Introduction
Follicular lymphoma (FL) and marginal zone lymphoma (MZL) are two of the most common indolent B-cell lymphomas in China [
1,
2]. However, these diseases are rarely seen in the pediatric population, accounting for only 1–2% of B-cell lymphomas in children and young adults [
3‐
5]. In 1979, Frizzera et al. first reported that FLs in children are different from those in adults [
6]. Similarly, in 2003, Taddesse-Heath et al. proposed that nodal MZLs in children are different from those in adults, too [
7]. Pediatric-type follicular lymphoma (PTFL) and pediatric nodal marginal zone lymphoma (PNMZL) were thus proposed as two special distinct entities in the 4th (2008), revised 4th (2017), and the updated 5th edition (2022) of the World Health Organization (WHO) Classification of Hematolymphoid Tumors [
3‐
5].
PTFL and PNMZL have many similar clinical characteristics: they both predominantly involve the head and neck lymph nodes of young males aged 15–18 years, most patients are diagnosed with stage I–II disease, the prognosis is usually excellent after resection of the lesions by surgery, and B symptoms are infrequent [
8‐
10].
Microscopically, PTFL features enlarged follicles, thin or absent mantle zones, and serpiginous margins. The tumor cells are usually blastoid, containing no prominent nucleoli, with expression of CD10 and BCL6. Mitotic figures are readily seen. In contrast, PNMZL tends to exhibit a proliferation of small lymphoid cells that surround follicles and expand into the interfollicular areas, with some follicles resembling progressively transformed germinal centers (PTGC). However, during routine diagnostic practice, we notice that quite a lot of patients have composite PTFL and PNMZL, even within one lymph node. These mixed-type cases present with both histologic and immunophenotypic features characteristic of typical PTFL and PNMZL cases, respectively. Genetically, previous studies have revealed that PTFL usually lacks the
BCL2 translocation, but possesses recurrent
TNFRSF14,
MAP2K1, and
IRF8 mutations [
11‐
15], which differs from the conventional form of FL occurring in adult patients.
Recently, some mixed-type cases with histologic features of both PTFL and PNMZL have been described, and no significant differences in mutational profiles have been found when compared with PTFL and PNMZL [
17‐
19].
Due to the rarity of PTFL and PNMZL, fewer than 100 cases have been documented in the English literatures with available molecular profiles. In this study, we analyzed the clinicopathologic features of 59 young patients, including 39 PTFL, 5 PNMZL, and 15 mixed-type tumor (MTT) cases. Next-generation sequencing (NGS) analysis was performed in 3 PTFL, 3 PNMZL, and 3 MTT cases with sufficient materials. In addition, a comprehensive literature review summarizing the genetic features of PTFL and PNMZL was presented.
Methods
Case selection
Thirty-nine PTFL, 5 PNMZL, and 15 MTT cases diagnosed between 2013 and 2019 at Fudan University Shanghai Cancer Center were collected. Slides were stained with hematoxylin and eosin (H&E) and subjected to immunohistochemical (IHC) procedure. All the slides were reviewed by 3 of the authors (HGL, XNJ, and XQL), and a final diagnosis was confirmed based on the criteria given in the 5th edition of the WHO Classification of Hematolymphoid Tumors [
5]. For the MTT cases, the composition of PTFL or PNMZL element was required to be more than 10% each of the whole lesion in this study.
Immunohistochemistry and flow cytometry
Immunohistochemical staining was performed according to the manufacturers’ protocol. Formalin-fixed paraffin-embedded (FFPE) sections were dewaxed, rehydrated, and stained with primary antibodies against CD20 (clone L26, Roche), CD10 (clone MX002, Maixin), BCL6 (clone LN22, Leica), BCL2 (clone SP66, Roche), MUM1 (clone MUM1P, Maixin), Kappa (multiclonal, Dako), Lambda (multiclonal, Dako), CD43 (clone L260, Roche), and Ki-67 (clone 30-9, Roche) on a BenchMark XT automated immunostainer (Ventana Medical System Inc., Roche Tucson, AZ, USA) according to the EnVision method.
