved in the B-cell receptor pathway, CARD11 and KLHL6 in particular, can lead to rouge B cells, which develop autoantibodies. These rouge B cells are regarded as preneoplastic stage in human lymphomagenesis.21Singh M. Jackson K.J.L. Wang J.J. Schofield P. Field M.A. Koppstein D. Peters T.J. Burnett D.L. Rizzetto S. Nevoltris D. Masle-Farquhar E. Faulks M.L. Russell A. Gokal D. Hanioka A. Horikawa K. Colella A.D. Chataway T.K. Blackburn J. Mercer T.R. Langley D.B. Goodall D.M. Jefferis R. Gangadharan Komala M. Kelleher A.D. Suan D. Rischmueller M. Christ D. Brink R. Luciani F. Gordon T.P. Goodnow C.C. Reed J.H. Lymphoma driver mutations in the pathogenic evolution of an iconic human autoantibody.Cell. 2020; 180: 878-894.e819Abstract Full Text Full Text PDF PubMed Scopus (51) Google Scholar During the process of hypermutation, the B-cell receptor of PCNSL tumor cells gains autoreactivity/polyreactivity against CNS antigens.22Montesinos-Rongen M. Purschke F.G. Brunn A. May C. Nordhoff E. Marcus K. Deckert M. Primary central nervous system (CNS) lymphoma B cell receptors recognize CNS proteins.J Immunol. 2015; 195: 1312-1319Crossref PubMed Scopus (26) Google Scholar,23Montesinos-Rongen M. Terrao M. May C. Marcus K. Blümcke I. Hellmich M. Küppers R. Brunn A. Deckert M. The process of somatic hypermutation increases polyreactivity for central nervous system antigens in primary central nervous system lymphoma.Haematologica. 2020; ([Epub ahead of print])https://doi.org/10.3324/haematol.2019.242701Crossref PubMed Scopus (7) Google Scholar These results imply that particularly in the analyses of hot spot mutations in cfDNA from malignant tumors derived from germinal center/post-germinal center lymphocytes, which have experienced the process of somatic hypermutation, acquired clonal alterations not (yet) leading to overt lymphoma need to be considered. This hypothesis that mutations at the hot spots of CD79B and MYD88 may occur in B cells under physiological conditions was herein tested for MYD88. In contrast to the CD79B Y196X mutation, the MYD88 L265P mutation in the current study always yields an exchange from CTG to CCG. Sixteen amino acids downstream (ie, L281P), the MYD88 gene harbors an identical triplet; however, an MYD88 L281P mutation has so far not been detected in PCNSL. Furthermore, one cannot exclude the possibility that the single case in which an MYD88, but not a CD79B, mutation was detectable in cfDNA, both of which were present in the respective PCNSL biopsy, might still be falsely positive. A definite answer to this question requires larger series of PCNSL biopsies and corresponding cfDNA, which should also include unique molecular identifier to improve the sensitivity and limit possible artifacts. Despite these limitations, cfDNA may be an attractive tool to monitor disease and to track a therapeutic response. Among 15 patients, including nine with PCNSL, six with secondary CNS lymphoma, and nine (60%) presenting with cerebrospinal fluid/meningeal involvement, cerebrospinal fluid–derived cfDNA with tumor-specific genetic alterations was detected in eight (53%) by NGS. Interestingly, on complete/near-complete response to therapy, these alterations were no longer detectable in cerebrospinal fluid–derived cfDNA.24Grommes C. Tang S.S. Wolfe J. Kaley T.J. Daras M. Pentsova E.I. Piotrowski A.F. Stone J. Lin A. Nolan C.P. Manne M. Codega P. Campos C. Viale A. Thomas A.A. Berger M.F. Hatzoglou V. Reiner A.S. Panageas K.S. DeAngelis L.M. Mellinghoff I.K. Phase 1b trial of an ibrutinib-based combination therapy in recurrent/refractory CNS lymphoma.Blood. 2019; 133: 436-445Crossref PubMed Scopus (108) Google Scholar Similarly, longitudinal analysis showed a rapid clearance of DLBCL mutations from plasma cfDNA in treatment-responsive, but not resistant, patients.2Rossi D. Diop F. Spaccarotella E. Monti S. Zanni M. Rasi S. Deambrogi C. Spina V. Bruscaggin A. Favini C. Serra R. Ramponi A. Boldorini R. Foa R. Gaidano G. Diffuse large B-cell lymphoma genotyping on the liquid biopsy.Blood. 2017; 129: 1947-1957Crossref PubMed Scopus (144) Google Scholar Thus, the strategy of cerebrospinal fluid–derived cfDNA analysis could develop into a valuable clinical and prognostically relevant tool. One may even speculate whether the detection of mutations in blood-derived cfDNA in the setting of CNS lymphoma may rather indicate extracerebral DLBCL, thus suggesting that the cerebral tumor in fact corresponds to secondary CNS lymphoma. If this can be proved in patients with CNS metastasis of systemic DLBCL, cfDNA mutational analysis may become a highly valuable parameter in the differential diagnosis of primary versus secondary DLBCL of the CNS. However, it remains to be elucidated whether in extracerebral relapse, albeit rare, tumor DNA may circulate in the peripheral blood and may indicate relapse. Nevertheless, both NGS and ddPCR indicated the potential presence of low-level MYD88/CD79B mutations in healthy individuals. This observation indicates the necessity to include normal samples for calibration of target-sequencing approaches, in particular in the analyses of hot spot mutations in cfDNA from malignant tumors derived from germinal center/post-germinal center B lymphocytes. Overall, these overt discrepancies between the various studies highlight the importance of a standardized stringent experimental design and the inclusion of healthy controls. The data presented herein suggest that at least at initial diagnosis and before surgery, PCNSL is not associated with a similar mutational pattern of CD79B and MYD88 mutations in biopsy and blood-derived cfDNA. We thank Marc Eßer and Diana Rudakova for expert technical assistance. Download .pdf (.1 MB) Help with pdf files Supplemental Figure S1TapeStation analysis of cell-free DNA (cfDNA). Samples were analyzed with a 2200 TapeStation (Agilent Technologies, Santa Clara, CA). A single, strong peak at 173 bp indicates purity of cfDNA. The absence of peak(s) >500 bp excludes contamination of cfDNA by cellular DNA. Markers indicate the range of analysis that detects DNA between 15 and 10,000 bp (lower and upper values, respectively). Case 4 is shown exemplarily. Similar data were obtained for all cases analyzed. FU, fluorescence units. Download .docx (.02 MB) Help with docx files Supplemental Table S1
Referência(s)