Page 127 - 2021_10-Haematologica-web
P. 127
B-cell evolutionary patterns in NLPHL
differ from Hodgkin-Reed-Sternberg cells of classical Hodgkin lymphoma in terms of morphology and immunophenotype (sIg+, CD20+, BCL6+, CD79a+, CD45+, CD30–, CD15–, EBER–).4,5 LP cells have been described to express IgG B-cell receptors (BCR),6,7 while in a distinct subgroup of NLPHL, LP cells have been found to be IgD- positive by immunohistochemical staining (~25%). This subgroup has attracted interest recently because of emerg- ing evidence of shared immunoglobulin characteristics and the role of Moraxella bacterial antigens in its patho- genesis.6,8,9 Histologically, most NLPHL cases show LP cells located within large nodules composed of reactive B cells. A few cases present with variant growth patterns in which reactive T cells and/or histiocytes predominate over the reactive B cells.10 These cases are more often diag- nosed in an advanced stage and relapse earlier.3,11 In gener- al, relapses occur in around 15% of patients and can arise even after many years.12 Still, NLPHL - both at diagnosis and at relapse - has a rather indolent clinical behavior with a favorable prognosis and there are effective treatment options. However, an important complication of NLPHL is transformation into a more aggressive B-cell non-Hodgkin lymphoma (NHL) such as diffuse large B-cell lymphoma (DLBCL), which is observed in 10-30% of cases.13,14 Past studies have identified both cases with NHL clones related to the LP-cell clone and cases emerging from different cells of origin.15-17 Cases with disease transformation have a sig- nificantly higher mortality and faster progression, thus requiring a different treatment approach. Markers that are able to predict the risk of transformation early would, therefore, be beneficial.
Here we used next-generation immunoglobulin heavy chain sequencing (IGH NGS) to investigate B-cell reper- toire metrics and evolutionary B-cell trajectories in NLPHL cases that later relapsed or transformed to high-grade NHL. As the most informative controls, we used non- malignant lymphoproliferations that can precede or fol- low NLPHL or Hodgkin lymphoma.18-21 Our major goals were to characterize better the LP-cell clone and its bystander B cells in these NLPHL subsets, to address the question of clonal identity at diagnosis, relapse and trans- formation, and to find B lineage signatures potentially pre- dictive of subsequent transformation with a technique that can be applied to paraffin-embedded tissue and that is robust enough to be used in routine diagnostics. Our analysis revealed pathogenic clues in NLPHL showing that the majority of NLPHL relapses and some transformations arise from the same cell of origin and that this cell of origin is subject to ongoing antigenic selection in many of these lymphomas. Moreover, the data presented here provide a set of molecular determinants for transformation that may be derived from targeted NGS on paraffin-embedded lym- phoma tissue.
Methods
Patients’ characteristics
Eighty-six paraffin-embedded tissues from patients and con- trols were collected between 1987 and 2020 at the University Medical Centers of Frankfurt, Halle, Tübingen, Kiel, Würzburg, Stuttgart, Lübeck (Germany) and Tampere (Finland). Three of 81 tissues were previously used for LP-cell dissection and immunoglobulin heavy chain (IGH) and light chain (IGL) sequencing as published previously9 (sample denomination case
3 = NLPHL01-initial diagnosis, case 6 = NLPHL02-initial diagno- sis, case 8 = NLPHL10-initial NLPHL). Our cohort included five cases of NLPHL that did not relapse or undergo transformation within 5 years after diagnosis (cohort 1), 16 NLPHL cases with paired samples at initial diagnosis and relapse(s) (cohort 2), ten cases with paired samples of initial NLPHL with transformation to DLBCL at relapse (cohort 3) and 28 control cases. These con- trol cases comprised ten cases of nonspecific lymphadenitis, four cases of progressive transformation of germinal centers, ten cases of DLBCL (activated B-cell subtype) and four cases of T-cell/his- tiocyte-rich large B-cell lymphoma. The clinical characteristics of cohorts 1-3 are listed in Online Supplementary Table S1. NLPHL was diagnosed according to the World Health Organization (WHO) classification 2017.4 All variants were classified by one of the authors (SH) according to the histopathological patterns described by Fan et al.10 Pattern A was considered ‘typical’, all other patterns were termed ‘atypical’ in Figures 1 and 5. Example pictures of histopathological CD20 staining of the most relevant cases are shown in Online Supplementary Figure S1. The study was approved by the local ethics committee of the University of Frankfurt (n. 157/17) and Halle (n. 2017-81) and was performed in accordance with the Declaration of Helsinki of 1975.
IgD staining
IgD immunostaining was performed as previously described.22 In brief, paraffin slides were deparaffinized and heat-pretreated at pH 8. They were subsequently incubated with a polyclonal IgD antibody (Agilent, Santa Clara, CA, USA) for 30 min at room temperature. The Envision-FLEX Kit (Agilent) was used for detec- tion.
Next-generation sequencing of IGH repertoires
IGH repertoires were obtained from bulk lymphoid tissue as described elsewhere.23-28 The rearranged IGH locus was amplified in a multiplex polymerase chain reaction from 250 ng of genomic DNA and BIOMED2-FR1 or -FR3 primer pools.29 NGS and demultiplexing were performed on an Illumina MiSeq sequencer (601-cycle single indexed, paired-end run, V3-chemistry). The MiXCR framework30 was used for data processing with the IMGT library31 as reference for sequence alignment. Only pro- ductive reads were used and all repertoires were normalized to 20,000 reads. All analyses and data plotting were performed using RStudio version 3.5.1. and the tcR,32 ade4,33 ggplot2,34 bub- bles35 and tidyverse36 packages.
Using FR3 polymerase chain reactions, the CDR3 sequence as well as the IGHD- and IGHJ-gene segments could be reliably identified, while the alignment of the IGHV-gene segment was less robust because of the small size of the amplified segment. To account for this weakness, we decided to indicate only the IGHV1-7 families, but not the individual IGHV genes.
A detailed description of the NGS and data analysis is included in the Online Supplementary Information.
Results
Clinical and pathological characteristics of cases of nodular lymphocyte-predominant Hodgkin lymphoma that did or did not subsequently transform into non-Hodgkin lymphoma
NLPHL has a high rate of transformation into subtypes of NHL such as DLBCL. We investigated three unselected subcohorts to identify determinants of transformation: five NLPHL cases (5 samples) without relapse or transfor- mation with a mean/median follow-up time of 16/17 (6 to
haematologica | 2021; 106(10)
2655