Light chain restriction was further demonstrated by flow cytometric analysis of fine–needle aspirates in 4 PTFL, 1 PNMZL, and 2 MTT cases by using a 6-color BD FACS-Canto II flow cytometer (Becton Dickinson Biosciences, San Jose, CA, USA) and the immunofluorescent antibodies Kappa (TB28-2, BD) and Lambda (1-155-2, BD).
Detection of immunoglobulin gene rearrangements
Immunoglobulin (IG) heavy chain (IGH) and light chain (IGK and IGL) gene rearrangements were analyzed by using standardized BIOMED-II polymerase chain reaction assays.
Targeted NGS
Targeted NGS was performed by using a customized panel of 571 lymphoma-related genes. Patients were genotyped using high-depth panel sequencing with a mean depth of coverage, and 800 × amplicon sequencing was performed using the customized panel. As input, genomic DNA was extracted from tissue specimens using the TIANamp Genomic DNA Kit DP304-02 (TIANGEN Company product) from the FFPE samples. Genomic DNA quality and quantity were determined using a Nanodrop 8000 UV–Vis spectrometer (Thermo Scientific, Waltham, MA, USA), a Qubit 2.0 Fluorometer (Life Technologies Inc., Grand Island, NY, USA), and a 2200 TapeStation Instrument (Agilent Technologies, Santa Clara, CA, USA). Genomic DNA (50 ng) from each tissue sample was sheared with Covaris S220 (Covaris, Woburn, MA, USA) and used for the construction of a library with lymphoma-related detection kits provided by Shanghai Rightongene Biotechnology Co., Ltd. Briefly, gDNA was sheared, end-repaired, subjected to A-tailing, paired-end adaptor ligation, and amplification. After the enriched libraries were multiplexed, the 2 × 150 libraries were sequenced on an Illumina HiSeq 2500 sequencing platform (Illumina, San Diego, CA, USA) or an Illumina NovaSeq 6000 sequencing platform (Illumina, San Diego, CA, USA) using the modes of the TruSeq Rapid PE Cluster Kit and the TruSeq Rapid SBS Kit (Illumina).
The sequencing data in FASTQ format were cleaned by Trimmomatic (version 0.39), and FastQC software was used to evaluate the quality of the sequencing data. All clean reads were aligned to the human genome (release hg19) using the Burrows–Wheeler Alignment tool (BWA) (version 0.7.17). SAMTools (version 1.3) was used to convert and sort the resulting SAM files to binary format (bam). Indel realignment was performed by using GATK (version 3.8) with the Smit-Waterman alignment algorithm for areas near known Indels (from the dbSNP142 database, 1 KG Indels) to remove errors due to alignment. Base recalibration was performed to correct the quality score generated from the sequencing run by using GATK (version 3.8). FreeBayes (version V1.3.1–17) was used for gene annotation, and ANNOVAR was used for variant calling. The variant allele frequency (VAF) of gene mutations was defined as the number of variant reads divided by the number of total reads and reported as a percentage.
Statistical analysis
All the statistical analyses were performed using SPSS version 23.0 (IBM Corp., Armonk, NY). Categorical variables were compared by Fisher’s exact test or the Pearson chi-square test. Continuous variables were compared by the Kruskal‒Wallis test. Two-sided P values < 0.05 were considered to indicate statistical significance.
Discussion
PTFL and PNMZL are known to share many common clinicopathologic features, which are different from those of their adult counterparts, classic FL and nodal MZL [
21‐
23]. Both of the pediatric forms affect most frequently teens with a male predominance (with the male to female ratio of approximately 7:1), presenting most often with localized disease of lymphadenopathy involving the head and neck region. In the current study, we analyzed the clinicopathologic features of 39 PTFL, 5 PNMZL, and 15 MTT patients, and found no remarkable difference among the 3 groups except for histologic findings. Patients with these diseases usually feature an excellent prognosis. There is no apparent difference on outcome between the patients receiving a “watch and wait” approach and receiving chemotherapy or radiotherapy after resection [
24,
25], and the rate of recurrence is usually low. Some recent studies have revealed that some pediatric-type indolent B-cell lymphoma cases may feature hybrid or overlapping histologic characteristics of PTFL and PNMZL [
17‐
19], such a phenomenon has also been frequently noted in our daily diagnostic practice, and those cases are often labeled as “MTT” in terminology. Actually, in addition to those apparent MTT cases with remarkable elements of PTFL and PNMZL, quite a number of cases diagnosed as PTFL may show some histologic changes reminiscent of PNMZL, too. For example, it is not uncommonly seen that focal marginal zone distention or occasional follicles with a PTGC-like pattern are present in a PTFL lesion. Furthermore, the originally believed PTFL-related
MAP2K1,
IRF8, and
TNFRSF14 mutations can be detected in PNMZL and MTT patients, too [
12‐
15].
The incidence of PTFL is relatively higher than that of PNMZL. And the mutational features of PTFL have been better characterized. Ozawa et al. first reported a recurrent somatic variant encoding p.K66R in the transcription factor interferon regulatory factor 8 (
IRF8), which affects the DNA binding domain, in 3 of 6 PTFL patients [
12]. Similar findings were revealed by Schmidt et al. later [
15]. Besides,
IRF8 mutations involving another hot spot, p.Y23H, had been reported in PNMZL and MTT cases, too [
17]. It is noteworthy that 1 PNMZL and 2 MTT cases in our series had
IRF8 mutation at the hotspot p.K66R (c.197A > G), which indicates that PNMZL and MTT patients may have the same genetic features as PTFL patients.
IRF8 mutations at p.K66R and p.Y23H might be specific to PTFL and PNMZL, which are different from the
IRF8 variants described in classic FL and DLBCL; the latter are frequently indel and missense mutations predominantly located in the C-terminal domain [
15,
26]. Another PTFL-associated abnormality,
MAP2K1 mutations, was identified in our PNMZL and MTT cases, too, although the mutation spots differed in different subtypes. We noted 2 PTFL patients in our series harbored
MAP2K1 p.F53U and p.K57E hotspot mutations, while one patient each with PNMZL and MTT had a
MAP2K1 mutation at exon 6 p.S200F and exon 3 p.C121S, respectively. Tumor necrosis factor receptor superfamily member 14 (
TNFRSF14) gene mutations are among the most common genetic abnormalities in germinal center lymphomas, which are variably associated with prognosis [
27]. Different from
CREBBP,
TNFRSF14 was mutated in PTFLs at a similar frequency to that in limited-staged classic FL patients [
13]. Another remarkable finding in the current study was that two cases (PTFL3 and MTT2) had
TNFRSF14 mutations at the same site (exon 1 c.68 T > C; p.L23P), while the remaining 2 patients with
TNFRSF14 mutations mainly involved exon 2 (c.136G > C in PTFL1 and c.136G > A in MTT1, respectively).
Given the fact that PTFLs and PNMZLs are characterized by so many similar or overlapping histologic and genetic features, one may wonder whether these two different terminologies represent two distinct disease entities or just the same disease with a varying morphologic and phenotypic spectrum. Taking the frequent observations of the coexistence of PTFL and PNMZL within the same lymph node into consideration, we would prefer the opinion that they are probably closely related, if not the same. It has been proposed in the International Consensus Classification of Mature Lymphoid Neoplasms, too, that these two diseases may be related, although these findings have not been incorporated in the updated 5th edition of the WHO Classification of Hematolymphoid Tumors, yet [
5,
28]. Actually, not only pediatric forms of FL and NMZL but also some classic FL and nodal MZL cases may sometimes exhibit hybrid or ambiguous morphologic or phenotypic features, for which no gold standard exists to distinguish them. Examples include FL with marginal zone differentiation [
29,
30], in which the marginal zone-like component aberrantly expressing germinal center B-cell markers [
31,
32], and the genuine composite form of FL and nodal MZL [
33]. There are accumulated evidences indicating overlapping genetic aberrations between the follicular and marginal zone components in those cases, too. For instance, aberrations of chromosome 3, features of nodal MZL, frequently occur in FL with marginal zone differentiation [
34]. Furthermore, microdissection demonstrates the presence of t (14;18) in both follicular and marginal zone components [
35]. Sequencing analysis investigating the clonal relationship between follicular center and monocytoid B-cell components shows they have the same clonal origin of the follicular center [
36]. Recently, Tzioni et al. found metachronous extranodal MZL and FL are clonally related, that is, the two components from the same origin can occur simultaneously or metachronously [
37]. A conventional view of B-cell development lies that immature B cells may differentiate into follicular or marginal zone B-cells in the spleen after acquiring B cell receptor (BCR) rearrangement [
38]. But recently, Babushku et al. demonstrated that follicular B-cells can act as precursors for marginal B-cells, too, and the transition between them can be triggered by B-cell activation [
39]. Based on these findings, it seems very likely that the coexistence of PTFL and PNMZL, or classic FL and nodal MZL, may actually reflect a recapitulation of normal B-cell development.
As for differential diagnosis, PTFLs often need to be distinguished from florid reactive follicular hyperplasia. The immune architecture of the lymph node is usually preserved in reactive follicular hyperplasia, that is, no invasion of neoplastic B-cells into the interfollicular regions can be observed. Clonality, such as immunoglobulin light chain restriction, or genetic aberrations, such as
TNFRSF14,
MAP2K1, and
IRF8 mutations, cannot be demonstrated in reactive processes, too. A second differential diagnosis is classic FL, especially high-grade ones, which differs from PTFLs by the presence of molecular t(14;18)/bcl-2 rearrangements and BCL2 protein overexpression. Besides, mutations of some epigenetic modifier genes, such as
EP300,
CREBBP,
EZH2,
KMT2D, and
ARID1A, are often present in classic FL, too, which are rarely seen in PTFL in contrast [
40]. The pure form of PNMZL is less commonly seen, which needs to be distinguished from conventional nodal MZL. The two diseases resemble each other histologically and phenotypically, except for the prevalence of PTGC-like lesions existent in PNMZLs. Besides, some conventional nodal MZLs may show marked plasmacytic differentiation, which are generally lacking in PNMZLs. Genetically, PNMZL is akin to PTFL by frequent mutations of
MAP2K1,
IRF8, and
TNFRSF14, whereas classic nodal MZL features trisomies 3, 12, 18 and mutations of
KMT2D,
PTPRD, and
NOTCH2 [
41,
42]. Practically, pure PNMZLs are relatively rare, which are more often accompanied by a composition of PTFL. Actually, the coexistence of PTFL and PNMZL may serve as a useful diagnostic clue for recognizing this group of pediatric indolent B-cell lymphomas. Differential diagnosis with atypical marginal zone hyperplasia may be challenging by morphology merely. Clonality and genetic analysis may aid in establishing a proper diagnosis [
17].
In conclusion, we analyzed the clinicopathologic features of 59 patients with PTFL, PNMZL, and mixed-type tumors. In a small fraction of them, NGS analysis was also performed. We found no significant difference in the mutational frequency of TNFRSF14, MAP2K1, and IRF8 genes among the three groups. These findings may support the opinion that both PTFL and PNMZL might represent one single disease entity with a broad morphologic and phenotypic spectrum.
Acknowledgements
The authors appreciate Dr. Jia-Hao Liu (Department of Pathology, Second Affiliated Hospital of Soochow University, Jiangsu, China), Dr. Qiang Liu (Department of Pathology, Shanghai Jiao Tong University School of Medicine Affiliated Renji Hospital, Shanghai, China), Dr. Quan Zhou (Department of Pathology, The Second Affiliated Hospital of Jiaxing University, Zhejiang, China), Dr. Min-Zhi Yin (Department of Pathology, Shanghai Children’s Medical Center, School of Medicine, Shanghai Jiaotong University, Shanghai, China), Dr. Jian-Chen Fang (Department of Pathology, Ningbo Clinical Pathology Diagnosis Center, Zhejiang, China.), Dr. Fang Yu (Department of Pathology, The First Affiliated Hospital of Zhejiang University, Zhejiang, China.), Dr. Li-Yu Cao (Department of Pathology, Fuyang People’s Hospital, Anhui, China), and doctors from the First Affiliated Hospital of Fujian Medical University, the Second People’s Hospital of Wuxi, the First Affiliated Hospital of Anhui Medical University, Sir Run Run Shaw Hospital of Zhejiang University School of Medicine, Yancheng City Dafeng People’s Hospital, and Shanghai Jiao Tong University School of Medicine Affiliated Xinhua Hospital, for providing tissue materials for this study.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.